Genomic Messages: How the Evolving Science of Genetics Affects Our Health, Families, and Future - George Annas, Sherman Elias (2015)
Chapter 4. Pharmacogenomics
We must not see any person as an abstraction. Instead,
we must see in every person a universe with its own
secrets, with its own treasures, with its own sources
of anguish, and with some measure of triumph.
—Elie Wiesel, preface to The Nazi Doctors
and the Nuremberg Code (1993)
Rebecca Skloot opens her immensely popular book, The Immortal Life of Henrietta Lacks, with this statement from Elie Wiesel reflecting on the uniqueness of every human being. Genomics, of course, underlines that uniqueness. Skloot’s book is about Henrietta Lacks, the woman whose cervical cancer tumor was the source of the famous HeLa cells, an immortal cell line that has been the backbone of laboratory research around the world for more than sixty years (figure 4.1). Portions of her tumor were removed for study at the segregated Johns Hopkins Hospital in 1951. Her story is a compelling portrait of an extended black family and their race- and poverty-based interactions with the white medical profession from the 1940s to the present. Genetics is never far from the surface. The physician who is sometimes called the “father of American genetics,” Victor McKusick, for example, began his own research on the Lacks family in 1973. McKusick wanted blood samples from the surviving Lacks family members to try to locate genetic markers that could be used to identify HeLa cells in various laboratory experiments. No informed consent for the blood drawing was obtained from the family members, who believed their blood was being tested to see if they had the same cancer Henrietta had died of more than twenty years previously. Lacks was thirty-one years old when she died. No consent had been obtained to use her cancerous cells in 1951 either.
4.1 Henrietta Lacks. With permission of Rebecca Skloot, The Immortal Life of Henrietta Lacks (New York: Crown Publisher, 2010).
For her book, Rebecca Skloot interviewed Susan Hsu, the member of McKusick’s team who drew the blood samples from the Lackses. Hsu said she felt very bad that the Lacks family didn’t understand why she was collecting their blood more than twenty-five years ago. She added that she could learn much more today from their blood, because of advances in DNA technology, than she could in 1970. She remarkably asked Skloot if she would be willing to ask them if they’d give blood again for her research: “If they are willing, I wouldn’t mind to go back and get some more blood.”
Lacks’s daughter told Skloot that she thought it unfair that even though her mother’s cells helped create the extremely lucrative biotechnology industry, her family lacked health insurance and struggled to pay their medical bills. Law professor and expert on racial discrimination Dorothy Roberts describes the Lacks story as a reflection of the “horrible history of medical exploitation and neglect of African Americans” on the basis that they are genetically inferior to whites. But, as she properly underlines, the Henrietta Lacks story also refutes that belief: “Her cells, although they came from a black woman, helped to improve the health of human beings the world over and testify to our common humanity.”
The tendency to treat differences in socially constructed racial groupings as genetics based has haunted genetic medicine since its beginnings, and we continue to be plagued by racism in genomics. In this chapter, we attempt to place pharmacogenomics—the use of genomics to determine which drug at which dose to give an individual patient—in the context of racism and racial disparities in American medicine. We then compare the way foods and drugs are affected by our genomes and discuss the usefulness of genomic screening to determine what drugs may be helpful for your condition.
Pharmacogenomics can usefully be described as selecting a drug and its dose on the basis of an individual patient’s genomic profile. The goal is to maximize the drug’s efficacy and minimize its toxicity. It is not, however, “personal” in the sense that we are trying to select a specific drug and dose just for you. Instead, pharmacogenomics is about selecting a drug and dose for the group of patients with whom you share specific genetic characteristics. These genetic similarities can be used to categorize or stratify patients in groups that are likely to react the same way to specific drugs. Thus stratified medicine is a more accurate term than personalized medicine for genomic medicine. Identifying the genes that control drug effectiveness also permits clinicians to improve treatment by using genomics rather than physical characteristics, such as skin color. This is important because of the history of racism in American medicine, complete with unscientific views of the differences among people of different colors.
