Identification of NM23-H2 as a tumour-associated antigen in chronic myeloid leukaemia.
ABSTRACT Therapeutic effects of haematopoietic stem cell transplantation are not limited to maximal chemoradiotherapy and subsequent bone marrow regeneration, but include specific as well as unspecific immune reactions known as graft-versus-leukaemia (GvL) effects. Specific immune reactions are likely to be particularly relevant to the long-term treatment of diseases, such as chronic myeloid leukaemia (CML), in which residual cells may remain quiescent and unresponsive to cytotoxic and molecular therapies for long periods of time. Specific GvL effects result from the expression on leukaemic cells of specific tumour-associated antigens (TAAs) in the context of HLA proteins. As human leukocyte antigen (HLA) types vary widely, the development of broadly applicable tumour vaccines will require the identification of multiple TAAs active in different HLA backgrounds. Here, we describe the identification of NM23-H2 as a novel HLA-A32-restricted TAA of CML cells and demonstrate the presence of specifically reactive T cells in a patient 5 years after transplantation. As the NM23 proteins are aberrantly expressed in a range of different tumours, our findings suggest potential applications beyond CML and provide a new avenue of investigation into the molecular mechanisms underlying CML.
- SourceAvailable from: Barbara Guinn[Show abstract] [Hide abstract]
ABSTRACT: Cancer now affects 1 in 3 people prompting an expansion in new and improved diagnostic techniques. Treatment modes include ways to boost the immune system, which is already adept at destroying infected tissues, and which may be guided to target cancerous cells. This is more commonly known as cancer immunotherapy and it has the potential to provide an effective targeted and personalised medicine for the elimination of cancer cells. We know that some specific proteins are expressed at very low levels in healthy cells but are overexpressed in cancer cells. Such differential expression can be studied in the laboratory (on the bench) and after much evidencing of their clinical utility may be used as therapy targets for cancer patients (at the bedside). To this end we describe how tumour antigens are researched in the literature (in silico) or identified and validated through broader (immuno-) screening programmes. We discuss how animal models are utilised to show the effect of targeting tumour antigens on the development of cancer. Finally we overview how the target antigen will need to stimulate the immune system in a way which best induces effective responses. This chapter overviews current practises with regards to antigen and peptide identification and the journey from bench to bedside.Gene Therapy, 1 01/2013: chapter 12: pages 298-325; INTECH.
- [Show abstract] [Hide abstract]
ABSTRACT: Cancer now affects 1 in 3 people prompting an expansion in new and improved diagnostic techniques. Treatment modes include ways to boost the immune system, which is already adept at destroying infected tissues, and which may be guided to target cancerous cells. This is more commonly known as cancer immunotherapy and it has the potential to provide an effective targeted and personalised medicine for the elimination of cancer cells. We know that some specific proteins are expressed at very low levels in healthy cells but are overexpressed in cancer cells. Such differential expression can be studied in the laboratory (on the bench) and after much evidencing of their clinical utility may be used as therapy targets for cancer patients (at the bedside). To this end we describe how tumour antigens are researched in the literature (in silico) or identified and validated through broader (immuno-) screening programmes. We discuss how animal models are utilised to show the effect of targeting tumour antigens on the development of cancer. Finally we overview how the target antigen will need to stimulate the immune system in a way which best induces effective responses. This chapter overviews current practises with regards to antigen and peptide identification and the journey from bench to bedside.
- [Show abstract] [Hide abstract]
ABSTRACT: For solid tumors of a malignant origin, the expression of the nm23-H1 gene is a positive prognostic factor. However, for chronic myeloid leukemia (CML), the prognostic role of nm23-H1 gene expression is unknown. The present study investigated the impact of nm23-H1 gene expression on the proliferation and migration of the CML K562 cell line to elucidate the association between nm23-H1 gene expression and CML cell survival. An RNAi lipo-recombinant plasmid of the nm23-H1 gene (pGCsi-nm23-H1) was constructed and transfected into the K562 cells. RT-PCR and western blotting were used to detect nm23-H1 mRNA and protein expression, respectively. The anchorage-independent growth ability of the transfected cells was observed in soft agar culture and the ability of the K562 cells to migrate was determined using a Transwell assay. Following the successful construction and transfection of the pGCsi-nm23-H1 plasmid into the K562 cells, nm23-H1 mRNA and protein expression levels were significantly lower compared with the control group. The stably-transfected pGCsi-nm23-H1 K562 cells exhibited a markedly increased ability to form colonies and the number and sizes of the colonies were significantly increased compared with those of the control. In vitro, the cells migrated through a Matrigel-coated membrane during incubation for 20 h. The Transwell assay revealed that the quantitative number of pGCsi-nm23-H1 K562 cells that migrated into the lower compartment of the invasion chamber was markedly increased compared with the control. In conclusion, nm23-H1 gene expression may inhibit K562 cell proliferation and migration. nm23-H1 may be a cancer suppressor gene and play a significant role in inhibiting the survival of CML cells.Oncology letters 10/2013; 6(4):1093-1097. · 0.24 Impact Factor
Identification of NM23-H2 as a tumour-associated antigen in chronic myeloid leukaemia
S Tschiedel1, C Gentilini1, T Lange1, C Wo ¨lfel2, T Wo ¨lfel2, V Lennerz2, S Stevanovic3, H-G Rammensee3, C Huber2, M Cross1
and D Niederwieser1
1Department of Hematology and Internal Oncology, University of Leipzig, Leipzig, Germany;2Department of Medicine III,
Johannes Gutenberg University Mainz, Mainz, Germany and3Department of Immunology, Eberhard Karls University Tu ¨bingen,
Tu ¨bingen, Germany
Therapeutic effects of haematopoietic stem cell transplantation
are not limited to maximal chemoradiotherapy and subsequent
bone marrow regeneration, but include specific as well as
unspecific immune reactions known as graft-versus-leukaemia
(GvL) effects. Specific immune reactions are likely to be
particularly relevant to the long-term treatment of diseases,
such as chronic myeloid leukaemia (CML), in which residual
cells may remain quiescent and unresponsive to cytotoxic and
molecular therapies for long periods of time. Specific GvL
effects result from the expression on leukaemic cells of specific
tumour-associated antigens (TAAs) in the context of HLA
proteins. As human leukocyte antigen (HLA) types vary widely,
the development of broadly applicable tumour vaccines will
require the identification of multiple TAAs active in different
HLA backgrounds. Here, we describe the identification of
NM23-H2 as a novel HLA-A32-restricted TAA of CML cells and
demonstrate the presence of specifically reactive T cells in a
patient 5 years after transplantation. As the NM23 proteins are
aberrantly expressed in a range of different tumours, our
findings suggest potential applications beyond CML and
provide a new avenue of investigation into the molecular
mechanisms underlying CML.
