Aberrant N-linked glycans promote the malignant potential of cells by enhancing the epithelial-to-mesenchymal transition and the invasive phenotype. To identify small molecule inhibitors of N-glycan biosynthesis, we developed a chemical screen based on the ability of the tetravalent plant lectin L-phytohemagglutinin (L-PHA) to bind and crosslink surface glycoproteins with β1,6GlcNAc-branched complex type N-glycans and thereby induce agglutination and cell death. In this screen, Jurkat cells were treated with a library of off-patent chemicals (n = 1,280) to identify molecules that blocked L-PHA–induced death. The most potent hit from this screen was the cardiac glycoside (CG) dihydroouabain. In secondary assays, a panel of CGs was tested for their effects on L-PHA–induced agglutination and cell death. All of the CGs tested inhibited L-PHA–induced death in Jurkat cells, and the most potent CG tested was digoxin with an EC50 of 60 ± 20 nmol/L. Digoxin also increased the fraction of some concanavalin A–binding N-glycans. Using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, digoxin specifically increased GlcNAc1Man3GlcNAc2Fuc1 and GlcNAc2Man3GlcNAc2Fuc1 oligosaccharides demonstrating an impairment of the N-glycan pathway. Consistent with this effect on the N-glycan pathway, digoxin inhibited N-glycosylation–mediated processes of tumor cell migration and invasion. Furthermore, digoxin prevented distant tumor formation in two mouse models of metastatic prostate cancer. Thus, taken together, our high throughput screen identified CGs as modifiers of the N-glycan pathway. These molecules can be used as tools to better understand the role of N-glycans in normal and malignant cells. Moreover, these results may partly explain the anticancer effect of CGs in cardiovascular patients. [Cancer Res 2008;68(16):6688–96]

Malignant cells display characteristic changes in N-glycan structures on surface glycoproteins including receptors and transporters that contribute to cancer progression and metastasis (reviewed in ref. 1). Oncogene activation stimulates increased expression of the Golgi enzymes that generate β1,6GlcNAc-branched tetraantennary N-glycans. These N-glycans are found on proteins including growth factor receptors and integrins, and have been shown to enhance growth signaling in motile tumor cells (24). Chemical or genetic disruption of the N-glycosylation pathway in cancer cells decreases these malignant features (5, 6). Therefore, targeting defective glycosylation pathways may be a novel approach for the treatment of malignancy and a strategy to prevent metastasis.

The N-glycosylation pathway begins in the lumen of the rough endoplasmic reticulum, and remodeling in the Golgi apparatus generates structural diversity. To initiate the pathway, the oligosaccharide precursor Glc3Man9GlcNAc2 is transferred en bloc from dolichol-PPi onto asparagine (Asn) residues in the sequence Asn-X-Ser/Thr (where X can be any amino acid except for proline) to form an Asn-linked glycan (N-glycan). Once attached to the protein, this precursor is modified through a well-defined pathway leading to the sequential removal of the three glucoses and one mannose by the actions of rough endoplasmic reticulum glycosidases to form Man8GlcNAc2-Asn. In the Golgi, Man8GlcNAc2-Asn is further modified by the removal of mannoses via Golgi mannosidases and by the addition of GlcNAc via GlcNAc transferases, leading to the generation of hybrid and complex N-glycans. Finally, other sugars such as fucose, galactose, and sialic acid are added to the N-glycans to increase their diversity (Fig. 1; reviewed in refs. 7, 8).

Figure 1.

Scheme of the N-glycan biosynthesis pathway. The N-glycan biosynthesis pathway is shown. The biosynthesis of N-glycans involves the following steps: (a) sequential conversion of the high mannose N-glycan M9Gn2 to M8Gn2, M7Gn2, M6Gn2, and M5Gn2 (oligosaccharides before α-mannosidase II in Table 1); (b) addition of β2-linked GlcNAc to the terminal α3-linked Man of M5Gn2 by GlcNAc-TI to form Gn1M5Gn2 (oligosaccharides before α-mannosidase II in Table 1); (c) removal of two Man residues from Gn1M5Gn2 by α-mannosidase II to form Gn1M3Gn2 (oligosaccharides after α-mannosidase II-Hybrid structures in Table 1); (d) addition of a second β2-linked GlcNAc to the terminal α6-linked Man of Gn1M3Gn2 by GlcNAc-TII to form Gn2M3Gn2 (oligosaccharides after α-mannosidase II-Complex structures in Table 1); (e) addition of l-Fucose (F) and Galactose (Hx) residues. Swainsonine (SW) is a known inhibitor of the α-mannosidase II enzyme. OT, oligosaccharyltransferase; GI and GII, the α-glucosidases; TI, TII, TIV, and TV, β-N-acetylglucosaminyltransferases; MI, the α-1,2mannosidases; MII and MIII, α-1,3/6mannosidases; Gal-T, β1,2-galactosyltransferases; Sa T, α-sialyltransferases; UDP-GlcNAc, UDP-N-acetylglucosamine; UDP-Gal, UDP-galactose; CMP-SA, CMP-sialic acid. The Golgi apparatus subdomains (Cis, Medial, and Trans) are shown by separate boxes. Gray highlights, the N-glycan structures that bind ConA and L-PHA.

Figure 1.

Scheme of the N-glycan biosynthesis pathway. The N-glycan biosynthesis pathway is shown. The biosynthesis of N-glycans involves the following steps: (a) sequential conversion of the high mannose N-glycan M9Gn2 to M8Gn2, M7Gn2, M6Gn2, and M5Gn2 (oligosaccharides before α-mannosidase II in Table 1); (b) addition of β2-linked GlcNAc to the terminal α3-linked Man of M5Gn2 by GlcNAc-TI to form Gn1M5Gn2 (oligosaccharides before α-mannosidase II in Table 1); (c) removal of two Man residues from Gn1M5Gn2 by α-mannosidase II to form Gn1M3Gn2 (oligosaccharides after α-mannosidase II-Hybrid structures in Table 1); (d) addition of a second β2-linked GlcNAc to the terminal α6-linked Man of Gn1M3Gn2 by GlcNAc-TII to form Gn2M3Gn2 (oligosaccharides after α-mannosidase II-Complex structures in Table 1); (e) addition of l-Fucose (F) and Galactose (Hx) residues. Swainsonine (SW) is a known inhibitor of the α-mannosidase II enzyme. OT, oligosaccharyltransferase; GI and GII, the α-glucosidases; TI, TII, TIV, and TV, β-N-acetylglucosaminyltransferases; MI, the α-1,2mannosidases; MII and MIII, α-1,3/6mannosidases; Gal-T, β1,2-galactosyltransferases; Sa T, α-sialyltransferases; UDP-GlcNAc, UDP-N-acetylglucosamine; UDP-Gal, UDP-galactose; CMP-SA, CMP-sialic acid. The Golgi apparatus subdomains (Cis, Medial, and Trans) are shown by separate boxes. Gray highlights, the N-glycan structures that bind ConA and L-PHA.

Close modal

To better understand the N-glycosylation pathway and identify strategies to target the aberrant N-glycosylation in neoplastic cells, we designed, automated, and conducted a chemical screen of off-patent drugs and chemicals. This high-throughput screen identified cardiac glycoside (CG) Na+/K+-ATPase inhibitors that altered N-glycosylation. Subsequent evaluation showed that inhibition of the Na+/K+-ATPase increased the fraction of concanavalin A (ConA)-binding N-glycans. Moreover, the effects of CG on N-glycosylation seemed functionally important as these compounds inhibited the glycosylation-mediated processes of cell migration and invasion as well as decreased distant tumor formation in vivo.

