BOSTON – Ten years after he returned to his native Iceland to build a biopharma company, Kari Stefansson says deCODE Genetics’ intense search for genes underlying common diseases is not only pushing promising new drug candidates into the clinic but is also revealing new insights into the very basis of common disease.

In a keynote address at Bio-IT World’s Life Sciences Conference + Expo Tuesday, Stefansson acknowledged, “There’s great enthusiasm for human genetics, but it has yet to deliver anything of great significance.” But he believes that “the genetics of common disease is the genetics of gene expression.”

Kari Stefansson

Since 1996, deCODE has launched research programs to study 50 common diseases, mapping and isolating susceptibility genes for 30 diseases and turning nine of those into drug discovery and development programs. Three drugs are already in the clinic, a number Stefansson says will increase to five before the end of 2006.

Complex traits are confounded by environmental factors. Cases of lung cancer in Iceland are invariably triggered by smoking, but 10–15 percent of cancer deaths cluster in families, suggesting a genetic susceptibility that doesn’t penetrate until subject to an environmental trigger.

DeCODE’s search for the genetic causes of common disease have revealed several common themes. DNA sequence variants that predispose an individual to common disease rarely fall in the gene coding sequence; typically influence gene expression, e.g. alternative splicing; almost always provide a good drug target; and allow for the “intelligent design of clinical trials.” 

DeCODE’s data mining relies on three key criteria – phenotype, genealogy, and genome data. “Genealogy is important in quality control and is the key to success in genome-wide association studies,” said Stefansson. Stefansson proudly noted how his personal genealogy tree extends back 1000 years to one Egil Skallagrimmson – “A great poet, great warrior, and said to be the ugliest man alive.”

Testing Times
DeCODE’s genotyping programs typically involve screens of 300–500,000 SNPs, but that raises the statistical problem of multiple testing. DeCODE’s standard practice is to test any Icelandic gene variants in a second population. “Every single discovery we report is done after we have replicated it in at least several other populations,” said Stefansson.

Variations in the PDE4D genotype influence gene expression in opposite directions.

“Biochemical pathways exist for a reason. If you shift from one extreme to another, it will confer risk to another problem,” said Stefansson, citing statins as a prime example. “You have to be cognizant that evolution has left the pathway for a reason.”

The gene deCODE identified for myocardial infarction – FLAP – acts at the beginning of the leukotriene inflammatory pathway. The variant, which has been confirmed in British and American cohorts, increases production of LTB4, one of two branches of the LT pathway.

“Do we have environmental components that work through the same pathway?” asked Stefansson, pointing to reports of a link between gingivitis and heart disease. “Can we use the same measure to control genetic and environmental components [of the disease]?”

DeCODE licensed a FLAP inhibitor, DG031, from Bayer—one of several candidates previously identified by Big Pharma (for asthma) with the best safety profile. DG031 downregulates production of LTB4 and is heading for Phase III trials.

Meanwhile, variants in a second leukotriene gene, LTA4 hydrolase, greatly increases risk in African-Americans compared to Caucasians. Interestingly, the variant is not found in Africa at all, suggesting it has arisen since man left Africa 50–100,000 years ago. “Why is this variant so much more dangerous in African-Americans?” Stefansson speculates that exposure to different microorganisms as man migrated into Europe resulted in the upregulation of the leukotriene pathway, but life expectancy was low. Stefansson suggests that the mutation was introduced into the African-American population relatively recently via admixture, leaving insufficient time for adjustment. 

In diabetes, deCODE has found an increased risk of 50 percent associated with a SNP in a transcription factor, TCF7L2. This finding has been confirmed by 5–10 groups. Another gene haplotype, found in 10 percent of Africans and 95 percent of individuals in East Asia, is associated with increased body mass. Stefansson agreed “this is a bit twisted,” as obesity is usually associated with Type 2 diabetes, but this gene appears to confer protection.

Stefansson also said that deCODE had identified a gene variant on chromosome 8 that increases the risk of prostate cancer by 50 percent in European and African-American patients. The variant is 1.6 times more prevalent in African-Americans than Caucasians, which curiously is also the increased prevalence of the disease in African-Americans.