Physiologically based pharmacokinetic models of small molecules and therapeutic antibodies: a mini-review on fundamental concepts and applications

The mechanisms of absorption, distribution, metabolism and elimination of small and large molecule therapeutics differ significantly from one another and can be explored within the framework of a physiologically based pharmacokinetic (PBPK) model. This paper briefly reviews fundamental approaches to PBPK modeling, in which drug kinetics within tissues and organs are explicitly represented using physiologically meaningful parameters. The differences in PBPK models applied to small/large molecule drugs are highlighted, thus elucidating differences in absorption, distribution and elimination properties between these two classes of drugs in a systematic manner. The absorption of small and large molecules differs with respect to their common extravascular routes of delivery (oral versus subcutaneous). The role of the lymphatic system in drug distribution, and the involvement of tissues as sites of elimination (through catabolism and target mediated drug disposition) are unique features of antibody distribution and elimination that differ from small molecules, which are commonly distributed into the tissues but are eliminated primarily by liver metabolism. Fundamental differences exist in the ability to predict human pharmacokinetics based upon preclinical data due to differing mechanisms governing small and large molecule disposition. These differences have influence on the evolving utilization of PBPK modeling in the discovery and development of small and large molecule therapeutics.

A strategy for risk mitigation of antibodies with fast clearance

A majority of human therapeutic antibody candidates show pharmacokinetic properties suitable for clinical use, but an unexpectedly fast antibody clearance is sometimes observed that may limit the clinical utility. Pharmacokinetic data in cynomolgus monkeys collected for a panel of 52 antibodies showed broad distribution of target-independent clearance values (2.4-61.3 mL/day/kg), with 15 (29%) having clearance > 10 mL/day/kg. Alteration in the interaction with the recycling FcRn receptor did not account for the faster than expected clearance observed for the antibodies; off-target binding was presumed to account for the fast clearance. We developed an assay based on ELISA detection of non-specific binding to baculovirus particles that can identify antibodies having increased risk for fast clearance. This assay can be used during lead generation or optimization to identify antibodies with increased risk of having fast clearance in both humans and cynomolgus monkeys, and thus increase the likelihood of obtaining a suitable drug candidate.

Maximizing tumour exposure to anti-neuropilin-1 antibody requires saturation of non-tumour tissue antigenic sinks in mice

BACKGROUND AND PURPOSE

Neuropilin-1 (NRP1) is a VEGF receptor that is widely expressed in normal tissues and is involved in tumour angiogenesis. MNRP1685A is a rodent and primate cross-binding human monoclonal antibody against NRP1 that exhibits inhibition of tumour growth in NPR1-expressing preclinical models. However, widespread NRP1 expression in normal tissues may affect MNRP1685A tumour uptake. The objective of this study was to assess MNRP1685A biodistribution in tumour-bearing mice to understand the relationships between dose, non-tumour tissue uptake and tumour uptake.

EXPERIMENTAL APPROACH

Non-tumour-bearing mice were given unlabelled MNRP1685A at 10 mg·kg(-1) . Tumour-bearing mice were given (111) In-labelled MNRP1685A along with increasing amounts of unlabelled antibody. Blood and tissues were collected from all animals to determine drug concentration (unlabelled) or radioactivity level (radiolabelled). Some animals were imaged using single photon emission computed tomography - X-ray computed tomography.

KEY RESULTS

MNRP1685A displayed faster serum clearance than pertuzumab, indicating that target binding affected MNRP1685A clearance. I.v. administration of (111) In-labelled MNRP1685A to tumour-bearing mice yielded minimal radioactivity in the plasma and tumour, but high levels in the lungs and liver. Co-administration of unlabelled MNRP1685A with the radiolabelled antibody was able to competitively block lungs and liver radioactivity uptake in a dose-dependent manner while augmenting plasma and tumour radioactivity levels.

CONCLUSIONS AND IMPLICATIONS

These results indicate that saturation of non-tumour tissue uptake is required in order to achieve tumour uptake and acceptable exposure to antibody. Utilization of a rodent and primate cross-binding antibody allows for translation of these results to clinical settings.

Anti-neuropilin-1 (MNRP1685A): unexpected pharmacokinetic differences across species, from preclinical models to humans

PURPOSE

To compare the pharmacokinetics (PK) of MNRP1685A, a human monoclonal antibody (mAb) against neuropilin-1 (NRP1), in mice, rats, monkeys, and cancer patients from a Phase I study to model with parallel linear and nonlinear clearances.

METHODS

Binding characteristics of MNRP1685A in different species were evaluated using surface plasmon resonance technology. PK profiles of MNRP1685A after single and/or multiple doses in different species were analyzed using population analysis. PK parameters were compared across species.

RESULTS

MNRP1685A binds to NRP1 in all four species tested. Consistent with the wide expression of NRP1, MNRP1685A demonstrated pronounced non-linear PK over a wide dose range. PK profiles are best described by a two-compartment model with parallel linear and nonlinear clearances. Model-derived PK parameters suggest similar in-vivo target expression levels and binding affinity to target across all species tested. However, compared to typical human/humanized mAbs, non-specific clearance of MNRP1685A was faster in mice, rats, and humans (60.3, 19.4, and 8.5 ml/day/kg), but not in monkeys (3.22 ml/day/kg).

