Pharmacogenomics of Antidepressants:
Drug Discovery, Treatment, and Ethical Considerations
Julio Licinio, MD
Jonas Hannestad, MD, PhD
Ma-Li Wong, MD
Dr. Licinio is professor of psychiatry and biobehavioral sciences and medicine/endocrinology at the University of California School of
Medicine in Los Angeles. He is also director of the University’s Interdepartmental Clinical Pharmacology Center. Dr. Hannestad is postdoctoral fellow, and Dr. Wong is professor of psychiatry and biobehavioral sciences and director, both at the Laboratory of Pharmacogenomics at Neuropsychiatric Institute of the University of California School of Medicine, Los Angeles.
Pharmacogenomics is a new area of medicine that uses the data emerging from the sequencing of the human genome to predict drug responses and to identify new targets for treatment. Pharmacogenomics is of great relevance to depression, a common and complex disorder of unknown cause for which prediction of treatment response and identification of new targets for therapeutics are of crucial importance. The clinical reality is that weeks of continued antidepressant treatment are required before therapeutic effects occur. Moreover, it is not possible to know in advance if a patient is likely to respond to a specific drug. While therapeutic effects take long to emerge, adverse reactions manifest themselves rather
soon. This impacts negatively on compliance, leading to incomplete and failed treatments, and potentially disastrous outcomes, such as suicide, which is now the eighth leading cause of death in the United States. The identification of the genomic substrates underlying antidepressant treatment would facilitate not
only the formulation of individualized treatment approaches, but also the development of new classes of drugs that would affect those genomic targets more directly than existing compounds, leading to more rapidly effective clinical responses. As the research needed to achieve these goals is being conducted, myriad ethical questions emerge: Which ethnic groups will participate in—and consequently benefit from—such research? How will genomic information related to drug response be handled in ethical, legal, economic, and social terms? Progress in the pharmacogenomics of depression needs to be paralleled by thoughtful consideration of the implications of such work at the individual and societal levels.
“Doctors are men who prescribe medicines of which they know little, to cure diseases of which they know less, in human beings of whom they know nothing.”
But now that the human genome is an open book, personalized medicines—drugs tailored to our genetic idiosyncrasies—will soon be possible. In a decade or so, pills geared to particular “genotypes” are expected to begin arriving in pharmacies along with tests to show who should get them. Eventually we’ll look back in wonder at how we used to play guinea pig in the primitive therapeutic experiments our doctors carried out each time they wrote new prescriptions for us, just as we now shake our heads about the poor chumps who got blood transfusions before it was possible to match donors’ and recipients’ blood types (The adverse reactions included kidney failure and fatal clots).1
Pharmacogenomics is a new area of medicine in which information databases that are emerging from the sequencing
of the human genome and related high throughput technology are used to improve therapeutics. The term “psychogenomics”2 has been used to describe the process of applying the powerful tools of genomics and proteomics to achieve a better understanding of the biological substrates of normal behavior and of diseases of the brain that manifest themselves as behavioral abnormalities. This article will cover two topics that are of direct relevance to genomics in neuroscience. The first is the application of genomics to the treatment of a specific psychiatric disorder. We will discuss depression, which is a common and complex disorder of unknown case with high prevalence that costs the US economy over $50 billion per year, and which is the fourth cause of disability worldwide (second in developed countries).3-6 The second topic we address in this article is the issue of conducting research in ethnically identified groups.Pharmacogenomics can offer identification of novel therapeutic targets and individualization of treatment
Pharmacogenomics of Depression: Drug Discovery
The psychosocial substrates of depression are fully acknowledged and recognized. For example, it is undeniable that stress and loss can precipitate episodes of depression. For example, the work of Lerer’s group in Jerusalem has shown that loss (divorce being worse than death) of a parent (mother more than father) before age 9 years was highly associated with depression in adulthood.7 Additionally, structured psychosocial treatments, such
as cognitive or interpersonal psychotherapies, can be highly effective in the treatment of depression.8,9 Thus, one cannot deny the impact of psychosocial approaches to depression treatment. Nevertheless, clinical experience also supports a biological basis for treatment response. Many patients in clinical trials who receive only medication achieve full remission in a few weeks. It is widely accepted that there is a biological substrate to depression and that interventions at
the level of that substrate can lead to remission. A key question in the field of depression research is: What are the final therapeutic targets of antidepressants?
