Letters to the Editor
175
Implementation of array based whole-genome high-resolution technologies confirms the absence of secondary copy-number alterations in MLL-AF4-positive infant ALL patients Leukemia (2011) 25, 175–178; doi:10.1038/leu.2010.232; published online 14 October 2010
The current opinion in cancer research is that human tumors are the results of the progressive accumulation of subsequent multiple genetic lesions.1,2 In many leukemias, the first transforming event has been demonstrated to occur prenatally in utero,1,3 followed, several years later, by the subsequent postnatal acquisition of additional genetic mutations, leading to the overt leukemia development.1 In support to this hypothesis, several independent groups have identified the existence of ‘genetic hot spots’ in the human genome, regardless of the age of onset and the cytogenetic subgroups to which they belong, targeting non-random key genes with a crucial role in leukemogenesis.4,5 However, for some subtypes of leukemia, such as Miked lineage leukemia (MLL) gene rearranged infant acute lymphoblastic leukemia (ALL), the pathogenetic mechanism remains still unclear. By applying the Affymetrix Human Mapping single-nucleotide polymorphism
Table 1
List of tumor-associated genomic aberrations in MLL-AF4 positive infant ALL patients
Pat.ID Region 1
(SNP; Affymetrix, Santa Clara, CA, USA) 100K or 500K Array technology, we and others have recently reported an exceptionally limited number of copy-number alterations in high-risk infant ALL with MLL gene rearrangement, compared with other form of leukemia occurring in older children subsets.4,6 This suggested that MLL rearrangement per se, as a first and sole hit, has a major role in driving and hastening the leukemogenic process in infant ALL, without the need of secondary cooperating mutations. It can be argued that, as a major limitation of this study, small lesions under the detection limit of the method applied might have potentially been missed. Nowadays, the recent technology advancements give us the possibility to drastically increase the sensitivity, in a way that the re-evaluation of the same cases and the implementation of the data previously obtained with different, less sensitive platforms might be very helpful in reinforcing the recently published results. The new released Affymetrix Cytogenetics Whole-Genome 2.7 M Array (Affymetrix) provides, at present, the greatest power to detect known and novel aberrations across the entire genome, as the best available proxy to the high
Size Marker Confidence Genesa (Mb) count (%)
Type
Start
End
(2)(q11.2)
LOH
96.701.218
98.224.741
1.5
96
99,89
(16)(q21) (17)(p13.3)
LOH 60.198.487 LOH 667.764
61.736.238 2.301.157
1.5 1.6
315 98
100 99,99
(X)(p21.1)
CNNM3, CNNM4, COX5B, ZAP70 CDH8 OVCA2, MNT mir-212, HBII-296B, mir-22, mir-132, HBII-296A No genes
Loss
36.653.789
36.704.848
0.5
38
89,38
4 10
(17)(q21.2q21.31) LOH
37.719.964
39.224.666
1.5
98
97,23
11
(7)(q34q36.3)
LOH 141.976.845 158.811.205 16.8
1590
100
12
(X)(p11.22p11.23) LOH
49.659.688
51.361.250
1.7
149
100
BRCA1, MLX, STAT3, EZH1, ETV4 CDK5, ABCB8, WDR60, CUL1, MLL3 CCNB3, BMP15
(X)(q21.1q21.2)
LOH
82.752.745
84.952.581
2.2
309
100
CYLC1, HDX
(4)(p16.1) LOH (9)(q22.32q22.33) LOH (17)(q23.1q23.2) LOH
8.640.336 97.868.331 55.400.858
10.297.183 99.371.560 56.942.635
1.7 1.5 1.5
131 103 135
98,38 99,99 99,94
(1)(p35.1p35.2)
LOH
31.480.444
33.677.272
2.2
113
100
(2)(q35) (4)(q13.2)
LOH 217.945.526 219.531.742 LOH 68.797.839 70.333.311
1.6 1.5
128 84
100 100
(6)(p11.1p12.1)
LOH
1.7
98
100
16 18
21
57.091.948
58.835.468
sno/micro RNAa
WDR1, CLNK CDC14B TBX2, TBX4, BCAS3 LCK, HDAC1, BAI2, EIF3I, WNT6 TMPRSS11 family, UGT2B family BAG2, RAB23
SNP 100K or 250Kb No No No No No
mir-548f-4
Yes
mir-501, mir-660, mir-500, mir-502, mir-362 mir-548i-4
No No
mir-548i-2 mir-1302-8 ACA66
Yes No No No
mir-26b
No No Yes
