A

nig(gto©

K

1

gipilDsm[s

0d

Leukemia Research 31 (2007) 33–37

mRNA expression of MAGE-A3 gene in leukemia cells

A. Martınez a,e, I. Olarte a, M.A. Mergold a, M. Gutierrez a, E. Rozen a, J. Collazo a,O. Amancio-Chassin a, R.M. Ordonez a, J.J. Montesinos b, H. Mayani b, D.K. McCurdy c,

P. Ostrosky-Wegman d, E. Garrido-Guerrero e, E.I. Miranda a,∗a Laboratorio de Biologıa Molecular, Servicio de Hematologıa, Hospital General de Mexico, Dr. Balmis 148, Col. Doctores,

06726 Mexico City D.F., Mexicob Unidad de Investigacion Oncologica, Hospital de Oncologıa, Centro Medico Nacional, IMSS, Mexico D.F., Mexico

c Department of Pediatric Rheumatology, Childrens Hospital of Orange County, CA, USAd Departamento de Medicina Genomica y Toxicologıa Ambiental, Instituto de Investigaciones Biomedicas, UNAM, Mexico D.F., Mexico

e Departamento de Genetica y Biologıa Molecular, CINVESTAV-IPN, Mexico D.F., Mexico

Received 20 February 2006; received in revised form 6 April 2006; accepted 13 May 2006Available online 27 June 2006

bstract

Leukemia-associated antigens such as proteins encoded by MAGE genes might provide tools for immunotherapy of leukemia. Positive andegative results of MAGE-A gene expression in hematological malignancies have been reported. This led us to study MAGE-A gene expressionn human leukemias using RT-PCR. Among 115 leukemias from various subtypes, 14/34 (41.17%) AML were positive for one of the threeenes analyzed (MAGE-A1 1/32; MAGE-A3 10/32; MAGE-B2 3/12). Expression was also detected in 23/76 (30.26%) B-cell ALL patientsMAGE-A1 2/53; MAGE-A3 20/53; MAGE-B2 1/32). One of these patients expressed both MAGE-A1 (weak signal) and -A3 (strong signal)

enes. Other patient with CML were positive for MAGE-B2 (1/5, 20%). MAGE-A3 expression data were corroborated by real time RT-PCRhrough determination of MAGE-A3 transcript levels. We concluded that the MAGE-A3 gene is expressed at the mRNA level in a proportionf human leukemias.

2006 Elsevier Ltd. All rights reserved.

b[o

ibwaea

eywords: MAGE-A; Leukemia; AML; ALL; CML

. Introduction

Identification of immunogenic leukemia-associated anti-ens as target structures is mandatory for specificmmunotherapy of leukemia. MAGE gene family encodeseptides recognized by autologous cytotoxic T lymphocytesn a MHC class-I restricted fashion [1]. MAGE-A genes areocated in the chromosome Xq whereas MAGE-B (also calledAM-6) are located on chromosome Xp [2]. MAGE genes are

ilent in healthy tissues with the exception of placenta and

ale germ-line cells that do not bear HLA class-I molecules

3]. MAGE-A1, -A3 and -B2 are widely expressed in cancersuch as melanomas, lung carcinomas, head and neck tumors,

∗ Corresponding author. Tel.: +52 55 27892000x1163.E-mail address: pmiranda@servidor.unam.mx (E.I. Miranda).

iMiMft

145-2126/$ – see front matter © 2006 Elsevier Ltd. All rights reserved.oi:10.1016/j.leukres.2006.05.009

ladder carcinomas [2,4], ovarian tumors [4,5], seminomas6], neuroblastomas [7], hepatocellular carcinomas [8], col-rectal carcinomas [9] and osteosarcomas [10].

