Chapter 12. Pharmacogenetics

The genetic basis underlying variation in drug response among individuals has become evident with the introduction of modern analytical methods for the analysis of gene sequence and expression. The goal of pharmacogenetics is to stratify drug therapy into groups of individual patients based on their genetic makeup. Additional factors such as the environment, diet, age, lifestyle, and state of health can influence a person's response to medicine. An understanding of an individual's genetic makeup is thought to be the key to drug selection, drug design, and dosage regimen development. Greater efficacy and safety in drug therapy is based on stratification of patients into groups based on their relevant phenotypes (Phillips et al, 2001; Mancinelli et al, 2000). Pharmacogenetics is the study of the genetic basis of interindividual patient variability in the response to drug therapy. Pharmacogenetics allows for individualization of drug therapy. In contrast, pharmacokinetics provides a means for estimating kinetic parameters of the drug absorption, distribution, metabolism, and excretion in various population subgroups and then applying the information to drug therapy for the average patient.

Pharmacogenomics is closely related to pharmacogenetics and is considered to be an equivalent or overlapping field. Pharmacogenomics involves study of the role of genes and their genetic variations (DNA, RNA level) in the molecular basis of disease and the resulting pharmacologic impact of drugs on that disease. Pharmacogenomics is sometimes defined to include drug design aimed at variants of a pharmacologic target. Pharmacogenetics and pharmacogenomics are both important disciplines involved in the study of genes that code for drug-metabolizing enzymes (see Chapter 11), drug receptors (see Chapter 19), drug transporters, and ion channels or efflux systems (see Chapter 13). Many of the above are relatively new factors involved in determining how genetic variation contributes to diversity in the response to drugs, including the ultimate fate of the drug and its ability to exert a therapeutic response without undue side effects.

Application of pharmacogenetics to pharmacokinetics and pharmacodynamics helps in development of models that predict an individual's risk to an adverse drug event and therapeutic response. With some drugs, pharmacogenetics allows the recognition of subgroups or strata based on differences in genetic makeup in drug receptor sequences and are therefore characterized by predictable changes in pharmacodynamic response to drugs. Understanding the genetic and molecular differences in disease etiology, individual pharmacokinetics, and drug mechanism produce insight on how a patient will respond to a given drug. For example, the monoclonal antibody Herceptin was designed to treat a subset of breast cancer patients who overexpress the HER-2 (human epidermal growth factor receptor-2) gene. Patients who lack HER-2 overexpression are considered to be nonresponders to Herceptin therapy. In the past, such differences would be apparent only after a trial-and-error period. This genetic knowledge improves our ability to select or design the proper drug for individuals suffering from a disease with a varying range of molecular defects.

Pharmacogenetics provides justification for individualized dosing for many drugs known to be highly variable in their effects. The outcome of disease, resistance to treatment, and adverse reactions are increasingly recognized as an interaction of the individual's genes and the environment. A review of the role of genes in human susceptibility to infection predicted that recent advances in genetics and high-throughput genotyping technology will enable screening of the whole genome for genetic factors that determine susceptibility to HIV and AIDS, malaria, and tuberculosis (Kwiatkowski, 2000). Kwiatkowski (2000) listed many genes and their encoded proteins that play roles in immunity and disease fighting. Examples include genes with alleles that affect susceptibility to hepatitis B, HIV, and other known infections. Susceptibility to the AIDS virus includes proteins that bind the receptor (such as HLA, human leukocyte antigen) and immune response amplification such as chemokine and cytokine response (Kaur and Mehra, 2009). Besides host factors, progress has been even faster for genes of microbes that play roles in the efflux of drugs out of the organism, a principal factor for antibiotic resistance in many pathogens. The study of efflux drug pumps in microbes and other organisms have been reviewed by Franke et al (2010).

Pharmacogenetics (PGt) or pharmacogenomics (PGx), a term preferred by some researchers, is of interest to the pharmaceutical industry and regulatory agencies. These groups tried to generate a consensus on how and to what extent should PGx or PGt information be applied to improve drug therapy and safety of both old and new drugs (Lesko and Woodcock, 2002). The FDA has since provided guidance on PGx (FDA Guidance, 2005) to help sponsors determine whether submission of PGx data will likely be required or voluntary for investigational, biological, or new drug applications including the type of PGt data needed in the product label. Drugs such as irinotecan, mercaptopurine, and warfarin contain pharmacogenetic information on their labels.