Genetics and Racism
Racism continues to plague pharmacogenomics at the clinical level. For example, a 2014 study concluding that black women with breast cancer were 40 percent more likely to die of the disease than white women showed that survival differences were not because of any genetic difference, but resulted from “systematic racism,” which delayed diagnosis and treatment. This is not just true for breast cancer. Racial disparities in health care and health outcomes in almost all areas of medicine are a function not of biological differences between minority and majority populations, but rather differing environmental factors, including racism, poverty, lack of health insurance, and lack of access to decent medical care. Nonetheless, because genetics has been widely misused to suggest a biological basis for racial differences, it is critical that doctors and patients alike understand the relationship between genes and skin color. The stories of the Tuskegee Syphilis Study, like the story of Henrietta Lacks, are emblematic of past medical practices that continue to incite distrust on the part of minorities in medicine, distrust that threatens to undermine even the promise of pharmacogenomics.
“The United States government did something that was wrong—deeply, profoundly, morally wrong. It was an outrage to our commitment to integrity and equality for all our citizens.” President Bill Clinton spoke these words at a White House ceremony—this one held three years before a more celebratory White House ceremony announcing the completion of the draft of the human genome. The occasion, in 1997, was a formal apology to the surviving victims of what is known as the Tuskegee Syphilis Study. The study was based on racism: a belief that black men were so different biologically from white men that syphilis in blacks was a different disease than syphilis in whites. The study followed the progression of syphilis in a cohort of black males until they died. The subjects were lied to about the study and neither informed of their disease nor treated for it. Medical historian James Jones described the study as the “longest nontherapeutic experiment on human beings in medical history.” It ran from 1932 to 1972. The Tuskegee Syphilis Study remains a hideous blot on American medicine. In words that remind us of Elie Wiesel’s commentary on the Nuremberg Doctors’ Trial, quoted to open this chapter, survivor Herman Shaw reflected, “We were treated unfairly and to some extent like guinea pigs. We were not pigs. We were not dancing boys as we were projected in the movie, Miss Evers Boys. We were all hard working men, not boys, and citizens of the United States. The wounds that were inflicted upon us cannot be undone.”
There are no genes that code for race, but there are genes that appear with greater frequency in different ethnic groups. It is disturbing that race and ethnicity are often used interchangeably in the scientific literature. Race, a social construct, not a scientific or genetic one, and is a poor proxy for genetics. Even for genes that appear frequently in specific ethnic groups, it is far more scientific and accurate to directly identify a gene or genes that affect drug metabolism. A leading geneticist, Arno Motulsky, is credited with opening the race/gene/drug interaction exploration in a 1957 article on “genetically conditioned drug reactions.” He concluded (using race and ethnicity interchangeably) that “since a given gene may be more frequent in certain ethnic groups, any drug reaction that is more frequently observed in a given racial group, when other environmental variables are equal, will usually have a genetic basis.” The Tuskegee study was still going on, so it is probably not surprising that the major study Motulsky cited for the proposition that blacks responded differently than whites to the malaria drug primaquine was one done exclusively on black prisoners in the early 1950s. With so little known about genetics at the time, Motulsky thought that observing drug reactions in humans could “contribute to the progress of human genetics in general.” Science is now rightly in the process of trying to reverse this order, using human genetics to predict drug reactions.
Pharmacogenomics got a big boost from the 2000 White House announcement of the draft human genome. A number of the world’s most prominent geneticists announced that the project would once and for all put an end to assertions that human groupings are fundamentally different at the genetic level. Chris Stinger of London’s Natural History Museum observed, “We are all Africans under the skin.” Other geneticists said, “Race is only skin deep” and “There is nothing scientific about race: no genes of any sort pattern along racial lines.” Craig Venter summed it up: “Race is a social concept, not a scientific one. We all evolved in the last 100,000 years from the same small number of tribes that migrated out of Africa and colonized the world.” Geography matters, because people who live in the same area are more likely to have children together and perpetuate common genes. Race does not matter, because two black people are no more likely than a white person and a black person to share the same genes. The most scientifically valid approach is therefore to directly identify the relevant genes or variants.
