Hsr Cytogenetics

20q13 (EDN3)

has been described (Verloes et al. 1993). In addition, there have been reports of the coexistence of neuroblastoma and neurofibromatosis type 1, including the coincidence of these disorders in familial neuroblastoma (Maris et al. 1997,2002). Indeed, homozygous inactivation of the NF1 gene in primary neuroblastomas has also been described (Origone et al. 2003; Martinsson et al. 1997). These data suggest that the genes implicated in the genesis of Hirschsprung disease (RET, EDNRB, EDN3, GDNF, ECE1, and ZFHX1B), central hypoventilation (RET, GDNF, EDN3, BDNF, and PHOX2B) and/or NF1 may be causally involved in the initiation or progression of human neuroblastoma, especially in the context of a neurocristopathy (Table 3.1). A recent study reported a germline mutation in PHOX2B in a patient with neuroblastoma (Amiel et al. 2003), although previous reports have found no evidence for linkage at the 4p12 PHOX2B locus (Maris et al. 2002). GDNF and related molecules, neurturin (NRTN), artemin (ARTN), and persephin (PSPN), signal through a unique multicomponent receptor system consisting of RET tyrosine kinase and glycosyl-phos-phatidylinositol-anchored coreceptor (GFRalpha1-4) (Sariola and Saarma 2003; Takahashi 2001); however, other than RET and GDNF, mutations in the genes encoding these ligands and coreceptors have not yet been implicated in the pathogenesis of HD or CCHS.

3.3 Constitutional Chromosomal Abnormalities

Discovery of cancer predisposition genes has been facilitated by the identification of rare patients with constitutional genomic DNA aberrations. Although neuroblastoma patients with de novo karyotypic abnormalities are rare, detailed analyses of these cases have been informative. Satge and colleagues recently reviewed 51 cases of constitutional karyotypic aberrations in neuroblastoma patients and confirmed recurrent constitutional deletions at chromosomal regions 1p36, 2p23, 3q, 11q14-23, and 15q (Satge et al. 2003). High-resolution genetic mapping of some of these deletions has aided in determining the location of putative neuroblastoma suppressor genes at chromosomes 1p36 and 11q14-23 (White et al. 2001; Mosse et al. 2003). The three children with constitutional 1p36 interstitial deletions all had profound neurocognitive deficits and were diagnosed with neuroblastoma during infancy. The constitutional deletions overlap the location of a putative 1p36 tumor suppressor gene (see Chaps. 4 and 5), suggesting that germline absence of a gene within this region may predispose to the development of neuroblas-toma. Constitutional balanced translocations have been identified rarely in neuroblastoma patients, and no common region is apparent. Whole chromosome gains or losses are also rare in neuroblastoma pa tients. Interestingly, there appears to be an excess incidence of neuroblastoma in patients with 45,X (Turner) syndrome and perhaps trisomy 13, whereas trisomy 21 (Down syndrome) appears to be associated with a decreased risk for developing neuroblastoma (Satge et al. 2003; Blatt et al. 1997; Satge et al. 1998).

3.4 Hereditary Neuroblastoma

Neuroblastoma, like retinoblastoma and Wilms' tumor, is an embryonal malignancy that is notable for both a sporadic and hereditary form of the disease. Knudson and Strong provided the first genetic hypothesis of neuroblastoma tumorigenesis in 1972. In a comparison of 29 cases of hereditary neuroblastoma (from 13 families) to 504 unselected cases, they showed that 56% of familial cases were diagnosed at less than lyear of age compared with 26% of the nonfamilial cases. In addition, 23% of the familial cases had multiple primary tumors documented, compared with 5% of the nonfamilial cases. Analysis of the pedigree structures was consistent with an autosomal-dominant mode of inheritance with incomplete penetrance that they calculated to be 0.63 (63% chance carriers will be affected). These data strongly suggested that the genetics of neuroblastoma initiation are similar to retinoblastoma, and subsequent studies indicated that, like RB1, a hereditary neuro-blastoma predisposition gene should be a tumor suppressor. Nevertheless, for many of the reasons listed below, the genetic etiology of neuroblastoma has remained elusive over three decades since Knudson and Strong's original observations.

