Genome-wide single-nucleotide polymorphism analysis in juvenile myelomonocytic leukemia identifies uniparental disomy surrounding the NF1 locus in cases associated with neurofibromatosis but not in cases with mutant RAS or PTPN11

Abstract

Juvenile myelomonocytic leukemia (JMML) is a malignant hematopoietic disorder whose proliferative component is a result of RAS pathway deregulation caused by somatic mutation in the RAS or PTPN11 oncogenes or in patients with underlying neurofibromatosis type 1 (NF-1), by loss of NF1 gene function. To search for potential collaborating genetic abnormalities, we used oligonucleotide arrays to analyse over 116 000 single-nucleotide polymorphisms across the genome in 16 JMML samples with normal karyotype. Evaluation of the SNP genotypes identified large regions of homozygosity on chromosome 17q, including the NF1 locus, in four of the five samples from patients with JMML and NF-1. The homozygous region was at least 55 million base pairs in each case. The genomic copy number was normal within the homozygous region, indicating uniparental disomy (UPD). In contrast, the array data provided no evidence for 17q UPD in any of the 11 JMML cases without NF-1. We used array-based comparative genomic hybridization to confirm 17q disomy, and microsatellite analysis was performed to verify homozygosity. Mutational analysis demonstrated that the inactivating NF1 lesion was present on both alleles in each case. In summary, our data indicate that a mitotic recombination event in a JMML-initiating cell led to 17q UPD with homozygous loss of normal NF1, provide confirmatory evidence that the NF1 gene is crucial for the increased incidence of JMML in NF-1 patients, and corroborate the concept that RAS pathway deregulation is central to JMML pathogenesis.

Main

Juvenile myelomonocytic leukemia (JMML) is a clonal hematopoietic disorder of infancy and early childhood characterized by excessive proliferation of the monocytic and granulocytic lineages. Typical features include marked hepatosplenomegaly, skin infiltration, increased white blood cell count, absolute monocytosis, anemia and thrombocytopenia (Niemeyer et al., 1997). A key feature of JMML progenitor cells is hypersensitivity to granulocyte–macrophage colony-stimulating factor (Emanuel et al., 1991) as a result of deregulated signaling through the RAS pathway (reviewed in Flotho et al., 2007). The molecular defect behind the hyperactivity of this pathway in JMML is a somatic PTPN11 point mutation in 35% of cases or a somatic NRAS/KRAS point mutation in 25% of cases. Of the JMML patients, 11% have constitutional neurofibromatosis type 1 (NF-1). In these cases, RAS hyperactivity is caused by inactivation of both alleles of the NF1 tumor suppressor gene, a negative RAS regulator (Side et al., 1997). Cytogenetic studies of leukemic cells reveal monosomy 7 or large 7q deletions in approximately 30% of JMML cases, and other chromosomal abnormalities are found in 10% (Niemeyer et al., 1997). These aberrations occur independently of the mutational status of PTPN11, NRAS/KRAS or NF1. Although the contribution of cytogenetic aberrations to the pathogenesis of JMML is largely unclear, the frequency of their occurrence suggests a role for genetic events that collaborate with RAS pathway lesions in JMML.

Single-nucleotide polymorphism (SNP) microarrays offer the opportunity to analyse simultaneously large numbers of polymorphic loci across the genome, providing high-density information on regional copy number changes and/or loss of heterozygosity (LOH) (reviewed in Engle et al., 2006). Allelic imbalance may play an important role in leukemogenesis through the loss of tumor suppressor genes or the amplification of oncogenes. In this study, we used SNP arrays to assess whether submicroscopic allelic gains or losses are a frequent phenomenon in JMML cells, which exhibit a normal karyotype when examined by standard cytogenetics. In addition, we set out to identify regions of LOH with potential significance in the pathogenesis of JMML.