In 2001, at the World Conference on Racism in South Africa, George suggested that it would be cold comfort to simply replace racism (the belief that fundamental human characteristics are determined by race) with “genism” (the belief that fundamental human characteristics are determined by genes). We both continue to believe that this is a danger, but neither of us was ready for the first round of pharmacogenomics, which was dominated by using race rather than genetics as a rationale for developing the drug BiDil (discussed later). As Eduardo Bonilla-Silva, author of Racism Without Racists, observed, “Race, like Freddy Krueger, keeps coming back after we believe we killed it.” In a climate in which racial disparities continue to plague contemporary medicine and black patients continue to be suspicious of the white medical establishment, we should not be shocked.
The death of race-based genetic medicine has been hard for some to accept. In 2000, for example, a small Massachusetts biotech firm, NitroMed, announced that it had approval from the FDA to conduct a clinical trial with its new drug, BiDil. What made BiDil controversial was that it was a heart failure medication designed exclusively for use by African Americans. Following a clinical trial that involved only blacks (how is this possible in the twenty-first century?), the FDA approved BiDil “as the first drug with a race-specific indication: to treat heart failure in a ‘black’ patient.” BiDil was not a new biotech drug but simply two (“bi”) generic blood vessel dilators (“dil”) combined into one pill. Nonetheless, its approval was hailed as a major milestone in pharmacogenomics. As Science put it, “By backing BiDil, the FDA panel gave another push to pharmacogenomics, an approach that promises to revolutionize both drug discovery and patient care.”
Jonathan Kahn, the historian of BiDil (and author of Race in a Bottle, from which much of our material on the drug is taken), has emphasized that BiDil is in no way a genomic or pharmacogenomic drug—as no related genes were even looked for, let alone identified. Instead, as the chair of the FDA advisory panel, Steven Nissen, put it (without irony), “We’re using self-identified race as a surrogate for genetic markers.” He continued, “We are moving forward to genome-based medicine. It’s going to happen.” Kahn notes that the lack of any genetic basis for the drug actually helped to market it: instead of having to screen patients for a gene or genes, it could simply be marketed to African Americans. In his words, “Medical researchers might see race as a surrogate to get at biology in drug development, but corporations use biology as a surrogate to get at race in drug marketing.” Although ultimately a failure in the medical marketplace, as the most discussed race-based drug in the world, BiDil teaches us that racial stereotyping is not dead, even in medicine and science, and that genomics has sufficient public appeal that it can serve as a cover story to promote even nongenomic-related pharmaceuticals to both the FDA and the public. This is fake genomics.
Another attempt to inject race into genomics was New York Times reporter Nicholas Wade’s A Troublesome Inheritance: Genes, Race and Human History. He argues, for example, that culture is at least partly genetic, so genetics can help explain why, for example, “American institutions do not transplant so easily to tribal societies like Iraq or Afghanistan.” This implausible argument is in the same category as the argument that genes (rather than access to medical care, poverty, and other factors) can explain why black women have worse outcomes from breast cancer treatment than white women. As reviewer H. Allen Orr put it, “What about all those other differences [i.e., other than genes] in history, language, distribution of wealth, religion, educational attainment, ravages of war, arable land, resentment toward perceived invaders, and so on? Among these factors, I suspect that genes are perhaps the one most similar between American and Afghan societies.” Bad or unscientific genomics is relatively easy to spot and debunk. But what does good genomics look like in the prescription drug field? The story of warfarin answers this question.
Pharmacogenomics and Warfarin
On NBC’s Today Show, Serena Williams, one of the all-time greatest professional tennis players, described how on February 18, 2011, she was rushed to a hospital suffering from swelling of her leg and severe shortness of breath. (Serena had cut her foot on a piece of glass the year before, a cut that had required two surgeries and her being in a leg cast for ten weeks.) The doctors did a CAT scan that showed Serena had experienced pulmonary embolisms—blood clots in the lungs—a potentially life-threatening condition. The pulmonary embolisms were caused by blood clots in her leg breaking loose and traveling to her lungs. She was discharged on anticoagulant drug injections and was able to attend parties at the Academy Awards on February 27. The next day she was readmitted to the hospital because she had developed a grapefruit-sized collection of blood underneath her skin (a hematoma) that required surgical drainage. Later, Serena was switched to an oral anticoagulant drug, warfarin, and had to wear a boot for an additional ten weeks. Thanks in part to warfarin, she was able to resume her spectacular tennis career (figure 4.2).