CONCLUSIONS

Monkey PK properly predicted the target-mediated clearance of MNRP1685A but underestimated its non-specific clearance in humans. This unique PK property warrants further investigation of underlying mechanisms.

Subcutaneous bioavailability of therapeutic antibodies as a function of FcRn binding affinity in mice

The neonatal Fc receptor (FcRn) plays an important and well-known role in immunoglobulin G (IgG) catabolism; however, its role in the disposition of IgG after subcutaneous (SC) administration, including bioavailability, is relatively unknown. To examine the potential effect of FcRn on IgG SC bioavailability, we engineered three anti-amyloid β monoclonal antibody (mAb) reverse chimeric mouse IgG2a (mIgG2a) Fc variants (I253A.H435A, N434H and N434Y) with different binding affinities to mouse FcRn (mFcRn) and compared their SC bioavailability to that of the wild-type (WT) mAb in mice. Our results indicated that the SC bioavailability of mIgG2a was affected by mFcRn-binding affinity. Variant I253A.H435A, which did not bind to mFcRn at either pH 6.0 or pH 7.4, had the lowest bioavailability (41.8%). Variant N434Y, which had the greatest increase in binding affinity at both pH 6.0 and pH 7.4, had comparable bioavailability to the WT antibody (86.1% vs. 76.3%), whereas Variant N434H, which had modestly increased binding affinity at pH 6.0 to mFcRn and affinity comparable to the WT antibody at pH 7.4, had the highest bioavailability (94.7%). A semi-mechanism-based pharmacokinetic model, which described well the observed data with the WT antibody and variant I253A.H435A, is consistent with the hypothesis that the decreased bioavailability of variant I253A.H435A was due to loss of the FcRn-mediated protection from catabolism at the absorption site. Together, these data demonstrate that FcRn plays an important role in SC bioavailability of therapeutic IgG antibodies.

Pharmacokinetics and toxicology of therapeutic proteins: Advances and challenges

Significant progress has been made in understanding pharmacokinetics (PK), pharmacodynamics (PD), as well as toxicity profiles of therapeutic proteins in animals and humans, which have been in commercial development for more than three decades. However, in the PK arena, many fundamental questions remain to be resolved. Investigative and bioanalytical tools need to be established to improve the translation of PK data from animals to humans, and from in vitro assays to in vivo readouts, which would ultimately lead to a higher success rate in drug development. In toxicology, it is known, in general, what studies are needed to safely develop therapeutic proteins, and what studies do not provide relevant information. One of the major complicating factors in nonclinical and clinical programs for therapeutic proteins is the impact of immunogenicity. In this review, we will highlight the emerging science and technology, as well as the challenges around the pharmacokinetic- and safety-related issues in drug development of mAbs and other therapeutic proteins.

Minipig as a potential translatable model for monoclonal antibody pharmacokinetics after intravenous and subcutaneous administration

Subcutaneous (SC) delivery is a common route of administration for therapeutic monoclonal antibodies (mAbs) with pharmacokinetic (PK)/pharmacodynamic (PD) properties requiring long-term or frequent drug administration. An ideal in vivo preclinical model for predicting human PK following SC administration may be one in which the skin and overall physiological characteristics are similar to that of humans. In this study, the PK properties of a series of therapeutic mAbs following intravenous (IV) and SC administration in Göttingen minipigs were compared with data obtained previously from humans. The present studies demonstrated: (1) minipig is predictive of human linear clearance; (2) the SC bioavailabilities in minipigs are weakly correlated with those in human; (3) minipig mAb SC absorption rates are generally higher than those in human and (4) the SC bioavailability appears to correlate with systemic clearance in minipigs. Given the important role of the neonatal Fc-receptor (FcRn) in the PK of mAbs, the in vitro binding affinities of these IgGs against porcine, human and cynomolgus monkey FcRn were tested. The result showed comparable FcRn binding affinities across species. Further, mAbs with higher isoelectric point tended to have faster systemic clearance and lower SC bioavailability in both minipig and human. Taken together, these data lend increased support for the use of the minipig as an alternative predictive model for human IV and SC PK of mAbs.

