Utilizing a Dual Human Transporter MDCKII-MDR1-BCRP Cell Line to Assess Efflux at the Blood Brain Barrier

To facilitate the design of drugs readily able to cross the blood brain barrier (BBB), a Madin-Darby canine kidney (MDCK) cell line was established that over expresses both P-glycoprotein (Pgp) and breast cancer resistance protein (BCRP), the main human efflux transporters of the BBB. Proteomics analyses indicate BCRP is expressed at a higher level than Pgp in this cell line. This cell line shows good activity for both transporters [BCRP substrate dantrolene efflux ratio (ER) 16.3 ± 0.9, Pgp substrate quinidine ER 27.5 ± 1.2], and use of selective transporter inhibitors enables an assessment of the relative contributions to overall ERs. The MDCKII-MDR1-BCRP ER negatively correlates with rat unbound brain/unbound plasma ratio, Kpuu Highly brain penetrant compounds with rat Kpuu ≥ 0.3 show ERs ≤ 2 in the MDCKII-MDR1-BCRP assay while compounds predominantly excluded from the brain, Kpuu ≤ 0.05, demonstrate ERs ≥ 20. A subset of compounds with MDCKII-MDR1-BCRP ER < 2 and rat Kpuu < 0.3 were shown to be substrates of rat Pgp using a rat transfected cell line, MDCKII-rMdr1a. These compounds also showed ERs > 2 in the human National Institutes of Health (NIH) MDCKI-MDR1 (high Pgp expression) cell line, which suggests that they are weak human Pgp substrates. Characterization of 37 drugs targeting the central nervous system in the MDCKII-MDR1-BCRP efflux assay show 36 have ERs < 2. In drug discovery, use of the MDCKII-MDR1-BCRP in parallel with the NIH MDCKI-MDR1 cell line is useful for identification of compounds with high brain penetration. SIGNIFICANCE STATEMENT: A single cell line that includes both the major human efflux transporters of the blood brain barrier (MDCKII-MDR1-BCRP) has been established facilitating the rapid identification of efflux substrates and enabling the design of brain penetrant molecules. Efflux ratios using this cell line demonstrate a clear relationship with brain penetration as defined by rat brain Kpuu.

A PKPD Case Study: Achieving Clinically Relevant Exposures of AZD5991 in Oncology Mouse Models

Capturing human equivalent drug exposures preclinically is a key challenge in the translational process. Motivated by the need to recapitulate the pharmacokinetic (PK) profile of the clinical stage Mcl-1 inhibitor AZD5991 in mice, we describe the methodology used to develop a refined mathematical model relating clinically relevant concentration profiles to efficacy. Administration routes were explored to achieve target exposures matching the clinical exposure of AZD5991. Intravenous infusion using vascular access button (VAB) technology was found to best reproduce clinical target exposures of AZD5991 in mice. Exposure-efficacy relationships were evaluated, demonstrating that dissimilar PK profiles result in differences in target engagement and efficacy outcomes. Thus, these data underscore the importance of accurately ascribing key PK metrics in the translational process to enable clinically meaningful predictions of efficacy.

Interspecies evaluation of a physiologically based pharmacokinetic model to predict the biodistribution dynamics of dendritic nanoparticles

The exposure of a dendritic nanoparticle and its conjugated active pharmaceutical ingredient (API) was determined in mouse, rat and dog, with the aim of investigating interspecies differences facilitating clinical translation. Plasma area under the curves (AUCs) were found to be dose proportional across species, while dose normalized concentration time course profiles in plasma, liver and spleen were superimposable in mouse, rat and dog. A physiologically based pharmacokinetic (PBPK) model, previously developed for mouse, was evaluated as a suitable framework to prospectively capture concentration dynamics in rat and dog. The PBPK model, parameterized either by considering species-specific physiology or using alternate scaling methods such as allometry, was shown to capture exposure profiles across species. A sensitivity analysis highlighted API systemic clearance as a key parameter influencing released API levels. The PBPK model was utilized to simulate human exposure profiles, which overlaid dose-normalized data from mouse, rat and dog. The consistency in measured interspecies exposures as well as the capability of the PBPK model to simulate observed dynamics support its use as a powerful translational tool.

Challenges and Opportunities for Improved Drug-Drug Interaction Predictions for Renal OCT2 and MATE1/2-K Transporters

Transporters contribute to renal elimination of drugs; therefore drug disposition can be impacted if transporters are inhibited by comedicant drugs. Regulatory agencies have provided guidelines to assess potential drug-drug interaction (DDI) risk for renal organic cation transporter 2 (OCT2) and multidrug and toxin extrusion 1 and 2-K (MATE1/2-K) transporters. Despite this, there are challenges with translating in vitro data using currently available tools to obtain a quantitative assessment of DDI risk in the clinic. Given the high number of drugs and new molecular entities showing in vitro inhibition toward OCT2 and/or MATE1/2-K and the lack of translation to clinically significant effects, it is reasonable to question whether the current in vitro assay design and modeling practice has led to unnecessary clinical evaluation. The aim of this review is to assess and discuss available in vitro and clinical data along with prediction models intended to provide clinical context of risk, including static models proposed by regulatory agencies and physiologically-based pharmacokinetic models, in order to identify best practices and areas of future opportunity. This analysis highlights that different in vitro assay designs, including substrate and cell systems used, strongly influence the derived concentration of drug producing 50% inhibition values and contribute to high variability observed across laboratories. Furthermore, the lack of sensitive index substrates coupled with specific inhibitors for individual transporters necessitates the use of complex models to evaluate clinical DDI risk.

