A priori prediction of tissue:plasma partition coefficients of drugs to facilitate the use of physiologically-based pharmacokinetic models in drug discovery

The tissue:plasma (P(t:p)) partition coefficients (PCs) are important drug-specific input parameters in physiologically based pharmacokinetic (PBPK) models used to estimate the disposition of drugs in biota. Until now the use of PBPK models in early stages of the drug discovery process was not possible, since the estimation of P(t:p) of new drug candidates by using conventional in vitro and/or in vivo methods is too time and cost intensive. The objectives of the study were (i) to develop and validate two mechanistic equations for predicting a priori the rabbit, rat and mouse P(t:p) of non-adipose and non-excretory tissues (bone, brain, heart, intestine, lung, muscle, skin, spleen) for 65 structurally unrelated drugs and (ii) to evaluate the adequacy of using P(t:p) of muscle as predictors for P(t:p) of other tissues. The first equation predicts P(t:p) at steady state, assuming a homogenous distribution and passive diffusion of drugs in tissues, from a ratio of solubility and macromolecular binding between tissues and plasma. The ratio of solubility was estimated from log vegetable oil:water PCs (K(vo:w)) of drugs and lipid and water levels in tissues and plasma, whereas the ratio of macromolecular binding for drugs was estimated from tissue interstitial fluid-to-plasma concentration ratios of albumin, globulins and lipoproteins. The second equation predicts P(t:p) of drugs residing predominantly in the interstitial space of tissues. Therefore, the fractional volume content of interstitial space in each tissue replaced drug solubilities in the first equation. Following the development of these equations, regression analyses between P(t:p) of muscle and those of the other tissues were examined. The average ratio of predicted-to-experimental P(t:p) values was 1.26 (SD = 1.40, r = 0.90, n = 269), and 85% of the 269 predicted values were within a factor of three of the corresponding literature values obtained under in vivo and in vitro conditions. For predicted and experimental P(t:p), linear relationships (r > 0.9 in most cases) were observed between muscle and other tissues, suggesting that P(t:p) of muscle is a good predictor for the P(t:p) of other tissues. The two previous equations could explain the mechanistic basis of these linear relationships. The practical aim of this study is a worthwhile goal for pharmacokinetic screening of new drug candidates.     

Hepatocyte Composition-Based Model as a Mechanistic Tool for Predicting the Cell Suspension: Aqueous Phase Partition Coefficient of Drugs in In Vitro Metabolic Studies

This study is an extension of a previously published microsome composition-based model by Poulin and Haddad (Poulin and Haddad. 2011. J Pharm Sci 100:4501–4517), which was converted to the hepatocyte composition-based model. The first objective was to investigate the ability of the composition-based model to predict nonspecific binding of drugs in hepatocytes suspended in the incubation medium in in vitro metabolic studies. The hepatocyte composition-based model describes the cell suspension-aqueous phase partition coefficients, which were used to estimate fraction unbound in the incubation medium (fu_inc) for each drug. The second objective was to make a comparative analysis between the proposed hepatocyte composition-based model and an empirical regression equation published in the literature by Austin et al. (Austin RP, Barton P, Mohmed S, Riley RJ. 2004. Drug Metab Dispos 33:419–425). The assessment was confined by the availability of experimentally determined in vitro fu_inc values at diverse hepatocyte concentrations for 92 drugs. The model that made use of hepatocyte composition data provides comparable or superior prediction performance compared with the regression equation that relied solely on physicochemical data; therefore, this demonstrates the ability of predicting fu_inc also based on mechanisms of drug tissue distribution. The accuracy of the predictions differed depending on the class of drugs (neutrals vs. ionized drugs) and species (rat vs. human) for each method. This study for hepatocytes corroborates a previous study for microsomes. Overall, this work represents a significant first step toward the development of a generic and mechanistic calculation method of fu_inc in incubations of hepatocytes, which should facilitate rational interindividual and interspecies extrapolations of fu_inc by considering differences in lipid composition of hepatocytes, for clearance prediction in the physiologically-based pharmacokinetics (PBPK) models. © 2013 Wiley Periodicals, Inc. and the American Pharmacists Association J Pharm Sci 102:2806–2818, 2013     

The prediction of drug metabolism, tissue distribution, and bioavailability of 50 structurally diverse compounds in rat using mechanism-based absorption, distribution, and metabolism prediction tools.