In contrast to data on solid cancers, MAGE gene expressionn hematological malignancies has not been clear. Cham-ost et al. [11] reported that in none of 48 leukemiasas the MAGE-1 gene expressed. In contrast, Shichijo et

l. [12,13] demonstrated that MAGE-A1 was frequentlyxpressed (∼50%) in adult T-ALL. However Chambost etl. [14] had extended their initial study to 154 patients find-ng that none expressed MAGE-A1. Since the expression of

AGE genes in human leukemias would be of concern for

mmunotherapy, we have now investigated the expression of

AGE-A1, -A3 and -B2 in a panel of fresh blast samplesrom 115 patients. MAGE-A3 gene is expressed in a substan-ial proportion of leukemias.

3 mia Re

2

2

lotItw(B17ls2(

2

gTsTp(Afncpwc419b

TP

O

MMMMMM11�

�

2

eMr

2

cr81GN(mwH6rp1wcvcar

2

cpPw

4 A. Martınez et al. / Leuke

. Materials and methods

.1. Leukemic cells from patients

Peripheral blood from 115 leukemia patients was col-ected (at diagnosis) into a heparinized tube and centrifugedver Lymphoprep (Nycomed Pharmacia AS, Oslo, Norway)o obtain peripheral blood mononuclear cells (PBMC).nformed consent was obtained from the patients beforehe blood was collected for the experiments. These patientsere diagnosed as follows: 34 acute myelogenous leukemias

AML) including six subtypes of the French-American-ritish (FAB) classification (2 AML-M1; 9 AML-M2;2 AML-M3; 7 AML-M4; 3 AML-M5; 1 AML-M6),6 acute lymphoblastic leukemias (ALL) from B or Tineages (71 B-lineage, 5 T-lineage) including the threeubtypes of the FAB classification (2 ALL-L1; 72 ALL-L2;

ALL-L3) and five chronic myelogenous leukemiasCML).

.2. RT-PCR assays

For the determination of the expression of MAGEenes, total RNA was extracted using TRIzol reagent (Lifeechnologies, Paisley, UK) and 1 �g of RNA was used toynthesize cDNA by M-MLV reverse transcriptase (Lifeechnologies, Paisley, UK) [15]. PCR amplification waserformed with the following previously described primersTable 1) [2,13,14]: CHO-14 and CHO-12 for MAGE-A1;B-1197 and BLE-5 for MAGE-A3; LQR1 and LQF1

or MAGE-B2 [16]. RNA of K562 cells [2,12,14,17] andormal PBLs [2,12,17] were used as positive and negativeontrols. Each cDNA was also tested by PCR using specificrimers for �2-microglobulin (Table 1) to exclude samplesith strongly degraded RNA [15]. The PCR reaction was

arried out for 35 cycles: 1 min at 94 ◦C, 1 min at 60 ◦C andmin at 72 ◦C for MAGE-A1 and -A3 genes; 1 min at 94 ◦C,

min at 58 ◦C and 1 min at 72 ◦C for MAGE-B2; 1 min at4 ◦C, 1 min at 55 ◦C and 1 min at 72 ◦C for �2-microglo-ulin.

able 1rimer sets used for the RT-PCR and for the real time PCR quantification

ligonucleotide set Sequence 5′ → 3′

AGE-A1 forward primer CGGCCGAAGGAACCTGACCCAGAGE-A1 reverse primer GCTGGAACCCTCACTGGGTTGCCAGE-A3 forward primer TGGAGGACCAGAGGCCCCCAGE-A3 reverse primer GGACGATTATCAGGAGGCCTGCAGE-B2 forward primer CTGACTTCCGCTTTGGAGGCAGE-B2 reverse primer GCACCCCCAGAAACAGAAGAGGAACA

8S forward primer GTAACCCGTTGAACCCCATT8S reverse primer CCATCCAATCGGTAGTAGCG-2 microglobulin forwardprimer

CCTCCATGATGCTGCTTACATGTC

-2 microglobulin reverseprimer

ATGTCTCGCTCCGTGGCCTTAGCT

3

t(a3A1spRwhr

s

search 31 (2007) 33–37

.3. Sequencing

Amplified cDNA products (three different samples forach gene) were sequenced and analyzed by an ABI Prismodel 310 Sequencer (Perkin-Elmer, New Jersey, USA) as

eported previously [15].