23 25 Abbreviations: LOH, loss of heterozygosity; SNP, single-nucleotide polymorphism; sno, small nucleolar RNA. a Genes and sno/micro RNA mapping within each regions are reported. b Last column indicates whether the genomic aberrations were also detected by Affymetrix Human mapping SNP 100K or 250K array platforms from our previous study.6 The heterozygous deletion of chromosome Xp21.1 in patient 1, the only copy-number alteration detected in the cohort of investigated patients, is indicated in italic. Leukemia
Letters to the Editor
176
Leukemia
Letters to the Editor
177 Figure 1 Heatmap of copy-number analysis. A total of 10 MLL-AF4-positive infant ALL cases (patients IDs correspond to our previously published cases6) were compared with 12 childhood ALL patients older than 1 year, including two TEL-AML1-positive cases (a, b), one common ALL patient negative for the translocations commonly associated to leukemia (c), and nine high-risk Ph þ ALL carrying the BRC-ABL fusion (d–l). Chromosome (Chr) 7 (a), 9 (b) and 12 (c) are shown as the most representative, and the genes more commonly altered in human leukemias are indicated. The weighted log2 copy-number ratio is displayed, ranging from –2 to 2. Gray bars: diploid DNA content (log2 ratio ¼ 0, copy numbe ¼ 2). Red bars: copy-number loss (deletion). Blue bars: copy-number gain (amplification). The colour reproduction of this figure is available on the html full text version of the manuscript.
throughput whole genome sequencing. With unbiased, wholegenome coverage and the highest density of 2.7 million markers, including not only 400 000 SNPs, but also 2.3 millions of nonpolymorphic copy-number markers, this array provides dense coverage of the annotated genes, cancer genes, micro RNA regions and haploinsufficiency genes, which are of particular relevance to cytogenetists, enabling to detect smaller submicroscopic aberrations as well as copy-number neutral loss of heterozygosity (LOH). As a major advantage, besides providing an high coverage of human genome, the array offers a better approach and a superior resolution, with a median distance between markers of 735 bp (compared with 2.5 Kb for the 500K arrays), that can be narrowed to 294 bp within the cancer genes loci, thus enabling the detection of very small structural changes. We had the opportunity to apply for the first time the Affymetrix Cytogenetics Whole-Genome 2.7 M Array (Affymetrix) technology for a more sensitive analysis of MLL rearranged infant ALL cases, with the aim to validate our previously published results.6 By applying this new platform, we have reanalysed 10 diagnostic-remission paired MLL-AF4 positive infant ALL samples from our 28 published cases 6 (patient 1, 4, 10, 11, 12, 16, 18, 21, 23 and 25) which were negative for secondary chromosomal abnormalities by SNP 100K or 250K. Briefly, for each investigated sample, 100 ng of genomic DNA was amplified by whole-genome amplification reaction, purified by magnetic beads and fragmented to generate small (o300 bp) products, which were subsequently labeled and loaded into a single array. After hybridization, the chip was washed, stained and scanned. Raw data were analyzed with Affymetrix Chromosome Analysis Suite 1.0.1 software (Affymetrix). We found 15 tumor-associated genomic aberrations, including 14 segmental LOH regions (11 of which were not previously identified with less sensitive SNP 100K or 250K platforms), and only one small deletion of chromosome Xp21.1 in patient 1 containing no genes (Table 1). Our analysis also confirmed the presence of numerous constitutional LOH (detectable at both diagnosis and remission, therefore not tumor associated), as the LOH(14q21) containing the FANCM gene in patient 4 and 21 (and additionally in patient 8, for whom only the diagnostic sample was tested), which was the most recurrent lesion observed by SNP 100K analysis being detected in 4 out of 28 patients.6 Furthermore, by comparing MLL-AF4 positive infant ALL samples with a cohort of 12 patients affected by childhood ALL who were older than 1 year at the time of disease presentation, including two TEL-AML1 positive cases, one common ALL patient negative for the chromosomal translocations commonly associated to leukemia, and nine high-risk Ph þ ALL carrying the BCR-ABL fusion, we observed that the copy-number alterations typically found in older children were absent in MLL-AF4-positive infant ALL patients (Figure 1). Therefore, through the re-evaluation and the implementation of our previous data, we have confirmed that infant MLL-rearranged cases lack secondary chromosomal abnormalities, at the best of the knowledge available. Next generation high-throughput whole-genome sequencing will allow to
