GALNT9 Gene Expression Is a Prognostic Marker in Neuroblastoma ...

GALNT9 Gene Expression Is a Prognostic Marker inNeuroblastoma Patients

Nora Berois,1*† Charles-Henry Gattolliat,2† Enrique Barrios,3 Laura Capandeguy,1 Setha Douc-Rasy,2

Dominique Valteau-Couanet,4 Jean Benard,2,5* and Eduardo Osinaga1,6

BACKGROUND: The enzymes encoded by the GALNT[UDP-N-acetyl-�-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase (GALNAC-T)] genefamily catalyze the first step of O-glycosylation. Little isknown about the link between expression of the genesencoding GALNAC-T enzymes and tumor progressionin neuroblastoma, a pediatric cancer that can be classi-fied as either low or high risk. We assessed the expres-sion of genes in the GALNT family in a large cohort ofneuroblastoma patients and characterized members ofthis family that might be used as new prognosticmarkers.

METHODS: Reverse-transcription PCR analysis of 14GALNT genes with a panel of neuroblastoma cell linesidentified the GALNT9 gene as playing a potential rolein disease progression. We used the log-rank test andthe multivariable Cox proportional hazards modelwith a cohort of 122 neuroblastoma patients to analyzethe relationship between GALNT9 expression andoverall survival or disease-free survival.

RESULTS: In the high-risk neuroblastoma experimentalmodel IGR-N-91, GALNT9 expression was present inneuroblasts derived from primary tumors but not inneuroblasts from metastatic bone marrow. Moreover,GALNT9 in neuroblastoma cell lines was expressed insubstrate adherent (S)-type cell lines but not in neuro-nal (N)-type lines. In the tumor cohort, GALNT9 ex-pression was associated with high overall survival,independent of the standard risk-stratification covari-ates. GALNT9 expression was significantly associatedwith disease-free survival for patients currently classi-fied as at low risk (P � 0.0007).

CONCLUSIONS: GALNT9 expression correlates with bothimproved overall survival in low- and high-risk groupsand an improved clinical outcome (overall and disease-free survival) in low-risk patients. Thus, the GALNT9 ex-pression may be a prognostic marker for personalizedtherapy.© 2012 American Association for Clinical Chemistry

Neuroblastoma, the most common malignant solid tu-mor diagnosed during infancy (median age at diagno-sis, 17 months), accounts for 10% of all childhood can-cers (1 ). Its biology may be extremely variable: Certaintumors regress spontaneously, whereas others arehighly aggressive. The combination of age at diagnosis,tumor burden, histopathology, DNA index, andMYCN7 [v-myc myelocytomatosis viral related onco-gene, neuroblastoma derived (avian)] gene status isused to stratify risk categories (2 ). The low-risk groupconsists of non-MYCN–amplified tumors that exist ei-ther as localized forms (stages 1, 2, and 3) or as meta-static forms in children younger than 18 months (stage4S). The high-risk group comprises all MYCN-amplified neuroblastomas, regardless of stage and ageof the child, plus non-MYCN–amplified stage 4 neuro-blastomas for children older than 18 months. The low-risk group has a survival rate of up to 90%, but thehigh-risk group consists of aggressive tumors thatmore frequently lead to death (5-year survival rate, ap-proximately 35%).

Alterations in glycan profiles are a hallmark ofcancer development linked to the expression oftumor-associated carbohydrate antigens (3 ). Ab-normal O-glycans produced by cancer cells contrib-ute to the malignant phenotype and play an impor-

1 Laboratorio de Glicobiologıa e Inmunologıa Tumoral, Institut Pasteur de Mon-tevideo, Montevideo, Uruguay; 2 CNRS UMR 8126, Universite Paris-Sud 11,Institut Gustave Roussy, Villejuif, France; 3 Departamento de Metodos Cuanti-tativos, Universidad de la Republica, Montevideo, Uruguay; Departement de4 Pediatrie Oncologique and 5 Biologie et Pathologie Medicales, Institut GustaveRoussy, Villejuif, France; 6 Departamento de Inmunobiologıa, Universidad de laRepublica, Montevideo, Uruguay.

* Address correspondence to: N.B. at Laboratorio de Glicobiologıa e InmunologıaTumoral, Institut Pasteur de Montevideo, Mataojo 2020, Montevideo CP 11400,Uruguay. Fax �598-25-22-41-85; e-mail [email protected]. J.B. at De-

partement de Biologie et Pathologie Medicales, Institut Gustave Roussy, 39 rueCamille Desmoulins, 94805 Villejuif Cedex, France. Fax �33-1-42-11-54-94;e-mail [email protected].

† These authors contributed equally to this work.Received July 6, 2012; accepted October 15, 2012.Previously published online at DOI: 10.1373/clinchem.2012.1923287 Human genes: MYCN, v-myc myelocytomatosis viral related oncogene, neuro-

blastoma derived (avian); GALNT3, UDP-N-acetyl-�-D-galactosamine:polypep-tide N-acetylgalactosaminyltransferase 3 (GALNAC-T3); B2M, �2-microglobulin.

