Defining pharmacogenetics

Whilst recently watching Sir Paul Nurse present an episode of Horizon, he touched on a very interesting issue in science and science reporting, that of uncertainty.  Quite importantly he makes a distinction between the two principle types of uncertainty in science, that of the uncertainty of knowledge at the beginning of a research project, and that of a probabilistic nature.  The example he uses for the second of these two distinctions is that of a treatment for a disease which may work for an individual, but when applied across a hundred patients 20 will respond and 80 will not.  Straight away I thought about how that applies to my own research in pharamcogenetics.  I had a little epiphany of sorts.  Pharmacogenetics seeks to reduce that uncertainty in treatment across populations of patients to a point where it becomes clinically negligible. 

For a while I’ve been struggling to find a simplistic, but accurate, way of explaining what exactly pharmacogenetics is; thanks to the wonderful Sir Paul Nurse I now have that explanation.

A first attempt at science writing for the public

Recently I entered a writing competition held by an internet forum I frequent, and I would like to replicate the entry here.  The subject of the essay was to debunk or refute a popular misconception in science; I went for genetic determinism and the "Gene for X" fallacy. 

Please feel free to add your criticisms and comments, afterall, I know it isn't perfect and I'm always looking for opportunities to improve my writing.  The criteria for the competition can be found on RationalSkepticisn.org.

Not in my Genes! - A common misconception in human genetics

Setting the scene

In Andrew Nicol’s 1997 film Gattaca the protagonist sets the scene of a dystopian society ruled by genetic determinism. What makes the film so terrifying to many is the beginning sequence that describes this none-too-pleasant world as “...the very near future...” suggesting the hard reality of a world ruled, and our lives defined, by our genomes.
Vincent, Ethan Hawke’s character, we soon find out, is part of an underclass of citizens defined not by their socioeconomic status, but by the very chemical they owe their existence too. As our hero states with frightening indifference, they have discrimination down to a science.

Gattaca shocked audiences when it was released and prompted many to question the ethics of human genetics research. I aim to show why those fears are misplaced, and dispel the myth of genetic determinism, including its cousin, the common misconception known as “The Gene for X”. Hopefully along the way I will be able to shine a little light on the beautifully complex world of how our genomes shape our very form, and our fates...

I often hear or read about genetic discoveries in the media accompanied by the catchphrase “scientists have discovered the gene for [X]”. In fact just whilst writing this I’ve heard this same old phrase trotted out with latest research on a gene called DRD4 and an apparent link with promiscuity in human adults1. So apparently, and this is only according to the media I must stress, if you have a particular variant of this gene you are more likely to have problems with fidelity! Cue the many claims by less-than-faithful partners “but my genes made me do it!” It is this kind of reporting that misleads the public about how our genomes really work. Worryingly it is not just the public that are suckered in by this grossly simplistic view, many practising doctors have only a rudimentary understanding of how our genetic make-up really works, and are thus likely to pass on this mis-information to their patients.

The bottom line is that our genetic constitution does not define us; it influences us to varying degrees. At this juncture I must point out that this applies to all aspects of genetics, including what are called Mendelian disorders; the classically defined genetic diseases such as cystic fibrosis and sickle cell anaemia. I know what you’re thinking – “He’s lost the plot! I know that genetic mutations cause these diseases!” and of course you would be correct, however, there is a very important, if somewhat subtle detail that must be taken into account. A little background in molecular biology will hopefully shed some light on this conundrum.

Genetics 101

There is phenomenon or concept, if you will, in molecular biology that explains how information encoded in a DNA molecule is translated to a functional protein. It is proteins that do most of the work in our cells and bodies. They turn genes on, and off, they help us breakdown foodstuffs and extract energy from what we digest, as well as controlling various aspects of our immune systems and how we grow as a foetus. Needless to say they are rather important. This phenomenon is known as the central dogma, and it states that information flows from DNA to a related molecule called RNA, before it is finally translated into a sequence of amino acids which are the building blocks of proteins. So we can say that the passage of information follows this sequence:

DNA->RNA->Protein

Whilst there are some important exceptions to this rule, information does not flow the opposite way from a protein to create a new DNA sequence. This is very important because if we change information in the DNA molecule that encodes the protein (i.e. a gene) we can change a part of the protein itself. Et voila! A mutation.