Despite the unequivocal data supporting a genetic hypothesis for the initiation of neuroblastoma tu-morigenesis, it is relatively uncommon to obtain a positive family history of the disease for an individual neuroblastoma patient. Shojaei-Brosseau and colleagues recently used an epidemiological approach to show that only 5 of 426 consecutive neuroblastoma patients (1.2%) at a single institution had documentation of at least one first- or second-degree relative with neuroblastoma (Shojaei-Brosseau et al. 2004). This translates to a relatively high standardized inci dence ratio (SIR) of 11.4 (95% confidence interval of 3.7-26.5) for the development of neuroblastoma among index-case relatives, but the risk to siblings was estimated at only 0.2%. Patients who present with multiple primary tumors or congenital neurob-lastoma are more likely to harbor a germline mutation in a predisposition gene, but in many cases these may be de novo mutations. Taken together, these data suggest that heritable neuroblastoma is a rare phenomenon, and the pediatric oncologist should reassure parents of any newly diagnosed patient that the risk to siblings (particularly in the absence of high-risk features such as multifocal primary tumors) is very low.

The vast majority of reported neuroblastoma pedigrees are small, and large, multiplex, or three-generation families have been identified only rarely. Analyses of the published pedigrees in the past three decades strongly support the original conclusion of an autosomal-dominant mode of inheritance with incomplete penetrance. Although some families show multiple-affected individuals with few unaffected individuals between generations (obligate carriers), other families show multiple-affected individuals in the same generation (i.e., cousins) with no disease detected in intervening relatives (Maris et al. 2002; Perri et al. 2002; Lemire et al. 1998). Therefore, it is very difficult to determine precisely the penetrance of a mutant hereditary neuroblastoma gene segregating within a family, and this interfamilial heterogeneity may suggest that there is more than one heritable predisposition gene with different likelihoods of initiating neuroblastoma tumorigenesis.

Similar to patients with sporadic neuroblastoma, the clinical course in familial cases is also extremely variable (intrafamilial heterogeneity), with often striking contrast in the ages at presentation, disease stage, biological features of the tumor, and disease outcome. In addition, there are several reports of asymptomatic obligate carriers with elevated urinary catecholamines or in whom clinically occult tumors have been detected (Maris et al. 1997); therefore, reduced penetrance secondary to clinically occult or spontaneously regressing tumors, on the one hand, and the lethality of the condition prior to reproductive age, on the other, may both contribute to the

Neuroblastoma Facts

The pedigrees from seven neuroblastoma families with evidence for linkage to chromosome bands 16p12-13. Filled symbols indicate individual affected with neuroblastoma, ganglioneuroblastoma, or ganglioneuroma. Genotyping data are arranged into probable haplo-types based on minimization of recombination events for 16p polymorphic markers listed at bottom right and are displayed for each individual with an available DNA sample. Gray box indicates common haplotype segregating with disease in each family and shows genetic homogeneity at 16p. Arrowheads indicate haplotype lost when LOH was detected in corresponding tumor specimen rarity of familial neuroblastoma. These facts have also contributed to the difficulty in approaching this disease with classic genetic approaches in order to isolate genes that predispose to the development of neuroblastoma when mutated in the germline.

3.5 Genetic Studies of Familial Neuroblastoma

There are two published studies that used classic genetic linkage methods to localize hereditary neuro-blastoma predisposition genes. In a genome-wide search for linkage in seven pedigrees with at least two first-degree relatives affected with neuroblastoma, convincing evidence was discovered that a hereditary neuroblastoma predisposition gene (HNB1) is located on the distal short arm of chromosome 16 (16p12-13; Fig. 3.1) (Maris et al. 2002). Subsequent identification of a three-generation family with seven individuals affected with neuroblastoma appeared to confirm linkage to 16p with a cumulative LOD score of 3.7 (Maris et al. 2003). Loss of heterozygosity has been observed in 13% of sporadic neuroblastomas, suggesting that somatic inactivation of a 16p tumor suppressor gene might contribute to neuroblastoma initiation or progression in at least a subset of non-familial cases (Furuta et al. 2000). The genomic region likely to harbor HNB1 remains relatively large and the positional cloning of this gene is ongoing.

