mRNA expression of leukemia-associated antigens in patients with acute myeloid leukemia for the development of specific immunotherapies

Authors


Abstract

Specific immunotherapies for patients with acute myeloid leukemia (AML) using leukemia-associated antigens (LAA) as target structures might be a therapeutic option to enhance the graft-vs.-leukemia effect observed after allogeneic stem cell transplantation or to prolong a complete remission (CR) achieved by chemotherapy. Significant mRNA expression of LAA is a prerequisite for such immunotherapies. Here, previously characterized antigens associated with solid tumors (TAA) and newly characterized LAA were investigated for their expression in up to 60 AML patients and in leukemia cell lines. To investigate their specificity for leukemic blasts, the mRNA expression was also characterized in PBMN and CD34 positive cells of healthy volunteers and in a panel of normal tissues. The following antigens showed high mRNA expression in AML patients: MPP11 was detected in 43/50 (86%), RHAMM in 35/50 (70%), WT1 in 40/60 (67%), PRAME in 32/50 (64%), G250 in 18/35 (51%), hTERT in 7/25 (28%) and BAGE in 8/30 (27%) of AML patients. Real-time RT-PCR showed a tumor-specific expression of the antigens BAGE, G250 and hTERT, as well as highly tumor-restricted expression for RHAMM, PRAME and WT1. The antigen MPP11 was overexpressed. These antigens might be candidates for immunotherapies of leukemia patients and, because of their simultaneous expression, also for polyvalent vaccines. © 2003 Wiley-Liss, Inc.

While 65–80% of adult patients with acute myeloid leukemia (AML) will enter complete remission (CR) with anthracycline based induction chemotherapy, only 20–30% of patients will maintain durable remission with standard consolidation therapies.1 Novel treatment strategies for AML patients might include more specific immunostimulatory elements. The immune system plays an important role in the control of proliferating leukemic cells: interleukin 2 (IL-2), a T cell stimulating cytokine, was shown to be effective in inducing complete remission in a patient with AML.2 For younger patients with AML or CML (chronic myeloid leukemia), allogeneic hematopoietic stem cell transplantation (HSCT) constitutes an optional treatment.1, 3 The therapeutic concept of HSCT employs not only the effects of total body irradiation and high dose chemotherapy, but also the induction of the so-called graft-vs.-leukemia (GVL) effect.4 Patients with CML and an increasing number of patients with AML who relapse from leukemia after HSCT were administered donor lymphocyte infusions (DLI), often resulting again in a complete remission.1, 5, 6, 7 This fact demonstrates the efficacy of the GVL reaction and suggests that also leukemia-associated antigens (LAA) are recognized by the donor lymphocytes in the framework of the GVL reaction. For the amplification and expansion of the GVL effect in the context of HSCT, specific immunotherapies using LAA as targets might be a successful approach. Moreover, these immunotherapies might prolong the CR in patients achieved by polychemotherapy.

To define antigens for such specific immunotherapies, we examined in our study the mRNA expression of tumor/leukemia-associated antigens (TAA/LAA) with expected or known expression in AML and with already described immunogenicity. Candidates are at first cancer/testis (CT) antigens, now called cancer/germline (CG) antigens, like BAGE,8 GAGE,9 MAGE,10 NY-ESO-111 and SSX2,12, 13 being members of families of TAA expressed in tumor cells and in normal testis tissue but not in other normal tissues. The CG antigen PRAME is expressed in different cancer tissues and also at a very low level in some other tissues like adrenals, endometrium and ovary.14 TAA were hitherto characterized in solid tumors; some of these antigens like PRAME and WT1 are known to also be expressed in leukemias.14, 15, 16, 17, 18, 19, 20

