1. Drug Therapy in Psychiatric Disorders

Treatment of psychiatric disorders is mainly based on drug therapy. A wide range of classical and new drugs is available for their treatment: antidepressants [including tricyclic compounds, selective serotonin reuptake inhibitors (SSRIs), serotonin-nor-adrenaline inhibitors (SNRIs), and monoamine-oxidase inhibitors (MAOIs)] and antipsychotic drugs (dopaminergic blockers and multitarget drugs with strong affinity for serotonergic receptors) are widely used. However, treatment failure is poorly understood and medication is selected on a trial and error basis. An average 30 to 40% of treated patients show no improvement after treatment with antipsychotics and 60 to 70% of those who respond present severe adverse effects.[1,2] The newer range of anti-psychotic drugs minimise some of the most severe adverse effects. Their superiority in symptom improvement, however, is still under discussion, with the exception of the potent antipsychotic clozapine.[3] Therefore, the growing interest in understanding the mechanisms determining psychiatric treatment response, and in particular of clozapine, is well justified. The purpose of this article is to review the latest advances in discerning the genetic influence on the mechanism of action of psychiatric drugs.

In spite of the lack of epidemiologic studies illustrating their exact contribution to clinical improvement, genes are hypothesized to play an important role in the variability observed in response to psychiatric treatment. The first studies in the field focused on single gene mutations related to toxic responses.[4]

Metabolic mutations directly responsible for altered drug transformation have been identified in cytochrome P450 (CYP) enzymes and are well documented, as discussed in section 3. Genetic variants in metabolic enzymes can cause accumulation or rapid elimination of a drug, leading to a toxic reaction or to lack of effect at therapeutic doses.[5–7] However, the variability observed in response to psychiatric treatment cannot be explained by variation in single genes alone. Furthermore, the therapeutic efficacy of psychiatric drugs may also be affected by alterations in target sites. Recent studies linking mutations in targeted neurotransmitter receptors with clinical outcome suggest that response to psychiatric treatment is determined by a combination of variants in several genes, including metabolic enzymes and/or drug targeted receptors.[8,9] In this review we aim to summarize the existing evidence supporting this hypothesis and give an overview of pharmacogenetic studies on psychiatric drugs.

2. Identifying Genes Related to Treatment Response

Before determining the entire combination of genes which may influence the therapeutic efficacy of a particular drug, studies are performed to single out which individual genes may exhibit DNA alterations affecting response. A variable number of genes may be implicated in determining response, some having a major role and others coding for proteins with minor contributions to response. The preferred strategy to search for such genes is case-control (or nonresponder/responder) association studies.[10] This strategy is more sensitive for the detection of genes with a minor influence in a particular phenotype when compared to other possible strategies such as linkage studies. Linkage studies are more powerful for the detection of genes with a major effect but very unlikely to find genes of weak effect. In the last decade a vast number of association studies have been performed with the goal of correlating polymorphisms in candidate genes with individual’s response to treatment.[8] The candidate genes are first screened for mutations using a variety of time-consuming techniques such as single strand conformation polymorphism (SSCP),[11] denaturing high performance liquid chromatography (dHPLC),[12,13] and mismatch cleaving, followed by characterization of DNA change by sequencing.[14,15] This process has been greatly accelerated in recent years with the development of high-throughput automated sequencing and DNA arrays which identify and characterize mutations directly.[16–18] These techniques are generally used for the identification of genes related to particular traits. In pharmacogenetics, the identification of novel polymorphisms in candidate genes is followed by association studies where the distribution of genotype and allele frequencies are compared between patients responding versus those not responding to a particular drug. Any significant difference in the distribution of frequencies between the 2 groups is considered an indication of involvement in the mechanism controlling treatment response.

One of the main limitations of these studies is their susceptibility to producing type I and II errors (false positives or negatives) if the samples are small or not balanced (demographically and/or in terms of patient numbers). In addition, the complexity of the diagnosis and assessment of psychiatric disorders adds to the difficulties in replicating positive associations. Often large interstudy differences are observed, depending on the criteria and assessment tools used to characterise the sample patients, their illness status and their treatment response. Comparison of extreme phenotypes (very good response versus total lack of improvement),[19,20] standardization of response assessment criteria,[10] and collaborative studies involving large and balanced clinical samples will help resolve these difficulties. In spite of these problems, association studies remain useful for the discovery of treatment-related genes.