Perri and colleagues studied two families in which > third-degree relatives (cousins) were affected with neuroblastoma. They showed no evidence for linkage to 16p (Perri et al. 2002), in agreement with the original 16p linkage report in which two families consisting of cousins with neuroblastoma also showed no evidence of linkage to 16p (Maris et al. 2002). However, using a candidate-locus approach, they did show evidence for linkage to the distal short arm of chromosome 4p that overlapped a common region of hemizygous deletion observed in some primary neuroblastomas (Perri et al. 2002). Of note,the seven families linked to 16p showed strong evidence refuting linkage to 4p (J.M. Maris, unpublished data). Taken to-gether,these observations support the hypothesis that at least two hereditary neuroblastoma predisposition genes exist, and that the penetrance of the two predisposition genes is different. The literature also strongly suggests that each of the major candidate loci and/or genes listed in Table 3.1 have been excluded as harboring a hereditary neuroblastoma predisposition gene through candidate-locus and/or genome-wide analyses (Maris et al. 1996,2002; Tonini et al. 2001).

3.6 Conclusions

The rare neuroblastoma patients with a family history of the disease, associated genetic disorder, and/ or constitutional chromosomal abnormality offer unique insights into the molecular pathogenesis of this enigmatic tumor. The identification of at least two putative familial neuroblastoma predisposition loci supports the assumption that neuroblastoma is a complex disease genetically, with multiple pathways to tumor initiation. Although identification of hereditary neuroblastoma predisposition genes would be of immediate benefit to those rare families that show evidence for predisposition to the disease, it is likely that the larger impact will be drawn from the insights these discoveries will provide for neuroblastoma tumorigenesis in general.

Acknowledgements. This work was supported in part by National Institutes of Health grants CA78545 and CA87847 (J.M.M.), and by NIH grants CA39771 and CA94194, and by the Audrey E. Evans Endowed Chair (G.M.B.).

References

Amiel J, Laudier B, Attie-Bitach T, Trang H, de Pontual L, Gener B et al. (2003) Polyalanine expansion and frameshift mutations of the paired-like homeobox gene PHOX2B in congenital central hypoventilation syndrome. Nat Genet 33: 459-461

Blatt J, Olshan AF, Lee PA, Ross JL (1997) Neuroblastoma and related tumors in Turner's syndrome (published erratum appears in J Pediatr 1998,133:312). J Pediatr 131:666-670 Dodge HJ, Benner MC (1945) Neuroblastoma of the adrenal medulla in siblings. Rocky Mt Med 42:35-38 Furuta S, Ohira M, Machida T, Hamano S, Nakagawara A (2000) Analysis of loss of heterozygosity at 16p12-p13 (familial neuroblastoma locus) in 470 neuroblastomas including both sporadic and mass screening tumors. Med Pediatr Oncol 35:531-533 Knudson AG Jr, Strong LC (1972) Mutation and cancer: neu-roblastoma and pheochromocytoma. Am J Hum Genet

24:514-532

Kushner BH, Gilbert F, Helson L (1986) Familial neuroblas-toma. Case reports, literature review, and etiologic considerations. Cancer 57:1887-1893 Lemire EG, Chodirker BN, Williams GJ, Seargeant LE, Israels SJ, Phillips SM et al (1998) Familial neuroblastoma: report of a kindred with later age at presentation. J Pediatr Hematol Oncol 20:489-493 Maris JM, Brodeur GM (2001) Genetics of neuroblastoma. In: Cowell JK (ed) Molecular genetics of cancer, 2nd edn. BIOS, Oxford, pp 335-361 Maris JM, Kyemba SM, Rebbeck TR, White PS, Sulman EP, Jensen SJ et al. (1996) Familial predisposition to neuroblastoma does not map to chromosome band 1p36. Cancer Res

56:3421-3425

Maris JM, Chatten J, Meadows AT, Biegel JA, Brodeur GM (1997) Familial neuroblastoma: a three generation pedigree and a further association with Hirschsprung disease. Med Pediatr Oncol 28:1-5

Maris JM, Weiss MJ, Mosse Y, Hii G, Guo C, White PS et al. (2002) Evidence for a hereditary neuroblastoma predisposition locus at chromosome 16p12-13. Cancer Res 62:66516658

Maris JM, Mosse Y, Hii G, Dahia P (2003) Familial neuroblas-toma: Further evidence for linkage to 16p13 and refinement of the HNB1 locus. Med Pediatr Oncol 41:293