A further candidate antigen is proteinase 3. This gene is expressed in AML patients, and a cytotoxic T lymphocyte (CTL) epitope designated PR1 has been characterized.21, 22, 23 By the method of serologic screening of recombinant expression libraries (SEREX), several LAA in AML have been identified, such as RHAMM,24 MPP11 and the heat-shock proteins HSJ2 and HSP70,25 showing humoral immune responses only in tumor patients but not in healthy volunteers (HV). Humoral immune responses to MAZ were found in AML patients but also in HV.20 The SEREX defined CML-antigens RBPJκ and RFAFTK showed humoral immune responses in CML patients.26

Significant mRNA expression of these antigens would be a prerequisite for immunotherapies in AML patients. Simultaneous expression of different antigens might a basis for the efficacy of polyvalent vaccines. To investigate their specificity, the expression of TAA/LAA was evaluated in samples from AML patients in contrast to peripheral blood mononuclear cells (PBMN) from HV, CD34 positive cells and a panel of normal tissues.

Abbreviations

AML, acute myeloid leukemia; CG, cancer/germline; CR, complete remission; CT, cancer/testis; FAB, French-American-British; GVL, graft-vs.-leukemia; HSCT, hematopoietic stem cell transplantation; HV, healthy volunteer; IL, interleukin; LAA, leukemia-associated antigens; PBMN, peripheral blood mononuclear cells; SEREX, serologic screening of recombinant expression; TAA, tumor-associated antigens; TBP, TATA-box binding protein.

MATERIAL AND METHODS

Tumor cell lines

The human CML cell line K562 and the human AML cell lines HL60 (FAB M2), Kasumi-1 (FAB M2), KG1 (FAB M7) and Oci5 (FAB M4) were obtained from the ATCC (Manassas, VA). Human leukemia cell lines were cultured in RPMI 1640 medium (Biochrom, Berlin, Germany), FCS (10% vol/vol), L-glutamine (2 mM), penicillin (100 units/ml) and streptomycin (100 units/ml).

Cell samples

Pretreatment PBMN cell samples were collected from patients at our institution with de novo AML; the samples contained 60–90% blasts. The mean age of the AML patients was 50.1 years (range 19.1–74.6 years), and samples of 36 male and 24 female patients were examined. The AML cases were subtyped according to the French-American-British (FAB) classification as follows: 2 cases of M0, 6 cases of M1, 18 cases of M2, 5 cases of M3, 13 cases of M4, 15 cases of M5, 1 case of M6 and no case of M7. PBMN were also obtained from HV. These samples were prepared by ficoll separation and stored for RNA preparation at –132°C in liquid nitrogen. CD34 positive cells were obtained from HV by stimulation with granulocyte colony stimulating factor (G-CSF) followed by leukapheresis. Leukapheresis products were separated by CliniMacs columns (Miltenyi, Bergisch Gladbach, Germany) and showed > 98% purity in flow cytometry analysis. After separation, CD34 positive cells were cryopreserved and stored at −132°C in liquid nitrogen.

Informed consent

Samples were taken from HV or AML patients treated at our institution according to clinical study protocols approved by the local ethics committee. Informed consent was obtained from all patients and HV concerning the use of their blood for scientific purposes.

RT-PCR

Total RNA was isolated by the method phenol/chloroform isolation described by Chomczynski and Sacchi27 from PBMN. mRNA was prepared using the mRNA QuickPrep Micro purification kit (Amersham Pharmacia Biotech, Little Chalfont, UK). Each sample (2 μg) was subjected to cDNA synthesis (Superscript II Gibco BRL, Frederick, MD). For RT-PCR the equivalent of 0.1 μg mRNA was used. The sequences of the primers for RT-PCR are listed in Table Ia, as are the temperatures of annealing, MgCl2 concentration and the cycle numbers. For all antigens, denaturation temperature was 94°C and elongation temperature 72°C.