Clinical Chemistry 59:1225–233 (2013)

Cancer Diagnostics

225

tant functional role in cell adhesion, invasion, andmetastasis (4 ). The most abundant form of O-linkedglycosylation in higher eukaryotes, termed “mucin-type,” is initiated in the Golgi apparatus by thecovalent linkage of an �-N-acetylgalactosamine(GalNAc)8 residue to the hydroxyl group of Ser andThr residues (5 ), a reaction catalyzed by the UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferases(GALNAC-Ts) (EC 2.4.1.41) (6 ). These enzymes are acomplex family of isoenzymes, with 20 members hav-ing been characterized to date (7 ). The genes encodingcertain isoforms are broadly expressed, whereas otherisoforms are restricted in their distributions and activ-ities to certain cells or tissues (6). Individual GALNAC-Tshave distinct activities (8), and the O-glycosylation pro-cess in a given cell is regulated by the repertoire of theisoenzyme genes expressed in that cell (4).

The various genes encoding GALNAC-Ts aredifferentially expressed in malignant tissue, com-pared with normal tissue (9 –11 ). Overexpressionof the GALNT3 [UDP-N-acetyl-�-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase 3(GALNAC-T3)] gene promotes the growth of pancre-atic cancer cells (12 ), whereas somatic and germlinemutations in GALNT12 that encode a nonfunctionalenzyme are associated with colon cancer development(13 ). Increasing evidence suggests that GALNAC-Tsmight be useful tumor markers. GALNAC-T3 enzymeproduction, for example, has been shown to correlatewith prognosis in patients with colorectal and gall blad-der cancer (14, 15 ). The gene encoding GALNAC-T6 isexpressed in breast cancer but not in normal breastepithelium, and this difference is apparent at both themRNA (16 ) and protein (17 ) levels. In previous work,we used a reverse-transcription PCR (RT-PCR) assayto identify an association between GALNAC-T6 pro-duction in bone marrow samples and poor clinical out-come in lymph node–negative breast cancer patients(16 ). We have also shown that GALNT13, the gene en-coding the GALNAC-T13 isoenzyme (18 ), might be anovel molecular marker of bone marrow involvementin neuroblastoma patients (19 ). When we applied mi-croarray gene expression analysis to the xenograft-derived cell model of human neuroblastoma [the so-called IGR-N-91 cell line (20 )], we also found thatGALNT13 was the gene most strongly upregulated inmetastatic malignant neuroblasts, compared with a

primary-tumor xenograft (19 ). In the same study, wemeasured tyrosine hydroxylase, ganglioside D2 syn-thase, DOPA decarboxylase, and GALNT13 transcriptsin bone marrow aspirates from the same neuroblas-toma patients to evaluate whether GALNT13 might beuseful for detecting disseminated disease. We foundthat GALNT13 expression in bone marrow at diagnosiswas the strongest predictor of a poor clinical outcome.This evidence prompted the present study, in which weanalyzed the expression of 14 GALNT genes at the mRNAlevel in a panel of neuroblastoma cell lines, including theIGR-N-91 model. Because we found the brain-specificisoform, encoded by GALNT9 (21), to be produced pri-marily in neuroblasts displaying a low aggressive behav-ior, we assessed the production of this enzyme in a cohortof 122 primary neuroblastoma tumors. Our findings in-dicate that GALNT9 gene expression is associated with abetter clinical outcome in neuroblastoma patients.

Materials and Methods

PATIENTS AND CLINICAL SAMPLES

Tumor samples from a retrospective cohort of patients(n � 122) who had been staged according to the Inter-national Neuroblastoma Staging System (22 ) were col-lected between 1987 and 2009 at the Institut GustaveRoussy with the approval of the appropriate ethicscommittees and according to the national law applica-ble to people taking part in biomedical research.Primary-tumor tissues obtained from patients eithervia Tru-Cut (CareFusion Corporation) biopsy or aftersurgery were immediately snap-frozen and stored inliquid nitrogen until nucleic acid extraction. The mainclinical features of the patients and the tumor charac-teristics are described in Table 1 in the Data Supple-ment that accompanies the online version of this articleat http://www.clinchem.org/content/vol59/issue1.

TUMOR CRYOSECTIONS AND EXTRACTION OF TOTAL RNA

The first and last cryosections were used to select tumortissues with a malignant tumor cell content of �60%(23, 24). Total RNA was isolated with TRIzol Reagent(Invitrogen) according to the manufacturer’s proto-cols. Nucleic acid concentration and purity were mea-sured with a Nanodrop ND-1000 spectrophotometer(NanoDrop/Thermo Scientific), and nucleic acid qual-ity was checked with a 2100 Bioanalyzer (AgilentTechnologies).

CELL LINES

The neuroblastoma metastatic-cell model, IGR-N-91,was derived from a high-risk neuroblastoma throughin vitro culture of malignant neuroblasts that had beencollected from the bone marrow of an 8-year-old boyand then xenografted into nude mice, as has been de-

8 Nonstandard abbreviations: GalNAC, �-N-acetylgalactosamine; GALNAC-T,UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase; RT-PCR, reverse-transcription PCR; PTX, primary-tumor xenograft; BM, bone marrow metastasis;N, neuronal; S, substrate-adherent; dNTP, deoxynucleoside triphosphate;Apo2L/TRAIL, apolipoprotein 2 ligand/tumor necrosis factor–related apoptosis-inducing ligand.

