Volume 92, Issue 6 p. 797-804
Research Article
Free Access

Overexpression of lysosomal-type sialidase leads to suppression of metastasis associated with reversion of malignant phenotype in murine B16 melanoma cells

Takehito Kato

Takehito Kato

Division of Biochemistry, Research Institute, Miyagi Prefectural Cancer Center, Natori, Miyagi, Japan

2nd Department of Surgery, Tohoku University, School of Medicine, Sendai, Miyagi, Japan

Search for more papers by this author
Yan Wang

Yan Wang

Division of Biochemistry, Research Institute, Miyagi Prefectural Cancer Center, Natori, Miyagi, Japan

Search for more papers by this author
Kazunori Yamaguchi

Kazunori Yamaguchi

Division of Biochemistry, Research Institute, Miyagi Prefectural Cancer Center, Natori, Miyagi, Japan

Search for more papers by this author
Caroline M. Milner

Caroline M. Milner

MRC Immunochemistry Unit, Department of Biochemistry, University of Oxford, Oxford, United Kingdom

Search for more papers by this author
Ryuzaburo Shineha

Ryuzaburo Shineha

2nd Department of Surgery, Tohoku University, School of Medicine, Sendai, Miyagi, Japan

Search for more papers by this author
Susumu Satomi

Susumu Satomi

2nd Department of Surgery, Tohoku University, School of Medicine, Sendai, Miyagi, Japan

Search for more papers by this author
Taeko Miyagi

Corresponding Author

Taeko Miyagi

Division of Biochemistry, Research Institute, Miyagi Prefectural Cancer Center, Natori, Miyagi, Japan

Fax: +81-22-381-1195

Division of Biochemistry, Research Institute, Miyagi Prefectural Cancer Center, Medeshima-shiode, Natori, Miyagi 981-1293, JapanSearch for more papers by this author
First published: 03 May 2001
Citations: 53

Abstract

Increased sialylation in cell surface glycoproteins is one characteristic feature of cancer cells, particularly related to their metastatic potential and invasiveness. Expression of lysosomal-type sialidase, which plays a major role in hydrolysis of such sialo-glycoproteins, is therefore considered to have a great influence on malignant properties of cancer cells. To investigate whether the sialidase expression level is linked to the malignant phenotype, we transfected B16-BL6 murine melanoma cells, a highly invasive and metastatic line, with an expression vector harboring a rat lysosomal sialidase cDNA; then clones were isolated and examined for changes in biological character. Sialidase-overexpressing cells showed suppression of experimental pulmonary metastasis and tumor progression. The transfectants exhibited diminished cell growth, anchorage-independent growth and increased sensitivity to apoptosis induced by suspension culture or serum depletion in vitro, but no significant alterations in invasiveness, cell motility and cell attachment to fibronectin, collagen IV and laminin. Flow cytometric analysis with either peanut agglutinin (PNA) or Ricinus communis agglutinin (RCA) lectin revealed that desialylated forms of glycoproteins on the cell surfaces were increased. In particular, a desialylated form of a cell surface glycoprotein of 83 kDa was prominent in the transfectants, as determined by galactose oxidase labeling. These observations indicate that sialidase expression is inversely associated with metastatic potential and tumor growth in cancer cells, probably through a regulation mechanism that suppresses cell growth and anchorage-independent growth and promotes apoptosis with deprivation of cell anchorage. © 2001 Wiley-Liss, Inc.

Sialidase (EC 3.2.1.18) is a key enzyme in catabolism of glycoproteins and glycolipids. Desialylation of these glycoconjugates is a crucial event leading to modulation of cellular functions in numerous physiological and pathological processes.1, 2 Alterations in sialylation during malignant transformation have been observed to be closely associated with the malignant phenotype in terms of metastatic potential and invasiveness.3-7 To understand how sialidases are involved in this aberrant sialylation, our study has focused on mammalian forms in cancer cells.

Mammalian sialidases are classified into 3 to 4 forms based on their subcellular localization and substrate preference. Cytosolic, lysosomal and membrane forms of the sialidases have been cloned; their primary sequences revealed that they are proteins encoded by distinct genes with different major subcellular locations and enzymatic properties.8 We previously demonstrated that metastatic potential is inversely correlated with lysosomal-type sialidase activity in transformed rat 3Y1 cells9 and in murine B16 melanoma cells,10 whereas it did not have a significant relation to sialic acid levels. We also investigated the role of sialidase in metastatic potential by cloning and transfecting a rat cytosolic sialidase cDNA into highly metastatic and invasive B16 melanoma variants (B16-BL6 cell lines). The result was decreased pulmonary metastasis, invasiveness and cell motility with an associated decrease in GM3 content.11