Quantitative Evaluation of Dendritic Nanoparticles in Mice: Biodistribution Dynamics and Downstream Tumor Efficacy Outcomes

A physiologically based pharmacokinetic model was developed to describe the tissue distribution kinetics of a dendritic nanoparticle and its conjugated active pharmaceutical ingredient (API) in plasma, liver, spleen, and tumors. Tumor growth data from MV-4-11 tumor-bearing mice were incorporated to investigate the exposure/efficacy relationship. The nanoparticle demonstrated improved antitumor activity compared to the conventional API formulation, owing to the extended released API concentrations at the site of action. Model simulations further enabled the identification of critical parameters that influence API exposure in tumors and downstream efficacy outcomes upon nanoparticle administration. The model was utilized to explore a range of dosing schedules and their effect on tumor growth kinetics, demonstrating the improved antitumor activity of nanoparticles with less frequent dosing compared to the same dose of naked APIs in conventional formulations.

Current Practices, Gap Analysis, and Proposed Workflows for PBPK Modeling of Cytochrome P450 Induction: An Industry Perspective

The International Consortium for Innovation and Quality (IQ) Physiologically Based Pharmacokinetic (PBPK) Modeling Induction Working Group (IWG) conducted a survey across participating companies around general strategies for PBPK modeling of induction, including experience with its utility to address various questions, regulatory interactions, and regulatory acceptance. The results highlight areas where PBPK modeling is used with high confidence and identifies opportunities where confidence is lower and further evaluation is needed. To enhance the survey results, the PBPK-IWG also collected case studies and analyzed recent literature examples where PBPK models were applied to predict CYP3A induction-mediated drug-drug interactions. PBPK modeling of induction has evolved and progressed significantly, proving to have great potential to accelerate drug discovery and development. With the aim of enabling optimal use for new molecular entities that are either substrates and/or inducers of CYP3A, the PBPK-IWG proposes initial workflows for PBPK application, discusses future trends, and identifies gaps that need to be addressed.

Mechanistic physiology-based pharmacokinetic modeling to elucidate vincristine-induced peripheral neuropathy following treatment with novel kinase inhibitors

Purpose

Limited information is available regarding the drug–drug interaction (DDI) potential of molecular targeted agents and rituximab plus cyclophosphamide, doxorubicin (hydroxydaunorubicin), vincristine (Oncovin), and prednisone (R-CHOP) therapy. The addition of the Bruton tyrosine kinase (BTK) inhibitor ibrutinib to R-CHOP therapy results in increased toxicity versus R-CHOP alone, including higher incidence of peripheral neuropathy. Vincristine is a substrate of P-glycoprotein (P-gp, ABCB1); drugs that inhibit P-gp could potentially cause increased toxicity when co-administered with vincristine through DDI. While the combination of the BTK inhibitor acalabrutinib and R-CHOP is being explored clinically, the DDI potential between these therapies is unknown.

Methods

A human mechanistic physiology-based pharmacokinetic (PBPK) model of vincristine following intravenous dosing was developed to predict potential DDI interactions with combination therapy. In vitro absorption, distribution, metabolism, and excretion and in vivo clinical PK parameters informed PBPK model development, which was verified by comparing simulated vincristine concentrations with observed clinical data.

Results

While simulations suggested no DDI between vincristine and ibrutinib or acalabrutinib in plasma, simulated vincristine exposure in muscle tissue was increased in the presence of ibrutinib but not acalabrutinib. Extrapolation of the vincristine mechanistic PBPK model to other P-gp substrates further suggested DDI risk when ibrutinib (area under the concentration–time curve [AUC] ratio: 1.8), but not acalabrutinib (AUC ratio: 0.92), was given orally with venetoclax or digoxin.

Conclusion

Overall, these data suggest low DDI risk between acalabrutinib and P-gp substrates with negligible increase in the potential risk of vincristine-induced peripheral neuropathy when acalabrutinib is added to R-CHOP therapy.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00280-021-04302-5.

Discovery of AZD9833, a Potent and Orally Bioavailable Selective Estrogen Receptor Degrader and Antagonist

Herein we report the optimization of a series of tricyclic indazoles as selective estrogen receptor degraders (SERD) and antagonists for the treatment of ER⁺ breast cancer. Structure based design together with systematic investigation of each region of the molecular architecture led to the identification of N-[1-(3-fluoropropyl)azetidin-3-yl]-6-[(6S,8R)-8-methyl-7-(2,2,2-trifluoroethyl)-6,7,8,9-tetrahydro-3H-pyrazolo[4,3-f]isoquinolin-6-yl]pyridin-3-amine (28). This compound was demonstrated to be a highly potent SERD that showed a pharmacological profile comparable to fulvestrant in its ability to degrade ERα in both MCF-7 and CAMA-1 cell lines. A stringent control of lipophilicity ensured that 28 had favorable physicochemical and preclinical pharmacokinetic properties for oral administration. This, combined with demonstration of potent in vivo activity in mouse xenograft models, resulted in progression of this compound, also known as AZD9833, into clinical trials.

