Professor of Medicine in Psychiatry and Biopharmaceutical Sciences
University of California, San Francisco
Name: Neal Benowitz
Position: Professor of Medicine in Psychiatry and Biopharmaceutical Sciences, and Chief of the Division of Clinical Pharmacology and Experimental Therapeutics, University of California, San Francisco
Background: Faculty member of the Univerisity of California San Francisco, 1974 — the present
Education: MD, University of Rochester School of Medicine, 1969
As the principal investigator of the Pharmacogenetics of Nicotine Addiction and Treatment project within the Pharmacogenetics Research Network, Neal Benowitz is heading one of the three new projects included in the PGRN during this second round of funding by the US National Institutes of Health through its member institutes.
The National Institute of General Medical Sciences heads the PGRN, and alongwith the US National Cancer Institute and the US National Institute on Drug Addiction is providing Benowitz' project with $10 million. NIDA is a recent addition to the list of the PGRN's funding organizations, and according to Rochelle Long, chief of pharmacological and physiological sciences at the US National Institute of General Medical Sciences, its participation signals the agency's increased confidence in the network's research.
Last week, Pharmacogenomics Reporter spoke to Benowitz — who is also principal investigator of the Pharmacogenetics of Nicotine Addiction Research Consortium and Program Leader of the Tobacco Control Program of the UCSF Comprehensive Cancer Center — to learn more about the details of pharmacogenomics in nicotine addiction.
Ten million dollars provided by NIDA, NIGMS, and the NCI for finding genetic variation related to differences in response to medication for treating tobacco dependence.
What are the medications for treating tobacco dependence, and what are the reasons for thinking they might have a genetic basis?
Well, currently there are two medications that are [US Food and Drug Administration]-approved, two classes. One is nicotine itself, in various forms … and then bupropion, or Zyban, which was originally developed as an antidepressant, but which is also effective for smoking cessation. There are also effective, but not-FDA-approved drugs: Clonidine, which is a drug used for treating hypertension; and Nortripiline, which is another antidepressant. So those have been shown in clinical trials to be effective, but are not FDA approved.
And then on the horizon, there are three treatments that are being investigated in phase III trials. One of them is Vereniclean, which is a nicotine partial agonist — so it has some nicotine effects, but it also blocks nicotine effects. And then there is a drug called Ramonobont, which is a cannabinoid-receptor blocker that actually works on cannabinoid circuits. And then the last thing that's under investigation is a nicotine vaccine.
So those are the treatments. The reason that we think that genetics may be involved is that if you look at twin studies on smoking, smoking prevalence itself is heritable. It averages about 50 to 60 percent. If you also do subanalyses and look at how much [people] smoke, how much they smoke is heritable, and the likelihood of quitting in those people who are smokers is also heritable to a high degree. So there's good evidence that's there's heritability in smoking behaviors and being able to quit.
Based entirely on twin studies?
Then there has been some preliminary data with different genes in association with nicotine addition and nicotine addition treatment. There have been a number of studies looking at dopamine genes and various genes involved in breaking down norepinephrine and dopamine. And also on nicotine metabolism genes — in fact the strongest evidence comes from genes for the liver enzyme CYP2A6, which is involved in nicotine metabolism. And studies have shown that people who are genetically slow metabolizers tend to smoke fewer cigarettes because they need less nicotine intake to have a certain level, because they metabolize more slowly. There's one study that suggests that if you are a genetically slow metabolizer, you are more likely to transition from experimentation to dependence. Plus other studies say if you are a slow metabolizer, it's easier to quit.
So there's a bunch of data that have been collected over several years relating the 2A6 gene to smoking behavior. And then there have been a few studies of nicotine treat looking at specific candidate genes. And some that have shown promise have been the dopamine-2 receptor gene, the opioid mu receptor gene, the COMP gene, which is a catechol-O-methyltransferase that is involved in breaking down catecholamines, and then one study looking at the CYP2B6 gene.
Dr. Swann and I did a big twin study on nicotine metabolism, and found that if you look at that, it's about 60 percent heritable. And then we also found that if you look in Causcasians at the current known variance of the 2A6 gene, that accounts for only a small amount of the heritability. So we think there are still more genes to be discovered.
But we do have a non-invasive marker of the rate of nicotine metabolism, which in nicotine patches was a strong predictor of response. So people who are slow metabolizers did much better than people who are fast metabolizers — presumably, fast metabolizers are not getting enough nicotine dose.
So we have a number of lines of evidence that make us think that certainly nicotine metabolism is probably involved, and then perhaps some of the other genes involved in nicotine response — dopamine genes, opioid genes, others — may also predict response. But to look at any single candidate gene is sort of simplistic, because people's behavior is complex, and is determined by the effects and interactions of many genes. So the approach of our project is to take existing clinical trial data, where we have DNA samples, and do extensive genotyping — high-throughput genotyping and extensive genotyping of maybe 200 genes, we're not sure how many, but something like that — and then do Bayesian-type modeling to try to identify gene clusters or gene interactions that predict response.
So [this] is really a collaborative project with UC San Francisco, where I am; SRI International, where our genetics core and bioinformatics core are; our statistics core is at USC; and then we also have a clinical core at Penn; a clinical core at Nebraska; and a genetics investigator at Toronto.
Have you finalized the list of genes to investigate?
No. We have a list of about 20 to 30 genes that have been looked at by other people for nicotine addiction, but we're exploring pathway analysis, biological plausibility studies, microarray expression studies, and we're going to try to expand that list to look at a number of candidates that haven't been looked at by other investigators.
Have you already got the equipment you need to carry out this project?
That's going to be done through the Illumina platform. For some of them, like the drug-metabolizing genes, they're not amenable — they're too complex to do by the Illumina [platform]. So those are going to be done by more traditional TaqMan-type approaches.
For specific genes, we probably will. It's not clear which ones we're going to do that for yet.
How many patients' samples are you going to be looking at?
Between 4,000 and 5,000. We're taking pre-existing studies that have been completed or are ongoing, for which we have DNA samples. This large approach, we think, will work because we have a large number of samples, which you need when you look at multiple genes and gene interactions. So it's power that, if we can look at several thousand patients, we can get pretty good confidence in finding associations, if they exist.
How can the results be turned into molecular diagnostics?
The idea is that if we could have gene clusters that would predict which medications a person would respond to — and so we have some trials of one medication versus another, that would help select the drug — and then we might have drug metabolism genes that will help us predict the dose that's required.
So, in theory, we could screen for a certain number of genes and identify the best drug and the best dose for the patient. That would be the ultimate goal.
Whom does the intellectual property belong to? Are NIDA, NIGMS, and NCI going to own it, and then license it out from there?
That's my impression.