Table IA. Primers and Conditions for Conventional RT-PCR of Immunogenic Antigens in PBMN from AML Patients and for RT-PCR of the House-Keeping gene TBP
AntigenForward primerReverse primerAccession no.TcCyclesC(MgCl2)
BAGE5′ TGG CTC GTC TCA CTC TGG 3′5′ CCT CCT ATT GCT CCT GTT G 3′NM_001187.164321.5 mM
GAGE1/2/85′ GAC CAA GAC GCT ACG TAG 3′5′ CCA TCA GGA CCA TCT TCA 3′G1 NM_001468.156351.5 mM
   G2 NM_001472.1   
   G8 NM_012196.1   
G 250 (CA9)5′ ACT GCT GCT TCT GAT GCC TGT 3′5′ AGT TCT GGG AGC GGC GGG A 3′NM_001216.168352.5 mM
HER2/neu5′ CAT CAA CTG CAC CCA CTC CT 3′5′ GCA GCA GTC TCC GCA TCG TG 3′NM_004448.165321.5 mM
HSJ25′ AGG AGC AGT AGA GTG CTG TC 3′5′ GAC AGC ACT CTA CTG CTC CT 3′NM_008298.156351.5 mM
HTERT5′ CCT CTG TGC TGG GCC TGG ACG ATA 3′5′ ACG GCT GGA GGT CTG TCA AGG TAG 3′NM_003219.168350.5 mM
MAGE A15′ CGG CCG AAG GAA CCT GAC CCA G 3′5′ GCT GGA ACC CTC ACT GGG TTG CC 3′NM_004988.264351.5 mM
MAGE A3/5/65′ AAG GAG AAG ATC TGC CAG TGG GTC TC 3′5′ ACA GTC GCC CTC TTT TGC GAT TAT GG 3′A3 NM_005362.272351.5 mM
   A5 NM_021049.2   
   A6 NM_005363.2   
MAZ5′ CCT TCC GCG ACG TCT ACC ACC TGA 3′5′ CTA CTG CTG CCG CTG CCG CTG 3′NM_002383.168351.5 mM
MPP115′ AAG ATC ATT ATG CAG TTC TTG GAC 3′5′ CCA ATA ACA TCT TTG GCA GTT CT 3′X9826060351.5 mM
NY-ESO-15′ ATG GAT GCT GCA GAT GCG G 3′5′ GGC TTA GCG CCT CTG CCC TG 3′NM_13925060351.5 mM
NY-CO-385′ TGT CCG GCT CCT ACG CAT CA 3′5′ AGT GGC CCG AGC TGT TCC TTA TCT 3′NM_005709.160351.5 mM
NEWREN605′ GAA TCG CCC CAG CCT CTT TG 3′5′ ACT CTG CGC ATC CAC TTT CTT CAG 3′NM_032582.258302.5 mM
PINCH5′ GCC AAC TGC GGG AAG GAG 3′5′ GGA AGC AAA CAT CAC CAA ATA 3′NM_004987.256321.5 mM
PRAME5′ GTC CTG AGG CCA GCC TAA GT 3′5′ GGA GAG GAG GAG TCT ACG CA 3′NM_00611564351.5 mM
Proteinase35′ ACC TCA GTC CAG CTG CCA 3′5′ GAA AGT GCA AAT GTT ATG 3′NM_002777.252351.5 mM
RAGE15′ GTG TCT CCT TCG TCT CTA CTA 3′5′ GGT ATT CCT GAT CCT GTT 3′U46191.157302.5 mM
RBPJk5′ CTG CAG TCT CCA CGT ACG T 3′5′ CAT AGC TTC CCT AGT AAG TC 3′L07875.156350.5 mM
RHAMM5′ CAG GTC ACC CAA AGG AGT CTC G 3′5′ CAA GCT CAT CCA GTG TTT GC 3′NM_012484.160351.5 mM
SCP15′ GTA CAG CAG AAA GCA AGC AAC TGA ATG 3′5′ GAA GGA ACT GCT TTA GAA TCC AAT TTC C 3′NM_003176.160351.5 mM
SSX25′ ATG AAC GGA GAC GAC GCC 3′5′ TGA GGG GAG TTA CTC GTC ATC 3′XM_088714.268351.5 mM
Syntaxin5′ CAG TGG GCA AAG CGA GGT GTT 3′5′ ACT GTG ACG CCA ATG ATG ACT GCT 3′NM_004604.258351.5 mM
WT15′ ATG AGG ATC CCA TGG GCC AGC A 3′5′ CCT GGG ACA CTG AAC GGT CCC CGA 3′NM_000378.264351.5 mM
House-keeping gene      
TBP5′ CAC GAA CCA CGG CAC TGA TT 3′5′ TTT TCT TGC TGC CAG TCT GGA C 3′NM_003194.268241.5 mM

