The story began when Tanzi and colleagues tested nine known single nucleotide polymorphisms (SNPs) in the ADAM10 gene for association with AD using 1,439 DNA samples from 436 families in the NIMH AD Genetics Initiative collection. First author Minji Kim and colleagues found that one SNP (rs2305421) showed evidence of association, and this was confirmed by a subsequent genomewide association study (see Bertram et al., 2008). Though they found no association using samples from the Consortium of Alzheimer’s Genetics (CAG) samples, possibly due to the lower power of this sample, suggest the authors, the association held up when analysis was conducted on a combined NIMH/CAG sample set. Kim and colleagues then sequenced the entire ADAM10 gene in samples from 32 families in the NIMH dataset that showed the strongest association with this one SNP.

This sequencing uncovered two rare, amino acid-coding changes in exon 5 of the ADAM10 gene. Looking at the complete NIMH database, Kim and colleagues found five families affected, three with one substitution (a histidine for glutamine at position 170) and two with a second (a glycine for arginine at position 181). Screening of 1,111 individuals from 351 affected families in the National Institute on Aging Study sample revealed one additional family with each mutation.

Because the mutations are extremely rare, they will not impact as many individuals as the genes for early-onset AD or ApoE, the strongest genetic risk factor for late-onset AD. However, they are the only ones for late-onset AD to date that can be ascribed a function. “We have functional links for early-onset AD, but I don’t know of any for late-onset,” said Tanzi. “These mutations clearly affect function of ADAM10 in vitro. There is almost 70 percent loss of function conferred by either of these mutations, which is on par with a known artificial dominant negative mutation.” Kim and colleagues found less α-secretase products (sAPPα and α C-terminal fragment) in Chinese hamster ovary (CHO) cells producing amyloid precursor protein (APP) when the cells expressed ADAM10 mutants compared to cells expressing wild-type ADAM10. Cells expressing the mutants also produced more amyloid-β than cells making wild-type ADAM10. The CHO cells expressing the mutant ADAM10 also failed to mount a strong α-secretase response to phorbol esters, which can induce the secretase. The results indicated that the mutations might increase susceptibility to AD by reducing α-secretase activity in favor of the amyloidogenic β-secretase pathway.

Tanzi admits that one thing that weakens the story slightly is that some of the affected members in some families didn’t carry either mutation, suggesting other factors explain their susceptibility. The researchers also found the mutations may not be fully penetrant because some people carrying them are unaffected. “That’s not to say they don’t carry other mitigating protective factors and that if they live long enough they might get AD, so you can’t say they are escapees,” said Tanzi. “But the fact that they are around the same age range of affected individuals and that they carry the mutation but are not affected mean we have to assume the mutations are incompletely penetrant.” One reason for the lack of penetrance might be that other α-secretases in the brain, including ADAM9 and TACE (tumor necrosis factor-α convertase), compensate for loss of ADAM10 activity. That might be worth studying, said Tanzi. His group also plans to follow the unaffecteds to see if they eventually do get AD, and he is also working with transgenic models to investigate how the functional loss of ADAM10 activity plays out in vivo.—Tom Fagan.

Reference:
Kim M, Suh J, Romano D, Truong MH, Mullin K, Hooli B, Norton D, Tesco G, Elliott K, Wagner SL, Moir RD, Becker KD, Tanzi RE. Potential late-onset Alzheimer's disease-associated mutations in the ADAM10 gene attenuate {alpha}-secretase activity. Human Molecular Genetics 2009 July 15. Abstract