Rare autosomal copy number variations in early-onset familial Alzheimer's disease.

Over 200 rare and fully penetrant pathogenic mutations in amyloid precursor protein (APP), presenilin 1 and 2 (PSEN1 and PSEN2) cause a subset of early-onset familial Alzheimer's disease (EO-FAD). Of these, 21 cases of EO-FAD families carrying unique APP locus duplications remain the only pathogenic copy number variations (CNVs) identified to date in Alzheimer's disease (AD). Using high-density DNA microarrays, we performed a comprehensive genome-wide analysis for the presence of rare CNVs in 261 EO-FAD and early/mixed-onset pedigrees. Our analysis revealed 10 novel private CNVs in 10 EO-FAD families overlapping a set of genes that includes: A2BP1, ABAT, CDH2, CRMP1, DMRT1, EPHA5, EPHA6, ERMP1, EVC, EVC2, FLJ35024 and VLDLR. In addition, CNVs encompassing two known frontotemporal dementia genes, CHMP2B and MAPT were found. To our knowledge, this is the first study reporting rare gene-rich CNVs in EO-FAD and early/mixed-onset AD that are likely to underlie pathogenicity in familial AD and perhaps related dementias.

Identification of LMX1B as a novel oncogene in human ovarian cancer.

Ovarian cancers are thought to result from the accumulation of multiple genetic aberrations that transform ovarian and/or fallopian tube surface epithelial cells, allowing for their abnormal growth, proliferation and metastasis. In the report presented here, we carried out genome-wide copy-number analysis using comparative genomic hybridization on a panel of mouse ovarian cancer (OVCA) cell lines previously established in our laboratory. We identified a recurrent focal amplification on mouse chromosomal region 2qB, which contains the LIM-homeodomain-containing transcription factor 1B (Lmx1b) gene. LMX1B is not expressed in normal human ovary, but is expressed in many human OVCA cell lines and primary tumors. High expression of LMX1B correlates with poor outcome. To clarify the role of LMX1B in ovarian carcinogenesis, we transduced LMX1B into a panel of mouse and human OVCA cell lines and demonstrated that LMX1B strongly promotes migration of cancer cells in culture and promotes xenograft growth in nude mice. Conversely, knockdown of LMX1B in a human cell line with endogenous high expression of LMX1B inhibits cell migration in vitro and tumor growth in vivo. Microarray analysis of cells overexpressing LMX1B identified the nuclear factor (NF)-κB pathway as a potential mediator of tumor progression and subsequent treatment of NFκB inhibitor decreased the migratory capacity of these cells. Thus, our data demonstrate that LMX1B is a novel oncogene in OVCA pathogenesis.

Role of common and rare APP DNA sequence variants in Alzheimer disease.

OBJECTIVES: More than 30 different rare mutations, including copy number variants (CNVs), in the amyloid precursor protein gene (APP) cause early-onset familial Alzheimer disease (EOFAD), whereas the contribution of common APP variants to disease risk remains controversial. In this study we systematically assessed the role of both rare and common APP DNA variants in Alzheimer disease (AD) families. METHODS: Families with EOFAD genetically linked to the APP region were screened for missense mutations and locus duplications of APP. Further, using genome-wide DNA microarray data, we examined the APP locus for CNVs in a total of 797 additional early- and late-onset AD pedigrees. Finally, 423 single nucleotide polymorphisms (SNPs) in the APP locus, including 2 promoter polymorphisms previously associated with AD risk, were tested in up to 4,200 individuals from multiplex AD families. RESULTS: Analyses of 8 21q21-linked families revealed one family carrying a nonsynonymous mutation in exon 17 (Val717Leu) and another family with a partially penetrant 3.5-Mb locus duplication encompassing APP. CNV analysis in the APP locus revealed an additional family carrying a fully penetrant 380-kb duplication, merely spanning APP. Last, contrary to previous reports, association analyses of more than 400 different SNPs in or near APP failed to show significant effects on AD risk. CONCLUSION: Our study shows that APP mutations and locus duplications are a very rare cause of EOFAD and that the contribution of common APP variants to AD susceptibility is insignificant. Furthermore, duplications of APP may not be fully penetrant, possibly indicating the existence of hitherto unknown protective genetic factors.

Gain of chromosome arm 1q in atypical meningioma correlates with shorter progression-free survival.