As we have now been able to obtain a rat lysosomal sialidase cDNA and stable transfectants by introduction of this gene into B16-BL6 cell lines, the effects of sialidase expression on malignant phenotype, including metastasis and tumor growth, could be shown to be due to completely different mechanisms from the cytosolic sialidase case. It is generally known that the major role of lysosomal sialidase is in glycoconjugate catabolism because of its subcellular location and because of sialo-oligosaccharide storage in sialidosis cases lacking this sialidase.12 The present study for the first time provides evidence that the sialidase is able to participate in regulation of important cellular functions such as cell proliferation and apoptosis, with increased expression leading to reversion of the malignant phenotype of cancer cells.

MATERIAL AND METHODS

Cell lines

The murine melanoma variant B16-BL6, a highly metastatic and invasive subclone derived from the B16 murine melanoma, the origins and properties of which were described by Poste et al.13 and by Fidler,14 was kindly provided by Dr. S. Taniguchi (Shinshu University School of Medicine, Shinshu, Japan). COS-1 cells were obtained from the RIKEN Cell Bank. They were grown in DMEM supplemented with 10% FBS in 5% CO2. Cells in the exponential growth phase within four passages were used for this investigation.

Lysosomal sialidase cDNA transfection

A rat lysosomal sialidase cDNA was isolated from a rat liver cDNA library by hybridization using a 1.2 kb cDNA with the entire coding sequence for human lysosomal sialidase (accession number X78687)15 as probe, according to the procedure described previously.16 The nucleotide sequence has been submitted to the DDBJ, EMBL and GenBank nucleotide sequence databases with the accession number AB035722.

To examine whether the cDNA encodes a lysosomal sialidase, the 1.2 kb fragment (nucleotides 33 to 1269) of the complete cDNA in Bluescript was amplified by PCR and subcloned into the EcoRI site of pCAGGS, a eukaryote expression vector with the chicken β-actin promoter, a generous gift from Dr. Jun-ichi Miyazaki (Osaka University School of Medicine, Osaka, Japan). The expression plasmid (pCArlSD) was transfected into COS-1 cells grown in DMEM supplemented with 10% FBS by electroporation. After 48 hr of growth in culture, the cells were harvested and used for sialidase assays. To obtain stable transfectants, the neomycin-resistant plasmid pCV107 was co-transfected with the expression vector pCArlSD into B16-BL6 cells (1 × 105/35 mm dish) using Lipofectamine (GIBCO BRL, Grand Island, NY) or Effectene (QIAGEN, Hilden, Germany) as recommended by the manufacturers. Colonies resistant to G418 (Geneticin, GIBCO BRL) at 700 μg/ml for 25 to 30 days were isolated.

Sialidase assay

Cells were washed with PBS and sonicated on ice for 8 sec in 9 vol of ice-cold PBS containing 0.2 mM phenylmethylsulfonyl fluoride, (10 μg/ml) leupeptin and pepstatin (0.5 μg/ml) at an intermediate setting (Sonifier 250; Branson, Danbury, CT). The mixture was centrifuged at 600g for 10 min at 4°C, and the supernatant (crude homogenate) was routinely used as the enzyme source. In some cases, the homogenate was further centrifuged at 100,000g, and the resulting pellet was used as particulate fraction. The assay mixture contained 60 nmol of the substrate 4-methylumbelliferyl N-acetylneuraminic acid (4MU-NeuAc) (Nakarai, Kyoto, Japan), 5 μmol of sodium acetate (pH 4.7), 100 μg of BSA and the enzyme fractions (40 to 80 μg protein) in a final volume of 0.1 ml. After incubation at 37°C for 1 hr, the reaction was terminated by addition of 2.5 ml of 0.25 M glycine-NaOH (pH 10.4), and the amount of 4-methylumbelliferone released was determined fluorometrically. When sialyllactose, glycoproteins and gangliosides were substrates, the released sialic acid was determined by the thiobarbituric acid method of Aminoff17 as described previously.18