Martinsson T, Sjoberg RM, Hedborg F, Kogner P (1997) Ho-mozygous deletion of the neurofibromatosis-1 gene in the tumor of a patient with neuroblastoma. Cancer Genet Cyto-genet 95:183-189 Mosse Y, Greshock J, King A, Khazi D, Weber BL, Maris JM (2003) Identification and high-resolution mapping of a constitutional nq deletion in an infant with multifocal neu-roblastoma. Lancet Oncol 4:769-771 Origone P, Defferrari R, Mazzocco K, Lo Cunsolo C, Bernardi B de, Tonini GP (2003) Homozygous inactivation ofNF1 gene in a patient with familial NF1 and disseminated neuroblas-toma. Am J Med Genet 118A:309-313 Perri P, Longo L, Cusano R, McConville CM, Rees SA, Devoto M et al. (2002) Weak linkage at 4p16 to predisposition for human neuroblastoma. Oncogene 21:8356-8360 Sariola H, Saarma M (2003) Novel functions and signalling pathways for GDNF. J Cell Sci 116:3855-3862 Satge D, Sasco AJ, Carlsen NL, Stiller CA, Rubie H, Hero B et al. (1998) A lack of neuroblastoma in Down syndrome: a study from 11 European countries. Cancer Res 58:448-452 Satge D, Moore SW, Stiller CA, Niggli FK, Pritchard-Jones K, Bown N et al. (2003) Abnormal constitutional karyotypes in patients with neuroblastoma: a report of four new cases and review of 47 others in the literature. Cancer Genet Cyto-genet 147:89-98 Shojaei-Brosseau T, Chompret A, Abel A, de Vathaire F, Raquin MA, Brugieres L et al (2004) Genetic epidemiology of neu-roblastoma: a study of 426 cases at the Institut Gustave-Roussy in France. Pediatr Blood Cancer 42:99-105 Takahashi M (2001) The GDNF/RET signaling pathway and human diseases. Cytokine Growth Factor Rev 12:361-73 Tonini GP, McConville C, Cusano R, Rees SA, Dagnino M, Longo L et al. (2001) Exclusion of candidate genes and chromosomal regions in familial neuroblastoma. Int J Mol Med 7:85-89

Verloes A, Elmer C, Lacombe D, Heinrichs C, Rebuffat E, Demarquez JL et al. (1993) Ondine-Hirschsprung syndrome (Haddad syndrome). Further delineation in two cases and review of the literature. Eur J Pediatr 152:75-77 White PS, Thompson PM, Seifried BA, Sulman EP, Jensen SJ, Guo C et al. (2001) Detailed molecular analysis of 1p36 in neuroblastoma. Med Pediatr Oncol 36:37-41

Molecular Cytogenetics

Manfred Schwab

Contents

4.1 Introduction 27

4.2 Classical Cytogenetics 27

4.3 Oncogene Expression Profiling 28

4.4 "Neuroblastoma Suppressor Genes"

and Loss of Heterozygosity 29

4.4.1 Chromosome 1p Deletion 29

4.4.1.1 One or More "Tumor Suppressor

Gene"Loci in 1p 31

4.4.2.1 Chromosome 11 Deletion and 17q Gain ... 32

4.4.3 LOH of Additional Chromosomes 32

4.4.4 LOH and Tumor Suppressor Genes: an Evasive Connection or Flawed Hypothesis? 32

4.5 Comparative Genomic Hybridization 34

4.6 Tumor Cell Ploidy 34

4.7 Conclusion 35

References 35

4.1 Introduction

It a major tenet in cancer research that alterations in cellular genes lead to the malignant transformation of normal cells. Two major classes of cancer-related genes have been identified: (a) oncogenes,which contribute to cancer "dominantly" by positive modulation of cellular growth; and (b) tumor suppressor genes, which are thought to control normal cellular growth and differentiation and act in a "recessive" negative way, contributing to cancer through functional inactivation. Both sporadic and familial genetic factors contribute to the pathogenesis of most types of cancer and, as reviewed in Chap. 3, a small subset of neuroblastoma cases have an apparent heritable genetic etiology; however, the vast majority of patients appear to develop neuroblastoma through spontaneously acquired somatic events rather than germline aberrations.

This chapter reviews our current understanding of the somatic genetic events that are associated with neuroblastoma pathogenesis and with clinical phe-notype.

