| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Experimental Therapeutics |
2,6-Sialylation of Cell-Surface N-Glycans Inhibits Glioma Formation in Vivo1
Chicago Institute for Neurosurgery and Neuroresearch, Chicago, Illinois 60614
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2,3-linked terminal sialic acids on glycoproteins bearing N-linked oligosaccharides, most notably on 3ß1 integrin, which is the predominant integrin found in these tumors. 2,6-linked terminal sialic acids, however, are not expressed. Two stable transfectants were made using a tumorigenic human glioma cell line, U-373 MG. Galß1,4GlcNAc 2,6-sialyltransferase (ST6Gal I) transfectants were made to replace the endogenous 2,3-linked sialic acids with 2,6-linked sialic acids. And Galß1,3(4)GlcNAc 2,3-sialyltransferase (ST3Gal III) transfectants were made to increase further the expression of cell-surface, N-glycan, 2,3-linked sialic acids. Although ST3Gal III transfection resulted in increased invasivity when compared with parental U-373 MG and vector-transfected control cells in vitro, ST6Gal I transfection abolished invasion in vitro and induced alterations in both cell morphology, cell-spreading, and adhesion-mediated protein tyrosine phosphorylation. Furthermore, the ST6Gal I transfectants produced no intracranial tumors in severe combined immunodeficient mice, whereas parental U-373 MG cells, the vector-transfected control cells, and ST3Gal III-transfected U-373 MG cells did. These results suggest that both the linkage and expression levels of the terminal sialic acids of 3ß1 integrin N-glycans play an important role in glioma cell-extracellular matrix interactions. Thus, manipulating ST6Gal I gene expression may have therapeutic potential for the treatment of malignant gliomas. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
3ß1 integrin was found to be the most predominant integrin expressed. Cell migration on extracellular matrix proteins correlated well with tumor grade and could be inhibited by antibodies against 3 integrin (7)
and ß1 integrin (8)
. In clinical specimens, 3 integrin mRNA expression was quantitatively related to the grade of malignancy in both gliomas and medulloblastomas (9
, 10)
.
Changes in the N-glycan structures of integrins can affect cell-cell and cell-extracellular matrix interactions, thereby affecting cell adhesion, cell migration, and tumor malignancy (11, 12, 13, 14, 15)
. Modification of tumor cell-surface carbohydrate expression by specific glycosylation inhibitors leads to a decrease in tumor formation and metastasis in vivo (16, 17, 18)
. In epithelial cells, a shift in integrin N-glycans to highly branched ß1,6-GlcNAc types leads to decreased cell adhesion, resulting in an increase in both cell motility and tumorigenicity by altering the function of 5ß1 and vß3 integrins (19)
. Additionally, studies suggest that the expression of ß1,6-linked GlcNAc-bearing N-glycans on 3ß1 integrin in human glioma specimens and in glioma cell lines plays a major role in glioma cell migration and invasion (20)
.
Changes in terminal sialylation of N-glycans, too, have been reported to be closely associated with cellular adhesion, migration, and metastasis in tumor cells. Terminal sialylation of complex-type N-linked oligosaccharides is either 2,6-linked or 2,3-linked and is dependent on the expression of two distinct sialyltransferases, ST6Gal I3
(EC 2.4.99.1) and ST3Gal III (Ref. 4
; EC 2.4.99.6), respectively (21)
. These enzymes are developmentally regulated, tissue-specific, and involved in the regulation of a number of important cell processes (22)
. ST6Gal I plays a key role in oncogenic transformation (23)
, metastatic potential (24
, 25) , and differentiation of colon carcinomas (26
, 27)
. Neither ST6Gal I mRNA expression nor 2,6-linked sialoglycoproteins were observed in any of the glioma specimens or glioma cell lines examined (28
, 29)
; however, we have found a marked increase in 2,3-linked sialoglycoprotein and ST3Gal III mRNA expression in human glioma specimens and glioma cell lines (29)
. Furthermore, ST6Gal I transfectants made with the human tumorigenic glioma cell line, U-373 MG, led to decreased adhesivity to fibronectin and collagen matrices and to a marked inhibition of 3ß1 integrin-mediated invasivity in vitro (30)
. Inhibition of in vitro invasivity was correlated with the levels of ST6Gal I enzyme activity in the transfectants (30)
.
