Klaus Lindpaintner has coordinated Roche's efforts in genomics, genetics, and proteomics since 1998, when he moved there from Harvard Medical School. Senior Informatics Editor Malorye Branca discusses pharmacogenomics with him.
Q. Has pharmacogenomics evolved the way it was anticipated?
A. There is a growing sense that pharmacogenetics is important. But it won't be applicable to every medicine or every disease. We need to be aware of that fact and keep our eyes open. The slogan "the right medicine for the right person" is a huge overstatement that makes it sound black and white. In reality, it won't be like that at all.
Q. So what will it be like?
A. Medicines will be more likely to work in some patients but less likely to work in others. In the future, with pharmacogenetic tests, we will become gradually better in predicting likely responses to a drug, but these tests will not give us black-and-white answers, rather probabilities. With every medicine you develop, you're playing the odds and hoping a significant number of patients will benefit and that you can easily identify them. But there are so many other things besides genetics that play a role — like diet and age. It will not be straightforward.
Q. Will pharmacogenomics be most useful to identify responders, or to avoid adverse events?
A. I don't think we will see the earliest benefits with adverse events, with the one exception: drugs that address lethal illnesses. For example, with abacavir [an anti-HIV drug from GlaxoSmith-Kline], there is a 6 or 7 percent incidence of serious adverse hypersensitivity, and they have just shown they can cut that in half by genotyping.
Q. What is the most common way you use pharmacogenomics at Roche?
A. To avoid the "surprise" factor. You're making a drug, you put it into people, and suddenly you realize some patients aren't responding. If this happens at the end of Phase II it's a nightmare. You have a whole program ready to roll out that is suddenly stopped in its tracks. To avoid this, you want to front-load as much as possible. By studying the mechanisms of the drug you may be alerted early on that it may work better in a subpopulation, and you can initiate genetic studies then. That was the case with Herceptin [Genentech Inc.'s anti-cancer drug, which is marketed by Roche outside the United States]. We had a sense very early on that it would work only in a subpopulation.
We also try to collect samples for genotyping during every trial. Nine out of 10 times we don't need them, and they stay in the freezer. But if we run into a nasty surprise, we have the samples on hand and can investigate. For medicines that make it to market, we keep the samples in case we need to validate or invalidate any subsequent claims made about the drug. At the other end of the spectrum, in basic research, we are looking for new mechanisms, such as through our [$200-million] deal with deCODE Genetics. There, we usually start right away studying genetic variations.
Q. Are you finding interesting new mechanisms through genetic studies?
A. Here's one interesting example: 1 to 2 percent of type II diabetes patients have a condition called MODY (mature onset diabetes of the young) that is clearly inherited. These patients have a mutation in a gene that produces a glucose-metabolizing enzyme called glucokinase. We've been able to find chemical compounds that can activate this enzyme and lower glucose levels in patients with MODY. It turns out that these compounds not only correct the cause of diabetes in MODY patients, but also lower blood sugar in anyone, so they could be used broadly for type II diabetes.
Q. How much of your budget do you devote to pharmacogenomics?
A. We look at genetics as one tool in the toolkit, no more than that, so we are investing cautiously. It's currently no more than 5 percent of our total R&D expenditure, but that's no pittance, either. There are some opportunities we feel we cannot afford to miss, but there are many other things that are as important as genetics, like toxicology and pharmacokinetics. It needs to be taken into perspective.
Q. What's the most important thing for people to understand about pharmacogenomics?
A. It's no revolution, no paradigm shift. It's just part of the same slow, evolutionary shift in the way we find medicines. Penicillin made a huge difference to medicine, and people keep expecting another magic bullet like that. Pharmacogenetics is just one more step down the path of medical progress. I warn people against all the hype and excitement surrounding the Human Genome Project. By hyping it up, we are creating expectations that won't be met, as well as uncalled for fears and concerns because people get misconceptions about what it all means.
Q. How has genomics affected drug discovery and development overall?
A. Show me one medicine that's been developed in the last 10 to 15 years in which the spectrum of molecular biology and molecular genetics — all these new biotechnologies — have not had a significant impact. It's used to design screening assays, to create recombinant molecules, and in many other things we do every day. You simply can't do drug discovery and development without molecular biology and genomics. The tools aren't the problem. The real reason some pharma company pipelines look dire is that before, we could lean on targets other people had worked on. Now, as we embark on our own targets, we must take extra steps to ascertain they are valid and druggable. We can't force it, and you can't turn it on like you were flipping a switch. It is hard work, diligent science, dedicated research, and you can't do it overnight.
Back to The PATH to Personalized Medicine