evaluate whether point mutations or submicroscopic genomic lesions might be associated to MLL leukemogenesis. The significance of tumor associated and constitutional LOH, and whether they could have a role in the pathogenesis or predisposition to leukemia still remain open questions. Our results, together with the evidences of a prenatal origin of the rearrangement,3 the exceptionally high concordance rate in monochorionic twins, and the frequent association with short latency infant disease in patients,1 suggest that, unlike other form of leukemia occurring in children, adolescents and adults, infant ALL is a single-hit leukemia, as MLL fusions genes is sufficient, per se, for the overt disease. This implies that the pathogenetic mechanism of MLL-positive infant ALL leukemia may rather be different from that of other leukemias, pointing to the uniqueness of this aggressive disease. In conclusion, these new findings not only extend our knowledge on MLL leukemia, but also have an impact on our thinking concerning the concept of the requirement of secondary mutations in the multistep pathogenetic process leading to the overt leukemia, with MLL-positive leukemia representing a rare exception to the rule. Indeed, recent data suggest that the MLL-positive leukemia can be considered an epigenetic disease, in which the profound pleiotropic chromatin remodeling effect of the fusion protein and the broad epigenetic methylation activity on a wide variety of target genes rapidly leads to the deregulation of multiple signaling pathways in the cellular network responsible for the malignant transformation.7
Conflict of interest The authors declare no conflict of interest.
Acknowledgements This study was supported by the FIRB-MIUR grant, Associazione Italiana per la Ricerca sul Cancro (AIRC), Fondazione Cariplo, Fondazione ‘M Tettamanti’ and Associazione ‘Sciare per la Vita’.
M Bardini1, M Galbiati1, A Lettieri1, S Bungaro1, TA Gorletta2,3, A Biondi1,4 and G Cazzaniga1,4 1 Centro Ricerca Tettamanti, Clinica Pediatrica Universita` Milano-Bicocca, Ospedale San Gerardo, Monza, Italy and 2 Consorzio Genopolis, BTBS, Universita` di Milano-Bicocca, Milano, Italy E-mail:
[email protected] 3 Current address: Dipartimento di Biotecnologie e Bioscienze, Universita` di Milano-Bicocca, Milano, Italy. 4 These authors share the seniorship of this paper. References 1 Greaves MF, Wiemels J. Origins of chromosome translocations in childhood leukaemia. Nat Rev Cancer 2003; 3: 639–649. 2 Gilliland DG, Jordan CT, Felix CA. The molecular basis of leukemia. Hematology 2004; 80–97. Leukemia
Letters to the Editor
178 3 Ford AM, Ridge SA, Cabrera ME, Mahmoud H, Steel CM, Chan LC et al. In utero rearrangements in the trithorax-related oncogene in infant leukaemias. Nature 1993; 363: 358–360. 4 Mullighan CG, Goorha S, Radtke I, Miller CB, Coustan-Smith E, Dalton JD et al. Genome-wide analysis of genetic alterations in acute lymphoblastic leukaemia. Nature 2007; 446: 758–764. 5 Paulsson K, Cazier JB, Macdougall F, Stevens J, Stasevich I, Vrcelj N et al. Microdeletions are a general feature of adult and adolescent
acute lymphoblastic leukemia: unexpected similarities with pediatric disease. Proc Natl Acad Sci USA 2008; 105: 6708–6713. 6 Bardini M, Spinelli R, Bungaro S, Mangano E, Corral L, Cifola I et al. DNA copy-number abnormalities do not occur in infant ALL with t(4;11)/MLL-AF4. Leukemia 2010; 24: 169–176. 7 Krivtsov AV, Feng Z, Lemieux ME, Faber J, Vempati S, Sinha AU et al. H3K79 methylation profiles define murine and human MLL-AF4 leukemias. Cancer Cell 2008; 14: 355–368.