4.2 Serena Williams hitting a backhand shot during the 2012 Wimbledon Championships on Day 10 in the Semifinals versus Victoria Azarenka. Wikimedia Images, July 5, 2012.
Approximately half a million people have a pulmonary embolism in the United States annually. If untreated, 30 percent of them would die. Drugs that diminish the ability of blood to clot, called anticoagulants, are a mainstay of treatment for pulmonary embolism. Like Serena, patients are usually given an injectable anticoagulant as an initial treatment, then converted to pills. The most commonly prescribed anticoagulant pill is warfarin. Then secretary of state Hillary Clinton was prescribed warfarin when she was hospitalized at the end of 2012. She had a blood clot just behind her left ear (a right transverse sinus thrombosis) as a result of hitting her head after she fainted and fell a few weeks earlier. Her physicians announced, “She will be released once the medication dose [of warfarin] has been established.”
Warfarin works by inhibiting an enzyme that uses vitamin K, the vitamin which is necessary for the production of several blood-clotting proteins in the liver. It was initially marketed as a rat poison and is still used for this purpose. Saying that warfarin is a “blood thinner” is a misnomer, since it does not affect the thickness (viscosity) of the blood; rather it reduces the degree of clotting. Warfarin is not only used to prevent and treat embolisms. It is also used for people with certain types of irregular heartbeats, people with replacement or mechanical heart valves, and people who have suffered a heart attack. In each of these conditions, warfarin is used to help prevent a blood clot from developing and traveling to the brain to cause a stroke, or to the lung to cause respiratory failure.
Determining the optimal dose of warfarin is tricky because of marked and often unpredictable dosing variation among individuals. Giving too little warfarin does not protect against the development of clots, whereas giving too much can cause bleeding problems. The primary way a physician monitors the anticoagulant effect of warfarin is by trying to maintain the drug dose within a narrow therapeutic range. The difficulty of maintaining the right dose of warfarin is evidenced by warfarin being one of the drugs most often responsible for emergency room visits, most commonly hemorrhage (too much warfarin) or blood clots and stroke (too little warfarin). It has been established that two genes primarily affect how individuals respond to warfarin.
It should be emphasized that even though genetic testing can provide physicians with relevant dosing information, there is no evidence that genetic testing was done for either Serena Williams or Hillary Clinton. The point is not that such testing is irrelevant; rather, it is time consuming, and its impact on treatment is currently marginal. With or without the genetic profile, physicians will still use trial and error to establish the proper dosage of warfarin for a particular patient. Use of genetic information to help inform the dosage decision will only routinely occur only after patients have their whole-genome profile included in their electronic health records. Access to relevant genetic information will then be easy and quick.
The FDA approved a labeling change for warfarin (Coumadin) in 2007, explaining that people’s genetic makeup may influence how they respond to the drug and highlighting “the opportunity for healthcare providers to use genetic tests to improve their initial estimate of what is a reasonable warfarin dose for individual patients. Testing may help optimize the use of warfarin and lower the risk of bleeding complications of the drug.” The FDA commissioner said, “Today’s approved labeling change is one step in our commitment to personalized medicine. By using modern science to get the right drug in the right dose for the right patient, the FDA will further enhance the safety and effectiveness of the medicines Americans depend upon [emphasis added].” No requirement for genetic testing was included in the labeling change. Shortly thereafter, the FDA cleared the way for the first genetic tests for warfarin metabolism.
Medicare announced in 2009 that it would not pay for genetic tests for warfarin metabolism (which cost $50 to $300) because there was not enough evidence that use of the tests actually improved patients’ health. The agency did say that it would pay for the tests as part of clinical trials to gather such evidence. The evidence has not yet been found. For example, a study published in 2014 randomly assigned one thousand patients to warfarin on the basis of clinical variables only or on the basis of clinical variables plus genotype information. Genotyping did not improve anticoagulation control during the first month of therapy. We think, nonetheless, that effectiveness will change when physicians become more comfortable using genetic information to inform dosing decisions.