.4. Quantitative real time RT-PCR

Amplifications of the MAGE-A3 transcript for 22 leukemiaell samples and the negative controls were performed byeal time RT-PCR using 2 �l of cDNA, 20 mM Tris–HCl, pH.3, 50 mM KCl, 2 mM MgCl2, 200 �M dNTP, 1 �M AB-197 and BLE-5 primers for MAGE-A3 (Table 1), 1.25 unitsold Amplitaq DNA Polymerase (Roche, Branchburg,J, USA), 10 nM fluorescein and SYBR green-I 1:50,000

Roche, Indianapolis, IN, USA) in a 25 �l reaction. Theixture was heated to 95 ◦C for 7 min; then amplificationas performed in a i-Cycler iQ Detection System (Bio-Rad,ercules, CA, USA) for 50 cycles (1 min at 94 ◦C, 1 min at0 ◦C and an extension of 4 min at 72 ◦C). All reactions wereun in triplicate, and a dynamic range was built with eachroduct of PCR on copy number serial dilutions of 1 × 1010,× 108, 1 × 106, 1 × 104 and 1 × 102. Standard curvesere then constructed plotting the threshold cycle (the PCR

ycle at which a specific fluorescence becomes detectable)ersus the log of each cDNA dilution step (coefficient oforrelation 0.995) (data not shown). Results were expresseds the number of copies of the target gene normalized to 18SRNA (see Table 1 for primer sequences).

.5. Statistical analysis

To assess the relation between the MAGE-A3 transcriptopy number and the WBC and BM blast percentages in sam-les, Pearson’s correlation coefficients and the corresponding-values were reported. The tests were two-sided and resultsere considered significant at the 0.01 level.

. Results

Healthy donors were negative for the MAGE genesested. Among 115 leukemias from various subtypes, 14/3441.17%) AML were positive for one of the three genesnalyzed (MAGE-A1 1/32; MAGE-A3 10/32; MAGE-B2/12). Expression was also detected in 23/76 (30.26%) B-cellLL patients (MAGE-A1 2/53; MAGE-A3 20/53; MAGE-B2/32). One of these patients expressed both MAGE-A1 (weakignal) and -A3 (strong signal) genes (Fig. 1, lane 4). Otheratient with CML were positive for MAGE-B2 (1/5, 20%).epresentative results of the gels are shown in Figs. 1 and 2,here the K562 erythroleukemic cell line or PBLs from a

ealthy donor was used as a positive or a negative controlespectively. Summarized results are shown in Table 2.

Sequence analyses were performed to validate the RT-PCRpecificity. Nucleotide sequences that were obtained were

A. Martınez et al. / Leukemia Research 31 (2007) 33–37 35

Fig. 1. Expression of MAGE-A genes in acute leukemia. Lane 1, 100 bp DNA ladder; lane 2, K562 cell line; lane 3, PBLs from a healthy donor; lanes 4, 7, 8,10 and 13, ALL-L2 representative samples; lane 5, ALL-L1; lanes 6, 11 and 12, AML-M2; lane 9, AML-M3.

Table 2Expression of MAGE-A1, MAGE-A3 and MAGE-B2 genes in fresh blats from leukemia patients

Positive samples (samples tested) MAGE-A1 (+) MAGE-A3 (+) MAGE-B2 (+) Positive samples

Acute myelogenous leukemia 1/32 10/32 3/12 14/34Acute lymphoblastic leukemia 2/53 20/53 1/32 23/76Chronic myelogenous leukemia ND ND 1/5 1/5

All patients 3/85 30/85 5/49 38/115

ND, no determination.

