Altered mRNA Expression of Sialyltransferase in Squamous Cell Carcinomas of the Cervix

Gynecologic Oncology 83, 121–127 (2001) doi:10.1006/gyno.2001.6358, available online at http://www.idealibrary.com on

Altered mRNA Expression of Sialyltransferase in Squamous Cell Carcinomas of the Cervix 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.,* Ying-Chieh Tsai, Ph.D.,† ,3 and Chiou-Chung Yuan, M.D.* ,2,3 *Department of Obstetrics and Gynecology, Veterans General Hospital-Taipei and Institute of Clinical Medicine, National Yang-Ming University; and †Department of Biochemistry, National Yang-Ming University, Taipei, Taiwan Received November 21, 2000

INTRODUCTION

Objective. Increased sialylation has been reported in various kinds of cancers, but to date, sialylation of cervical carcinoma has never been evaluated. This study of the changes in messenger ribonucleic acid (mRNA) expression of the four sialyltransferases (ST3Gal I, ST3Gal III, ST3Gal IV, and ST6Gal I) in a normal cervix and that with FIGO stage IB1 squamous cell carcinoma was undertaken to assess the extent of sialylation associated with establishment of the carcinoma. Methods. Alterations in ST mRNA expression in FIGO IB1 cervical cancer (n ⴝ 30) and normal cervixes (n ⴝ 30) were examined by semiquantitative reverse transcription–polymerase chain reaction (RT-PCR). Results. ST6Gal I expression was enhanced in squamous cell carcinoma of the cervix (P ⴝ 0.026, Mann–Whitney U test), but mRNA expression from the other three STs (ST3Gal I, ST3Gal III, and ST3Gal IV) was significantly down-expressed in squamous cell carcinoma of the cervix compared to the normal cervix (P ⴝ 0.003, P < 0.001, and P ⴝ 0.001, respectively). High ST6Gal I expression was associated with more invasive properties of cervical cancer, such as deep stromal invasion, lymph or vascular space involvement, and poor differentiation (P ⴝ 0.010, P < 0.001, P < 0.001, respectively). Conclusions. A combination of enhanced ST6Gal I mRNA expression and decreased mRNA expression from ST3Gal I, ST3Gal III, and ST3Gal IV might be important in cervical cancer. Future studies will investigate whether RT-PCR detection of the expression of these enzymes can be helpful for prognostic purposes. © 2001 Academic Press Key Words: reverse transcription–polymerase chain reaction (RT-PCR); sialyltransferase; squamous cell carcinoma of the cervix.

Sialic acids and their derivatives are ubiquitous at the terminal positions of oligosaccharides of glycoproteins that play roles in a variety of biological processes, such as cell– cell communication, cell–matrix interaction, adhesion, and protein targeting [1, 2]. The transfer of the sialic acids from guanosine monophosphate–sialic acid to an acceptor carbohydrate is catalyzed by sialyltransferases (STs) [2]. Recent experimental evidence indicates that the activity of STs is closely related to tumor formation and metastasis [3–5]. To our knowledge, no investigation into the expression of STs in human cervical carcinoma has been carried out. In this study, we assess the differences between these expressions in the normal cervix and that with squamous cell carcinoma. MATERIALS AND METHODS

1 This work was supported in part by grant 89VGH-225 from the Taipei Veterans General Hospital and grants (NSC-89-2314-B-075-047 and NSC-892314-B-075-088) from the National Science Council, Taiwan, ROC. 2 These authors contributed equally. 3 To whom correspondence should be addressed at Veterans General Hospital, Taipei, 201, Section 2, Shih-Pai Road, Taipei 112, Taiwan. Fax: ⫹8862-28734101. E-mail: [email protected].