RT-PCR analysis for lysosomal sialidase and protective protein expression

To confirm the expression of rat lysosomal sialidase exogenously introduced, the mRNA level was evaluated in the transfectants by quantitative RT-PCR. The cDNA competitor was prepared by BbsI digestion of the Bluescript vectors containing entire open reading frames (ORFs) of lysosomal sialidase. Primers were sense (5′-CACGTGGTCCTCTACGGCTT TC-3′, nucleotides 341 to 360 from the start codon) and antisense (5′-AATGTGCCATTG CTGAAGCTCC-3′, nucleotides 1025 to 1046). First-strand cDNAs were synthesized from total RNA of the transfectants by reverse transcription and then used as templates for PCR under the following conditions: 1 min at 94°C, 1 min at 57°C and 72°C for 37 cycles, followed by 10 min at 72°C. To normalize for sample variation, the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH)19 was also measured as an internal control using a GAPDH competitor, prepared by digestion of the 0.5 kb cDNA (nucleotides 566 to 1017) with StyI and BspMI. The mRNA level of protective protein was also estimated in the transfectants by semi-quntitative RT-PCR, since protective protein has been described to be essential for the sialidase activity expression.20 A set of primers, sense (nucleotides 268 to 292 from the start codon) and antisense (nucleotides 1298 to 1321), were synthesized according to the mouse cDNA sequence,21 and PCR was conducted using cDNA as mentioned above as the template at an annealing temperature of 55°C.

Determination of cellular sialic acids

Cells (107) at subconfluency were harvested, washed with PBS and lyophilized. The glycolipids were extracted in sequence with 5 ml chloroform/methanol (C/M, 1:1, v/v), 2.5 ml C/M (2:1,v/v) and 2.5 ml C/M(1:2, v/v) and then evaporated to dryness. To determine the cellular sialic acids, the glycolipid extracts (lipid-bound sialic acid) and the residues after extraction (protein-bound sialic acid) were treated in 0.1 N H2SO4 at 80°C for 1 hr and assayed for sialic acids by the thiobarbituric acid method of Aminoff.17 To estimate small amounts of sialic acids, fluorometric high-performance liquid chromatography with 1, 2-diamino-4, 5methylene dioxybenzene was used.22

Analysis of metastatic ability in vivo

The experimental metastatic ability of the cell lines was determined by injecting 2 × 105 viable cells into the tail vein of male C57BL/6 mice and male athymic BALB/c nude mice. Colonies visible on the lung surface were counted and the lung weights were measured following death, 18 to 21 days after injection.

Cell and tumor growth

In vitro cell growth was determined by MTT assay with the WST-1Cell Proliferation Assay System (Takara, Tokyo, Japan). The seeding cells in 96-well tissue culture plates at 2.5 × 103 cells/well were assayed every 24 hr. In vivo tumor growth was evaluated by measuring the dimensions of tumors formed by inoculation of cells (5 × 105) into the subcutaneous space in the back of mice. Tumor size was calculated from the equation: V = (a × b2) × 0.5, where a is the length and b is the width.

Soft agar colony formation assay

The ability of the cell lines to grow in soft agar was determined. Hard agar was prepared in 35 mm dishes with 2 ml of DMEM containing 10% FBS and 0.72% bactoagar (DIFCO, Detroit, MI). Cells (2 to 5 × 104) were rapidly suspended in 2 ml of DMEM containing 10% FBS and 0.36% Bactoagar at 37°C and spread onto the hard agar layer. After incubation for 3 weeks at 37°C, colonies and single cells in agar were counted. Colony ratios were assessed as percentages of colonies from the sums of colonies and single cells.

Attachment inhibition with polyhydroxyethyl-methacrylate coating

The culture plates were coated with dilutions of poly-hydroxyethyl-methacrylate (poly HEMA, Sigma, St. Louis, MO). Poly-HEMA (6 g) was dissolved in 50 ml 95% ethanol, and the mixture was stirred slowly overnight at 37°C to dissolve the polymer completely. After centrifugation for 30 min at 2500 rpm to remove undissolved particles, this stock was then diluted with 95% ethanol from 10−1 to 10−7, and 0.1 ml/cm2 was pipetted into the culture plates. The plates were allowed to dry for 48 hr at 37°C with the lids off. Cells were plated at 15,000/cm2 and analyzed at 72 hr of culture morphologically and flow cytometrically.

In vitro migration, invasion and cell attachment

The assays for cell motility and invasion were performed as previously described.11 Cell motility was determined by measuring phagokinetic tracks of the cells on colloidal gold-coated coverslips; for the assay of invasive potential, Biocoat Matrigel Invason Chambers (Becton Dickinson Labware, Braintree, MA) was used. Cell attachment assays were carried out using 3′-O-acetyl-2′,7′-bis(carboxyethyl)-4 or 5-carboxyfluorescein diacethoxymethyl ester (BCECF-AM; Wako, Osaka, Japan). BCECF-AM in DMSO solution was incubated at 37°C for 30 min in 1 ml of serum-free DMEM containing 1% BSA, and after washing, the cells (1 × 105) were plated in triplicate onto 24-well test plates of BIOCOAT Cellware coated with laminin, fibronectin or collagen IV (Becton Dickinson Labware). At 30 min intervals, the cells were washed with PBS and incubated in 1 ml of PBS containing 1% Triton X-100 at 37°C for 2 hr. The amount of BCECF hydrolyzed by attached cells was assayed fluorometrically.

