Send Orders for Reprints to reprints@benthamscience.

ae

Current Molecular Medicine 2017, 17, 1-11 1

RESEARCH ARTICLE

Expression and Prognostic Relevance of GAGE1 and XAGE1

Cancer/Testis Antigens in Head and Neck Squamous Cell
Carcinoma
B.T.R. Karia1,†, F.T. Zamunér1,†, V. Carlin1, C.Z. de Oliveira2,3, A.L. Carvalho*,4 and
A.L. Vettore*,1

1
Laboratory for Molecular Cancer Biology, Department of Biological Sciences, Universidade Federal de São
2
Paulo, Diadema, Brazil; Biostatistics Department, PIO XII Foundation, Hospital de Cancer de Barretos,
3
Barretos, Brazil; Education and Research BP, A Beneficência Portuguesa de São Paulo, São Paulo, Brazil;
4
Department of Head and Neck Surgery, PIO XII Foundation, Hospital de Cancer de Barretos, Barretos,
Brazil

Abstract: Background: In the last decades, survival rates in Head and Neck squamous
cell carcinoma (HNSCC) have not changed, with a five-year survival of only 50%. Thus,
there is a great need for the identification of new molecular targets and development of
novel therapeutic strategies. Cancer-testis antigens (CTAs) are expressed in various
types of tumor but rarely in healthy normal tissues. Therefore, they appear as ideal
targets for immunotherapy approaches, as well as, unique markers for cancer
diagnosis/prognosis.
Objective: This study evaluated the expression pattern of cancer/testis antigens (CTA)
ARTICLE HISTORY
in HNSCC samples and correlated the expression data with the clinicopathological
prognostic variables.
Received: September 29, 2017 Methods: An in silico screening was performed using all CTA genes cataloged on the
Revised: March 9, 2018
Accepted: March 10, 2018 CTDatabase and the expression of the eight CTA genes (ARMC3, DDX53, FTHL17,
GAGE1, MAGEA11, SYCE1, TCP11, and XAGE1) was examined in 89 HNSCC and 20
DOI: normal mucosa samples using RT-PCR analysis.
10.2174/1566524018666180322162145
Results: GAGE1 (48.3%), XAGE1 (40.4%) and MAGEA11 (19.1%) were frequently and
specifically expressed in HNSCC samples and 68.5% of the cases expressed at least
one of these antigens. Moreover, GAGE1 and XAGE1 mRNA positivity was significantly
associated with the presence of metastasis in the lymph nodes (p=0.038 and p=0.023,
respectively) and, by multivariate analysis, male gender (p=0.032), advanced clinical
stage (p=0.018) and mRNA positivity for GAGE1 (p=0.010) were independent prognostic
factors for overall survival.
Conclusion: These findings suggest GAGE1 and XAGE1 expressions to be useful as
prognostic markers for HNSCC.
Keywords: Head and neck squamous cell carcinoma, cancer/testis antigens, mRNA expression, prognostic
markers, GAGE1, XAGE1.

1. INTRODUCTION Head and neck cancers are mostly squamous cell

carcinomas and arise from the mucosa of distinct head
Squamous cell carcinoma of the head and neck
(HNSCC) is one of the most frequent cancers in the and neck topologies including oral cavity, pharynx and
larynx. Over the past decades, diagnosis and treatment
world, with nearly 650,000 new cases per year [1].
of HNSCC patients have advanced due to combined
efforts in surgery, radiotherapy and chemotherapy, but
*Address correspondence to these authors at the (A.L. Vettore) long-term survival rates have changed only marginally
Laboratório de Biologia Molecular do Câncer, UNIFESP, Rua Pedro and the 5-year overall survival rate is still around 50-
de Toledo, 669 – 11° Floor, São Paulo, SP 04039-032, Brazil; Tel:
+55 11 5576-4848, Ext. 2002; E-mail: [email protected];
60% [2]. Therefore, there is a great need for the
(A.L. Carvalho) Department of Head and Neck Surgery, PIO XII identification of new molecular targets for diagnosis
Foundation, Barretos Cancer Hospital, Barretos, Brazil, Rua Antenor and prognosis of HNSCC, as well as, for the
Duarte Villela, 1.131, Barretos, SP, Brazil; Tel: +55 17 3321-6600; E- development of novel therapeutic approaches to
mail: [email protected] reduce the need for mutilating surgeries and morbid
†
Both authors contributed equally to this work.
adjuvant therapies. Interestingly, the use of tumor

