Drugs used for psychiatric applications span a wide
range from mood-altering drugs like antidepressants, to anti-psychotics
used in schizophrenia. Drugs for anxiety, attention
deficit/hyperactivity, and delirium are also included in this class. We
also know from first- or second-hand experience that reports of efficacy
with these medications can vary widely. Many psychiatric patients are
familiar with the notion of “trying” new medications and new treatment
regimens under the direction of their physicians to see if the drugs
improve their conditions. Pharmacotherapy in psychiatry is almost always
performed on a trial-and-error basis; a given drug or treatment’s
efficacy is assessed individually and empirically by patients and their
physicians. The inter-individual differences in response to treatments
as well as the variability in side effects experienced by the
psychiatric patient are a well-known phenomenon. Treatment failure in
psychiatry is common.
Pharmacogenomics
Psychiatry could turn out to be one of the strongest
benefactors of pharmacogenomics (PGx) testing. Pharmacogenomics is the study
of how individual variations in the human genome affect disposition and
response to medications. Although there are multiple factors which can
affect drug response, a predominant factor is the inter-individual
differences in the function and/or expression of drug-metabolizing enzymes.
In PGx, polymorphisms can help predict drug response.
Polymorphisms are alterations in a gene (allele) affecting at least 1% of
the population. Such alterations may or may not affect the function of the
resultant protein product. For those variant proteins that do alter drug
disposition or response, if they can be uncovered a priori, that
knowledge should help predict drug response.
Although polymorphisms in a gene’s coding for drug
receptors and transporters have been shown to affect drug response, the most
well-studied proteins in PGx are the cytochrome P450 enzymes (CYP450), a
large group of heme-containing enzymes which are found predominantly in the
liver. CYP450 enzymes work on a variety of substrates, altering their
molecular structures to facilitate excretion. These enzymes account for much
of the liver’s detoxifying action, altering endogenous and foreign compounds
so they can be excreted from the body.
There are many different CYP450 enzymes which metabolize
psychiatric drugs. CYP1A2, CYP2C9, and CYP2C19 are all associated with
genetic polymorphisms which can affect psychiatric-drug performance. The
CYP450 enzyme of most interest in psychiatric applications, however, is
CYP2D6, as it is thought to be responsible for the metabolism of at least
25% of all drugs.1 CYP2D6 accounts for only about 1% of all
CYP450 enzymes but is important in the metabolism of about 100 drugs, many
of which are used in psychiatric applications.2,3 There are
dozens of CYP2D6-variant alleles which can arise as the result of mutations
and polymorphisms. The normal wild-type allele displays average metabolic
activity, whereas some variants have enhanced or diminished activity. Some
clinical reference laboratories now offer CYP2D6 genotyping, which will
indicate whether a patient is a poor, intermediate, extensive, or
ultra-rapid metabolizer.
By knowing a patient’s disposition to drugs, he could be started on appropriate dosing regimens without the extensive trial-and-error period that is common with psychiatric medications.
By measuring a dozen or so of the most common and
significant variant sites, a laboratory can detect approximately 98% of
known variant CYP2D6 alleles. How common are these variants? The prevalence
varies from <1% to as much as 21%. It is estimated, for example, that 8% to
10% of the population has a complete deficiency of CYP2D6.4,5 The
prevalence of poor metabolizers is thought to be around 6% to 10% for white
populations but is lower in other ethnic groups such as Asians and African
Americans.6,7,8 CYP2D6 activity in the general population of
‘normal’ metabolizers is also known to be highly variable, ranging as much
as ten thousandfold between patients.9 Clearly, the
pharmacogenomics associated with CYP2D6 is relevant to millions of people.
Clinical implications
What are the clinical implications for those who have
abnormal CYP2D6 metabolism? Poor metabolizers (PM) are persons who carry two
deficient CYP2D6 alleles and, as a result, exhibit decreased metabolism of
drugs. These patients require lower dosages to establish therapeutic levels
of drug in vivo. A PM patient who receives a standard dose is more
likely to experience unwanted side effects — or even toxicity — since he
will metabolize and clear the drug more slowly. This is important when
considering the high rate of unwanted side effects experienced by patients
on mood-altering drugs — side effects which often lead to discontinuation of
therapy.
Poor metabolizers can also experience diminished effects
with drugs that need to be metabolized to active compounds by CYP450s. Using
an analgesic example, the prodrug codeine, which is converted to morphine
in vivo, will not be adequately transformed in PM patients, leading to
diminished pain relief or the appearance of tolerance/addiction.
Intermediate metabolizers (IM) have one wild-type copy of the gene and one
absent, or dysfunctional, copy. The IM group is very heterogeneous. Persons
with normal enzyme activity who carry two functional alleles are referred to
as extensive metabolizers (EM). These patients should respond to standard
dosages although, as with the IM group, the EM response is also wide
ranging. Ultra-rapid metabolizers (UM) have more than two functional alleles
due to gene duplication or multiplication. Because of this, they require
higher doses than normal, since drug metabolism and clearance is enhanced.
