Expression of Sialyltransferase Family Members in Cervix Squamous Cell Carcinoma Correlates with Lymph Node Metastasis

Gynecologic Oncology 86, 45–52 (2002) doi:10.1006/gyno.2002.6714

Expression of Sialyltransferase Family Members in Cervix Squamous Cell Carcinoma Correlates with Lymph Node Metastasis 1 Peng-Hui Wang, M.D.,* Ywan Feng Li, Ph.D.,* Chi-Mou Juang, M.D.,* Yan-Ru Lee, M.B.,* Hsiang-Tai Chao, M.D., Ph.D.,* Heung-Tat Ng, M.D.,* Ying-Chieh Tsai, Ph.D.,† ,2 and Chiou-Chung Yuan, M.D.* ,2,3 *Department of Obstetrics and Gynecology, Taipei Veterans General Hospital and Institute of Clinical Medicine, and †Department of Biochemistry and Institute of Biochemistry, National Yang-Ming University, Taipei 112, Taiwan Received November 29, 2001

INTRODUCTION

Objective. Altered messenger ribonucleic acid (mRNA) expression of the four sialyltransferases (STs including ST3Gal I, ST3Gal III, ST3Gal IV, and ST6Gal I) is important in squamous cell carcinoma of the cervix. This study further investigates their changes in mRNA expression of the four STs in FIGO stage IB1 squamous cell carcinoma to assess the extent of sialylation associated with lymph node metastases. Methods. Alterations in ST mRNA expression in FIGO IB1 cervical squamous cell carcinomas (n ⴝ 79) were examined by semiquantitative reverse transcription–polymerase chain reaction. Results. Both ST6Gal I mRNA and ST3Gal III mRNA expressions were significantly increased in patients with lymph node metastases compared to those without lymph node metastases (P ⴝ 0.002 and P ⴝ 0.001, respectively, Mann–Whitney U test). Using receiver operating characteristic curves of ST ratio index for accuracy comparison of lymph node metastases, ST3Gal III and ST6Gal I were observed to be fairly interchangeable (area under the curve (AUC) of 3Gal I ⴝ 0.810; AUC of 6Gal I ⴝ 0.786, significance of difference between AUC ⴝ 0.810). High ST6Gal I expression was associated with other invasive properties of cervical cancer, such as deep stromal invasion and presence of lymph– vascular space involvement. ST6Gal I expression seemed to be more enhanced in bigger tumors. Conclusions. Our results suggested that ST3Gal III and ST6Gal I were of importance for the lymph node metastases in FIGO IB1 cervical cancer patients; more specifically, overexpression of ST6Gal I was of crucial relevance for the presence of poor prognostic factors, such as deep stromal invasion and lymph–vascular space involvement and lymph node metastases. © 2002 Elsevier Science

Sialic acids including a number of their derivatives are ubiquitous at the terminal positions of oligosaccharides of glycoproteins [1, 2]. Due to their acidic nature, they impart a net negative charge to the cell surface and are important in cell– cell or cell–matrix interaction [3, 4]. The transfer of the sialic acids from cystidine-5-monophospho-N-acetylneuraminic acid (CMP-NeuAc) to the terminal position of carbohydrate group of glycoproteins and glycolipids is catalyzed by a family of sialyltransferases (STs) [5]. There is a large body of evidence to suggest that tumor cells have changed surface properties from their normal counterparts, that these changes are partially due to altered sialoglycoconjugates expressed on the plasma membrane [4], and that altered sialylation (change in glycoprotein expression), which occurs during certain pathological processes, such as oncogenic transformation, tumor metastases, and invasion, is associated with enhanced ST activity [6 –20]. Increased ␤-1,6-branching, increased SialylLewis epitopes, increased Sialyl-Tn antigen (Sialyl-␣-Ser/ Thr), or the general increase in sialylation of cell surface glycoproteins is commonly observed in N-lined and O-lined oligosaccharides of carcinoma cells [6, 17, 21]. Carbohydrate changes occur common in breast cancer [15, 22–24], colorectal cancer [1, 12, 21, 25–29], lung cancer [30, 31], hepatic carcinoma [32, 33], gastric carcinoma [34], head and neck squamous cell carcinoma [35], brain tumor [36, 37], choriocarcinoma [38], prostate cancer [39], and squamous cell carcinoma of the cervix [13]. More specifically, these changes in glycoprotein expression are reported to play important roles in tumor grade, invasion, metastatic ability, and poor clinical outcome [7, 12, 15, 21, 23–25, 29 –31, 33, 35, 37]. Some of these studies also evaluated the mRNA expression of STs and found that increased mRNA expression of STs is correlated with poor outcome in breast cancers [15, 16, 23–25] and colon carcinoma [1, 12]. There are some studies that only evaluated the relationship between the change of ST mRNA expression and their clinical correlation, including colorectal cancer [27] and gastric cancer [34]. So far, altered expression of mRNA of STs in

(USA)

Key Words: reverse transcription–polymerase chain reaction (RT-PCR); sialyltransferase; squamous cell carcinoma of the cervix. 1

This work was supported in part by a grant 90VGH-279 from the Taipei Veterans General Hospital and grant (NSC-90-2314-B-075-137) from the National Science Council, Taiwan. 2 These authors contributed equally. 3 To whom reprint requests should be addressed at Department of Obstetrics and Gynecology, Veterans General Hospital–Taipei, 201, Section 2, Shih-Pai Road, Taipei 112, Taiwan. Fax: ⫹886-2-28734101. E-mail: phwang@ vghtpe.gov.tw. 45

0090-8258/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.

