Cr515s Minitutorial--pharmacogenomics

CR515S—Mini-Tutorial by Nancy Reinhold

Personalized Medicine: The Promise of Pharmacogenomics Nancy Reinhold A Swiss-born chemist named Johannes Friedrich Miescher first isolated a substance he called nuclein (later named nucleic acids) from white blood cells in the late 1860’s, but was unaware of its purpose. Approaching a full century after Miescher’s discovery, Oswald Avery, Maclyn McCarty and Colin MacLeod performed the experiments that, in 1943, revealed deoxyribonucleic acid (DNA) as the carrier of genetic information, a role assumed to be the work of a protein at that time in history. Taking from the work of Rosalind Franklin and Maurice Wilkins, James Watson and Francis Crick published the molecular structure of DNA in a 1953 issue of Nature, thereby paving the way for the development of the biotechnology industry and precipitating further evolution of modern genetics research. In 1990, the US Department of Energy and National Institutes of Health officially launched the Human Genome Project with an aim to map the entire human genetic blueprint. More than a decade later, this project culminated in the publication of the complete sequencing of all twenty-four human chromosomes by 20031. From the extensive databases generated in the collaborative effort of the Human Genome Project, we stand to advance our understanding of the role of genes in innumerable metabolic pathways and biochemical processes as well as achieve greater insight into genetic risk factors for disease. This expanded knowledge brings with it a possibility of personalized medicine or pharmacogenomics--predicting how individuals will respond (or not) to various drug therapies based on their genetic makeup. This paper will explore the concept of personalized medicine by explaining how pharmacogenomics may revolutionize healthcare and what potential impact it may have on the conduct of clinical trials. Additionally, examples of current pharmacogenomic breakthroughs will be discussed along with a brief appraisal of the potential hurdles to making personalized medicine a reality.

Basis of personalized medicine and its potential The safety and efficacy of drugs can be influenced by numerous factors, including lifestyle, age, environment, diet, health status and concomitant use of other medicines just name a few. Beyond these factors, there is also a genetic basis for the proper function of pharmaceuticals and how they are processed in the body. Single nucleotide polymorphisms or “SNPs” are often mentioned when discussing pharmacogenomics. A SNP refers to substitution of a single nucleotide in the base pair sequence of a gene that may impact the function or expression of the gene. They occur throughout the human genome at frequency of about 1 per 1000 base pairs2 (to learn more about

CR515S—Mini-Tutorial by Nancy Reinhold SNPs, an illustrated tutorial can be found at http://www.genome.gov/Pages/Education/DNADay/TeachingTools/MakingSNPsMakeSense.htm l.) Genetic polymorphisms, defined as “a variation in DNA that is too common to be due merely to new mutation,”3 can influence drug efficacy or toxicity through variation in cell signaling, drug receptors, drug transporters and drug metabolism.4 Currently, doctors choose which medications to prescribe for their patients based on their overall familiarity with different medications, which they may have gained through their own clinical experience or through education by various other means. As one article puts it, “the final choice of medications remains empirical, if not arbitrary, for the great majority of patients.”5 Ultimately, the clinician does not know in advance if the drug they prescribe will have the desired effect or if perhaps the patient may even have an adverse drug reaction. It is in this clinical decision that the potential of pharmacogenomics is most greatly valued. One major focus of pharmacogenomics is determining which genes affect the action or bioavailability of drugs, understanding gene function in key biochemical pathways and the impact of variation in those genes. Increased knowledge about the effects genes have on absorption, distribution, metabolism and/or excretion of medicines could guide treatment decisions by allowing predictions of effectiveness or toxicity in populations possessing a particular genetic profile. A patient’s genetic makeup may manifest as an enzyme deficiency that precludes the necessary conversion of a drug into its active forms within the body, rendering the drug ineffective. Conversely, the patient’s genotype may be such that they metabolize the drug too quickly such that standard dosing is not effective, thus requiring the patient to receive larger doses in order to achieve a therapeutic level of the drug. Yet another scenario is when the patient fails to excrete drug metabolites appropriately, leading to accumulation of chemicals that are toxic and giving rise to a serious adverse reaction to the medication. Through greater understanding of the relationships between gene function and pharmacokinetics, physicians could avoid administering a drug that is unlikely to benefit a patient, modify dosing regimens to improve effectiveness of the treatment and limit toxicity or adapt their strategy for patient care (e.g. more closely monitor patients known to have a higher risk of adverse drug reaction). In addition to enhancing our ability to predict drug response, pharmacogenomics also offers a more directed approach to drug development. Understanding the mechanism of disease as dictated by gene expression enables the identification of targets for therapy and aids basic research into the development of products that are anticipated to have specific action in critical

