599Workshop W03: Tumours

W04-01 Implication of Neurofibromatosis Type 2(NF2) Gene in Development of Brain TumorSaya, HKumamoto University School of Medicine, Japan

Neurofibromatosis 2 (NF2) is an autosomal domi-nantly inherited disorder which is characterized by apredisposition to multiple intracranial benign tumorsincluding bilateral vestibular schwannomas and menin-giomas. The protein encoded by the NF2 gene, termedmerlin, has striking similarities to ezrin, radixin andmoesin (ERM) proteins which link membrane proteinsto the cytoskeleton. Therefore, it can be speculated thatthe disruption of cytoskeletal organization and/or apop-totic signal by the alternations of the NF2 gene isinvolved in genesis of benign intracranial tumors. Weperformed mutation analysis of NF2 gene in tumors ofboth NF2 and non-NF2 patients, and found that themajority of alterations were nonsense mutations orexon-missing mutations in the ERM-homology domain.Furthermore, we found that merlin interacts withpoly(ADP-ribose) polymerase (PARP), Ku86 and Ku70,all those molecules are involved in the DNA repair, andthat the ERM homology domain of merlin is essentialfor these interactions. These findings suggest that NF2mutations impair the merlin-related complex formation,and it potentially affects genomic integrity. We alsoidentified that merlin is cleaved by the calpain and con-siderable activation of the calpain results in the loss ofmerlin expression in some meningiomas and schwanno-mas which lack detectable NF2 mutations. All thesefindings suggest that loss-of-function of merlin by notonly genetic alterations but also proteolysis induces dis-regulation of the genetic integrity and leads to genera-tion of intracranial benign tumors.

W04-02 Von Hippel-Lindau Disease Maher, EUniversity of Birmingham, UK

Von Hippel-Lindau (VHL) disease is a dominantlyinherited familial cancer syndrome characterised by apredisposition to retinal and central nervous system hae-mangioblastomas, renal cell carcinoma, phaeochromo-cytoma and, less frequently, pancreatic islet cell andendolymphatic sac tumours. Although rare, VHL dis-ease provides a paradigm for illustrating the clinical fea-

tures and significance of familial cancer syndromes andhow the molecular genetic analysis of rare inheritedcancers can provide critical insights into the mechanismof tumourigenesis of common sporadic cancers. VHLdisease is caused by germline mutations in the VHLtumour suppressor gene (TSG). Phenotypic expressionof VHL disease is influenced by allelic heterogeneityand modifier effects. The VHL TSG has a critical “gate-keeper” role in regulating growth and differentiation oftarget tissues. Thus inactivation of the VHL TSG is themost frequent genetic event in sporadic haemangioblas-tomas and sporadic renal cell carcinoma. The functionof the VHL gene product is under intense investigationbut recent work suggests that pVHL is involved in tar-geting proteins for ubiquitylation and proteosomaldegradation. Hypoxia inducible factors (HIF-1 and HIF-2) are important targets for pVHL and inactivation ofthe VHL gene in tumours causes a HIF-mediatedincrease in VEGF expression and tumour angiogenesisconsistent with the pathological features of haeman-gioblastomas.

Kaelin WG, Maher ER (1998) The VHL Tumour suppressorgene paradigm. Trends in Genetics 14:423-425

W04-03 The Role of the Neurofibromatosis Type 1(NF1) Gene in TumorigenesisUpadhyaya, MInstitute of Medical Genetics, UK

Neurofibromatosis type 1 (NF1) is a neurogeneticdisorder that affects 1 in 3000 individuals worldwide.The clinical hallmarks of the disease include café-au-laitspots, benign neurofibromas and hamartomas (Lischnodules) of the iris. NF1 shares features common to anumber of heritable cancer syndromes that are due tomutations in a tumour suppressor gene. These key fea-tures include, (I) a tissue-restricted pattern of primarycancers in neural crest and myeloid lineage cells; (ii) theonset of malignancy at an earlier age than in the gener-al population; (iii) the occurrence of multiple primarytumours in some individuals. The NF1 gene product,neurofibromin, has been thought to be involved in thecontrol of cellular growth and differentiation by at leastthree mechanisms: (i) as an upstream downregulator ofp21ras; (ii) as a downstream effector of p21ras; (iii) as alink between tubulin and p21ras. genetic and biochemi-cal evidence indicate that NF1 is a tumour suppressor

