The following article from www.theheart.org provides a good discussion about the pros and cons of genetic testing for people who may have a family history of premature coronary artery disease:
Letting the genome out of the bottle: Unraveling the genetics of heart disease
June 23, 2008 | Lisa Nainggolan
London, UK – At the AHA Scientific Sessions in Orlando last November, deCODE Genetics offered its test for 9p21—a variant that is recognized to be the strongest genetic predictor of early MI discovered to date [1,2,3]—retailing at around $200, for free.
“All the cardiologists wanted to have it done on themselves. We brought 500 deCODE MI testing kits, and they ran out in the first morning,” says deCODE CEO Dr Kari Stefansson.
Whether this eagerness was due to the well-recognized conference phenomenon of physicians wanting to avail themselves of as much for free as possible or whether they genuinely wanted to know their 9p21 status remains an open question. But one thing is certain, they are not alone.
Dr Eric Topol
Top cardiologist and genomics expert Dr Eric Topol (Scripps Translational Science Institute, La Jolla, CA) is one of a number of people who has had his genome scanned—a step further than just a simple test for one genetic variant. A number of companies are now offering genomewide scans that search for numerous variants affecting the risk of many diseases at the same time.
Enthusiasm for testing is not universal, however, with many critics believing it is way too premature. Geneticist and cardiologist Dr James Scott (Imperial College London, UK) is particularly wary: “These companies offering such genetic tests have to very cautious and careful with the hype—if they overplay it, they will crash and burn.” And he believes that genomewide scans, costing from $1000 to $2500, “are a complete waste of space. It’s crookery. Robbery. It just is.”
Dr James Scott
One of the major problems, say Scott and other critics, is that the risks revealed represent hazards in isolation, not taking into account the effects of lifestyle, environment, and other factors. Also, the scientific accuracy of some of these tests is in question; results need to be validated in prospective trials, they stress. And finally, short of changing lifestyle, there is very little else that can currently be done with the results, and those of an anxious disposition will merely worry more than ever about things they can do little about, warn the naysayers.
Topol himself has mixed feelings about the current utility of such tests but, he says, the “genome is now out of the bottle—some have predicted there will be tens of thousands of individuals who will have their genomes scanned this year.” Whether it be single gene tests for 9p21 or other similar assays for diseases such as diabetes and cancer, at a couple of hundred dollars a pop, or the more expensive full-genome scans, he believes the phenomenon is here to stay.
Even a year ago, the idea that a saliva test could yield genomic information that provided estimates of the probabilities of getting diseases such as prostate and breast cancer, diabetes, and age-related macular degeneration would have seemed like something from a science-fiction movie to most. Not anymore. During the past 12 months there has been an explosion in genomics, and this is only the beginning, say experts in the field.
So how has this happened seemingly so quickly? The answer is genomewide-association studies (GWAS), which have been unleashed en masse upon the public in the past year, with many new studies published on novel loci for diseases such as breast and prostate cancer, coronary artery disease, MI, diabetes, and obesity. Every week, it seems that new papers turn up in journals and headlines scream, “New genetic risk factor discovered” for this disease or that.
And although these new studies have many shortcomings, there is a palpable excitement surrounding them. Most researchers in the field believe this new direction in genetics represents a revolution that will ultimately yield information of an unparalleled nature when it comes to complex diseases. GWAS represent a novel way of screening for common polymorphisms—any variation in the sequence of DNA among individuals. Single nucleotide polymorphisms (SNPs) are the most common polymorphisms encountered, and this is what scientists are scanning for—SNPs linked with common diseases. If something is identified in a very large population and replicated in multiple other cohorts, it starts to become convincing.
“2007 was a watershed year,” says Topol, who works on genomics along with colleagues at Scripps Translational Science Institute. “This is an extraordinary movement, the likes of which we have never seen before.”
Perhaps understandably, doctors in general are nervous about this new era. How will they cope with the influx of patients who present themselves at their office with their genetic information, when for many physicians the terminology of modern genetics resembles a foreign language? They will need to learn the new genetic lingo as the science evolves and recognize which findings are most important as these studies begin to offer unprecedented insights into complex diseases. And they will need to fully understand the pros and cons of genetic tests and GWAS to be able to fully discuss them with their patients.