CR515S—Mini-Tutorial by Nancy Reinhold steps of pathogenesis. Not only is this kind of information significant for improving the ability to treat existing disease, but through this advanced learning, an individual’s predisposition to certain illnesses might be revealed, allowing opportunity to take preemptive measures that would delay or prevent onset of disease. It may even be possible to institute early treatments that interrupt the progression to disease rather than merely reacting after symptoms manifest when the illness may already be in advanced stages.

Modern-day examples of personalized medicine While the advantage of pharmacogenomics may seem like something that will only be attained in the distant future, there are already a number of examples where personalized medicine has become a reality even if its application is still somewhat rudimentary. A review by Tsai and ∗

Hoyme and a publication by McLeod and Evans list the following genes and the findings of each gene’s effect on responses to certain drugs:

•

TPMT (thiopurine methyltransferase)—Individuals with inherited deficiency of TPMT experience higher incidence of hematopoietic toxicity with standard dosing of azathioprine, mercaptopurine and thioguanine (drugs used to treat various conditions like rheumatoid arthritis, inflammatory bowel disease and childhood leukemia).

•

CYP2D6 (debrisoquin hydroxylase)—Patients with mutation causing low functioning of CYP2D6 have increased occurrence of side effects from certain drugs (e.g. clozapine, amiodarone, flecainide) used for psychiatric, cardiovascular and neurological conditions. Reduced activity has also been associated with ineffective analgesia with morphine.

•

ApoE4 genotype—“83% of [Alzheimer’s] patients without an apoE4 genotype had an improvement in total response and cognitive response to tacrine therapy, compared with 40% in patients with apoE4”5

Impact to pharmaceutical industry, regulatory agencies and healthcare It is yet unknown just how extensively medical science will be reshaped by pharmacogenomics. One anticipated benefit is that safer drugs will more quickly become available while potentially being more cost-effective. If pharmacogenomics delivers on that promise, patients will surely reap the rewards, but what are the ramifications for the pharmaceutical industry, regulatory



agencies and healthcare?

The literature highlights other additional examples, but this list has been condensed to include the cases that were less technically complex.

CR515S—Mini-Tutorial by Nancy Reinhold Pharmaceutical companies and regulatory agencies have been heavily criticized for insufficient post-market surveillance in recent years, leaving many to question if we are trading safety for faster availability of new therapies. Given the troubling statistic that 10% of new drugs approved in the timeframe of 1975 to 1999 were pulled from the US market due to safety reasons6, the concern may be justified. Could pharmacogenomics be the answer to the problem? Having a firmer grasp on the mechanism of action for both drug and disease through pharmacogenomics may diminish the need for post-market studies as the safety profiles of new drugs would be more predictably established pre-approval through scientific and theoretical reasoning. Minimizing post-market surveillance would decrease the resource burden for sponsors and regulators alike and, ostensibly, still promote better drug safety than proffered by the current system. Additionally, pharmaceutical companies may see a decrease in product liability claims as the “one-size-fits-all” approach to prescribing a medication becomes less commonplace. The anticipated contribution of pharmacogenomics to improve discovery of new tools to fight disease is substantial, but its implications for clinical research are no less propitious. It is believed that prospective screening of subjects for favorable SNP profiles and excluding those for whom an investigational drug is suspected to be ineffective or harmful would be useful in lessening the cost of performing clinical trials7. Presumably, the expense of generating clinical trial data would be reduced since fewer subjects would need to be tested to show efficacy and data analyses could be achieved more quickly. Cost reduction would surely be welcomed by the pharmaceutical industry and may even make them less reluctant to perform additional clinical trials if/when requested by regulatory authorities. Another way that pharmaceutical companies and consumers may benefit from pharmacogenomics is through expansion of product pipelines. Drug candidates that do not progress past Phase I testing by today’s standards may be more likely to continue as indications are narrowed to a smaller demographic for whom the therapy shows promise with fewer risks. New indications for existing drugs may be more easily discerned as we discover different diseases with shared biochemical/molecular pathways. Products previously withdrawn from the market due to questionable safety profiles could be revived if more information is gathered about which patients can safely use the medication.

