Perspectives in pharmacokinetics. Physiologically based pharmacokinetic modeling as a tool for drug development.

Since the pioneering work of Haggard and Teorell in the first half of the 20th century, and of Bischoff and Dedrick in the late 1960s, physiologically based pharmacokinetic (PBPK) modeling has gone through cycles of general acceptance, and of healthy skepticism. Recently, however, the trend in the pharmaceuticals industry has been away from PBPK models. This is understandable when one considers the time and effort necessary to develop, test, and implement a typical PBPK model, and the fact that in the present-day environment for drug development, efficacy and safety must be demonstrated and drugs brought to market more rapidly. Although there are many modeling tools available to the pharmacokineticist today, many of which are preferable to PBPK modeling in most circumstances, there are several situations in which PBPK modeling provides distinct benefits that outweigh the drawbacks of increased time and effort for implementation. In this Commentary, we draw on our experience with this modeling technique in an industry setting to provide guidelines on when PBPK modeling techniques could be applied in an industrial setting to satisfy the needs of regulatory customers. We hope these guidelines will assist researchers in deciding when to apply PBPK modeling techniques. It is our contention that PBPK modeling should be viewed as one of many modeling tools for drug development.

Cyclosporine disposition and metabolite profiles in renal transplant patients receiving a microemulsion formulation.

Several lines of evidence suggest that cyclosporine may undergo prehepatic metabolism, the possible contribution of which to the overall biotransformation of the drug is, however, unclear. A recently developed oral formulation of cyclosporine, Sandimmune Neoral, which incorporates the drug in a microemulsion preconcentrate, exhibits a faster rate of absorption and a shorter residence time in the gastrointestinal tract compared to the currently marketed formulation, Sandimmune. If prehepatic metabolism plays an important role, this could, theoretically, have an impact on the metabolite profile from the microemulsion formulation. Therefore, the pharmacokinetics of cyclosporine and its major metabolites were assessed in 13 clinically stable renal transplant patients receiving the commercial and the new formulations at steady state. Whole blood samples were collected over a dosing interval and analyzed by high-performance liquid chromatography (HPLC). Model-fitting of the concentration-time profiles of the parent compound indicated that while the systemic disposition was similar between formulations, absorption-related pharmacokinetic differences were evident. This was manifested in patients at steady state as shorter lag time and faster rate of absorption of the parent compound from the microemulsion formulation. The metabolite-to-parent area under the curve (AUC) ratios for the major metabolites AM1, AM4N, and AM9 were comparable between formulations. Specifically for metabolites AM1 and AM9, which predominated in whole blood and could, therefore, be fully characterized, the area ratios were bioequivalent when comparing the two formulations. Hence, absorption-related differences between the two oral formulations does not affect the systemic metabolite profile during steady-state administration in patients.

Physiologically based pharmacokinetic study on a cyclosporin derivative, SDZ IMM 125.

The immunosuppressant, SDZ IMM 125 (IMM), is a derivative of cyclosporin A (CyA). The disposition kinetics of IMM in plasma, blood cells, and various tissues of the rat was characterized by a physiologically based pharmacokinetic (PBPK) model; the model was then applied to predict the disposition kinetics in dog and human. Accumulation of IMM in blood cell is high (equilibrium blood cell/plasma ratio = 8), although the kinetics of drug transference between plasma and blood cell is moderately slow, taking approximately 10 min to reach equilibrium, implying a membrane-limited distribution into blood cells. A local PBPK model, assuming blood-flow limited distribution and tissue/blood partition coefficient (KP) data, failed to adequately describe the observed kinetics of distribution, which were slower than predicted. A membrane transport limitation is therefore needed to model dynamic tissue distribution data. Moreover, a slowly interacting intracellular pool was also necessary to adequately describe the kinetics of distribution in some organs. Three elimination pathways (metabolism, biliary secretion, and glomerular filtration) of IMM were assessed at steady state in vivo and characterized independently by the corresponding clearance terms. A whole-body PBPK model was developed according to these findings, which described closely the IMM concentration-time profiles in arterial blood as well as 14 organs/tissues of the rat after intravenous administration. The model was then scaled up to larger mammals by modifying physiological parameters, tissue distribution and elimination clearances; in vivo enzymatic activity was considered in the scale-up of metabolic clearance. The simulations agreed well with the experimental measurements in dog and human, despite the large interspecies difference in the metabolic clearance, which does not follow the usual allometric relationship. In addition, the nonlinear increase in maximum blood concentration and AUC with increasing dose, observed in healthy volunteers after intravenous administration, was accommodated quantitatively by incorporating the known saturation of specific binding of IMM to blood cells. Overall, the PBPK model provides a promising tool to quantitatively link preclinical and clinical data.