226 Clinical Chemistry 59:1 (2013)

scribed (20 ). Derived sublines were established fromthe primary-tumor xenograft (PTX) and bone marrowmetastasis (BM) in nude mice. Other neuroblastomacell lines [IMR-32, SH-SY5Y, SK-N-AS, SK-N-BE (2 ),LAN-1, and LAN-5] showing the typical phenotypes ofestablished malignant neuroblasts [i.e., neuronal (N)and substrate-adherent (S)] were obtained from vari-ous sources (25 ). The cells were maintained in DMEMat 37 °C in a 5% CO2 humidified atmosphere. TotalRNA was isolated with TRIzol Reagent according to themanufacturer’s protocols.

RT-PCR ANALYSIS

Total RNA (1 �g) prepared from cell lines and primarytumors was reverse transcribed by Moloney murineleukemia virus reverse transcriptase (Amersham/GEHealthcare Life Sciences). The reaction mixture con-sisted of 200 U enzyme, 2 �L of 10 mmol/L of eachdeoxynucleoside triphosphate (dNTP), and 1 �L of250 ng random hexamers in a 20-�L total reaction vol-ume. Different RT-PCR reactions were run with theappropriate negative controls to amplify GALNT fam-ily members. The primer sequences are shown in Table2 in the online Data Supplement. The B2M (�2-microglobulin) gene was amplified to verify cDNAquality. We added 1 �L cDNA to a final 25-�L PCRreaction volume containing 1� provided enzyme buf-fer, 2 mmol/L MgCl2, 200 �mol/L dNTPs, 400 nmol/Lof each primer, and 1 U Taq DNA polymerase (Fer-mentas). The amplification conditions consisted of 35cycles of 30 s at 95 °C, 30 s at the annealing tempera-ture, and 1 min at 72 °C. The annealing temperaturewas 60 °C for GALNT1, 2, 4, 5, 6, 7, 9, and 15; 56 °C forGALNT3, 10, 11, 12, and 14; and 62 °C for GALNT13.The PCR products (15 �L) were analyzed by electro-phoresis on 2% agarose gels and by direct visualizationafter ethidium bromide staining.

STATISTICAL ANALYSIS

We performed several analyses to determine whetherthe expression levels of the GALNT genes of interestwere significantly related to overall and disease-freesurvival of the patients (end points of cancer-specificdeath and relapse, respectively). Overall survival wasdefined as the time from diagnosis to the date of deathor last follow-up; disease-free survival was defined asthe time from diagnosis to the date of death or the firstappearance of relapse. SAS software (SAS Institute) wasused to construct Kaplan–Meier survival curves and toperform univariate and multivariable Cox propor-tional hazards analyses. To generate survival curves, westratified patients into groups of GALNT gene expres-sion (i.e., expression, �; lack of expression, �). Sur-vival curves were compared by the log-rank test andjudged significantly different at P values �0.05. We

undertook univariate Cox proportional hazards re-gression analyses with SAS software and defined statis-tical significance as P values �0.05. To ascertainwhether expression of the GALNT genes of interestwere independent prognostic factors, we used Coxproportional hazards regression analyses to examinethe joint effects of covariates. Patients were dichot-omized by age into patients younger than 18 monthsand patients 18 months of age and older. Stage wasdichotomized for the entire neuroblastoma cohortwith respect to metastatic progression (i.e., stage 1,2, and 3 vs stage 4. Stage 4S was included with stages1, 2, and 3, because the prognosis for these patients isusually considered good. For the GALNT gene ofinterest, patients were categorized into groups as de-scribed above. All variables with P values �0.05 inunivariate analyses were selected for inclusion in themultivariable model.

Results

EXPRESSION OF THE GALNT GENE FAMILY IN HUMAN

NEUROBLASTOMA CELL LINES

We investigated the production of GALNAC-Ts in theIGR-N-91 neuroblastoma experimental model, whichoffered the opportunity to compare neuroblastoma celllines established from a primary tumor (PTX) andmatched metastatic neuroblasts (BM). An RT-PCR as-say was designed for 14 members of the GALNT genefamily (GALNT1 to GALNT15 except GALNT8). Incomparing the 2 cell sublines, we found that severalGALNT genes were expressed in both, (PTX and BMneuroblasts expressed GALNT1, 2, 4, 6, 7, 12, and 14),but the expression of 3 genes (GALNT3, 5, and 10) wereundetectable in these 2 cell lines. We found a markeddifference between PTX and BM malignant neuro-blasts in the expression GALNT9, 11, 13, and 15. TheGALNT11 and GALNT13 genes were expressed in theBM cell line but not in the PTX cell line, and conversely,GALNT9 and GALNT15 genes were expressed in thePTX line but not in the BM cell line (Fig. 1). Conse-quently, we measured the expression of these genesin a panel of 6 neuroblastoma cell lines. GALNT11,13, and 15 were found in almost all the cells lines andhence were broadly expressed. In sharp contrast,GALNT9 displayed a very restricted expression pat-tern, being found only in SK-N-AS and SK-N-BE (2 )cell lines (Fig. 2). These results led us to chooseGALNT9 gene for further evaluation of clinicalsamples.