Mutations can occur in all sorts of places in DNA, and can subsequently have all sorts of different effects on not just the protein itself; it can change the shape of the protein, and thus how it works, or it can stop it from interacting with other molecules, or it can give it a new function entirely. As a result a mutation can cause catastrophic damage to our cells, and our bodies, leading to a genetic disease such as cystic fibrosis or sickle cell anaemia. What is very important here is to understand that each protein does not work in isolation; it fits into a very complex machine with hundreds, often thousands, of finely tuned interacting components. Much like a watch that relies on a number of different cogs turning at different rates, thus a cellular network of proteins relies on the careful regulation of the function of interacting networks of protein molecules. The area of biology tasked with understanding these interacting networks is termed systems biology, and is a relatively new player on the field of molecular biology. Previously scientists, and in fact they still do, study proteins and biological molecules in isolation, an approach called reductionism. This approach allows scientists to break apart complex problems and study them a little bit at a time. Unfortunately it this approach that has had a knock-on effect; our interpretations are also affected by this reductionist methodology.

So what has this got to do with genetics and genetic diseases? A few cogent examples may help to explain this. There are many mutations in a particular gene that cause cystic fibrosis (CF); the cystic fibrosis transmembrane conductance regulator, CFTR. Cystic fibrosis, like any disease, is a collection of phenotypes (symptoms) that manifest themselves together in a person because of a common cause, in this case a mutation in our gene CFTR. This is where we need to be careful; this gene does not cause CF. Specific mutations within this gene lead to the formation of multiple phenotypes that co-segregate together. That doesn’t mean necessarily that any mutation within this gene will always cause CF, in fact some mutations lead to a related, but less severe condition called chronic bilateral absence of the vas deferens (but only in males of course!). We find a similar scenario in sickle cell anaemia. Mutations in one of the proteins that make up haemoglobin, the protein that is responsible for carrying oxygen in red blood cells so that it can be fed to all of the tissues of our bodies, leads to the formation of stiff fibres in red blood cells that causes them to take on a characteristic “sickle” shape. These sickled cells do not carry oxygen to tissues and just clog up small capillaries due to their inflexibility. These cells get damaged and die, and also rupture small blood vessels, the result of which is anaemia and chronic internal bleeding. This condition is not the same in everyone though. There are a number of different factors that can affect how severe the disease becomes, some of which may also be genetic, but importantly some of these can be environmental.

In both of these scenarios mutations can be associated with the disease as a whole, but very rarely is the phenotype mapped 1:1 with the mutation; this is one of the major difficulties of molecular pathology. Each of these mutations are generally rare (the examples of CF and sickle cell anaemia are more common in certain populations for other interesting reasons that cannot be explained here), so what about the 10 million or so common variants we all carry around with us?

There are a number of different mutations that are so common within the human gene pool (or any species gene pool for that matter) that they are instead referred to as polymorphisms (meaning many bodies). Some of these have an effect on protein function and thus the observed physiological phenotype, whilst others may affect the regulation of a gene. Importantly some of these may have no function at all, a fine example of truly neutral mutations.


From ref[2]

The diagram above illustrates the link between how common a variant is and how large the effect it has on the phenotype, called its penetrance. Mutations that have a high penetrance are those that cause Mendelian diseases like sickle cell anaemia and cystic fibrosis. If a mutation does not cause a trait to be expressed and only influences it modestly then we say it has an incomplete, or lower, penetrance.

Enter the Post-Genomics Era

In recent years a hypothesis was developed within human genetics research that perhaps it would be possible to investigate the genetic component of complex characteristics, polygenic traits that are the summation of multiple interacting genetic variants all with small individual influences on the trait of interest. This trait may be a continuous trait such as adult height, or intelligence, or it may be a binary outcome such as the occurrence of a disease such as cancer or type-2 diabetes (called adult-onset diabetes). Thanks to technological advances the tool used to investigate this hypothesis is called the genome-wide association study; the aim being to investigate as much of the genome at once to try and capture any potential influence on the trait of interest.

Thousands of these genome wide studies have been performed in the last 5 years or so attempting to discover the genetic component of a wide range of diseases and characteristics. The result has been a mountain of associated genetic variants, with very little functional consequence to compare them to. It is though that partly this may be due to type I errors, false positive results. An alternative, more...biological explanation, is that because each variant has such a small impact on the trait or disease that changes in the way they are classified will be able to highlight clearer associations.

Another major issue with this common disease-common variant hypothesis is that very little of the variation between individuals can be explained with these genetic variants alone. Some have suggested that perhaps they have very little role to play - cue the Nature vs Nurture debate. Importantly it is how our environments and genomes interact that modulates any influence either may have on the outcome of a particular characteristic.