2007: The year of the genomics gold rush
The first genetic studies took unusual diseases with simple Mendelian inheritance patterns and isolated the rare deleterious polymorphisms, also known as mutations, responsible. Then 20 years ago, scientists began searching for candidate genes that predisposed to cardiovascular and other complex diseases, looking this time for common polymorphisms. With the sequencing of the human genome at the beginning of the new millennium came the hope that advancement would be rapid. But although the candidate-gene approach and linkage studies identified many gene variants said to confer susceptibility to different diseases, the work was painful; findings have been inconsistent and difficult to replicate and validate.
“The candidate-gene approach is exhausted, it has by and large run out of steam because most of the effects of these variants are terribly small, and you can’t look at enough candidate genes,” Scott explains.
This is where GWAS have come into their own, allowing scientists to trawl enormous sections of the genome at previously unthinkable speeds and relate certain SNPs to clinical conditions and measurable traits. This is not sequencing of the human genome—only two individuals worldwide have so far had their entire genomes sequenced and published: Drs Craig Venter (J Craig Venter Institute, Rockville, MD) and James Watson, of double-helix fame—but rather they are genome screens looking for up to a million SNPs.
All of this has been made possible by two fundamental breakthroughs in the past couple of years, explains Topol, who along with colleagues has dubbed this trend the “genomics gold rush” .
First, large-scale efforts led to the discovery of a substantial fraction of the estimated 3.2 million base pairs of nucleotides that commonly vary between individuals, he explains. These base-pair variations are not inherited independently but as units, referred to as a haplotype or “bin.” In 2005, haplotype maps were created “that showed the vast majority of common variation in the human genome could be reduced to approximately 250 000 to 500 000 bins,” Topol notes. “Effectively, these bins serve as zip codes for subsequent tagging of variations that influence a disease or trait.”
Second, ultrahigh-throughput genotyping became possible. “The technology is so extraordinary now, and it’s accelerating,” he says. “We are on a very steep curve. The bar is being raised every few months. We can analyze a million SNPs in each person, and within two to three years we might even be able to sequence tens of thousands of people and assemble every nucleotide in their genome. You start off with a null hypothesis and you let the genome talk. If you find something in a very large population and can replicate it in multiple other cohorts, then that is compelling.”
Expensive studies, many limitations
Dr Marc Sabatine
Yet there are still many limitations. Despite the convenience of GWAS, they are exceedingly expensive—known as the “million-dollar plots,” because they cost so much to perform—and most SNPs that are identified still suffer from the same fate as those discovered in candidate-gene studies: their effects in any complex disease are relatively small. Dr Marc Sabatine (Brigham and Women’s Hospital, Boston, MA) explains: “Unlike simple Mendelian disorders, something like coronary heart disease is obviously far more complex, and the effects of any one gene will be subtle.
“The fact is that most publications pertaining to new loci for complex diseases are going to be of very little value until there are a whole bunch of them,” says Scott. “We are looking through a really new window on how complex genetics works. It’s going to be quite challenging, as the effects are so relatively small in terms of the total variance in phenotype.”
For example, the typical odds ratio for premature MI if a person carries one copy of the 9p21 variant—which most people agree is the most impressive marker to date for heart disease—is 1.2, “a relatively small effect of a common variant,” says Topol. And while almost 50% of individuals of European ancestry have a least one copy of the marker, the population-attributable risk is still only 21%, “leaving room for other genes to exert a substantial effect,” he notes. But he points out that when homozygous individuals are considered—about 20% of Europeans carry two copies of 9p21—the risk leaps up, with an almost doubling of the risk of early heart attack.
Other issues include some important statistical nuances. In a recent commentary in the International Journal of Epidemiology , Dr Timothy Frayling (Peninsula Medical School, University of Exeter, UK) outlines some of the confounding issues with regard to GWAS, including: potential sources of bias; the fact that they require more stringent statistical standards than conventional studies; the importance of sample size; and the issue of population stratification. Fortunately, many of these concerns are being addressed, he says.
On the issue of population stratification, however, Frayling says this can be confounding if not properly controlled for. For example, if there are two background populations and disease frequency and allele frequencies are different between the two, then false-positive and false-negative results can occur. Topol expands upon this: “So many of the things we have observed are ancestry- or ethnic-specific, and when you start to try to replicate in different populations, there can be problems, because populations in San Diego are different from populations in Atlanta.” And he points out that the vast majority of GWAS so far have looked only at populations of European ancestry: “We know little with respect to those of African or Asian ancestry,” he observes.