The aim of this study was to assess a physiologically based modeling approach for predicting drug metabolism, tissue distribution, and bioavailability in rat for a structurally diverse set of neutral and moderate-to-strong basic compounds (n = 50). Hepatic blood clearance (CL(h)) was projected using microsomal data and shown to be well predicted, irrespective of the type of hepatic extraction model (80% within 2-fold). Best predictions of CL(h) were obtained disregarding both plasma and microsomal protein binding, whereas strong bias was seen using either blood binding only or both plasma and microsomal protein binding. Two mechanistic tissue composition-based equations were evaluated for predicting volume of distribution (V(dss)) and tissue-to-plasma partitioning (P(tp)). A first approach, which accounted for ionic interactions with acidic phospholipids, resulted in accurate predictions of V(dss) (80% within 2-fold). In contrast, a second approach, which disregarded ionic interactions, was a poor predictor of V(dss) (60% within 2-fold). The first approach also yielded accurate predictions of P(tp) in muscle, heart, and kidney (80% within 3-fold), whereas in lung, liver, and brain, predictions ranged from 47% to 62% within 3-fold. Using the second approach, P(tp) prediction accuracy in muscle, heart, and kidney was on average 70% within 3-fold, and ranged from 24% to 54% in all other tissues. Combining all methods for predicting V(dss) and CL(h) resulted in accurate predictions of the in vivo half-life (70% within 2-fold). Oral bioavailability was well predicted using CL(h) data and Gastroplus Software (80% within 2-fold). These results illustrate that physiologically based prediction tools can provide accurate predictions of rat pharmacokinetics.     

Prediction of volume of distribution at steady state in humans: comparison of different approaches

INTRODUCTION: Reasonable prediction of volume of distribution at steady state (Vd(ss)) in humans is required for screening drug candidates, evaluating drug safety, and estimating first-in-human doses. AREAS COVERED: This review summarizes methods for the prediction of human Vd(ss) and tissue plasma partition coefficients (K(p)). The methods reviewed includes allometric scaling, physiologically based models, correlation with animal Vd(ss) and in silico models. The assumptions, equations, input data required, advantages, and limitations of each approach are discussed. Due to the variations among test datasets, some studies have reached inconsistent conclusions. Hence, a comprehensive comparison of various approaches using a large and exhaustive dataset is warranted to address the controversies in human Vd(ss) prediction. EXPERT OPINION: Compared with allometric scaling, the Oie-Tozer method and correlations between human and animal Vd(ss) are more accurate. All the three methods can be used for accurate predictions of human Vd(ss) just prior to first-in-human studies. Although in vivo animal data are not required, tissue composition-based approaches and inter-tissue correlation of K(p) provide reasonable predictive accuracies and are promising for physiologically based pharmacokinetic modeling. The in silico models are most suitable for high-throughput screening of compounds, which are at an early stage of development.     

Thermodynamic study of the transfer of acetanilide and phenacetin from water to different organic solvents

The molar (K(C)(o/w)) and rational (K(X)(o/w)) partition coefficients in the octanol/buffer, i-propyl myristate/buffer, chloroform/buffer, and cyclohexane/buffer systems were determined for acetanilide and phenacetin at 25.0, 30.0, 35.0, and 40.0 degrees C. In all cases except for cyclohexane, the K(C)(o/w) and K(X)(o/w) values were greater than unity. This demonstrates that these two drugs have predominantly lipophilic behavior. Gibbs and van't Hoff thermodynamic analyses have revealed that the transfer of these drugs from water to organic solvents is spontaneous and that it is mainly driven enthalpically for i-propyl myristate and chloroform, and entropy-driven for octanol and cyclohexane.     

Prediction of in vivo hepatic clearance and half-life of drug candidates in human using chimeric mice with humanized liver

Accurate prediction of pharmacokinetics (PK) parameters in humans from animal data is difficult for various reasons, including species differences. However, chimeric mice with humanized liver (PXB mice; urokinase-type plasminogen activator/severe combined immunodeficiency mice repopulated with approximately 80% human hepatocytes) have been developed. The expression levels and metabolic activities of cytochrome P450 (P450) and non-P450 enzymes in the livers of PXB mice are similar to those in humans. In this study, we examined the predictability for human PK parameters from data obtained in PXB mice. Elimination of selected drugs involves multiple metabolic pathways mediated not only by P450 but also by non-P450 enzymes, such as UDP-glucuronosyltransferase, sulfotransferase, and aldehyde oxidase in liver. Direct comparison between in vitro intrinsic clearance (CL(int,in vitro)) in PXB mice hepatocytes and in vivo intrinsic clearance (CL(int,in vivo)) in humans, calculated based on a well stirred model, showed a moderate correlation (r2 = 0.475, p = 0.009). However, when CL(int,in vivo) values in humans and PXB mice were compared similarly, there was a good correlation (r2 = 0.754, p = 1.174 × 10??). Elimination half-life (t(1/2)) after intravenous administration also showed a good correlation (r2 = 0.886, p = 1.506 × 10??) between humans and PXB mice. The rank order of CL and t(1/2) in human could be predicted at least, although it may not be possible to predict absolute values due to rather large prediction errors. Our results indicate that in vitro and in vivo experiments with PXB mice should be useful at least for semiquantitative prediction of the PK characteristics of candidate drugs in humans.     