Leukemia (2008) 22, 1542–1550; doi:10.1038/leu.2008.107;
published online 22 May 2008
Keywords: CML; Nm23-H2; TAA
Chronic myeloid leukaemia (CML) is treated with a steady
improvement in outcome having been achieved in recent years
both by the introduction of specific inhibitors of the oncogenic
BCR/ABL fusion protein and by improving the outcome of
haematopoietic cell transplantation (HCT).1–3Both therapeutic
options have different underlying mechanisms, one being a
molecular approach using tyrosine kinase inhibitors and the
other an immunological approach with a more or less intensive
cytostatic component using HCT. For a stem cell disease such as
CML, in which residual populations of tumour cells may remain
in a quiescent state unresponsive either to cytotoxic/radiation
therapy or to specific BCR/ABL inhibitors,4–6it is very likely that
immune surveillance by the donor-derived immune system
plays a decisive role in the long-term control of the disease.
Graft-mediated immune reactions, however, are not restricted
to tumour cells (graft-versus-leukaemia reaction) but are also
directed to normal host cells (graft-versus-host disease).7There is
growing interest in harnessing the antitumour potential of the
incoming graft to maximize reactions against tumour cells while
limiting those against normal host tissues. The molecular basis
for the antigen-specific immune reaction lies in the presentation
of a repertoire of short peptides derived from the degradation of
intrinsic proteins. These peptides are routinely displayed by
human leukocyte antigen (HLA) molecules as ‘self-antigens’ on
the surface of normal cells and/or on the surface of tumour
cells.8Either mutation (neoantigens) or aberrant expression (self-
antigens) of a particular protein in tumour cells can result in the
display of such antigens in the context of HLA molecules and in
tumour-specific cytotoxic T-cell reaction. Alloantigens restricted
to tumour cells or to a specific cell system (for example,
haematopoietic cells) might be interesting for GvL effects
without GvHD. Clearly, the ultimate aim is not only to identify
the target structures of GvL reactions but also to develop
vaccination protocols that maximize specific reactions against
tumour-specific antigens (to increase GvL) while minimizing
those against normal tissues (to control GvH).
Here, we describe, for the first time, the identification of
non-metastasis protein 23-H2 (NM23-H2) as an immunogenic
tumour-associated protein in Phþcells and demonstrate the
detection of specific ex vivo T-cell reactivity in a patient 5 years
Materials and methods
In 1979, a 25-year-old patient was diagnosed with Philadelphia
chromosome-positive (Phþ) CML. A HCT with minimal
conditioning was performed in 2001 from his HLA-identical
sister. The patient was treated with fludarabin (30mg/m2, days
?4 to ?2) and total body irradiation of 2Gy, followed by
immunosuppression with cyclosporine A and mycofenolat
mofetil. Leukaemic peripheral blood mononuclear cells (PBMC)
were collected before HCT during the chronic phase of the
disease. At the time of writing, more than 6 years post-HCT, the
patient is in complete molecular remission.
Cell lines and antigen-presenting cells
An Epstein–Barr virus (EBV)-transformed lymphoblastoid cell
line (EBV-LCL) was established by culturing 5?106PBMC
isolated from the patient pre-transplant in 500ml culture
supernatant from the EBV producer line B95-8 and 1mg/ml
cyclosporine A (Sandimmun; Novartis, Basel, Switzerland). The
EBV-LCL was maintained thereafter in RPM1640 (Biochrom
AG, Berlin, Germany), 10% fetal calf serum (Invitrogen,
Carlsbad, CA, USA), 1% glutamine and penicillin/streptomycin
Received 14 January 2008; revised 17 March 2008; accepted 28
March 2008; published online 22 May 2008
Correspondence: Professor Dr D Niederwieser, Department of
Hematology and Internal Oncology, University of Leipzig, Johannisallee
32A, Leipzig 04103, Germany.
Leukemia (2008) 22, 1542–1550
& 2008 Macmillan Publishers Limited All rights reserved 0887-6924/08 $30.00
(Invitrogen). Fluorescence in situ hybridization revealed the
EBV-LCL to be 100% Phþ.
Phythaemagglutinin (PHA) blasts were generated by stimulat-
ing 0.5?106pre-transplant patient PBMC with 1mg/ml PHA
(Sigma-Aldrich, St Louis, MO, USA) for 72h in AIM-V medium
(Gibco BRL, Gaithersburg, MD, USA) supplemented with 10%
pooled human serum (Sigma-Aldrich) and 250U/ml IL-2
(Chiron-Behring, Marburg, Germany). The cells were further
expanded in the same medium without PHA. At day 20, the
cells were frozen in aliquots and thawed at the day of
the experiment. PHA blasts were Ph?as assessed by reverse
transcription-PCR (Table 1).
Monocyte-derived dendritic cells were used as APC. Mono-
nuclear cells (MNCs) from 100ml blood of the HLA-identical
donor were purified by density gradient centrifugation over
Biocoll (Biochrom AG, Berlin, Germany). Further isolation of
the monocyte fraction was then performed by using a
mini-MACS-positive selection system as described by the
manufacturer (Miltenyi Biotech, Bergisch Gladbach, Germany).
Monocytes were seeded at 7?105cells per ml in CellGro DC
medium (CellGenix, Freiburg, Germany) supplemented with
1000U/ml IL-4 (Cell Concepts, Umkirch, Germany) and
800U/ml granulocyte/macrophage colony-stimulating factor
(Leukomax; Schering-Plough, Kenilworth, NJ, USA) for 6 days.
Maturation of immature dendritic cells was promoted by the
addition of 10ng/ml tumour necrosis factor-alpha (TNFa),
1000U/ml IL-6, 10ng/ml IL-1b (all from Cell Concepts) and
1mg/ml prostaglandin E2(PGE2) (Sigma-Aldrich) for 24–48h.9
Mixed lymphocyte/leukaemic cell culture and IFN-g
Leukaemia-specific mixed lymphocyte/leukaemic cell culture
(MLLC) was performed by co-culturing on 24-well plates 106
donor PBMC and 106irradiated Phþpatient PBMC per well
in 2ml AIM-V medium (Gibco BRL) supplemented with 10%
human serum and recombinant human IL-2 (from day 4, 250
U/ml). MLLC was stimulated on days 7 and 14 with 1?106
irradiated EBV-LCL (Phþ) cells per well and 250U/ml IL-2.