But in conclusion, Frayling says that GWAS are some of the most robust research methods ever devised, and that epidemiologists generally feel confident about them: “GWAS are making life a lot easier. If [they] . . . meet what are now routine quality criteria, they offer an unprecedented increase to our understanding of common diseases and conditions.”
How do the SNPs exert their effects?
Another criticism of GWAS is that they generate meaningless lists of numbers with as-yet little clinical applicability: likened by some to a telephone directory without addresses—some numbers are in it and maybe some names, but where the people live remains unknown. So, despite identifying SNPs with reasonably strong associations to a particular disease, scientists are no nearer to knowing how they exert their effects.
Dr Ali Marian (Source: Center for Cardiovascular Genetic Research, the Brown Foundation Institute of Molecular Medicine, University of Texas)
Dr Ali Marian (University of Texas, Houston) is a geneticist and a self-acknowledged critic of GWAS. He is particularly disturbed by the inability so far to delineate the biology underlying genomewide associations.
“Millions of dollars have been spent on GWAS. But it’s hit and run. They show something, and everybody follows them. It’s a herd effect. And then they move on to another 10 million SNPs. The burden should be on these investigators who claim these are the risk factors to show how they work. If you truly believe in your data, prove it. Unless you do that, you haven’t done your job properly. In order to cure a disease or make an impact, you’ve got to understand the pathogenesis of the disease. In order for me to believe that 9p21 is a risk factor for MI, they will have to figure out how it exerts its effects,” he says.
And, he continues, “Even for single gene disorders, genes are not the whole picture, they are not the sole determinants of the phenotype. When it comes to complex diseases, there are a huge number of competing factors.”
One of the reasons it has proven difficult to peg meaning to some of the associations coming out of GWAS is that, in many cases, although a specific “bin” is implicated, the actual functionally relevant, culprit DNA variant(s) have not been defined. “Only 1% to 2% of the whole genome is coding genes, there are only approximately 20 000 genes coding nucleotides, so most of the genome is not genes,” Topol explains.
Again, 9p21 provides a good example. In this case, the marker does not even lie in or near a gene. “The point is, you don’t even know which SNP is really the culprit, and if it’s not in a gene, it isn’t coding any protein, that’s for sure. So you don’t know how in the world in this whole genome the marker is exerting its effect. It may be some kind of modulating effect on a gene somewhere, but it doesn’t have to be close by, it can be very far away,” he adds.
However, Topol believes this is one of the reasons the field is so exciting, because it keeps turning up surprises. In type 2 diabetes, for instance, “not one gene associated with insulin resistance has been discovered, but genes relating to insulin secretion and transport, zinc binding to insulin, and pancreatic islet beta-cell development have been identified. Someday there may be at least 10 subtypes of diabetes mellitus based on specific individual biological variation,” he suggests.
“And if you have a marker, you have a risk independent of how it works. No one understands how the marker is exerting its effect—that’s important more for a specific drug-discovery program—but it doesn’t mean it’s not a valuable insight.”
Sabatine says: “It used to be that we would labor in labs and discover some pathway that was important and then eventually we would say, let’s look for genetic variants. Now we’ve turned it around. Now the genotyping far outstrips our ability to understand the biology, but it underlines the area we need to study.”
And Scott stresses that the mechanistic studies will be done. “First of all, you have to make sure the statistics and replications are correct, and if they are, you’ve got to do the biology. Naturally, people will want to find out what these things [such as 9p21] do, but it takes a lot of time, and different types of biologists tend to do that—they will go and make knockout mice, do cell studies and structural studies, and the like.”
Nevertheless, even Topol concedes that a break may be required from more GWAS. “The gene hunt has been remarkable in turning up these bins, but we almost need a time-out to get into functional studies,” he says.
And he is at pains to stress that there remain huge gaps in knowledge that are beyond the scope of this article. “All we have so far in this field are SNP variants that are statistically compelling associations. There are several inconvenient truths. We have incomplete coverage of the genome, we don’t know much about structural variants, copy number variants, or insertions and deletions. We need to know about metabolomics, proteomics, glycomics, and epigenomics. Once you get this information, all the pieces will come together, and a lot of the difficulties will be pushed aside.”