Workshop W04: Hereditary Tumour Syndromes

gene. Information derived from NF1 knockout mice,together with cellular and molecular data derived fromneurofibroma culturing, suggest that Schwann cells areprimarily responsible for neurofibroma development.The evaluation of the genetic factors involved in theNF1 tumorigenesis would be a key step in an under-standing of the fundamental molecular mechanismsunderlying the disease. Determining the factors that areinvolved in the process of tumorigenesis in NF1 isimportant for understanding the disease mechanismsthat will potentiate more appropriate genetic coun-selling. The NF1 germ-line mutational spectrum is welldefined however little similar information is yet avail-able on somotic mutations of the gene. The sequentialsteps in gene alterations associated with NF1 tumourprogression are also not clear. Mutations of the NF1gene have been identified in other tumours not usuallyassociated with NF1 patients; indicating that neurofi-bromin must have functional importance in differentcell types. Identification of mutations in other genesmay also provide a prognostic marker for the develop-ment of NF1 tumours. Since the genetic events leadingto cancer are usually complex, it will be important toassess the roles of other genes which may interact withthe NF1 gene in the tumorigenic process. The advent ofhigh-throughput cDNA microarray technology nowmakes this type of study feasible. Gene expression pro-filing of tumour tissue should facilitate the recognitionof specific pathways and identify potential candidategenes responsible for tumorigenesis

W04-04 Tuberous Sclerosis Complex: MolecularGenetic Insights Into PathogenesisSampson, JRUniversity of Wales College of Medicine, UK

Tuberous sclerosis complex (TSC) is an autosomaldominant mutlisystem hamartoma syndrome. Centralnervous system manifestations include cortical tubers,abnormal white matter migration tracks and sub-ependymal giant cell astrocytomas. Two TSC determin-ing genes have been identified by positional cloning,TSC1 and TSC2, encoding previously unknown pro-teins now termed hamartin and tuberin. Loss of het-erozygosity in TSC hamartomas and occassional malig-nant tumours indiates a tumour suppressor function forTSC1 and TSC2. TSC2 is also mutated in the Eker ratthat exhibits dominantly transmitted predisposition torenal and other tumours and in the gigas Drosophilamutant that exhibits cellular overgrowth and endoredu-plication of DNA, suggesting a role for TSC2 in cell

cycle control. However, hamartin and tuberin interactdirectly to form a predominantly cytosolic complex.Although tuberin contains a domain related to rap1GAPthat appears essential for its tumour suppressor activity,the significance of its modest reported GTPase activityis uncertain. Cellular roles for hamartin also remainunclear, but recent evidence suggests an affect on celladhesion mediated via rho. TSC1 and TSC2 knockoutmice have been engineered and the avaialble animalmodels and natural TSC mutants in man are being expo-lited to dissect the growth control pathways that are dys-regulated in TSC.

W04-05C Mutational and Expression Analysis of theTSC1 and TSC2 Genes in GangliogliomasBecker, A1; Loebach, M1; Normann, S1; Klein, H1; Noethen, M2;von Deimling, A1; Mizuguchi, M3; Elger, CE4; Schramm, J5;Wiestler, OD1; Blumcke, I1

1Dept. of Neuropathology, Bonn, Germany; 2Dept. of Genetics,Bonn, Germany; 3Dept. of Pediatrics, Japan; 4Dept. of Epileptol-ogy, Bonn, Germany; 5Dept. of Neurosurgery, Bonn, Germany