CD8þT lymphocytes were isolated using a CD8þT-cell
Isolation Kit (Miltenyi Biotech). CD8þMLLC was continued by
weekly stimulation of 106T cells with 106irradiated EBV-LCL
(Phþ) and 2?105irradiated CD8?autologous PBMC. CD8þ
MLLC responders were frozen in aliquots on day 20. For
functional assays, CD8þ
MLLCs were restimulated with
irradiated EBV-LCL (Phþ) cells for 4 days after thawing and
then tested in 50ml of AIM-V medium (1?104per well) against
pre- and post-transplant recipient and donor PBMC, recipient
EBV-LCL (Phþleukaemic cells 1?105per well), recipient PHA
blasts (Ph?cells 1?105per well), 293T transfectants (2?104
per well), donor APC (1?104per well) and subpopulations
of cells (1?105per well) in 50ml of AIM-V medium. Sub-
populations were isolated from the blood (CD3þ, CD4þ,
CD8þ, CD14þ) or bone marrow cells (CD34þ) from recipients
and third-party patients using a mini-MACS-positive selection
system as described by the manufacturer (Miltenyi Biotech).
HLA-A*3201-negative as well as HLA-A*3201-positive PBMC
from third-party patients with CML were used to test specificity.
Assays were incubated for 24 or 48h (293T transfectants) at
371C in 5% CO2in air and then developed. The Vectastain Elite
ABC Kit (Axxora Inc., San Diego, CA, USA) was used for
colourimetric detection of spots, which were counted using an
automated image analysis ELISPOT reader system (AID, Strab
HLA expression plasmids
cDNAs encoding HLA-A*3101, HLA-A*3201, HLA-B*4401,
HLA-B*6001, HLA-Cw*0301 and HLA-Cw*0501 were cloned
by reverse transcription-PCR from patient leukaemic cells into
pcDNA3.1 (Invitrogen) as described.10
Construction and screening of the cDNA library
A cDNA library of patient Phþleukaemic cells was constructed
in the vector pcDNA3.1 using a cDNA construction kit
(Invitrogen) and divided into pools of 100 cDNAs. 293T cells
(20000 per well) were then co-transfected with the HLA-cDNA
(100ng per well) and mixed cDNAs prepared from the pools of
leukaemic cells (100 pools, 100–200ng per well) in MultiScreen
ELISPOT MAIPSWU10 plates (Millipore, Bedford, MA, USA)
using PolyFect transfection reagent (Qiagen, Basel, Switzerland)
and tested in an IFN-g ELISPOT assay after 48h. Positive cDNA
pools, which were recognized by the specific cytotoxic T-cell
line (CTL), were then subdivided into 400 pools of 10 colonies
each and retested in the same way. Positive pools were finally
diluted to 400 single clones each and subjected to a final round
cDNA fragments and synthetic peptides
Antigen-coding cDNA fragments were amplified by PCR and
cloned into pcDNA3.1/V5-His TOPO (Invitrogen). 293T cells
were transiently co-transfected with these plasmids together
with HLA-A*3201 cDNA and then tested for recognition by T
cells. Peptides were synthesized by S Rothemund, University of
Leipzig, and then solubilized in PBS or PBS/0.5% dimethyl-
sulphoxide and stored at ?201C. For experiments with peptide
loading, 106APCs per ml in AIM-V medium without serum were
incubated with the indicated concentrations of peptides for 2h
Ex vivo testing of CD8þsorted PBMC before and after
CD8þcells (0.5–2.0?105per well) of the donor and of the
patient before HCT and at different time points after HCT were
obtained from PBMC and tested directly in the ELISPOT assay
against 2?104293T cells per well transfected with HLA-
A*3201 cDNA and either loaded with 10mM peptide or co-
transfected with full-length Nm23-H2 (Figure 6).
revealed myeloid (CD34+and CD14+) but not T-lymphoid (CD4+
and CD8+) cells are affected by the bcr/abl mutation in CML
Sorting of recipient PBMC population pre-transplant
Patient cells (A) Quantitative RT-PCR
Abbreviations: CML, chronic myeloid leukaemia; EBV-LCL, Epstein–
Barr virus-transformed lymphoblastoid cell line; FISH, fluorescence
in situ hybridization; ND, not determined; PBMC, peripheral blood
mononuclear cells; PHA; RT-PCR, reverse transcription-PCR.
*846 total ABL, **1567 total ABL.
NM23-H2 is a TAA in CML
S Tschiedel et al
Fluorescence in situ hybridization, quantitative reverse
transcription-PCR and NM23-H2 protein expression
The percentage of Phþ
interphases was determined by
fluorescence in situ hybridization of at least 30 interphases
with the LSI bcr/abl ES probe (Vysis, Stuttgart, Germany). For
quantitative reverse transcription-PCR, 0.2–8.5?105cells of the
patient’s PBMC and CD4þ, CD8þ, CD14þ
subpopulations were lysed in guanidine isocyanate solution11
and stored at ?201C until used. RNA was extracted with the
RNeasy Micro Kit (Qiagen) and total RNA yield was reverse
transcribed into cDNA with random hexamer primers. The
quantitative measurement of bcr/abl fusion gene transcripts was
carried out by a TaqMan-based real-time quantitative PCR
analysis as described by Gabert et al.12mRNA levels of NM23-
H2 were measured using the QuantiTect Primer Assay (Qiagen).
Results were normalized to RPLP0 and glyceraldehyde-3-
phosphate dehydrogenase (GAPDH). Protein expression of
NM23-H2 was investigated by immunocytochemistry. A total
of 1?105cells per slide were methanol-fixed and stained with
the NM23-H2-specific antibody L-16 (Santa Cruz) and b-actin
antibody (Sigma-Aldrich). Evaluation was carried out using a
confocal laser scanning microscope LSM 510 (Zeiss).
