Portal de Periódicos da CAPES

Targeted expression of MYCN causes neuroblastoma in transgenic mice

1997; Springer Nature; Volume: 16; Issue: 11 Linguagem: Inglês

10.1093/emboj/16.11.2985

1460-2075

William A. Weiss, Kenneth Aldape, Gayatry Mohapatra, Burt G. Feuerstein, J. Michael Bishop,

Virus-based gene therapy research

Article1 June 1997free access Targeted expression of MYCN causes neuroblastoma in transgenic mice William A. Weiss Corresponding Author William A. Weiss G.W.Hooper Foundation, San Francisco, CA, 94143-0552 USA Departments of Neurology (Division of Child Neurology), San Francisco, CA, 94143-0552 USA Search for more papers by this author Ken Aldape Ken Aldape Pathology (Neuropathology), San Francisco, CA, 94143-0552 USA Search for more papers by this author Gayatry Mohapatra Gayatry Mohapatra Laboratory Medicine, San Francisco, CA, 94143-0552 USA Neurological Surgery (Brain Tumor Research Center), San Francisco, CA, 94143-0552 USA Search for more papers by this author Burt G. Feuerstein Burt G. Feuerstein Laboratory Medicine, San Francisco, CA, 94143-0552 USA Neurological Surgery (Brain Tumor Research Center), San Francisco, CA, 94143-0552 USA Search for more papers by this author J.Michael Bishop J.Michael Bishop G.W.Hooper Foundation, San Francisco, CA, 94143-0552 USA Microbiology and Immunology, University of California, San Francisco, CA, 94143-0552 USA Search for more papers by this author William A. Weiss Corresponding Author William A. Weiss G.W.Hooper Foundation, San Francisco, CA, 94143-0552 USA Departments of Neurology (Division of Child Neurology), San Francisco, CA, 94143-0552 USA Search for more papers by this author Ken Aldape Ken Aldape Pathology (Neuropathology), San Francisco, CA, 94143-0552 USA Search for more papers by this author Gayatry Mohapatra Gayatry Mohapatra Laboratory Medicine, San Francisco, CA, 94143-0552 USA Neurological Surgery (Brain Tumor Research Center), San Francisco, CA, 94143-0552 USA Search for more papers by this author Burt G. Feuerstein Burt G. Feuerstein Laboratory Medicine, San Francisco, CA, 94143-0552 USA Neurological Surgery (Brain Tumor Research Center), San Francisco, CA, 94143-0552 USA Search for more papers by this author J.Michael Bishop J.Michael Bishop G.W.Hooper Foundation, San Francisco, CA, 94143-0552 USA Microbiology and Immunology, University of California, San Francisco, CA, 94143-0552 USA Search for more papers by this author Author Information William A. Weiss 1,2, Ken Aldape3, Gayatry Mohapatra4,5, Burt G. Feuerstein4,5 and J.Michael Bishop1,6 1G.W.Hooper Foundation, San Francisco, CA, 94143-0552 USA 2Departments of Neurology (Division of Child Neurology), San Francisco, CA, 94143-0552 USA 3Pathology (Neuropathology), San Francisco, CA, 94143-0552 USA 4Laboratory Medicine, San Francisco, CA, 94143-0552 USA 5Neurological Surgery (Brain Tumor Research Center), San Francisco, CA, 94143-0552 USA 6Microbiology and Immunology, University of California, San Francisco, CA, 94143-0552 USA The EMBO Journal (1997)16:2985-2995https://doi.org/10.1093/emboj/16.11.2985 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info 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. Introduction Neuroblastoma is a tumor of neural crest origin that comprises 8-10% of all childhood malignancies and causes 15% of cancer-related deaths in children. Ninety percent of cases strike children <10 years old (Matthay, 1995). The tumor arises anywhere in the sympathetic nervous system, most frequently in the adrenal medulla and in paraspinal ganglia (Brodeur and Castleberry, 1993). The best characterized genetic abnormality in neuroblastoma is amplification of MYCN, a gene originally isolated from neuroblastoma cells (Kohl et al., 1983; Schwab et al., 1983). Amplification of MYCN occurs in approximately one-third of neuroblastomas and correlates with advanced disease, suggesting that amplification of MYCN is a late event in tumorigenesis (Brodeur et al., 1984; Kohl et al., 1984). An alternative interpretation of these data is that MYCN-dependent and MYCN-independent pathways exist early in neuronal transformation, and that the MYCN-dependent transformation leads to more aggressive tumors with a worse prognosis. In either event, it is not known whether the genetic events that give rise to neuroblastoma must occur in a particular chronological sequence. The evidence that MYCN participates in the genesis of neuroblastoma has until now been entirely circumstantial. However, there are experimental demonstrations that the gene does have oncogenic potential. For example, overexpression of MYCN can cooperate with a mutationally activated H-RAS gene to transform rat embryo cells (Schwab et