46

WANG ET AL.

gynecologic cancers is still unavailable, except one [13]. In our earlier study [13], we found that the combination of enhanced ST6Gal I mRNA expression and decreased mRNA expression from ST3Gal I, ST3Gal III, and ST3Gal IV is important in cervical cancer. Because lymph node metastasis in cervical cancers is the most important and independent risk factor for patients’ recurrence and survival, in this study, we assess the differences between these expressions in the FIGO stage IB1 squamous cell carcinomas of the cervix with lymph node metastases and that without lymph node metastases to search their roles for lymph node metastases in these cervical cancers. MATERIALS AND METHODS This prospective study after excluding initial 6 patients without consent forms for this study involved 88 patients, who could sign their consent forms according to the guidelines of the Human Ethics Committee in the Department of Obstetrics and Gynecology, Taipei Veterans General Hospital (the National Medical Center of Taiwan), for study, undergoing radical hysterectomy and pelvic lymph node dissection for FIGO stage IB1 squamous cell carcinoma of the cervix at the same institute between January and December 2000. Excluding 9 patients with parametrial invasion (n ⫽ 7), vaginal invasion (also the presence of parametrial invasion), or uterine body invasion (n ⫽ 3), 68 patients did not have any evidence of extracervical metastases and 11 patients had only pelvic lymph node metastases without other extracervical invasion, and these were proven by pathological examination. All specimens were collected and processed as usual [13].

the SuperScript preamplification system for first-strand cDNA Synthesis (Life Technologies), according to the manufacturer’s protocol with oligo(deoxythymidine) as the initiation primer in a final reaction volume of 20 ␮l. The synthesized cDNA in 1 ␮l was subjected to PCR amplification using four pairs of ST-specific primers, respectively, and the GAPDH-specific primers. Meanwhile, the extracted RNA from the individual sample was subjected to the following PCR reaction as negative control. The condition used for the PCR was set up based on the titration experiment conducted, which was described in detail before [13]. The PCR mixture consisted of 0.1 U Super-Therm DNA polymerase (JMR Holdings), 1x PCR buffer (JMR Holdings), 1.5 mM MgCl 2, 320 nM dNTPs, 0.8 –1.6 ␮l of 10 ␮M ST-specific primers, 0.8-1.6 ␮l of 10 ␮M GAPDH-specific primers, 1 ␮l of cDNA, and extradistilled water added to 50 ␮l. The volume of ST-specific primers (10 ␮M) used was 1.6 ␮l for ST3Gal I and ST3Gal III, 1.2 ␮l for ST3Gal IV, and 0.8 ␮l for ST6Gal I. The volume of GAPDH primers (10 ␮M) used was 1.6 ␮l for ST3Gal I and ST3Gal III and 0.8 ␮l for ST3Gal IV and ST6Gal I. All PCR experiments were conducted in the GeneAmp PCR system 2400 (Applied Biosystem). Sialyltransferase mRNA expression of each sample was determined semiquantitatively in at least three independent experiments using Alpha-Imager 2000. Using GAPDH as an internal standard, the deviation between triplicate measurements was on average 21%. In all experiments, a negative control reaction was performed by replacing the cDNA template with sterile water, and positive controls were performed with HepG2 cDNA. Sialyltransferase Assay

Semiquantitative Reverse Transcription–Polymerase Chain Reaction (RT-PCR) for mRNA Expression Tissue was snap-frozen in liquid nitrogen until RNA extraction. In total, 79 samples of FIGO stage IB1 squamous cell carcinoma of the cervix were studied. Tissues were powdered in liquid nitrogen. The method of RNA extraction and semiquantitative RT-PCR has been described in detail before [13]. Briefly, RNA was extracted using StrataPrep total RNA miniprep kit (Stratagene, La Jolla, CA). RNA yield and quality were determined by spectrophotometer (Hitachi U-3210, Tokyo, Japan) and agarose gel electrophoresis, while viability of the RNA in each tissue sample was confirmed by amplification of complementary DNA (cDNA) for GAPDH. Total cellular RNA (1 ␮g) and oligo(deoxythymidine) 18 (1 ␮g) were heated at 70°C for 10 min and placed on ice for at least 1 min. Seven microliters of the reaction mixture (1⫻ first-strand buffer, 10 mM dithiothreitol, 0.5 mM deoxyribonucleotide triphosphate (dNTP)) was added, mixed gently, and incubated at 42°C for 5 min. One microliter (200 U) of SuperScript II was added and mixed by gentle pipetting, and then it was incubated 50 min at 42°C. The reaction was inactivated by heating at 70°C for 15 min. Reverse transcription into cDNA was achieved by using