3. The Influence of Metabolic Enzymes on Response

The first steps in the metabolism of psychiatric drugs are performed by CYP enzymes.[21] Genetic variations affecting the metabolic rate of these enzymes can have direct implications for the biotransformation of drugs.[22] Mutations in the CYP2D6, CYP2C9 and CYP2C19 genes have already been shown to cause altered drug metabolism[7] (a full list of polymorphisms in metabolic enzymes with updated nomenclature can be found in http://www.imm.ki.se/CYPalleles [23]). Alterations disrupting the normal functioning of these enzymes have a direct impact in the assimilation of drugs, which can cause severe adverse reactions (see table I).[9,24] Individuals carrying these mutations may show toxic accumulation (poor metabolisers, PM) or rapid elimination (ultrarapid metabolisers, UM) of various drugs. The frequencies of these mutations vary among populations, being relatively more common in Caucasians [with 7% of individuals showing CYP2D6 (PM) deficient phenotype] than in Asians (1% for the same deficiency).[36,37]

Table I
figure Tab1

List of common mutations with functional effects in CYP metabolic enzymes

Full size image

Many psychiatric drugs are metabolized by CYP enzymes (see table II). Most drugs can be metabolized by more than one enzyme, which may dilute the effect of a deficient (PM) CYP subtype. The effect of the rarer UM enzyme variants is unlikely to be compensated for by other enzymes. Those drugs metabolized by only one specific CYP enzyme are at higher risk of producing adverse reactions in patients carrying the deficient variants. In addition, treatment with two or more drugs metabolized by the same subtypes may result in interactions such as competition for the enzyme and decreased rates of metabolism.[6,46,40] Important interactions have been reported between SSRIs and other psychiatric drugs due to the effects of CYP450 enzymes on their metabolism.[51,52] Recent studies have shown that fluvoxamine inhibits clozapine N-demethylation and causes a pronounced increase in clozapine plasma concentration.[53] However, several studies have failed to find clear association between drug metabolism and therapeutic response.[54–56] More importantly, CYP2D6 variants have been related to abnormal movements observed during neuroleptic treatment,[57] and CYP2D6 and CYP1A2 alleles have been associated with drug induced tardive dyskinesia in schizophrenic patients.[58,59]

Table II
figure Tab2

Commonly used antipsychotic and antidepressant drugs and their metabolic enzymes

Full size image

In view of the adverse reactions caused by deficient drug metabolism, previous knowledge of a patient’s metabolic status and drug pharmacokinetics are vital to avoid such responses. There are a variety of methods for the rapid detection of common metabolic mutations including multiplex PCR,[60,61] allele-specific amplification coupled with allele-specific fluorogenic 5t? nuclease chain reaction (Taqman®) technology[62] and DNA-chip arrays for the identification of CYPs polymorphisms in large samples that have been devised by several companies.[63,64] Pre-treatment characterization of metabolic status will allow the adjustment of therapeutic doses or the selection of alternative medication according to the patient’s metabolic rate. The expected result will be a reduction of drug interactions and toxic responses caused by deficient enzymes.

4. Candidate Genes in Target Sites: The Usual Suspects

Aside from the metabolic enzymes responsible for the biotransformation of drugs, therapeutic activity can be affected by alterations on the site of action. In recent years, pharmacogenetic research has investigated genetic variants of receptors targeted by antipsychotic and antidepressant drugs. The candidate genes for these studies are the usual suspects: genes coding for neurotransmitter systems known to be altered in psychiatric disorders, such as dopaminergic and serotonergic receptors, which are strongly targeted by psychiatric drugs. In the last decade numerous association studies have tried to correlate genetic variants in drug targeted receptors with antipsychotic and antidepressant efficacy. In particular, the genes coding for dopaminergic and serotonergic receptors have been thoroughly analyzed for mutations, and the numerous variants detected prompted extensive investigation of their role in psychiatric disorders and psychiatric treatment.

