a study on acute myeloid leuk - Nature

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


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


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


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,


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


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, Lo¨ffler 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.