British Journal of Haematology, 2000, 109, 842±846
SHORT REPORT
Genetic instability in myelodysplastic syndrome: detection of
microsatellite instability and loss of heterozygosity in bone
marrow samples with karyotype alterations
Lienhard Maeck, 1 Detlef Haase, 1 Claudia Schoch, 2 Wolfgang Hiddemann 2 and Frauke Alves 1 1 Department
of Haematology and Oncology, Georg-August-University, GoÈttingen, Germany, and 2 Department of Internal Medicine III,
University Hospital Grosshadern, Ludwig-Maximilians-University, MuÈnchen, Germany
Received 2 February 2000; accepted for publication 28 February 2000
Summary. Using a polymerase chain reaction (PCR)-
based approach, we examined the prevalence of loss of
heterozygosity (LOH) and microsatellite instability (MSI) in
relation to chromosomal imbalances in myelodysplastic
syndrome (MDS). Two of 26 patients displayed MSI (8%),
one of them at five loci. LOH was detected in six out of 26
cases (23%), predominantly involving markers IRF1 [5q31]
and WT1 [11p]. Two patients displayed a corresponding
chromosomal deletion by conventional cytogenetics. Sup-
porting the mutator phenotype hypothesis, a significant
The initial genetic event of the presumed multistep
pathogenesis of myelodysplastic syndrome (MDS) is an
unknown submicroscopic stem cell defect, which, in a
subset of cases, is followed by non-clonal karyotype
instability (Haase et al, 1992). In approximately half of the
patients, clonal karyotypic abnormalities, which are detect-
able at the stem cell level, occur (Haase et al, 1997). By
unravelling loss of heterozygosity (LOH) and microsatellite
instability (MSI), microsatellite polymorphic markers are
useful to test the status of DNA mismatch repair (MMR) and
cell cycle-conducting tumour-suppressor genes, both of
which, if deficient, promote genetic instability (Loeb, 1998).
LOH describes the homozygous state of a distinct
chromosomal region and points to the presence of a closely
located inactivated tumour-suppressor gene that might be
involved in malignant transformation. MSI is defined by the
occurrence of novel microsatellite alleles in neoplastic DNA
when compared with DNA from non-malignant tissue of the
same individual. It is a hallmark of patients with hereditary
non-polyposis colorectal cancer (HNPCC) and related
Correspondence: F. Alves, Department of Haematology and
Oncology, Georg-August-University, Robert-Koch-Str. 40, 37075
GoÈttingen, Germany. E-mail: falves@gwdg.de
842
coincidence of LOH, MSI and chromosome abnormalities
was observed (P , 0´025). Moreover, our data suggest that
LOH represents an initial rather than a secondary genetic
event in MDS, promoting genetic instability in a subset of
patients.
Keywords: myelodysplastic syndrome, microsatellite
instability, loss of heterozygosity, interferon regulatory
factor 1, cytogenetics.
malignancies as the consequence of a defective DNA MMR
machinery. MSI has also been observed in sporadic tumours
and distinct haematological disorders, suggesting that DNA
proofreading failure may be a common pathomechanism in
neoplastic transformation.
To verify whether both mechanisms might be important
at the onset of de novo MDS, bone marrow samples were
analysed for LOH and MSI using a panel of 18 microsatellite
repeats located on 15 different chromosomal arms. From
each patient, buccal smears were taken as a source for
constitutional DNA. The results of the microsatellite
analyses were related to corresponding cytogenetic data.
PATIENTS AND METHODS
Patients. Twenty three patients with de novo MDS, one
patient with therapy-related MDS (t-MDS) and two patients
with acute myeloid leukaemia secondary to MDS (sAML)
secondary to MDS (sAML) were recruited for microsatellite
analyses (Table IA). Bone marrow samples and buccal
smears were collected after informed consent.
Cytogenetics. Chromosome analyses were performed using
a modified GAG-banding technique after short-term cultures
of bone marrow samples as described previously (Stollmann
q 2000 Blackwell Science Ltd
 Table IA. Karyotypes of the patients examined and results of the microsatellite analyses.
q 2000 Blackwell Science Ltd, British Journal of Haematology 109: 842±846