These patients may be resistant to treatments, and more time may be required
to adjust the dosage before therapy is achieved. In theory, identifying a
CYP2D6 UM upfront, would decrease the time needed to adjust a dosage upward,
helping to achieve therapeutic success faster.
Other variables besides mutations affect CYP450 enzymes.
Many drugs are known to induce the expression of CYP450 enzymes. CYP450
enzyme levels change in the presence of certain chemicals. Smoking, for
example, can induce CYP1A2. St. John’s wort can induce CYP3A4. Additionally,
the drugs carbamazepine, rifampin, and phenobarbital can induce many
CYP450s, including CYP2D6. Of course, CYP450s can also be inhibited by
substances and substrates. CYP2D6 can be inhibited by many drugs including
common medications like cimetidine (Tagamet) and fluoxetine (Prozac).
Because most patients are on multiple medications and since dietary and
environmental factors can change CYP450 expression levels, the metabolic
capacity of patients cannot be solely predicted based on their genotypes. To
consider both when ascribing the predictive power of a genotype and when
trying to justify the cost of genotyping is important.
Genotype vs. phenotype
Genotyping can give a definitive profile of CYP2D6
alleles. But because there are more than 70 known mutations and
polymorphisms that can occur on CYP2D6, even large, well-equipped reference
laboratories are not likely to offer complete screening or sequencing. Most
genotyping approaches only test for the most common or best characterized
alleles. Without complete sequencing of the entire allele, one may not be
able to completely rule out a relevant mutation or polymorphism in a patient
who shows none of the more common alleles. Yet, complete sequencing methods
are currently cost-prohibitive and if less-common variants are uncovered,
the significance of these will likely be unknown. Thus, complete sequencing
may only confound the interpretation, since the significance of all possible
mutations has not yet been ascertained.
When one considers the number of variants, the fact that
multiple CYP450s may be involved in a drug’s metabolism and the possible
presence of inducing/inhibiting substances in the patient, phenotyping for
drug metabolism can sound more attractive than genotyping. For CYP2D6, this
can be accomplished with the use of probe drugs like dextromethorphan or
debrisoquine. Dextromethorphan is a common over-the-counter cough
suppressant which is extensively metabolized by CYP2D6. The primary
metabolite is the demethylated compound dextrorphan. By measuring the parent
drug and the metabolite in urine, the metabolic capacity of CYP2D6 can be
estimated. This test is, however, labor intensive, requiring chromatography
and/or mass spectrometry techniques; questions remain concerning its
sensitivity.
Simple therapeutic drug monitoring of the drug in
question can also reveal a good deal about a drug’s metabolism and, like
probe-drug analysis, would take into account the influence of external
factors like co-medications, diet, smoking, and impaired organ function. But
using therapeutic drug monitoring requires that an assay be available for
the psychiatric drug in question. Immunoassays for most CYP2D6-metabolized
psychiatric medications are not widely available for automated analyzer
platforms. When antibody assays are available, they are more often employed
as qualitative toxicology assays rather than quantitative drug-monitoring
assays. Gas chromatography with mass spectrometry (G-MS) can be used for
nearly all psychiatric drugs, but such equipment is common only in larger
reference laboratories and, again, any GC-MS assay must be set up as a
quantitative assay rather than a more simple qualitative drug screen if the
results are to have value in gauging drug metabolism.
Pharmacogenomics is the study of how individual variations in the human genome affect disposition and response to medications.
Also consider the complications surrounding specimen
collection. Multiple serum and/or urine specimens are needed to assess a
drug’s metabolism in vivo. This would be a significant pre-analytical
challenge compared with genotyping which requires only one blood sample for
the entire life of the patient.
It is important to point out that probe-drug analysis is
not, in principle, synonymous with therapeutic-drug monitoring. Unlike
genotyping and probe-drug testing, therapeutic-drug monitoring is performed
during therapy, and, thus, it is not predictive of drug disposition.
Instead, it simply reports drug disposition after therapy has started and
reached steady state. Given the complexities of probe-drug analysis,
genotyping of CYP2D6 is likely to be the method of choice for predicting
patient drug response, despite its limitations. Yet, if quantitative
therapeutic drug monitoring methods for psychiatric drugs are in place, they
can be used alongside genotyping. Ideally, a genotype would first predict
response, and then therapeutic-drug monitoring could be used to verify
appropriate serum drug levels.