Gal␤1,3GalNAc-acetyllactosamine)-Obzl (acceptor for ST2, 3Gal I), Gal␤1,3GlcNAc␤1,3Gal␤1,4GlcNAc (acceptor for ST2,3Gal III), Gal␤1,4GlcNAc (acceptor for ST2,3Gal IV), and asialo-fetuin (acceptor for ST6Gal I) were from Sigma (St. Louis, MO). CMP-[ 14C]NeuAc (100 nCi/␮l) was from Amersham (Amersham Pharmacia Biotech, UK). Specimens were washed with ice-cold phosphate-buffered saline, pH 7.3, at 4°C. Cells were then pelleted and incubated on a rocking plate for 30 min at 4°C in 500 ␮l 0.25 M sodium cacodylate buffer, pH 6.5, containing 0.4% Triton X-100, and then sonicated with ten 1-s pulses on ice [1, 40]. Nuclei were removed by low-speed centrifugation and supernatant was centrifuged at 30,000g at 4°C for 30 min. The pellet was suspended in ice-cold 100 mM sodium cacodylate buffer (pH 6.0) and the supernatant was concentrated 10-fold Amicon 30 filters and used as the enzyme source. Protein content in diluted supernatants was determined by the Bio-Rad protein assay kit (Munich, Germany) using bovine serum albumin as standard. The enzyme assay method was modified from Kurosawa et al.’s original design [40] and had been described before [41]. Briefly, the enzyme assays with glycoproteins, oligosaccharides, and glycolipids as acceptors were performed in the

SIALYLTRANSFERASE AND CERVICAL CARCINOMA

presence of 0.1 M sodium cacodylate buffer, pH 6.0, 10 mM MgCl 2, 0.5% Triton CF-54, 100 ␮M CMP-[ 14C]NeuAc (100 nCi/␮l), 2 ␮l of acceptor (1 mM oligosaccharide or 4 mg/ml glycoprotein), and 5 ␮l of individual cell line– cell lysate in a final volume of 10 ␮l with incubation at 37°C for 1 h. At the end of the incubation period, the reaction mixtures were subjected to SDS–polyacrylamide gel electrophoreses (SDSPAGE) for glycoproteins as acceptors or were subjected chromatography on high-performance thin-layer chromatography plates (Merck, Darmstadt, Germany) with a solvent system of ethanol/1-butanol/pyridine/acetic acid/water (100/10/10/2/30) for glycolipids and glycoprotein as acceptors and with a solvent system of ethanol/pyridine/n-butanol/water/acetic acid (100/10/10/30/5) for oligosaccharides as acceptors. The radioactivity of the corresponding products and total radioactivity were detected and quantified with a BAS 1500 radioimage analyzer (Fuji Film, Japan). To obtain the quantitative data, experiments were reexamined under the linear condition with time and enzyme concentration (enzyme activity unit, nanomol sialic acid/mg protein/h). Statistics Statistical analysis of data was carried out using the statistical software SPSS, version 10.0 (SPSS Inc., Chicago, IL). Mann–Whitney U test was used to test differences between the expression of a given ST in cervical cancer tissues on the patients with lymph node metastases and those without lymph node metastases. Test accuracy between expression of a given ST and different poor pathological parameters including lymph node metastases, deep stromal invasion (equal to or more than 1/2 of the cervix in thickness), lymph or vascular space involvement, and tumor size based on our previous report [42– 44] was also tested using receiver operating characteristic (ROC) curve. Differences between area under curve (AUC) were compared with modification of Wilcoxon rank sum test [45– 47]. Multiple regression models were used to determine which pathological risk factors had more influence in predicting ST mRNA expression. Probability values below 5% were considered statistically significant. RESULTS Four ST mRNAs expression in squamous cell carcinomas of the cervix without lymph node metastases (n ⫽ 68) and those without lymph node metastases (n ⫽ 11) were examined by RT-PCR with coamplification of GAPDH. The RNA quality was assessed by running it on an agarose gel, while viability of the RNA in each tissue sample was confirmed by amplification of complementary DNA for GAPDH. The standard deviation between triplicate measurements was 21%. Carcinomas with lymph node metastases showed increased ST3Gal III (median 1.918) and ST6Gal I (median 1.841) expression compared to those without lymph node metastases, with median 1.053, P ⫽

