Is Predictive Prevention The Future Of Medicine
Genetically predicting which drug is more harmful to whom and who is more likely to develop which disease is becoming a reality — a primer on pharmacogenomics and GWAS
Betty’s joints have been acting up recently. She was enjoying life as usual when first her right toe and then the left knee started troubling her. The joints would become red, stiff, swollen, and unbearably painful. Her physician explained that crystals of uric acid have started collecting inside the joints and prescribed a medicine— allopurinol — to keep her uric acid levels in check. The condition is called gout and is fairly common.
A week later, large blisters appeared on her abdomen which, within days, spread to involve her legs, face, mouth, and arms. More than half of her body surface area blistered and started to peel off, leaving fleshy red underskin behind. A respiratory infection followed soon and her kidneys took a hit too, requiring 6 days on dialysis. She survived but took more than two months to recover fully.
Betty can consider herself lucky, as nearly half of patients that develop toxic epidermal necrolysis — TEN for short — succumb to the devastating condition. TEN is a rare but potentially fatal allergic skin condition caused by a number of prescription medicines. Allopurinol is one of them. Not everyone who takes the drug develops the complication though. The estimated frequency is less than 1 case per 1000 persons taking the medicine. Can we predict who is more prone? The answer is an encouraging YES, thanks to pharmacogenomics — an emerging field of medical genetics and the cornerstone of personalized medicine.
Predicting Adverse Drug Outcomes
How you react to a particular medication is, in part, determined by your genetic makeup. This is expected since genes are the ultimate drivers of our molecular destiny. Continuing with our example of allopurinol, genetic studies have shown that the presence of a variant of the HLA-B gene (known as HLA-B*5801) increases the risk of adverse skin reactions. HLA-B*5801 is found in Han Chinese at a higher frequency than other ethnic groups and, not unexpectedly, the adverse events are more common in Han Chinese. Testing for this gene before starting allopurinol is found to be a cost-effective way of preventing TEN and other skin reactions in this ethnic group. The Clinical Pharmacogenetics Implementation Consortium (CPIC) recommends such testing and it is being incorporated into clinical practice in Taiwan and Singapore, among other countries.
Allopurinol is not the only drug whose adverse effects are known to be, at least in part, genetically determined. CPIC, a project of the US Department of Health, currently provides guidelines on 25 gene-medicine pairs for which there is evidence that prior testing is useful. Pharmacogenomics, the science of tailoring patients medications based on their genetic makeup, is increasingly making its way from research laboratories to clinical practice. The ability to prescribe ‘the right medicine for the right patient’ is an essential step towards personalized medicine.
The CPIC drug list includes such commonly used medicines as aspirin, the blood thinner warfarin, and anti-depressants called SSRIs. Routine susceptibility testing for them is not common practice currently, but with mounting evidence and refinements in testing, pharmacogenomics is expected to play an increasing role in clinical practice in the near future. Moreover, the application is not limited to susceptibility testing for adverse effects only. Genetic testing can be used to predict responsiveness, or the lack of it, to specific medications…
Predicting Response To Treatment
Hypertension or raised blood pressure is ubiquitous in humans, irrespective of race or ethnicity. The preferred treatment option for a newly diagnosed 47-year-old black patient, however, may be different than a 52-year-old white European, as the former is likely to respond better to one group of BP-lowering drugs than another. This difference is undoubtedly due to underlying genetic differences though we don’t exactly know them — yet. For hypertension and other common medical conditions, the role of genetics is limited mainly to theory currently. On the other hand, for cancer, genetic risk stratification and response prediction is a reality and an established clinical practice. For example, a patient with acute myeloid leukemia — the nastiest of blood cancers — with a mutated FLT3 gene is expected to have a worse outcome when treated with standard chemotherapy compared to a patient who doesn’t have the mutation. Such predictive genetic testing is routinely used in oncology to guide treatment plans.
As our understanding of the genetic basis of commonplace disorders like hypertension, diabetes, and heart diseases improves, it is plausible that such genetics-based prescription will make its way into clinical practice for these disorders as well.
Genetics and Susceptibility to Diseases
Nancy Wexler was 21 years old in 1978 when her mother, Leonore Wexler, in her fifties, was diagnosed with Huntington’s Disease (HD) — a devastating, hereditary disorder affecting the nervous system that renders the patient incapable of performing day to day activities. Children of a patient with Huntington's have a 50% chance of developing the disease. In other words, their fate is essentially decided by a coin toss. What makes the matter worse is the fact that most of the individuals carrying the culprit gene do not show symptoms of the disease until the third or fourth decade of their lives. You will only know if the odds are not in your favor when you develop the disease. Thus for children of patients with Huntington’s, life turns ‘into a grim roulette’. Or as one patient puts it ‘the terrible waiting game, wondering about the onset and the impact.’