1566-5240/17 $58.00+.00 © 2017 Bentham Science Publishers

2 Current Molecular Medicine, 2017, Vol. 17, No. 10 Karia et al.

antigens as targets for immunotherapy is emerging as undergoing odontological and preprosthetic surgeries
a new treatment strategy for different tumor types [3, 4] were used as normal controls.
and identifying antigens for specific types of cancer is
The study protocol had been reviewed and received
critical to optimizing the therapeutic efficacy.
approval by the ethics committee of the Barretos
Furthermore, tumor antigens can also act as markers Cancer Hospital and it was performed in accordance
for diagnosis and prognosis of cancer. Of note, few
with the ethical guidelines of the 1975 Declaration of
antigenic peptides have been identified in HNSCC thus
Helsinki. Written consent was obtained from all patients
far.
and healthy donors. Socio-demographic, clinical, and
Among the various classes of tumor-associated anatomo-pathological data were collected from the
antigens identified, Cancer/Testis Antigens (CTAs) are patients’ medical records. Smoking was defined as the
particularly interesting targets because they are use of tobacco, chewable or smoked, for at least 1 year
expressed in tumors of different histological topology continuously. Alcohol consumption was defined as
and present limited expression in normal tissues, with intake of more than 2 alcoholic drinks per day, for at
high levels being confined to immature cells and least 1 year continuously.
spermatogonia, oogonia and trophoblasts [5]. Due to
their capacity to induce an immune response and their 2.2. RNA Extraction and cDNA Synthesis
almost exclusive expression in different types of
CTA expression was determined by reverse
cancer, CTAs appear as ideal targets for
transcription-polymerase chain reaction (RT-PCR).
immunotherapy approaches, as well as unique markers
Total RNA from tissue samples (tumor and normal)
for cancer diagnosis/prognosis [6].
was extracted using the TRIzol reagent (Invitrogen)
Up to now, around 250 CTA genes have been following the manufacturer's recommendations. The
described and they are all cataloged in the RNA concentration was measured spectrophoto-
CTDatabase (http://www.cta.lncc.br) [7], but the metrically (NanoDrop 1000 Spectrophotometer,
expression patterns of most of them remain largely Thermo Fischer Scientific) and the quality of the RNAs
uncharacterized in various cancers. Although some was checked by electrophoresis in Bioanalyser
studies have reported the expression analysis of CTA (Agilent).
in HNSCC, there is little data regarding their
importance as prognostic factors related to HNSCC Two micrograms of each RNA sample were
subjected to cDNA synthesis using the SuperScript III
clinical outcome [8-17]. Besides the purpose of
First-Strand Synthesis System (Invitrogen) according to
providing novel therapeutic targets for HNSCC, the aim
the manufacturer's instructions. The cDNA obtained
of this study was to evaluate the expression pattern of
was diluted 10X before use.
eight CTA genes in a relatively large group of HNSCC
samples and examine the relationship between this
expression pattern and recognized clinicopathological 2.3. RT-PCR Analyses
variables in an attempt to determine whether the The mRNA expression analyses of the selected
presence of these antigens could also be relevant to CTA were performed as described previously [18].
the malignancy of head and neck squamous cell Reaction conditions and primer sequences were the
carcinomas. same as available at CTDatabase
(http://www.cta.lncc.br/index.php) (Table 1). Testicular
2. MATERIALS AND METHODS tissue was used as a positive control and the β-actin
gene was used as an internal quantity control.
2.1. Patients Electrophoresis was performed by loading 12 µL of
each sample on a 2% agarose gel and visualized by
This retrospective study involved tissue specimens Sybr Safe staining using an EP AlphaImager (Alpha
from 89 HNSCC patients who underwent tumor Innotech). Each RT-PCR experiment was performed in
resection between 2007 and 2010 at the Department of duplicate using the same cDNA sample. All cases with
Head and Neck Surgery of the Barretos Cancer a detectable band on both replicates were considered
Hospital, São Paulo, Brazil. These samples were positive. The intensities of the PCR products were
available at the tumor bank of the Barretos Cancer heterogeneous, and some specimens yielded only faint
Hospital. Only patients diagnosed with primary bands. These were scored positive only if the result
HNSCC, not previously treated, that were over 18 could be reproduced by a repeated RT-PCR. Cases
years of age, treated with curative intent and with very low transcript levels that were not
presenting with tumors at oral cavity, pharynx or larynx reproducibly positive were regarded as negative.
were included in the study. The HNSCC diagnoses
were confirmed by a pathologist and all samples were 2.4. Selection of CTA Genes
checked microscopically for the presence of neoplastic
tissue and the absence of contaminating normal We used the same strategy described by Zamuner
mucosa. Tissue samples were snap-frozen in liquid et al. [19] to select CTA-candidates as potential
nitrogen within 30 minutes after resection and stored markers for HNSCC among all CTAs present in the
o
at -80 C until the extraction of RNA. Additionally, 20 CTDatabase. Basically, we first mined publicly
oral mucosa tissue samples from healthy donors available data on CTA expression in HNSCC samples
from the Oncomine (http://www.oncomine.org),
Prognostic Relevance of GAGE1 and XAGE1 Expression in HNSCC Current Molecular Medicine, 2017, Vol. 17, No. 10 3

NExtBio (www.nextbio.com) and SAGE Anatomic patients, respectively. Primary tumor sites included:
Viewer (http://cgap.nci.nih.gov/SAGE/AnatomicViewer). oral cavity (68.6%), pharynx (15.7%) and larynx
After that, we examined the high-throughput expression (15.7%). Regarding stage, the majority of the patients
data (EST, MPSS, CAGE) provided by Hofmann et al. was classified as T3/T4 tumors (72.7%), N0 necks
[20] and the published data of CT mRNA expression in (55.2%), presenting as advanced clinical stage III/IV
cell lines (available in CTDatabase). In this way, we (81.8%).Thirty-nine patients presented positive lymph
were able to select the most promising CTA genes to nodes, and 21 cases presented extracapsular spread.
have their expression evaluated in the HNSCC Perineural invasion was detected in 24.7% of the 85
samples. cases with data available, while lymphovascular
invasion was noted in 14 cases (17.1%) among 82
2.5. Cell Lines Treatment with 5-Aza-dC patients with data available.
FaDu, JHU12, JHU13 and JHU28 human HNSCC
3.2. Selection of CTA Candidates
cell lines were grown in DMEM (Invitrogen)
supplemented with 10% fetal bovine serum, 100 U/mL Using all CTA genes cataloged on the CTDatabase,
penicillin and 100 µg/mL streptomycin. Cells were an in silico screening was performed to interrogate
maintained at 37°C in a humid atmosphere containing SAGE and NCBI-CGAP EST databases, microarray
5% CO2. Cells were counted and plated in 10-cm data (Oncomine and NextBio collections), available
dishes 1 day prior to treatment when the media was data of CTA expression in cancer cell lines
replaced by fresh media containing 5µM of 5-aza-2’- (CTDatabase) and data provided by Hofmann et al.
deoxycytidine (decitabine, DAC; Sigma) and kept for 3 [20] about a high-throughput survey of CTA expression
or 5 consecutive days. JHU12, JHU13 and JHU28 in cancer and normal tissues. This analysis allows the
human HNSCC cell lines were kindly provided by Dr. identification of CTA genes with high probability to be
Joseph Califano (Otolaryngology–Head and Neck expressed in HNSCC.
Surgery, Johns Hopkins Medical Institutions, Baltimore,
Since one of the main goals of this study was the
USA).
identification of new prognosis markers for HNSCC,
CTA genes previously described as frequently
2.6. Statistical Analyses expressed in this malignancy (as MAGE1, MAGE4,
Statistical analysis was performed using the MAGE10, PRAME, among others) were not included in
software SPSS 19.0 for Windows. A Pearson the further analysis. As a result of this mining process,
correlation was used to quantify correlations between we were able to select eight CTA genes (ARMC3,
CTA’s expression and clinic-demographic variables. DDX53, FTHL17, GAGE1, MAGEA11, SYCE1, TCP11
Survival curves were calculated by Kaplan-Meier and XAGE1) for gene expression evaluation in a
method and differences between groups were context of elevated expression in HNSCC samples and
compared using the log-rank test. Overall survival absence in normal mucosa.
(OS) was measured as the time interval between the
date of the of initial treatment for the primary tumor 3.3. CTA Expression in HNSCC Specimens
and the date of the last follow up or death, while the
In the first series, the expression of the eight
Disease-free survival (DFS) was defined as the time
selected CTA genes was evaluated in 20 normal
interval between the date of initial treatment and the
mucosa samples. This analysis showed that ARMC3
date of the diagnosis of the first recurrence. For
(40.0%), SYCE1 (40.0%), TCP11 (40.0%), DDX53
evaluation of the independent contribution of (100.0%) and FTHL17 (100.0%) were frequently
significant clinical and molecular variables on
expressed in normal controls, showing low specificity.
disease-free survival or overall survival, all factors with
In contrast, GAGE1 (0%), MAGEA11 (0%) and XAGE1
significance in Kaplan-Meier analysis (p-value≤0.2)
(5%) were found rarely expressed in normal mucosa
were tested in the multivariate analyses using the
and were considered the most appropriate genes for
Cox proportional hazard model. Results were
the additional analyses (Table 3). So, in a second
calculated with 95% confidence intervals (CI). For all series, the presence of GAGE1, MAGEA11 and
analysis, we considered statistical significance when p-
XAGE1 transcripts was examined in 89 HNSCC
value≤0.05.
samples (Supplementary Table 1). By the end, GAGE1
was expressed in 48.3% (43/89) of all HNSCC cases
3. RESULTS evaluated, XAGE1 in 40.4% (36/89) and MAGEA11 in
19.1% (17/89) (Table 3).
3.1. Patient Characteristics and Clinical Predictors
Notably, 66.2% of the HNSCC samples showed
Table 2 summarizes the demographic and clinical expression of at least one member of the 3-CTA gene
characteristics of HNSCC patients included in the panel. Among them, 6.7% expressed all three CTA
study. Eighty-nine primary tumor samples were genes, while coexpression of two CTAs was found in
collected from untreated HNSCC patients (77 males 25.8% of the examined HNSCC cases. Further, the
and 12 females). Median age was 58.9 years (range expression of at least one CTA member of the panel
32-82 years). Tobacco or alcohol consumption (current was detected in 33.7% of the cases evaluated.
or former) was reported by 90.7% and 73.5% of the
4 Current Molecular Medicine, 2017, Vol. 17, No. 10 Karia et al.