Flow cytometry

Cells (105) were stained with fluorescein isothiocyanate (FITC)-labeled lectins (Honen, Tokyo, Japan) on ice for 30 min, washed twice with PBS, suspended in 0.5 ml of cold PBS and then analyzed using a FACScan (Becton Dickinson, San Jose, CA). When anti-GM3 mAb (M2590; Snow Brand, Tokyo, Japan) was employed, the cells were incubated for 1 hr with the unlabeled antibody, washed twice and stained with FITC-goat anti-mouse IgG F(ab′)2 (Tago). For cell cycle and apoptosis analyses, 5 × 105 cells were fixed in cold 70% ethanol, treated with 100 μg/ml RNase A for 30 min at 37°C and stained with 50 μg/ml propidium iodide in the dark. The extent of apoptosis induced by serum depletion was also determined after 48 hr culture in DMEM without FBS. The cells were double-stained with annexin V-Fluorescein (Roche Molecular Biochemicals, Mannheim, Germany) and propidium iodide and subjected to flow cytometric analysis.

Lectin blotting

The lectins used were Sambucus siebaldiana agglutinin (SSA), Maackia amurens mutagen (MAM), wheat-germ agglutinin (WGA), peanut agglutinin (PNA) and Ricinus communis agglutinin (RCA; Honen, Tokyo, Japan). Lectin blot analysis was conducted as described previously.11 Briefly, cell homogenates were resolved by 10% SDS-PAGE and transferred to Hybond C membranes (Amersham Pharmacia, Buckinghamshire, UK). Blots were washed in TBST solution [10 mM Tris-HCl (pH 7.5), 100 mM NaCl and 0.1% Tween 20] and incubated with TBST containing biotinylated lectins. After washing, lectin-binding glycoproteins were visualized with horseradish peroxidase-streptavidin (Vector, Burlingame, CA).

Galactose oxidase surface labeling

The cells (4 × 105) grown in 35 mm dishes were washed twice with PBS and incubated with 10 units of galactose oxidase (Sigma) in 100 μl PBS for 30 min at room temperature. The oxidized cells were washed, collected in PBS containing 0.02% EDTA and tritiated with 1 mCi NaB3[H]4 for 30 min at room temperature. Further reduction was then conducted with 1.25 mg nonradioactive NaBH4 (5 mg/ml of 0.01 N NaOH) for 15 min. Excess radioactivity was removed by washing with PBS three times, and the cell pellets were subjected to 7% SDS-PAGE (16 × 14 cm gel size). Labeled compounds were detected by enhanced autoradiography using EN3HANCE (NEN Life Science Products, Boston, MA).

Immunoprecipitation and Western blot analysis

Cell homogenates (35 μg protein) were electrophoresed on 7% polyacrylamide gels under reducing conditions and transferred to PVDF membranes (Hybond P; Amersham Pharmacia). After blocking with TBST containing 1% BSA for 1 hr at room temperature, the membranes were incubated for 1 hr with the first antibody in TBST, washed extensively and then incubated for 1 hr with peroxidase-conjugated goat anti-mouse IgG or anti-rabbit IgG (Promega, Madison, WI). After washing several times, they were developed with an enhanced chemiluminescence system (ECL; Amersham Pharmacia), according to the manufacturer's procedures. In some cases cell homogenates (400 μg protein) were subjected to immunoprecipitation with the first antibody, and the immune complexes recovered with protein G-Sepharose (Amersham Pharmacia) were applied to electrophoresis as described above. Tyrosine phosphorylation level was estimated for growth factor receptors and for focal adhesion kinase (FAK) using an anti-phosphotyrosine peroxidase conjugate (ECL; Amersham Pharmacia) after immunoprecipitation. Antibodies for the following proteins were purchased: insulin receptor β, insulin-like growth factor receptor β, epidermal growth factor (EGF) receptor, transferrin receptor and FAK (Santa Cruz Biotechnology, Santa Cruz, CA); and PDGF receptor and E-cadherin (Transduction Laboratories, San Diego, CA).