Abstract 1042: Preclinical pharmacokinetic and metabolic characterization of the next generation oral SERD AZD9833

AZD9833 is a potent, orally delivered, non-steroidal selective estrogen receptor degrader (SERD) that both antagonizes and degrades ERα. It is currently in clinical testing for the treatment of ER+ metastatic breast cancer (SERENA-1; NCT03616587). While ER is a clinically validated target, sustained inhibition of the target via oral delivery has proven an elusive goal. Despite intensive research, no oral SERD options are currently approved. Poor pharmacokinetic (PK) properties and/or short half-life (t ½) are often underlying features that limit oral SERD candidates from once or twice daily dose regimens. AZD9833 is a clinical stage next generation oral SERD that may overcome these challenges.

The pre-clinical PK properties of AZD9833 are consistent with once daily oral dosing in patients and are reported herein. AZD9833 is a low molecular weight (&lt;500 Da), low lipophilicity (logD &lt;3) base with good permeability characteristics. High volume of distribution (6-13 L/kg) and high bioavailability were observed in preclinical species. The in vitro hepatic metabolic clearance (CLint) for AZD9833 varied across species and were found to be nonlinear as a function of concentration. This nonlinear behaviour most strongly manifested in dog where in vivo bioavailability was &lt;2 or 95 %F depending on dose. To explain these observations, detailed in vitro mechanistic studies were conducted and physiologically based PK (PBPK) models were created. Cytochrome P450 phenotyping studies suggest dog clearance is mediated by CYP2D15 whereas CYP3A is the dominant P450 isoform responsible for clearance in human. In vitro CLint vs. concentration studies exhibited a low Km in dog (0.2 μM) relative to human (2 μM) and other species. To further our understanding, we constructed PBPK models which demonstrated that the dose dependent exposures in dog were predictable. The human model suggests that nonlinear behaviour will manifest well above predicted efficacious doses. Modeling also allowed us to assess the potential for drug-drug interactions with likely co-medications such as CDK 4/6 inhibitors. The PBPK predictions of good human exposure and long t ½ are being explored in the ongoing clinical trial.

Citation Format: Eric T. Gangl, Roshini Markandu, Pradeep Sharma, Andy Sykes, Petar Pop-Damkov, Pablo Morentin Gutierrez, James S. Scott, Dermot F. McGinnity, Adrian J. Fretland, Teresa Klinowska. Preclinical pharmacokinetic and metabolic characterization of the next generation oral SERD AZD9833 [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1042.

Improving the Accuracy of Predicted Human Pharmacokinetics: Lessons Learned from the AstraZeneca Drug Pipeline Over Two Decades

During drug discovery and prior to the first human dose of a novel candidate drug, the pharmacokinetic (PK) behavior of the drug in humans is predicted from preclinical data. This helps to inform the likelihood of achieving therapeutic exposures in early clinical development. Once clinical data are available, the observed human PK are compared with predictions, providing an opportunity to assess and refine prediction methods. Application of best practice in experimental data generation and predictive methodologies, and a focus on robust mechanistic understanding of the candidate drug disposition properties before nomination to clinical development, have led to maximizing the probability of successful PK predictions so that 83% of AstraZeneca drug development projects progress in the clinic with no PK issues; and 71% of key PK parameter predictions [64% of area under the curve (AUC) predictions; 78% of maximum concentration (C_max) predictions; and 70% of half-life predictions] are accurate to within twofold. Here, we discuss methods to predict human PK used by AstraZeneca, how these predictions are assessed and what can be learned from evaluating the predictions for 116 candidate drugs.

Prediction of human renal clearance from preclinical species for a diverse set of drugs that exhibit both active secretion and net reabsorption

Identifying any extrahepatic excretion phenomenon in preclinical species is crucial for an accurate prediction of the pharmacokinetics in man. This understanding is particularly key for drugs with a small volume of distribution, because they require an especially low total clearance to be suitable for a once-a-day dosing regimen in man. In this study, three animal scaling techniques were applied for the prediction of the human renal clearance of 36 diverse drugs that show active secretion or net reabsorption: 1) direct correlations between renal clearance in man and each of the two main preclinical species (rat and dog); 2) simple allometry; and 3) Mahmood's renal clearance scaling method. The results show clearly that the predictions to man for the methods are improved significantly when corrections are made for species differences in plasma protein binding. Overall, the most accurate predictions were obtained by using a direct correlation with the dog renal clearance after correcting for differences in plasma protein binding and kidney blood flow (r² = 0.84), where predictions, on average, were within 2-fold of the observed renal clearance values in human.

