Svein G. Dahl

Mechanism-Based Pharmacokinetic–Pharmacodynamic Modeling—A New Classification of Biomarkers

In recent years, pharmacokinetic/pharmacodynamic (PK/PD) modeling has developed from an empirical descriptive discipline into a mechanistic science that can be applied at all stages of drug development. Mechanism-based PK/PD models differ from empirical descriptive models in that they contain specific expressions to characterize processes on the causal path between drug administration and effect. Mechanism-based PK/PD models have much improved properties for extrapolation and prediction. As such, they constitute a scientific basis for rational drug discovery and development. In this report, a novel classification of biomarkers is proposed. Within the context of mechanism-based PK/PD modeling, a biomarker is defined as a measure that characterizes, in a strictly quantitative manner, a process, which is on the causal path between drug administration and effect. The new classification system distinguishes seven types of biomarkers: type 0, genotype/phenotype determining drug response; type 1, concentration of drug or drug metabolite; type 2, molecular target occupancy; type 3, molecular target activation; type 4, physiological measures; type 5, pathophysiological measures; and type 6, clinical ratings. In this paper, the use of the new biomarker classification is discussed in the context of the application of mechanism-based PK/PD analysis in drug discovery and development.

Plasma Level Monitoring of Antipsychotic Drugs Clinical Utility

SummaryThe steady-state plasma concentrations of antipsychotic drugs show large interpatient variations but remain relatively stable from day to day in each individual patient. Monitoring of antipsychotic drug concentrations in plasma might be of value provided the patients are treated with only 1 antipsychotic drug. Some studies have reported a relationship between therapeutic response and serum antipsychotic drug ‘concentrations’ as measured using the radioreceptor assay (RRA) method, which measures dopamine receptor-blocking activity in plasma. Most studies, however, have failed to demonstrate such a relationship, and the RRA does not seem to provide the generally useful tool for plasma concentration monitoring of antipsychotic drugs that was hoped for initially. A lack of correlation between dopamine receptor-blocking activity in plasma and therapeutic response may be due to differences in the blood-brain distribution of both antipsychotic drugs and their active metabolites.Chemical assay methods (e.g. GLC and HPLC) have been used in studies which examined the relationships between therapeutic response and antipsychotic drug concentrations in red blood cells and in plasma. It seems that for these drugs, measuring red blood cell concentrations does not offer any significant advantage over measuring plasma concentrations. Reasonably controlled studies of plasma concentration-response relationships using randomly allocated, fixed dosages of chlorpromazine, fluphenazine, haloperidol, perphenazine, sulpiride, thioridazine and thiothixene have been published but often involve relatively few patients. A correlation between therapeutic response and plasma concentrations of thioridazine and its metabolites has not been demonstrated, and plasma level monitoring of thioridazine and its metabolites therefore appears to have no clinical value. Clinical behavioural deterioration concomitant with high plasma concentrations of chlorpromazine and haloperidol have been reported. A dosage reduction might be considered after 2 to 4 weeks’ treatment in non-responders who have plasma chlorpromazine concentrations above 100 to 150 µg/L or plasma haloperidol concentrations above 20 to 30 µg/L. Non-responders and good responders to chlorpromazine treatment, however, have plasma drug concentrations in the same range, and a therapeutic range of plasma chlorpromazine levels has not been defined. Therapeutic plasma haloperidol concentrations (i.e. ‘window’) in the range of 5 to 20 µg/L have been reported by some investigators, but others have found no such relationship. A generally valid therapeutic plasma concentration range for haloperidol has not yet been defined.Sulpiride appears to have a therapeutic ‘window’ with an upper limit around 1.5 µmol/L. The interindividual variations in plasma sulpiride concentrations appear to be relatively small compared with other antipsychotic drugs and hence may limit the clinical requirements, if any, of plasma level monitoring.Measuring steady-state plasma drug concentrations may help to obtain an optimal therapeutic response with fluphenazine, perphenazine and thiothixene. Perphenazine appears to have a therapeutic window of 1.2 to 2.4 µg/L; levels greater than 2.4 µg/L may be associated with extrapyramidal side effects. The reported therapeutic ranges for fluphenazine and thiothixene are 0.2 to 2.0 µg/L and 2 to 15 µg/L, respectively. Dosages of fluphenazine and thiothixene may be adjusted from plasma level measurements after 2 weeks of treatment.Unfortunately, there is no evidence that plasma concentration monitoring of antipsychotic drugs may significantly reduce the incidence of tardive dyskinesia.