Patient no. Sex Age (years) FAB classification Cytogenetics LOH/MSI

1 M 62 RAEB 46,XY [22]

2 M 57 RAEB 46,XY,del(1)(p33),i(14)(q10) [18]/46,XY [2]
3 F 73 RA 47,XX, 18 [2]/47,XX, 1 11 [1]/46,XX [42]
4 M 61 RAEB 44,XY,dic(1; 22)(p36; p11)(dup(1)(q11p35),del(5)(q13q33),-7,add(17)(p1?1) [9]/
44,XY,dic(1; 22)(p36; p11)(dup(1)(q11p35),del(1)(q11),del(5)(q13q33),-7,add
(17)(p1?1) [3]/46,XY [15]
5 M 62 RAEB 46,XY [25]
6 F 46 t±RAEB 45,XX, 27 [18]/45,XX, 23, 27, 18 [2]/46,XX, 27, 121 [4]/90,XXXX, 27, 27 [1] CACNL1A3, WT1
sporadic chromosomal alterations [12/25]
7 M 72 RAEB 46,XY [1]/45,X,-Y [18]/44,X,-Y,-22 [4] LPL, D10S197,
D12S89, RB1, DCC
8 M 67 RARS 47,XY,t(9; 21)(q13; q22), 1der(9)t(9; 21)(q13; q22) [24]/46,XY [1]
sporadic chromosomal alterations in 5 metaphases
9 F 60 RA 47±50,XX,del(5)(q13),der(?)ins(?; 5)(?; q?) 2,-13, 1der(13)(?), 1del(14)(q13), IRF1, TP53
11 2 3mar [cp24]/additional sporadic chromosomal alterations in 10 metaphases/46,XX [1]
10 M 75 RAEB 46,XY [4]/47,XY, 18 [8]
11 M 64 RAEB-T 46,XY [23]
12 F 62 RAEB 46,XX [10]/44,XX,del(5)(q15q31),-7,-20 [2]/44,XX,del(5)(q15q31),-7,del(17)(p11.2),-20 [2]/
44,XX,del(5)(q13),-7,-20 [4]/44,XX,del(5)(q13),-7,dic(14; 17)(p11; p11), 117,-20 [1]/
80±90,XXXX,idemx2 [cp3]/46,XX,dic(14; 15)(p11; p11) [2]
13 M 59 RARS 46,XY [25]
14 F 73 sAML 46±48,XX, 1der(1), 216, 221, 222, 12 23 mar[cp16]/46,XX [4]
15 M 78 RA 46,XY [25]
16 M 48 46,XY [25]/sporadic chromosomal alterations [9/25] IRF1
17 M 59 CMML 46,XY [25]/sporadic chromosomal alterations [5/25]
18 M 65 sAML 47,XY, 18 [19]/46,XY [1] IRF1
19 M 48 CMML 46,XY [25]; sporadic chromosomal alterations [3/25] D12S89
20 M 80 RAEB±T 46,XY [5]/41±56,XY, ±Y,del(2)(q31),del(3)(p21),del(5)(q15),add(6)(q?), 27,der(9)(q?), 210, 111, 112, IRF1
add(12)(p13),t(12; 15)(p13; q11.2),-16,-18,der(20)t(3; 20)(q21; p13), 11±4Mar [cp19]/occurrence of
polyploid metaphases (10%)
21 M 70 RAEB-T 46,XY [29]/46,XY,inv(1)(p36q21) [1]/continued as clonal abnormality at initial diagnosis in 1993 WT1
22 M 61 RAEB 46,XY del(12)(p11.2) [19]/46,XY [1]
23 M 61 RA 46,XY [25]/sporadic chromosomal alterations [3/25]
24 M 66 RA 46,XY [20]
25 F 65 RA 46,XX [18]

Short Report
26 F 56 RARS 46,XX [25]

FAB, French±American±British classification; RA, refractory anaemia; RARS, refractory anaemia with ring sideroblasts; RAEB, refractory anaemia with excess blasts; CMML. chronic
myelomonocytic leukaemia; RAEB-T, refractory anaemia with excess blasts in transformation; t-RAEB, therapy-related refractory anaemia with excess blasts; sAML, acute myeloid leukaemia
secondary to myelodysplastic syndrome.

843
844 Short Report
Table IB. Loci and primers.

