Abstract
Ovarian cancer (OVCA) is the most lethal gynecological malignancy. The high mortality rate associated with this disease is due in large part to the development of resistance to chemotherapy; however, the biological basis of this remains unclear. Gemcitabine is frequently used for the treatment of patients with platinum-resistant OVCA. We report molecular signaling pathways associated with OVCA response to gemcitabine. Forty-one OVCA cell lines were subjected to gene expression analysis; in parallel, IC50 values for gemcitabine were quantified using CellTiter-Blue viability assays. Pearson’s correlation coefficients were calculated for gene expression and gemcitabine IC50 values. The genes associated with gemcitabine sensitivity were subjected to pathway analysis. For the identified pathways, principal component analysis was used to derive pathway signatures and corresponding scores, which represent overall measures of pathway expression. Expression levels of the identified pathways were then evaluated in a series of clinico-genomic datasets from 142 patients with stage III/IV serous OVCA. We found that in vitro gemcitabine sensitivity was associated with expression of 131 genes (p<0.001). These genes include significant representation of three molecular signaling pathways (p<0.02): O-glycan biosynthesis, Cell cycle_Role of Nek in cell cycle regulation and Immune response_Antiviral actions of Interferons. In an external clinico-genomic OVCA dataset (n=142), expression of the O-glycan pathway was associated with overall survival, independent of surgical cytoreductive status, grade and age (p<0.001). Expression levels of Cell cycle_Role of Nek in cell cycle regulation and Immune response_Antiviral actions of Interferons were not associated with survival (p=0.3107 and p=0.5411, respectively). Collectively, expression of the O-glycan biosynthesis pathway, which modifies protein function via post-translational carbohydrate binding, is independently associated with overall survival from OVCA. Our findings shed light on the molecular basis of OVCA responsiveness to gemcitabine and also identify a signaling pathway that may influence patient survival.
Keywords: gemcitabine chemosensitivity, genomic study, O-glycan pathway, principal component analysis, ovarian cancer survival
Introduction
Ovarian cancer (OVCA) is the leading cause of gynecologic cancer mortality and the sixth most common cancer diagnosed in women in the United States. Advanced-stage epithelial OVCA is highly heterogeneous at a clinical, biologic, and genetic level, but patients are currently treated in a uniform fashion with cytoreductive surgery and platinum/taxane-based combination chemotherapy. Unfortunately, most patients ultimately succumb to persistent or recurrent platinum-resistant disease (1,2). Currently, efforts to develop therapeutic agents with greater efficacy against platinum-resistant disease are limited because of incomplete understanding of the molecular determinants of OVCA drug response.
Gemcitabine (2′,2′-difluorodeoxycytidine), a synthetic nucleoside analog of cytidine, is frequently used as a second-line therapy for patients with relapsed OVCA (3). As a pyrimidine analogue, gemcitabine replaces the nucleic acid cytidine during DNA replication, blocking processing and chain elongation by the DNA polymerase complex, resulting in G1 arrest and a subsequent cytostatic effect. Additionally, the gemcitabine triphosphate metabolite is incorporated into RNA, thus inhibiting RNA synthesis (4). Gemcitabine efficacy has been evaluated extensively both in vitro and in vivo against OVCA (5–8). Gemcitabine has demonstrated single-agent activity against OVCA cell lines (9) and synergistic activity with several other antineoplastic agents, including platinum compounds, topotecan, and etoposide (10). In animal tumor models, the gemcitabine effect has been shown to be schedule dependent, and continuous infusions over 24 h appear to enhance gemcitabine cytotoxicity (11). Phase II and III studies of gemcitabine (800–1250 mg/m2/week) in patients with recurrent OVCA have demonstrated response rates up to 19% (12–14). Despite such data, the molecular determinants of gemcitabine activity remain to be fully elucidated. In this study, we sought to determine the molecular underpinnings of OVCA response to gemcitabine at a genome-wide level. We investigated the genes and molecular signaling pathways associated with the response of OVCA cells in vitro to gemcitabine and explored how these pathways influence in vivo clinical outcomes for patients with this disease.
Materials and methods
Overview
We subjected 41 OVCA cell lines to gene expression analysis and, in parallel, measured gemcitabine sensitivity (IC50). Genes associated with baseline gemcitabine sensitivity, identified by Pearson’s correlation analysis, were subjected to molecular pathway analysis. We evaluated expression of identified pathways using a series of clinico-genomic datasets from 142 patients with stage III/IV serous OVCA. All 142 patients had signed the IRB-approved, written informed consent forms.
Cell culture
OVCA cell lines were obtained from the American Type Culture Collection (Manassas, VA; CAOV3, OV90, OVCAR3, SKOV3), from the European Collection of Cell Cultures (Salisbury, UK; A2780CP, A2780S), from Kyoto University (Kyoto, Japan; CHI, CHIcisR, M41, M41CSR, Tyknu, and TyknuCisR), or as kind gifts from Dr Patricia Kruk, Department of Pathology, College of Medicine, University of South Florida, Tampa, FL, and Susan Murphy, PhD, Department of OBGYN/Division of GYN Oncology, Duke University, Durham, NC (A2008, C13, CAOV2, HeyA8, IGR-OV1, IMCC3, IMCC5, MCAS, OV2008, OVCA420, OVCA429, OVCA432, OVCA433, FUOV1, PEO1, PEO4, SK-OV-6, T8, TOV-112D, TOV-21-G, Dov13, BG1, Ovary1847, OVCAR10, OVCAR8, OVCAR5, OVCAR4, OVCAR2, SK-OV-4). Cell lines were maintained in RPMI-1640 medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (Fisher Scientific, Pittsburgh, PA), 1% sodium pyruvate, 1% penicillin/streptomycin (Cellgro, Manassas, VA), and 1% non-essential amino acids (HyClone, Hudson, NH). Mycoplasma testing was performed every 6 months, in accordance with the manufacturer’s protocol (Lonza, Rockland, ME).
RNA extraction and microarray expression analysis
RNA from 41 OVCA cell lines was extracted using the RNeasy kit following manufacturer’s recommendations (Qiagen, Valencia, CA). Quality of the RNA was measured using an Agilent 2100 Bioanalyzer. The targets for Affymetrix DNA microarray analysis were prepared according to the manufacturer’s instructions, and targets were hybridized to customized Human Affymetrix HuRSTA gene chips (HuRSTA-2a520709), which include 60,607 probe sets and representation of 19,308 genes (Gene Expression Omnibus accession number GSE34615).