Genomic DNA of bone marrow or peripheral blood granulocytes from 16 children with JMML was used for SNP array analysis (Table 1). We preferred granulocytes to the mononuclear cell fraction, because of the unclear clonality of lymphocytes in JMML (Flotho et al., 1999). Four samples carried a RAS gene mutation, seven cases were mutant for PTPN11, and five patients had the clinical diagnosis of NF-1. We first screened the 16 JMML samples for genomic amplifications or deletions. For this purpose, the DNA copy number at each SNP locus was estimated from the corresponding signal intensity on the array. The SNP array data revealed several sporadic aberrations that had previously been unrecognized: loss of 6q16–q27 material coupled with gain of 3q22–q29 in sample D003 (Figure 1a) and gain of 13q12–q34 in sample D119 (not shown). To verify these findings in an independent assay, we performed array-based comparative genomic hybridization of the two samples involved. The results were concordant with the SNP array data in that all three abnormalities were confirmed and no additional regions of allelic imbalance were found (Figure1b and data not shown). Interestingly, loss of 6q material is a frequent aberration in non-Hodgkin lymphoma and is also occasionally found in myelodysplastic syndromes (MDS) (Steinemann et al., 2003; Mitelman et al., 2006). Gain of 3q is associated with clonal evolution from Fanconi anemia to MDS (Tönnies et al., 2003). 13q amplifications have recently been linked to overexpression of an oncogenic microRNA cluster in B-cell lymphomas (He et al., 2005). Thus, we have not identified any recurring regions of genomic amplification or deletion in JMML. Specifically, the SNP array data provided no evidence of a cytogenetically cryptic segment on 7q that was commonly affected by allelic imbalance or LOH. However, the detection of several sporadic aberrations that were not identified by conventional cytogenetics serves to validate SNP arrays as a tool for the analysis of regional copy number abnormalities in the genome, particularly when cell material suitable for standard karyotyping is unavailable.

Table 1 Patient characteristics and molecular analysis of leukemic cells
Figure 1
figure1

Analysis of allelic imbalance in JMML sample D003. (a) For SNP array analysis, granulocytes were purified by density centrifugation on a Ficoll gradient. DNA was extracted using the Trifast kit (Peqlab, Erlangen, Germany) and checked for integrity by agarose gel electrophoresis. The oligonucleotide arrays (GeneChip Mapping 100K array set, Affymetrix, Santa Clara, CA, USA) were prepared according to the manufacturer's instructions. Briefly, genomic DNA was digested using HindIII or XbaI, before ligation to adapters and PCR-based amplification of adapter-ligated fragments. Amplified products were fragmented, end-labeled and hybridized to the probe arrays. Signals corresponding to a total of 116 204 probe sets were read using a confocal scanner (Affymetrix). DNA copy numbers were calculated from raw signal values using dChip (Zhao et al., 2004). A reference data set of 60 acute myeloid leukemia remission samples processed at the same facility was used as standard for diploidy at each locus. Genomic copy numbers at 3955 SNP loci on chromosome 3 and 3959 SNP loci on chromosome 6 are displayed. Values were inferred by median smoothing with window size 50. The data indicate gain of a 63-Mbp segment comprised of chromosome bands 3q22–q29 and heterozygous deletion of a 72-Mbp segment corresponding to bands 6q16–q27. (b) Confirmation of copy number abnormalities by array-based comparative genomic hybridization. We used a DNA chip containing more than 8000 bacterial artificial chromosome/P1-derived artificial chromosome clones (manufactured at DKFZ, Heidelberg, Germany). Clone selection, spotting, labeling and hybridization were performed as described previously (Zielinski et al., 2005). Image data were acquired using a dual laser scanner and GenePix Pro 4.0 software (Axon Instruments, Union City, CA, USA). Further analysis was carried out using R software (‘marray’ and ‘aCGH’, http://www.r-project.org). Raw fluorescence intensity values were normalized with the print-tip LOESS function, and spot quality criteria were set as foreground to background >3.0 and standard deviation of triplicates <0.2. GLAD software was used for breakpoint calling (Hupe et al., 2004). Signal ratios of 388 probes on chromosome 3 and 318 probes on chromosome 6 are displayed.

