Electronic Databases in the Life Sciences

During the past ten years, the life sciences (i.e., biology, medicine, and pharmaceutical research) have evolved from being largely low-throughput, observational disciplines to primarily high-throughput, data-driven disciplines. In other words, life sciences are becoming a “data science.” Thanks to advances in DNA sequencing, medical imaging, robotic sample handling, and high-throughput screening, it is possible to generate as much data in a day-long experiment as it might have taken for an entire scientific career. For instance, a single eight-hour sequencing run on a DNA pyrosequencer can generate enough sequence data to fill a 1,000-page book (1, 2). The resulting genome sequence could be automatically annotated in a few hours yielding an enormous volume of information that could easily occupy ten large telephone books (3, 4).

Our capacity to generate gigabytes of information on a daily basis is having a profound impact on the way that scientific information is being disseminated or delivered. Although most scientific data are still presented in scientific journals and most high-level scientific knowledge is still published in textbooks, it is becoming increasingly obvious that the paper-publishing industry cannot keep up with the pace of scientific advancement and the quantity of data that the scientific community would like to publish. Fortunately the World Wide Web (i.e., the Web) has come to the rescue. The Web makes it possible to publish and disseminate huge quantities of information quickly and inexpensively. Not only has the Web helped to save scientific publishing, it has also led to the development of a new and very important kind of scientific archive: the electronic database. Electronic databases are Web-accessible archives that contain scientifically important data that are either too voluminous to publish in a book or journal or in a format that is incompatible with paper publication. Electronic databases such as GenBank (5), the Protein Data Bank (6), or PubMed allow information to be continuously updated through the contributions of thousands of scientists or dozens of curators who continuously upload and deposit data into these resources. Electronic databases also allow their data to be searched, accessed, or displayed in ways that were simply not possible through a paper journal or a leather-bound book. Indeed, the emergence of electronic, Web-accessible databases has to be considered one of the more significant developments in the field of life sciences.

Variability in Drug Response and Pharmacokinetics

The variability in the response of individual patients to drugs was recognized long before the advent of our earliest understanding of the influence of genetics on drug response more than fifty years ago (1–3). Several reports have reviewed the topic of variability in some detail, attempting to identify and quantify the sources of variability (4, 5). Harter and Peck (4) estimated the relative contribution of different sources of variability to the cumulative total variability, and thought that pharmacokinetic variability accounted for about half of the total variability (4). Genetic factors contribute to the variability in pharmacokinetics and, in particular, they affect the enzymes and transporters involved in the absorption, distribution, metabolism, and elimination of drugs. This chapter will outline the basic principles of pharmacokinetics as they apply to the disposition of drugs in humans.

Pharmacokinetics is defined as the study of the time course of drug concentrations in the body, and can be separated into components describing the absorption, distribution, metabolism, and excretion of a drug, often abbreviated as ADME. The term pharmacodynamics refers to the study of the relationship between drug dose or concentration and the intensity and time course of pharmacological, clinical, or toxicological responses. In its simplest concept, pharmacokinetics can be thought of as “what the body does to the drug” and pharmacodynamics is “what the drug does to the body.”

Peptic ulcer and gastroesophageal reflux disease (GERD) are common benign upper gastrointestinal disorders. The major causes of peptic ulcer are Helicobacter pylori (H. pylori) infection and nonsteroidal anti-inflammatory drugs (NSAIDs), including aspirin. No matter whether caused by H. pylori, NSAIDs, or both, however, gastric acid plays a markedly important role in the pathogenesis of peptic ulcer. Furthermore, the major cause of GERD is the reflux of gastric acid from the stomach to the esophagus. Gastric acid therefore acts as the most important pathogenic factor of upper gastrointestinal disorders, and acid inhibition with a proton pump inhibitor (PPI) is the major strategy against them. PPIs in combination with one or two antibiotics are also used for the eradication of H. pylori.