PGt's and PGx's central role in medicine has been enabled by high-throughput technology that allows rapid screening of tens of thousands of genes rapidly and simultaneously. For example, the DNA chip is a microchip that uses hybridization technology to concurrently detect the presence of tens of thousands of sequences in a small sample. The probes (of known sequence) are spotted onto discreet locations on the chip, so cDNA (complementary DNA) hybridization from the patient's sample to a probe residing in a defined location indicates the presence of a specific sequence. Other rapid and low-cost sequencing technologies such as ULCS (ultra-low-cost sequencing) or cyclic array technologies will also permit rapid and high-volume sequencing and/or sequencing of individual genomes. These technologies usually rely on some combination of miniaturization, multiplex or parallel assays, analyte amplification and/or concentration, and detection signal amplification.

This chapter will focus on variations in pharmacokinetics due to PGx. However, variation in drug response is also in large part due to nongenetic factors, as listed in Table 12-1. Variations in drug response due to polymorphisms (genetic variations) in the drug's receptor or downstream processes can also be identified using PGx principles and screening, with the number of examples growing rapidly each year (Table 12-2). The challenge in PGx is to distinguish clinically significant haplotypes (common patterns of polymorphic variation) from the thousands of co-existing polymorphisms that have been identified in either disease etiology or treatment. Polymorphisms in a gene may have several possible outcomes including no change in phenotype (as in a silent mutation), a nonclinically significant change in phenotype, a clinically significant change in phenotype that may or may not represent a possible drug target or potential biomarker for diagnosis, prognosis, or other types of treatment. For example, Brouwers et al (2006) examined over 34,000 genes in pheochromocytoma, and found that 2246 genes had differential gene expression between noradrenergic and adrenergic tumors, and 636 genes were differentially expressed when comparing malignant and benign tumors. An important follow-up from this study is to determine which of these polymorphisms are most relevant for diagnosis, prognosis, and treatment and how to use this information most effectively for treatment design.

Scientists compared the genetic signatures of drugs in patients who are or are not responsive to drug treatments (such as hormone therapy for prostate cancer, or imatinib for acute lymphocytic leukemia) to find which genetic determinants enable a positive (or even a negative) drug response. Chemicals with genetic signatures that are similar to effective drugs also suggest that the compound may have potential therapeutic activity. Such discoveries may someday improve the efficiency of drug selection and discovery.

A well-studied example of genetic polymorphism is the variable metabolic acetylation of drugs among patients resulting in different drug responses. Patients' ability to metabolize certain drugs such as hydralazine, procainamide, and isoniazid are categorized as either fast acetylators, normal acetylators, or slow acetylators. Acetylation status is dependent on the patient's genetic composition, which determines the activity of the acetylation enzyme N-acetyltransferase. Acetylation status determines whether a patient is dosed with a correspondingly higher or lower dose compared to “normal acetylators.”

Although acetylation status in humans has been well described, other metabolic or pharmacologic variations have been poorly understood until recently. Genetic variations are well known in bacteria and other microorganisms because rapid changes in these organisms are easily observed. In humans, mutations and related changes occur to different degrees in thousands of proteins and other macromolecules.

Polymorphisms, genetic variations with a frequency of greater than 1% of the population, or mutations, a change in genetic sequences in less than 1% of the population, can affect patient therapeutic response or metabolism of a given drug (Meyer, 2000). Genetic polymorphism may also refer to the occurrence in a population of more than one allele at the same locus with the least frequent allele occurring more frequently than can be accounted for by mutation alone. Less precisely, genetic polymorphism is described as differences in DNA sequence among individuals, groups, or populations that gives rise to different phenotypic forms such as the human blood groups or other body characteristics. Genetic polymorphism can occur with drug metabolism enzymes, drug transporters, and even drug receptors that govern therapeutic responses of drugs. Therefore, it is rational to individualize dose regimens based on pharmacokinetic data obtained from different polymorphic groups. Many alleles encoding different drug receptors are being discovered and studied with increasing frequency. Pharmacokinetic parameters influenced by genetic differences include drug bioavailability, distribution, metabolism, and tissue binding. Environmental factors (eg, smoking, drug–drug interactions, nutritional status) and pathophysiology of the patient (eg, cardiovascular disease, age, hepatic and renal disease) can also influence the pharmacokinetics of a drug. Our understanding of the impact of these genetic differences on clinical pharmacokinetics and pharmacodynamics is in its infancy.

Polymorphism in cytochrome isozymes has been demonstrated in drug metabolism along with their corresponding allelegenes. The clinical significance of certain polymorphic forms with regard to drug metabolism has been reported for many drugs. Genetic tests are available to screen polymorphisms for cytochrome P-450 drug-metabolizing enzymes in an individual. Prior knowledge of an individual's metabolic capability can reduce the risk of adverse drug reactions by adjustment of the dose regimens according to an individual patient's metabolic capability. The types of genetic mutations affecting metabolic enzymes are illustrated in Fig. 12-1.

