Image: Courtesy of Stephen B. Liggett
The promise of pharmacogenetics will not be realized easily. To date, most studies have focused on individual SNPs (single nucleotide polymorphisms), or perhaps a few, but none have considered the potentially complex interactions between SNPs on the same gene. That one-at-a-time approach could cause researchers to miss important clues, says Stephen B. Liggett, professor of medicine and molecular genetics, University of Cincinnati College of Medicine. Liggett and his team published results in this Hot Paper showing that multiple SNPs on a single gene can cause functional changes.1
Liggett says he was puzzled by studies such as one involving the a2-adrenergic receptor, which resides on the surfaces of the smooth muscle cells of the respiratory airway, where it regulates airway diameter. It is also the target of beta-agonists that are used to treat asthma. But much individual variability exists in treatment response, prompting researchers to investigate whether SNPs in the receptor's gene might account for the differences.
"There were a series of clinical studies ... and sometimes associations were found [between an individual SNP and a response pattern] and sometimes they were not. And sometimes they were the opposite of what we thought they would be," says Liggett.
Liggett realized one potential explanation existed: Multiple SNPs on the same DNA strand could be interacting with one another, producing effects in concert that would be difficult to evaluate by looking at one SNP at a time. "We already knew that there were two functional SNPs in the coding region, and we suspected that there were multiple polymorphisms in the promoter and 5' upstream regions. It is a combination of SNPs that happens in nature, so the approach was to extensively explore those areas."
Before conducting the analysis, Liggett and his team separated the chromosomes to group the SNPs together within the same gene, which is also known as the haplotype. In a survey of Caucasian, African-American, Asian, and Hispanic-Latino individuals, the team discovered 13 SNPs. Of the 8,192 possible combinations, they found greater than 20-fold differences in the frequencies of the four most common haplotypes.
The team then investigated the relevance of the five most common haplotypes. In vivo, they found greater than two-fold differences in cellular responses to a-agonists. The responses varied by haplotype rather than by individual SNPs, providing strong evidence that it was the SNP combination that led to functional changes.
The team also created a vector containing two haplotypes differing at eight different SNP positions. These two variations were associated with divergent treatment responses. In transfected cells, they observed that the haplotype associated with the greater physiological response in patients had a greater than 50% increase in both mRNA levels and receptor density on the cells' surfaces.
The paper was a first, Liggett says. "We discovered the SNPs, made the haplotypes, looked at the ethnic populations, and did the clinical studies and the in vitro studies, all in one paper." The group, he continues, is trying to mature the thought process around common genetic variants. "They need to be thought of in context with other common variants, until proven otherwise. It could be that there is a dominant SNP that controls everything, and you'll find that out when you do your clinical work or in vitro study." Still, he believes that these sorts of interactions will be common explanations for variations in clinical response. "We're going to be forced to look at haplotypes throughout the gene to improve our predictive capacity."
Focusing on SNP complexity increases the amount of work, but the payoff will be well worth it, Liggett says. "It's obviously more trouble analytically ... but you will ultimately need fewer patients in your study. Studies are going to require more rational design" to allow for the removal of patients who have, for example, rare haplotypes. "But it removes that person from the study ... and if you can eliminate a portion of people who are nonresponders, you'll have a dramatic change in your overall effectiveness."
Jim Kling (firstname.lastname@example.org) is a freelance writer in Washington, DC.