Reagents. The LOPAC chemical library, L-PHA, digoxin, digitoxin, dihydrooaubain, and fibronectin were purchased from Sigma-Aldrich.

Cell culture. Jurkat human leukemia, WRO human thyroid carcinoma, PPC-1 human prostate cancer, and Colo320 colorectal adenocarcinoma cells were maintained in RPMI 1640. HT1080 human fibrosarcoma, 5637 human bladder carcinoma, and HeLa human cervical cancer cells were maintained in DMEM. All cells were supplemented with 10% fetal bovine serum (FBS; Hyclone) and antibiotics. WRO cells were also supplemented with 1 mmol/L sodium pyruvate. All cell lines were cultured in a standard humidified incubator at 37°C in a 5% CO2 atmosphere.

High-throughput screen for inhibitors of L-phytohemagglutinin–induced cell death. Liquid handling was performed by a Biomek FX Laboratory Automated Workstation (Beckman Coultera). Jurkat cells (5,000 cells per well) were seeded in 96-well plates followed by the addition of aliquots from the LOPAC library of 1,280 off-patent drugs and chemicals with a final DMSO concentration of 0.05%. Jurkat cells were selected for this assay as their growth in suspension conditions facilitated the automated nature of this screen. Twenty-four hours after addition of the compound library, L-phytohemagglutinin (L-PHA) was added at a final concentration of 20 μg/mL (9). Forty eight hours after the addition of L-PHA, cell viability was measured using the 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxphenyl)-2-(4-sulfophenyl)-2H-tetrazolium inner salt (MTS) reduction assay according to the manufacture's protocols (Promega) and as previously described (10). Cell viability was calculated relative to vehicle-treated (0.1% DMSO) control cells on each plate. To identify statistically significant hits, we calculated the Z score for each hit relative to the negative control. These Z scores were transformed into P values based on the standard-normal distribution. To control for multiple testing, we used a false discovery rate (FDR) correction to generate a q value for each compound (11). Statistically significant hits were selected as those with a q value of <5%, at which level four compounds were identified.

Measurement of ConA-binding glycopeptides. Quantification of ConA-binding glycans after swainsonine and digoxin treatment was performed as follows. Briefly, Jurkat cells (1 × 107) were treated with 2 μmol/L swainsonine, 100 nmol/L digoxin, or buffer control for 24 h. After treatment, cells were harvested, washed, and sonicated. The homogenate was centrifuged and the pellet was solubilized in 6 mol/L guanidine-HCl in 0.1 mol/L Tris buffer (pH 8.0) containing 20 mmol/L DTT to reduce disulfide bonds. Solid iodoacetamide was then added to a final concentration of 60 mmol/L, and the solution was incubated for 1 h to block free sulfhydryl groups. The solution was then centrifuged to remove any remaining insoluble residue. The solubilized, reduced, and alkylated proteins were precipitated with 10 volumes of cold absolute ethanol-glacial acetic acid.

The pellet was collected, dried, and suspended in 50 mmol/L NH4HCO3. Trypsin (Promega) at 1:100 (w:w), relative to total protein, was added (half at the beginning of the digestion and the other half 3 h later), and after an overnight digestion, the enzyme was denatured by boiling. After proteolysis and centrifugation, the supernatant was dried and redissolved in PBS. The peptide mixture was loaded onto a ConA-Sepharose column equilibrated with PBS buffer. After washing the column using PBS, the ConA-bound glycopeptides were eluted with 15% methyl-α-d-mannoside in PBS. A SepPak-C18 cartridge was then used to remove the methyl-α-d-mannoside. The glycopeptides were eluted from SepPak-C18 with 50% acetonitrile in 0.1% trifluoroacetyl, and the solution was lyophilized. Equal amounts of material were dissolved in H2O, and the amount of hexose in the sample was determined by the phenol-sulfuric acid method (12).

ConA binding to cell surface glycans. Cells (500 cells per well) were seeded in 96-well plates. After adhering overnight, cells were treated with increasing concentrations of digoxin for 48 h. After treatment, cells were fixed with 3.7% formaldehyde and washed. Surface N-glycans were stained with ConA (20 μg/mL) conjugated to tetramethylrhodamine B isothiocyanate (TRITC; EY Laboratories) and nuclei were stained with Hoechst 33342. The total intensity of ConA staining on each cell was quantified using Cellomics ArrayScan II (Cellomics; ref. 2).

siRNA transfections. Cells (1 × 105) were seeded in 6-well plates and transfected the next day using Lipofectamine 2000 (Invitrogen) and double-stranded siRNAs targeting either the human Na+/K+-ATPase, or Non-Targering siRNA (siControl; Smartpool, Dharmacon). Cells were harvested 72 h posttransfection and then assayed for ConA binding to cell surface glycans.

Reverse-transcriptase real-time PCR. First-strand cDNA was synthesized from 1 μg of DNase-treated total cellular RNA using random primers and SuperScript II reverse transcriptase (Invitrogen) according to the manufacturer's protocols. Real-time PCR assays were performed in triplicate with 5 ng of RNA equivalent cDNA, SYBR Green PCR Master mix (Applied Biosystems), and 400 nmol/L of gene-specific primers. Reactions were processed and analyzed on an ABI 7900 Sequence Detection System (Applied Biosystems). Forward/reverse PCR primer pairs for human cDNAs were as follows: murine Na+/K+-ATPase, Forward 5′-TGT GAT TCT GGC TGA GAA CG-3′ and Reverse 5′-TCT TGC AGA TGA CCA AGT CG-3′; 18S, Forward 5′-AGG AAT TGA CGG AAG GGC AC-3′ and Reverse 5′-GGA CAT CTA AGG GCA TCA CA-3′. Relative mRNA expression was determined using the ΔΔCT method as described (13).

DNA constructs and generation of stable cell lines. Stable cell lines expressing the α subunit of the murine Na+/K+-ATPase were engineered by transfecting PPC1 human prostate cancer cells with cDNA corresponding to the murine Na+/K+-ATPase in pcDNA3 vector using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Cells stably expressing murine Na+/K+-ATPase were selected with 800 μg/mL G418 (Invitrogen).

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry to semiquantify N-glycan expression. Jurkat cells (1 × 108) were treated with 100 nmol/L digoxin, 2 μmol/L Swainsonine, or buffer control. After incubation, cells were lysed in 35 mmol/L Tris, 8 mol/L urea, 4% CHAPS, and 65 mmol/L DTT (pH 8.0) followed by sonication and freezing. Equal amounts of protein were dialyzed with cassettes (Slide-A-Lyzer; molecular weight cutoff of 7,000; Pierce) in 10 mmol/L NH4HCO3/0.02% SDS. After dialysis, samples were concentrated via vacuum centrifugation. Samples were deglycosylated at 37°C for 48 h with PNGase F (New England Biolabs). Oligosaccharides were purified via C18 and porous graphitized carbon solid phase extraction. The composition of N-glycans was semiquantified using mass spectrometry via matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF), using DHB as matrix, on a MALDI-CFR (Shimadzu Biotech) as previously described (14). The intensities of the peaks were arbitrarily normalized to the level of (Hexose)2(HexNAc)2(Deoxyhexose)1.

