CYP2D6 variation, behaviour and psychopathology: implications for pharmacogenomics-guided clinical trials

Abstract

Individual and population differences in polymorphic cytochrome P450 enzyme function have been known for decades. The biological significance of these differences has now been deciphered with regard to drug metabolism, action and toxicity as well as disposition of endogenous substrates, including neuroactive compounds. While the cytochrome P450 enzymes occur abundantly in the liver, they are expressed in most tissues of the body, albeit in varying amounts, including the brain. The latter location of cytochrome P450s is highly pertinent for susceptibility to neuropsychiatric diseases, not to mention local drug metabolism at the site of psychotropic drug action in the brain. In the current era of personality medicine with companion theranostics (i.e. the fusion of therapeutics with diagnostics), this article underscores that such versatile biological roles of cytochrome P450s offer multiple points of entry for personalized medicine and rational therapeutics. We focus our discussion on CYP2D6, one of the most intensively researched drug and endogenous compound metabolism pathways, with a view to relevance for, and optimization of, pharmacogenomic-guided clinical trials. Working on the premise that CYP2D6 is related to human behaviour and certain personality traits such as serotonin and dopamine system function, we further suggest that the motivation of healthy volunteers to participate in clinical trials may in part be influenced by an under-or over-representation of certain CYP2D6 metabolic groups.

CYP2D6 genetic polymorphisms and neuroactive endogenous compound metabolism

Among cytochrome P450s, CYP2D6 is one of the most intensively researched and clinically important enzymes involved in the metabolism of a large number of widely used central nervous system (CNS) drugs. CYP2D6 is highly polymorphic. Multiple allelic variants of the CYP2D6 gene have been identified, which are associated with an absent or increased enzyme activity in individuals who are respectively so called poor (PMs) and extensive metabolizers (EMs), including in the latter group a subgroup of ultrarapid metabolizers (UMs). This polymorphic enzyme is involved in the metabolism of many drugs of relevance to psychiatry, neurology and addiction medicine such as antidepressants, antipsychotics and opioids 1,2. Moreover, CYP2D6 has also been shown to contribute to endogenous metabolism of neuroactive substrates, which can explain the associations hitherto observed with human behaviour and disease susceptibility (e.g., personality, neurocognition and neuropsychiatric disorders) 3,4

CYP2D6 was identified in both animal and human brain tissues in the late 1980s 5. More than a decade later, dextromethorphan to dextrorphan metabolism was demonstrated in human brain microsomes 6. CYPD6 has been described in neurons in the human cerebral cortex, hippocampus and cerebellum 7, basal ganglia and midbrain 8,9. Other studies have provided additional evidence of CYP2D6 expression in these brain regions and the thalamus 10 as well as in glial cells in those brain areas 11.

In addition to the mapping of CYP2D6 in the brain, this enzyme has also been involved in the metabolism of tyramine to dopamine (DA) in vitro 12,13 and in the regeneration of serotonin (5-HT) from 5-methoxytryptamine 14,15. Interestingly, in vivo studies have shown that UMs display higher serotonin concentrations in platelets than EMs and PMs 16. A potential influence of CYP2D6 polymorphism in the balanced functioning and physiological crosstalk of the DA and 5-HT endogenous systems has been proposed 17,18. This hypothesis was based on the results from another in vivo study suggesting that PMs might have a higher DA tone in the pituitary 19, which might be in combination with lower serotonin tone since the serotonin systems exerts a tonic inhibitory control on the dopaminergic circuits. Recently, it has been proposed that a potential mechanism for the interaction of the serotonin and DA system would be the synthesis of 5-HT in DA neurones 20.

Additionally, CYP2D6 has been implicated in the endogenous metabolism of the ligand for the cannabinoid receptor CB1, anandamide 21. The possibility that CYP2D6 may be involved in the regulation of endogenous neuroactive steroids, such as progesterone and its derivatives in brain tissues has been suggested 22,23. The trace amines betacarbolines, pinolines, harmaline and harmine have been also related 24.

Taken together, the relationships observed between CYP2D6 variation and personality that are to be detailed in the next section could be mediated by the influence of this enzyme activity in the serotonergic/dopaminergic tone plus other neurotransmitters or neuromodulators.