Effects of anti-VEGF on pharmacokinetics, biodistribution, and tumor penetration of trastuzumab in a preclinical breast cancer model

Both human epidermal growth factor receptor 2 (HER-2/neu) and VEGF overexpression correlate with aggressive phenotypes and decreased survival among breast cancer patients. Concordantly, the combination of trastuzumab (anti-HER2) with bevacizumab (anti-VEGF) has shown promising results in preclinical xenograft studies and in clinical trials. However, despite the known antiangiogenic mechanism of anti-VEGF antibodies, relatively little is known about their effects on the pharmacokinetics and tissue distribution of other antibodies. This study aimed to measure the disposition properties, with a particular emphasis on tumor uptake, of trastuzumab in the presence or absence of anti-VEGF. Radiolabeled trastuzumab was administered alone or in combination with an anti-VEGF antibody to mice bearing HER2-expressing KPL-4 breast cancer xenografts. Biodistribution, autoradiography, and single-photon emission computed tomography-X-ray computed tomography imaging all showed that anti-VEGF administration reduced accumulation of trastuzumab in tumors despite comparable blood exposures and similar distributions in most other tissues. A similar trend was also observed for an isotype-matched IgG with no affinity for HER2, showing reduced vascular permeability to macromolecules. Reduced tumor blood flow (P < 0.05) was observed following anti-VEGF treatment, with no significant differences in the other physiologic parameters measured despite immunohistochemical evidence of reduced vascular density. In conclusion, anti-VEGF preadministration decreased tumor uptake of trastuzumab, and this phenomenon was mechanistically attributed to reduced vascular permeability and blood perfusion. These findings may ultimately help inform dosing strategies to achieve improved clinical outcomes.

Impact of drug conjugation on pharmacokinetics and tissue distribution of anti-STEAP1 antibody-drug conjugates in rats

Antibody-drug conjugates (ADCs) are designed to combine the exquisite specificity of antibodies to target tumor antigens with the cytotoxic potency of chemotherapeutic drugs. In addition to the general chemical stability of the linker, a thorough understanding of the relationship between ADC composition and biological disposition is necessary to ensure that the therapeutic window is not compromised by altered pharmacokinetics (PK), tissue distribution, and/or potential organ toxicity. The six-transmembrane epithelial antigen of prostate 1 (STEAP1) is being pursued as a tumor antigen target. To assess the role of ADC composition in PK, we evaluated plasma and tissue PK profiles in rats, following a single dose, of a humanized anti-STEAP1 IgG1 antibody, a thio-anti-STEAP1 (ThioMab) variant, and two corresponding thioether-linked monomethylauristatin E (MMAE) drug conjugates modified through interchain disulfide cysteine residues (ADC) and engineered cysteines (TDC), respectively. Plasma PK of total antibody measured by enzyme-linked immunosorbent assay (ELISA) revealed ∼45% faster clearance for the ADC relative to the parent antibody, but no apparent difference in clearance between the TDC and unconjugated parent ThioMab. Total antibody clearances of the two unconjugated antibodies were similar, suggesting minimal effects on PK from cysteine mutation. An ELISA specific for MMAE-conjugated antibody indicated that the ADC cleared more rapidly than the TDC, but total antibody ELISA showed comparable clearance for the two drug conjugates. Furthermore, consistent with relative drug load, the ADC had a greater magnitude of drug deconjugation than the TDC in terms of free plasma MMAE levels. Antibody conjugation had a noticeable, albeit minor, impact on tissue distribution with a general trend toward increased hepatic uptake and reduced levels in other highly vascularized organs. Liver uptakes of ADC and TDC at 5 days postinjection were 2-fold and 1.3-fold higher, respectively, relative to the unmodified antibodies. Taken together, these results indicate that the degree of overall structural modification in anti-STEAP1-MMAE conjugates has a corresponding level of impact on both PK and tissue distribution.

Potential errors in the volume of distribution estimation of therapeutic proteins composed of differently cleared components

The volume of distribution at steady state (Vss) of therapeutic proteins is usually assessed by non-compartmental or compartmental pharmacokinetic (PK) analysis wherein errors may arise due to the elimination of therapeutic proteins from peripheral tissues that are not in rapid equilibrium with the sampling compartment (usually blood). Here we explored another potential source of error in the estimation of Vss that is linked to the heterogeneity of therapeutic proteins which may consist of components (e.g. glycosylation variants) with different elimination rates. PK simulations were performed with such hypothetical binary protein mixtures where elimination was assumed to be exclusively from the central compartment. The simulations demonstrated that binary mixtures containing a rapid-elimination component can give rise to pronounced bi-phasic concentration-time profiles. Apparent Vss observed with both non-compartmental and 2-compartmental PK analysis, increased with increasing fraction as well as with increasing elimination rate k ( 10 ) of the rapid-elimination component. Simulation results were complemented by PK analysis of an in vivo study in cynomolgus monkeys with different lots of lenercept, a tumor necrosis factor receptor-immunoglobulin G1 fusion protein, with different heterogeneities. The comparative Vss data for the three lenercept lots with different amounts of rapidly cleared components were consistent with the outcome of our simulations. Both lots with a higher fraction of rapidly cleared components had a statistically significant higher Vss as compared to the reference lot. Overall our study demonstrates that Vss of a therapeutic protein may be overestimated in proteins with differently eliminated components.

Effects of anti-VEGF on predicted antibody biodistribution: roles of vascular volume, interstitial volume, and blood flow

Background

The identification of clinically meaningful and predictive models of disposition kinetics for cancer therapeutics is an ongoing pursuit in drug development. In particular, the growing interest in preclinical evaluation of anti-angiogenic agents alone or in combination with other drugs requires a complete understanding of the associated physiological consequences.