PPIs are substitutes of benzimidazole and are mainly metabolized by the cytochrome P450 (CYP) system in the liver (1). The principal enzyme in this metabolism is CYP2C19 (1), although CYP3A4 also plays a role (2–6). Given the presence of interindividual differences in CYP2C19 activity, the pharmacokinetics (PKs) and pharmacodynamics (PDs) of PPIs largely depend on polymorphisms in CYP2C19 (7).

Discovering how individual genetic variation influences why some people respond to therapeutics but others show little improvement or have serious side effects lies at the heart of the promise of pharmacogenomics to improve health care. Realizing this goal, however, will depend on building a robust infrastructure that supports genomic medicine. These changes include developing affordable and ubiquitous genomic sequencing capability, creating expansive DNA sample sets annotated with detailed phenotypic information, educating health care providers on developments in genomic medicine, integrating diagnostic tools into clinical decision making, and attending to a complex array of social, economic, and ethical concerns that accompany each of these shifts in biomedicine. Achieving these major endeavors will depend on ensuring that the public not only understands the goals of pharmacogenomics, but also that the public actively supports and participates in basic and clinical research.

Public approval of pharmacogenomics research will depend on assurances that the risks will be minimal and that the benefits of its integration into clinical care will result in improved health care. A major issue of concern over the collection, storage, and distribution of human DNA is the risk associated with potential breaches of confidentiality and privacy and the threat of stigmatization and/or discrimination against individuals by insurance companies, employers, and others. The principle of protecting an individual's genetic information in the course of research and/or clinical care is a deeply held principle in the United States. Personal medical history has remained a sensitive and legally protected category of information. The special status of privacy of medical information has led to confidentiality protections as in physician-patient privilege. To attend to these requirements, researchers and health providers must develop strategies that secure individual genetic information and personal medical history that also enable basic and translation research.

With the release of the Human Genome Mapping Project data (1) and the subsequent International HapMap Project (2), a wealth of genotype information is now publically available to researchers. Pharmacogenomics can utilize this information to its advantage by screening patient samples for known functional or tagging polymorphisms and deriving associations with drug outcome and toxicity. In addition, where no known functional polymorphisms have been identified in genes involved in drug pathways, technologies have emerged to perform whole-genome screens to find novel genome regions for further study.

Often considered the “gold standard” of genotyping, Sanger sequencing performed on the same DNA region in multiple individuals (resequencing) can be used to identify both new and previously reported polymorphisms. However, this process is not cost-effective, and analysis time can be slow. Consequently, it is often used as a quality-control step to confirm genotypes reported through the various technologies discussed later in this chapter.

Obstetric pharmacology, in many respects, is still in its infancy. Because of valid concerns about the safety of drugs to both the mother and developing fetus, not as much data exist for drugs in pregnancy. There is also a distinct paucity of data regarding pharmacogenetics in pregnancy. At this time, many of the potential pharmacogenetic applications to pregnancy drug therapy are theoretical. This chapter summarizes the drug therapy for five of the major pregnancy conditions/complications, along with a brief discussion of potential pharmacogenetic or individualized pharmacotherapy applications for each. These conditions are preterm labor, depression, diabetes, the nausea and vomiting of pregnancy (NVP), and hypertension.

Preterm Labor

Preterm birth is the leading cause of morbidity and mortality in newborns in the United States. More than 12.5 percent of all live births are preterm, accounting for a large amount of health care spending for infants. The number of preterm births in the United States, as well as in many other industrialized countries, continues to rise. Preterm labor is the most common cause of hospitalization of pregnant women (1). Although the incidence and burden of preterm birth are relatively clear, the causes of preterm labor are not well understood (2).