Fig. 2. Expression of MAGE-B2 in leukemia patients. Lane 1, 100 bp marker; lane 2, K562 cells; lane 3, PBLs from a healthy donor; lanes 4 and 5, AML-M2samples; lane 6, AML-M4; lane 7, AML-M5; lanes 8 and 9, CML.

36 A. Martınez et al. / Leukemia Research 31 (2007) 33–37

Table 3Patients’ clinical data and quantitation of the MAGE-A3 transcript

Patient number Sex Age (years) % BM blastsat diagnosis

WBC (×109/l) Diagnosis Normalizeda MAGE-A3transcript copy no.

1 Female 27 80 25 ALL-L2 50382 Female 26 75 37 ALL-L2 151983 Male 50 90 28 ALL-L2 177494 Female 55 50 15 ALL-L2 86765 Male 46 92 18 AML-M4 278736 Male 22 68 20 ALL-L2 07 Female 28 50 10 AML-M3 39248 Male 68 43 5 ALL-L2 25439 Male 42 50 6 AML-M3 151810 Male 35 72 70 ALL-L2 011 Male ND 53 3.5 ALL-L2 248712 Male 23 62 17 AML-M2 250013 Male 39 60 16 ALL-L2 014 Female 51 73 25 ALL-L2 015 Female 61 ND 170 CML 016 Female 49 55 13 ALL-L2 250817 Male 35 83 17 ALL-L2 1250018 Male 22 47 6 ALL-L2 150919 Female 38 69 7 ALL-L2 663620 Male 39 68 10 ALL-L2 021 Male 28 55 20 ALL-L2 122522 Male 36 53 17 ALL-L2 2521

K

Nopies)

caM

aqlf1actimen

Fa

aba(dbe

4

562

D, no determination.a Normalized MAGE-A3 copy number defined as (MAGE-A3 copies/18S c

ompared with the reported sequences for the MAGE genes,nd as expected, the amplification bands corresponded to theAGE-A1, -A3 or -B2, respectively (not shown).Expression of MAGE-A3 was detected in 30/85 (35%)

cute leukemia patients, so we decided to perform auantification of the transcripts of this gene. Twenty twoeukemia samples (positives or negatives) were availableor this analysis. The MAGE-A3 transcript copy number in6 samples (previously detected positive by the RT-PCRnalyses) ranged between 1225 and 27,873 copies per 106

opies of 18S, with a median value of 3234 (Table 3). Theranscript copy number in the K562 cells was 133,168. One

mportant factor that may influence the level of expression

easured is the proportion of leukemic cells present inach blood sample. Therefore the MAGE-A3 transcript copyumber was correlated with WBC and BM blast percentage

ig. 3. Plot of MAGE-A3 transcript copy numbers vs. BM blast percentaget diagnosis.

iophMcaocsTdirMne

133168

× 106.

t diagnosis. In positive samples, a correlation was foundetween a high transcript copy number and a high percent-ge of BM blasts (Pearson’s correlation = 0.65; P = 0.01)Fig. 3). In contrast, in six leukemia samples (previouslyetected negative by RT-PCR), as well as in eight peripherallood mononuclear cell samples from healthy volunteers noxpression was detected by real time RT-PCR.

. Discussion

This report has demonstrated that the MAGE-A3 genes expressed at the mRNA level in a substantial proportionf human leukemias. Our data are consistent in part withublished results of mRNA MAGE expression in a series ofuman leukemias [12,13]. Shichijo et al. [12,13] reported thatAGE-A1 is preferentially expressed in T-cell leukemia. In

ontrast, we found that MAGE-A3 is expressed in B-cell ALLnd in AML. This difference may be due to the different typesf leukemia used for the different studies. Seventy two per-ent ALL patients were classified as T-cell leukemia in thetudy of Shichijo et al. [12,13]. Instead, we only had 6.6% of-ALL patients. Additionally we included in our series theifferent types of AML whereas Shichijo et al. [12,13] stud-ed only myeloid-monocytic leukemias. Chambost et al. [14]