The experiments reported here were undertaken to study further the role of N-glycan terminal sialic acids in the cell-extracellular matrix interactions and to explore therapeutic potential of manipulating sialyltransferase gene expression for the treatment of malignant gliomas. First, ST3Gal III transfectants were made to overexpress endogenous glioma-associated N-linked terminal 2,3-sialic acids, and in vitro invasivity was examined. Then, the effects of altered terminal sialylation on cell-extracellular interactions were examined using both ST3Gal III- and ST6Gal I-transfected U-373 MG glioma cells. Finally, in vivo tumor formation of the transfectants was examined using an intracranial tumor model. Increased expression of glioma-associated 2,3-linked sialic acids by overexpression of ST3Gal III resulted in increased glioma invasivity in vitro. Replacement of glioma-associated 2,3-linked sialic acids with 2,6-linked sialic acids by ST6Gal I gene transfection disrupted glioma cell-extracellular matrix interactions in vitro. Although the changes in terminal sialic acids expression by sialyltransferase gene transfection induced alterations in cell-extracellular interactions in vitro, the most drastic change was found in tumor formation in vivo. These changes were attributable in part to the alteration of terminal sialic acids on the 3ß1 integrin expressed in these cells.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Stable Transfection of ST3Gal III into U-373 MG Human Glioma Cells.
Human glioma U-373 MG cells were used for the stable transfection of the ST3Gal III and ST6Gal I genes. For overexpression of ST3Gal III and 2,3-linked sialoglycoconjugates, 1.2-kb rat ST3Gal III cDNA (Ref. 22
; kindly provided by Dr. James Paulson, The Scripps Research Institue, La Jolla, CA) was inserted into the pcDNA3 expression vector (Invitrogen, San Diego, CA) at the EcoRI site, and the orientation of the insert was confirmed by restriction digestion with PstI. The ST3Gal III/pcDNA3 plasmid was transfected into U-373 MG cells using the cationic liposome system DOTAP (Boeringer Mannheim, Indianapolis, IN) according to previously described methods (20
, 30)
. After 3 weeks in culture in selection medium containing 800 µg/ml of G418, transfected cells were subcloned to isolate individual clones. These clones were cultured for another 4 weeks in selection medium and analyzed for gene expression by Northern analysis. Stable transfection of the ST6Gal I gene into U-373 MG cells was reported previously (30)
.
Northern Analysis.
Rat ST3Gal III cDNA (1.2 kb) was isolated after EcoRI restriction digestion and used as a cDNA probe for Northern analysis. Total RNA, 15 µg per each ST3Gal III-transfected clone, was electrophoresed in an agarose-formaldehyde gel and transferred to Duralon nylon membrane (Stratagene, La Jolla, CA). After UV cross-linking, the blots were hybridized with the 32P-radiolabeled cDNA probe in ExpressHyb solution (Clontech, Palo Alto, CA). The blots were then exposed to X-OMAT film (Kodak, Rochester, NY), and the films were developed.
In Vitro Invasion Assays.
Invasivity of the ST3Gal III-transfected subclones was examined using a commercial membrane invasion culture system (6
, 20
, 31) . Biocoat Matrigel Invasion Chambers (Collaborative Research, Bedford, MA) consist of two compartments separated by a filter precoated with Matrigel, which contains laminin, type IV collagen, entactin and heparan sulfate. Cell invasion, which is the result of cell adhesion to the extracellular matrix, degradation of the matrix protein, and cell migration, is measured by counting the number of cells that have migrated to the under side of the filter via 8-µm pores. Cells (4 x 104) were plated into the upper chamber and incubated for 24 h. Ten percent FBS containing medium (0.5 ml) was placed in the lower compartment to facilitate chemoattraction (31)
. Cells that have migrated through the Matrigel and the filter were fixed in 10% formalin and stained with hematoxylin. The membranes were mounted on glass slides, and the cells were counted. Parental U-373 MG, pcDNA3 vector-transfected cells and ST6Gal I-transfected U-373 MG cells were used as controls.
Adhesion-mediated Protein Tyrosine Phosphorylation.