AIMS: Atypical (World Health Organization grade II) meningiomas have moderately high recurrence rates; even for completely resected tumours, approximately one-third will recur. Post-operative radiotherapy may aid local control and improve survival, but carries the risk of side effects. More accurate prediction of recurrence risk is therefore needed for patients with atypical meningioma. Previously, we used high-resolution array comparative genomic hybridization to identify genetic variations in 47 primary atypical meningiomas and found that approximately 60% of tumours show gain of 1q at 1q25.1 and 1q25.3 to 1q32.1 and that 1q gain appeared to correlate with shorter progression-free survival. This study aimed to validate and extend these findings in an independent sample. METHODS: Eighty-six completely resected atypical meningiomas (with 25 recurrences) from two neurosurgical centres in Ireland were identified and clinical follow-up was obtained. Utilizing a dual-colour interphase fluorescence in situ hybridization assay, 1q gain was assessed using Bacterial Artificial Chromosome probes directed against 1q25.1 and 1q32.1. RESULTS: The results confirm the high prevalence of 1q gain at these loci in atypical meningiomas. We further show that gain at 1q32.1 and age each correlate with progression-free survival in patients who have undergone complete surgical resection of atypical meningiomas. CONCLUSIONS: These independent findings suggest that assessment of 1q copy number status can add clinically useful information for the management of patients with atypical meningiomas.

In vitro phosphorylation of BRCA2 by the checkpoint kinase CHEK2.

Germline mutations in both BRCA2 and CHEK2 are associated with an increased risk for male breast cancer. To search for potential interactions between the products of these breast cancer susceptibility genes, we undertook systematic mapping of BRCA2 for potential phosphorylation sites by CHEK2. In vitro kinase assays and mass spectrometric analysis identified a 50 amino-acid fragment within the N-terminus of BRCA2 potentially targeted by CHEK2, containing two major phosphopeptides. Inducible overexpression of this peptide, but not a derivative with mutated phosphorylation sites, leads to increased chromosome fragmentation and suppression of cellular proliferation. These results suggest a link between CHEK2 and BRCA2 pathways, which may contribute to the spectrum of cancers associated with germline CHEK2 mutations.

Genetic subgroups of anaplastic astrocytomas correlate with patient age and survival.

Astrocytomas are brain tumors with variable responses to radiation and chemotherapy. Tumor grade and patient age are important prognostic factors but do not account for the variability in clinical outcome. We hypothesized that genetic subgroups play a role in the outcome of grade III astrocytomas and studied 80 grade III astrocytomas by comparative genomic hybridization. Some chromosomal aberrations (+7p/q, -9p, -10q, -13q, +19q) were related to aberrations that are frequent in grade IV astrocytoma, whereas others (+10p, -11q, +11p, -Xq) were more frequent in grade III astrocytoma. +7p, +19 and -4q were more frequent in tumors from older patients while -11p was more frequent in tumors from younger patients. Finally, gains of 7p and 7q were associated with shorter patient survival, independent of age. Our results indicate that genetic events underlie the well-known effects of age on survival in grade III astrocytoma and demonstrate the importance of molecular classification in astrocytic tumors.

HYAL1LUCA-1, a candidate tumor suppressor gene on chromosome 3p21.3, is inactivated in head and neck squamous cell carcinomas by aberrant splicing of pre-mRNA.

The hyaluronidase first isolated from human plasma, Hyal-1, is expressed in many somatic tissues. The Hyal-1 gene, HYAL1, also known as LUCA-1, maps to chromosome 3p21.3 within a candidate tumor suppressor gene locus defined by homozygous deletions and by functional tumor suppressor activity. Hemizygosity in this region occurs in many malignancies, including squamous cell carcinomas of the head and neck. We have investigated whether cell lines derived from such malignancies expressed Hyal-1 activity, using normal human keratinocytes as controls. Hyal-1 enzyme activity and protein were absent or markedly reduced in six of seven carcinoma cell lines examined. Comparative genomic and fluorescence in situ hybridization identified chromosomal deletions of one allele of HYAL1 in six of seven cell lines. Initial RT - PCR analyses demonstrated marked discrepancies between levels of HYAL1 mRNA and protein. Despite repeated sequence analyses, no mutations were found. However, two species of transcripts were identified when primers were used that included the 5' untranslated region. The predominant mRNA species did not correlate with protein translation and contained a retained intron. A second spliced form lacking this intron was found only in cell lines that produced Hyal-1 protein. Inactivation of HYAL1 in these tumor lines is a result of incomplete splicing of its pre-mRNA that appears to be epigenetic in nature. Oncogene (2000) 19, 870 - 877.

