Download Genomic gains and losses influence expression levels of genes

Leukemia (2005) 19, 1224–1228
& 2005 Nature Publishing Group All rights reserved 0887-6924/05 $30.00
www.nature.com/leu
Genomic gains and losses influence expression levels of genes located within
the affected regions: a study on acute myeloid leukemias with trisomy 8, 11,
or 13, monosomy 7, or deletion 5q
C Schoch1, A Kohlmann1, M Dugas1, W Kern1, W Hiddemann1, S Schnittger1 and T Haferlach1
1
Laboratory for Leukemia Diagnostics, Department of Internal Medicine III, University Hospital Grosshadern,
Ludwig-Maximilians-University, Munich, Germany
We performed microarray analyses in AML with trisomies 8
(n ¼ 12), 11 (n ¼ 7), 13 (n ¼ 7), monosomy 7 (n ¼ 9), and deletion
5q (n ¼ 7) as sole changes to investigate whether genomic
gains and losses translate into altered expression levels of
genes located in the affected chromosomal regions. Controls
were 104 AML with normal karyotype. In subgroups with
trisomy, the median expression of genes located on gained
chromosomes was higher, while in AML with monosomy 7 and
deletion 5q the median expression of genes located in deleted
regions was lower. The 50 most differentially expressed genes,
as compared to all other subtypes, were equally distributed
over the genome in AML subgroups with trisomies. In contrast,
30 and 86% of the most differentially expressed genes
characteristic for AML with 5q deletion and monosomy 7 are
located on chromosomes 5 or 7. In conclusion, gain of whole
chromosomes leads to overexpression of genes located on the
respective chromosomes. Losses of larger regions of the
genome translate into lower expression of the majority of
genes represented by only one allele. The reduced expression
of these genes is the most characteristic difference in gene
expression profiles between AML with monosomy 7 and AML
with deletion 5q, respectively, and other AML subtypes. Therefore, these data provide evidence that gene dosage effects gene
expression in AML with unbalanced karyotype abnormalities.
Losses of specific regions of the genome determine the gene
expression profile more strongly than the gain of whole
chromosomes.
Leukemia (2005) 19, 1224–1228. doi:10.1038/sj.leu.2403810
Published online 19 May 2005
Keywords: acute myeloid leukemia; gene expression; gene dosage
Introduction
Acute myeloid leukemia (AML) is a heterogeneous group of
diseases. From a genetic point of view, three subgroups can be
distinguished: (1) AML with normal karyotype, (2) AML with a
primary balanced chromosome aberration, and (3) AML with
unbalanced karyotype abnormalities characterized by gains
and/or losses of usually larger regions of the genome and no
known primary balanced abnormality.1 The important pathogenetic role of leukemia-specific fusion transcripts such as PMLRARA, AML1-ETO, and CBFB-MYH11 has been proven. The
role of genomic gains and losses in AML is less clear. It has been
assumed that gene dosage effects may play a role in the
pathogenesis of this AML subgroup. Virtaneva et al supported
this hypothesis showing that AML with trisomy 8 as the sole
karyotype abnormality overexpressed genes located on chromosome 8 compared to AML with normal karyotype.2 It was the
aim of this study to investigate whether gains and losses on the
genomic level translate into altered genes expression also in
other areas of the genome in AML.
Materials and methods
Samples
Bone marrow samples of AML patients at diagnosis were
analyzed: 12 cases with trisomy 8 (AML-TRI8), seven with
trisomy 11 (AML-TRI11), seven with trisomy 13 (AML-TRI13),
and nine cases with monosomy 7 (AML-MO7), and seven cases
with a deletion of the long arm of chromosome 5 (5q) (AMLDEL5q), as well as 104 AML with normal karyotype (AML-NK).
All karyotype abnormalities were the sole aberrations. All cases
were analyzed by cytomorphology, chromosome-banding
analysis and fluorescence in situ hybridization (FISH) on
interphase nuclei using loci-specific probes for EGR1 (5q31),
RB (13q14), the centromeric regions of chromosomes 7, 8, and
11.3 The proportion of cells carrying the genetic abnormality
was above 60% in all cases as determined by FISH analyses. The
studies abide by the rules of the local Internal Review Board and
the tenets of the revised Helsinki protocol.