A database of mutants and effects of site-directed mutagenesis experiments on G protein-coupled receptors

A database system and computer programs for storage and retrieval of information about guanine nucleotide‐binding protein (G protein) ‐coupled receptor mutants and associated biological effects have been developed. Mutation data on the receptors were collected from the literature and a database of mutants and effects of mutations was developed. The G protein‐coupled receptor, family A, point mutation database (GRAP) provides detailed information on ligand‐binding and signal transduction properties of more than 2130 receptor mutants. The amino acid sequences of receptors for which mutation experiments have been reported were aligned, and from this alignment mutation data may be retrieved. Alternatively, a search form allowing detailed specification of which mutants to retrieve may be used, for example, to search for specific amino acid substitutions, substitutions in specific protein domains or reported biological effects. Furthermore, ligand and bibliographic oriented queries may be performed. GRAP is available on the Internet (URL: http://www‐grap.fagmed.uit.no/GRAP/homepage.html) using the World‐Wide Web system.

Molecular modeling of serotonin, ketanserin, ritanserin and their 5-HT2C receptor interactions

Molecular modeling techniques were used to build a three-dimensional model of the rat 5-HT2C receptor, which was used to examine receptor interactions for protonated forms of serotonin, ketanserin and ritanserin. Molecular dynamics simulations which were started with the fluoro benzene moiety of ketanserin and ritanserin oriented towards the cytoplasmic side of the receptor model, produced the strongest antagonist-receptor interactions. The fluoro bezene ring(s) of the antagonists interacted strongly with aromatic residues in the receptor model, which predicts slightly different orientations and ligand-receptor interactions of ketanserin and ritanserin at a putative binding site. The model suggests that Asn333 (transmembrane helix 6) is involved in a hydrogen-bonding interaction with ketanserin, but not with ritanserin. The model also also suggests that the position corresponding to Cys362 (transmembrane helix 7) may be an important determinant for specifying 5-HT2A receptor selectivity in ketanserin binding.

Antinociceptive effects of neuropeptide Y and related peptides in mice

This study compares the antinociceptive and orexigenic activities of NPY and analogs after intracerebroventricular administration in mice. NPY had an antinociceptive action in the mouse writhing test which was not affected by prior treatment with naltrexone, yohimbine, idazoxan or reserpine. A detailed examination revealed that NPY (0.023-0.7 nmol), PYY (0.007-0.07 nmol), NPY2-36 (0.023-0.23 nmol) and the Y1 agonist [Leu31, Pro34]-NPY (0.07-0.7 nmol) all produced a dose-dependent and complete suppression of acetic acid-induced writhing. In contrast, the Y2 agonist, NPY13-36, had little or no antinociceptive effect. As shown by their ED50 values, the relative potency of the peptides was PYY > NPY2-36 > or = NPY > [Leu31, Pro34]-NPY > > NPY13-36, suggesting that a Y1 rather than a Y2 or Y3 receptor subtype was implicated in the antinociceptive action. Thereafter, all peptides were assessed for their effects on food intake. With respect to dose and peptide specificity, the hyperphagic effects of NPY and related peptides paralleled those on nociception, suggesting a common receptor mechanism. However, a purported NPY antagonist, [D-Trp32]-NPY, attenuated NPYs effect on feeding yet this same peptide elicited a dose-dependent inhibition of acetic acid-induced writhing, suggesting some molecular distinction between antinociception and stimulation of food intake.

Reversal by K-agonists of peritoneal irritation-induced ileus and visceral pain in rats

Peritoneal irritation in rats induced by i.p. administration of acetic acid produces abdominal contractions reflecting visceral pain, and gastrointestinal ileus characterized by inhibition of gastric emptying and small intestine transit. In this study, gastric emptying (GE) and intestinal transit, calculated by the geometric center (GC) method, were estimated using a test meal labeled with 51Cr-EDTA. Visceral pain was assessed by counting abdominal contractions. Acetic acid produced abdominal contractions (80.8 +/- 3.3) and inhibition of GE (-54%) and GC (-63%) during the test-period. The kappa-opioid receptor agonists, CI-977 (+/-)-U-50,488H, (+/-)-bremazocine, PD-117,302, (-)-cyclazocine, and U-69,583, reversed abdominal contractions and inhibitions of gastrointestinal transit in a dose-related manner. The mu-opioid receptor agonists and potent analgesics, morphine and fentanyl did not restore normal gastric emptying and intestinal transit. These data suggest that selective kappa-opioid receptor agonists might be used to treat abdominal pain associated with motility and transit impairment during postoperative ileus.

