Prediction of Drug-Drug Interactions Arising from CYP3A induction Using a Physiologically Based Dynamic Model.

Using physiologically based pharmacokinetic modeling, we predicted the magnitude of drug-drug interactions (DDIs) for studies with rifampicin and seven CYP3A4 probe substrates administered i.v. (10 studies) or orally (19 studies). The results showed a tendency to underpredict the DDI magnitude when the victim drug was administered orally. Possible sources of inaccuracy were investigated systematically to determine the most appropriate model refinement. When the maximal fold induction (Indmax) for rifampicin was increased (from 8 to 16) in both the liver and the gut, or when the Indmax was increased in the gut but not in liver, there was a decrease in bias and increased precision compared with the base model (Indmax = 8) [geometric mean fold error (GMFE) 2.12 vs. 1.48 and 1.77, respectively]. Induction parameters (mRNA and activity), determined for rifampicin, carbamazepine, phenytoin, and phenobarbital in hepatocytes from four donors, were then used to evaluate use of the refined rifampicin model for calibration. Calibration of mRNA and activity data for other inducers using the refined rifampicin model led to more accurate DDI predictions compared with the initial model (activity GMFE 1.49 vs. 1.68; mRNA GMFE 1.35 vs. 1.46), suggesting that robust in vivo reference values can be used to overcome interdonor and laboratory-to-laboratory variability. Use of uncalibrated data also performed well (GMFE 1.39 and 1.44 for activity and mRNA). As a result of experimental variability (i.e., in donors and protocols), it is prudent to fully characterize in vitro induction with prototypical inducers to give an understanding of how that particular system extrapolates to the in vivo situation when using an uncalibrated approach.

Evaluation of Time Dependent Inhibition Assays for Marketed Oncology Drugs: Comparison of Human Hepatocytes and Liver Microsomes in the Presence and Absence of Human Plasma.

PURPOSE: To evaluate an alternative in vitro system which can provide more quantitatively accurate drug drug interaction (DDI) prediction for 10 protein kinase inhibitors for which DDI risk was over-predicted by inhibition data generated in human liver microsomes (HLM). METHODS: Three cryopreserved human hepatocyte (hHEP) systems: 1) plated hHEPs; 2) hHEPs suspended in Dulbecco's Modified Eagle Medium (DMEM) and 3) hHEPs suspended in human plasma (plasma hHEPs) were developed to detect CYP3A time dependent inhibition, and the static mechanistic model was used to predict clinical outcomes. RESULTS: A general trend was observed in the CYP3A inactivation potency (k inact /K I, app ) as HLM > plated > DMEM ≥ plasma hHEPs. Using the static mechanistic model, DDIs predicted using parameters estimated from plated, DMEM and plasma hHEPs had 84, 74 and 95% accuracy (out of 19 clinical interaction studies) within 2-fold of the reported interaction, respectively. They demonstrated significant improvement compared to the DDIs predicted using parameters estimated from HLMs where 58% accuracy was obtained. CONCLUSIONS: Based on 19 DDIs, plasma hHEPs demonstrate a more reliable clinical DDI prediction for 10 protein kinase inhibitors and prototypical CYP3A time dependent inhibitors.

α-Aryl pyrrolidine sulfonamides as TRPA1 antagonists.

A series of α-aryl pyrrolidine sulfonamide TRPA1 antagonists were advanced from an HTS hit to compounds that were stable in liver microsomes with retention of TRPA1 potency. Metabolite identification studies and physicochemical properties were utilized as a strategy for compound design. These compounds serve as starting points for further compound optimization.

Synthesis and evaluation of a series of 4-azaindole-containing p21-activated kinase-1 inhibitors.

A series of 4-azaindole-containing p21-activated kinase-1 (PAK1) inhibitors was prepared with the goal of improving physicochemical properties relative to an indole starting point. Indole 1 represented an attractive, non-basic scaffold with good PAK1 affinity and cellular potency but was compromised by high lipophilicity (clogD=4.4). Azaindole 5 was designed as an indole surrogate with the goal of lowering logD and resulted in equipotent PAK1 inhibition with a 2-fold improvement in cellular potency over 1. Structure-activity relationship studies around 5 identified additional 4-azaindole analogs with superior PAK1 biochemical activity (Ki <10nM) and up to 24-fold selectivity for group I over group II PAKs. Compounds from this series showed enhanced permeability, improved aqueous solubility, and lower plasma protein binding over indole 1. The improvement in physicochemical properties translated to a 20-fold decrease in unbound clearance in mouse PK studies for azaindole 5 relative to indole 1.