CellTiter-Blue cell viability assays
Drug activity was evaluated using a high-throughput CellTiter-Blue cell viability assay. Cells (2.5×103 per well) were plated in 384-well plates using complete media with 10% fetal bovine serum and allowed to adhere overnight. After cell adherence, increasing concentrations of gemcitabine were added to appropriate wells using an automated pipetting station. Four replicate wells were used for each drug concentration and for vehicle controls. Drug dilutions initially consisted of 1.5-fold serial dilutions from a maximum concentration of 100 μM. The cells were incubated with the drug for 72 h, and 5 μl of CellTiter-Blue reagent (Promega Corp.) were added to each well. Fluorescence was read at 579-nm excitation/584-nm emission using a Synergy 4 microplate reader (Bio-Tek Instruments, Inc., Winooski, VT). IC50 values were determined using a sigmoidal equilibrium model fit (XLfit 5.2, ID Business Solutions Ltd.). The IC50 was defined as the concentration of drug required for a 50% reduction in growth/viability.
Statistical analysis
Expression data from 41 OVCA cell lines were subjected to background correction and normalization using the ‘Robust Multichip Average’ algorithm in the Affymetrix Expression Console (http://www.affymetrix.com/estore/index.jsp). Pearson’s correlation test was performed on individual gene expression and IC50 values. Probe sets with p<0.001 were considered to have significant correlations with IC50 values and were uploaded to MetaCore GeneGo for pathway analysis (http://www.genego.com/metacore.php). Pathways with p<0.05 were considered significant, based upon the GeneGo/MetaCore™ statistical test for significance.
Building signatures of pathway activity
The principal component analysis (PCA) methodology was used to derive a gene expression signature for each pathway. A corresponding ‘pathway score’ was thus generated that quantifies the overall level of pathway gene expression in a summary value. That is, the PCA score is a numeric value that summarizes the level of expression of the entire pathway. First, data were reduced into a small set of uncorrelated principal components. This set of principal components was generated based on its ability to account for variation. The first PCA was used to represent the overall expression level for the pathway as it accounts for the largest variability in the data. That is, the pathway score is equal to Σwixi, a weighted average expression among the pathway genes, where xi represents gene i expression level, wi is the corresponding weight (loading coefficient) with Σw2i=1, and the wi values maximize the variance of Σwixi. Details of this methodology have been previously reported by our group (15,16).
Validation of signatures in primary OVCA datasets
The pathway gene expression signature scores were evaluated in an independent publicly available clinico-genomic dataset from 142 OVCA samples (16). In brief, all 142 samples were known to have been resected from patients with advanced-stage (III/IV), serous epithelial OVCA, who underwent primary cytoreductive surgery followed by primary therapy with a platinum-based regimen (+/− taxane or cyclophosphamide). Response to this primary therapy [complete response (CR) versus incomplete response (IR)] has previously been described for these patients (16). In brief, patients who demonstrated a CR had no evidence of disease on physical examination, serum tumor marker monitoring, or radiographic imaging. The IR category included all other patients. Log-rank tests with Kaplan-Meier survival curves were used to test any association between the pathway scores (‘high’ versus ‘low’ based on a median value cut-off) and overall survival for patients with OVCA.
Results
Forty-one OVCA cell lines were treated with increasing concentrations of gemcitabine, and the IC50 values were determined (Table I)
Table I.
Gemcitabine IC50.
| Cell line | IC50 (mean) | IC50 (SD) | No. |
|---|---|---|---|
| A2008 | 163.9E-9 | 309.0E-9 | 12 |
| A2780CP | 366.4E-9 | 696.6E-9 | 9 |
| A2780S | 51.8E-9 | 46.5E-9 | 4 |
| BG1 | 30.4E-6 | 27.6E-6 | 8 |
| C13 | 418.1E-9 | 804.0E-9 | 9 |
| CAOV2 | 3.9E-6 | 8.5E-6 | 12 |
| CAOV3 | 2.2E-9 | 1.2E-9 | 5 |
| CHI | 268.6E-9 | 522.6E-9 | 26 |
| CHIcisR | 23.7E-9 | 58.8E-9 | 13 |
| Dov13 | 6.0E-9 | 3.5E-9 | 4 |
| FUOV1 | 59.4E-6 | 6.9E-6 | 3 |
| HeyA8 | 1.5E-6 | 2.4E-6 | 6 |
| IGR-OV1 | 531.2E-9 | 1.3E-6 | 14 |
| IMCC3 | 942.6E-9 | 1.1E-6 | 15 |
| IMCC5 | 105.1E-9 | 159.6E-9 | 20 |
| M41 | 39.5E-9 | 18.7E-9 | 5 |
| M41CSR | 37.4E-9 | 34.4E-9 | 12 |
| MCAS | 56.4E-6 | 99.8E-6 | 8 |
| OV2008 | 383.8E-9 | 1.1E-6 | 15 |
| OV90 | 18.9E-9 | 11.9E-9 | 9 |
| Ovary1847 | 864.6E-9 | 2.7E-6 | 19 |
| OVCA420 | 12.7E-6 | 22.0E-6 | 5 |
| OVCA429 | 22.2E-9 | 24.2E-9 | 5 |
| OVCA432 | 14.9E-6 | 25.7E-6 | 3 |
| OVCA433 | 9.9E-9 | 10.9E-9 | 5 |
| OVCAR10 | 671.9E-9 | 2.5E-6 | 16 |
| OVCAR2 | 22.0E-9 | 30.8E-9 | 17 |
| OVCAR3 | 6.2E-6 | 14.0E-6 | 14 |
| OVCAR4 | 655.3E-9 | 870.8E-9 | 5 |
| OVCAR5 | 278.3E-9 | 721.8E-9 | 17 |
| OVCAR8 | 272.6E-9 | 681.6E-9 | 12 |
| PEO1 | 134.7E-9 | 244.5E-9 | 9 |
| PEO4 | 536.2E-9 | 868.6E-9 | 10 |
| SK-OV-3 | 16.1E-6 | 30.5E-6 | 11 |
| SK-OV-4 | 3.3E-9 | 1.6E-9 | 12 |
| SK-OV-6 | 3.5E-6 | 10.4E-6 | 11 |
| T8 | 255.6E-9 | 456.6E-9 | 9 |
| TOV-112D | 44.5E-6 | 67.3E-6 | 9 |
| TOV-21G | 764.1E-9 | 1.5E-6 | 11 |
| Tyknu | 4.6E-9 | 2.8E-9 | 4 |
| TyknuCisR | 8.5E-9 | 8.4E-9 | 8 |
Pearson’s correlation test using gemcitabine IC50 and OVCA cell line gene expression data identified 131 unique genes to be associated with gemcitabine sensitivity (p<0.001; Table II). GeneGo MetaCore™ analysis identified three biological pathways that were represented within the 131 genes associated with gemcitabine sensitivity (p<0.02). These molecular signaling pathways included O-glycan biosynthesis (p=0.001), Cell cycle_Role of Nek in cell cycle regulation (p=0.005), and Immune response_Antiviral actions of Interferons (p=0.01). Statistical significance was derived from the total number of genes input into the pathway analysis software, the number of input genes present in a specific pathway, and the actual number of genes in that pathway. Thus, the p-value represents the probability that mapping a set of genes to a particular pathway occurs by chance. The O-glycan pathway demonstrated the highest level of statistical significance in its association with sensitivity to gemcitabine (p=0.001) (Fig. 1).
