Pickering-Brown, speaking on 15 July, described an investigation into the genotypes of nearly 400 people with FTLD. The normal number of repeats in C9ORF72 is 23 or fewer; 8 percent of the Manchester cohort possessed larger stretches of up to 1,500 repeats. The C9ORF72 expansions cause a specific pathology consisting of TDP-43-positive inclusions in the brain or spinal cord, as well as ubiquitin-positive but TDP-43-negative aggregates in the cerebellum.

The mystery of the two brothers arose when pathologist David Mann examined their autopsy samples and did see the telltale cerebellar pathology, but, curiously, the genetic analysis showed only half a dozen repeats in C9ORF72. Pickering-Brown recalled that Mann queried, “Are you sure these do not have a repeat expansion?” By repeat-primed polymerase chain reaction, the method of choice to identify the expansion in this difficult-to-sequence location, they were not. The PCR output showed six peaks, corresponding to six repeats. But when the researchers scrutinized their results more closely, they observed “really tiny peaks” that could indicate longer expansions, Pickering-Brown told Alzforum. Southern blots confirmed Mann’s hunch: The C9ORF72 gene was, in fact, 800-950 repeats long in the brothers. “Something is stopping the repeat-primed PCR from working efficiently,” Pickering-Brown concluded.

A mutation in one of the sequences to which the PCR primers bind could interfere with the subsequent chain reaction, he reasoned. One primer binds outside the repeat region, but the sequence there was normal. The team surmised that the mutation must be present inside the hexanucleotide repeats themselves. They designed primers with nucleotide substitutions at each of the six sites in the GGGGCC code. When they altered each of the last two cytosines individually, the PCR worked “a little bit better,” but still not as well as it ought to. Although the team is not sure, they suspect that both of those final bases are altered in the DNA from the two brothers.

“Stuart’s data demonstrated that extensive Southern blot analysis is going to be required to answer key issues related to C9ORF72,” commented Leonard Petrucelli of the Mayo Clinic in Jacksonville, Florida, in an e-mail to ARF. “Variation in the repeat sequence in C9ORF72 suggests that this cause of FTD and ALS might be more common than we first thought,” added Petrucelli, who was not involved in the study. For his part, Pickering-Brown thinks this altered repeat sequence is quite rare, although he will not be able to confirm his hunch until he knows the right code to look for in other samples. How did one family acquire the unusual repeats? Perhaps, Pickering-Brown speculated in an interview with Alzforum, the polymorphism arose in this particular kindred even before the repeat expansion occurred.

People who carry C9ORF72 expansions share a common haplotype, leading many researchers to suspect that the repeats proliferated in a single founder (Smith et al., 2012). Pickering-Brown’s work, albeit incomplete, could support an alternative theory. It might be that something about this haplotype promotes the hexanucleotide expansion, which would explain how the repeats expanded separately in the family he studied. Corroborating this idea, Pickering-Brown and others have found that the length of affected people’s repeat regions varies widely among their tissues. He presented Southern blots of brain homogenates, which show that the C9ORF72 gene appears not as a single-sized band but a faded smear ranging from several hundred to a few thousand repeats long. Thus, the repeat region may be inherently unstable and prone to shrink or expand with every cell division. As Petrucelli concluded, “The elusive chromosome 9 locus is still not giving up its secrets easily.”—Amber Dance.