Table 1. Primers used in the RT-PCR assays.

Gene Forward Primer Reverse Primer Amplificon Size (bp)

ARMC3 CCTCCTTGAACTTGGAGCTG TTCTTTACATCCGGGTCAGG 322

DDX53 GGTGCCGATACTCCCACTAT TTGCTTCAGATTCCCCGTTT 300

FTHL17 TGGCCCTGGAGAACTTCTT ATGGTCTTGACTTGCTCGTG 297

GAGE1 GACCAAGACGCTACGTAG CCATCAGGACCATCTTCA 243

MAGEA11 TCTGCCAAGAGTCCAGGTTT CCTCTAGAGTGCCCACATTCA 308

SYCE1 GCTGGCAAAGGAGATTTGTG TCACTGGTCTACCTGGTCTGG 404

TCP11 CACCAGTTCGAGATGAAGCA GGTCCCAGAGAAGGAGGTTC 396

XAGE1 TCCCAGGAGCCCAGTAATGGAGA CAGCTTGTCTTCATTTAAACTTGTGGTTGC 275

Table 2. Demographic and clinical characteristics of the 89 HNSCC patients included in the study.

Characteristic Category n %

Gender Male 77 86.5

Female 12 13.5
Age < 55 years 31 34.8
≥ 55 years 58 65.2
Tobacco Consumption Yes 78 90.7
No 8 9.3
Alcohol Consumption Yes 61 73.5
No 22 26.5
Tumor Site Oral Cavity 61 68.6
Pharynx 14 15.7
Larynx 14 15.7
T Stage T1/T2 24 27.3
T3/T4 64 72.7
N Stage N0 48 55.2
N1/N2/N3 39 44.8
Clinical Stage CS I/II 16 18.2
CS III/IV 72 81.8
Perineural Invasion Yes 21 24.7
No 64 75.3
Lymphvascular Invasion Yes 14 17.1
No 68 82.9

Pathological Lymph Yes 39 50.0

Node Metastasis* No 39 50.0
#
Extracapsular Spread Yes 21 60.0
No 14 40.0
* Includes only patients submitted to neck dissection.
#
Includes only patients with pathological lymph node metastasis.
Prognostic Relevance of GAGE1 and XAGE1 Expression in HNSCC Current Molecular Medicine, 2017, Vol. 17, No. 10 5

Table 3. Comparison of CTA expression on HNSCC and normal control samples.

Normal Control HNSCC

Sensitivity, % Specificity, %
Genes
Cases Expression Cases Expression [95%CI] [95%CI]
(n) n (%) (n) n (%)

GAGE1 20 0 (0) 89 43 (48.3) 48.3 [38.2; 58.5] 100.0 [83.9; 100]

XAGE1 20 1 (5) 89 36 (40.4) 40.4 [30.8; 50.8] 95.0 [76.4; 99.1]

MAGEA11 20 0 (0) 89 17 (19.1) 19.1 [12.3; 28.5] 100.0 [83.9; 100]

ARMC3 20 8 (40) NE NE NE 60.0 [38.7; 78.1]

SYCE1 20 8 (40) NE NE NE 60.0 [38.7; 78.1]

TCP11 20 8 (40) NE NE NE 60.0 [38.7; 78.1]

DDX53 10 10 (100) NE NE NE 0.0 [0.0; 27.75]

FTHL17 10 10 (100) NE NE NE 0.0 [0.0; 27.75]

NE: Not Evaluated

Table 4. Association among GAGE1, MAGEA11 and XAGE1 expression and demographic, clinic-pathological features
of HNSCC patients.