Methodology/Principal Findings

Technescan™ PYP™, a clinically utilized radiopharmaceutical, was used to measure tissue vascular volumes in beige nude mice that were naïve or administered a single intravenous bolus dose of a murine anti-vascular endothelial growth factor (anti-VEGF) antibody (10 mg/kg) 24 h prior to assay. Anti-VEGF had no significant effect (p>0.05) on the fractional vascular volumes of any tissues studied; these findings were further supported by single photon emission computed tomographic imaging. In addition, apart from a borderline significant increase (p = 0.048) in mean hepatic blood flow, no significant anti-VEGF-induced differences were observed (p>0.05) in two additional physiological parameters, interstitial fluid volume and the organ blood flow rate, measured using indium-111-pentetate and rubidium-86 chloride, respectively. Areas under the concentration-time curves generated by a physiologically-based pharmacokinetic model changed substantially (>25%) in several tissues when model parameters describing compartmental volumes and blood flow rates were switched from literature to our experimentally derived values. However, negligible changes in predicted tissue exposure were observed when comparing simulations based on parameters measured in naïve versus anti-VEGF-administered mice.

Conclusions/Significance

These observations may foster an enhanced understanding of anti-VEGF effects in murine tissues and, in particular, may be useful in modeling antibody uptake alone or in combination with anti-VEGF.

Projecting human pharmacokinetics of therapeutic antibodies from nonclinical data: what have we learned?

The pharmacokinetics (PK) of therapeutic antibodies is determined by target and non-target mediated mechanisms. These antibody-specific factors need to be considered during prediction of human PK based upon preclinical information. Principles of allometric scaling established for small molecules using data from multiple animal species cannot be directly applied to antibodies. Here, different methods for projecting human clearance (CL) from animal PK data for 13 therapeutic monoclonal antibodies (mAbs) exhibiting linear PK over the tested dose ranges were examined: simple allometric scaling (CL versus body weight), allometric scaling with correction factors, allometric scaling based on rule of exponent and scaling from only cynomolgus monkey PK data. A better correlation was obtained between the observed human CL and the estimated human CL based on cynomolgus monkey PK data and an allometric scaling exponent of 0.85 for CL than other scaling approaches. Human concentration-time profiles were also reasonably predicted from the cynomolgus monkey data using species-invariant time method with a fixed exponent of 0.85 for CL and 1.0 for volume of distribution. In conclusion, we expanded our previous work and others and further confirmed that PK from cynomolgus monkey alone can be successfully scaled to project human PK profiles within linear range using simplify allometry and Dedrick plots with fixed exponent.

A hybrid approach to advancing quantitative prediction of tissue distribution of basic drugs in human

A general toxicity of basic drugs is related to phospholipidosis in tissues. Therefore, it is essential to predict the tissue distribution of basic drugs to facilitate an initial estimate of that toxicity. The objective of the present study was to further assess the original prediction method that consisted of using the binding to red blood cells measured in vitro for the unbound drug (RBCu) as a surrogate for tissue distribution, by correlating it to unbound tissue:plasma partition coefficients (Kpu) of several tissues, and finally to predict volume of distribution at steady-state (V(ss)) in humans under in vivo conditions. This correlation method demonstrated inaccurate predictions of V(ss) for particular basic drugs that did not follow the original correlation principle. Therefore, the novelty of this study is to provide clarity on the actual hypotheses to identify i) the impact of pharmacological mode of action on the generic correlation of RBCu-Kpu, ii) additional mechanisms of tissue distribution for the outlier drugs, iii) molecular features and properties that differentiate compounds as outliers in the original correlation analysis in order to facilitate its applicability domain alongside the properties already used so far, and finally iv) to present a novel and refined correlation method that is superior to what has been previously published for the prediction of human V(ss) of basic drugs. Applying a refined correlation method after identifying outliers would facilitate the prediction of more accurate distribution parameters as key inputs used in physiologically based pharmacokinetic (PBPK) and phospholipidosis models.

Development and evaluation of a novel method for preclinical measurement of tissue vascular volume

Identification of clinically predictive models of disposition kinetics for antibody therapeutics is an ongoing pursuit in drug development. To encourage translation of drug candidates from early research to clinical trials, clinical diagnostic agents may be used to characterize antibody disposition in physiologically relevant preclinical models. TechneScan PYP was employed to measure tissue vascular volumes (V(v)) in healthy mice. Two methods of red blood cell (RBC) labeling were compared: a direct in vivo method that is analogous to a clinical blood pool imaging protocol, and an indirect method in which radiolabeled blood was transfused from donor mice into recipient mice. The indirect method gave higher precision in RBC labeling yields, lower V(v) values in most tissues, and lower (99m)Tc uptake in kidneys and bladder by single photon emission computed tomographic (SPECT) imaging relative to the direct method. Furthermore, the relative influence of each method on the calculated area under the first 7 days of the concentration-time curve (AUC(0-7)) of an IgG in nude mice was assessed using a physiologically based pharmacokinetic model. The model was sensitive to the source of V(v) values, whether obtained from the literature or measured by either method, when used to predict experimental AUC(0-7) values for radiolabeled trastuzumab in healthy murine tissues. In summary, a novel indirect method for preclinical determination of V(v) offered higher precision in RBC labeling efficiency and lower renal uptake of (99m)Tc than the direct method. In addition, these observations emphasize the importance of obtaining accurate physiological parameter values for modeling antibody uptake.

