Pharmacogenomics and Drug Development: The Impact of US FDA Postapproval Tracking on Clinical Pharmacology
Erin Farrell; Jonathan Usuka
Personalized Medicine. 2008;5(2):133-139. ©2008 Future Medicine Ltd.
Posted 06/25/2008
Abstract and Introduction
Abstract
Severe adverse drug reactions to commonly prescribed drugs such as Vioxx® have led to a call for increased scrutiny in deciding which patients are given which drugs, and how much drug they should receive. A personalized approach to medicine offers a larger variety of drugs and doses that would be prescribed only to a subgroup of patients. Pharmacogenomics could help divide patients into these subgroups based on variation in the genes either causing the disease or encoding the principle drug-metabolizing enzymes. Given the cost and infrastructure associated with assembling genetic data, drug sponsors, regulatory agencies and clinicians each play a role in the collection, storage and oversight of pharmacogenetic information. The 110th Congress is in the process of making changes to the drug-approval process and the role of genetics in that process.
Introduction
Drug development is changing. The biology of modern disease is increasingly intricate and subgroups of the population often experience widely varied responses to treatment. As a result of unexpected adverse drug reactions (ADRs), many drugs are withdrawn from the market by their sponsors ( Table 1 ).
Recent incidents of ADRs from drugs such as Vioxx® have led to criticism of our drug-approval process and, increasingly, our lack of systematic postapproval monitoring of drugs. Much of the variation in drug response is rooted in our genome, and the SNPs that underlie our individuality can also control the effects of our drugs.[1] Pharmacogenomic research seeks to gain insight into the individual metabolism of drugs, and should be included as a vital part of the new critical path to drug development. Pending legislation in the 110th Congress could improve cooperation between drug sponsors, the US FDA and clinicians to ensure that genetic data are methodically considered during the approval and postapproval stages of drug development.
A System in Crisis?
Vioxx® & the State of the FDA
When Merck voluntarily pulled Vioxx® from the market in September 2004 owing to adverse cardiovascular events, it exposed a fear that our current system of drug safety was in dire need of improvement.[2] There were a number of contentious issues including a discrepancy in the number of heart attacks reported in the New England Journal of Medicine in May 2000, the incidence of other adverse cardiovascular events, and the statistical analysis of reported data.[101] It is estimated that 88,000 Americans had heart attacks while receiving the drug in its 5-year lifespan, 38,000 of whom died.[3] The incident drew attention to the importance of postapproval monitoring, as well as weaknesses in the current system of ADR reporting.
As of March 2007, Merck had won nine cases and lost five in the resulting lawsuits from Vioxx.[4] But regardless of their outcome, pressure on the FDA led them to examine their system of drug safety through an external review by the Institute of Medicine (IOM).[5] The Future of Drug Safety and its summary Action Steps for Congress contained 25 recommendations that aimed to strengthen the pre- and post-approval phases of drug development to promote a ‘lifecycle' approach to the study, regulation and communication regarding the risks and benefits of drugs. The report highlights the lack of FDA authority in enforcing sponsor compliance throughout the process, particularly postapproval, and suggests fines, injunctions and drug withdrawal as necessary enforcement tools. It recommends that the agency add risk-management programs that control direct-to-consumer marketing, labeling practices for newly approved drugs, and distribution contingent on physician training and maintenance of an active adverse event surveillance system. The IOM report also calls for Congress to pass legislation to ensure compliance by both the FDA and drug sponsor, as well as adequate resources for the increased FDA surveillance. This would lessen the FDA dependence on funds from the controversial Prescription Drug User Fee Act and increase credibility. Whether the state of affairs at the FDA has been exaggerated by the media or not, it is clear that premarket drug-safety evaluation through gold-standard randomized, controlled trials should be coupled with equally objective postmarket analysis of safety and efficacy, and the funding for such evaluation must be sufficient.[2]
Pharmacogenomics & the Critical Path to safe & Effective Drug Development
The inclusion of pharmacogenomic data collection and analysis at each stage of a drug's lifecycle would directly address many of the weaknesses in the drug-safety system addressed in the IOM report. Premarket assessments consider small, select groups of patients where the full range of ADRs might not be obvious. The Human Genome Project highlighted the genetic similarity of all humans, finding that two random individuals from any one ethnic or geographical group are almost as genetically different as any two random individuals from the entire world.[6] However, many analyses demonstrate that genetic data can be used to accurately classify humans into populations using genetic markers such as microsatellites,[7] SNPs[8] and other mutations. An understanding of the distribution of interindividual genetic variation underlying phenotypes, disease susceptibility, and responses to treatment could reduce the risk of ADRs and keep some drugs from being pulled from the market. It can be the case where a drug that elicits a serious ADR in some or most of the population can still be an important treatment option for other patients, as is the case with BiDil®, warfarin and Herceptin®.