Tissue culture plates (100 mm) were coated with 10 µg/ml human fibronectin in PBS for 16 h and then incubated with 1% BSA in PBS for 1 h to block nonspecific cell adhesion. The plates were then washed twice with PBS. ST3Gal III- and ST6Gal I-transfected U-373 MG cells were gently treated with x0.5 trypsin-EDTA (Life Technologies) in PBS for 5 min at 37°C and neutralized with DMEM containing 0.2% BSA. Cells were washed once with 0.2% BSA-DMEM and resuspended in protein-free DMEM before plating on the fibronectin-coated plates. After incubation for 30 min at 37°C in the presence of 10% CO2, unattached cells and culture medium were removed by aspiration. The remaining attached cells were immediately solubilized with 200 µl of lysis buffer containing 10 mM Tris-HCl (pH 7.4), 1% SDS, and 1 mM sodium orthovanadate. The lysates were centrifuged at 12,000 x g for 5 min to eliminate nonsoluble material. Thirty µg of protein from each sample were loaded onto an 8% SDS-polyacrylamide gel. After electrophoresis, the proteins were transferred to a polyvinylidene fluoride membrane, and the membrane was incubated with 4% BSA in PBS at room temperature for 30 min. horseradish peroxidase-conjugated antiphosphotyrosine antibody (RC-20; Transduction Laboratories) was then added at 1:2,500 dilution, and the blot was incubated for 1 h at room temperature. The membrane was then washed three times with PBS containing 0.05% Tween 20, and the antibody-bound proteins were detected using an enhanced chemiluminescence kit (Amersham Pharmacia Biotech, Inc., Piscataway, NJ).
Immunoprecipitation of 3ß1 Integrin; Lectin and Western Blotting.
Effects of sialyltransferase transfection on overall glycoprotein sialylation and terminal sialylation of 3ß1 integrin N-glycans were examined by lectin and Western blot analyses. Approximately 1 x 108 parental or transfected U-373 MG cells were washed once with PBS, harvested by scraping and pelleted by centrifugation. The cell pellet was homogenized in 1.5 ml of 10 mM Tris-HCl (pH 7.4) containing 1% Triton X-100, 0.15 M NaCl, 10 µg/ml leupeptin and 1 mM phenylmethylsulfonyl fluoride (buffer A) using a Teflon-glass homogenizer. After 10,000 x g centrifugation for 10 min, 1.2 ml of the supernatant was collected for immunoprecipitation of 3ß1 integrin. Twenty-five µg of antihuman 3ß1 integrin mouse monoclonal antibody (clone M-KD102, Chemicon International, Temecula, CA) was added to 1.2 ml of the supernatant in an Eppendorf tube and gently inverted for 3 h at 4°C. Fifty µl of protein G agarose (Sigma Chemical Co., St. Louis, MO) was then added and mixed gently for 1 h to absorb the integrin-antibody complex. The protein G agarose was spun down by brief centrifugation, and the supernatant was discarded. The agarose gel was washed three times each with 1 ml of buffer A. After a final washing, the gel was briefly centrifuged, and the buffer was aspirated. The gel was then resuspended in 100 µl of denaturation buffer and boiled for 10 min to release the integrin from the gel. Protein extracts and immunoprecipitated protein samples were loaded onto 8% SDS-polyacrylamide gels. After electrophoresis, proteins were transferred to an Immobilon-P membrane (Millipore, Bedford, MA), and the blots for lectin staining were blocked in 0.5% casein in TBS [150 mM NaCl and 50 mM Tris-HCl (pH 7.5)] for 45 min at room temperature. Each blot was washed twice for 10 min with TBS, then once with lectin vehicle (1 mM MgCl2, 1 mM MnCl2, and 1 mM CaCl2 in TBS) before a 1-hour incubation with DIG-labeled lectin. DIG-labeled SNA (1:1,000) or DIG-labeled MAA (1:200; Roche Molecular Biochemicals, Indianapolis, IN) was used to detect 2,6-linked or 2,3-linked terminal sialic acids of N-glycans, respectively. Blots were washed three times for 10 min each in TBS, and were then incubated 1 h with DIG-AP (1:1,000 dilution, Roche Molecular Biochemicals) in TBS. After three additional TBS washes (10 min each), the blots were overlaid with 5-bromo-4-chloro-3-indolylphosphate (BCIP)/nitro blue tetrazolium (NBT) Liquid Substrate System (Sigma Chemical Co.). Blots were washed in several changes of water to stop staining.
Membranes for Western blotting were blocked with 4% BSA in TBS overnight at 4°C, then incubated at room temperature for 1 h with polyclonal rabbit anti-3 integrin antibody (dilution, 1:1000; Chemicon, Temecula, CA) or monoclonal antihuman ß1 integrin antibody (clone HB1.1; dilution, 1:250) in TBS containing 2% BSA, 0.1% Tween 20 (buffer B). The membranes were washed once with buffer B and twice in TBS containing 0.1% Tween 20 for 10 min/wash. They were then incubated for 1 h in either peroxidase-conjugated antirabbit IgG or antimouse IgG antibody (dilution, 1:10,000 or 1:5,000, respectively; Amersham Pharmacia Biotech, Inc.) in buffer B. The blots were then washed as described above and developed with the enhanced chemiluminescence Western Blotting Detection System (Amersham Pharmacia Biotech, Inc.).