Marker Chromosomal location Length (bp) Repeat motif Primer sequences

BAT-40 1p13.1 , 80±100 TTTT.TT. (T7)¼¼¼.TTTT. (T)40 5 0 -ATTAACTTCCTACACCACAAC-3 0

5 0 -GTAGAGCAAGACCACCTTG-3 0
CACNL1A3 1q31/32 141±157 Not reported 5 0 -GCTGAGCTAGCGAGGGGCAGGGT-3 0
5 0 -CCCAGCAAAAACTGAGTGTGGATG-3 0
BAT-26 2p , 80±100 (T)5¼. (A)26 5 0 -TGACTACTTTTGACTTCAGCC-3 0
5 0 -AACCATTCAACATTTTTAACCC-3 0
D4S171 4q33/35 43±161 (AC)20AG(AGAC)5AGA 5 0 -TGGGTAAAGAGTGAGGCTG-3 0
5 0 -GGTCCAGTAAGAGGACAGT-3 0
APC 5q21/22 96±122 (CA)n 5 0 -ACTCACTCTAGTGATAAATCGGG-3 0
5 0 -AGCAGATAAGACAGTATTACTAGTT-3 0
IRF1 5q31 , 243 (CA)n 5 0 -CTGGAATCTCTATGGCAGATAGGTC-3 0
5 0 -GTGCCCAGGTAGGAAGGGGCTTTAC-3 0
D7S522 7q31 217±229 (CA)n 5 0 -GCCAAACTGCCACTTCTC-3 0
5 0 -ACGTGTTATGCCACTCCC-3 0
D8S87 8p12 145±157 Not reported 5 0 -GGGTTGTTGTAAATTAAAAC-3 0
5 0 -TGTCAAATACTTAAGCACAG-3 0
LPL 8p22 106±134 Not reported 5 0 -TAGAGCACACTATCCAGGTGA-3 0
5 0 -CAGTGGGTTATTTGTGGGATA-3 0
D10S197 10qter 161±173 CACCAGA(CA)7.A.A(CA)12(AGAAA)2 5 0 -ACCACTGCACTTCAGGTGAC-3 0
5 0 -GTGATACTGTCCTCAGGTCTCC-3 0
WT1 11p13 40±148 (CA)n 5 0 -AATGAGACTTACTGGGTGAGG-3 0
5 0 -TTACACAGTAATTTCAAGCAACGG-3 0
D12S89 12p13.2 254±288 (CA)n 5 0 -ATTTGAGAGCAGCGTGTTTT-3 0
5 0 -CCATTATGGGGAGTAGGGGT-3 0
RB1 13q14.3 266±306 (CTTT[T])n 5 0 -CTCCTCCCTACTTACTTGT-3 0
5 0 -AATTAACAAGGTGTGGTGGTACACG-3 0
NF1 17q11.2 171±187 (CA)n 5 0 -CAGAGCAAGACCCTGTCT-3 0
5 0 -CTCCTAACATTTATTAACCTTA-3 0
NM23±H1 17q22 94±104 (CA)n 5 0 -TTGACCGGGGTAGAGAACTC-3 0
5 0 -TCTCAGTACTTCCCGTGACC-3 0
TP53 17p13.1 103±135 (A)14GAAAAGAAAAAGAAAAGAA 5 0 -AGGGATACTATTCAGCCCGAGGTG-3 0
A(.)51(CA)24 5 0 -ACTGCCACTCCTTGCCCCATTC-3 0
DCC 18q21 , 215 Not reported 5 0 -GATGACATTTTCCCTCTAGA-3 0
5 0 -TTTAGTGGTTATTGCCTTGAA-3 0
PLCpr 20q12/13.1 150±184 (CA)n 5 0 -AACCAGTCTGCTCTTCCGGTG-3 0
5 0 -CTGCCTTCAACTGATCTCAATGG-3 0