Exploratory analysis of population pharmacokinetic data from clinical trials with application to isradipine.

Drug level monitoring during routine clinical visits in the course of phase III trials provides a means to document pharmacokinetic variability in a patient population. Such a pharmacokinetic screen was performed for the new calcium antagonist isradipine. A total of 697 blood samples were collected at any time after the morning dose from 252 patients who had received oral doses of isradipine. Three approaches of data analysis based on exploratory (graphical and statistical) techniques were used to relate plasma level to patient demographic data and laboratory parameters. The pharamacokinetics of isradipine seemed to be influenced by the demographic variables of age (already detected in conventional studies) and weight, as well as by the blood serum levels of inorganic phosphorous, uric acid, alkaline phosphatase, and bilirubin, but only to a small, clinically irrelevant extent. The findings from the three approaches were complementary. They suggest that a pharmacokinetic screening in clinical trials is feasible at reasonable experimental cost and effort and provides useful data on interindividual and intraindividual pharmacokinetic variability in patients.

Quantification of absorption, intestinal and hepatic first pass effect in a chronic dog model.

A chronic dog model was used to measure the absorption and to elucidate the site and extent of presystemic metabolism of a selective D1-agonist (CY 208-243). The dog was instrumented with portal vein and carotid artery catheters together with an electromagnetic flow measuring device around the portal vein. After administering [14C]CY 208-243 intrajejunally, absorption rate was defined as the product of porto-arterial substrate difference and portal venous blood flow. The extent of absorption amounted to 34% for total radioactivity and 31% for unchanged drug, this indicating a gastrointestinal first-pass of 9%. In an additional study [14C]CY 208-243 was injected intravenously to the dog; the absorption (29%) and the bioavailability (5%) of CY 208-243 were calculated from the ratio of dose normalized, oral versus intravenous AUC values for total radioactivity and unchanged drug, respectively. These data confirm the absorption value found with the chronic dog model and indicate a global presystemic, i.e. intestinal and hepatic, first-pass effect of 83%. In conclusion, this chronic dog model allows an accurate assessment of drug absorption and a quantification of the gastrointestinal and hepatic first-pass effects.

On the dose dependency of cyclosporin A absorption and disposition in healthy volunteers.

The pharmacokinetics of Cyclosporin A (CyA, Sandimmune) was studied in 12 healthy male volunteers after oral dosing of 350 mg, 700 mg, and 1400 mg as a drinking solution. Blood samples were collected over 96 hr and analyzed by high pressure liquid chromatography. Concentration data were evaluated with model-independent and model-based linear pharmacokinetic concepts. Individual CyA concentration-time profiles in whole blood were well described by a two-compartment open model with zero-order absorption for all three doses. Comparison of pharmacokinetic parameters across doses indicates that both absorption and disposition are dose-dependent. Nonlinear disposition is suggested by the significant increase of the terminal half-life from 8.9 +/- 4.9 hr to 11.9 +/- 4.9 hr (mean +/- SD) after a 350 mg and a 1400 mg dose, respectively. Changes in the metabolic activity of the liver with concentration might be responsible for this phenomenon. In addition, the modeling approach indicated that bioavailability decreases with increasing dose. Moreover, the dependence of the rate of CyA absorption (zero-order rate constant) versus dose was well described by a hyperbola. The limited solubility of the drug in the gastrointestinal tract might be responsible for this behavior. The lag time (0.2-0.8 hr) was independent of dose. This value is similar to the time of gastric emptying in fasting volunteers. The duration of absorption for 11 of 12 subjects was in the range 2.5-3.5 hr over all doses and agrees well with the small intestine transit time. Some subjects showed a marked secondary peak at one or two doses, which could be adequately fitted by a model with two successive zero-order inputs. This double-peak behavior was ascribed to the influence of the food on gastric emptying. Dose dependency of disposition and absorption counterbalance each other in the usual dose range. This leads to an almost proportional increase of area under the blood CyA concentration-time profile with increasing dose.

A simple method for the quantitation of 14C-whole-body autoradiograms.

A simple method for the quantitation of 14C-whole-body autoradiograms by comparative densitometry is described. If processed under the same conditions as the samples of tissues to be investigated, a radioactive blood scale can be used as standard, with the exception of samples from bones, eyes and fat. This method is demonstrated to be simple, accurate, reproducible and precise.

Relationship between plasma concentrations and cardiac beta-adrenoceptor blockade--a study with oral and intravenous bopindolol.