XAGE1 MAGEA11 GAGE1

Characteristic Category
negative positive p-value* negative positive p-value* negative positive p-value*

< 55 years 17 (54.8%) 14 (45.2%) 24 (77.4%) 7 (22.6%) 16 (51.6%) 15 (48.4%)

Age 0.513 0.547 0.992
≥ 55 years 36 (62.1%) 22 (37.9%) 48 (82.8%) 10 (17.2%) 30 (51.7%) 28 (48.3%)

Male 48 (62.3%) 29 (37.7%) 63 (81.8%) 14 (18.2%) 41 (53.2%) 36 (43.8%)

Gender 0.179 0.581 0.461
Female 5 (41.7%) 7 (58.3%) 9 (75.0%) 3 (25.0%) 5 (41.7%) 7 (58.3%)

Tobacco No 5 (62.5%) 3 (37.5%) 7 (87.5%) 1 (12.5%) 4 (50.0%) 4 (50.0%)

0.903 0.593 0.892
Consumption Yes 47 (60.3%) 31 (39.7%) 62 (79.5%) 16 (20.5%) 41 (52.6%) 37 (47.4%)

Alcohol No 13 (59.1%) 9 (40.9%) 18 (81.8%) 4 (18.2%) 8 (36.4%) 14 (63.6%)

0.995 0.988 0.093
Consumption Yes 36 (59.0%) 25 (41.0%) 50 (82.0%) 11 (18.0%) 35 (57.4%) 26 (42.6%)

Oral Cavity 35 (57.4%) 26 (42.6%) 48 (78.7%) 13 (21.3%) 32 (52.5%) 29 (47.5%)

Tumor Site Pharynx 5 (35.7%) 9 (64.3%) 0.034 10 (71.4%) 4 (28.6%) 0.098 6 (42.9%) 8 (57.1%) 0.839

Larynx 13 (92.9%) 1 (7.1%) 14 (100.0%) 0 (0.0%) 8 (57.1%) 6 (42.9%)

T1/T2 15 (62.5%) 9 (37.5%) 19 (79.2%) 5 (20.8%) 13 (54.2%) 11 (45.8%)

T Stage 0.793 0.828 0.830
T3/T4 38 (59.4%) 26 (40.6%) 52 (81.3%) 12 (18.8%) 33 (51.6%) 31 (48.4%)

N0 32 (66.7%) 16 (33.3%) 38 (79.2%) 10 (20.8%) 27 (56.3%) 21 (43.8%)

N Stage 0.149 0.739 0.354
N1/N2/N3 20 (51.3%) 19 (48.7%) 32 (82.1%) 7 (17.9%) 18 (46.2%) 21 (53.8%)

CS I/II 10 (62.5%) 6 (37.5%) 12 (75.0%) 4 (25.0%) 8 (50.0%) 8 (50.0%)

Clinical Stage 0.840 0.530 0.843
CS III/IV 43 (49.7%) 29 (40.3%) 59 (81.8%) 13 (18.1%) 38 (52.8%) 34 (47.2%)

Lymphvascular No 43 (63.2%) 25 (36.8%) 55 (80.9%) 13 (19.1%) 37 (54.4%) 31 (45.6%)

0.361 0.433 0.207
Invasion Yes 7 (50.0%) 7 (50.0%) 10 (71.4%) 4 (28.6%) 5 (35.7%) 9 (64.3%)

Perineural No 38 (59.4%) 26 (40.6%) 52 (81.3%) 12 (18.8%) 35 (54.7%) 29 (45.3%)

0.840 0.620 0.352
Invasion Yes 13 (61.9%) 8 (38.1%) 16 (76.2%) 5 (23.8%) 9 (42.9%) 12 (57.1%)

Pathological No 28 (71.8%) 11 (28.2%) 28 (71.8%) 11 (28.2%) 26 (66.7%) 13 (33.3%)

Lymph Node 0.038 0.095 0.023
Metastasis^ Yes 19 (48.7%) 20 (51.3%) 34 (87.2%) 5 (12.8%) 16 (41.0%) 23 (59.0%)

Extracapsular No 4 (28.6%) 10 (71.4%) 14 (100.0%) 0 (0.0%) 7 (50.0%) 7 (50.0%)

# 0.102 0.087 0.500
Spread Yes 12 (57.1%) 9 (42.9%) 17 (81.0%) 4 (19.0%) 8 (38.1%) 13 (61.9%)
* Pearson correlation was used to calculate the p-values.
^ Includes only patients submitted to neck dissection.
#
Includes only patients with positive pathological lymph node metastasis.
6 Current Molecular Medicine, 2017, Vol. 17, No. 10 Karia et al.

(Table 4). No significant correlation between CTA gene

3.4. The Relationship Between CTA Expression and expression and other clinicopathologic features, such
Clinicopathological Characteristics as gender, age, tumor stage, tobacco and alcohol con-
The expression profile of GAGE1, XAGE1 and sumption, perineural infiltration, vascular embolization
MAGEA11 genes in HNSCC were evaluated for or extracapsular spread could be observed (Table 4).
potential correlation with clinicopathological Additionally, we evaluated the association of clinical
characteristics. This analysis demonstrated that the and molecular variables with the overall survival (OS)
expression of XAGE1 was significantly correlated with and the disease-free survival (DFS). As shown in Table
the tumor site (p = 0.034) and the presence of lymph 5, the OS was better for those patients with early
nodes metastasis (p = 0.038). The expression of clinical stage (91.7% CS I/II vs. 46.0% CS III/IV,
GAGE1 also showed a significant correlation with the p=0.006), early T stage (86.6% T1/T2 vs. 47.6% T3/T4,
presence of metastasis in the lymph nodes (p = 0.023) p=0.002) and with no perineural invasion (65.4%

Table 5. Overall survival and disease-free survival rates according to the clinical and molecular variables.

Variables Categories 5-Year Overall P-Value 5-Year Disease-Free P-Value

Survival (%) Survival (%)