al., 1985) and can convert established rat cell lines to tumorigenicity (Small et al., 1987). Transgenic mice that overexpress MYCN under control of the immunoglobulin heavy chain enhancer develop lymphoid tumors (Dildrop et al., 1989; Rosenbaum et al., 1989). These data suggest that MYCN has a role in modulating cellular proliferation, but the evidence that MYCN actually contributes to neuroblast transformation remains inferential. In an effort to explore the role of MYCN in neuroblastoma, we created transgenic mice that overexpress MYCN in cells derived from the neural crest. Such animals develop neuroblastoma several months after birth, demonstrating that overexpression of MYCN can initiate tumorigenesis. The latent period prior to neuroblastoma development suggests that additional genetic events contribute to tumor formation. By comparative genomic hybridization (CGH), neuroblastomas from the transgenic mice demonstrate a number of consistent chromosomal gains and losses in regions syntenic with those observed in human neuroblastoma. These mice represent a model for neuroblastoma that may be useful both to identify additional genes involved in tumorigenesis, and to test new therapies in preclinical trials. Results Targeting MYCN overexpression to the neural crest During development, neuroblasts appear as a normal constituent of the neural crest. To target expression of human MYCN to these cells, we used the tyrosine hydroxylase promotor (Figure 1). This promotor is active in migrating cells of the neural crest early in development (Banerjee et al., 1992), and has been shown to direct expression of chloramphenicol acetyl transferase to sympathetic ganglia and the adrenal, in which neuroblastomas often arise (Banerjee et al., 1992). As hoped, mice carrying the transgenic construct showed tissue-specific expression of the MYCN protein product (Mycn). By Western analysis, Mycn was expressed at high levels in the adrenal gland, which is derived from the neural crest (Figure 2a). There was marginal expression in brain, heart, testes and spleen (data not shown). Figure 1.Construct used for generating tissue-specific expression of the MYCN transgene. A cDNA for human MYCN was ligated downstream of the rat tyrosine hydroxylase promotor. The rabbit β-globin enhancer was used to enhance expression, and a herpes simplex virus thymidine kinase gene sequence was used as a transcription terminator. The transgene was cleaved from vector sequences using NsiI. E, EcoRI; S, SalI; N, NsiI. Download figure Download PowerPoint Figure 2.Expression of the MYCN transgene. (a) Western blot analysis of total protein from various transgenic organs. Protein extracts prepared from equal amounts of organ tissues were immunoprecipitated, fractionated by PAGE and hybridized with an anti-Mycn monoclonal antibody (Boehringer Mannheim, Indianapolis, IN). A, adrenal; B, brain; H, heart; I, intestine; K, kidney; Li, liver; Lu, lung; M, muscle; Sk, skin; Sp, spleen; T, testes. The positive control was from Kelly cells (Schwab et al., 1983) and represented approximately one-quarter of the protein loaded in other lanes. The level of Mycn expressed in the adrenal gland varied among mice in the same line. (b) Mycn protein levels in tumor samples. Equal quantities of protein from tumor and control tissues were subjected to SDS-PAGE, transferred to nitrocellulose membranes and probed with a Mycn monoclonal antibody. The positive control was from Kelly cells. The negative control was kidney tissue from a transgenic mouse. Hemizygous and homozygous refer to mice with one or two alleles of the MYCN transgene insertion. Download figure Download PowerPoint Neuroblastomas in mice overexpressing the MYCN transgene Three independent MYCN transgenic lines developed thoracic paraspinous masses (Table I). The histology of these tumors was consistent with neuroblastoma. Tumors showed varying degrees of neuronal differentiation, as evidenced by ganglionic cells and neuropile (Figure 3). We do not believe that the presence of ganglionic cells can be explained by the adventitious inclusion of normal tissue in the tumors: first, ganglionic tumors were often found in regions where ganglia are not known to reside; and second, many ganglionic tumors showed signs of proliferation, such as mitotic figures and double nuclei (data not shown). Figure 3a shows a typical small round blue cell tumor adjacent to a paraspinous ganglion, Figure 3b a tumor with both small round blue cells and more differentiated elements. Figure 3.Histopathology of tumors in MYCN transgenic mice. Tissues were processed for staining and electron microscopy as detailed in Materials and methods. Sections are shown at high (×250) power. (a) Anaplastic tumor. Arrows denote the normal paraspinous ganglion in a section of the tumor. (b) Tumor showing a broader spectrum of differentiation. The short arrow denotes more primitive small round blue cells, while the longer arrow denotes malignant cells with a more ganglionic appearance. (c) Synaptophysin staining of a tumor section. (d) Neuron-specific enolase staining of a tumor section. (e and f) Neuronal elements seen on electron microscopy of tumor sections (×25 000). The arrow in (e) denotes synapse formation in a tumor section. The arrows in (f) denote neurosecretory granules. Download figure Download PowerPoint Table 1. Transgenic lines produced with the tyrosine hydroxylase-MYCN transgene Mouse strain No. of founders Lines established Lines with tumors C57B6/J×Balb/c 8 3a 2b FVB/N 9 6c 1d C57B6/J 7 3e 1f a All of these lines overexpressed MYCN in adrenal tissue. b These mice developed thoracic paraspinous masses with the histology of neuroblastoma. One died as a sterile founder, and the other occurred in a line that expresses MYCN, and transmits both the transgene and tumor susceptibility. c Three of these six lines overexpressed MYCN. d This mouse died of a thoracic paraspinous mass with the histology of neuroblastoma. No additional tumors have been seen in this line. e One of these three lines overexpressed MYCN. f This sterile founder died of a brain tumor with neuronal elements as evidenced by ganglionic differentiation and positive cell staining for synaptophysin and neuron-specific enolase. The histology was consistent with that of a primitive neuro-ectodermal tumor with neuronal differentiation (cerebral neuroblastoma). Tumors stained positively for the neuronal markers synaptophysin and neuron-specific enolase (Figure 3c and d). Tumor cells with ganglionic differentiation were positive for the marker S100 (Nakagawara et al., 1986), while anaplastic tumor cells remained negative. Tumor cells did not react with antibodies for chromogranin A and O13 (data not shown). Electron microscopy of tumor tissue demonstrated synapse formation (Figure 3e) and neurosecretory granules (Figure 3f). Western blotting showed that in some tumors, the levels of the Mycn were equivalent to those found in the human neuroblastoma cell line Kelly (Figure 2b) which has high levels of MYCN amplification (Schwab et al., 1983). The relative amounts of Mycn varied from one tumor to another, but in every instance they far exceeded those found in normal tissue(s). One line of mice has developed tumors in multiple animals. This line had ∼4 copies of the MYCN transgene by Southern analysis (data not shown). Mice from this line were back-crossed twice to C57B6/J, and 50 offspring were followed for tumor development. Fourteen of these animals developed tumors, with an incidence of ∼5% at 3 months and 20% at 6 months (Figure 4a). Six animals had large abdominal tumors incorporating the adrenal glands, kidney, intestine and other organs. Six animals had thoracic paraspinous tumors and two had both abdominal and thoracic tumors. Figure 4.Survival of mice carrying the MYCN transgene. Wild-type mice had 100% tumor-free survival in these experiments (data not shown). (a) Survival of mice hemizygous and homozygous for the MYCN transgene. The outbred founder line (C57BL6/J×Balb/c) was back-crossed twice to C57BL6/J. Fifty such mice were followed (⋄). F1 offspring from the above cross were interbred and 13 offspring homozygous for the MYCN transgene were followed (□). (b) Survival of MYCN transgenic mice carrying hemizygous inactivating mutations in p53, NF1 and RB1. The outbred founder line was back-crossed once to C57BL6/J, and then to a p53+/− mouse in an FVB/N strain background. Twenty three mice hemizygous for the MYCN transgene and heterozygous for the p53 inactivating mutation were followed (⊞). Twenty seven littermates transgenic for MYCN but wild-type at p53 were followed to control fo strain effects (▵). The outbred founder line was back-crossed once to C57BL6/J, then crossed to a 129S/v mouse heterozygous for NF1 or RB1 inactivating mutations. Fifteen MYCN transgenics heterozygous for the NF1 mutation (⋄), 20 MYCN transgenic mice heterozygous for the RB1 mutation (□) and 25 MYCN transgenic littermates wild-type at these loci (○) were followed for 1 year. Download figure Download PowerPoint Clinical presentations varied among animals. Some animals became emaciated. These had either thoracic tumors (Figure 5a) or small abdominal masses. Others developed very large palpable abdominal masses (Figure 5b) or a progressive paralysis of the lower extremities. The paralyzed mice had either abdominal or thoracic paraspinous masses, with tumor