Relative contribution of CYP3A to amitriptyline clearance in humans: in vitro and in vivo studies.

The relative contribution of cytochrome P450 3A (CYP3A) to the oral clearance of amitriptyline in humans has been assessed using a combination of in vitro approaches together with a clinical pharmacokinetic interaction study using the CYP3A-selective inhibitor ketoconazole. Lymphoblast-expressed CYPs were used to study amitriptyline N-demethylation and E-10 hydroxylation in vitro. The relative activity factor (RAF) approach was used to predict the relative contribution of each CYP isoform to the net hepatic intrinsic clearance (sum of N-demethylation and E-10 hydroxylation). Assuming no extrahepatic metabolism, the model-predicted contribution of CYP3A to net intrinsic clearance should equal the fractional decrement in apparent oral clearance of amitriptyline upon complete inhibition of the enzyme. This hypothesis was tested in a clinical study of amitriptyline (50 mg, p.o.) with ketoconazole (three 200 mg doses spaced 12 hours apart) in 8 healthy volunteers. The RAF approach predicted CYP2C19 to be the dominant contributor (34%), with a mean 21% contribution of CYP3A (range: 8%-42% in a panel of 12 human livers). The mean apparent oral clearance of amitriptyline in 8 human volunteers was decreased from 2791 ml/min in the control condition to 2069 ml/min with ketoconazole. The average 21% decrement (range: 2%-40%) was identical to the mean value predicted in vitro using the RAF approach. The central nervous system (CNS) sedative effects of amitriptyline were slightly greater when ketoconazole was coadministered, but the differences were not statistically significant. In conclusion, CYP3A plays a relatively minor role in amitriptyline clearance in vivo, which is consistent with in vitro predictions using the RAF approach.

Serotonin transporter polymorphisms: no association with response to antipsychotic treatment, but associations with the schizoparanoid and residual subtypes of schizophrenia.

The human serotonin transporter gene (5-HTT) demonstrates two polymorphisms with possible functional impact: a 44-bp insertion/deletion polymorphism of the promoter region and a 17-bp variable number of tandem repeat polymorphism (VNTR) in intron 2 (STin2). Such genetic polymorphisms in the serotoninergic system may increase the susceptibility to schizophrenia or may serve as predictors of therapeutic response. We therefore analyzed these polymorphisms as susceptibility factors for schizophrenia by comparison of 684 schizophrenic inpatients with 587 healthy controls. We furthermore compared the therapeutic outcome of schizophrenic patients differentiated by the 5-HTT genotypes. Schizo-affective patients were more frequently homozygous for the 44-bp insertion allele (Odds ratio, OR: 1.6, 95% confidence interval, CI: 1.1--2.3, P < 0.03) than were all other schizophrenic patients and controls. The 17-bp VNTR alleles found were: STin2.7, 9, 10, and 12. Sequence analysis revealed seven different sequence motifs with an invariable arrangement. Patients with schizo-paranoid schizophrenia were more frequently homozygous for the STin2.12 allele than were controls (OR: 1.4, CI: 1.1--1.8, P < 0.007) and all other schizophrenic patients (OR: 1.6, CI: 1.2--2.3). The STin2.9 allele represented a risk factor for the residual subtype of schizophrenia (OR: 6.4, CI: 2.5--16.2, P < 0.001). On the basis of global clinical impressions, as well as measurements with the positive and negative syndrome scale we found no association of the polymorphisms with therapeutic response. In conclusion, the 44-bp polymorphism may be associated with the schizo-affective and the 17-bp VNTR with the residual and schizo-paranoid subtype of schizophrenia, findings which require further biochemical and epidemiological confirmation.

Cytochrome P-450 enzymes and FMO3 contribute to the disposition of the antipsychotic drug perazine in vitro.

