Small-Molecule Inhibition of GCNT3 Disrupts Mucin ... · Chinthalapally V. Rao1, Naveena B. Janakiram1,Venkateshwar Madka1, Gaurav Kumar1, Edgar J. Scott 2 , Gopal Pathuri 1 ,Taylor

Therapeutics, Targets, and Chemical Biology

Small-Molecule Inhibition of GCNT3 DisruptsMucin Biosynthesis and Malignant CellularBehaviors in Pancreatic CancerChinthalapally V. Rao1, Naveena B. Janakiram1, Venkateshwar Madka1, Gaurav Kumar1,Edgar J. Scott2, Gopal Pathuri1, Taylor Bryant1, Hannah Kutche1, Yuting Zhang1,Laura Biddick1, Hariprasad Gali3, Yan D. Zhao4, Stan Lightfoot1, and Altaf Mohammed1

Abstract

Pancreatic cancer is an aggressive neoplasm with almostuniform lethality and a 5-year survival rate of 7%. Severaloverexpressed mucins that impede drug delivery to pancreatictumors have been therapeutically targeted, but enzymesinvolved in mucin biosynthesis have yet to be preclinicallyevaluated as potential targets. We used survival data fromhuman patients with pancreatic cancer, next-generationsequencing of genetically engineered Kras-driven mouse pan-creatic tumors and human pancreatic cancer cells to identify thenovel core mucin-synthesizing enzyme GCNT3 (core 2 b-1,6N-acetylglucosaminyltransferase). In mouse pancreatic cancertumors, GCNT3 upregulation (103-fold; P < 0.0001) wascorrelated with increased expression of mucins (5 to 87-fold;P < 0.04–0.0003). Aberrant GCNT3 expression was also asso-ciated with increased mucin production, aggressive tumorigen-

esis, and reduced patient survival, and CRISPR-mediatedknockout of GCNT3 in pancreatic cancer cells reduced prolif-eration and spheroid formation. Using in silico small moleculardocking simulation approaches, we identified talniflumate as anovel inhibitor that selectively binds to GCNT3. In particular,docking predictions suggested that three notable hydrogenbonds between talniflumate and GCNT3 contribute to a dock-ing affinity of �8.3 kcal/mol. Furthermore, talniflumate aloneand in combination with low-dose gefitinib reduced GCNT3expression, leading to the disrupted production of mucins invivo and in vitro. Collectively, our findings suggest that targetingmucin biosynthesis through GCNT3 may improve drug respon-siveness, warranting further development and investigation inpreclinical models of pancreatic tumorigenesis. Cancer Res; 76(7);1965–74. �2016 AACR.

IntroductionPancreatic cancer is a lethal disease, and its management is an

ongoing challenge. Pancreatic cancer is the fourth leading cause ofdeaths due to cancer in the United States (1). It is a highlyaggressive cancer that is usually diagnosed at an advanced stageand has the worst prognosis of any malignancy, with a 7% 5-yearsurvival rate due to high chemoresistance. This chemoresistance is

due in part to altered expressions of mucins, which form a meshthat makes target sites inaccessible to drugs (2–9). Clinical andpreclinical studies have shown aberrant expression of mucinsduring pancreatic cancer development. The mucins may preventdrugs fromaccessing their sites of action. The expression of severalmucins, including MUC1, MUC4, MUC5AC, and MUC16, ishighly upregulated in pancreatic ductal adenocarcinoma (PDAC),pancreatic intraepithelial neoplasia (PanIN), intrapapillarymucinous neoplasms, and mucinous cystic neoplasms frompatients with pancreatic cancer (2–18). The extent to which thedense mucin mesh influences the antiproliferative activity of5-fluorouracil (5-FU) was investigated using human pancreaticcancer cells (2, 3). These studies have led to increasing recognitionof mucins as potential diagnostic markers and therapeutic targetsin pancreatic cancer. However, there are no reports evaluating themucin-synthesizing genes as potential therapeutic targets.

Over 95%of human pancreatic cancers are associatedwith Krasmutations. Activation of the Kras and EGFR signaling pathways isinvolved in increased cell proliferation, angiogenesis, metastasis,and decreased apoptosis. Clinical studies also suggest a correla-tion between overexpression of mucins and pancreatic cancers(10–13). One important issue is whether synthesis of mucins bycore enzymes is regulated coordinately. We have shown thatp48Cre/þ-LSL-KrasG12D/þ genetically engineeredmice (GEM)withKras mutations display significant dysregulation of EGFR andupregulation of mucin biosynthesis in PanIN lesions (19). Wehave observed that inhibition of PDAC with the EGFR inhibitor

1Center for Cancer Prevention and DrugDevelopment, Department ofMedicine, Hem-Onc Section, University of Oklahoma Health SciencesCenter, Oklahoma City, Oklahoma. 2Department of Microbiology andImmunology, University of Oklahoma Health Sciences Center, Okla-homa City, Oklahoma. 3College of Pharmacy, University of OklahomaHealth Sciences Center, Oklahoma City, Oklahoma. 4Biostatistics andEpidemiology, College of Public Health, University of OklahomaHealth Sciences Center, Oklahoma City, Oklahoma.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

C.V. Rao and N.B. Janakiram contributed equally to this article.

Corresponding Authors: Chinthalapally V. Rao, University of Oklahoma HealthSciences Center, 975NE 10th Street, BRC 1203,OklahomaCity, OK 73104. Phone:405-271-3224;Fax: 405-271-3225;E-mail: [email protected]; orAltafMohammed,University of Oklahoma Health Sciences Center, 975 NE 10th Street, BRC North,Room # 209, Oklahoma City, OK 73104. Phone: 405-271-8000, ext. 32533; Fax:405-271-3225; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-15-2820

�2016 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org 1965

on September 8, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 15, 2016; DOI: 10.1158/0008-5472.CAN-15-2820

http://cancerres.aacrjournals.org/

gefitinib is associated with suppression of mucin synthesis inp48Cre/þ-LSL-KrasG12D/þ GEM (19). Although EGFR inhibitors(gefitinib) are effective, higher doses lead to significant skintoxicity. Hence, new approaches to target mucin synthesis areurgently needed to improve the efficacy of existing drugs or todesign new drugs. Although several mucins, likeMuc1 andMuc5,have been targeted, we have little knowledge about potentialintervention approaches involving mucin-synthesizing genes.