Relationship between CYP2D6 genetic variation and personality

With the aim to clarify existing data, the key studies on the relationship between CYP2D6 and personality are summarized in Tables 1 and 2, and discussed subsequently in the same order. For further clarity, we separated those studies conducted in healthy volunteers from those in patients.

Table 1.

YP2D6 and personality studies

Reference	Population	n	Nationality	CYP2D6 debrisoquine hydroxylation capacity (Phenotype)	CYP2D6 alleles (Genotype)	Personality measures	Main results
Bertilsson et al. 1989 [26]	Healthy volunteers	769	White Swedish	PMs (MR>12,6) and EMs (MR< 12,6)		KSP	PMs scored lower in psychastenia subscale and had a higher frequency of extreme responses than EMs.
LLerena et al. 1993 [25]	Healthy volunteers (students)	225	White Spanish	EM1 (MR≤0.30), EM2 (0.30 < MR≥0,87); EM3 (0,87 < MR≤12,6) and PM (MR>12,6)		KSP	PMs scored lower in socialization measures, and higher in psychic anxiety, somatic anxiety, psychastenia, and inhibition of aggression.
González et al. 2008 [28]	Healthy volunteers (students)	246	Mestizo Cuban	EM1 (MR≤0.28), EM2 (0.28 < MR≥0,77); EM3 (0,77 < MR≤12,6) and PM (MR>12,6)	CYP2D6 allelic variants with abolished (*3, *4, *4xN, *5 and *6), decreased (*10 and *17) and increased (*1xN and *2xN) activity.	KSP	PMs (phenotype) scored lower in socialization and higher in psychic anxiety, and irritability.
Peñas-LLedó et al., 2009 [32]	Healthy volunteers (students)	144	White Spanish		CYP2D6 allelic variants with abolished (*3, *4, *4xN, *5 and *6), decreased (*10 and *17) and increased (*1xN and *2xN) activity.	KSP	PMs (genotype) presented higher ‘impulsivity’ in both KSP and TCI-R scalesand lower distress (SCL-90-R).
						TCI-R
Kirchheiner et al., 2006 [29]	Healthy volunteers	222	White German		CYP2D6 allelic variants with abolished (*3, *4, *5 and *6) activity and the duplication allele	NEO-Five Factor Inventory	PM (genotype) females showed higher conscientiousness
Suzuki et al., 2003 [30]	Healthy students	255†	Asian Japanese		CYP2D6 *10 (*1/*1 and *1/*10 vs. *10/*10)	TCI	No association with CYP2D6*10 genotype frequencies in any of the personality traits.
							Absence of UMs (genotype) and PMs (genotype).
Iwashima et al., 2007 [31]	Healthy volunteers (students)	342	Asian Japanese		CYP2D6 allelic variants with abolished (*3, *4, *4xN, *5, *6, *7, *8, *11, *14, *15, *19, *20 and *40), decreased (*9, *10, *17, *29, *36 and *41), normal (*2 and *35) and increased (*1xN and *2xN) activity	TCI	No association with CYP2D6.
							Low frequency of UMs (genotype) and PMs (genotype).
Roberts et al., 2004 [36]	Depressed patients	121	White (95% New Zealand)		CYP2D6 *1, *2, *4, *5, *9, *10, *13, *16, and the duplication alleles.	TCI	PMs (genotype) scored lower in harm avoidance (fear of uncertainty, fatigability and shyness)
Gan et al., 2004 [35]	Patients‡	48	Asian Malay (44), Chinese (3) and Indian (1)		EM1 (CYP2D6*1/*1), EM2 (CYP2D6*1/*4,*1/*5, *1/*9 and *1/*10), EM3 (CYP2D6*4/*10, *5/*10, *10/*10 and *10/*17).	Type A and B personality questionnaire	EM2 and EM2 were found to be mainly of personality type B compared with EM1 (predominance of type A personality).
							Absence of UMs (genotype) and PMs (genotype).