Table II.
Genes associated with in vitro gemcitabine chemoresistance.
| Probe set ID | Gene name | Gene description | Score | p-value |
|---|---|---|---|---|
| ENST00000376242_at | PSORS1C3 | PSORS1C3, AB023059.1 | 0.785 | 1.22E-09 |
| Ak123047_a_at | NR3C2 | NR3C2, MGC133092, MLR, MR, MCR | 0.749 | 1.72E-08 |
| ENST00000366558_a_at | KMO | KMO, dJ317G22.1 | 0.728 | 6.82E-08 |
| NM_152772_at | TCP11L2 | t-complex 11 (mouse) like 2 | 0.700 | 3.61E-07 |
| NM_003890_at | FCGBP | Human IgG Fc binding protein | 0.688 | 6.62E-07 |
| NM_021936_at | PAPPA2 | Pregnancy-associated plasma preproprotein-A2 | 0.680 | 9.89E-07 |
| NM_139173_s_at | NHEDC1 | Na+/H+ exchanger domain CG10806-like | 0.676 | 1.25E-06 |
| NM_152888_s_at | COL22A1 | Collagen, type XXII, alpha 1 | 0.656 | 3.29E-06 |
| NM_016242_at | EMCN | Endomucin, endomucin-2 | 0.654 | 3.60E-06 |
| AL133118_at | MAPKSP1 | MAPKSP1, MAPBP, MP1, MAP2K1IP1 | 0.638 | 7.26E-06 |
| NM_030923_s_at | TMEM163 | Transmembrane protein 163 | 0.636 | 8.00E-06 |
| NM_024013_at | IFNA1 | IFNA1, IFL, IFN, IFN-α, IFNA13, IFN α-D, LeIF D | 0.631 | 9.60E-06 |
| NM_199235_at | COLEC11 | Collectin sub-family member 11 | 0.626 | 1.18E-05 |
| NM_003585_at | DOC2B | Double C2-like domains, β | 0.620 | 1.55E-05 |
| NM_005472_at | KCNE3 | Cardiac voltage-gated K channel accessory | 0.618 | 1.65E-05 |
| NM_194309_at | C21orf125 | C21orf125, PRED49, FLJ38036 | 0.618 | 1.67E-05 |
| ENST00000260323_at | UNC13C | unc-13 homolog C | 0.616 | 1.82E-05 |
| ENST00000234725_at | TMEM48 | Transmembrane protein 48 | −0.612 | 2.09E-05 |
| NM_198058_at | ZNF266 | Zinc finger protein 266 | −0.603 | 3.06E-05 |
| AW510703_at | SLC15A4 | Solute carrier family 15, member 4 | 0.601 | 3.20E-05 |
| NM_020119_at | ZC3HAV1 | Zinc finger antiviral protein | −0.597 | 3.76E-05 |
| NM_019104_s_at | LIN37 | lin-37 homolog | 0.596 | 3.91E-05 |
| NM_022774_at | DEM1 | DEM1, FLJ11445, FLJ13183, FLJ21144, C1orf176 | −0.596 | 3.92E-05 |
| AA723953_at | SGCD | Sarcoglycan, delta (35 kDa dystrophin-associated glycan) | 0.591 | 4.82E-05 |
| NM_012253_s_at | TKTL1 | Transketolase-like 1 | 0.590 | 4.87E-05 |
| NM_175613_a_at | CNTN4 | Axonal cell adhesion molecule contactin 4 | 0.590 | 4.97E-05 |
| NM_006198_at | PCP4 | Purkinje cell protein 4 | 0.589 | 5.01E-05 |
| NM_012391_at | SPDEF | Human prostate specific Ets, PDEF | 0.588 | 5.25E-05 |
| AK124251_at | LHFPL3 | LHFP-like protein 3 | 0.586 | 5.65E-05 |
| AK024279_a_at | WIPI2 | WD repeat domain, phosphoinositide interacting 2 | −0.583 | 6.38E-05 |
| N25888_a_at | GDF15 | Growth differentiation factor 15 | 0.581 | 6.72E-05 |
| NM_000705_at | ATP4B | ATPase, H+/K+ transporting, beta polypeptide | 0.578 | 7.57E-05 |
| AK097996_at | GALNT2 | Polypeptide N-acetylgalactosaminyltransferase 2 | −0.577 | 7.71E-05 |
| NM_014848_at | SV2B | Synaptic vesicle protein 2B | 0.577 | 7.96E-05 |
| AL049464_at | THSD4 | Thrombospondin, type I, domain containing 4 | 0.576 | 8.03E-05 |
| BM668558_at | SART1 | Squamous cell carcinoma antigen recognized by T c | −0.576 | 8.23E-05 |
| CR606639_a_at | ZFP57 | Zinc finger protein 57 | 0.574 | 8.62E-05 |
| NM_018053_at | XKR8 | X Kell blood group precursor-related family | −0.574 | 8.73E-05 |
| NM_002239_at | KCNJ3 | Subfamily, potassium inwardly-rectifying channel J3 | 0.573 | 9.09E-05 |
| BC009808_at | NBR1 | Neighbor of BRCA1 gene 1 protein | 0.573 | 9.18E-05 |
| ENST00000360944_s_at | RBAK | RB-associated KRAB repressor | −0.572 | 9.19E-05 |
| AK023318_s_at | CBARA1 | Calcium binding atopy-related autoantigen 1 | 0.570 | 9.94E-05 |
| BQ574912_s_at | TOMM5 | TOMM5, C9orf105, RP11-263I4.1, Tom5, bA613M10.3 | −0.565 | 0.0001177 |
| ENST00000361262_at | SMC5 | Structural maintenance of chromosomes 5 | −0.563 | 0.0001256 |
| ENST00000369578_a_at | ZNF292 | Zinc-finger domain protein | 0.563 | 0.000126 |
| BC050372_a_at | OLAH | Oleoyl-ACP hydrolase | 0.563 | 0.0001284 |
| NM_172238_at | TFAP2D | Transcription factor AP-2 β | 0.563 | 0.0001288 |
| NM_134266_at | SLC26A7 | Solute carrier family 26, member 7 | 0.562 | 0.0001305 |
| BC027487 at | C15orf62 | 0.561 | 0.0001364 | |
| DC311076_a_at | PIP4K2A | Phosphatidylinositol-4-phosphate 5-kinase type-2 | 0.561 | 0.0001374 |