The mRNA expression of the antigens was evaluated in AML patients, in PBMN from HV and in leukemia cell lines (Table II) and in CD34 positive selected samples from HV (Table III). The mRNA expression of the house-keeping gene TATA-box binding protein (TBP), of which hitherto no retro-pseudogenes are known,28 was adjusted to the same expression level. After this standardization procedure for TBP, the expression of the genes of interest was assessed by RT-PCR and classified into one of the following categories: +, mRNA expression of the antigen; –, no detectable mRNA expression of the gene of interest and (+), very low, but detectable mRNA expression in conventional RT-PCR (up to 35 cycles). mRNA isolated from the cell lines K562 and Kasumi-1 testing positive for the genes evaluated served as a positive control. Distilled water (DW), negative cell lines or negative PBMN samples from healthy donors served as negative controls. Furthermore, the mRNA expression of interesting LAA was evaluated in different normal tissues (Table IV) with a normalized standard cDNA tissue panel (Multiple Tissue cDNA Panels BK1420-1 and BK1421-1, Clontech, Palo Alto, CA).

Table II. mRNA Expression Pattern of Immunogenic Antigens in Pretreatment Samples from AML Patients, Samples from Healthy Volunteers and Leukemia Cell Lines (K562, HL-60, KASUMI-1, OCI-5, KG-1) Serving as Positive Controls by Conventional RT-PCR
 AML patientsPBMN of HVLeukemia cell lines1
Antigen+/all%+/all%K562HL-60K1Oci-5KG-1
  • 1

    +, expression; −, no expression.

BAGE8/3027%0/250%++
GAGE 1/2/80/100%0/250%
G 25018/3551%0/250%+
HER2neu0/100%0/250%
HSJ220/20100%25/25100%+++++
hTERT7/2528%0/250%+++
MAGE A10/350%0/250%+
MAGE A3/5/60/100%0/250%+
MAZ15/4038%13/2552%++++
MPP1143/5086%0/250%+++++
NY-ESO-10/100%0/250%
NY-CO-380/100%0/250%
NEWREN6025/3571%7/2528%+++++
PINCH31/3589%17/2568%+++++
PRAME32/5064%0/250%++++
Proteinase 320/3067%7/2528%+++
RAGE10/100%0/250%
RBPJk17/2085%10/2540%++++
RHAMM35/5070%0/250%+++++
SCP10/100%0/250%+
SSX20/200%0/250%+
Syntaxin20/20100%25/25100%+++++
WT140/6067%0/250%+++
Table III. mRNA Expression Pattern of Immunogenic Antigens in CD34 Positive Selected Cells from Healthy Donors
CD 34 positive cells
Antigen+/all1%
  • 1

    +, expression; −, no expression.