Scratch wound–healing assay. Scratch wound–healing assay was performed as previously described (15, 16). Briefly, HT1080 (1 × 106) cells were treated with digoxin or buffer control. After 24 h, cells were harvested and seeded (1 × 105) on fibronectin-coated 4-well chamber slides. After adhering overnight, the monolayer was scratched with a plastic pipette tip. Migration of the cells over 6 h was captured with a digital camera (Nikon) mounted on an inverted microscope (Nikon). Cellular migration was measured in relative units (pixels).

Migration and invasion assays. Invasion and migration assays were performed as previously described (15). Briefly WRO cells (2 × 105) were treated with digoxin or buffer control for 24 h. After treatment, cells were harvested and seeded in uncoated invasion chambers for migration assay, or BioCoat Matrigel Invasion Chambers (BD Biosciences) in serum-free RPMI 1640 containing 0.2% bovine serum albumin were used for invasion assays. Growth medium containing 5% FBS was use as a chemoattractant in the bottom well. After 24 h of incubation, cells that had migrated or invaded the lower surface of the membrane were stained with Diff-Quik Stain (BD Biosciences). The number of migrating or invading cells were imaged and counted using the Aperio ScanScope CS whole slide Scanner (Aperio Technologies) and Image-Pro Plus Software (version 4.5; Media Cybernetics, Inc.).

In vivo studies. The effects of digoxin on distant tumor formation in vivo were evaluated as previously described (10). Briefly, dsRED-PPC-1 cells that stably express dsRed2 fluorescent protein were treated in culture with digoxin (100 nmol/L) or buffer control for 20 h. After treatment, 3.5 × 106 viable cells (as determined by trypan blue exclusion assay) were either injected via the tail vein or s.c. into the hind limbs into sublethally irradiated (3.5 Gy) male severe combined immunodeficient (SCID) mice between ages 5 and 7 wk. The number of cells injected was at least 3-fold above the minimum threshold required for distal tumor formation (data not shown). Mice injected with tumor cells s.c. were maintained for 2 wk and then sacrificed via carbon dioxide inhalation. Tumors were excised and weighed. Mice injected with tumor cells i.v. were maintained for 4 wk after injection or until moribund, at which time, the animals were sacrificed via carbon dioxide inhalation for complete examination.

In a separate model, DsRed-labeled PPC-1 cells (3 × 106) were injected i.v. into sublethally irradiated SCID mice. Mice were then treated with digoxin or buffer control daily for 2 wk. Four weeks after injection of the cells, the mice were sacrificed via carbon dioxide inhalation and dissected.

Red fluorescent tumors were detected via whole body imaging and whole organ imaging using a Leica MZ FLIII fluorescent stereomicroscope with a 100 W mercury lamp, a 560/40 excitation filter, and a 610 long-pass emission filter. Images were acquired using a Olympus DP70 digital camera at ×0.8 magnification and analyzed using Image Pro Plus 6.0 (MediaCybernetics). A single common threshold was applied to identify and measure fluorescence in each organ (17). The number of fluorescent spots was recorded for each lung lobe. All quantification was performed on unmanipulated images.

Mice were obtained from an in-house breeding program and housed in laminar-flow cage racks under standardized environmental conditions with ad libitum access to food and water. All experiments were performed according to the regulations of the Canadian Council on Animal Care.

Inhibitors of L-PHA–induced cell death identified via a high-throughput screen. Phytohemagglutinin is a tetravalent plant lectin that binds and crosslinks complex N-glycoproteins on cell surfaces and induces agglutination and cell death (18). To identify compounds that alter N-glycan remodeling in malignant cells, we developed, automated, and conducted a chemical screen for inhibitors of L-PHA–induced cell death in Jurkat cells. In the optimized assay, Jurkat cells were seeded in 96-well plates treated with aliquots of the LOPAC (n = 1,280) library of off-patent drugs and chemicals (final concentration, ∼5 μmol/L and 0.05% DMSO). Twenty-four hours after incubation, cells were treated with L-PHA (final concentration, 20 μg/mL), and 48 hours later, cell viability was measured by the MTS assay (Fig. 2A). As controls, cells received L-PHA or buffer alone. To identify molecules that inhibited L-PHA–induced cell death, we used a statistically robust methodology. For each compound in the LOPAC library, we calculated a Z score using the extensively replicated L-PHA control wells. These Z scores were converted into P values using the standard normal distribution. As we screened 1,280 distinct compounds, we then applied a FDR correction for multiple hypothesis testing to the vector of P values (11). Statistically significant hits were defined as compounds with a q value (adjusted P value) of <0.05, indicating a false-positive rate of at most 5%. Four statistically significant hits were identified in this manner. Of the four statistically significant hits, secondary screening validated one compound: the Na+/K+-ATPase inhibitor dihydroouabain. This compound restored viability of L-PHA–treated cells to 70% or over of untreated cells.

Figure 2.

A chemical screen identifies cardiac glycosides as inhibitors of L-PHA–induced cell death. A, Jurkat cells (5,000) were seeded in 96-well plates and treated with aliquots of the LOPAC chemical library (final concentration, ∼5 μmol/L). Twenty four hours after treatment, L-PHA (final concentration, 20 μg/mL) was added. Forty eight hours after L-PHA addition, cell viability was measured by an MTS assay. Viability was expressed as a percentage of buffer treated control cells. Statistically significant hits were selected as those with a q value of <5%. Four statistically significant hits were identified in this manner, which restored viability of L-PHA–treated cells to 70% or over of untreated cells (dashed line). B, Jurkat cells (5,000) were seeded in 96-well plates and treated with increasing concentrations of dihydroouabain, digitoxigenin, digitoxin, and digoxin. Twenty four hours after treatment, L-PHA (final concentration, 20 μg/mL) was added. Forty eight hours after L-PHA addition, cell viability was measured by an MTS assay. Viability was expressed as a percentage of buffer treated control cells. Points, mean percent viable cells; bars, SD.

Figure 2.

A chemical screen identifies cardiac glycosides as inhibitors of L-PHA–induced cell death. A, Jurkat cells (5,000) were seeded in 96-well plates and treated with aliquots of the LOPAC chemical library (final concentration, ∼5 μmol/L). Twenty four hours after treatment, L-PHA (final concentration, 20 μg/mL) was added. Forty eight hours after L-PHA addition, cell viability was measured by an MTS assay. Viability was expressed as a percentage of buffer treated control cells. Statistically significant hits were selected as those with a q value of <5%. Four statistically significant hits were identified in this manner, which restored viability of L-PHA–treated cells to 70% or over of untreated cells (dashed line). B, Jurkat cells (5,000) were seeded in 96-well plates and treated with increasing concentrations of dihydroouabain, digitoxigenin, digitoxin, and digoxin. Twenty four hours after treatment, L-PHA (final concentration, 20 μg/mL) was added. Forty eight hours after L-PHA addition, cell viability was measured by an MTS assay. Viability was expressed as a percentage of buffer treated control cells. Points, mean percent viable cells; bars, SD.

