CONVERSATION · KARI STEFANSSON
January 13, 2003 | Interviewing Kari Stefansson — the profane, philosophical, and hot-tempered founder, president, and CEO of Icelandic genomic company deCODE Genetics — is not for the faint-hearted. His intense irritation over a clattering laptop almost prematurely curtailed our recent conversation. Once calm was restored, however, Stefansson discussed his company's latest efforts in tracking down genes associated with dozens of common diseases. With government cooperation, deCODE utilizes a genealogy database of most of the country's 280,000 citizens to track disease patterns and then analyze DNA samples from affected family members. DeCODE says it has pinpointed genes associated with 20 major diseases, including schizophrenia. But Stefansson is feeling the heat as deCODE's stock price hovers around $2 from a high that tipped $30 a share since it went public in July 2000. In September 2002, deCODE cut 200 jobs and posted a third-quarter loss of more than $85 million.
Bio·IT World editors John Dodge and Mark Uehling interviewed Stefansson in the company's office in Waltham, Mass.
Q: What proportion of the population in Iceland has contributed DNA at this point?
A: About 80,000 people, or 33 percent of our adult population ... We have a facility that generates about 15 million genotypes a month. The biggest output in the States [is the Marshfield Clinic in Wisconsin] with about 8 million genotypes a year. [Soon] we will be generating about 30 million to 40 million genotypes a month. We are currently genotyping more than the rest of the world put together. And the reason we can do that is we have put together software systems that allow us to do automatic allele calling.
Q: Can you describe the software systems?
A: I have no interest in discussing software systems with you. If you want to see it, just come to Iceland. Don't ask me silly questions like this. I am not going to describe to you [expletive] software systems!
Q: If I was going to volunteer my DNA and I was an Icelandic citizen, how would I do it?
A: This is a fairly complex question you are asking. It is a question with innuendos that do not interest me.
Q: I was not looking for innuendos ...
A: They come naturally to you!
When we are working on specific diseases, we establish relationships [with those physicians] who are taking care of patients with a particular disease. They put together a list of patient names. That list is encrypted by [the Icelandic government's] Data Protection Commission. The encrypted IDs are sent to us. We call in a representative from the Data Protection Commission [who encrypts deCODE's genealogy databases using the same code]. The list [of patients] is cross-referenced with the genealogy database, and out come family trees with encrypted IDs. We send that list with the encrypted IDs back to the physicians. They call in a representative from the Data Protection Commission who encrypts our genealogy database ...
One day, the Americans are going to put together a system like that.
When the physicians get the list of patients who we feel could help us solve the problem, they call in the representatives [from the Data Protection Commission] who decode the list. The physicians who have names send out letters to patients, asking them to participate. If they are willing, they are called in and given information on that study and [asked if] they want to sign an informed consent form. The tubes [of blood contributed by consenting patients] are labeled with the same encrypted IDs that we have before, and the tubes are sent to us. This is a fairly arduous process.
Q: What is the percentage of those not consenting?
A: Nine percent.
Q: Could you do what you're doing in any other country in the world, given the privacy and social barriers?
A: Yes, I believe we could do this in any country in the world.
The notion that patients and relatives are concerned about privacy when it comes to biomedical research is simply not correct. I've worked in [the United States] for 20 years, and I don't remember a single study where I didn't get at least 90 percent participation rate. Patients and their relatives are concerned about focus and research on the disease. They want to see research progress. One of the interesting things is that when patient organizations visit our company, the first thing they complain about is the complex third-party encryption system to protect privacy.
Q: DeCODE is also a drug discovery company. Is that a new direction, or has that always been part of your charter since deCODE's inception?
A: The question is, how are you going to turn these discoveries into monetary value? The idea was always to isolate disease genes to use them on [their] biological pathways ... The idea has always been to market some of this to big pharmaceutical companies and then take some of this and develop it ourselves.
Q: Where does F. Hoffmann-La Roche Ltd. (which signed a $200-million deal with deCODE in 1998) factor into the equation?
A: Roche is just one of these pharmaceutical companies that we want to work with. Roche has absolutely no influence on what we do except that they pay some money for the things we do. We feel that they should pay more for it, but they are not of the same opinion. We will be signing more deals in the future with big pharmaceutical companies, but that is not necessarily the goal. The goal is to bring products to the market. We will have pharmacogenomics tests on the market within two years. It's going to take us seven to 10 years to get drugs on the market.
Q: Are other partners on an equal standing with Roche?
A: We don't have other partners on equal standing with Roche. We are working with Pharmacia [Corp.], which will now be part of Pfizer [Inc.]. We're working with [Applied Biosystems Group] and Applera [Corp.]. We sold them a software system to do allele-calling on their machine that is 10 times faster than the system they are selling now. We are working with Vertex [Pharmaceuticals Inc.], and we will be signing more deals in the future with big pharmaceutical companies (deCODE has since allied with Merck & Co. to find treatments for obesity — Eds.), but that is not necessarily the goal. The goal is to bring products to the market.