Age < 55 y.o 67.7 55.7

0.084 0.616
≥ 55 y.o 45.9 65.4
Gender Male 56.4 0.075 62.8 0.115
Female 37.0 54.5
Tobacco Yes 58.6 0.569 63.2 0.255
Consumption No 62.5 57.1
Alcohol Yes 57.9 0.912 71.9 0.130
Consumption No 65.0 44.6
Tumor Site Oral Cavity 49.7 0.461 51.2 0.114
Pharynx 64.3 76.9
Larynx 70.5 91.7
T Stage T1/T2 86.6 0.002 55.3 0.542
T3/T4 41.4 65.6
N Stage N0 64.1 0.065 60.9 0.355
N1/N2/N3 43.7 58.0
Clinical Stage CS I/II 91.7 0.006 63.0 0.556
CS III/IV 46.0 61.8
Perineural Yes 31.7 0.038 58.3 0.150
Invasion No 65.4 64.7
Lymphovascular Yes 50.0 0.172 70.7 0.632
Invasion No 59.3 63.7
Pathological Lymph Yes 46.3 0.085 59.3 0.115
Node Metastasis* No 62.0 65.9
Extracapsular Yes 42.9 0.655 56.6 0.346
#
Spread No 53.6 64.9
GAGE1 Yes 44.5 0.012 72.6 0.905
No 62.7 58.9
XAGE1 Yes 49.2 0.174 66.7 0.886
No 58.0 57.9
MAGEA11 Yes 64.7 0.595 58.9 0.960
No 57.9 61.9
*Includes only patients submitted to neck dissection.
#
Includes only patients with pathological lymph node metastasis.
Prognostic Relevance of GAGE1 and XAGE1 Expression in HNSCC Current Molecular Medicine, 2017, Vol. 17, No. 10 7

absent vs. 31.7% present, p=0.038). The OS was also Table 6. Results of multivariate analysis (Cox
poor for cases expressing GAGE1 (p=0.012, Fig. 1). regression model) of selected prognostic
No significant association between clinical or factors for overall survival (Gender, Clinical
molecular features and DFS was observed (Table 5). Stage and GAGE1 expression).

Categories HR 95% CI P

Gender
Male 1 Ref.
Female 2.54 1.1 ─ 6.0 0.032
Clinical Stage
CS I/II 1 Ref.
CS III/IV 11.07 1.5 ─ 81.3 0.018
GAGE1
No 1 Ref.
Yes 2.56 1.3 ─ 5.2 0.010

Fig. (1). Overall rate according to the GAGE1 expression

pattern in HNSCC samples (p=0.012).

Besides, targeting the identification of independent

predictors of survival, we conducted a multivariate
survival analysis involving the significant clinical
features and the molecular variable associated with
survival probability (GAGE1 mRNA-positivity). This
analysis of prognostic variables revealed that male Fig. (2). RT-PCR for GAGE1 and XAGE1 expression on
gender (HR=2. 54, 95%CI=1. 1-6.0; p=0.032), HNSCC cell lines treated with decitabine (DAC). FaDu,
advanced clinical stage (HR= 11.07, 95%CI=1. 5-81.3; JHU12, JHU13 and JHU28 cell lines were treated with 5µM
p=0.018), and GAGE1 expression (HR= 2.56, DAC for 3 and 5 days. (0d) untreated; (3d) three-day DAC
95%CI=1. 3-5.2; p=0.010) remained as independent treatment; (5d) five-day DAC treatment. β−actin (as
predictors of poor overall survival (Table 6). endogenous control) was equally expressed on all cell lines
before and after DAC treatment (data not shown).
3.5. Upregulation of GAGE1 and XAGE1 Expression
on HNSCC Cells by DAC
4. DISCUSSION
In normal tissues, it is well accepted that CTA
expression is prevented by epigenetic silencing of their The prognosis for patients with HNSCC is mainly
regulatory elements and the aberrant expression of established by the stage of the disease at presentation
CTAs in tumors is, normally, associated with obtained by clinical examination, cytology of lymph
demethylation of their promoters. To verify if the nodes, imaging exams, and definite histopathology
treatment with the demethylating agent DAC could alter after surgery. Regrettably, the survival rate for HNSCC
the GAGE1 and XAGE1 expression patterns in patients did not present a relevant improvement in the
HNSCC, four cell lines (FaDu, JHU12, JHU13 and last decades principally because patients frequently
JHU28) were treated with DAC and the presence of develop locoregional recurrences, distant metastases
GAGE1 and XAGE1 transcripts was evaluated by RT- and second primary tumors [21]. So, the identification
PCR. of new markers for diagnosis or prognosis and the
development of new therapies and their integration into
These four cell lines present a slight expression of the current forms of treatment of this malignancy is
GAGE1 and XAGE1 antigens, but after the exposure to overriding. Immunotherapy may provide an alternative
the demethylating agent they showed either increase or to the established model, since the natural immune
new onset expression of these CT antigens (Fig. 2). response can be stimulated by an active
8 Current Molecular Medicine, 2017, Vol. 17, No. 10 Karia et al.