We next evaluated the SNP genotypes of the 16 JMML samples in search of genomic regions with LOH. We identified large regions of copy-neutral homozygosity on chromosome 17 in four of five samples from patients with JMML and NF-1 (Figure 2 and Table 1). The genomic region spanned a range of approximately 55 million base pairs (Mbp) in each case and involved more than 1400 contiguous SNPs. Importantly, in each case, the homozygous area included the locus of the NF1 tumor suppressor gene on 17q11.2. The absence of corresponding genomic deletion in the regions of homozygosity suggested the presence of uniparental disomy (UPD) surrounding the NF1 tumor suppressor gene in leukemic cells of four JMML patients with constitutional neurofibromatosis. By contrast, our data did not indicate LOH on 17q in the 11 JMML samples from children without NF-1. In all four cases with 17q UPD, the recombination breakpoints appeared to be located within a 400-kbp region on 17q11.1–17q11.2; however, the unavailability of non-leukemic or parental DNA precluded precise mapping of the breakpoints. Normal genomic copy number of 17q was confirmed by array-based comparative genomic hybridization in all cases (data not shown). To confirm homozygosity of the affected region, we genotyped eight microsatellite polymorphisms selected on the basis of high heterozygote frequency and independent segregation. All markers were homozygous in the four samples as demonstrated by the appearance of only one allele in microsatellite analysis (Table 2). Considering the known allele frequencies of the microsatellites in combination, it is highly unlikely that inherited haplotypes would account for the uniform homozygosity observed. The results are therefore consistent with the presence of 17q UPD. We also performed microsatellite analysis on the leukemic DNA of one JMML patient who did not have NF-1 (sample D119). As expected, and consistent with the SNP array data, all markers but one were heterozygous in this sample.

Figure 2
figure2

Genome-wide LOH analysis in JMML. The SNP array preparation is described in the legend to Figure 1. Genotypes were derived from the scanned array image using proprietary software (GCOS/GTYPE, Affymetrix) at default settings except for a more stringent cutoff level for equivocal genotypes (call confidence P-value <0.05 instead of <0.25 as preselected in the software). Genotype information was generated for 85 425–109 880 polymorphic markers in each sample. (Hybridization of sample D102 to the XbaI array failed; therefore, 45 654 HindIII SNPs were used for further analysis of D102.) Information on genomic position was missing for 851 SNPs; these SNPs were excluded from further analysis. Since paired constitutional DNA for each leukemia sample was not available, LOH was analysed through the statistical inference model implemented in dChip software (Lin et al., 2004). The inferred probability of LOH at 1955 SNP loci on chromosome 17 is displayed for JMML samples D003, D102, D115, D126 and D127. Each of the five samples was obtained from a patient with NF-1. Four of the five samples show LOH of a large segment on chromosome arm 17q. The genomic position of the NF1 gene is indicated with an arrow.

Table 2 Microsatellite analysis of chromosome arm 17q in five leukemic samples from patients with JMML

The observations that 17q UPD was limited to patients with NF-1 and that the isodisomic region included the NF1 locus is consistent with the concept that the lack of a functional NF1 gene is critical for the emergence of tumors or leukemias in individuals with neurofibromatosis (reviewed in Dasgupta and Gutmann, 2003; Lauchle et al., 2006). Patients with NF-1 carry one intact and one deficient NF1 allele in the germline, with the deficient allele resulting from inheritance or de novo mutation. To demonstrate that the mutant, not the normal, NF1 allele was isodisomic in the leukemias with 17q UPD, we performed DNA sequencing for mutational analysis of the NF1 gene in leukemic granulocytes from the five patients with JMML and NF-1. Sample D003 contained an intragenic NF1 deletion, c.3861_3862delCT (Figure 3a). This frameshift mutation results in a truncated protein due to a premature termination codon at amino-acid position 1312. Similarly, sample D115 was found to carry a deletion c.2066delT (frameshift; protein truncated at position 747). Sample D102 exhibited a point mutation c.574C>T, substituting arginine at position 192 with a premature stop codon. An analogous point mutation, c.7699C>T, was found in sample D116, replacing glutamine 2567 with a termination codon. Importantly, the sequencing data showed that each NF1 mutation was homozygous in these four samples with 17q UPD, confirming that the normal NF1 allele had been lost from leukemic cells in these cases. By contrast, the absence of 17q LOH from sample D127 suggested a compound-heterozygous mechanism of NF1 inactivation in this case. As expected, we identified a c.2288_2295dupTGAGGCGC duplication on one allele in combination with a c.3366delT deletion on the other allele (Figure 3b). Both changes cause a frameshift and introduce premature termination codons at amino-acid positions 766 and 1141, respectively. Our results validate the concept that somatic inactivation of the normal NF1 allele in a myeloid progenitor cell is the key process leading to the occurrence of JMML in individuals with a heterozygous germline NF1 defect, even in a case that did not undergo reduction to homozygosity at the NF1 locus.

