ST3Gal.I sialyltransferase relevance in bladder cancer tissues and cell lines
- First Online:
- Cite this article as:
- Videira, P.A., Correia, M., Malagolini, N. et al. BMC Cancer (2009) 9: 357. doi:10.1186/1471-2407-9-357
The T antigen is a tumor-associated structure whose sialylated form (the sialyl-T antigen) involves the altered expression of sialyltransferases and has been related with worse prognosis. Since little or no information is available on this subject, we investigated the regulation of the sialyltransferases, able to sialylate the T antigen, in bladder cancer progression.
Matched samples of urothelium and tumor tissue, and four bladder cancer cell lines were screened for: ST3Gal.I, ST3Gal.II and ST3Gal.IV mRNA level by real-time PCR. Sialyl-T antigen was detected by dot blot and flow cytometry using peanut lectin. Sialyltransferase activity was measured against the T antigen in the cell lines.
In nonmuscle-invasive bladder cancers, ST3Gal.I mRNA levels were significantly higher than corresponding urothelium (p < 0.001) and this increase was twice more pronounced in cancers with tendency for recurrence. In muscle-invasive cancers and matching urothelium, ST3Gal.I mRNA levels were as elevated as nonmuscle-invasive cancers. Both non-malignant bladder tumors and corresponding urothelium showed ST3Gal.I mRNA levels lower than all the other specimen groups. A good correlation was observed in bladder cancer cell lines between the ST3Gal.I mRNA level, the ST activity (r = 0.99; p = 0.001) and sialyl-T antigen expression, demonstrating that sialylation of T antigen is attributable to ST3Gal.I. The expression of sialyl-T antigens was found in patients' bladder tumors and urothelium, although without a marked relationship with mRNA level. The two ST3Gal.I transcript variants were also equally expressed, independently of cell phenotype or malignancy.
ST3Gal.I plays the major role in the sialylation of the T antigen in bladder cancer. The overexpression of ST3Gal.I seems to be part of the initial oncogenic transformation of bladder and can be considered when predicting cancer progression and recurrence.
List of abbreviations used
polymerase chain reaction
Reverse transcriptase-polymerase chain reaction.
Bladder cancer is one of the most common cancers in humans and its incidence has been increasing during the past years . 70-80% of all bladder cancers are nonmuscle-invasive and have a low mortality rate. However, despite the complete resection of the primary lesion, and the success of treatments such as the intravesical instillation with bacillus Calmette-Guérin (BCG) , 30-50% of patients with nonmuscle invasive bladder cancer experience recurrence within the first year after resection and 15% of these patients manifest worsening of tumor grade and stage. For those reasons, bladder cancers are becoming a serious public health problem and a social and economic burden.
The oligosaccharide chains of glycoproteins and glycolipids are often decorated by sialic acids, a family of nine carbon sugars derived from neuraminic acid. In humans, sialylation of glycoconjugates is mediated by different sialyltransferase enzymes which, depending on their nature, may establish different types of linkages [linkage through an α2-3- or an α2-6-bond to galactose (Gal); through an α2-6-bond to N-acetylgalactosamine (GalNAc) or N-acetylglucosamine (GlcNAc); or through an α2-8-bond to another sialic acid, forming polysialic acid] (reviewed in ). During neoplastic transformation, the activity of sialyltransferases may be altered and, as a consequence, cancer cells express more heavily sialylated glycans at the surface . This aberrant sialylation may mediate key pathophysiological events during the various steps of tumor progression, including invasion and metastasis formation. This is due to the fact that sialylated structures can prevent cell-cell interactions through non-specific charge repulsion, but they can be specifically bound by cell adhesion molecules, such as selectins . On the other hand, the addition of sialic acids may mask the underlying sugar structure, thus avoiding recognition by other specific glycan binding molecules, such as galectins . Moreover, specific sialylated structures, aberrantly expressed in carcinomas, have been used as targets for cancer immunotherapy in preclinical and clinical studies [7, 8, 9].