IDH2 somatic mutations in chronic myeloid leukemia patients in blast crisis Leukemia (2011) 25, 178–181; doi:10.1038/leu.2010.236; published online 21 October 2010
We read with interest a series of manuscripts,1–6 reporting somatic mutations of the isocitrate dehydrogenase 1 and 2 enzyme isoforms (IDH1, IDH2) in patients with de novo acute myeloid leukemia (AML), in patients with chronic and blast phase Philadelphia chromosome-negative (Ph–) myeloproliferative neoplasms (MPNs) and in patients with early and accelerated phase myelodysplastic syndromes (MDSs). We have recently screened, by massively parallel sequencing, the transcriptome of a Philadelphia chromosome-positive (Ph þ ) chronic myeloid leukemia (CML) patient. The patient was a 62-year-old female diagnosed with Ph þ p210BCRABL-positive CML, high risk according to both Sokal (2.4) and Euro (1704.39) scores. No additional chromosomal abnormalities were detected at diagnosis by chromosome banding analysis. The patient was enrolled on a phase 2 study of nilotinib 400 mg twice daily as first line treatment of CML.7 She achieved a major molecular response (BCR-ABL/ABL ¼ 0.1% according to the International Scale) after 3 months, but suddenly progressed to lymphoid blast crisis (BC) after 6 months from diagnosis. At that time, mutation screening of the Abl kinase domain performed by Sanger sequencing showed evidence of a nilotinib-resistant T315I mutation, whereas no additional chromosomal abnormalities were detectable. The patient died of her disease 1 month, thereafter. After having obtained written informed consent from her next of kin, we retrieved the samples collected at diagnosis, at the time of major molecular response and at the time of disease progression, isolated messenger RNA and proceeded to massively parallel sequencing on a Solexa Illumina Genome Analyzer II platform according to manufacturer’s instructions. An IDH2 R140Q heterozygous mutation deriving from a G to A nucleotide substitution on chromosome 15, position 88432938 (hg18, NCBI build 36.1) was found in the sample collected at the time of progression to lymphoid BC, but this variant was not seen in the sample collected at the time of diagnosis nor in the sample collected at the time of remission. Conventional Sanger sequencing performed on genomic DNA confirmed massively parallel sequencing findings (Figures 1a–c). Other sequence changes were also detected that will be reported in detail elsewhere. A number of recent studies have reported this same IDH2 R140Q mutation in patients with various forms of leukemia, but data in CML and acute lymphoblastic leukemia (ALL) were still lacking. This prompted us to investigate the frequency of IDH1 and IDH2 mutations in a large series of CML and ALL patients. IDH1 and IDH2 exon 4, where all the mutations so far Leukemia
reported map, were screened by direct sequencing in 50 patients with newly diagnosed chronic phase (CP) CML, 5 patients with accelerated phase CML, 30 patients with lymphoid BC CML, 30 patients with myeloid BC CML, 5 patients with mixed myeloid/lymphoid BC CML, 34 patients with Ph þ ALL and 23 patients with Ph– ALL (Table 1) (methods detailed in Supplementary Information), but no mutations were found. Given
Figure 1 Presence of the R140Q IDH2 mutation at the time of progression to blast crisis (BC) (c), but neither at the time of diagnosis (a) nor at the time of remission (b). In the original lymphoid BC patient sample, analyzed by massively parallel sequencing, conventional Sanger sequencing confirmed that the G to A nucleotide substitution at position 88432938 (hg18, NCBI build 36.1), corresponding to the R140Q mutation, was present in the sample collected at the time of disease progression, but not before. Unfortunately, for the three additional myeloid BC patients found to harbour the same R140Q mutation, no paired chronic phase samples were available to confirm that, similar to the case shown here, the mutation was associated with disease progression.