This prospective study involved 34 selected patients undergoing radical hysterectomy and pelvic lymph node dissection for FIGO stage IB1 squamous cell carcinoma of the cervix and 30 selected patients undergoing a total hysterectomy for uterine myoma in the Veterans General Hospital–Taipei (the National Medical Center of Taiwan) between January and December 1999. Excluding four patients with lymph node metastasis (n ⫽ 3), parametrial invasion (also the presence of lymph node metastasis), vaginal invasion, or uterine body invasion (n ⫽ 1), no patient (n ⫽ 30) had any evidence of extracervical metastases, and this was proved by pathological examination. Patient consent for collection of cervical tissues was obtained from all women according to the guidelines of the Human Ethics Committee in the Department of Obstetrics and Gynecology, Veterans General Hospital–Taipei. All specimens were collected and processed within 20 min of surgery. Sections of the tumors and normal cervix were cut under pathological guidance and rapidly frozen in liquid nitrogen. To verify that the sample consisted of carcinoma tissue, cryostat sections (at 5 ␮m) were first examined by a pathologist. Whole

121

0090-8258/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

122

WANG ET AL.

tissue pieces containing necrotic areas, fatty tissue, or a large amount of normal epithelium or stroma tissue were discarded. Tissue was snap-frozen in liquid nitrogen until RNA extraction. In total, 30 samples of squamous cell carcinoma of the cervix and 30 cases of normal cervix were studied. Tissues were powdered in liquid nitrogen. RNA was extracted using StrataPrep Total RNA Miniprep Kit (Stratagene, La Jolla, CA). Following the supplier’s recommendations, an appropriate amount of lysis buffer containing 0.1 M ß-mercaptoethanol was added to the tissues. The tube was vortexed to homogenize the lysate. Following cell lysis, the sample was prefiltered in a spin cup to remove particulates and to reduce the amount of DNA. The filtrate was then transferred to a spin cup with a silica-based fiber matrix where the RNA could bind to the fiber matrix. Treatment with low-salt wash buffer and digestion with DNase removed the remaining DNA. Serial washes removed DNase and other proteins. Highly pure RNA was eluted from the fiber matrix with a small volume of elution buffer (10 mM Tris-HCl, pH 7.5, and 1 mM ethylenediaminetetraacetic acid) and captured in a microcentrifuge tube. To verify if the total RNA extracted by StrataPrep was free of DNA, at least one the following procedures was tested for each sample. First, glyceraldehyde phosphate dehydrogenase (GAPDH) (L1/L2) polymerase chain reaction (PCR) primers encompass introns in the gene; therefore, ⬃1 kb (exons only) versus ⬃1.7 kb (exons plus introns) can identify the presence of genomic DNA. Second, another pair of GAPDH primers resulted in 371 bp versus 563 bp, serving the same purpose (data not shown). Third, RNA from the corresponding samples was used in the PCR as control. Any of the above-mentioned experiments identified that the DNA was undetectable in the RNA samples reported in this study. 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 [6]. GAPDH serves as an internal control for gene expression, as documented in several reports [6 – 8]. The gene is also well established in our laboratory for other gene expression experiments. 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, then it was incubated 50 min at 42°C. The reaction was inactivated by heating at 70°C for 15 min. Reverse transcription (RT) into cDNA was achieved by using 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 [6], respec-