47

0.001; and median 1.0465, P ⫽ 0.002, respectively (Fig. 1). Using ROC curves of ST ratio index for accuracy comparison of lymph node metastases, ST3Gal III and ST6Gal I were observed to be fairly interchangeable with AUC of 0.810 and 0.786, respectively, P ⫽ 0.810 (Fig. 2). Using multiple regression models for predicting ST3Gal III mRNA expression and for predicting ST6Gal I mRNA expression, we found that the presence of lymph node metastases could be accurately predicted by both ST6Gal I mRNA expression and ST3Gal III mRNA expression (Tables 1 and 2). Only lymph node metastases could predict the increased ST3Gal III mRNA expression, but nearly all conventionally poor prognostic factors including poorly differentiation, deep stromal invasion, and lymph–vascular space involvement could also predict the increased ST6Gal I mRNA expression (Table 2). In addition, using receiver operating characteristic curve model, high ST6Gal I expression was associated with more invasive property of cervical cancer such as lymph–vascular space involvement with AUC of 0.918 and deep stromal invasion with AUC of 0.839. Finally, ST6Gal I expression showed a marginal significant linear correlation for larger tumor size, with r ⫽ 0.223, P ⫽ 0.055. ST mRNA expression in cervical cancer specimens correlated well (correlation coefficient r 2 ⫽ 0.8726 in ST3Gal I, 0.9036 in ST3Gal IV, and 0.9313 in ST6Gal I, respectively) with enzyme activity, but ST3Gal III mRNA expression correlated marginally well (correlation coefficient r 2 ⫽ 0.7631) with ST3Gal III enzyme activities. Both ST6Gal I enzyme and ST3Gal III enzyme activities were also significantly increased in patients with lymph node metastases compared to those without lymph node metastases (P ⫽ 0.001 and P ⫽ 0.014, respectively, Mann–Whitney U test). DISCUSSION Lymph node metastasis in early stage cervical cancers is still the most important and independent risk factor, not only for recurrence but also for survival. Many pathological factors including bulky tumor size, poor differentiation, presence of lymph–vascular space involvement, presence of parametrial invasion, and deep stromal invasion are related with lymph node metastases. Besides these conventional pathological parameters for predicting prognosis in an early-stage cervical cancer, there are many biological factors (cellular molecules), which show close correlation with aggressive and invasive behaviors in tumors [12–16, 48 –53]. Among these biological factors, sialic acids are one of the most promising molecules because they involve cell– cell and cell–matrix interactions and cellular recognition [19, 20]. The levels of ␣-2,3-sialyltransferase ST3Gal I, ST3Gal III, ST3Gal IV, and ␣-2,6-sialyltransferase ST6Gal I messages were found to be significantly changed in cancer tissues [1, 13, 24, 26]. Although at least 18 distinct ST genes exist [54, 55], all involve tumor-associated changes in the expression of cellsurface sialoglycoconjugates. However, we did not study the

48

WANG ET AL.

FIG. 1. Messenger RNA expression of the four sialyltransferases in cervical cancer tissues without lymph node metastases and those with lymph node metastases. Data are presented as a box-plot diagram displaying median values (represented by the thick horizontal lines), 25th percentile, 75th percentile, and data within the 10th and 90th percentiles.

FIG. 2. Receiver operating characteristic (ROC) curves of ST ratio index. Comparison for accuracy in predicting lymph node metastases on the patients with cervical cancer; ST3Gal III and ST6Gal I were observed to be fairly interchangeable (area under the curve (AUC) of 3Gal I ⫽ 0.810; AUC of 6Gal I ⫽ 0.786; significance of difference between AUC ⫽ 0.810). ST3Gal IV and ST3Gal I were found to be far less accurate.

49

SIALYLTRANSFERASE AND CERVICAL CARCINOMA

TABLE 1 Multiple Regression Model for Predicting ST3Gal III mRNA Expression Unstandarized coefficient

Differentiation Deep stromal invasion Lymph–vascular space involvement LN metastases

B

SE

Standarized coefficient ␤

0.075 ⫺0.079 0.180 0.566

0.161 0.189 0.191 0.225

0.057 ⫺0.054 0.125 0.285

other 14 STs yet, and we would like to expand the other STs in the same group such as ST3Gal II, V, and VI and ST6GalNAc I, II, and III in the near future. In this report, which was a continuous study based on findings in our earlier study [13], we found that ST6Gal I mRNA expression statistically increased in squamous cell carcinoma of the cervix compared with the normal cervix, high mRNA expression of ST6Gal I was positively correlated to high ST3Gal III expression either on normal cervix tissue or on cervical carcinoma tissue, and cancers with lymph–vascular space involvement seemed to present a trend of increasing mRNA expression of ST3Gal III. We also suspected that ST3Gal III might be used as a predictor in cervical cancer patients with high risk of distant metastases [13]; we expanded the study to search their roles for lymph node metastases in patients with FIGO IB1 cervical cancer. In addition, we further evaluated the enzyme activities of STs in these specimens and found a good correlation between the ST enzyme activities and the ST mRNA expression, except for a marginally good correlation between mRNA expression and enzyme activity in ST3Gal III. ␣-2,3-Sialyltransferase ST3Gal III or ST3Gal IV carries on transfer of sialic acid in the ␣-2,3-linkage to Gal␤1, (3)4GlcNAc on N-linked chains of glycoproteins. The latter is also capable of adding sialic acid to Gal␤1,3GalNAc found on O-linked chains of glycoproteins. The role of ST3Gal III or ST3Gal IV was also discussed before [30, 56]. ST3Gal III is involved in the biosynthesis of sLe (x) and sLe (a) that are known as selectin ligands and tumor-associated carbohydrate structures [56], which play an important role in tumor metastases because the E- or P-selectin expressed on the surface of vascular endothelial cells interacts with sialyl Lewis antigens,