Figure 3
figure3

Mutational analysis of the NF1 gene in JMML cells. All 60 coding exons including adjacent splice sites were PCR-amplified from genomic DNA of JMML granulocytes and sequenced bidirectionally. Mutations were confirmed by sequencing an independent, reamplified PCR product. (a) The analysis of sample D003 revealed a 2-bp deletion in exon 22 at nucleotide positions 3861 and 3862. This frameshift mutation results in synthesis of truncated neurofibromin. Homozygosity of the mutation is indicated by the absence of double peaks downstream of the frameshift. (b) Heterozygous single-nucleotide deletion at position 3366 in sample D127.

We used high-density SNP microarrays to assess simultaneously the genotypes and allelic dosage at more than 116 000 loci across the genome of leukemic cells from 16 children with JMML. This led to the identification of large isodisomic regions, encompassing the NF1 tumor suppressor gene, in four of five samples from patients with JMML and constitutional neurofibromatosis. With a median intermarker distance of 8.5 kb, this study represents the highest resolution analysis of copy number abnormalities and LOH in JMML thus far reported. It is noteworthy that the arrays failed to detect segmental LOH on chromosomal arm 7q in any of the 16 cytogenetically normal cases. However, it is possible that the cases examined harbor extremely focal abnormalities below the limit of resolution of the arrays used. Higher resolution approaches, such as with currently available 500 K SNP arrays or oligonucleotide tiling arrays, will be of interest.

In several aspects, the results illustrate the utility of SNP arrays as a screening tool for genetic aberrations in cancer or leukemia (Raghavan et al., 2005). First, the isodisomic regions would appear normal if standard cytogenetic techniques such as metaphase karyotyping or fluorescence in situ hybridization were applied. Second, conventional methods for genotyping, for example, the examination of short tandem repeats, identify LOH but have the disadvantage that they can only be performed on one polymorphic marker at a time and require additional tests to distinguish between isodisomy and deletion.

In individuals with NF-1, loss of the normal NF1 allele from somatic tissues is frequently associated with, if not mandatory for, the development of tumors and/or leukemia (Dasgupta and Gutmann, 2003; Arun and Gutmann, 2004; Lauchle et al., 2006). Others have very recently reported that uniparental disomy of large 17q segments is frequent in NF-1-associated leukemias (Stephens et al., 2006). In combination with our results, these findings indicate a prominent but as yet underappreciated role for mitotic recombination in NF1-mediated leukemogenesis. Interestingly, Stephens et al. (2006) demonstrated that the isodisomic region was interstitial in a subset of their samples, implying the occurrence of double recombination between both 17q homologues. In the samples investigated here, SNP array data and microsatellite analyses do not support the idea of a double event because in each case the markers were homozygous throughout 17q distal to the rearrangement. However, our data may still be reconcilable with interstitial isodisomy, as constitutional homozygosity of the telomeric markers, although highly unlikely, cannot be ruled out in the absence of parental or non-leukemic genotypes.

To our knowledge, this is the first report to define at the mutational level a compound-heterozygous NF1 gene lesion in an NF-1-associated leukemia without LOH at the NF1 locus. In addition, our data indicate that 17q UPD was not present in any of the 11 JMML cases unrelated to NF-1. These findings underscore that isodisomy is not a coincidental observation in the leukemic genome of patients with NF-1 who develop JMML. Furthermore, the results presented here corroborate the concept that gene lesions interacting with the RAS pathway are central to the pathogenesis of JMML.

References

  1. Arun D, Gutmann DH . (2004). Recent advances in neurofibromatosis type 1. Curr Opin Neurol 17: 101–105.

  2. Dasgupta B, Gutmann DH . (2003). Neurofibromatosis 1: closing the GAP between mice and men. Curr Opin Genet Dev 13: 20–27.

  3. Emanuel PD, Bates LJ, Castleberry RP, Gualtieri RJ, Zuckerman KS . (1991). Selective hypersensitivity to granulocyte–macrophage colony-stimulating factor by juvenile chronic myeloid leukemia hematopoietic progenitors. Blood 77: 925–929.

  4. Engle LJ, Simpson CL, Landers JE . (2006). Using high-throughput SNP technologies to study cancer. Oncogene 25: 1594–1601.

  5. Flotho C, Kratz CP, Niemeyer CM . (2007). Targeting RAS signaling pathways in juvenile myelomonocytic leukemia. Curr Drug Targets in press.

  6. Flotho C, Valcamonica S, Mach-Pascual S, Schmahl G, Corral L, Ritterbach J et al. (1999). RAS mutations and clonality analysis in children with juvenile myelomonocytic leukemia (JMML). Leukemia 13: 32–37.