The mucin-synthesizing core 2 b-1,6 N-acetylglucosaminyl-transferase (GCNT3/C2GNT) plays a significant role in mucinbiosynthesis. Aberrant GCNT3 expression leads to overexpressionof mucins (17, 18, 20). GCNT3 activity plays important roles inphysiologic processes, including inflammatory and immuneresponses. Thus, GCNT3 is an attractive novel target for pancreaticcancer treatment (17–18, 20–21). To date, nearly all reports onpancreatic cancer mucins have focused on mucin peptide geneexpression or on aspects of the glycosylated mucins, with rela-tively few reports addressing the enzymes that catalyze mucinbiosynthesis.

In this article, using human pancreatic cancer tissues, GEMtissues, transcriptome analysis, in silico experiments, ClusteredRegularly Interspaced Short Palindromic Repeats (CRISPR),spheroid assay, and in vitro and in vivo studies, we identifiedGCNT3 as a novel target for pancreatic cancer intervention. Then,we determined that talniflumate effectively binds to GCNT3 withlow binding energy. Finally, we found that talniflumate decreasedGCNT3 expression alone or in combination with gefitinib in vitroand in vivo.

Materials and MethodsCell lines and reagents

Human pancreatic cancer cell lines (MIA PaCa, BxPC-3, andPANC-1) were obtained from the ATCC. The cell lines wereauthenticated by short-tandem repeat analysis by the ATCC. Celllines were initially expanded and cryopreserved within 1 monthof receipt andwere typically used for 3months, and at that time, afresh vial of cryopreserved cellswas used (2011–2015). These cellswere maintained in culture as an adherent monolayer in RPMI1640 medium or DMEM supplemented with 10% FBS. Antibo-dies for GCNT3, Muc4, Muc13, PCNA, cyclin D1, and actin werepurchased from Abcam, Novus, and Santa Cruz Biotechnology.

Patient specimensA human pancreatic cancer tissue array (HPan-Ade180Sur-01),

consisting of 90 cases of tumor and matched normal adjacenttissue (1 core/case) with survival data, was procured from USBiomax, Inc. The tumors were categorized as clinical stages I, II,and IV (seventh edition of the American Joint Committee onCancer criteria). Survival informationwas available; patients werefollowed for 3 to 7 years. Patients' history of smoking, drinking,diabetes, hepatitis B information, and family medical historywere recorded. In addition, we used OD-CT-DgPan03-001 PDAC(62 cases; US Biomax, Inc.). These PDACs (1 core/case) werepathology grade 1, 1–2, 2, and 2–3. Ninety percent of cases hadtumor–node–metastasis detail information. Further, we usedpatient tissue slides from our Stephenson Cancer Center core.

IHC analysisThe paraffin-embedded tissue microarrays and GEM tissue

slides were used for IHC staining of GCNT3, following our

previously published method (19). Mucin staining was per-formed per our earlier publication (19).

Mouse model, diet, and handlingAll animal research was performed under the auspices of

animal protocols approved by the University of OklahomaHealth Sciences Center (OUHSC, Oklahoma City, OK) Institu-tional Animal Care and Use Committee. Animals were housed inventilated cages under standardized conditions (21�C, 60%humidity, 12-hour light/12-dark cycle, 20 air changes/hour) inthe university rodent barrier facility. Semipurified modified AIN-76A diet ingredients were purchased from Bioserv, Inc. Genera-tion of p48Cre/þ-LSL-KrasG12D/þ mice expressing the activatedKrasG12D oncogene has been described previously (22–27). TheEGFR inhibitor gefitinib was procured from the center for cancerprevention and drug development drug repository. Mice wereallowed ad libitum access to the respective diets and to automatedtap water purified by reverse osmosis.

Breeding and genotyping analysisLSL-KrasG12D/þ and p48Cre/þ mice were maintained on a

C57BL/6 heterozygous genetic background. LSL-KrasG12D/þ andp48cre/þmice were bred, and offspring of activated p48Cre/þ.-LSL-KrasG12D/þ and C5BL/6 wild-type mice were generated at therequired quantities as previously described (22–27).

Tissue processing and histologic analysis of PanIN lesions andPDAC

After euthanizing the mice, pancreata and other key organs,including liver, spleen, kidney, and lung, were collected andweighed. Tissues were fixed in 10% formalin for 24 hours andwere routinely processed and embedded in paraffin. Formalin-fixed 4-mm tissue sections from each pancreas were stained withhematoxylin & eosin and were histologically evaluated by apathologist blinded to the experimental groups. PanIN lesionsand carcinoma were classified according to histopathologic cri-teria, as recommended elsewhere (19, 22–27). To quantify theprogression of PanIN lesions, the total number of ductal lesionsand their grades were determined (19, 22–27). Similarly, pan-creatic carcinoma and normal-appearing pancreatic tissue wereevaluated in all animals.

Transcriptome analysis/next-generation sequencing/RNA-seq,and data analyses

Pancreatic tumor tissues and normal pancreas were collectedfrom at least 6 10-month-old Kras mice and 6 10-month-oldwild-type mice. Pancreatic tumor tissues were also collectedfrom at least 3 untreated control mice and 3 gefitinib-treatedmice, after 25 weeks of treatment. Total RNA was extractedwith Trizol reagent (Life Technologies) and subjected to cDNAlibrary construction, followed by next-generation sequencing(NGS), per the NGS protocol, with a Solid or Illuminasequencer in the OUHSC core facility (Microgen). Thesequence data were deposited to the Geospiza/Perkin Elmercompany server and analyzed with GeneSifter bioinformaticssoftware.

Synthesis of talniflumateTalniflumate was synthesized with the synthetic method and

characterized by comparison of spectral data; it was fully in

Rao et al.