Integrated in vitro analysis for the in vivo prediction of cytochrome P450-mediated drug-drug interactions

Unbound IC(50) (IC(50,u)) values of 15 drugs were determined in eight recombinantly expressed human cytochromes P450 (P450s) and human hepatocytes, and the data were used to simulate clinical area under the plasma concentration-time curve changes (deltaAUC) on coadministration with prototypic CYP2D6 substrates. Significant differences in IC(50,u) values between enzyme sources were observed for quinidine (0.02 microM in recombinant CYP2D6 versus 0.5 microM in hepatocytes) and propafenone (0.02 versus 4.1 microM). The relative contribution of individual P450s toward the oxidative metabolism of clinical probes desipramine, imipramine, tolterodine, propranolol, and metoprolol was estimated via determinations of intrinsic clearance using recombinant P450s (rP450s). Simulated deltaAUC were compared with those observed in vivo via the ratios of unbound inhibitor concentration at the entrance to the liver to inhibition constants determined against rP450s ([I](in,u)/K(i)) and incorporating parallel substrate elimination pathways. For this dataset, there were 20% false negatives (observed deltaAUC >or= 2, predicted deltaAUC < 2), 77% correct predictions, and 3% false positives. Thus, the [I](in,u)/K(i) approach appears relatively successful at estimating the degree of clinical interactions and can be incorporated into drug discovery strategies. Using a Simcyp ADME (absorption, metabolism, distribution, elimination) simulator (Simcyp Ltd., Sheffield, UK), there were 3% false negatives, 94% correct simulations, and 3% false positives. False-negative predictions were rationalized as a result of mechanism-based inhibition, production of inhibitory metabolites, and/or hepatic uptake. Integrating inhibition and reaction phenotyping data from automated rP450 screens have shown applicability to predict the occurrence and degree of in vivo drug-drug interactions, and such data may identify the clinical consequences for candidate drugs as both "perpetrators" and "victims" of P450-mediated interactions.

From ATP to AZD6140: the discovery of an orally active reversible P2Y12 receptor antagonist for the prevention of thrombosis

Starting from adenosine triphosphate (ATP), the identification of a novel series of P2Y(12) receptor antagonists and exploitation of their SAR is described. Modifications of the acidic side chain and the purine core and investigation of hydrophobic substituents led to a series of neutral molecules. The leading compound, 17 (AZD6140), is currently in a large phase III clinical trial for the treatment of acute coronary syndromes and prevention of thromboembolic clinical sequelae.

Evaluation of Human Pharmacokinetics, Therapeutic Dose and Exposure Predictions Using Marketed Oral Drugs

In this article approaches to predict human pharmacokinetics (PK) are discussed and the capability of the exemplified methodologies to estimate individual PK parameters and therapeutic dose for a set of marketed oral drugs has been assessed. For a set of 63 drugs where the minimum efficacious concentration (MEC) and human PK were known, the clinical dose was shown to be well predicted or in some cases over-estimated using a simple one-compartment oral PK model. For a subset of these drugs, in vitro potency against the primary human targets was gathered, and compared to the observed MEC. When corrected for plasma protein binding, the MEC of the majority of compounds was < or=3 fold over the respective in vitro target potency value. A series of in vitro and in vivo experiments were conducted to predict the human PK parameters. Metabolic clearance was generally predicted well from human hepatocytes. Interestingly, for this compound set, allometry or glomerular filtration rate (GFR) ratio methods appeared to be applicable for renal CL even where CL(renal) > GFR. For approximately 90% of compounds studied, the predicted CL using in vitro-in vivo (IVIV) extrapolation together with a CL(renal) estimate, where appropriate, was within 2-fold of that observed clinically. Encouragingly volume of distribution at steady state (V(ss)) estimated in preclinical species (rat and dog) when corrected for plasma protein binding, predicted human V(ss) successfully on the majority of occasions--73% of compounds within 2-fold. In this laboratory, absorption estimated from oral rat PK studies was lower than the observed human absorption for most drugs, even when solubility and permeability appeared not to be limiting. Preliminary data indicate absorption in the dog may be more representative of human for compounds absorbed via the transcellular pathway. Using predicted PK and MEC values estimated from in vitro potency assays there was a good correlation between predicted and observed dose. This analysis suggests that for oral therapies, human PK parameters and clinical dose can be estimated from a consideration of data obtained from in vitro screens using human derived material and in vivo animal studies. The benefits and limitations of this holistic approach to PK and dose prediction within the drug discovery process are exemplified and discussed.

The pivotal role of hepatocytes in drug discovery

This review promotes the value of isolated hepatocytes in modern Drug Discovery programmes and outlines how increased understanding, particularly in the area of in vitro-in vivo extrapolation (IVIVE), has led to more widespread use. The importance of in vitro metabolic intrinsic clearance data for predicting in vivo clearance has been acknowledged for several years and the greater utility of hepatocytes, compared with hepatic microsomes and liver slices, for this application is discussed. The application of hepatocytes in predicting drug-drug interactions (DDIs) resulting from reversible and irreversible (time-dependent) inhibition is relatively novel but affords the potential to study both phase I and phase II processes together with any impact of drug efflux and/or uptake (cellular accumulation). Progress in this area is reviewed along with current opinions on the comparative use of primary hepatocytes and higher throughput reporter gene-based systems for studying cytochrome P450 (CYP) induction. The appreciation of the role of transporter proteins in drug disposition continues to evolve. The study of hepatic uptake using isolated hepatocytes and the interplay between drug transport and metabolism with respect to both clearance and DDIs and subsequent IVIVE is also considered.