What’s the story in heart disease? It’s the prologue
The flurry of GWAS published last year means that six conditions now have very good genomic definition: age-related macular degeneration, type 2 diabetes, prostate cancer, breast cancer, Crohn’s disease, and systemic lupus erythematosus.
The results for these six diseases are “remarkable” and show, relatively speaking, how primitive the knowledge still is for cardiovascular disease, says Topol.
Apart from 9p21, the importance of which has been further boosted by the discovery in January that it is also a marker for abdominal aortic aneurysm and intracranial aneurysm—seemingly implicating some kind of vascular role —there have been only a few other robust findings in the field of cardiovascular disease.
One is the identification of a major genomic maker for atrial fibrillation (AF), 4q25 , and there have also been three GWAS identifying seven new genetic loci associated with lipids, representing the first attempts to report on the applicability of common polymorphisms to blood cholesterol levels [8,9,10]. There have also been important findings in the related fields of type 2 diabetes and obesity.
Of course, there have been many more studies than these, with other polymorphisms implicated in MI, and a number of other lipid, diabetes, and obesity genes identified, to name just a few. And therein lies the difficulty for many physicians: how to identify which research is most important. Topol advises doctors to “look in the prestigious journals. They raised their standards in the course of 2007 because they got flooded with genetics papers, not just cardiovascular but across the board.”
One particular website is also extremely useful—A Catalog of Published Genome-Wide Association Studies . The GWAS publications listed here include only those attempting to assay at least 100 000 SNPs in the initial stage. Papers are organized from most to least recent date of publication, indexing from online publications if available. The easy-to-read tables list full details of the author, date of publication, and journal; the disease or trait for which SNPs have been identified; the initial sample size and replication sample size; the region and gene identified; the strongest SNP-risk allele; and p value and odds ratio per copy.
To test or not to test: That’s the question
The relative paucity of information on heart disease points to the question of just how much use it is to take a single gene test or full genome scan. Many experts believe that, for cardiovascular disease, such testing is far too premature, not least because there has been no good research to show that people are going to behave differently knowing they have an increased genetic risk of any particular disease.
The fact that the companies marketing such tests are advertising directly to consumers in the US is of concern and means that inevitably some people will be keen to have the latest “gene test.” While a single test for a specific disease costs in the region of $200 to $300, many predict it is only a matter of time before the new genomewide scans launched at the end of last year will supersede many of these single gene tests.
Three companies are now offering genomewide scans. Navigenics offers the service for around $2500, including unlimited phone discussions with a genetic counselor , while “23 and Me”  and deCODE genetics “deCODE me”  offer their services for just under $1000: “A bargain, this is the whole kit and caboodle in a $1000 assessment that transcends all these other tests, $200 here, $200 there,” says Topol.
During a recent state-of-the-art lecture at the ACC meeting, Topol asked the expert physician panel whether they would have their genome scanned. Dr David Crossman (University of Sheffield, UK) said he would, “but only in the context of a research protocol.” Dr Ray E Hershberger (University of Miami Miller School of Medicine, FL) said he probably wouldn’t: “What are you going to do with the data? I’m not sure I would do anything with it.”
But Dr Marc E Shelton (Southern Illinois University School of Medicine, Springfield) felt it was a decision that would come down to personality: “If I had an anxious patient and their results were going to freak them out, I’d say, wait a few years. But if it were me and I could get out of one more colonoscopy, then I’d do it!”
Topol himself is ambivalent about the utility of such information at the current time. During his own genome scan he unexpectedly found he carries two copies of 9p21, making his risk of an early MI almost double that of a person who doesn’t carry the variant, but—as he concedes—”there’s not much you can do with this information right now. True, you might want to fix up your lifestyle, and it’s nice for cocktail-party chatter, but that’s about all.”
And he is concerned that there are too many data and not enough people to help interpret it. “We are facing an information overload the likes of which we have never seen before. The biggest rate-limiting step is being able to handle all of that data—it’s a runaway train,” he notes.
“The availability of online tools such as SNPedia  means we are now in the position where the patient often knows more about their risk implications than their doctor,” he continues. “You can now walk into your physician’s office with a million SNPs and say to your cardiologist, ‘What should I do?’ And your cardiologist says, ‘What’s a SNP?’ They have no clue what any of this stuff is.”
