Pharmacogenetics

Published on 07/02/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 1478 times

Chapter 6 Pharmacogenetics

The concept of interindividual variations in response to drug therapy is not new. The fact that some individuals will respond well, some will not respond, and still others will experience toxicity from a given drug has been an accepted part of medicine since the advent of widespread use of pharmaceuticals to treat human disease.

There are many reasons for this variability among patients, including factors such as age, gender, and health status. However, as we enter the era of personalized medicine, what has changed is the willingness to simply accept a trial-and-error approach to prescribing, rather than attempting to predict which drugs will be successful in a given patient. With the advent of the information and data collection age, clinicians have ready access to patient information such as laboratory values, concomitant medications, and medical conditions that may influence response to drug therapy. However, what is becoming increasingly obvious as all other factors are accounted for is that likely one of the most important predictors of an individual’s response to drug therapy is, in fact, his or her genes.

The Promise of Personalized Medicine

Adverse events caused by drug therapy are a significant burden to the healthcare system and society in general. Many of these adverse events are preventable, a consequence of an individual’s genetic makeup as described in the next sections. Personalized medicine certainly has a significant role to play in the avoidance of harm; however, the real promise of pharmacogenetics may be in optimal prescribing. From a health economics perspective, the use of personalized medicine may be one of the most effective ways to manage the seemingly intractable problem of rising drug costs and access to life-saving pharmaceuticals.

Take the example of a new anticancer drug, considered to be a breakthrough therapy for lung cancer. As a collection of diseases rooted in genetics, cancer has become one of the clearest examples of the potential for pharmacogenomics in predicting pharmacodynamics of a given drug in a given patient, as outlined in the following example.

An anticancer drug may be considered successful if 10% to 20% of patients respond to it. From a clinician’s perspective and from the perspective of those 10% to 20% of patients, this is a success, but from a health economist’s perspective, a $30,000-per-year drug that fails in 80% to 90% of patients is not cost-effective. If the annual cost for a new anticancer drug is going to be around $30,000 to $60,000, the only way to make these promising new drugs cost-effective is to administer them to only the 10% to 20% of patients who will benefit from them.

Pharmacogenetics and Pharmacogenomics

Pharmacogenetics and pharmacogenomics are often used interchangeably but do have distinct meanings. Although either term refers to the use of genetic information to guide therapeutic decision-making:

Pharmacogenetics focuses on the influence of single genes on drug response.

Pharmacogenomics takes a broader view of the influence of an individual’s entire genome on his or her response to drug therapy.

One can consider pharmacogenetics to be the original term for this field, coined in the days when tools such as microarrays did not exist and the human genome had not been mapped. In those early days it seemed unlikely or impossible that an entire genome could be analyzed in an efficient enough manner to become a useful tool in everyday medicine. Now that we have the means and the knowledge to envision a genome-wide approach to everyday prescribing, the term pharmacogenomics is often used to refer to this field. Table 6-1 lists some of the terms commonly used in pharmacogenetics.

TABLE 6-1 Definitions of Some Common Terms Used in Pharmacogenetics

Gene	A sequence of nucleotides that corresponds to a sequence of amino acids in an entire protein or part of a protein. Genes are typically found at a specific location on a chromosome.
Genome	The full genetic complement of an individual.
Allele	Any of the alternative forms of a gene at a particular locus. These alternative forms may or may not result in different phenotypes.
Null allele	A mutation in a gene that leads to a loss of function. Either the gene is not expressed at all (i.e., no protein or RNA) or the product is not functional.
Polymorphism	Variation in DNA sequence present at a specific locus within a population.
Genotype	The genetic makeup of an individual.
Phenotype	The observable physical or biochemical characteristics of an organism. Determined by genotype and environment.
Monogenic	Related to or controlled by a single gene
Polygenic	Related to or controlled by multiple genes. A polygenic trait is a phenotype that is determined by multiple genes rather than a single gene (monogenic).
Germ line	Cellular lineage; genetic information that is passed from one generation to the next.
Haplotype	A group of alleles of different genes on a single chromosome that are so closely linked that they are inherited as a unit.
Somatic cell	Any cell in the body, with the exception of those involved in reproduction.