CML-reactive T cells
Peripheral blood mononuclear cells isolated from a CML patient
in the chronic phase of the disease shortly before transplantation
were irradiated and used to stimulate PBMC from the HLA-
identical donor in a MLLC. Fluorescence in situ hybridization
revealed the bone marrow mononuclear cells of the patient to
0 50100 150200250 300
PBMC donor (Ph−, B)
EBV-LCL (Ph+; A)
PHA-stimulated PBMC (Ph−, A)
PBMC (Ph+; A)
EBV-LCL (Ph+, A)
A32− PBMC (K, Ph+)
A32− PBMC (I, Ph+)
A32− PBMC (H, Ph+)
A32− PBMC (G, Ph+)
A32− PBMC (F, Ph+)
A32− PBMC (E, Ph+)
A32+ CD3+ (D, Ph−)
A32+ CD14+ (D, Ph+)
A32+ PBMC (D, Ph+)
mononuclear cell (PBMC), PHA-stimulated PBMC (Ph?), 100% PhþEpstein–Barr virus-transformed lymphoblastoid cell line (EBV-LCL, Phþ) and
subpopulations of chronic myeloid leukaemia (CML) mononuclear cells (CD4þ, Ph?; CD8þ, Ph?; CD14þ; Phþand CD34þ, Phþ) from the
patient pre-transplant (A), as well as PBMC from the donor (B), were tested in 20h IFN-g ELISPOT assays. The specificity of the blocking antibodies
GAP-A3, SFR8-B6, B1.23.2 and W6.32 are HLA-A3; HLA-B60 and HLA-Cw1, HLA-Cw3, HLA-Cw7, HLA-Cw8, HLA-Cw12, HLA-Cw13, HLA-
Cw14, HLA-Cw16; HLA-A31, HLA-A32, HLA-B, HLA-C and HLA-class I, respectively. Patient cells expressed HLA-A*3101, HLA-A*3201, HLA-
B*4401, HLA-B*6001, HLA-Cw*0301, HLA-Cw*0501. (b) HLA-A*3201þCML cell populations, but not HLA-A*3201?CML cells from third-party
patients. T-cell reactivity against HLA-A32þ(D) and HLA-A32?(E–K) cell populations from third-party CML patients was tested in 20h IFN-g
ELISPOT assays. In patient D, Ph?and Phþsubpopulations were used to test specificity. Patient samples D–K are all from third-party patients and
were not used in the MLLC. CTL, cytotoxic T-cell line; MLLC, mixed lymphocyte/leukaemic cell culture.
Specificity of CD8þMLLC (CTL) responder cell population for (a) Philadelphiaþ(Phþ) cells. T-cell reactivity against peripheral blood
NM23-H2 is a TAA in CML
S Tschiedel et al
be 100% Phþ. On day 14 of the MLLC, reactive CD8þT cells
were purified from the culture using a magnetic bead affinity
procedure and expanded for a further 6 days in the continued
presence of irradiated target cells. This procedure generated
over 2?108T cells that were frozen in aliquots for further
The expanded donor T cells were highly reactive against
PBMC of the recipient before HCT. The specificity and HLA
restriction of antigen recognition were assessed by exposure to a
range of cell types and subtypes of the original patient, a range
of cells from third-party patients with CML and by blocking
antibodies to HLA antigens expressed by the patient (Figure 1a).
T cells were found to recognize PBMC and EBV-LCL (both Phþ)
but not PHA-stimulated blasts (Ph?; see Table 1). Testing of
sorted fractions of recipient blood and bone marrow cells
pre-transplant revealed reactivity against myeloid (CD34þ
and CD14þ) but not T lymphoid (CD4þand CD8þ) cells,
consistent with the specific recognition of cells affected by the
bcr/abl mutation. Although reactivity was efficiently blocked by
the W6.32 antibody against all HLA-A, HLA-B and HLA-C
heavy chains14and by the B1.23.2 antibody, which blocks
HLA-B and HLA-C, as well as HLA-A31 and HLA-A32,15the
irrelevant HLA-A3 antibody GAP-A316had no effect. Partial
inhibition was achieved with the SFR8-B6 antibody, which
recognizes the Bw6 determinant on HLA-B and HLA-C
molecules.17Taken together, the blocking experiments con-
firmed HLA-class I restriction of the CTL. HLA restriction was
confirmed using CML cells from unrelated patients with and
without HLA-A*3201. Only HLA-A*3201-positive, bcr/abl-
positive cells (PBMC, CD14þ) but not bcr/abl-negative cells
(CD3þ) from the third-party patient were recognized by our CTL
A cDNA expression library was constructed using mRNA from
the leukaemic population. Approximately 1000 pools of 100
EBV-LCL (Ph+; A)
0 50 100
EBV-LCL (Ph+; A)
150 200250 300
EBV-LCL (Ph+; A)
peptide. CD8þMLLC (CTL) was specific for cDNA pools and single
clones co-expressed with HLA-A*3201 cDNA on 293T cells. The
ELISPOT results are shown for the positive results only from the
sequential screening of pools containing 100 clones (a), 10 clones (b)
and single clones (c). Stimulator cells without tumour antigen (293T/
HLA-A32; 0) were used as negative and EBV-LCL (Phþ; A from patient
pre-transplant) as positive controls. CTL, cytotoxic T-cell line; EBV-
LCL, Epstein–Barr virus-transformed lymphoblastoid cell line; MLLC,
mixed lymphocyte/leukaemic cell culture.
Sequential library screening to identify the immunogenic
0 50 100150 200250
EBV-LCL (Ph+; A)
H1. NM23-H1 and NM23-H2 were cloned into expression vector
pcDNA3.1 and transfected together with HLA-A*3201 cDNA into
293T cells. Transfectants were tested for recognition by the MLLC in a
48-h IFN-g ELISPOT assay. Patient’s pre-transplant EBV-LCL (Phþ; A)
is used as positive control and cells without tumour antigen (293T/
HLA-A32; 0) as negative control. CTL, cytotoxic T-cell line; EBV-LCL,
Epstein–Barr virus-transformed lymphoblastoid cell line; MLLC, mixed
lymphocyte/leukaemic cell culture.
CD8þMLLC (CTL) recognizes NM23-H2 but not NM23-
0 100 200
300400 500 600
EBV-LCL (Ph+, A)
MLLC. The columns indicate the results of the IFN-g ELISPOT assays of
the CD8þMLLC against APC cells of the donor loaded with each of
the three NM23-H2 peptides NM23-H270?78; NM23-H2125?133and
NM23-H2141?149. Patient’s pre-transplant EBV-LCL (Phþ; A) was used
as positive and unloaded APC (APC(B)þ0) as negative control.
