David A. Mrazek, MD

While the utilization of clinical genotyping to determine drug response and dosage has been anticipated for many years, the actual utilization of screening for atypical drug metabolizers has only recently become a reality. The key factor making this innovation possible has been the development of low-cost, high-throughput genotyping that can accurately establish the precise genotype of key drug metabolizing enzymes (DME). In the very near future, more complex methodologies to ascertain the influence of both DME genes and genes that code for key drug targets will be available. However, today it is possible to improve the likelihood of a positive therapeutic outcome as well as minimize adverse drug effects for select children and adolescents who are being treated for depression.

At this point in the development of psychiatric pharmacogenomics, the most frequently screened DME gene is cytochrome P450 (CYP) 2D6. Genotyping of the 2D6 gene is available through most large reference laboratories. For example, clinical testing of the 2D6 genotype was initiated at the Mayo Clinic in February 2003.

One important consideration in pharmacogenomic clinical genotyping is the accuracy and comprehensiveness of the assessment. The 2D6 gene is highly variable, having 180 reported variants. The vast majority of these variants are extremely rare, but there are 11 different alleles that occur in significant numbers of patients throughout the world.

This gene is located on the long arm of the 22nd chromosome in an area that is referred to as a "hot spot." It is a hot spot because there is an increased probability for the occurrence of "uneven crossovers" at this location, and as a consequence, there has been an accumulation of multiple copies of this gene on this chromosome. The wide variation in the 2D6 gene demonstrates that the protein product is not a critical enzyme required for survival.

In white populations, 11 polymorphisms account for about 99% of the allelic frequency. 

 lists the most common alleles, as well as estimates of their population frequency. While these estimates are generally true for most U.S. white communities, there are a number of interesting exceptions where specific polymorphisms have accumulated in homogeneous ethnic populations.

In order to use genotypic data clinically, a practical systematic method of categorizing genotypes has been developed. Essentially, there are four broad categories. The first category is normal or "extensive" metabolizers. These individuals have two effective copies of the 2D6 gene. The second category is "intermediate" metabolizers, which describes individuals who have one effective copy of the 2D6 gene and one defective copy. The third category is "poor" metabolizers, which describes those individuals who have two defective copies of the gene. Finally, there are individuals who are categorized as ultrarapid metabolizers who have three or more effective copies of the gene due to uneven crossovers during meiosis.

To maximize the clinical value of 2D6 genotyping, the reference laboratory must be able to accurately identify all of the relevant polymorphisms. For the most part, rare polymorphisms code for inactive forms of the enzyme. However, if these rare polymorphisms are not explicitly identified by the assay, microarray analyses can misread the aberrant polymorphism as normal. For example, if only four common forms of the gene (*1, *2, *3 and *4) are screened for in an assay, it would be estimated that 21% of genotypic readings will contain an error. The two most common errors using too few alleles are to misread intermediate metabolizers as normal metabolizers or poor metabolizers as intermediate metabolizers. This problem can be virtually eliminated by screening for the 11 most common forms of the 2D6 gene.

The initial motivation to screen patients using 2D6 genotyping was to identify poor metabolizers who were at high risk of developing adverse side effects and toxic levels of 2D6 substrate medications. In extreme cases, unrecognized poor metabolizers have been treated with moderate-to-high levels of medication resulting in severe side effects and, in rare cases, death (Sallee et al., 2000).

The first application of 2D6 genotyping in clinical practice was to genotype patients who complained of side effects at low doses of 2D6 substrate medication with the goal of being able to confirm the etiology of their adverse effects. More recently, the cost of genotyping has decreased dramatically while at the same time, concerns regarding the safety of antidepressant medication have become more clearly articulated. In this climate of increased concern for safety, the prescreening of all patients who are to be given 2D6 substrate medications has become a standard component of the workup prior to administration of antidepressant medication in some clinical practices.

 illustrates some examples of commonly prescribed medications that are primarily metabolized by the 2D6 enzyme. These medications should either be avoided completely or used in very low doses in patients who have been demonstrated to be poor metabolizers (Kirchheiner et al., 2001). Other medications--in addition to antidepressants and other psychotropic medications--are metabolized by the 2D6 enzyme. One important example is codeine, which is best conceptualized as a "pro-drug" that requires metabolism to an active metabolite. Specifically, the analgesic effect of codeine is the result of its metabolism to its first metabolic product, which is morphine. In patients with poor 2D6 metabolism, little of the prescribed dose of codeine is transformed into morphine. Consequently patients receive inadequate analgesia. Another example is the use of cough syrups with dextromethorphan in patients who are poor 2D6 metabolizers. Large or even moderate doses of dextromethorphan may result in high serum levels of the active ingredient in patients with poor 2D6 metabolism. These high blood levels can result in acute agitation and, in rare cases, psychotic symptoms.

