The identification of two families with homozygous APP variants (Tomiyama et al., 2008 and now Di Fede et al., 2009) is interesting, but the linkage analysis in neither pedigree is sufficient for us to be sure that the mutations are pathogenic, let alone sufficient for us to tell whether they act in a co-dominant or recessive fashion. The homozygosity in both families is almost certainly caused by unrecognized consanguinity and this is common even in ostensibly outbred populations (Nalls et al., 2009). Thus, in both cases it is still possible that the variants are simply harmless polymorphisms or that they additively increase risk of disease. In both families, it might be informative to carry out whole genome genotyping to see which other loci are homozygous. Interpretation of these variants as recessive is premature.

View all comments by John Hardy

I recommend the primary paper.

View all comments by J. Lucy Boyd

There may be a reversible formation of dimers in the APP molecule, before being cleaved by γ-secretase and the subsequent formation of Aβ dimers, thus preserving the molecule for cleavage by α-secretase.

Data provided by this extraordinary article based on a mutation of the amino-terminal region of Aβ in its precursor molecule APP (changes in its primary sequence trigger peptide assembly and fibril formation) point to this possibility, since the relationship between the carboxy-terminal fragments generated by β- and α-secretase (1.9 ± 0.2 increase in C99: C83 ratio in patient's fibroblasts) implies double the rate of β product versus α in the mutant APP.

View all comments by Miguel Rodríguez-Manotas