a particular type of DNA sequence at a specific location in the genome. Such a sequence can be a wild-type or variant sequence.
DNA or deoxyribonucleic acid
the double-stranded molecule found in the nucleus of cells that carries the genetic blueprint of an organism.
the biological process of making an RNA copy from the DNA sequence of the gene. The RNA produced is called the messenger RNA (mRNA) which provides information for making the protein coded by the gene.
a basic unit of heredity made of DNA and found on chromosomes in the nucleus.
broadly, refers to the entire genetic makeup of an individual. More specifically, it refers to the pair of alleles found at a particular genomic location.
the presence of two different alleles at a particular genomic location.
the presence of identical alleles (either wild-type or variant) at a particular genomic location.
proteins that facilitate, enhance, or accelerate chemical reactions in the body. These enzymes convert the drug molecules to either inactive or active products in the body.
the study of how the entire genome affects the response to a drug therapy.
the physical or functional manifestation of the genotype in an organism.
a gene variant that occurs at a frequency greater than 1% in a population.
RNA or ribonucleic Acid
a single-stranded molecule with multiple functions, including allowing genes that make up the DNA to be expressed in the form of proteins.
Single nucleotide polymorphism or SNP
a sequence variation at a single position in a DNA sequence.
a specific region of the genetic material that differs from the most common DNA sequence and that may or may not lead to altered function.
phenotype or allele that is in the form most commonly found in the population, and that is assumed to result in the “typical” function for an organism.
After completing this chapter, you should be able to
Define pharmacogenomics and key terms related to this subject.
Outline basic forms of genetic variability that affect medication action.
Identify common gene variants in drug-metabolizing enzymes and drug transporters.
Describe how genetic variability influences enzyme activities and metabolizer status in individual patients.
Discuss how prodrugs are affected differently than standard bioactive drugs by the influence of genetic variants in metabolizing enzymes.
Describe the pharmacogenetic interaction between at least three different medication-gene pairs and the appropriate dose or therapy adjustments needed in such interactions.
Efficacy and safety are the defining characteristics of any drug approved for human administration. While the drug development process entails the evaluation of drug candidates in a sizable patient population, there can still be patient-specific or subpopulation-specific pharmacological effects that may be observed only after the drug has been deployed in clinical practice. A complex interplay of factors influences the therapeutic window within which the drug produces its intended pharmacological effect. However, a one-size-fits-all drug therapy frequently fails. Drug therapy is often modified based on several patient-specific factors (Table 3-1). Each person’s genetic makeup is a unique biological blueprint, and it can profoundly influence not only susceptibility to specific diseases, but also response to drug therapies. The modification of drug therapies after inefficacy or adverse events are evident has always been possible. However, this trial-and-error approach can come at the expense of delay in appropriate treatment and, at times, serious health consequences. Pharmacogenomics provides preemptive ways to avoid these scenarios and increases the likelihood of successful drug therapy for a given condition and its treatment.
Common Factors Influencing Modification or Tailoring of Drug Therapy
Presence of other diseases or conditions
Hepatic or renal function
Other co-administered drugs
Disease severity or progression
Cost of therapy
Pharmacogenomics and the related term “pharmacogenetics” refer to the study of how genes in an individual or in a population affect the response to drug therapy. Another term that has commonly been used is “precision medicine” which refers to the use of a systematic approach to disease treatment and prevention that takes into account a patient’s genes, environment, and lifestyle. Pharmacogenomics is a component of the broader precision medicine approach. Several factors have helped to increase the attention on pharmacogenomics as a tool to improve the clinical outcomes of drug therapy. Despite an increase in targeted drug therapies, the health and economic burden of ineffective drug treatments remain high. In addition, drug therapies are often associated with frequent and predictable adverse effects that are responsible for significant mortality, especially in the hospital setting. The cost of prescription drugs has increased rapidly over the past few decades which has justifiably increased patient expectations of safe, proven, and effective drug treatments. The advent of large-scale gene-sequencing technologies and data analysis tools have also fueled the rise of pharmacogenomics in mainstream medicine. This chapter provides a basic explanation of the science of pharmacogenomics and its application in clinical practice, using discrete examples.
THE FUNDAMENTALS OF GENETICS
DNA is organized into chromosomes and packaged into the nucleus of cells, encoding the genetic information that serves as the “blueprint” of life. Adult human cells have 23 pairs of chromosomes, including a pair of sex chromosomes. One chromosome in each pair is inherited from the father and one from the mother. Fertilization of the egg with the sperm followed by the pairing of the chromosomes from the mother and father results in the regeneration of the two sets of chromosomes. DNA is made of four nucleotide bases—two purine bases (adenine [A] and guanine [G]) and two pyrimidine bases (cytosine [C] and thymine [T]). In the RNA molecule, a nucleotide called uracil (U) replaces T. The nucleotide bases in DNA are attached to a sugar (deoxyribose)-phosphate backbone. Two such strands of polymeric chains coil together to form the classical double helical form of DNA. In the DNA double helix, the nucleotide bases on the two strands are joined to each other following the base-pairing rules (A pairs with T, and G pairs with C; A pairs with U in RNA). The organization of nucleotides in a specific sequence forms the genetic message that form genes, which encode proteins responsible for carrying out all the bodily functions—from structural proteins of the body to enzymes, transporters, receptors, etc. The basic structure of DNA is illustrated in Figure 3-1.
Genes are the parts of the human genome that carry the information that affects an individual’s characteristics and/or phenotype. Humans have two copies of each gene, one inherited from each parent. Overall, the human genome is believed to code for approximately 20,000 to 25,000 proteins, a number far lower than was once believed. Using the genetic information carried in the DNA to form proteins entails the molecular processes of transcription and translation.