APC, antigen-presenting cells; EBV-LCL, Epstein–Barr virus-trans-
formed lymphoblastoid cell line; MLLC, mixed lymphocyte/leukaemic
Specific recognition of peptide NM23-H270?78by CD8þ
NM23-H2 is a TAA in CML
S Tschiedel et al
clones each were picked. Each pool was then transfected into
293T cells together with plasmids expressing each of the
HLA class I proteins present on patient MNCs (HLA-A*3101,
HLA-Cw*0501), making a total of 6000 independent transfec-
tions. The presence of antigen-coding cDNA was determined
by ELISPOT assay (Figure 2a). Among the positive reactions, we
analyzed the one restricted by HLA-A*3201 in detail.
This initial round of screening identified 3 pools of 100 clones
(B2, E10 and E12), which produced a significant ELISPOT
reaction. Each of the pools was then divided into 400 pools of
10 clones each and screened again, resulting in the identifica-
tion of one positive pool (B2.A1; Figure 2b), which originated
from the original pool B2. We prepared and tested 400 single
clones of this pool and found 3 of them to be recognized by the
MLLC (Figure 2c).
Restriction analysis and sequencing showed all three positively
reacting cDNAs to be the same, 600bp long, and to be identical
to NM23-H2.18As many of the tumour antigens characterized to
date result from tumour-specific mutations,19we performed an
intensive mutation screen using multiple NM23-H2-specific
primers. All positive clones were sequenced entirely on both
strands and were found to contain the whole open reading frame
of NM23-H2 without any mutations, compared to the published
human sequence (Gene ID: 4831). To confirm this finding, the
entire open reading frames of NM23-H2 and NM23-H1 (88%
homology to H2) were amplified independently by PCR, cloned
into an expression vector and transfected together with HLA-
A*3201 cDNA into 293T cells. Spot formation of the MLLC
responders clearly confirmed the specific recognition of NM23-
H2 (Figure 3).
Identification of the peptide antigen
Although the anchor amino acids for HLA-A32 have yet to be
defined, a comparison of the two HLA-A*3201-restricted
peptides identified to date revealed a common tryptophan
residue at the C terminus of each nonapeptide.20The NM23-H2
coding region contains three candidate tryptophan residues. In
an attempt to identify HLA-A32-restricted antigenic peptides
from NM23-H2, nonameric peptides corresponding to the
NM23-H2 sequences up to and including each of these
tryptophan residues (NM23-H270?78; NM23-H2125?133 and
NM23-H2141?149) were synthesized. These peptides were
loaded at concentrations of 10mM separately onto HLA-
A*3201-expressing APCs and tested in the ELISPOT assay with
the CD8þMLLC. Under these conditions, significant reactivity
was seen only to NM23-H270?78(SGPVVAMVW, Figure 4).
To obtain a more quantitative measure of the specificity and
avidity of the CD8þ-CTL for the antigenic NM23-H2 peptides,
the ELISPOT was repeated using serial 10-fold dilutions of each
peptide from 10mM down to 1nM (Figure 5). This revealed a very
high level of specificity and avidity for the peptide NM23-
H270?78, such that recognition was still 50% at the lowest
concentration tested (1nM). It is not surprising that the maximal
recognition reached was 85% of EBV-LCL recognition, as the
MLLC showed additional specificities to HLA-B and HLA-C
molecules (see blocking experiments, Figure 1a).
NM23-H270?78-reactive T cells ex vivo
Peripheral blood mononuclear cells of the donor and of the
patient before and at different time points after HCT were
collected and tested directly ex vivo. Interestingly, CD8þT cells
showed low reactivity in the recipient before HCT, but not in the
donor and in the patient up to 3 months after HCT against
NM23-H270?78. Later on, a distinct reactivity against NM23-
H270?78and endogenously expressed full-length NM23-H2 was
found concomitant with bcr/abl negativity (Figure 6).
Expression of NM23-H2 was investigated on the protein and
mRNA level in the pre-transplant PBMC of the patient,
compared with PBMC of the HLA-identical donor. As shown
in Figure 7a, NM23-H2 protein expression was clearly increased
in the CML cells of the patient where NM23-H2 was colocalized
with b-actin in the cytoplasm. In contrast, there was no
upregulation of NM23-H2 at the mRNA level (Figure 7b).
Peptide concentration (nM)
100 1000 10000100000
Specific recognition (%)
HLA-A32þAPC from the donor pulsed with serial dilutions of each NM23-H2 peptide. Specific recognition is shown as the percent of that
seen with patients EBV-LCL (A, Phþ) from pre-transplant cells. APC, antigen-presenting cells; EBV-LCL, Epstein–Barr virus-transformed
lymphoblastoid cell line; MLLC, mixed lymphocyte/leukaemic cell culture.
Dose-dependent peptide recognition by CD8þMLLC. Twenty hours Elispot assays were carried out on the CD8þMLLC stimulated with
NM23-H2 is a TAA in CML
S Tschiedel et al
In the present analysis, we identify for the first time an
immunogenic peptide derived from NM23-H2 as a tumour-
associated antigen in cells from a patient with CML and
demonstrate the presence of NM23-H2-reactive T cells after
HCT. Phþcells (EBV-LCL; CD34þand CD14þ) but not Ph?
cells (CD4þand CD8þ) were recognized by the CD8þ-CTL. In
addition, CML cells from other HLA-A*3201 but not from HLA-
A*3201-negative patients were identified by the CTL.
As no mutations were apparent in the antigenic NM23
sequences cloned from the CML cells, we assume that the
observed immunogenicity resulted from a tumour cell-specific
expression pattern of NM23-H2. An alternative but less likely
possibility is that a mutation in donor NM23-H2 is responsible
for specific recognition of patient cells by donor T cells.