tively, and the GAPDH-specific primers (5⬘-TTTGGTCGTATTGGGCGCCTGGTCA-3⬘ and 3⬘-TGGTGGTCGGGGTCGTTCTCGTGTT-5⬘). The GAPDH primers encompass introns in the genomic gene; therefore, 1 kb (exons only) versus 1.7 kb (plus introns) was able to identify whether the genomic DNA was present in the RNA preparation. Meanwhile, the extracted RNA from the individual sample was subjected to the following PCR reaction as a negative control. The condition used for the PCR was set up based on the titration experiment conducted, first by various amounts of cDNA and then second by various numbers of PCR cycles. The amount of cDNA obtained from human hepatoblastoma cells (HepG2) (as the positive control) was first fivefold diluted and volumes of 1.0, 2.5, 5.0, and 10.0 ␮l were used to conduct the PCR at 30 cycles for ST3Gal I, ST3Gal IV, and ST6Gal I (group A) and 40 cycles for ST3Gal III. A volume of 2.5 ␮l was then used to conduct PCR with cycle numbers of 26, 28, 30, 32, and 34 for primers in group A. The same amount of cDNA was used to conduct PCR with cycle numbers of 36, 38, 40, 42, and 44 for ST3Gal III. Reactions with PCR cycle number 30 for group A primers and 40 for ST3Gal III were in the linear range and thus were chosen as the assay conditions. Due to fewer ST transcripts expressed by the tissue samples, 1 ␮l of the undiluted cDNA was used for the PCR. The PCR mixture consisted of 0.1 U Super-Therm DNA polymerase (JMR Holdings), 1⫻ 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 extra-distilled water added to 50 ␮l. The volumes of ST-specific primers (10 ␮M) used were 1.6 ␮l for ST3Gal I and ST3Gal III, 1.2 ␮l for ST3Gal IV, and 0.8 ␮l for ST6Gal I. The volumes of GAPDH primers (10 ␮M) used were 1.6 ␮l for ST3Gal I and ST3Gal III and 0.8 ␮l for ST3Gal IV and ST6Gal I. For PCR of ST3Gal I, ST3Gal IV, and ST6Gal I, the reaction was conducted at 1 min at 94°C, 50 s at 65°C, and 2 min 40 s at 72°C for five cycles and then 50 s at 94°C, 50 s at 67°C, and 2 min 40 s at 72°C for 25 cycles, followed by 10 min at 72°C and 10 min at 4°C. PCR of ST3Gal III was conducted at 2 min at 94°C, 1 min at 94°C, 1 min at 65°C, and 2 min at 72°C for 40 cycles followed by 2 min at 72°C and 10 min at 4°C. All PCR experiments were conducted in the GeneAmp PCR system 2400 (Applied Biosystems, Foster City, CA). Reaction products obtained were electrophoresed in 2% agarose-containing ethidium bromide. Density measurements were made using an Alpha-Imager 2000. The density of each ST band was compared with that of GAPDH which was coamplified within the same tube, and the ratio (density units of ST band/density units of GAPDH) was calculated. The detection limit was 1 ng of double-stranded DNA. The linear portion of the assay ranged up to 25 ng. ST messenger RNA (mRNA) expression of each sample was determined in at least two independent experiments. Using GAPDH as an internal standard, the deviation between duplicate measurements was on average 18%. In all experiments, a negative control reaction was per-

SIALYLTRANSFERASE AND CERVICAL CARCINOMA

formed by replacing the cDNA template with sterile water, and positive controls were performed with HepG2 cDNA (kindly provided by Dr. Chin-Wen Chi, Laboratory of Biochemical Oncology, Department of Medical Research, Taipei Veterans General Hospital, Taipei, Taiwan), a cell line that expresses the four STs [7, 8]. Statistical analysis of data was done using the statistical software SAS, version 6.12 (SAS Institute, Cary, NC) [9] and SPSS, version 10.0 (SPSS Inc., Chicago, IL). Both a paired t test and Mann–Whitney test were used to test differences between the expression of a given ST in the carcinoma and in normal tissue of the cervix. Pathological parameters including deep stromal invasion (equal to or more than half of the cervix in thickness), lymph or vascular space involvement, and differentiation based on our previous report [10 –12] were also tested using Student’s t test for two groups and one-way analysis of variance (ANOVA) for three groups. Probability values below 5% were considered statistically significant. RESULTS Four ST mRNA expressions in squamous cell carcinoma of the cervix (n ⫽ 30) and in normal cervixes (n ⫽ 30) 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 cDNA for GAPDH. The standard deviation between duplicate measurements was 18%. The PCR condition was set based on the titration experiment conducted, first by various amounts of cDNA and then various numbers of reaction cycles (data not shown). The final condition for PCR, which was determined to enable the amplification of both GAPDH and ST, occurred in a somewhat linear manner. Therefore, a semiquantitative analysis of mRNA expression