The most common basis for genetic variation, and thus the basis for a pharmacogenomic approach to drug therapy, is the single nucleotide polymorphism, or SNP (Figure 6-1). An SNP occurs when a single nucleotide is exchanged for another at a point in an individual’s genome. It is estimated that the human genome consists of approximately 3 billion nucleotides, which in specific combinations form 25,000 to 40,000 genes and encode approximately 100,000 proteins (at last count). SNPs that occur in coding regions of the genome have the potential to influence protein expression by altering an amino acid within the protein.

Nonsynonymous coding SNPs (cSNPs) or missense SNPs change the identity of an amino acid. Substitution of this amino acid may change the structure or function of the protein.

Synonymous coding SNPs or sense SNPs do not change the identity of an amino acid and are not likely to affect the structure or function of the protein.

Nonsense mutations are SNPs that lead to a stop codon. A stop codon will halt translation of nucleotides into a protein.

Figure 6-1 Single nucleotide polymorphism.

These variants are often indicated with an asterisk followed by a number indicating the specific mutation in that allele (e.g., CYP450 2D6*4). Certain alleles occur more commonly in some ethnic groups than in others. The impact of these SNPs on phenotypes and their subsequent clinical consequences can again be divided into the two fundamental branches of pharmacology: those that influence pharmacokinetics and those that influence pharmacodynamics.

Pharmacokinetics

The field of pharmacogenomics began with the observations of dramatic differences in the way that certain individuals metabolize drugs. This, along with the fact that adverse drug reactions (ADRs) were the first and most obvious application of pharmacogenomics, has meant that the influence of pharmacogenomics on pharmacokinetics has been much more extensively studied than the impact of pharmacogenomics on pharmacodynamics.

Until recently, the identification of genetically based aberrant metabolism would invariably begin with the observation of an ADR. An ADR can occur because of either higher- (more frequent) or lower-than-expected plasma levels of a given drug. A physician would note toxicities to therapeutic doses of a given drug and discover that plasma levels of that drug were much higher than expected. When other factors such as drug interactions or liver or renal disease were factored out, the clinician was left with genetics as the most viable option for explaining the unexpectedly high plasma levels.

A classic example occurred with the antituberculosis drug isoniazid (Figure 6-2). Isoniazid is acetylated (metabolized) by N-acetyltransferase 2 (NAT2), and it was observed that some patients metabolized this drug slowly (i.e., were slow metabolizers), whereas others were rapid metabolizers.

The clinical consequence is that rapid metabolizers would have low plasma levels of the drug, increasing the potential for therapeutic failure, whereas slow metabolizers were predisposed to toxicities associated with the drug.

This wide variation in response to isoniazid led researchers to speculate that genetics might be playing a role. Isoniazid became one of the first drugs to be identified as having a genetic component to its pharmacologic actions.

Figure 6-2 Influence of genetics on isoniazid metabolism.

(Modified from Meyer UA: Pharmacogenetics: five decades of therapeutic lessons from genetic diversity, Nat Rev Genet 2004 5:669, 2004. Reprinted by permission from Macmillan Publishers.)

Further confirmation of a genetic role would come with the identification of other family members that share the same trait.

With regulatory bodies now either strongly suggesting or requiring that pharmacogenetic studies be submitted with new drug applications, the sequence of events noted earlier has been reversed. Once a new drug enters the market, the pharmacogenetic profile has typically already been characterized, and if an individual has undergone genotyping, many ADRs can be avoided, it is hoped, with adjustments to the drug regimen. The key is to be able to identify the polymorphisms in a given patient before initiation of therapy.

Phase I Reactions, CYP450, and Genetic Polymorphisms

The cytochrome P-450 superfamily of enzymes plays a key role in phase I drug biotransformation. The CYP450 family is itself a rather large collection of closely related enzymes, with a system of nomenclature that reflects these relationships—for example, CYP3A4 and CYP3A5 are more closely related than CYP3A4 and CYP2D6. However, with time it has become apparent that there is heterogeneity within these subfamilies and that this heterogeneity is genetically based.