RESULTS

Transfection of a lysosomal sialidase cDNA to B16-BL6 cells

Since enzymatic properties of rat lysosomal sialidase have been characterized using partially purified enzyme fractions from rat liver,18 we obtained a corresponding cDNA clone (accession number AB035722) and employed the sialidase for the present investigation. When the primary sequence of the rat lysosomal sialidase was compared with human15,23,24 and mouse25-27 lysosomal sialidases so far cloned, a high identity was found (81% and 94% in nucleotide sequence and 82% and 97% in amino acid sequence, respectively). To confirm that the cDNA encodes lysosomal sialidase, an expression plasmid containing the entire ORF of the gene (pCArlSD) was transiently transfected into COS-1cells. The transfectants showed over 20-fold increase in sialidase activity with 4MU-NeuAc, a good synthetic substrate for the enzyme, at acidic pH compared with vector-harboring cells (253 ± 35 versus 12.3 ± 1.3 units/mg protein).

Highly metastatic B16-BL6 cells were then transfected with either control plasmid (pCAGGS) or with the plasmid containing the sialidase cDNA (pCArlSD). Four clones (BL6-lSD2, -5, -6 and -11) derived from the transfection were examined for sialidase activity as summarized in Table I. They showed 2- to 13- and 2- to 17-fold increase in activity toward 4MU-NeuAc compared with the parent cells (B16-BL6) and G418-resistant control cells (BL6-neo), respectively. The activity toward sialyllactose was also increased similarly in the sialidase transfectants. More than 75% of the activities in crude homogenates of all the transfectants was recovered in the particulate fractions. The transfectant BL6-lSD2 showed a certain amount of hydrolytic activity toward fetuin (15% to 20% relative to sialyllactose) and hardly any gangliosides. However, gangliosides GM3 and GM2 became sensitive to sialidase in the presence of 0.1% sodium cholate, and their hydrolysis rates were 15% and 10% relative to sialyllactose, respectively.

Table I. Expression of Lysosomal Sialidase Activity in Sialidase-Transfected B16-BL6 Cells1
Cell line Sialidase activity (units/mg protein)
4MU-NeuAc Sialyllactose
B16-BL6 11.2 ± 0.5 3.8 ± 0.7
BL6-neo 14.4 ± 0.4 4.4 ± 0.4
BL6-1SD2 115.3 ± 7.2 30.8 ± 1.1
BL6-1SD5 194.2 ± 15.3 55.6 ± 3.6
BL6-1SD6 70.2 ± 12.7 21.6 ± 4.0
BL6-1SD11 25.3 ± 7.0 7.5 ± 0.6
  • 1 Enzyme activity was determined at pH 4.5 in the crude extracts. Each value represents a mean of data from 3 experiments.

We then examined expression of lysosomal sialidase by competitive RT-PCR, as shown in Figure 1. The level was found to be nearly parallel to the activity level as determined by densitometric analysis, although low levels of endogenous sialidase were detected in control cells due to the structural similarity to the endogenous mouse sialidase. To observe whether any change occurred in the expression of protective protein, an essential factor for sialidase activity, the mRNA level in the transfectants was also evaluated by RT-PCR. There was no significant difference in the expression among the cells before and after transfection of the sialidase gene (data not shown). We next compared the cellular sialic acid contents between the transfectants and control cells. Protein-bound sialic acids were determined to be 48.2, 49.9, 44.6 and 45.4 nmol, and lipid-bound sialic acids were 18.7, 19.3, 22.7 and 19.9 nmol in 107 cells of parental cells, BL6-neo, BL6-lSD2 and BL6-lSD5, respectively. These values indicate that sialidase overexpression did not bring about a drastic quantitative change in cellular sialic acid content. The sialidase transfection did not give rise to major morphological changes in the transfectants.

Details are in the caption following the image

Quantitative measurement of lysosomal sialidase mRNA by competitive PCR. Control and sialidase-transfected cells were harvested for total RNA preparation. The mRNA level was quantified as described in the Material and Methods.

Suppression of lung metastasis

The effects of overexpression of lysosomal sialidase on metastatic potential were examined by injecting the transfectants i.v. into syngeneic C57BL/6 mice. Evaluation by comparing whole lung weights (Fig. 2a), with extensive infiltration of control cells, showed reduction of metastasis by 40% to 76% in the sialidase transfectants, although the extent was not closely correlated to exogenous sialidase activity. The average number of lung metastatic nodules produced by the transfectants was also decreased compared with the control cell case (Fig. 2b). The discrepancy between lung weight and the number of nodules, especially in BL6-lSD-2 and 11, may be partly due to size difference in nodules, since it was observed that the nodules in the former transfectant were somewhat larger than those in control cells and smaller in the latter. A similar reduction in lung metastasis was observed in both athymic BALB/c nude mice and SCID mice, suggesting a negligible involvement of the immune system in the reduced metastasis.

Details are in the caption following the image

Experimental lung metastasis of control and sialidase-transfected B16-BL6 cells. Extent of metastasis was assessed by weight of lungs (a) and number of lung metastatic colonies (b). Bars indicate standard deviations for 3 independent experiments. Significantly different from BL6-neo (*,p < 0.05; **, p < 0.01 evaluated by Student's t-test).