| NM_006786_at | UTS2 | Human urotensin II | 0.559 | 0.0001448 |
| BC036592_at | GABRB2 | Gamma-aminobutyric-acid receptor beta-2 subunit | 0.557 | 0.0001555 |
| NM_018667_at | SMPD3 | Sphingomyelin phosphodiesterase 3 | 0.554 | 0.0001718 |
| NM_014717_at | ZNF536 | Zinc finger protein 536 | 0.552 | 0.000184 |
| NM_014629_s_at | ARHGEF10 | Rho guanine nucleotide exchange factor 10 | −0.552 | 0.000185 |
| NM_001005212_at | OR9Q1 | Olfactory receptor, family 9, subfamily Q | −0.552 | 0.0001851 |
| CR603904_s_at | EIF2AK2 | Protein kinase RNA-regulated (EIF2AK1) | −0.550 | 0.0001982 |
| BC050525_at | USP1 | Ubiquitin-specific processing protease 1 | −0.549 | 0.0002021 |
| AK024011_at | TOE1 | Target of EGR1 | −0.547 | 0.0002146 |
| NM_001037165_s_at | FOXK1 | Forkhead box K1 | −0.547 | 0.0002163 |
| DW432944_at | C4orf36 | C4orf36, hypothetical protein LOC132989, MGC26744, Hs.507712 | 0.547 | 0.0002164 |
| NM_001551_at | IGBP1 | Immunoglobulin-binding protein 1 | 0.546 | 0.0002189 |
| BX091412_at | KLHL34 | KLHL34, kelch-like 34, MGC125650, RP11-450P7.3, FLJ34960 | −0.546 | 0.0002246 |
| R37641_at | CA10 | Carbonic anhydrase-related protein 10 | 0.545 | 0.0002306 |
| NM_000343_at | SLC5A1 | Human Na+/glucose cotransporter 1 mRNA | 0.545 | 0.0002323 |
| BG776661_at | C10orf104 | C10orf104, FLJ33728 | 0.543 | 0.0002473 |
| BC122561_at | LIN7A | Lin-7 homolog A | 0.542 | 0.0002521 |
| NM_016486_at | TMEM69 | Transmembrane protein 69 | −0.541 | 0.0002638 |
| M18414_at | TRDV1 | TRDV1, hDV101S1 | 0.541 | 0.000264 |
| NM_014503_at | UTP20 | UTP20, down-regulated in metastasis | −0.539 | 0.0002797 |
| AY153484_at | PAX2 | Paired box gene 2 | 0.537 | 0.0002962 |
| BU589560_at | CLDN12 | CLDN12, claudin 12 | 0.536 | 0.0003033 |
| NM_001422_s_at | ELF5 | ELF5, ESE2, ESE-2 | 0.536 | 0.0003043 |
| BC038514_a_at | DPP10 | Dipeptidyl peptidase 10 | 0.536 | 0.0003078 |
| BX649183_at | IVNS1ABP | Influenza virus NS1A binding protein | 0.531 | 0.0003515 |
| NM_032588_at | TRIM63 | Muscle specific ring finger protein 1 | 0.531 | 0.0003543 |
| NM_153705_at | KDELC2 | KDELC2, MGC33424, KDEL (Lys-Asp-Glu-Leu) containing 2 | −0.531 | 0.0003585 |
| BX647977_a_at | RNMT | Human RNA (guanine-7-)methyltransferase | 0.530 | 0.0003622 |
| NM_032525_at | TUBB6 | Tubulin beta-6 chain | −0.530 | 0.0003656 |
| NM_017983_at | WIPI1 | Human WD-repeat protein interacting with phosphol | 0.530 | 0.0003692 |
| NM_003101_at | SOAT1 | Sterol O-acyltransferase 1 | −0.530 | 0.0003695 |
| NM_182538_at | SPNS3 | SPNS3, spinster homolog 3, MGC29671 | 0.529 | 0.0003766 |
| BU730580_at | RHO | Rhodopsin | 0.528 | 0.0003879 |
| AL713688_s_at | hCG_2009921 | hCG_2009921, LOC441204 | 0.527 | 0.0003942 |
| NM_016426_at | GTSE1 | GTSE1, G-2 and S-phase expressed 1 | −0.526 | 0.000407 |
| DB377031_x_at | PSG4 | Pregnancy specific β-1-glycoprotein 4 | 0.526 | 0.0004144 |
| BC101614_a_at | WDR72 | WD repeat domain 72 | 0.523 | 0.0004451 |
| BI761936_a_at | C12orf69 | 0.522 | 0.0004594 | |
| NM_021808_at | GALNT9 | Polypeptide N-acetylgalactosaminyltransferase 9 | 0.521 | 0.0004817 |
| NM_022127_at | SLC28A3 | Concentrative Na+-nucleoside cotransporter | 0.520 | 0.00049 |
| AK098151_at | PDK4 | Pyruvate dehydrogenase kinase 4 | 0.519 | 0.0005028 |
| NM_174900_at | ZFP42 | Zinc finger protein 42 | 0.519 | 0.0005127 |
| BC035128_a_at | MXI1 | MAX interacting protein 1 | 0.519 | 0.0005142 |
| NM_001085_at | SERPINA3 | Serine proteinase inhibitor, clade A, member 3 | 0.516 | 0.0005467 |
| AL564246_at | ZNF277 | Zinc finger protein 277 | 0.516 | 0.0005478 |
| NM_002813_at | PSMD9 | Proteasome 26S non-ATPase subunit 9 | −0.515 | 0.0005733 |
| NM_005318_at | H1F0 | H1 histone family, member 0 | −0.515 | 0.0005758 |
| AL136587_at | AGPAT5 | 1-acylglycerol-3-phosphate O-acyltransferase 5 | −0.514 | 0.0005852 |
| NM_015474_at | SAMHD1 | SAM domain- and HD domain-containing protein 1 | −0.514 | 0.0005942 |
| AV708719_at | FAM65C | FAM65C, dJ530I15.2, FLJ00360, FLJ32230, C20orf175 | 0.513 | 0.0006068 |
| AF313619_at | PAQR8 | Lysosomal membrane protein in brain 1 | 0.513 | 0.0006086 |