As clinicians and patients know, not everybody responds well to each of the more than 20 antidepressants approved
by the Food and Drug Administration. Approximately 60% to 70% of patients respond to any specific antidepressant. Therefore, there is a need to develop new treatments for the substantial minority of patients who are currently labeled as “treatment-resistant” or “refractory.” The greatest challenge to the development of new drugs is the identification of new therapeutic targets. For example, the reason fluoxetine was such a blockbuster was because it was the first selective
serotonin reuptake inhibitor (SSRI) on the market. Potential new targets for depression treatment include neuropeptides. They have been thought to have a role in the biology of depression, and consequently, there has been an effort to develop drugs that modify neuropeptidergic function. Drugs acting at the level of the receptor for the neuropeptides
corticotropin-releasing hormone (CRH) and substance P have been used experimentally in the treatment of depression with success, but have not made their way to the market yet.10,11 What other targets are there for depression? The simple answer is that we do not know.
It is obvious that antidepressants act on the brain to affect the substrate of depressive symptoms. Their initial and acute targets are central monoaminergic systems. Drugs that at least initially affect one or more of three monoamines (serotonin, norepinephrine, dopamine)12,13 can be fully effective in treating depression. The effects of antidepressants on one or more of those amines takes place within hours, yet the clinical response to antidepressants takes weeks to occur. Therefore, some targets that are still unknown,
and common to various classes of drugs, are being activated and causing the antidepressant effect. Many decades of psychopharmacology research have failed to elucidate the identity of those targets that, once activated, can lead to remission of depressive symptoms. Genomics opens up new doors to such efforts. The use of advanced molecular biology techniques coupled with our enhanced understanding of genomic sequences will facilitate the search for genom
ic targets of antidepressant drugs. As an example, our laboratory has used techniques such as differential messenger mRNA display and DNA microarrays (“DNA chips”) to identify new transcripts that are expressed in the brain in response to chronic treatment with both fluoxetine and impramine (Figure 1). Such transcripts may be involved in the pathways that are modulated by antidepressants as they exert their therapeutic effects. We are currently validating those findings by a variety of independent methods and characterizing those genes. Once these genes are fully characterized they will
be natural new targets for antidepressant drug development.
Pharmacogenomics of Depression:
In any area of medicine, including psychiatry, the outcome of treatment is never certain. All of us who practice
medicine work with patients who respond in the most diverse ways to the same
treatment. After receiving the same dose of the same medication, some patients have no response whatsoever, others have only severe side effects, while others can experience complete remission of their symptoms. In some cases, it is possible to predict drug response based on the patient’s personal or family history of treatment response. Physicians also try to balance the patient’s health status and possible drug interactions with the side-effect profile of various antidepressants.
Current work conducted by our group and others on the clinical pharmacogenomics of depression is aimed at using genetic markers to identify predictors of treatment response. For this to occur, it is necessary that rigorous clinical research be developed to thoroughly examine the relation between genotype and the phenotype of drug response. This way, in well-conducted clinical trials, favorable clinical responses and adverse events are related to specific genetic polymorphisms. The goal of such work is to identify markers associated with treatment responses. This line of investigation offers enormous promise for the individualization of treatment. We would all benefit from knowing the likelihood of favorable responses or adverse reactions before taking a drug. In spite of enormous promise, a variety of problems emerge in the conduction of such work. These include clinical factors such the confounding variable presented by the placebo response, the issues of sample size, patients’ genetic background, ethnic stratification, use of continual versus categorical outcome measures, choice of drugs and treatment strategy, treatment compliance, and environmental contributions to treatment outcome.6
The challenges presented by the genetic approaches to such studies are also considerable and have been discussed elsewhere.6 A key dilemma is the choice of which genetic polymorphisms to choose for associations with drug responses, and how to statistically ascertain small effects in
the context of multiple comparisons of variables that have a varying (but not fully characterized) degree of partial relationship to one another.
