JOURNAL OFAvailable online at www.sciencedirect.com

PSYCHIATRIC

RESEARCHJournal of Psychiatric Research 42 (2008) 89–97

www.elsevier.com/locate/jpsychires

Associations between MDR1 gene polymorphisms andschizophrenia and therapeutic response to olanzapine

in female schizophrenic patients

Nada Bozina a,*, Martina Rojnic Kuzman b, Vesna Medved b, Nikolina Jovanovic b,Jadranka Sertic a, Ljubomir Hotujac b

a Clinical Institute of Laboratory Diagnosis, Zagreb University Hospital Centre, Salata 2, 10000 Zagreb, Croatiab Department of psychiatry, Zagreb University Hospital Centre and Zagreb University School of Medicine, Kispaticeva 12, 10000 Zagreb, Croatia

Received 15 July 2006; received in revised form 27 September 2006; accepted 2 October 2006

Abstract

Multidrug resistant protein (MDR1) gene, which codes for P-glycoprotein and functions as an efflux transporter in different cells, iswidely localized in normal tissues including the gastrointestinal tract, blood cells, biliary tract, kidney and brain and plays a major role inabsorption, distribution and elimination of various xenobiotics. Therefore, MDR1 gene variants were proposed as potential susceptibil-ity factors for diseases and as determinants of treatment response to various drugs. We investigated the relationships between exon 21G2677T and exon 26 C3435T genetic variants of MDR1 gene with susceptibility and treatment response in female schizophrenic patients.The study was conducted in two steps. We first compared allele, genotype and haplotype distributions between 117 female schizophrenicpatients and 123 control female subjects. Afterwards, we studied treatment response to olanzapine, in 87 out of 117 previously unmed-icated female patients. Overall, we found lower representation of G2677/C3435 haplotype in schizophrenic female patients compared tocontrols. Test result for linkage disequilibrium between loci was found to be significant. Furthermore, we found significant associationsbetween MDR1 exon 21 G2677T genotypes and treatment response measured with positive PANSS percentage changes, with T alleleand TT genotype being associated with significantly better treatment response. A borderline, non-significant statistical associationwas found between MDR1 exon 26 C3435T genotypes and treatment response, with TT genotype being associated with better treatmentresponse. Our data support functional importance of the MDR1 mutations for the susceptibility and treatment response in femaleschizophrenic patients.� 2006 Elsevier Ltd. All rights reserved.

Keywords: Schizophrenia; MDR1; Olanzapine; P-glycoprotein; Treatment response; Genetic polymorphism

1. Introduction

Antipsychotic drugs are considered the core treatment ina variety of schizophrenic spectrum disorders (AmericanPsychiatric Association, 2004). Compared to typical anti-psychotics, second generation antipsychotics (SDAs) havemore acceptable side effect profiles and possibly broadersymptom efficacy, especially concerning negative symp-

0022-3956/$ - see front matter � 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jpsychires.2006.10.002

* Corresponding author.E-mail address: nbozina@net.hr (N. Bozina).

toms, thus bringing considerable progress in the treatmentof schizophrenic patients (Bagnall et al., 2003). Still, obvi-ous interindividual differences in treatment response toSDAs indicate that genetic factors may be relevant.

Amongst the best understood mediators of drug resis-tance is the multi drug resistant (MDR1 or ABCB1) P-gly-coprotein (P-gp) (Ling, 1997). The transmembrane effluxtransporter P-gp is widely localized in normal tissuesincluding the apical membrane of the gastrointestinal tract,blood cells, the biliary canalicular membrane of hepato-cytes, the luminal membranes of proximal tubular epithe-lial cells in the kidney and the luminal membranes of

90 N. Bozina et al. / Journal of Psychiatric Research 42 (2008) 89–97

endothelial cells in cerebral capillaries forming the blood-brain barrier (Cordon-Cardo et al., 1989; Hitzl et al.,2001), and thus limits cellular uptake of xenobiotics byexcreting these compounds into bile, urine and intestinallumen, and limits accumulation in the brain. Therefore itplays a major role in absorption, distribution and elimina-tion of drugs (Lin and Yamazaki, 2003). P-gp is highlypolymorphic and its expression in endothelial cells in cere-bral capillaries forming the blood–brain barrier limits sub-strate access to the brain (Thompson et al., 2000; El Elaet al., 2004). Studies using mice models showed the impactof P-gp on blood–brain barrier; olanzapine penetrationinto brain was greater in transgenic abcb1a (mdr1), P-gly-coprotein-deficient mice than in FVB1 (wild-type) animals(Wang et al., 2004). P-glycoprotein-deficient mice were alsoreported to have 10 times higher drug concentrations (Uhrand Grauer, 2003; Uhr et al., 2006). Reports using in vitromethods suggest that risperidon, sertraline and paroxetinemay be P-gp substrate, unlike clozapine, haloperidol,chlorpromazine, citalopram and venlafaxine (Schinkelet al., 1997; Mahar Doan et al., 2002; Weiss et al., 2003).Using ATP-ase activity as a marker of P-gp binding affin-ity, the authors demonstrated that P-gp may in variousdegrees influence the access of antipsychotics to the brain,where olanzapine was ranked as an intermediate P-gp sub-strate (Boulton et al., 2002). According to data from ani-mal studies using knock-out mice, olanzapine is anintermediate P-pg substrate (Wang et al., 2004).

Considering the localization of the P-pg and its role inabsorption, distribution and elimination of various xenobi-otics, it might act as a barrier to different exogenous noxa.Functional variants with altered gene expression mighttherefore constitute a susceptibility factor for complex dis-eases, where both genetic and environmental factors wereshown to affect the disease risk (Furuno et al., 2001). Asschizophrenia is a complex disease, and environmental fac-tors are important for its expression (Kaplan and Sadock,2003), MDR1 gene could play an important role.

Functional polymorphisms of the MDR1 gene wereextensively studied through their influence on expressionof MDR1 (Hoffmeyer et al., 2000; Nakamura et al.,2002), their association with pharmacokinetics and bio-availability of some drugs (Uhr et al., 2006) and their asso-ciations with clinical effects (Loscher and Potschka, 2002;Roberts et al., 2002; Yamauchi et al., 2002; Eichelbaumet al., 2004). Two single nucleotide polymorphisms(SNP), silent mutation C3435T in exon 26 and exon 21SNP G2677T generated greatest interest, yet were foundto be contradictorily associated with different changes inexpression of the MDR1 protein and plasma drug concen-tration (Hoffmeyer et al., 2000; Sakaeda et al., 2001).