Immunostaining of EGF-Receptor in U-373 MG Gliomas.
Immunostaining of EGF-receptor expressed on intracranial U-373 MG glioma cells was carried out using the BioGenex Super Sensitive Immunodetection System (BioGenex, San Ramon, CA). Six-µm frozen brain sections were mounted on glass microscope slides and immediately fixed by immersion in 95% ethanol. The brain sections were washed with agitation at room temperature for 5 min in PBS before and between each of the following immunostaining procedures. The specimens were first blocked for 5 min at room temperature with Peroxide Block and incubated for 10 min at room temperature in 2.5 µg/ml mouse monoclonal antibody against epidermal growth factor receptor (clone 528; Calbiochem-Novabiochem, San Diego, CA) diluted in PBS. Normal mouse immunoglobulins were used as a control (BioGenex). After biotinylated secondary antibody (BioGenex) was applied for 10 min at room temperature, the sections were incubated for 10 min in streptavidin-horseradish peroxidase complex (BioGenex), and staining was developed for 2 min. After the slides were washed with PBS for 5 min, sections were counterstained for 20 min with Methyl Green (BioGenex). Finally, slides were briefly washed in PBS, then 95% ethanol, two changes of 100% ethanol, and three changes of Histo-Clear histological clearing agent (National Diagnostics, Manville, NJ), and the coverslips were mounted with Permount Mounting Medium (Fisher Scientific, Pittsburgh, PA).
Intracranial Tumorigenicity of the ST3Gal III- and ST6Gal I-transfected U-373 MG Glioma Cells.
Intracranial tumorigenicity of ST3Gal III-, ST6Gal I-, pcDNA3 vector-transfected cells, and parental U-373 MG cells was examined in SCID mice. All mice were maintained in the animal facility of Children’s Memorial Institute for Education and Research at Northwestern University, Chicago, IL. This study was approved by their Institutional Animal Care and Use Committee. Ten µl of a 1.25 x 106 glioma cell suspension were injected stereotactically into the right basal ganglia of anesthetized SCID mice (C.B-17 scid/scid, 6 weeks of age; Charles River Laboratories), and the brains were harvested after 6 weeks. The brains were mounted on cryostat pedestals and serial 6-µm thick coronal sections were cut through the basal ganglia at 20-µm intervals. The sections were used to determine tumor size by H&E staining or antihuman EGF-receptor antibody staining. The maximum cross-sectional area of the tumors was determined by computer-assisted image analysis using the Microcomputer Imaging Device (MCID) software package of Imaging Research (Brock University, St. Catherines, Ontario, Canada). Ten mice/transfectant were used in each of four groups (parental U-373 MG glioma cells; three different ST3Gal III-transfected U-373 MG glioma clones, J2, J8, and J22; three different ST6Gal I-transfected U-373 MG glioma clones, J11, J20, and J22; and pcDNA3 vector-transfected U-373 MG cells as a control) for a total of 80 mice. Tumor size was determined by measuring the maximum cross-sectional area of the tumors using the Microcomputer Imaging Device (MCID) software package (Imaging Research, St. Catherines, Ontario, Canada).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
|
. In culture, little difference was noted in cell morphology among ST3Gal III, parental U-373 MG cells, and the vector-transfected cells, all of which express ST3Gal III and terminal 2,3-linked sialic acids of N-glycans: cells piled up as they became confluent in culture. On the contrary, ST6Gal I-transfected cells grew in a monolayer (Fig. 3)
. Morphological differences in cell-spreading and adhesion were examined further. Little difference was found in cell adhesion and spreading among the ST3Gal III transfectants, parental U-373 MG cells, and the vector-transfected cells; whereas, spreading was observed in few ST6Gal I-transfected cells at 1 h. At 24 h, ST6Gal I-transfected cells were well-spread and rounded as compared with the bipolar, triangular, or fan-shaped cell morphology in cells expressing 2,3-linked terminal sialic acids (Fig. 4)
.