Gene expression and statistical analysis
We performed gene expression analysis using oligonucleotide
microarrays covering 33 000 transcripts (Affymetrix U133 set)
according to the protocol of the manufacturer applying quality
controls as described elsewhere.4,5 For further analysis, only
probe sets with a known chromosomal localization (www.affy
metrix.com/analysis/) and those expressed at least in one sample
were used. All probe sets used in this study together with the
gene they represent and their chromosomal localization are
shown in Tables S1a–S1e of the Supplementary Information. For
each probe set, the mean expression value was calculated for
each AML subgroup. Then for each probe set, a ratio ‘mean
expression level AML unbalanced/mean expression level AML
normal’ was determined. In addition, differentially expressed
genes were identified by means of the two-sample t-statistic
score with correction for unequal variances.6 The analysis was
programmed using R (version 1.7.1; www.r-project.org/).
Results and discussion
Correspondence: Dr C Schoch, Laboratory for Leukemia Diagnostics,
Department of Internal Medicine III, University Hospital Grosshadern,
Ludwig-Maximilians-University of Munich, Marchioninistr. 15,
Munich 81377, Germany; Fax: þ 49 89 7095 4971;
E-mail: [email protected]
Received 30 November 2004; accepted 12 April 2005; Published
online 19 May 2005
Gene copy number abnormalities are characteristic for malignant diseases and are also found in association with developmental abnormalities and/or mental retardation. It is reasonable
to propose that the transcription of genes mapping to
chromosomal regions gained or deleted is altered due to a
Gene dosage effects in AML
C Schoch et al
1225
dosage effect. Differences in the transcriptional regulation of
genes may explain why a change in copy number does not
always result in a higher or lower expression of the affected
gene. It was the aim of this study to determine the effect of gene
dosage in AML with trisomies 8, 11, 13, monosomy 7, or
deletion 5q as a sole abnormality.
However, DEFA1 was not confirmed (ratio: 0.74). In our series
the 3 genes most significantly upregulated in AML-TRI8 were:
PTK2B, CNOT7, and LYPLA1. PTK2B was identified also by
Virtaneva et al as upregulated in AML-TRI8 in a comparison of
AML-TRI8 vs AML-NK vs CD34 þ normal stem cells, while the
other two genes were not mentioned.
Trisomy 11 and 13
Trisomy 8
A total of 1052 probe sets cover sequences localized on
chromosome 8 (Table S1a, supplemental material). In all, 835 of
them were expressed in at least one sample and were thus used
for further analyses. The mean expression level of 607 of these
835 (72.6%) probe sets was higher in cases with trisomy 8 as
compared to AML with normal karyotype (Figure 1a). The
median expression level of probe sets on chromosome 8 was
1.27-fold higher in AML-TRI8 cases than in AML-NK (Figure 2a
and b) confirming a gene dosage effect as described by
Virtaneva et al.2 In that series, seven out of 29 genes most
significantly upregulated in AML-TRI8 for which a chromosomal
assignment could be made, map to chromosome 8: FABP5,
PTDSS1, LY6E, COX6C, LYN, SIAHBP1, and DEFA1. In
contrast, none of the 25 genes upregulated in AML-NK map to
chromosome 8. The following genes identified by Virtaneva et al
as up-regulated in AML-TRI8 were confirmed in our series:
FABP5 (ratio: 1.73), PTDSS1 (ratio: 1.65), LY6E (ratio: 2.0),
COX6C (ratio: 1.34), LYN (ratio: 1.4), and SIAHBP1 (ratio: 1.37).
a
In all, 1263 of 1693 and 440 of 530 probe sets covering
sequences localized on chromosomes 11 and 13, respectively,
were expressed in at least one sample (Table S1b and S1c,
Supplementary Information). The mean expression of 828 out of
1263 (65.6%) probe sets was higher in cases with trisomy 11,
and of 275 out of 440 (62.5%) higher in cases with trisomy 13 as
compared to AML with normal karyotype, respectively
(Figure 1b). The median expression of probe sets on chromosome 11 was 1.25-fold higher in AML-TRI11 cases than in AMLNK and the expression of probe sets on chromosome 13 was
1.14-fold higher in AML-TRI13 cases than in AML-NK,
demonstrating a gene dosage effect also in AML with trisomy
11 and 13 (Figure 2b).