RATIONALE: Perazine (PER) is a phenothiazine antipsychotic drug frequently used in Germany that undergoes extensive metabolism. OBJECTIVES AND METHODS: To anticipate metabolic drug interactions and to explore the relevance of polymorphisms of metabolic enzymes, perazine-N-demethylation and perazine-N-oxidation were investigated in vitro using human liver microsomes and cDNA expressed enzymes. RESULTS: CYP3A4 and CYP2C9 were identified as the major enzymes mediating PER-N-demethylation. At 10 microM PER, a concentration consistent with anticipated in vivo liver concentrations, CYP3A4 and CYP2C9 contributed 50% and 35%, respectively, to PER-N-demethylation. With increasing PER concentrations, contribution of CYP2C9 decreased and CYP3A4 became more important. In human liver microsomes, PER-N-demethylation was inhibited by ketoconazole (>40%) and sulfaphenazole (16%). Allelic variants of recombinant CYP2C9 showed differences in PER-N-demethylase activity. The wild type allele CYP2C9*1 was the most active variant. Maximal activities of CYP2C9*2 and CYP2C9*3 were 88% and 18%, respectively, compared to the wild type activity. Perazine-N-oxidation was mainly mediated by FMO3. In the absence of NADPH, heat treatment of microsomes abolished PER-N-oxidase activity. Methimazole inhibited PER-N-oxidation, while CYP specific inhibitors had no inhibitory effect. Perazine is a potent inhibitor of dextromethorphan-O-demethylase, S-mephenytoin-hydroxylase, alprazolam-4-hydroxylase, phenacetin-O-deethylase and tolbutamide-hydroxylase activity in human liver microsomes. CONCLUSIONS: Alterations in the activity of CYP3A4, CYP2C9 and FMO3 through genetic polymorphisms, enzyme induction or inhibition bear the potential to cause clinically significant changes in perazine clearance. PER may alter the clearance of coadministered compounds metabolized by CYP2D6, CYP2C19, CYP2C9, CYP3A4 and CYP1A2.

Long-term therapeutic drug monitoring of clozapine and metabolites in psychiatric in- and outpatients.

RATIONALE: Clozapine is a unique antipsychotic drug, outstanding for its lack of extrapyramidal side-effects and its superior efficacy in refractory schizophrenia. However, an unambiguous concentration-response relationship has not yet been established. OBJECTIVE: We investigated serum concentrations of clozapine, norclozapine and clozapine-N-oxide in psychiatric in- and outpatients to identify particular metabolic patterns in clozapine responders and non-responders and putative threshold levels for clozapine response. METHODS: Psychiatric assessments, CYP2D6 genotype, and weekly serum concentrations of clozapine, norclozapine and clozapine-N-oxide were obtained in 34 adult schizophrenic in-and outpatients (18 men, 16 women) during 10 weeks of clozapine treatment with a naturalistic dose design. RESULTS: Responders (n=21) displayed significantly lower serum concentrations of clozapine corrected for dose compared to non-responders (n=13; P<0.05), while none of the other parameters (absolute clozapine concentration, metabolite ratios, gender) were different. Smokers had significantly lower dose-corrected clozapine concentrations. A positive correlation was observed between age and average steady state clozapine concentrations. CONCLUSIONS: These findings indicate a possible link between CYP activity and response to clozapine that is not mediated through differences in serum concentrations. No clinically meaningful pattern in serum parameters could be identified that differentiates responders from non-responders. Thus, clozapine TDM seems ineffective for predicting clinical response. Smoking behavior is a major determinant of clozapine clearance while CYP2D6 genotype does not impact clozapine disposition.

Nefazodone, meta-chlorophenylpiperazine, and their metabolites in vitro: cytochromes mediating transformation, and P450-3A4 inhibitory actions.