Overview of How Genetic Diversity Arises

Genetic diversity arises from differences in the genome of humans. Mapping of the human genome has revealed that humans are approximately 99.9 percent identical relative to their DNA sequences, with differences occurring at the rate of one change in every 100 to 300 bases along the sequence of 3 billion bases that comprise the human genome (1, 2). These differences in coding sequences at these sites are termed single-nucleotide polymorphisms (SNPs) and are considered significant when they occur with a minimum frequency of 1 percent in a given population. Whereas genetic differences of 0.1 percent among individual humans appear negligible, occurrences of ∼10 million SNPs have been cataloged to date. These SNPs may occur in coding or noncoding regions of the genome and may affect gene expression or disease susceptibility. Whereas SNPs are estimated to account for 90 percent of all genetic variability, other genetic differences have been detected and may stem from errors during DNA replication, including copy number variations, insertions, deletions, inversions of bases, or other mutational events caused by environmental factors (3). This genetic variability contributes to genetic diversity among populations as well as its individual members. Genetic diversity has an impact on all manner of human traits from external appearance to disease susceptibility and response to pharmacological agents.

To gain a greater appreciation of the magnitude of the impact that genetic diversity has on human phenotypes, it is best to begin with an exploration of its historical development in modern humans. In general, race and ethnicity, which are largely defined culturally by phenotypic traits or defined by man-made designations of geographic origin, become much less distinct at the genomic level. Mutations can occur within alleles that occupy a distinct site within a given gene or genomic region and thereby help to define traits. It is thought that population genetic diversity is largely due to a combination of allelic mutation at specific sites along the genome and the selective pressure from population segregation that occurred during the migration of humans out of Africa over a period of about 200,000 years (4). Each offspring inherits two alleles, one from each parent. Thus, if either parent is not a carrier of the wild-type allele originally encoded in the genome due to local mutations within the allele, the offspring may inherit the mutant allele and be heterozygous for a given trait. The mutation may be dominant (i.e., expressed), recessive (carried silently), or, in some cases, coexpressed. When a new mutation is associated with a beneficial trait, it is thought that positive selection occurs, allowing those carrying the beneficial allele to survive, reproduce, and pass on the trait to their offspring at a frequency dictated by its pattern of inheritance. However, the major contribution to genetic diversity occurred because of the geographical isolation that resulted from tribal resettlement following migration that produced colonies of individuals with a reduced genetic pool representative of the founders of each new colony.

Differential response to the standard dose of drug therapy in patients is commonly seen in clinical practice. Environmental factors, diet, age, gender, disease severity, interacting drugs, and genetic variation all contribute to the variability in drug response. The importance of genetic variation in drug response has become more prominent with the emergence of pharmacogenomics and personalized medicine. Pharmacogenomics is the study of how genetic differences affect responses to drugs. By knowing more about a person's genetic makeup, clinicians will be better able to assess the risks and benefits associated with medications to maximize treatment success. Pharmacogenomics has the potential to have an impact on many steps of medical care, from diagnosis to tailored drug prescription, and from basic drug discovery to clinical trial design.

In the past two decades, substantial knowledge has accumulated about genetic events contributing to differences in drug responses, some resulting in drug-labeling changes and clinical practices (1, 2). U.S. drug labels for 6-mercaptopurine, warfarin, and irinotecan contain dosage adjustments based on TPMT, CYP2C9 and VKORC1, and UGT1A1 genotypes (3–5). As the U.S. Food and Drug Administration (FDA) requires more pharmacogenomic information to be included on drug labels, pharmacogenomics will be increasingly accepted and integrated into mainstream clinical practice. We are already observing a steady shift from the “trial and error” approach to a more knowledge-guided personalized approach toward drug therapy. To realize the full potential of pharmacogenomics, many formidable challenges still need to be overcome. These challenges include the ethical, economical, and legal issues associated with the ever expanding field of genetic testing, as well as the strong need to increase genetic literacy among patients and health care providers. Easier access to well-characterized clinical outcome and biometric data on patients under treatment are also crucial to identify and track genotype-phenotype relationships. With the combined efforts of researchers and health care providers to use the knowledge of pharmacogenomics in drug treatment and diagnosis, the barriers between scientific discovery and the clinical application of pharmacogenomics will diminish over time to realize the full benefits of the field.

Historically, submissions to the Food and Drug Administration (FDA) have been based on safety and efficacy data obtained from clinical trials conducted in adults with limited or no data from children. As a result, pediatricians and other health care professionals have relied on empiric therapeutic strategies, largely the consequence of treatment on a trial-and-error basis. In essence, the absence of information in the product label forces pediatricians to choose between avoiding the use of a medication that may be beneficial and using a medication “off-label” in the absence of evidence-based safety and efficacy data with the accompanying potential for ineffective and harmful outcomes.

