Professor of Molecular Pharmacology and Experimental Therapeutics and Internal Medicine
Mayo Medical School and Mayo Clinic
Name: Richard Weinshilboum
Position: Professor of Molecular Pharmacology and Experimental Therapeutics and Internal Medicine, Mayo Medical School and Mayo Clinic in Rochester, Minnesota
Education: MD, University of Kansas
Residency in Internal Medicine, Massachusetts General Hospital
Richard Weinshilboum gave the introductory lecture opening a four-speaker symposium at the American Association for Cancer Research in Anaheim, Calif., last week. With a focus on breast cancer, the symposium supplied a primer on the state of pharmacogenomics in cancer research, and its movement away from one-gene approaches to larger, multi-genic and genomic methods.
We called up Weinshilboum this week for overview of the session and an update on the activities of the US National Institutes of Health's Pharmacogenetics Research Network, in which he currently serves as a network researcher and as chair.
The PGRN database at Stanford University the Pharmacogenetics-Pharmacogenomics knowledge base or PharmGKB can be found here.
Can you tell me a bit about the opening-day symposium you and four other PGRN members took part in?
The symposium was meant to highlight both pharmacogenomics as a discipline in this case because it was at the AACR meeting, with a focus on topics that related to cancer and to highlight this national effort by the National Institutes of Health to put together an interactive collaborative network that will collaborate not just with members of the network, but with groups outside, with a focus on pharmacogenomics.
Now, several of the centers within the network have a focus on pharmacogenomic issues that relate to the treatment of neoplasia and they were highlighted particularly at the AACR meeting.
My introductory comments in my presentation utilized the thiopurine methyltransferase polymorphism, which has become a prototypic example with cytochrome P450 2D6, and that's not just my selection, those are two examples that were selected by the Food and Drug Administration in their initial draft, and then final guidelines as valid biomarkers for pharmacogenomics to emphasize where the field has come from. What you heard at the symposium in presentations that followed mine were that, in addition to these classical examples of monogenic genetic variants that affect pharmacokinetics. That is, metabolism or drug transport. Increasingly and I think this was highlighted by Howard McLeod from Washington University in his presentation […] we are thinking about the application of pharmacogenomics well beyond the single-gene variants, and are focusing on pharmacokinetic and pharmacodynamic pathways. That is, genetic variations involving all the genes that encode for the proteins that have an influence on the final drug concentration to which the target is exposed. That's the pharmacokinetic side. On the pharmacodynamic side, there is the drug target itself, and [everything] upstream and downstream signaling that is associated with drug effect on that target. Increasingly I think we'll be seeing that type of approach supplemented and informed by the application of a variety of genome-wide screens to try and give us opportunities to learn about pharmacogenomics effects that are beyond what we currently understand at the pK or pD level.
I think it should be emphasized that the focus in pharmacogenomics, I think, is placed upon the translational applications. There's nothing like basic science, and the kinds of things sponsored by the network involve both basic pharmacogenomics, and applied or translational pharmacogenomics, so a good deal of our work in our center here at Mayo involves gene-resequencing functional genomics, and then the rapid application of the outcomes of those high-throughput resequencing efforts and the depositing of all these data in a publicly available database at pharmGKB. To try to understand the functional implications of the genetic variation, our focus has been on non-synonymous coding SNPs, and what's become clear -and what continues to surprise people is that a major mechanism by which these open reading frame-based SNPs have their functional implications is by altering the quantity of protein, not by altering the protein function, although clearly that can occur.
What I'm basically talking about is, for example, the fact that in subjects who have two copies of the most common allele for thiopurine methyl-transferase the so-called *3A variant that we first described in 1996 that those individuals have virtually none of that protein in any of their tissues. The Mary Rellings' laboratory at St. Jude's [Childrens' Research Hospital] demonstrated that that was because the protein was rapidly degraded. Liewei Wang in our laboratory in 1993 showed that that's probably because of protein misfolding and chaperone protein recognition of the misfolding protein. And work that I presented at the AACR, and which is in press at the Proceedings of the National Academy of Sciences, has demonstrated that in addition to misfolding leading to targeting for degradation, one can have the formation of protein aggregation and aggresome formation for the *3A variant of TPMT. Now, we demonstrated repeatedly for a number of genes, that for changing a single amino acid out of hundreds in a protein, the most common mechanism on a genetic basis based on common polymorphisms [leads to] a decrease in quantity of protein. Most commonly as a result of rapid degradation and clearly, only for TPMT in pharmacogenetics do we have good evidence of protein misfolding. But I think that may well prove to be a common mechanism.
At the symposium, Erin Schuetz from St. Jude's at a basic pharmacogenomic level had a presentation with regard to alternative splicing and pharmacogenomics. Howard McLeod was talking about pathway-based high-throughput techniques, as applied to the translational side. And there was a presentation that came out of a University of Indiana group led by David Flockhart, which involved the application of pharmacogenomics to understanding an individual's response to tamoxifen.
So, we had a couple of translational presentations, a couple of relatively more basic presentations highlighting work that comes from this network, but with an emphasis on cancer the thiopurine drugs are a mainstay in the treatment of acute lymphoblastic leukemia of childhood, tamoxifen [in the] treatment of breast cancer, the treatment of colon cancer, and mechanisms that affect alternative splicing the cytochrome P450 3A4 and 3A5.
Who makes up the network?
Mayo is one of centers of the network I'm currently the chair of the steering committee for the network. The University of Indiana that PGRN site is led by David Flockhart. Howard McLeod is the principal investigator for Washington University center. And Erin Schuetz is at St. Jude, and that group is led by Mary Relling.
How close are we to seeing pharmacogenomics in the clinic?
The treatment of neoplasia has been one of the places where pharmacogenomics has had a major impact already. It was a demonstration of the principles the reasons are that this is a life-threatening or life-taking illness, and the drugs that are used typically have a narrow therapeutic index that is, the difference between the therapeutic dose and the toxic dose is quite small. An example is ironotecan, one of the drugs that Howard McLeod was talking about, which is metabolized to form an active metabolite, [which is] then itself inactivated by further metabolized and inactivated by glucoronidation. One of the isoforms that participates in glucoronidation is UGT 1A1. There is a common genetic variation in the promoter for UGT 1A1 that means that some people, perhaps 10 percent of the Caucasian-European population, will be homozygous for the variant within the core promoter for that gene, such that they make less UGT 1A1. Therefore, they are less able to metabolize ironotecan. Side effects of that drug are both myelosuppression and life-threatening diarrhea, and there is increasing evidence that those side effects are related to the UGT 1A1 polymorphism. Both TPMT and thiopurines and ironotecan and UGT 1A1 were selected by the FDA for public hearings with regard to the inclusion of pharmacogenetic information in labeling. TPMT [was] first, ironotecan a lot of that work came out of Mark Ratain's lab at the University of Chicago (he is also a principal investigator for one of the PGRN centers).
The fact that the therapeutic index for anti-neoplastic drugs is so narrow, has meant that this has been an extremely valuable test bed for the principles of pharmacogenetics. What we're seeing is increasingly that those principles are moving beyond single genes and are focusing on the entire pathway by which a drug is activated, inactivated, or has its pharmacodynamic effects.
The treatment of cancer has been an important area for testing the principles of pharmacogenetics, developing our thinking regarding pharmacogenetics and pharmacogenomics, and clearly based on the kinds of presentations made at the AACR meeting will continue to be a major area for these principles as they evolve [to address entire genomes].