Patients who are poor metabolizers routinely develop higher levels of antidepressant medication than were intended by their clinicians. This inadvertent "overdose" can be particularly problematic in children and young adults who have prodromal bipolar psychiatric symptoms that include depression. The use of selective serotonin reuptake inhibitors that are primarily metabolized by 2D6, such as paroxetine (Paxil) and fluoxetine (Prozac), can be problematic at even moderate doses, since higher than anticipated serum levels can be obtained that result in an activation of manic symptomatology.

The second major indication for clinical 2D6 genotyping is to identify ultrarapid metabolizers. In most white populations, the frequency of this condition is between 1% and 2% of the population (Cascorbi, 2003). However, in select populations in Europe, such as Spain, the incidence of ultrarapid metabolizers reaches 10%. The rate of ultrarapid metabolizers in the Middle East ranges between 15% and 20%, with the rate in Ethiopia exceeding 20% (Aklillu et al., 1996). This interesting geographic distribution of individuals with ultrarapid metabolism suggests that there may be some selective advantage for this genotype in these specific geographical areas.

The primary difficulty in treating ultrarapid metabolizing patients with 2D6 substrate medications is that therapeutic serum levels of these medications are difficult to achieve. One possible strategy for managing these patients is to gradually increase their dose of medication to higher than recommended dosage levels. However, some caution in this strategy is warranted, given the limited information available about the primary and secondary metabolites of many medications, and the reasonable concern that accumulation of these metabolic products may be problematic. A more prudent clinical approach is to consider an alternative antidepressant medication that is metabolized by either multiple CYP enzymes or through an entirely different mechanism.

Currently, there is little empirical evidence to guide clinicians in the proper dosage of patients with an intermediate rate of metabolism. The demonstration of the pharmacokinetic responses of patients with intermediate 2D6 genotypes should be the focus of future research.

Relatively few technical innovations in psychiatry have modified clinical practice over the past decade. However, the current availability of increasingly sophisticated and accurate genotype screening promises to have widespread implications for pharmacotherapy. In addition to genotyping the 2D6 gene, there is increasing interest in genotyping the entire family of CYP genes. For psychiatrists, genotyping of the 2C19 gene to complement information derived from 2D6 genotyping has obvious clinical benefit.

While clinical genotyping of the serotonin transporter gene has not been widely adopted, it is a likely candidate for introduction into clinical practice after the confirmation of preliminary findings that specific genotypes are linked to variable drug response. Specifically, white patients who are homozygous for the long form of the genotype have been demonstrated to have more effective responses to fluoxetine and paroxetine. It is also anticipated that additional receptor genotyping will further improve the accuracy of pharmacogenetic prediction.

Dr. Mrazek is chair of the department of psychiatry and psychology at the Mayo Clinic and professor of psychiatry and pediatrics at the Mayo Clinic College of Medicine.

Aklillu E, Persson I, Bertilsson L et al. (1996), Frequent distribution of ultrarapid metabolizers of debrisoquine in an Ethiopian population carrying duplicated and multiduplicated functional CYP2D6 alleles. J Pharmacol Exp Ther 278(1):441-446.

Cascorbi I (2003), Pharmacogenetics of cytochrome p4502D6: genetic background and clinical implication. Eur J Clin Invest 33(suppl 2):17-22.

Kirchheiner J, Brosen K, Dahl ML et al. (2001), CYP2D6 and CYP2C19 genotype-based dose recommendations for antidepressants: a first step towards subpopulation-specific dosages. [Published erratum Acta Psychiatr Scand 104(6):475.] Acta Psychiatr Scand 104(3):173-192.

Sallee FR, DeVane CL, Ferrell RE (2000), Fluoxetine-related death in a child with cytochrome P-450 2D6 genetic deficiency. J Child Adolesc Psychopharmacol l0(1):27-34.

Articles

While the utilization of clinical genotyping to determine drug response and dosage has been anticipated for many years, the actual utilization of screening for atypical drug metabolizers has only recently become a reality. What is the promise of this technology, as well as the limitations?

David A. Mrazek, MD

Articles

Pharmacogenomic Screening for Depressed Children and Adolescents