123

FIG. 2. Semiquantitative RT-PCR analysis of four sialyltransferases in FIGO IB squamous cell carcinoma of the cervix.

was feasible. Figures 1 and 2 show examples of RT-PCR results of normal cervix and cervical squamous cell carcinoma, respectively. Carcinomas showed increased ST6Gal I expression compared to normal cervixes (P ⬍ 0.05, Mann–Whitney U test) but mRNA expressions from the other three STs (ST3Gal I, ST3Gal III, and ST3Gal IV) were significantly down-expressed in squamous cell carcinoma of the cervix compared to the normal cervix, with P ⫽ 0.003, P ⬍ 0.001, and P ⫽ 0.001, respectively (Fig. 3). A paired t test also showed significant differences of mRNA expression of four STs between the normal cervix tissue and carcinoma tissue (ST3Gal I, P ⬍ 0.001; ST3Gal III, P ⬍ 0.001; ST3Gal IV, P ⫽ 0.003; ST6Gal I, P ⫽ 0.010). Although altered expression of ST6Gal I in cervical carcinoma was not so impressively significant, high ST6Gal I expression was associated with more invasive properties of cervical cancer, such as deep stromal invasion and lymph or vascular space involvement (Table 1). In addition, ST6Gal I expression showed a positive correlation with worse cancer grading (P ⬍ 0.001, one-way ANOVA). Figures 4 and 5 show the relationship of mRNA expression of four STs in normal cervical tissue and cervical carcinoma tissue, respectively. We found that each tissue shows dissociation of ST3Gal III mRNA expression compared to the other three variables in normal control subjects but this dissociation of ST3Gal III was not significant in cervical cancer patients. Figure 6 shows that this phenomenon might result from wide distribution of mRNA expression of ST3Gal III, especially in normal control subjects. DISCUSSION

FIG. 1. Semiquantitative RT-PCR analysis of four sialyltransferases in a normal cervix.

Sialic acids include of a number of derivatives of the ninecarbon acid amino sugar neuraminic acid. The amino group of neuraminic acid may be substituted with an acetyl or glycoloyl

124

WANG ET AL.

FIG. 3. Messenger RNA expression of the four sialyltransferases in normal cervix tissue and cervical carcinoma. 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.

moiety, while the hydroxyl groups can be methylated from the ester with acetyl, lactyl, sulfate, or phosphate groups to form over 20 naturally occurring derivatives [13]. The sialic acids are widely distributed in nature as terminal sugars on oligosaccharides attached to protein or lipid moieties. Due to their acidic nature, they impart a net negative charge to the cell surface and are important in cell– cell or cell–matrix interactions. For example, cancer cells have the ability to mask specific cellular recognition sites [14]. There is a large body of evidence suggesting that tumor cells have changed surface properties from their normal counterparts and that these changes are partially due to altered sialo-glycoconjugates expressed on the plasma membrane [14]. These altered surfaces

affect the behavior of the cells and change their ability to invade and metastasize [13]. Modifications of cellular glycosylation are a common phenotypic change in malignancy. However, only a limited number of biosynthetic pathways are frequently altered in cancers [7]. Increased ␤1,6-branching, increased sialyl-Lewis epitopes, or the general increase in sialylation of cell surface glycoproteins are commonly observed in N-lined and O-lined oligosaccharides of carcinoma cells [3]. These changes in glycosylation are reported to play important roles in tumor grade, invasion, metastatic ability, and poor clinical outcome [4]. Carbohydrate changes also occur in breast cancer [7, 15–17], colorectal cancer [1, 18 –21], lung cancer [22, 23], and hepatic carcinoma [24].