The activity of other drugs has also been linked to specific genes. In the summer of 2012, for example, the FDA issued a warning when three children died and another child had life-threatening respiratory depression after taking routinely prescribed doses of codeine following surgery to remove tonsils. Health care professionals and parents were cautioned that when prescribing codeine-containing drugs, the lowest effective dose for the shortest time should be used. Codeine is used to treat mild to moderate pain but is also found in some cough suppressants. Once in the body, codeine is converted into morphine in the liver by an enzyme called CYP2D6. The CYP2D6 gene that encodes this enzyme has at least eighty known variants. These variants are grouped to classify how well individual patients will metabolize codeine. At one extreme, “poor metabolizers” may not get the pain relief expected from codeine. At the other extreme are “ultrarapid metabolizers.” After receiving normal doses of codeine, their livers quickly convert codeine into dangerously high blood levels of morphine that can lead to death. The three children who died after receiving codeine were likely ultrarapid metabolizers. In one tragic (and highly unusual) case, a mother who was subsequently found to be an ultrarapid metabolizer took codeine for postpartum episiotomy pain while she was breast-feeding. Her thirteen-day-old baby died of morphine “poisoning” with a serum concentration over thirty times higher than typically seen in babies whose mothers who are breast-feeding while receiving codeine.
Genetic testing can also help prevent organ rejection. This was true even in the early 1950s when transplanted kidneys were consistently rejected until identical twins were used as donor and recipient. Genetic tests to confirm that twins were identical did not exist at the time. Instead, genetic identity was determined much less scientifically on the basis of blood groups, fingerprints, and skin grafts between the twins. About 18,000 Americans annually receive kidney transplants, one-third from living donors. For people with end-stage kidney disease, a kidney transplant offers enhanced quality of life. The main obstacle to more transplants is the shortage of cadaver organs, which would be helped if more Americans signed organ donor cards—something you should seriously consider.
Whichever protocol a transplant center uses, all kidney transplant recipients require lifelong immunosuppression to prevent rejection, just as many patients who have had blood clots require lifelong use of an anticoagulant. Perhaps the most commonly used drug for immune suppression is azathioprine (Imuran). An enzyme called thiopurine methyltranferase (TPMT), encoded by the TPMT gene, controls azathioprine activity. Approximately 10 percent of the population inherit one nonfunctional TMPT gene variant, conferring intermediate TMPT enzyme activity, and 0.3 percent inherit two nonfunctional variants for low or absent activity. Patients with intermediate TPMT activity may be at increased risk of bone marrow suppression—a decrease in cells responsible for providing immunity, carrying oxygen, and ensuring normal blood clotting—if receiving conventional doses of azathioprine. Patients with low or absent TPMT activity are at increased risk of developing severe, life-threatening bone marrow toxicity when receiving conventional doses. Analysis of TPMT gene variants can provide physicians with valuable guidance for either not prescribing azathioprine or adjusting its dose. In clinical practice, enzyme activity and genotyping are often used as complementary tests. Genotyping can tell us how we will likely react to specific drugs. It can also tell us how we might react to specific foods.
Food and Genes
Drugs and food are absorbed, metabolized, and excreted by bodily processes involving multiple genes, as well as environmental influences, including the types and amounts of food we eat, and the types and doses of drugs we take. The emerging field of nutrigenomics, which aims to identify the genetic factors that influence the body’s response to diet and studies how bioactive ingredients of food affect gene expression, helps us understand the current state of pharmacogenomics. For centuries it has been recognized that tastes in food are not always inherited. Renaissance writer Michel de Montaigne, for instance, observed about his own appetite: “I am not excessively fond of either salads or fruits, except melons. My father hated all kinds of sauces; I love them all.” He was also against regimented diets, writing, “Let us leave the daily diets to the almanac-makers and the doctors.” He argued that the whole point of food was enjoyment, and was opposed to treating food like medicine. But foods can directly affect our health.
Some Mediterranean, African, and Asian people who ingest fava beans become seriously ill with high fever, weakness, jaundice, vomiting, and diarrhea. The condition is called favism. More than fifty years ago, it was determined that favism was caused by deficiency of an enzyme called glucose-6-phosphate dehydrogenase (G6PD), which is encoded by a gene on the X chromosome; hence the condition almost exclusively affects males. More than 140 mutations of the G6PD gene have been described that can result in episodes of severe acute hemolytic anemia (breakdown of red blood cells) and rarely in kidney failure and death.