Gangliogliomas are the most frequent tumor entityencountered in patients with temporal lobe epilepsy. Thecharacteristic histopathological admixture of glial andneuronal elements, their differentiated phenotype andbenign biological behaviour suggest an origin from adevelopmentally compromized or hamartomatous pre-cursor lesion. Here, we analyzed the tuberous sclerosisgenes TSC1 and TSC2 as potential candidates involvedin the pathogenesis of this neoplasm. Recent data sug-gest that both genes play a role in cortical differentiationand growth control. In the present study, DNA sequence,mRNA expression profiles as well as the immunohisto-chemical distribution of TSC1 and TSC2 were analyzedin 20 patients with ganglioglioma. Fifteen of thesetumors carried polymorphisms in the TSC2 gene. Thefrequency of these polymorphisms was significantlyincreased in exon 5 (12.5%) and exon 41 (15%) com-pared to control individuals (8,1% and 6,5% respective-ly, n=100). One somatic mutation in exon 33 of theTSC2 gene was encountered. In the TSC1 gene, sevenpolymorphisms (exon 14) but no mutation wereobserved. Semiquantitative mRNA expression analysisrevealed a significant 66.8 % reduction of TSC2 tran-script levels in ganglioglioma compared to normal braintissue (n=7). Immunohistochemistry using TSC2(tuberin) specific polyclonal antibodies demonstratedreduced signals in tumor cells. Preliminary data on lasermicrosdissection and harvesting of individual glial andneuronal elements identified the exon 33 mutation with-in the glial portion but not in dysplastic neurons of the

600 Workshop W03: Tumours

tumor. Our data reveal numerous polymorphisms aswell as a novel TSC2 mutation in tissue obtained frompatients with gangliogliomas. The functional relation-ship between abundant TSC2 polymorphisms,decreased mRNA levels and reduced protein expressionremains to be determined. The selective detection of aTSC2 mutation within the glial component is compati-ble with the hypothesis that ganglioglioma develop froma glioneuronal precursor lesion thru neoplastic transfor-mation of a glial element. Our work is supported by theDeutsche Forschungsgemeinschaft (SFB 400, Be2078/1-1 and the Graduate Program).

W04-06C Cerebral Lesions of the Eker Rat Model ofTuberous SclerosisMizuguchi, M1; Takashima, S2; Nozaki, Y1; Yamanouchi, H3;Nakazato, Y4

1Jichi Medical School, Japan; 2National Center Hospital for Men-tal, Nervous and, Japan; 3Dokkyo University School of Medicine,Japan; 4Gunma University School of Medicine, Japan

Background: Tuberous sclerosis (TSC) is caused byvariable mutations in either the TSC1 or TSC2 gene.Such genetic heterogeneity complicates genotype-phe-notype analyses in human TSC patients.

Hypothesis: The Eker rat, an animal model for inher-ited renal cancer, is caused by a mutation in the rat Tsc2gene. In this study we examined whether the neu-ropathology of Eker rats simulates that of human TSC.

Methods: Brains of Eker (n = 19) and control (n = 6)rats were studied histologically. Cerebral lesions foundwere immunostained for neuronal and glial markers.

Results: In the 19 Eker rat brains, there were one cor-tical tuber, two subcortical hamartomas, twosubependymal hamartomas and one anaplastic gangli-oglioma, whereas no abnormalities were found in the 6control brains. The rat tuber and ganglioglioma werenovel lesions found first in this study. The cortical tuberwas composed exclusively of cytomegalic neuronswhich were positive for neurofilament proteins. Theother lesions showed immunoreactivities to both neu-ronal and glial antigens. The rat anaplastic gangli-oglioma showd pleomorphism, nuclear atypism, manymitotic figures and a high Ki-67 labeling index (9.2%).

Conclusions: The Eker rat cortical tuber resembledhuman tubers. Its presence confirmed the value of thisanimal model as a tool for investigating the molecularneuropathology of TSC. On the other hand, the ratanaplastic ganglioglioma had features of a malignantneoplasm which are absent from human subependymalgiant cell astrocytomas.

601Workshop W03: Tumours