Growth suppression in vivo and in vitro

To test the effect on tumor formation, we inoculated exponentially growing cells of BL6-neo, BL6-lSD2 and BL6-lSD5 subcutanously into the syngeneic mice at 5 × 105 cells/site and followed progression of tumor formation (Fig. 3a). Although all these cells produced tumors 19 days after injection, the two transfectants formed no detectable tumors until 15 days and thereafter only gradually. BL6-lSD5 with higher sialidase activity formed considerably smaller tumors compared with BL6-lSD2, the higher expression of the sialidase leading to a larger reduction in tumorigenicity as well as a longer delay in tumor formation. To examine the possibility that resumption of tumor growth might be due to decreased expression of the sialidase, the tumors excised were assayed for sialidase activity at 22 days. The tumors produced by BL6-lSD2 and BL6-lSD5 exhibited marked decrease in sialidase activity (4.12 ± 0.94, and 1.38 ± 0.48) down to the control level (4.59 ± 0.67 units/mg protein), suggesting that increased tumor growth in later stages was due to reduction or loss of the transfected sialidase expression.

Details are in the caption following the image

Tumor growth and cell proliferation of control and sialidase-transfected B16-BL6 cells. Curves for tumor growth (a) and cell growth (b) were determined, respectively, by measuring tumors in two dimensions and by MTT assay as described in Material and Methods. Each point represents a mean of data from 3 experiments.

Growth curves of the transfectants were generated to determine whether sialidase expression affects cell growth in tissue culture. Sialidase transfection dramatically reduced cell growth, as shown in Figure 3b. Overexpression of sialidase was thus found to have a marked influence on cellular growth of melanoma cells both in vivo and in vitro.

Suppression of anchorage-independent growth

We next analyzed anchorage-independent growth of the cells in soft agar, since this property is generally considered to be a marker of transformation that correlates well with tumorigenicity. The results are summarized in Figure 4a. The numbers of colonies formed were significantly decreased with the sialidase-overexpressing compared with parental and BL6-neo cells. The transfectants produced fewer colonies and remained as single cells for up to 3 weeks, in contrast to the large colonies formed with control cells (Fig. 4b). The transfectants expressing high sialidase activity exhibited a great reduction in anchorage-independent growth, whereas BL6-lSD11, with the lowest activity, showed the least effect.

Details are in the caption following the image

Colony formation in soft agar containing DMEM and 10% FBS. In the upper part of the figure (a), percentages of colonies are given, compared with that of parental B16-BL6 cells. Photographs (b) show typical examples. Significantly different from BL6-neo (*, p < 0.05; **, p < 0.01 evaluated by Student's t-test).

Increased sensitivity to apoptosis

The sensitivity regarding apoptosis induction by a stress condition was compared between the transfectants and control cells. Both were able to proliferate to a considerable extent in monolayer culture even under serum-depleted conditions (0.1%). The apoptotic cells following serum deprivation were detected by annexin V and propidium iodide staining (Fig. 5). Two positive transfectants, BL6-lSD2 and BL6-lSD6, showed increased staining to annexin V compared with control BL6 cells, whereas there was no significant difference in propidium iodide. The cells were then deprived of their anchorage by suspension culture in poly-HEMA-coated dishes, which inhibits cell attachment (Fig. 6). Parental and BL6-neo cells generated multicellular aggregates, survived and replicated, whereas the transfectants were significantly less efficient in forming aggregates and demonstrated markedly reduced survival. The cells were further analyzed by flow cytometry, as shown in Figure 6a. The population in the subG1 phase (apoptotic cells), expressed as a percentage of the total population, was increased in the sialidase transfectants following poly-HEMA-induced suspension culture. The percentages in transfectants BL6-lSD5 and BL6-lSD6 were 18.2% and 14.4%, respectively. These results indicate that sialidase overexpression resulted in increased susceptibility to apoptotic stress.

Details are in the caption following the image

Extent of apoptosis following serum deletion. At 48 hr after serum depletion, control and sialidase-transfected B16-BL6 cells were stained with propidium iodide and annexinV and analyzed by flow cytometry. Negative control (open area) was performed without any staining. The values in histograms indicate the median of the intensity. Two separate experiments gave comparable results.

Details are in the caption following the image

Induction of apoptosis by suspension culture in poly-HEMA coated plates. Cells are floating and either form aggregates or float in the medium as single cells. Flow cytometry (a) was performed for cells labeled with propidium iodide and analyzed with CellFIT. Photographs (b) show typical examples.