In an episode of MASH, Max Klinger, a corporal of Lebanese descent, is trying to get a psychiatric discharge from the army. He is accused of faking exhaustion and back pain but is ultimately diagnosed as having G6PD deficiency resulting from the antimalarial drug primaquine. The black prisoners in the primaquine study (the study cited by Motulsky at the beginning of this chapter) who developed severe anemia at large doses also likely had G6PD deficiency. The study was based on the mistaken belief that black skin color, rather than ancestral geography, determined the medication’s side effects. G6PD deficiency affects more than 400 million people worldwide. It is most common in areas where malaria is endemic, suggesting that the deficiency is protective against malaria. Knowledge that an individual has G6PD deficiency is important when considering the use of commonly prescribed drugs, high doses of vitamin C, and even aspirin.
Taste for certain foods, such as sushi or gefilte fish, may be acquired, but taste may also be genetic. Some of us really can’t stand the taste or even the odor of particular foods, such as brussels sprouts, cilantro, or broccoli, no matter how well they are prepared or how hard we try. These and other food and odor aversions may be due, at least in part, to genetic factors. Differences among individuals in perception of the bitter taste of vegetables, as well as bitter beverages—coffee, grapefruit juice, quinine in tonic water, and alcohol—have been associated with variants in more than two dozen bitter receptor genes, collectively called TAS2Rs. Some people cannot perceive table sugar (sucrose) in liquids, because of variants in the sweet receptor gene TAS1R3.
After eating asparagus, some of us smell a varying pungent sulfurous odor in our urine, like cooked cabbage. This effect is very rapid, taking as little as fifteen to thirty minutes. Two reasons for variations in this odor have been proposed. The first is that there are differences in the amount of odorant people produce. The second is that people vary in their ability to smell. Both reasons likely play a role. Differences in how people digest and metabolize asparagus and variations in the ability to smell the odor are due to differences in the DNA sequence at a single nucleotide in the gene OR2M7, which encodes a receptor for smell. Taste and smell are both complex. Genetic variations can only provide part of the answer, because of the contribution of environmental, behavioral, and cultural influences—which means we can’t explain, for example, why Montaigne liked all the sauces his father hated.
What does all this have to do with pharmacogenomics? We have learned a lot about genes and nutrition but not enough to make health claims that genetic testing can give you enough information to develop a personalized diet. In early 2014, for example, the Federal Trade Commission (FTC) reached a consent agreement with GeneLink, in which the company agreed to stop marketing its “genetically customized nutritional supplements,” which it claimed it could personalize to fit “each consumer’s unique genetic profile” (based on an assessment of DNA obtained from a cheek swab provided by the consumer). The FTC announced under the proposed settlement that GeneLink was prohibited from claiming that any food would treat, prevent, or reduce “the risk of any diseases, including diabetes . . . by modulating the effect of genes, or based on a consumer’s customized genetic assessment—unless the claim is true and supported by at least two adequate and well-controlled studies.” Someday in the near future you may be able to get your genome sequenced and find useful information that could guide your diet and improve your health. But that day has not yet arrived, and when someone says it has, make sure you ask to see the scientific studies on which the claim is based.
Preprescription Gene Tests
Should your doctor order genetic testing prior to prescribing a drug or suggesting a diet for you? Not yet. Unless there is a good reason to believe you might have a genetic profile that makes a specific drug dangerous for you to use, trial and error will continue to be used by most physicians when prescribing for most patients. And this is, as we have emphasized, unlikely to change until you and patients like you have your entire genome sequenced and saved in or linked to your electronic health record.