EMs, extensive metabolizers; KSP, Karolinska Scales of Personality; n, number of subjects; PMs (genotype), individuals with two or more CYP2D6 allelic variants with abolished activity; TCI, temperament and Character Inventory; UMs (genotype), individuals with more than two active allelic variants; PMs, poor metabolizers; TCI-R, Temperament and Character Inventory-Revised.

^†255 students (222 = females);

^‡Patients were individuals hospitalized for orthopaedic surgery, but otherwise healthy by medical history and physical examination.

Table 2.

Studies about the relationship between CYP2D6 and neurocognition and psychopathology

Reference	Population	Nationality	CYP2D6 alleles (Genotype)	Measures	Main results
Peñas-LLedó et al., 2009 [32]	144 healthy volunteers mostly students	White Spanish	CYP2D6 allelic variants with abolished (*3, *4, *4xN, *5 and *6), decreased (*10 and *17) and increased (*1xN and *2xN) activity.	CANTAB	PMs (genotype) presented better Rapid visual processing or sustained attention as well as lower distress/psychopathology.
				SCL-90-R GSI
Stingl et al., 2011 [50]	113 healthy volunteers	White German	CYP2D6 allelic variants with inactived (*3, *4, *5 and *6), decreased (*9, *10, *17 and *41), fully functional (*2 and *35) activity, and the gene duplication.	Functional brain imaging: a n-back memory task and an implicit emotional face matching task	A significant effect of CYP2D6 genotype was found in the precuneus and the cuneus. In both tasks activation increased with increasing CYP2D6 activity.
LLerena et al., 2007 [40]	128 schizophrenia patients and 142 healthy volunteers	White Spanish	CYP2D6 allelic variants with abolished (*3, *4, *4xN, *5 and *6), decreased (*10 and *17) and increased (*1xN and *2xN) activity.		Frequency of PMs (genotype) was lower in schizophrenic patients than in healthy volunteers.
Peñas-LLedó et al., 2012 [55]	267 patients with ED and 285 healthy volunteers	White Spanish	CYP2D6 allelic variants with abolished (*3, *4, *4xN, *5 and *6), decreased (*10 and *17) and increased (*1xN and *2xN) activity.		Frequency of UMs (genotype) was higher in patients with ED than in healthy volunteers.
Peñas-LLedó et al., 2012 [54]	Healthy Volunteers	White and Mestizo Cubans	CYP2D6 and debrisoquine hydroxylation phenotype.	EDI	Higher frequency of UMs among bulimia scorers >5.
Zackrisson et al., 2010 [57]	242 individuals who died of fatal intoxication, 262 of suicide and 212 of natural death	White Swedish	CYP2D6 allelic variants with abolished (*3, *4, *4xN, *5 and *6) and normal (*2) and increased (*1xN and *2xN) activity.		A higher number of UMs (genotype) among individual who died of suicide than in the other groups.
			CYP2C19 allelic variants with abolished (*2, *3 and *4) activity.
Peñas-LLedó et al., 2011 [56]	203 patients with ED (165 without suicidal behavior and 38 with suicidal behaviour†)	White Spanish	CYP2D6 allelic variants with abolished (*3, *4, *4xN, *5 and *6)and increased (*1xN and *2xN) activity.		Frequency of UMs (genotype) was higher in patients with lifetime suicidal behaviour than in patients without suicidal behaviour.
Peñas-LLedó et al., 2012 [58]	342 suicide attempters and 377 healthy controls.	White Spanish	CYP2D6 allelic variants with abolished (*3, *4, *4xN, *5 and *6), decreased (*10 and *17) and increased (*1xN and *2xN) activity.	Beck-SIS§	A higher number of ‘severe’ suicide attempters carrying two or more than two active CYP2D6 genes as compared with the rest of the patient population or the healthy control group was found.
Blasco-Fontecilla et al., 2013 [59]	342 suicide attempters	White Spanish	CYP2D6 allelic variants with abolished (*3, *4, *4xN, *5 and *6), decreased (*10 and *17) and increased (*1xN and *2xN) activity.	MINI	Suicide attempters with two or more CYP2D6 active genes were more likely to be diagnosed with personality disorders.
				IPDE-SQ
Stingl & Viviani, 2011 [50]	285 depressed inpatients.	White German	CYP2D6*3, *4, *5, *6, *9, *10, and *35 alleles, and the duplication of gene.	MINI‡	In UMs (genotype) the risk of a high suicidality score was elevated as compared to those with other genotypes.
				HAM-D	No association between UM status and basal HAM-D scores at study entry.
Höfer et al., 2013 [74] discussed in Peñas-LLedó et al., 2013 [75]	243 MDD patients	White European¶	Genotyping was performed for all relevant variations of the CYP1A2 gene (*1A, *1F, *1C, *1 J, *1 K), the CYP2C9 gene (*2, *3), the CYP2C19 gene (*2, *17) and the CYP2D6 gene (*3, *4, *5, *6, *9, *19, *XN).	MINI	No association between both suicide risk and personal history of suicide attempts, and the CYP1A2, CYP2C9, CYP2C19 and CYP2D6 metabolic profiles were found.‡‡
				HAM-D††