P

95% for B

0.644 0.677 0.348 0.014

⫺0.246–0.395 ⫺0.456–0.298 ⫺0.200–1.724 0.116–1.015

such as sLe (x), sDLe (x), and sLe (a), expressed on the surface of malignant cells, and mediates the adhesion of malignant cells to the vascular endothelium [57, 58]. The role of ST3Gal III mRNA expression in cervical cancer tissues seemed to be confusing and difficult to interpret because our earlier study [13] demonstrated that down-regulation of ST3Gal III mRNA expression often occurred in the cervical cancer specimens, but in this study, we found that overexpression of ST3Gal III mRNA expression was closely correlated to lymph node metastases, just like the findings from studying breast cancer and lung cancer [15, 31]. The expression of ST3Gal III mRNA in normal cervical tissue specimens showed a much wider range of discrepancy but this dissociation of ST3Gal III was not significant in cervical cancer specimens [13]. The former could statistically significantly affect the mean value of data interpretation in our earlier study [13]. In addition, the dissociation between the ST3Gal III mRNA expression and ST3Gal III enzyme activities would partly contribute explain this discrepancy. We suspected that using more sensitive detection methods, for example, real-time RT-PCR for quantifying mRNA or monoclonal antibody for quantifying ST3Gal III enzyme protein, would clarify this conflicted result, although we did not perform these experiments. Finally, because the expression of ST3Gal III mRNA involved the adhesion between the surrounding tissue, reduced cancer cell attachment to the surrounding stroma of the cervix might increase the ability of “drop of the cancer cells” or “escape of the cancer cells” from the surrounding tissue when down-regulation of ST3Gal III mRNA expression occurs in oncogenesis (earlier stage of cancer development), which is likely to help their release into the peripheral blood or lymphatic circulation. After dissemination,

TABLE 2 Multiple Regression Model for Predicting ST6Gal I mRNA Expression Unstandarized coefficient

Differentiation Deep stromal invasion Lymph–vascular space involvement LN metastases

B

SE

Standarized coefficient ␤

0.431 0.239 0.503 0.389

0.090 0.106 0.107 0.126

0.367 0.183 0.388 0.220

P

95% for B

0.000 0.027 0.000 0.003

0.251–0.611 0.027–0.450 0.289–0.716 0.138–0.641

50

WANG ET AL.

adhesion to the endothelial cells became more important so increased expression of ST3Gal III mRNA might enhance the interaction between cancer cells and vascular endothelium to help establishing successful metastases. We suspected that increased expression of ST3Gal III might be a late event in cancer development, at least followed by increased expression of ST6Gal I mRNA. ST6Gal I is responsible for the addition of sialic acid in the ␣-2,6-linkage to Gal␤1,4GlcNAc (N-acetyllactosamine), a sequence commonly found in N-linked chains of glycoproteins. Much evidence supports that enhanced ST6Gal I expression possibly is important in cancer development and progression [15, 36, 37, 40, 42, 43]. The majority of the studies showed high ST6Gal I expression was associated with poor histopathological parameters, especially in solid tumors, such as grade III poor differentiation [15], absence of progesterone receptor [15], and invasive potential [28, 32, 34, 35]. Only the hematogeneous cell lines showed the diverse characteristics which showed overexpression of the ST6Gal I in a low metastatic variant of a murine lymphoblastoid cell line [60]. Hui and Le Marer showed that high cell-surface sialylation reduces matured myeloid attachment to the bone marrow stroma, which is likely to help their release into the peripheral blood, and further found that up-regulation of ST6Gal I together with ␣-2,6sialylation of cell-surface glycoproteins is likely to modulate cellular adherence and survival of progenitor cells [61]. In addition, the different sialylation of tumor-associated lymphocytes might be also important in survival of the cancer patients. Elpek et al. showed that CDw75 antigen, which is a sialylated carbohydrate epitope generated by the enzyme ␣-2,6-sialyltransferase, expression of the B lymphocytes was correlated with larger tumor size, infiltrative growth pattern, advanced stage, and positive lymph nodes [62]. Every kind of chemotherapy has its specific killing or inhibiting points on cancer cells. Azuma et al. found that the sialic acids on the cancer cell surface were also significantly affected after chemotherapy treatment [59]. In our laboratory, we also successfully identify soyasaponin I, which is a potent and specific sialyltransferase inhibitor for ST3Gal I, but not against the other glycosyltransferases and glycosidases [14]. With advancing biotechnology, more specific molecules against different kinds of STs might be found in the near future and this strategy might provide a vision of possibly synergistic therapy in managing cancer patients. In conclusion, an early event in tumorigenesis of the squamous cell carcinoma of the cervix is up-regulation of ST6Gal I and its positive feedback control was associated with growing and spreading of tumors. A following event might be overexpression of ST3Gal III, which showed a significant association with lymph node metastases in FIGO IB1 cervical squamous cell carcinoma patients. Although the clinically long-term significance of increased expression of ST6Gal I and ST3Gal III in patients with early stage cervical cancers has not been clarified yet, regarding survival, this study provides a biolog-

ical model for monitoring lymph node metastases and finding factors to overcome this problem, which may change our therapeutic approach. ACKNOWLEDGMENTS The authors appreciate Miss Shu-Ching Liang, Miss Wen-Yuann Shyong, and Miss Yu-Ying Ho for their help in preparation of this paper.