4.2 Classical Cytogenetics

In 1965, minute chromatin bodies, now referred to as double minutes (DMs; Fig. 4.1a), were first discovered in neuroblastoma cells (Cox et al. 1965). Subsequently, another novel chromosome abnormality, homogenously staining chromosomal region (HSR; Fig. 4.1b), was identified in human neuroblastoma cells as well as in antifolate-resistant hamster cells (Biedler and Spengler 1976); however, the biological

Figure 4.1 a,b

Cytogenetic manifestations of amplified DNA in human neuroblastoma cells.a Double minutes (DMs). b Homogeneously staining chromosomal region (HSRs; arrowheads)

Figure 4.1 a,b

Cytogenetic manifestations of amplified DNA in human neuroblastoma cells.a Double minutes (DMs). b Homogeneously staining chromosomal region (HSRs; arrowheads)

significance of these cytogenetic aberrations remained unclear for many years. Some investigators speculated that DMs and HSRs were chromosomal manifestations of multiplicated drug-resistance genes; others hypothesized that DMs may inhibit neo-plas-tic growth (Sandberg et al. 1972), or that loss of DMs through fragmentation from an HSR might be associated with the loss of the malignant pheno-type of that cell (Balaban-Malenbaum and Gilbert 1977).

The advent of chromosome banding techniques in 1968 led to the unequivocal identification of all human chromosomes (Caspersson et al. 1968). The first systematic search for neuroblastoma-associated chromosomal alterations dates back to 1977, when Brodeur and co-workers noted the presence of chromosome 1p deletion in a conspicuous number of neuroblastoma cell lines and primary tumors (Brodeur et al. 1977). The high incidence of 1p deletions was confirmed in larger studies (Brodeur et al. 1981; Gilbert et al. 1982), and the authors speculated that this deletion represented the first "hit" in the two-step genetic sequence of tumor development proposed by Knudson (Knudsson 1971; Brodeur et al. 1977,1981; Gilbert et al. 1982).

4.3 Oncogene Expression Profiling

The discovery of retroviral oncogenes (v-onc) and their cellular homologues (c-onc) in the early 1980s (Bishop 1982; Varmus 1982) quickly led to the identification of mutated c-oncs in human cancer cells (Der et al. 1982; Parada et al. 1982; Santos et al. 1982). In many types of cancer, these genes were found to be altered in structure (Groffen et al. 1984; Heisterkamp et al. 1983) or expression (Dalla-Favera et al. 1982; Neel et al. 1982) as a result of non-random chromosomal translocation. The first mRNA expression array (Onco-Array) used v-onc cDNAs spotted on filter membranes (Schwab et al. 1983b) to which a complex reverse-transcribed, radioactively labeled cDNA from neuroblastoma cells was hybridized. This principle technology was the forerunner of more recent large-scale expression array platforms (see Chap. 8). Profiling of neuroblastoma cell lines quickly established the strong expression of a c-onc, seemingly the MYC gene, the cellular homologue of the chicken retroviral gene (Fig. 4.2).

Subsequent DNA analyses quickly established an increased DNA copy number of a gene that was not

Figure 4.2

Expression profiling of oncogenes in neuroblastoma cell line Kelly to detect oncogene overexpression. Oncogene specific DNAs, many as cDNA of retroviral oncogenes, were spotted on a nitrocellulose filter, which subsequently was probed with radioactively labeled cDNA generated by reverse transcription of total polyadenylated RNA extracted from the tumor cells. Under conditions of reduced-stringency hybridization, a strong signal was seen for MYC which, by DNA analysis, turned out to result from the enhanced expression, consequent to DNA amplification,of a MYC-relative,the MYCN gene