Anything for a quick buck?

No discovery would be complete without someone trying to make a quick buck or two from it. Genetics research is no different and in recent years several companies have sprung up that will offer you a read out of parts of your genome for the modest sum of $500. These companies copped a lot of flak recently from the FDA, not because of the potential problems with interpreting the results of the tests they sell, but that they are associated with diseases and thus come under the guise of medical diagnostics which require strict FDA regulation, i.e. the FDA wanted to police them so they attempted to use the flimsiest possible reason for classifying them as diagnostic tools. As a result these direct to consumer genetic tests have had a bit of a rocky infanthood for political, not scientific, reasons.

What about the scientific reasons?

Importantly, and this has been a little overshadowed by the FDA’s heavy handedness, what these companies are offering is not so far removed from the scenario we see moments after the birth of Vincent in Gattaca. They are providing discerning customers with a prediction of their health on the basis of their genetic make-up. Whoah! Hang on! Does that mean we have already entered our Gattaca-style world without realising? In short, no. In long, what these companies provide is a very crude prediction of the impact of individual impacts on the customers health for a range of different diseases, and a number of superficial traits, such as wet or dry earwax, ability to taste a bitter chemical called phenylthiocarbamide (PTC), and whether you have curly hair or not (I’m not sure if I need a genetic test for two of those), but hey! It’s just a bit of fun right? Well, yes and no. Yes because it helps to bring the complex world of human genetics into the public domain, but also no, because people who take these tests may take the results to their local GP and find out that their knowledge of these tests is inferior to their respective Wikipedia page.

Let’s take an example. A variant within an important gene involved in the functioning of our immune systems called IL7RA has been associated with a slight increased risk of developing multiple sclerosis. Now that sounds quite serious to be at an increased risk of developing such a debilitating autoimmune condition. No need to go and get a wheel chair quite yet, put the phone down, there’s no need to book an appointment with your GP, we need to look at the level of the risk first. In the study itself people who carried two copies of this particular polymorphism were 1.08-fold more likely to develop MS compared to a population of healthy individuals. One...point...zero...eight fold. The average risk of Joe Bloggs on the street is 0.70 in 100, i.e. a 0.07% lifetime chance of developing MS in the absence any other information that might increase your risk. Carrying two copies of this variant increases that lifetime risk to, wait for it...0.096%! Even these risks are dependent on your age and your ethnicity. Not all human populations even carry around the same genetic variants. Of course if you do end up developing MS you can be sure that those little mutations may, not will, have had an influence on when you develop the disease, and potentially how mild or severe it is.

This example is fairly typical of most genetic variants that affect our chances of developing a particular disease or reaching a particular height. Even if we manage to discover all of the genetic influence on diseases like MS we still won’t be able to give a 100% certitude that you will or won’t develop the disease because the genetic component probably only makes up 30-40% of the factors that influence risk of disease. Or for a continuous trait they only explain a portion of the variation between individual people.

Tying up Loose Ends

How does this relate to our “Gene for X” misconception? Hopefully the more astute amongst you will have noticed I have not said that a particular gene is the cause for any disease or trait, rather it is specific variants that we find associated with characteristics. So next time you hear that scientists have discovered the gene for coffee drinking, or some other equally inane human behaviour, take a second to recall that obscure essay you read on an internet forum once, and think about how likely it is that there is a single gene that controls whether you like your caffeine from coffee or tea.

All of our behaviours, characteristics and risks of disease are of course going to be influenced by our genetic make-up. I hope I have been able to paint a picture of how the summation of our genetic variation we carry with us plays such a major role during the course of our lives, but that it does not define who we are.

References
[1] Garcia et al (2010) Associations between dopamine D4 Receptor gene variation with both infidelity and sexual promiscuity. PLoS ONE 5(11): e14162. doi:10.1371/journal.pone.0014162

[2] McCarthy et al (2008) Genome-wide association studies for complex traits: consensus, uncertainty and challenges. Nat Res Genet 9, 356-369

[3] Zhang et al (2005) Two genes encoding immune-regulatory molecules (LAG3 and IL7R) confer susceptibility to multiple sclerosis. Genes Immun 6(2):145-52.


The other competition entries can also be found on RatSkep (this is starting to look like an advert for them), some of which were very impressive.  If you find the time they are well worth reading, I imagine nearly everyone who reads them will learn something new.

Personalized medicine - hype or hope?