EVALUATION OF TIME-DEPENDENT CYTOCHROME P450 INHIBITION USING CULTURED HUMAN HEPATOCYTES

Primary human hepatocytes in culture are commonly used to evaluate cytochrome P450 (P450) induction via an enzyme activity endpoint. However, other processes can confound data interpretation. To this end, the impact of time-dependent P450 inhibition in this system was evaluated. Using a substrate-cassette approach, P450 activities were determined after incubation with the prototypic inhibitors tienilic acid (CYP2C9), erythromycin, troleandomycin, and fluoxetine (CYP3A4). Kinetic analysis of enzyme inactivation in hepatocytes was used to describe the effect of these time-dependent inhibitors and derive the inhibition parameters kinact and KI) which generally were in good agreement with the values derived using recombinant P450s and human liver microsomes (HLMs). Tienilic acid selectively inhibited CYP2C9-dependent diclofenac 4'-hydroxylation activity, and erythromycin, troleandomycin, and fluoxetine inhibited CYP3A4-dependent midazolam 1'-hydroxylation in a time- and concentration-dependent manner. Fluoxetine also inhibited CYP2C19-dependent S-mephenytoin 4'-hydroxylation in a time- and concentration-dependent manner in hepatocytes, HLMs, and recombinant CYP2C19 (KI 0.4 microM and kinact 0.5 min(-1)). As expected, the effect of fluoxetine on CYP2D6 in hepatocytes was consistent with potent yet reversible inhibition. A very weak time-dependent CYP2C9 inhibitor (AZ1, a proprietary AstraZeneca compound; KI 30 microM and kinact 0.02 min(-1)) effectively abolished CYP2C9 activity over 24 h at low (micromolar) concentrations in primary cultured human hepatocytes. This work demonstrates that caution is warranted in the interpretation of enzyme induction studies with metabolically stable, weak time-dependent inhibitors, which may have dramatic inhibitory effects on P450 activity in this system. Therefore, in addition to enzyme activity, mRNA and/or protein levels should be measured to fully evaluate the P450 induction potential of a drug candidate.

PREDICTION OF CYP2C9-MEDIATED DRUG-DRUG INTERACTIONS: A COMPARISON USING DATA FROM RECOMBINANT ENZYMES AND HUMAN HEPATOCYTES

The IC50 values of 14 drugs were determined in recombinantly expressed CYP2C9 (rCYP2C9) and human hepatocytes and the data used to simulate clinical area under the plasma concentration-time curve (AUC) changes upon coadministration with prototypic CYP2C9 substrates. There was an excellent correlation between IC(50, apparent) values determined using diclofenac and naproxen as CYP2C9 substrates (r2 = 0.82, p < 0.0001), with values being generally higher in the naproxen assay. After correcting for nonspecific binding, the IC(50, unbound) values were similar between the assays, for the majority of compounds. Two compounds, amiodarone and benzbromarone, demonstrated substrate-specific differences, activating naproxen O-demethylase to approximately 250% of control activity at 1 mM and 1 microM, respectively, while inhibiting diclofenac 4'-hydroxylation with IC(50, apparent) values of 3 microM and 0.04 microM, respectively. CYP2C9 IC(50, apparent) values generated in human hepatocytes were systematically higher than those determined with rCYP2C9. After correcting for nonspecific binding, there was an excellent correlation of IC(50, unbound) values generated in the different milieu (r2 = 0.88, p < 0.0001). The ratio of inhibitor concentration at the entrance to the liver to the inhibition constant ([I]in/Ki) was used to simulate clinical deltaAUC changes and compared with that observed in vivo. Where [I]in, total/Ki, apparent) was used, there were zero false negatives (observed deltaAUC >or=2, predicted deltaAUC <2), eight correct assignations, and seven false positives (observed deltaAUC <or=2, predicted deltaAUC >2. Where [I]in, unbound/Ki, unbound was used, there was one false negative, 14 correct assignations, and zero false positives. In summary, the data presented here suggest that for CYP2C9 interactions, the use of total liver inhibitor concentrations may indeed avoid false negatives, but more realistic predictions may be achieved using unbound liver inhibitor concentrations and unbound in vitro inhibition parameters.

A UNIFIED MODEL FOR PREDICTING HUMAN HEPATIC, METABOLIC CLEARANCE FROM IN VITRO INTRINSIC CLEARANCE DATA IN HEPATOCYTES AND MICROSOMES

The aim of this study was to evaluate a unified method for predicting human in vivo intrinsic clearance (CL(int, in vivo)) and hepatic clearance (CL(h)) from in vitro data in hepatocytes and microsomes by applying the unbound fraction in blood (fu(b)) and in vitro incubations (fu(inc)). Human CL(int, in vivo) was projected using in vitro data together with biological scaling factors and compared with the unbound intrinsic clearance (CL(int, ub, in vivo)) estimated from clinical data using liver models with and without the various fu terms. For incubations conducted with fetal calf serum (n=14), the observed CL(int, in vivo) was modeled well assuming fu(inc) and fu(b) were equivalent. CL(int, ub, in vivo) was predicted best using both fu(b) and fu(inc) for other hepatocyte data (n=56; r(2)=0.78, p=3.3 x 10(-19), average fold error=5.2). A similar model for CL(int, ub, in vivo) was established for microsomal data (n=37; r(2)=0.77, p=1.2 x 10(-12), average fold error=6.1). Using the model for CL(int, ub, in vivo) (including a further empirical scaling factor), the CL(h) in humans was also calculated according to the well stirred liver model for the most extensive dataset. CL(int, in vivo) and CL(h) were both predicted well using in vitro human data from several laboratories for acidic, basic, and neutral drugs. The direct use of this model using only in vitro human data to predict the metabolic component of CL(h) is attractive, as it does not require extra information from preclinical studies in animals.