BiDil & a Small Step Towards Personalized Drugs
BiDil (isosorbide and hydralazine in a fixed-dose combination) is the controversial, first FDA-approved drug that is marketed for a single ethnic group, in this case African–Americans for the treatment of congestive heart failure (reviewed in[9,10]). In 1997, the FDA rejected the New Drug Application for BiDil, as the statistics from the study run by the Veterans Administration for Medco, the drug's sponsor, were too disordered and inadequate.[10] Years later, the data were re-examined in the context of race, and since then BiDil has been indicated in self-identified black patients to improve survival and symptoms of heart failure.[11,12] Although, in principle, opening the door to treatment of a subgroup of the population that did not respond to other drugs is a step forward, the process that led to BiDil's approval was likely biologically oversimplified[13].When the most distinct populations are considered and hundreds of loci are used, individuals are often more similar to members of other populations than to members of their own population.[14,15] Therefore, using geography or rudimentary ancestry to divide humans into populations is very crude and is in no way a substitute for real genetic profiling.[7,16]
Warfarin Pharmacogenetics & Dosing
Genetic profiling early in a drug's development to study variations in pharmacokinetic and pharmacodynamic profiles would help correlate genetic biomarkers with drug response. Identifying potential fast and slow metabolizers is one of the ways in which a number of treatments today are catered towards the patient. Mutation and variation in phase I, II and III drug-metabolizing enzymes in the liver can be directly related to differences in the metabolism of commonly prescribed drugs such as warfarin.[17–19] Currently, this anticlotting drug demonstrates how companion diagnostic genetic tests can help physicians modify dosing regimens to reflect pharmacogenetic differences among patients. Warfarin is the most commonly prescribed oral anticoagulant for the treatment and prevention of blood clots and is prescribed to more than 1 million patients annually in the USA.[20] The dosing of warfarin is critical, as too high a dose carries a high risk of life-threatening hemorrhage,[18] and too low a dose will be ineffective, and there is considerable interindividual variation.[21,22] Approximately half of the dose variability is related to variations in CYP2C9 and vitamin K epoxide reductase complex subunit 1 (VKORC1), which causes costly adverse events.[18,23] Human CYP2C9 polymorphisms are associated with an increased risk of overanticoagulation, and screening for CYP2C9 variants may allow clinicians to develop dosing protocols and surveillance techniques to reduce the risk of ADRs in patients receiving warfarin.[18]
Pharmacogenetic approaches can be instrumental for predicting individual differences in response to a therapeutic intervention. Recent success using a murine haplotype-based computational method to rapidly identify a genetic factor, cyp2c29, regulating the metabolism of warfarin shows that genetic variants responsible for interindividual pharmacokinetic differences in drug metabolism can be identified by computational genetic analysis in mice.[24] As a rapid alternative to extensive quantitative trait loci analysis, computational methods incorporating a database containing SNP information from 15 well-characterized inbred mouse strains mapped data from an in vivo warfarin study. This same SNP-based method correctly identified uridine 5´-diphosphate glucuronosyltransferase 1 family polypeptide A1 as the polymorphic gene responsible for differential irinotecan metabolism in mice by mapping in vitro data.[25] If other potential drugs can be so quickly mapped using pharmacokinetic or phenotypic data, polymorphic genes affecting their metabolism will be identified earlier, resulting in targeted treatments to patients, as in the case of warfarin dosing.
Herceptin: a Pharmacogenomic Model
In addition to the use of genetic markers to identify differences in a drug's metabolism, disease states can be subgrouped according to genetic markers, thus dictating which treatments will be efficacious. The most popular example of pharmacogenetic classification of a disease is Herceptin (trastuzumab) (Genentech, CA, USA). In 1987, UCLA scientists made the connection between Her2 overexpression and the aggressive cancer found in 25% of breast cancer patients.[26] Genentech cloned an antibody for Her2 and subsequently announced an agreement with DAKO to develop a diagnostic kit to screen breast cancer patients for HER2-protein overexpression. Over a decade later, the FDA approved Herceptin and Herceptest® as treatment for metastatic breast cancer, the first time that the FDA required that a diagnostic laboratory test kit used to predict patient response be made available for use with a drug.[27] Herceptin offers a clear illustration of the power of personalized medicine and highlights the importance of incorporating genetic analysis in the development and application of new therapies.
Her2-overexpression testing prior to treatment with Herceptin is an outstanding example of pharmacogenetics, but remains one of the only examples of predictive diagnostic test and drug codevelopment. The next major example will likely come from Pfizer, with their soon-to-be-approved Maraviroc®, a drug designed to bloc HIV entry into T cells, specifically the subtype of virus requiring the CCR5 coreceptor.[102] Noticing that several companies were developing CCR5-targeting drugs, Monogram Biosciences (CA, USA) independently developed a test to detect the CCR5 tropism. Diagnostic tests would more efficiently facilitate a personalized approach to medicine if FDA oversight was strengthened to include a structured submission protocol for genetic data during clinical trials and after drug approval as a part of a comprehensive postmarket surveillance program. Legislation aiming to do that is currently before the 110th Congress.