B A G E0/100%
h T E R T0/100%
G 2 5 00/100%
M A Z1/138%
M P P 1 11/617%
P IN C H11/1385%
P R A M E0/130%
R H A M M0/130%
WT17/1354%
Table IV. Differential mRNA Expression of LAA in a Standardized Tissue Panel
AntigenSkeletal musclePlacentaPancreasLungLiverKidneyHeartBrainColonOvaryPeripheral bloodProstateSpleemTestisThymus
  1. BAGE and WT1 showed expression only in testis and/or placenta tissue like a cancer/germline antigen, hTERT and G250 were not expressed in any normal tissue. mRNA expression of the antigen PRAME was found in testis, but also in normal ovary and pancreas tissue. The antigens RHAMM and MPP11 showed also very low mRNA expression in pancreas and lung respectively in lung and kidney at a high cycler number of 35 cycles (conventional RT-PCR). MAZ, NewRen60 and proteinase 3 were expressed in different normal tissues, Syntaxin was expressed in all normal tissues.–+, expression; (+), very weak expression;–, no expression.

BAGE+
G250
hTERT
MAZ++++++++++
MPP11(+)(+)+
NEWREN60+++++(+)+++
PRAME++++
Proteinase3++++++++
RHAMM+(+)(+)+
Syntaxin+++++++++++++++
WT1+

Real-time RT-PCR

For the quantitative measurement of the mRNA expression of MPP11, PRAME, RHAMM and WT1 in PBMN from HV and AML patients, as well as in leukemia cell lines, real-time RT-PCR was performed using the light cycler SYBR Green I technology according to the manufacturer's protocol outlined in Table Ib. mRNA (0.1 μg) was used for RT-PCR. To quantify the mRNA expression of these antigens, a conventional PCR was performed and the amount of product cDNA was measured by photometry. A serial dilution of cDNA was subjected to PCR to obtain standard curves. For quantification of the mRNA expression, the amount of mRNA in femtogram (fg) was standardized to the amount of one picogram (pg) of TBP.28 RT-PCR for TBP was performed using the same primers as described for conventional RT-PCR in Table IA and under the conditions outlined in Table IB.

Table IB. Conditions for Quantitative RT-PCR of Immunogenic Antigens in PBMN from AML Patients and for Real-Time RT-PCR of the House-Keeping Gene TBP
GeneDenaturationAnnealingElongationCyclesc (MgCl2)
  1. The conditions for conventional RT-PCR are indicated in Table Ia. For all antigens denaturation temperature was 94°C and elongation temperature 72°C. For the quantitative measurement of the mRNA expression of MPP11, PRAME, RHAMM and WT1 against TBP, real-time RT-PCR was performed using the light cycler SYBR Green I technology according to the manufacturer's protocol and under the conditions described in Table Ib.

MPP1194°C (15 s)62°C (15 s)72°C (60 s)402.5 mM
PRAME95°C (20 s)64°C (10 s)72°C (40 s)402.5 mM
RHAMM95°C (15 s)62°C (10 s)72°C (40 s)453 mM
WT195°C (15 s)62°C (10 s)72°C (22 s)452.25 mM
TBP95°C (10 s)62°C (15 s)72°C (20 s)402 mM

Western blot analysis

Western blot analysis was performed for the antigens WT1 and RHAMM. For the antigens MAZ, PRAME (personal communication with Y. Kawakami, Tokyo and H. Ikeda, Sapporo, Japan) and PINCH, no antibodies were available. For all Western blots, 20 μg protein of all samples was used. For WT1 Western blot analysis, cells (5 × 106 to 1 × 107) were lysed in 1 ml lysis buffer (20 mM Tris base, 2 mM EDTA, 137 mM NaCl, 0.5% triton ×100, pH 7.4) and protease inhibitors phenyl methyl sulfonamide (30 mmol/l) and P.I. cocktail (leupeptin, pepstatin and aprotenin, 10 μg/ml). Equal amounts of protein (20 μg/lane) were separated for SDS-PAGE. Separated proteins were blotted onto nitrocellulose membranes, and additional protein-binding sites on the membrane were blocked with 0.5% defatted milk in Tris-buffered saline. Blots were incubated with primary WT1 c19 antibody (a rabbit polyclonal antibody) specific for the 52 kDa Wilms' tumor nuclear protein WT1 (Santa Cruz, CA) for 1 hr at room temperature, washed and incubated with peroxidase-conjugated goat anti-rabbit immunoglobulin (Ig)G (1:5,000 dilution) and then visualized with chemiluminescence (ECL; Amersham, Quebec, Canada) according to the manufacturer's instructions. For RHAMM, Western blots were incubated with the primary anti-RHAMM antibody, a rabbit polyclonal antibody (generously contributed by Dr. Aβmann, Karlsruhe, Germany) specific for the human 85–90 kDa RHAMM protein recognizing all splice variants. Western blot analysis was performed as described for WT1.