Whatever the basis of the immune recognition, our findings
suggest that an immune response to aberrantly expressed
NM23-H2 may be involved in the long-term GvL effect
following HCT and that the successful stimulation of this
reaction in vivo may provide additional protection against
Non-metastasis protein 23 was first described as a tumour
marker associated with reduced metastatic activity in melano-
ma.21Aberrant NM23 expression (as opposed to mutation) has
since been found to be of clinical significance in a wide range of
cancers, although the consequences of overexpression vary
widely.22In melanoma and in breast cancer, for example, high
mRNA levels are associated with a reduced likelihood of
metastatic progression and a favourable outcome, whereas
increased expression in prostate cancer, neuroblastoma or non-
Hodgkin’s lymphoma is associated with a poor outcome.23–26
Similarly, the overexpression of either of the closely related
nm23-H1 or nm23-H2 genes in AML has been reported to be
20 4060 80
EBV-LCL (Ph+; A)
293T/HLA-A32; NM23-H2(full length)
CD8+ (patient post-HCT) :
cells/10×5 CD8+ (C)
0 5 10 15 2025 30 3540 4550 5560 657075 80 85
Time post-HCT (months)
HCT were tested against HLA-A*3201-transfected 293T (293T/HLA-A32) cells pulsed with peptides NM23-H270?78, NM23-H2125?133,
NM23-H2141?149or cotransfected with full-length NM23-H2 in a 48-h ELISPOT assay. EBV-LCLs (Phþ; A) were used as a positive control and
293T/HLA-A32 without peptide as a negative control. (b) CD8þPBMC of the patient before (A) and at different time points after HCT (C), as well as
of the donor before HCT were tested against HLA-A*3201-transfected 293T cells, pulsed with 10mM peptide NM23-H270?78in the ELISPOT assay.
bcr/abl was measured in the peripheral blood of the patient and recipient using reverse transcription-PCR as described in the Materials and
methods section. EBV-LCL, Epstein–Barr virus-transformed lymphoblastoid cell line; MLLC, mixed lymphocyte/leukaemic cell culture; PMBC,
peripheral blood mononuclear cell.
Ex vivo reactivity of CD8þPBMC against NM23-H2 peptides and full-length NM23-H2. (a) CD8þPBMC of the patient 62 months after
NM23-H2 is a TAA in CML
S Tschiedel et al
associated with reduced survival.27,28However, the evidence to
date for an involvement of NM23-H2 in CML is very scant, with
two reports of an increase in mRNA levels during blast
crisis.29,30Despite the implication of a general involvement of
NM23 proteins in tumourigenesis, there is currently no clear
consensus concerning the underlying mechanisms. Both NM23-
H1 and NM23-H2 are nucleoside diphosphate kinases and are
likely to play multiple roles in coordinating the balance of
nucleoside diphosphates and triphosphates available for gene
expression, DNA synthesis, signalling and metabolism. Further-
more, NM23-H1 and NM23-H2 each interact directly with
other regulatory proteins controlling cell fate.22Specifically,
NM23-H1 interacts directly with a wide range of proteins
involved in the regulation of signalling, metabolism, apoptosis
and cell division,22whereas NM23-H2 is a transcriptional
activator of c-myc,18,31,32which in turn can activate expression
of both the nm23-H1 and nm23-H2 genes.33
As overexpression of c-Myc protein is closely associated with
CML progression,34,35alterations in NM23-H2 expression in
CML seem likely. It is therefore somewhat surprising that we can
find no evidence for a general upregulation of NM23-H2 mRNA
in CML cells, including those from which the antigenic
sequences were cloned and which are recognized by the
specific T cells. This suggests that the CML-specific changes in
regulation may be at a post-translational level. Indeed, both the
c-myc and nm23-H2 genes are subject to regulation at the level
of mRNA translation, with c-myc translation being dependent
on signalling through the ras/map kinase pathway, whereas
NM23-H2 translation appears to respond to the Akt/mTOR
pathway.36–38This would be consistent with a recent study
indicating that a number of peptides presented predominantly
on tumour tissue showed no or only minor changes in mRNA
expression levels compared with normal tissue.39
Regulatory networking of the c-myc and nm23 genes
involving decisive changes at the translational or post-transla-
tional level may help to explain why the correlations made
between mRNA levels and prognosis in various cancers are
highly context-dependent. For these reasons, we are currently
testing the possibility that CML cells accumulate NM23-H2
protein without significant increases in the levels of specific
mRNA, and preliminary results support this theory. Although it
is possible that BCR/ABL exerts specific effects on NM23-H2
expression by the activation of signalling pathways and c-Myc
expression, recent reports demonstrate a broad effect of
BCR/ABL activity on tumour antigen expression. Hence, the
increased expression of survivin, adipophilin, hTERT, WT-1,
Bcl-x1 and Bcl-2 is mediated by BCR/ABL and reversed by
imatinib mesylate,40whereas BCR/ABL-transfected dendritic
cells elicit T-cell responses with specificities very similar to
those found in CML patients in major molecular remission
following imatinib mesylate treatment.40,41
Regardless of the molecular mechanisms affecting NM23-H2
expression in CML, the detection of a specific T-cell reaction in
a transplanted patient implies that NM23 can mediate clinically
relevant immunogenicity and may therefore be a candidate
leukaemia-specific antigen for therapeutic manipulation of the
immune response. A variety of leukaemia-specific antigens have
been identified in recent years. One of the most obvious
candidates for CML is the BCR/ABL protein itself, and as BCR/
ABL-derived peptides have indeed been shown to elicit a
cytotoxic T-cell response in vitro,42their potential as tumour
vaccines is currently being assessed.43,44Primary granule
proteins, which act as autoantigens in the autoimmune reactions
of Wenger’s Granuloma and related vasculitis, have also been
identified as promising candidate for myeloid leukaemia-
specific antigens. Two of these, PR3 and NE, have been studied
in detail. PR3 was originally identified as an HLA-A0201-
restricted peptide that induced myeloid-specific CTL responses.