Abstract 696: Anti-NRP1 tumor uptake is impacted by normal tissue sinks in a tumor bearing mouse model

Neuropilin-1 (NRP1) is a VEGF165 and semaphorin receptor widely expressed by vascular endothelial cells in normal tissues and on some tumor cells. Blocking NRP1 with a human IgG1 against the b domain of NRP1 has been shown to inhibit tumor growth in several preclinical models. However, due to the wide expression of NRP1 in normal tissues, the uptake of anti-NRP1 by tumor cells may be impacted by the uptake in normal tissues. The objective of this study was to determine the normal tissue uptake of anti-NRP1 and assess its impact on tumor tissue uptake in a tumor bearing mouse model. Nude mice bearing tumors with either low, HM7, or high, SKMES, NRP1 expression were dosed with [111In]-DOTA labeled anti-NRP1 with or without excess amounts of unlabeled anti-NRP1 via single IV injection. Following dosing, blood, tumor, and several normal tissues were collected at pre-selected time points for the determination of radioactivity levels using a gamma counter. In the absence of unlabeled anti-NRP1, 3% Injected Dose (ID)/ml of radioactivity was observed in the blood at 15 minutes post-dose and decreased quickly. There was minimal radioactivity seen in tumor tissue. In contrast, the majority of radioactivity was detected in the lungs and liver. With increasing amount of unlabeled anti-NRP1, the radioactivity in the lungs/liver dropped, while increasing in blood and tumor tissue. At ∼10 mg/kg unlabeled anti-NRP1, there was no additional decrease of radioactivity in these normal tissues whereas the tumor radioactivity appeared to reach its peak. Taken together, these data suggest that the lungs/liver are tissue sinks that sequester a majority of the anti-NRP1 when given at lower doses, thus hampering tumor uptake. It is important to optimize the dose and dose regimen in order to overcome this tissue sink binding and allow for a sufficient amount of antibody to be available for tumor uptake.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 696.

Development of a novel method for predicting human volume of distribution at steady-state of basic drugs and comparative assessment with existing methods

The parameters characterizing tissue distribution refer to the tissue/plasma partition coefficients (Kp), which can be used to derive volume of distribution at steady-state (V(ss)). The effort for predicting drug distribution in human has been further expanded to calculation methods using in vitro-based algorithms. The objective of the present study was to develop a novel prediction method to estimate human V(ss) for moderate-to-strong bases. The predictive performance of the novel method was compared with other well established in vitro-based methods available in the literature. Relevant information collected from previous prediction studies of V(ss) facilitated the development of the novel method. This was based on the calculation of V(ss) from data on Kp, which were estimated by correlating the unbound tissue/plasma ratio in vivo (Kpu) with the unbound red blood cells partitioning (RBCu) determined in vitro. The comparative assessment of the novel correlation method with existing prediction methods of human V(ss) was done using a literature dataset of 61 basic drugs (at least one pK(a) > or = 7). The five existing V(ss) prediction methods published in the literature are comprised of four versions of tissue composition-based models along with the model of Lombardo using the principle of Oie-Tozer. The statistical analysis of the prediction performance indicated that the novel method demonstrated a greater degree of accuracy compared to all other published methods. The maximum percentage of predicted values that fall within a twofold-error range is 77% for the basic drugs tested. Overall, the present study describes the development and the assessment of the predictive performance of the novel prediction method of human V(ss) based upon in vitro data, which appears to be superior based on the current dataset studied for basic drugs.

Whole body physiologically-based pharmacokinetic models: their use in clinical drug development

BACKGROUND

Whole-body physiologically-based pharmacokinetic (WB-PBPK) models mathematically describe an organism as a closed circulatory system consisting of compartments that represent the organs important for compound absorption, distribution, metabolism and elimination.

OBJECTIVES

To review the current state of WB-PBPK model use in the clinical phases of drug development.

METHODS

A qualitative description of the WB-PBPK model structure is included along with a review of the varying methods available for input parameterisation. Current and potential WB-PBPK model application in clinical development is discussed.

CONCLUSIONS

This modelling tool is at present used for small and large molecule drug development primarily as a means to scale pharmacokinetics from animals to humans based on physiology. The pharmaceutical industry is active in employing these models to clinical drug development although the applications in use now are narrow in comparison to the potential. Expanded integration of WB-PBPK models into the drug development process will only be achieved with staff training, managerial will, success stories and regulatory agency openness.