eported that one preB-ALL of 43 ALL patients expressed the

AGE-A3 gene. Expression of MAGE genes on a substantialumber of leukemic cell lines (from B-, T- or myeloid lin-ages) suggest that these different results might not be largely

mia Re

dbtaaA

l[TaM1bpt

fbtgpitMicibmAs(dwrMeMdi

me

mrdtb

A

f

mM

R

[

[

[

[

[

[

[

[

[

A. Martınez et al. / Leuke

ue to the lineage of cells [12,14]. The fact that MAGE haseen reported as not expressed in AML patients [12,13] andhat we did not find transcripts of this gene in a high percent-ge of these patients suggest that this gene might not provideuseful tool for T-cell based vaccines/immunotherapies inML patients.The prevalence of HTLV-1 infection (a cause of T-cell

eukemia development) in Japan, is the highest in the world18]. MAGE gene mRNA was preferentially expressed in-cell leukemia in the Shichijo et al. series [12]. Zacut etl. [19] also reported that EBV-infected cell lines expressedAGE genes. However the prevalence of EBV or HTLV-is unknown in our population, so the possible correlation

etween MAGE gene expression and viral infection is not sup-orted by our data. Additional studies are essential to clarifyhis issue.

Our results did not match with the negative results in blastsrom 154 leukemia patients, previously reported by Cham-ost et al. [14]. Different methods might be responsible forhis discrepancy. Careful comparison of the used methodolo-ies showed that we used (for MAGE-A3) the same specificrimers published in their report. It is of note that their PCRnvolved 30 cycles, while ours used 35 cycles suggestinghat our PCR conditions could mathematically amplify the

AGE gene 32 times more than theirs could. However, theyncreased the sensitivity of their PCR assays by adding fiveycles of amplification and only weak signals were foundn T-cell lines [14], previously reported as MAGE positivey Shichijo et al. [12]. Therefore, the conflicting resultsight not be due to the number of PCR amplification cycles.lternatively, Chambost et al. [14] used 1/40th of the cDNA

olutions for the PCR reactions, whereas we used 1/10thas Shichijo’s et al. method [12]). This could be a crucialifference that caused the observed discrepancies. Further,e also used the Real Time RT-PCR analyses to confirm our

esults. Sixteen of 22 leukemia samples were positive for theAGE-A3 gene with the same specific primers. Chambost

t al. [14] also reported that K562 was positive for theAGE-A3 gene, in agreement with our results. This may be

ue to the high MAGE-A3 transcript copy number observedn these cells, as detected by the Real Time RT-PCR analyses.

The findings of our investigation are based on data from theRNA analysis and require future studies including protein

xpression by western blotting or immunocytology.We conclude that MAGE-A3 gene is expressed at the

RNA level in a proportion of human leukemias. Theseesults may contribute in the clarification of the controversialata concerning MAGE expression in leukemia. Additionally,his increases the archive on differential expression of genesetween different populations in the world.

cknowledgments

AMT, IOC, MM and ER were supported by CONACyTellowships. We thank I. Lechuga for critical reading of the

[

search 31 (2007) 33–37 37

anuscript. This work was supported by CONACyT 30759-and 48015-M.

eferences

[1] Gaugler B, Van den Eynde B, Van der Bruggen P, Romero P, Gaforio JJ,De Plaen E, et al. Human gene MAGE-3 codes for an antigen recognizedon a melanoma by autologous cytolytic T lymphocytes. J Exp Med1994;179:921–30.

[2] De Plaen E, Arden K, Traversari C, Gaforio JJ, Szicora JP, De Smet C,et al. Structure, chromosomal localization and expression of 12 genesof the MAGE family. Immunogenetics 1994;40:360–9.