Localization of common deletion regions on 1p and 19q in human gliomas and their association with histological subtype.

Allelic alterations of chromosomes 1 and 19 are frequent events in human diffuse gliomas and have recently proven to be strong predictors of chemotherapeutic response and prolonged survival in oligodendrogliomas (Cairncross et al., 1998; Smith et al., submitted). Using 115 human diffuse gliomas, we localized regions of common allelic loss on chromosomes 1 and 19 and assessed the association of these deletion intervals with glioma histological subtypes. Further, we evaluated the capacity of multiple modalities to detect these alterations, including loss of heterozygosity (LOH), fluorescence in situ hybridization (FISH), and comparative genomic hybridization (CGH). The correlation coefficients for detection of 1p and 19q alterations, respectively, between modalities were: 0.98 and 0.87 for LOH and FISH, 0.79 and 0.60 for LOH and CGH, and 0.79 and 0.53 for FISH and CGH. Minimal deletion regions were defined on 19q13.3 (D19S412-D19S596) and 1p (D1S468-D1S1612). Loss of the 1p36 region was found in 18% of astrocytomas (10/55) and in 73% (24/33) of oligodendrogliomas (P < 0.0001), and loss of the 19q13.3 region was found in 38% (21/55) of astrocytomas and 73% (24/33) of oligodendrogliomas (P = 0.0017). Loss of both regions was found in 11% (6/55) of astrocytomas and in 64% (21/33) of oligodendrogliomas (P < 0.0001). All gliomas with LOH on either 1p or 19q demonstrated loss of the corresponding FISH probe, 1p36 or 19q13.3, suggesting not only locations of putative tumor suppressor genes, but also a simple assay for assessment of 1p and 19q alterations as diagnostic and prognostic markers.

Comparative genomic hybridization in patients with supratentorial and infratentorial primitive neuroectodermal tumors.

BACKGROUND: Intracranial primitive neuroectodermal tumors (PNETs) occur in the supratentorial and infratentorial regions of the brain. Although histologically similar, the natural history of the tumor at each site differs. The study goal was to determine whether there was evidence of a genetic difference between supratentorial and infratentorial PNETs. METHODS: Using comparative genomic hybridization (CGH), 53 PNETs were analyzed to determine copy number aberrations. Forty-three tumors were located in the cerebellum (IPNETs), and ten were supratentorial PNETs (SPNETs). All samples were reviewed to confirm the diagnosis. Each specimen had at least 50% tumor. RESULTS: Six of the 43 cases of IPNET had no copy number aberrations. In contrast, each case of SPNET had copy number aberrations detected by CGH. Statistically significant differences in copy number aberrations of chromosomes 14, 17, and 19 were detected in the two groups. The most common copy number aberration in the IPNETs was gain of chromosome 17q, which was observed in 16 of 43 cases (37%). However, no case of SPNET had gain of 17q. Loss of 14q was detected in four of ten SPNETs but was not detected in any of the IPNET cases. Loss of 19q was detected in 4 of 10 SPNETs and in only 1 of 43 IPNETs. CONCLUSIONS: These results indicate that the genetic aberrations of IPNETs differ from the genetic aberrations of SPNETs. Although they are similar histologically, SPNETs and IPNETs appear to be biologically distinct entities.

Chromosomal abnormalities in glioblastoma multiforme by comparative genomic hybridization: correlation with radiation treatment outcome.