These CTL can be expanded in vitro and are cytotoxic to CML
cells.45Encouraging results in patients with refractory or
progressing myeloid leukaemia have recently led to the
initiation of immunotherapy studies with PR3 in less advanced
patients. Finally, a screen of previously characterized leukae-
mia-specific antigens with specific relevance to CML has
identified further candidates for CML immunotherapy, at least
one of which (RHAMM/CD168-R3) appears to have elicited
specific T-cell responses in transplanted patients.46There is,
therefore, good reason to believe that immunotherapies based
on TAAs could improve long-term outcome in CML patients.47It
should be noted that, whereas the antigenic peptides previously
described were identified by ‘reverse immunology’, the NM23
peptide reported here was identified from an MLLC by cDNA
To assess the ultimate therapeutic potential of peptide
vaccines derived from NM23, it will be necessary to determine,
firstly, whether or not aberrant NM23-H2 expression is a
widespread feature of CMLs and, secondly, whether the protein
generates peptides that can act as functional antigens in HLA
backgrounds other than HLA-A32. Given the widespread
involvement of NM23 proteins in tumourigenesis, it will also
be interesting to investigate the potential relevance of NM23-H2
as a therapeutic TAA in other cancers. In the meantime, the
regulatory interdependence of NM23-H2 and c-myc provides
a basis from which to design specific studies to elucidate the
(A, 100% Ph+)
PBMC (B, Ph−)
PBMC (A, Ph+)
Donor PBMC (B, Ph−)
(a) Protein expression of NM23-H2 is increased in patient’s pre-
transplant PBMC (Phþ) as assessed by immunocytochemistry. Cells
were stained with NM23-H2 (red) and b-actin (green). Merge indicates
overlaying of the staining patterns. (b) mRNA expression of NM23-H2
is not upregulated in patient’s pre-transplant PBMC (Phþ), compared
to donor PBMC (Ph?; P¼NS). Assays were performed in triplicate and
normalized to the mRNA level of five healthy donors. PBMC,
peripheral blood mononuclear cell.
Protein and mRNA expression of NM23-H2 in CML cells.
NM23-H2 is a TAA in CML
S Tschiedel et al
cells, which should contribute to our understanding of the
progression of CML.
of NM23 proteinsin normaland leukaemic
We thank A Jilo for technical assistance. This work was supported
by Grant 70-3344 (TP IV C/D) from the Deutsche Krebshilfe and
by Grant D08 from the IZKF Leipzig to DN.
1 Gratwohl A, Brand R, Apperley J, Crawley C, Ruutu T, Corradini P
et al. Allogeneic hematopoietic stem cell transplantation for
chronic myeloid leukemia in Europe 2006: transplant activity,
long-term data and current results. An analysis by the Chronic
Leukemia Working Party of the European Group for Blood
and Marrow Transplantation (EBMT). Haematologica 2006; 91:
2 Martinelli G, Iacobucci I, Soverini S, Cilloni D, Saglio G, Pane F
et al. Monitoring minimal residual disease and controlling drug
resistance in chronic myeloid leukaemia patients in treatment with
imatinib as a guide to clinical management. Hemat Oncol 2006;
3 Ohno R. Treatment of chronic myeloid leukemia with imatinib
mesylate. Int J Clin Oncol 2006; 11: 176–183.
4 Holtz MS, Bhatia R. Effect of imatinib mesylate on chronic
myelogenous leukemia hematopoietic progenitor cells. Leuk
Lymphoma 2004; 45: 237–245.
5 Holtz MS, Forman SJ, Bhatia R. Nonproliferating CML CD34+
progenitors are resistant to apoptosis induced by a wide range of
proapoptotic stimuli. Leukemia 2005; 19: 1034–1041.
6 Copland M, Jorgensen HG, Holyoake TL. Evolving molecular
therapy for chronic myeloid leukaemiaFare we on target?
Hematology 2005; 10: 349–359.
7 Hogan WJ, Deeg HJ. Stem cell transplantation: graft-mediated
antileukemia effects. Methods Mol Med 2005; 109: 421–444.
8 Cohen CJ, Sarig O, Yamano Y, Tomaru U, Jacobson S, Reiter Y.
Direct phenotypic analysis of human MHC class I antigen
presentation: visualization, quantitation, and in situ detection of
human viral epitopes using peptide-specific, MHC-restricted
human recombinant antibodies.
9 Romani N, Reider D, Heuer M, Ebner S, Ka ¨mpgen E, Eibl B et al.
Generation of mature dendritic cells from human blood: an
improved method with special regard to clinical applicability.
J Immunol Methods 1996; 196: 137–151.
10 Ennis PD, Zemmour J, Salter RD, Parham P. Rapid cloning of HLA-
A, B cDNA by using the polymerase chain reaction: frequency and
nature of errors produced in amplification. Proc Natl Acad Sci USA
1990; 87: 2833–2837.
11 Chomczynski P, Sacchi N. Single-step method of RNA isolation by
acid guanidinium thiocyanate-phenol-chloroform extraction. Anal
Biochem 1987; 162: 156–159.
12 Gabert J, Beillard E, van der Velden VH, Bi W, Grimwalde D,
Pallisgaard N et al. Standardization and quality control studies of
‘real-time’ quantitative reverse transcriptase polymerase chain
reaction of fusion gene transcripts for residual disease detection in
leukemiaFa Europe Against Cancer program. Leukemia 2003; 17:
13 Lennerz V, Fatho M, Gentilini C, Frye RA, Lifke A, Ferel D et al.
The response of autologous T cells to a human melanoma is
dominated by mutated neoantigens. Proc Natl Acad Sci USA 2005;
14 Barnstable CJ, Bodmer WF, Brown G, Galfre G, Milstein C,
Williams AF et al. Production of monoclonal antibodies to group A
erythrocytes, HLA and other human cell surface antigensFnew
tools for genetic analysis. Cell 1978; 14: 9–20.
15 Rebai N, Malissen B. Structural and genetic analyses of HLA class I
molecules using monoclonal xenoantibodies. Tissue Antigens
1983; 22: 107–117.
16 Berger AE, Davis JE, Cresswell P. Monoclonal antibody to HLA-A3.
Hybridoma 1982; 1: 87–90.
17 Radka SF, Kostyu DD, Bernard D, Amos B. A monoclonal antibody
directed against the HLA-Bw6 epitope. J Immunol 1982; 128:
18 Postel EH, Berberich SJ, Flint SJ, Ferrone CA. Human c-myc
transcription factor PuF identified as nm23-H2 nucleoside dipho-
sphate kinase, a candidate suppressor of tumor metastasis. Science
1993; 261: 428–429.
19 Van den Eynde BJ, van der Bruggen P. T cell defined tumor
antigens. Curr Opin Immunol 1997; 9: 684–693.
20 Harrer T, Harrer E, Kalams SA, Barbosa P, Trocha A, Johnson RP
et al. Cytotoxic T lymphocytes in asymptomatic long-term
21 Steeg PS, Bevilacqua G, Kopper L, Thorgeirsson UP, Talmadge JE,
Liotta LA et al. Evidence for a novel gene associated with low
tumor metastatic potential. J Natl Cancer Inst 1988; 80: 200–204.
22 Roymans D, Willems R, Van Blockstaele DR, Slegers H. Nucleo-
side diphosphate kinase (NDPK/NM23) and the waltz with
multiple partners: possible consequences in tumor metastasis.