Bopindolol, an esterified beta-adrenoceptor blocking drug, was administered to nine healthy male volunteers in oral (1 mg and 4 mg) and intravenous (1 mg) doses. Plasma concentrations determined using a radio-receptor assay (RRA) and high pressure liquid chromatography (h.p.l.c.) yielded almost identical results, indicating that hydrolysed bopindolol, the major metabolite, is responsible for the pharmacological activity of the drug. After intravenous administration the half-life for the formation of the hydrolysis product was about 0.3 h. The elimination of hydrolysed bopindolol from the plasma, determined with a one-compartment model occurred with a half-life of about 4 h. There were indications for a longer beta phase of elimination with a half-life of about 8 h, which, owing to the relative insensitivity of the method for concentrations present after more than 24 h, could not be determined exactly. The absolute bioavailability of the active compound is about 70%. Cardiac beta-adrenoceptor blockade was determined as the reduction in exercise-induced tachycardia. With oral doses the maximum effect was reached after 3 h (-29 beats min-1 after 1 mg, -40 beats min-1 after 4 mg). After intravenous administration most of the effect was present after 0.5 h but the maximum effect (-33 beats min-1) was only reached at 3 h. Bopindolol possesses a long duration of action: after 48 h 33% of the maximum effect of the oral dose of 4 mg was still present.(ABSTRACT TRUNCATED AT 250 WORDS)

Simultaneous modeling of bopindolol kinetics and dynamics.

Bopindolol has beta-blocking effects for 96 hr despite a 4-hr t1/2. To investigate the concentration-effect relationship after single and repeated doses. 2-mg oral doses were given once and then daily for 13 days to six healthy subjects. In plasma, no unchanged drug, only the hydrolysis product of bopindolol (referred to as bopindolol concentration) was detectable or could be measured up to 24 hr. Chemical assay by HPLC and determination of total active beta-adrenoceptor blocking material by radioreceptor assay gave identical results. The t1/2 was 4 to 5 hr. Effects, measured as reduction in exercise-induced tachycardia (REIT) and as the isoproterenol dose ratio (DR - 1), were followed for 96 hr. The concentration of bopindolol in plasma (predicted with a one-compartment body model) could be related to the measured effects by classic effect models for 20 t1/2s. Parameter estimates for kinetic end effect models did not differ after single and repeated doses. With the parameters from the single-dose experiment, the time course of the plasma concentration and the effects after the multiple-dose experiment could be adequately predicted for 24 and 96 hr. A deep compartment, an active metabolite, or irreversible destruction of the receptor (accounting for the persistence of the effect) could be excluded. The "dissociation constant" of 100 pmol/l (from DR -1/concentration) and the minimal effective plasma concentration (from REIT/log concentration) of 1 pmol/l suggest that enough receptors are occupied at chemically unmeasurable levels in plasma to induce an effect. The "dissociation constant" determined in vivo is of the same order as that from in vitro radioligand studies.

Intravenous cyclosporine kinetics in renal failure.

Kinetics of the novel immunosuppressive cyclosporine were determined in four patients with terminal renal failure. After a short intravenous infusion (2.05 to 3.5 mg/kg in 4 hr), blood and plasma concentrations were measured (HPLC and radioimmunoassay [RIA] up to 36 hr. After infusion, concentration curves of the drug were characterized by a rapid initial fall (t 1/2 alpha 0.10 +/- 0.03 hr), followed by a biphasic elimination phase with corresponding t 1/2s of 1.08 +/- 0.25 hr (t 1/2 beta) and 15.8 +/- 8.4 hr (t 1/2 gamma). The volumes of distribution, calculated from whole blood concentrations (HPLC), were 0.140 +/- 0.48 l/kg (volume of the central compartment) and 3.49 +/- 2.65 l/kg (volume of distribution at steady state), whereas systemic clearances were 0.369 +/- 0.08 l/hr/kg. Blood levels measured by RIA exceeded the HPLC values after the fourth hour by up to 100%, indicating the production of cross-reacting cyclosporine metabolites. Plasma concentrations were considerably lower than in whole blood. Elimination of unchanged cyclosporine in patients with renal failure appears to be of the same order as in those with normal kidney function. Modification of the initial dosage regimens is therefore probably not required.

Cyclosporin A: correlation between HPLC and RIA serum levels.

Serum levels of cyclosporin A were analysed in parallel with an HPLC and a RIA method for six patients who received repeated intramuscular and oral doses of this immunosuppressive drug after bone marrow transplantation. A good correlation was found between both methods with a similar time course of the serum curves. Due to cross-reacting metabolites, the values from the RIA assay were on average 30 to 100% higher than with the HPLC assay which is specific for the parent drug.

Water orientation and motion in phospholipid bilayers: a comparison between 17O- and 2H-NMR.

17O- and 2H-NMR spectra were obtained of a lamellar phase of dipalmitoyl-3sn-phosphatidylcholine (DPL) and D2 17O with water content of 3--15 moles water/mole DPL, in the temperature range 20 to 80 degrees C. In every case, the quadrupole splittings observed for 17O were 6.6 times larger than those for 2H. Therefore the two methods contain in principle the same information, but with less details from 17O. On the other hand, dynamic information is easily obtained from 17O linewidth data and complements the deuterium results.