Evaluation of fresh and cryopreserved hepatocytes as in vitro drug metabolism tools for the prediction of metabolic clearance

The intrinsic clearances (CLint) of 50 neutral and basic marketed drugs were determined in fresh human hepatocytes and the data used to predict human in vivo hepatic metabolic clearance (CLmet). A statistically significant correlation between scaled CLmet and actual CLmet was observed (r2 = 0.48, p < 0.05), and for 73% of the drugs studied, scaled clearances were within 2-fold of the actual clearance. These data have shown that CLint data generated in human hepatocytes can be used to provide estimates of human hepatic CLmet for both phase I and phase II processes. In addition, the utility of commercial and in-house cryopreserved hepatocytes was assessed by comparing with data derived from fresh cells. A set of 14 drugs metabolized by the major human cytochromes P450 (P450s) (CYP1A2, 2C9, 2C19, 2D6, and 3A4) and uridine diphosphate glucuronosyltransferases (UGT1A1, 1A4, 1A9, and 2B7) have been used to characterize the activity of freshly isolated and cryopreserved human and dog hepatocytes. The cryopreserved human and dog cells retained on average 94% and 81%, respectively, of the CLint determined in fresh cells. Cryopreserved hepatocytes retain their full activity for more than 1 year in liquid N2 and are thus a flexible resource of hepatocytes for in vitro assays. In summary, this laboratory has successfully cryopreserved human and dog hepatocytes as assessed by the turnover of prototypic P450 and UGT substrates, and both fresh and cryopreserved human hepatocytes may be used for the prediction of human hepatic CLmet.

Predicting drug pharmacokinetics in humans from in vitro metabolism studies

The pharmaceutical industry is committed to market safer drugs with fewer side effects, predictable pharmacokinetic properties and quantifiable drug-drug interactions. There is an increasing need to develop robust, enhanced-throughput in vitro assays, which accurately extrapolate to humans. The major drug metabolizing human hepatic cytochrome P450s (CYPs; CYP1A2, 2C9, 2C19, 2D6 and 3A4) have been co-expressed functionally in Escherichia coli with human NADPH-cytochrome P450 reductase and validated as surrogates to their counterparts in human liver microsomes (HLM) with respect to their kinetic and inhibition properties. Using these recombinant enzymes, fully automated in vitro assays to assess CYP inhibition and determine the enzymology of drug oxidation have been developed and validated. IC(50) values determined for a series of test compounds in HLM and recombinant CYPs were similar (r(2)=0.9, P<0.001). There was a good correlation between the sum of individual CYP intrinsic clearance (Cl(int)) and HLM Cl(int) (r(2)=0.8, P<0.001) for ten prototypic substrates for which clearance was CYP-dependent. Several in vitro incubation milieu (e.g. CYPs, HLM, human hepatocytes) are routinely used and the level of non-specific binding was investigated with respect to effects on K(m) and K(i) determinations. There were clear correlations between binding and lipophilicity (logD(7.4)) for a selection of bases (r(2)=0.98, P<0.001) and acids (r(2)=0.79, P<0.001) that may allow prediction of this property. Our laboratory has shown that recombinant enzymes are suitable for "frontline" predictive human metabolism studies in early drug discovery.

Automated definition of the enzymology of drug oxidation by the major human drug metabolizing cytochrome P450s

A fully automated assay to determine the enzymology of drug oxidation by the major human hepatic cytochrome P450s (CYPs; CYP1A2, -2C9, -2C19, -2D6, and -3A4) coexpressed functionally in Escherichia coli with human NADPH-P450 reductase has been developed and validated. Ten prototypic substrates were chosen for which clearance was primarily CYP-dependent, and the activities of these five major CYPs were represented. A range of intrinsic clearance (CL(int)) values were obtained for substrates in both pooled human liver microsomes (HLM; 1-380 microl. min(-1)mg(-1)) and recombinant CYPs (0.03-7 microl. min(-1)pmol(-1)) and thus the percentage contribution of individual CYPs toward their oxidative metabolism could be estimated. All the assignments were consistent with the available literature data. Tolbutamide was metabolized by CYP2C9 (70%) and CYP2C19 (30%), diazepam by CYP2C19 (100%), ibuprofen by CYP2C9 (90%) and CYP2C19 (10%), and omeprazole by CYP2C19 (68%) and CYP3A4 (32%). Metoprolol and dextromethorphan were primarily CYP2D6 substrates and propranolol was metabolized by CYP2D6 (59%), CYP1A2 (26%), and CYP2C19 (15%). Diltiazem, testosterone, and verapamil were metabolized predominantly by CYP3A4. In addition, the metabolite profile for the CYP-dependent clearance of several markers determined by mass spectroscopy was as predicted from the literature. There was a good correlation between the sum of individual CYP CL(int) and HLM CL(int) (r(2) = 0.8, P <.001) for the substrates indicating that recombinant CYPs may be used to predict HLM CL(int) data. This report demonstrates that recombinant human CYPs may be useful as an approach for the prediction of the enzymology of human CYP metabolism early in the drug discovery process.