During the past fifteen years, new federal laws and regulations have increased the level of scientific and clinical rigor of investigations aimed at ensuring that the use of medications by children is, indeed, safe and effective. Interested readers are referred to a detailed chronology of events occurring between 1994 and 2002 (1), and a contemporary discussion of the issues surrounding more recent legislative activities, such as the Pediatric Research Equity Act (PREA) of January 2003 (2). In general, it is recognized that growth and development are accompanied by changes in the physiological and biochemical processes determining drug disposition and response, for example, drug absorption, distribution, metabolism, excretion, and targets of drug response (3). However, the acquisition of information that can inform safe and efficacious use of medications in children of different ages or developmental stages has been a relatively recent development, increasing dramatically in recent years as a consequence of the various legislative initiatives. It is now apparent that extrapolation of adult data to pediatric populations is quite inappropriate because drug clearance may be greater than or, in some cases, less than that observed in adults (4). Thus, even though weight-based dosing strategies are becoming more sophisticated and have improved abilities to use adult data to infer drug clearance in children (5), they are unlikely to provide consistent dosing guidelines across all pediatric age groups or chemical classes. This is largely due to the variability in the developmental patterns of expression of the various drug-metabolizing enzymes and transporters involved in the disposition of individual compounds. Furthermore, evidence that the response to some medications may be different in children relative to adults despite comparable drug exposure is beginning to accrue (e.g., buspirone [4]), implying that age-related differences in drug targets and downstream signal transduction pathways may also be present.

Introduction

β-Hydroxy-β-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors (statins) are highly efficacious in the prevention of coronary artery disease. Although statins are generally considered safe, their use may be associated with musculoskeletal complaints that limit tolerance to treatment and, in the most extreme case, can lead to rhabdomyolysis. Both candidate gene and genome-wide association studies are being used to assess possible genetic susceptibility to statin-induced myopathy, with the recognition that this phenotype represents a broad spectrum of syndromes influenced by other drugs and disease states. In addition to potentially guiding statin therapy, the results of such studies may provide mechanistic insight into the critical cellular events linking statin use to muscle pathology in patients at risk.

Clinical Background

Multiple large clinical trials have demonstrated that statins (HMG-CoA reductase inhibitors) reduce the incidence of both primary and secondary coronary artery disease in patients at risk (1–4). Primary prevention trials have demonstrated that statin use can reduce the risk of a first major coronary event by more than 30 percent (3, 5). Secondary prevention trials reveal a risk reduction of similar magnitude (2). Aggressive intervention trials suggest that greater lipid lowering is associated with further reduction in risk (6).

Health care is moving toward a more individualized approach that has been termed “personalized medicine.” The underlying causes for this transition are many; they include the ability to genotype and sequence DNA, the increasing emphasis on consumerism among patients, and changes in the pharmaceutical industry worldwide and particularly at the U.S. Food and Drug Administration (FDA) and its sister regulatory agencies around the world. Pharmacogenetics and pharmacogenomics both involve the study of how genetics exerts an impact on drug response phenotype. For our purposes, the term “pharmacogenetics” connotes single genes that dominate the effects on a drug response, whereas “pharmacogenomics” connotes systems of many genes that create complex drug response phenotypes. It is clear that pharmacogenetics and pharmacogenomics are the core elements of the future of personalized medicine.

The emergence of robust and effective patient advocacy groups over the past thirty years has led to organized demands by patients for medicines that are more effective and that have fewer side effects. This was fueled by the Institute of Medicine “To Err is Human” report of 1999, which estimated that more than 50,000 Americans die each year because of medical errors, in particular, involving prescription drugs. Health care organizations have registered the clinical and financial dangers inherent in medication errors, and more precise prescribing is now a central part of quality control and even part of the marketing campaigns for hospitals in the United States. The pharmaceutical industry is experiencing the death of the “blockbuster” model of drug development in which one dose of a single medication can be used to treat everyone, including men and women, people of all races, infants, adolescents, adults, and the elderly.