Glioblastoma multiforme (GM) is the most common and most malignant astrocytoma in adults. After surgery, radiation therapy extends patient survival; however, in vivo response to radiation therapy is variable. The purpose of this investigation was to determine whether the cytogenetic abnormalities of GM differ according to patient response to radiation therapy. Radiation response was defined by either progression [radiation-resistant (RR)] or resolution [radiation-sensitive (RS)] of tumor at the first postradiation radiographic imaging evaluation. Twenty RR and 10 RS frozen tissue specimens were subjected to cytogenetic analysis by comparative genomic hybridization. RS and RR specimens had different cytogenetic aberrations that mapped predominantly to chromosomes 7, 9, 10, 13, and 19. Relative gain of 7 occurred in 70% of the RR and 30% of the RS cases and was the most significant difference involving a single change between the two groups (P = 0.06). RR and RS specimens also differed in their patterns of simultaneous cytogenetic aberrations. A simultaneous gain of chromosomes 7 and 19 was found in 30% of the RR cases but was absent in the RS group. Concurrent loss of 9p23-24 and 13q14 regions was absent in the RS cohort but occurred in 30% of the RR series. This latter cytogenetic pattern was also associated with older age. Amplifications were more common in the RR series, but the difference did not reach statistical significance. The data suggest that GM with different in vivo responses to radiation therapy also differ cytogenetically.

Multiple genetic aberrations including evidence of chromosome 11q13 rearrangement detected in pituitary adenomas by comparative genomic hybridization.

OBJECT: This study was conducted to determine whether comparative genomic hybridization (CGH) is a more sensitive method for detecting genetic aberrations than other tests currently in use. METHODS: The authors used CGH to examine 40 primary and 13 recurrent adenomas obtained from 52 patients for loss and gain of genetic material. Copy number aberrations (CNAs) were detected in 25 (48%) of the 52 patients studied. The chromosomes affected were, in order of decreasing frequency, 11, 7, X, 1, 8, 13, 5, 14, 2, 6, 9, 10, 12, 3, 18, 21, 4, 16, 15, 19, 22, and Y. Endocrinologically active adenomas were more likely to contain (p = 0.009) and had a greater number (p = 0.003) of CNAs. Of 26 adenomas with CNAs, 18 showed multiple aberrations involving entire chromosomes or chromosome arms. The most frequent CNA involving a chromosome subregion, which was present in four (8%) of 53 adenomas, was the loss of all chromosome 11 material except for a preserved common segment containing 11q13. Immunoperoxidase staining did not detect cyclin D1 expression in those four cases, making cyclin D1 an unlikely target of this rearrangement. CONCLUSIONS: These findings indicate that genetic abnormalities are present in pituitary adenomas at a higher rate than previously reported, are associated with endocrinological activity, and often involve several chromosomes. Rearrangement at 11q13 may inactivate a tumor suppressor gene or activate an oncogene that is important in the initiation or progression of sporadic pituitary adenomas.

Prader-Willi syndrome is caused by disruption of the SNRPN gene.

A Prader-Willi syndrome patient is described who has a de novo balanced translocation, (4;15)(q27;q11.2)pat, with breakpoints lying between SNRPN exons 2 and 3. Parental-origin studies indicate that there is no uniparental disomy and no apparent deletion. This patient expresses ZNF127, SNRPN exons 1 and 2, IPW, and D15S227E (PAR1) but does not express either SNRPN exons 3 and 4 or D15S226E (PAR5), as assayed by reverse transcription-PCR, of peripheral blood cells. Methylation studies showed normal biparental patterns of inheritance of loci DN34/ZNF127, D15S63, and SNRPN exon 1. Results for this patient and that reported by Sun et al. support the contention that an intact genomic region and/or transcription of SNRPN exons 2 and 3 play a pivotal role in the manifestations of the major clinical phenotype in Prader-Willi syndrome.

Genetic analysis of glioblastoma multiforme provides evidence for subgroups within the grade.

We analyzed 72 primary and 25 recurrent glioblastoma multiforme (GBM) samples for DNA sequence copy number abnormalities (CNAs) by comparative genomic hybridization (CGH). The number of aberrations per tumor ranged from 2 to 23 in primary GBM and 5 to 25 in recurrent GBM. There were 26 chromosome regions with CNAs in more than 20% of tumors. 7q22-36 was the most common gain and 10q25-26 was the most common loss; each occurred in more than 70% of tumors. Of 27 amplification sites, epidermal growth factor receptor (EGFR) was the most common; it was observed in 25% of primary GBMs. Statistical analysis based on pairwise correlation of CNAs indicated that there is more than one class of primary GBM.

The hyaluronidase gene HYAL1 maps to chromosome 3p21.2-p21.3 in human and 9F1-F2 in mouse, a conserved candidate tumor suppressor locus.