An early example of genetic differences in CYP450 enzymes occurred with CYP2D6 and debrisoquine, an older sympatholytic once used for hypertension. A landmark study conducted in 1000 Swedish subjects clearly demonstrated the distinction between poor metabolizers, extensive metabolizers, and ultrarapid metabolizers in this population. CYP2D6 has since become a prime example of the impact of genetic polymorphisms on pharmacokinetics. Poor metabolizers have two null alleles of CYP2D6, the most frequent being *4, which occurs in 20% to 25% of Caucasians and is responsible for 79% to 90% of all poor metabolizers.

Conversely, ultrarapid metabolizers have a gene duplication or multiple duplications, meaning that they have excess enzyme.

Figure 6-3 demonstrates how these phenotypes can influence the dosage of a drug, in this case the antidepressant nortriptyline. Note that most patients fall into the extensive or intermediate metabolizer category.

It is the outliers, the poor metabolizers and the ultrarapid metabolizers, however, that require significant changes to the normal dose. Note that although there are fewer people with either of these phenotypes, poor metabolizers and ultrarapid metabolizers are still present in 10% to 20% of this population.

Figure 6-3 Influence of phenotype on dosing.

(Modified from Meyer UA: Pharmacogenetics: five decades of therapeutic lessons from genetic diversity, Nat Rev Genet 2004 5:669, 2004. Reprinted by permission from Macmillan Publishers.)

The clinical impact of CYP450 polymorphisms varies among the various isoenzymes of this superfamily; the effects are summarized in the following sections.

CYP2D6

CYP2D6 is the CYP enzyme that is most associated with genetic polymorphisms. The frequency of polymorphisms varies greatly among ethnicities. For example, 5% to 10% of Caucasians are poor metabolizers, whereas only a small percentage (<1%) of Africans and Asians are poor metabolizers. Much of the phenotype in Caucasians is caused by the *4 null allele.

Table 6-2 illustrates how much the frequency of CYP2D6 ultrarapid metabolizers can vary within Europe. For example, 10% of Greek citizens are ultrarapid metabolizers, whereas only 1% of people from Finland have this phenotype.

CYP2D6 is most commonly associated with metabolism of antidepressants, including the tricyclics (e.g., nortriptyline), selective serotonin reuptake inhibitors (SSRIs) (e.g., paroxetine), and venlafaxine. Other notable drugs using this isozyme include β-blockers (e.g., metoprolol) and antipsychotics (e.g., risperidone). Other examples include the following.

The conversion of codeine to morphine is also mediated by this isozyme, and ultrarapid metabolizers of CYP2D6 are prone to opioid toxicity owing to excess conversion of codeine to morphine.

TABLE 6-2 Estimation of the Number of Ultrarapid CYP2D6 Metabolizers in Western Europe^*

Phase II Reaction Polymorphisms

Phase II reactions are also prone to genetic polymorphisms, although phase II reactions are less important than CYP450 in metabolism. Therefore these polymorphisms also have less clinical significance than those involving CYP450.

Pharmacodynamics

One of the most common ways that polymorphisms alter drug response is through mutations that alter the structure or even the presence of drug targets. Some examples include the following:

Serotonin reuptake transporters in depression. Patients with polymorphisms that lead to reduced expression of serotonin transporters may be more resistant to treatment with SSRIs. The problem with trying to draw a definitive link between this phenotype and treatment failure is the influence of other factors such as psychotherapy and even other genetic factors that may predispose the patient to treatment-resistant depression.

β₂ receptors in asthma. Polymorphisms in β₂ receptors can lead to either reduced or enhanced efficacy of β₂ agonists. One study found that children with a specific polymorphism leading to more rapid response had shorter stays in the ICU after an asthma attack.