Cancer Res; 76(7) April 1, 2016 Cancer Research1966




agreement with those reported in the literature (28, 29). 1H-NMR spectroscopy was performed on a Varian Mercury VX-300NMR Spectrometer at the NMR Facility (University of Okla-homa). High-resolution electrospray mass (ES-HRMS) spectralanalysis was performed at the OUHSC Molecular Biology-Proteomics Facility using a Qstar Elite Q-TOF (Applied Biosys-tems) mass spectrometer. High-performance liquid chromatog-raphy (HPLC) solvents consisted of water containing 0.1%trifluoroacetic acid (solvent A) and acetonitrile containing0.1% trifluoroacetic acid (solvent B). A Sonoma C18 (ESIndustries, 10 mm, 100 Å, 4.6 � 250 mm) column was usedwith a flow rate of 1.5 mL/min. The HPLC gradient systembegan with an initial solvent composition of 95% A and 5% Bfor 2 minutes, followed by a linear gradient to 5% A and 95% Bin 30 minutes, after which the column was reequilibrated. Theabsorption detector was set at 254 nm.

In silico molecular docking studiesTalniflumate was obtained from NCBI's Pubchem database

(Pubchem CID 48229). The crystal structure for GCNT3 wasdownloaded from the RCSB Protein Data Bank (PDBID 2GAM).Chain A and the reference ligand, GALB1,3GALNAC, were sepa-rated from the pdb file into two separate files for docking, and allwater molecules were removed. Using AutoDockTools, nonpolarhydrogens and partial charges were added to the receptor GCNT3,and to the ligands GALB1,3GALNAC and talniflumate. For flex-ible ligand docking, rotatable bonds were defined for bothGALB1,3GALNAC and talniflumate, with a total of 12 and 6rotatable bonds, respectively. To accomplish a blind dockingsimulation, six search grids were created with AutoGrid4 tosurround the entire protein with the following grid center coor-dinates: (84.029, �5.002, 65.752), (51.029, �5.002, 65.752),(67.529, 11.498, 65.752), (67.529, �21.502, 65.752), (67.529,�5.002, 68.752), (67.529, �5.002, 62.752). Each grid was set to126� 126� 126 points with a spacing of 0.35 Å. The Lamarckiansearch algorithm was used to dock the ligands in all six energy

grids. The GA population size was increased to 500, energyevaluations to 25,000,000, number of generations to10,000,000, and the number of runs to 500. Other default searchparameters were unchanged. All docking results were concatenat-ed and sorted by binding free energies to identify the conforma-tion with the lowest binding free energy. Figures were createdusing Pymol.

MTT assay, Western immunoblotting, and real-time PCR3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide

(MTT) assay for viability, Western immunoblotting, and real-timePCR techniques were used to analyze the expression of GCNT3,Muc1, Muc13, Muc4, Muc5ac, PCNA, and cyclin D1.

CRISPR knockout for GCNT3 and spheroid assayCRISPR/Cas9 KO plasmid transfection. In a 6-well tissue cultureplate, 1.5 � 105 to 2.5 � 105 cells were seeded in 3 mL ofantibiotic-free standard growth medium per well, 24 hours priorto transfection. Cells were grown to 40% to 80% confluency.Healthy and subconfluent cells are used for successful humanGCNT3 CRISPR/Cas9 KO Plasmid transfection (Santa Cruz Bio-technology) as per the manufacturer's instructions. After incuba-tion, successful transfection of CRISPR/Cas9 KO Plasmid wasvisually confirmed by detection of theGFP via fluorescentmicros-copy. Cell viability was determined using the MTT method.

Spheroid assay. The AlgiMatrix 3D Culture System (Life Tech-nologies) was used for multicellular tumor spheroid assays,following the vendor's instructions. After GCNT3 transfection,untransfected and transfected cells were trypsinized and resus-pended (1 � 106 cells/mL) in standard cell culture medium.Then, 10% (v/v) AlgiMatrix Firming Buffer was added to thissuspension. An AlgiMatrix 3D Culture System plate was inoc-ulated with the suspension, which was added to the top of eachdry sponge with a pipette. Approximately 5 minutes after

A

Mucins PanINs

>50% 5–50% ≤5%

PanINs 4/57 21/55 32/57

(7%) (36.8%) (56.2%)

>50% 5–50% ≤5%

Male

High Low

25/103 28/103 50/103(24.2%) (27.18%) (48.54%)

Female 9/43 13/43 21/43(20.9%) (30.23%) (48.83%)

C

B

D

Months

90807060504030201000.0

0.2

0.4

0.6

0.8

1.0

Sur

viva

l pro

babi

lity

PanINs PDACGCNT3 GCNT3

Figure 1.Correlations between GCNT3expression and pancreatic tumorprogression. A, Kaplan–Meier analysisof patients' survival by GCNT3expression level. Patients with higherGCNT3 expression (>5.5%) showeddecreased overall survival. B and C,GCNT3 expression in patient tumors(B) and PanIN lesions (C). D, Alcianblue (left) for mucin and IHC (middleand right) for GCNT3 staining andmucin expression in human PanINsand pancreatic cancer.

Impact of GCNT3 in Pancreatic Cancer

www.aacrjournals.org Cancer Res; 76(7) April 1, 2016 1967




rehydration, additional volumes of culture medium were addedto the top of each sponge, without firming buffer, and wereincubated at 37�C in a humidified atmosphere of 5% CO2 in air.Culture medium was changed based on cell proliferation orwhen the medium began to change color. Spheroid formationwas monitored for 3 weeks, and the number and size ofspheroids were recorded.

Statistical analysisSurvival time was compared between men and women and

between high and low expressions of GCNT3 using the log-rank test. Survival rates across time by gender and by GCNT3expression were estimated and plotted using the Kaplan–Meiermethod. The other data are presented as mean � SE. Differ-ences in body weights were analyzed using ANOVA. Statisticaldifferences between control and treated groups were evaluatedusing the Fisher exact test for PDAC incidence. Unpaired t testwith the Welch correction was used for PanIN and PDAClesions. Differences between groups were considered signifi-cant at P < 0.05.

ResultsCorrelation between GCNT3 expression and patient survival inhuman pancreatic cancer

To determine the association between GCNT3 expressionand patient survival, we performed an IHC analysis for GCNT3expression on a tissue microarray containing 180 specimensfrom individuals with pancreatic cancer (90) or matchedcontrols (90). The median age was 62 years, and 36.1% ofpatients were female. Based on the seventh edition of theAmerican Joint Committee on Cancer criteria, there were 4stage IA, 35 stage IB, 13 stage IIA, 33 stage IIB, and 2 stage IVtumors. Pathologically, the tumors were adenocarcinoma,ductal adenocarcinoma, mucinous adenocarcinoma, and ade-nosquamous carcinoma. There were 51.3% poorly differenti-ated tumors and 48.63% well-to-moderately differentiatedtumors. All tumors that were tested for GCNT3 expressionwere evaluated by an independent pathologist.