[3] Boon T. Human tumor antigens recognized by T lymphocytes. J ExpMed 1996;183:725–9.

[4] Lurquin C, De Smet C, Brasseur F, Muscatelli F, Martelange V, DePlaen E, et al. Two members of the human MAGE B gene familylocated in Xp21.3 are expressed in tumors of various histological ori-gins. Genomics 1997;46:397–408.

[5] Yamada A, Kataoka A, Shichijo S, Imai Y, Nishida T, Itoh K. Expresionof MAGE-1, MAGE-2, MAGE-3/-6, and MAGE-4a/-4b genes in ovariantumors. Int J Cancer 1995;64:388–93.

[6] Dabovic B, Zanaria E, Bardoni B, Lisa A, Bordignon C, Russo V, et al.A family of rapidly evolving genes from the sex reversal critical regionin Xp21. Mamm Genome 1995;6:571–80.

[7] Corrias MV, Scaruffi P, Occhino M, De Bernardi B, Tonini GP, Pistoia V.Expression of MAGE-1, MAGE-3 and MART-1 genes in neuroblastoma.Int J Cancer 1996;69:403–7.

[8] Karillama K, Higashi T, Kobayashi Y, Nouso K, Nakatsukasa H,Yamano T, et al. Expression of MAGE-1 and -3 genes and gene prod-ucts in human hepatocellular carcinoma. Br J Cancer 1999;81:1080–7.

[9] Mori M, Inoue H, Mimori K, Shibuta K, Baba K, Nakashima H, et al.Expression of MAGE genes in human colorectal carcinoma. Ann Surg1996;224:183–8.

10] Sudo T, Kuramoto T, Komiya S, Inoue A, Itoh K. Expression of MAGEgenes in osteosarcoma. J Orthop Res 1997;15:128–32.

11] Chambost H, Brasseur F, Coulie P, De Plaen E, Stoppa AM, Baume D,et al. A tumor-associated antigen expression in human haematologicalmalignancies. Br J Haematol 1993;84:524–6.

12] Shichijo S, Tsunosue R, Masuoka K, Natori H, Tamai M, MiyajimaJ, et al. Expression of the MAGE gene family in human lymphocyticleukaemia. Cancer Immunol Immunother 1995;41:95–103.

13] Shichijo S, Sagawa K, Brasseur F, Boon T, Itoh K. MAGE-1 gene isexpressed in T-cell leukemia. Int J Cancer 1996;65:709–10.

14] Chambost H, van Baren N, Brasseur F, Olive D. MAGE-A genesare not expressed in human leukemias. Leukemia 2001;15:1769–71.

15] Martınez M, Gutierrez M, Zafra G, Garcıa-Carranca A, Gariglio P,Miranda E. Younger age and shorter chronic phase in b2a2-positivechronic myeloid leukemia adults with high white blood cell count atdiagnosis. Haematologica 2002;87:666–8.

16] McCurdy DK, Tai LQ, Nguyen J, Wang Z, Yang HM, Udar N, et al.MAGE Xp-2: a member of the MAGE gene family isolated from anexpression library using systemic lupus erythematosus sera. Mol GenetMetab 1998;63:3–13.

17] Greiner J, Ringhoffer M, Simikopinko O, Szmaragowska A, HuebschS, Maurer U, et al. Simultaneous expression of different immuno-genic antigens in acute myeloid leukemia. Exp Hematol 2000;28:1413–22.

18] Tajima K. The 4th nation-wide study of adult-T cell leukemia/

lymphoma (ATL) in Japan: estimates of risk of ATL and its geograph-ical and clinical features. Int J Cancer 1990;45:237–43.

19] Zacut R, Topalian SL, Kawakami Y, Mancini M, Eliyahu S, RosenbergSA. Differential expression of MAGE-1, -2, and -3 messenger RNA intransformed and normal human cell lines. Cancer Res 1993;53:5–8.