| NM_005656_at | TMPRSS2 | Transmembrane protease, serine 2 catalytic chain | 0.513 | 0.0006113 |
| CN310658_s_at | FXYD6 | FXYD domain-containing ion transport regulator 6 | 0.512 | 0.000615 |
| NM_032609_s_at | COX4I2 | Cytochrome c oxidase subunit 4 isoform 2 | 0.509 | 0.0006842 |
| NM_007168_at | ABCA8 | ATP-binding cassette, sub-family A member 8 | 0.507 | 0.00071 |
| NM_012478_at | WBP2 | WW domain binding protein 2 | 0.507 | 0.0007162 |
| AK125857_at | NUP62 | Nuclear pore glycoprotein p62 | −0.507 | 0.0007163 |
| NM_000078_at | CETP | Cholesteryl ester transfer protein | 0.507 | 0.0007259 |
| NM_001102610_a_at | TUBGCP5 | Tubulin, gamma complex associated protein 5 | −0.505 | 0.0007604 |
| NM_005773_at | ZNF256 | Zinc finger protein 256 | −0.505 | 0.0007613 |
| CB852298_at | CHORDC1 | Chord domain-containing protein 1 | −0.505 | 0.0007663 |
| NM_024306_at | FA2H | Fatty acid hydroxylase domain containing 1 | 0.504 | 0.0007716 |
| NM_031891_a_at | CDH20 | Cadherin 20 | 0.503 | 0.000811 |
| NM_020380_at | CASC5 | Cancer susceptibility candidate 5 | −0.502 | 0.0008354 |
| NM_003417_at | ZNF264 | Zinc finger protein 264 | −0.501 | 0.000841 |
| NM_018840_at | C20orf24 | C20orf24, PNAS-11, RAB5-interacting protein, RIP5 | −0.501 | 0.0008587 |
| NM_021269_s_at | ZNF708 | Zinc finger protein 15-like 1 (KOX 8) | −0.500 | 0.0008837 |
| NM_020167_at | NMUR2 | Neuromedin U receptor 2 | 0.499 | 0.0008887 |
| NM_001112724_at | STK32A | Serine/threonine kinase 32A | 0.499 | 0.0008995 |
| AK075129_s_at | RHOBTB1 | Rho-related BTB domain containing 1 | 0.498 | 0.0009097 |
| ENST00000357899_a_at | ZBTB44 | Zinc finger and BTB domain containing 44 | 0.498 | 0.0009199 |
| CR456455_s_at | SERHL | Serine hydrolase-like | −0.498 | 0.0009238 |
| NM_080717_at | TBX5 | T-box transcription factor TBX5 | 0.498 | 0.0009268 |
| BC098116_at | ABCA11P | FLJ14297, MGC120309, MGC120310, MGC120312, MGC132744, MGC138274 | −0.498 | 0.0009313 |
| AK026107_a_at | RBM25 | RNA binding motif protein 25, RNA-binding motif protein 25 | −0.497 | 0.0009449 |
| LIT1500_s_at | NOL5A | Nucleolar protein 5A | −0.497 | 0.0009486 |
| AF233261_a_at | OTOR | OTOR, fibrocyte-derived protein, melanoma inhibitory activity-like B protein | 0.497 | 0.0009601 |
| CR610033_a_at | TOM1L1 | Target of Myb-like protein 1 | 0.496 | 0.0009693 |
| AI144436_at | SF3A3 | Spliceosome-associated protein 61, Splicing factor 3A subunit 3 | −0.495 | 0.0009879 |
| AB053232_at | GAL3ST3 | Galactose 3′-sulfotransferase, galactose-3-O-sulfotransferase 3 | 0.495 | 0.0009897 |
| NM_206915_s_at | NGFRAP1 | NGFRAP1, BEX3, DXS6984E, HGR74, NADE, Bex | −0.495 | 0.0009941 |
Figure 1.
O-glycan biosynthesis/human version pathway.
Expression of the O-glycan pathway is associated with OVCA clinical outcome
Based on the above findings, we utilized PCA to develop gene expression signature scores for the pathways associated with gemcitabine sensitivity in vitro (15). In this way, a 34-gene ‘O-glycan biosynthesis pathway signature’ (OGBPS) (Table III) was generated and evaluated in an independent OVCA genomic dataset (16). Using the median value as a threshold to define high versus low OGBPS score, we identified an association between high OGBPS score and favorable survival (p=0.003; Fig. 2A). A similar association between high OGBPS score and favorable survival was observed in patients who underwent optimal (p=0.002) and suboptimal (approaching significance, p=0.07) cytoreduction (Fig. 2B). Most importantly, OVCA patients with a high OGBPS score who underwent suboptimal cytoreduction had a survival superior to patients with a low OGBPS score who underwent optimal cytoreduction (p=0.003). Interestingly, patients who demonstrated a CR to primary platinum-based therapy but had a high OGBPS score had superior survival compared with those patients who demonstrated a CR but had a low OGBPS score (p=0.003) (Fig. 2C). Patients who had an IR to primary therapy had no difference in survival associated with tumor OGBPS score (p=0.653) (Fig. 2D). When evaluated with cytoreductive status, grade, and age, the OGBPS score was an independent variable associated with survival (p<0.001).
Table III.
OGBPS 34 genes.