Beck-SIS, Beck Suicide Intent Scale; ED, eating disorders; IPDE-SQ, International Personality Disorder Examination Screening Questionnaire; KSP, Karolinska Scales of Personality; MDD, major depressive disorder.; MINI, Mini-International Neuropsychiatric Interview; PMs (genotype), individuals with two or more CYP2D6 allelic variant; s with abolished activity; TCI-R, Temperament and Character Inventory-Revised; UMs (genotype), individuals with more than two active allelic variants.

^†A suicide attempt was defined as a self-destructive act with some degree of intent to end one's life. Thus, to be considered an attempt, the attempt was required to have two components, an action that was self-destructive and an acknowledgement of intent to die.

^‡MINI includes six specific items on suicidality.

^§Those individuals scoring above percentile 75 in the objectives circumstances section of the Beck-SIS were classified as ‘severe’ suicide attempters.

^¶Individuals were recruited from Austria, Israel, Belgium, France and Italy.

^††Suicidality was assessed using two items: Presence or absence of current suicidal risk in the MINI item on suicidality (A current suicidal risk was defined by the presence ofat least one of the following suicide-related items: having in the past month thought that it would be better being dead or wishing to die, wanting to harm oneself, thinking about suicide, having a suicide plan, attempting suicide and ever attempted suicide at least once in the lifetime) and the HAM-D item 3 on suicidality (score 0–4).

^‡‡Some aspects of the methodology and results of this study have been discussed by Peñas-LLedó et al., 2013) 75. This paper had limitations related to the inaccuracy and imprecision of the main measures (misclassification of metabolizer status and definition of suicidal phenotypes).

Studies about CYP2D6 and personality in healthy volunteers

The first hypothesis anticipating that CYP2D6 could have an endogenous neuroactive substrate or product such as a biogenic amine was put forward as early as 1993 25. However, a subtle but significant hint about the putative influence of CYP2D6 on personality traits came even earlier from a study in healthy volunteers in Sweden where the participants were phenotyped with debrisoquine and evaluated with Karolinska Scales of Personality (KSP). Curiously, the PMs were found to report lower levels of psychastenia than the rest of the healthy volunteers or EMs 26.

Later in Spain 25 a larger and independent population of healthy volunteers were phenotyped with debrisoquine and evaluated with KSP that had just been translated into Spanish 27. The difference from the previous study was that instead of comparing just the two broad groups of PMs and EMs, four groups were identified: PMs and other three relatively homogeneous groups of EMs separated on the basis of their CYP2D6 debrisoquine hydroxylation capacity. Yet, significant differences were again noted/replicated between PMs and the EM groups in this independent study sample of Spanish subjects, admittedly from a social and environmental context different from that of Sweden, thereby lending further support for linkages between CYP2D6 variation and personality or behavioural traits. PMs were shown to report greater levels of psychic anxiety and lower socialization than EMs. These two seminal observations and studies using KSP personality traits and CYP2D6 phenotype, as well as subsequent research trying to replicate this relationship about personality and CYP2D6 are summarized in Table 1.