To date, linkage disequilibrium (LD) analysis of differ-ent MDR1 polymorphisms showed C3435T to be in LDwith at least one other functional polymorphic locus(Kim et al., 2001; Tang et al., 2002). Considering thatC3435T is a silent mutation not resulting in amino acidchanges, the association found in the study suggests that

it may be in linkage disequilibrium with another, functionalpolymorphism of MDR1 gene. The non-synonymous exon21 SNP G2677T might seem particularly interesting. Itcould also explain the conflicting results.

Polymorphisms of different genes that could serve aspotential gene markers in the prediction of treatmentresponse to SDAs in schizophrenia have been extensivelystudied. However, to date only one study reported associa-tion of exon 26 C3435T and exon 21 G2677T with treat-ment response to bromperidol in schizophrenic patients(Yasui-Furukori et al., 2006), although most second gener-ation antipsychotics were shown to be P-gp substrates(Boulton et al., 2002; El Ela et al., 2004). Considering therole of P-gp in bioavailability and its expression in endo-thelial cells in cerebral capillaries forming the blood–brainbarrier which limits substrate access, we hypothesized thatfunctional polymorphisms of MDR1 gene might influencetreatment response. The aims of the study were to investi-gate the potential influence, of MDR1 polymorphismsexon 26 C3435T and exon 21 G2677T on treatmentresponse in schizophrenic patients treated with olanzapineand on their susceptibility to schizophrenia.

2. Methods

The study was conducted in two steps. First, we com-pared allele, genotype and haplotype distributions between117 female schizophrenic patients and 123 control femalesubjects. All patients and subjects were of Croatian des-cent. All patients were acutely exacerbated women aged18–65 and meeting DSM-IV criteria for schizophrenia orschizoaffective disorder, referred to the Department of Psy-chiatry, Zagreb University Hospital Centre. Patients withsignificant abnormalities on standard laboratory tests,EEG, physical examination, significant organic or neuro-logical disease, or history of mental disorders other thanschizophrenia (alcoholism, drug addiction and epilepsy)were excluded from the study. At screening, diagnosiswas made by two experienced psychiatrists (VM and LH)and each was kept blind to the diagnosis made by theother. The means ± SD age, years of education, and dura-tion of illness were 39.4 ± 11.5 years, 11.9 ± 2.8 years and7.8 ± 7.5 years, respectively. The control group consistedof 123 female subjects, blood donors, aged 35–65, withoutany history of neuropsychiatric disorders. The means ± SDage and years of education were 38.1 ± 11.3 years and12.7 ± 1.5 years, respectively.

Second, we studied the baseline symptomatology andtreatment response to olanzapine. However, 30 out of117 patients had already failed one or two antipsychoticmedications during treatment prior to the beginning ofthe study. Therefore, we studied baseline severity of symp-toms and treatment response to olanzapine in the sample of87 out of 117 female schizophrenic patients who hadreceived no antipsychotic medication prior to olanzapineadmission. Olanzapine was administered in open treat-ment, in fixed doses (olanzapine 10 mg/day (n = 77), or

Table 1Distribution of allele and genotype frequencies of MDR1 exon 21 G2677Tand exon 26 C3435T in female schizophrenic patients and control femalesubjects

Locus Schizophrenic

patients

n = 117

Control

subjects

n = 123

MDR 1 exon 21G2677T

Allelea G 143 160T 91 86

Genotypeb GG 44 54GT 55 52TT 18 17

MDR 1 exon 26C3435T

Allelea C 110 134T 124 112

Genotypeb CC 27 36CT 56 62TT 34 25

a Fischers’s exact 2 · 2: exon 21 G2677T, p = 0.344; exon 26 C3435,p = 0.101.

b Fischers’s exact 2 · 5: exon 21 G2677T, p = 0.628 and exon 26 C3435,p = 0.259.

N. Bozina et al. / Journal of Psychiatric Research 42 (2008) 89–97 91

20 mg/day (n = 10) on the basis of good clinical practice.The means ± SD age, years of education, and duration ofillness (defined by anamnestic data on the first occurrenceof the symptoms) were 36.4 ± 8.5 years, 12.1 ± 2.8 yearsand 4.1 ± 4.5 years, respectively. Concomitant medicationssuch as diazepam (up to 10 mg/day) or clonazepam (up to6 mg/day) for occasional insomnia were allowed (duringthe first week). No changes in olanzapine dose wereallowed. All patients were asked to report any adverseeffect (extrapyramidal symptoms, akathisia, andsomnolence) and smoking habits. The study protocolwas approved by the Hospital Ethical Committee of theZagreb University Hospital Centre. After receiving infor-mation about the study, each subject signed informedconsent.

2.1. Assessment

To assess and evaluate baseline symptomatology andimprovement of clinical psychotic symptoms as therapeuticresponse to antipsychotic, all patients were rated using thePositive and Negative Symptom Scale (PANSS) at twotime points (initially and three months afterwards), bytwo experienced psychiatrists (VM and LH). Interrater reli-ability score was high, above 0.86.

2.2. Genotyping procedures

Genomic DNA was extracted from peripheral lympho-cytes using salting out procedure (Miller et al., 1988). Analy-sis of 2677 G/T/A polymorphisms in exon 21 of MDR1 genewas performed according to the method described by Casc-orbi et al. (2001). For 3435C/T polymorphism in exon 26,the method described by Sakaeda et al. (2001) was applied.Substitution G2677T/A in exon 21 was detected by PCR-RFLP method with Ban I and BsrI restriction endonucleases,respectively. G, T and A alleles were represented by 198 bp,244 bp and 206 bp fragments, respectively.

PCR-RFLP method with Nde II restriction endonucle-ase was applied to detect MDR1-C3435T substitution.‘‘C’’ and ‘‘T’’ alleles were represented by 193 bp and144 bp fragments, respectively.

2.3. Statistical analysis

A test for Hardy–Weinberg equilibrium using Markovchain method (Guo and Thompson, 1992), as well as link-age-disequilibrium likelihood-ratio test between loci whosegametic phase is unknown (Slatkin and Excoffier, 1996)were performed, as implemented in Arlequin ver. 3.01(Excoffier et al., 2006). Haplotype frequencies were esti-mated using Expectation-Maximization algorithm imple-mented in the same program, leading to maximumlikelihood estimates of haplotype frequency. Fischer’sexact test was used for pair-wise comparisons of the allele(2 · 2) and genotype (2 · 5) frequencies between schizo-phrenic patients and controls. Log likelihood ratio tests

were performed to compare distributions of the estimatedhaplotypes between case and control groups.