Close modal

Other cardiac glycosides inhibit L-PHA–induced cell death. Dihydroouabain belongs to the CG family of Na+/K+-ATPase inhibitors, which are used clinically to treat patients with heart failure and atrial arrhythmias (19). Moreover, CGs have potential antitumor properties in patients (20, 21), but the mechanism by which they exert this effect is unknown. To determine whether the effects on N-glycan remodeling were specific to dihydroouabain or a class-effect of CGs, we evaluated a panel of CGs for inhibition of L-PHA–induced cell death. Jurkat cells were treated with L-PHA along with increasing concentrations of digoxin, digitoxigenin, digitoxin, and dihydroouabain, and cell viability was measured by the MTS assay (Fig. 2B). All CGs tested blocked L-PHA–induced cell death. The most potent CG tested was digoxin that inhibited L-PHA–induced cell death with an EC50 of 60 ± 20 nmol/L. This result is in keeping with digoxin being a more potent inhibitor of the Na+/K+-ATPase than dihydroouabain and digitoxigenin (22). In addition to inhibition of L-PHA–induced cell death, CGs also blocked L-PHA–mediated agglutination of the Jurkat cells (data not shown). Of note, at higher concentrations of the CGs or longer periods of incubation, the CGs directly induced cell death, consistent with previous reports (23, 24). Compared with the human isoform, CGs bind the α1 subunit of murine Na+/K+-ATPase with less affinity, thereby rendering murine cells resistant to any effects of CGs that are mediated through this ATPase (25, 26). Consistent with this prediction, the tested CGs did not block L-PHA–induced cell death nor agglutination of MDAY-D2 murine leukemia cells (data not shown), suggesting that CGs inhibition of L-PHA toxicity to Jurkat cells requires its known function as inhibitors of human Na+/K+-ATPase.

Digoxin increases the levels of total and cell surface ConA-binding glycoproteins. To explore the effect of inhibiting the Na+/K+-ATPase on the N-glycan profile of cancer cells, we measured the effects of digoxin on levels of ConA-binding high mannose and hybrid N-glycans. Jurkat cells were treated with digoxin (100 nmol/L), the known Golgi α-mannosidase II inhibitor, swainsonine (2 μmol/L; refs. 18, 27), or buffer control for 24 hours. After treatment, the abundance of ConA-binding glycopeptides was measured by the phenol-sulfuric acid method as described in the Materials and Methods section. Digoxin and swainsonine increased the abundance of ConA-binding glycopeptides 2.3- and 1.8-fold, respectively, compared with controls (Fig. 3A).

Figure 3.

Digoxin increases levels of ConA-binding glycoproteins through inhibition of the Na+/K+-ATPase. A, Jurkat cells (1 × 107) were treated with, 100 nmol/L digoxin, 2 μmol/L Swainsonine, or buffer control for 24 h. After treatment, cells were harvested and the levels of ConA-binds glycopeptides were measured by the colorimetric technique as described in the Materials and Methods. Columns, mean concentration of ConA-binding glycopeptides; bars, SD. B, Colo320, 5637, PPC-1, HeLa, WRO, and HT1080 cells (1 × 104 cells/mL) were seeded in 96-well plates and treated with increasing concentrations of digoxin for 48 h. After incubation, cells were stained with TRITC-labeled ConA and Hoechst 33342. The intensity of ConA staining was measured using the automated Arrayscan microscope as described in the Materials and Methods. Points, mean fluorescent intensity; bars, SD. C, PPC-1 cells stably transfected with cDNA corresponding to the α1 subunit of murine Na+/K+-ATPase or vector control were treated with digoxin (50 nmol/L) for 48 h. After incubation, cells were stained with TRITC-labeled ConA and Hoechst 33342 and the intensity of ConA staining was measured using the automated Arrayscan microscope as described in the Materials and Methods. Columns, mean fluorescent intensity; bars, SD. D, PPC-1 cells were transfected with siRNA corresponding to the human Na+/K+-ATPase or a control siRNA. Three days after transfection, cells were harvested and stained with TRITC-labeled ConA and Hoechst 33342 and the intensity of ConA staining was measured using the automated Arrayscan microscope as described in the Materials and Methods. Columns, mean fluorescent intensity; bars, SD.

Figure 3.

Digoxin increases levels of ConA-binding glycoproteins through inhibition of the Na+/K+-ATPase. A, Jurkat cells (1 × 107) were treated with, 100 nmol/L digoxin, 2 μmol/L Swainsonine, or buffer control for 24 h. After treatment, cells were harvested and the levels of ConA-binds glycopeptides were measured by the colorimetric technique as described in the Materials and Methods. Columns, mean concentration of ConA-binding glycopeptides; bars, SD. B, Colo320, 5637, PPC-1, HeLa, WRO, and HT1080 cells (1 × 104 cells/mL) were seeded in 96-well plates and treated with increasing concentrations of digoxin for 48 h. After incubation, cells were stained with TRITC-labeled ConA and Hoechst 33342. The intensity of ConA staining was measured using the automated Arrayscan microscope as described in the Materials and Methods. Points, mean fluorescent intensity; bars, SD. C, PPC-1 cells stably transfected with cDNA corresponding to the α1 subunit of murine Na+/K+-ATPase or vector control were treated with digoxin (50 nmol/L) for 48 h. After incubation, cells were stained with TRITC-labeled ConA and Hoechst 33342 and the intensity of ConA staining was measured using the automated Arrayscan microscope as described in the Materials and Methods. Columns, mean fluorescent intensity; bars, SD. D, PPC-1 cells were transfected with siRNA corresponding to the human Na+/K+-ATPase or a control siRNA. Three days after transfection, cells were harvested and stained with TRITC-labeled ConA and Hoechst 33342 and the intensity of ConA staining was measured using the automated Arrayscan microscope as described in the Materials and Methods. Columns, mean fluorescent intensity; bars, SD.

Close modal

To further explore the effects of Na+/K+-ATPase inhibition on ConA-binding glycans and to evaluate the generalizability of this finding, we measured the binding of ConA to the surface of a panel of solid tumor cell lines. Colo320, 5637, HeLa, WRO, PPC-1, and HT1080 cells were treated with increasing concentrations of digoxin for 48 hours. After treatment, cells were fixed and stained with TRITC-labeled ConA. The binding of fluorescently labeled ConA was quantified by the Cellomics Array scan (Fig. 3B). Consistent with modification of the N-glycan pathway, digoxin increased the binding of ConA to the surfaces of all the cell lines studied.

To determine whether the effects of digoxin on the surface expression of ConA-binding proteins were due to inhibition of its known target Na+/K+-ATPase, we overexpressed in PPC1 human prostate cancer cells the α1 subunit of murine Na+/K+-ATPase that is less sensitive to cardiac glycosides. Increased expression of the Na+/K+-ATPase was confirmed by reverse-transcriptase real-time PCR (data not shown). Overexpression of the α1 subunit of murine Na+/K+-ATPase blocked the effects of digoxin and prevented the increase in surface ConA staining (Fig. 3C). Likewise, we transfected PPC-1 cells with siRNA against the human Na+/K+-ATPase and knockdown of the Na+/K+-ATPase was confirmed by Q-RTPCR (data not shown). Compared with cells treated with control siRNA, PPC-1 cells transfected with siRNA against the Na+/K+-ATPase displayed increased ConA-binding proteins on the cell surface (Fig. 3D). Likewise, knockdown of the Na+/K+-ATPase with siRNA in HeLa cells showed a similar effect on ConA surface–binding proteins (data not shown). Thus, taken together, these results show that inhibition of the Na+/K+-ATPase has a significant modifying effect on the N-glycosylation pathway.