Serotonin Transporter Gene
A natural candidate gene for such studies is the one encoding the serotonin transporter. After serotonin is released in the synaptic cleft, the serotonin transporter (5-HTT) brings serotonin back into the presynaptic neuron, thereby decreasing the amount of bioavailable serotonin in the synapsis. Blocking the 5-HTT by drugs such the SSRIs leads to increased synaptic concentrations of serotonin.14 The 5-HTT displays a polymorphism in its regulatory region, the presence or absence of a 44 base-pair insertion. The short variant of the polymorphism reduces the transcriptional efficiency of the 5-HTT gene promoter, resulting in decreased 5-HTT expression and 5-HTT uptake
The group at the San Raffaele Hospital in Milano has showed that patients with the “long” form of the 5-HTT regulatory region show a higher response to fluvoxamine and paroxetine.16,17 Furthermore, a not uncommon problem with the
treatment of bipolar disorder with antidepressants is the precipitation of a manic phase. This risk is higher if the patient has the “short” form of the 5-HTT regulatory region.18 Additionally, the group from Milano also showed that antidepressant response to a nonpharmacologic intervention, namely sleep deprivation, was more likely to occur in patients with the long genotype.19 The short and long forms of the 5-HTT regulatory region can also influence the occurrence of extrapyramidal side effects and akathisia, which can be induced by SSRIs. These side effects are known to affect compliance.20 In Korea, Kim and colleagues21 studied the effects of the long and short forms of the 5-HTT regulatory region and they also examined the association of treatment response to the presence of an insert in the second intron of that gene.21
It is noteworthy that while the group from Milano found that the long (l/l or l/s) form of the 5-HTT gene was associated with better treatment response to an SSRI, the group from Korea found that the short (s/s) form was associated with better treatment response. These results suggest that factors other than a specific genotype may have important roles in determining the effect of gene-drug interactions. The differences between the two studies, including culture, diet, type of medical care, psychosocial support, and genetic background, are so vast that it is not possible to attribute a specific cause for such disparities in the data. These discrepancies raise the important
point that it in order to document replicability, it is essential that genomic and pharmacogenomic studies be conducted in the same manner in at least two different populations.
In conclusion, genetic variations in one target of antidepressant action, the 5-HTT, are associated with treatment response. However, these variations do not fully predict the response to treatment. It is highly likely that no single genetic marker will fully explain the genetic components of antidepressant treatment responses. To achieve full understanding of this topic, a variety of markers in multiple genes that contribute to treatment responses will have to be identified. Current research efforts are aimed at identifying genetic targets of antidepressant action. Polymorphisms in those genes will be natural candidates for future pharmacogenomics studies of
Cytochrome P450 system
A body of research has been conducted on the relation between the cytochrome P450 (CYP 450) superfamily23 and antidepressant response. The CYP 450 is a group of related enzymes located in endoplasmic reticulum. Those enzymes are expressed mainly in liver, and also in the gut and the brain. They use oxygen to transform endogenous (eg, steroids) or exogenous (eg, drugs) substances into more polar products that can be eliminated in
the urine. The electrons are supplied by reduced nicotinamide adenine
dinucleotide phosphate (NADPH) CYP 450 reductase, a flavoprotein that transfers electrons from NADPH to CYP 450.
The CYP superfamily is divided into 14 families and 17 subfamilies of enzymes defined on the basis of similarities in
their amino-acid sequences. The enzymes transforming drugs in humans belong to CYP families 1–4. The antidepressants are extensively metabolized by these enzymes. Consequently, genetic variations that affect enzyme activity will impact on the metabolism of antidepressant drugs and will affect clinical responses to treatment. Among the CYP 450 superfamily, CYP 2D6 has an important role in the metabolism of various antidepressants, as well as other commonly used drugs (Table 1).