Pharmacogenomic biomarkers have been touted as providing “fantastic opportunities for personalized medicine,” and we agree. If this potential were realized, it could have profound beneficial effects on our health care by improving drug efficacy, reducing costs, and most importantly, reducing adverse drug reactions. Adverse drug reactions, for example, account for about 7 percent of hospitalizations, 20 percent of readmissions to the hospital, and 100,000 deaths annually in the United States. On the other hand, most adverse reactions are the result of the wrong drug or the wrong dose, or drug interactions, not genetic variation. Matching drugs to genomes is not a magic solution to drug errors. Pharmaceutical and biotechnology companies have heavily invested in genomic-based strategies for developing new drugs, although their financial incentives may be different. Biotechnology companies primarily have sought ways to use genetic manipulation, including recombinant DNA methods, to make new drugs. This has, for example, been the major strategy of companies like Genentech, which produced recombinant human insulin, the first biotech product on the market. Genomic information is also expected to provide insights into the underlying biological mechanisms of disease and reveal biological targets and pathways that could lead to new drug discovery
It used to be argued that pharmaceutical companies are less interested in drugs designed for people with specific genetic sequences because this would fracture the market for their drugs, which are now sold for the entire population. Currently the opposite strategy is under way, as most major pharmaceutical companies see considerable profit in making expensive drugs for diseases that affect a small number of people. One example of this is Sanofi SA, the giant French pharmaceutical company, which recently purchased Genzyme, a U.S. company. Genzyme makes, among other products, a drug for people with rare enzyme deficiencies, such as Gaucher and Fabry diseases, which it sells for a very high price, making it unaffordable for people without health insurance.
Another example is Vertex, who with the financial support of the Cystic Fibrosis Foundation, produced the cystic fibrosis (CF) drug Kalydeco (ivacaftor). The drug, which the FDA commissioner touted as a sterling example of personalized medicine when it was approved in 2012, can only treat a very small percentage of CF patients (those with a specific genetic mutation)—about 1,200 in the United States. It is priced at more than $300,000 per year and needs to be taken for life. That means major profits for the drug company and healthy living for those lucky enough to get the drug. But as Barry Werth, the author of The Billion Dollar Molecule, asked in 2014, “How can we justify an eye-wateringly high-priced specialty drug for a few hundred patients that will cost as much as all other medications for everyone else with the disease combined?” He rightly worried not only about the cost to society of these specialized drugs, but also that working on rare diseases to develop similar specialized drugs means “less talent and money will be devoted to widespread scourges like diabetes” and tuberculosis. We applaud the development of a specialized drug market on behalf of the patients who will be helped. But pricing is a societal problem, and we must recognize that as long as specialized drugs are priced at exorbitant levels that only the rich and governments can afford, no matter how good and how genome-specific they are, they will serve primarily to make medicine even less affordable for society than it is today.
In addition to specialized drugs, other advances have occurred that translate pharmacogenomics into clinical practice. Today 10 percent of drug labels for FDA-approved drugs contain pharmacogenomics information. This information may describe variability in how individuals may respond to drugs, risks for adverse events, genotype-specific dosing, or information on how a particular drug works. Even among this small number of drugs where pharmacogenomic information has been approved for clinical practice guidance, controversies remain. A good example is clopidogrel (Plavix®), one of the most commonly prescribed drugs in the United States for prevention of heart attack and stroke in patients at increased risk for these problems. Clopidogrel works by preventing platelets from clumping together and forming blood clots. Uses of clopidogrel, particularly when combined with aspirin, include prevention of clot formation following placement of coronary artery stents and in a condition called acute coronary syndrome, a life-threatening form of coronary heart disease in which the heart muscle does not receive enough oxygen. This can occur during a heart attack or unstable angina when a person has severe chest pain, even at rest. Some individuals have specific variants in the CYP2C19 gene that result in poor metabolism of clopidogrel that decreases its effectiveness in clot prevention.
The FDA approved a new label for clopidogrel in 2010 with a “boxed warning” about the diminished effectiveness of the standard drug dosing in individuals with these gene variants—that is, “poor metabolizers.” As with warfarin, it was left to the clinician to decide whether genetic testing should be performed to guide therapeutic use of the drug. It also left vague how combinations of different variants of the gene should be taken into account (“intermediate metabolizers”). To make matters more confusing, two randomized trials suggest that there is no difference in outcomes of patients who have “poor metabolizer” variants and those who don’t.