MALDI-TOF mass spectrometry shows an alteration of the N-glycosylation pathway down stream of α-mannosidase II. To identify the site in the N-glycan pathway blocked by Na+/K+-ATPase inhibitors, Jurkat cells were treated with digoxin, the known Golgi α-mannosidase II inhibitor swainsonine, or buffer control. After treatment, the expression profile of intracellular oligosaccharides was measured by MALDI-TOF mass spectrometry (Table 1). Swainsonine increased the abundance of M9Gn2, M8Gn2, M7Gn2, M6Gn2, and M5Gn2, which is consistent with a block at Golgi α-mannosidase II and accumulation of these precursors. In contrast, treatment with digoxin decreased the abundance of these oligosaccharides but increased the levels of Gn1M3Gn2F1 hybrid and Gn2M3Gn2F1 biantennary complex N-glycans. Slight increases in Gn3M3Gn2F1 triantennary and Gn4M3Gn2F1 tetraantennary complex N-glycans were also seen. These results suggest that inhibition of the Na+/K+-ATPase by digoxin impairs the N-glycan pathway, thereby promoting the accumulation of hybrid and biantennary complex N-glycans. Digoxin therefore inhibits the N-glycan remodeling through a mechanism distinct from swainsonine.

Table 1.

Digoxin and swainsonine alter the expression of oligosaccharides

Oligosaccharides before α-mannosidase II
m/zStructureCtrlDigoxinSWConA binding
1,257 M5Gn2 37.39 28.88 67.63 ++ 
1,403 M5Gn2F1 1.35 49.71 ++ 
1,419 M6Gn2 25.22 15.06 36.41 ++ 
1,581 M7Gn2 16.21 11.11 26.01 ++ 
1,606 Gn1M5Gn2F1 1.35 0.987 21.96 ++ 
1,622 Hx1Gn1M5Gn2 1.8 1.4 4.62 ++ 
1,743 M8Gn2 20.72 10.86 26.01 ++ 
1,905 M9Gn2 13.96 8.64 16.76 ++ 
      
Oligosaccharides after α-mannosidase II—hybrid structures
 
     
m/z
 
Structure
 
Digoxin
 
Ctrl
 
SW
 
ConA binding
 
1,136 GnM3Gn2 2.25 4.19 2.31 
1,282 GnM3Gn2F1 8.11 17.78 6.94 
      
Oligosaccharides after α-mannosidase II—complex structures
 
     
m/z
 
Structure
 
Ctrl
 
Digoxin
 
SW
 
ConA binding
 
1,485 Gn2M3Gn2F1 4.5 29.87 5.20 ++ 
1,645 Hx1Gn2M3Gn2F1 ++ 
1,689 Gn3M3Gn2F1 1.8 5.43 2.89 − 
1,810 Hx2Gn2M3Gn2F1 1.8 1.48 2.89 − 
1,892 Gn4M3Gn2F1 2.47 2.31 − 
Oligosaccharides before α-mannosidase II
m/zStructureCtrlDigoxinSWConA binding
1,257 M5Gn2 37.39 28.88 67.63 ++ 
1,403 M5Gn2F1 1.35 49.71 ++ 
1,419 M6Gn2 25.22 15.06 36.41 ++ 
1,581 M7Gn2 16.21 11.11 26.01 ++ 
1,606 Gn1M5Gn2F1 1.35 0.987 21.96 ++ 
1,622 Hx1Gn1M5Gn2 1.8 1.4 4.62 ++ 
1,743 M8Gn2 20.72 10.86 26.01 ++ 
1,905 M9Gn2 13.96 8.64 16.76 ++ 
      
Oligosaccharides after α-mannosidase II—hybrid structures
 
     
m/z
 
Structure
 
Digoxin
 
Ctrl
 
SW
 
ConA binding
 
1,136 GnM3Gn2 2.25 4.19 2.31 
1,282 GnM3Gn2F1 8.11 17.78 6.94 
      
Oligosaccharides after α-mannosidase II—complex structures
 
     
m/z
 
Structure
 
Ctrl
 
Digoxin
 
SW
 
ConA binding
 
1,485 Gn2M3Gn2F1 4.5 29.87 5.20 ++ 
1,645 Hx1Gn2M3Gn2F1 ++ 
1,689 Gn3M3Gn2F1 1.8 5.43 2.89 − 
1,810 Hx2Gn2M3Gn2F1 1.8 1.48 2.89 − 
1,892 Gn4M3Gn2F1 2.47 2.31 − 

NOTE: Jurkat cells (1 × 107) were treated with 100 nmol/L digoxin, 2 μmol/L Swainsonine, or buffer control for 24 h. After incubation, oligosaccharides were isolated and analyzed by MALDI-TOF mass spectrometry. The ability of glycans to bind to ConA is based on previous studies using standard N-glycans and ConA-Sepharose columns (38, 39). The data represent the normalized intensity of each oligosaccharide relative to (Hexose)2(HexNAc)2(Deoxyhexose)1. Previous work from several groups (e.g., ref. 40) with mass spectrometry of N-glycans has shown that the m/z values in this table correspond to structures in which Hexose is D-Mannose (M) or unidentified (probably d-Galactose), HexNAc is N-Acetyl-d-glucosamine (Gn), and Deoxyhexose is l-Fucose.

Digoxin decreases migration and invasion of malignant cells. Increased GlcNAc branching of N-glycans on the cell surface promotes the malignant and metastatic potential of cells by altering cell migration and invasion (1, 28). As treatment with digoxin altered the N-glycan remodeling, we evaluated the effects of this compound on the N-glycosylation–mediated processes of cell migration and invasion. To determine the effects on cell migration, we evaluated the effect of digoxin in a scratch wound–healing assay. HT1080 human fibrosarcoma cells were treated with digoxin or buffer control and seeded in fibronectin-coated 4-well chamber slides. After adhering overnight, the cell monolayer was scratched to create a wound. Migration of cells to heal the wound was measured over time. Treatment with digoxin impaired cell migration and delayed wound healing (Fig. 4A). Of note, at the concentrations tested in these assays, digoxin-treated cells were >90% viable as measured by the MTS assay. Furthermore, the treated cells migrated sufficiently to completely heal the wound by 24 hours. Thus, the effects of digoxin on cell migration cannot be attributed to simple reductions in cell viability.

Figure 4.

Digoxin inhibits cell migration and invasion. A, HT1080 cells (1 × 106) were seeded in 10-cm tissue culture plates. The next day, cells were treated with digoxin 100 nmol/L (squares) or buffer control (Ctrl; circles) for 24 h and then seeded (1 × 105) on fibronectin-coated 4-well chamber slides. After adhering overnight, the monolayer was scratched with a plastic pipette tip. Migration was captured overtime; columns, mean distance of migrated cells in relative units (RU); bars, SD. *, P = 0.03 by the Student's t test. B, WRO cells (1 × 106) were treated with digoxin (50 nmol/L) for 24 h and then plated into invasion chambers with 8-μm pores. Five percent FBS was used as chemoattractant. After 24 h, cells that had migrated through the pores were fixed and stained. The number of cells that migrated through the pores was counted automatically as described in the Materials and Methods. Columns, mean of migrating cells compared with buffer treated cells; bars, SD. C, WRO cells (1 × 106) were treated with digoxin (50 nmol/L) for 24 h and then seeded into invasion chambers with Matrigel. Five percent FBS was used as chemoattractant at the bottom of the well. After 24 h, the cells that had invaded through the pores were fixed and stained. The number of cells that migrated through the pores was counted automatically as described in the Materials and Methods. Columns, mean of migrating cells compared with buffer treated cells; bars, SD.