Figure 4.2

Expression profiling of oncogenes in neuroblastoma cell line Kelly to detect oncogene overexpression. Oncogene specific DNAs, many as cDNA of retroviral oncogenes, were spotted on a nitrocellulose filter, which subsequently was probed with radioactively labeled cDNA generated by reverse transcription of total polyadenylated RNA extracted from the tumor cells. Under conditions of reduced-stringency hybridization, a strong signal was seen for MYC which, by DNA analysis, turned out to result from the enhanced expression, consequent to DNA amplification,of a MYC-relative,the MYCN gene the authentic MYC gene, but rather a close relative, initially referred to as N-myc (Schwab et al. 1983a; MYCN is the correct human gene nomenclature). Molecular cytogenetic analyses identified DMs and HSRs as the site of amplified MYCN (Schwab et al. 1984). Enhanced expression of the MYCN gene also contributed to tumorigenic cellular growth (Schwab et al. 1985). The activities of the MYCN protein, and the clinical significance of the amplified MYCN gene, have been the subject of previous reviews (Schwab 1998; Schwab et al. 2003). Amplified MYCN has been referred to as the "clinical debut of oncogenes," and because of the strong association between MYCN amplification and poor outcome, determining MYCN status in neuroblastoma tumors prior to initiating therapy is now considered an international clinical standard (Schwab et al. 2003). Array technology can now probe thousands of genes. While this technology is still evolving, large-scale expression profiling of neuroblastoma tumors has already begun (see Chap. 8; Alaminos et al. 2003; Berwanger et al. 2002; Fan et al. 2004; Khan et al. 2001; Mora et al. 2003; Sotiriou et al. 2002).

4.4 "Neuroblastoma Suppressor Genes" and Loss of Heterozygosity

The concept of tumor suppressor genes evolved from seminal observations made while studying retino-blastoma (Knudson 1971). The identification of the molecular pathway of retinoblastoma development by successive inactivation of the two RB1 alleles at a gene locus was seen as a paradigm for tumor suppressor gene inactivation in other human cancers. Statistical analyses indicated that a two-hit genetic pathway, similar to the one identified in retinoblastoma, would lead to the development of neuroblastoma (Knudson 1971). Furthermore,the consistent 1p deletion detected in neuroblastoma tumors suggested that loss of a putative "neuroblastoma suppressor gene" (NSG) on chromosome 1p may represent the first "hit." The second hit was presumed to be a point mutation - or other subtle alterations - of the NSG on the other allele (although the sequence of "hits" can be either way). To identify the candidate NSG, LOH studies on a large number of tumors have been performed to define smallest region of overlapping deletions (SRO).

4.4.1 Chromosome 1p Deletion

Overall, up to 35 % of neuroblastomas have LOH of chromosome 1p (Fong et al. 1989; Maris et al. 2000;

Figure 4.3

Comparison of 1p smallest region of overlapping deletions (SROs) identified by LOH studies in neuroblastoma.A Schwab et al. (1996); B Bauer et al.(2001); CCaron et al.(2001);D Martinsson et al.(1997),Ejeskar et al.(2001);E Maris et al.(2001b); F Hogarty et al.(2000),White et al.(2001). Open boxes at the end of bars represent the first non-deleted marker. Arrows give the distances between markers defining the SRO B (Bauer et al.2001) and between markers bounding the consensus region of deletion. Order of markers is from Bauer et al. (2001) and the UCSC genome browser (http://genome.ucsc.edu) and was confirmed using the NCBI resource UniSTS (http://www.ncbi.nlm.nih.gov)

Takayama et al. 1992; Takita et al. 1995). A large number of molecular analyses in primary tumors has refined the SRO, mainly detecting LOH with polymorphic markers mapped to 1p (Caron et al. 2001; Ejeskar et al. 2001; Fong et al. 1989,1992; Hogarty et al. 2000; Maris et al. 2001a; Martinsson et al. 1995; Schwab et al. 1996;Weith et al. 1989;White et al. 1995,2001). These efforts resulted in an SRO within 1p36 defined proximally by D1S244 and distally by D1S80 (Fig. 4.3) The low incidence of small interstitial deletions within 1p36 has made it difficult to further narrow the SRO, a prerequisite for positional cloning. Furthermore, although several 1p36 rearrangements have been identified in neuroblas-toma cell lines, along with a constitutional translocation t(1;17)(p36.31-36.13;q1 1.2-12) in a pa tient with multifocal neuroblastoma (Laureys et al. 1990), these chromosomal breakpoints are dispersed throughout a large genomic region.