Identification of nicotinamide phosphoribosyltransferase (NAMPT) inhibitors with no evidence of CYP3A4 time-dependent inhibition and improved aqueous solubility.

Herein we report the optimization efforts to ameliorate the potent CYP3A4 time-dependent inhibition (TDI) and low aqueous solubility exhibited by a previously identified lead compound from our NAMPT inhibitor program (1, GNE-617). Metabolite identification studies pinpointed the imidazopyridine moiety present in 1 as the likely source of the TDI signal, and replacement with other bicyclic systems was found to reduce or eliminate the TDI finding. A strategy of reducing the number of aromatic rings and/or lowering cLogD7.4 was then employed to significantly improve aqueous solubility. These efforts culminated in the discovery of 42, a compound with no evidence of TDI, improved aqueous solubility, and robust efficacy in tumor xenograft studies.

Mechanistic studies of the cationic binding pocket of CYP2C9 in vitro and in silico: metabolism of nonionizable analogs of tienilic acid.

Tienilic acid (TA) is selectively oxidized at the C-5 position of the thiophene ring by the human liver enzyme cytochrome P450 2C9 (CYP2C9). This oxidation is mediated by the proximal positioning of the thiophene over the heme iron, which is proposed to be coordinated by an interaction of the TA carboxylic acid to a cationic binding pocket in the enzyme active site. In this study, we investigated how chemical modification of TA influences the bioactivation by CYP2C9. For this investigation, nine analogs of TA were chosen with substitutions on either side of the molecule. We tested three parameters, including CYP2C9 inhibition, metabolic profiling, and in silico docking. Of the 10 compounds tested, only two (TA and a noncarboxyl analog) resulted in competitive and time-dependent inhibition of CYP2C9. Metabolic profiling revealed a trend in which substitution of the carboxylate with nonionizable functional groups resulted in metabolic switching from oxidation of the aromatic ring to dealkylation reactions at the opposite side of the structure. The in silico modeling predicted an opposite binding orientation to that of TA for many analogs, including the 3-thenoyl regio-isomer analog, which contradicts previous models. Together these data show that disrupting interactions with the cationic binding pocket of CYP2C9 will impact the sites of metabolism and inhibition of the enzyme.

Drug-drug interaction potential of marketed oncology drugs: in vitro assessment of time-dependent cytochrome P450 inhibition, reactive metabolite formation and drug-drug interaction prediction.

PURPOSE: To evaluate 26 marketed oncology drugs for time-dependent inhibition (TDI) of cytochrome P450 (CYP) enzymes. Evaluate TDI-positive drugs for potential to generate reactive intermediates. Assess clinical drug-drug interaction (DDI) risk using static mechanistic models. METHODS: Human liver microsomes and CYP-specific probes were used to assess TDI in a dilution shift assay followed by generation of K(I) and k(inact). Reactive metabolite trapping studies were performed with stable label probes. Static mechanistic model was used to predict DDI risk using a 1.25-fold AUC increase as a cut-off for positive DDI. RESULTS: Negative TDI across CYPs was observed for 13/26 drugs; the rest were time-dependent inhibitors of, predominantly, CYP3A. The k(inact)/K(I) ratios for 11 kinase inhibitors ranged from 0.7 to 42.2 ml/min/µmol. Stable label trapping agent-drug conjugates were observed for ten kinase inhibitors. DDI predictions gave no false negatives, one true negative, four false positives and three true positives. The magnitude of DDI was overestimated irrespective of the inhibitor concentration selected. CONCLUSIONS: 13/26 oncology drugs investigated showed TDI potential towards CYP3A, formation of reactive metabolites was also observed. An industry standard static mechanistic model gave no false negative predictions but did not capture the modest clinical DDI potential of kinase inhibitors.

A comparison of the roles of peroxisome proliferator-activated receptor and retinoic acid receptor on CYP26 regulation.

