Background: Proteomics is routinely used to type clinical amyloid deposits, and offers additional benefit of identifying genetic variants, which can be diagnostically useful. Reviewing the proteomics data for ATTR patients attending our Centre revealed an unusually large number of samples containing a rare pathogenic H90D TTR variant alongside the more common H90N variant.
Methods: These findings raised questions to their source. Proteomics was used to monitor the generation of H90D/H90N variants in fresh, frozen, stored samples during extraction and digestion, and also following Cu2+-mediated oxidation.
Results: There was no evidence that the variants were present in the circulation, except in one patient with genetically confirmed H90D TTR, in fresh fat aspirates or tissues from an ATTR amyloid mouse model. The variant could be generated in vitro from both wild-type TTR and ex vivo ATTR fibrils by non-enzymic oxidation of histidine at position 90. These data suggest that the H90D variant can be generated artefactually from wild-type 90H TTR through a radical-mediated oxidation of histidine, followed by its conversion to asparagine and aspartic acid. This probably occurs during storage.
Conclusions: In the absence of genetic data, the identification of H90D TTR in stored tissue by proteomics should be treated with caution.
Keywords: FFPE storage; H90D; TTR; proteomics; variant misidentification.