Genetic analysis of neurofibromatosis type 2 (NF2) patients and NF2-associated tumors with emphasis on chromosome 22 deletions

Abstract

Neurofibromatosis type 2 (NF2) is an autosomal dominant disease with the hallmark of bilateral vestibular schwannomas. NF2 patients may also develop schwannomas at other locations as well as meningiomas, neurofibromas and ependymomas. NF2 shows a distinct clinical variability ranging from very mild to severe forms. Since identification of the NF2 gene in 1993, mutation-screening analyses have been performed. Mutations were, however, not found in numerous cases and the mechanism behind tumor development in NF2 is not fully understood. The correlation between type of NF2 gene mutation and clinical phenotype of patients does not provide a clear-cut explanation of the clinical variability. Also, absence of NF2 gene mutations in constitutional tissue from a related disorder, such as schwannomatosis, points towards a possible existence of an additional NF2-related locus from chromosome 22, which may modify the disease phenotype. This implies the necessity of further clarification of the genetic factors involved in NF2. Deletions on chromosome 22 and the inactivation of NF2 tumor suppressor gene are critical steps for meningioma formation. However, 40% of the tumors do not show aberrations of chromosome 22 or mutations in the NF2 gene. This suggest that alternative mechanisms are responsible for the development of a large fraction of meningiomas. We analyzed 25 meningiomas, which do not display chromosome 22 deletions, for genetic abnormalities by CGH (comparative genomic hybridization) to metaphase chromosomes. Two tumors showed loss of chromosome 1p and 3p which suggest that deletions of both 1p and 3p may contribute to meningioma tumorigenesis (paper I). In schwannomas, detailed mutation analyses of the NF2 gene showed that 60% of the tumors carry inactivating mutations. Thus, the mechanism behind the development of 40% of schwannomas is unknown.

We studied 50 sporadic and NF2-associated schwannomas by high resolution LOH (loss-of-heterozygosity) on chromosome 22 and other chromosomes. Chromosome 22 deletions were detected in over 80% of the cases. Four tumors showed LOH not involving the NF2 locus. All exons of the NF2 gene were sequenced in these tumors and mutations were detected only in one case. Thus, additional regions chromosome 22 may harbor mutations possibly involved in schwannoma turnorigenesis (paper II). We also identified an early-onset NF2 patient with a large constitutional deletion on chromosome 22. Constitutional deletions of the entire NF2 gene were previously described in NF2 patients; two large deletions encompassing 700-800 kb have been reported in mildly affected subjects. Our severely affected case showed a much larger deletion stretching approximately 5 Mb towards the telomere (paper III). It may therefore be hypothesized that the severe phenotype in this patient is a result of a combined mutation in NF2 gene and in a putative modifier gene.

Further mapping of deletions in this candidate region was performed by analysis of 116 NF2 patients for deletions on 22q. Analysis was carried out using high-resolution microarray-CGH on a genomic array covering at least 90% of a 7.4 Mb interval of 22q, around and distal to the NF2 locus. This is a novel approach for high-resolution detection of chromosomal abnormalities, both in constitutional and tumor-derived DNA. Deletions were detected in nine severe, nine moderate and six mild patients. Deletions in severely and moderately affected patients varied in size, whereas cases with mild NF2 displayed deletions affecting the NF2 locus only (paper IV). This result indicates that the geneotype/phenotype correlation is unlikely to exist and supports the notion that the NF2 modifier gene may exist in the vicinity of the NF2 locus. Finally, the tumor suppressor gene SMARCB1 was tested for mutations in meningioma and schwannoma. SMARCB1 is mutated in malignant rhabdoid tumors and is located on 22q11.2, a region frequently deleted in meningiomas. Forty-three meningiomas and twenty-one schwannomas were tested for mutations in exons 2 through 8. However, no mutations were detected, suggesting that the SMARCB1 gene is not frequently involved in the pathogenesis of these tumors. We also identified the mouse ortholog and characterized different splice forms of this gene, both in human and in mouse (paper V).