| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
1 Division of Cancer Cell Research, Institute of Medical Science, The University of Tokyo, Tokyo, Japan;
2 Pharmaceutical Development Section, Research Institute, Daiichi Fine Chemical, Ltd., Takaoka, Japan; and
3 Department of Pathology, School of Medicine, Keio University, Tokyo, Japan
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
CD44 is a major receptor of HA and has been shown to play important roles in physiological and pathological processes such as lymphocyte homing, T-cell activation, wound healing, angiogenesis, and metastatic spread of cancer cells (8, 9, 10) . Outside of the cell, CD44 has a globular lectin-like domain for ligand binding and is connected to the cell surface through a stem sequence. Following the type I transmembrane domain, it has a cytoplasmic tail that acts as a binding site for multiple proteins, such as ezrin, radixin, and moesin (ERM) proteins and ankyrin, that associate with the actin cytoskeleton (1 , 3 , 8) . Regulators of small G-proteins such as Rho-GDI and Tiam1 are also recruited to the site (11) . Thus, the extracellular portion of CD44 mediates the interaction of cells with the tissue architecture through a relatively flexible adhesion, and its cytoplasmic portion interacts with and regulates the actin cytoskelton.
CD44 has many isoforms generated by alternative splicing (8) . The most abundant and basic form is the standard hematopoietic type, CD44H. Alternative splicing inserts additional sequences encoded by the extra exons into the stem region. At least 20 different CD44 transcripts with alternatively spliced exons have been detected, and each isoform is thought to have specific functions such as a possible difference in the ability to recognize HA (12, 13, 14) . In addition, expression of isoforms with variant sequences v6 and v8-v10 has been shown to correlate with metastatic spread of cancer cells (15 , 16) , and matrix metalloproteinase 7 binds to heparan sulfate proteoglycan that modifies the variant sequence v3 (17) .
Further diversification of the isoforms is generated by posttranslational modifications such as variable N- and O-linked glycosylations (18) . CD44H has a polypeptide with a predicted size of 37 kDa, but the observed molecular mass is around 85–95 kDa due to the extensive glycosylation. Four serine-glycine motifs in the stem region are also potential modification sites for glycosaminoglycan heparan sulfate or chondroitin sulfate. However, it is not easy to detect the glycosylation of CD44 at a particular site, and therefore, knowledge about site-specific glycosylation and its biological role is very limited, particularly with regard to the stem region.
In this study, we developed mAbs and polyclonal antibodies that recognize defined epitopes of the stem region of CD44. All of the antibodies were reactive against the polypeptides prepared in Escherichia coli by Western blotting. However, some of them failed to recognize native CD44 molecules produced by several of the tumor cell lines examined. The reactivity of such antibodies resumed when the samples were treated with glycosidases, or when CD44 mutants with site-specific mutations at the possible glycosylation sites were expressed. Thus, the reactivity of these antibodies is affected by the state of glycosylation at specific sites. Interestingly, such specific forms of CD44 were found to be expressed in a cell type-specific manner in tumor tissues. Although both tumor and stromal cells express CD44, one glycosylation-sensitive antibody specifically recognized CD44 in tumor cells but not in the stroma, whereas another glycosylation-sensitive antibody recognized CD44 in stromal cells but not in tumor cells. This is presumably the first indication that differential glycosylation of CD44 occurs in a cell type-specific manner in vivo. These set of antibodies are potentially useful as diagnostic tools for diseases including malignant tumor and to shed light on the complicated regulations and functions of CD44 in relation to glycosylation in the stem region.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Recombinant Proteins and Oligopeptides.
The cDNA encoding the stem region of CD44H (Thr130-Glu268) with a FLAG tag at the NH2 terminus and a His6 tag at the COOH terminus (rCD44HS) was expressed using a bacterial vector, pET3a (Stratagene, La Jolla, CA; Ref. 19
). The E. coli strain BL21(DE3)pLysS was transformed with the plasmid, and protein expression was induced by 0.4 mM isopropyl-1-thio-ß-D-galactopyranoside. Cells were collected and sonicated in TNC buffer (50 mM Tris-HCl, 150 mM NaCl, 10 mM CaCl2, and 0.02% NaN3) containing 2 mM phenylmethylsulfonyl fluoride. Supernatant was collected, and the His6-tagged protein was purified with chelating Sepharose and a gel filtration column using 
KTA explorer 10S systems (Amersham Pharmacia Biotech, Buckinghamshire, United Kingdom). Other CD44H stem fragments (RS1 tagged with glutathione S-transferase and RS2–RS7 tagged with His6) were designed to be expressed in E. coli similarly to rCD44HS and purified. The P1 oligopeptide contains DGHSHGSQEGGA (aa 243–254) derived from the CD44 stem sequence.