Personality differences between CYP2D6 PMs and EMs were later compared in several healthy volunteer populations. Three of the studies, the two previous one in Swedish and Spanish, and a third one in Cuban healthy volunteers used the same phenotyping procedure (debrisoquine test) and the same personality measure (KSP) 25,26,28. These studies were similar with regard to establishing differences in psychastenia between PMs and EMs 25,26,28. Above all, Spaniards and Cubans showed almost identical results for some personality traits 25,28. PMs presented higher psychic anxiety and lower socialization than the three groups of EMs. It is of note that these two studies (in Spain and Cuba) shared also the same recruiting procedure and type of volunteers (mostly university students) as well as cultural background. On the contrary, the Swedish population was older, with different education, and recruitment procedures since they were participants in clinical trials. Differences across studies may have highly influenced the different pattern of associations observed between Swedes and Latinos.

Posterior studies in German 29, Japanese 30,31 as well as in another group of Spanish volunteers 32 used different methods for determining CYP2D6 activity, which was assumed by genotypes (Table 1). They also used different personality measures, mostly the Temperament and Character Inventory (TCI), making it almost impossible to establish comparisons. Additional differences were related to recruitment methods of the German volunteers since they were a heterogeneous population from the community that responded to a newspaper advertisement 29. Furthermore, this study used a new measure, the NEO Five-Factor Inventory (NEO-FFI) 33, finding differences between PM and EM females. PMs reported higher ‘conscientiousness’ or responsibility, orderliness, and the pursuit for achievement through perseverance. This NEO-FFI trait has been consistently associated with hard working and reliable individuals, who demonstrate excellence in the workplace 34. Thus, this result of high ‘conscientiousness’ can be interpreted in the light of previous ones signalling higher anxiety among healthy PMs, which may indicate that anxiety levels help them to increase performance instead of being related to anxiety disorders and the pursuit of unrealistic standards.

In Japan, two other studies 30,31 also determined only the CYP2D6 genotype, and personality was measured with the TCI. Both studies found no association between CYP2D6 and personality traits, which is reasonable given the very low number of PMs and low CYP2D6 variability 30 included no PMs, only individuals homozygous or heterozygous for the CYP2D6*10 variant (with decreased activity) that were compared with individuals with apparently functional or wild-type alleles. Consistently, no differences emerged between EMs (*1/*1) and individuals with *10/*10 and *1/*10 genotypes. Nevertheless, those with reduced CYP2D6 activity presented a tendency to higher impulsivity (novelty seeking). Similarly, Iwashima et al. 31 only included one PMs carrying two alleles related to null enzyme activity (*5/*5). This PM was compared with the remaining 341 participants separated into ‘intermediate’ metabolizers (IM) with either two reduced activity alleles or one reduced activity allele and one no activity allele (e.g. *4/*10, *5/*10, *10/*10), EMs with one or more wild type alleles (e.g. *1/*1, *1/*2, *1/*5, *1/*10), and UMs considered those with more than two copies of wild type (e.g. *1/*1XN, *1/*2XN). As expected no differences were found. Just out of curiosity, the PM scored higher on impulsivity (novelty seeking) than the mean, minimum and maximum scores of EMs, IMs and UMs.

In summary, personality findings across heterogeneous populations appear difficult to replicate, since they rely on self-reports which may depend on cultural context, education, mood state, etc. Moreover, variability in recruitment may include, under the label of ‘healthy volunteers’, different types of participants (reviewed later).

Studies about CYP2D6 and personality in clinical settings

Two studies have explored the relationship between CYP2D6 and personality in clinical samples. The first of them was developed in Malaysian individuals, who were in hospital expecting to undergo orthopaedic surgery 35. This study used a completely novel classification of individuals according to their genotype and a different measure of personality. As in the previous Asian studies there were no PMs, and individuals were divided into those carrying alleles related to null or reduced activity, who were categorized as ‘slow’ (*4/*10; *5/*10; *10/*10; *10/*17), those with at least one null or reduced activity allele or ‘intermediate’ (*1/*4, *1/*5; *1/*9; *1/*10), and those considered normal (*1/*1). The evaluation of personality was done with a ‘Type A personality’ questionnaire, which measures a behavioural pattern characterized by tenseness, impatience, urgency and aggressiveness that often predicts stress-related disorders. The groups ‘slow’ and ‘intermediate’ were found to present lower scores on ‘Type A personality’, which suggests lower vulnerability to develop a stress related disorder in those carrying reduced or null function alleles 35.