The T antigen, or Thomsen-Friedenreich antigen, is a simple glycan, formed by the dissacharide, galactose (Gal) β1-3-linked to N-acetylgalactosamine (GalNAc), O-glycosidicaly-linked to serine or threonine . The addition of sialic acid to this core 1 structure, forming the sialyl-T antigen, inhibits any further chain extension (with the exception of the possible addition of another sialic acid to GalNAc). These antigens are largely increased in some types of cancers (colon, prostate, cervix, ovary, breast) and, according to the above mentioned role of sialylation, the sialylated form (sialyl-T) has been associated with worse prognosis (reviewed in ). According to the known specificities, three sialyltransferases, namely ST3Gal.I, ST3Gal.II and ST3Gal.IV can mediate the sialylation of the T antigen and the increase in the expression of these sialyltransferases has been shown to be one of the major mechanisms responsible for the sialylation of T antigen . There is growing evidence that these molecules can act as good markers in cancer. In fact, in breast cancer, ST3Gal.I has been found increased compared with normal tissue and its expression is related to the grade of the tumor . Altered mRNA expressions of these sialyltransferases were also shown to be of importance in malignant epithelial ovarian cancers  and in colon carcinoma [14, 15].
The relevance of sugar structures related with the T antigen in the control of bladder cell proliferation is indicated by the fact that the antiproliferative factor (AFP), a sialoglycopeptide comprised of the sialyl-T trisaccharide linked to a peptide derived from a membrane receptor , which is secreted by bladder cells of interstial cystitis patients, exerts a strong inhibitory effect on bladder cell proliferation, both in vivo and in vitro. Nevertheless, in spite of the few studies dealing with the expression of T and sialyl-T antigens [17, 18, 19] or other types of sialylated structures [20, 21] in bladder cancer and there are no published data about the expression of related sialyltransferases.
In this study, we attempted to investigate the relevance of the sialyltransferases ST3Gal.I, ST3Gal.II or ST3Gal.IV expression in bladder cancer progression, by comparing their expression in neoplastic tissue, from different grade bladder tumors, with those of normal urothelium. We found that the mRNA levels of ST3Gal.I were significantly higher in malignant bladder tumors. The analysis of four bladder cancer cell lines showed that this sialyltransferase gives the most relevant contribution for the sialylation of T antigens in bladder cancer cells.
Patient and tissue specimens
Stage and grade distribution of bladder cancer in our study patients
Patients groups and subgroupsa
Ib(n = 43)
II (n = 6)
IA (n = 29)/IB (n = 14)
Number of cases per group
14 (48.3%)/4 (28.5%)
15 (51.7%)/10 (71.4%)
9 (31.0%)/4 (28.5%)
5 (17.2%)/1 (7.1%)
7 (24.1%)/7 (50.0%)
3 (10.3%)/2 (14.3%)
After collection, samples were immediately divided: a fraction was immersed in RNAlater® RNA Stabilization Reagent (Sigma Aldrich Inc., St. Louis, MO, USA) and then preserved at -20°C and another fraction was stored at -80°C, until further processing.
RNA isolation and characterization
Samples preserved in RNAlater (20 to 50 mg) were homogenized and the total RNA was isolated using the GenElute Mammalian Total RNA Purification kit and DNAase treatment (Sigma), according to the manufacturer's instructions. RNA concentrations were measured and only samples with A260/A280ratios between 1.9 and 2.1 were considered further. 250 to 500 ng of total RNA (1 μg for cell lines) was reverse transcribed into cDNA, by using the random-primers-based High Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA, USA).