RATIONALE: Understanding of the mechanisms of biotransformation of antidepressant drugs, and of their capacity to interact with other medications, is of direct relevance to rational clinical psychopharmacology. OBJECTIVES: To determine the human cytochromes P450 mediating the metabolism of nefazodone, and the inhibitory activity of nefazodone and metabolites versus human P450-3A. METHODS: Biotransformation of nefazodone to its metabolic products, and of meta-chlorophenylpiperazine (mCPP) to para-hydroxy-mCPP, was studied in vitro using human liver microsomes and heterologously expressed human cytochromes. Nefazodone and metabolites were also tested as inhibitors of alprazolam hydroxylation, reflecting activity of cytochrome P450-3A isoforms. RESULTS: mCPP and two hydroxylated derivatives were the principal metabolites formed from nefazodone by liver microsomes. Metabolite production was strongly inhibited by ketoconazole or troleandomycin (relatively specific P450-3A inhibitors), and by an anti-P450-3A antibody. Only heterologously expressed human P450-3A4 mediated formation of nefazodone metabolites from the parent compound. Nefazodone, hydroxy-nefazodone, and para-hydroxy-nefazodone were strong 3A inhibitors, being more potent than norfluoxetine and fluvoxamine, but less potent than ketoconazole. The triazoledione metabolite and mCPP had weak or negligible 3A-inhibiting activity. Formation of parahydroxy-mCPP from mCPP was mediated by heterologously expressed P450-2D6; in liver microsomes, the reaction was strongly inhibitable by quinidine, a relatively specific 2D6 inhibitor. CONCLUSION: The complex parallel biotransformation pathways of nefazodone are mediated mainly by human cytochrome P450-3A, whereas clearance of mCPP is mediated by P450-2D6. Nefazodone and two of its hydroxylated metabolites are potent 3A inhibitors, accounting for pharmacokinetic drug interactions of nefazodone with 3A substrate drugs such as triazolam and alprazolam.

A double-blind comparison of lorazepam and oxazepam in psychomotor retardation and mutism.

BACKGROUND: An increasing number of case reports indicate a superior therapeutic response of catatonialike symptoms, such as severe psychomotor disturbance and mutism, associated with psychiatric disorder to the benzodiazepine lorazepam (LO). Equivocal results, however, are also reported with regard to other benzodiazepines for the treatment of this syndrome. The purpose of this study was to compare the effects of LO and oxazepam (OX), benzodiazepines with comparable pharmacokinetics, on psychomotor retardation and mutism associated with psychiatric disorder. METHODS: Twenty-one hospitalized patients with severe psychomotor retardation and mutism were treated with 2 mg LO and 60 mg OX in a double-blind crossover study design. RESULTS: Both benzodiazepines significantly reduced psychomotor symptoms. When administered for the first time, 4 of 7 patients with LO and 6 of 10 patients with OX improved at least 50% on visual analog scale (VAS) rating. Reduction in symptoms was significant with LO and OX treatment on either day of treatment. The second time, however, LO was significantly better compared with OX in alleviating the target symptoms. CONCLUSIONS: Both OX and LO are effective for the treatment of psychomotor retardation. Thus, the beneficial effect of LO on psychomotor retardation and mutism is not a unique pharmacodynamic property but more likely due to its pharmacokinetic profile. The differential effect of the two benzodiazepines on the second day of treatment warrants further clarification. Several hypotheses are evaluated.

Testosterone, gonadotropin, and cortisol secretion in male patients with major depression.

OBJECTIVE: Previous studies of sex hormone concentrations in depression yielded inconsistent results. However, the activation of the hypothalamic-pituitary-adrenal system seen in depression may negatively affect gonadal function at every level of regulation. The objective of this study was to explore whether major depressive episodes are indeed associated with an alteration of gonadal function. METHODS: Testosterone, pulsatile LH secretion, FSH, and cortisol were assessed using frequent sampling during a 24-hour period in 15 male inpatients with major depression of moderate to high severity and in 22 healthy comparison subjects (age range 22-85 years). RESULTS: An analysis of covariance model showed that after adjustment for age only, daytime testosterone (p < .01), nighttime testosterone (p < .05), and 24-hour mean testosterone secretion (p < .01) were significantly lower in the depressed male inpatients. There was also a trend for a decreased LH pulse frequency in the depressed patients (p < .08). CONCLUSIONS: Gonadal function may be disturbed in men with a depressive episode of moderate to high severity.