| NM_020981_at | B3GALT1 | Beta-1,3-galactosyltransferase |
| NM_003783_at | B3GALT2 | Beta-1,3-galactosyltransferase, beta-3-galt2 |
| NM_003782_a_at | B3GALT4 | Beta-1,3-galactosyltransferase 4 |
| NM_033171_at | B3GALT5 | GlcNAc-beta-1,3-galactosyltransferase 5, GLCT5, homolog of C |
| NM_138706_at | B3GNT6 | Beta-1,3-N-acetylglucosaminyltransferase protein |
| U10474_at | B4GALT1 | B4GALT1 |
| NM_003780_at | B4GALT2 | B4GALT2 |
| NM_003779_at | B4GALT3 | Beta4Gal-T3 |
| NM_020156_at | C1GALT1 | Core 1 synthase, glycoprotein-N-acetylgalactosamine |
| AW798875_at | GALNT1 | Polypeptide N-acetylgalactosaminyltransferase 1 |
| NM_024564_at | GALNT10 | Polypeptide N-acetylgalactosaminyltransferase 10 |
| NM_022087_at | GALNT11 | Polypeptide N-acetylgalactosaminyltransferase 11 |
| AI638649_at | GALNT12 | Polypeptide N-acetylgalactosaminyltransferase 12 |
| AK131195_a_at | GALNT13 | UDP-N-acetyl-alpha-D-galactosamine:polypeptide |
| NM_024572_s_at | GALNT14 | Polypeptide N-acetylgalactosaminyltransferase 14 |
| AK097996_at | GALNT2 | Polypeptide N-acetylgalactosaminyltransferase 2 |
| BX647473_a_at | GALNT3 | Polypeptide N-acetylgalactosaminyltransferase 3 |
| NM_003774_at | GALNT4 | Polypeptide N-acetylgalactosaminyltransferase 4 |
| BX097451_s_at | GALNT5 | Polypeptide N-acetylgalactosaminyltransferase 5 |
| BU542820_at | GALNT6 | Polypeptide N-acetylgalactosaminyltransferase 6 |
| NM_017423_at | GALNT7 | Polypeptide N-acetylgalactosaminyltransferase 7 |
| BM719843_a_at | GALNT8 | N-acetylgalactosaminyltransferase 8 |
| NM_021808_at | GALNT9 | Polypeptide N-acetylgalactosaminyltransferase 9 |
| NM_020692_at | GALNTL1 | Polypeptide N-acetylgalactosaminyltransferase 16 |
| BC030625_at | GALNTL2 | Polypeptide N-acetylgalactosaminyltransferase 13 |
| NM_198516_at | GALNTL4 | UDP-N-acetyl-alpha-D-galactosamine |
| NM_145292_at | GALNTL5 | UDP-N-acetyl-alpha-D-galactosamine |
| NM_001490_at | GCNT1 | Beta-1,6-N-acetylglucosaminyltransferase 1 |
| NM_145649_s_at | GCNT2 | Beta-1,6-N-acetylglucosaminyltransferase 2 |
| NM_004751_at | GCNT3 | Beta1,6-N-acetylglucosaminyltransferase 3 |
| CR619813_at | ST3GAL1 | 3-Sialyltransferase, Gal-NAc6S |
| AK127322_at | ST3GAL2 | Beta-galactoside alpha-2,3-sialytransferase |
| NM_018414_at | ST6GALN | 6-Sialyltransferase I alpha-N-acetylgalactosaminide alpha-2 |
| BC067524_a_at | WBSCR17 | Polypeptide N-acetylgalactosaminyltransferase, Williams-Beuren syndrome chromosome region 17 |
Figure 2.
High OGBPS PCA score is associated with favorable clinical outcome. Kaplan-Meier curves depicting the association between OGBPS PCA score and (A), overall survival from OVCA; (B), overall survival and complete response to platinum therapy median cut-off; (C), overall survival and incomplete response to platinum therapy; and (D), overall survival and cytoreductive status. ^Information available for 141 of 142 samples from MCC and Duke datasets. The numbers at risk are shown at the top of graphs. Log-rank test p-values indicate significance. CR, complete response; IR, incomplete response, O, optimal; S, suboptimal.
No associations with survival were observed for the first PCA score for the Cell cycle_Role of Nek in cell cycle regulation (59 genes, p=0.3107) or the Immune response_Antiviral actions of Interferons pathway (66 genes, p=0.5411).
Discussion
In this analysis, we applied an in vitro and in vivo genome-wide approach to define the molecular underpinnings of OVCA gemcitabine sensitivity. We identified genes and molecular signaling pathways associated with OVCA sensitivity to gemcitabine and, in doing so, have identified the OGBPS to be associated with in vitro gemcitabine response and also overall survival from OVCA.
Previous efforts to define the molecular basis of gemcitabine resistance have identified molecules such as deoxycytidine kinase (dCK) (8,17–19), ribonucleotide reductase (20–22), and human equilibrative nucleoside transporter-1 (hENT1) (4,23–26). Decreased activity of dCK, which phosphorylates gemcitabine to its monophosphate form, has previously been reported to be associated with resistance to gemcitabine (8,17–19). Consistent with these data, in our analysis, we demonstrated a negative correlation between OVCA cell line mRNA expression of the dCK gene and increasing gemcitabine resistance (Pearson’s correlation: −0.33, p=0.05). Previously, overexpression of the M1 and M2 subunits of ribonucleotide reductase (RRM1 and RRM2) has been demonstrated to be associated with gemcitabine resistance in gastrointestinal cancer cells (27,28). In our analysis, we observed no association between gemcitabine resistance and expression of RRM1, although we observed an association between low levels of RRM2 expression (using median expression as a threshold) and high gemcitabine IC50 (p<0.02). It is unclear why our findings are contradictory to those of Davidson et al (27); however, they may be due to differences in cancer types studied. Inhibition of hENT1 was previously reported to be associated with gemcitabine chemoresistance (4,25). This correlates with our findings in which we demonstrated a negative correlation between OVCA cell line mRNA expression of the hENT1 gene and increasing gemcitabine resistance (Pearson’s correlation: −0.3, p = 0.06).
The process of glycosylation involves the enzymatic addition of carbohydrates to proteins or lipids and is the most common form of post-translational modification. Three categories of protein-linked glycans exist, including those linked to the amide group of asparagine (N-linked), those linked to the hydroxyl group of serine, threonine, or hydroxylysine 3 (O-linked), and those linked to a carboxyl group of tryptophan (C-linked) (29). The main pathway for complex O-glycan biosynthesis is located in the endoplasmic reticulum and Golgi compartments, restricting glycosylation largely to the endoplasmic reticulum, Golgi, lysosomal, plasma membrane, and secretory proteins, with the exception of nuclear and cytosolic proteins, which can be modified with a single O-linked GlcNAc (30). O-glycans have been reported to have a broad range of functions in protein structure and stability, immunity, receptor-mediated signaling, non-specific protein interactions, modulation of the activity of enzymes and signaling molecules, and protein expression and processing (30,31). Although these biological roles range in importance, they can be critical for development, growth, function, and survival. Moreover, a specific O-glycan may influence a range of functions at different locations and times within an organism (31). Previously, limited access to endoglycosidases to cleave intact O-glycans from their protein backbone, as well as the extreme diversity of their structures, has limited research relative to study of N-linked glycan pathway-linked diseases (historically considered the congenital disorders of glycosylation). More recently, in human cancers, O-glycans have been shown to play important roles in cancer cell attachment, signaling, invasion (32–35), and survival in the bloodstream. Inhibition of the O-glycan pathway in colorectal cancer cell lines has been shown to inhibit cell growth and induce apoptosis (36). Down-regulation of the N-glycan biosynthesis pathway was also reported to be associated with chemoresistance in cholangiocarcinoma cell lines (37).