RESULTS

mRNA expression analysis of LAA by conventional RT-PCR

The mRNA expression of different immunogenic antigens was evaluated in PBMN from AML patients, in human leukemia cell lines and in PBMN (Table II), in CD34 positive selected cell samples from HV (Table III), as well as in a panel of normal tissues (Table IV). Moreover, the simultaneous expression of 2 and more antigens was examined (Figs. 1 and 2).

Figure 1.

Conventional RT-PCR for simultaneously expressed LAA in 2 representative patients, A and B. The genes of interest are indicated below the gels and the size of the PCR product is highlighted on the left side. M is the DNA molecular weight marker VI (Roche, Mannheim, Germany) and M* is the 100-bp ladder marker (Roche). On each gel, lane 1 shows the result for PBMN from a HV as a negative control, lane 2 the result for the patient sample and lane 3 the result for the positive control, i.e., for the cell line K562, and for Kasumi-1 in the case of G250. The amount of cDNA tested was normalized by the house-keeping gene TBP (for details see the Material and Methods section).

Figure 2.

Simultaneous mRNA expression of the differentially expressed LAA BAGE, G250, hTERT, MPP11, PRAME, RHAMM and WT1 in AML patients. More than 80% of the AML patients simultaneously expressed 2 or more of these antigens. Only 2 of the AML patients did not show mRNA expression of any of the antigens tested.

Several antigens with described expression in solid tumors showed no mRNA expression in leukemia cell lines and in PBMN of leukemia patients and HV: GAGE-1, -2 and -8; HER2; RAGE1; NY-CO-38 and NY-ESO-1. Other TAA tested showed only positive mRNA expression in leukemia cell lines but not in samples from AML patients: MAGE A1, -A3, -A5, -A6; SSX2 and SCP1 were only expressed in K562 cells.

The information on mRNA expression of the other TAA tested is given in detail in Tables II and III.

No significant differences were detected as for the expression of the LAA in the FAB subtypes M0 to M5.

Expression analysis of LAA by real-time RT-PCR

Quantitative measurement by real-time RT-PCR of mRNA coding for MPP11, PRAME, RHAMM and WT1 was performed in PBMN from 10 AML patients. The amount of mRNA for these genes in PBMN from AML patients and from HV is shown in Figure 3. No mRNA expression of PRAME, RHAMM and WT1 was found in PBMN from HV. MPP11 was overexpressed in AML patients in contrast to PBMN from HV.

Figure 3.

Real-time RT-PCR for the 4 LAA indicated in PBMN from AML patients (n = 10) and HV (n = 10). The antigens RHAMM, PRAME and WT1 showed a leukemia-specific expression in AML. The antigen MPP11 was found to be overexpressed in AML samples. The amounts are indicated in fg of LAA mRNA per 1 pg of mRNA of the housekeeping gene TBP (for details see the Material and Methods section).

mRNA expression analysis of LAA in different normal tissues

Antigens with an interesting expression pattern were examined for their mRNA expression in different normal tissues using a normalized tissue panel (Table IV). The antigens BAGE and WT1 showed an expression pattern as a CG antigen with expression only in testis and/or placenta tissue. hTERT and G250 were not expressed in any tissue examined. mRNA expression of the antigen PRAME was found in testis but also in normal ovary and pancreas tissue. The antigens RHAMM and MPP11 showed a very low mRNA expression in pancreas, in lung and kidney only at a high cycler number of 35 cycles. MAZ, NewRen60 and proteinase 3 were expressed in different normal tissues. Syntaxin was expressed in all normal tissues.