REFERENCES 1. Petretti T, Kemmner W, Schulze B, Schlag PM. Altered mRNA expression of glycosyltransferase in human colorectal carcinomas and liver metastases. Gut 2000;46:359 – 66. 2. Lowe JB. Carbohydrate recognition in cell-cell interaction. In: Fukuda M, Hindsgaul O, editors. Molecular glycobiology. Oxford: Oxford University Press; 1994. p. 163–94. 3. Thomas P. Cell surface sialic acid as a mediator of metastatic potential in colorectal cancer. Cancer J 1996;9:1–10. 4. Schauer R. Sialic acids and their role as biological masks. Trends Biochem Sci 1985;10:357– 60. 5. Kim YJ, Kim KS, Kim SH, Kim CH, Ko JH, Choe IS, Tsuji S, Lee YC. Molecular cloning and expression of human Gal␤1,3GalNac␣2,3-sialyltransferase (hST3GalII). Biochem Biophys Res Commun 1996;228:324 –7. 6. Fukuda M. Possible roles of tumor-associated carbohydrate antigens. Cancer Res 1996;56:2237– 44. 7. Hakomori S. Tumor malignancy defined by aberrant glycosylation and shingo-(glyco)-lipid metabolism. Cancer Res 1996;56:5309 –18. 8. Dorudi S, Kinrade E, Marshall NC, Feakins R, Williams NS, Bustin SA. Genetic detection of lymph node micrometastases in patients with colorectal cancer. Br J Surg 1998;85;98 –100. 9. Aubert M, Panicot L, Crotte C, Gibier P, Lombardo D, Sadoulet MO, Mas E. Restoration of ␣(1,2)fucosyltransferase activity decreases adhesive and metastatic properties of human pancreatic cancer cells. Cancer Res 2000; 60:1449 –56. 10. Zhu Y, Srivatana U, Ullah A, Gagneja H, Berenson CS, Lance P. Suppression of a sialyltransferase by antisense DNA reduces invasiveness of human colon cancer cells in vitro. Biochim Biophys Acta 2001;1536:148 – 60. 11. Yamamoto H, Oviedo A, Sweeley C, Saito T, Moskal JR. ␣2,6-sialylation of cell-surface N-glycans inhibits glioma formation in vivo. Cancer Res 2001;61:6822–9. 12. Schneider F, Kemmner W, Haensch W, Franke G, Gretschel S, Karsten U, Schlag PM. Overexpression of sialyltransferase CMP-sialic acid: Gal␤1,3GalNAc-R␣6-Sialyltransferase is related to poor patient survival in human colorectal carcinomas. Cancer Res 2001;61:4605–11. 13. Wang PH, Li YF, Juang CM, Lee YR, Chao HT, Tsai YC, Yuan CC. Altered mRNA expression of sialyltransferase in squamous cell carcinomas of the cervix. Gynecol Oncol 2001;83:121–7. 14. Wu CY, Hsu CC, Chen ST, Tsai YC. Soyasaponin I, a potent and specific sialyltransferase inhibitor. Biochem Biophys Res Commun 2001;284: 466 –9. 15. Recchi MA, Hebbar M, Hornez L, Harduin-Lepers A, Peyrat JP, Delannoy P. Multiplex reverse transcription polymerase chain reaction assessment of sialyltransferase expression in human breast cancer. Cancer Res 1998;58: 4066 –70. 16. Recchi MA, Harduin-Lepers A, Boilly-Marer Y, Verbert A, Delannoy P. Multiplex RT-PCR method for the analysis of the expression of human sialyltransferase: application to breast cancer cells. Glycoconj J 1998;15: 19 –27.

SIALYLTRANSFERASE AND CERVICAL CARCINOMA 17. Brockhausen I, Yang J, Lehotay M, Ogata S, Itzkowitz S. Pathways of mucin O-glycosylation in normal and malignant rat colonic epithelial cells reveal a mechanism for cancer-associated Sialyl-Tn antigen expression. Biol Chem 2001;382:219 –32. 18. Kannagi R. Carbohydrate-mediated cell adhesion involved in hematogenous metastasis of cancer. Glycoconj J 1997;14:577– 84. 19. Harduin-Lepers A, Recchi MA, Delannoy P. 1994, the year of sialyltransferase. Glycobiology 1995;5:741–58.

51

teins of head and neck squamous-cell carcinomas. Eur Arch Oto-RhinoLaryngol 1997;254:437– 41. 36. Yamamoto H, Saito T, Kaneko Y, Kersey D, Yong VW, Bremer EG, Mkrdichian E, Cerullo L, Leestma J, Moskal JR. ␣2,3-sialyltransferase mRNA and ␣2,3-linked glycoprotein sialylation are increased in malignant gliomas. Brain Res 1997;755:175–9.

20. Tsuji S. Molecular cloning and functional analysis of sialyltransferases. J Biochem (Tokyo) 1996;120:1–13.

37. Kaneko Y, Yamamoto H, Kersey DS, Colley KJ, Leestma JE, Moskal JR. The expression of Gal␤1,4GlcNAc ␣2,6 sialyltransferase and ␣2,6-linked sialoglycoconjugates in human brain tumors. Acta Neuropathol 1996;91: 284 –92.