Firstly, apologies for the unannounced month-long sabbatical.  I've recently moved house and set up shop in my new residence in Leeds, in addition to starting my summer studentship working on the wonderfully complex FCGR locus (its the copy number variation that makes it so complex btw).  All this in combination with a little bit of first-blog burnout means I've been far too lax of late.

So I thought I'd return from self-imposed exile with a little bit of a discussion of a short article published online in the Financial Times.  An article written by a GP in Glasgow in which she claims that personalized medicine is just a load of hype.  This got me thinking a little bit, is personalized medicine truly hype, or is their really hope for a health-care revolution?  Naturally I'm a little bit biased in favour of personalized medicine, afterall I'm due to start a PhD in pharmacogenetics this coming October, so there is a little bit of vested interest in the field, as I feel that it truly can revolutionize global healthcare.

So to the article in question first of all.  Dr McCartney expresses her opinion that the genomic medicine revolution is all hype because of its inherent uncertainty.  I'd say that is a valid point to make, after all the known genetic associations to date are largely of very small effect sizes; odds ratios in the range of 1.2-1.6 for most associated variants.  She also mentions that smoking is one of the biggest, if not the biggest, risk factor for developing lung cancer, and thus the environment has a very important role to play.  Another perfectly valid point.  So why am I so bothered about this article?  Well, it's because of what she doesn't mention, and qutie markedley leaves unsaid.  The interactions between our genome and our life-time environment are very complex and poorly understood.  That is not a valid reason to malign personalized medicine when it has not even reached its infancy.  The current known associations have very little clinical utility because of their small effect size and individual minor influence.  Their possible interactions, that is geneXgene interactions, have not even been fully investigated.  Who is to say that a panel of 50 genetic markers doesn't have clinical utility?  I'm not saying that it definitely will, but we cannot say until such time that this route is explored in a rigorous and meangiful way.

Dr McCartney goes onto point out the pitfalls of HER2 testing as a predictor of response to Herceptin (trastuzumab).  Lack of sensitivity and specificity in a single test that relies on histology is not a basis for lambasting all of personalized medicine.  Response to Herceptin is likely under the influence of other genetic loci, after all there are pharmacodynamic factors to take into account alongside the usual clinical covariates of disease stage, patient age, BMI, dosage, etc that must be taken into account.  Afterall, personalized medicine is about the individual thus we are going to need to incorporate as many individual factors into the equation as possible to individualise each treatment.

She concludes that we may run the risk of creating more problems if we rely solely on genetic determinants (that's a misnomer, but this is not the place to discuss genetic determinism), and that prevention is ultimately better than cure - well said.  However, knowledge of our genetic make up is potentially a very useful tool in our arsenal against disease because it may highlight disease susceptibilities that we can overcome by manipulating our own environments, i.e. our lifestyles.  Overall, I'd say that Dr McCartney has some valid points to make about what barriers we need to overcome to reach the personalized medicine era, others she does not mention (perhaps due to pre-publication editing), however, the article is simplistic and comes across as very naive and ill informed.  After all, where is the mention of Warfarin dosing?  The known genetic influence on hepatitis treatment?  The role of the cytochrome P 450's and drug metablolism, not to mention adverse pharmacological reactions?  These are some of the targets of personalized medicine, and pharmacogenetics.  For instance ~30-40% or rheumatoid arthritis sufferers fail to respond to initial biologic treatment with anti-TNF therapeutics.  Does Dr McCartney believe that a genetic understanding, in collaboration with biochemical knowledge of the influences on biologic treatment as a tool for predicting treatment response, thus saving the patient from months of uncertainty and further pain, not to mention the potential savings for the health service where a single course of ineffectual treatment can cost upwards of £10,000.  Is that all hype?

We have not reached the personalized medicine era, we are still exploring the role of genetic susceptibility in disease and therapeutic response, but it is moving at a fast pace.  I agree that perhaps we need to be a little cautious, not succumb to zealotry, this is people's lives at stake, but the potential benefits are staggering.  Imagine being able to walk into a clinic in 10-15 years time and have the clinicians be able to predict any potential adverse side effects, and whether or not you are likely to need an altered dosage regime.  That requires more than just a knowledge of the underlying genetics and biology, it requires political and societal changes, the implementation of an infrastructure that can support and utlise such vast quantities of information, not to mention the educational requirements for both physicians and the general public.