To date, we are unaware of any reports suggesting that the O-glycan pathway influences OVCA cell response to therapeutic interventions or overall survival. In this study, expression of the O-glycan pathway (quantified by a OGBPS score) was associated with OVCA overall survival when we analyzed: a) all patients with OVCA, b) patients who underwent optimal OVCA surgical cytoreduction, and c) patients who experienced a CR to primary surgery plus platinum-based therapy. The association between OGBPS score and overall survival for patients who underwent suboptimal surgical cytoreduction did not reach statistical significance (p=0.07), and no association was identified in patients who experienced an IR to primary surgery plus platinum-based therapy. When evaluated with cytoreductive status, grade, and age, the OGBPS score was an independent variable associated with survival (p<0.001). The explanation for the associations between OGBPS and OVCA survival is likely complex. Although in this study we identified the O-glycan pathway by its association with in vitro gemcitabine sensitivity, we do not believe that the impact of the pathway on overall survival is driven by its effect of gemcitabine sensitivity. In fact, high OGBPS score was associated with resistance to gemcitabine, yet showed a more favorable outcome for patients with OVCA. As noted above, O-glycans are known to influence cancer cell attachment, signaling, invasion, and survival in the bloodstream (32–35). It is likely that the effect on OVCA clinical outcome is associated with one or more of these important oncologic processes. It will therefore be essential in future studies to investigate associations between OGBPS score, activity of individual members of the O-glycan pathway, and OVCA cell phenotypic behavior.
Our discovery of associations between O-glycan pathway expression and gemcitabine sensitivity and patient survival is novel. These findings potentially have substantial implications for future clinical management of patients with OVCA. In the future, empiric-based treatment decision-making must be replaced with a more tailored strategy that stratifies patients based on their molecular fingerprints. Such an approach will identify those patients with the ‘highest risk’ disease, those who may benefit from additional pathway-targeted therapy added to standard of care cytotoxic regimens, and potentially those who may (or may not) benefit from aggressive surgical interventions.
Acknowledgments
Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the funding agencies. We thank Rasa Hamilton (Moffitt Cancer Center) for editorial assistance. We also like to thank Merck pharmaceuticals for their contributions. This study was supported in part by the Hearing the Ovarian Cancer Whisper, Jacquie Liggett Foundation, the National Cancer Institute Grant R21 CA-110499-01A2, the Ocala Royal Dames for Cancer Research Inc., the Phi Beta Psi Sorority, the Ovarian Cancer Research Fund, and the US Army Medical Research and Materiel Command under award no. DAMD17-02-2-0051.
References
- 1.Gadducci A, Sartori E, Maggino T, et al. Analysis of failures after negative second-look in patients with advanced ovarian cancer: an Italian multicenter study. Gynecol Oncol. 1998;68:150–155. doi: 10.1006/gyno.1997.4890. [DOI] [PubMed] [Google Scholar]
- 2.McGuire WP, Hoskins WJ, Brady MF, et al. Cyclophosphamide and cisplatin compared with paclitaxel and cisplatin in patients with stage III and stage IV ovarian cancer. N Engl J Med. 1996;334:1–6. doi: 10.1056/NEJM199601043340101. [DOI] [PubMed] [Google Scholar]
- 3.Ozols RF. The current role of gemcitabine in ovarian cancer. Semin Oncol. 2001;28:18–24. doi: 10.1016/s0093-7754(01)90274-x. [DOI] [PubMed] [Google Scholar]
- 4.Mackey JR, Mani RS, Selner M, et al. Functional nucleoside transporters are required for gemcitabine influx and manifestation of toxicity in cancer cell lines. Cancer Res. 1998;58:4349–4357. [PubMed] [Google Scholar]
- 5.Distefano M, Ferlini C, De Vincenzo R, Gaggini C, Mancuso S, Scambia G. Antagonistic effect of the combination gemcitabine/topotecan in ovarian cancer cells. Oncol Res. 2000;12:355–359. doi: 10.3727/096504001108747819. [DOI] [PubMed] [Google Scholar]
- 6.Peters GJ, Bergman AM, Ruiz van Haperen VW, Veerman G, Kuiper CM, Braakhuis BJ. Interaction between cisplatin and gemcitabine in vitro and in vivo. Semin Oncol. 1995;22:72–79. [PubMed] [Google Scholar]
- 7.Ruiz van Haperen VW, Veerman G, Boven E, Noordhuis P, Vermorken JB, Peters GJ. Schedule dependence of sensitivity to 2′,2′-difluorodeoxycytidine (Gemcitabine) in relation to accumulation and retention of its triphosphate in solid tumour cell lines and solid tumours. Biochem Pharmacol. 1994;48:1327–1339. doi: 10.1016/0006-2952(94)90554-1. [DOI] [PubMed] [Google Scholar]
- 8.Ruiz van Haperen VW, Veerman G, Eriksson S, et al. Development and molecular characterization of a 2′,2′-difluorodeoxycytidine-resistant variant of the human ovarian carcinoma cell line A2780. Cancer Res. 1994;54:4138–4143. [PubMed] [Google Scholar]
- 9.Ruiz van Haperen VW, Veerman G, Eriksson S, Stegmann AP, Peters GJ. Induction of resistance to 2′,2′-difluorodeoxycytidine in the human ovarian cancer cell line A2780. Semin Oncol. 1995;22:35–41. [PubMed] [Google Scholar]
- 10.Van Moorsel CJ, Veerman G, Bergman AM, et al. Combination chemotherapy studies with gemcitabine. Semin Oncol. 1997;24 (Suppl 7):17–23. [PubMed] [Google Scholar]
- 11.Braakhuis BJ, Ruiz van Haperen VW, Boven E, Veerman G, Peters GJ. Schedule-dependent antitumor effect of gemcitabine in in vivo model system. Semin Oncol. 1995;22:42–46. [PubMed] [Google Scholar]
- 12.Friedlander M, Millward MJ, Bell D, et al. A phase II study of gemcitabine in platinum pre-treated patients with advanced epithelial ovarian cancer. Ann Oncol. 1998;9:1343–1345. doi: 10.1023/a:1008469212268. [DOI] [PubMed] [Google Scholar]
- 13.Lund B, Hansen OP, Theilade K, Hansen M, Neijt JP. Phase II study of gemcitabine (2′,2′-difluorodeoxycytidine) in previously treated ovarian cancer patients. J Natl Cancer Inst. 1994;86:1530–1533. doi: 10.1093/jnci/86.20.1530. [DOI] [PubMed] [Google Scholar]