Gene expression measurements
Reverse transcriptase-polymerase chain reaction (RT-PCR) in real time was performed using Taqman probes methodology, mainly as we described [22, 23]. For each primer/probe set, the Assay ID (Applied Biosystems), the detected reference sequences and the location on the gene were the following: ST3Gal.I (Hs00161688_m1; NM_173344.2NM_003033.3; exons 5, 6), ST3Gal.II (Hs00199480_m1; NM_006927.2; exons 1, 2) and ST3Gal.IV (Hs00272170_m1; NM_006278.1; exons 1, 2). The mRNA expression was normalized using the geometric mean of the expression of the endogenous controls, β-actin and GAPDH genes . The relative mRNA levels were calculated by using the adapted formula 2-ΔCt *1000, which infers the number of mRNA molecules of a certain gene per 1000 molecules of the endogenous controls. ΔCt stands for the difference between the cycle threshold of the target gene and that of the endogenous control genes. The efficiency of the amplification reaction for each primer-probe was above 95% (determined by the manufacturer).
The human bladder cancer cell lines, HT1374, MCR, T24 and 5637, were grown in Dulbecco's modified Eagle medium (DMEM) (Sigma), supplemented with 10% foetal calf serum (Sigma), 2 mM L-glutamine and 100 μg/ml penicillin/streptomycin.
Sialyltransferase activity was assayed in whole cell homogenates as we previously described [25, 26], using Galβ1,3GalNAcα1-O-benzyl (Sigma), as acceptor. The enzyme reaction was performed at 37°C for 2 hours and then the products were isolated by hydrophobic chromatography, eluted and counted by liquid scintillation. Controls without exogenous acceptors were run in parallel and their incorporation was subtracted. The activity was measured as the amount of radioactive sialic acid transferred from the donor to the acceptor substrate, per hour and per amount of protein (nmol/h* μg protein).
Cells were resuspended in DMEM medium (5 × 106/ml) and treated with 200 mU/ml neuraminidase from Clostridium perfringens (Roche Diagnostics GmbH, Mannheim, Germany) in serum-free medium at 37°C for 90 min. In the case of dot blots, 25 μg of protein homogenate were treated with 0.8 mU/μl neuraminidase at 37°C for 6 hours. In parallel, identical samples were mock-treated in the same conditions with heat-inactivated neuraminidase (neuraminidase previously submitted 15 min at 100°C).
Analysis of the phenotypic expression of T antigen by flow cytometry
Staining with Arachis hypogaea lectin, a peanut agglutinin (PNA) was used to examine the expression of T antigens in the cell lines. 105 cells, treated or not with neuraminidase, were resuspended in serum free RPMI medium and incubated with 50 μg/ml of Fluorescein isothiocyanate (FITC)-labelled PNA (EYLaboratories, San Mateo, CA), for 15 min, at 18°C in the dark. Cells were then washed and resuspended in FACS Flow supplemented with 0.1% BSA. Cells were then analyzed in a FacsCalibur Flow cytometer using CellQuest software (Becton Dickinson), following standard procedures.
Lectin blot analysis
The expression of T antigens in tissue samples was assessed by PNA lectin dot blot. 12 μg of protein from neuraminidase treated and mock-treated samples were transferred to a Hybond nitrocellulose membrane (GE Healthcare). After blocking, the membranes were stained with PNA-digoxigenin conjugate (1 μg/ml) for 1 h and subsequently incubated with 7.5 U/μl anti-digoxigenin antibody conjugated with horse-radish peroxidase (Roche). After washing, the membranes were incubated with Immobilon™ Western Chemiluminescent Substrate (Millipore) according to manufacturer's instructions and detected with autoradiography film.
Analysis of ST3Gal.Itranscript variants
Comparison between ST3Gal.I mRNA (Reference Sequences: NM_003033.2 and NM_173344.2) and genomic sequences [NCBI http://www.ncbi.nlm.nih.gov and Ensembl http://www.ensembl.org databases] were performed using Basic Local Alignment Search Tool (BLAST) and refined manually. In order to estimate the size of ST3Gal.I transcript and distinguish its two variants in bladder cancer cells, 2 μl of cDNA (obtained as referred above) were amplified by conventional RT-PCR, using forward primers located in exon Y (F1: 5'-CTATGCCAGACAGTTTCGACA-3'), exon ×1 (F2: 5'-CAAGAGCTGCAGTGAGCAAA-3') and exon 1 (F3: 5'-CGACAGGATGGGAAAGAAAA-3'); and reverse primers located in exon 1 (R1: 5'-TGTGGGAGTAGTTCAGGAAGAA-3'), exon 2 (R2: 5'-CTCCAGCATAGGGTCCACAT-3') and exon 6 (R3: 5'-AGACATGCTCTGCCACGC-3'). PCR amplification was carried out with a 5 min of preincubation at 95°C, followed by 35 cycles of the following profile: denaturation at 95°C for 60 s, annealing at 55°C for 60 s, and extension at 72°C for 60 s. PCR products were analyzed on a 2% agarose gel stained with ethidium bromide. In some cases their identity was confirmed by DNA sequencing.