Genetic polymorphism within a specific genotype may occur with different frequencies depending on racial or population factors, which evolved from selective geographic, regional, and ethnic factors. The probability of carrying a specific allele varies among different subjects depending on whether the dominating factor is geographic or ethnic. In practice, genetic polymorphisms with higher frequencies are more important because they are likely to affect more people. However, some rare mutations are important because they cause extreme medical consequences or may be fatal for the individual.

Using genetic polymorphism considerations, drugs may be developed for groups of patients who are genetically comparable to improve efficacy and reduce risk of an adverse event. Drugs may also be developed only for patients who express a particular genetic profile to provide a desired level of safety or efficacy. In addition, doses for patients can be based on their metabolic capacity by determining whether they are “poor metabolizers” or “ultrarapid metabolizers.” Molecular studies in pharmacogenetics began with cloning of CYP2D6 and now have been extended to more than two dozen drug-metabolizing enzymes and several drug transport systems (Meyer, 2000). The most important isozymes with genetic polymorphism involved in drug metabolism are shown in Fig. 12-2.

Recognition of the genotypes associated with drug disposition and metabolism and the ability to obtain a “specific pharmacogenetic profile” for the patient will allow individualization of drug therapy and reduction of drug interactions (Evans and Relling, 1999; Roses, 2000, 2001). PGx research aims to elucidate these polygenic (multiple-gene) determinants of drug effects. The interplay of genetic polymorphism with interindividual differences in pharmacokinetics and pharmacodynamics has been well reviewed (Evans and Johnson, 2001; McLeod and Evans, 2001). The ultimate goal is to provide new strategies for optimizing drug efficacy and toxicity based on groups of patient's genetic determinants.

Pharmacogenomics, PGx, developed rapidly as a result of advances in molecular genetics and genomics employing high-throughput tests such as micro-array technology. The previously held notion of the monogenic nature of disease (one gene causing one disorder) is yielding to the concept of polygenic disorders, by which dozens or even thousands of genes may be differentially expressed compared to normal, healthy tissue. As such new information arises, new challenges and opportunities also emerge for novel drug development.

The coordinated goal to sequence the 30,000 or more human genes via the Human Genome Project has also fueled progress in PGx. Many new genes or gene products have emerged from the Human Genome Project as new potential drug targets. These new targets may be receptors, membrane proteins, enzymes, or ion channels that may be directly or indirectly involved in disease pathogenesis. While these newly discovered receptors or target enzymes can be exploited as new drug targets, polymorphic variations of those genes must be considered when developing new drugs that target these proteins.

Unique genetic sequences have been identified for thousands of proteins and endogenous substances that support normal cellular functions. The genetic information is coded in the two helical strands of DNA. DNA consists of four basic nitrogenous substances or bases (C, cytosine; A, adenine; T, thymine; and G, guanine), which combine with deoxyribose and phosphate to form the respective nucleotides. The four nucleotides are combined in unique sequences for each gene. Genes are coded in a special region or locus in the DNA.

A change or mutation in gene sequence may or may not result in a change in phenotype such as chronically reduced or increased level or activity of a protein or an essential enzyme. In some cases, such changes result in an exaggerated or reduced therapeutic response to a drug. The cell is homozygous if the genetic sequences occupying the locus are the same on the maternal and paternal chromosome; if they are different, the cell is heterozygous. When more than one alternative forms of a gene exist, the alternative genes are referred to as alleles of the gene. The identity of the alleles carried by an individual at a given gene locus is referred to as the genotype. Alleles that vary by a single nucleotide change can now be characterized rapidly at the DNA level by single-nucleotide polymorphism (SNP) screening. SNPs occur in about one of every 100 to 1500 base pairs (bp) between two unrelated individuals. Any two individuals may differ by 0.1% of their more than 3 billion base pairs. Common or informative SNPs are those that occur at frequencies of greater than 1% (Meyer, 2000). The physical effect observed as a result of genotype difference is referred to as phenotype. Haplotypes, in contrast, are groups of SNPs that vary together to influence a phenotype. The International HapMap Project is an international collaboration to provide a respository of haplotype information (www.hapmap.org). Haplotypes are identified by examining gene expression patterns for commonality. For example, high-throughput data may be converted to visual information (eg, red for higher expression, green for lower expression) and organized by patient, expression level, phenotype and/or gene in a two- or three-dimensional grid in attempt to detect patterns in gene expression. The most relevant genes involved in disease etiology or classification have been determined for tumors such as AML (acute myeloid leukemia) and menangioma.

Genes are considered functionally polymorphic when allelic variants exist stably in the population and their gene products exhibit altered activity in relation to the wild type (“normal”). Ideally, variation in drug response can be predicted by monitoring differences in phenotype or genotype for a single patient or a group of patients. Often, a surrogate chemical measure for a physiologic condition or response called a biomarker may be used for this purpose. Biomarkers may also be used to identify disease susceptibility and diagnose or monitor disease or response to treatment.