Monosomy 7
A total of 1441 probe sets cover sequences localized on
chromosome 7 (Table S1d, Supplementary Information). In all,
b
Trisomy 8
Trisomy 11
50%
100
10
ratio
10
ratio
50%
100
1
1
0.1
66%
73%
0.1
0.01
0.001
0.01
1263 probe sets on chr. 11
835 probe sets on chr. 8
c
d
Monosomy 7
100
10
10
1
1
ratio
ratio
100
50%
0.1
87%
0.01
0.1
Deletion 5q
50%
67%
0.01
0.001
0.001
1088 probe sets on chr. 7
502 probe sets in 5q13q31
Figure 1
Ratios of mean expression of probe sets affected by change of gene dosage between the respective subtype with gain or loss of one
chromosome/chromosomal region in comparison to AML with normal karyotype. (a) In all, 73% of the 835 probe sets located on chromosome 8
show a higher mean expression (ratio41) in cases with trisomy 8 than in AML with normal karyotype. (b) In all, 65% of the 1263 probe sets located
on chromosome 11 show a higher mean expression (ratio41) in cases with trisomy 11 than in AML with normal karyotype. (c) In all, 87% of the
1088 probe sets located on chromosome 7 show a lower mean expression (ratioo1) in cases with monosomy 7 than in AML with normal
karyotype. (d) In all, 67% of the 502 probe sets located on the long arm of chromosome 5 within the chromosomal bands 5q13–5q31 show a lower
mean expression (ratioo1) in cases with deletion 5q than in AML with normal karyotype.
Leukemia
Gene dosage effects in AML
C Schoch et al
1226
a
1.8
1.6
1.4
Ratio
1.2
1.0
0.8
0.6
0.4
0.2
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22
Chromosome
b
1.8
1.6
1.4
Ratio
1.2
1
0.8
0.6
0.4
0.2
Chr. 8
AML-TRI8
vs. AML-NK
Chr. 11
AML-TRI11
vs. AML-NK
Chr. 13
AML-TRI13
vs. AML-NK
Chr. 7
AML-MO7
vs. AML-NK
Chr. 5q13q31
AML-DEL5q
vs. AML-NK
Figure 2
Gene dosage influences levels of gene expression. (a) AML samples with trisomy 8 (n ¼ 12) exhibit higher expression for genes
located on chromosome 8 than AML samples with normal karyotype (n ¼ 104). The median expression levels and the second and third quartile for
probe sets by chromosome in AML with þ 8 relative to AML with normal karyotype are shown. (b) The median expression levels and 25 and 75%
interval for probe sets on chromosomes 8, 11, 13, 7, and on chromosomal bands 5q13–5q31 in AML with the respective gain or loss of the
chromosome/chromosomal region relative to AML with normal karyotype are shown.
1088 were expressed in at least one sample. The mean
expression of 948 out of 1088 (87.1%) probe sets was lower
in cases with monosomy 7 than in AML with normal karyotype
(Figure 1c). The median expression of probe sets on chromosome 7 was 0.57-fold in AML-MO7 cases compared to AMLNK, demonstrating a strong effect of gene dosage on gene
expression also in AML with monosomy 7 (Figure 2b). In Figure
3a, the median expression ratios are separately shown for genes
located on 7p and the chromosomal bands 7q11, 7q21, 7q22,
7q31, 7q32, 7q33, 7q34, 7q35, and 7q36. The expression ratios
are clearly below 1 in all chromosomal regions ranging from
0.48 to 0.63.
5q31 was 0.82-fold in AML-DEL5q cases compared to AML-NK
demonstrating a gene dosage effect also in AML with 5q deletion
(Figure 2b). In Figure 3b, the median expression ratios are
separately shown for genes located on 5p and the chromosomal
bands 5q11, 5q12, 5q13, 5q14, 5q15, 5q21, 5q22, 5q23, 5q31,
5q32, 5q33, 5q34, and 5q35. The expression ratios are only
clearly below 1 for those chromosomal bands located within the
most commonly deleted region (5q14–5q33), showing the
lowest ratio for genes located within the chromosomal band
5q15 (median ratio 0.61). Genes located on the short arm of
chromosome 5 or in chromosomal bands 5q11, 5q12, and
5q13, showed on average a slightly higher expression ratio in
cases with 5q deletion than in AML with normal karyotype.