Cancer is perhaps the most promising application for pharmacogenetics, particularly with respect to pharmacodynamics. As a disease that is rooted in genetics, polymorphisms that influence drug response are ubiquitous in cancer and contribute to the relatively low response rates to treatment after many decades of concerted effort in drug development.

It is not unusual to see response rates of 10% to 20% for successful cancer regimens, meaning that patients must endure considerable trial and error before an effective regimen is found. Unfortunately in many cases, these effective regimens are found too late. One of the promises of pharmacogenetics in cancer is that identification of polymorphisms may allow targeting of drug therapy to the correct patient.

One recent success story occurred with a drug called gefitinib, for advanced non–small cell lung cancer. The response rate, measured by a significant reduction in tumor size, was only 10% in this difficult-to-treat population. However, it was also observed that women of Asian descent with no history of smoking responded particularly well to this drug.

Eventually it was discovered that patients whose tumors had mutations in the epidermal growth factor receptor (EGFR), the target for gefitinib, responded to the drug with remarkable predictability, whereas patients who lacked these mutations rarely if ever responded to treatment.

Another promising feature of the interaction between pharmacodynamics and pharmacogenetics is in drug development. Again using cancer as an example, it is now possible to readily identify mutations in tumors that can then be used to direct the design of a given drug. For example, targeted therapies such as imatinib and dasatinib for chronic myelogenous leukemia (CML) have been designed with specific mutations in mind.

CML is an ideal disease for personalized medicine, as it is caused by a single oncogene, known as BCR-ABL. BCR-ABL has a very specific marker that is part of the diagnostic process, and results in a gain of function (i.e., excess cell division).

Imatinib was specifically designed to address this mutation. When it was discovered that cancers were mutating, resulting in loss of response to imatinib, another drug (dasatinib) was subsequently designed to specifically address patients with this new mutation, and so on.

This use of pharmacogenetics has lead to CML being one of the most ‘treatable’ cancers, with >95% of patients responding to therapy, and possibility of a 100% response rate in the near future.

The Next Step: from Genetics to Genomics

One of the challenges with interpreting the influence of genetic polymorphisms on drug response is the interplay of multiple factors.

For example, warfarin levels are influenced by alterations in CYP450 expression, as described earlier, but another key polymorphism that influences warfarin response occurs in the enzyme vitamin K epoxide reductase (VKOR), which is the pharmacodynamic target for warfarin.

VKOR converts vitamin K to its reduced form, which is the form that activates several clotting factors in the coagulation cascade. Warfarin inhibits this enzyme, therefore inhibiting activation of clotting factors and inhibiting coagulation.

Depending on the VKOR phenotype of the patient, he or she may require either higher or lower doses of warfarin. Patients with low VKOR levels require lower doses, and patients with high VKOR expression need higher-than-average doses. This polymorphism is actually thought to have a greater role in determining variations in warfarin dosage than polymorphisms in CYP450.

Complications in determining clinical impact therefore arise when a patient has both the CYP450 and VKOR polymorphisms, particularly when they occur in opposing directions: one leading to an increased dose requirement, and one to a decreased requirement. In these cases, it would be difficult to predict whether there is no net effect (i.e., the two cancel each other out) or the net direction of effect.

Given the large number of patients with polymorphisms that affect pharmacokinetics, and a similarly large number with polymorphisms that affect pharmacodynamics, it is likely that the scenario just described will be identified frequently in the future. Addressing the impact of a patient’s genome rather than a single genetic polymorphism on drug response requires tools that will manage this overwhelming amount of data for clinicians. Perhaps the most important tool that will allow the routine use of pharmacogenomic data is the microarray.

Detection of Genetic Polymorphisms

A microarray is a collection of immobilized single-stranded DNA fragments that contain a known nucleotide sequence that is used to identify and sequence DNA samples (Figure 6-4). Microarrays can be used in the analysis of gene expression. A microarray provides an automated means for identifying genes in a given sample. For example, cancerous and healthy cells may be analyzed from a single patient in order to determine the differences in gene expression between these two types of cells.