No significant alterations in invasiveness, cell motility and cell attachment

When invasiveness through Matrigel-coated filters in a Boyden chamber and cell motility on colloidal gold-coated coverslips were assessed, no significant difference was detectable in these properties between the sialidase-transfected cells and control cells (data not shown). Cell attachment to type IV collagen, laminin or fibronectin was not changed by sialidase transfection (data not shown).

Qualitative changes in glycoproteins in the sialidase-transfected cells

To elucidate the molecular basis of the reduced metastasis and tumorigenicity, we investigated whether specific glycoprotein molecules exist as targets for the expressing sialidase causing these phenomena. As no significant quantitative change was detected in protein-bound or lipid-bound sialic acid contents, as described earlier, we examined qualitative changes in intracellular and cell surface glycoproteins and glycolipids by flow cytometry and lectin blotting. When sialic acid-recognizing lectins (SSA binding to α2-6 linkages and MAM recognizing preferentially α2-3 linkages) were used for flow cytometric analysis, cell surface changes were not detected in the transfectants compared with control cells. On the other hand, analyses with RCA and PNA, which recognize mainly Gal-GlcNAc and Gal-GalNAc, respectively, revealed considerable increase in the binding affinity in the transfectants (Fig. 7), indicating removal of sialic acids from some cell surface molecules as a result of sialidase overexpression. Lectin blotting analysis for cellular glycoproteins, however, showed no reproducible change with MAM or SSA or with WGA having affinity for clustered sialyloligosaccharides groups as well as for terminal N-acetylglucosamines.

Details are in the caption following the image

Flow cytometry with fluorescence-conjugated lectins. The control and sialidase-transfected B16-BL6 cells were labeled with fluorescence-conjugated PNA or RCA lectin and analyzed by flow cytometry.

We then carried out qualitative analysis of exposed cell surface glycoproteins by galactose oxidase treatment followed by reduction with NaB3H4. Several bands were detected in both cells, and a 83 kDa band was found to be increased in the sialidase transfectants compared with control cells (Fig. 8). This indicates that it is a major target protein for the expressing sialidase, and it may be a result of desialylation by the sialidase. Consistent with the data for lipid-bound sialic acid contents, the cell surface expression of GM3, a major ganglioside in B16-BL6 cells, was similar in both transfectants and control cells, as assessed by flow cytometric analysis using anti-GM3 antibody (data not shown). Taken together, these results indicate that some glycoproteins including the 83 kDa glycoprotein are targets for the expressing sialidase, and that their desialylation may lead to reduced tumorigenicity and metastasis.

Details are in the caption following the image

Galactose oxidase labeling of cell surface glycoproteins. The cells labeled with galactose residues and NaB3H4 were analyzed by SDS-gel electrophoresis and detected by enhanced autoradiography (a) as described in Material and Methods. Lane 1, BL6-neo; lane 2, BL6-lSD 2; lane 3, BL6-lSD 5. Densitometric scans are shown in the lower panel (b).

To obtain further insights into the molecular mechanisms of alteration of metastasis and tumor growth, we then tested whether any functional change occurred in cell surface glycoproteins including some cell adhesion molecules and growth factor receptors (data not shown). The level of E-cadherin was first evaluated by Western and RCA lectin blotting, since it has been described to regulate anchorage-independent growth28 as well as cell adhesion, and its carbohydrate modification affects metastatic ability of tumor cells.29 However, no significant changes were detected with sialidase transfection. The phosphorylation level on FAK did not alter as expected, with no difference in attachment to the extracellular matrix. The tyrosine phosphorylation levels of growth factor receptors [insulin receptor, insulin-like growth factor receptor, EGF receptor, and platelet-derived growth factor (PDGF) receptor] were also measured as candidate molecules for cell growth suppression caused by the transfection. Similar levels were observed before and 15 min after addition of respective growth factors in the sialidase transfectants and control cells, although bands for the last two receptors were scarcely detected. Finally we examined the possibility of an involvement of transferrin receptor in the growth suppression due to the sialidase transfection, as suggested by the report30 that cell surface transferrin receptor is related to metastasis of B16 melanoma cells in serum-depleted media (0.1%). Transferrin slightly stimulated cell growth of the control and transfectants by 17% and 11% to 15%, respectively, after 24 to 48 hr of the treatment, indicating no significant change in transferrin effect by the sialidase transfection.