tracking along the peripheral nerve and encasing the spinal cord (Figure 6). Figure 5.Tumors induced by the MYCN transgene. (a) Thoracic paraspinous mass. Mice with thoracic paraspinous masses became emaciated, dyspneic and hunched. (b) Abdominal mass. Mice with abdominal masses appeared otherwise normal. Download figure Download PowerPoint Figure 6.Tumor tracking along radicular nerves and encasing the spinal cord. A number of mice developed progressive paraparasis of the lower extremities. These animals had paraspinous masses that infiltrated neural foramina and encased the spinal cord. The tumor and spinal column were fixed, decalcified, sectioned and stained with hematoxylin and eosin. (a) Low power micrograph of the spinal column and tumor in a decalcified axial photograph. Note the tracking of the tumor around the spinal cord in the center of the bony canal. (b) High power view (×250) of the boxed region in (a). The thin arrow denotes the peripheral nerve. The medium arrow denotes the dorsal root ganglion. The thick arrow shows the tumor. Download figure Download PowerPoint Metastatic spread of mouse neuroblastoma Occasional animals had gross metastases to liver, lung or ovary. This observation prompted us to analyze organ histology from eight animals that died of neuroblastoma, but in which no gross metastases were seen. Five such animals showed microscopic metastases to liver, lung, lymphatics, kidney, ovary, testes, brain and muscle. Selected examples are shown in Figure 7a and b. No mice showed metastases to cortical bone, and a single mouse had bone marrow involvement. Figure 7.Metastatic spread of tumor. (a) Hematogenous spread to the kidney. Thin arrows denote the tumor in renal venules. The thick arrow shows the renal glomerulus. The structure at the upper left is a renal arteriole. (b) Lymphatic spread to regional lymph nodes. The top of the structure shows the tumor effacing the normal lymph node architecture. The lower area of the photograph shows the normal lymphatic ultrastructure. Download figure Download PowerPoint Tumor formation is dependent on MYCN gene dosage To determine whether transgene dosage is related to tumor incidence, transgenic mice were crossed to produce homozygosity for the MYCN transgene. Homozygotes displayed both increased incidence and decreased latency of tumor formation, approaching 100% at 4 months (Figure 4a). To test whether γ irradiation could increase tumor penetrance, mice hemizygous for the MYCN transgene were irradiated at 3-4 Gy before day 3 of life. No increase in tumor incidence was seen in these animals. Additional genetic lesions contribute to tumorigenesis In the limited studies performed to date, loss of NF1 occurs commonly, and loss of RB1 rarely in human neuroblastoma cell lines or tumor tissue (Nakamura et al., 1991; Johnson et al., 1993; The et al., 1993). To test whether loss of these tumor suppressor genes could increase tumor penetrance, mice lacking NF1 and RB1 were crossed to mice that overexpressed MYCN. Mice homozygous for NF1 and RB1 deletions die at midgestation (Jacks et al., 1992, 1994). In order to avoid this complication, mice were generated that were hemizygous for the MYCN transgene and heterozygous for deletions of NF1 or RB1. Such mice had a decreased latency, and an increased incidence of tumors, to ∼75% by 10 months (Figure 4b). The NF1 and RB1 mutant mice were in a 129S/v strain background, so littermates wild-type at the tumor suppressor loci and carrying the MYCN transgene were also followed to control for strain effects. Littermates had a tumor penetrance of 40% at 10 months (Figure 4b). No neuroblastomas were seen in mice heterozygous for deletions of RB1 or NF1, but not carrying the MYCN transgene. These data suggest that loss of either NF1 or RB1 can contribute to tumorigenesis in mice overexpressing MYCN. Structural mutations in p53 are rare in human neuroblastoma (Imamura et al., 1993; Hosoi et al., 1994). To test the role of p53 in mouse neuroblastoma, we crossed the MYCN transgenic to an FVB/N mouse heterozygous for a p53 deletion. Mice hemizygous for the MYCN transgene and heterozygous for deletion of p53 showed a tumor incidence similar to that of littermates hemizygous for the MYCN transgene and wild-type at the p53 locus (Figure 4b). The tumor penetrance in mice hemizygous for the MYCN transgene and wild-type at the p53 locus was less than that seen in similar mice back-crossed to a C57BL6/J background (Figure 4), even though these mice were back-crossed to FVB/N only once. MYCN transgenic mice lacking both p53 alleles were also generated, and did not show increased tumor penetrance (data not shown). Analysis of tumors by comparative