Another study evaluated this relationship in patients with major depressive disorder in New Zealand 36. PMs vs. EMs showed lower levels of anxiety. PMs scored lower than EMs on the TCI-R scale of ‘harm avoidance’, which has been related to the risk for depression and other illness and to their outcome 36.

Previous findings about CYP2D6 and personality traits in clinical samples suggest that PMs might have some protective personality features against disease severity. While it is difficult to draw conclusions from these data, other studies relying on more objective measures may help our understanding of the risks or benefits of the different CYP2D6 metabolic groups for the development of diseases that are mostly treated with CYP2D6 substrates.

Table 1 summarizes the main methodological aspects such as number of participants, nationality, CYP2D6 debrisoquine hydroxylation capacity (phenotype), CYP2D6 alleles (genotype), the personality measures and most significant results. The sources of variability in the relationship between CYP2D6 activity and personality might be due to participants' differences in factors such as age, gender and ethnic background, culture, education, due to recruitment procedures and population characteristics. Differences may be also due to CYP2D6 and personality evaluation procedures. Furthermore, there might also be a population selection bias linked to personality across volunteers participating in biomedical studies involving an invasive procedure and/or psychological measures (reviewed later).

CYP2D6 and neurocognitive function in healthy volunteers

We decided to explore whether PMs and EMs presented differences in a more objective measure than personality such as cognitive functioning 32. Participants were evaluated using computerized, non-linguistic and culturally blind cognitive tests (http://www.CANTAB.com). PMs showed a different performance in comparison with the rest in rapid visual information processing (RVP), a test of sustained attention. PMs appeared to have a better capacity than EMs to a task of vigilance or alertness for a period of time 32. When controlling for distress by using a measure of overall psychopathology (SCL-90-R 37, PMs also showed better performance on spatial working memory 32. Thus, PMs seem to perform more accurately in tasks that demand sustained attention or vigilance.

Moreover, CYP2D6 activity has been related to blood flow in brain regions underlying sustained attention or alertness such as the thalamus during resting 38 and the cuneus and precuneus during cognitive demands 39.

CYP2D6 and vulnerability to psychopathology: studies in patients

Table 2 summarizes studies about the relationship between CYP2D6 and neurocognition and psychopathology in patients.

Schizophrenia

A lower frequency of PMs than expected has been observed in schizophrenia inpatients 40, which is in agreement with previous studies 41,42. Other studies did not find these differences 43–47. However, these studies should be interpreted with caution since they only analyzed two or three defective CYP2D6 variant alleles (*3, *4, and/or *10) as causing PM status, and/or used a diverse control group.

Anxiety and depression

A study testing whether CYP2D6 homozygous carriers of the two null allele *4 were more predisposed to anxiety and depression disorders in the elderly found no associations 48, although an increased frequency of UMs has been reported among women with late pregnancy or post-partum depressive symptoms 49. Among German UMs (genotype) depressive patients, the risk of a high suicidality was elevated as compared with those with other genotypes 50. We recently demonstrated among UMs an earlier dropout from fluoxetine or amitriptyline treatment in major depressive patients, which might be related to suicide risk as discussed below 51. Also, other clinically relevant CYPs have been related to depression, such as CYP2C9, 52 or depressive levels, such as CYP2C19 53.

Eating disorders

A higher frequency of CYP2D6 UMs (evaluated by debrisoquine test) was found among the group of the population with higher scores on a scale measuring symptoms of bulimia 54. Similarly, CYP2D6 allele distribution in patients with eating disorders was related to higher enzyme activity than in healthy controls 55. However this latter finding may be biased by the number of UM patients with a lifetime history of suicidal behaviour 56 as described next.

UMs and suicide

The presence of a higher frequency of CYP2D6 UMs among individuals who died by suicide than in those with a natural death has been reported 57 and also among patients with eating disorders with a history of suicidal attempts compared with those who had never attempted suicide 56. In agreement with these findings, it was later reported that among suicide attempters, those with a higher number of CYP2D6 active genes presented a greater severity of the suicide attempt 58 and greater personality psychopathology 59.