Q: How much will you rely on partners?
A: The issue is how we are going to survive during these long years until we have significant products on the market. That is the fundamental question. We are going to do it in two ways. One is to work with partners like Roche so they take most of the risk and they also get most of the upside when there is a drug on the market. The other is by selling services. We are selling medicine and chemistry services. We are selling pharmacogenomics studies. We are selling clinical trials. We are selling software systems that function as services. What you have is two ways of sustaining ourselves during the time that it takes to develop drugs for the market. We are going to break even next year because we have to break even.
Q: What kind of impact did the downward spiral of ImClone Systems Inc. have on drug discovery?
A: I don't think that ImClone had any impact whatsoever. How can I put this? There is a problem for our industry and a problem for most industries because of the place where the market is now. It is much more difficult to raise money. It is particularly difficult to raise money for early development because the market does not seem to have any interest. I'm convinced this is going to have a very negative effect on biotechnology in the drug industry.
Q: Did ImClone [create this situation]?
A: I don't think ImClone [did]. [Former ImClone CEO Sam Waksal] was one loser [who was] less than candid. He was not particularly [effective] in the way he responded to tremendous stress. I don't think that means he is a bad man or particularly dishonest ... he just cracked. There are many other things that have happened in biotech and in the pharmaceutical industry across the board that have undermined the confidence.
Q: How important is deCODE's Neuregulin 1 finding in terms of treating schizophrenia? (See American Journal of Human Genetics, October 2002, vol. 71, pp. 877-892. )
A: We have now reproduced this in a lot of populations. It is the most solid piece of work of the genetics of complex diseases so far. It's a very, very difficult disease.
Q: What percentage of schizophrenia cases does Neuregulin 1 affect?
A: Of the haplotypes that we have pulled out, it's about 30-some percent of patients with schizophrenia where we can [trace] the abnormality to Neuregulin 1.
Q: Will there be a drug out in seven to 10 years that might exploit this discovery?
A: I don't know. Let's say that is at least seven to 10 years out. The beauty and importance of this is if you look at the medications that are currently used with schizophrenia, they are not likely to influence the disease process. They attack [only] the symptoms. Schizophrenia is indeed a progressive disease. It's a disease where the patient deteriorates over years. The beauty of this discovery is that it may lead us to not only attack the science and symptoms, but to influence the focus of the disease.
Q: How far along are you with developing diagnostic products?
A: DNA-based diagnostics is a whole sort of new chapter because these are the tests that allow us to measure the probability that someone is going to develop a disease. I'm absolutely convinced there are tests that are going to dramatically [change] the health-care system — shift it from intervention to preventive health care.
Diagnostics is always going to be half of therapeutics, because you have to be able to diagnose before you can treat. I think that DNA diagnostics are not going to lead us into any kind of genetic determinism. They are not going to decrease our freedom — they are going to increase our freedom.
Q: What have you learned about aging in the Icelandic population?
A: Aging is inherited. [There is] an allele that confers the risk of becoming more than 90 years of age, but we have not isolated it. That is one thing we can do in Iceland that you cannot do anywhere else.
We took individuals over 90 years of age and asked, 'Are they related more to each other than the control groups?' It turned out people 90 years of age were much more related to each other than the control groups. When you are looking at an entire nation, it basically means there is a basic component to [longevity] ... Once you are 65 years of age, if one of your parents became 90 years of age, you have statistically significantly less probability of dying in the next year than someone who had parents who died before the age of 90.
Then the question comes up, what the hell have you inherited? Are you ducking the disease genes, or are you inheriting a positive asset? If you are avoiding the disease, it would have to be a very complex condition because the disease genes are so many. So it looks like in Iceland that, when you look at population of genomics, longevity can be conferred by a single gene.
We have mapped the gene ... and are in the process of [isolating] the gene. If you stepped in front of a truck, the longevity gene is not going to protect you. If you have a mutation of BRCA1 and BRCA2 [breast cancer genes], the longevity gene is not going to help. So what we are talking about here is a gene that sends a very strong signal. It's extremely difficult to isolate a gene like that.
Q: How close are you to pulling out the gene?
A: I don't know that. I've so often been wrong. I have always predicted that we would be much quicker at isolating the genes than we have been. That's not because I am trying to mislead people. It's just that I am so stupid.
Q: If you isolate that gene, could you create drugs to disrupt the aging process?
A: Behind this is the question of death. There is a playwright by the name of William Shakespeare, who was talking about Julius Caesar, who had a premonition that he would be killed that day. And he answers with the following passage (gaps filled in courtesy of "Bartlett's Familiar Quotations" — Eds.):
Cowards die many times before their deaths.
The valiant never taste of death but once.
Of all the wonders that I have heard,
It seems to me most strange that men should fear;
Seeing that death, a necessary end, will come when it will come.
... The longevity [gene] is not going to change that. You're going to die!
PHOTO CREDIT JOHN SOARES