Search

International audience

e21592 Background: While 10% of melanomas occur in a context suggesting hereditary predisposition, a clear molecular explanation has only been established for approximately 20% of families. In the course of clinical care, we identified a new CTNNA1 truncating germline mutation in a family affected by multiple early-onset melanomas. Methods: NGS and CGH-array were performed on the index case’s melanoma, followed by Sanger sequencing of the germline DNA of relatives. Immunochemistry (IHC) was employed to evaluate the level of αE-catenin (encoded by CTNNA1) in the family's samples. Stable CTNNA1 knockout human melanoma cell lines were generated to investigate the functional effects of CTNNA1 loss. Functional assays, including colony formation, 3-D tumor spheroid formation, wound healing, and transwell invasion were performed, as well as electron microscopy and RNAsequencing (RNAseq). CTNNA1 mutational status was determined in several databases and further sequencing of CTNNA1 in a DNA bank of families with multiple melanomas was done. Results: While the allele frequency in the index patient’s tumoral DNA was compatible with a germline mutation, the CTNNA1 F611fs*10 mutation was subsequently found cosegregating with individuals affected by melanoma in the family. CGH array on tumor DNA identified a segmental loss on chromosome 5, leading to a loss of heterozygosity of CTNNA1, resulting in a loss of αE-catenin observed by IHC. Clinically, the mean age of first melanoma diagnosis was 29,7 years (range: 18-56), 2 were metastatic, and the others were SSM (superficial spreading melanoma) in situ (n = 3) or Breslow index 0,56 mm (n = 1). Functional assays performed on CTNNA1 KO melanoma cell lines showed a loss of cell-to-cell adhesion phenotype, in accordance with the altered adherens junctions observed by electron microscopy, and the specific pathway enrichments observed by RNAseq. This specific phenotype could be rescued by transfection with a plasmid containing wild-type CTNNA1, as opposed to the CTNNA1 F611fs* plasmid. Germline CTNNA1 mutations are rare as none could be further identified in a DNA bank of 27 multiple melanoma families. In a database of 4743 melanomas somatically sequenced for CTNNA1, 131 of them had a CTNNA1 alteration (2,76%), with a median tumor mutational burden of 44 mut/MB of DNA (range from 0 to 451). Among them, 15 alterations were predicted to be inactivating, including 7 associated with a BRAF or NRAS activating mutation. Conclusions: Altogether, our results strongly support that CTNNA1 loss of function predisposes to melanoma formation characterized by a decreased cell adhesion. Since germline CTNNA1 alterations have already been implicated in lobular breast cancers and hereditary diffuse gastric cancers, CTNNA1 likely constitutes a tumor suppressor gene involved in familial melanoma, thus broadening the spectrum of syndromes associated with this gene.