Physiologically based pharmacokinetic (PBPK) modeling of disposition of epiroprim in humans

The objective of this study was to use in synergy physiologically based and empirical approaches to estimate the drug-specific input parameters of PBPK models of disposition to simulate the plasma concentration-time profile of epiroprim in human. The estimated input parameters were the tissue:plasma partition coefficients (Pt:p) for distribution and the blood clearance (CL) for the in vivo conditions. Epiroprim represents a challenge for such methods, because it shows large interspecies differences in its pharmacokinetic properties. Two approaches were used to predict the human Pt:p values: the tissue composition model (TCM) and the "Arundel approach" based on the volume of distribution at steady state (Vdss) determined in vivo in the rat. CL in human was predicted by (1) conventional allometric scaling of in vivo animal clearances (CAS), (2) physiologically based direct scaling up of in vitro hepatocyte data (DSU), and (3) allometric scaling of animal intrinsic in vivo blood CL normalized by the ratios of animal:human intrinsic clearances determined in vitro with hepatocytes (NAS). The performance of prediction was assessed by comparing separately the above pharmacokinetic parameters (Vdss estimated from the Pt:p values and blood CL) with the corresponding in vivo data obtained from the plasma kinetic profiles. These input parameters were used in PBPK models, and the resulting plasma concentration-time profiles of epiroprim were compared with those observed in rat and human. Previously to the construction of the human PBPK model, a model for the rat was also developed to gain more confidence on the model structure and assumptions. Overall, using the TCM and the NAS for the parameterization of the distribution and clearance, respectively, the PBPK model gave the more accurate predictions of epiroprim's disposition in human. This study represents therefore an attractive approach, which may potentially help the clinical candidate selection.

New antibacterial agents derived from the DNA gyrase inhibitor cyclothialidine

Cyclothialidine (1, Ro 09-1437) is a potent DNA gyrase inhibitor that was isolated from Streptomyces filipinensis NR0484 and is a member of a new family of natural products. It acts by competitively inhibiting the ATPase activity exerted by the B subunit of DNA gyrase but barely exhibits any growth inhibitory activity against intact bacterial cells, presumably due to insufficient permeation of the cytoplasmic membrane. To explore the antibacterial potential of 1, we developed a flexible synthetic route allowing for the systematic modification of its structure. From a first set of analogues, structure-activity relationships (SAR) were established for different substitution patterns, and the 14-hydroxylated, bicyclic core (X) of 1 seemed to be the structural prerequisite for DNA gyrase inhibitory activity. The variation of the lactone ring size, however, revealed that activity can be found among 11- to 16-membered lactones, and even seco-analogues were shown to maintain some enzyme inhibitory properties, thereby reducing the minimal structural requirements to a rather simple, hydroxylated benzyl sulfide (XI). On the basis of these "minimal structures" a modification program afforded a number of inhibitors that showed in vitro activity against Gram-positive bacteria. The best activities were displayed by 14-membered lactones, and representatives of this subclass exhibit excellent and broad in vitro antibacterial activity against Gram-positive pathogens, including Staphylococcus aureus, Streptococcus pyogenes, and Enterococcus faecalis, and overcome resistance against clinically used drugs. By improving the pharmacokinetic properties of the most active compounds (94, 97), in particular by lowering their lipophilic properties, we were able to identify congeners of cyclothialidine (1) that showed efficacy in vivo.

Utility of physiologically based pharmacokinetic models to drug development and rational drug discovery candidate selection

The present paper proposes a modeling and simulation strategy for the prediction of pharmacokinetics (PK) of drug candidates by using currently available in silico and in vitro based prediction tools for absorption, distribution, metabolism and excretion (ADME). These methods can be used to estimate specific ADME parameters (such as rate and extent of absorption into portal vein, volume of distribution, metabolic clearance in the liver). They can also be part of a physiologically based pharmacokinetic (PBPK) model to simulate concentration-time profiles in tissues and plasma resulting from the overall PK after intravenous or oral administration. Since the ADME prediction tools are built only on commonly generated in silico and in vitro data, they can be applied already in early drug discovery, prior to any in vivo study. With the suggested methodology, the following advantages of the mechanistic PBPK modeling framework can now be utilized to explore potential clinical candidates already in drug discovery: (i) prediction of plasma (blood) and tissue PK of drug candidates prior to in vivo experiments, (ii) supporting a better mechanistic understanding of PK properties, as well as helping the development of more rationale PK-PD relationships from tissue kinetic data predicted, and hence facilitating a more rational decision during clinical candidate selection, and (iii) the extrapolation across species, routes of administration and dose levels.