Data from healthy urothelium were paired with data from tumor tissue and statistically analyzed using Kruskal-Wallis (one-way ANOVA) test. The correlation between the sialyltransferase gene expression and activity, in the cell lines, was analyzed using Spearman and Pearson methods. All statistical tests were considered statistically significant when p < 0.05. All data are expressed as the mean ± SEM. Analyses were conducted using GraphPad Prism software, version 5.0 (GraphPad Software, La Jolla, CA).
ST3Gal.Iexpression is increased in bladder cancer
The expression of ST3Gal.II was not statistically different between tumor tissue and normal urothelium in all the analyzed groups of patients. However, in the group II of patients, ST3Gal.II mean mRNA levels were twice higher [ranging from 47.9 ± 20.9 (urothelium) to 38.8 ± 11.2 (tumors)] than in group I [ranging from 27.8 ± 4.1 (urothelium) to 23.4 ± 4.0 (tumors)] and five times higher than group III [ranging from 9.4 ± 3.6 (urothelium) to 5.1 ± 2.0 (tumors)] (Fig 1B).
Similarly, while ST3Gal.IV expression was also not statistically different between tumor tissue and normal urothelium in all the analyzed groups of patients, in samples from patients with non malignant tumors (group III) its mRNA expression was 3 to 4 times lower [ranging from 1.9 ± 1.1 (urothelium) to 3.5 ± 0.9 (tumors)] than samples from other patient's groups. Taken as a whole, the ST3Gal.IV mRNA levels were quite inferior to the mRNA levels of the ST3Gal.I and ST3Gal.II (Fig 1C).
Although, not statistically significant, we observed unexpected differences mRNA sialyltransferase expression levels among the normal urothelium of patients from the three groups. This is most evident for ST3Gal.I, which is expressed at the lowest level by patients who developed non malignant tumors (group III), at an intermediate level by those who developed malignant but nonmuscle-invasive cancers (group I) and at the highest level by those who developed invasive cancers (group II). These data can be interpreted as the result of a "field effect" of the tumor, whose presence and degree of malignancy would affect sialyltransferase expression also in apparently uninvolved tissues .
Nonmuscle-invasive bladder cancer with tendency for recurrence show pronounced ST3Gal.I mRNA increase and decreased ST3Gal.IImRNA levels
33% of the patients with nonmuscle-invasive bladder cancer, included in this study, have experienced recurrence in less than 12 months after TURBT plus intravesical BCG. To elucidate whether the expression of ST3Gal.I, ST3Gal.II or ST3Gal.IV genes was dissimilar among patients without recurrence (Group IA) and those with subsequent recurrence (Group IB), we examined the transcripts levels of these sialyltransferases in these subgroups of patients.
ST3Gal.I is responsible for the sialylation of T antigens in bladder cancer cell lines
Sialyl-T antigens are expressed in bladder tumor tissue and urothelium
Bladder cancer cells express two ST3Gal.Itranscript variants
In the first paper describing the molecular cloning of human ST3Gal.I , it was reported the presence of two types of ST3Gal.I cDNAs which differ in the 5'untranslated region; and the shorter of the two transcript variants misses one or more exons. While the sequences of these two ST3Gal.I transcript variants were deposited at GenBank (Reference Sequences: NM_003033.2 and NM_173344.2 for the longer and shorter form, respectively), there was no further reference in the literature concerning this issue. In addition, it was not consensual whether there were six or seven exons in the coding region.