The relationship between the baseline severity of symp-toms and treatment responses measured by percentagechanges of total PANSS and PANSS subscales (Positive,Negative and General Psychopathology) with genetic poly-morphisms were studied with analysis of variance. p-Valuesof 0.05 or less were considered statistically significant.Comparisons regarding age, years of education, durationof illness, dose of antipsychotic, smoking habits andreported adverse effects between genotypes were performedwith ANOVA, v2, and Fisher’s exact test. Binary logisticregression was used to predict baseline severity of symp-toms and treatment response from sociodemographic data(age, years of education), duration of illness, adverseeffects, smoking habits and dose of antipsychotics neededto achieve the remission of symptoms. Comparisonsregarding age, years of education, duration of illness,smoking habits, severity of initial symptoms and reportedadverse effects and genotype distributions in different dosegroups (10 mg and 20 mg) were performed using v2 andFisher’s exact tests. All statistical analyses were carriedout using SPSS 11.5 (SSPS inc., Chicago, IL, USA) statis-tical software package. The sequential Bonferroni adjust-ments (Holm, 1979; Rice, 1989) were applied to correctfor the effect of multiple tests using SAS Release 8.02.

3. Results

Allele and genotype frequencies of the MDR1 exon 21G2677T and MDR1 exon 26 C3435T genotypes in femaleschizophrenic patients and controls are shown in Table 1.No significant deviations from the expected Hardy–Wein-berg proportions were observed in the sample of schizo-phrenic patients (exon 21 G2677T, p = 0.68; exon 26C3435T, p = 0.46), or in controls (exon 21 G2677T,

92 N. Bozina et al. / Journal of Psychiatric Research 42 (2008) 89–97

p = 0.43; exon 26 C3435T, p = 0.99), and in total sample(exon 21 G2677T, p = 0.47; exon 26 C3435T, p = 0.99).Test results for linkage disequilibrium between loci werefound to be significant in schizophrenic patients (p <0.001, v2 = 43.45, df = 2), controls (p < 0.001, v2 = 78.89,df = 1) and total sample (p < 0.001, v2 = 114.93, df = 2).Pair-wise linkage profiles of the two SNPs present in ourpopulation determined using the EM algorithm showedminor linkage (defined as <10% occurrence) only for2677T/C3435 (Table 2). Pair-wise comparisons of the allelefrequency between schizophrenic patients and controlsrevealed no statistical differences for both loci (Fischers’sexact 2 · 2 test: exon 21 G2677T, p = 0.344; exon 26C3435, p = 0.101). Also, no significant difference in the dis-tribution of exon 21 G2677T or exon 26 C3435T genotypewas observed between schizophrenic patients and controls(Fischers’s exact 2 · 5 test: exon 21 G2677T, p = 0.628and exon 26 C3435, p = 0.259) (Table 1).

No statistical differences were found in distributions ofthe estimated haplotypes between schizophrenic patientsand controls, (v2 = 6.121, df = 3, p = 0.105). However,comparison of carriers of G2677/C3435 haplotype vs. oth-ers showed G2677/C3435 to be significantly less repre-sented in the schizophrenic group (v2 = 5.31, df = 1,p = 0.040), while the haplotype G2677/3435T showed bor-derline non-significant overrepresentation in the schizo-phrenic group (Table 2).

Analyzing associations between treatment responsemeasured with percentages of changed total PANSS andall PANSS subscales and MDR1 exon 26 genotype, weobserved differences that were, however, not statisticallysignificant (changed total, positive and general PANSS inTT > CT > CC). Nevertheless, we found significant associ-ations between MDR1 exon 21 genotype and treatmentresponse measured with positive PANSS percentagechange in genotype-based analysis and in allele-based anal-ysis, with TT genotype and T allele being associated withsignificantly better treatment response explaining 8.9%and 3.4% of variance, respectively. Also, borderline statis-tical significance was observed between MDR1 exon 26genotype and total PANSS scale and general PANSS

Table 2Haplotype frequencies (proportions) of the two SNP loci: exon 21 G2677Tand exon 26 C3435T in female schizophrenic patients and controls

Haplotype Exon 21

G2677T

Exon 26

C3435T

Schizophrenic patients

n = 117Control

subjects

n = 123

H1 G C 99 (41) 128 (52)a

H2 G T 45 (20) 32 (13)b

H3 T C 9 (5) 6 (3)H4 T T 79 (33) 80 (32)

Haplotype frequency was determined using the statistical program basedon the Expectation Maximization (EM algorithm).v2 = 6.121, df = 3, p = 0.105.a G2677/C3435 haplotype vs. others, v2 = 4.202, df = 1, p = 0.040.b G2677/3435T haplotype vs. others, v2 = 3.615, df = 1, p = 0.057.

subscale percentage changes with TT genotype being asso-ciated with better treatment response (Table 3).

We found no statistically significant associationsbetween MDR1 exon 26 and MDR1 exon 21 genotypeswith baseline PANSS values (measured with total PANSSscale and all PANSS subscales).

There were no statistically significant differences in clin-ical profiles such as age, years of education, duration of ill-ness, smoking habits, and dose of antipsychotics or adverseeffects (only somnolence was reported) following olanza-pine administration between MDR1 exon21 and MDR1exon 26 genotypes (data not shown). Binary logistic regres-sion analysis also established that there were no statisticallysignificant confounding effects of age, duration of illness,level of education, reported side effects, smoking habitsand dose of antipsychotics administered, on treatmentresponse measured by the percentage changes of totalPANSS and all PANSS subscales.

To test the possible differences between the group ofpatients who received 10 mg of olanzapine per day andthe group of patients who received 20 mg of olanzapineper day, we performed additional analyses. As expected,patients receiving 10 mg/day had lower positive initialPANSS scores at admission to our hospital compared tothe group receiving 20 mg/day (mean values ± SD were54 ± 19 vs. 62 ± 14), which confirmed that dose was deter-mined according to the basis of good clinical practice.There were no differences in clinical profiles such as age,years of education, duration of illness, reported adverseeffects and smoking habits and genotype distributionsbetween those groups (data not shown).

4. Discussion

Several conclusions could be drawn from these data: (1)MDR1 exon 21 variants might be associated with treat-ment response to olanzapine in schizophrenic femalepatients, (2) MDR1 exon 26 variants might have onlyminor effects on treatment response to olanzapine inschizophrenic female patients, (3) exon 21 G2677T andexon 26 C3435T might produce combined effects for sus-ceptibility to schizophrenia in females, (4) exon 21G2677T and exon 26 C3435T were found to be in linkagedisequilibrium in the studied population.