During transcription, the genetic information in the DNA is copied to generate a single-stranded messenger RNA (mRNA) molecule by the enzyme RNA polymerase. This mRNA molecule generated in the nucleus then leaves the nucleus to enter the cytoplasm, where it is used in the process of protein synthesis or translation. The genetic information in the mRNA molecule forms a script that is read three letters (A, T, G, or C) at a time (known as codons) in a process involving organelles called ribosomes and another specialized RNA molecule called the transfer RNA or tRNA. Each of the 64 codons (the number of possible combinations of four bases in a three-letter format) codes for a specific amino acid. The mRNA sequence information is “read” in a specific order of three-letter codons known as the open reading frame, defined by a start codon (typically AUG, which codes for the amino acid methionine) and that ends with a stop codon (UGA, UAA, or UAG). The tRNA transfers the appropriate amino acid coded by a specific codon to the ribosome for assembly and thereby the mRNA sequence is translated into the polypeptide sequence of amino acids to form proteins. There is redundancy in the genetic code, with multiple possible codons for a specific amino acid. For example, the amino acid arginine (Arg) is coded by codons CGU, CGC, CGA, CGG, AGA, and AGG. Overall, this process of transcription and translation is called the “central dogma of molecular biology,” which is simply defined as DNA makes RNA, and RNA makes protein.
There is a vast similarity in the DNA sequence among the 6 billion bases that make up the human genome. In fact, 99.5% of the DNA sequence is believed to be identical across the human race. At the scale of billions of bases, the 0.5% difference still accounts for millions of bases and is sufficient for meaningful differences in the genetic predisposition to diseases and variable drug response. Differences occur at the rate of one change in every 1,000 bases. Various types of genetic variations have been identified (Figure 3-2). Among the most important types of sequence variations or mutations are the single nucleotide polymorphisms or SNPs (pronounced as “snips”). SNPs refer to variations in a sequence at a single position in a DNA sequence. For example, rather than a stretch of sequence such as AATGCTAG, an individual may have AAAGCTAG. The T in the third position, in this case, is changed to an A. A population frequency of >1% is used as a threshold to define variations at a single nucleotide position as a polymorphism. When population frequency is <1%, the variations are classified as mutations. Approximately 12 million SNPs have been identified to date. Importantly, a vast number of these SNPs are “silent”—that is, they do not alter or affect the function or expression of genes. However, in certain cases, these SNPs can have a profound effect on the expression or activity of the specific protein coded by the gene harboring the SNP.
The SNPs are of different types, broadly classified as:
Synonymous—these SNPs do not result in a change in the encoded amino acid. For example, an AÆG substitution in AGA to AGG still results in the Arg amino acid in the polypeptide chain.
Nonsynonymous: These SNPs alter the amino acid encoded by the codon. For example, a CÆA substitution in CCG to CAG results in a change in the amino acid encoded from proline to glutamine. This type of SNP is also sometimes called a missense substitution. Another type of nonsynonymous SNP is a nonsense substitution wherein a single nucleotide change in the DNA sequence results in the generation of a stop codon, which causes the termination of the transcription process.
SNPs can also be in noncoding parts of the DNA, such as in the promoter regions of the genes. The promoter region of a gene is the DNA sequence that flanks the gene and to which critical proteins called transcription factors and RNA polymerase bind, ultimately resulting in the expression of the gene. So, while variants in these regions do not alter the sequence of the encoded protein, they have the potential to affect the expression levels of the proteins by affecting the normal transcriptional rate of a gene. It is important to note that based on the presence or absence of variant(s) at a particular site, a gene can be present in one of the two or more alternate forms. Each of these forms of a gene is called an allele. An allele can be a wild-type allele, which refers to the most common version of the gene sequence in a population, or it can be a variant allele. A copy of a gene inherited from a parent can be in wild-type or variant form. Each pair of alleles represents the genotype of a specific gene. Genotypes are described as homozygous if there are two identical alleles at a particular genetic location and as heterozygous if the two alleles differ. The genotype that an individual has for a specific gene (wild type/wild type; wild type/variant; variant/variant) eventually determines the phenotype that the patient exhibits for that gene function.
Because of the existence of millions of SNPs, a naming system is in place to ensure that individual SNPs are accurately referred to in the scientific literature and clinical practice. A few different nomenclature conventions have been used for designating SNPs. The two most common ones are—a star-allele-based convention and a RefSNP-based naming convention (Table 3-2).
Commonly Used Nomenclature System for Single Nucleotide Polymorphisms (SNPs)
Uses the standard gene symbol followed by a star symbol and an arbitrary number. The wild-type allele is generally designated as *1. Variants are assigned other designations (*2, *3, *4,. . . , etc.).
CYP2C9*2 is a SNP in the CYP2C9 gene that results in an arginine to cysteine substitution at the 144th position of the CYP2C9 protein
Uses accession numbers in a specific database of SNPs called the dbSNP. The dbSNP accession numbers are written as the prefix “rs” followed by a number.
rs1799853 refers to the CYP2C9*2 SNP
While SNPs are the most common types of genetic variations, another type of genetic variation are the copy number variations (CNVs). The CNVs in genes result in an individual carrying multiple copies of a gene on a chromosome. A significant example of CNV is observed in the CYP2D6 gene, with multiple functional copies of the CYP2D6 gene reported in certain individuals. The typical annotation of gene duplication or multiplication is, for example, *1xN or *2xN where xN is the number of copies. These multiple copies can be of a normal function allele, an increased function allele, or a decreased function allele.