Although the intensities of the PCR products differed among the analyzed cell lines, all samples expressed both transcripts (Fig 6B). The two ST3Gal.I transcripts were also observed in the patient's specimens, with no significant differences regarding their nature (urothelium or tumor tissue) and the tumor grade (Fig 6B).
Considering the augmented ST3Gal.I mRNA expression which we observed in bladder cancer, it was not possible to correlate it with one of the transcripts. Although further studies should be conducted, these results suggest that, in these bladder cancer tissue and urothelium, both ST3Gal.I transcripts are randomly expressed, with no particular association with the overall expression of the gene and the type of tissue.
It is relevant to note that during the in silico analysis we have observed that the shorter cDNA had the highest identity (89%) with a pseudogene located at chromosome 4, followed by homology with a chimpanzee pseudogene, also chromosome 4, and then by genomic ST3Gal.I sequence, at chromosome 8. This sequence has no homology with any viral or bacterial gene and it is not flanked by any insertion element, which excludes the hypothesis of a pathogenic chromosome insertion.
A great part of the aberrant glycosylation found in cancer cells is due to the altered expression of sialyltransferases. These enzymes are considered as prognostic factors in some cancers  and recent data has revealed that they are important targets for therapeutic approaches [30, 31]. Investigating the relevance of sialyltransferases in cancer is however a complex task, due the diversity expressed by human cells and for presenting some overlapping substrate specificities, tissue-restricted patterns of expression and individual variations.
Our aim was to investigate, in bladder cancers, the expression of the sialyltransferases, ST3Gal.I, ST3Gal.II and ST3Gal.IV, which are proposed to have a decisive role in the T antigen sialylation . Our study with different groups of patients with bladder tumors, demonstrate markedly that the ST3Gal.I mRNA levels increases during the malignant transformation of urothelium. In fact, malignant bladder tumors show increased ST3Gal.I mRNA levels when compared with non malignant tumors. Particularly, tumors from patients with nonmuscle-invasive bladder cancers show significant increased ST3Gal.I mRNA levels compared to normal urothelium and, in patients where cancer penetrates the muscle layer (muscle-invasive), the increased ST3Gal.I mRNA is also observed in the apparent normal mucosa. This unexpected observation can be explained by at least two non mutually exclusive mechanisms. First, it can be hypothesized that the tumor exerts a kind of "field effect" which through the release of soluble mediators or by other means alters the basic sialyltransferases expression level in normal tissues. Interestingly, this effect is positively related with malignancy, since the level of ST3Gal.I in urothelium is the lowest in patients bearing non malignant tumors, intermediate in the group with nonmuscle-invasive cancer and highest in patients with invasive cancer. Second (but less likely), it is possible that individuals with a high basic level of ST3Gal.I in urothelium, are more prone to the development of invasive bladder cancers.
Interestingly, when further analysing specimens from patients with nonmuscle-invasive bladder cancer, subgrouped according to the development of recurrence, we found that tumors of patients who experienced subsequent recurrence showed higher increased expression of ST3Gal.I and decreased expression of ST3Gal.II (Fig 2). Therefore, our findings suggest that alterations in ST3Gal.I and ST3Gal.II transcript levels may anticipate recurrence propensity. Nevertheless, since we could not obtain higher number of patients within the defined inclusion criteria, the relevance of these data should be further substantiated. Even so, to our knowledge, this is the first report demonstrating that, likewise to other cancers types [13, 15], the bladder cancer tissue also presents a deregulated expression of sialyltransferases.