To determine whether a patient is a rapid or slow metabolizer, the patient is given a known substrate for that enzyme and the patient's intrinsic clearance (see Chapter 11) is measured. Traditionally, intersubject variation in metabolism has been investigated by the use of known substrates followed by in vitro verification of the enzyme level. Alternatively, the metabolic-status genotype may be determined directly from subjects' DNA. The latter approach is more definitive and offers much insight during drug development into how genes affect the metabolism of drugs. The mechanism of metabolism has been elucidated for many drugs using both approaches. In practice, fragments of DNA samples are compared based on SNPs or haplotypes. If the patient's genotype and its functional activity are known, the probable individual drug response can be predicted.

In the future, rapid sequencing, use of biomarkers, and/or measuring associated phenotypes will play a major role in detecting unusual variations in heritable clinical phenotypes of drug response. Once a large number of these genetic variations and their frequencies in different populations are known, they can be used to correlate a patient's genetic “fingerprint” to the patient's probable individual drug response. This genetic or biomarker information can then be used to group patients into genetic groups or “strata” to personalize drug treatment for these patient subgroups. For example, patients with chronic myeloid leukemia who are BCR-ABL (a tyrosine kinase) positive are good candidates for Gleevec (imatinib mesylate), a tyrosine-kinase inhibitor.

SNP and haplotype variations in coding regions of genes (about 30,000–100,000 per genome) cause variations in amino acid sequence and protein function. Genetic variations in gene regulatory regions can cause differences in protein expression that may affect drug response. Genetic profiles of individuals may be analyzed to determine disease susceptibility as well as predisposition to PGt considerations in drug efficacy or toxicity.

Variations in drug pharmacokinetics and pharmacodynamics are due largely to genetic polymorphism in genes involved in drug metabolism, absorption, disposition, and disease pathogenesis. As early as the 1950s, researchers realized that some adverse drug reactions were caused by genetically determined variations in enzyme activity. Adverse drug reactions of debrisoquin have led to the discovery of the genetic polymorphism of this drug-metabolizing enzyme, debrisoquine hydroxylase CYP2D6. More recently, a review of the PGt literature showed that a sizable portion of ADRs (˜30%) involved in drug therapy implicated drug metabolism polymorphisms in CYP2D6 (Sadee, 2002). In CYP2D6-deficient patients, dosing of anti-arrhythmic drug sparteine causes more nausea, diplopia, and blurred vision adverse reactions. The presence of four phenotypic subpopulations for CYP2D6 ensures genetic variation in the population and thus in part accounts for interindividual differences in adverse drug response between patients: 3% to 10% of whites, 6% to 8% of African Americans, and less than 2% of Asians are poor 2D6 metabolizers. As a result, these groups are more prone to adverse effects from 2D6 substrates such as SSRIs (selective serotonin reutake inhibitors), tricyclic antidepressants, and metoprolol, and are more likely to be unresponsive to codeine, which requires 2D6 activation to its active metabolite. Other examples of adverse reactions caused by genetic polymorphisms include prolonged muscle relaxation in some subjects suffering from an inherited deficiency of a plasma cholinesterase after receiving a cholinergic drug. Hemolysis caused by antimalarial drugs is recognized as being caused by inherited variants of glucose 6-phosphate dehydrogenase. Slow acetylation status of isoniazid in some patients causes increased risk for peripheral neuropathy.

It is important to determine whether the variation in ADR is truly genetic or due to other factors. A method used to distinguish hereditary and environmental components of variability is the comparison of monozygotic and dizygotic twins, or pharmacokinetically by repeated drug administration and comparison of the variability of the responses within and between individuals. The National Institutes of Health (NIH), wants to identify genetic risk factors and biomarkers that predict drug response. NIH has organized the Pharmacogenomics Knowledge Base, a nationwide collaboration to study genes, drugs, and disease (www.pharmgkb.org). Examples of genetic polymorphism affecting drug response/side effect are listed in Table 12-3.

As discussed in Chapter 11, cytochrome P-450 enzymes (CYPs) are a family of oxidative enzymes involved in phase I metabolism of many drugs. The most important families for drug metabolism are CYP1, CYP2, and CYP3, with CYP3 arguably the most important for drug elimination. CYP-metabolized drugs often act as a substrate for more than one CYP enzyme, resulting in redundancy if a specific CYP enzyme is a slow metabolizer due to genetics (CYP3A varies approximately fivefold) and/or inhibitors (can increase variability to 400-fold).