We recently cloned and expressed the major hyaluronidase activity from human plasma, HYAL1, and found that the protein is 40% identical to the testicular hyaluronidase, PH-20. The HYAL1 mRNA sequence was used in a homology search of the mouse database of expressed sequence tags (dbEST). Two ESTs were obtained and, in combination with 5'RACE-PCR, were used to clone the mouse HYAL1 ortholog (Hyal1). Hyal1 codes for a protein of 462 amino acids that is 73% identical to the human sequence. Hyal1 stably expressed in human embryonic kidney cells resulted in a 20,000-fold increase of hyaluronidase activity. Sequence-tagged sites derived from the HYAL1 gene from both species were used to isolate P1 genomic clones that were used as probes for fluorescence in situ hybridization. The human gene was localized to chromosome 3p21 and the mouse gene to a syntenic region on chromosome 9F1-F2. In mouse, serum hyaluronidase polymorphism has previously been mapped by an interspecific backcross to 60 cM from the centromere of chromosome 9, which corresponds to a cytogenetic location of 9F1-F2. The mouse Hyal1 gene is therefore very likely to be responsible for the hyaluronidase polymorphism linked to this locus. We also present evidence that human HYAL1 is identical to an uncharacterized gene positionally cloned by others from chromosome 3p21.3 that is homozygously deleted in several small-cell lung carcinoma cell lines.

Gliomas in families: chromosomal analysis by comparative genomic hybridization.

Gliomas that aggregate in otherwise unremarkable families may have a heritable genetic basis. To determine the spectrum of genetic alterations in glioma-susceptible families, we examined tumor DNA from familial cases for regions of chromosomal gain or loss using comparative genomic hybridization (CGH). We compared chromosomal alterations within and among glioma families to those found in sporadic gliomas. A specific chromosomal abnormality common to the tumors of multiple unrelated probands with glioma or a specific chromosomal abnormality common to multiple affected persons in a single glioma-prone family would support the hypothesis of an inherited predisposition to glioma and at the same time identify specific regions of the genome harboring putative glioma susceptibility genes. Tumor DNA from 11 patients from seven families with two or more individuals with glioma was analyzed, including three members of a remarkable family having 10 affected individuals. We found no chromosomal abnormality common to all tumors of all probands nor did we find family-specific abnormalities in two of three glioma-prone kindreds. There were frequent copy number aberrations (CNAs) on chromosomes 7, 10, 19, and the sex chromosomes; other CNAs included +3q(13.3-29), -4q, +5q, -9q34, +12, -13q(21-->33), -15, -16p, +17qter, -18, -21, and -22. Amplifications occurred at +2 7p(11.1-->12), +2 7q(21.2-->33), +2 12q(13.2-->14), and +2 12p(11-->12). Although there were several novel CNAs [-16p, and +2 12p(11-p12)], none could readily explain the inheritance of these tumors.

Analyses of brain tumor cell lines confirm a simple model of relationships among fluorescence in situ hybridization, DNA index, and comparative genomic hybridization.

Several techniques are commonly used for genetic analysis of interphase nuclei. Flow cytometry assays the distribution of DNA content in populations of nuclei stained with a DNA-specific fluorochrome. Fluorescence in situ hybridization (FISH) quantifies the number of copies of a specific DNA sequence in single nuclei. Comparative genomic hybridization (CGH) assesses the relative copy number of DNA sequences throughout a test genome by comparing the signal intensities of test and reference DNA samples hybridized to a template of normal metaphase chromosomes. In principle, there are specific relationship among data obtained from these measurements, and combined measurements should provide a more comprehensive view of the sample that is analyzed. We applied these three techniques to nine brain tumor cell lines and find that a model of CGH that includes unsuppressed repeat sequences describes the data well. We estimate that up to 35% of the fluorescence intensity in well-blocked CGH preparations may not represent unique sequences. Taking these factors into account, our results are, in general, mutually consistent, and highlight issues critical for interpreting CGH preparations.

FISH probes for mouse chromosome identification.

P1 clones near the telomeres and centromeres of each mouse chromosome except Y have been selected from a mouse genomic library and mapped using fluorescence in situ hybridization (FISH). Each clone was selected to contain a genetically mapped polymorphic DNA sequence as close as possible to the centromere or telomere of a chromosome. The genetic distance from the various P1 clones to the most distal genetically mapped polymorphic sequence ranged from 0 for about half of the clones to 6.7 cM for the probe at the telomere of chromosome 14. The average distance to the most distal or proximal chromosome marker was 1.5 cM. The use of FISH with these probes for mouse chromosome identification during comparative genomic hybridization is illustrated.