|
|
|
3ß1 integrin (30)
. Cells were plated on a fibronectin matrix to induce adhesion-mediated protein tyrosine phosphorylation, which was examined by Western blotting using an antiphosphotyrosine antibody (Fig. 5)
. Two major tyrosine phosphorylated proteins with molecular masses of 110 kDa and 125 kDa were observed in ST3Gal III cells; whereas, the 110-kDa phosphorylated protein was not seen in ST6Gal I-transfected cells. The identity of the 110-kDa protein is not known. These results suggest that ST6Gal I expression alters cell-extracellular matrix interactions.
|
3ß1 integrin terminal sialylation by ST3Gal III and ST6Gal I transfection in U-373 MG cells. MAA or SNA lectin was used to detect 2,3-linked or 2,6-linked terminal sialic acids of N-glycans, respectively. As shown in Fig. 6
, MAA lectin staining was greater in ST3Gal III-transfected cells than in controls, suggesting increased 2,3-linked sialylation of N-glycans. Contrary to our expectation, levels of MAA lectin staining in ST6Gal I-transfected cells were similar to that of parental U-373 MG cells and were even higher than in the vector-transfected cells. On the other hand, SNA lectin staining of 2,6-linked sialic acids was only seen in ST6Gal I-transfected cells. We also looked at lectin staining in immunoprecipitated 3ß1 integrin from these transfectants. On the basis of MAA lectin staining, ST3Gal III transfectants expressed higher levels of 2,3-linked sialic acids on 3ß1 integrin than did parental U-373 MG or vector-transfected cells. Only weak MAA lectin staining was seen in the integrin from ST6Gal I-transfected cells. On the other hand, SNA lectin staining was strong in 3ß1 integrin from ST6Gal I-transfected cells, but was negative in the ST3Gal III transfectants and the controls (Fig. 6)
. These data suggest that transfection of ST6Gal I almost completely replaces endogenous N-glycans 2,3-linked sialic acids with 2,6-linked sialic acids on 3ß1 integrin. However, overall levels of 2,3-linked sialic acid expression in ST6Gal I-transfected cells remained similar to levels of expression in parental U-373 MG cells. As shown in Fig. 6
, the 3 and ß1 integrin subunits were coimmunoprecipitated using an antibody to 3ß1 integrin.
|
. The sections were also stained with an antibody against EGF-receptor, which is highly expressed in U-373 MG glioma cells (32)
but not in normal brain tissue (Fig. 7D, E, and F)
. Both parental and vector-transfected U-373 MG cells formed large tumors in vivo, with the vector-transfected controls forming 50% larger tumors than the parental cells. Tumors formed by the ST3Gal III-transfected U-373 MG clones were, on the average, one-tenth the size of those formed by vector-transfected U-373 MG cells. And ST6Gal I-transfected U-373 MG glioma clones formed virtually no tumors. Differences in tumor size among the animal groups are shown in Fig. 8
.
|
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Integrins are a superfamily of transmembrane receptors that participate in cell-cell and cell-matrix interactions (3, 4, 5)
. Aberrant Nglycosylation of integrins alters cell adhesion or motility and has been implicated in malignant phenotypes such as invasiveness and metastasis (reviewed in Ref. 11
). The two most commonly observed forms of aberrant N-glycosylation in experimental tumor models are a shift to more highly branched N-linked oligosaccharides and an increase in terminal sialylation. For example, a shift of integrin N-glycans to highly branched ß1,6-GlcNAc in premalignant mink lung epithelial cells leads to decreased cell adhesion resulting in an increase in cell motility and tumorigenicity by altering the functions of 5ß1 and vß3 integrins (19)
. We have recently reported that 3ß1 integrin is the predominant glycoprotein carrying ß1,6-GlcNAc-linked Nglycans in human glioma cells, and that increased expression of ß1,6-GlcNAc-linked N-glycans results in increased invasivity in vitro (20)
. Increased sialylation of the ß1 integrin subunit has been correlated with decreased adhesiveness and metastatic potential (14
, 36)
.
Sialyltransferases, which are responsible for the biosynthesis of sialoglycoconjugates, comprise a structurally related family of enzymes that display substrate specificity and tissue specificity, and are all developmentally regulated (21)
. ST6Gal I and ST3Gal III are two sialyltransferases which add terminal 2,6-linked and 2,3-linked sialic acids to N-linked oligosaccharide chains, respectively.