Polymorphisms in CYP and other metabolic enzymes can affect the clearance of a drug and thus the overall pharmacokinetics for a given patient. Advanced knowledge of a patient's polymorph status could therefore be used to design optimum dosing regimens before an adverse reaction or subtherapeutic response occurs clinically. However, given the redundancy of drug metabolism and the contribution from environmental factors (such as diet, other drugs, age, weight, etc), the use of enzyme status data may be difficult to translate to a clinical decision. For example, warfarin metabolism is highly variable and results from a combination of factors including CYP2C9 (2%–10%), VKORC1 (10%–25%), and environmental factors (20%–25%). Algorithms for warfarin dosing are available which are particularly effective for patients requiring either less than 3 mg per day or more than 7 mg per day (Limdi and Veenstra, 2010). The significance of CYPs and transporters, and potential for interaction have great impact on drug efficacy and safety (Huang et al, 2008).

CYP2D6

CYP2D6 (see Chapter 11) is a large isozyme family that affects metabolism of many drugs. CYP2D6 is highly polymorphic. More than 70 variant alleles of the CYP2D6 locus have been reported. The metabolism of the tricyclic antidepressants amitriptyline, clomipramine, desipramine, imipramine, and nortriptyline, and the tetracyclic compounds maprotiline and mianserin is influenced by the CYP2D6 polymorphism to various degrees. Genetic polymorphism of CYP2D6 was first investigated with debrisoquine (Gonzalez et al, 1988). Poor metabolizers often carry two nonfunctional alleles of this gene, resulting in reduced drug clearance.

Since about 10% of the population are poor CYP2D6 metabolizers, drug candidates that are CYP2D6 substrates are often dropped from further development by researchers in favor of others (Sadee, 2002). Poor metabolizers have increased plasma concentrations of tricyclic antidepressants when given recommended doses of the drug. Adverse effects may occur more frequently in poor metabolizers and may be misinterpreted as symptoms of depression and may further lead to erroneous increases in the dose. When determining CYP2D6 metabolic status (slow vs fast metabolizers) in patients on tricyclic antidepressants, co-administration of other CYP2D6 substrates such as serotonin selective reuptake inhibitors may result in erroneously concluding poor CYP2D6 metabolic status.

In contrast, ultrarapid metabolizers have relatively fast drug metabolism due to the presence of more enzyme or increased enzyme activity. Patients in this group are prone to therapeutic failure due to the resulting subtherapeutic drug concentrations when “normal” doses are given. PGx studies have revealed that some fast metabolizers of CYP2D6 are the result of gene duplication present among different racial groups. Depending on the population studied, 5% to 20% of patients can be classified as either rapid or poor metabolizers.

Polymorphic drug metabolism is found in a large number of drugs used in psychiatric patients. Retrospective analysis of psychiatric patients treated with substrates of CYP2D6 strongly indicates that genotyping can improve the likelihood of preventing adverse drug reactions and decrease the costs of therapy.

The polymorphic O-demethylation of codeine is of clinical importance for drug activation when this drug is given as an analgesic. About 10% of codeine is O-demethylated by CYP2D6 to morphine, and this conversion is deficient in poor metabolizers. Poor metabolizers therefore experience no analgetic effects of codeine.

CYP1A2

Another isozyme, CYP1A2, which metabolizes 5% of randomly selected drugs, may also be considered during development, since up to 15% of a patient population can be classified as poor metabolizers, according to Sadee (2002). Fluvoxamine is a substrate and potent inhibitor of CYP1A2, causing important interactions with drugs such as amitriptyline, clomipramine, imipramine, clozapine, and theophylline that are partly metabolized by this cytochrome P-450 enzyme.

CYP2C9

Still another example of a clinically important drug metabolism polymorphism is the association of variant alleles of CYP2C9 with the requirement for lower warfarin dose. In a retrospective study of a population from an anticoagulant clinic, the CYP2C9 alleles associated with decreased enzyme activity (*2 and *3) are overrepresented in patients stabilized on low doses of warfarin. These patients had an increased incidence of major and minor hemorrhage.

CYP2C19

The 4′-hydroxylation of the (S)-enantiomer of mephenytoin is catalyzed by CYP2C19. The polymorphic enzyme has a poor metabolizer (PM) frequency of about 3% in Caucasians, 15% to 25% among Asians, and 4% to 7% among Black Africans (Dahl, 2002). The major defective allele responsible for the PM phenotype is CYP2C19*2, which is found in 13% and 32% of Caucasians and Asians, respectively. A second allele, CYP2C19*3, is found mostly among Asians and rarely in Caucasians.

Interestingly, few polymorphisms are reported for the isozyme subfamily CYP3A. This isozyme is involved in the metabolism of endogenous steroid and testosterone. Mutations of this vital enzyme may not be compatible with life.