In a recent themed issue of the Journal of the American Medical Association (JAMA), a review by the US Department of Veterans Affairs and the RAND Corporation found that doctors generally feel “woefully underprepared” to integrate genetics into their practice . The authors conclude that there is a need for a large-scale effort to educate both health professionals and the public about genomic medicine and its current limitations. To this end, a number of papers in the same issue addressed different areas within the field, including a Special Communication on how to interpret a genomewide association study .
And so-called genetic experts will likely be of little help, Topol adds. “There are only 2200 practicing genetic counselors in the US, and even they haven’t been exposed to this challenge yet.”
One of the major concerns is the prospect of adverse consequences to genetic testing, particularly loss of privacy and discrimination by health insurers or employees, the JAMA authors point out. Topol says he has already seen many patients in California who have had their genomes scanned, and the issue of insurance, particularly in the US, “is one of grave concern to all of us and to our patients.”
Fortunately, a bill has just been passed in the US Congress that will allow people to take advantage of the promise of personalized medicine without fear of discrimination. The Genetic Information Nondiscrimination Act—which has been debated in Congress for 13 years—was passed May 1, 2008 and signed into law by the president on May 21 . The bill means that neither employers nor health insurers will be allowed to use genetic information to make employment or insurance decisions. Topol notes, however, that the bill will not come into force for around a year, so there will still be a time period whereby those who have had their genomes scanned will technically be in limbo. Furthermore, it does not apply to life-insurance or long-term-disability-insurance carriers, he adds.
The pots of gold at the end of the rainbow: Designer drugs and tailored therapy
[Source: US Department of Energy Human Genome Project]
While the debate about the utility of genetic tests and genome scans at the current time will rumble on, in the future there will undoubtedly be other, arguably more important, implications of GWAS.
The information gleaned from GWAS promises to yield new drug development targets, often via pathways that were never previously dreamed of, as illustrated by the surprising findings in type 2 diabetes. In fact, GWAS have already enabled the rapid development of new treatment strategies for age-related macular degeneration, and novel treatments for other diseases that are well characterized in GWAS will likely follow.
And individualized therapy tailored to a person’s specific genetic makeup may once have been the stuff of science fiction, but its day is about to dawn, and sooner than some may think. While acknowledging that true personalized, tailored medicine is still a number of years away, many experts believe this is one of the ultimate goals of all these GWAS.
“I think the idea of doing genetic testing to tailor therapy is ultimately the pot of gold of all this work for the Human Genome Project. That’s what we want to do, to be able to tailor our therapy,” says Sabatine.
“In the past, we’d ask about a family history of heart disease. That was kind of a crude metric. Now we hope genotyping will provide a quantitative metric. And what would make that measurement particularly compelling would be if it would clearly lead us down a different pathway for therapy—not only for withholding treatment, but potentially to know which other therapies would be more beneficial,” he says excitedly.
While this is still in its infancy, there have been some encouraging developments. One of the most advanced fields in this area is the pharmacogenetics of warfarin. Doctors are desperate to be able to tailor warfarin therapy, because treatment is so challenging due to the wide variation in individual response to the drug and the ongoing monitoring that is required during therapy. Studies have already shown that patients with certain common genetic variants require a lower dose of warfarin and a longer time to reach a stable dose and are at higher risk of overanticoagulation and serious bleeding. Last year, the US FDA approved a labeling change for warfarin that describes the reported effects of two particular genotypes on dose requirements.
Dr Sharon Cresci [Source: Washington University Center for Cardiovascular Research]
And just recently, US researchers identified a genetic variant that appears to act like a natural beta blocker in 40% of African Americans . While white patients with heart failure participating in clinical studies of beta blockers have shown clear benefit from these drugs, the effects of beta blockers in African Americans have been ambiguous. The new findings may help explain why beta blockers don’t appear to benefit some blacks, say the researchers.
Author Dr Sharon Cresci (Washington University, St Louis, MO) told heartwire: “I believe this gene has significantly contributed to the discrepancies that have been found in such studies. If researchers were able to go back and genotype patients from their studies, it would probably help clarify some of their findings.”
Although she and her coworkers stressed that it is too early, on the basis of just this one study, to advocate not prescribing beta blockers to blacks who carry the genetic variant, she believes a prospective study may help answer this question.
“The research is a step toward individualized therapy tailored to personal genetic makeup. I believe we are on the threshold of that, but we are not there yet.”