Originally published Nov. 4
NEW YORK (GenomeWeb) – By performing a meta-analysis of several genome-wide association studies comprising more than 40,000 patients, UK researchers have found strong evidence of two novel loci associated with statin response and confirmed two other previously identified regions of the genome known to influence the effectiveness of cholesterol-lowering drugs.
The findings, published last week in Nature Communications explain approximately five percent of the variation in response to statin drugs, Mark Caulfield, the study's senior author and co-director of the Queen Mary University of London's William Harvey Research Institute, told PGx Reporter.
In addition to these markers, if the researchers can use their broad GWAS analysis approach to find more variants with similar contributions to statin response variation — linking these regions not only to cholesterol lowering, but also to patients' actual cardiovascular outcomes — then it could lead to the development of a panel test that has clinical utility for a large number of statin users.
While GWAS and other genomic screening efforts have uncovered some important and useful PGx variants, progress has sometimes been checkered due to the challenges of studying drug response in heterogeneous populations that have been treated with multiple medications, and in small cohorts.
As part of an effort called the Genomic Investigation of Statin Therapy (GIST) consortium, Caulfield and his colleagues set out to combine many datasets together in hopes that a much larger pool could help them discover more about the genetic component of statin response.
"Some people given the same dose don't respond as well as others, and while this isn't a massive problem and we can measure it by testing [blood] cholesterol, it can go undetected for a long time because of how we usually use these medicines," Caulfield said.
"The effects of statins on cardiovascular endpoints are fairly rapid in onset," he explained. "So knowing if someone will not respond as well as expected at the outset could give you a much better stratification of patients and avoid the risk that patients might languish for months exposed to a high risk of [adverse cardiovascular events]."
In their meta-analysis, the researchers combined data from both randomized controlled trials, and observational studies of statin-treated patients – six controlled trials, and ten observational studies in total, covering more than 18,000 statin recipients.
The team then selected 246 SNPs from 158 loci for a second investigation in three more studies comprising more than 22,000 statin-treated subjects.
Overall, the study identified four gene regions that were linked to statin response, two that had been previously known definitively, and two others that were novel.
The two known loci, in the genes APOE, and LPA, impact proteins that carry cholesterol. These regions have been previously linked to variation in the effect of statins, but the current study lends even more evidence to their implication in lower-than-expected response to the drugs.
Overall, nine SNPs in the APOE gene region reached genome-wide significance for LDLC lowering, the authors reported. For the lead SNP, the group found that the minor allele was associated with a larger LDLC-lowering response than the major allele, with a similar magnitude to that seen in previous studies.
Three more SNPs in the LPA gene region were significantly associated with LDLC response in the meta-analysis cohort, again seeing a similar magnitude in terms of the impact of major and minor alleles of the lead SNP.
In addition to confirming and expanding these previous findings, the large analysis also identified two new loci associated with variations in response to statins. One was in the gene SLCO1B1, variants of which are associated with increased risk of muscle damage due to statin treatments.
According to Caulfield, the same mechanism that explains SLCO1B1's role in statin-induced muscle injury may also explain how the marker impacts the cholesterol-lowering effects of the drugs.
"Previous analysis has found that SLCO1B1 is a transporter of statins into the liver, where they principally act," he said. "If you can't transport them in to the liver you also can't break them down, leading to accumulation in muscle tissue and muscle damage."
"We found variants in the same gene that appear to be associated with reduced response and we suspect the mechanism is similar," he explained. Essentially the same biology that leads to statins accumulating where they shouldn't suggests they are not reaching and acting on their intended target.
The other novel locus identified by the group is a region associated with three genes, SORT1, CELSR2, and PSRC1.
Caulfield said that all three genes are plausible locations for an effect on statin response, but one in particular, SORT1, which is involved in intra-cellular cholesterol transport, is the group's top contender.
In aggregate, variants represented by the four loci the group discovered affect the observed variation in statin LDLC lowering by about five percent, Caulfied said.
What the researchers do not yet know, and what is a central focus on the work they have been doing to follow up this discovery, is whether this influence on LDLC lowering translates to increased cardiovascular events in those who have diminished response.
Caulfield said that the team is already advanced in its efforts to link the effect of these four loci to clinical outcomes. In addition, the group is expanding the study, both with more subjects and by searching more areas of the genome for potential influential variants, using a strategy of imputing the presence of variants using data on patterns of association between SNPs from the 1,000 Genomes Project.
"Even though you may only have typed 1 million variants, because we know how they correlated across the genome, we can actually use anchor points to impute SNPs we haven't actually typed," Caulfield explained. "It takes you from only about 2.5 million gene variants to almost 9 million," he added.
Whether the findings can be clinically useful in the future – as a way to predict patients' ability to respond and to inform dosing or alternate treatments – depends on whether the team can find additional variants with similar effects, but most importantly, on whether any of these loci are related not only to cholesterol lowering effectiveness, but also to cardiovascular outcomes.
"If they are, the impetus to use them as a [PGx] test would be much greater," Caulfield said.
The GIST consortium is not the only group trying to track down genetic markers associated with variations in statin response. A recent study by researchers at the Children's Hospital Oakland Research Institute identified 100 genes whose expression levels differed between people who are high and low responders to statins. These differences could explain some 12 percent of the variation in response, according to the authors.
Commercial diagnostics developer AutoGenomics and personalized medicine firm Genomas are also co-developing a gene-expression based assay for predicting statin response and risk of muscle damage.
Caulfied said that he and his colleagues have also been looking at how gene expression correlates with the genotype influences they have discovered.