Rapid characterization of the major drug-metabolizing human hepatic cytochrome P-450 enzymes expressed in Escherichia coli

The major drug-metabolizing human hepatic cytochrome P-450s (CYPs; CYP1A2, 2C9, 2C19, 2D6, and 3A4) coexpressed functionally in Escherichia coli with human NADPH-P-450 reductase have been validated as surrogates to their counterparts in human liver microsomes (HLM) using automated technology. The dealkylation of ethoxyresorufin, dextromethorphan, and erythromycin were all shown to be specific reactions for CYP1A2, CYP2D6, and CYP3A4 that allowed direct comparison with kinetic data for HLM. For CYP2C9 and CYP2C19, the kinetics for the discrete oxidations of naproxen and diazepam were compared to data obtained using established, commercial CYP preparations. Turnover numbers of CYPs expressed in E. coli toward these substrates were generally equal to or even greater than those of the major commercial suppliers [CYP1A2 (ethoxyresorufin), E. coli 0.6 +/- 0.2 min(-1) versus B lymphoblasts 0.4 +/- 0.1 min(-1); CYP2C9 (naproxen), 6.7 +/- 0.9 versus 4.9 min(-1); CYP2C19 (diazepam), 3.7 +/- 0.3 versus 0.2 +/- 0.1 min(-1); CYP2D6 (dextromethorphan), 4.7 +/- 0.1 versus 4.4 +/- 0.1 min(-1); CYP3A4 (erythromycin), 3 +/- 1.2 versus 1.6 min(-1)]. The apparent K(m) values for the specific reactions were also similar (K(m) ranges for expressed CYPs and HLM were: ethoxyresorufin 0.5-1.0 microM, dextromethorphan 1.3-5.9 microM, and erythromycin 18-57 microM), indicating little if any effect of N-terminal modification on the E. coli-expressed CYPs. The data generated for all the probe substrates by HLM and recombinant CYPs also agreed well with literature values. In summary, E. coli-expressed CYPs appear faithful surrogates for the native (HLM) enzyme, and these data suggest that such recombinant enzymes may be suitable for predictive human metabolism studies.

Fully automated analysis of activities catalysed by the major human liver cytochrome P450 (CYP) enzymes: assessment of human CYP inhibition potential

1. Fully automated inhibition screens for the major human hepatic cytochrome P450s have been developed and validated. Probe assays were the fluorometric-based ethoxyresorufin O-deethylation for CYP1A2 and radiometric analysis of erythromycin N-demethylation for CYP3A4, dextromethorphan O-demethylation for CYP2D6, naproxen O-demethylation for CYP2C9 and diazepam N-demethylation for CYP2C19. For the radiometric assays > 99.7% of 14C-labelled substrate was routinely extracted from incubations by solid-phase extraction. 2. Furafylline, sulphaphenazole, omeprazole, quinidine and ketoconazole were identified as specific markers for the respective CYP1A2 (IC50 = 6 microM), CYP2C9 (0.7 microM), CYP2C19 (6 microM), CYP2D6 (0.02 microM) and CYP3A4 (0.2 microM) inhibition screens. 3. For the radiometric methods, a two-point IC50 estimate was validated by correlating the IC50 obtained with a full (seven-point) assay (r2 = 0.98, p < 0.001). The two-point IC50 estimate is useful for initial screening, while the full IC50 method provides more definitive quantitation, where required. 4. IC50 determined for a series of test compounds in human liver microsomes and cytochrome P450 cDNA-expressed enzymes were similar (r2 = 0.89, p < 0.001). In particular, the CYP1A2, CYP2D6 and CYP3A4 screens demonstrated the flexibility to accept either enzyme source. As a result of incomplete substrate selectivity, expressed enzymes were utilized for analysis of CYP2C9 and CYP2C19 inhibition. Good agreement was demonstrated between IC50 determined in these assays to IC50 published by other laboratories using a wide range of analytical techniques, which provided confidence in the universality of these inhibition screens. 5. These automated screens for initial assessment of P450 inhibition potential allow rapid determination of IC50. The radiometric assays are flexible, sensitive, robust and free from analytical interference, and they should permit the identification and eradication of inhibitory structural motifs within a series of potential drug candidates.

Multiple activated oxygen species in P450 catalysis: contributions To specificity in drug metabolism

A hypervalent iron-oxene species has been widely proposed as the "active oxygen" in cytochrome P450 (P450)-catalyzed reactions. We recently examined the effect of mutation of the highly conserved threonine residue in P450s 2B4 and 2E1 to alanine, a change that is believed to interfere with proton delivery to the active site, and have determined the change in rates of deformylation of aldehydes, epoxidation of olefins, and hydroxylation of various substrates. The results support the concept that three distinct oxidants are functional in P450 catalysis: nucleophilic peroxo-iron, nucleophilic or electrophilic hydroperoxo-iron, and electrophilic oxenoid-iron. The occurrence of multiple oxidizing species may contribute to the remarkable versatility of the P450 family of isozymes in the modification of drugs and other substrates. Furthermore, the relative concentrations of these oxidants in a particular P450 isozyme may contribute to substrate specificity and govern the type of reaction catalyzed.