GALNT9 EXPRESSION IN PRIMARY HUMAN NEUROBLASTOMA

TUMORS

On the basis of the in vitro results, we used RT-PCR toassess GALNT9 expression in samples of primary neu-

GALNT9 Gene Is a Prognostic Marker in Neuroblastoma

Clinical Chemistry 59:1 (2013) 227

roblastoma tumors from a tumor cohort from 122 pa-tients (Table 1; see Table 1 in the online Data Supple-ment). We found GALNT9 expression in 77 (63%) ofthe 122 tumors. We then investigated whether the lackof GALNT9 expression was associated with variousclinical features (Table 2). The relationship betweendisease stage and GALNT9 was not significant (P �0.0596). GALNT9 was expressed not only in most pa-tients with localized and 4S stages (47 of 66 patients,71%) but also in 30 (54%) of 56 patients with stage 4disease. GALNT9 was expressed in most patientswith no MYCN amplification (66 of 96 patients,69%) and in patients younger than 18 months atdiagnosis (42 of 56 patients, 75%); however,GALNT9 expression occurred in only 42% of pa-tients (11 of 26) with MYCN amplification and 53%of patients (35 of 66) �18 months. Thus, the lack ofGALNT9 expression was associated with MYCN am-plification (P � 0.0209) and age at diagnosis (P �0.0147), indicating a link between lack of GALNT9expression and the high-risk phenotype.

EXPRESSION OF GALNT9 AS A PROGNOSTIC MARKER OF PATIENT

SURVIVAL

Statistical analyses were performed to determine thesignificance of GALNT9 expression as a prognostic fac-tor for the entire cohort (n � 122). Disease outcome(survival and relapse events) was documented for allpatients. The final follow-up date was February 2011,with a median follow-up period of 51 months. Anevent (death or relapse) was registered for 54 patients(44%), with death recorded for 42 patients (34%). A uni-variate Cox proportional hazards regression analysisshowed that patients with a lack of GALNT9 expression intheir tumors had an overall and disease-free survival ratesignificantly lower than patients whose tumors expressedthat gene (Table 3). Consistent with the results of the uni-variate Cox regression analysis, the Kaplan–Meier sur-vival analysis highlighted that patients with GALNT9 ex-pression displayed a better prognosis than patients withno expression, in terms of both overall survival anddisease-free survival (P � 0.0001, and P � 0.0004, respec-tively; Fig. 3A). Analyses with the multivariable Cox pro-

Fig. 1. GALNT gene expression in the neuroblastoma metastatic model IGR-N-91 cell lines.

GALNT genes are presented as follows: GALNT1 is T1, GALNT2 is T2, and so forth. (A), Summary table of the expression of 14GALNT genes in the parental cell line IGR-N-91, as well as in a PTX cell line and a BM cell line. (B), Agarose gel electrophoresisimages showing representative results: 1, 100-bp molecular weight marker; 2, negative control; 3, IGR-N-91 cell line; 4, PTXcell line; 5, BM cell line; 6, positive control (IMR-32 cell line for GALNT1, T2, T4, T6, T7, T11, T12, T13, T14, and T15; HT29cell line for GALNT3; K562 cell line for GALNT5; brain for GALNT9; and LS174T cell line for GALNT10).

228 Clinical Chemistry 59:1 (2013)

portional hazards model assessed whether GALNT9 ex-pression was an independent prognostic factor (Table 3)and found it to be an independent prognostic factor foroverall survival (hazard ratio, 2.066; 95% CI, 1.078–

3.963; P � 0.0290) but not for disease-free survival (haz-ard ratio, 1.446; 95% CI, 0.824–2.538; P � 0.1986).

GALNT9 EXPRESSION PREDICTS SURVIVAL FOR BOTH HIGH- AND

LOW-RISK NEUROBLASTOMA PATIENTS

We first investigated the association of GALNT9 ex-pression with survival in the group of high-risk patients(n � 58). Kaplan–Meier curves showed a significantassociation of GALNT9 expression with overall sur-vival (P � 0.0391) but not with disease-free survival(P � 0.4178) (Fig. 3B).

GALNT15 368 bp MW 500 bp

GALNT9 223 bp

GALNT13 425 bp MW 500 bp

MW 500 bpGALNT11 440 bp

MW 500 bp A

B

C

D

ββ2-microglobulin 600 bp MW 500 bp

E

MW 2 3 4 5 6 7 8 9 10 11 121

Fig. 2. Representative results of GALNT RT-PCR analyses of neuroblastoma cell lines and tumors.

RT-PCR for GALNT9 (A), GALNT11 (B), GALNT13 (C), GALNT15 (D), and �2-microglobulin (“housekeeping” gene control) (E).MW, molecular weight marker (100 bp); 1, negative control; 2, IGR-N-91 cell line; 3, PTX cell line; 4, BM cell line; 5, IMR-32cell line; 6, SH-SY5Y cell line; 7, SK-N-AS cell line; 8, SK-N-BE (2) cell line; 9, LAN-1 cell line; 10, LAN-5 cell line; 11,neuroblastoma primary tumor 1; 12, neuroblastoma primary tumor 2.