TABLE 1 The Relationship between Different Sialytransferases and Pathological Factors in Cervical Carcinoma (n ⴝ 30 Patients) a Variable DSI ⬍1⁄2 (n ⫽ 19) ⱖ1⁄2 (n ⫽ 11) P value LVSI No (n ⫽ 18) Yes (n ⫽ 12) P value a

ST3Gal I

ST3Gal III

ST3Gal IV

ST6Gal I

0.610 ⫾ 0.384 0.810 ⫾ 0.496 0.265

1.502 ⫾ 1.095 2.105 ⫾ 1.599 0.284

0.332 ⫾ 0.097 0.398 ⫾ 0.229 0.379

1.047 ⫾ 0.526 1.700 ⫾ 0.633 0.010

0.647 ⫾ 0.466 0.738 ⫾ 0.387 0.566

1.348 ⫾ 0.946 2.285 ⫾ 1.599 0.086

0.320 ⫾ 0.103 0.412 ⫾ 0.210 0.180

0.849 ⫾ 0.289 1.944 ⫾ 0.418 0.001

DSI, deep stroma invasion; LVSI, lymph or vascular space involvement. Data are presented as mean ⫾ SD unless otherwise noted.

SIALYLTRANSFERASE AND CERVICAL CARCINOMA

FIG. 4. Distribution of values of ST/GAPDH among normal control subjects. Relative dissociation of ST3Gal III to the other three variables is evident.

Recchi et al. [7] evaluated five types of ST expression in human breast cancer. In addition, the levels of ␣2,3-sialyltransferase ST3Gal I, ST3Gal II, ST3Gal III, and ␣2,6-sialyltransferase ST6Gal I messages were found to be significantly changed in cancer tissues [1, 18, 20]. Due to their promising reports, we pioneered the study of the role of ST expression in cervical carcinoma because cervical cancer is the most prevalent female cancer in Taiwan. Therefore, for the cervical cancer studied here, we focused on these several genes first and will expand to more detail in the future. Initially, we evaluated five types of STs, including ST3Gal I, II, III, and IV and ST6Gal I, because we used HepG2 cDNA, a cell line that expresses the five STs, as a positive control. However, when ST3Gal II

FIG. 5. Distribution of values of ST/GAPDH among cervical cancer patients. Cases show relatively high values of ST3Gal III but these are less evident than those of normal controls.

125

FIG. 6. Bar graphs show four types of ST/GAPDH. The ratios of ST3Gal III/GAPDH in normal control subjects are in a wide distribution but those in cervical cancer patients are in a relatively limited distribution.

RT-PCR was conducted for the tissues collected, unlike other genes, the resulting band intensities of ST3Gal II were mostly weak and fluctuated among the reactions even within the normal control. Therefore, it was excluded in this study. Among these four STs, we found that only ST6Gal I mRNA statistically increased in squamous cell carcinoma of the cervix compared with the normal cervix. Although our data about ST6Gal I showed less prominent significance than the other three STs, high mRNA expression of ST6Gal I was associated with more invasive properties of cancers, such as deep stromal invasion, lymph or vascular space involvement, and poor differentiation. ST6Gal I is responsible for the addition of sialic acid in the ␣2,6-linkage to Gal␤1,4GlcNAc (N-acetyl-lactosamine), a sequence commonly found in N-linked chains of glycoproteins. In fact, much evidence supports enhanced ST6Gal I expression’s possibly being important in cancer development and progression [7, 21, 24 –26], although there are some conflicting data [27]. In breast cancer, high ST6Gal I expression was associated with histoprognostic grade III and was negatively correlated with expression of progesterone receptor [7]. Dall’Olio et al. [21] found that increasing expression of ST6Gal I was positively correlated with an invasive phenotype of colon cancer. This increase of metastatic potential was explained by the study of Bosch et al. [25], in which they found that ST6Gal I is responsible for reduced homotypic aggregation of colorectal carcinoma cells and may thus facilitate the release of single cells from the primary tumor. In a hepatic cancer model, enhanced expression of ST6Gal I was also found during tumor development and progression [24]. In squamous cell carcinomas of the oropharynx and oral cavity, Bergler et al. [26] used an indirect method to imply that