Western blot analysis

Western blot analysis of RHAMM and WT1 was performed in leukemia cell lines and PBMN from AML patients and HV.

Protein expression of WT1 was examined in 5 human leukemia cell lines, in PBMN from 15 AML patients and from HV. K562 cells exhibited high protein expression of WT1 and was used as a positive control. Ten of 15 (66%) samples of AML patients showed positive protein expression. AML cell samples with mRNA expression also exhibited protein expression of WT1. PBMN from HV showed no WT1 protein expression in accordance with the missing mRNA expression of WT1 in RT-PCR assays.

The cell line K562 exhibited high protein expression of RHAMM and was used as a positive control. High protein expression was also detected in all other human leukemia cell lines (5/5) tested. The RHAMM protein was expressed in 10/15 (66%) of PBMN of AML patients. In all PBMN samples from HV, no protein expression of the antigen RHAMM was found.

DISCUSSION

Treatment of patients with AML became more effective during the past decades, but a CR is often not durable and a high percentage of AML patients relapse.1, 7 Therefore, complementary therapeutic approaches are under exploration for the prevention of relapse, the reason for the unfavorable prognosis in AML patients. Immunotherapy directed to LAA might elicit specific CTL responses effectively eliminating the minimal residual disease after chemotherapy or enhancing the GVL effect after HSCT. An ideal antigen for immunotherapies in leukemias should be expressed preferentially in leukemic blasts, but neither on stem cells of normal hematopoiesis nor on normal tissues. This expression pattern was investigated in our study for a panel of immunogenic TAA/LAA.

The results showed 3 categories of antigens with different expression patterns:

  • 1Antigens with no mRNA expression in leukemia patients and HV;
  • 2Antigens with mRNA expression in leukemias but also in PBMN of HV; and
  • 3Antigens with tumor-restricted mRNA expression.

Antigens with no mRNA expression in leukemia patients and HV

The CG antigens of the MAGE family and SSX2 are expressed in different solid tumors.10, 12, 13 In contrast, we found no mRNA expression of these antigens in leukemic blasts of AML patients. Also, the CG antigen, SCP1, and the genes GAGE-1, -2 and -8;9 HER2neu;29, 30 RAGE131 and NY-CO-3832 expressed in solid tumors or in cutaneous T cell lymphomas33, 31 were not expressed in the leukemia cell samples we tested. The lack of expression of TAA in leukemia might be due to the methylation that plays an important role in the regulation of these genes. All of the antigens of this category do not constitute targets for immunotherapies in AML.

Antigens with expression in leukemias but also in PBMN of HV

For the AML antigen proteinase 3, a CTL epitope designated PR1 has been characterized as inducing specific T cell responses.21, 22, 23 However, proteinase 3 is expressed in several normal tissues as well as the genes PINCH, HSJ2, syntaxin, MAZ and RBPJκ.20, 25, 35, 36 Specific immunotherapy targeting ubiquitously expressed genes might attack normal tissues, including the normal hematopoietic system, by triggering autoimmune processes. Proteinase-3 is another designation for c-ANCA that is a well-known antigen in Wegener's granulomatosis, an autoimmune vasculitis.37 Therefore its use for clinical vaccination trials is rather questionable.

Antigens with tumor-restricted expression pattern

The CG antigen BAGE was shown to be expressed in melanomas (22%), bladder carcinoma (15%), breast cancer (10%) and head and neck carcinomas.8 BAGE was more frequently expressed in metastatic melanomas than in primary lesions.8 CTL specific to BAGE have been characterized.8 In our experiments, BAGE showed a tumor-specific expression pattern. Whereas we found BAGE expression in 8/30 AML patients, Adams et al.38 could not detect mRNA for BAGE in 26 patients. This might be due to the time of storage of samples until mRNA was extracted, as well as due to different mRNA extraction methods and PCR conditions.