Extrapolating in vitro data on drug metabolism to in vivo pharmacokinetics: evaluation of the pharmacokinetic interaction between amitriptyline and fluoxetine.

Recently, models have been proposed to extrapolate in vitro data on the influence of inhibitors on drug metabolism to in vivo decrement in drug clearance. Many factors influence drug clearance such as age, gender, habits, diet, environment, liver disease, heredity, and other drugs. In vitro investigation of hepatic cytochrome P450 activity has generally centered on genetic influences and interactions with other drugs. This group of enzymes is involved in many, although not all, drug interactions. The interaction of amitriptyline and fluoxetine is an example. Of the different in vitro paradigms, interaction studies utilizing human liver microsomal preparations have proved to be the most generally applicable for in vitro scaling models. Assuming Michaelis-Menten conditions and applying nonlinear regression, a hybrid inhibition constant (Ki) can be generated that allows classification of the inhibitory potency of an inhibitor toward a specific reaction. This constant is largely independent of the substrate concentration, but in vivo relevance is critically dependent on the inhibitor concentration in the site of metabolic activity, the liver cell cytosol. Many lipophilic drugs are extensively bound to plasma protein but, nonetheless, demonstrate extensive partitioning into liver tissue. This is not compatible with diffusion only of the unbound drug fraction into liver cells. The introduction of a partition factor, based on data from a number of possible sources, provided a reasonable basis for the scaling of in vitro data to in vivo conditions. Many interactions could be reconstructed or predicted with greater accuracy and clinical relevance for interactions such as terfenadine or midazolam and ketoconazole. Even for less marked interactions such as amitriptyline and fluoxetine, this model provides a forecast consistent with the clinically observed range of 22-45% reduction in oral clearance, although this interaction is complicated by the presence of two inhibitors, fluoxetine and norfluoxetine. The concept of in vitro-in vivo scaling is promising and might ultimately yield a fast and more cost-effective screening for drug interactions with reduced human drug exposure and risk.

O- and N-demethylation of venlafaxine in vitro by human liver microsomes and by microsomes from cDNA-transfected cells: effect of metabolic inhibitors and SSRI antidepressants.

The biotransformation of venlafaxine (VF) into its two major metabolites, O-desmethylvenlafaxine (ODV) and N-desmethylvenlafaxine (NDV) was studied in vitro with human liver microsomes and with microsomes containing individual human cytochromes from cDNA-transfected human lymphoblastoid cells. VF was coincubated with selective cytochrome P450 (CYP) inhibitors and several selective serotonin reuptake inhibitors (SSRIs) to assess their inhibitory effect on VF metabolism. Formation rates for ODV incubated with human microsomes were consistent with Michaelis-Menten kinetics for a single-enzyme mediated reaction with substrate inhibition. Mean parameters determined by non-linear regression were: Vmax = 0.36 nmol/min/mg protein, K(m) = 41 microM, and Ks 22901 microM (Ks represents a constant which reflects the degree of substrate inhibition). Quinidine (QUI) was a potent inhibitor of ODV formation with a Ki of 0.04 microM, and paroxetine (PX) was the most potent SSRI at inhibiting ODV formation with a mean Ki value of 0.17 microM. Studies using expressed cytochromes showed that ODV was formed by CYP2C9, -2C19, and -2D6. CYP2D6 was dominant with the lowest K(m), 23.2 microM, and highest intrinsic clearance (Vmax/K(m) ratio). No unique model was applicable to the formation of NDV for all four livers tested. Parameters determined by applying a single-enzyme model were Vmax = 2.14 nmol/min/mg protein, and K(m) = 2504 microM. Ketoconazole was a potent inhibitor of NDV production, although its inhibitory activity was not as great as observed with pure 3A substrates. NDV formation was also reduced by 42% by a polyclonal rabbit antibody against rat liver CYP3A1. Studies using expressed cytochromes showed that NDV was formed by CYP2C9, -2C19, and -3A4. The highest intrinsic clearance was attributable to CYP2C19 and the lowest to CYP3A4. However the high in vivo abundance of 3A isoforms will magnify the importance of this cytochrome. Fluvoxamine (FX), at a concentration of 20 microM, decreased NDV production by 46% consistent with the capacity of FX to inhibit CYP3A, 2C9, and 2C19. These results are consistent with previous studies that show CYP2D6 and -3A4 play important roles in the formation of ODV and NDV, respectively. In addition we have shown that several other CYPs have important roles in the biotransformation of VF.