Figure 4.

Digoxin inhibits cell migration and invasion. A, HT1080 cells (1 × 106) were seeded in 10-cm tissue culture plates. The next day, cells were treated with digoxin 100 nmol/L (squares) or buffer control (Ctrl; circles) for 24 h and then seeded (1 × 105) on fibronectin-coated 4-well chamber slides. After adhering overnight, the monolayer was scratched with a plastic pipette tip. Migration was captured overtime; columns, mean distance of migrated cells in relative units (RU); bars, SD. *, P = 0.03 by the Student's t test. B, WRO cells (1 × 106) were treated with digoxin (50 nmol/L) for 24 h and then plated into invasion chambers with 8-μm pores. Five percent FBS was used as chemoattractant. After 24 h, cells that had migrated through the pores were fixed and stained. The number of cells that migrated through the pores was counted automatically as described in the Materials and Methods. Columns, mean of migrating cells compared with buffer treated cells; bars, SD. C, WRO cells (1 × 106) were treated with digoxin (50 nmol/L) for 24 h and then seeded into invasion chambers with Matrigel. Five percent FBS was used as chemoattractant at the bottom of the well. After 24 h, the cells that had invaded through the pores were fixed and stained. The number of cells that migrated through the pores was counted automatically as described in the Materials and Methods. Columns, mean of migrating cells compared with buffer treated cells; bars, SD.

Close modal

To further evaluate the effects of digoxin on cell migration, WRO human thyroid carcinoma cells were treated with digoxin, and migration through uncoated invasion chambers with an 8-μm pore size was measured (Fig. 4B). Again, digoxin inhibited cell migration but did not reduce cell viability as measured by MTS assay.

Finally, we assessed the effects of digoxin on cell invasion (Fig. 4C). WRO cells were treated with digoxin and seeded into Matrigel-containing invasion chambers in serum-free medium. Medium with 5% FBS was placed in the lower chamber as a chemoattractant. Twenty four hours after seeding, cell invasion through the Matrigel was measured. Digoxin treatment decreased cell invasion but did not reduce cell viability as measured by the MTS assay.

Thus, taken together, concentrations of digoxin that blocked N-glycan remodeling also inhibited the N-glycosylation–mediated processes dependent on GlcNAc-branched N-glycans: cell migration and invasion. These results suggest that the effects of digoxin on N-glycan expression are functionally important.

Digoxin decreases distant tumor formation in vivo. As aberrant GlcNAc branching of N-glycans on the cell surface promotes metastases, we tested the effects of digoxin on distant tumor formation in mouse models of metastatic prostate cancer. In the first model, dsRed-labeled PPC-1 cells were treated with digoxin (100 nmol/L), or buffer control in culture. After 20 hours of treatment, cells were injected i.v. into sublethally irradiated SCID mice. Three weeks after injection, mice were sacrificed and distant tumor formation in the organs was imaged with fluorescent microscopy.

Invasion of the prostate cancer cells was detected in the lung, bone, and liver, clinically relevant sites of metastases in prostate cancer. In particular, metastasis of dsRed-PPC-1 cells to the lung was readily quantifiable using image-based analysis (10, 17, 29, 30). Compared with buffer control, mice injected with digoxin-treated cells had decreased mean tumor number (142 ± 71 tumors versus 20 ± 19 tumors; P = 0.0004 by Student's t test) within the lung (Fig. 5A). Median tumor number was also significantly attenuated by drug treatment (data not shown). It is important to note that both treated and control cells were >90% viable at the time of injection. To determine whether the decreased distant tumor formation was simply due to decreased proliferation, DsRed-PPC-1 cells were treated in culture with digoxin or buffer control and injected s.c. into mice. In contrast to its effects on distant tumor formation, digoxin did not significantly alter the growth of dsRedPPC-1 cells injected s.c. compared with cells treated with buffer control (227 ± 50 mg versus 331 ± 81 mg; P = 0.24 by Student's t test; Fig. 5B). Therefore, the reduction in tumor burden after digoxin treatment is not solely due to reductions in cellular proliferation.

Figure 5.

Digoxin inhibits distant tumor formation in vivo. A, fluorescent dsRed-PPC-1 cells were treated with digoxin (100 nmol/L; n = 8) or buffer control (n = 7). Twenty hours after treatment, cells were harvested and 3.5 × 106 viable cells were injected into the tail veins of sublethally irradiated SCID mice. Three weeks after injection, mice were sacrificed and their organs were imaged using a fluorescent microscope. The numbers of distant tumors from all five lobes of the lungs were quantified using image analysis software. Columns, mean number of tumors; bars, SD. P values are based on the Student's t test. B, fluorescent dsRed-PPC-1 cells were treated with digoxin (100 nmol/L; n = 6) or buffer control (n = 5). Twenty hours after treatment, cells were harvested and injected s.c. into sublethally irradiated SCID mice. Two weeks after injection, mice were sacrificed, and the tumors were excised and weighed. Columns, mean tumor weight; bars, SD. P values are based on the Student's t test. C, fluorescent dsRed-PPC-1 cells (3.5 × 106 viable cells) were injected into the tail veins of sublethally irradiated SCID mice. Daily for 14 d, the mice were treated with digoxin [0.675 mg/kg/d (n = 9) or 1.35 mg/k/d (n = 9)] or buffer control (n = 5). Four weeks after injection, mice were sacrificed and their organs were imaged using a fluorescent microscope. The number of distant tumors from all five lobes of the lungs were quantified using image analysis software. Columns, mean number or area of tumors; bars, SD. P values are based on the Student's t test.

Figure 5.

Digoxin inhibits distant tumor formation in vivo. A, fluorescent dsRed-PPC-1 cells were treated with digoxin (100 nmol/L; n = 8) or buffer control (n = 7). Twenty hours after treatment, cells were harvested and 3.5 × 106 viable cells were injected into the tail veins of sublethally irradiated SCID mice. Three weeks after injection, mice were sacrificed and their organs were imaged using a fluorescent microscope. The numbers of distant tumors from all five lobes of the lungs were quantified using image analysis software. Columns, mean number of tumors; bars, SD. P values are based on the Student's t test. B, fluorescent dsRed-PPC-1 cells were treated with digoxin (100 nmol/L; n = 6) or buffer control (n = 5). Twenty hours after treatment, cells were harvested and injected s.c. into sublethally irradiated SCID mice. Two weeks after injection, mice were sacrificed, and the tumors were excised and weighed. Columns, mean tumor weight; bars, SD. P values are based on the Student's t test. C, fluorescent dsRed-PPC-1 cells (3.5 × 106 viable cells) were injected into the tail veins of sublethally irradiated SCID mice. Daily for 14 d, the mice were treated with digoxin [0.675 mg/kg/d (n = 9) or 1.35 mg/k/d (n = 9)] or buffer control (n = 5). Four weeks after injection, mice were sacrificed and their organs were imaged using a fluorescent microscope. The number of distant tumors from all five lobes of the lungs were quantified using image analysis software. Columns, mean number or area of tumors; bars, SD. P values are based on the Student's t test.