The antigen G250 was first characterized in renal cell carcinoma, and specific CTL recognizing a specific epitope of G250 were generated,39 but no data about the expression in leukemias have been published. We found mRNA expression of G250 in AML patients that we could demonstrate to be tumor-specific.

The telomerase catalytic subunit (hTERT) is a TAA expressed in different human cancers, and its expression correlates with telomerase activity.40 hTERT peptides binding to HLA-A*0201 and inducing specific CTL responses41 were characterized. In accordance with the literature,38 we did not detect hTERT mRNA expression neither in PBMN from HV nor in normal tissues.

The murine homologue to the human m-phase phosphoprotein 11 (MPP11), the mouse Id associating one gene (MIDA1), is known to regulate cell growth.42, 43, 44 Immunization with a plasmid encoding MIDA1 resulted in significant suppression of tumor growth, whereas no autoimmune effects in this system were observed.45, 46 In rats, Northern blotting showed high expression of MIDA1 in thymus, testis and in tumor, but no expression in normal tissues.45, 46 The high expression level of MPP11 might be responsible for its high immunogenicity and makes this TAA, in synopsis with the high frequency of MPP11 expression in AML patients, an attractive target for immune therapies.

Expression of the gene PRAME was described in melanoma, sarcoma, lung squamous cell carcinoma, renal cell carcinoma14, 33 and 35% of in AML patients,15, 20 whereas no expression was found in PBMN and bone marrow cells from HV.15, 20 Different peptides inducing specific responses by PRAME-specific CTL were characterized.47 Because of the lack of expression in control groups and high frequency of PRAME expression in AML, PRAME seems to be a favorable candidate for further vaccination studies and immunotherapy in AML, despite its expression at a very low level in other tissues like pancreas and ovary.

RHAMM showed tumor-specific humoral immune responses in leukemia and different solid tumors and it was overexpressed in leukemias.24 RHAMM is a receptor for hyaluronic acid mediated motility, playing a fundamental role in cell growth, differentiation and motility,48, 49 and it is overexpressed in different tumors.50, 51 WT1 is expressed in AML blasts, but also in CD34 positive cells of the early hematopoesis.16, 17, 18 WT1 is a gene expressed during embryonic cell differentiation, and it is associated with a poor prognosis in AML.16 Because of the expression in CD34 positive cells, which has been a matter of debate for several years14, 15, 16, 17, 18, 19 and which was confirmed in the present study, immunotherapies using WT1 derived peptides might induce CTL against progenitor cells of the normal early hematopoiesis.

In summary, we characterized the expression of several TAA/LAA in AML patients by the method of conventional RT-PCR, and for selected TAA/LAA also by real-time PCR and Western blotting. The antigens BAGE, G250 and hTERT showed tumor-specific mRNA expression, and MPP11, RHAMM and PRAME showed highly tumor-restricted mRNA expression in leukemic cells. When compared to PRAME, the quantitative mRNA expression level of RHAMM and MPP11 was significantly higher. The frequency of mRNA expression in AML patients was the highest for MPP11, followed by RHAMM, PRAME, G250, hTERT and BAGE. BAGE and hTERT might not be appropriate for clinical studies because of the low frequency of their expression. The antigens RHAMM, PRAME, MPP11 and G250 might therefore be candidates for immunotherapy of leukemias. Because of their simultaneous expression in AML, these antigens also constitute targets for a polyvalent vaccine.

Acknowledgements

We thank Ms. A. Szmaragowska and Ms. S. Braun for their excellent technical work in this project, and Ms. M. Götz for taking care of this article.

Ancillary