21. Itzkowitz SH, Yuan M, Montgomery CK, Kjeldsen T, Takahashi HK, Bigbee WL, Kim YS. Expression of Tn, sialosyl-Tn, and T antigens inhuman colon cancer. Cancer Res 1989;49:197–204.

38. Fukushima K, Hara-Kuge S, Seko A, Ikehara Y, Yamashita K. Elevation of ␣236 sialyltransferase and␣132 fucosyltransferase activities in human choriocarcinoma. Cancer Res 1998;58:4301– 6.

22. Ura Y, Dion AS, Williams CJ, Olsen BD, Redfield ES, Ishida M, Herlyn M, Major PP. Quantitative dot blot analyses of blood-group-related antigens in paired normal and malignant human breast tissues. Int J Cancer 1992;50:57– 63.

39. Okajima T, Fukumoto S, Miyazaki H, Ishida H, Kiso M, Furukawa K, Urano T, Furukawa L. Molecular cloning of a novel␣2,3-sialyltransferase (ST3Gal IV) that sialylates type II lactosamine structures on glycoproteins and glycolipids. J Biol Chem 1999;274:11479 – 86.

23. Burchell J, Poulsom R, Hanby A, Whitehouse C, Cooper L, Clausen H, Miles D, Taylor-Papadimitriou J. An ␣2,3sialyltransferase (ST3Gal I) is elevated in primary breast carcinomas. Glycobiology 1999;9:1307–11.

40. Kurosawa N, Kojima N, Inoue M, Hamamoto T, Tsuji S. Cloning and expression of Gal␤1,3GalNAc-specific GalNAc␣2,6-sialyltransferase. J Biol Chem 1994;269:19048 –53.

24. Matsuura N, Narita T, Hiraiwa N, Hiraiwa M, Murai H, Iwase T, Funahashi H, Imai T, Takagi H, Kannagi R. Gene expression of fucosyl- and sialyl-trans ferases which synthesize sialyl Lewisx, the carbohydrate ligands for E-selectin, in human breast cancer. Int J Oncol 1998;12:1157– 64.

41. Wang PH, Lo WL, Hsu CC, Lin TW, Lee WL, Wu CY, Yuan CC, Tsai YC. Different enzyme activities of sialyltransferases in gynecological cancer cell lines. Eur J Gynecol Oncol 2002;23:221– 6.

25. Ito H, Hiraiwa N, Sawada-Kasugai M, Akamatsu S, Tachikawa T, Kasai Y, Akiyama S, Ito K, Takagi H, Kannagi R. Altered mRNA expression of specific molecular species of fucosyl- and sialyl-transferases in human colorectal cancer tissues. Int J Cancer 1997;71:556 – 64.

42. Yuan CC, Wang PH, Lai CR, Yen MS, Chen CY, Juang CM. Prognosispredicting system based on factors related to survival of cervical carcinoma. Int J Gynecol Obstet 1998;63:163–7. 43. Yuan CC, Wang PH, Lai CR, Tsu EJ, Yen MS, Ng HT. Recurrence and survival analyses of 1,115 cervical cancer patients treated with radical hysterectomy. Gynecol Obstet Invest 1999;47:127–32.

26. Yamada N, Chung YS, Takatsuka S, Arimota Y, Sawada T, Dohi T, Sowa M. Increased sialyl sialyl lewis a expression and fucosyltransferase activity with acquisition of a high metastatic capacity in a colon cancer cell line. Br J Cancer 1997;76:582–7.

44. Juan CM, Wang PH, Yen MS, Lai CR, Ng HT, Yuan CC. Application of tumor markers CEA, TPA, and SCC-Ag in patients with low-risk FIGO stage IB and IIA squamous cell carcinoma of the uterine cervix. Gynecol Oncol 2000;76:103– 6.

27. Kudo T, Ikehara Y, Togayachi A, Morosumi K, Watanabe M, Nakamura M, Nishihara S, Narimatsu H. Up-regulation of a set of glycosyltransferase genes in human colorectal cancer. Lab Invest 1998;78:797– 811.

45. Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839 – 43.

28. Dall’Olio FD, Chiricolo M, Lau JTY. Differential expression of the hepatic transcript of ␤-galactoside ␣2,6 sialyltransferase in human colon cancer cell lines. Int J Cancer 1999;81:243–7.

46. McNeil BJ, Hanley JA. Statistical approaches to the analysis of receiver operating characteristic (ROC) curves. Med Decision Making 1984;4: 137–50.

29. Bosch J, Brossmer R, Kemmner W, Schlag P. Preparation and characterization of differently aggregated colorectal carcinoma cell subpopulations from surgical specimens. Cancer Detect Prev 1998;22:319 –29.