Effects of Genetic Variations
As mentioned earlier, the genotype for a specific gene influences the observed phenotype. Broadly, the variants produce either a loss-of-function or a gain-of-function effect on the gene activity. The loss-of-function effect is more common. An individual may have inherited one or both copies of a gene in its variant form. Based on the possible combinations of inheritance patterns for the wild-type and variant allele, significant differences in gene function can be observed. In the case of a loss-of-function variant the following phenotypical effects would be observed for the corresponding genotypes:
Wild type/Wild type (homozygous for wild type): normal function
Wild type/Variant (heterozygous): reduction in function
Variant/Variant (homozygous for variant): significant reduction in function
Thus, while normal gene function is expected when both the inherited alleles are wild-type, a significant reduction in function is seen when both the inherited alleles are loss-of-function variants. Analogously, carrying both alleles which are gain-of-function variants is expected to produce a significant increase in gene function compared to the homozygous wild-type inheritance. It is also important to acknowledge the relative prevalence of a gene variant in a given population or allele frequency. There can be significant differences in the prevalence of a variant based on a specific genetic ancestry. So, while some variants may be more common in the Caucasian population, they may be missing or less prevalent in populations of African or Asian ancestry and vice versa. Population-wide genetic screening has yielded important observational data on the prevalence patterns of specific gene variants, and in some cases, have shaped population-specific medication therapy guidelines and recommendations. However, ethnicity only predicts the likelihood of carrying specific genetic variants. Individual patient sequencing is still necessary to confirm the presence or absence of the variants of interest.
The effect of having a genetic variant is also largely dependent on the gene that carries that variant and the natural function of the gene. While the variants can occur in any gene, for the purpose of pharmacogenomics, the following discussion will only consider the genes that have a role in regulating a medication’s drug-body interactions. Broadly, the drug-body interactions can be grouped according to either the pharmacokinetic or the pharmacodynamic properties of the medication. Recall from Chapter 2 that pharmacokinetics describes the effect that the body has on a medication. For the medication to produce its effects on the human body it must navigate through several membrane barriers (absorption or A), effectively use the circulatory system to travel through the body (distribution or D), and survive enzymatic transformation (metabolism or M), and the processes that are designed to remove it from the body (elimination or E). These ADME characteristics of a given medication comprise its pharmacokinetic profile and determine the levels of the active medication found at the target action site.
Once the medication is at the active site (for example, the heart cells or myocytes for an anti-arrhythmic medication), the medication has a chance to exert its therapeutic effect. These effects that the medication produces on the body are classified as its pharmacodynamic effects. The pharmacodynamics of a medication characterize the biochemical and physiological effects it exerts via its ability to bind and modify the function of various target proteins in the body, including receptors, enzymes, ion channels, or transporters. In the context of pharmacogenomics, the vast majority of currently known and clinically important variants are found in the genes encoding for proteins that determine the pharmacokinetics of medications.
To understand the implications of gene variants for pharmacokinetics, it is important to review which genes and their corresponding proteins govern the concentrations of drug molecules in the body. The concentration of the drug in the body determines not only the therapeutic efficacy of the drug but also the adverse effect potential of the treatment. Therefore, any increase or decrease in the levels of the drug in the body beyond the therapeutic range has the potential to compromise the efficacy or exacerbate the adverse effects of the medication. A major role in regulating these levels is ascribed to the metabolizing enzymes, which are proteins that promote or accelerate chemical reactions in the body. Human beings have hundreds of drug-metabolizing enzymes. The chemical reactions catalyzed by the drug-metabolizing enzymes can be classified into either phase I metabolism or phase II metabolism.
Phase I Metabolism
The phase I reaction is an enzymatic reaction that often occurs in the liver wherein polar modifications are introduced to the molecules of the drug by the processes of oxidation, reduction, or hydrolysis. Phase I reactions facilitate the conversion of lipophilic (dissolving in lipids or fats) drug molecules to hydrophilic (dissolving in water). This group of reactions is catalyzed predominantly by the cytochrome P450 superfamily of mixed-function oxidases (CYPs).
The sequencing of the human genome has identified 57 CYP genes. Together, they account for the metabolism of >75% of prescription pharmaceuticals. The majority of CYP genes are polymorphic (exist in variant forms) and these polymorphisms are found at significant levels in the population. These polymorphisms result in either a decrease or increase in enzymatic activity leading to an increase or decrease in the levels of the drugs that are metabolized by these enzymes. Among the numerous CYP enzymes, the most important polymorphisms are seen in CYP2C9, CYP2C19, and CYP2D6. Collectively, the genes that code for these enzymes are responsible for the way biotransformation of 60% to 70% of all medications occurs. The following discussion will cover the relevant genetic variants found in each of these phase I metabolizing enzymes (Figure 3-3).
This is one of the most abundant enzymes in the human liver and is responsible for the metabolism of 15% to 20% of prescribed and over-the-counter drugs (Table 3-3). The most clinically important variants of CYP2C9 are *2 and *3, both of which have a significantly lower frequency in the African and Asian populations than the Caucasian population.
Common Drug-Metabolizing Enzymes and the Medications They Metabolize
CYP2C9*2: A variant seen in 10% to 20% for Caucasian populations and considerably less in African (0 to 6%) and Asian (1% to 3%) populations. This variant results in a 50% reduction in enzymatic activity compared to *1 (wild type).
CYP2C9*3: This variant is seen at a frequency of 4% to 10% in Caucasians, 1% in Africans, and 3% in Asians. This variant results in an 80% reduction in enzymatic activity compared to *1.
This enzyme is another major polymorphic CYP450 enzyme. There are three clinically important variants in the CY2C19 gene: two loss-of-function variants (*2 and *3) and one gain-of-function variant (*17).
CYP2C19*2: The allele frequencies of the *2 variant are 12% in Caucasians, 15% in Africans, and 29% to 35% in Asians.