Health Impact

Diabetes mellitus has become a major public health epidemic, affecting more than 250 million individuals worldwide in 2008, increasing to 380 million in 2025. Each year 3.8 million deaths are attributable to diabetes. An even greater number die of cardiovascular disease made worse by diabetes-related lipid disorders and hypertension (1).

Type 2 diabetes is much more common than type 1 diabetes, and accounts for approximately 90 percent of all diabetes worldwide (2). Type 1 diabetes is characterized by a lack of production of insulin in the body, whereas type 2 diabetes is due to the body's diminished insulin secretion from pancreatic β-cells and the resistance of tissues to insulin (3).

Adverse drug reactions occur during drug development and in clinical practice with approved medicines. They are responsible for the termination of approximately 20 percent of investigational drugs in the pharmaceutical pipeline. Approximately 1 percent of marketed drugs are withdrawn or restricted postmarketing because of safety-related issues. Adverse drug reactions affect millions of people every year, are responsible for a significant fraction of hospitalizations, and are a leading cause of death in developed countries. Thus, patients, the medical community, health care providers, regulatory agencies, and the pharmaceutical industry have a compelling interest to understand these adverse reactions and identify factors that influence them.

In this chapter we define adverse drug reactions and several related and commonly used terms; evaluate their impact on drug development, public health, and individual patient well-being; provide an overview of the contribution of known genetic variants to adverse drug reaction risk; and discuss efforts to identify genetic adverse drug reaction risk factors and incorporate them into development and clinical practice.

Basic Concepts in HIV Therapeutics

Considerable progress has been made during the past twenty years in the development of effective treatment against HIV-1, the agent of AIDS. Today, there are more than twenty agents commercialized that target various steps in the viral life cycle (Figure 21.1). These drugs, used in combination, inhibit two or more steps of the viral cycle: viral entry (fusion inhibitors, coreceptor antagonists), reverse transcription (nucleoside analog and nonnucleoside inhibitors), integration into the host genome (anti-integrases), and viral polyprotein processing (protease inhibitors). The need for combination antiretroviral therapy (ART) is determined by the ease of viral escape from the selective pressure when confronted with single antiretroviral agents. The ability to escape from drug pressure is, in turn, the result of rapid mutation rates due to a viral polymerase (the reverse transcriptase) that lacks proofreading activity.

HIV treatment is highly standardized, with defined criteria for its initiation and detailed knowledge on comparative efficacy (1). Decisions to initiate ART are primarily based on the degree of immunosuppression (generally defined by the numbers of CD4+ T cells – the target of HIV-1 – in blood), with additional consideration given to the levels of viral replication (defined by the level of viremia – viral genome copy numbers per milliliter), and the presence of clinical symptoms. In the absence of treatment, most individuals infected with HIV-1 will have progression of immunosuppression, leading to AIDS-related opportunistic infections and cancer. Individuals progress to AIDS over a mean of eight years after infection; however, rates of progression are highly variable among individuals.

Informed consent is an ethical and legal requirement for research on human subjects and the treatment of patients, as well. The process of informed consent provides an opportunity for the disclosure of material information and a structure for shared decision making that respects the autonomy of subjects and patients to decide whether to participate in research and to plan the course of their treatment. This chapter will describe the key elements of informed consent in research and clinical care in the context of pharmacogenetics and pharmacogenomics.

It is important to distinguish research from clinical care. Research is defined as “a systematic investigation, including research development, testing and evaluation, designed to develop or contribute to generalizable knowledge” (45 CFR § 46.102(d)). By contrast, clinical care or “practice” has been defined as “interventions that are designed solely to enhance the well-being of an individual patient or client and that have a reasonable expectation of success” (Belmont Report [1], 3). Although the definitions have been in place for several decades, new technologies and methodologies increasingly are blurring the distinctions. For example, translational research, case studies, outcomes research, quality improvement, and public health applications often present definitional challenges.