Site-directed mutations of CD44H at SHTT237 were generated using a GeneEditor site-directed mutagenesis system (Promega, Madison, WI) and subcloned into a eukaryotic expression vector, pcDNA3.1(+)Zeo (Invitrogen, Carlsbad, CA).
Preparation of Antibodies.
Two 6-week-old female BALB/c mice were immunized three times with either rCD44HS, RS1, or the keyhole limpet hemocyanin-conjugated oligopeptide P1 after emulsification with an equal volume of Freund’s complete adjuvant. Three days after the last injection, splenocytes were isolated for fusion with SP-2 mouse myeloma cells (SP-2/0-Ag14). The hybridization and subsequent culturing and cloning of hybrids were carried out as described previously (20)
. The antibodies (IgG1) obtained were purified from ascitic fluids by 40% saturated ammonium sulfate fractionation, followed by protein A-Cellulofine column chromatography (Seikagaku Corp., Tokyo, Japan). An isotype for each mAb obtained using a MonoAb-ID EIA kit (Zymed Laboratory, Inc., San Francisco, CA) showed that all of the light chains are of the 
type, but the heavy chains are variable (three g1 clones and one g2b clone). To obtain polyclonal antibody for oligopeptide P1, two female rabbits were immunized with oligopeptide P1-keyhole limpet hemocyanin conjugate emulsified with an equal volume of Freund’s complete adjuvant. A week after the last injection, antiserum was bled from the animals and treated with 40% saturated ammonium sulfate fractionation, followed by DEAE-Sephacel column chromatography (Amersham Pharmacia Biotech) and rCD44HS binding Sepharose 4B column chromatography (Amersham Pharmacia Biotech). The anti-CD44H antibody clone 2C5 was obtained from R&D Systems (Minneapolis, MN).
RT-PCR and Primers.
Total RNA samples (3 µg) were reverse transcribed using 0.3 µg of random primer. Then, an aliquot of the cDNA product was subjected to amplification with Taq DNA polymerase for 30 cycles (20 cycles for GAPDH). The primers were as follows: CD44, 5'-AGACATCTACCCCAGCAAC-3' and 5'-CGTTGAGTCCACTTGGCTTTC-3'; and GAPDH; 5'-AAGGCTGAGAACGGGAAGCTTGTCATCAAT-3' and 5'-TTCCCGTCTAGCTCAGGGATGACCTTGCCC-3'.
Preparation and Introduction of Recombinant Adenovirus.
The cDNA encoding CD44H, tagged with a c-Myc epitope at the NH2 terminus and with a FLAG tag at the COOH terminus (CD44MF), was generated by a PCR-based method and subcloned into the vector plasmid pcDNA3.1(+)Zeo (Invitrogen). The cDNA insert was transferred to an adenoviral type 5 genome using the Adeno-X Expression System (Clontech, Palo Alto, CA). Recombinant adenovirus was amplified in human embryonic kidney 293 cells and purified using CsCl gradient centrifugation. Virus titers (infectious units/ml) were determined using an Adeno-X Rapid Titer Kit (Clontech).
ZR-75-1 cells (3 x 106) were seeded in 100-mm dishes and cultured in RPMI 1640 supplemented with 10% FBS and kanamycin for 24 h. The cells were infected with recombinant adenovirus for 1 h at 37°C in DMEM containing 5% FBS, and then the medium was exchanged with FBS-free RPMI 1640. After 48 h, cells were solubilized in radioimmunoprecipitation assay buffer [50 mM of Tris-HCl (pH 7.5), 1% Triton X-100, 1% deoxycolic acid, 0.1% SDS, and 150 mM of NaCl] in the presence of a protease inhibitor mixture (Roche Diagnostics, GmbH, Mannheim, Germany). Cell lysates were subjected to immunoprecipitation with anti-FLAG M2 antibody-conjugated agarose beads (Sigma Aldrich) as described previously (21) .
Deglycosylation of CD44H Molecules.
CD44MF was expressed in ZR-75-1 cells by infecting the cells with an adenovirus carrying the gene. The product was recovered by immunoprecipitation using anti-FLAG M2 antibody-conjugated agarose beads and suspended with 50 µl of 20 mmol of phosphate buffer (pH 7.0). Samples were then treated with 2 milliunits of neuraminidase, 0.2 milliunit of O-glycosidase, and/or 0.3 unit of N-glycosidase F (Roche Diagnostics, GmbH) in 20 mM of phosphate buffer (pH 7.0) at 37°C for 16 h. To express CD44 mutants in the S/T-rich region, the cells were transfected with the expression plasmids using FuGENE 6 (Roche Diagnostics, GmbH) according to the manufacturer’s instructions.