CYP2C19*3: This variant is most common in Asians (2% to 9%), with <1% allelic frequency in Caucasians and Africans.
CYP2C19*17: This variant results in increased transcription and expression of CYP2C19. It has a frequency of 21% in Caucasians, 16% in Africans, and 3% in Asians.
This major member of the P450 family has been shown to be involved in the metabolism of 25% of prescribed drugs. Several variants of CYP2D6 have been identified, including some that exist in multiple copies ranging from 2 to 13.
Fully functional CYP2D6 variants: *1 and *2
Partially functional CYP2D6 variants: *10, *17, and *41
Nonfunctional CYP2D6 variants: *3, *4, *5, and *6
Among the nonfunctional CYP2D6 variants, the most common expression is seen for *4 in the Caucasian population (˜20%); *3 is more common in the Caucasian population than others, although found at low frequencies (˜1%); *5 is more common in African and Asian populations (5% to 7%) than Caucasian populations (<1%); and *6 is more common in African (3%) than other populations.
Phase II Metabolism
Among the phase II metabolizing enzymes, important pharmacogenetic associations have been shown for thiopurine S-methyltransferase (TPMT) and uridine diphosphate (UDP) glucuronosyltransferases 1A1 (UGT1A1).
This enzyme causes S-methylation of thiopurine drugs (6-mercaptopurine [6-MP] and azathioprine), a modification that leads to their inactivation. Because 6-MP and azathioprine are narrow therapeutic index drugs used in the treatment of cancers and autoimmune and inflammatory disorders, the variants in the TPMT gene can have a profound impact on their pharmacological effects. The most clinically relevant TPMT variants are *2, *3A, *3B, and *3C.
TPMT*2: This is a rare variant that results in a 100-fold reduction in TPMT activity.
TPMT*3A: This is a combination of two SNPs and results in a significant decrease in TMPT activity. This is the most common variant in the Caucasian population.
TPMT*3B: This is a rare variant commonly inherited together with a SNP in exon 10 (*3C) leading to the generation of the *3A variant.
TPMT*3C: This variant is most common in populations of African or East Asian ancestry.
UGT enzymes attach a glucuronic acid moiety to the drugs and endogenous substances (like bilirubin) that enhances their solubility in water, thereby allowing them to be excreted in bile or urine. The variants of UGT1A1 known to be clinically relevant include *28 and *6.
UGT1A1*28: This variant has significant expression in Caucasians, with approximately 10% being homozygous and about 40% heterozygous. Also, it is present at 42% to 56% population frequency in African and 9% to 16% in Asian populations.
UGT1A1*6: This variant is common in populations with East Asian ancestry (˜20%).
This background in drug-metabolizing enzymes is useful to understand the functional consequences of the presence of genetic variants in these enzymes. Individuals can be classified into the following standard phenotypes of metabolizer status for each metabolizing enzyme gene.
Importantly, the metabolizer phenotypes are determined by genotype for a particular enzyme that an individual may have. As mentioned earlier, this is dependent on the different combinations of allele pairs (wild type and variant) that an individual has for a specific gene. An increase in metabolic activity or UM phenotype is usually a result of gain-of-function variants. In contrast, the decrease in metabolic activity of an enzyme or PM is usually a result of loss-of-function variants. The amplitude of gain or loss of function is dependent on whether an individual is homozygous or heterozygous for the gain- or loss-of-function variants. Individuals homozygous for gain-of-function variants are usually UM and those homozygous for loss-of-function are usually PM. The exact genotype-phenotype relationship may vary for different genes (Table 3-4).
The Assignment of Metabolizer Phenotype for the CYP Enzymes Based on Genotype
xN denotes the number of copies of the CYP2D6 gene.
An important concept to understand regarding drug metabolism is the effect it has on the biological activity of the drug molecules. Typically, the drugs are bioactive molecules capable of producing the desired therapeutic effect in their native unmodified forms. The modifications of the chemical structure of the parent drug molecule in the process of metabolism often result in a reduction in the activity of the drug. However, this is not always true. For certain drugs, the process of undergoing metabolism in the body results in the activation of the drug molecule. These drugs are called prodrugs—they are administered in their inactive precursor forms and get converted to active moieties by virtue of the chemical modification that occurs in the body. Therefore, the effect of pharmacogenetic variants in the drug-metabolizing enzyme is dependent on the type of drug molecule being examined (Figure 3-4).
For a standard drug (eg, warfarin) with a bioactive parent moiety, a loss-of-function variant results in decreased conversion to the inactive form and an increase in the levels of the active drug molecule that can exceed the intended therapeutic range. This situation can result in increased toxicity of the drug directly related to increased levels of the drug in the circulation. On the other hand, a gain-of-function variant for an enzyme metabolizing the bioactive parent moiety will result in increased metabolism of the drug to inactive forms and as a result, there is a risk of loss of therapeutic efficacy.
For a prodrug (eg, clopidogrel), wherein the drug metabolite is the active form, a loss-of-function variant results in decreased conversion to the active form and as a result, a loss of efficacy. On the other hand, a gain-of-function variant for an enzyme metabolizing the prodrug will result in increased metabolism of the drug to active forms and as a result, an increased risk of causing a toxic drug response.
Both scenarios are observed in clinical practice. Therefore, a knowledge of the basic pharmacological characteristics of the drug is critical to understanding the potential pharmacogenetic interactions.