ST6Gal I has been suggested to play an important role in the oncogenic transformation of colon mucosa (37)
and to be an indicator of metastatic and invasive potential in colon cancer (24)
. We have reported the expression of ST6Gal I mRNA, ST6Gal I protein, and 2,6-linked sialic acids in meningiomas, chordomas, and craniopharyngiomas, but no 2,6-ST expression was found in malignant gliomas or in medulloblastomas (28)
. On the other hand, ST3Gal III mRNA and 2,3-linked sialic acids were expressed in all malignant glioma specimens and glioma cell lines examined (29)
. In this paper, we have shown that the 3ß1 integrin bears terminal 2,3-linked sialic acids, the product of ST3Gal III. ST3Gal III-transfectants are indistinguishable from parental or vector-transfected cells in morphology and proliferation, spreading, and adhesion-mediated protein tyrosine phosphorylation (data not shown) but are more invasive in vitro. And overexpression of ST3Gal III resulted in only a slight increase in 2,3-linked sialylation of total glycoproteins and 3ß1 integrin, suggesting that most N-linked glycoproteins in parental U-373 MG glioma cells contain terminal 2,3-linked sialic acids. On the other hand, stable transfection of ST6Gal I resulted in the replacement of 2,3-linked sialic acids with 2,6-linked sialic acids on 3ß1 integrin; however, overall cellular 2,3-linked sialic acid expression was not diminished. These results suggest that the glycoprotein substrates for ST3Gal III and ST6Gal I may be different, and that other factors, such as protein structure or substrate accessibility, may have an effect on terminal sialylation of N-glycans on glycoproteins. Alteration of N-glycans terminal sialylation by ST6Gal I transfection resulted in marked changes in cell morphology, cell-cell interaction, and adhesion-mediated signaling. Although the exact mechanism of these changes brought by ST6Gal I transfection is not clear, it may be, at least, due to the replacement of 2,3-linked sialic acids with 2,6-linked sialic acids on ß1 integrin. Recently, it was reported that integrin functions are modulated by not only N-glycosylation, but also by surrounding gangliosides and tetraspan membrane glycoproteins (38
, 39)
. Gangliosides GT1b and GD3 inhibit keratinocyte adhesion to fibronectin by carbohydrate-carbohydrate interactions with 5 integrin (38)
. Glycosylation affects translocation of integrin, Src, and caveolin, thus controlling cell adhesion and signaling (40)
. 3ß1 integrin is a part of multimolecular complexes including tetraspans, CD9, CD81, and CD151 (39
, 41)
. 2,6-linked terminal sialylation of ß1 integrin in the ST6Gal I-transfectants might affect its interaction with other molecules, thus resulting in altered cell signaling. Changes in terminal sialylation of other glycoproteins, which interact with 3ß1 integrin, may also play a role. Nevertheless, changes in N-linked terminal sialic acids have marked effects on glioma cells in vitro and in vivo. Although levels of in vitro invasivity were positively correlated with the levels of 2,3-linked sialylation (ST3Gal III > parental U-373 MG cells > pcDNA3 vector-transfected cells) as shown in this study, 2,6-linked sialylation of 3ß1 integrin N-glycans by ST6Gal I transfection abolished in vitro invasivity (30)
.
In general, increased cell surface sialylation is associated with reduced cellular adhesivity and, thus, with increased metastasis (14
, 24
, 36)
. Theoretical models have shown an inverted U-shaped relationship between cellular adhesivity and migration (42
, 43) . Changes in cell adhesion brought about by an alteration in integrin N-glycosylation could enhance cell migration depending upon the initial strength of adhesion between a cell and its substratum. For example, an optimal adhesiveness for muscle cell migration on collagen has been identified (44)
. And concentration-dependent, inhibitory, and enhancing effects of an integrin-binding inhibitor on cell motility have been observed (45)
. Additionally, cell motility of mammary cells across collagen-coated filters was increased only in those clones with intermediate levels of adhesion to collagen (46)
. We have shown that in vitro invasion of U-373 MG glioma cells is mediated by 3ß1 integrin (30)
. On the basis of the theoretical models on cell adhesion and migration, increased invasivity of ST3Gal III transfectants is likely attributable to reduced adhesivity resulting from increased expression of endogenous 2,3-linked sialic acids. 2,6-linked sialylation of 3ß1 integrin N-glycans also resulted in reduced adhesivity in ST6Gal I transfectants (30)
; however, there was a marked difference in cell-extracellular signaling among ST3Gal III and ST6Gal I transfectants as shown in this study. Opposing effects on in vitro invasivity by ST3Gal III and ST6Gal I transfection are likely attributable to altered signaling rather than to mere reduced adhesivity.