There were 88 patients (32 females, 56 males) with pancreaticcancer in the study. The median follow-up time was 12 months.Fifty-nine patients died during the study. Women had a

GCNT3

P < 0.0001

N N T TGCNT3GAPDH

GCNT3 Mucins

PanINs

PDAC

PanINs

PDAC

MucinsGCNT3

A B

C

D150

250

1.5

1.0

0.5

0.0Normal

P < 0.0001

Tumor

200

150

100

50

Pan

INs

mu

ltip

licit

y (m

ean

±± S

E)

Pan

crea

s w

ts (

g)

mea

n ±

SE

0PanIN 1 PanIN 2 PanIN 3

100

50

0Normal Tumor

Exp

ress

ion

E F

Figure 2.GCNT3 expression in pancreatic cancer of GEM. A, weights of normal pancreas (wild-type mice) and tumor (GEM). B, number of PanIN lesions in 10-month-oldGEM. C, mRNA expression of GCNT3 in normal tissue vs. tumors. D and E, NGS transcriptome analysis of GCNT3 and mucins in tumors vs. normal pancreasfrom GEM. F, IHC showing GCNT3 expression in PanINs and PDAC (left), and Alcian blue staining showing mucin expression in PanINs and PDAC (right). Asignificant increase in GCNT3 and mucin expression was seen in the tumors.

Rao et al.





marginally nonsignificant longer survival time thanmen (mediansurvival: 30 vs. 12months, P¼ 0.053; Supplementary Fig. S1A). A5.5% staining score as a cutoff point was used to determine lowGCNT3 (�5.5%) versus high GCNT3 (>5.5%) expression.Patients with low expression of GCNT3 had a longer survivaltime than patients with high expression of GCNT3 (mediansurvival: 17.5 vs. 10.5 months, P ¼ 0.036; Fig. 1A). In patientswith high expression of GCNT3, there was no significant differ-ence in survival time between women andmen (median survival:17 vs. 10, P ¼ 0.2894; Supplementary Fig. S1B). Importantly,those patients with history of smoking, drinking, or diabetescomprised 62.9% of the sample with high GCNT3. These dataindicate that GCNT3 may be an important prognostic factor forpatientswithpancreatic cancer. Further, after analyzing additionalsamples (total 103 pancreatic cancer samples) for the percentpositive cells stained forGCNT3, 24.2%(25/103)ofmale patientswith PDAC showed >50% GCNT3 expression, 27.18% (28/103)showed 5% to 50% GCNT3 expression, and 48.54% (50/103)patients showed <5%GCNT3 expression. In female patients with

PDAC, 20.9% (9/43) showed >50% GCNT3 expression, 30.23%(13/43) showed 5% to 50% GCNT3 expression, and 48.83%(21/43) showed <5% GCNT3 expression (Fig. 1B). Similarly,high expression of GCNT3 was observed in human PanINs(Fig. 1C). Figure 1D shows the expression of GCNT3 and mucinin PanINs and PDAC.

Pancreatic cancer progression in GEMOver the past few years, we have demonstrated the utility of the

p48Cre/þ-LSL-KrasG12D/þ (Kras) mouse model for testing theefficacy of pancreatic cancer chemopreventive agents (19, 22–27). In this study, we used pancreata from10-month-old Kras andwild-type mice (N ¼ 6/group). A significant increase in pancreasweight was observed in Kras mice compared with wild-type mice(Fig. 2A). As summarized in Fig. 2B, the Kras mice developedPanIN lesions (mean � SEM; PanIN1: 190 � 28, PanIN2: 172 �23, PanIN3: 160� 18), whereas no evidence of PanIN lesions wasobserved in wild-type mice. Out of 6 Kras mice, 4 mice (66%)showed carcinoma incidence, with 32.5% carcinoma spread

1H-NMR spectrum of talniflumate

Synthesis of talniflumate

ESI-HRMS spectrum of talniflumate

HPLC chromatogram of talniflumate

Reten�on �me– 29:52 (mm:ss)

A

B

C

D

001122334455667788991010

Talniflumate_3g_8132015_DMSO-d6

Calculated mass – 414.08Observed mass – 415.0943

Figure 3.Synthesis and purification of talniflumate. A, scheme showing synthesis of talniflumate. B, 1H-NMR spectrum of talniflumate. C, ESI-HRMS spectrum of talniflumate.D, HPLC chromatogram of talniflumate.






within the pancreas. Only 6% of the pancreata from Kras micewere free from PanIN lesions and PDAC, whereas 100% ofpancreata were normal in wild-type mice.

Gene expression profiling in pancreatic tumor tissues bynext-generation transcriptome analysis

We next collected the pancreata and determined GCNT3expression using PCR. We observed that tumors showed signif-icantly increased GCNT3 compared with normal pancreas.GCNT3 mRNA levels are significantly higher in the pancreatictumors of GEM compared with normal pancreas (Fig. 2C). Wethen determined GCNT3, EGFR, and top hits related to mucinexpression using transcriptome analysis (Fig. 2D and E). Weobserved a significant increase in the GCNT3 (103.16-folds;P < 0.0001) in correlation with increased mucins Muc4 (50-fold;P < 0.04), Muc5ac (87-fold; P < 0.01), Muc6 (67-fold; P < 0.008),Muc 1 (5-fold; P < 0.009), Muc 16 (5-fold; P < 0.0003), and Muc20 (17-fold; P < 0.007), along with EGFR (P < 0.001), in thepancreatic tumors from 10-month-old mice compared with pan-creata from wild-type mice, as determined by NGS/RNAseqanalysis (Fig. 2E). Further, IHC and mucin staining revealed highexpression of GCNT3 and mucins in PanIN lesions and PDAC(Fig. 2F). Importantly, the GCNT3 expression in the pancreatictissue was observed to increase progressively as the GEM age(Supplementary Fig. S2A and S2B)