We have for the first time analyzed a group of previouslyunmedicated patients, limiting the influence of the mostconfounding factor, i.e. previous medications, on studyresults. According to our results, MDR1 exon 21 polymor-phic locus G2677T might be associated with therapeuticresponse to olanzapine in schizophrenic female patients.We found significant association between 2677T/T alleleand genotype and better treatment response measured withpositive PANSS percentage changes, which explains about9% of variance and borderline statistical associations withtotal PANSS percentage change, probably accounting forpositive and general PANSS percentage changes. The lackof associations between G2677T genotypes and negative

Table 3Genotype comparisons of the severity of psychotic symptoms at admission and after 3-month olanzapine treatment in previously unmedicated femaleschizophrenic patients

Genotype p (R2)

GG (n = 35) GT (n = 41) TT (n = 11)

MDR 1 exon 21 G2677T

Total PANSS score at admission 102 ± 20 93 ± 21 94 ± 20 0.143Positive subscore at admission 28 ± 9 24 ± 8 27 ± 5 0.259Negative subscore at admission 27 ± 7 25 ± 7 22 ± 7 0.112General subscore at admission 48 ± 10 44 ± 10 45 ± 8 0.290Percentage change in PANSS score 47 ± 13 41 ± 14 49 ± 9 0.085Changed positive subscore 59 ± 17 52 ± 19 68 ± 9 0.022 (0.089)Changed negative subscore 39 ± 20 35 ± 20 24 ± 19 0.114Changed general subscore 42 ± 14 36 ± 15 47 ± 12 0.061

MDR 1 exon 26 C3435T CC (n = 23) CT (n = 38) TT (n = 26)Total PANSS score at admission 97 ± 21 96 ± 19 96 ± 17 0.977Positive subscore at admission 24 ± 10 27 ± 8 26 ± 7 0.4040Negative subscore at admission 24 ± 8 25 ± 7 25 ± 7 0.856General subscore at admission 43 ± 10 47 ± 10 46 ± 8 0.155Percentage change in PANSS score 38 ± 15 45 ± 14 47 ± 15 0.079Changed positive n subscore 47 ± 18 55 ± 19 58 ± 18 0.112Changed negative subscore 30 ± 24 39 ± 19 34 ± 19 0.247Changed general subscore 35 ± 13 39 ± 15 43 ± 14 0.157

Genotype p (R2) Genotype p (R2)

MDR 1 exon 21 G2677T aGG/GT (n = 76) bGT/TT (n = 52)

Total PANSS score at admission 97 ± 19 0.312 94 ± 20 0.079Positive subscore at admission 26 ± 9 0.790 25 ± 8 0.218Negative subscore at admission 25 ± 7 0.001 23 ± 6 0.013General subscore at admission 46 ± 10 0.413 44 ± 9 0.165Percentage change in PANSS score 44 ± 14 0.205 43 ± 14 0.331Changed positive subscore 55 ± 18 0.031 (0.034) 55 ± 18 0.129Changed negative subscore 37 ± 20 0.107 33 ± 20 0.277Changed general subscore 39 ± 15 0.131 39 ± 15 0.158

MDR 1 exon 26 C3435T cCC/CT (n = 61) dCT/TT (n = 64)Total PANSS score at admission 96 ± 20 0.610 96 ± 18 0.339Positive subscore at admission 26 ± 9 0.900 26 ± 7 0.171Negative subscore at admission 24 ± 7 0.800 25 ± 7 0.212General subscore at admission 46 ± 10 0.824 47 ± 9 0.098Percentage change in PANSS score 42 ± 15 0.159 46 ± 14 0.214Changed positive subscore 53 ± 19 0.247 58 ± 18 0.286Changed negative subscore 36 ± 20 0.660 37 ± 19 0.127Changed general subscore 37 ± 14 0.071 41 ± 15 0.305

Values are means ± SD.a GG/GT vs. TT.b GT/TT vs. GG.c CC/CT vs. TT.d CT/TT vs. CC.

N. Bozina et al. / Journal of Psychiatric Research 42 (2008) 89–97 93

symptoms could be due to the fact that time course of clin-ical improvement of negative symptoms requires a periodlonger than 3 months, which was the duration of the study(Hatta et al., 2003). Although we found no statistically sig-nificant association between 26 C3435T genotype andtreatment response measured with changed PANSS, somedifferences were observed (changed total PANSS inTT > CT > CC), which reached borderline significance,indicating a possible weak influence of this polymorphismon treatment response.

To date, two recent studies examined associationsbetween exon 21 G2677T and exon 26 C3435T polymor-phisms and therapeutic response in schizophrenic patients,

both performed in Japan (Takao et al., 2006; Yasui-Furuk-ori et al., 2006). Yasui-Furukori et al. reported subjectstreated with bromperidole, assessed with brief psychiatricrating scale (BPRS) and found only exon 21 2677T variantto be associated with worse treatment response measuredwith improvement on total score, but not with percentagechange. A few issues could explain the discrepancy withour results. First, considering the difference in initialPANSS values, it is possible that it could have influencedtreatment response measured with the absolute PANSSchange and resulted in false positive associations, as theauthors discussed themselves. Second, the small sample size(33 patients) and large standard deviations could have

94 N. Bozina et al. / Journal of Psychiatric Research 42 (2008) 89–97

influenced the obtained statistical significances of theirresults. Takao et al., found no differences in allele distribu-tion of exon 21 and exon 26 C3435T between schizophrenicpatients and treatment resistant schizophrenic patients.The differences between these studies might be due to thedifferent methodology applied. Also, ethnic diversity ofour subjects compared to subjects of both studies (Takaoet al., 2006; Yasui-Furukori et al., 2006) makes our resulthardly comparable. Allele frequencies of both C3435Tand G2677T polymorphisms of MDR1 show large varia-tions in populations (Tang et al., 2002; Cascorbi et al.,2001). In Caucasians, C3435 frequency is 43–54% (Bernalet al., 2003; Kurzawski et al., 2006), 34–63% in Asians(Nakamura et al., 2002) and 73–90% in African population(Chelule et al., 2003). Allelic frequency of SNP in exon 21G2677 is 57% in Caucasian, 43% in Japanese, and 34% inIndian population (Ieiri et al., 2004). Interethnic differencesin allelic distribution were proposed as an underlying fac-tor of observed interethnic variability in bioavailability ofdifferent drugs, i.e. oral bioavailability of cyclosporinewas lower in African Americans compared to Caucasiansor Hispanics (Lindholm et al., 1992). Allele and genotypefrequencies in the studied population did not differ fromreported results in other European regions (Hoffmeyeret al., 2000; Fromm, 2002; Ieiri et al., 2004). However, con-sidering significant variations of allele and genotype fre-quencies, this ethnic diversity between our studies andother studies might be important.

