Heterogeneity of mesothelioma cell lines as defined by altered genomic structure and expression of the NF2 gene.

Germ-line mutations in the neurofibromatosis 2 (NF2) gene cause a susceptibility to the development of schwannoma and meningioma, 2 mostly benign tumors of neural crest origin. Bi-allelic inactivation of this gene has been observed in sporadic schwannomas and meningiomas. The NF2 gene may also be somatically inactivated in human malignant mesotheliomas (HMMs). Surprisingly, patients with an NF2 germ-line mutation have not been reported to be at an increased risk for this highly invasive tumor of mesodermal origin. To investigate in HMMs the silencing mechanism of the NF2 gene, we have analyzed its structure and expression in a series of 18 cell lines derived from HMMs. NF2 gene alterations were identified at a genomic level in 7 cell lines and were associated with a marked decrease in the concentration of the NF2 transcript. This decrease was also observed in 4 additional cell lines with no identified NF2 mutation. The 11 cell lines presented evidence suggesting deletion of one NF2 allele. None of these enabled the detection of normal or truncated forms of the NF2 protein by immunoprecipitational immunoblot analyses. In the 7 remaining cell lines, NF2 mRNA and NF2 protein were easily detectable. Among the latter, 4 lines were heterozygous for several chromosome 22 microsatellite loci, suggesting the presence of 2 NF2 alleles. Taken together, our data indicate that silencing of the NF2 gene is restricted to a subset of mesothelioma cell lines. The availability of established cell lines with different characterized NF2 status provides a powerful tool to explore the mechanism by which the NF2 protein exerts its tumor suppressive activity.

Impaired interaction of naturally occurring mutant NF2 protein with actin-based cytoskeleton and membrane.

Although schwannomin, the product of the neurofibromatosis type 2 gene, shares homology with three cytoskeleton-to-membrane protein linkers defining the ERM family, the mechanism by which it exerts a tumor suppressive activity remains elusive. Based on the knowledge of naturally occurring mutations, a functional study of schwannomin was initiated. Constructs encoding the two wild-type isoforms and nine mutant forms were transfected into HeLa cells. Transiently expressed wild-type isoforms were both observed underneath the plasma membrane. At this location they were detergent insoluble and redistributed by a cytochalasin D treatment, suggesting interaction with actin-based cytoskeletal structures. Proteins with single amino acid substitutions at positions 219 and 220 demonstrated identical properties. Three different truncated schwannomins, that are prototypic for most naturally occurring NF2 mutations, were affected neither in their location nor in their cytochalasin D sensitivity. However, they were revealed to be detergent soluble, indicating a relaxed interaction with the actin-based structures. An increased solubility was also observed for a mutant with a single amino acid substitution at position 360 in the C-terminal half of the protein. Mutant proteins with either a single amino acid deletion at position 118 or an 83 amino acid deletion within the N-terminal domain had lost the submembraneous localization and tended to accumulate in perinuclear patches that were unaffected by cytochalasin D treatment. A similar behavior was observed when the N-terminal domain was entirely deleted. Taken together these observations suggest that the N-terminal domain is the main determinant that localizes the protein at the membrane where it interacts weakly with actin-based cytoskeletal structures. The C-terminal domain potentiates this interaction. With rare exceptions, most naturally occurring mutant schwannomins that have lost their tumor suppressive activity are impaired in an interaction involving actin-based structures and are no longer firmly maintained at the membrane.