KIRKLAND, Wash., Dec 12 - Maynard Olson thinks genomics is heading in the wrong direction.
Instead of hunting for disease-associated SNPs and genes – as companies like Decode genetics, Gemini Genomics, and Incyte are currently doing – researchers should be looking for health-associated alleles, the University of Washington geneticist told a group of genomics executives and journalists assembled by Rosetta Inpharmatics CEO Stephen Friend last week.
“Many of the pathological mechanisms being studied are dismally poor candidates for therapeutic intervention,” said Olson as he projected a slide to the group entitled “Gene Today… Gone Tomorrow.” (The slide showed a website screen announcing the not-so-recent discovery of the gene for Rhett’s syndrome, for which no therapeutic has yet been discovered.)
Many of these gene-based mechanisms, he said, involve protein malfunction or deficiencies. And since most pharmacological agents antagonize proteins rather than replace them, it will be difficult to translate these genetic discoveries into therapeutic agents.
“The information you would like is where there is a protein that when you get rid of it can fix disease,” Olson said. “Instead of [discovering] proteins on the basis of the deleterious phenotype, we need to acquire them on the basis of the advantageous phenotype – genes if when mutated made you healthier.”
One example of this healthy mutation, Olson said, is an allele that makes people unable to produce the CCR5 chemokine receptor required to infect immune cells with HIV. The HIV hooks onto this receptor and the CD4 receptor to infect a cell.
But when people have a homozygous genetic mutation that makes them unable to produce this receptor – as do a number of people in Northern Europe – they are completely resistant to HIV infection, as reported by David Ho in a November 1996 Nature Medicine article.
“It is not unlikely that a significant proportion of all human genes are in this category,” Olson said.
Olson blamed the exclusive focus on finding disease-associated alleles on the leadership at National Institute for General Medical Sciences for what he called the “disastrous spin put on genetic epidemiology,” that encourages the idea that “everyone’s got bad genes inside them.”
Olson’s talk sent ripples of disagreement through his assembled audience.
Before researchers go treasure hunting for “healthy genes,” they need to define what actually constitutes health, said Roger Ulrich, director of toxicology programs at Abbott Labs. This definition of health could be quite arbitrary, he argued.
“A healthy 100 year-old is different from a healthy 21 year-old,” he said.
Further, he added: “Most major pharmaceutical companies don’t believe diseases in general are caused by single gene defects. There are regular elements that control expression of genes and companies are pursuing therapies that intervene downstream or upstream.”
Abbott is also using genomic research (in partnership with Rosetta) to identify alleles that cause toxic responses to drugs, as well as to find drugs – a form of genomics that would not be affected by difficulties in translating genes to drugs. “We’re just as interested in what paths not to go down,” he said.
Even if a disease-associated gene does not lead directly to a therapeutic, “you can intervene with genetic testing,” pointed out William Efcavitch, Applied Biosystems’ director of synthesis and arrays.
Arthur Sands, CEO of Lexicon Genetics, painted a more optimistic picture about the ability of disease-associated gene discoveries to lead to drug targets. His company recently launched the LexVision program to use knockout mice to discover thousands of genes that are drug targets and has entered into research collaborations with Rockefeller University and The Scripps Research Institute to find schizophrenia-associated genes using knockout mice.
Olson’s argument “is restating the obvious,” Sands asserted. “When you have a mutation you lose function. Most drugs are antagonists.” If a genetic change in a knockout mouse removes a receptor for a disease cell, as in the HIV example, this is a valuable piece of information.
But in a case where a mutation makes a DNA repair gene fall apart, an antagonist to the DNA repair side of equation could be a valuable antitumor agent, if the replicating DNA is that of tumor cells. “We have to think beyond health and disease models,” he said.
In response to this flurry of criticism and discussion, Olson admitted there was merit to studying disease-associated genes.
“I am taking the extreme position for emphasis,” he said. “ But the information is clearly valuable. The history of medicine has focused overwhelmingly on the deleterious tail [of gene expression.] The other tail is essentially unexplored.”