Some studies reported polymorphic variant in exon 263435T to be associated with lower MDR1 expression inthe duodenum and increased plasma levels of digoxin(Hoffmeyer et al., 2000), while a study conducted by Saka-eda et al. found lower plasma concentration of the digoxinin 3435T carriers (Sakaeda et al., 2001). Considering thatexon 26 C3435T is a silent mutation, it was suggested thatthis polymorphism might be in linkage disequilibrium withanother functional MDR1 polymorphism, which couldexplain the conflicting result from genetic studies. Indeed,our results confirmed its linkage with exon 21 G2677T.Studies reporting the influence of the exon 21 C2677T poly-morphism on MDR1 expression and effect on drug bloodlevels also produced conflicting results. Compared to2677G allele, 2677T was found to be associated withenhanced, although non-significant efflux of digoxin (Kimet al., 2001) and found higher, non-significant expressionof P-gp in placenta in 2677G compared to 2677T variant(Tanabe et al., 2001). The same author (Kim et al.,2001), however, reported that the subjects homozygousfor the exon 21 G2677 had higher (borderline statistical sig-nificance) fexofenadine area under the plasma concentra-tion (AUC) than the T2677T subjects. As 2677T alleleswere associated with better treatment response in ourstudy, our results are compatible with the studies byTanabe et al. (2001) and Hoffmeyer et al. (2000), which alsoassumed T variants to be associated with lower P-gp func-tions. The discrepancies between studies might be due tovarious factors, e.g. possible significantly different regula-

tion of P-gp expression in body tissues or different drugsstudied. Further, other factors not included in the studymight have influenced our results. There are reports aboutrelevant interactions between P-gp and CYP3A4 occurringat the intestinal level (Hesselink et al., 2003; Bai et al.,2004), which might have influenced drug disposition. Olan-zapine is a P-gp and CYP1A2 and CYP2D6 substrate andsuch interactions at the periphery, and possibly centrally,could not be excluded (Prior and Baker, 2003; El Elaet al., 2004; Wang et al., 2006), especially since it was con-firmed that CYP enzymes are also localized in the brain(Hedlund et al., 2001; Funae et al., 2003).

The most reported association studies of MDR1 geneand neuropsychiatric disorders have their origin in studiesof epilepsy (Kwan and Brodie, 2005). Some studiesreported the possible role of P-gp in treatment-resistantepilepsy; patients with treatment-resistant epilepsy weresignificantly more likely to have 3435C allele (Siddiquiet al., 2003) or overexpression of MDR1 (Dombrowskiet al., 2001). Subsequent studies failed to find the sameassociations (Sills et al., 2005; Kim et al., 2006). Our resultsare compatible with the results of Siddiqui et al. (2003) asTT genotype could be assumed to be associated with lowP-gp expression, possibly allowing higher drug concentra-tions in the brain. However, although the effects of poly-morphisms in MDR1 in the peripheral cells (affectingdrug distribution) as well as centrally in endothelial cellsaffecting blood–brain barrier, are well studied (Ling,1997), drug response might not be influenced only by P-pg effects on drug availability-plasma levels or brain levelsof olanzapine since MDR1 is also expressed in neurons andastrocytes (i.e. in epileptogenic lesions) (Sisodiya et al.,2002). A question arises whether the alternate MDR1 geneexpression in neurons occurs during the psychosis as welland affects the clinical outcome of a psychotic episode.

Considering the localization of the P-pg and its role inabsorption, distribution and elimination of various xenobi-otics, MDR1 gene variants were proposed as potential sus-ceptibility factors for diseases where both genetic andenvironmental factors have been shown to play a role indisease risk. Animal studies suggest this hypothesis as plau-sible as MDR1 knock-out mice are susceptible to develop-ment of severe spontaneous intestinal inflammations(Panwala et al., 1998). Panwala et al. suggested that a phys-iological role of P-gp might be to constitute a barrier toexogenous xenobiotics, even bacteria toxins. It has beensuggested that P-pg plays a role in susceptibility to epi-lepsy, as according to animal studies, acute seizures ingenetically epilepsy-prone rats can induce expression ofMDR1 (Kwan et al., 2002). The most convincing studyof MDR1 polymorphic variants and susceptibility for neu-ropsychiatric disorders (Furuno et al., 2001) revealed theassociations of G3435T genotypes with Parkinson disease.Higher frequency of MDR1 3435TT genotype carriers wasfound with early-onset Parkinson disease, in comparison tolower frequency in late-onset Parkinson disease and controlsubjects.

N. Bozina et al. / Journal of Psychiatric Research 42 (2008) 89–97 95

To our knowledge, this is the first association study ofMDR1 gene and schizophrenia. Schizophrenia is a com-plex disease and environmental factors are important forits expression (Kaplan and Sadock, 2003). According toour results, MDR1 gene might be a relevant factor whichmediates the efflux of some unknown xenobiotic (Panwalaet al., 1998). For example, a meta-analysis from prospec-tive studies indicated cannabis to be implicated in the aeti-ology of schizophrenia (Henquet et al., 2005). The use ofcannabis is more frequent in schizophrenics compared toother psychiatric disorders even prior to the beginning ofthe illness (Boydell et al., 2006), and patients with life his-tory of cannabis use developed schizophrenia at an earlierage than schizophrenic patients with no history of cannabisuse (Arendt et al., 2005). Interestingly, cannabinol (CBN),one of the major marijuana constituents was found to sig-nificantly inhibits P-gp-mediated drug transport (Zhuet al., 2006), while cannabidiol (CBD) and D9-tetrahydro-cannabinol (THC) inhibited P-gp expression (Hollandet al., 2006). If physiological role of P-gp might be to con-stitute a barrier to exogenous xenobiotics as suggested byPanwala et al. (1998), it might be possible that subjectswith altered P-pg function might show susceptibility forschizophrenia (at least induced by cannabis). In the presentstudy, our results lead to the conclusion that MDR1 genemight play a minor role in susceptibility to schizophreniain females. We found lower representation of the G2677/C3435 haplotype in the schizophrenic group, which couldsuggest an additive effect of both polymorphisms. Ourresults might imply a relative resistance of the control sub-jects considering that G2677/C3435 might be associatedwith higher P-pg function, which is compatible with thestudy of the role of C3435T polymorphism in Parkinsondisease conducted by Furuno et al. (2001). However, sincewe found no differences in either allele or genotype fre-quencies, it is possible that both loci might be in linkagedisequilibrium with another MDR1 mutation, i.e. exon12 (Hoffmeyer et al., 2000; Johne et al., 2002). Therefore,although we determined suggestive factors for susceptibil-ity to schizophrenia in females, these results require replica-tion, also including the analysis of other MDR1polymorphic loci.