126

WANG ET AL.

increasing ST6Gal I may contribute to cervical lymph node metastases. Kimura et al. [27] found that alteration of glycoconjugates by ST6Gal I in granulosa cells during atresia is involved in some of the processes of ovarian follicular atresia and granulosa cell apoptosis. Enhanced ST6Gal I enzyme activity was found in the choriocarcinoma cell line, but ST3Gal III and ST3Gal IV activities did not change as a result of tumorigenesis [28]. In a brain tumor, ST6Gal I expression was increased in meningiomas, chondromas, and craniopharyngiomas but showed no difference in any tumors of glial origin or in medulloblastoma [29]. In a murine lymphoblastoid cell line model, the data were controversial because Lo et al. [30] found overexpression of ST6Gal I in a low-metastatic variant. Transfer of sialic acid in the ␣2,3-linkage to Gal␤1,(3) 4GlcNAc on N-linked chains of glycoproteins is carried out by ␣2,3-sialyltransferase ST3Gal III or ST3Gal IV. The latter is also capable of adding sialic acid to Gal␤1,3GalNAc found on Olinked chains of glycoproteins. The role of ST3Gal III or ST3Gal IV was also discussed before [22, 31]. Cancer cells have to adhere to vascular endothelial cells and then extravasate from the bloodstream into the surrounding tissues in order to metastasize hematogenously [22]. Altered glycosylation of malignant cell-surface lipids and proteins plays an important role in tumor cell adherence to endothelial cells by interaction mediated by E selectin [22, 32, 33]. ST3Gal III is involved in the biosynthesis of sialyl-dimeric Lewis sLe x and sLe a, which are known as selectin ligands and tumor-associated carbohydrate structures [31]. Although our study demonstrated low mRNA expression of ST3Gal III in the cervical carcinoma tissue compared to those in the normal cervix tissue, we found that high mRNA expression of ST6Gal I was positively correlated to high ST3Gal III expression either on normal cervix tissue or on cervical carcinoma tissue. In addition, cancers with lymph or vascular space involvement seemed to present a trend of increasing mRNA expression of ST3Gal III (P ⫽ 0.086), which might be used as a predictor in patients with a high risk of distant metastases in the future, especially hematogenous spread to lung, bone, or liver in cervical cancer patients. In fact, overexpression of ST3Gal III was correlated with a poor prognosis in breast cancer [7], and increased expression of sLex correlates with recurrence [23]. However, due to studying a small number of patients and having a very short follow-up period, we could not observe this phenomenon. Expression of ST3Gal IV was significantly enhanced in gastric carcinoma tissue compared to normal tissue [34]. Increased expression of ST3Gal III or ST3Gal IV may play a role in glial tumorigenesis [35]. In our study, we failed to demonstrate the over-expression of ST3Gal III mRNA and ST3Gal IV mRNA in squamous cell carcinoma of the cervix. On the contrary, down-regulation of both STs was found in squamous cell carcinoma of the cervix. ST3Gal I’s being responsible for the ␣2,3-sialylation of Gal␤1,3GalNAc on O-linked chains of glycoproteins and glycolipids was proposed to have a decisive role in sialyl-Le x/Le a biosynthesis [36] and sialylation of the Thomsen-Friedenreich