Close modal

To further evaluate the effects of digoxin on distant tumor formation, sublethaly irradiated SCID mice were injected i.v. with dsRed-PPC-1 cells and then treated i.p. with digoxin (1.35 and 0.675 mg/kg) or buffer control daily for 14 days. Four weeks after the injection of the cells, mice were sacrificed and analyzed as above. Compared with buffer control, both concentrations of digoxin decreased the mean number of tumors [48 ± 32 tumors (0.65 mg/kg) versus 28 ± 19 tumors (1.35 mg/kg) versus 171 ± 83 tumors (control)] in the lungs (Fig. 5C). Similar reductions in median tumor number were also observed (data not shown). Thus, taken together, digoxin inhibits distant tumor formation in vivo.

To better understand the effect of modulating N-glycosylation in malignant cells, we used a chemical biology screen for inhibitors of L-PHA–induced cell death to identify novel regulators of glycosylation. From this screen, we identified the CG dihydroouabain and determined that structurally related CGs also blocked L-PHA–induced cell death. Of the tested CGs, digoxin was the most potent inhibitor. These results, coupled with the lack of effect on murine cells suggest that CGs alter N-glycosylation by inhibiting their known target, the Na+/K+-ATPase. Further supporting a mechanism linked to the known binding target, overexpression of the murine Na+/K+-ATPase that is less sensitive to CG inhibition abrogated the effects of digoxin on surface ConA-binding N-glycoproteins. Moreover, knockdown of the Na+/K+-ATPase using siRNA recapitulated the effects of digoxin. Thus, taken together, we conclude that inhibiting the Na+/K+-ATPase alters the N-glycosylation pathway.

In this report, we used four independent approaches to show the ability of CGs to modify the N-glycosylation pathway. First, L-PHA induces cell death by binding primarily to cell surface N-glycans that occur downstream of the GnTII in the N-glycan synthesis pathway, and CGs blocked L-PHA–induced cell death. Second, CGs increased total cellular ConA-binding glycoproteins. Third, CGs increased cell surface ConA–binding N-glycoproteins. Finally, mass spectrometry revealed that digoxin increased the ConA-binding GnM3Gn2F1 hybrid and Gn2M3Gn2F1 biantennary N-glycans.

The profile of oligosaccharides after digoxin treatment indicates that digoxin alters the N-glycosylation. These results also indicate that digoxin alters the pathway through a mechanism distinct from the α-mannosidase II inhibitor swainsonine.

In this report, we showed that digoxin inhibited cellular migration and invasion, which are known functional consequence of blocking the N-glycan branching. For example, Mgat5−/− mice have reduced cancer growth and metastasis (31). Likewise, siRNA knockdown of GlcNAc-TV in malignant cells impairs cell migration and invasion (28). Although digoxin's inhibition of cell migration and invasion are consistent with its effects on the N-glycan remodeling, we cannot exclude that these effects are related to other pathways effected by the molecule.

CGs can induce cell death in malignant cells (23, 24) through multiple mechanisms including activation of Cdk5 (32), Src kinase (33), and p21 (33). However, the effects of CG on N-glycan pathway and cellular processes of migration and invasion were not artifacts of cell death, as the concentrations of digoxin and times of incubation required to alter N-glycan remodeling were lower than those required to induce cell death. Furthermore, as the concentrations of CG required to alter the N-glycan pathway are lower than the concentrations associated with activation of these other pathways, we suspect that the effects of CG on N-glycosylation are not related to activation of these other pathways.

Serum concentrations of 2 nmol/L digoxin can be achieved in humans without significant toxicity (34). Although the concentration of digoxin required to inhibit N-glycosylation exceeded 2 nmol/L, potentially, chronic daily dosing of digoxin could sufficiently reduce N-glycan branching to affect the function of downstream effectors. Interestingly, patients with breast carcinoma who were coincidentally receiving CGs for cardiac dysfunction had a lower rate of relapse and metastasis than patients not receiving CGs (20, 21). Our results suggest that CGs could have an antitumor effect in patients by altering the N-glycosylation profile in malignant cells. To assess the effects of digoxin on distant tumor formation and mimic some of the processes of metastasis, we tested digoxin in two mouse models. In both of these models, digoxin decreased distant tumor formation. Thus, the decreased metastases supports a mechanism of action linked to the inhibition of glycosylation. Thus, digoxin could be a lead for a novel therapeutic agent for the treatment of malignancy and an adjunct to prevent metastases. A limitation to the xenograft studies, however, is that we cannot be certain that digoxin prevented metastases through a mechanism related to ability to inhibit glycosylation. Potentially inhibition of the Na+/K+ATPase may have antitumor effects through mechanisms distinct from impairing glycosylation.

Interestingly, other glycosylation inhibitors have entered clinical trials for the treatment of malignancy. Swainsonine, a small molecule inhibitor of Golgi α-mannosidase II, has anticancer activity in preclinical models (27, 35). Given these results, swainsonine was advanced into a phase I clinical trial for patients with refractory malignancy. In the context of this trial, tumor regression was noted in one patient with head and neck cancer (36, 37). Thus, inhibiting α-mannosidase II may be a clinically effective anticancer strategy and inhibiting targets downstream of α-mannosidase II might also produce antitumor effects without untoward toxicity.

In summary, we used a chemical biology approach to investigate the Golgi N-glycan pathway. Our high-throughput screen identified Na+/K+-ATPase inhibitors that altered the N-glycan branching at a point distinct from the α-mannosidase II inhibitor. Na+/K+-ATPase inhibitors also blocked the N-glycan–dependent processes of cell migration and invasion of cancer cells as well as distant tumor formation in mouse models. Thus, our results help explain previously reported anticancer effects of these compounds and highlight new strategies for the development of anticancer therapies.

No potential conflicts of interest were disclosed.

Grant support: Canadian Institutes of Health Research.