immunotherapeutic intervention, which can assist in et al. [34], studying melanoma cell lines, detected an
controlling the progression and tumor recurrence. In inhibition of cell migration and invasion after XAGE1
light of these considerations, our results showed that knockdown, while Zhou et al. [35] detected an increase
almost 70% of the HNSCC cases presented expression in the proliferation and metastasis of adenoid cystic
of at least one of the three selected CTA genes carcinoma cells when XAGE is overexpressed.
(GAGE1, MAGEA11 and XAGE1), suggesting that
In several malignant diseases, as multiple myeloma
these patients could take benefit from an immuno-
gastric, lung and ovarian cancers, the expression of
therapeutic approach using a multivalent vaccine
CTAs has been correlated with advanced disease and
based in these antigens. Although our results showed a
poor outcomes [36-41]. Probably, due to the small
rare mRNA positivity for these CTAs in normal mucosa,
sample sizes analyzed, all except two of the previous
according to the CTDatabase, the expression of studies failed in establishing a clear correlation
GAGE1 is restricted to testis, while MAGEA11 and
between CTA expression and HNSCC prognosis.
XAGE1 transcripts could be detected in few normal
Cuffel et al. [9] showed that patients with tumors
tissues.
expressing MAGEA4 or multiple CTA genes had a
Some studies have reported a relatively frequent, poorer overall survival and MAGEA4 mRNA positivity
but sometimes conflicting, expression of CTA genes in was a prognostic marker of poor outcome independent
HNSCC [8-16]. According to them, MAGEA3 is of clinical parameters, while Zamuner et al. [19]
expressed in a high proportion of HNSCC (43-72%), as showed MAGEA3/6 positivity associated with
well as, MAGEA4 (27-60%), PRAME (42-49%), SSX1 significantly better disease-free survival for HNSCC
(45%), MAGEC1 (28-48%), MAGEA1 (30-40%), NY- patients. Laban et al. [42] conducted immuno-
ESO-1 (6-33%), SSX4 (30%), and MAGEC2 (10-33%). histochemistry assays in 552 HNSCC primary tumors
BAGE, MAGEA2, SSX2, SCP1 and MAGEA10 are also using different monoclonal antibodies to multiple MAGE
reported as occasionally expressed in HNSCC. family members, observed that patients expressing
pan-MAGE or MAGE-A3/A4 presented a significantly
lower overall survival. Our results reveal, for the first
Cuffel et al. [9] performed a semi-quantitative RT- time, a significant association between GAGE1 and
PCR analysis to examine tumor expression of 12 CTA XAGE1 expression and the presence of metastasis in
genes and reported the presence of GAGE1 in 38% of the lymph nodes. Besides previously shown for
HNSCC specimens examined, while Figueiredo et al. melanoma cells [34], our results also demonstrated that
[11] found GAGE1 expression in 15% of the HNSCC knockdown of XAGE1 is able to reduce the
samples evaluated. Atanackovic et al. [8] also reported invasiveness of HNSSC cell (data not shown).
a high proportion of GAGE1 positivity in tumor samples Furthermore, the GAGE1 expression also has been
(56%); however, they also detected GAGE1 transcripts demonstrated to correlate with poor prognosis in
in 34.6% of the healthy adjacent tissue samples. We esophageal tumors, neuroblastoma and gastric cancer
identified a specific and frequent mRNA positivity for [43, 44]. Likewise, our results also detected a
GAGE1 gene in HNSCC (48.3%). significant association between GAGE1 mRNA
The GAGE1 gene presents multiple copies in the positivity and poor prognosis for HNSCC patients and a
genome, which are structured in a cluster at Xp11.23. multivariate analysis indicated GAGE1 expression as
The sequences of the family members are highly an independent prognostic factor for overall survival of
conserved but they differ by sparse nucleotide HNSCC patients.
substitutions [22]. The GAGE family is composed by It is well recognized that patients without metastasis
small acidic proteins, which have been detected in a have a better prognosis than patients with lymph node
variety of tumors, including non-small cell lung cancers, metastasis. Only 25-40% of patients with lymph node
bladder cancers, melanoma, and head and neck metastasis at presentation will reach a 5-year survival
tumors [23]. When expressed in tumor cells, GAGE1 period [45]. So, the occurrence lymph node metastases
protein can elicit both humoral and cellular immune is important prognostic indicators for HNSCC [46].
responses, indicating that they are appropriate targets Thereby, the discovery of a correlation between the
for cancer immunotherapy [23]. metastatic behavior of cancer cells and the expression
To the best of our knowledge, this is the first study pattern of specific molecules might help in the
to detect a high expression of MAGEA11 and XAGE1 identification of HNSCC patients that could take benefit
in HNSCC. MAGEA11, previously reported as from an aggressive treatment approach even at an
frequently expressed in prostate cancers [24], interacts early stage of the disease. In this context, the
with the androgen receptor and increases its activity by evaluation of GAGE1 and XAGE1 expression pattern
modulating the interdomain interaction [25]. Therefore, may further provide new information about patient
it is suggested that the increased expression of prognosis and be useful for planning appropriate
MAGEA11 facilitates prostate cancer progression by treatment strategies.
enhancing androgen receptor-dependent tumor growth Cancer/testis antigens are suitable targets for
[26]. XAGE1 expression is also upregulated in immunotherapy of human malignancies, and different