In summary, the involvement of CYP2D6 in endogenous metabolism could mediate both the pharmacological treatment and risk and evolution of the diseases that are being treated.

CYP2D6 and psychological functioning: Implications for clinical research

Besides the previously discussed potential double influence on the variability of both pharmacological treatment and vulnerability to certain diseases, CYP2D6 genetic polymorphism may be of relevance for clinical trials as discussed below. Our proposal is that clinical trial participants must be stratified according to pharmacogenomics in general, and to CYP2D6 in particular, since it could mediate a) a population selection bias and b) an inaccurate drug evaluation effect influencing both drug pharmacokinetics (PK) and pharmacodynamics (PD). Therefore, Drug Regulatory Agencies may consider using pharmacogenetics to stratify the subjects involved in human research protocols.

CYP2D6 and personality studies: relevance for population selection bias

Our original hypothesis for a role of CYP2D6 in the metabolism of endogenous biogenic neurotransmitter amines 25 was partially based on previous observations 60 of a greater frequency of CYP2D6 PMs among unrelated volunteers involved in a clinical trial. Therefore, the relationship between CYP2D6 and human behaviour may lead to a population selection bias of humans involved in clinical research.

Another factor to take into account would be the psychological status of homogeneous individuals enrolled in clinical trials. Self-proclaimed healthy volunteers, even if they report no history of psychopathology, might in fact include both healthy and subclinical individuals with vulnerability to distress. In support of this idea it has been shown that a proportion of healthy volunteer subjects participating in studies about personality present personality features related to mental disorders 61–67. As an example in our last study about CYP2D6 and personality 32, the subjects' psychological status was evaluated with Symptom Checklist-90-Revised (SCL-90-R) 37. Interestingly, all PMs reported very low scores suggesting a bias. Indeed, the individuals with higher scores on the SCL-90-R (EMs) showed differences in most personality scales when compared with those with low scores. Thus, the psychological status of volunteers, despite being ‘healthy’ and passing the filter of a psychiatric interview, may influence the differences found in the relationship of personality and CYP2D6 across studies (Table 1). This bias might be reduced if volunteers were screened for overall psychological status, otherwise it may have an effect.

Another potential bias in biomedical research is possibly an over-representation of certain personality features among participants. Because biomedical studies usually involve an invasive procedure (injections, the intake of drugs, etc.), volunteers tend to be altruistic and/or show higher than normative scores on personality traits such as impulsivity, novelty seeking or monotony avoidance 68,69.

In summary, in light of the above findings there seems to exist a need for behavioural and pharmacogenetic stratification of humans involved in clinical research.

CYP2D6 and psychological functioning: implication for drug evaluation (PK/PD) in clinical trials

Certain CYP enzymes beyond their contribution to the metabolism of psychotropic drugs (i.e. antidepressants, antipsychotics) are involved in the biotransformation of neuroactive endobiotics, which may influence physiological processes. In the light of this, our hypothesis is that genetic polymorphisms of these CYPs could mediate therapeutic responses to a given drug by two different mechanisms: influencing pharmacological response (drug metabolism, PK) but also mediating the metabolism of endogenous products related to therapeutic response (placebo or nocebo effects) (PD). Thus, this could be relevant for clinical research, and the pharmacogenetic stratification of subjects for drug evaluation effects should be considered.

To finalize, it is highlighted that major Drug Regulatory Agencies (EMA, FDA and the Japanese Drug Regulatory Agency) support the use of pharmacogenetics in the drug development phase as recently published 70. Therefore, pharmacogenetics should be taking into consideration for clinical trials as well as for drugs cost-effectiveness studies 71 and drug–drug interactions 72.

Conclusions and outlook

This review summarizes the findings on CYP2D6 polymorphisms and their role in endogenous substrate metabolism as well as the large body of association studies reporting the possible functional consequences of the polymorphism.