A few factors could have influenced the results. Giventhe heterogeneous nature of the illness itself, it is possiblethat the influence of MDR1 genes is significant in particu-lar subgroups of schizophrenia, i.e. in more severe forms ofschizophrenia. Also, interactions of different genes thatwere not addressed in this study, together with their poten-tial involvement in susceptibility to schizophrenia (i.e. sero-tonin and dopamine transporters or receptors) might haveinfluenced the results.

There are several methodological considerations andlimitations of the study that should be acknowledged.First, olanzapine is an MDR1 substrate and the plasmalevels of olanzapine could vary among patients; as olanza-pine blood concentrations were not measured in this study,we are not able to present direct associations between the

MDR1 polymorphisms and drug concentration, althoughconsidering that the majority of patients were receivingthe same dose (10 mg/day) and since olanzapine plasmalevels show positive correlation with olanzapine dosage(Mauri et al., 2005) we do not expect significant deviationsin plasma olanzapine levels among those patients. How-ever, we carefully observed the factors that could haveinfluenced olanzapine dose as smoking, metabolic distur-bances, renal functions, and found no differences betweengenotypes of the studied loci. Also, we found no differenceswhen comparing groups of patients on 10 mg of olanzapineper day and 20 mg of olanzapine per day, regarding all con-trolled factors, with the exception of higher initial positivePANSS values. However, as only 10 subjects were treatedwith 20 mg of olanzapine per day, the reliability of theanalysis is relatively weak and therefore studies measuringplasma levels of olanzapine are more adequate to give rel-evant answers in regard to the influence of associations ofolanzapine dose and drug concentration to treatmentresponse. Second, as olanzapine is an intermediateMDR1 substrate, replication studies with strong substratesas risperidone would provide stronger evidence. However,as a strong side of the study, we point out that all analyseswere performed on subjects who received no antipsychoticmedication prior to our study, which is a major confound-ing factor in most association studies of treatmentresponses and genetic polymorphisms. This enables us toconclude more freely about the influence of MDR1 allelesand genotypes on treatment response to olanzapine. Third,all study subjects were women, which set a limitation togeneralizing the findings to male subjects with schizophre-nia. However, at the same time this contributes to thehomogeneity of the study sample since gender differencesin the course of illness have been described (Lewineet al., 1981; Babigian, 1985; DeLisi, 1992).

5. Conclusion

The presence of MDR1 exon21 2677T variant and2677TT genotype suggested good treatment response mea-sured with positive PANSS subscale percentage changeswhile borderline associations were found with total andgeneral PANSS percentage changes. The presence ofMDR1 exon26 3435T variant and 3435TT genotypeshowed similar, yet not statistically significant associations.We found less representation of the G2677/C3435 haplo-type in schizophrenic female patients. Test result for link-age disequilibrium between loci was found to besignificant. However, because of the limitations of our sam-ple and weak statistical significance of the findings, theseresults should be taken as suggestive and replication studiesare warranted for further conclusions. Considering thatallele frequencies of both polymorphisms show large varia-tions across populations, along with differences in reportedlinkage disequilibrium between the two loci, replicationwith samples of different ethnic origin would provide valu-able results.

96 N. Bozina et al. / Journal of Psychiatric Research 42 (2008) 89–97

References

American Psychiatric Association. Practise guidelines for treatment ofpatients with schizophrenia, 2nd ed. American Journal of Psychiatry2004;161:1–56.

Arendt M, Rosenberg R, Foldager L, Perto G, Munk-Jorgensen P.Cannabis-induced psychosis and subsequent schizophrenia-spectrumdisorders: follow-up study of 535 incident cases. British Journal ofPsychiatry 2005;187:510–5.

Babigian HM. Schizophrenia. Epidemiology. In: Kaplan HY, Sadock BJ,editors. Textbook of psychiatry. 4th ed. Baltimore: Williams andWilkins Co.; 1985. p. 643–50.

Bagnall AM, Jones L, Grnnelly L, Lewis R, Glanville J, Gilbody S, et al.A systematic review of antipsychotic drugs in schizophrenia. HealthTechnology Assessment 2003;7:1–193.

Bai S, Stepkowski SM, Kahan BD, Brunner LJ. Metabolicinteraction between cyclosporine and sirolimus. Transplantation2004;77:1507–12.

Bernal ML, Sinues B, Fanlo A, Mayayo E. Frequency distribution ofC3435T mutation in exon 26 of the MDR1 gene in a Spanishpopulation. Therapeutic Drug Monitoring 2003;25:107–11.

Boulton DW, DeVane CL, Liston HL, Markowitz JS. Human recombi-nant P-glycoprotein specificity for atypical antipsychotics. Life Sci-ences 2002;71:163–9.

Boydell J, VAN Os J, Caspi A, Kennedy N, Giouroukou E, Fearon P,et al. Trends in cannabis use prior to first presentation with schizo-phrenia, in South-East London between 1965 and 1999. PsychologicalMedicine 2006;36:1441–6.

Cascorbi I, Gerloff T, Johne A, Meisel C, Hoffmeyer S, Schwab M, et al.Frequency of single nucleotide polymorphism in the P-glycoproteindrug transporter MDR1 gene in white subjects. Clinical Pharmacologyand Therapeutics 2001;69:169–74.

Chelule PK, Gordon M, Palanee T, Page T, Mosam A, Coovadia HM,et al. MDR1 and CYP3A4 polymorphisms among African, Indian,and white populations in KwaZulu-Natal, South Africa. ClinicalPharmacology and Therapeutics 2003;74:195–6.

Cordon-Cardo C, O’Brien JP, Casals D, Rittman-Grauer L, Biedler JL,Melamed MR, et al. Multidrug-resistance gene (P-glycoprotein) isexpressed by endothelial cells at blood–brain barrier sites. Proceedingsof the National Academy of Sciences of the United States of America1989;86:695–8.

DeLisi LE. The significance of age of onset for schizophrenia. Schizo-phrenia Bulletin 1992;18:209–15.