Conclusion
Progressive genetic technology is starting to change the approach to drug discovery for treatment of complex diseases. Thus, it is likely that efficiency in drug development, and safety and efficacy in drug administration, can be gained from coordinated collection of pharmacogenetic data. The pace at which new drugs will be discovered and brought safely to market is going to depend on the transparency, quality and completeness of genetic data collected from patients. A large amount of coordination between the FDA, research institutes, clinicians and pharmaceutical companies is needed. Pending legislation in the US Congress holds great potential to facilitate a faster and more transparent approval process for new drugs, while providing improved corresponding diagnostic tests.
Future perspective
Existing legislation addresses potential discrimination and ethical issues through the Genetic Nondiscrimination Act of 2007, which was passed in the US House of Representatives by a vote of 420 to three. The act will protect individuals against discrimination based on their genetic information with respect to health insurance and employment. These protections are intended to encourage Americans to take advantage of genetic testing as part of their medical care. Pharmacogenomic data submission is loosely encouraged by the FDA. The next step is to equip the FDA with the resources necessary to organize and analyze pharmacogenomic data so that drug companies will more often consider subgroups of a diseased group for treatment. Biomarkers, certain gene traits or SNPs that predict biological predispositions can distinguish between groups and even individuals.[28] Early identification of these in the drug-development process will lead to increased drug efficacy and safety, and an improved understanding of disease. The amount of collaboration necessary to promote the widespread use of pharmacogenomic data in the drug-development process is on the scale of the Human Genome Project. Pending legislation focuses on postmarket drug surveillance and the importance of genomic consideration throughout drug development ( Table 2 ).
Some experienced members of Congress are taking notice of the IOM report. Senators Enzi (Republican [R]; WY, USA) and Kennedy (Democrat [D]; MA, USA) reintroduced the Drug Safety and Innovation Act in February 2007, which supports many of the IOM recommendations, giving the FDA the authority to fine drug companies for not adhering to pre-agreed postmarket commitments from their proposed Risk Evaluation and Mitigation Strategy process (S 484).[29] Also included in the extensive bill is the creation of a nonprofit organization to be known as the Reagan–Udall Institute for Applied Biomedical Research, which is to advance the agenda of the Critical Path Initiative. This program was launched in 2004 to improve the science of developing, manufacturing and evaluating the safety of drugs, and pays particular attention to the need for better diagnostics for biomarkers.[103] The House of Representatives (HR) version of the Enzi/Kennedy bill, introduced by congressmen Waxman (D; CA, USA) and Markey (D; MA, USA) (HR 1561), includes some additional resources for the FDA to increase their postmarket drug-safety authority.
Senators Grassley (R; IA, USA) and Dodd (D; CT, USA) (S.468) and congressmen Tierney (D; MA, USA) and Ramstad (R; MN, USA) introduced a bill with similar goals (HR 788), although this was through a manner more independent of the current FDA structure. This bill creates a separate Center for Post-Market Evaluation and Research, whose role is to order and monitor studies even after drugs are marketed. It is noteworthy that under the Grassley bill, the pre- and post-market arms of the FDA might operate more autonomously and increase the objectivity and transparency of the drug-review process. The Grassley–Dodd plan might require more drastic changes for the pharmaceutical industry and the FDA, and would be more costly than the Kennedy and Enzi approach.
The role of genetic information in drug safety is ignored in both sets of bills. Adverse-event reporting during and after clinical trials should be routinely recorded along with relevant SNP information for methodical biomarker evaluation. In order to be effective, both bills will require a considerable increase in funding for the FDA. Drug companies might also expect a hefty increase in the costs associated with getting a drug to market and ensuring its safety afterwards, making it near impossible for smaller drug companies to pursue drug leads. In the end, it is certain that the government will ultimately be subsidizing more astronomical drug costs.
However, a third piece of legislation might be the key to unleashing a new wave of drugs designed for subgroups of the population based on genetics. Senators Obama (D; IL, USA) and Burr (R; NC, USA) introduced the Genomics and Personalized Medicine Act of 2007 (S 976).[30] It emphasizes the need for standardized terminology, systematic data storage and collection, and habitual use of data throughout clinical trials and in the doctors office. The bill recognizes the need for a central harmonized biobanking initiative through which institutes and industry can more efficiently access and analyze large amounts of genetic information. In addition, Obama proposes funding for training in applied genomics and economic incentives for development of pharmacogenomic diagnostics, making provisions for US$150 million to sponsor research on genomics, and a 100% tax credit for the development of companion diagnostic tests that would expectantly decrease the incidence of unpredictable drug reactions.[31–33] The most promising picture for the next generation of drug development is an FDA that integrates parts from each of these bills, although the Obama legislation is critical in order to modernize drug safety. No clinical trial can predict each and every potential adverse reaction, but pharmacogenomic data can help us learn when and how to prescribe our drugs if scrutinized in the way that other clinical-trial data are.
AI