4.1 Dopaminergic Receptors

Classical antipsychotic drugs used for the treatment of schizophrenia, the most severe psychiatric disorder, strongly target dopamine D2 receptors in mesolimbic and striatal areas of the brain. Their action on mesolimbic D2 receptors is directly related with antipsychotic efficacy. However, their affinity for striatal D2 receptors is also associated with production of severe extrapyramidal adverse effects (EPS).[2]

Association studies investigating polymorphisms in dopaminergic receptors have thus far produced inconclusive results. Reports of association between D2 polymorphisms and antipsychotic response[65,66] contradict previous studies which failed to find any association.[67,68] Association between a D3 structural polymorphism, Ser9Gly, andresponse[69–71] was contradicted by Malhotra and collaborators.[72] However, reports of association between this polymorphism and tardive dyskenisia[73–75] suggest that the D3 receptor may mediate improvement of specific symptoms. Reports of D4 variants associated with antipsychotic response[76,77] seemed to confirm their potential as therapeutic targets. However, several studies failed to confirm this finding.[78–81]

4.2 Serotonergic Receptors

Newer antipsychotics, termed atypical for their lack of EPS production, are multitarget drugs with strong affinities for serotonin (5-hydroxytryptamine; 5-HT) receptors, among others. Although it is not clear which receptor(s) mediates antipsychotic activity, it is thought that the 5-HT2/D2 affinity ratio is an important determinant of therapeutic success.[82,83] Antidepressant therapy includes drugs which also target genes coding for proteins of the serotonergic system (SSRIs and SNRIs). It is clear that serotonergic receptors are major therapeutic targets for psychiatric drugs, and therefore alterations in genes coding for these receptors may play a major role in the interindividual differences observed during psychiatric treatment.

The first report of association between a 5-HT2A variant, 102- T/C, and response to the atypical antipsychotic clozapine[84] was contradicted by several studies which failed to observe significant association, although this polymorphism was found to be associated with neuroleptic response.[85] However, a meta-analysis of 6 studies that had published data on clozapine response showed a clear trend of association between this polymorphism and clozapine response.[19] The same meta-analysis included studies investigating the relation between a 5-HT2A His452Tyr polymorphism[86,87] and clozapine response, and confirmed the correlation between genetic variants of the receptor and clinical outcome. These results constitute the strongest evidence of association between receptor variants and treatment response, suggesting an important role of the 5-HT2A receptors in mediating the therapeutic activity of clozapine. The reported association between a 5-HT2C structural polymorphism, Cys23Ser, and clozapine response[88] was not confirmed by other investigators.[89,90] However, a later metaanalysis[91] and a finding of association between a 5-HT2C polymorphic repeat (-330-GT/-244-CT) and clozapine response[92] seem to support involvement of this receptor in mediating therapeutic activity. Although no association was found between 5-HT5A and antipsychotic response,[93] a report of association between a silent 5-HT6 polymorphism (267-C/T) by Yu and collaborators[94] suggests that other 5-HT receptors may be involved in determining response.

4.3 Neurotransmitter Transporters

Genes coding for transporters regulating neurotransmitter levels have also been investigated. Their regulatory action on neurotransmitter systems makes them strong candidates in the etiology and treatment of psychiatric disorders. A positive finding of association between a polymorphism in the serotonin transporter (5-HTTLPR) and fluvoxamine[95] was followed by studies showing minor contribution of 5-HTT alleles to antipsychotic response.[92,96] Alleles of the dopamine transporter gene (DAT1) were observed to be associated with behavioural response to methylphenidate, a drug used in children for the treatment of attention-deficit hyperactivity disorder (ADHD).[97]

4.4 Other Candidate Genes

Other neurotransmitter receptors including histaminergic, muscarinic and adrenergic receptors have also been investigated. However, several investigators have failed to detect strong associations between these receptors and response.[98–100] Further investigation of genetic variation in these genes is needed to confirm their role in psychiatric disorders. In addition, recent reports of association have related mutations in genes encoding the G-protein β3 subunit (G- β3), methylenetetrahydrofolate reductase (MTHFR), and in the brain-derived-neurotrophic-factor (BDNF) with treatment response (see table III).[101–103]

Table III
figure Tab3

Significant associations between genetic polymorphisms in neurotransmitter receptors and antipsychotic/antidepressant response

Full size image

In summary, these studies show that mutations in genes targeted by psychiatric drugs may influence therapeutic success. However, one of the problems of association studies is the difficulty in replicating results by different groups. Most of the positive association findings reported here have been contradicted by negative reports. The causes for these discrepancies could be several, including chance findings, sample size, sample ethnic origin, and clinical characterization.[104] Nevertheless, replication by independent investigators is vital to confirm the reliability of the findings.