Dombrowski SM, Desai SY, Marroani M, Cucullo L, Goodrich K,Bingaman W, et al. Overexpression of multiple drug resistance genesin endothelial cells from patients with refractory epilepsy. Epilepsia2001;42:1501–6.

Eichelbaum M, Fromm MF, Schwab M. Clinical aspects of the MDR1(ABCB1) gene polymorphism. Therapeutic Drug Monitoring2004;26:180–5.

El Ela AA, Hartter S, Schmitt U, Hiemke C, Spahn-Langguth H,Langguth P. Identification of P-glycoprotein substrates and inhibitorsamong psychoactive compounds-implications for pharmacokinetics ofselected substrates. The Journal of Pharmacy and Pharmacology2004;56:967–75.

Excoffier L, Laval G, Schneider S. Arlequin ver. 3.01: an integratedsoftware package for population genetic data analysis. Switzer-land: University of Bern; 2006.

Fromm MF. Genetically determined differences in P-glycoprotein func-tion, implications for disease risk. Toxicology 2002;181–182:299–303.

Funae Y, Kishimoto W, Cho T, Niwa T, Hiroi T. CYP2D in the Brain.Drug Metabolism and Pharmacokinetics 2003;18:337–49.

Furuno T, Landi MT, Ceroni M, Caporaso N, Bernucci I, Nappi G, et al.MDR1 polymorphism and Parkinson’s disease. Drug MetabolismReviews 2001;33:129.

Guo SW, Thompson EA. Performing the exact test of Hardy-Weinbergproportions for multiple alleles. Biometrics 1992;48:361–72.

Hatta K, Nakamura H, Matsuzaki I. Acute-phase treatment in generalhospitals: clinical psychopharmacologic evaluation in first-episodeschizophrenia patients. General Hospital Psychiatry 2003;25:39–45.

Hedlund E, Gustafsson JA, Warner M. Cytochrome P450 in the brain.Current Drug Metabolism 2001;2:245–63.

Henquet C, Murray R, Linszen D, van Os J. The environment andschizophrenia: the role of cannabis use. Schizophrenia Bulletin2005;31:608–12.

Hesselink DA, van Schaik RH, van der Heiden IP, van der Werf M,Gregoor PJ, Lindemans J, et al. Genetic polymorphisms of theCYP3A4, CYP3A5 and MDR1 genes and pharmacokinetics of thecalcineurin inhibitors cyclosporine and tacrolimus. Clinical Pharma-cology and Therapeutics 2003;74:245–54.

Hitzl M, Drescher S, van der Kuip H, Schaffeler E, Fischer J, Schwab M,et al. The C3435T mutation in the human MDR1 gene is associatedwith altered efflux of the P-glycoprotein substrate rhodamine 123 fromCD56+ natural killer cells. Pharmacogenetics 2001;11:293–8.

Hoffmeyer S, Burk O, von Richter O, Arnold HP, Brockmoller J, Johne A,et al. Functional polymorphism of the human multidrug-resistancegene: multiple sequence variations and correlations of one allele withP-glycoprotein expression and activity in vivo. Proceedings of theNational Academy of Sciences of the United States of America2000;97:3473–8.

Holland ML, Panetta JA, Hoskins JM, Bebawyc M, Roufogalis BD, AllenJD, et al. The effects of cannabinoids on P-glycoprotein transport andexpression in multidrug resistant cells. Biochemical Pharmacology2006;71:1146–54.

Holm S. A simple sequentially rejective multiple test procedure.Scandinavian Journal of Statistics 1979;6:65–70.

Ieiri I, Hiroshi T, Otsubo K. The MDR1 (ABCB1) Gene polymorphismand its clinical implications. Clinical Pharmacokinetics2004;43:553–76.

Johne A, Kopke K, Gerloff T, Mai I, Rietbrock S, Meisel C, et al.Modulation of ready-state kinetics of digoxin by haplotypes of the P-glycoprotein MDR1 gene. Clinical Pharmacology and Therapeutics2002;72:584–93.

Kaplan HI, Sadock BJ. Synopsis of psychiatry. 9th ed. Philadelphia: Lip-pincott Williams & Wilkons; 2003.

Kim RB, Leake BF, Choo EF, Dresser GK, Kubba SV, Schwarz UI,et al. Identification of functionally variant MDR1 alleles amongEuropean Americans and African Americans. Clinical Pharmacologyand Therapeutics 2001;70:189–99.

Kim DW, Kim M, Lee SK, Kang R, Lee SY. Lack of association betweenC3435T nucleotide MDR1 genetic polymorphism and multidrug-resistant epilepsy. Seizure 2006;15:344–7.

Kurzawski M, Pawlik A, Gornik W, Drozdzik M. Frequency of commonMDR1 gene variants in a Polish population. Pharmacological Reports:PR 2006;58:35–40.

Kwan P, Brodie MJ. Potential role of drug transporters in the pathogen-esis of medically intractable epilepsy. Epilepsia 2005;46:224–35.

Kwan P, Sills GJ, Butler E, Gant TW, Meldrum BS, Brodie MJ. Regionalexpression of multidrug resistance genes in genetically epilepsy-pronerat brain after a single audiogenic seizure. Epilepsy 2002;43:1318–23.

Lewine RR, Strauss JS, Gift TE. Sex differences in age at first hospitaladmission for schizophrenia: fact or artifact? American Journal ofPsychiatry 1981;138:440–4.

Lin JH, Yamazaki M. Role of p-glycoprotein in pharmacokinetics: clinicalimplications. Clinical Pharmacokinetics 2003;42:59–98.

Lindholm A, Welsh M, Alton C, Kahan BD. Demographic factorsinfluencing cyclosporine pharmacokinetic parameters in patients withuremia: racial differences in bioavailbility. Clinical Pharmacology andTherapeutics 1992;52:359–71.

Ling V. Multidrug resistance: molecular mechanism and clinical relevance.Cancer Chemotherapy and Pharmacology 1997;40(Suppl.):S3–8.

Loscher W, Potschka H. Role of multidrug transporters in pharmacore-sistance to antiepileptic drugs. The Journal of Pharmacology andExperimental Therapeutics 2002;301:7–14.

N. Bozina et al. / Journal of Psychiatric Research 42 (2008) 89–97 97

Mahar Doan KM, Humphreys JE, Webster LO, Wring SA, Shampine LJ,Serabjit-Singh CJ, et al. Passive permeability and P-glycoprotein-mediated efflux differentiate central nervous system (CNS) and non-CNS marketed drugs. The Journal of Pharmacology and ExperimentalTherapeutics 2002;303:1029–37.