A.D. Schimmer is the recipient of a Canadian Institutes of Health Research Clinician Scientist Award and is a Leukemia and Lymphoma Society Scholar in Clinical Research.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Dennis JW, Granovsky M, Warren CE. Glycoprotein glycosylation and cancer progression.
Biochim Biophys Acta
1999
;
1473
:
21
–34.
2
Lau KS, Partridge EA, Grigorian A, et al. Complex N-glycan number and degree of branching cooperate to regulate cell proliferation and differentiation.
Cell
2007
;
129
:
123
–34.
3
Lagana A, Goetz JG, Cheung P, Raz A, Dennis JW, Nabi IR. Galectin binding to Mgat5-modified N-glycans regulates fibronectin matrix remodeling in tumor cells.
Mol Cell Biol
2006
;
26
:
3181
–93.
4
Lajoie P, Partridge EA, Guay G, et al. Plasma membrane domain organization regulates EGFR signaling in tumor cells.
J Cell Biol
2007
;
179
:
341
–56.
5
Goss PE, Baker MA, Carver JP, Dennis JW. Inhibitors of carbohydrate processing: A new class of anticancer agents.
Clin Cancer Res
1995
;
1
:
935
–44.
6
Hakomori S. Tumor malignancy defined by aberrant glycosylation and sphingo(glyco)lipid metabolism.
Cancer Res
1996
;
56
:
5309
–18.
7
Kornfeld R, Kornfeld S. Assembly of asparagine-linked oligosaccharides.
Annu Rev Biochem
1985
;
54
:
631
–64.
8
Schachter H. Biosynthetic controls that determine the branching and microheterogeneity of protein-bound oligosaccharides.
Adv Exp Med Biol
1986
;
205
:
53
–85.
9
Datti A, Donovan RS, Korczak B, Dennis JW. A homogeneous cell-based assay to identify N-linked carbohydrate processing inhibitors.
Anal Biochem
2000
;
280
:
137
–42.
10
Mawji IA, Simpson CD, Gronda M, et al. A chemical screen identifies anisomycin as an anoikis sensitizer that functions by decreasing FLIP protein synthesis.
Cancer Res
2007
;
67
:
8307
–15.
11
Storey JD. A direct approach to false discovery rates.
J R Stat Soc B
2002
;
64
:
479
–98.
12
Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F. Colorimetric method for determination of sugars and related substances.
Anal Chem
1956
;
28
:
350
–6.
13
Schimmer AD, Thomas MP, Hurren R, et al. Identification of small molecules that sensitize resistant tumor cells to tumor necrosis factor-family death receptors.
Cancer Res
2006
;
66
:
2367
–75.
14
Sarkar M, Leventis PA, Silvescu CI, Reinhold VN, Schachter H, Boulianne GL. Null mutations in Drosophila N-acetylglucosaminyltransferase I produce defects in locomotion and a reduced life span.
J Biol Chem
2006
;
281
:
12776
–85.
15
Liu W, Wei W, Winer D, et al. CEACAM1 impedes thyroid cancer growth but promotes invasiveness: a putative mechanism for early metastases.
Oncogene
2007
;
26
:
2747
–58.
16
Demetriou M, Nabi IR, Coppolino M, Dedhar S, Dennis JW. Reduced contact-inhibition and substratum adhesion in epithelial cells expressing GlcNAc-transferase V.
J Cell Biol
1995
;
130
:
383
–92.
17
Cairns RA, Hill RP. Acute hypoxia enhances spontaneous lymph node metastasis in an orthotopic murine model of human cervical carcinoma.
Cancer Res
2004
;
64
:
2054
–61.
18
Green ED, Baenziger JU. Oligosaccharide specificities of Phaseolus vulgaris leukoagglutinating and erythroagglutinating phytohemagglutinins. Interactions with N-glycanase-released oligosaccharides.
J Biol Chem
1987
;
262
:
12018
–29.
19
Doherty JE. Clinical use of digitalis glycosides. An update.
Cardiology
1985
;
72
:
225
–54.
20
Stenkvist B, Bengtsson E, Eklund G, et al. Evidence of a modifying influence of heart glucosides on the development of breast cancer.
Anal Quant Cytol
1980
;
2
:
49
–54.
21
Stenkvist B, Pengtsson E, Dahlqvist B, Eriksson O, Jarkrans T, Nordin B. Cardiac glycosides and breast cancer, revisited.
N Engl J Med
1982
;
306
:
484
.
22
Brown L, Erdmann E, Thomas R. Digitalis structure-activity relationship analyses. Conclusions from indirect binding studies with cardiac (Na+ + K+)-ATPase.
Biochem Pharmacol
1983
;
32
:
2767
–74.
23
Johansson S, Lindholm P, Gullbo J, Larsson R, Bohlin L, Claeson P. Cytotoxicity of digitoxin and related cardiac glycosides in human tumor cells.
Anticancer Drugs
2001
;
12
:
475
–83.
24
Raghavendra PB, Sreenivasan Y, Ramesh GT, Manna SK. Cardiac glycoside induces cell death via FasL by activating calcineurin and NF-AT, but apoptosis initially proceeds through activation of caspases.
Apoptosis
2007
;
12
:
307
–18.
25
Thomas R, Gray P, Andrews J. Digitalis: its mode of action, receptor, and structure-activity relationships.
Adv Drug Deliv Rev
1990
;
19
:
313
–550.
26
Price EM, Lingrel JB. Structure-function relationships in the Na K-ATPase α subunit: site-directed mutagenesis of glutamine-111 to arginine and asparagine-122 to aspartic acid generates a ouabain-resistant enzyme.
Biochemistry
1988
;
27
:
8400
–8.
27
Dennis JW. Effects of swainsonine and polyinosinic:polycytidylic acid on murine tumor cell growth and metastasis.
Cancer Res
1986
;
46
:
5131
–6.
28
Zhang Y, Zhao JH, Zhang XY, Guo HB, Liu F, Chen HL. Relations of the type and branch of surface N-glycans to cell adhesion, migration and integrin expressions.
Mol Cell Biochem
2004
;
260
:
137
–46.
29
Mawji IA, Simpson CD, Hurren R, et al. Critical role for Fas-associated death domain-like interleukin-1-converting enzyme-like inhibitory protein in anoikis resistance and distant tumor formation.
J Natl Cancer Inst
2007
;
99
:
811
–22.
30
Hoffman RM. The multiple uses of fluorescent proteins to visualize cancer in vivo.
Nat Rev Cancer
2005
;
5
:
796
–806.
31
Granovsky M, Fata J, Pawling J, Muller WJ, Khokha R, Dennis JW. Suppression of tumor growth and metastasis in Mgat5-deficient mice.
Nat Med
2000
;
6
:
306
–12.
32
Lin H, Juang JL, Wang PS. Involvement of Cdk5/p25 in digoxin-triggered prostate cancer cell apoptosis.
J Biol Chem
2004
;
279
:
29302
–7.
33
Kometiani P, Liu L, Askari A. Digitalis-induced signaling by Na+/K+-ATPase in human breast cancer cells.
Mol Pharmacol
2005
;
67
:
929
–36.
34
The effect of digoxin on mortality and morbidity in patients with heart failure. The Digitalis Investigation Group.
N Engl J Med
1997
;
336
:
525
–33.
35
Dennis JW, Koch K, Yousefi S, VanderElst I. Growth inhibition of human melanoma tumor xenografts in athymic nude mice by swainsonine.
Cancer Res
1990
;
50
:
1867
–72.
36
Goss PE, Baptiste J, Fernandes B, Baker M, Dennis JW. A phase I study of swainsonine in patients with advanced malignancies.
Cancer Res
1994
;
54
:
1450
–7.
37
Goss PE, Reid CL, Bailey D, Dennis JW. Phase IB clinical trial of the oligosaccharide processing inhibitor swainsonine in patients with advanced malignancies.
Clin Cancer Res
1997
;
3
:
1077
–86.
38
Narasimhan S, Wilson JR, Martin E, Schachter H. A structural basis for four distinct elution profiles on concanavalin A-Sepharose affinity chromatography of glycopeptides.
Can J Biochem
1979
;
57
:
83
–96.
39
Cummings RD. Use of lectins in analysis of glycoconjugates.
Methods Enzymol
1994
;
230
:
66
–86.
40
Dell A, Reason AJ, Khoo KH, Panico M, McDowell RA, Morris HR. Mass spectrometry of carbohydrate-containing biopolymers.
Methods Enzymol
1994
;
230
:
108
–32.