5. Individualization of Treatment: Can Response be Predicted by Looking at Genes?

The overall picture given by association studies is that there is no major gene controlling psychiatric response. Treatment response is controlled by a variety of genes, in agreement with the multitarget profile of most psychiatric drugs. Variants in metabolic genes may influence the development of toxic reactions and adverse effects; reports of association between polymorphisms in CYP enzymes and drug induced adverse effects[58,59] confirm this hypothesis. Contrary to expectations, no strong association has yet been observed between polymorphisms in dopamine receptors and response to dopamine blockers. However, new polymorphisms detected in the promoter regions of D3 and D4 receptor genes[105,106] with a potential effect on receptor expression should be investigated. The strongest associations reported in section 4 referred to 5-HT2 polymorphisms and clozapine response, in agreement with the high affinity of the drug for those receptors. However, minor contributions from other receptor variants are not discounted as the variability observed in response to clozapine treatment could not be explained by 5-HT2 polymorphisms alone. In the case of multitarget drugs such as clozapine, clinical response is likely to be mediated primarily through one or two systems, with other receptor systems having minor effects.

In spite of the important information provided by individual association studies, these studies are of limited use for the prediction of treatment outcome. A combination of variations in key genes is more likely to determine an individual’s therapeutic response. We investigated 33 polymorphisms in 19 genes including metabolic enzymes, neurotransmitter transporters and receptors targeted by clozapine in an effort to identify a combination of polymorphisms that influence both positive and negative response to the drug. We found that a combination of six polymorphisms in genes coding for the 5-HT2A, 5-HT2C, histamine H2 receptors and the 5-HTT predicted clozapine response with nearly 80% success in a sample of patients with schizophrenia.[92] Although we are hoping to improve the prediction levels with the incorporation of newly detected polymorphisms, this is the first evidence suggesting that response to psychiatric treatment can be predicted by looking at an individual’s genes, and opens the possibility for future personalization of antipsychotic therapy. We are currently performing similar studies investigating genetic combinations that would predict response to a variety of psychiatric drugs.

Although still not a diagnostic tool, studies like this will facilitate the early selection of beneficial treatments at the right doses for each individual according to their pharmacogenetic propensities. Pre-treatment genotyping of key polymorphisms will indicate which drugs are more likely to be well metabolized and have a therapeutic effect. This will have a direct impact in clinical therapy by reducing adverse reactions and increasing patient compliance and clinical improvement.

6. Psychiatric Treatment in the 21st Century

Finding drug-response related polymorphisms and analysing their influence is a time consuming process which has been greatly accelerated by high-throughput techniques. Automated sequencing, DNA chips and microarrays are being used for the rapid detection of mutations in the human genome and to investigate their effect.[63,107] As a result, single nucleotide polymorphism (SNP) databanks have already been created for use in medical research.[108] Several pharmaceutical companies have already entered the race and are producing kits for the genetic detection of disease susceptibility and selection of appropriate drugs for their treatment.[63,64] It is hoped that current research will lead to the development of kits for the individualization of psychiatric treatment.

7. Conclusion

Pharmacogenetic objectives in psychiatric research have changed dramatically in the last decade. Genetic association studies have provided information on the contribution of gene mutations to interindividual variability in treatment response. Polymorphisms in metabolic enzymes and in neurotransmitter systems may influence the therapeutic efficacy of psychiatric drugs. Mutations in genes encoding metabolic CYP450 enzymes, neurotransmitter receptors and neurotransmitter transporters have been singled out as responsible for interindividual variation in response to psychopharmacotherapy. With increasing knowledge of the human genome and rapid discovery of mutations, studies of single gene mutations are being replaced by investigations of polymorphism combinations that will ultimately lead to pre-treatment prediction of psychopharmacotherapeutic response. In spite of the difficulties in diagnosis and assessment of psychiatric disorders, pharmacogenetic research has produced the first successful studies correlating genetic profile with antipsychotic response. Similar studies investigating genetic determinants of response to treatment with a variety of psychiatric drugs are currently underway. These studies will ultimately allow for pre-treatment selection of psychiatric drugs most likely to have a positive effect according to the individual patient’s pharmacogenetic profile. It is not too optimistic to think that in the next decade psychiatric drugs could be chosen on the basis of individual’s genetic background.