Table 1. Characteristic of patients and tissues.

Patients/tissues,n (%)

No. of tissues 122

Stage

1, 2, 3, 4S 66 (54)

4 56 (46)

Age at diagnosis

�18 months 56 (46)

�18 months 66 (54)

MYCN status

Not amplified 96 (79)

Amplified 26 (21)

No. of patients 122

Present status

Alive 80 (66)

Death from disease 42 (34)

Disease-free status

Disease free 68 (56)

Disease 54 (44)

Table 2. Analysis of the relationship betweenvarious clinical characteristics, survival, and GALNT9

expression with the 2-sided Fisher exact test(n � 122).

Characteristics Cases GALNT9 (�) GALNT9 (�) P

Stage

1, 2, 3, 4S 66 47 19 0.0596

4 56 30 26

MYCN status

Not amplified 96 66 30 0.0209

Amplified 26 11 15

Age at diagnosis

�18 months 56 42 14 0.0147

�18 months 66 35 31

GALNT9 Gene Is a Prognostic Marker in Neuroblastoma

Clinical Chemistry 59:1 (2013) 229

Finally, we investigated the prognostic value ofGALNT9 expression in low-risk patients (n � 64).Kaplan–Meier survival curves revealed that the clinicaloutcome (overall and disease-free survival) of patientswith GALNT9 expression was strongly correlated withbetter survival than patients with no GALNT9 expres-sion (P � 0.0001, and P � 0.0007, respectively) (Fig.3C).

Discussion

Altered glycosylation is a universal feature of cancercells, and some glycan structures are well-known mark-ers of tumor progression. Several studies have linkedspecific ganglioside changes in human neuroblastomatumors to differences in biological behavior and clini-cal outcome (26, 27 ). We previously reported thatGALNT13 is highly expressed (12 times higher) in met-

astatic neuroblasts than in primary tumors (19 ). Thepresent study has shown that GALNT9 is expressed inPTX cells but not in BM cells in the IGR-N-91 model.This gene is also expressed in less aggressive S-type neu-roblasts (SK-N-AS cell line), but not in N-type neuro-blasts (SH-SY5Y, IMR-32, LAN-1, and LAN-5 celllines). These findings could be of particular impor-tance, because S-type cells have been reported to ex-hibit weaker invasiveness and metastatic properties, alower growth capability in vivo (28 ), and a higherspontaneous apoptosis rate, compared with N-typecells (29 ).

GALNT9 expression was initially thought to be ex-clusive to the brain, with mRNA concentrations beingmore abundant in the cerebellum than in the cerebralcortex, frontal lobe, temporal lobe, and putamen (21 ).GALNT9, GALNT8, GALNT18, and GALNT19 consti-tute a subfamily that differs markedly in sequence from

Table 3. Univariate and multivariable Cox regression analysis of survival in all neuroblastoma patients (n � 122).

Univariate analysis Multivariable analysis

Hazard ratio (95% CI) P Hazard ratio (95% CI) P

Overall survival

Age at diagnosis

�18 months 1.0 (reference) �0.0001 1.0 (reference) 0.0100

�18 months 5.757 (2.550–12.998) 3.061 (1.307–7.168)

Stage

1, 2, 3, 4S 1.0 (reference) �0.0001 1.0 (reference) 0.0005

4 8.683 (3.822–19.723) 4.565 (1.939–10.751)

MYCN status

Not amplified 1.0 (reference) �0.0001 1.0 (reference) 0.0013

Amplified 3.867 (2.080–7.190) 2.912 (1.521–5.576)

GALNT9 expression

(�) 1.0 (reference) �0.0001 1.0 (reference) 0.0290

(�) 3.857 (2.064–7.208) 2.066 (1.078–3.963)

Disease-free survival

Age at diagnosis

�18 months 1.0 (reference) �0.0001 1.0 (reference) 0.0035

�18 months 4.577 (2.351–8.911) 2.901 (1.418–5.933)

Stage

1, 2, 3, 4s 1.0 (reference) �0.0001 1.0 (reference) 0.0027

4 5.011 (2.691–9.333) 2.777 (1.424–5.415)

MYCN status

Nonamplified 1.0 (reference) �0.0001 1.0 (reference) 0.0010

Amplified 3.314 (1.886–5.823) 2.692 (1.489–4.869)

GALNT9 expression

(�) 1.0 (reference) 0.001 1.0 (reference) 0.1986

(�) 2.548 (1.190–4.357) 1.446 (0.824–2.538)

230 Clinical Chemistry 59:1 (2013)

the other members of the GALNT family (30 ). Inter-estingly, the GALNAC-T8, -T9, and -T18 enzymes donot have catalytic activity toward classic peptide sub-strates, which are derived from mucin 1 (MUC1) vari-ant MUC1/A, MUC5AC, and mono-GalNAc/Thr7th-EA2 (a glycopeptide containing 1 GalNAc in thethreonine at the 7th position, named EA2) (31 ), sug-gesting that the GALNAC-T members of this subfamilymay glycosylate very specific substrates. Functionalanalysis has shown that GALNAC-T9 is able to glyco-sylate a synthetic peptide (SDC284) derived fromsyndecan-3 (expressed in neuronal cells), but it pos-sesses no activity for the transfer of GalNAc to 3 other

peptides (SDC106, SDC155, and SDC165) derivedfrom syndecan-3 (18 ). These results indicate thatGALNT9 displays a very restricted specificity for thisacceptor substrate, in contrast to that observed forGALNAC-T13. These functional differences betweenGALNAC-T9 and GALNAC-T13, and the fact that thegenes encoding these 2 enzymes display an opposingexpression pattern, lead to the hypothesis that theycould be of biological relevance in neuroblastomabehavior.