We do need to see both sides of the story with regards to decisions that affect something as important as our personal and societal health.  A more balanced approach that investigate both the barriers still to overcome and the limitations is called for.  Sometimes it does seem like there is only ever fanfare surrounding the predictions of personalized medicine, however, a closer read of the blogosphere shows that this fanfare is well measured with caution and an understanding of the pitfalls of predicting disease susceptibility.

Gene patents - stifling innovation?

There have been a couple of events that have spurred me into voicing my own opinions with regard to gene patenting. The first is the Myriad gene patent of BRCA1 and BRCA2 which was overturned some weeks ago, and is now being contested (hardly a surprise when there is such a monetary incentive!). The other is an interview with ex-UK HGP director, John Sulston, in which he lays out his own concerns about gene patenting. We have to remember he was around at the start of the idea of being able to patent a gene or genome, when Craig Venter started up Celera in a bid to beat the public sequencing effort to the prize, and thus charge researchers for the privilege of accessing our own genomes.

I’m a fan of open access and transparency in science, in fact I would go so far as to say that it is a necessity of good scientific practise, and upholds and empowers the scientific method and the concept of science itself. So the idea of patenting a gene, or its specific variants, which was not invented, and is present in numerous persons within a given population, is anathema to the open access and transparent nature of good scientific practise. It prevents researchers from unhindered research into the mechanisms of mutations within the gene in question, and the ability to use it for clinical applications, such as diagnostic indicators in disease, or as prognostic indicators of therapeutic response or adverse side effects.

This is one of the major problems that can crop up when some of the drivers of pharmacological research have financially vested interests, as is the case with Myriad and other pharmaceutical companies. Don’t get me wrong, this isn’t a dig at BigPharma, it’s a dig at the money-grabbers, bureaucrats and lawyers for thinking they could ever place a patent on a naturally occurring biological molecule. If they are going to patent something to protect their intellectual property, do it in a responsible manner that does not stifle progress and innovation in the name of financial gain. Patent the diagnostic test and specific protocol itself if needs be; surely that protects their intellectual property sufficiently?

Now, I’m not a patent lawyer (I once heard a talk given by a chap going through the training, I wouldn’t want to subject myself to that level of tedium and mind-numbing law-talk, despite the excellent remuneration), so I don’t know the inner-workings of patent law, and the loop holes and requisites required, but the Myriad patent case and others in Australia need to set a precedent that naturally occurring biological molecules, complexes and machinery are not patentable, and that any artificial or derived variations on the natural theme are shown to be, with sufficient supporting evidence, sufficiently different that they are unlikely to exist, or occur within nature itself.

That should rule out any unscrupulous companies trying to patent rare variants too.

I understand the importance of patents, I believe (correct me if I’m wrong here), the patent offices were originally set up to promote innovation whilst protecting the rights of the inventors themselves. So we need to keep that original basis in mind when we consider the patenting of biological materials. Will this patent promote, or stifle, innovation?

Gene patenting can only ever stifle innovation that arises from competition, thus it is untenable, and should be rejected outright.

The Weekly Round-Up

This week there have been several stories of note that I've not had time to blog about, so I thought I'd start a weekly round-up of interesting news pieces.

[1] Starting with the Sanger Institutes announcement of their own 10,000 genomes project. The WTSI is performing its own 10,000 genomes project, its aim to help uncover the various genetic elements that predispose to various diseases and disorders by full resequencing of 4,000 individuals, and the exomes of another 6,000.

[2] This announcement from the WTSI comes in the same week that the original 1000 genomes project announces its release of their pilot data, prior to the start of their database for public and research use.

[3] I've already covered this story, but it seems, to me at least, a step forward towards the implementation of clinical sequencing by the Royal Brompton Hospital

[4] An interview with Francis Collins in The Times about how he sees the future of genomic medicine panning out, with an emphasis on the hurdles still to overcome, including the education of physicians and the general public about personal genomics and personalized medicine.

[5] 23andMe have published their first paper in PLoS Genetics, a GWAS of various phenotypic traits. Whilst most of these may seem superficial (hair colour, eye colour, etc), it is their research framework that is the focus of this paper.

[6] Its 10 years since Craig J. Venter and Francis Collins stepped out on the White House Lawn with President Bill Clinton to announce the complete draft of the Human Genome.

There has been much speculation about the predicted impact of genomic medicine, and whether or not it has, or will be able to, deliver on all its promises. It's true that most of the general public won't have noticed the significant advances made in genomic medicine, but I think we are within 10 years of routine clinical sequencing and the start of an era in personalized medicine. Watch this space!