DISCUSSION

We previously demonstrated that endogenous lysosomal sialidase activity is inversely correlated with metastatic potential of malignant transformed rat 3Y1 fibroblasts.31 This tendency was also observed in B16 melanoma variants, even with a low level of endogenous activity.10 To understand lysosomal sialidase function in cancer metastasis, we attempted to obtain a stable transfectant by introduction of this sialidase into highly metastatic cells. In the present study we succeeded for the first time in obtaining stable transfectants and provided evidence that lysosomal sialidase overexpression inhibits the metastatic potential of B16 melanoma, at least partially through reduction of cell growth and sensitization to apoptosis. In particular, the reduction in malignancy including decreased anchorage-independent growth and tumorigenicity appeared to be partially, if not completely, dependent on the increased sialidase expression, indicating that the sialidase is involved in regulation of cellular functions affecting malignant properties of cancer cells as well as in catabolism of cellular wastes.

Restoration of anchorage dependency following transfection was shown with both colony formation in agar and deprivation of cell anchorage induced by poly-HEMA treatment. The transfectants generated many fewer cell aggregates than the control cells. Although we were not able to show any clear difference in E-cadherin expression, some other adhesion molecules might participate in this phenomenon.

The stable transfectant showed considerably higher sialidase activity than control cells with 4MU-NeuAc or sialyllactose as substrate at acidic pH without co-transfection of protective protein cDNA. This is probably due to sufficient endogenous expression of the protective protein for sialidase activity in the cells, as suggested by PCR results for the protein, or less dependence of the rat enzyme on the protein, although it has been described to be required by human and mouse enzymes for activity.20, 21 Consistent with our data on B16 melanoma transfectants, transient expression in COS-1 cells also resulted in high sialidase activity in spite of no exogenous protective protein.

Our previous work on introduction of rat cytosolic sialidase cDNA into B16BL6 cells demonstrated that overexpression led to marked suppression of metastasis accompanied by decrease in invasion and cell motility, probably based on GM3 metabolic reduction.11 Overexpression of lysosomal sialidase also resulted in decreased metastasis but with reduced anchorage-dependent growth rather than significant changes in invasion and cell motility. Consistent with the substrate specificity of this sialidase observed previously,9 the sialidase overexpressed showed the highest hydrolytic ability toward sialyllactose among natural substrates tested here, and it is somewhat more active on fetuin than on gangliosides even with 0.1% sodium cholate. It is therefore not contradictory that a major target molecule of the overexpressed sialidase was found to be a 83 kDa cell surface glycoprotein, although there is no evidence at present that this glycoprotein contributes to the reduced metastasis and malignancy. It should be noticed that, unlike these two sialidases, membrane-associated sialidase almost specific for gangliosides, which we recently cloned,32 did not cause suppression of metastasis in B16-BL6 cells (M. Sawada, T. Kato and T. Miyagi, unpublished data). The distinct effects of the overexpressed sialidases on malignant properties of the cells may be due to differences in substrate specificity and subcellular localization, implying that each form of mammalian sialidase has distinct cellular functions.

A number of studies have indicated that cell surface glycoproteins are involved in determination of the metastatic phenotype of B16 melanoma cells. Among these glycoproteins, there are several candidates with a similar size to the 83 kDa protein. Reduced sialylation of PNA binding sugar chains on a 75 to 85 kDa protein33 and increased expression of sialylated PNA oligosaccharides of a 78 kDa glycoprotein have been reported to be related to increased metastasis.34, 35 As the 78 kDa form was identified to function as an autocrine motility factor,36 it would not be expected to have been involved here because of the lack of any effect on cell motility by the sialidase transfection. Furthermore, 115, 90 and 82 kDa cell surface sialo-glycoproteins reactive to RCA lectins37 and 90 to 95 kDa transferrin-related glycoprotein have also been suggested to be associated with metastasis.30 Although it is certain that a desialylated 83 kDa protein detected by cell surface labeling was a major target molecule for the expressing sialidase, we are not able to present evidence for their direct involvement in metastatic suppression. It is still very feasible that desialylation of various glycoproteins as a whole participates in reversion of the malignant phenotype. As functional alteration of several growth factor receptors such as insulin receptor, insulin-like growth factor-I receptor, PDGF receptor and EGF receptor was not detected in the transfectants, other molecules related to cell growth and cell-cell contact may be involved in the reduced metastasis and malignancy. We are currently engaged in identifying the molecule(s) acting as a target for the expressing sialidase causing the above-described changes.

Taken together, the present results strengthen our previous data31 showing that lysosomal sialidase affects malignant properties including metastatic ability of the cancer cells; modulation of this gene expression would open up the possibility of potential applications in cancer therapy.

Acknowledgements

We appreciate the skillful technical assistance of Ms. Setsuko Moriya, Miyagi Prefectural Cancer Center.

      The full text of this article hosted at iucr.org is unavailable due to technical difficulties.