Through the study of different bladder cancer cell lines it was possible to establish a functional correlation between ST3Gal.I and the manifestation of sialyl-T antigens at cell surface. Nevertheless, in patient specimens, it was not possible to confirm this correlation. Surprisingly, bladder tumor tissues, independently of their malignancy or grade, show higher PNA reactivity than the corresponding urothelium, suggesting the expression of higher content of non sialylated T antigens. One hypothesis to explain this discrepancy between ST3Gal.I expression and PNA reactivity is the fact that, in bladder cancer tissue, the ST3Gal.I is acting specifically on a limited number of molecules and any resulting increased content of sialyl-T containing glycans, due to its activity, is not perceptible by the adopted methodology, the dot blot analysis. Another possibility is that in cancer tissues there is a strong expression of T antigen molecules which cannot be fully sialylated neither in the presence of increased ST3Gal.I activity. Thirdly, there may exist translational regulation mechanisms of ST3Gal.I or the competition with other glycosyltransferase, such as core 2 β1-6 GlcNAc transferases, which compete with ST3Gal.I for core 1, generating the core 2 structures instead of sialyl-T antigens .
For this study, we selected the staining with PNA lectin which recognizes non sialylated T antigens and the expression of sialyl-T antigens is inferred after desialylating the tissue homogenates with neuraminidase treatment. This staining method is largely used for the detection of T and sialyl-T antigens in human normal and malignant tissues because it requires little amount of samples [32, 33]. The reactivity of PNA lectin with bladder cancer specimens has been already studied by other authors [34, 35], but the conclusions were controversial: while some  only observed PNA reactivity after neuraminidase treatment of bladder tumors, others  observed positive reactivity of PNA lectin in tumors without neuraminidase treatment, which was roughly consistent with the results here presented.
Interestingly, in a great part of other cancers, where a deregulated sialyltransferase expression is observed, the enzyme/product relationship is not direct (reviewed in ). Also, the use of cell line experimental models to correlate the overexpression of a given sialyltransferase and cancer-associated phenotypes provide results sometimes conflicting, mostly because of the differences between the glycosidic acceptors of each tumor type and each cell line .
ST3Gal.I gene is known to code for two transcript variants encoding the same protein . It was suggested that a longer ST3Gal.I transcript could play an important role in the translational control of the expression of ST3Gal.I, since its longest 5' untranslated region contains multiple upstream ATG codons and open reading frames ("minicistrons") which could repress translation . In this study, we observed that both transcripts were expressed in the majority of all the analyzed samples, both in healthy urothelium and tumor tissue, suggesting that the differential expression of these transcripts is not dictated by the malignant transformation of the cells. Apparently, the two transcripts were expressed in bladder cancer cells expressing different levels of ST3Gal.I mRNA and having different levels of ST3Gal.I activities. Albeit further functional investigations should be addressed, we suggest that, in bladder cancer, both transcripts have a redundant role and none of them have a particular role in the translational control of the expression of ST3Gal.I.
While we revised the characteristics of ST3Gal.I transcripts, two interesting features were observed: 1) the fact that the shorter ST3Gal.I transcript lacks a sequence within exon 1, at the 5' untranslated region, and 2) the identification of a pseudogene with high homology with ST3Gal.I cDNA. Similar chromosome insertion of "cDNA" was not found in any other sialyltransferase member and its origin is not explicable by the present genomic paradigms. Probably, some gene transcripts were processed by internal or external reverse transcriptases and inserted in genome during evolution. Nevertheless, it remains to clarify whether this sequence plays any role in regulating ST3Gal.I expression.
Collectively, the data presented in this paper supports the hypothesis that aberrant expression of the sialyltransferases is part of initial oncogenic transformation of bladder. ST3Gal.I plays the major role in the T antigen sialylation, and its expression is associated with bladder cancer malignancy and recurrence. However, further work is required to identify the involved molecular events and assess the ST3Gal.I role, aiming the identification of potential immunotherapeutic targets. Nevertheless, our findings suggest that the sialyltransferase mRNA levels, in patient's bladder tissue, are parameters to be considered when predicting progression and recurrence of cancer.
This work was partially supported by Astellas Pharma (Yamanuchi Price). We acknowledge University of Bologna for supporting stay of PAV at this University (Early Stage Research Fellowship from Institute for Advanced Studies). We thank David Lopes for the assistance in the transcript analysis.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.