Recently, an SRO was refined to a 1 Mb region within 1p36.3 defined by LOH in a primary tumor that extends distally from D1S214, and by a constitutional deletion between D1S468 and D1S2826 in a patient with neuroblastoma (White et al. 2001). Independently, a smaller candidate region of approximately 1 Mb (between D1S2731 and D1S2666) was mapped to 1p36.3 (Bauer et al. 2001). Both regions appear to overlap in the vicinity of marker D1S214. Neuroblastoma cell line NGP has a translocation t(1;15)(36.2;q24), including a 2-Mb DNA duplication at 1p36.2 (Amler et al. 1995). Although the proximal breakpoint defined by the duplication appears to map outside the 1p36.3 SRO, the distal breakpoint, which maps to between D1S160 and D1S214, probably lies within the 1-Mb SRO. Several new genes mapping near this breakpoint region were identified recently that are currently being further characterized (Amler et al. 2000; K.O. Henrich et al., submitted). A homozygous deletion spanning approximately 500 kb at D1S244 has been reported in two neuroblastoma cell lines (Ohira et al. 2000); however, this homo-zygous deletion is localized proximal to the refined 1 Mb SRO, which would make a single tumor suppressor gene within 1p36.3 unlikely. Also, it has not been established that the two cell lines have been derived from different patients.

A terminal 1p36 deletion syndrome has been described which is associated with mental retardation and craniofacial features (Shaffer and Heilstedt 2001; Shapira et al. 1997). The prevalence of this deletion (1p36.3) is estimated to be 1 in 5000, making it the most common terminal deletion (Shaffer and Heilst-edt 2001). The deletion is distal to D1S228, and in some cases the large deletions include the 1-Mb SRO within 1p36.3 (Wu et al. 1999). To date, 2 patients with terminal 1p36.3 deletion syndrome have developed neuroblastoma (Biegel et al. 1993; White et al. 2001); however, neuroblastoma has not been detected in any of the originally published cases (Wu et al. 1999), suggesting that neuroblastoma is not a common feature of this syndrome. It remains unclear whether some rare patients with 1p36.3 deletion syndrome may have a predisposition to neuroblastoma depending on their specific deleted regions, or whether the two published cases were simply due to coincidence.

4.4.1.1 One or More "Tumor Suppressor Gene" Loci in 1p

Several observations suggest that more than one 1p locus may be affected in neuroblastoma. Outcome has reported to be poorer in patients with tumors that have large 1p deletions than patients with short or interstitial deletions (Takeda et al. 1994). Furthermore, while tumors with large 1p deletions were associated with adverse prognostic factors, such as diploidy or tetraploidy, and amplified MYCN, tumors with small interstitial deletions had DNA content in the triploid range and a high proportion of tumors were detected by mass screening. The existence of two distinct deleted regions was also suggested by LOH at polymorphic loci in clinically identified neuroblastomas (Caron et al. 1995; Schleiermacher et al. 1994). Additional studies have demonstrated that tumors with and without MYCN amplification show different types of SRO (Cheng et al. 1995; Gehring et al. 1995; Caron et al. 1993; Fong et al. 1989). In MYCN-amplified tumors, 1p deletions are very common and are large, always at least including a region from 1p35-1p36 to telomere. In contrast, 1p deletions occur in only 15-20 % of tumors that lack MYCN amplification, and the deletions are consistently smaller and commonly map to 1p36.3; thus, a second tumor suppressor locus inactivated by the 1p deletions in MYCN-non-amplified neuroblastomas has been postulated (Caron et al. 1995; Schleiermacher et al. 1996). This TSG was suggested to be localized at 1p35-36.1, just distal to the deletion border of the smallest 1p deletion found in MYCN-amplified cases (Caron et al. 1995; Spieker et al. 2001). The smallest SRO of the MYCN single-copy tumors is included into the larger SRO of MYCN-amplified tumors, implying that a distal suppressor locus in 1p36.2-3 must also be deleted in MYCN-amplified tumors.

The genomic complexity of the 1p region and the large size of its deletions have made it difficult to identify a neuroblastoma TSG. Although several candidate genes have been proposed, none has been shown to contain tumor-specific mutations, indicating that alternate mechanisms of TSG inactivation, such as epigenetic silencing or haploinsufficiency, may have to be considered. In addition, structural alterations of chromosome 1 have to be evaluated together with coincident genetic changes in other genomic regions, such as amplified MYCN, 17q gain, and diploidy/triploidy.