Simultaneous assessment of CYP3A4 and CYP1A2 activity in vivo with alprazolam and caffeine.

Alprazolam (ALP) and caffeine (CAF) were suggested as probe drugs for the activities of CYP3A4 and CYP1A2, respectively. We investigated the disposition of oral ALP (1 mg) and CAF (100 mg) in 17 normal volunteers to establish and validate a procedure for the simultaneous assessment of CYP3A4 and CYP1A2 enzyme activity. Nine received ALP alone, ALP and CAF and CAF alone in an open three-way crossover study to test for pharmacokinetic interaction. Four received ALP after a 2-day pretreatment with ketoconazole, an inhibitor of CYP3A4, and four normal volunteers received ALP after 4 days of rifampin, an inducer of CYP3A4. AUC values of ALP and CAF administered alone were not different from AUC values when both drugs were administered combined, indicating that there is no metabolic interaction. The ratio formed of paraxanthine and CAF correlated significantly with systemic CAF clearance at 3, 4, 6, 8, 10 and 24 h. There was a strong correlation between AUC values of ALP and CAF and the plasma concentration obtained 6, 8, 10, or 24 h after ingestion of the drug. Ketoconazole and rifampin pretreatment significantly changed AUC values of ALP (mean AUC values in microg/l h: ALP = 242.2, ALP + ketoconazole = 426.2, ALP + rifampin = 28.4, ANOVA F = 17.7, P < 0.001). We conclude that ALP and CAF can be administered simultaneously for the assessment of CYP activity. Plasma concentrations 6, 8, 10, and 24 h after drug ingestion reflect AUC of ALP and CAF and therefore in-vivo CYP3A4 and CYP1A2 activity, respectively.

Multiple human cytochromes contribute to biotransformation of dextromethorphan in-vitro: role of CYP2C9, CYP2C19, CYP2D6, and CYP3A.

Cytochromes mediating the biotransformation of dextromethorphan to dextrorphan and 3-methoxymorphinan, its principal metabolites in man, have been studied by use of liver microsomes and microsomes containing individual cytochromes expressed by cDNA-transfected human lymphoblastoid cells. In-vitro formation of dextrorphan from dextromethorphan by liver microsomes was mediated principally by a high-affinity enzyme (Km (substrate concentration producing maximum reaction velocity) 3-13 microM). Formation of dextrorphan from 25 microM dextromethorphan was strongly inhibited by quinidine (IC50 (concentration resulting in 50% inhibition) = 0.37 microM); inhibition by sulphaphenazole was approximately 18% and omeprazole and ketoconazole had minimal effect. Dextrorphan was formed from dextromethorphan by microsomes from cDNA-transfected lymphoblastoid cells expressing CYP2C9, -2C19, and -2D6 but not by those expressing CYP1A2, -2E1 or -3A4. Despite the low in-vivo abundance of CYP2D6, this cytochrome was identified as the dominant enzyme mediating dextrorphan formation at substrate concentrations below 10 microM. Formation of 3-methoxy-morphinan from dextromethorphan in liver microsomes proceeded with a mean Km of 259 microM. For formation of 3-methoxymorphinan from 25 microM dextromethorphan the IC50 for ketoconazole was 1.15 microM; sulphaphenazole, omeprazole and quinidine had little effect. 3-Methoxymorphinan was formed by microsomes from cDNA-transfected lymphoblastoid cells expressing CYP2C9, -2C19, -2D6, and -3A4, but not by those expressing CYP1A2 or -2E1. CYP2C19 had the highest affinity (Km = 49 microM) whereas CYP3A4 had the lowest (Km = 1155 microM). Relative abundances of the four cytochromes were determined in liver microsomes by use of the relative activity factor approach. After adjustment for relative abundance, CYP3A4 was identified as the dominant enzyme mediating 3-methoxymorphinan formation from dextromethorphan, although CYP2C9 and -2C19 were estimated to contribute to 3-methoxymorphinan formation, particularly at low substrate concentrations. Although formation of dextrorphan from dextromethorphan appears to be sufficiently specific to be used as an in-vitro or in-vivo index reaction for profiling of CYP2D6 activity, the findings raise questions about the specificity of 3-methoxymorphinan formation as an index of CYP3A activity.