3 and ß1 integrin subunits contain 13 and 12 potential N-glycosylation sites respectively. MAA and SNA lectin staining of 3ß1 integrin is primarily a result of heavy N-glycosylation on the ß1 subunit (data not shown). It is reported that glycosylation of and ß integrin chains play functionally distinct roles in cell-matrix interactions (47)
. It is also reported that ß1 integrin interacts with signaling molecules, including focal adhesion kinase, and with cytoskeletal proteins (2)
. We have found the lack of 110 kDa protein phosphorylation in ST6Gal I transfectants, which might be a result of 2,6-linked sialylation of ß1 integrin. We reported previously that ST6Gal I transfection into U-373 MG glioma cells resulted in inhibition of 3ß1 integrin-mediated glioma invasion in vitro, changes in cell morphology and focal adhesions, reduction in cell adhesion on extracellular matrix proteins, and an increase in the expression of p125 focal adhesion kinase mRNA (30)
. These changes may be attributable to altered ß1 integrin sialylation and signaling.
Integrin-extracellular matrix interactions also play a major role in anchorage-dependent tumor growth (3, 4, 5
, 48)
. Thus, we examined the effects of altered terminal sialylation on in vivo tumor formation. Both parental and vector-transfected U-373 MG cells formed large intracranial tumors. Small tumors were formed by the ST3Gal III-transfected U-373 MG clones, and ST6Gal I-transfected U-373 MG glioma clones formed virtually no tumors. Glioma cells with higher levels of 2,3-linked sialic acid expression (ST3Gal III transfectants > parental U-373 MG cells > the vector-transfected U-373 MG cells) formed smaller tumors (the vector-transfected U-373 MG cells > parental U-373 MG cells > ST3Gal III transfectants). These marked differences in intracranial tumor formation are likely attributable to differences in glioma cell-extracellular interactions rather than to differences in mere cell proliferation rate, inasmuch as all cultured glioma cells used in this study showed little variation in rates of cell proliferation (data not shown). These results can be explained by the idea proposed by Bissell et al. (49)
that the interactions between cells and extracellular matrix govern gene expression via biochemical signals transduced by membrane-bound receptors and, thus, the ultimate decision of a cell to proliferate, differentiate, or apoptose is dependent on the cellular microenvironment. More recently, they have shown that the malignant phenotype of a human breast tumor cell can be reverted by correcting integrin-extracellular matrix signaling (reviewed in Ref. 50
). ST3Gal III transfectants were more invasive in vitro but not in vivo. This discrepancy can also be explained by cell-extracellular matrix interactions within the microenvironment. Data presented in this study demonstrate that both terminal sialylation and its linkage to integrin N-glycans play critical roles in cell-cell and cell-extracellular interaction.
Clearly integrins play a complex role in modulating cell-cell interactions and signal transduction. Recent studies have also demonstrated that the 3ß1 integrin plays a role in the malignant behavior of gliomas and medulloblastomas (6
, 7 , 10)
. The results presented here show that by changing cell-surface sialic acids by sialyltransferase gene transfection, particularly the addition of 2,6-linked sialic acids to the 3ß1 integrin, malignant glioma formation was inhibited.
The manipulation of cell-surface carbohydrate expression by specific glycosyltransferase gene delivery may provide a new therapeutic approach for the treatment of malignant tumor cells. ST6Gal I expression, by altering 3ß1 integrin function, may be useful as both an ortho-adhesion and an ortho-signaling therapy in malignant brain tumors.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This work was supported in part by grants from the Illinois division of the American Cancer Society (to H. Y.), the Buchanan Foundation (to J. M.), the Grant Foundation (to J. M.), the Brach Foundation (to J. M.), and the Falk Foundation (to J. M.). 
2 To whom requests for reprints should addressed, at Chicago Institute for Neurosurgery and Neuroresearch, 430 West Deming Place, Chicago, IL 60614. Phone: (773) 388-7860: Fax: (773) 388-7866: E-mail: j-moskal{at}nwu.edu
3 The abbreviations used are: ST6Gal I, CMP-NeuAc: Galß1, 4GlcNAc 2,6sialytransferase; ST3Gal III, CMP-NeuAc: Galß1,3 GlcNAc 2,3-sialytransferase; GlcNAc, N-acetylglucosamine; DIG, digoxygenin; SNA, Sambucus nigra agglutinin; MAA, Maackia amurensis agglutinin; SCID, severe combined immunodeficient; EGF, epidermal growth factor.