4.4.2 Deletion of 11q

Cytogenetic analyses have demonstrated the presence of 11q deletions in about 15% of neuroblastoma tumors (Mertens et al. 1997). In LOH studies,11q loss has been detected in 5-32% of the tumors (Takeda et al. 1996). Loss of the whole chromosome 11 appears to be strongly associated with low stage tumors, whereas unbalanced deletion of 11q is predominantly observed in high-stage tumors without amplified MYCN (Guo et al. 1999, 2000; Maris et al. 2001b). Deletion events affecting 11q are predominantly large and terminal. A single region of 2.1 cM within 11q23.3, flanked by markers D11S1340 and D11S1299, was deleted in all tumors with 11q LOH (Guo et al. 1999). Constitutional rearrangements of 11q have been observed in some neuroblastoma patients, including a deletion of 11q23-qter, balanced translocations involving 11q21 and 11q22, and an inversion of 11q21-q23 (Bown et al. 1993; Hecht et al. 1982; Koiffmann et al. 1995). The role of these constitutional changes is not clear, but it has been speculated that disruption of one or more 11q genes may predispose to the development of neuroblas-toma.

4.4.2.1 Chromosome 11 Deletion and 17q Gain

Fluorescence in situ hybridization (FISH) analyses have demonstrated that, after 1p, chromosome arm 11q is the second most common partner for 17q translocations (van Roy et al. 1994). Such translocations, resulting in concurrent loss of distal 11q and gain of 17q, account for approximately half of the 11q deletion cases (Vandesompele et al. 2001); thus, LOH studies assessing the prognostic value of chromosome losses must take into account the 17q status of each individual tumor.

4.4.3 LOH of Additional Chromosomes

Genome-wide surveys at randomly selected loci have revealed several chromosomal regions with LOH including. 9p21 (Marshall et al. 1997), 14q32 (Hoshi et al. 2000; Thompson et al. 2001), and others (Westermann and Schwab 2002). Although numerous investigators have speculated that TSGs may reside in these sites, to date, in spite of laborious efforts, not one neuroblastoma TSG has been identified.

4.4.4 LOH and Tumor Suppressor Genes: an Evasive Connection or Flawed Hypothesis?

There are several possible explanations for the failure to identify a neuroblastoma TSG. Firstly, the two-hit model, in its original form, may not be applicable to neuroblastoma. In addition, the current logic of utilizing LOH studies to determine the SRO and then surveying the chromosomally intact homologue for genes in this region and for mutations may be flawed. One possibility is that the loss of a single allele by deletion may be sufficient to produce a biological effect. Evidence for haploinsufficiency is accumulating (Goss et al. 2002; Gruber et al. 2002; Kucherlapati et al. 2002; Spring et al. 2002; Venkatachalam et al. 1998) for a number of genes, including BLM, Fen1, TP53, ATM, and others (Table 4.1). It is also possible that deletion of a single allele, such as in 1p, alone or in combination with deletion at another genetic locus, may contribute to tumorigenesis simply by dosage effect, without any mutational or epigenetic change of the remaining allele. Evidence is also emerging that slight gene dosage changes,like segmental duplications, can contribute to human malignant and non-malignant disorders (Corvi et al. 1995; Gratacos et al. 2001; Savelyeva et al. 2001). Genetic imbalance for 1p36 (at least 2 copies of chromosome 1 present with additional 1p36-deleted chromosome 1 copies) may also be associated with poor prognosis, similar to that seen in patients with tumors with 1p deletion or amplified MYCN (Spitz et al. 2002).

Another problem could be genetic heterogeneity for particular LOH regions among neuroblastoma subtypes, and thus, the strategy of analyzing the genes of a consensus region deduced from a larger number of tumors is flawed. One scenario, hypothetical but in principle suggested earlier (Takeda et al. 1994), could be that one biological or clinical group (group 1) of tumors results from the inactivation of one gene, while another clinical group of neuroblas-tomas (group 2) depends on the inactivation of another gene. Both genes may be in 1p36,but the group-2 gene may be several megabases away from the group-I gene. When LOH data are combined from these two groups, the SRO will be extremely unlikely to harbor the damaged second allele (Fig. 4.4).

Table 4.1. Genes and haploinsufficiency in tumorigenesis

Gene

Function

Reference

AML1/CBFA2

Transcription factor

Barton and Nucifora (2000); Song et al. (1999)

Cdh1

Cell-cell adhesion

Smits et al. (2000)

Dmpl

Cell-cycle control

Inoue et al.(2001)

Lkb1

Kinase with unknown target

Miyoshi et al.(2002)

10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook


Post a comment