Mauri MC, Steinhilber CP, Marino M, Invernizzi E, Fiorentini A, CerveriG, et al. Clinical outcome and olanzapine plasma levels in acuteschizophrenia. European Psychiatry 2005;20:55–60.

Miller SA, Dykes DD, Polesky HF. A simple salting out procedure forextracting DNA from human nucleated cell. Nucleic Acids Research1988;16:1215.

Nakamura T, Sakaeda T, Horinouchi M, Tamura T, Aoyama N,Schirakawa T, et al. Effect of the mutation (C3435T) et exon 26 ofthe MDR1 gene on expression level of MDR1 messenger ribonucleicacid in duodenal enterocytes of healthy Japanese subjects. ClinicalPharmacology and Therapeutics 2002;71:297–303.

Panwala CM, Jones JC, Viney JL. A novel model of inflammatory boweldisease: mice deficient for the multiple in drug resistance gene, mdr1a,spontaneously develop colitis. Journal of Immunology1998;161:5733–44.

Prior TI, Baker GB. Interactions between the cytochrome P450 systemand the second-generation antipsychotics. Journal of Psychiatry andNeuroscience 2003;28:99–112.

Rice WR. Analyzing tables of statistical tests. Evolution 1989;43:223–5.Roberts RL, Joyce PR, Mulder RT, Begg EJ, Kennedy MA. A common

P-glycoprotein polymorphism is associated with nortriptyline inducedpostural hypotension in patients treated for major depression. ThePharmacogenomics Journal 2002;2:191–6.

Sakaeda T, Nakamura T, Horinouchi M, Kakumoto M, Ohmoto N,Sakai T, et al. MDR1 Genotype-related pharmacokinetics of digoxinafter single oral administration in healthy Japanese subjects. Pharma-ceutical Research 2001;18:1400–4.

Schinkel AH, Mayer U, Wagenaar E, Mol CA, van Deemter L, Smit JJ,et al. Normal viability and altered pharmacokinetics in mice lackingmdr1-type (drug-transporting) P-glycoproteins. Proceedings of theNational Academy of Sciences of the United States of America1997;94:4028–33.

Siddiqui A, Kerb R, Weale ME, Brinkmann U, Smith A, Goldstein DB,et al. Association of multidrug resistance in epilepsy with a polymor-phism in the drug-transporter gene ABCB1. New England Journal ofMedicine 2003;348:1442–8.

Sills GJ, Mohanraj R, Butler E, McCrindle S, Collier L, Wilson EA, et al.Lack of association between the C3435T polymorphism in the humanmultidrug resistance (MDR1) gene and response to antiepileptic drugtreatment. Epilepsia 2005;46:643–7.

Sisodiya SM, Lin WR, Harding BN, Squier MV, Thom M. Drugresistance in epilepsy: expression of drug resistance proteins incommon causes of refractory epilepsy. Brain 2002;125:22–31.

Slatkin M, Excoffier L. Testing for linkage disequilibrium in genotypicdata using the EM algorithm. Heredity 1996;76:377–83.

Takao T, Tachikawa H, Kawanishi Y, Katano Y, Sen B, Homma M,et al. Association of treatment-resistant schizophrenia with theG2677A/T and C3435T polymorphisms in the ATP-binding cassettesubfamily B member 1 gene. Psychiatric Genetics 2006;16:47–8.

Tanabe M, Ieiri I, Nagata N, Inoue K, Ito S, Kanamori Y, et al.Expression of P-glycoprotein in human placenta: relation to geneticpolymorphism of the multidrug resistance (MDR)-1 gene. The Journalof Pharmacology and Experimental Therapeutics 2001;297:1137–43.

Tang K, Ngoi SM, Gwee PC, Chua JM, Lee EJ, Chong SS, et al. Distincthaplotype profiles and strong linkage disequilibrium at the MDR1multidrug transporter gene locus in three ethnic Asian populations.Pharmacogenetics 2002;12:437–50.

Thompson JS, Koszdin K, Bernadsrs CM. Opiate-induced analgesia isincreased and prolonged in mice lacking P-glycoprotein. Anaesthesi-ology 2000;92:1392–9.

Uhr M, Grauer MT. Abcb1ab P-glycoprotein is involved in the uptake ofcitalopram and trimipramine into the brain of mice. Journal ofPsychiatric Research 2003;37:179–85.

Uhr M, Grauer MT, Yassouridis A, Ebinger M. Blood-brain barrierpenetration and pharmacokinetics of amitriptyline and its metabolitesin p-glycoprotein (abcb1ab) knock-out mice and controls. Journal ofPsychiatric Reserch 2006; January 4 [Epub ahead of print].

Wang JS, Taylor R, Ruan Y, Donovan JL, Markowitz JS, Lindsay DeVane C. Olanzapine penetration into brain is greater in transgenicAbcb1a P-glycoprotein-deficient mice than FVB1 (wild-type) animals.Neuropsychopharmacology 2004;29:551–7.

Wang JS, Zhu HJ, Markowitz JS, Donovan JL, Devane CL. Evaluation ofantipsychotic drugs as inhibitors of multidrug resistance transporter P-glycoprotein. Psychopharmacology (Berl) 2006;187:415–23.

Weiss J, Gregor Dormann SM, Martin-Facklam M, Kerpen CJ, Ketabi-Kiyanvash N, Haefeli WE. Inhibition of P-glycoprotein by newerantidepressants. The Journal of Pharmacology and ExperimentalTherapeutics 2003;305:197–204.

Yamauchi A, Ieiri I, Kataoka Y, Tanabe M, Nishizaki T, Oishi R, et al.Neurotoxicity induced by tacrolimus after liver transplantation.Relation to genetic polymorphisms of the ABCB1 (MDR1) gene.Transplantation 2002;74:571–8.

Yasui-Furukori N, Saito M, Nakagami T, Kaneda A, Tateishi T, KanekoS. Association between multidrug resistance 1 (MDR1) gene polymor-phisms and therapeutic response to bromperidol in schizophrenicpatients: a preliminary study. Progress in Neuro-Psychopharmacologyand Biological Psychiatry 2006;30:229–86.

Zhu HJ, Wang JS, Markowitz JS, Donovan JL, Gibson BB, Gefroh HA,et al. Characterization of P-glycoprotein inhibition by major cannab-inoids from marijuana. Journal of Pharmacology and ExperimentalTherapeutics 2006;317:850–7.