Mapping out the future

Pharmacogenomics is the stepping stone to the promise that, in the not too distant future, drugs could be tailor-made for each person. There are other influences, such as environment, diet, age, lifestyle and state of health that can affect an individual’s response to medicines. But by understanding our own genetic makeup, the possibility is there to create personalized drugs with greater efficacy and safety. But just how close is wishful thinking to the reality? Jonathon Edgley spoke to Dr Allen Roses, Senior Vice President of Genetics Research at GSK, to find out more.
JE. What exactly is ‘personlized’ medicine and what promises does it hold?
I don’t use the expression ‘personalized medicine, I think the correct term would be ‘informed’ medicine.

AR. When lay people read about personlized medicine they get the impression that medicine is being personalized for them. In fact, what we’re doing for adverse effects is identifying individuals who perhaps shouldn’t take a drug. We’re also looking at groups of individuals with some prediction about their level of efficacy.

JE. With this in mind, would you be able to provide a little more detail as to what pharmacogenetics is and its role within personalized medicine?

AR. Pharmacogenetics has been around as a discipline for a while. With the genome elucidated it allows us to use it in more generally applicable situations. It’s primary use falls into one of two fairly distinct categories. First is adverse events or side effects. In that category, what we want to know is not what the percentage of people who may have an adverse event is, but whether this individual would be someone who would suffer an adverse event – could we make that type of prediction?

For efficacy, the way it usually works is you look at the medicine. If it works, fine; if it doesn’t, you take the next medicine. And if that doesn’t work, you take the next one. Therefore, we can make a prediction based on the patients genetics and the relationship to that particular drug, and the information we have about it during development – which drug it is that the patient would most likely respond to.

JE. How is pharmacogenetics currently being used at GSK?

AR. In order to do this during development, the single most important thing you can have is the DNA from the person who is taking part in the drug trial, as opposed to looking at it and then trying to get the DNA retrospectively. When we enroll people in drug trials, for the most part – and certainly for the last couple of years – we’ve got informed consent and IRB approval for doing pharmacogenetics on nominally all of our drug trials.

JE. Is it evolving in the way you thought? How close is it to being an ‘everyday’ experience – what is there left to come?

AR. You ask when is it going to have an effect – at GSK it already is! We have several examples of both adverse events and efficacy, which we’ve actually presented publicly to demonstrate it is working – it’s not a question of when.

Let me give you an example. We have a drug in development for cancer therapy. Cancer therapies tolerate adverse events better than most drugs because it’s such a serious indication. What we did is look at our phase 1 and phase 2a patients. Phase 1 are normal subject volunteers who get the drug and we look at safety. Phase 2a are the patients we are working with and looking at the potential for efficacy.

We added together the initial phase 1 and phase 2 sub studies that we collected. We had 107 patients, of which 15 percent (16 patients) had either diarrhea or skin rash, or both. For the most part, those are tolerable kinds of things, but what we did is ask the question ‘could we identify whether the expectation of skin rash or diarrhea was there before we gave a drug’. We took some of the candidate genes, including metabolizing enzymes, because we know how the drug is metabolized from our preclinical development. We then did very dense snip maps of these genes.

In general, we knew how the drug was metabolizing and which were the major and which were the minor metabolizing genes. We took these five genes that we had and had snips about 1000-1500 bases – very tightly mapped snips within these genes. We then carried out association studies with the patients that had the adverse events compared to those that did not. We found a very interesting phenomenon: every single patient who had both diarrhea and rash, or one or the other, had abnormalities identifiable – not mutations but polymorphisms – in one enzyme. There is a polymorphism in that gene that is called a Star 2 mutation and all 16 patients had the Star 2 mutation, and none of the other patients had it. The diarrhea was so severe for two of the normal subjects and one of the phase two patient subjects that they had to withdraw from taking the drug. Those three patients all had two copies of the Star 2 mutation.

Now we know that for people who carry Star 2, we need to make some adjustment as to the doses they receive.

JE. Are dietary and lifestyle behaviors still likely to play a role and affect the safety and efficacy of medicines for particular individuals, or will pharmacogenetics make this an irrelevance?
AR. Sure, they run hand in hand, but there is no such thing as a general diet or lifestyle. Basically, there are going to be various interactions with the environment like smoking and eating – the kinds of things people do irrespective or the medicine they’re taking for a particular abnormality.

You know if you don’t take your drug it isn’t going to help you. Those kinds of things are always going to be interactive, but the measurement of those variables in a scientific context is a lot more difficult than the measurement of genetic variations. What we’re going to get are genetic variations that give us some predictability. It is a much more exact science than pleading with lifestyle or dietary type factors – though they too could be measured, but not with the same degree of exactness.

JE. What are the major obstacles, if any, to the continuing development of pharmacogenetics? For example, I read a short piece questioning the current regulatory environment’s ability to support the genetic testing pipeline and achieve the promise of pharmacogenetics.

AR. I think the regulators are way ahead of the companies. Both the FDA and the EMEA are well aware of the data for pharmacogenetics, much more so than many of the companies who are not actively involved in doing these studies.

We are probably the leading company in doing this and we’ve presented numerous cases, like the one I mentioned previously, to show what it can do. Very few other companies have that degree of involvement but the FDA is very aware of this. For instance, when you want to present or use these kinds of data in the context of a registration package for a drug there is a voluntary genomic data submission. This is not a decision-making meeting but a discussion with the FDA on how you would propose to use it so they can ask questions. It prevents the situation arising where you go in with a hard and fast “this is what we’re going to do” and they come back at you with a “we don’t know what you’re talking about”!

JE. Despite its promise, some people are raising ethical concerns about pharmacogenetics: invasion of medical privacy, unequal distribution of benefits amongst them. Do you acknowledge these and how do you counter them?

AR. There are always ethical concerns. My biggest ethical concern is why would we continue to give a medicine that would have a known rate of adverse events and not have the data to make them predictable. In an age where we can do that, is it ethical not to? In an age where we can tell that a certain group of patients will respond to a drug while others may just have adverse events, is it ethical not to do the testing?

This is not about the worries some people might have that some specific mutations that are inherited might be measured and therefore they would get a diagnosis they don’t want to get. There are so many snips in the genome that the ones that we use are not involved in those kinds of diagnosis.

JE. Finally, what would you say is the most important thing people need to understand about pharmacogenetics?

AR. Pharmacogentics is going to affect the safety and predictable efficacy of drugs. In countries (including most European countries) in which there is a reluctance on the part of the governments to provide drugs on their formulary, if you can predict who is most likely to respond then the differentiation of these drugs will make them available on the formularies in a much wider context. But if a drug only works on 10 percent of the population, countries have decided not to put that kind of drug on their formularies.

If you could pick out that 10 percent then that’s a differentiation that will allow modern drugs to get to more people and it would be those people it would help.