It is important to consider the information about genetic variants in drug-metabolizing enzymes in the context of the specific role that the enzyme plays in the metabolism of a given drug. Often, drug molecules may have multiple metabolic pathways, each contributing to the breakdown of a specific fraction of the drug. If the variant is found in an enzyme that does not contribute significantly to the breakdown of the drug, then having genetic variants in that enzyme will not have a measurable clinical impact on the pharmacological effect of the drug. A variation of this rule is when the minor metabolic pathway may be more important for the pharmacological effect of the drug. For example, the antiplatelet drug clopidogrel is broken down by enzymes called esterases (accounting for 85% of clopidogrel metabolism) and by CYP2C19 (accounting for 15% of clopidogrel metabolism). In this case, although CYP2C19 is responsible for a much smaller fraction of clopidogrel metabolism, this metabolic pathway converts the prodrug clopidogrel to its active moiety responsible for antiplatelet action; esterases convert clopidogrel to inactive metabolites. Therefore, in this case, loss-of-function variants in CYP2C19 are clinically important although the contribution of CYP2C19 to the metabolic fate of clopidogrel is relatively minor.
While metabolizing enzymes are a major focus of pharmacogenetic interactions, variants in other types of proteins such as transporters and the proteins of the immune system are also clinically relevant.
Transporters are proteins that allow the uptake of drugs and other molecules into cells. Genetic variants that affect the functional activity of transporters can prevent a drug from accumulating in the target tissue and lead to greater off-target adverse effects of drugs. Among transporters, organic anion transporting polypeptide 1B1 (OATP1B1), the protein product of a gene called SLCO1B1, is especially important in the context of pharmacogenetics. OATP1B1 is found in the liver cells and is responsible for the transport of widely used drugs such as statins, angiotensin-converting enzyme (ACE) inhibitors, and angiotensin II (AT II) receptor blockers into the liver cells (Figure 3-5). The two known variants of SLCO1B1 that are associated with reduced OATP1B1 transport function are *5 and *15.
The immune response to drug therapy can have significant implications. Drug-induced hypersensitivity reactions or drug allergies account for 10% of all adverse drug reactions. For a given drug, these reactions would occur only in a small fraction of patients. Predicting which patients would show a hypersensitivity reaction is not easy but certain genetic markers can be useful in this determination. The human leukocyte antigen B (HLA-B) protein is a component of the immune system that has been implicated in drug-induced hypersensitivity. Normally, HLA-B is expressed on the surface of cells and presents cellular peptides to T cells, a major type of immune cell. Certain HLA-B alleles are associated with susceptibility to severe skin reactions (Stevens-Johnson syndrome [SJS] or toxic epidermal necrolysis [TEN]) to drugs such as abacavir or carbamazepine. These include variants *57:01 and *15:02. The *15:02 variant is reported at a higher allelic frequency in those with Han Chinese ancestry.
EXAMPLES OF PHARMACOGENOMICS IN ACTION
The application of complex genetic information to guide therapeutic decisions in clinical practice remains a challenge. A consortium of pharmacogenetic experts called the Clinical Pharmacogenetics Implementation Consortium (CPIC), founded in 2009, has led the efforts to formulate guidelines and recommendations for the use of genetic information in clinical practice guidelines for drug therapy. So far, guidelines for 37 drug/drug class-gene pairs have been published. On the regulatory side, the Food and Drug Administration (FDA) also includes pharmacogenomic information in the various sections of the drug label. Currently, 457 FDA-approved therapeutic products contain pharmacogenomic information in different sections of the drug label, although not all of them are associated with a specific action required based on genetic information. While pharmacogenetic interaction evidence exists for several drug-gene pairs, the actual number of such pairs where that information is “actionable” and requires therapy modifications is much smaller. Even in this group of actionable pharmacogenetic associations, fewer drug-gene pairs require or mandate pharmacogenetic testing; the majority provide recommendations with the assumption that the sequence of the specific interacting gene is already known for a patient, which is very often not the case. These recommendations are based on the strength of the pharmacogenetic evidence reported in the literature. Naturally, pharmacogenomics application is still in its early stages. However, with an increased focus on gene sequencing from healthcare systems and the end consumers, an enhanced utilization of pharmacogenetic evidence in therapeutic decisions is expected to occur.
The power of pharmacogenomics to affect medication therapy is best illustrated by reviewing the drug-gene pair associations. The scope of this chapter does not allow a description of all known drug-gene pharmacogenetic associations. However, the following examples will serve to provide a good overview of some well-established drug-gene pharmacogenetic pairs (Table 3-5).