Immunohistochemistry.
Human tumor cell lines (MIAPaCa-2 and ZR-75-1) were cultured on a glass coverslip for 20 h. The cells were fixed with 3% paraformaldehyde and immunostained for CD44H using anti-CD44H antibodies. Alexa 488-conjugated goat antimouse IgG (Molecular Probes) was used as a secondary antibody. The signal was analyzed by confocal microscopy (Bio-Rad). Total cells were visualized by staining F-actin with Alexa 568-conjugated phalloidin (Molecular Probes). Human oral and breast carcinoma tissues were fixed with a periodate-lysine-paraformaldehyde fixative and embedded in paraffin wax. Sections were reacted with the antibodies (10 µg/ml), followed by biotinylated horse antimouse IgG (Vector Laboratories, Burlingame, CA) and avidin-biotin-horseradish peroxidase complex solution (DAKO, Glostrup, Denmark). Signals were visualized using 3,3'-diaminobenzidine tetrahydrochloride (Sigma Aldrich).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Additional oligopeptides from the NH2 terminus (RS1) and COOH-terminal region (P1) were also used as immunogens. Among several hybridoma cell lines that produce mAbs, four clones (285-2F12, 284-43F1, 268-1F5, and 294-6F2) were selected by Western blotting assay using E. coli cell lysate expressing CD44HS (Fig. 1B)
. mAbs 285-2F12 and 294-6F2 were obtained from mice immunized with the RS1 and P1 peptides, respectively, and mAbs 284-43F1 and 268-1F5 were obtained from mice immunized with CD44HS. Other antibodies used in Fig. 1B
were a rabbit polyclonal antibody (C6) against the P1 peptide and a commercially available mAb, 2C5.
|
, RS2–RS7) and analyzed similarly (Fig. 1D)
. mAb 2C5 reacted with the RS2 fragment as shown in the figure, but not with RS1 (data not shown). Thus, the extended aa sequence in RS2 is important as the epitope for mAb 2C5. Both mAbs 284-43F1 and 268-1F5 reacted with RS5 but not RS4, indicating that the sequence between 206 and 230 aa is important for reactivity. Using two polypeptides covering this region, the epitope of 284-43F1 was mapped to aa 205–219, and that of 268-1F5 was mapped to aa 220–230 (data not shown). Both mAb 294-6F2 and polyclonal antibody C6 were raised using P1 peptide, but they recognized different portions of the P1 peptide because 294-6F2 reacted with RS6, but C6 did not (Fig. 1D)
. In summary, the recognition sequence of each antibody is indicated on the CD44 sequence as in Fig. 1E
.
Reactivity of the Antibodies to the Native CD44 Expressed by Cell Lines.
To test the reactivity of the antibodies to the naturally modified CD44 molecule, 10 human tumor cell lines were examined. CD44 mRNA was detected by RT-PCR in seven cell lines but was undetectable or very weakly expressed in the remaining three (Fig. 2)
. All of the antibodies detected a major band of 85–95 kDa in pancreatic carcinoma MIAPaCa-2 cells and astrocytoma U-251 cells. Antibodies 285-2F12, 2C5, and 284-43F1 reacted to the CD44 expressed in all of the cell lines that express the mRNA, and the size of the major band was similar to that of MIAPaCa-2 and U-251 cell lines, although they differed slightly, depending on the cell lines. Polyclonal antibody C6 showed a similar recognition pattern, but it reacted weakly with the CD44 expressed in breast carcinoma MDA-MB231 cells.
|
Effect of Glycosylation on the Reactivity of the Antibodies.