47. Dawson B, Trapp RC. Basic & clinical biostatistics. 3rd ed. New York: McGraw–Hill; 2001. p. 141–2.

30. Ogawa JI, Inoue H, Koide S. ␣-2,3-sialyltransferase type 3N and␣-1,3focosyltransferase type VII are related to sialyl Lewis synthesis and patient survival from lung carcinoma. Cancer 1997;79:1678 – 85. 31. Ogawa J, Tsurumi T, Yamada S, Koide S, Shohtsu A. Blood vessel invasion and expression of sialyl Lewis and proliferation cell nuclear antigen in stage I non-small cell lung cancer. Cancer 1993;73:1177– 843. 32. Pousset D, Piller V, Bureaud N, Monsigny M, Piller F. Increased ␣-2,6 sialylation of N-Glycans in a transgenic mouse model of hepatocellular carcinoma. Cancer Res 1997;57:4249 –56. 33. Liu F, Qi HL, Chen HL. Regulation of differentiation and proliferationinducer on lewis antigens, ␣-fucosyltransferase and metatstic potential in hepatoccarcinoma cells. Br J Cancer 2001;84:1556 – 63.

48. Wolf JK, Ramirez PT. The molecular biology of cervical cancer. Cancer Invest 2001;19:621–9. 49. Schindl M, Oberhuber G, Obermair A, Schoppmann SF, Karner B, Birner P. Overexpression of Id-1 protein is a marker for unfavorable prognosis in early-stage cervical cancer. Cancer Res 2001;61:5703– 6. 50. Wei LH, Kuo ML, Chen CA, Cheng WF, Cheng SP, Hsieh FJ, Hsieh CY. Interleukin-6 in cervical cancer: the relationship with vascular endothelial growth factor. Gynecol Oncol 2001;82:49 –56. 51. Ueda M, Terai Y, Kumagai K, Ueki K, Yamaguchi H, Akise D, Ueki M. Vascular endothelial growth factor C gene expression is closely related to invasion phenotype in gynecological tumor cells. Gynecol Oncol 2001; 82:162– 6.

34. Petretti T, Schulze B, Schlag PM, Kemmner W. Altered mRNA expression of glycosyltransferases in human gastric carcinomas. Biochim Biophys Acta 1999;1428:209 –18.

52. Tjalma WA, Weyler JJ, Bogers JJ, Pollefliet C, Baay M, Goovaerts GC, Vermorken JB, van Dam PA, van Marck EA, Buytaert PM. The importance of biological factors (bcl-2, bax, p53, PCNA, MI, HPV and angiogenesis) in invasive cervical cancer. Euro J Obstet Gynecol Reprod Biol 2001;97:223–30.

35. Bergler W, Riedel F, Schwartz-Albiez R, Gross HJ, Hormann K. A new histobiochemical method to analyze sialylation on cell-surface glycopro-

53. Taniguchi A, Kaneta R, Morishita K, Matsumoto K. Gene structure and transcriptional regulation of human Gal␤1,4(3)GlcNAc ␣2,3-sialyltrans-

52

54.

55.

56.

57. 58.

WANG ET AL. ferase VI (hST3Gal VI) gene in prostate cancer cell line. Biochim Biophys Res Commun 2001;287:1148 –56. Kim YJ, Kim KS, Kim HS, Kim CH, Ko JH, Choe IS, Tsuji S, Lee YC. Molecular cloning and expression of human Gal␤1,3GalNAc␣2,3-sialyltransferase (hST3Gal II). Biochem Biophys Res Commun 1996;228: 324 –7. Lee YC, Kkjima N, Wada E, Kurosawa N, Nakaoka T, Hamamoto T, Tsuji S. Cloning and expression of cDNA for a new type of Gal␤1,3GalNAc␣2,3-sialyltransferase. J Biol Chem 1994;269:10028 –33. Ji MY, Lee YC, Kim KS, Cho JW, Jung KY, Kim CH, Choo YK. Developmental patterns of Gal␤1,3(4)GlcNAc ␣2,3-sialyltransferase (ST3Gal III) expression in the mouse: in situ hybridization using DIGlabeled RNA probes. Arch Pharmacal Res 1999;22:243– 8. Kansas GS. Selectins and their ligands: current concepts and controversies. Blood 1996;88:3259 – 87. Takada A, Ohmori K, Yoneda T, Tsuyuoka K, Hasegawa K, Kiso M,

Kannagi R. Contribution of carbohydrate antigens sialyl Lewis A and sialyl Lewis X to adhesion of human cancer cells to vascular endothelium. Cancer Res 1993;53:354 – 61. 59. Azuma Y, Taniguchi A, Matsumoto K. Decreased in cell surface sialic acid in etoposide-treated Jurkat cells and the role of cell surface sialidase. Glycoconj J 2000;17:301– 6. 60. Lo NW, Dennis JW, Lau JT. Overexpression of the ␣2,6-sialyltransferase, ST6Gal I, in a low metastatic variant of a murine lymphoblastoid cell line is associated with appearance of a unique ST6Gal I mRNA. Biochem Biophys Res Commum 1999;264:619 –21. 61. Hui N, Le Marer N. ␣2,6-sialylation regulation in cd34⫹ progenitor cells in the human bone marrow and granulocyte colony-stimulating factor mobilization. J Hematother Stem Cell Res 2001;10:661– 8. 62. Elpek GO, Gelen T, Karpuzoglu G, Karpusoglu T, Keles N. Clinicopathologic evaluation of CDw75 antigen expression in patients with gastric carcinoma. J Pathol 2001;193:169 –74.