Cerebrospinal fluid concentrations of corticotropin-releasing hormone, vasopressin, and somatostatin in depressed patients and healthy controls: response to amitriptyline treatment.

The effect of amitriptyline upon hypothalamic-pituitary-adrenal [HPA]-system-regulating neuropeptides (corticotropin-releasing hormone [CRH], vasopressin, somatostatin) was studied in a group of depressed elderly patients and controls. A first lumbar puncture was performed in 37 depressed in-patients. This was followed by a 6-week medication phase with amitriptyline. Upon its completion a second cerebrospinal fluid (CSF) sample was obtained in 18 of these 37 patients. In 25 healthy controls a first lumbar puncture was done eleven of these individuals agreed to take 75 mg/d amitriptyline for 6 weeks and to participate in the follow-up CSF study. Within the group of depressed patients amitriptyline led to a significant decrease of CSF CRH in treatment responders only (F1, 16 = 5.2; P < 0.02). Also, in normal controls CSF CRH concentration tended to decrease with amitriptyline treatment (t-test; P < 0.09). No effects of amitriptyline upon vasopressin or somatostatin were observed. In normal controls (r = 0.4; P < 0.02) and in patients (r = 0.4; P < 0.03) age correlated positively with baseline CSF somatostatin. A trend for CSF CRH to increase with aging was found only in controls (r = 0.3; P < 0.09); patients did not show a significant association here. Finally, CSF neuropeptide concentration at baseline did not differ between the group of depressed patients and healthy controls. Our study corroborates the evolving concept that antidepressants effect various components of the HPA system with the net result of a reduction in its activity. In addition, we found CSF CRH and CSF somatostatin concentrations to be better reflections of age than of depression and, finally, that during aging and during depression the HPA system changes in similar directions.

Endurance training and its effect upon the activity of the GH-IGFs system in the elderly.

There is an age-associated decline in the activity of the GH-IGFs system. However, so far, it has not been studied, whether this decline in somatotrophic activity might be preventable by intensive exercising. We studied 11 elderly male (50-78 years) marathon runners and 10 age-matched male (52-73 years) sedentary controls to evaluate plasma concentrations of GH, total and free IGF-I and IGF-II and of IGF-binding protein-1 (IGFBP-1), IGFBP-2, IGFBP-3 and insulin. When comparing the two groups (runners vs controls) no differences were found in GH (1.0 +/- 1.2 vs 1.3 +/- 1.3 microg/l [mean +/- SD]), IGF-1 (115 +/- 23 vs 113 +/- 21 microg/l), IGF-II (961 +/- 192 vs 864 +/- 125 microg/l), free IGF-1 (227 +/- 80 vs 318 +/- 146 ng/l), free IGF-II (563 +/- 249 vs 492 +/- 108 ng/l), IGFBP-3 (2403 +/- 515 vs 2307 +/- 326 microg/l) or insulin (26 +/- 13 vs 27 +/- 18 mU/l). However, IGFBP-1 (4.44 +/- 2.61 vs 2.28 +/- 0.93 microg/l) and IGFBP-2 (493 +/- 143 vs 340 +/- 186 microg/l) were found to be significantly increased in marathon runners. In conclusion, our findings do not support the hypothesis that the age-associated decline in GH, IGF-1 and IGFBP-3 may be preventable by intensive endurance training. However, marathon running affects the regulation of the GH-IGFs system activity at the level of IGFBP-1 and -BP-2.