Clin Exp Metastasis 2002; 19: 465–476.
23 Aryee DN, Simonitsch I, Mosberger I, Kos K, Mann G, Schlogl E
et al. Variability of nm23-H1/NDPK-A expression in human
lymphomas and its relation to tumour aggressiveness. Br J Cancer
1996; 74: 1693–1698.
24 van Noesel MM, Versteeg R. Pediatric neuroblastomas: genetic
and epigenetic ‘danse macabre’. Gene 2004; 325: 1–15.
25 Hartsough MT, Steeg PS. Nm23/Nucleoside diphosphate kinase in
human cancers. J Bioenerg Biomem 2000; 32: 301–307.
26 Kauffman EC, Robinson VL, Stadler WM, Sokoloff MH, Rinker-
Schaeffer CW. Metastasis suppression: the evolving role of
metastasis suppressor genes for regulating cancer cell growth at
the secondary site. J Urol 2003; 169: 1122–1133.
27 Okabe-Kado J, Kasukabe T. Physiological and pathological
relevance of extracellular NM23/NDP kinases. J Bioenerg Biomem
2003; 35: 89–93.
28 Okabe-Kado J, Kasukabe T, Honma Y. Differentiation inhibitory
factor Nm23 as a prognostic factor for acute myeloid leukemia.
Leuk Lymphoma 1998; 32: 19–28.
29 Wakimoto N, Yokoyama A, Mukai Y, Kuwada N, Yamashita T,
Matsumura T et al. Elevated expression of differentiation inhibitory
factor nm23 mRNA in monoblastic crisis of a patient with chronic
myelogenous leukemia. Int J Hematol 1998; 67: 313–318.
30 Yokoyama A, Okabe-Kado J, Wakimoto N, Kobayashi H, Sakashita
A, Maseki N et al. Evaluation by multivariate analysis of the
differentiation inhibitory factor nm23 as a prognostic factor in
acute myelogenous leukemia and application to other hematologic
malignancies. Blood 1998; 91: 1845–1851.
31 Postel EH. Multiple biochemical activities of NM23/NDP kinase in
gene regulation. J Bioenerg Biomem 2003; 35: 31–40.
32 Postel EH, Berberich SJ, Rooney JW, Kaetzel DM. Human NM23/
nucleoside diphosphate kinase regulates gene expression through
DNA binding to nuclease- hypersensitive transcriptional elements.
J Bioenerg Biomem 2000; 32: 277–284.
33 Godfried MB, Veenstra M, v Sluis P, Boon K, v Asperen R, Hermus
MC et al. The N-myc and c-myc downstream pathways include the
chromosome 17q genes nm23-H1 and nm23-H2. Oncogene
2002; 21: 2097–2101.
34 Gopal V, Kadam P, Preisler H, Hulette B, Li YQ, Steele P et al.
Abnormal regulation of the myc gene in myeloid leukemia. Med
Oncol Tumor Pharmacother 1992; 9: 139–147.
35 Handa H, Hegde UP, Kotelnikov VM, Mundle SD, Dong LM,
Burke P et al. BCL-2 and C-MYC expression, cell cycle kinetics
and apoptosis during the progression of chronic myelogenous
leukemia from diagnosis to blastic phase. Leuk Res 1997; 21:
36 Gupta S, Seth A, Davis RJ. Transactivation of gene expression by
Myc is inhibited by mutation at the phosphorylation sites Thr-58
and Ser-62. Proc Natl Acad Sci USA 1993; 90: 3216–3220.
37 Vervoorts J, Lu ¨scher-Firzlaff J, Lu ¨scher B. The ins and outs of MYC
regulation by posttranslational mechanisms. J Biol Chem 2006;
38 Joosten M, Bla ´zquez-Domingo M, Lindeboom F, Boulme ´ F,
Van Hoven-Beijen A, Habermann B et al. Translational control
JImmunol 1996; 156:
NM23-H2 is a TAA in CML
S Tschiedel et al
of putative protooncogene Nm23-M2 by cytokines via phosphoi-
nositide 3-kinase signaling. J Biol Chem 2004; 279: 38169–38176.
39 Weinzierl AO, Lemmel C, Schoor O, Mu ¨ller M, Kru ¨ger T,
Wernet D et al. Distorted relation between mRNA copy
number and corresponding major histocompatibility complex
ligand density on the cell surface. Mol Cell Proteomics 2007; 6:
40 Brauer KM, Werth D, von Schwarzenberg K, Bringmann A, Kanz L,
Gru ¨nebach F et al. BCR-ABL activity is critical for the immuno-
genicity of chronic myelogenous leukemia cells. Cancer Res 2007;
41 Scheich F, Duyster J, Peschel C, Bernhard H. The immunogenicity
of Bcr-Abl-expressing dendritic cells is dependent on the Bcr-Abl
kinase activity and dominated by Bcr-Abl-regulated antigens.
Blood 2007; 110: 2556–2560.
42 Bocchia M, Korontsvit T, Xu Q, Mackinnon S, Young Yang S,
Sette A et al. Specific human cellular immunity to bcr-abl
oncogene-derived peptides. Blood 1996; 87: 3587–3592.
43 Rusakiewicz S, Molldrem JJ. Immunotherapeutic peptide vaccina-
tion with leukemia-associated antigens. Curr Opin Immunol 2006;
44 Bocchia M, Abruzzese E, Forconi F, Ippoliti M, Trawinska MM,
Pirrotta MT et al. Imatinib does not impair specific antitumor T-cell
immunity in patients with chronic myeloid leukemia. Leukemia
2006; 20: 142–143.
45 Barrett AJ, Rezvani K. Translational mini-review series on
vaccines: peptide vaccines for myeloid leukaemias. Clin Exp
Immunol 2007; 148: 189–198.
46 Schmitt M, Li L, Giannopoulos K, Chen J, Brunner C, Barth T
et al. Chronic myeloid leukemia cells express tumor-associated
antigens eliciting specific CD8? T-cell responses and are lacking
costimulatory molecules. Exp Haematol 2006; 34: 1709–1719.
47 Dermime S, Gilham DE, Shaw DM, Davidson EJ, Meziane EK,
Armstrong A et al. Vaccine and antibody-directed T cell tumour
immunotherapy. Biochimica Et Biophysica Acta 2004; 1704:
NM23-H2 is a TAA in CML
S Tschiedel et al