genomic hybridization To examine genetic aberrations at the chromosomal level, we used CGH to analyze 21 tumors from hemizygous and homozygous mice. This technique is based on two-color fluorescence in situ hybridization and scans the entire genome for gains and losses of chromosomal material. Tumors from five homozygous mice and three hemizygous mouse showed no chromosomal changes. Thirteen tumors (all from hemizygous mice) showed gains and losses of chromosomal regions (Figure 8, Table II). In mice hemizygous for the MYCN transgene, chromosomal regions were most commonly gained on chromosomes 11 (5/16) and 17 (6/16). These regions are syntenic with human chromosomes 6 and 17 respectively (Copeland et al., 1993), both of which commonly are gained in human neuroblastoma (Plantaz et al., 1997). Chromosomal loss was most often detected on chromosomes 5 (4/16), 9 (3/16), 16 (4/16) and X (4/16). These chromosomes are syntenic with human chromosomes 4, 11, 3 and X respectively (Copeland et al., 1993), all of which commonly are lost in human neuroblastoma (Plantaz et al., 1997). Figure 8.Chromosomal changes in tumors analyzed by comparative genomic hybridization. Tumors were analyzed as in Materials and methods. Five tumors showed no chromosomal changes. Download figure Download PowerPoint Table 2. Comparative genomic hybridization results from individual tumors Tumora Copy number change by CGH 1b +11(A2−B3), +15(E−F3) 2b +3, +11, +Y 3b −5, −9, −16 4b −4(A2−D3), −5, −9, −16 5c −5(A2−E5), +11(B1−E2), −16 6b −3(C−F1) 7d +17, −X 8b +15, +17, +18, −X(D−F3) 9c +11 (B2−E), −X(D−F3) 10d +1, +2, +3, −4(A4−C7), −4(E), −5(A2−E2), −9, −10, +11, −13, −14, −15(A2−D3), −16, +17, −18, −X (A7−F) 11d +17(A−D) 12d −3, +17, +18 13d +12, +17 a Eight tumors, including all five tested from mice homozygous for the MYCN transgene, had no changes by CGH. b Undifferentiated tumors. c Tumors showing mixed differentiation. d Tumors with predominantly ganglionic histology. Gain of chromosome 17 also correlated with lack of ganglionic differentiation. Five tumors showed predominantly ganglionic differentiation, and all showed gain of chromosome 17. Six tumors were predominantly poorly differentiated, and only one of these showed gain on chromosome 17. Two tumors were intermediate in differentiation, and showed no abnormalities of chromosome 17. Discussion Overexpression of MYCN in the neuroectoderm predisposes mice to neuroblastoma Our data demonstrate that targeted expression of MYCN causes neuroblastoma in transgenic mice and represent the first demonstration that MYCN can contribute to the transformation of neuroblasts in vivo. This murine model mirrors human neuroblastoma in that: (i) tumors arise in appropriate locations (abdomen and thorax); (ii) mice may present with spinal cord involvement; (iii) tumorigenesis is affected by MYCN gene dosage; (iv) tumors have the characteristic histology of human neuroblastoma (small round blue cells, ganglionic differentiation and neuropile); (v) tumor cells stain positively for synaptophysin and neuron-specific enolase; (vi) tumors form synapses and contain neurosecretory granules; and (vii) tumors show gains and losses of chromosomes in regions syntenic with those observed in human neuroblastoma. MYCN is amplified in one-third of human neuroblastomas and is associated with an unfavorable prognosis. These observations suggest that MYCN acts late in tumor progression. Alternatively, MYCN amplification may occur early in a subset of neuroblastomas, generating more aggressive tumors than those transformed through MYCN-independent pathways. It is unlikely, however, that amplification of MYCN is the initial event in tumorigenesis, since amplification is alleged not to occur in normal mammalian cells (Tlsty, 1990; Wright et al., 1990). Thus, it is probable that early events in the genesis of neuroblastoma promote genetic instability and allow cells to tolerate such instability. Genomically unstable cells can then be transformed by lesions such as MYCN amplification in conjunction with other genetic events. By overexpressing MYCN in a transgenic mouse, we bypass the need for destabilization of the genome which is a likely prerequisite for MYCN amplification, while maintaining a requirement for other genetic events that contribute to neuroblastic transformation. For some human tumors, one or more genetic lesions typically occur early during the course of tumorigenesis, whereas others occur later (Kinzler and Vogelstein, 1996). Must the multiple mutations culminating in malignancy occur in a particular order? The data presented here demonstrate that overexpression of MYCN, although unlikely to be the initial event in the genesis of human