Detection of p16 gene deletions in gliomas: a comparison of fluorescence in situ hybridization (FISH) versus quantitative PCR.

The p16 protein plays a key role in cell cycle control by preventing CDK4 from inactivating the retinoblastoma protein (pRb). The corresponding tumor suppressor gene (p16/MTS1/CDKN2) has recently been implicated in malignant progression of astrocytomas and could potentially serve as an important marker for patient prognosis and for guiding specific therapeutic strategies. We have undertaken a study to evaluate 2 methods of detecting p16 deletion. Thirty diffuse gliomas were analyzed for p16 gene dosage. Dual color fluorescence in situ hybridization (FISH) was performed on cytologic preparations using paired centromeric (CEN) and locus-specific probes for CEN9/p16, CEN8/RB, and CEN12/CDK4. Quantitative PCR was performed using primers for p16, MTAP, and reference genes. Eleven cases were also studied using comparative genomic hybridization (CGH). Abnormalities of the p16-CDK4-RB pathway were identified in 21 (70%) cases by FISH and/or PCR. These included 15 (50%) with p16 deletion, 9 of which were detected by both techniques, 3 by FISH alone, and 3 by PCR alone (concordance rate = 81%). FISH analysis further revealed tetraploidy/aneuploidy in 14 (47%), RB deletion in 11 (37%) and CDK4 amplification in 1 (3.3%). There were 94% and 100% concordance rates between CGH and FISH or PCR, respectively. Quantitative PCR was noninformative in 4 cases. Although FISH and quantitative PCR are both reliable techniques, each has limitations. PCR is likely to miss p16 deletions when there is significant normal cell contamination or clonal heterogeneity, whereas the p16 YAC probe used for FISH analysis may miss small deletions. Replacement of the latter with a cosmid probe may improve the sensitivity of FISH in future experiments.

Targeted expression of MYCN causes neuroblastoma in transgenic mice.

The proto-oncogene MYCN is often amplified in human neuroblastomas. The assumption that the amplification contributes to tumorigenesis has never been tested directly. We have created transgenic mice that overexpress MYCN in neuroectodermal cells and develop neuroblastoma. Analysis of tumors by comparative genomic hybridization revealed gains and losses of at least seven chromosomal regions, all of which are syntenic with comparable abnormalities detected in human neuroblastomas. In addition, we have shown that increases in MYCN dosage or deficiencies in either of the tumor suppressor genes NF1 or RB1 can augment tumorigenesis by the transgene. Our results provide direct evidence that MYCN can contribute to the genesis of neuroblastoma, suggest that the genetic events involved in the genesis of neuroblastoma can be tumorigenic in more than one chronological sequence, and offer a model for further study of the pathogenesis and therapy of neuroblastoma.

Gain of chromosome 17 is the most frequent abnormality detected in neuroblastoma by comparative genomic hybridization.

Neuroblastoma behavior is variable and outcome partially depends on genetic factors. However, tumors that lack high-risk factors such as MYCN amplification or 1p deletion may progress, possibly due to other genetic aberrations. Comparative genomic hybridization summarizes DNA copy number abnormalities in a tumor by mapping them to their positions on normal metaphase chromosomes. We analyzed 29 tumors from nearly equal proportions of children with stage I, II, III, IV, and IV-S disease by comparative genomic hybridization. We found two classes of copy number abnormalities: whole chromosome and partial chromosome. Whole chromosome losses were frequent at 11, 14, and X. The most frequent partial chromosome losses were on 1p and 11q. Gains were most frequent on chromosome 17 (72% of cases). The two patterns of gain for this chromosome were whole 17 gain and 17q gain, with 17q21-qter as a minimal common region of gain. Other common gains were on chromosomes 7, 6, and 18. High level amplifications were detected at 2p23-25 (MYCN region), at 4q33-35, and at 6p11-22. Chromosome 17q gains were associated with 1p and/or 11q deletions and advanced stage. The high frequency of chromosome 17 gain and its association with bad prognostic factors suggest an important role for this chromosome in the development of neuroblastoma.