Classical Pharmacogenetic Interaction Examples
Effect on Pharmacology
Clopidogrel is a prodrug that is converted to its active form by CYP2C19
Carriers of CYP2C19 loss-of-function variants are susceptible to increased cardiovascular events due to inadequate antiplatelet effect of clopidogrel
Using alternate antiplatelet agents (like ticagrelor or prasugrel) in CYP2C19 IM and PM carriers; standard clopidogrel dosing suggested in CYP2C19 EM and UM patients
*1 and *2 (fully functional) *10, *17, and *41 (partially functional), *3, *4, *5, and *6 (nonfunctional). Multiple copies of functional genes (hyper-functional)
Codeine is a prodrug that is bioactivated to morphine by CYP2D6
CYP2D6 PM would have inadequate pain relief from codeine due to a decrease in its conversion to morphine; CYP2D6 UM are more susceptible to morphine’s analgesic effect and serious adverse effects of drowsiness and respiratory depression
Avoid codeine in both CYP2D6 UM and PM
*5 and *15 (loss-of-function)
OATP1B1 (the protein product of SLOC1B1) is responsible for the transport of statin drugs into liver cells where statins produce their pharmacological effect
Loss of function of SLCO1B1 causes an increase in plasma levels of simvastatin; carriers of SLCO1B1 loss-of-function variants are more likely to experience the muscle toxicity of simvastatin
Using low-dose simvastatin or alternate statin agent (rosuvastatin or pravastatin) in SLCO1B1 loss-of-function carriers
HLA-B proteins are responsible for the presentation of endogenous proteins and certain drug molecules on the surface of cells for recognition by the immune system
HLA-B*5701 is able to bind to abacavir and presents it on the cell surface; using abacavir in HLA*57:01 carriers results in an immune-mediated hypersensitivity reaction particularly in the form of Stevens-Johnson syndrome
HLA-B*57:01 screening in all patients starting abacavir for the first time before the initiation of therapy is recommended; in patients lacking the HLA-B*57:01, abacavir can be used at standard doses; in patients carrying HLA-B*57:01, abacavir is not recommended
Warfarin is metabolized by CYP2C9 to inactive metabolites
Loss of function of CYP2C9 causes an increase in sensitivity to warfarin as its metabolism to inactive metabolites is decreased
Patients with CYP2C9 *2 and *3 or VKORC1 1639G>A SNP require less warfarin dose to achieve target INR. Patients with both CYP2C9 *2 and *3 and VKORC1 1639G>A SNP have the lowest warfarin dose requirements; using validated algorithms for warfarin dosing in the non-African carriers of VKORC1 –1639G>A and CYP2C9*2 and *3 variants is recommended
–1639G>A (decrease in expression)
VKORC1 is the molecular target of the anticoagulant action of warfarin and is inhibited by warfarin
Patients carrying the VKORC1 1639G>A SNP have a greater sensitivity to warfarin and thereby lower dose requirement
CYP = cytochrome P450; EM = extensive metabolizer; HLA = human leukocyte antigen; INR = international normalized ratio; OATP1B1 = organic anion transporting polypeptides 1B1; PM = poor metabolizers; SLCO1B1 = solute carrier organic anion transporter family member 1B1; SNP = single nucleotide polymorphism; UM = ultra-rapid metabolizers; VKORC1 = vitamin K epoxide reductase 1.
Clopidogrel is a widely prescribed antiplatelet drug that functions as an irreversible inhibitor of a receptor found on the surface of platelets. Clopidogrel binding to this receptor prevents adenosine diphosphate (ADP) from binding to it and activating the platelets. The key feature of clopidogrel pharmacology is that it is administered as an inactive prodrug that must be metabolized to its active form by the action of the enzyme CYP2C19. This process is affected by genetic polymorphisms in CYP2C19, as indicated in an earlier section. Variants that decrease the CYP2C19 activity decrease the formation of the active metabolite of clopidogrel and thereby decrease its antiplatelet effect. Increased variability in clopidogrel response is observed in such individuals and therapy modifications may be required.
Carriers of CYP2C19 loss-of-function alleles are at risk of more cardiovascular events (myocardial infarction [MI], stroke, stent thrombosis, etc.) due to inadequate antiplatelet effect with clopidogrel usage. Based on their genotypes, they can be classified as EM (*1/*1), UM (*1/*17, *17/*17), IM (*1/*2, *1/*3, *2/*17), and/or PM (*2/*2, *3/*3, *2/*3). Current guidelines suggest using standard clopidogrel dosing in UM and EM but considering an alternative antiplatelet agent such as ticagrelor or prasugrel for IM and PM patients. These recommendations are most beneficial and applicable for patients with acute coronary syndrome (ACS) and those undergoing percutaneous coronary intervention (PCI). Patients with stable ischemic heart disease are less likely to derive benefit from these recommendations.
Codeine is an analgesic that belongs to the opioid class of drugs. It is indicated for mild to moderately severe pain and also finds use as an antitussive medication. Codeine is a prodrug that is metabolized by CYP2D6 leading to its bioactivation to morphine in the liver. The analgesic effect of codeine is almost entirely dependent on the generation of morphine in this reaction. As indicated earlier, CYP2D6 is a polymorphic phase I enzyme, with known fully functional (*1 and *2), partially functional (*10, *17, and *41), and nonfunctional (*3, *4, *5, and *6) variants. Based on the inherited alleles, the possible phenotypes of CYP2D6 activity include UM, EM, IM, and PM.
CYP2D6 UM carriers possess multiple functional copies of the CYP2D6 gene, with reports of 2 to 13 copies present in some individuals. In patients with PM CYP2D6 phenotype, codeine is not activated to morphine. Such patients would experience inadequate pain relief from codeine and fewer uncomfortable gastrointestinal side effects although the sedation, nausea, and dry mouth are reported to be comparable between PM and EM patients. On the opposite end of the activity spectrum, the UM phenotype will result in an increased generation of morphine, making patients more susceptible to codeine’s analgesic effect as well as its serious adverse effects of drowsiness and respiratory depression, even at standard doses. The CPIC guidelines suggest avoiding codeine in both CYP2D6 UM and PM patients. The FDA label for codeine, while describing the pharmacogenetic influence on codeine pharmacology, does not recommend a specific course of action.
Statins are prescribed to reduce serum cholesterol, specifically the low-density lipoprotein cholesterol (LDL-C). Due to their LDL-C-lowering property, statins have a significant effect of decreasing cardiovascular disease risk. The mechanism of action of statins involves inhibition of HMG-CoA reductase in the liver. HMG-CoA is the rate-limiting enzyme that catalyzes an early step in the synthesis of cholesterol. In response to the reduced free cholesterol content within the liver cells, the expression of the LDL receptor gene is increased. An increase in the number of LDL receptors on the surface of liver cells causes increased removal of LDL from the blood, thereby lowering LDL-C levels. While statins enjoy widespread usage, statin-associated muscle symptoms (SAMS) can limit their use, at least in certain groups of patients. The incidence of SAMS is reported to range between 7% and 29%. These muscle toxicities often lead to noncompliance with statin therapy, resulting in an increased risk of cardiovascular disease and mortality.