The activity of CYP 2D6 is bimodal, some people (6% of Caucasians) have no copy of the gene, while others have gene duplication. One third of Ethiopians have such gene duplication. Overall, the CYP 2D6 cluster has 48 mutations and
50 alleles.24-28 A dramatic case report illustrates the clinical relevance of this
gene cluster. A 9-year-old diagnosed with attention-deficit/hyperactivity disorder, obsessive-compulsive disorder, and Tourette’s disorder was treated with a combination of methylphenidate, clonidine, and fluoxetine. After treatment was initiated, the patient had generalized seizures that evolved to status epilepticus followed by cardiac arrest and death. The medical examiner’s report indicated death caused by fluoxetine toxicity. At autopsy, blood, brain, and other tissue concentrations of fluoxetine and norfluoxetine were several-fold higher than expected based on literature reports for overdose situations. This led authorities to charge the parents with murder and prompted juvenile authorities to take away their other two children pending the outcome of a homicide investigation. Subsequent testing of autopsy tissue revealed the presence of a gene defect at the CYP 450 CYP 2D locus, which is known to result in poor metabolism of fluoxetine.29 Criminal charges to the parents were then dismissed. The fact that the population frequency of such clinically relevant mutant alleles and duplicated genes is dependent on ethnicity raises critical ethical considerations.
The frequency of polymorphisms in genes that are relevant to antidepressant treatment response may vary among ethnic groups. It is therefore important to consider ethnicity in clinical research in this area. Our work on the clinical pharmacogenomics of antidepressants in the Los Angeles Mexican-American population has brought to our attention a number of important issues and considerations that emerge when studying an ethnically-
identified group. First, we are often asked about the rationale for studying a specific group. Our rationale has been that because we will examine polymorphisms in genes associated with drug response, including CYP 450 genes, we want to avoid in our study factors other than drug response that could be responsible for variations in the frequency of polymorphisms. One such factor would be ethnic stratification.
Ethnic stratification is particularly important in Los Angeles, which is said to be the world’s most ethnically diverse metropolitan area. For example, it is possible that if we study a general population we may identify a genetic polymorphism X that could be 15% more prevalent in the antidepressant responder group than in nonresponders.
Let us imagine that the responder group is predominantly of ethnicity A and the nonresponder group is mostly of ethnicity B. If there is ethnically-related variation in the frequency of polymorphism X, it would be very hard to
determine if the variation in the rates of polymorphism X between the two groups is due to their differences in drug response profiles or to their different ethnic composition. To avoid this type of confounding scenario, we opted to study one population group, so that all patients (responders and nonresponders) would belong to the same group, and therefore would not stratify divergently during the course of the study. Such measures do not fully eliminate genetic stratification as we are not studying an ethnic homogeneous population. Different individuals of Mexican origin have varying rates of European, Native-American, and, to a smaller extent, African backgrounds. Nevertheless, even though Mexican-Americans are not a homogenous group, they are certainly less heterogeneous than the population of Los Angeles at large.
During the course of this study, several topics emerged. There are sensitive issues when one recruits patients not only because they have a disease, but also because of who they are ethnically.30 The potential for stigma is ever present. According to the Census Bureau, Mexican-Americans are the most rapidly-growing minority group in the US. Today, 1 in every 14 Americans is Mexican-American. This group is among the poorest and least educated in the nation. What will happen if we find that this population is far less responsive to a common treatment than the general population? Could our data become an obstacle to obtaining health insurance for a group that already has considerable difficulty gaining access to adequate medical care? We have developed strategies to consult the community in order to inform them about the research and obtain their impressions. Such a complex and logistically-challenging process involves the following considerations:
(A) Whom does one consult?
(B) In a large county that has 3 million Mexican-Americans, where does one go to develop such a process of community consultation?
(C) How does one deal with dissent? For example, what should the investigator do if some members
of the community are highly supportive of the research effort while others are against it?
(D) What if the research results can potentially worsen existing stigma?
(E) How does one motivate community members and community leaders to participate in a process that will be of no immediate benefit to them?
(F) What is the cost of such consultation and who pays for it?
In general, we have found that multiple meetings in different locations are better than one large and highly engineered meeting. The creation of a community advisory group can facilitate an ongoing dialogue with the community and help in dealing with conflicting views. Research results should be reported in a sensitive and careful manner to avoid ethnic stereotypes and stigma. As the study of ethnically identified groups evolves, strategies to consult with communities will be refined and become more widespread.