neuroblastoma, can in fact initiate the sequence of events that lead to neuroblastoma in mice. Additional genetic lesions contribute to tumorigenesis The prolonged latent period for tumor formation and the additional chromosomal lesions seen by CGH both argue that tumor formation in the mouse, as in humans, requires genetic lesions in addition to overexpression of MYCN. Five mice homozygous for the MYCN transgene had no chromosomal gains or losses seen by CGH, suggesting that neuroblasts with increased MYCN gene dosage may require fewer additional genetic events for transformation than those with lower levels of MYCN. Three of 16 mice hemizygous for the MYCN transgene also showed no chromosomal copy number abnormalities. This number is comparable with 4/29 human neuroblastomas examined by CGH which also showed no abnormalities (Plantaz et al., 1997). It has been suggested that the amplicon carrying MCYN in neuroblastomas may contain other genes that also contribute to tumorigenesis (Manohar et al., 1995). On the other hand, a survey of amplified DNA in human neuroblastomas found that MYCN may be the only genetic element held in common among the amplicons in the various tumors (Reiter and Brodeur, 1996). In the mouse model presented here, we encountered one tumor that was trisomic for chromosome 12 (which carries mycn). Otherwise, we found no evidence for amplification of endogenous mycn (which could serve as an indicator for amplification of linked genes as well). In mice hemizygous for the MYCN transgene, the genetic losses that occurred most often were of chromosomes 5 (4/16 mice), 9 (3/16 mice), 16 (4/16 mice) and X (4/16 mice). The involved chromosomes are syntenic with human chromosomes 4 (lost in 7/29 human neuroblastoma samples), 11 (lost in 11/29 human tumors), 3 (lost in 8/29 human tumors) and X (lost in 10/29 human tumors) respectively (Plantaz et al., 1997). The chromosomal region 1p36.1-36.2 may carry a tumor suppressor gene involved in the genesis of human neuroblastoma (White et al., 1995). Loss of this region may correlate with amplification of MYCN (Fong et al., 1992; Caron et al., 1993), although this view is controversial (Weith et al., 1989). In contrast, CGH detected loss of the syntenic region of chromosome 4 in only two of the 16 mouse tumors studied here. We cannot presently account for this apparent discrepancy. It is possible that lesions will be found when the syntenic region is studied at higher resolution. Alternatively, providing an excess of Mycn early in tumor progression may alter the requirements for subsequent genetic events. The genetic gains that occurred most often were of chromosome 17 (in 6/16 mice) and chromosome 11 (in 5/16 mice). The chromosomes involved are syntenic with human chromosomes 6 (gained in 10/29 human neuroblastoma samples) and 17 (gained in 21/29 human tumors) respectively (Plantaz et al., 1997). Gain of mouse chromosome 17 also correlated with ganglionic differentiation in 5/6 tumors showing predominantly ganglionic differentiation by histology. In human studies, gain of chromosome 17 was the most commonly found genetic abnormality, and the minimal region of chromosome 17 which showed gain was 17q21.3-qter (Plantaz et al., 1997). In the data presented here, the five tumors showing gain of mouse chromosome 11 shared an area defined by the much smaller region between bands B2 and B3. Much of this mouse interval is syntenic with the short arm of human chromosome 17, and only a very limited region is syntenic with 17q21.3-ter (Mouse Genome Database, 1996). The region of chromosome 11 gained in mouse tumors potentially provides a better definition of the interval gained in human neuroblastoma. Experiments to better characterize this region by fluorescence in situ hybridization are in progress. It is intriguing that the chromosomal regions most commonly affected in this mouse tumor sample are all syntenic to those affected in human tumors. These data suggest that the genetic pathways contributing to mouse neuroblastoma are similar to those observed in the human disease. In an effort to identify specific genetic lesions that might contribute to the genesis of neuroblastoma in the transgenic mice, we explored the effects of deficiencies in the tumor suppressor genes NF1, RB1 and p53. We found that heterozygous deficiencies in either NF1 or RB1 augmented tumorigenesis. While we cannot exclude the possibility that increased tumorigenesis results from haploinsufficiency for NF1 and RB1, we presume that spontaneous damage inactivated or eliminated the remaining alleles of the genes. In contrast, even homozygous defects in p53 had no apparent effect on tumorigenesis. These results are in at l