Summary
The ultimate goal in measuring CYP2D6 (or any other
CYP450) function or identifying variant alleles is to predict effective
therapeutic doses and responses in patients. This is the promise of
individualized medicine. By knowing a patient’s disposition to drugs, he
could be started on appropriate dosing regimens without the extensive
trial-and-error period that is common with psychiatric medications. We could
also avoid drugs whose metabolism may prove to be problematic, choosing
second-line therapies which are metabolized by different, unaffected
enzymes. Of course, knowing a genotype is not very useful unless we couple
the genotype findings with clinically validated dosing algorithms.
Dosing recommendations for PM, EM, IM, and UM patients
are beginning to appear in the literature for various classes of drugs but,
at present, there are no well-accepted guidelines available. The Food and
Drug Administration does encourage the incorporation of pharmacogenomic
testing for investigational compounds in the development process. As the
notion of PGx becomes more familiar and more clinical trials are completed,
evidence-based dosing adjustments should be forthcoming.
Although the biochemistry and pharmacology of CYP450 drug
metabolism has made huge strides, the application of pharmacogenomics has
not yet become commonplace for a number of reasons. Drug metabolism is a
complex process, and CYP2D6 may not be the only polymorphic protein involved
in a given drug’s metabolism. Also, both primary and secondary metabolic
pathways exist for drugs, the latter of which may be utilized when other
drugs or endogenous compounds occupy the principle pathway. Given the
possibility of multiple metabolic pathways, the presence of co-medications,
inducers, and inhibitors in the diet and disease changes, predicting drug
metabolism in a person remains difficult even when a given CYP450 genotype
is obtained.
Also there are multiple variants which can be present and
consideration must be given as to which variant allele(s) to look for while,
at the same time, always considering the cost/benefit ratio of a possible
testing algorithm. Despite the numerous uncontrolled variables involved in
drug metabolism and the inability for pharmacogenomics to address them all,
there remains some promise for CYP2D6 genotyping to at least help physicians
hone in on appropriate dose ranges before therapy is initiated or in
identifying individuals at metabolic extremes who are at the highest risk
for adverse outcomes. Ultimately, if CYP2D6 characterization is shown to be
an evidence-based improvement in the practice of psychiatric medicine,
laboratorians will need to be prepared for an influx of requests for these
tests.
Kevin F. Foley, PhD, D(ABCC), MT, a director of clinical chemistry at Kaiser Permanente NW in Portland, OR,
has served as academic faculty at the University of Vermont, Mayo Clinic,
and Northern Michigan University. His training in medical technology was
followed by graduate training in pharmacology. Denise I. Quigley, PhD,
F(ACMG), is board certified in clinical cytogenetics and clinical molecular genetics and
directs the Cytogenetics and Molecular Genetics Laboratories at Kaiser
Permanente Northwest-Portland Her PhD in molecular and medical genetics was
granted by Oregon Health and Science University, and she completed her
post-graduate training in cytogenetics and molecular genetics at the
University of North Carolina-Chapel Hill.
References
- Ingelman-Sundberg M, Genetic polymorphisms of
cytochrome P450 2D6 (CYP2D6): clinical consequences, evolutionary
aspects and functional diversity.Pharmacogenomics J.
2005;5(1):6-13. - DeVane CL, Nemeroff CB. 2002 guide to psychotropic
drug interactions. PrimPsychiatry. 2002;9:28-57. - Gough AC, Smith CA, Howell SM, Wolf CR, et al.
Localization of the CYP2D gene locus to human chromosome 22q13.1 by
polymerase chain reaction, in situ hybridization, and linkage analysis.
Genomics. 1993;15:430-432. - Sachse C, Brockm”oller J, Bauer S, Roots I. Cytochrome
P450 2D6 variants in a Caucasian population: allele frequencies and
phenotypic consequences. Am J Hum Genet. 1997;60(2):284-295. - Sistonen J, Sajantila A, Lao O, Corander J, Barbujani
G, Fuselli S. CYP2D6 worldwide genetic variation shows high frequency of
altered activity variants and no continental structure.Pharmacogenet
Genomics. 2007;17(2):93-101. - de Leon J, Barnhill J, Rogers T, Boyle J, et al.
Pilot study ofthe cytochrome P450-2D6 genotype in a psychiatric state
hospital. Am J Psychiatry. 1998;155:1278-1280. - Daly AK, Brockmoller J, Broly F, et al. Nomenclature
for human CYP2D6 alleles. Pharmacogenetics. 1996;6: 193-201. - Belpaire FM, Bogaert MG. Cytochrome P450: genetic
polymorphism and drug interactions. Acta Clin Belg.
1996;51:254-260. - Bertilsson L, Dahl M-L, Ekqvist B, Jerling M, Lierena A. Genetic
regulation of the disposition of psychotropic drugs. In: Meltzer HY, Nerozzi
D, eds. Current Practices and Future Developments in the Pharmacotherapy
of Mental Disorders. Amsterdam: Elsevier Scientific
Publications; 1991:73-80.