Received 2/23/01. Accepted 7/18/01.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
3 integrin antibodies. Int. J. Cancer, 67: 777-784, 1996.[Medline]
v and ß1 integrins. Cancer Res., 56: 1939-1947, 1996.3 in gliomas and medulloblastomas. Br. J. Cancer, 79: 333-339, 1999.[Medline]
1–6Manß1- branched asparagine-linked oligosaccharides in murine tissues and human breast carcinomas. Cancer Res., 49: 945-950, 1989.2,6-sialyltransferase in rat fibroblast (FR3T3) cells. Glycobiology, 2: 49-56, 1992.2,6-sialylation of N-acetyllactosamine sequences in ras-transformed rat fibroblasts correlates with high invasive potential. Glycobiology, 5: 219-226, 1995.2,6-sialylated sugar chains by transfection with rat ßgalactoside 2,6-sialyltransferase cDNA. Biochem. Biophys. Res. Comm., 211: 554-561, 1995.[Medline]
2,6-sialyltransferase in small intestine epithelium. Dev. Biol., 165: 126-136, 1994.[Medline]
2,6-sialyltransferase and 2,6-linked sialoglycoconjugates in human brain tumors. Acta Neuropathol. (Berl.), 9: 284-292, 1996.
2,3 Sialyltransferase mRNA and 2,3-linked glycoprotein sialylation are increased in malignant gliomas. Brain Res., 755: 175-179, 1997.[Medline]
2,6-sialyltransferase gene transfection into a human glioma cell line (U-373MG) results in decreased invasivity. J. Neurochem., 68: 2566-2576, 1997.[Medline]
2,6-linked sialic acid residues in neoplastic but not in normal human colonic mucosa. Am. J. Pathol., 139: 1435-1448, 1991.[Abstract]
5 modulates 5ß1 function. J. Biol. Chem., 276: 8436-8444, 2001.4ß1, CD151/3ß1, CD151/6ß1) under conditions disrupting tetraspan interactions. Biochem. J., 340: 103-111, 1999.
2 integrin mRNA in mammary cells. J. Cell Sci., 108: 595-607, 1995.[Abstract]
and ß integrin chains in cell-matrix interactions. Proc. Natl. Acad. Sci. USA, 90: 1795-1799, 1993.This article has been cited by other articles:
|
|
 |
|
  M. Lee, H.-J. Lee, S. Bae, and Y.-S. Lee Protein Sialylation by Sialyltransferase Involves Radiation Resistance Mol. Cancer Res., August 1, 2008; 6(8): 1316 - 1325. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Li, K. Ishihara, T. Yokota, T. Nakagawa, N. Koyama, J. Jin, Y. Mizuno-Horikawa, X. Wang, E. Miyoshi, N. Taniguchi, et al. Reduced {alpha}4 1 Integrin/VCAM-1 Interactions Lead to Impaired Pre-B Cell Repopulation in Alpha 1,6-Fucosyltransferase Deficient Mice Glycobiology, January 1, 2008; 18(1): 114 - 124. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Zhao, S. Itoh, X. Wang, T. Isaji, E. Miyoshi, Y. Kariya, K. Miyazaki, N. Kawasaki, N. Taniguchi, and J. Gu Deletion of Core Fucosylation on {alpha}3beta1 Integrin Down-regulates Its Functions J. Biol. Chem., December 15, 2006; 281(50): 38343 - 38350. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Zhao, T. Nakagawa, S. Itoh, K.-i. Inamori, T. Isaji, Y. Kariya, A. Kondo, E. Miyoshi, K. Miyazaki, N. Kawasaki, et al. N-Acetylglucosaminyltransferase III Antagonizes the Effect of N-Acetylglucosaminyltransferase V on {alpha}3beta1 Integrin-mediated Cell Migration J. Biol. Chem., October 27, 2006; 281(43): 32122 - 32130. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Chiricolo, N. Malagolini, S. Bonfiglioli, and F. Dall'Olio Phenotypic changes induced by expression of {beta}-galactoside {alpha}2,6 sialyltransferase I in the human colon cancer cell line SW948 Glycobiology, February 1, 2006; 16(2): 146 - 154. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
2 analysis of the maximum cross-sectional area of each tumor. Ten mice/transfectant were used in each of 4 groups: parental U-373 MG glioma cells, three ST6Gal I-transfected U-373 MG clones (J11, J20, and J22), three ST3Gal III-transfected U-373 MG clones (J8, J22, and J2) and pcDNA3 vector-transfected U-373 MG cells as a control.