et al, 1985). Chromosomes were classified according to the
1995 International System for Human Cytogenetic Nomen-
clature. The term sporadic chromosomal alterations is used
in cases with unequivocal chromosome and/or chromatid
breaks occurring in at least two metaphases, as well as in
cases with sporadic deletions, translocations and inversions
occurring in only one metaphase.
DNA isolation. Genomic DNA from bone marrow samples
and buccal smears was isolated by ethanol precipitation
following proteinase K digestion using the QIAmp Blood Kit
(Qiagen, Hilden, Germany).
Polymerase chain reaction (PCR). Primer sequences and
chromosomal locations of the corresponding microsatellites
are given in Table IB. Primer pairs were synthesized using
an Expedite Model 8009 DNA/RNA synthesizer (Perceptive
Biosystems, Wiesbaden, Germany). Amplifying a GT repeat
inside the seventh intron of the IRF1 gene, we introduced a
new primer pair which was designed according to the
octamer frequency disparity method as described previously
(Griffais et al, 1991). The forward primer was designed from
the IRF1 genomic DNA sequence (exon 8).
Aliquots of amplified DNA were mixed with 5 ml of
formamide dye solution, heated to 968C for 5 min and then
chilled on ice before being subjected to electrophoresis on a
10% polyacrylamide standard denaturing sequencing gel
(National Diagnostics, Atlanta, GA, USA). The gels were
silver-stained using standard procedures.
Genescan analysis. Analysis was performed according to
the Genescan 500 TAMRA protocol (ABI, Weiterstadt,
Germany), using the manufacturer's analysis program
(ABI).
RESULTS
The results of the microsatellite analyses are summarized in
Table IA. Six patients showed LOH (23%) and two patients
displayed MSI (8%). Five of them had advanced stage MDS
and one patient had sAML. Case 7, classified as refractory