Our analysis of 122 neuroblastoma primary tu-mors revealed an association between GALNT9 geneexpression and survival. For the high-risk patients,

Fig. 3. Kaplan–Meier analysis of overall survival and disease-free survival of neuroblastoma patients.

(A), Analysis of the entire cohort (n � 122): GALNT9 expression (�), n � 77; lack of expression (�), n � 45. Overall survival,P � 0.0001; disease-free survival, P � 0.0004. (B), Analysis of the high-risk group (n � 58): GALNT9 expression (�), n � 31;lack of expression (�), n � 27. Overall survival, P � 0.0391; disease-free survival, P � 0.4178. (C), Analysis of the low-riskgroup (n � 64): GALNT9 expression (�), n � 46; lack of expression (�), n � 18. Overall survival, P � 0.0001; disease-freesurvival, P � 0.0007.

GALNT9 Gene Is a Prognostic Marker in Neuroblastoma

Clinical Chemistry 59:1 (2013) 231

GALNT9 expression was correlated with overall sur-vival but not with disease-free survival. In contrast, thesurvival analysis of the low-risk neuroblastoma sub-groups clearly highlighted the prognostic value ofGALNT9 expression for both overall and disease-freesurvival of patients. In other words, patients with tu-mors having GALNT9 expression displayed higher sur-vival rates than patients with no expression. More spe-cifically, within the low-risk neuroblastoma subgroup,a relapse can be predicted for tumors that lack GALNT9expression, whereas a positive outcome can be pre-dicted for tumors with GALNT9 expression. Therefore,assessment of GALNT9 gene expression in these pa-tients could provide an additional method of biologicalclassification of tumors. Expression of the GALNT9gene therefore appears to have predictive value beyondthat for risk stratification based on age, disease stage,and MYCN amplification, and its expression in partic-ular may constitute a marker of relapse for neuroblas-tomas classified as low risk. Very recently, nonrandomchromosomal abnormalities (numerical and segmen-tal changes) have been shown to improve the standardrisk stratification (32, 33 ). In this regard, studies ofGALNT9 gene expression with independent tumor co-horts are necessary to refine the forthcoming risk-stratification algorithm (age, stage, MYCN gene status,and genomic changes).

The low expression of the genes encoding someGALNAC-T isoenzymes has been associated with moreaggressive tumors. GALNAC-T6 is reportedly pro-duced in fibroblasts, placenta, and normal pancreascells. The absence of this enzyme in pancreatic tumorswas correlated with a poor clinical outcome (34 ). Sim-ilarly, a loss of GALNAC-T3 was correlated with ahigher metastatic potential in a mouse colon cancermodel (35 ), and strong GALNAC-T3 protein produc-tion, as demonstrated in human colorectal carcinomaswith immunochemistry staining, significantly en-hanced the likelihood of patient survival (36 ).One possible explanation for neuroblastoma is that

GALNT9 expression is a marker for more-maturestages of neuroblastic tumor cells, which are associatedwith less aggressive tumor cells. In neuroblastoma,S-type cells are more susceptible to apolipoprotein2 ligand/tumor necrosis factor–related apoptosis-inducing ligand (Apo2L/TRAIL)-mediated apoptosisthan N-type cells, owing to their expression ofcaspase-8 (29 ). O-glycosylation pathways have beenfound to play a role in the regulation of cell growth bymodifying the rate of apoptosis of cancer cells (37, 38 ).

In conclusion, our data support the use ofGALNT9 as a prognostic marker that is independent ofthe conventional risk stratification, and it may be avaluable marker that may be helpful in guiding specifictherapy in low-risk patients.

Author Contributions: All authors confirmed they have contributed tothe intellectual content of this paper and have met the following 3 re-quirements: (a) significant contributions to the conception and design,acquisition of data, or analysis and interpretation of data; (b) draftingor revising the article for intellectual content; and (c) final approval ofthe published article.

Authors’ Disclosures or Potential Conflicts of Interest: Upon man-uscript submission, all authors completed the author disclosure form.Disclosures and/or potential conflicts of interest:

Employment or Leadership: None declared.Consultant or Advisory Role: None declared.Stock Ownership: None declared.Honoraria: None declared.Research Funding: Comision Honoraria de Lucha Contra el Cancer,Montevideo, Uruguay, Programa Grupos de Investigacion CSIC,Montevideo, Uruguay, and Societe Francaise des Cancers del’Enfant, La Ligue contre le Cancer Comite Oise.Expert Testimony: None declared.Patents: None declared.