The combined dexamethasone/corticotropin-releasing hormone stimulation test is more closely associated with features of diurnal activity of the hypothalamo-pituitary-adrenocortical system than the dexamethasone suppression test.

BACKGROUND: The dexamethasone suppression test (DST) is a widely used endocrine test in psychiatry, but was reported to not allow reliable inferences with regard to the basal activity of the hypothalamo-pituitary-adrenocortical (HPA) system. We compared the association of the standard DST and the combined dexamethasone/corticotropin-releasing hormone (DEX/CRH) challenge with parameters of diurnal cortisol profiles. METHODS: We performed a DEX/CRH challenge and 24-hour cortisol profiles in 25 depressed patients (mean age: 47.4 +/- 16.0 years) and 33 age-matched healthy controls (mean age: 51.4 +/- 19.3 years). RESULTS: A path analysis showed cortisol area under the curve (AUC) after CRH (= DEX/CRH status) to be dependent upon minimal 24-hour cortisol and evening frequency of pulsatile cortisol release. In contrast, postdexamethasone cortisol (= DST status) was related to 24-hour mean cortisol. Simple linear regressions supported an association of cortisol AUC with several parameters of the diurnal cortisol profiles, which was not true for the standard DST. CONCLUSIONS: We conclude that the combined DEX/CRH challenge test is more closely associated with the activity of the HPA system than the standard DST in healthy and depressed subjects.

Five distinct human cytochromes mediate amitriptyline N-demethylation in vitro: dominance of CYP 2C19 and 3A4.

The human cytochromes P450 (CYPs) mediating amitriptyline N-demethylation have been identified using a combination of enzyme kinetic and chemical inhibition studies. Amitriptyline was N-demethylated to nortriptyline by microsomes from cDNA transfected human lymphoblastoid cells expressing human CYPs 1A2, 2C9, 2C19, 2D6, and 3A4. CYP 2E1 showed no detectable activity. While CYP 2C19 and CYP 2D6 showed high affinity, CYP 3A4 showed low affinity; CYP 2C9 and 1A2 showed intermediate affinities. Based on these kinetic parameters and estimated relative abundance of the different CYPs in human liver, CYP 2C19 was identified as the major amitriptyline N-demethylase at low (therapeutically relevant) amitriptyline concentrations, whereas CYP 3A4 may be more important at higher amitriptyline concentrations. Chemical inhibition studies with ketoconazole and omeprazole indicate that CYP 3A4 is the major amitriptyline N-demethylase at 100 mumol/L amitriptyline, while CYP 2C19 is equally important at a substrate concentration of 5 mumol/L. The CYP 1A2 inhibitor alpha-naphthoflavone and the CYP 2C9 inhibitor sulfaphenazole produced much less inhibition of amitriptyline N-demethylation at both substrate concentrations. Quinidine produced no detectable inhibition. The kinetics of amitriptyline N-demethylation by human liver microsomes were consistent with a two enzyme model, with the high affinity component exhibiting Michaelis Menten kinetics and the low affinity component exhibiting Hill enzyme kinetics. No difference was apparent in the kinetics of amitriptyline N-demethylation in two liver samples with low levels of CYP 2C19 activity compared with two other samples with relatively normal 2C19 activity. This may reflect the importance of higher substrate concentration values in estimation of kinetic parameters in vitro.

In vitro approaches to predicting drug interactions in vivo.