Prediction of pharmacokinetics prior to in vivo studies. II. Generic physiologically based pharmacokinetic models of drug disposition

Many in vitro data on physicochemical properties and specific absorption, distribution, metabolism, and elimination (ADME) processes are already available at early stages of drug discovery. These data about new drug candidates could be integrated/connected in physiologically based pharmacokinetic (PBPK) models to estimate a priori the overall plasma and tissue kinetic behaviors under in vivo conditions. The objective of the present study was to illustrate that generic PBPK models integrating such data can be developed in drug discovery prior to any in vivo studies. This approach was illustrated with three example compounds, including two lipophilic bases (diazepam, propranolol) and one neutral more hydrophilic drug (ethoxybenzamide). Distribution and liver metabolism were the processes integrated in the generic rat PBPK models of disposition. Tissue:plasma partition coefficients (P(t:p)s) used for description of distribution were estimated from established tissue composition-based equations, which need only in vitro data on drug lipophilicity and plasma protein binding as sole input parameters. Furthermore, data on intrinsic clearance (CL(int)) determined in vitro with hepatocytes were scaled to the in vivo situation to estimate hepatic metabolic clearance. These prediction approaches were both incorporated in the PBPK models to enable automated estimation of distribution and liver metabolism for each drug studied. The generic PBPK models suggested can simulate a priori concentration-time profiles of plasma and several tissues after intravenous administrations to rat. The results indicate that most of the simulated concentration-time profiles of plasma and 10 tissues are in reasonable agreement with the corresponding experimental data determined in vivo (less than a factor of two). However, some more relevant deviations were observed for specific tissues (brain and gut for diazepam; liver and gut for ethoxybenzamide; lung for propranolol) because of important ADME processes were probably neglected in the PBPK models of these drugs. In this context, generic PBPK models were also used for mechanistic evaluations of pharmacokinetics for generating research hypotheses to understand these deviations. Overall, the present generic and integrative PBPK approach of drug disposition suggested as a tool for a priori simulations and mechanistic evaluations of pharmacokinetics has the potential to improve the selection and optimization of new drug candidates.

Prediction of pharmacokinetics prior to in vivo studies. 1. Mechanism-based prediction of volume of distribution

In drug discovery and nonclinical development the volume of distribution at steady state (V(ss)) of each novel drug candidate is commonly determined under in vivo conditions. Therefore, it is of interest to predict V(ss) without conducting in vivo studies. The traditional description of V(ss) corresponds to the sum of the products of each tissue:plasma partition coefficient (P(t:p)) and the respective tissue volume in addition to the plasma volume. Because data on volumes of tissues and plasma are available in the literature for mammals, the other input parameters needed to estimate V(ss) are the P(t:p)'s, which can potentially be predicted with established tissue composition-based equations. In vitro data on drug lipophilicity and plasma protein binding are the input parameters used in these equations. Such a mechanism-based approach would be particularly useful to provide first-cut estimates of V(ss) prior to any in vivo studies and to explore potential unexpected deviations between sets of predicted and in vivo V(ss) data, when the in vivo data become available during the drug development process. The objective of the present study was to use tissue composition-based equations to predict rat and human V(ss) prior to in vivo studies for 123 structurally unrelated compounds (acids, bases, and neutrals). The predicted data were compared with in vivo data obtained from the literature or at Roche. Overall, the average ratio of predicted-to-experimental rat and human V(ss) values was 1.06 (SD = 0.817, r = 0.78, n = 147). In fact, 80% of all predicted values were within a factor of two of the corresponding experimental values. The drugs can therefore be separated into two groups. The first group contains 98 drugs for which the predicted V(ss) were within a factor of two of those experimentally determined (average ratio of 1.01, SD = 0.39, r = 0.93, n = 118), and the second group includes 25 other drugs for which the predicted and experimental V(ss) differ by a factor larger than two (average ratio of 1.32, SD = 1.74, r = 0.42, n = 29). Thus, additional relevant distribution processes were neglected in predicting V(ss) of drugs of the second group. This was true especially in the case of some cationic-amphiphilic bases. The present study is the first attempt to develop and validate a mechanistic distribution model for predicting rat and human V(ss) of drugs prior to in vivo studies.

Prediction of adipose tissue: plasma partition coefficients for structurally unrelated drugs