melanoma, Ewing's sarcomas, as well as, in breast and clinical trials are ongoing. However, the heterogeneous
lung cancers, but the contribution of this CTA to intra-tumor expression of CTA may hamper the
tumorigenesis is poorly understood [27-33]. Caballero effectiveness of CTA-directed vaccination through the
Prognostic Relevance of GAGE1 and XAGE1 Expression in HNSCC Current Molecular Medicine, 2017, Vol. 17, No. 10 9

emergence of CTA-negative neoplastic clones. So it marker for HNSCC patients. However, further studies
would be important to induce antigen expression in evaluating larger HNSCC cohorts are required for
CTA-negative tumors prior to immunization [18]. validation of these results.
Epigenetic mechanisms are at the base of CTA
restricted expression pattern in healthy somatic tissues. ETHICS APPROVAL AND CONSENT TO
The activation of the CTA expression in tumors cells PARTICIPATE
should be the result of aberrant DNA demethylation
[47]. The treatment with demethylating agents, such as The study protocol had been reviewed and received
DAC, promotes a decrease in DNA methylation level approval by the Ethics Committee of the Barretos
resulting in the re-expression of CTA genes in cancer Cancer Hospital.
cells. Although DAC can induce the CTA expression on
malignant cells, this demethylating agent has no effect HUMAN AND ANIMAL RIGHTS
in cell lines derived from normal tissues, such as
All the procedures involving human subjects were
endothelial cells, melanocytes, fibroblasts,
performed in accordance with the ethical guidelines of
keratinocytes and myoblasts [47-49]. In light of these
the 1975 Declaration of Helsinki. No animals were used
considerations, combining CTA-based immunotherapy
for studies that are base of this research.
approaches and DAC therapy could constitute an
interesting alternative treatment option for cancer
patients. CONSENT FOR PUBLICATION
Although DNA methylation has become a Written consent was obtained from all patients and
therapeutic target for human cancers and the use of healthy donors.
demethylating agents in the treatment of different
malignancies has received FDA approval, this CONFLICT OF INTEREST
therapeutic approach should be taken with precaution
The authors declare no conflict of interest, financial
for the HNSCC patients. Our findings suggest that
or otherwise.
some genes associated with patient poor prognosis,
such as GAGE1 and XAGE1, could have their
expression induced by the treatment with ACKNOWLEDGEMENTS
demethylating drugs. In light of these considerations, This work was supported by a grant from Fundação
studies evaluating the effect of demethylating agents in de Amparo à Pesquisa do Estado de São Paulo -
the tumorigenesis of HNSCC should be conducted. FAPESP. B.T.R.K. and V.C. were recipients of a
Several points in this report warrant emphasis. First, scholarship from Coordenação de Aperfeiçoamento de
in this study, we show that MAGE11, GAGE1 and Pessoal de Nível Superior - CAPES and F.T.Z was the
XAGE1 CTAs are frequently expressed in HNSCC. recipient of scholarship from Fundação de Amparo à
Second, GAGE1 or XAGE1 mRNA positivity is Pesquisa do Estado de São Paulo – FAPESP. A.L.V.
significantly associated with the presence of metastasis has a scholarship from National Counsel of
in the lymph nodes. Third, almost 70% of the HNSCC Technological and Scientific Development - CNPq.
patients expressed MAGE11, GAGE1 or XAGE1. As
stated by Shiraishi et al. [50], cancer immunotherapy is SUPPLEMENTARY MATERIAL
emerging as a promising modality for cancer treatment
Supplementary material is available on the
and many CTAs are immunogenic and their use as
publisher's website along with the published article.
therapeutic cancer vaccines is being evaluated in
ongoing clinical trials.
REFERENCES
On the other hand, there are some limitations to the
[1] Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer
present study as well. The patient number was limited
statistics, 2002. CA Cancer J Clin 2005; 55(2): 74-108.
and these patients were not consecutively and [2] Carvalho AL, Nishimoto IN, Califano JA, Kowalski LP. Trends
prospectively collected for this study. Moreover, we in incidence and prognosis for head and neck cancer in the
evaluated a frozen tissue cohort in which selection bias United States: a site-specific analysis of the SEER database.
existed as a result of the picking of specimens with Int J Cancer 2005; 114(5): 806-16.
large volume tumors appropriate for frozen tissue [3] Krishnadas DK, Shusterman S, Bai F, et al. A phase I trial
combining decitabine/dendritic cell vaccine targeting MAGE-
collection. A1, MAGE-A3 and NY-ESO-1 for children with relapsed or
therapy-refractory neuroblastoma and sarcoma. Cancer
CONCLUSION Immunol Immunother 2015; 64(10): 1251-60.
[4] Schwab CL, English DP, Roque DM, Pasternak M, Santin
In conclusion, the current study indicates that AD. Past, present and future targets for immunotherapy in
MAGEA11, GAGE1 and XAGE1 are upregulated in ovarian cancer. Immunotherapy 2014; 6(12): 1279-93.
HNSCC. The expression of GAGE1 and XAGE1 is [5] Scanlan MJ, Simpson AJ, Old LJ. The cancer/testis genes:
review, standardization, and commentary. Cancer Immun
significantly associated with the presence of metastasis 2004; 4: 1.
in the lymph nodes and GAGE1 mRNA positivity is an [6] Simpson AJ, Caballero OL, Jungbluth A, Chen YT, Old LJ.
independent marker for overall survival, suggesting this Cancer/testis antigens, gametogenesis and cancer. Nat Rev
CTA may serve as specific and novel prognostic Cancer 2005; 5(8): 615-25.
10 Current Molecular Medicine, 2017, Vol. 17, No. 10 Karia et al.