Not all therapeutic variations and side effects result from genetic differences in the receptor or drug metabolism. Drug response (including therapeutic and unintended side effects) is influenced by many direct and indirect factors, including modifying effects from environmental factors on the disease process and drug disposition. As a result, some researchers are unsure whether prescribing drugs based on a PGt profile will significantly reduce side effects for most drugs, since many side effects and therapeutic failures may be the result of incorrect diagnosis or failure to account for other influencing variables such as the nature and severity of the disease, the individual's age and race, organ function, concomitant therapy, drug interactions, and concomitant illnesses.

Transporterpharmacogenetics is a rapidly developing field that is concerned with drug uptake and efflux into or through tissues. Significant problems in the clinical application of drugs result from poor or variable oral drug bioavailability and high intra- and interindividual variation in pharmacokinetics. Several membrane transporter proteins are involved in the absorption of drugs from the intestinal tract into the body, into nonintestinal tissues, or into specific target sites of action (Borst et al, 2000).

Drug efflux is an important cause of drug resistance in certain types of cells. In cytotoxic chemotherapy for several human cancerous diseases, drugs are generally very effective, but in the case of intrinsic or acquired multidrug resistance, usually highly effective antineoplastic compounds, eg, vincristine, vinblastine, daunorubicin, or doxorubicin, fail to produce response. One of the many causes of such multidrug resistance (MDR) is the appearance of special efflux proteins, eg, the P-glycoprotein can transport drug out in some cells and result in low drug levels in targeted cells. The commonly known P-gp is an example of many MDR-associated proteins. P-gp is genetically designated as ABCB1 or MDR1. Genetic variants of P-gp/ABCB1 are found in approximately 5% of Caucasians (Sharom et al, 1999) and usually associated with impaired function in vitro. Transporter polymorphism has important effects on in vivo drug pharmacokinetics. ABCB1 haplotypes have been proposed to predict the kinetics of drugs such as digoxin, cyclosporine, and fexofenadine. P-gp substrates and inhibitors are discussed in Chapter 13.

ABC Transporters

The multidrug resistance-associated proteins (MRPs) are members of the ATP-binding cassette (ABC) superfamily with six members currently, of which MRP1 (ABCC1), MRP2 (ABCC2), and MRP3 (ABCC3) are commonly known to affect drug disposition. MRP1 is ubiquitous in the body. Substrates for MRP1 include glutathione, glucuronide, adefovir, ciprofloxacin, paclitaxel, saquinavir, and sulfate; inhibitors include cyclosporine, glyburide, indomethacin, sulfinpyrazone, and verapamil. MRP1 is expressed basolaterally in the intestine, although its role in extruding drugs out of the enterocytes is still uncertain. There is some substrate overlap between MRP1 and apically located P-gp (see Chapter 13).

Solute Carrier Transporters

Another important class of drug carriers is the solute carriers (SLCs) such as the organic anion transporter protein (OATP) and organic cation transporter (OCT). SLCs are membrane transporters that may function as passive transporters, ion-coupled transporters, or exchangers with organic anions and cations, respectively, as their substrates. For example, OATP1B1 (SLCO1B1) has at least 17 SNPs identified and transports drugs such as statin drugs, rifampicin, enalapril, and methotrexate, but is inhibited by erythromycin, thyroxine, and rifampin. Changes in AUC and plasma concentrations have been reported for patients expressing OATP1B1 variants.

In the future, proteins involved in disease will increasingly become identified as important biomarkers for pharmacodynamic studies. Genomics has led to the development of proteomics, which involves the study of biologically interesting proteins and their variants. Proteins such as cell surface proteins (eg, COX-2, D-2R), intracellular proteins (eg, troponin I), and secreted proteins (eg, MCP-l) can be used as probes for drug discovery or as biomarkers for drug safety.

The physiologic response of the body to a drug is generally the result of interaction of the drug at a specific target site in the body. It is estimated that about 50% of drugs act on membrane receptors, about 30% act on enyzmes, and about 5% act on ion channels (Meyer, 2000). Many of the genes encoding these target proteins exhibit polymorphisms that may alter drug response. Clinically relevant examples of polymorphism leading to variable responses are listed in Table 12-4. For example, the β2-adrenergic receptor, and its common mutation of Arg→Gly at amino acid 16, greatly reduces the bronchodilator response of albuterol (Ligget, 2000). In addition, mutations in the angiotensin-converting enzyme (ACE) gene have been proposed to account for variations in the response to ACE inhibitors. Another study has shown that a combination of two mutations in the gene encoding a high-affinity sulphonylurea receptor leads to a 40% reduction in the insulin response to tolbutamide (Hansen et al, 1998). The response to clozapine in patients with schizophrenia appears to involve genetic polymorphisms in the 5-hydroxytryptamine (serotonin) receptor, HTR2A.