- 14.Markman M, Webster K, Zanotti K, Kulp B, Peterson G, Belinson J. Phase 2 trial of single-agent gemcitabine in platinum-paclitaxel refractory ovarian cancer. Gynecol Oncol. 2003;90:593–596. doi: 10.1016/s0090-8258(03)00399-8. [DOI] [PubMed] [Google Scholar]
- 15.Chen DT, Nasir A, Culhane A, et al. Proliferative genes dominate malignancy-risk gene signature in histologically-normal breast tissue. Breast Cancer Res Treat. 2010;119:335–346. doi: 10.1007/s10549-009-0344-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Marchion DC, Cottrill HM, Xiong Y, et al. BAD phosphorylation determines ovarian cancer chemo-sensitivity and patient survival. Clin Cancer Res. 2011;17:6356–6366. doi: 10.1158/1078-0432.CCR-11-0735. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Galmarini CM, Mackey JR, Dumontet C. Nucleoside analogues: mechanisms of drug resistance and reversal strategies. Leukemia. 2001;15:875–890. doi: 10.1038/sj.leu.2402114. [DOI] [PubMed] [Google Scholar]
- 18.Jordheim L, Galmarini CM, Dumontet C. Drug resistance to cytotoxic nucleoside analogues. Curr Drug Targets. 2003;4:443–460. doi: 10.2174/1389450033490957. [DOI] [PubMed] [Google Scholar]
- 19.Van der Wilt CL, Kroep JR, Bergman AM, et al. The role of deoxycytidine kinase in gemcitabine cytotoxicity. Adv Exp Med Biol. 2000;486:287–290. doi: 10.1007/0-306-46843-3_56. [DOI] [PubMed] [Google Scholar]
- 20.Cory JG, Sato A. Regulation of ribonucleotide reductase activity in mammalian cells. Mol Cell Biochem. 1983;53–54:257–266. doi: 10.1007/BF00225258. [DOI] [PubMed] [Google Scholar]
- 21.Thelander L, Berg P. Isolation and characterization of expressible cDNA clones encoding the M1 and M2 subunits of mouse ribonucleotide reductase. Mol Cell Biol. 1986;6:3433–3442. doi: 10.1128/mcb.6.10.3433. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Zhou BS, Tsai P, Ker R, et al. Overexpression of transfected human ribonucleotide reductase M2 subunit in human cancer cells enhances their invasive potential. Clin Exp Metastasis. 1998;16:43–49. doi: 10.1023/a:1006559901771. [DOI] [PubMed] [Google Scholar]
- 23.Garcia-Manteiga J, Molina-Arcas M, Casado FJ, Mazo A, Pastor-Anglada M. Nucleoside transporter profiles in human pancreatic cancer cells: role of hCNT1 in 2′,2′-difluorodeoxycytidine-induced cytotoxicity. Clin Cancer Res. 2003;9:5000–5008. [PubMed] [Google Scholar]
- 24.Mackey JR, Yao SY, Smith KM, et al. Gemcitabine transport in xenopus oocytes expressing recombinant plasma membrane mammalian nucleoside transporters. J Natl Cancer Inst. 1999;91:1876–1881. doi: 10.1093/jnci/91.21.1876. [DOI] [PubMed] [Google Scholar]
- 25.Rauchwerger DR, Firby PS, Hedley DW, Moore MJ. Equilibrative-sensitive nucleoside transporter and its role in gemcitabine sensitivity. Cancer Res. 2000;60:6075–6079. [PubMed] [Google Scholar]
- 26.Ritzel MW, Ng AM, Yao SY, et al. Recent molecular advances in studies of the concentrative Na+-dependent nucleoside transporter (CNT) family: identification and characterization of novel human and mouse proteins (hCNT3 and mCNT3) broadly selective for purine and pyrimidine nucleosides (system cib) Mol Membr Biol. 2001;18:65–72. doi: 10.1080/09687680010026313. [DOI] [PubMed] [Google Scholar]
- 27.Davidson JD, Ma L, Flagella M, Geeganage S, Gelbert LM, Slapak CA. An increase in the expression of ribonucleotide reductase large subunit 1 is associated with gemcitabine resistance in non-small cell lung cancer cell lines. Cancer Res. 2004;64:3761–3766. doi: 10.1158/0008-5472.CAN-03-3363. [DOI] [PubMed] [Google Scholar]
- 28.Jung CP, Motwani MV, Schwartz GK. Flavopiridol increases sensitization to gemcitabine in human gastrointestinal cancer cell lines and correlates with down-regulation of ribonucleotide reductase M2 subunit. Clin Cancer Res. 2001;7:2527–2536. [PubMed] [Google Scholar]
- 29.Hofsteenge J, Muller DR, De Beer T, Loffler A, Richter WJ, Vliegenthart JF. New type of linkage between a carbohydrate and a protein: C-glycosylation of a specific tryptophan residue in human RNase Us. Biochemistry. 1994;33:13524–13530. doi: 10.1021/bi00250a003. [DOI] [PubMed] [Google Scholar]
- 30.Wells L, Hart GW. O-GlcNAc turns twenty: functional implications for post-translational modification of nuclear and cytosolic proteins with a sugar. FEBS Lett. 2003;546:154–158. doi: 10.1016/s0014-5793(03)00641-0. [DOI] [PubMed] [Google Scholar]
- 31.Varki A. Biological roles of oligosaccharides: all of the theories are correct. Glycobiology. 1993;3:97–130. doi: 10.1093/glycob/3.2.97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Fuster MM, Brown JR, Wang L, Esko JD. A disaccharide precursor of sialyl Lewis X inhibits metastatic potential of tumor cells. Cancer Res. 2003;63:2775–2781. [PubMed] [Google Scholar]
- 33.Gabius HJ. Cell surface glycans: the why and how of their functionality as biochemical signals in lectin-mediated information transfer. Crit Rev Immunol. 2006;26:43–79. doi: 10.1615/critrevimmunol.v26.i1.30. [DOI] [PubMed] [Google Scholar]
- 34.Huet G, Gouyer V, Delacour D, et al. Involvement of glycosylation in the intracellular trafficking of glycoproteins in polarized epithelial cells. Biochimie. 2003;85:323–330. doi: 10.1016/s0300-9084(03)00056-7. [DOI] [PubMed] [Google Scholar]
- 35.Ulloa F, Real FX. Benzyl-N-acetyl-alpha-D-galactosaminide induces a storage disease-like phenotype by perturbing the endocytic pathway. J Biol Chem. 2003;278:12374–12383. doi: 10.1074/jbc.M211909200. [DOI] [PubMed] [Google Scholar]
- 36.Patsos G, Hebbe-Viton V, Robbe-Masselot C, et al. O-glycan inhibitors generate aryl-glycans, induce apoptosis and lead to growth inhibition in colorectal cancer cell lines. Glycobiology. 2009;19:382–398. doi: 10.1093/glycob/cwn149. [DOI] [PubMed] [Google Scholar]
- 37.Sato J, Kimura T, Saito T, et al. Gene expression analysis for predicting gemcitabine resistance in human cholangiocarcinoma. J Hepatobiliary Pancreat Sci. 2011;18:700–711. doi: 10.1007/s00534-011-0376-7. [DOI] [PubMed] [Google Scholar]