From a candidate region to gene characterization : analysis of three new genes with respect to meningioma tumorigenesis

Abstract

Meningioma is a common and usually benign tumor arising from the meninges covering the central nervous system (CNS). The yearly incidence of symptomatic meningiomas is 1.2 cases per 100 000 individuals but autopsy findings reveal an occurrence ten times higher. In adults, intracranial meningiomas represent 15-20% of all primary brain neoplasms, while spinal meningiomas account for 25% of all spinal tumors. The vast majority (90%) of meningiomas are sporadic but rare familial aggregations can be found. These families indicate that a genetic predisposition for meningiomas exists. However, genetic linkage analysis is restricted as these families often contain only two affected members.

Meningiomas are also found in 50% of the patients suffering from neurofibromatosis type 2 (NF2), which is an autosomal dominant disease with the hallmark of bilateral vestibular schwannomas. The NF2 gene is well characterized and is frequently inactivated in NF2-associated and sporadic meningiomas. However, approximately 50% of sporadic meningiomas do not display NF2 gene mutations or aberrant gene expression. It is therefore clear that additional genes responsible for the tumorigenesis of meningioma need to be identified.

Monosomy 22 is a primary event involved in meningioma development which is detected in up to 70% of the tumors. We examined a series of 170 sporadic meningiomas with 45 chromosome 22 specific markers in an attempt to identify a candidate chromosomal region. Interstitial deletions were found in six tumors (3.5%), revealing that at least four regions from this autosome may harbor meningioma tumor suppressor genes (TSGs). Physical maps, based on YAC and cosmid contigs, were constructed in three of these candidate regions, encompassing a total of approximately 4 Mbp.

One candidate region was further studied to identify active genes. Two new genes were isolated using exon-trapping: a functionally anonymous gene, named pKI.3, and a new member of the adaptin family, the B'-adaptin gene (GDB symbol, ADTBI ). ADTBI was totally inactivated in one tumor and therefore represented a putative meningioma TSG. Our meningioma series was searched for additional genomic rearrangements, point mutations and/or aberrant mRNA expression. The ADTB1 gene was found to be totally inactivated in 12% of 71 sporadic meningiomas, either via homozygous deletion (one case) or via total loss of gene expression (eight cases). We therefore named it beta-adaptin-meningioma-chromosome 22 (BAM22). The genomic structure of the ADTBI/BAM22 gene is composed of 23 exons, one being alternatively spliced, and has a genomic size of 60 kb. We also cloned the mouse ortholog of the BAM22 gene, adtbl, and protein sequence comparison of the murine and human genes revealed 96.5% identity.

Adaptins are important subunits of heterotetrameric complexes, termed adaptors or assembly proteins (APs), which are components of clathrin-coated vesicles (CCVs). CCVs are responsible for endocytosis and intracellular transport of membrane-spanning receptors from the plasma membrane or the trans-Golgi network (TGN). The AP-1 type of adaptor is composed of a B'- and a y-adaptin, a small and a medium subunit, AP19 and AP47, respectively. The total lack of a B'-adaptin in meningioma cells is likely to disrupt intracellular transport involving AP-1, i.e. transfer of lysosomal hydrolases from the TGN to Iysosomes via mannose-6-phosphate receptors. It can therefore be assumed that a dysfunction of other components of the AP-1 adaptor will also inactivate this transport pathway.

We therefore determined whether the genes encoding for the y-adaptin, API9 and AP47 subunits would also be located on chromosome 22. The human y-adaptin gene (ADTG) was characterized and mapped to chromosome 16q23, a region deleted in prostate and hepatocellular carcinomas as well as in Wilms' tumor. The API9 and AP47 genes were localized to human chromosomes 7 and 19, respectively.