Tissue:plasma (P(t:p)) partition coefficients (PCs) are important parameters describing tissue distribution of drugs. The ultimate goal in early drug discovery is to develop and validate in silico methods for predicting a priori the P(t:p) for each new drug candidate. In this context, tissue composition-based equations have recently been developed and validated for predicting a priori the non-adipose and adipose P(t:p) for neutral organic solvents and pollutants. For ionizable drugs that bind to different degrees to common plasma proteins, only their non-adipose P(t:p) values have been predicted with these equations. The only compound-dependent input parameters for these equations are the lipophilicity parameter, such as olive oil-water PC (K(vo:w)) or n-octanol-water PC (P(o:w)), and/or unbound fraction in plasma (fu(p)) determined under in vitro conditions. Tissue composition-based equations could potentially also be used to predict adipose tissue-plasma PCs (P(at:p)) for ionized drugs. The main objective of the present study was to modify these equations for predicting in vivo P(at:p) (white fat) for 14 structurally unrelated ionized drugs that bind substantially to plasma macromolecules in rats, rabbits, or humans. The second objective was to verify whether K(vo:w) or P(o:w) provides more accurate predictions of in vivo P(at:p) (i.e., to verify whether olive oil or n-octanol is the better surrogate for lipids in adipose tissue). The second objective was supported by comparing in vitro data on P(at:p) with those on olive oil-plasma PC (K(vo:p)) for five drugs. Furthermore, in vivo P(at:p) was not only predicted from K(vo:w) and P(o:w) of the non-ionized species, but also from K*(vo:w) and P*(o:w), taking into account the ionized species in addition. The P(at:p) predicted from K*(vo:w), P*(o:w), and P(o:w) differ from the in vivo P(at:p) by an average factor of 1.17 (SD = 0.44, r = 0.95), 15.0 (SD = 15.7, r = 0.59), and 40.7 (SD = 57.2, r = 0.33), respectively. The in vitro values of K(vo:p) differ from those of P(at:p) by an average factor of 0.86 (SD = 0.16, r = 0.99, n = 5). The results demonstrate that (i) the equation using only data on fu(p) as input and olive oil as lipophilicity surrogate is able to provide accurate predictions of in vivo P(at:p), and (ii) olive oil is a better surrogate of the adipose tissue lipids than n-octanol. The present study is an innovative method for predicting in vivo fat partitioning of drugs in mammals.

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.

Antimalarial activity of the bisquinoline trans-N1,N2-bis (7-chloroquinolin-4-yl)cyclohexane-1,2-diamine: comparison of two stereoisomers and detailed evaluation of the S,S enantiomer, Ro 47-7737

The S,S enantiomer of the bisquinoline trans-N1,N2-bis(7-chloroquinolin-4-yl)cyclohexane-1,2-diamine, Ro 47-7737, is significantly more potent against chloroquine-resistant Plasmodium falciparum than the R,R enantiomer and the previously described racemate. Both the enantiomers and the racemate are more potent inhibitors of heme polymerization than chloroquine, and their activities are probably mediated by inhibition of this parasite-specific process. The S,S enantiomer, Ro 47-7737, was studied in more detail and proved to be a potent antimalarial in the treatment of P. vivax ex vivo and P. berghei in vivo. Its suppression of P. berghei growth in a mouse model (50% effective dose, 2.3 mg/kg of body weight) was equal to that of chloroquine and mefloquine, and Ro 47-7737 was found to be more potent than these two drugs in the Rane test, in which the curative effect of a single dose is monitored. The dose at which 50% of animals were permanently cured (34 mg/kg) was markedly superior to those of chloroquine (285 mg/kg) and mefloquine (> 250 mg/kg). When administered orally at 50 mg/kg, Ro 47-7737 also showed a faster clearance of parasites than either chloroquine or mefloquine, and unlike the other two compounds, Ro 47-7737 showed no recrudescence. In a study to compare prophylactic efficacies of oral doses of 50 mg/kg, Ro 47-7737 provided protection for 14 days compared to 3 days for mefloquine and 1 day for chloroquine. The good curative and prophylactic properties of the compound can be explained in part by its long terminal half-life. The ability to generate parasite resistance to Ro 47-7737 was also assessed. With a rodent model, resistance could be generated over eight passages. This rate of resistance generation is comparable to that of mefloquine, which has proved to be an effective antimalarial for many years. Toxicity liabilities, however, ruled out this compound as a candidate for drug development.

[In vitro and in vivo photochemical reactivity of quinine- and quinidine-N-oxide]

By irradiation with UV-light quinine- and quinidine-1,1'-dioxide in polar solvents are rearranged to 2'-oxo derivatives and in nonpolar solvents to formylindol derivatives. Rats being exposed to UV-light after oral administration of quinine- or quinidine-1,1'-dioxide showed in blood plasma the 2'-oxo derivates too. In accordance with the results of the different fotochemical reactivity of quinoline-1-oxide-, chlordiazepoxide, methaqualone-1-oxide and some pyrido-pyrimidine-8-oxides a first drafting is developed due to the supposed toxic effects of the exited imino-N-oxides.

[In vitro and in vivo photochemical reactivity of methaqualone-1-oxide]

Methaqualone-1-oxide (1) exhibits photochemical reactivity. By irradiation of 1 with solar light the oxaziridin 3 is formed at first, which reacts in vitro (human proteins) and in vivo (rats) with macromolecules. As result of the photochemical in vitro and in vivo reactions of 1 the photoproduct 2-acetamidobenzoic acid-2'-methylanilide (6), involved in oxidation of protic compounds, was detected and after hydrolysis of proteins it appears that the short-lived 3 was adding to proteins.

[Photochemical reactivity of methaqualone-1-oxide]

The photoreactivity of methaqualone-1-oxide a main metabolite of the hypnotic methaqualone has been studied in polar and apolar solvents using UV- and daylight. Five photoproducts were isolated and identified by their analytical behaviour (TLC, UV, IR, high-resolution MS). The structure of the compounds 4, 6, 7, and 8 refer to unstable, reactive intermediates (oxaziridine, biradical) during the photolysis.