CYP2D6 metabolism of endobiotics and xenobiotics

CYP2D6 genetic polymorphism presents interindividual variation of the enzyme hydroxylation capacity. CYP2D6 may cause absent (PMs), decreased and normal enzyme activity (EMs). In addition, the latter group of EMs also includes individuals with increased enzyme hydroxylation capacity or UMs 1. CYP2D6 is involved in the metabolism of many CNS drugs such as antidepressants, antipsychotics and opioids, as well as in the metabolism of endogenous neuroactive substrates (i.e. neuroactive monoamines, endocannabinoids and endomorphines). Moreover, CYP2D6 is located in several brain regions (i.e. cortex, hippocampus and cerebellum), which seem of relevance for human behaviour and psychopathology. Therefore, interindividual variability in CYP2D6 hydroxylation capacity may have implications not only for the metabolism of several psychotropic drugs but also for explaining differences in human behaviour and psychopathology. However, considering that most CYP2D6 endogenous neuroactive substrates show low affinity, new studies are necessary to identify the definitive role of CYP2D6 in different human brain tissues. Furthermore, CYP2D6 expression patterns in different brain regions and their functional implications in brain metabolism might be considered.

There are reports of associations between psychological factors and CYP2D6 variation. Most of them are pure association studies. Thus, the mechanistic link between the functional role of CYP2D6 in neurons and behaviour needs to be clarified.

CYP2D6, personality, neurocognition and psychopathology

In healthy volunteers, PMs have been associated with a personality profile characterized by higher impulsivity and anxiety than EMs 25,28,32,73. Nevertheless, the PM personality profile does not appear related to an increased vulnerability to psychopathology since PMs seem to have a better capacity than EMs to the cognitive function of sustaining attention, which is deficient in individuals with anxious and impulsive related disorders 32. Moreover, CYP2D6 activity has been related to blood flow in brain regions underlying sustained attention or alertness 38,39.

In support of this hypothesis, a lower frequency of PMs than expected has been observed in schizophrenia patients 40. Moreover, a higher frequency of CYP2D6 UMs has been found among subjects who died by suicide 57, and among patients with eating disorders with a history of suicidal attempts 56. In agreement with these findings, a higher number of CYP2D6 active genes was related to a greater severity of the suicide attempt 58, and greater personality psychopathology 59. In conclusion, there is evidence supporting the relationship between CYP2D6 variability and psychological and neurocognitive functioning as well as to psychopathology.

Nevertheless, new studies are necessary to support a direct involvement of CYP2D6 in psychiatric disorders since there are only small studies reporting associations, which increase the risk for false-positive findings.

CYP2D6 and clinical trials

Subjects with a different response to CNS drugs that are CYP2D6 substrates may also be those with a different sensitivity to mental disorders. In other words, CYP2D6 might be of relevance for theranostics. Consequently, the current model of drug clinical research can be put into question. Since CYP2D6 genetic polymorphism might be related to behaviour, this could influence a population selection bias when enrolling healthy volunteers, which will be particularly relevant for the study of CYP2D6 substrates as we found 60. Furthermore, with regard to the evaluation of drug therapeutic effects, the effects of endogenous substrates could be also influencing the variability in the therapeutic response (placebo or nocebo effect) since the therapeutic response goes beyond the pharmacological effect.

Besides, the evaluation of CYP2D6 in clinical trials may be of importance for CYP2D6 substrates given its involvement in drug metabolism, in particular for active drugs in the CNS, because local brain metabolism may be important for the interaction with neuroactive substances and their biotransformation.

Overall, CYPs enzymes such as CYP2D6, which have been linked to behaviour, may lead to a population selection bias in clinical trials and to an over-representation of certain metabolic groups affecting the drug effect evaluation itself. Therefore, pharmacogenetic evaluation must be included in clinical research.

Competing Interests

All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf and declare no support from any organization for the submitted work, no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years and no other relationships or activities that could appear to have influenced the submitted work.

The study has been partly supported by European Union (FEDER), Institute of Health Carlos III-FIS and the (PI10/02010) and Gobierno de Extremadura (BS10023). The contributionof M.E.G. Naranjo (supported by Gobierno de Extremadura, Consejería de Empleo, Empresa e Innovación and Fondo Social Europeo EU, Grant PD10199) is gratefully acknowledged.