In situ and in vivo efficacy of peroral absorption enhancers in rats and correlation to in vitro mechanistic studies

The present investigation attempts to increase intestinal permeability and hence absorption of biopharmaceutic classification system (BCS) Class III (cefotaxime sodium (CX)) and Class IV (cyclosporin A (CSA)) drugs by employing certain absorption enhancers. Drugs were co-perfused with sodium caprate (SC, 0.25% w/v), piperine (P, 0.004% w/v) and sodium deoxycholate (SD, 1.0% w/v) separately in rat in situ single pass intestinal perfusion model. These additives increased intestinal permeability (P(app)) and absorption rate constant (K(a)) up to two and fourfold, respectively. SC exhibited substantial absorption enhancement of both CX and CSA, while SD and P enhanced absorption of CX and CSA, respectively. Co-administration of SC significantly enhanced peroral bioavailability of CX (from 29.4 +/- 1.7 to 69.6 +/- 3.2) and CSA (from 18.4 +/- 15.6 to 49.6 +/- 25.1) in rats, while P increased bioavailability of CSA (from 18.4 +/- 15.6 to 33.1 +/- 17.7). Transmission electron microscopy of intestinal mucosa revealed that SC and SD act on lipid and protein domains of absorptive membrane. P showed no effect on intestinal P(app) and oral bioavailability of CX but has a profound effect on CSA, a known P-glycoprotein (P-gp) substrate. These results indicated that P enhances intestinal absorption of CSA by modulating P-gp mediated efflux transport. Release of lactate dehydrogenase in situ from intestinal mucosa in the presence of absorption enhancer was taken as index of its local toxicity. All the absorption enhancers showed significantly less release of LDH compared to positive control, sodium dodecyl sulfate (60% w/v). Overall, the data indicate that the features of these commonly used food ingredients or endogenous bile salts can effectively improve bioavailability of various BCS Class III and Class IV drugs.     

Effect of 1-O-ethyl-3-butylcyclohexanol on the skin permeation of drugs with different physicochemical characteristics

The effects of 1-O-ethyl-3-butylcyclohexanol (OEBC) on the in vitro skin permeation of ten model drugs with different physicochemical properties across excised rat skin were evaluated. The results showed that the addition of OEBC significantly improved the in vitro skin permeation of the model drugs compared with the control (without OEBC). To clarify the promoting mechanism of OEBC, a multiple regression analysis was employed. When the permeation study was performed without OEBC, the permeability coefficient was quantitatively predicted as a linear function of molecular weight (log MW) and their lipophilicity (partition coefficient of drugs between octanol and water (log K(o/w)) with a sufficiently high correlation coefficient (r=0.842). It was suggested that skin permeation of drugs without OEBC was explained as a function of diffusion of drugs through the skin and partitioning of drugs to the skin. Although OEBC was administered, the permeability coefficient of drugs cannot be predicted as a linear function of log MW and log K(o/w) (r=0.572).     

Air to muscle and blood/plasma to muscle distribution of volatile organic compounds and drugs: linear free energy analyses

Distribution coefficients, K(mus), from the gas phase to the muscle have been collected for volatile organic compounds (VOCs). For 114 VOCs, a linear free energy relationship (LFER) yields an equation for log K(mus) with R(2) = 0.944 and SD = 0.267; construction of a training and test set shows that the LFER can predict further values to around 0.30 log units. The combination of the log K(mus) values with values for air to blood yields distribution coefficients from blood to muscle, log P(mus), for 110 VOCs; the corresponding LFER has R(2) = 0.537 and SD = 0.207 and a predictive capability of 0.22 log units. We also collected data on the distribution of drugs from blood or plasma to muscle and showed that the two sets of data can be combined. A LFER for blood/plasma to muscle for 59 drugs has R(2) = 0.745 and SD = 0.253 and a predictive capability of 0.25 log units. Finally, we show that the in vitro data on VOCs and the in vivo data on drugs can be combined; a LFER on the total data for 163 compounds has R(2) = 0.595, SD = 0.220, and a predictive capability of about 0.25 log units.     