The photochemical cycle of bacteriorhodopsin.

The reaction cycle of bacteriorhodopsin in the purple membrane isolated from Halobacterium halobium has been studied by optical absorption spectroscopy using low-temperature and flash kinetic techniques. After absorption of light, bacteriohodopsin passes through at least five distinct intermediates. The temperature and pH dependence of the absorbance changes suggests that branch points and/or reversible steps exist in this cycle. Flash spectroscopy in the presence of a pH-indicating dye shows that the transient release of a proton accompanies the photoreaction cycle. The proton release occurs from the exterior and the uptake is on the cytoplasmic side of the membrane, as required by the function of bacteriorhodopsin as a light-driven proton pump. Proton translocating steps connecting release and uptake are indicated by deuterium isotope effects on the kinetics of the cycle. The rapid decay of a light-induced linear dichroism shows that a chromophore orientation change occurs during the reaction cycle.

Photoreactions of bacteriorhodopsin.

Bacteriorhodopsin is a membrane-bound light energy transducer which generates an electrochemical proton gradient. It undergoes a cyclic photoreaction in which five intermediates have been identified. During the cycle it releases a proton from one surface of the membrane and takes up a proton on the opposite surface. The active chromophore consists of retinal bound through a Schiff base to the protein. The Schiff base is deprotonized during the photoreaction cycle and appears to be involved in the transport of protons through the membrane. The retinal may also undergo an isomerization.

Kinetics and stoichiometry of light-induced proton release and uptake from purple membrane fragments, Halobacterium halobium cell envelopes, and phospholipid vesicles containing oriented purple membrane.

We have used flash spectroscopy and pH indicator dyes to measure the kinetics and stoichiometry of light-induced proton release and uptake by purple membrane in aqueous suspension, in cell envelope vesicles and in lipid vesicles. The preferential orientation of bacteriorhodopsin in opposite directions in the envelope and lipid vesicles allows us to show that uptake of protons occurs on the cytoplasmic side of the purple membrane and release on the exterior side. In suspensions of isolated purple membrane, approximately one proton per cycling bacteriorhodopsin molecule appears transiently in the aqueous phase with a half-rise time of 0.8 ms and a half-decay time of 5.4 ms at 21degreesC. In cell envelope preparations which consist of vesicles with a preferential orientation of purple membrane, as in whole cells, and which pump protons out, the acidification of the medium has a half-rise time of less than 1.0 ms, which partially relaxes in approx. 10 ms and fully relaxes after many seconds. Phospholipid vesicles, which contain bacteriorhodopsin preferentially oriented in the opposite direction and pump protons in, show an alkalinization of the medium with a time constant of approximately 10 ms, preceded by a much smaller and faster acidification. The alkalinization relaxes over many seconds. The initial fast acidification in the lipid vesicles and the fast relaxation in the envelope vesicles are accounted for by the misoriented fractions of bacteriorhodopsin. The time constants of the main effects, acidification in the envelopes and alkalinization in the lipid vesicles correlate with the time constants for the release and uptake of protons in the isolated purple membrane, and therefore show that these must occur on the outer and inner surface respectively. The slow relaxation processes in the time range of several seconds must be attributed to the passive back diffusion of protons through the vesicle membrane.

Conformation and motion of the choline head group in bilayers of dipalmitoyl-3-sn-phosphatidylcholine.

The conformation and motion of the choline head group in lipid bilayers above and below the gel-to-liquid crystal transition point are studied by means of deuterium and phosphorus magnetic resonance. For this purpose dipalmitoyl-3-sn-phosphatidylcholine is selectively deuterated at various positions on the choline and glycerol constituents. The residual deuteron quadrupole couplings and the phosphorus chemical-shift anisotropy of the corresponding lipid-water mixtures yield quantitative information on the segmental motions. The choline methyl group is only slightly hindered in its movement, but the motional freedom becomes increasingly restricted the closer the segment is located to the glycerol backbone. The average value of the OC-CN bond rotation angle changes with temperature. Increasing the temperature rotates the choline methyl group into the vicinity of the phosphorus atom. The choline group as a whole is thus characterized by a flexible, temperature-dependent structure. Its orientation in space is not fixed, either parallel or perpendicular to the bilayer surface. Instead all segments execute angular oscillations with varying degrees of restriction around the normal on the bilayer surface. The gel-to-liquid crystal phase transition at 41 degrees is clearly reflected in the deuterium and phosphorus resonance spectra of the choline moiety, while no change is observed at 34 degrees. The calorimetric pretransition at 34 degrees seems not to be associated with a conformational change in the choline group.