Talniflumate is a novel drug that targets GCNT3We synthesized talniflumate for our studies using the following

method (Fig. 3A; refs. 28, 29).Quality andpurityweredeterminedusing NMR, HRMS, and HPLC (Fig. 3B–D). Using bioinformaticstools, the binding confirmation of the noncompetitive inhibitortalniflumate (Fig. 4A) to GCNT3was explored using in silico smallmolecular docking simulations. Our blind docking simulationsrevealed that talniflumate binds to GCNT3 with a better affinity

than does the reference ligand GALB1,3GALNAC. Although talni-flumate is positioned in the same pocket of GCNT3 that GALB1,3GALNAC occupies, they do not occupy the same exact space(data not shown). Talniflumate's best docking affinity of �8.3kcal/mol is achieved deeper in the pocket of GCNT3, whereasGALB1,3GALNAC's best docking affinity of �7.51 kcal/mol isachieved closer to the surface of GCNT3. The docking predic-tions suggest three notable hydrogen bonds between talniflu-mate and GCNT3: Arg192 (3.0 Å), Try288 (3.5 Å), and Ala287(2.9 Å). Although side chain hydrogen bond contacts aresuggested between both the hydrophilic residues, Arg192 andTyr288, the docking suggests a main chain hydrogen bondcontact between talniflumate and the hydrophobic Ala287(Fig. 4B–D). One drawback with performing rigid receptordocking simulations is that the side chains do not adjust tothe docked ligand, which could explain the longer hydrogenbonds between talniflumate and both Arg192 and Tyr288. Invivo, these side chains could move closer to the ligand, short-ening the hydrogen bond and, therefore, making this a strongerinteraction.

Talniflumate inhibits GCNT3 and mucinsWe observed a consistent expression of GCNT3 and mucins in

the humanpancreatic cancer specimens, GEM tissues, and humanpancreatic cancer cell lines that were analyzed (Figs. 1 and 2;Supplementary Figs. S2–S4).We then sought to determinewheth-er talniflumate inhibits GCNT3 and mucins in GEM in vivo.Pancreata from 6-week-old Kras mice treated with talniflumatefor 1 week showed a significant decrease in GCNT3 and mucinexpression in PanIN lesions (Fig. 5A). Further, GCNT3 proteinexpression was significantly decreased in the pancreas of talni-flumate-treated Kras mice compared with untreated controlmouse pancreas (Fig. 5B). mRNA expression of GCNT3 was alsoobserved to be lower in pancreatic tissues from talniflumate-treated mice (Fig. 5C). These results support that talniflumate

Figure 4.In silico molecular docking studies toshow binding of talniflumate toGCNT3. A, crystal structure oftalniflumate. B–D, binding oftalniflumate to GCNT3. Talniflumatebinds toGCNT3with a docking affinityof �8.3 kcal/mol and deeper in thepocket of GCNT3 (B and D). Thedocking predictions suggest threenotable hydrogen bonds betweentalniflumate and GCNT3: Arg192(3.0 Å), Try288 (3.5 Å), and Ala287(2.9Å; C).

Rao et al.





(400 ppm) can inhibit GCNT3, and, thereby, mucins, in GEM invivo. Further, we also used next-generation Illumina sequencing toevaluate pancreatic cancer in gefitinib-treated mice and controls(Fig. 5D–G). The results clearly demonstrate that gefitinib treat-ment (100 ppm) significantly decreases EGFR, GCNT3, andmucins in the pancreas. Gefitinib significantly reduced PanINlesions (PanIN1, 2, and 3) and carcinoma (Fig. 5F). Mice treatedwith gefitinib showed more normal pancreatic tissues (i.e., freefrom PanIN lesions and carcinoma) than did untreated mice(Supplementary Fig. S2C and S2D).

Combination of gefitinib and talniflumate inhibits GCNT3 andmucins

We further determined the expression ofmucins andGCNT3 inthe human pancreatic cancer Mia PaCa cells treated with talni-flumate, gefitinib, and the combination of talniflumate andgefitinib (Fig. 6). A significant decrease in cell viability, GCNT3expression, and number of mucin-expressing cells was seen after

treatment compared with untreated cells (Fig. 6). We alsoobserved a decrease in cell viability upon treatment with talni-flumate and gefitinib in PANC-1 andBxPc-3 cells (SupplementaryFigs. S3–S5). Along with GCNT3, individual and combinationtreatments also reduced mucins, PCNA, and cyclin D1 (Supple-mentary Fig. S4). These results justify the use of talniflumate formucin synthesis disruption, thereby enhancing the efficacy ofEGFR inhibitors for pancreatic cancer treatment.

CRISPRGCNT3-KO leads to reduced cell viability and spheroidformation

We next sought to determine whether knocking out GCNT3inhibits cell growth and spheroid formation. Pancreatic cancercells were knocked out for GCNT3 using CRISPR technologyfollowing the vendor's instruction. Transfection efficiency wasdetermined by GFP fluorescence in GCNT3-KO cells (Fig. 7A).After 48 hours, cells were harvested and evaluated for cell viabilityviaMTT assay (Fig. 7B) and spheroid formation in algimatrix 3D

GCNT3 - C GCNT3 - T

Mucins - C Mucins - T

GCNT3-mRNA C

T

A B D

F

E

mRN

A ex

pres

sion

EGFR5.45 foldP < 0.001

Control Gefi�nib

CC C T T

GCNT3

β-Ac�n

Control Talniflumate0.0

0.5

1.0

P < 0.04

Rel

ativ

e p

rote

in e

xpre

ssio

n (

GC

NT

3/A

ctin

)

G

80

60

UntreatedRel

ativ

e m

RN

A e

xpre

ssio

n (G

CN

T3)

1.0

0.5

0.0Talniflumate

P < 0.03

40

20

400

Nu

mb

er o

f P

anIN

Mea

n ±±

SE

in p

ancr

eas

300

200

100

0

0 ppm100 ppm

P <

0.0

01

P <

0.0

01

P <

0.0

1

PanIN 1 PanIN 2 PanIN 3

0

Figure 5.Effect of talniflumate and gefitinib on progression of PanIN lesions. A, GEM fed with talniflumate (400 ppm) for 1 week showed a decrease in GCNT3 and mucinexpression as determined by IHC (top) and Alcian blue staining (bottom). B, talniflumate decreased GCNT3 protein expression significantly as determined byWestern immunoblotting and quantified by Image-J software (N ¼ 6/group; C, control; T, talniflumate treated). C, talniflumate decreased mRNA expressionof GCNT3 in the pancreatic tissues of GEM. D and E, whole-genome transcriptome analysis by Illumina sequencing showing decreased mRNA expressionof EGFR, GCNT3, and mucins in pancreata from GEM mice treated with gefitinib. F, inhibition of PanIN lesions by gefitinib. G, heat map of the untreated andtreated groups showing differential gene expressions.






culture system (Fig. 7C). We observed a decrease in cell prolifer-ation by 32%onday 1 and by 55%onday 2. Further, talniflumatesignificantly reduced spheroid formation, and GCNT3-KO cellsdid not form significant numbers of spheroids (Fig. 7C). Further,implantingGCNT3-KO cells into nudemice for in vivo studies willhelp us to evaluate the tumor-forming ability of these cells.