Organ growth functions in maturing male Sprague-Dawley rats

Growth equations can be used in physiologically based pharmacokinetic (PBPK) modeling to provide physiological parameters (e.g., body weight, tissue/organ volumes) for maturing rodents. No diligent systematic exercise was found in the literature dealing with growth equations for developing rats' tissues. A generalized Michaelis-Menten (GMM) model, originally developed to fit body weight vs. age data, was chosen to estimate different physiological compartment sizes. The GMM model has the functional form: Wt = (Wt(o).K(gamma) + Wt(max).Age(gamma))/(K(gamma) + Age(gamma)) where Wt is organ/tissue weight at a specified age, Wt(o) and Wt(max) are weight at birth and maximal growth respectively, and K and gamma are constants. Weights of freshly collected organs (liver, spleen, kidneys, heart, lungs, brain, gastrointestinal tract and adipose tissue), measured in male Sprague-Dawley rats of different ages (1-280 d) in our laboratory, were used to evaluate this model's performance. The GMM model was fitted to the organ weights, and the resulting parameters were statistically significant for all organs and tissues. Organ weights were highly correlated with their respective ages. GMM-derived organ growth and percent body weight (%BW) fractions of different tissues were plotted against animal age and compared with experimental values. The GMM-based organ growth and %BW fraction profiles were in general agreement with our empirical data as well as previous studies. The GMM model gave adequately precise weight predictions at all ages for all the tissues/organs examined.     

Structure-permeability relationship analysis of the permeation barrier properties of the stratum corneum and viable epidermis/dermis of rat skin

The purpose of this study was to evaluate structure-permeability relationships for chemicals through stratum corneum (SC) and viable epidermis/dermis (VED). In vitro skin permeation of ten compounds through excised rat skin was analyzed based on a two-layer diffusion model and the diffusion coefficients in SC (D(SC)) and VED (D(VED)) were determined. The relationships between the permeation parameters and the physicochemical parameters (octanol-water partition coefficient (log K(o/w)), and hydrogen bond donor number (HBD)) of the compounds were analyzed. D(SC) increased as lipophilicity increased, whereas D(VED) decreased for log K(o/w) > 2. Increases in log K(o/w) caused a decrease in the permeability coefficient from SC through VED (P(VED/SC)) for log K(o/w) > 1. The simulation study suggests that the in vitro skin permeation of a highly lipophilic compound is strongly controlled by skin thickness due to low diffusivity in VED. The present study suggests that VED act as a considerable permeation barrier for highly lipophilic compounds due to low diffusivity.     

Use of hepatocytes to assess the contribution of hepatic uptake to clearance in vivo.

The wealth of information that has emerged in recent years detailing the substrate specificity of hepatic transporters necessitates an investigation into their potential role in drug elimination. Therefore, an assay in which the loss of parent compound from the incubation medium into hepatocytes ("media loss" assay) was developed to assess the impact of hepatic uptake on unbound drug intrinsic clearance in vivo (CL(int ub in vivo)). Studies using conventional hepatocyte incubations for a subset of 36 AstraZeneca new chemical entities (NCEs) resulted in a poor projection of CL(int ub in vivo) (r2 = 0.25, p = 0.002, average fold error = 57). This significant underestimation of CL(int ub in vivo) suggested that metabolism was not the dominant clearance mechanism for the majority of compounds examined. However, CL(int ub in vivo) was described well for this dataset using an initial compound "disappearance" CL(int) obtained from media loss assays (r2 = 0.72, p = 6.3 x 10(-11), average fold error = 3). Subsequent studies, using this method for the same 36 NCEs, suggested that the active uptake into human hepatocytes was generally slower (3-fold on average) than that observed with rat hepatocytes. The accurate prediction of human CL(int ub in vivo) (within 4-fold) for the marketed drug transporter substrates montelukast, bosentan, atorvastatin, and pravastatin confirmed further the utility of this assay. This work has described a simple method, amenable for use within a drug discovery setting, for predicting the in vivo clearance of drugs with significant hepatic uptake.     

Advancing prediction of tissue distribution and volume of distribution of highly lipophilic compounds from a simplified tissue-composition-based model as a mechanistic animal alternative method

It has been reported that values of tissue-plasma ratios (K(p)) and resulting volume of distribution at steady state (V(ss)) are substantially overpredicted for several highly lipophilic drugs. This effect was observed particularly with the published version of the tissue-composition-based model, which used experimentally determined unbound fraction in plasma (fu(p)) as input for drugs. The reasons for the unreasonably high V(ss) predictions were investigated in this study for 14 highly lipophilic compounds with a log n-octanol-water partition coefficient (log P(ow)) of at least 5.8. Here, we argue that the experimentally determined fu(p) is inaccurate for these compounds, which affected the prediction of K(p) and V(ss). Alternatively, the tissue-plasma ratio of neutral lipids (nl) equivalent was used as the main factor governing K(p), and hence V(ss), in addition to log P(ow). The average fold error of deviation between the predicted and observed human V(ss) is 124 for the published model, whereas it significantly decreased to 1.5 for the proposed model. The sensitivity analysis confirmed the importance of nl content and drug lipophilicity. Overall, this study proposes a generic and simplified tissue-composition-based model for highly lipophilic drugs and chemicals, which is a step forward toward improving prediction of V(ss) into physiologically based pharmacokinetic (PBPK) models.