antigen [37]. Enhanced ST3Gal I expression was found in lung cancers [22, 23], in breast cancer [16, 17, 38], and in colon cancer [18]. In our study, we found down-regulation of ST3Gal I expression in squamous cell carcinoma of the cervix. Changes in ST expression have been observed in cancer tissues and cells, and the regulation of their expressions is achieved mainly at the transcriptional level. For example, transfection of rate fibroblasts with the ras oncogene leads to an increase of ␤-galactoside␣2,6 sialyltransferase mRNA and of the invasion potential of these cells [39]. Based on this study, up-regulation of ST6Gal I and downregulation of ST3Gal III, IV, and I were found in cervical cancer. High mRNA expression of ST6Gal I was associated with more invasive properties of cancers, such as deep stromal invasion, lymph or vascular space involvement, and poor differentiation. Although high expression of ST6Gal I was associated with high expression of ST3Gal III but was reversed with the expression of ST3Gal I and ST3Gal IV, their roles were not clear. Because the final sialyl– glycan structure is determined by concerted action of all expressed STs and so far, at least 18 distinct ST genes exist, the remaining 14 STs might need further evaluation. For example, ST3Gal II shares almost identical specificity with ST3Gal I; moreover, and ST3Gal VI shares acceptor specificity with ST3Gal III and ST3Gal IV; in addition, ST3Gal VI competes with ST6Gal [43], but we did not examine expression of these candidate STs. Future studies will expand the remaining 14 STs and investigate whether RT-PCR detection of the expression of these enzymes can be helpful for prognostic purposes. ACKNOWLEDGMENTS The authors appreciate Miss Shu-Ching Liang and Miss Wen-Yuann Shyong for their help in sample collection.

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. Fukuda M. Possible roles of tumor-associated carbohydrate antigens. Cancer Res 1996;56:2237– 44. 4. Hakomori S. Tumor malignancy defined by aberrant glycosylation and shingo-(glyco)-lipid metabolism. Cancer Res 1996;56:5309 –18. 5. Aubert M, Panicot L, Crotte C, Gibier P, Lombardo D, Sadoulet MO, Mas E. Restoration of alpha(1,2) fucosyltransferase activity decreases adhesive and metastatic properties of human pancreatic cancer cells. Cancer Res 2000;60:1449 –56. 6. 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. 7. Recchi MA, Hebbar M, Hornez L, Harduin-Lepers A, Peyrat JP, Delannoy P. Multiplex reverse transcription polymerase chain reaction assessment of

SIALYLTRANSFERASE AND CERVICAL CARCINOMA

8.

9. 10.

11.

12.

13. 14. 15.

16. 17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

sialyltransferase expression in human breast cancer. Cancer Res 1998;58: 4066 –70. 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. SAS Institute. SAS/STAT user’s guide, release 6.12. Cary, NC: SAS Institute; 1996. 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. 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. 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. Thomas P. Cell surface sialic acid as a mediator of metastatic potential in colorectal cancer. Cancer J 1996;9:1–10. Schauer R. Sialic acids and their role as biological masks. Trends Biochem Sci 1985;10:357– 60. 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. Soares R, Marinho A, Schmitt F. Expression of sialyl-Tn in breast cancer. Correlation with prognostic parameters. Pathol Res Pract 1996;192:1181– 6. Burchell J, Poulsom R, Hanby A, Whitehouse C, Cooper L, Clausen H, Miles D, Taylor-Papadimitriou J. An alpha2,3 sialyltransferase (ST3Gal I) is elevated in primary breast carcinomas. Glycobiology 999;9:1307–11. 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. Yamada N, Chung YS, Takatsuka S, Arimoto Y, Sawada T, Dohi T, Sowa M. Increased 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. Kudo T, Ikehara Y, Togayachi A, Morozumi 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. Dall’Olio FD, Chiricolo M, Lau JTY. Differential expression of the hepatic transcript of beta-galactoside alpha 2,6 sialyltransferase in human colon cancer cell lines. Int J Cancer 1999;81:243–7. 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. Ogawa J, Tsurumi T, Yamada S, Koide S, Shohtsu A. Blood vessel invasion and expression of sialyl Lewisx and proliferation cell nuclear antigen in stage I non-small cell lung cancer. Cancer 1994;91:1177– 83. Pousset D, Piller V, Bureaud N, Monsigny M, Piller F. Increased alpha-2,6 sialylation of N-glycans in a transgenic mouse model of hepatocellular carcinoma. Cancer Res 1997;57:4249 –56. 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. Bergler W, Riedel F, Schwartz-Albiez R, Gross HJ, Hormann K. A new histobiochemical method to analyze sialylation on cell-surface glycoproteins of head and neck squamous-cell carcinomas. Eur Arch Otorhinolaryngol 1997;254:437– 41.