Additional Ethical Considerations
There are multiple other ethical considerations in pharmacogenomics. A detailed review of this topic by Robertson discusses issues of confidentiality (which is always a problem in any type of genetics test) and labeling patients as “nonresponders.”31 Such a label could affect the patient’s perception of self, future medical care, and ability to obtain insurance or employment. An important possible complication of drugs that are tested and approved for people with specific genetic markers is the issue of what to do with those who do not have the markers that are associated with favorable outcome. For example, if a new antidepressant is approved for patients with specific genetic markers, it might be difficult to use that drug in individuals who do not have that genotype. A chronically depressed patient who is refractory to existing treatment may not have those genetic markers. Her physician may still think that the new drug is the best option for that patient, despite the risk of adverse reactions or low efficacy. Will such
“off-label” use be covered by the patient’s insurance? In that scenario, what would the physician’s liability be if the patient experiences severe adverse reactions?
Pharmacogenomics, a new area of investigation that integrates genomics and therapeutics, has much to offer the field of psychiatry. While the promise of individualized therapeutics is considerable, the obstacles cannot be overlooked. Those include clinical, technical, and ethical issues that are only now being fully addressed. For all groups to benefit from progress in pharmacogenomics, it is crucial to include members of various ethnically identified groups in such studies. However, as we do that, additional ethical issues emerge. Future research will determine whether the concepts of race or ethnicity are relevant to pharmacogenomics. Even if they are not, it is necessary to conduct studies to achieve that conclusion. Those studies are themselves fraught with ethical issues. Careful consideration of the
interplay of genomics, psychiatry, and ethics should guide a conscientious effort to advance the science of pharmacogenomics in a manner that maximizes the translation of scientific advances into better health car
e for all, while avoiding stigma and stereotypes.
Acknowledgment: Dr. Licinio received grants GM61394, DK58851, and HL04526 from the National Institute of Health (NIH), and a monetary award from the Dana Foundation. Dr. Hannestad is supported by a UCLA Norman Cousins Center PNI fellowship. Dr. Wong received NIH grants MH/NS62777, GM61394, HL04526, and AT00151 and a monetary award from the National Alliance for Research on Schizophrenia and Depression. �
1 Stipp D. A DNA tragedy. Fortune. 2000;142:170-180.
2. Nestler EJ. Psychogenomics: opportunities for understanding addiction.
J Neurosci. 2001;21:8324-8327.
3. Weissman MM, Bland RC, Canino GJ, et al. Cross-national epidemiology of major depression and bipolar disorder. JAMA. 1996;276:293-299.
4. Greenberg PE, Stiglin LE, Finkelstein SN, Berndt ER. The economic burden of depression in 1990. J Clin Psychiatry. 1993;54:405-418.
5. Murray CJL, Lopez AD. The Global Burden of Disease: A Comprehensive Assessment of Mortality and Disability From Diseases, Injuries, and Risk Factors in 1990 and Projected to 2020. Cambridge, MA: Harvard University Press; 1996.
6. Wong M-L, Licinio J. Research and treatment approaches to depression. Nature Rev Neurosci. 2001;2:343-351.
7. Agid O, Shapira B, Zislin J, et al. Environment and vulnerability to major psychiatric illness: a case control study of early parental loss in major depression, bipolar disorder and schizophrenia. Mol Psychiatry. 1999;4:163-172.
8. Weissman MM, Markowitz J, Klerman GL. Comprehensive Guide to Interpersonal Psychotherapy. New York, NY: Basic Books; 2000.
9. Beck JS. Cognitive Therapy: Basics and Beyond. New York, NY: Guilford
10. Kramer MS, Cutler N, Feighner J, et al. Distinct mechanism for
antidepressant activity by blockade of central substance P receptors. Science. 1998;281:1640-1645.
11. Zobel AW, Nickel T, Kunzel HE, et al. Effects of the high-affinity corticotropin-releasing hormone receptor 1 antagonist R121919 in major
depression: the first 20 patients treated. J Psychiatr Res. 2000;34:171-181.