Mutations in five genes involved in the cardiac ion channels affect the risk of drug-induced long-QT syndrome (Priori et al, 1999), a potential cause of sudden cardiac death in young individuals without structural heart disease. The prevalence of long-QT syndrome is about 1 in 10,000. All five genes code for membrane ion channels affecting sodium or potassium transport and are influenced by antiarrhythmics and other drugs (Evans and Relling, 1999). Examples such as these are reported in the biomedical literature, yet the data are often conflicting, and the clinical implications of such findings are often unclear.

The systematic identification and functional analysis of human genes is changing the study of disease processes and drug development. PGx potentially enables clinicians to make reliable assessments of an individual's risk of acquiring a particular disease, increase specificity in drug targeting, and account for individual variation of therapeutic response and toxicity of drugs. Polymorphic alleles are the best-studied individual risk factors for adverse drug reactions, including the genes for N-acetyltransferase, thiopurine methyltransferase, dihydropyrimidine dehydrogenase, and the cytochrome P-450 isozymes. Genotyping can predict phenotype extremes in these situations and identify metabolic status in individual patients. Genomics is providing the information and technology needed to analyze the complex multifactorial situations involved in drug therapy (Fig. 12-3). Awareness of inherited variations of drug response can lead to dose adjustment on the basis of the patient's genetic makeup and provides a promising approach to reducing adverse drug reactions (Evans and Johnson, 2001; Roses, 2000). Dahl (2002), deLeon (2009), and others have reviewed the issues involving PGx and the prescribing of antipsychotic drugs.

The study of drug interactions and PGt/PGx has revealed that many unexpected pharmacodynamic responses and pharmacokinetic variations among individuals during drug therapy involve genetic factors. These genetic factors contribute to variation at many levels, including drug transport, metabolism, and interaction at the receptor site.

Applying genetics to study interindividual variations in drug response may greatly improve drug therapy. Understanding and monitoring the underlying PGx factors will allow physicians to better individualize patient therapy to optimize drug efficacy and minimize side effects. The labeling of many new drugs now contains more information on drug interactions and metabolism based on our understanding of the effects of PGt/PGx, and on pharmacokinetics and pharmacodynamics (PK and PD) of the drug. Population pharmacokinetics has utilized some of these changes by enriching some simple or empirical PK/PD models to simulate the quantitative aspect of drug distribution and drug action.

Models that predict dosing and an individual's drug clearance more accurately will require more details concerning the individual patient's genetic profile, demographic, and pathologic information, as well as the drug's pharmacokinetic profile.

If one of the patient's parents is homozygous for a specific isozyme that metabolizes the drug, how is that information linked to a clearance prediction for the patient? Recognizing the allele that predisposes the patient to a severe adverse reaction or toxic plasma concentration may allow the dose to be reduced or the drug to be avoided entirely. Similarly, if a patient is known to be a nonresponder or at high risk for adverse reactions due to genetic variation, that information can be used to select an alternate drug a priori.Table 12-5 lists FDA black box warnings for drugs that have been identified with a serious adverse reaction associated with a polymorphism.

In Chapter 20, a mixed-effect model is described that takes into account the effect of enzyme induction due to concomitant administration of a drug that induces enzyme. Should new models be developed, or will an extension of some of the mixed-effect models suffice? Advances in PGt will no doubt stimulate developments in pharmacokinetics and pharmacodynamics.

The successful application of genetic screening tests to identify patients with specific risks in drug response or drug toxicity depends on many factors. Large amounts of relevant genetic information might be monitored. Robust, high-throughput, high-positive and low-negative predictive tests must be developed and implemented. Such an endeavor will also involve considerable training, adaptation, and acceptance of the new technology by physicians and other healthcare personnel. With genetic diagnostic tests becoming more common and affordable, it is expected that individual drug dosing will become more accurate and ultimately result in vast improvements in therapeutic response and better drug tolerance. Researchers have high expectations that the use of diagnostic DNA microarrays or gene chips will simplify and expand testing and have clinical applications in diagnosis, disease prevention, drug selection, and dose calculation. The challenge to pharmacokinetics is to integrate all the relevant information into a model that is accurate and simple enough for practical application.

How useful are the average pharmacokinetic parameters obtained from a pharmacokinetic model in a drug study population that does not incorporate parameters accounting for genomic determinants?