127

27. Kimura Y, Manabe N, Nishihara S, Matsushita H, Tajima C, Wada S, Miyamoto H. Up-regulation of the alpha2,6-sialyltransferase messenger ribonucleic acid increases glycoconjugates containing alpha2, 6-linked sialic acid residues in granulosa cells during follicular atresia of porcine ovaries. Biol Reprod 1999;60:1475– 82. 28. Fukushima K, Hara-Kuge S, Seko A, Ikehara Y, Yamashita K. Elevation of alpha236 sialyltransferase and alpha132 fucosyltransferase activities in human choriocarcinoma. Cancer Res 1998;58:4301– 6. 29. Kaneko Y, Yamamoto H, Kersey DS, Colley KJ, Leestma JE, Moskal JR. The expression of Gal beta 1,4GlcNAc alpha 2,6 sialyltransferase and alpha 2,6-linked sialoglycoconjugates in human brain tumors. Acta Neuropathol 1996;91:284 –92. 30. Lo NW, Dennis JW, Lau JT. Overexpression of the alpha2,6-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 Commun 1999;264:619 –21. 31. Ji MY, Lee YC, Kim KS, Cho JW, Jung KY, Kim CH, Choo YK. Developmental patterns of GalBeta1,3(4)GlcNAc alpha2,3-sialyltransferase (ST3Gal III) expression in the mouse: in situ hybridization using DIG-labeled RNA probes. Arch Pharm Res 1999;22:243– 8. 32. Phillips ML, Nudelman E, Gaeta FC, Perez M, Singhal AK, Hakomori S, Paulson JC. ELAM-1 mediates cell adhesion by recognition of a carbohydrate ligand, sialyl-Le x. Science 1990;250:1130 –2. 33. Walz G, Aruffo A, Kolanus W, Bevilacqua M, Seed B. Recognition by ELAM-1 of the sialyl-Le x determinant on myeloid and tumor cells. Science 1990;250:1132–5. 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. 35. Yamamoto H, Saito T, Kaneko Y, Kersey D, Yong VW, Bremer EG, Mkrdichian E, Cerullo L, Leestma J, Moskal JR. alpha2,3-sialyltransferase mRNA and alpha2,3-linked glycoprotein sialylation are increased in malignant gliomas. Brain Res 1997;755:175–9. 36. Kannagi R. Carbohydrate-mediated cell adhesion involved in hematogenous metastasis of cancer. Glycoconj J 1997;14:577– 84. 37. Whitehouse C, Burchell J, Gschmeissner S, Brockhausen I, Lloyd KO, Taylor-Papadimitriou J. A transfected sialyltransferase that is elevated in breast cancer and localizes to the medial/trans-Golgi apparatus inhibits the development of core-2-based O-glycans. J Cell Biol 1997;137:1229 – 41. 38. 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 sialyltransferases which synthesize sialyl Lewisx, the carbohydrate ligands for E-selectin, in human breast cancer. Int J Oncol 1998;12:1157– 64. 39. Le Marer N, Laudet V, Svensson EC, Cazlaris H, Van Hille B, Lagrou C, Stehelin D, Montreuil J, Verbert A, Delannoy P. The c-Ha-ras oncogen induces increased expression of beta-galactoside: alpha 2,6 sialyltransferase in rate fibroblast (FR3T3) cells. Glycobiology 1992;2:49 –56. 40. 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 (hST3Gal II). Biochem Biophys Res Commun 1996;228:324 –7. 41. 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. 42. Lee YC, Kojima 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. 43. Okajima T, Fukumoto S, Miyazaki H, Ishida H, Kiso M, Furukawa K, Urano T, Furukawa K. 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.