Role of Sponsor: The funding organizations played no role in thedesign of study, choice of enrolled patients, review and interpretationof data, or preparation or approval of manuscript.

Acknowledgments: We thank French surgeons and oncologists forproviding material and clinical reports, and Sabrina Cantais for tech-nical assistance.

References

1. Kim S, Chung D. Pediatric solid malignancies.Neuroblastoma and Wilms’ tumor. Surg ClinNorth Am 2006;86:469–87.

2. Maris JM, Hogarty MD, Bagatell R, Cohn SL.Neuroblastoma. Lancet 2007;369:2106–20.

3. Reis CA, Osorio H, Silva L, Gomes C, David L.Alterations in glycosylation as biomarkers forcancer detection. J Clin Pathol 2010;63:322–9.

4. Brockhausen I. Mucin-type O-glycans in humancolon and breast cancer: glycodynamics and func-tions. EMBO Rep 2006;7:599–604.

5. Hanisch F. O-glycosylation of the mucin type. BiolChem 2001;382:143–9.

6. Ten Hagen KG, Fritz TA, Tabak LA. All inthe family: the UDP-GalNAc:polypeptide

N-acetylgalactosaminyltransferases. Glycobiology2003;13:1R–16R.

7. Bennett EP, Mandel U, Clausen H, Gerken TA,Fritz TA, Tabak LA. Control of mucin-typeO-glycosylation: a classification of the polypep-tide GalNAc-transferase gene family. Glycobiol-ogy 2012;22:736–56.

8. Pratt M, Hang H, Ten Hagen K, Rarick J, GerkenTA, Tabak LA, Bertozzi CR. Deconvolutingthe functions of polypeptide N-�-acetylgalactosaminyltransferase family membersby glycopeptide substrate profiling. Chem Biol2004;11:1009–16.

9. Mandel U, Hassan H, Therkildsen MH, Rygaard J,Jakobsen MH, Juhl BR, et al. Expression of poly-

peptide GalNAc-transferases in stratified epithe-lia and squamous cell carcinomas: immunohisto-logical evaluation using monoclonal antibodies tothree members of the GalNAc-transferase family.Glycobiology 1999;9:43–52.

10. Nomoto M, Izumi H, Ise T, Kato K, Takano H,Nagatani G, et al. Structural basis for the regu-lation of UDP-N-acetyl-�-D-galactosamine: poly-peptide N-acetylgalactosaminyl transferase-3gene expression in adenocarcinoma cells. CancerRes 1999;59:6214–22.

11. Rajpert-De Meyts E, Poll S, Goukasian I, JeanneauC, Herlihy AS, Bennett EP, et al. Changes in theprofile of simple mucin-type O-glycans and poly-peptide GalNAc-transferases in human testis and

232 Clinical Chemistry 59:1 (2013)

testicular neoplasms are associated with germcell maturation and tumour differentiation. Vir-chows Arch 2007;451:805–14.

12. Taniuchi K, Cerny RL, Tanouchi A, Kohno K, Ko-tani N, Honke K, et al. Overexpression of GalNAc-transferase GalNAc-T3 promotes pancreatic can-cer cell growth. Oncogene 2011;30:4843–54.

13. Guda K, Moinova H, He J, Jamison O, Ravi L,Natale L, et al. Inactivating germ-line andsomatic mutations in polypeptide N-acetyl-galactosaminyltransferase 12 in human coloncancers. Proc Natl Acad Sci U S A 2009;106:12921–5.

14. Shibao K, Izumi H, Nakayama Y, Ohta R, NagataN, Nomoto M, et al. Expression of UDP-N-acetyl-�-D-galactosamine–polypeptide galNAcN-acetylgalactosaminyl transferase-3 in relationto differentiation and prognosis in patients withcolorectal carcinoma. Cancer 2001;94:1939–46.

15. Miyahara N, Shoda J, Kawamoto T, Furukawa M,Ueda T, Todoroki T, et al. Expression of UDP-N-acetyl-�-D-galactosamine-polypeptide N-acetylgalactosaminyltransferase isozyme 3 in thesubserosal layer correlates with postsurgical sur-vival of pathological tumor stage 2 carcinoma ofthe gallbladder. Clin Cancer Res 2004;10:2090–9.

16. Freire T, Berois N, Sonora C, Varangot M, BarriosE, Osinaga E. UDP-N-acetyl-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase 6(ppGalNAc-T6) mRNA as a potential new markerfor detection of bone marrow-disseminatedbreast cancer cells. Int J Cancer 2006;119:1383–8.

17. Berois N, Mazal D, Ubillos L, Trajtenberg F, Nicolas A,Sastre-Garau X, et al. UDP-N-acetyl-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase-6 as anew immunohistochemical breast cancer marker. J His-tochem Cytochem 2006;54:317–28.

18. Zhang Y, Iwasaki H, Wang H, Kudo T, Kalka TB,Hennet T, et al. Cloning and characterization of anew human UDP-N-acetyl-�-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase,designated pp-GalNAc-T13, that is specifically ex-pressed in neurons and synthesizes GalNAc�-serine/threonine antigen. J Biol Chem 2003;278:573–84.