q 2000 Blackwell Science Ltd, British Journal of Haematology 109: 842±846


Fig. 1. (A) MSI detected in patient 7 (RAEB). bm, amplified bone
marrow DNA; buc, amplified DNA from corresponding buccal
mucosa. The examined loci and their chromosomal locations are
designated at the bottom of each paired sample. Allelic gain in bone
marrow DNA is demonstrated by arrows. (B) Four patients
displaying LOH at the IRF1 locus. bm, bone marrow DNA; buc,
corresponding DNA from buccal epithelium. In contrast to patients
16, 18, and 20, patient 9 displayed loss of the shorter allele
(arrows). (C) Electropherogram showing amplification products of
the IRF1 primers in patient 20. (a) A normal heterozygous signal
amplified from buccal mucosa; (b) the corresponding bone marrow
sample which exhibits loss of the longer allele. The x-axis shows
fragment sizes in base pairs (bp), the y-axis shows peak height
values. The shorter allele is 248 bp and the longer allele 256 bp
long.
anaemia with excess blasts (RAEB), showed MSI for five
markers (LPL [8p], D10S197 [10qter], D12S89 [12p], RB1
[13q] and DCC [18q]; Fig 1A) and patient 19, who exhibited
chronic myelomonocytic leukaemia (CMML), displayed MSI
for marker D12S89. Four patients showed LOH at locus
IRF1, two at locus WT1 and one patient displayed allelic loss
for TP53. To verify the relatively high prevalence of LOH at
the IRF1 locus (Fig 1B), bone marrow DNA from patient 20
was submitted to a Genescan analysis. By means of this
additional method, loss of the longer allele (256 bp) in the
bone marrow DNA could be confirmed (Fig 1Ca and b).
Correlation of karyotype and microsatellite analysis
Patients with a normal karyotype showed neither LOH nor
q 2000 Blackwell Science Ltd, British Journal of Haematology 109: 842±846
Short Report 845
MSI. In contrast, we found a significantly positive correla-
tion among the occurrence of LOH, MSI and abnormal
karyotype, including sporadic chromosomal alterations
(exact Fisher test; P , 0´025). A correlation between loci
affected by cytogenetic anomalies and loci involved in LOH
could be observed in patients 9 and 20. Both displayed
chromosomal deletions within the critical region 5q13±
5q31 and showed LOH of IRF1 (5q31) in bone marrow DNA
(Table IA).
DISCUSSION
A disease-inherent genetic instability is supposed to be the
basis for an accumulation of somatic mutations in MDS
(Haase et al, 1992; Jacobs, 1992). In the present study,
analysing 18 microsatellite repeats located on 15 chromo-
somal arms, LOH was detected in bone marrow DNA from
six patients (23%), whereas two further patients displayed
MSI (8%). One of the latter, classified as RAEB (patient 7),
showed MSI at 5 of 17 informative loci (29%). It is therefore
tempting to suggest that a malfunctioning DNA MMR
machinery might have contributed to the pathogenesis of
MDS in this case.
Previous reports on colorectal cancer determined a
negative correlation between the occurrence of MSI and
other genetic events (Ionov et al, 1993) and suggested MSI
of marker BAT-26 to be a reliable indicator of a replication
error phenotype (Hoang et al, 1997). In our study, MSI and
LOH were associated with clonal karyotypic abnormalities
or sporadic chromosomal alterations, respectively, and none
of our patients presented with alterations of marker BAT-26.
Although DNA MMR deficiency might be involved in a
subset of patients with MDS, different pathogenetic path-
ways seem to be responsible for genetic instability in
colorectal cancer. This might also explain the considerably
higher prevalence of unstable markers in HNPCC (, 40%)
in comparison with MDS.
LOH was more frequently detected than MSI, mainly
involving loci IRF1 and WT1. After introducing a new primer
pair amplifying a GT repeat inside the IRF1 gene, LOH was
detected in four patients. Two of the patients displayed a
corresponding loss of chromosome band 5q31, representing
the minimal commonly deleted region (Le Beau et al, 1993).
Our results underline the possible significance of IRF1 gene
loss in MDS, perhaps as one initial step during pathogenesis.
They might also reflect a disease-inherent genetic instability
in the 5q31 region, where IRF1 might not be directly involved
but is located in the vicinity of another presumed tumour-
suppressor gene. Using a primer pair amplifying a CA repeat in
the 3 0 untranslated region of the WT1 gene, LOH was detected
in two patients, one with therapy-related refractory anaemia
with excess blasts (t-RAEB), which transformed into AML
later on, and one with refractory anaemia with excess blasts
in transformation (RAEB-T). Mutations in the WT1 gene have
been reported to occur in 15% of cases of AML and were
associated with a poor response to chemotherapy (King-
Underwood et al, 1996). Similarly, allelic loss of WT1 might be
predictive for a poor clinical prognosis in MDS. In view of the
predominant localization of LOH at only two loci (IRF1 and
846 Short Report
WT1), we suggest that the allelic losses detected in our
patients were initial rather than later genetic events.
Only one of our patients displayed allelic loss at the TP53
locus (case 9). Similarly, it has been shown that p53
mutations are extremely rare in de novo MDS (Sugimoto et al,
1993). We therefore conclude that both karyotype and
molecular genetic instability in de novo MDS are not
triggered by cell cycle dysregulation due to loss of the p53
gene function. This is an intriguing fact as the opposite has
been described in therapy-related leukaemia (Ben-Yehuda
et al, 1996).
Our data showed a significantly positive correlation
between the occurrence of LOH and MSI and an abnormal
karyotype, including sporadic chromosomal alterations