Involvement of prostaglandins and CGRP-dependent sensory afferents in peritoneal irritation-induced visceral pain

This study investigates the contribution of prostaglandins (PG) and calcitonin gene-related peptide (CGRP) pathways in visceral pain induced by peritoneal irritation in rats. Peritoneal irritation was produced by i.p. administration of acetic acid (AA: 0.06-1.0%, 10 ml/kg). Visceral pain was scored by counting abdominal contractions. The effect of CGRP (3-100 microg/kg, i.p.) was also evaluated. Like AA, CGRP induced abdominal pain. Neonatal pretreatment with capsaicin reduced abdominal contractions produced by AA (0.6%) and CGRP (20 microg/kg) with 64.6% and 45.6%, respectively. Abdominal contractions induced by AA and CGRP were blocked by two antinociceptive drugs, mu-and kappa-opioid agonists, morphine and (+/-)-U-50,488H, respectively. Indomethacin (3 mg/kg, s.c.) reduced the number of abdominal contractions produced by AA by 78.1%+/-6.4% but did not inhibit abdominal contractions produced by CGRP. The CGRP, receptor antagonist, hCGRP(8-37) (300 microg/kg, i.v.) inhibited AA- and CGRP-induced abdominal contractions with 57.5%+/-12.4% and 51.6%+/-11.3%, respectively. Concomitant i.p. administration of PGE1 and PGE2 (0.3 mg/kg of each) produced abdominal contractions which were inhibited 45.6%+/-9.3% by hCGRP(8-37) (300 microg/kg i.v.). Taken together, these results suggest that peritoneal irritation is likely to trigger the release of prostaglandins, which in turn produces a release of CGRP from primary sensory afferents.

A putative model of the dopamine transporter

A three-dimensional model of the human dopamine transporter was constructed by molecular modeling techniques from its amino acid sequence, based on sequence analysis of this and 9 other transporter proteins. The model has 12 membrane spanning alpha-helices arranged in two 7-helical bundles, loops between helices and amino- and carboxy termini. The molecular electrostatic potentials were mainly negative at the synaptic side and positive in the cytoplasmic domains of the transporter model, strongly positive in some of the transmembrane domains, and strongly negative in other transmembrane domains. The model suggests specific binding sites for dopamine and cocaine, a functional role for chloride ions, and accounts for known structure-activity relationships of cocaine analogs at the dopamine transporter.

Fedotozine blocks hypersensitive visceral pain in conscious rats: action at peripheral κ-opioid receptors

The effect of fedotozine on visceral hypersensitivity was evaluated in conscious rats. One hour after colonic irritation (0.6% acetic acid intracolonically), a 30 mmHg colonic distension was applied for 10 min. Irritation increased the number of abdominal contractions induced by colonic distension (23.4 +/- 4.1 versus 4.8 +/- 1.4 in saline-treated rats, P < 0.001). Facilitation of colonic pain was reversed in a dose-dependent manner by fedotozine ((+)-(-1R1)-1-phenyl-1-[(3,4,5-trimethoxy)benzyloxymethyl]-N ,N-dimethyl-n-propylamine), (+/-)-U-50,488H (trans-(+/-)-3,4-dichloro-N-methyl-N-(2-1-pyrrolidinyl]cyclohexyl)benzen eacetamide) and morphine (respective ED50 values 0.67, 0.51 and 0.23 mg/kg s.c.). The kappa-opioid receptor antagonist, nor-binaltorphimine, abolished the effects of fedotozine and (+/-)-U-50,488H but not those of morphine. Low doses of naloxone (30 microg/kg s.c.) blocked the effect of morphine but not of fedotozine or (+/-)-U-50,488H. After intracerebroventricular administration, morphine was very potent (ED50 1.7 microg/rat), (+/-)-U-50,488H poorly active (58% of antinociception at 300 microg/rat) and fedotozine inactive up to 300 microg/rat. These results show that fedotozine blocks hypersensitive visceral pain by acting on peripheral kappa-opioid receptors in animals.

Response heterogeneity of 5-HT3 receptor antagonists in a rat visceral hypersensitivity model

Subcutaneous administration of granisetron (BRL 43694, endo-1-methyl-N-(9-methyl-9-azabicyclo[3.3.1.]non-3-yl-1 H-indazole-3-carboxamide) and zacopride (4-amino-N-(1-azabicyclo[2.2.2.]oct-3-yl)-5-chloro-2-methoxybenzamide), two 5-HT3 receptor antagonists, at doses ranging from 3 to 1000 micrograms/kg, inhibited abdominal contractions induced by distension (30 mmHg, 10 min) of irritated colon (0.6% acetic acid) in conscious rats with a bell-shaped dose-response curve. The ED50 of granisetron and zacopride were 17.6 and 8.2 micrograms/kg, respectively. In contrast, both tropisetron (ICS 205-930, (3-a-tropanyl)t-indole-3-carboxylic ester) and ondansetron (GR38032F, 1,2,3,9-tetrahydro-9-methyl-3-[(2-methyl-1 H-imidazol-1-yl)methyl]-4 H-carbazol-4-one hydrocloride dihydrate) were inactive in this model. These data further support the concept of a heterogeneity in the potency of 5-HT3 receptor antagonists in modulating visceral hypersensitivity in conscious rats. This finding is in agreement with a reported efficacy of granisetron but not of ondansetron in patients with irritable bowel syndrome.