DiscussionPancreatic cancer has an aggressive nature and ranks among the

worst in survival rates of all epithelial cancers. The difficulties inearly diagnosis and the lack of effective therapies contribute to itspoor prognosis. One of the challenges in pancreatic cancer ther-apy is the failure of existing drugs to access their target sites due toexcess mucin production. Clinical and preclinical studies haveshown that various mucins (Muc1, Muc4, Muc13, and Muc5ac)are significantly overexpressed in pancreatic tumors. A large body

of evidence supports the role of these mucins in tumor progres-sion. Some have been studied as biomarkers or targeted forpancreatic cancer treatment. Clinical studies suggest a correlationbetween the overexpression of mucins and pancreatic cancer(10–13), and have demonstrated a link between the aberrantexpression anddifferential overexpressionofmucin glycoproteinsand the initiation, progression, and poor prognosis of the disease(14–16, 19, 30). However, little advancement has been made inthis area. We sought to evaluate whether a mucin-synthesizinggene could be an ideal target for pancreatic cancer intervention,rather than simply targeting individual mucins. In this direction,we identified GCNT3 as a potential target that is significantlyoverexpressed in human pancreatic cancer and is known todecrease patient survival by 7 months (Fig. 1A). Further, 62.9%of the patients with high GCNT3 expression had a history ofalcohol use, tobacco use, and/or diabetes. Our results suggest thatexcessive mucin synthesis may be associated with these factors.

0

20

40

60

80

C T G T+G

P < 0.0001

P < 0.0001

% M

uci

n p

osi

tive

cel

ls

Untreated T+G

GCNT3C T G T+G

C - Untreated T - TalniflumateG - Gefi�nib

Ac�n

A

C

B

DConcentration

MIA PaCa-2 cell line – 24 h100

80

60

40

20

0

Via

bili

ty (

% o

f co

ntr

ol)

(µµmol/L)

Tal 50 µmol/L

Tal 100 µmol/L

Tal 150 µmol/L

Tal 200 µmol/L

Gef 0.1 µmol/L Gef 1 µmol/L Gef 10 µmol/L Gef 25 µmol/L

Figure 6.Effect of talniflumate and gefitinib on human pancreatic cancer cells. A, effect of gefitinib and talniflumate on cell viability. B, Western immunoblot showingthe effect of talniflumate and gefitinib on protein expression of GCNT3. Decreased GCNT3 protein expression was seen with individual and combined treatments.C and D, a significant decrease in mucin-positive cells was observed upon treatment of MiaPaCa cells with talniflumate and gefitinib.

Rao et al.





We have shown that p48Cre/þ-LSL-KrasG12D/þ GEM with Krasmutations display significant dysregulation of EGFR and upregu-lation of mucin biosynthesis in PanIN lesions and PDAC (Fig. 2;ref. 19). Further, next-generation transcriptome analysis revealedthat GCNT3 is aberrantly expressed in pancreatic tumors fromGEM compared with normal pancreas, and has increased mucinsubtypes. GCNT3 is a mucin core-synthesizing gene, and itsoverexpression is associated with pancreatic cancer tumorigenesisand enhanced proliferation (Fig. 3). Jones and colleagues (inSupplementary Table S9; 20) demonstrated that normal pancreasdo not express GCNT3, whereas GCNT3 is overexpressed signif-icantly in tumors from patients with pancreatic cancer. GCNT3 isalso expressed highly in human pancreatic cancer cell lines. Ourstudies, preliminary reports, and the available literature supportthe notion that GCNT3 is involved in mucin synthesis in pan-creatic cancer, and is an attractive novel target for pancreatic cancertreatment. However, no mucin synthesis inhibitors that affectGCNT3 have been tested, and there are no published data thatsupport a direct role of GCNT3 in alteringmucin synthesis duringpancreatic cancer growth.

Talniflumate is an orally available, small-molecule inhibitor,and a potential muco-regulator that inhibits mucin synthesis,blocks mucous overproduction in patients with chronic mucousrespiratory problems, and reduces cystic fibrosis (31, 32). Talni-flumate is also reported as calcium-activated chloride channel(hCLCA1/mCLCA3) blocker; it reduces mucin synthesis in cellculture and animal models. Its anti-inflammatory actions are viainhibition of cyclooxygenases and inhibit Cl-/HCO3- exchangeractivity. Further, it increased survival in a cystic fibrosis mousemodel of distal intestinal obstructive syndrome (31, 32). Thecombinationof these properties in anorally tolerateddrugmaybeuseful in the treatment of pancreatic cancer. Although talniflu-mate was shown to exert its effects by an anti-inflammatorymechanism, studies have not supported its role as a mucousinhibitor in pancreatic cancer. Hence, after synthesizing talniflu-mate and fully characterizing it using NMR, HMR, and HPLCpurification, we used in silico molecular modeling approaches toinvestigate the binding capability of talniflumate to GCNT3.

Compared with known ligand GALB1,3GALNAC, talniflumateshowed a best docking affinity of �8.3 kcal/mol toward GCNT3,with three notable hydrogen bonds between GCNT3 and talni-flumate (Fig. 4; Supplementary Fig. S6). After confirming thattalniflumate effectively binds to GCNT3, we evaluated the effectsof talniflumate on GCNT3 expression in PanIN lesions fromGEM. Talniflumate significantly reduced GCNT3 protein expres-sion, mucins, and GCNT3 mRNA compared with untreated con-trols. These data further confirm that talniflumate has an effect onGCNT3.