19. Berois N, Blanc E, Ripoche H, Mergui X, Trajten-

berg F, Cantais S, et al. ppGalNAc-T13: a newmolecular marker of bone marrow involvement inneuroblastoma. Clin Chem 2006;52:1701–12.

20. Ferrandis E, Da Silva J, Riou G, Benard J. Coacti-vation of the MDR1 and MYCN genes in humanneuroblastoma during the metastatic process inthe nude mouse. Cancer Res 1994;54:2256–61.

21. Toba S, Tenno M, Konishi M, Mikami T,Itoh N, Kurosaka A. Brain-specific expression ofa novel human UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase (GalNAc-T9).Biochim Biophys Acta 2000;1493:264–8.

22. Brodeur GM. Neuroblastoma: biological insightsinto a clinical enigma. Nat Rev Cancer 2003;3:203–6.

23. Ambros IM, Benard J, Boavida M, Bown N, CaronH, Combaret V, et al. Quality assessment of ge-netic markers used for therapy stratification.J Clin Oncol 2003;21:2077–84.

24. Benard J, Raguenez G, Kauffmann A, Valent A,Ripoche H, Joulin V, et al. MYCN-non-amplifiedmetastatic neuroblastoma with good prognosisand spontaneous regression: a molecular portraitof stage 4S. Mol Oncol 2008;2:261–71.

25. Thiele C. Neuroblastoma cell lines. In: Masters J,ed. Human cell culture. Lancaster, UK: KluwerAcademic Publishers; 1989. p 21–53.

26. Hettmer S, Malott C, Woods W, Ladisch S, KaucicK. Biological stratification of human neuroblas-toma by complex “B” pathway ganglioside ex-pression. Cancer Res 2003;63:7270–6.

27. Ara T, DeClerck YA. Mechanisms of invasion andmetastasis in human neuroblastoma. Cancer Me-tastasis Rev 2006;25:645–57.

28. Piacentini M, Piredda L, Starace DT,Annicchiarico-Petruzzelli M, Mattei M, Oliverio S,et al. Differential growth of N- and S-type humanneuroblastoma cells xenografted into SCID mice.Correlation with apoptosis. J Pathol 1996;180:415–22.

29. Hopkins-Donaldson S, Bodmer JL, Bourloud KB,Brognara CB, Tschopp J, Gross N. Loss ofcaspase-8 expression in highly malignant humanneuroblastoma cells correlates with resistance totumor necrosis factor-related apoptosis-inducingligand-induced apoptosis. Cancer Res 2000;60:4315–9.

30. Raman J, Guan Y, Perrine CL, Gerken TA, Tabak

LA. UDP-N-acetyl-�-D-galactosamine:polypeptideN-acetylgalactosaminyltransferases: completionof the family tree. Glycobiology 2012;22:768–77.

31. Li X, Wang J, Li W, Xu Y, Shao D, Xie Y, et al.Characterization of ppGalNAc-T18, a memberof the vertebrate-specific Y subfamily ofUDP-N-acetyl-�-D-galactosamine: polypeptideN-acetylgalactosaminyltransferases. Glycobiology2012;22:602–15.

32. Janoueix-Lerosey I, Schleiermacher G, Michels E,Mosseri V, Ribeiro A, Lequin D, et al. Overallgenomic pattern is a predictor of outcome inneuroblastoma. J Clin Oncol 2009;27:1026–33.

33. Schleiermacher G, Janoueix-Lerosey I, Ribeiro A,Klijanienko J, Couturier J, Pierron G, et al. Accu-mulation of segmental alterations determinesprogression in neuroblastoma. J Clin Oncol 2010;28:3122–30.

34. Li Z, Yamada S, Inenaga S, Imamura T, WuY, Wang KY, et al. Polypeptide N-acetyl-galactosaminyltransferase 6 expression in pan-creatic cancer is an independent prognostic factorindicating better overall survival. Br J Cancer2011;104:1882–9.

35. Kato K, Takeuchi H, Kanoh A, Miyahara N, Nemoto-Sasaki Y, Morimoto-Tomita M, et al. Loss of UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase3 and reduced O-glycosylation in colon carcinomacells selected for hepatic metastasis. Glycoconj J2010;27:267–76.

36. Shibao K, Izumi H, Nakayama Y, Ohta R, NagataN, Nomoto M, et al. Expression of UDP-N-acetyl-�-D-galactosamine-polypeptide galNAc N-acetylgalactosaminyl transferase-3 in relation todifferentiation and prognosis in patients withcolorectal carcinoma. Cancer 2002;94:1939–46.

37. Valenzuela HF, Pace KE, Cabrera PV, White R,Porvari K, Kaija H, et al. O-glycosylation regulatesLNCaP prostate cancer cell susceptibility to apo-ptosis induced by galectin-1. Cancer Res 2007;67:6155–62.

38. Patsos G, Hebbe-Viton V, Robbe-Masselot C,Masselot D, San Martin R, Greenwood R, et al.O-glycan inhibitors generate aryl-glycans, induceapoptosis and lead to growth inhibition in colo-rectal cancer cell lines. Glycobiology 2009;19:382–98.

GALNT9 Gene Is a Prognostic Marker in Neuroblastoma

Clinical Chemistry 59:1 (2013) 233