Obstacles to making personalized medicine a reality

CR515S—Mini-Tutorial by Nancy Reinhold Although many remarkable milestones have been achieved in the field of pharmacogenomics, there is still a long way to go and many barriers to overcome in order to make personalized medicine an every day practice. First, the amount of information still left to interpret and understand is formidable. The human genome is estimated to contain well over 20,000 genes and the functions of over 50% of discovered genes are unknown1. As of 2001, scientists had already identified approximately 1.4 million SNPs1,6 (there are over 3 billion base pairs total in the human genome). Evaluating the role of each gene and what impact, if any, SNPs have in biological and physiological function is going to take a considerable amount of time. Add to these numbers the sheer complexity of genetic interactions and the timeline stretches. From what is currently known, it is clear that disease states are frequently attributed to a collection of SNPs as opposed to just one. Moreover, genes are not independent actors in disease processes and drug response. As Severino and Del Zompo state, “For most adverse reactions, predisposition seems to be multifactorial, involving not only defects at multiple gene loci but also environmental factors such as concomitant infections.” One example of a treatment that is influenced by both behavior factors (i.e. diet/nutrition) and by allelic variation is warfarin anticoagulation therapy. Inconsistent dietary intake of foods rich in Vitamin K can cause fluctuations of the nutrient’s concentration in the bloodstream, which has been known to complicate therapeutic maintenance of warfarin treatment. Now a relationship between warfarin dose response and two polymorphisms, CYP2C9 and VKORC1 has also been determined in several studies. The combination of these factors account for 55-60% of response variability.8 Elucidation of how much each factor contributes to dosing variability, both independently and together, requires additional investigation. Comparable to the warfarin example, it is likely that many other mixed relationships will have to be delineated before clinical application can be realized. Even once the mysteries of gene function are expounded, there exist several other potential obstacles to transcend for pharmacogenomics to become a useful tool. Analyses of cost, pragmatic application logistics and adoption into everyday clinical practice by physicians are just a few implicit impediments. On the subject of cost, we must consider that each patient’s genetic composition would need to be ascertained through testing as the starting point for personalized medicine. Presently, the cost of microarray testing may be exorbitant for some individuals to pay out-of-pocket. Some genomic profiling services run as much as $1000 or more, depending on the type and level of testing.9,10 Fortunately, genotyping would only need to be done once in each person’s lifetime, but the availability of insurance coverage for this testing would certainly impact accessibility and equitable application of pharmacogenomic technologies.

CR515S—Mini-Tutorial by Nancy Reinhold Further confounding the issue of fair distribution among affected patients is the question of how to practically apply the information that pharmacogenomics will bring. To date the ability of pharmacogenomic data to forecast drug response has not been absolute. Often scientific findings show strong correlations between genetic influences and therapeutic efficacy or drug response modulation, but few have been 100% predictive. Should a patient whose genetic profile indicates a 60% likelihood of serious adverse reaction to a drug be strictly denied access to that therapy? What if no other treatments are available to that patient for their condition? These questions may be no more easily answered in a future where pharmacogenomic application is more commonplace than it currently is. Of no trivial pertinence to the implementation personalized medicine is the matter of how pharmacogenomics is successfully incorporated into daily clinical practice. Roden, et al summarize this issue well, saying, “Even for common and well-defined genetic variants with reproducible and important consequences for disease or drug therapy, wide acceptance by the medical community has been slow.”11 Finding ways to remediate the reluctance of practitioners to integrate pharmacogenomics into patient care situations will be vital to making personalized medicine a reality. The Future & Other Considerations Besides the hindrances to the developing field of personalized medicine heretofore mentioned, are there any other considerations for making pharmacogenomics a routine step in our healthcare norms? With all the benefits we have to gain from personalized medicine, there should be no opposition to its progression, right? Again, this is a complicated matter that has yet to be clarified. Arguably, one of the biggest factors in making personalized medicine a reality is the willingness of the pharmaceutical industry to embrace pharmacogenomics. After all, a critical ingredient of personalized medicine is the continued development and testing of medical products that is typically undertaken by the pharmaceutical industry. Surely industry can appreciate the possibility of spending less to perform the clinical trials that are required to bring their products to the market. And having a higher probability of success for a larger number of drug candidates would be attractive as more marketed products might translate into bigger revenues. An added bonus is the reduction in financial losses due to product liability litigation as the safety profile of their products is more clearly defined. It would appear industry has nothing to lose, but is it really so simple? With pharmacogenomics as a new driving force to drug development, we might