The cytochrome P450 26 family is believed to be responsible for all-trans-retinoic acid (atRA) metabolism and elimination in the human fetus and adults. CYP26A1 and CYP26B1 mRNA is expressed in a tissue-specific manner, and mice in which the CPY26 isoform has been knocked out show distinct malformations and lethality. The aim of this study was to determine differences in CYP26A1 and CYP26B1 regulation and expression. Analysis of CYP26A1 and CYP26B1 expression in a panel of 57 human livers showed CYP26A1 to be the major CYP26 isoform present in the liver, and its expression to be subject to large interindividual variability between donors. CYP26A1 and retinoic acid receptor (RAR) beta were found to be greatly inducible by atRA in HepG2 cells, whereas CYP26B1, RARalpha, and RARgamma were induced to a much lesser extent. Based on treatments with RAR isoform-selective ligands, RARalpha is the major isoform responsible for CYP26A1 and RARbeta induction in HepG2 cells. Classic cytochrome P450 inducers did not affect CYP26 transcription, whereas the peroxisome proliferator-activated receptor (PPAR) gamma agonists pioglitazone and rosiglitazone up-regulated CYP26B1 transcription by as much as 209- +/- 80-fold and CYP26A1 by 10-fold. RARbeta was also up-regulated by pioglitazone and rosiglitazone. CYP26B1 induction by PPARgamma agonists was abolished by the irreversible PPARgamma antagonist 2-chloro-5-nitrobenzanilide (GW9662), whereas RARbeta and CYP26A1 induction was unaffected by GW9662. Overall, the results of this study suggest that CYP26B1 and CYP26A1 are regulated by different nuclear receptors, resulting in tissue-specific expression patterns. The fact that drugs can alter the expression of CYP26 enzymes may have toxicological and therapeutic importance.

Utility of Pooled Cryopreserved Human Enterocytes as an In vitro Model for Assessing Intestinal Clearance and Drug-Drug Interactions.

BACKGROUND: A recent advancement in isolation and cryopreservation has resulted in commercially available primary human enterocytes that express various drug metabolizing enzymes (DMEs) and transporters. The main objective of this study was to further evaluate the utility of pooled cryopreserved enterocytes, specifically MetMax™ cryopreserved human enterocytes (In vitro ADMET Laboratories), as an in vitro model for assessing intestinal clearance in comparison to hepatocytes. METHODS: It was found that, for CYP3A4/5 substrates such as midazolam, amprenavir and loperamide, in vitro metabolic clearance is generally lower in enterocytes compared to that of hepatocytes, which is consistent with the relative abundance of the enzyme between the intestine and liver. In contrast, raloxifene, a surrogate UGT activity substrate, showed 3-fold greater turnover in enterocytes than hepatocytes, which is likely attributed to the differential expression of individual UGTs in human liver and intestine. For procaine, a known CES2 substrate, the measured apparent clearance was higher in hepatocytes, but formation of 4-aminobenzoic acid, a CE2-specific metabolite, was more pronounced in enterocytes, suggesting that CE2 is more active in enterocytes. Salbutamol, a SULT1A3 substrate, showed little turnover in both enterocytes and hepatocytes, and more abundant sulfate conjugate was detected in enterocytes, indicating higher SULT activity in enterocytes than hepatocytes. As expected, ketoconazole inhibited CYP3A4/5-mediated metabolite formation in enterocytes for midazolam, amprenavir and loperamide, suggesting that cryopreserved enterocytes may be useful in determining intestinal CYP3A inhibition parameters. Interestingly, elacridar, a P-gp inhibitor, suppressed metabolite formation in enterocytes for loperamide, a substrate of CYP3A4 and P-gp, suggesting that enterocytes in suspension do not have active P-gp efflux functions, and the suppression of metabolism in enterocytes is probably caused by inhibition of CYP3A4/5 by elacridar. RESULTS: Our results suggest that pooled cryopreserved human enterocytes, specifically the MetMax™ cryopreserved human enterocytes, represent a valuable in vitro model for assessing first-pass clearance and potential drug interactions in human intestine.

Discovery of a Potent (4 R,5 S)-4-Fluoro-5-methylproline Sulfonamide Transient Receptor Potential Ankyrin 1 Antagonist and Its Methylene Phosphate Prodrug Guided by Molecular Modeling.

Transient receptor potential ankyrin 1 (TRPA1) is a non-selective cation channel expressed in sensory neurons where it functions as an irritant sensor for a plethora of electrophilic compounds and is implicated in pain, itch, and respiratory disease. To study its function in various disease contexts, we sought to identify novel, potent, and selective small-molecule TRPA1 antagonists. Herein we describe the evolution of an N-isopropylglycine sulfonamide lead (1) to a novel and potent (4 R,5 S)-4-fluoro-5-methylproline sulfonamide series of inhibitors. Molecular modeling was utilized to derive low-energy three-dimensional conformations to guide ligand design. This effort led to compound 20, which possessed a balanced combination of potency and metabolic stability but poor solubility that ultimately limited in vivo exposure. To improve solubility and in vivo exposure, we developed methylene phosphate prodrug 22, which demonstrated superior oral exposure and robust in vivo target engagement in a rat model of AITC-induced pain.