Use of the EGFR inhibitor gefitinib (Iressa) significantlyreduced tumorweight, PanINs, andPDAC,with anaccompanyingsignificant decrease in mucin synthesis and GCNT3, suggesting apotential role of EGFR in the regulation of mucin biosynthesisthrough GCNT3 (Fig. 3). Given that clinical doses of most FDA-approved EGFR inhibitors (gefitinib) are associated with skintoxicity, itwouldbebeneficial to use lower doses of gefitinib alongwith talniflumate to produce synergistic effects on inhibition oftumor cell proliferation. The combination of talniflumate andgefitinib showeda significant decrease inGCNT3andmucinswithreduced cell proliferation. Further, talniflumate inhibited spher-oid formation, and CRISPR transfection of GCNT3 decreased cellproliferation and spheroid formation.

However, further studies are warranted to evaluate the mech-anism of GCNT3 regulation and talniflumate's effects on pancre-atic cancer development in detail. Understanding the function ofGCNT3 and mucin synthesis in the setting of pancreatic cancer iscritically important to understanding the mechanisms of pancre-atic cancer tumorigenesis and to developing new targeted thera-pies that reduce chemoresistance.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: C.V. Rao, N.B. Janakiram, V. Madka, A. MohammedDevelopment of methodology: N.B. Janakiram, V. Madka, G. Kumar,A. Mohammed

GC

NT

3-K

O

24 h 48 hA B

CControl Talniflumate GCNT3-KO

Day 1

P < 0.001

P < 0.0001

Pan

c-1

cell

viab

ility

100

80

60

40

20

GC

NT

3-K

O

0Day 2

Figure 7.GCNT3-KO by CRISPR decreases cellviability and spheroid formation. A,CRISPR knockout of GCNT3 is shownby the green fluorescence after 24and 48 hours of transfection. B, adecrease in cell viability was seenwith GCNT3-KO. C, talniflumatesignificantly reduced the number ofspheroids after 3 weeks of treatment.The GCNT3-KO cells did not yieldspheroids.






Acquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): C.V. Rao, N.B. Janakiram, V. Madka, G. Kumar,E.J. Scott, G. Pathuri, T. Bryant, H. Kutche, Y. Zhang, H. Gali, A. MohammedAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): C.V. Rao, N.B. Janakiram, V. Madka, G. Kumar,E.J. Scott, Y.D. Zhao, S. Lightfoot, A. MohammedWriting, review, and/or revision of the manuscript: C.V. Rao, N.B. Janakiram,V. Madka, Y.D. Zhao, A. MohammedAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): C.V. Rao, G. Kumar, E.J. Scott, T. Bryant,H. Kutche, Y. Zhang, L. Biddick, A. MohammedStudy supervision: C.V. Rao, A. Mohammed

AcknowledgmentsThe authors thank the OUHSC Rodent Barrier Facility staff, Kathy Kyler for

valuable suggestions and editorial help, the Stephenson Cancer Center at theOUHSC, Oklahoma City, OK, for use of the Biostatistics Core, which providedstatistical analysis services, and the cancer tissue pathology core for providing

human pancreatic cancer tissues. Transcriptome analysis was supported by theNational Center for Research Resources and the National Institute of GeneralMedical Sciences of the NIH through grant number 8P20GM103447. Finally,they thank the Laboratory for Genomics and Bioinformatics at the OUHSC forproviding core services for sequencing of pancreatic tissues.

Grant SupportThis work was supported by the College of Medicine Alumni Association

Grant (A. Mohammed) and Kerley-Cade Endowed Chair Fund and NCINCI-N01-CN53300 (C.V. Rao).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received October 13, 2015; revised December 31, 2015; accepted January 18,2016; published OnlineFirst February 15, 2016.

References1. American Cancer Society. Cancer facts and figures. Atlanta, GA: ACS; 2015.2. Kalr AV, Campbell RB. Mucin impedes cytotoxic effect of 5-FU against

growth of human pancreatic cancer cells: overcoming cellular barriers fortherapeutic gain. Br J Cancer 2007;8:910–8.

3. Kalra AV, Campbell RB. Mucin overexpression limits the effectiveness of5-FU by reducing intracellular drug uptake and antineoplastic drug effectsin pancreatic tumours. Eur J Cancer 2009;45:164–73.

4. Jonckheere N, Skrypek N, Seuningen IV. Mucins and Pancreatic cancer.Cancers 2010;2:1794–812.

5. SkrypekN,Duchene B, HebbarM, Leteurtre E, Seuningen IV, Jonckheere N.The MUC4 mucin mediates gemcitabine resistance of human pancreaticcancer cells via the concentrative nucleoside transporter family. Oncogene2013;32:1714–23.

6. Jonckheere N, Skrypek N, Seuningen IV. Mucins and tumor resistance tochemotherapeutic drugs. Biochim Biophys Acta 2014;1846:142–51

7. Mimeault M, Johansson SL, Senapati S, Momi N, Chakraborty S, Batra SK.MUC4 down-regulation reverses chemoresistance of pancreatic cancerstem/progenitor cells and their progenies. Cancer Lett 2010;295:69–84.

8. Wissniowski TT, Meister S, Hahn EG, Kalden JR, Voll R, Ocker M. Mucinproduction determines sensitivity to bortezomib and gemcitabine inpancreatic cancer cells. Int J Oncol 2012;40:1581–9.

9. Nath S, Daneshvar K, Roy LD, Grover P, Kidiyoor A, Mosley L, et al. MUC1induces drug resistance in pancreatic cancer cells via upregulation ofmultidrug resistance genes. Oncogenesis 2013;2:e51.

10. Kaur S, Kumar S, Momi N, Sasson AR, Batra SK. Mucins in pancreatic cancerand itsmicroenvironment.NatRevGastroenterolHepatol 2013;10:607–20.

11. Nagata K, Horinouchi M, Saitou M, Higashi M, Nomoto M, Goto M, et al.Mucin expression profile in pancreatic cancer and the precursor lesions.J Hepatobiliary Pancreat Surg 2007;14:243–54.

12. Moniaux N, Andrianifahanana M, Brand RE, Batra SK. Multiple roles ofmucins in pancreatic cancer, a lethal and challenging malignancy. Br JCancer 2004;91:1633–8.

13. Torres MP, Chakraborty S, Souchek J, Batra SK. Mucin-based targetedpancreatic cancer therapy. Curr Pharm Des 2012;18:2472–81.