In vitro metabolic models using human liver microsomes can be applied to quantitative prediction of in vivo drug interactions caused by reversible inhibition of metabolism. One approach utilizes in vitro Ki, values together with in vivo values of inhibitor concentration to forecast in vivo decrements of clearance caused by coadministration of inhibitor. A critical limitation is the lack of a general scheme for assigning intrahepatic exposure of enzyme to inhibitor or substrate based only on plasma concentration; however, the assumption that plasma protein binding necessarily restricts hepatic uptake is not tenable. Other potential limitations include: flow-dependent hepatic clearance, "mechanism-based" chemical inhibition, concurrent induction, or a major contribution of gastrointestinal P450-3A isoforms to presystemic extraction. Nonetheless, the model to date has provided reasonably accurate forecasts of in vivo inhibition of clearance of several substrates (desipramine, terfenadine, triazolam, alprazolam, midazolam) by coadministration of selective serotonin reuptake-inhibitor antidepressants and azole antifungal agents. Such predictive models deserve further evaluation, since they may ultimately yield more cost-effective and expeditious screening for drug interactions, with reduced human drug exposure and risk.

Insulin-like growth factor-I (IGF-I) plasma concentrations are increased in depressed patients.

There is some evidence that the somatotrophic system in depression, as assessed by basal growth hormone (GH) concentrations and by GH releasing hormone (GHRH) challenge, might be dysfunctional. However, the rather limited data have been inconclusive so far and plasma concentrations of both insulin-like growth factor-1 (IGF-I) and binding proteins (IGFBP 1 to IGFBP-6) have not been measured simultaneously in depressed patients. We studied 24 severely depressed patients and 33 healthy controls and estimated 24-hour mean plasma cortisol, six-hour evening mean plasma growth hormone (GH), morning plasma IGF-I, IGFBP 2 and 3 and GH-binding protein (GH-BP). Twenty-four-hour mean cortisol (306 +/- 69 vs. 196 +/- 30 nmol/l, p < .001) and IGF-I (157 +/- 40 vs. 120 +/- 33 micrograms/l, p < .01) plasma concentrations were found to be significantly increased in depressed patients, while there was no difference in GH or binding proteins between both groups. MANOVA analysis revealed age and diagnosis to have main effects upon plasma IGF-I. Especially young age and a diagnosis of major depression are associated with higher plasma IGF-I. After treatment only patients in remission had attenuated IGF-I plasma concentrations. We conclude that plasma IGF-I is increased in acutely depressed patients similar to other states of hypercortisolemia.

Human cytochromes mediating N-demethylation of fluoxetine in vitro.

Biotransformation of the selective serotonin reuptake inhibitor antidepressant, fluoxetine, to its principal metabolite, norfluoxetine, was evaluated in human liver microsomes and in microsomes from transfected cell lines expressing pure human cytochromes. In human liver microsomes, formation of norfluoxetine from R,S-fluoxetine was consistent with Michaelis-Menten kinetics (mean K(m) = 33 microM), with evidence of substrate inhibition at high substrate concentrations in a number of cases. The reaction was minimally inhibited by coincubation with chemical probes inhibitory for P450-2D6 (quinidine), -1A2 (furafylline, alpha-naphthoflavone), and -2E1 (diethyldithiocarbamate). Substantial inhibition was produced by coincubation with sulfaphenazole (Ki = 2.8 microM), an inhibitory probe for P450-2C9, and by ketoconazole (Ki = 2.5 microM) and fluvoxamine (Ki = 5.2 microM). However, ketoconazole, relatively specific for P450-3A isoforms only at low concentrations, reduced norfluoxetine formation by only 20% at 1 microM, and triacetyloleandomycin (> or = 5 microM) reduced the velocity by only 20-25%. Microsomes from cDNA-transfected human lymphoblastoid cells containing human P450-2C9 produced substantial quantities of norfluoxetine when incubated with 100 microM fluoxetine. Smaller amounts of product were produced by P450-2C19 and -2D6, but no product was produced by P450-1A2, -2E1, or 3A4. Cytochrome P450-2C9 appears to be the principal human cytochrome mediating fluoxetine N-demethylation. P450-2C19 and -3A may make a further small contribution, but P450-2D6 is unlikely to make an important contribution.