5q deletion
Differences in the global gene expression profile
In all, 1374 probe sets cover sequences localized on chromosome 5 (Table S1e, Supplementary Information), 644 of these
are located in chromosome bands 5q13–5q31, the region most
commonly deleted in AML with 5q deletion. In all, 502 of them
were expressed in at least one sample. The mean expression of
337 out of 502 (67.1%) probe sets was lower in cases with 5q
deletion than in AML with normal karyotype (Figure 1d). The
median expression of probe sets located in the region 5q13–
Leukemia
In order to evaluate the effect of gene dosage on the subtypespecific gene expression signature, the top 50 differentially
expressed probe sets for each subtype vs all other subtypes were
determined using ANOVA and t-test statistics (Welch t-test). The
50 most differentially expressed probe sets were equally
distributed over the genome for each of the comparisons AMLTRI8, AML-TRI11, AML-TRI13 vs all other subtypes. However,
Gene dosage effects in AML
C Schoch et al
1227
a
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
7p
b
7q11
7q21
7q22
7q31
7q32
7q33
7q34
7q35
7q36
1.4
1.2
1
0.8
0.6
0.4
0.2
0
5p
5q11 5q12 5q13 5q14 5q15 5q21 5q22 5q23 5q31 5q32 5q33 5q34 5q35
Figure 3
(a) Median expression levels for probe sets located on chromosome 7 separately shown for 7p and the chromosomal bands 7q11,
7q21, 7q22, 7q31, 7q32, 7q33, 7q34, 7q35, and 7q36. Ratios have been calculated between cases with monosomy 7 and AML with normal
karyotype. (b) Median expression levels for probe sets located on chromosome 5 separately shown for 5p and the chromosomal bands 5q11, 5q12,
5q13, 5q14, 5q15, 5q21, 5q22, 5q23, 5q31, 5q32, 5q33, 5q34, and 5q35. Ratios have been calculated between cases with deletion 5q and AML
with normal karyotype.
comparing AML-DEL-5q with all other subtypes revealed that 10
of the 34 probe sets for which chromosomal location was
available are located on chromosome 5 within the region
affected by the deletion. These probe sets represent eight genes
involved in signal transduction (HINT1, PDE8B, SNX2,
CSNK1A1, ANXA6), radioadaptive response (HSPA4), and
suppression of invasion (CTNNA1), respectively. Recently,
CTNNA1 was reported to show a lower expression in
leukemia-initiating cells of cases with del(5q) MDS/AML than
in normal hematopoietic stem cells and leukemia-initiating cells
of AML/MDS without 5q deletion (Liu et al. Blood 2004; 104:
62a; abstract). In Figure 4, the CTNNA1 expression levels in
the different AML subgroups of our cohort are shown demonstrating the significantly lower CTNNA1 expression in cases
with 5q deletion.
For 43 of the top 50 probe sets differentially expressed
between AML-MO7 and all other subtypes, the chromosomal
location is known. Of these 43 probe sets, 39 representing 36
different genes are localized on chromosome 7. They are
involved in mismatch repair (PMS2L1, PMS2L3, PMS2L5,
PMS2L8, PMS2L9), apoptosis (TAX1BP1, CASP2, CARD4),
DNA replication (RIP60, SSBP1), and signal transduction
(AKAP9, CARD4). Also HOXA3 and HOXA9 were significantly
lower expressed in AML-MO7 than in all other subtypes.
In inborn diseases as well as in several malignant diseases, a
pathogenetic role of haploinsufficiency has been demonstrated.7,8 Especially for several tumor suppressor genes,
including TP53, data suggest that some genes, when expressed
at half normal levels, that is from only one functional allele,
cannot fully suppress tumor growth.9,10 Therefore, genes located
in deleted regions showing a reduced expression may be new
candidate genes playing an important role in the pathogenesis of
malignancies with unbalanced karyotypes. The concept of
halpoinsufficiency, especially in AML with 5q deletion is
Leukemia
Gene dosage effects in AML
C Schoch et al
1228
1400
1200
1000
800
600
400
200
0
AML del(5q)
Figure 4
AML -7
AML +8
AML +11
AML normal
The median expression levels and the 25 and 75% interval of CTNNA1 for the different AML subgroups are shown.
strengthened also by the fact that despite a lot of effort no tumor
suppressor gene was identified within the deleted region, being
critical for the pathogenesis of this AML subtype. Therefore, the
deletion of large regions of the long arm of chromosome 5 may
not initiate tumor formation in itself but could foster transformation by rendering the cell genetically unstable and therefore
more likely to sustain other mutations and chromosomal
rearrangements as is frequently observed, in addition to 5q
deletions in cases with complex aberrant karyotypes.1
In conclusion, gain of whole chromosomes leads to overexpression of genes located on the respective chromosomes.