If your doctor were to start you on clopidogrel, would genetic testing for CYP2C19 gene variants be useful in deciding the correct dosage? Unfortunately, “not yet” is still the right answer. This is because genetic information would likely be of only limited value. Why? First, in addition to at least twenty-five variants of the CYP2C19 gene, variants in other genes, such as ABCB1 and CYP3A4, have also been shown to affect the drug response to clopidogrel. Second, multiple other factors influence the effects of clopidogrel, such as drug-drug interactions (for instance, aspirin and proton pump inhibitors like omprazole, or Prilosec), smoking, diet (for example, caffeine intake), differences in platelet function tests, and diabetes. Third, for use in the emergency department, for instance when a person has an acute myocardial infarction (heart attack), either an existing genome in the medical record would have to be interrogated or a rapid and accurate genetic analysis would have to be performed. Neither is available today. Finally, further studies are needed to establish whether and which specific genetic variants affect clopidogrel response. On the other hand, in some cases gene testing prior to initiating drug treatment can be life saving. For example, HLA-B*1502 gene variant testing is now required by the FDA prior to starting Asian patients on the antiepileptic drug carbamazepine, as described in Olivia’s case in chapter 2.
Pharmacogenomics of the Future
Pharmacogenomics is only one of a series of new “-omics” technologies (e.g., proteomics, toxicogenomics, antibodyomics, infectomics) that will add to the ability of physicians to diagnose and characterize diseases, predict and improve drug efficacy, and reduce adverse drug reactions. These technologies have great potential if judiciously used in combination in the clinic. A letter to the editor in Science provides a reasonable caution: “The idealistic goal of personalized medicine and individualized drug therapy, which needs a holistic understanding of each individual patient’s unique ‘–omics read-out,’ is most likely unattainable—for the vast majority of complex traits—by advances in technology alone.”
Much more work needs to be done before you can expect to see pharmacogenomics routinely incorporated into your health care or diet. This was well stated by Steven Nissen from the Cleveland Clinic Foundation (although we think he was wrong about BiDil, he remains a credible source on evidence-based medicine): “Unfortunately, in the popular press, the concept of personalized medicine has taken on a nearly cult like following with pronouncements describing how future physicians will use therapies that reflect the specific genetic makeup of individual patients. No matter how promising, pharmacogenomic approaches to treatment must withstand the same scrutiny required of all therapeutic advances—careful evaluation through well-designed randomized clinical trials.”
This will require large numbers of individual research subjects willing to share their DNA sequence and medical records with researchers (on the scale of millions) in national and international collaborative efforts. How to decide whether to take part in such projects is a subject we explore in chapter 9. Much has been and is being learned, but we are still a long way from the day when we can use genomics to identify “the right drug in the right dose for the right patient.”
Notwithstanding the current limitations and challenges of using pharmacogenomics in clinical practice, genomics has been revolutionary in cancer research. For generations the mainstay of cancer diagnosis has been based on microscopic appearance of tumors followed by hit-or-miss, one-size-fits-all approaches to treatment. However, as we discuss in some detail in chapter 8, increasingly we are seeing such traditional approaches being replaced by integration of genomics for diagnosis, prognosis, and targeted therapeutic options. It seems fair to conclude that for the immediate future most of the promise of pharmacogenomics will be for patients with serious, even potentially fatal, cancers.
To avoid the temptation of prematurely concluding that genomics really is much more about promises and predictions of success rather than actual accomplishments, we delay our exploration of cancer genomics to look at the beginning of life, where genomics has already made major changes that directly affect, among others, all pregnant women and their fetuses and children: reproductive medicine, including genetic screening of potential parents, embryos, fetuses, and newborns. Sometimes termed the reprogenomics revolution, reproductive genomics is the subject of the next three chapters.
WHEN THINKING ABOUT
Racism and racial stereotyping have
plagued genomics since its beginnings.
Do what you can to prevent genism.
Pharmacogenomics is in its infancy, and it is likely
to grow only after your genome is digitalized
and linked to your electronic health record.
It is unlikely that you or your physician
will use genomics to determine your
drug or dose in the near future.
Foods are analogous to drugs, and we are still
learning about how our genes can determine
how we react to foods and drugs.