12. Charney DS. Monoamine dysfunction and the pathophysiology and treatment of depression. J Clin Psychiatry. 1998;59(suppl 14):11-14.
13. Lecrubier Y, Boyer P, Turjanski S, Rein W. Amisulpride versus imipramine
and placebo in dysthymia and major depression. Amisulpride Study Group.
J Affect Disord. 1997;43:95-103.
14. Iversen L. Neurotransmitter transporters: fruitful targets for CNS drug discovery. Mol Psychiatry. 2000;5:357-362.
15. Lesch KP, Bengel D, Heils A, et al. Association of anxiety-related traits with a polymorphism in the serotonin transporter gene regulatory region. Science. 1996;274:1527-1531.
16. Smeraldi E, Zanardi R, Benedetti F, Di Bella D, Perez J, Catalano M. Polymorphism within the promoter of the serotonin transporter gene and antidepressant efficacy of fluvoxamine. Mol Psychiatry. 1998;3:508-511.
17. Zanardi R, Benedetti F, Di Bella D, Catalano M, Smeraldi E. Efficacy of
paroxetine in depression is influenced by a functional polymorphism within the promoter of the serotonin transporter gene. J Clin Psychopharmacol. 2000;20:105-107.
18. Mundo E, Walker M, Cate T, Macciardi F, Kennedy JL. The role of serotonin transporter protein gene in antidepressant-induced mania in bipolar disorder: preliminary findings. Arch Gen Psychiatry. 2001;58:539-544.
19. Benedetti F, Serretti A, Colombo C, et al. Influence of a functional
polymorphism within the promoter of the serotonin transporter gene on
the effects of total sleep deprivation in bipolar depression. Am J Psychiatry. 1999;156:1450-1452.
20. Lane RM. SSRI-induced extrapyramidal side-effects and akathisia: implications for treatment. J Psychopharmacol. 1998;12:192-214.
21. Kim DK, Lim SW, Lee S, et al. Serotonin transporter gene polymorphism and antidepressant response. Neuroreport. 2000;11:215-219.
22. Licinio J, Wong M-L. The pharmacogenomics of depression. Pharmacogenomics J. 2001;1:175-177.
23. Raucy JL, Allen SW. Recent advances in P450 research. Pharmacogenomics J. 2001;1:178-186.
24. Bertilsson L, Aberg-Wistedt A, Gustafsson LL, Nordin C. Extremely rapid hydroxylation of debrisoquine: a case report with implication for treatment with nortriptyline and other tricyclic antidepressants. Ther Drug Monit. 1985;7:478-480.
25. Johansson I, Oscarson M, Yue QY, Bertilsson L, Sjoqvist F, Ingelman-Sundberg M. Genetic analysis of the Chinese cytochrome P4502D locus: characterization of variant CYP2D6 genes present in subjects with diminished capacity for debrisoquine hydroxylation. Mol Pharmacol. 1994;46:452-459.
26. Bertilsson L, Dahl ML, Tybring G. Pharmacogenetics of antidepressants:
clinical aspects. Acta Psychiatr Scand. 1997;391(suppl):14-21.
27. Bertilsson L, Lou YQ, Du YL, et al. Pronounced differences between native Chinese and Swedish populations in the polymorphic hydroxylations of debrisoquin and S-mephenytoin. Clin Pharmacol Ther. 1992;51:388-397.
28. Aklillu E, Persson I, Bertilsson L, et al. Frequent distribution of ultrarapid metabolizers of debrisoquine in an ethiopian population carrying duplicated and multiduplicated functional CYP2D6 alleles. J Pharmacol Exp Ther. 1996;278:441-446.
29. Sallee FR, DeVane CL, Ferrell RE. Fluoxetine-related death in a child with cytochrome P-450 2D6 genetic deficiency. J Child Adolesc Psychopharmacol. 2000;10:27-34.
30. Weijer C, Emanuel EJ. Ethics. Protecting communities in biomedical research. Science. 2000;289:1142-1144.
31. Robertson JA. Consent and privacy in pharmacogenetic testing. Nat Genet. 2001;28:207-209.