14. Jabbar KS, Verbeke C, Hyltander AG, Sj€ovall H, Hansson GC, Sadik R.Proteomic mucin profiling for the identification of cystic precursors ofpancreatic cancer. J Natl Cancer Inst 2014;106:djt439.

15. Strobel O, Rosow DE, Rakhlin EY, Lauwers GY, Trainor AG, Alsina J, et al.Pancreatic duct glands are distinct ductal compartments that react tochronic injury and mediate Shh-induced metaplasia. Gastroenterology2010;138:1166–77.

16. Besmer DM, Curry JM, Roy LD, Tinder TL, Sahraei M, Schettini J, et al. et al.Pancreatic ductal adenocarcinoma mice lacking mucin 1 have a profounddefect in tumor growth and metastasis. Cancer Res 2011;71:4432–42.

17. Yonezawa S, Nakamura A, Horinouchi M, Sato E. The expression of severaltypes of mucin is related to the biological behavior of pancreatic neo-plasms. J Hepatobiliary Pancreat Surg 2002;9:328–41.

18. Kim GE, Bae HI, Park HU, Kuan SF, Crawley SC, Ho JJ, et al. Aberrantexpression of MUC5AC and MUC6 gastric mucins and sialyl Tn antigen in

intraepithelial neoplasms of the pancreas. Gastroenterology 2002;123:1052–60.

19. MohammedA, JanakiramNB, LiQ,MadkaV, ElyM, Lightfoot S, et al. EGFRinhibitor gefitinib prevents progression of pancreatic lesions to carcinomain a conditional LSL-K-rasG12D transgenic mice model. Cancer Prev Res2010;3:1417–26.

20. Jones S, Zhang X, Parsons DW, Lin JC, Leary RJ, Angenendt P, et al. Coresignaling pathways in human pancreatic cancers revealed by global geno-mic analyses. Science 2008;321:1801–6.

21. BeumPV, Singh J, BurdickM,HollingsworthMA, Cheng PW. Expression ofCore 2 b-1,6-N-Acetylglucosaminyltransferase in a human pancreaticcancer cell line results in altered expression of MUC1 tumor-associatedepitopes. J Biol Chem 1999;274:24641–8.

22. Mohammed A, Qian L, Janakiram NB, Lightfoot S, Steele VE, Rao CV.Atorvastatin delays progression of pancreatic lesions to carcinoma byregulating PI3/AKT signaling in p48Cre/1 LSL-KrasG12D/1 mice. Int JCancer 2012;131:1951–62.

23. Mohammed A, Janakiram NB, Brewer M, Duff A, Lightfoot S, Brush RS,et al. Endogenous n-3 polyunsaturated fatty acids delay progression ofpancreatic ductal adenocarcinoma in fat-1.P48Cre/þ-LSLKrasG12D/þmice. Neoplasia 2012;14:1249–59.

24. RaoCV,MohammedA, JanakiramNB,Qian L, Ritchie RL, Lightfoot S, et al.Inhibition of pancreatic intraepithelial neoplasia progression to carcinomaby nitric oxide-releasing aspirin in p48Cre/þ-LSL-KrasG12D/þ Mice.Neoplasia 2012;14:778–87.

25. Hocker JR, Mohammed A, Aston CE, BrewerM, Lightfoot SA, Rao CV, et al.Mass profiling of serum to distinguishmicewith pancreatic cancer inducedby a transgenic Kras mutation. Int J Cancer 2013;133:2662–71.

26. Mohammed A, Janakiram NB, Brewer M, Ritchie RL, Marya A, Lightfoot S,et al. Antidiabetic drug metformin prevents progression of pancreaticcancer by targeting in part cancer stem cells and mTOR signaling. TranslOncol 2013;6:649–59.

27. Mohammed A, Janakiram NB, Madka V, Ritchie RL, Brewer M, Biddick L,et al. Eflornithine (DFMO) prevents progression of pancreatic cancer bymodulating ornithine decarboxylase signaling. Cancer Prev Res 2014;7:1198–209.

28. Los M. Process for the preparation of alkyl and aryl esters of 30-substitutedand 20, 30-disubstituted 2-anilino-3-pyridinecarboxylic acids. United Statespatent 4255581. 1981 Mar 10 .

29. LosM, Boned J E, Piccininali C. Nuevos esteresde acidos anilinonicotinicosY N-fenilantranilicos sustituidos (New esters of substituted anilinonico-tinic and phenylanthranilic acids). Farmaco Ed ScL 1981;36:372–85.

30. Hingorani SR, Petricoin EF, Maitra A, Rajapakse V, King C, Jacobetz MA,et al. Preinvasive and invasive ductal pancreatic cancer and its earlydetection in the mouse. Cancer Cell 2003;4:437–50.

31. Knight DTalniflumate (Genaera). Curr Opin Invest Drugs 2004;5:557–62.32. Walker NM, Simpson JE, Levitt RC, Boyle KT, Clarke LL. Talniflumate

increases survival in a cystic fibrosis mouse model of distal intestinalobstructive syndrome. J Pharmacol Exp Ther 2006;317:275–83.


Rao et al.




2016;76:1965-1974. Published OnlineFirst February 15, 2016.Cancer Res Chinthalapally V. Rao, Naveena B. Janakiram, Venkateshwar Madka, et al. and Malignant Cellular Behaviors in Pancreatic CancerSmall-Molecule Inhibition of GCNT3 Disrupts Mucin Biosynthesis

Updated version

10.1158/0008-5472.CAN-15-2820doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2016/02/13/0008-5472.CAN-15-2820.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/76/7/1965.full#ref-list-1

This article cites 30 articles, 5 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/76/7/1965.full#related-urls

This article has been cited by 1 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/76/7/1965To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-15-2820

http://cancerres.aacrjournals.org/content/suppl/2016/02/13/0008-5472.CAN-15-2820.DC1

http://cancerres.aacrjournals.org/content/76/7/1965.full#ref-list-1

http://cancerres.aacrjournals.org/content/76/7/1965.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/76/7/1965


Documents

Small-Molecule Inhibition of GCNT3 Disrupts Mucin ... · Chinthalapally V. Rao1, Naveena B. Janakiram1,Venkateshwar Madka1, Gaurav Kumar1, Edgar J. Scott 2 , Gopal Pathuri 1 ,Taylor