(P , 0´025). These data are in line with the mutator
phenotype hypothesis, suggesting that LOH and MSI might
reflect early mutative events which affect mutator genes and
thus induce genetic changes to a larger extent, i.e.
chromosome instability (Loeb, 1998). However, no correla-
tion was found between chromosomal breakpoints defined
by karyotype analysis and our patients displaying LOH and
MSI. These data reflect what might be expected when
regarding LOH and MSI as a result of DNA MMR failure
leading to a general increase in mutation rate. In only two
patients displaying LOH, both harbouring a partial deletion
of chromosome 5, a corresponding defect was found by
cytogenetic studies and microsatellite analysis. It is probable
that the allelic losses in the remaining four cases were on
the submicroscopic level.
Two patients with an alteration of the short arm of
chromosome 17 displayed no corresponding allelic loss of
marker TP53 detected by PCR (patients 4 and 12). Likewise,
four patients with a monosomy 7 detected by conventional
cytogenetics displayed no analogous LOH for marker
D7S522 by PCR (patients 4, 6, 12 and 20). These
discrepancies may originate from the high percentage of
cytogenetically normal cells which in patient 12 amounted
to 63%, in patient 4 to 56% and in patient 20 to 21%. The
frequency of LOH and MSI might therefore be under-
estimated in those cases.
In summary, LOH was more frequent than MSI (23% vs.
8%). Although a defective DNA MMR might be affected in a
small subset of patients with MDS, non-involvement of
marker BAT-26 and the association of further genetic events
point to a genetic pathway that is different from the
situation in HNPCC. The fact that LOH and MSI exclusively
occurred in those MDS patients harbouring cytogenetic
aberrations, including sporadic chromosomal alterations as
well as the predominant localization of LOH at only two loci,
may support the mutator phenotype hypothesis. Thus,
given that LOH and MSI occur at the onset of MDS, they
might represent an early predictor for chromosomal
instability. However, as in only two cases a precise
correlation between karyotypic lesions and submicroscopic
events such as LOH could be detected, LOH seems not to be
predictive for specific chromosomal lesions. With respect to
our results regarding IRF1 and WT1, the critical regions
affected should be subject to further investigations. More-
over, a microsatellite panel specific for MDS should be
q 2000 Blackwell Science Ltd, British Journal of Haematology 109: 842±846
evaluated systematically in order to uncover discrete genetic
lesions which would escape conventional cytogenetics. This
could be particularly useful in differentiating those MDS on
a molecular level that otherwise appear homogeneous by
conventional karyotyping. In these cases, it would be
interesting to see whether patients with MSI or LOH differ
in their prognosis from those with normal microsatellite
allelotypes.
ACKNOWLEDGMENTS
We thank Dr O. Knobloch from SEQLAB Sequence
Laboratories (GoÈttingen, Germany) for performing Genescan
analysis and V. Jassal for helpful comments regarding the
manuscript.
REFERENCES
Ben-Yehuda, D., Krichevsky, S., Caspi, O., Rund, D., Polliack, A.,
Abeliovich, D., Zelig, O., Yahalom, V., Paltiel, O., Or, R., Peretz, T.,
Ben-Neriah, S., Yehuda, O. & Rachmilewitz, E.A. (1996)
Microsatellite instability and p53 mutations in therapy-related
leukemia suggest mutator phenotype. Blood, 88, 4296±4303.
Griffais, R., AndreÂ, P.M. & Thibon, M. (1991) K-tuple frequency in
the human genome and polymerase chain reaction. Nucleic Acids
Research, 19, 3887±3891.
Haase, D., Fonatsch, C. & Freund, M. (1992) Karyotype instability in
myelodysplastic syndromes ± a specific step in pathogenesis
preceding clonal chromosome anomalies. Leukemia and Lym-
phoma, 8, 221±228.
Haase, D., Feuring-Buske, M., SchaÈfer, C., Schoch, C., Troff, C.,
Gahn, B., Hiddemann, W. & WoÈrmann, B. (1997) Cytogenetic
analysis of CD341 subpopulations in AML and MDS characterized
by the expression of CD38 and CD117. Leukemia, 11, 674±679.
Hoang, J.M., Cottu, P.H., Thuille, B., Salmon, R.J., Thomas, G. &
Hamelin, R. (1997) BAT-26, an indicator of the replication error
phenotype in colorectal cancers and cell lines. CancerResearch, 57,
300±303.
Ionov, Y., Peinado, M.A., Malkhosyan, S., Shibata, D. & Perucho, M.
(1993) Ubiquitous somatic mutations in simple repeated
sequences reveal a new mechanism for colonic carcinogenesis.
Nature, 363, 558±561.
Jacobs, A. (1992) Pathogenesis and clinical variations in the
myelodysplastic syndromes. Clinical Haematology, 15, 925±951.
King-Underwood, L., Renshaw, J. & Pritchard-Jones, K. (1996)
Mutations in the Wilms' tumor gene WT1 in leukemias. Blood,
87, 2171±2179.
Le Beau, M.M., Espinosa, R., Neuman, W.L., Stock, W., Roulston, D.,
Larson, R.A., Keinanen, M. & Westbrook, C.A. (1993) Cytoge-
netic and molecular delineation of the smallest commonly deleted
region of chromosome 5 in malignant myeloid diseases. Proceed-
ings of the National Academy of Sciences of the United States of
America, 90, 5484±5488.
Loeb, L.A. (1998) Cancer cells exhibit a mutator phenotype.
Advances in Cancer Research, 72, 25±56.
Stollmann, B., Fonatsch, C. & Havers, W. (1985) Persistent Epstein-
Barr virus infection associated with monosomy 7 or chromosome
3 abnormality in childhood myeloproliferative disorders. British
Journal of Haematology, 60, 183±196.
Sugimoto, K., Hirano, N., Toyoshima, H., Chiba, S.H., Mano, H.,
Takaku, F., Yazaki, Y. & &. Hirai, H. (1993) Mutations in p53
gene in myelodysplastic syndrome (MDS) and MDS-derived
leukemia. Blood, 81, 3022±3026.