  7. He L, Thomson JM, Hemann MT, Hernando-Monge E, Mu D, Goodson S et al. (2005). A microRNA polycistron as a potential human oncogene. Nature 435: 828–833.

  8. Hupe P, Stransky N, Thiery JP, Radvanyi F, Barillot E . (2004). Analysis of array CGH data: from signal ratio to gain and loss of DNA regions. Bioinformatics 20: 3413–3422.

  9. Lauchle JO, Braun BS, Loh ML, Shannon K . (2006). Inherited predispositions and hyperactive Ras in myeloid leukemogenesis. Pediatr Blood Cancer 46: 579–585.

  10. Lin M, Wei LJ, Sellers WR, Lieberfarb M, Wong WH, Li C . (2004). dChipSNP: significance curve and clustering of SNP-array-based loss-of-heterozygosity data. Bioinformatics 20: 1233–1240.

  11. Mitelman F, Johansson B, Mertens F . 2006. Mitelman Database of Chromosome Aberrations in Cancer. http://cgap.nci.nih.gov/Chromosomes/Mitelman.

  12. Niemeyer CM, Arico M, Basso G, Biondi A, Cantu RA, Creutzig U et al. (1997). Chronic myelomonocytic leukemia in childhood: a retrospective analysis of 110 cases. European Working Group on Myelodysplastic Syndromes in Childhood (EWOG-MDS). Blood 89: 3534–3543.

  13. Raghavan M, Lillington DM, Skoulakis S, Debernardi S, Chaplin T, Foot NJ et al. (2005). Genome-wide single nucleotide polymorphism analysis reveals frequent partial uniparental disomy due to somatic recombination in acute myeloid leukemias. Cancer Res 65: 375–378.

  14. Side L, Taylor B, Cayouette M, Conner E, Thompson P, Luce M et al. (1997). Homozygous inactivation of the NF1 gene in bone marrow cells from children with neurofibromatosis type 1 and malignant myeloid disorders. N Engl J Med 336: 1713–1720.

  15. Steinemann D, Gesk S, Zhang Y, Harder L, Pilarsky C, Hinzmann B et al. (2003). Identification of candidate tumor-suppressor genes in 6q27 by combined deletion mapping and electronic expression profiling in lymphoid neoplasms. Genes Chromosomes Cancer 37: 421–426.

  16. Stephens K, Weaver M, Leppig KA, Maruyama K, Emanuel PD, Le Beau MM et al. (2006). Interstitial uniparental isodisomy at clustered breakpoint intervals is a frequent mechanism of NF1 inactivation in myeloid malignancies. Blood 108: 1684–1689.

  17. Tönnies H, Huber S, Kuhl JS, Gerlach A, Ebell W, Neitzel H . (2003). Clonal chromosomal aberrations in bone marrow cells of Fanconi anemia patients: gains of the chromosomal segment 3q26q29 as an adverse risk factor. Blood 101: 3872–3874.

  18. Zhao X, Li C, Paez JG, Chin K, Janne PA, Chen TH et al. (2004). An integrated view of copy number and allelic alterations in the cancer genome using single nucleotide polymorphism arrays. Cancer Res 64: 3060–3071.

  19. Zielinski B, Gratias S, Toedt G, Mendrzyk F, Stange DE, Radlwimmer B et al. (2005). Detection of chromosomal imbalances in retinoblastoma by matrix-based comparative genomic hybridization. Genes Chromosomes Cancer 43: 294–301.

Download references

Acknowledgements

This work was carried out with the grant support from Kind Philipp Foundation T237/15893/2006 (CF), German Federal Ministry of Education and Research BMBF-DLR 01GM0307 (DS, BS); José Carreras Leukemia Foundation DJCLS R05-03 (DS, CPK, BS, CMN).

Author information

Correspondence to C Flotho.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Flotho, C., Steinemann, D., Mullighan, C. et al. Genome-wide single-nucleotide polymorphism analysis in juvenile myelomonocytic leukemia identifies uniparental disomy surrounding the NF1 locus in cases associated with neurofibromatosis but not in cases with mutant RAS or PTPN11. Oncogene 26, 5816–5821 (2007) doi:10.1038/sj.onc.1210361

Download citation

Keywords

  • leukemia
  • JMML
  • SNP
  • NF1