Wednesday, November 03, 2010

Mild alleles in severe diseases: an opportunity for enlightenment

Monday's Globe had a blurb about a book signing which re-kindled a previus interest of mine. The author, Michael Dana Kennedy, had quit his job as a medical researcher to write the novel, a tale of two brothers ending up on opposing forces in the Pacific during WW2. While the book concept might have enough appeal to go to the back of my infinite reading list, it's the author's backstory that really grabs me.

The reason Kennedy quit his job is that it was perceived as a health threat: he was diagnosed with cystic fibrosis in his mid-50s. This reminded me of an elderly female patient the Gene Sherpa had mentioned who he had diagnosed with CF.

Cystic fibrosis is a very difficult disease, and for a long time few patients made it out of their twenties. I understand that with modern care, including antibiotics and regular respiratory therapy, many patients live substantially longer. But, this underscores what I find interesting about these two patients -- without any treatment at all they have long outlived most of their peers afflicted with cystic fibrosis. Hence, they must have comparatively mild cases. And those should be interesting.

The key question is why are these cases so mild? The simplest answer would be that they carry at least one allele which retains substantial function of CFTR (the gene mutated in CF). The more complex answer would be that they carry other genetic variants which substantially moderate the impact of the defective allele(s). Either answer would be very enlightening, both for CFTR specifically and for better understanding protein function in general.

I would expect that with the PGP and 1000 genomes project and all the other human genome sequencing efforts public and private, many new alleles will be discovered in many well understood disease genes (or as well understood as any disease gene is). A key follow-up to execute, when possible, is to determine which of these alleles had health impact. CF is appealing from this angle because we know a biochemical phenotype (altered salt excretion) which can be measured and we know a possible medical issue to assess (history of frequent respiratory infections). BRCA1 would be another valuable case where we already know many disease alleles, though there the question is more complicated to answer. I'm sure there are many more.

Studying some of these easier cases will, with luck, help shed some light on the avalanche of novel genetic variants which are pouring from germline genome projects -- and an order of magnitude higher from cancer genome projects (since in many tumors there is some combination of deficient DNA repair/replication as well as significant historical exposure to mutagens such as cigarette smoke). Lacking good high-throughput ways to assess most of these functionally, it would behoove the community to leverage the ultimate functional tests -- human survival.