However, the altered expression of genes located on the gained
chromosome does not dominate the specific gene expression
signature of the specific subtype. These data suggest that trisomy
8, 11, or 13 may not be the primary disease determining
aberration but rather secondary events associated with progression. This is supported by the fact that especially trisomy 8
occurs in addition to a large variety of primary aberrations in AML
such as t(8;21)(q22;q22), inv(16)(p13q22), t(15;17)(q22;q12),
11q23/MLL rearrangements.11,12
In contrast, losses of larger regions of the genome translate not
only into lower expression of the majority of genes represented
by only one allele, but also is the most characteristic difference
in gene expression between AML with monosomy 7 or AML
with 5q deletion and other AML subtypes. Therefore, these data
provide evidence that gene dosage effects play an important role
in AML with unbalanced karyotype abnormalities, leading to
losses of genetic material. Genes located in the lost regions
showing also a reduced expression are candidate genes and
their role in pathogenesis of AML needs to be clarified.
2
3
4
5
6
7
8
9
10
11
Supplementary Information
Supplementary Information accompanies the paper on the
Leukemia website (http://www.nature.com/leu).
References
1 Schoch C, Haferlach T, Bursch S, Gerstner D, Schnittger S, Dugas
M et al. Loss of genetic material is more common than gain in
Leukemia
AML +13
12
acute myeloid leukemia with complex aberrant karyotype: a
detailed analysis of 125 cases using conventional chromosome
analysis and fluorescence in situ hybridization including 24-color
FISH. Genes, Chromosomes Cancer 2002; 35: 20–29.
Virtaneva K, Wright FA, Tanner SM, Yuan B, Lemon WJ, Caligiuri
MA et al. Expression profiling reveals fundamental biological
differences in acute myeloid leukemia with isolated trisomy 8 and
normal cytogenetics. Proc Natl Acad Sci USA 2001; 98:
1124–1129.
Schoch C, Schnittger S, Bursch S, Gerstner D, Hochhaus A, Berger
U et al. Comparison of chromosome banding analysis, interphaseand hypermetaphase-FISH, qualitative and quantitative PCR for
diagnosis and for follow-up in chronic myeloid leukemia: a study
on 350 cases. Leukemia 2002; 16: 53–59.
Schoch C, Kohlmann A, Schnittger S, Brors B, Dugas M,
Mergenthaler S et al. Acute myeloid leukemias with reciprocal
rearrangements can be distinguished by specific gene expression
profiles. Proc Natl Acad Sci USA 2002; 99: 10008–10013.
Kohlmann A, Schoch C, Schnittger S, Dugas M, Hiddemann W,
Kern W et al. Pediatric acute lymphoblastic leukemia (ALL) gene
expression signatures classify an independent cohort of adult ALL
patients. Leukemia 2004; 18: 63–71.
Altmann DG. Practical statistics for medical research. London:
CRC Press, 2004.
Seidman JG, Seidman C. Transcription factor haploinsufficiency:
when half a loaf is not enough. J Clin Invest 2002; 109: 451–455.
Santarosa M, Ashworth A. Haploinsufficiency for tumour suppressor genes: when you don’t need to go all the way. Biochim Biophys
Acta 2004; 1654: 105–122.
Largaespada DA. Haploinsufficiency for tumor suppression: the
hazards of being single and living a long time. J Exp Med 2001;
193: F15–F18.
Venkatachalam S, Shi Y-P, Jones SN, Vogel H, Bradley A, Pinkel D
et al. Retentio of wild-type p53 in tumore from p53 heterozygous
mice: reduction of p53 dosage can promote cancer formation.
EMBO J 1998; 17: 4657–4667.
Schoch C, Haase D, Fonatsch C, Haferlach T, Löffler H,
Schlegelberger B et al. The significance of trisomy 8 in de
novo acute myeloid leukaemia: the accompanying chromosome aberrations determine the prognosis. Br J Haematol 1997;
99: 605–611.
Schoch C, Schnittger S, Klaus M, Kern W, Hiddemann W,
Haferlach T. AML with 11q23/MLL abnormalities as defined by
the WHO classification: incidence, partner chromosomes, FAB
subtype, age distribution, and prognostic impact in an unselected
series of 1897 cytogenetically analyzed AML cases. Blood 2003;
102: 2395–2402.