• Average pharmacokinetic parameters are generally estimated by regression methods to minimize deviations from the data iteration within a group. These pharmacokinetic parameters are estimated for an average individual such that most subjects in the study population resemble the mean within a group. However, for a subject with the unusual allele, the underlying genetic factors contribute parameter differences that are the result of genetic expression. (Note: An attempt to force data fitting the mean would seriously error for a few individuals, ie, fit most subjects well, but will err badly for the few, on the basis of minimization of the sum of squared deviations.)
• Diazepam (Valium®) is highly metabolized in the body. Qin XP et al (1999) demonstrated that it is important to obtain individual PK parameters for patients using the drug diazepam. The presence of a single-nucleotide polymorphism (G681A) of the CYP2C19 gene cosegregates with the impaired metabolism of diazepam and desmethyldiazepam among Chinese subjects. The mean plasma elimination half-life value of diazepam is 84.0 ± 13.7 hours in homozygotes subjects. The corresponding half-life is much longer, 20.0 ± 10.8 hours for diazepam in wild-type homozygotes subjects. The longer half-life values in subjects of m1/m1 are also reflected by slow clearance of diazepam (2.8 ± 0.9 mL/min). However, diazapam disposition may be affected by protein binding besides CYP enzyme levels. This example shows the important contribution by pharmacogenetic polymorphism of CYPs.

Pharmacogenetics and pharmacogenomics involve the investigation and understanding of the role of genetics in drug therapy and therapeutic response, respectively. High-throughput technologies such as diagnostic microarrays have allowed the rapid sequencing of human genetic variability involved in drug transport, metabolism, and receptors, now exposing a complex interplay between these often highly variable genes. Clinically useful tools such as arrays and other sequencing technologies used to translate polymorphism data into effective recommendations are used to improve drug development and to individualize therapy for the patient.

Genetic polymorphism can occur with drug metabolic enzymes, drug transporters, and drug receptors that govern therapeutic responses of drugs. Genetic polymorphism that occurs with minor allele frequency of ≥1% in the population are very common (see Table 12-3). Phenotype is defined as the observable expression of a particular gene or genes. A deletion, insertion, or substitution can result in a loss or increase of a DNA sequence, which can lead to changes in copy-number or activity of the gene. Any of these genetic events can cause important changes in drug metabolism (eg, slow drug metabolizer or rapid-drug metabolizer phenotype), as shown in Fig. 12-1.

Allele: An alternative form of a gene at a given locus.

• Minor allele: A less common allele at a polymorphic locus.

Biological marker (biomarker): A characteristic that is objectively measured and evaluated as an indicator of normal biologic processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention.

Genetic polymorphism: Minor allele frequency of ≥1% in the population.

Genome: The complete DNA sequence of an organism.

Genotype: The alleles at a specific locus an individual carries.

Haplotype: A group of alleles from two or more loci on a chromosome, inherited as a unit.

Pharmacogenetic test: An assay intended to study interindividual variations in DNA sequence related to drug absorption and disposition (pharmacokinetics) or drug action (pharmacodynamics), including polymorphic variation in the genes that encode the functions of transporters, metabolizing enzymes, receptors, and other proteins.

Pharmacogenetics: A study of genetic causes of individual variations in drug response. In this review, the term “pharmacogenetics” is interchangeable with “pharmacogenomics.”

Pharmacogenomic test: An assay intended to study interindividual variations in whole-genome or candidate gene, single-nucleotide polymorphism (SNP) maps, haplotype markers, or alterations in gene expression or inactivation that may be correlated with pharmacological function and therapeutic response. In some cases, the pattern or profile of change is the relevant biomarker, rather than changes in individual markers.

Pharmacogenomics: Genome-wide analysis of the genetic determinants of drug efficacy and toxicity. Pharmacogenetics focuses on a single gene while pharmacogenomics studies multiple genes.

Phenotype: Observable expression of a particular gene or genes.

Promoter: A segment of DNA sequence that controls initiation of transcription of the gene and is usually located upstream of the gene.

1FDA Guidance for Industry: Pharmacogenomic Data Submissions, March 2005.

ABC transporters: ATP-binding cassette transporter

OATP: Organic anion transporter protein

OCT: Organic cation transporter

P-gp: P-glycoprotein, MDR1, ABCB1

PGt: Pharmacogenetics

PM: Poor metabolizer

SLC: Solute carrier transporter

SNP: A single-nucleotide polymorphism is a DNA sequence variation occurring when a single nucleotide—A, T, C, or G— in the genome (or other shared sequence) differs between members of a species or paired chromosomes in an individual.

The FDA has constructed a website called Genomics at FDA (www.fda.gov/cder/genomics) for individuals who would like to understand the agency's current thinking on pharmacogenetics. The site also provides education materials on pharmacogenetics.

The table of valid genomic biomarkers in the context of approved drug labels (www.fda.gov/cder/genomics/genomic_biomarkers_table.htm) summarizes all pharmacogenetic discoveries included on current drug labeling with a link to the approved package insert.