[7] Almeida LG, Sakabe NJ, deOliveira AR, et al. CTdatabase: a [28] Zendman AJ, Van Kraats AA, Weidle UH, Ruiter DJ, Van
knowledge-base of high-throughput and curated data on Muijen GN. The XAGE family of cancer/testis-associated
cancer-testis antigens. Nucleic Acids Res 2009; 37(Database genes: alignment and expression profile in normal tissues,
issue): D816-9. melanoma lesions and Ewing's sarcoma. Int J Cancer 2002;
[8] Atanackovic D, Blum I, Cao Y, et al. Expression of cancer- 99(3): 361-9.
testis antigens as possible targets for antigen-specific [29] Egland KA, Kumar V, Duray P, Pastan I. Characterization of
immunotherapy in head and neck squamous cell carcinoma. overlapping XAGE-1 transcripts encoding a cancer testis
Cancer Biol Ther 2006; 5(9): 1218-25. antigen expressed in lung, breast, and other types of
[9] Cuffel C, Rivals JP, Zaugg Y, et al. Pattern and clinical cancers. Mol Cancer Ther 2002; 1(7): 441-50.
significance of cancer-testis gene expression in head and [30] Nakagawa K, Noguchi Y, Uenaka A, et al. XAGE-1
neck squamous cell carcinoma. Int J Cancer 2011; 128(11): expression in non-small cell lung cancer and antibody
2625-34. response in patients. Clin Cancer Res 2005; 11(15): 5496-
[10] Eura M, Ogi K, Chikamatsu K, et al. Expression of the MAGE 503.
gene family in human head-and-neck squamous-cell [31] Liu XF, Helman LJ, Yeung C, et al. XAGE-1, a new gene that
carcinomas. Int J Cancer 1995; 64(5): 304-8. is frequently expressed in Ewing's sarcoma. Cancer Res
[11] Figueiredo DL, Mamede RC, Proto-Siqueira R, et al. 2000; 60(17): 4752-5.
Expression of cancer testis antigens in head and neck [32] Koizumi F, Noguchi Y, Saika T, et al. XAGE-1 mRNA
squamous cell carcinomas. Head Neck 2006; 28(7): 614-9. expression in prostate cancer and antibody response in
[12] Kienstra MA, Neel HB, Strome SE, Roche P. Identification of patients. Microbiol Immunol 2005; 49(5): 471-6.
NY-ESO-1, MAGE-1, and MAGE-3 in head and neck [33] Kikuchi E, Yamazaki K, Nakayama E, et al. Prolonged
squamous cell carcinoma. Head Neck 2003; 25(6): 457-63. survival of patients with lung adenocarcinoma expressing
[13] Pastorcic-Grgic M, Sarcevic B, Dosen D, et al. Prognostic XAGE-1b and HLA class I antigens. Cancer Immun 2008; 8:
value of MAGE-A and NY-ESO-1 expression in pharyngeal 13.
cancer. Head Neck 2010; 32(9): 1178-84. [34] Caballero OL, Cohen T, Gurung S, et al. Effects of CT-Xp
[14] Ries J, Mollaoglu N, Vairaktaris E, Neukam FW, Nkenke E. gene knock down in melanoma cell lines. Oncotarget 2013;
Diagnostic and therapeutic relevance of NY-ESO-1 4(4): 531-41.
expression in oral squamous cell carcinoma. Anticancer Res [35] Zhou B, Li T, Liu Y, Zhu N. Promoting effects on the
2009; 29(12): 5125-30. proliferation and metastasis of ACC tumor cell with XAGE-1b
[15] Mollaoglu N, Vairaktaris E, Nkenke E, Neukam FW, Ries J. overexpression. Oncol Rep 2013; 30(5): 2323-35.
Expression of MAGE-A12 in oral squamous cell carcinoma. [36] Andrade VC, Vettore AL, Felix RS, et al. Prognostic impact of
Dis Markers 2008; 24(1): 27-32. cancer/testis antigen expression in advanced stage multiple
[16] Montoro JR, Mamede RC, Neder Serafini L, et al. Expression myeloma patients. Cancer Immun 2008; 8: 2.
of cancer-testis antigens MAGE-A4 and MAGE-C1 in oral [37] Grigoriadis A, Caballero OL, Hoek KS, et al. CT-X antigen
squamous cell carcinoma. Head Neck 2012; 34(8): 1123-8. expression in human breast cancer. Proc Natl Acad Sci U S
[17] Laban S, Atanackovic D, Luetkens T, et al. Simultaneous A 2009; 106(32): 13493-8.
cytoplasmic and nuclear protein expression of melanoma [38] Gure AO, Chua R, Williamson B, et al. Cancer-testis genes
antigen-A family and NY-ESO-1 cancer-testis antigens are coordinately expressed and are markers of poor outcome
represents an independent marker for poor survival in head in non-small cell lung cancer. Clin Cancer Res 2005; 11(22):
and neck cancer. Int J Cancer 2014; 135(5): 1142-52. 8055-62.
[18] Freitas M, Malheiros S, Stavale JN, et al. Expression of [39] Napoletano C, Bellati F, Tarquini E, et al. MAGE-A and NY-
cancer/testis antigens is correlated with improved survival in ESO-1 expression in cervical cancer: prognostic factors and
glioblastoma. Oncotarget 2013; 4(4): 636-46. effects of chemotherapy. Am J Obstet Gynecol 2008; 198(1):
[19] Zamuner FT, Karia BT, de Oliveira CZ, et al. A 99.e1-7.
Comprehensive Expression Analysis of Cancer Testis [40] Svobodova S, Browning J, MacGregor D, et al. Cancer-testis
Antigens in Head and Neck Squamous Cell Carcinoma antigen expression in primary cutaneous melanoma has
Revels MAGEA3/6 as a Marker for Recurrence. Mol Cancer independent prognostic value comparable to that of Breslow
Ther 2015; 14(3): 828-34. thickness, ulceration and mitotic rate. Eur J Cancer 2011;
[20] Hofmann O, Caballero OL, Stevenson BJ, et al. Genome- 47(3): 460-9.
wide analysis of cancer/testis gene expression. Proc Natl [41] Velazquez EF, Jungbluth AA, Yancovitz M, et al. Expression
Acad Sci U S A 2008; 105(51): 20422-7. of the cancer/testis antigen NY-ESO-1 in primary and
[21] Leemans CR, Braakhuis BJ, Brakenhoff RH. The molecular metastatic malignant melanoma (MM)--correlation with
biology of head and neck cancer. Nat Rev Cancer 2011; prognostic factors. Cancer Immun 2007; 7: 11.
11(1): 9-22. [42] Laban S, Giebel G, Klumper N, et al. MAGE expression in
[22] Killen MW, Taylor TL, Stults DM, et al. Configuration and head and neck squamous cell carcinoma primary tumors,
rearrangement of the human GAGE gene clusters. Am J lymph node metastases and respective recurrences-
Transl Res 2011; 3(3): 234-42. implications for immunotherapy. Oncotarget 2017; 8(9):
[23] Gjerstorff MF, Ditzel HJ. An overview of the GAGE 14719-35.
cancer/testis antigen family with the inclusion of newly [43] Zambon A, Mandruzzato S, Parenti A, et al. MAGE, BAGE,
identified members. Tissue Antigens 2008; 71(3): 187-92. and GAGE gene expression in patients with esophageal
[24] Sang M, Wang L, Ding C, et al. Melanoma-associated squamous cell carcinoma and adenocarcinoma of the gastric
antigen genes - an update. Cancer Lett 2011; 302(2): 85-90. cardia. Cancer 2001; 91(10): 1882-8.
[25] Bai S, He B, Wilson EM. Melanoma antigen gene protein [44] Cheung IY, Chi SN, Cheung NK. Prognostic significance of
MAGE-11 regulates androgen receptor function by GAGE detection in bone marrows on survival of patients with
modulating the interdomain interaction. Mol Cell Biol 2005; metastatic neuroblastoma. Med Pediatr Oncol 2000; 35(6):
25(4): 1238-57. 632-4.
[26] Bai S, Wilson EM. Epidermal-growth-factor-dependent [45] Greenberg JS, El Naggar AK, Mo V, Roberts D, Myers JN.
phosphorylation and ubiquitinylation of MAGE-11 regulates Disparity in pathologic and clinical lymph node staging in oral
its interaction with the androgen receptor. Mol Cell Biol 2008; tongue carcinoma. Implication for therapeutic decision
28(6): 1947-63. making. Cancer 2003; 98(3): 508-15.
[27] Brinkmann U, Vasmatzis G, Lee B, Pastan I. Novel genes in [46] Jones AS, Cook JA, Phillips DE, Roland NR. Squamous
the PAGE and GAGE family of tumor antigens found by carcinoma presenting as an enlarged cervical lymph node.
homology walking in the dbEST database. Cancer Res 1999; Cancer 1993; 72(5): 1756-61.
59(7): 1445-8. [47] De Smet C, Loriot A, Boon T. Promoter-dependent
mechanism leading to selective hypomethylation within the 5'
Prognostic Relevance of GAGE1 and XAGE1 Expression in HNSCC Current Molecular Medicine, 2017, Vol. 17, No. 10 11

region of gene MAGE-A1 in tumor cells. Mol Cell Biol 2004; agent 5-aza-2'-deoxycytidine. Cancer Res 1994; 54(7): 1766-
24(11): 4781-90. 71.
[48] Schrump DS, Fischette MR, Nguyen DM, et al. Phase I study [50] Shiraishi T, Terada N, Zeng Y, et al. Cancer/Testis Antigens
of decitabine-mediated gene expression in patients with as potential predictors of biochemical recurrence of prostate
cancers involving the lungs, esophagus, or pleura. Clin cancer following radical prostatectomy. J Transl Med 2011; 9:
Cancer Res 2006; 12(19): 5777-85. 153.
[49] Weber J, Salgaller M, Samid D, et al. Expression of the
MAGE-1 tumor antigen is up-regulated by the demethylating

DISCLAIMER: The above article has been published in Epub (ahead of print) on the basis of the materials provided by the
author. The Editorial Department reserves the right to make minor modifications for further improvement of the manuscript.

PMID: 29577858