To examine the possible effect of glycosylation on the reactivity of the antibodies, we used ZR-75-1 cells, which do not express endogenous CD44H (Fig. 2)
, and an adenovirus carrying the cDNA for CD44H with a Myc tag at the NH2 terminus and a FLAG tag at the COOH terminus (CD44MF). The CD44MF expressed in the cells was detected by Western blotting using either anti-cMyc or anti-FLAG antibody (Fig. 3)
and the size of the band was similar to that of the native CD44 molecules expressed in the cell lines in Fig. 2
. The mAbs 285-2F12, 284-43F1, and 2C5 detected CD44H, but 268-1F5 and 294-6F2 did not do so. Polyclonal antibody C6 reacted with CD44MF very weakly. Thus, CD44MF expressed in the cells is thought to be modified similarly to the native CD44 molecules expressed in the five cell lines used in Fig. 2
.
|
was treated with different types of glycosidases (neuraminidase, O-glycosidase, and N-glycosidase F) before Western blotting (Fig. 4A)
. Neuraminidase is an exoglycosidase for sialic acids at the terminus of N- or O-linked oligosaccharide chains. O-Glycosidase releases the disaccharide Gal ß 1,3GalNAc from O-glycans that modify serine or threonine residues. Because the sialic acid at the terminus prevents O-glycosidase from functioning, a combination of neuraminidase and O-glycosidase is required for efficient hydrolysis of O-glycans. N-Glycosidase F is an endoglycosidase and cleaves all types of N-glycans modified at asparagines. A slight shift in the migration of CD44MF was observed after treatment with neuraminidase as detected with 285-2F12 (Fig. 4)
, and cotreatment with O-glycosidase caused a further shift, although O-glycosidase alone did not change the size of CD44MF. Treatment with N-glycosidase F reduced CD44MF to 70 kDa, but no further shift was observed in the presence of neuraminidase. The size of the CD44MF treated with all of the enzymes at the same time was around 69 kDa.
|
. Treatment with either neuraminidase or O-glycosidase alone did not change the reactivity, but treatment with both enzymes made CD44MF reactive to the mAb, indicating that O-glycosylation with sialic acid presumably near the recognition site of mAb 268-1F5 abrogates antibody binding. Similar results were obtained when the endogenous CD44 of HT1080 cells, which was not detected with mAb 268-1F5 (Fig. 2)
, was treated with glycosidases (data not shown). The reactivity of polyclonal antibody C-6 was greatly enhanced by treatment with N-glycosidase F but only slightly enhanced by treatment with a combination of neuraminidase and O-glycosidase. There is a potential N-glycosylation site (NTT257) immediately downstream of the C-6 epitope, and a modification at this site may affect the reactivity of the mAb.
In contrast, none of the treatments rendered CD44MF reactive to mAb 294-6F2. Because the treatment of CD44MF with all of the glycosydases did not remove the sugar modifications completely, as was apparent from the molecular size, it is likely that residual modifications still mask the reactivity to the antibody. Notably, O-glycosidase does not digest the GlucNac and GalNac portions of O-glycans. A S/T-rich sequence (SHTT237) that contains potential O-glycosylation sites is found near the recognition site of mAb 294-6F2. These residues were sequentially substituted with alanines as in Fig. 5
to evaluate the effect of modifications at the sites. All of the mutants were expressed in ZR-75-1 cells at equivalent levels as detected by Western blotting using mAb 285-2F12 (Fig. 5)
. Wild-type CD44MF and a mutant with threonine 237 replaced with alanine (SHAT) did not show a change in reactivity. However, CD44MF molecules with mutations in serine 234/threonine 236–237 (AHAA), threonine 236–237 (SHAA), or threonine 237 (SHTA) were recognized by mAb 294-6F2. Thus, the serine 234 and threonine 237 residues near the antibody recognition site are important for the reactivity and probably have been modified by O-glycosylation. The effect of these mutations was specific to mAb 294-6F2 because they did not affect the reactivity to mAb268-1F5 (data not shown).