Epoxidation of olefins by cytochrome P450: evidence from site-specific mutagenesis for hydroperoxo-iron as an electrophilic oxidant

P450 cytochromes (P450) catalyze many types of oxidative reactions, including the conversion of olefinic substrates to epoxides by oxygen insertion. In some instances epoxidation leads to the formation of products of physiological importance from naturally occurring substrates, such as arachidonic acid, and to the toxicity, carcinogenicity, or teratogenicity of foreign compounds, including drugs. In the present mechanistic study, the rates of oxidation of model olefins were determined with N-terminal-truncated P450s 2B4 and 2E1 and their respective mutants in which the threonine believed to facilitate proton delivery to the active site was replaced by alanine. Styrene epoxidation, cyclohexene epoxidation and hydroxylation to give 1-cyclohexene-3-ol, and cis- or trans-butene epoxidation (without isomerization) and hydroxylation to give 2-butene-1-ol were all significantly decreased by the 2B4 T302A mutation. Reduced proton delivery in this mutant is believed to interfere with the activation of dioxygen to the oxenoid species, as shown earlier by decreased hydroxylation of several substrates and enhanced aldehyde deformylation via a presumed peroxo intermediate. Of particular interest, however, the T303A mutation of P450 2E1 resulted in enhanced epoxidation of all of the model olefins along with decreased allylic hydroxylation of cyclohexene and butene. These results and a comparison of the ratios of the rates of epoxidation and hydroxylation support the concept that two different species with electrophilic properties, hydroperoxo-iron (FeO2H)3+ and oxenoid-iron (FeO)3+, can effect olefin epoxidation. The ability of cytochrome P450 to use several different active oxidants generated from molecular oxygen may help account for the broad reaction specificity and variety of products formed by this versatile catalyst.

Structural characterization of Paracoccus denitrificans cytochrome c peroxidase and assignment of the low and high potential heme sites

The amino acid sequence of the diheme cytochrome c peroxidase from Paracoccus denitrificans has been determined as the result of sequence analysis of peptides generated by chemical and enzymatic cleavages of the apoprotein. The sequence shows 60% similarity to the cytochrome c peroxidase from Pseudomonas aeruginosa, 39% similarity to an open reading frame encoding a putative triheme c-type cytochrome in Escherichia coli, and remote similarity to the MauG proteins from two methylotrophic bacteria. It is proposed, on the basis of the pattern of conserved residues in the sequences, that a change in iron coordination in the N-terminal heme domain may accompany reduction to the active mixed valence state, a change which may be accompanied by conformational adjustments in the highly conserved interface between the N- and C-terminal domains. These conformational adjustments may also lead to the appearance of a second Ca2+ binding site in the mixed valence enzyme. The exposed edge of the heme in the C-terminal domain is surrounded by several different patterns of charged residues in the Paracoccus and Pseudomonas enzymes, and this is consistent with the interaction of the former with the highly positively charged front face of the donor cytochrome c-550.

A single histidine is required for activity of cytochrome c peroxidase from Paracoccus denitrificans

The diheme cytochrome c peroxidase from Paracoccus denitrificans was modified with the histidine-specific reagent diethyl pyrocarbonate. At low excess of reagent, 1 mol of histidine was modified in the oxidized enzyme, and modification was associated with loss of the ability to form the active state. With time, the modification reversed, and the ability to form the active state was recovered. The agreement between the spectrophotometric measurement of histidine modification and radioactive incorporation using a radiolabeled reagent indicated little modification of other amino acids. However, the reversal of histidine modification observed spectrophotometrically was not matched by loss of radioactivity, and we propose a slow transfer of the ethoxyformyl group to an unidentified amino acid. The presence of CN- bound to the active peroxidatic site of the enzyme led to complete protection of the essential histidine from modification. Limited subtilisin treatment of the native enzyme followed by tryptic digest of the C-terminal fragment (residues 251-338) showed that radioactivity was located in a peptide containing a single histidine at position 275. We propose that this conserved residue, in a highly conserved region, is central to the function of the active mixed-valence state.

The affinity and specificity of Ca(2+)-binding sites of cytochrome-c peroxidase from Paracoccus denitrificans

The binding of Ca2+ to the dihaem cytochrome-c peroxidase from Paracoccus denitrificans was analysed by following perturbations in the visible and 1H-NMR spectra of both haem groups. The enzyme contains at least two types of Ca(2+)-binding site. Site I is occupied in the isolated enzyme, binds Ca2+ with a redox-state-independent Kd of 1.2 microM and accommodates neither Mg2+ nor Mn2+. Site II is unoccupied in dilute solutions of the isolated oxidised enzyme and binds Ca2+ cooperatively with a Kd of 0.52 mM. In the mixed valence form, the binding affinity increases to resemble that of site I. The cooperativity was shown by -Ca2+ binding to site II, the titration of haem methyl 1H-NMR resonances, and a half-of-sites effect observed for modification of an essential histidine with diethylpyrocarbonate. These are all consistent with site II being situated at the interface between two monomers of a dimeric enzyme. Thus the equilibrium of binding to site II is a reflection of the equilibrium for dimerisation and conditions which shift that equilibrium towards the dimer, such as increased ionic strength or high protein concentration, also increase Ca2+ affinity. Binding of Ca2+ to site II is required for formation of the active high spin state at the peroxidatic haem.