CR515S—Mini-Tutorial by Nancy Reinhold also anticipate more rigorous labeling restrictions which could reduce market share of a product. Other regulatory requirements could become more restrictive as FDA and other international agencies adapt their policies to growing pharmacogenomic technologies. Sales would be less a function of successful marketing strategies and more dependent on the predominance of favorable genetic profiles throughout the general population of consumers. Innovation could be delayed as basic research struggles to unravel the complexities of gene-drug interrelationships, making the research more laborious and costly. As pharmacogenomic knowledge expands, might there come a time when the current drug approval paradigm becomes unacceptable and no medical products would be approved without studying genetic correlations?

If not, the public may become more

reticent to use medical products that could not be developed with genetic compatibility in mind because pharmacogenomic progress is lagging. How all of these considerations affect industry’s perspective remains to be seen. Finally, what are the ethical considerations of personalized medicine? Knowledge of one’s genetic predispositions can be a sensitive issue, particularly if the information can be used for prejudicial motives (i.e. denying medical insurance coverage). Might certain groups of people with rare genetic profiles be ‘orphaned’ or have fewer treatment options available to them because development of drugs for them wouldn’t be profitable? Is it problematic to exclude certain genetic demographics from clinical trial participation based on a putative expectation that their genotype may predispose them to being refractory to therapy or prone to adverse reaction? Debate over these questions will undoubtedly ensue if/when pharmacogenomics and personalized medicine is mainstreamed. Conclusion Clearly, the field of pharmacogenomics presents an exciting array of possibilities for improving disease prevention and management, but the promise of personalized medicine warrants tempered optimism. While pharmacogenomics may lead to greater availability of safety information and enhanced ability to put drug response expectations in more quantifiable terms, can we confidently assume that medical products will be safer or that clinical decisions will necessarily be, in general, easier? After all, it is still entirely possible that several different diseases could share similar/related biochemical routes and the process of treating one condition may simply create another disease state. The benefits to be gained from pharmacogenomics are great, but the challenges are best not to be underestimated.

CR515S—Mini-Tutorial by Nancy Reinhold

1

US Department of Energy Human Genome Program. About the Human Genome Project page. Available at: http://www.ornl.gov/sci/techresources/Human_Genome/project/about.shtml. Accessed 13 November 2009. 2 Van Delden J, Bolt I, Kalis A, Derijks J, Leufkens H. Tailor-Made Pharmacotherapy: Future Developments and Ethical Challenges in the Field of Pharmacogenomics. Bioethics. 2004; 18: 303-321. 3 MedicineNet.com Web site. Available at: http://www.medterms.com/script/main/art.asp?articlekey=4992. Accessed 22 November 2009. 4 Tsai YJ, Hoyme HE. Pharmacogenomics: the future of drug therapy. Clinical Genetics. 2002; 62: 257-264. 5 McLeod HL, Evans WE. Pharmacogenomics: Unlocking the Humane Genome for Better Drug Therapy. Annu Rev Pharmacol Toxicol. 2001; 41: 101-21. 6 Severino G, Del Zompo M. Adverse drug reactions: role of pharmacogenomics. Pharmacological Research. 2004; 49: 363-373. 7 National Center for Biotechnology. http://www.ncbi.nlm.nih.gov/About/primer/pharm.html. Accessed 20 November 2009. 8 Ambry Genetics Web site. Available at: http://www.ambrygen.com/clinical_diagnostic_and_carrier_testing/test_Warfarin_Sensitivity.asp. Accessed on 24 November 2009. 9 SeqWright DNA Technology Services Web site. Available at: http://www.seqwright.com/personalgenomicprofiling/personalgenomicintro.html. Accessed on 22 November 2009. 10 GeneDx, Inc. online Test Information Sheet. Available at: http://www.genedx.com/site/uploaded_files/Info_sheet_GenomeDx_Aug09.pdf. Accessed on 22 November 2009. 11 Roden DM et al. Pharmacogenomics: Challenges and Opportunities. Annals of Internal Medicine. 2006; 145: 749-757