The LDL-C-lowering pharmacological effect of statins requires their uptake and concentration in the liver cells. The OATP1B1 transporter facilitates the liver uptake of statins in addition to several other medications (Figure 3.5). The OAT1B1 protein is encoded by the SLCO1B1 gene. The SCLO1B1*5 variant is associated with a decrease in the transport function of OATP1B1 attributed to a decrease in the expression of the transporter on the surface of liver cells. The heterozygous and homozygous carriers of the *5 variant have higher plasma concentrations of statins. This increased systemic exposure of statins is associated with an increase in the risk of development of myopathy. Among the statin drug class, this pharmacogenetic interaction is primarily ascribed to simvastatin. The risk of myopathy is higher with the 80 mg dose of simvastatin than for those on lower doses. Using a low dose of simvastatin or considering an alternative statin (such as pravastatin or rosuvastatin) is recommended by the CPIC guidelines in patients with intermediate or low function of OAT1B1. The availability of low-cost alternate statins as well as generally low incidence of adverse effects with statins has made the use of pharmacogenetic testing before statin initiation unnecessary. However, pharmacogenetic testing can help identify underlying reasons for the observation of muscle toxicity with simvastatin use and can help to guide statin switching in such patients.
Abacavir is a medication used in the treatment of acquired immunodeficiency syndrome (AIDS) caused by the human immunodeficiency virus (HIV). The use of abacavir is associated with skin hypersensitivity reactions, particularly SJS. A genetic mechanism for these reactions has been elucidated with the identification of the HLA-B gene variants. HLA-B belongs to the major histocompatibility complex (MHC) family of proteins that are expressed on the surface of almost all cells and participate in antigen presentation to the immune system. The protein breakdown products are attached to the MHC molecules for presentation on the cell surface. Physiologically, the presentation of “self” proteins or the body’s own proteins does not activate the immune system, whereas infection by pathogens results in the presentation of “non-self” proteins and elicitation of an immune response. Certain drug molecules also get processed in a manner similar to proteins and are presented on cell surfaces complexed with MHC molecules.
In the case of abacavir, such complexation with HLA protein occurs in patients with certain variants of HLA-B. In particular, the HLA-B*57:01 allele binds abacavir and presents the drug along with other peptide fragments on the surface of cells. This presentation results in the recognition of peptide fragments on the cell surface as non-self and causes the development of a hypersensitivity reaction due to immune system activation. This pharmacogenetic mechanism does not affect drug pharmacokinetics or pharmacodynamics; rather it affects only the risk of hypersensitivity reactions. The presence of even just one HLA-B*57:01 allele is associated with an increased risk of hypersensitivity reactions and there are no intermediate phenotypes. The evidence supporting the role of HLA-B*57:01 in abacavir-induced hypersensitivity reaction is quite strong; CPIC recommendations call for HLA-B*57:01 screening in all patients starting abacavir for the first time before the initiation of therapy. In patients lacking the HLA-B*57:01, abacavir can be used at standard doses. In patients carrying HLA-B*57:01, abacavir is not recommended.
Another HLA-B variant with known pharmacogenetic interaction is HLA-B*15:02 which is implicated in carbamazepine-associated SJS and TEN.
Warfarin is an oral anticoagulant that has been in clinical usage for more than six decades and is frequently called by its former brand name Coumadin. The initiation of warfarin therapy for its anticoagulation effect is characterized by frequent dose adjustments needed to achieve target international normalized ratio (INR) values in each patient. If the dose of warfarin is too high, it increases the risk of bleeding, whereas if the dose is too low, there is an increased risk of thromboembolism. The dose of warfarin required to achieve the therapeutic INR can be highly variable between patients; up to 10- to 20-fold differences (0.5 to 7 mg) in the warfarin dose required have been observed. Genetic factors have been shown to account for a significant portion of warfarin’s dose variability.
Variants in three genes have been shown to influence warfarin response: CYP2C9, vitamin K epoxide reductase 1 (VKORC1), and CYP4F2. The role of these genes in warfarin response is evident in examining warfarin’s mechanism of action (Figure 3-6). The S enantiomer of warfarin (the active stereoisomer) is metabolized by the enzyme CYP2C9 to inactive metabolites. Warfarin’s molecular target protein is VKORC1, a protein that warfarin inhibits. VKORC1 is responsible for converting vitamin K (in its epoxide form) to reduced vitamin K, which functions as a cofactor in the activation of clotting factors. The reduced vitamin K is converted back to epoxide vitamin K by the action of the enzyme gamma-glutamyl carboxylase (GGCX). These conversion reactions between epoxide and reduced vitamin K catalyzed by VKORC1 and GGCX are termed the vitamin K cycle. The enzyme CYP4F2 removes reduced vitamin K from the vitamin K cycle by converting it to hydroxy-vitamin K1. Overall, warfarin inhibition of VKORC1 prevents the generation of reduced vitamin K, which leads to the inhibition of clotting factor activation. CYP2C9 prevents warfarin activity by converting it to inactive metabolites. The action of CYP4F2 is consistent with the warfarin effect as it limits the amount of reduced vitamin K available for clotting factor activation.