|
. Two human breast and three oral cancer tissues were examined by immunostaining. Because the staining pattern was the same among tumors of the same type, representative results are given in Fig. 6B
. Tumor cells and some stromal cells were stained with mAb 284-43F1. The stromal cells look like hemocytes invading interstitial cancer tissue because the cell shape is not like that of fibroblasts, and CD44H is known to be expressed in lymphocytes and histocytes. A similar staining pattern was obtained with mAb 2C5 (data not shown). In contrast, mAbs 268-1F5 and 294-6F2 showed cell type-specific differences in staining. Only tumor cells were stained with mAb268-1F5 in both breast and oral carcinomas. On the other hand, mAb 294-6F2 stained tumor cells in breast carcinomas but only stromal cells in oral cancer. The staining was specific to the epitope sequence because the signals were absorbed with the oligopeptides (data not shown). Thus, the reactivity of mAbs 268-1F5 and 294-6F2 clearly differed between the tumor and stromal cells in tumor tissue. Because the reactivity of these antibodies is affected by glycosylation, it is most likely that the glycosylation of CD44H near the antibody recognition sites is regulated in a cell type-specific manner in tumor.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Glycosylation of protein is a complex process that requires multiple glycosyl transferases and trimming enzymes (22 , 23) . Depending on the enzymes expressed in cells and their regulation, the glycosylation patterns for the polypeptide chain and components of the sugar moiety vary greatly. The vital importance of the glycosylation of proteins has been demonstrated using mice defective in different glycosyl transferases (24, 25, 26) . However, the function of glycosylation in each protein and specific site is not well understood. This is also the case for CD44 because it is not clear why CD44 is glycosylated at multiple sites or how important these modifications are, particularly in the stem region. In general, the glycosylation of a protein is thought to limit the accessibility of regulatory proteins including proteases for degradation and also provides new sites of interaction for proteins that have affinity to the sugar moiety. From this point of view, it is interesting that CD44 in tumor tissue showed a cell type-specific difference in reactivity to the mAbs 268-1F5 and 294-6F2. Notably, mAb 268-1F5 reacted exclusively with the CD44 expressed in tumor cells but not that in the surrounding stroma in five tumors including two breast cancers and three oral cancers. The reactivity of mAb 294-6F2 was variable depending on the tumor type. In breast carcinoma, the antibody reacted with CD44 in tumor cells but not stromal cells, but the reverse was true in oral cancer. Although more extensive analysis is needed, it is most likely that the glycosylation of CD44 affecting the reactivity to mAbs 268-1F5 and 294-6F2 is regulated differently between tumor and stromal cells in tissue. It is conceivable that CD44 modified in a tumor-specific manner has specific functions that contribute to the malignant phenotype of the tumors. This is presumably the first indication that CD44 molecules are modified differentially by glycosylation at particular sites in a cell type-specific manner. Thus, antibodies, the reactivity of which is affected by glycosylation at specific sites, would be useful for monitoring the state of glycosylation at these sites in vivo, studying the regulation and functions of the modification, and finally identifying the enzymes responsible for the glycosylation. To date, various anti-CD44 antibodies have been prepared and used for immunostaining. However, our study also pointed out the possibility that some epitopes of the antibodies might be masked in a cell type-specific manner by glycosylation. This possibility is particularly important when inconsistent results are obtained by using different anti-CD44 antibodies.
CD44 is also known to be cleaved in the stem region by proteases and shed into the blood stream. This shedding presumably regulates the ability of the cell to adhere to HA and modifies signaling mediated by CD44. Multiple cell surface proteases presumably participate in the processing event (27) , and MT1-MMP is one protease having such an activity (19 , 28) . No other candidate proteases have been identified yet. If different proteases shed CD44, the cleavage sites will differ, depending on the proteinases responsible. However, the molecular sizes of the shed fragments do not provide definitive information about the cleavage sites, in that the molecular mass of CD44 is highly variable depending on the glycosylation. Antibodies with defined epitopes in the stem region will be useful for analyzing the CD44 fragments generated by different proteases. Malignant tumor cells show enhanced CD44 shedding (29) , and MT1-MMP is also expressed in such tumors (28 , 30) . Thus, these sets of antibodies can be used to detect the CD44 fragment shed by a particular enzyme, MT1-MMP for example. To pursue this possibility, we are currently developing a sandwich ELISA.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Motoharu Seiki, Division of Cancer Cell Research, Institute of Medical Science, The University of Tokyo, 4-6-1 Shirokane-dai, Minato-ku, Tokyo 108-8639, Japan. Phone: 81-3-5449-5255; Fax: 81-3-5449-5414; E-mail: mseiki{at}ims.u-tokyo.ac.jp
4 The abbreviations used are: HA, hyaluronan; MT1-MMP, membrane-type 1 matrix metalloproteinase; FBS, fetal bovine serum; mAb, monoclonal antibody; RT-PCR, reverse transcription-PCR; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; aa, amino acid(s); S/T, serine/threonine. 
Received 6/18/03. Revised 9/21/03. Accepted 10/14/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. K. Diaz, X. Zhou, E. T. Wright, M. Cristofanilli, T. Smith, Y. Yang, N. Sneige, A. Sahin, and M. Z. Gilcrease CD44 Expression Is Associated with Increased Survival in Node-Negative Invasive Breast Carcinoma Clin. Cancer Res., May 1, 2005; 11(9): 3309 - 3314. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Nakamura, N. Suenaga, K. Taniwaki, H. Matsuki, K. Yonezawa, M. Fujii, Y. Okada, and M. Seiki Constitutive and Induced CD44 Shedding by ADAM-Like Proteases and Membrane-Type 1 Matrix Metalloproteinase Cancer Res., February 1, 2004; 64(3): 876 - 882. [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 |