After their mother’s diagnosis and subsequent death, the waiting game had begun for Nancy and her sister. With no way to find out if she was carrying the culprit gene or not, life was in limbo. She, however, decided to embark on a life-defining project to find out the culprit gene of Huntington's disease and devise a way for its testing. Years of tiring yet inspiring research followed, detailed in the family memoirs by Nancy’s sister, Alice Wexler. Finally, in 1983, Nancy and James Gusella declared their findings of mapping the HD gene on chromosome 4, making it possible to genetically test the likelihood of a person developing the disease. This was one of the earliest genetic tests available for clinical use. With no way to prevent the development of HD if she was found to be a carrier of the HD gene, Nancy decided not to take the test herself. For a condition like coronary artery disease, however, such pre-knowledge can be potentially life-saving. Currently, she is 75 years old and a professor of Neuropsychology at Colombia University.
Huntington's Disease is rare. Other genetic disorders like thalassemia, cystic fibrosis, or sickle cell disease are more common. Today, it is common practice to do genetic testing for these conditions before a baby is born. These genetic disorders may be more common than HD but when considered in the whole spectrum of disease burden they make a minute proportion. The diseases responsible for the most number of deaths, for instance, are cardiovascular conditions and cancer. Moreover, in contrast to HD, if a person’s susceptibility to heart diseases is known, lifestyle modifications and perhaps medications can be helpful in reducing the chances of a fatal outcome. Our ability to predict a person’s genetic risk of developing cardiac disease or cancer, however, is limited to a handful of rare hereditary syndromes — until now that is.
GWAS: Guilty by association
Our idea of genetics and inheritance is shaped by high-school Mendelian genetics, where a single or a few known genes act in predictable ways to determine inheritance patterns of physical characteristics and diseases. This concept holds true for a few purely genetic diseases like sickle cell disease or cystic fibrosis. The vast majority of our diseases, however, are complex and multifactorial. Their increased incidence among family members nevertheless points towards a genetic component in these so-called non-genetic conditions.
Since genes are inherited at birth, we can — theoretically — predict the likelihood risk of these diseases years or even decades before their development. The problem, however, is that such diseases don't follow simple Mendelian inheritance since their genetic component is a product of variable contributions from a number of genes many of which are unknown, rather than a single known gene. In other words, these diseases are polygenic rather than monogenic.
Over the past 30 years, with genome-wide association studies (GWAS) a number of susceptibility gene variants have been discovered for common diseases like hypertension, inflammatory bowel disease, cardiac disease, and breast and colon cancer. GWAS scan a person’s whole genome (the whole set of DNA a person inherits) for an association between genetic variants and the development of a disease. Based on these studies, polygenic risk scores have been developed for various diseases including coronary artery disease, atrial fibrillation, type 2 diabetes, inflammatory bowel disease, and breast cancer. Such predictive scores can, in the near future, be used in clinical practice.
Take cardiovascular diseases for instance. Lifestyle factors like smoking, lack of physical activity, and high lipid intake are known to increase their risk. Not all individuals who smoke and eat steaks go on to get a heart attack, however. The difference in individuals’ response to environmental risks — as these factors are often called — is due to susceptibility gene variants or the lack of them. Researchers have identified more than 200 such gene loci for hypertension and more than 50 for coronary artery disease. A polygenic risk score based on these genetic variants for the European population showed that 8% of the population had inherited a genetic predisposition that conferred a threefold or higher increased risk for coronary artery disease, compared to the general population. The same study identified that 1.5% of the population had a genetically determined threefold or higher increased risk of developing breast cancer.
Such genetics-based predictive risk models, if incorporated into clinical practice, can help in devising individualized, risk-stratified preventive strategies. A person who is known to carry genetic polymorphs putting him at an increased risk of developing coronary artery disease will likely benefit from more extensive dietary modifications and will be more motivated to stay away from smoking.
The future of medicine is in prediction
Despite the fact that for many diseases prevention is far more effective than treatment, our current health model is geared towards treating a disease when it develops. This is in large part due to our inability to reliably predict the occurrence of diseases beforehand. With the uncovering of the genetic basis of commonplace diseases like hypertension, diabetes, coronary artery disease, predictive testing, and pre-emptive strategies are gradually becoming a reality. On the therapeutics front, similar predictive testing is paving way for personalized medicine.
Gene therapy and CRISPR are stealing the limelight for being the cool kids around, while the nerds — GWAS and pharmacogenomics — are silently making their way to mainstream medical practice. Either way, the future of medicine is exciting.