The CYP2C9 genetic variants have been indicated earlier. Especially important are the loss-of-function alleles—*2 and *3. Patients with these alleles have a higher sensitivity to warfarin due to decreased metabolism of warfarin to inactive metabolites. Such patients need lower doses of warfarin to achieve the therapeutic INR. With regard to VKORC1, the most common variant is seen in the promoter region of the gene. The SNP at the –1639 position causes a decrease in the transcription of the VKORC1 gene and hence, a decrease in the expression of the VKORC1 protein. Since the target protein of the warfarin effect is reduced, patients carrying the VKORC1 1639G>A SNP have a greater sensitivity to warfarin and thereby lower dose requirement. The allelic frequency of CYP2C9 and VKORC1 across different races shows important differences: CYP2C9 *2 and *3 are seen at higher levels in the white population (13% and 7%, respectively) as compared to Asian (0 and 4%) and Black populations (3% and 2%); VKORC1 –1639G>A is seen at a frequency of 40% in Whites, 91% in Asians, and 11% in Black populations. In patients with African ancestry, CYP2C9*5, *6, *8, and *11 are important in the determination of warfarin dose. A SNP in the CYP4F2 gene (rs2108622; designated as *3) has been shown to be associated with a higher warfarin dose requirement.
The recommendations from CPIC guidelines include using published and validated algorithms for warfarin dosing in the non-African carriers of VKORC1 –1639G>A and CYP2C9*2 and *3 variants. The carriers of either VKORC1 –1639G>A or CYP2C9*2 and *3 have lower dose requirements, whereas those with the loss-of-function CYP2C9 variants as well as VKORC1 –1639G>A have the lowest warfarin dose requirements. In contrast, wild-type CYP2C9 and VKORC1 carriers require higher warfarin doses. The incorporation of CYP4F2 genetic information has also been shown to improve the accuracy of warfarin dose prediction. The warfarindosing.org website contains the two major dosing algorithms developed for the determination of warfarin dose: the Gage algorithm and the International Warfarin Pharmacogenetics Consortium (IWPC) algorithm. These algorithms integrate genetic as well as nongenetic factors in the determination of warfarin dose. Warfarin dose determination shows that the pharmacogenetic factors, although important, are not always sufficient to explain the variability in drug response. Nongenetic factors such as the patient’s age, race, weight, height, and clinical factors such as the history of smoking, presence of liver disease, and use of other co-administered drugs (statins, amiodarone, antifungals, antibiotics, etc.) can have a significant effect on the warfarin dose required in each patient. This integrated approach that factors pharmacogenomic evidence into therapeutic decision making along with patient and clinical characteristics represents the spirit of precision medicine.
The pharmacogenetics of warfarin is unique in two different ways. First, it is one of the few examples showing how pharmacogenetic variants in pharmacodynamic proteins can influence drug response. The majority of the well-studied pharmacogenetic interactions are based on variants in pharmacokinetic proteins such as metabolizing enzymes and transporters. VKORC1 variants show that there likely are several more clinically meaningful variants in the pharmacodynamic proteins that have not been discovered yet. Second, the pharmacogenetics of warfarin illustrates the polygenic nature of warfarin response—ie, genetic variants in multiple genes influence response to warfarin. Frequently, how individual genes, in isolation, influence drug response has been the focus of attention. However, the drug response in the body is a composite of multiple proteins made by multiple genes, some affecting the active levels of the drugs and others affecting the pharmacological effect produced. These proteins/genes can be of the same functional class (e.g., metabolizing enzymes) or divergent protein classes (metabolizing enzymes vs. receptors). In a specific patient, the observed genetic influence on drug response is a net effect of all genes with which that drug interacts in the body to produce its pharmacological activity.
As illustrated by the above examples, pharmacogenomics-guided drug therapy has the potential to improve clinical outcomes of both efficacy and adverse effects. However, the application of this guidance is often dependent on the availability of a patient’s genetic sequence information. The genetic sequencing tests are available in different formats: single-gene tests, gene panels, and whole-genome sequencing. The cost of getting the genetic sequencing and the turnaround time for getting the genetic information to the prescribers often impede the full adoption of pharmacogenetic guidance. However, if this information were already available, then many of the pharmacogenomic recommendations could be readily applied in the selection of the right drug at the right dose. Such preemptive genetic testing would enable the application of pharmacogenomic-guided drug therapies over the lifetime of a patient. Unlike other diagnostic tests (blood sugar, lipid panels, etc.), genetic tests for each specific gene need to be performed only once and can provide immense value throughout a patient’s lifespan. The proliferation of such tests and services will expand the patient population who have their genetic information available to use in therapeutic decision making.
Efforts to lay the foundation for the application of pharmacogenomics in clinical practice have long been underway. The field of pharmacogenomics finally has enough momentum to gain mainstream prominence in medicine. The tailwinds helping the field are the converging forces of big data/data science, artificial intelligence, advances in sequencing technology, and an increase in consumer expectations of healthcare. All of these are, of course, propelled by an accumulating body of scientific evidence supporting drug-gene interactions, as well as the formation of a consortium of experts systematically evaluating this vast body of evidence and proposing appropriate recommendations and guidelines. At its very core, pharmacogenomics is a tool to maximize the benefit of drug therapy and minimize the associated risk. While there are still significant implementation barriers to overcome the promise of pharmacogenomics to unlock the full potential of targeted drug treatments is getting closer to fulfillment.
Pharmacogenomics knowledgebase or PharmGKB: This is a pharmacogenomics database that curates scientific and clinical evidence supporting drug-gene pharmacogenetic associations along with links to clinical guidelines, drug labels, individual drug and gene information, and visualization of drug pathways and mechanisms of pharmacogenetic interactions. Website: https://www.pharmgkb.org/ (accessed August 5, 2022).
CPIC or Clinical Pharmacogenetics Implementation Consortium: This is an international consortium of volunteers and experts who participate in the creation, curation, and dissemination of peer-reviewed, evidence-based detailed drug/gene clinical practice guidelines. These guidelines enable the translation of genetic laboratory test results into actionable prescribing decisions for affected drugs. Website: https://cpicpgx.org/ (accessed August 5, 2022).