Discovery of GDC-0853: A Potent, Selective, and Noncovalent Bruton's Tyrosine Kinase Inhibitor in Early Clinical Development.

Bruton's tyrosine kinase (Btk) is a nonreceptor cytoplasmic tyrosine kinase involved in B-cell and myeloid cell activation, downstream of B-cell and Fcγ receptors, respectively. Preclinical studies have indicated that inhibition of Btk activity might offer a potential therapy in autoimmune diseases such as rheumatoid arthritis and systemic lupus erythematosus. Here we disclose the discovery and preclinical characterization of a potent, selective, and noncovalent Btk inhibitor currently in clinical development. GDC-0853 (29) suppresses B cell- and myeloid cell-mediated components of disease and demonstrates dose-dependent activity in an in vivo rat model of inflammatory arthritis. It demonstrates highly favorable safety, pharmacokinetic (PK), and pharmacodynamic (PD) profiles in preclinical and Phase 2 studies ongoing in patients with rheumatoid arthritis, lupus, and chronic spontaneous urticaria. On the basis of its potency, selectivity, long target residence time, and noncovalent mode of inhibition, 29 has the potential to be a best-in-class Btk inhibitor for a wide range of immunological indications.

Discovery of 5-Azaindazole (GNE-955) as a Potent Pan-Pim Inhibitor with Optimized Bioavailability.

Pim kinases have been identified as promising therapeutic targets for hematologic-oncology indications, including multiple myeloma and certain leukemia. Here, we describe our continued efforts in optimizing a lead series by improving bioavailability while maintaining high inhibitory potency against all three Pim kinase isoforms. The discovery of extensive intestinal metabolism and major metabolites helped refine our design strategy, and we observed that optimizing the pharmacokinetic properties first and potency second was a more successful approach than the reverse. In the resulting work, novel analogs such as 20 (GNE-955) were discovered bearing 5-azaindazole core with noncanonical hydrogen bonding to the hinge.

GluN2A-Selective Pyridopyrimidinone Series of NMDAR Positive Allosteric Modulators with an Improved in Vivo Profile.

The N-methyl-d-aspartate receptor (NMDAR) is an ionotropic glutamate receptor, gated by the endogenous coagonists glutamate and glycine, permeable to Ca2+ and Na+. NMDAR dysfunction is associated with numerous neurological and psychiatric disorders, including schizophrenia, depression, and Alzheimer's disease. Recently, we have disclosed GNE-0723 (1), a GluN2A subunit-selective and brain-penetrant positive allosteric modulator (PAM) of NMDARs. This work highlights the discovery of a related pyridopyrimidinone core with distinct structure-activity relationships, despite the structural similarity to GNE-0723. GNE-5729 (13), a pyridopyrimidinone-based NMDAR PAM, was identified with both an improved pharmacokinetic profile and increased selectivity against AMPARs. We also include X-ray structure analysis and modeling to propose hypotheses for the activity and selectivity differences.

Discovery of GluN2A-Selective NMDA Receptor Positive Allosteric Modulators (PAMs): Tuning Deactivation Kinetics via Structure-Based Design.

The N-methyl-D-aspartate receptor (NMDAR) is a Na(+) and Ca(2+) permeable ionotropic glutamate receptor that is activated by the coagonists glycine and glutamate. NMDARs are critical to synaptic signaling and plasticity, and their dysfunction has been implicated in a number of neurological disorders, including schizophrenia, depression, and Alzheimer's disease. Herein we describe the discovery of potent GluN2A-selective NMDAR positive allosteric modulators (PAMs) starting from a high-throughput screening hit. Using structure-based design, we sought to increase potency at the GluN2A subtype, while improving selectivity against related α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPARs). The structure-activity relationship of channel deactivation kinetics was studied using a combination of electrophysiology and protein crystallography. Effective incorporation of these strategies resulted in the discovery of GNE-0723 (46), a highly potent and brain penetrant GluN2A-selective NMDAR PAM suitable for in vivo characterization.

Probing Mechanisms of CYP3A Time-Dependent Inhibition Using a Truncated Model System.

Time-dependent inhibition (TDI) of cytochrome P450 (CYP) enzymes may incur serious undesirable drug-drug interactions and in rare cases drug-induced idiosyncratic toxicity. The reactive metabolites are often generated through multiple sequential biotransformations and form adducts with CYP enzymes to inactivate their function. The complexity of these processes makes addressing TDI liability very challenging. Strategies to mitigate TDI are therefore highly valuable in discovering safe therapies to benefit patients. In this Letter, we disclose our simplified approach toward addressing CYP3A TDI liabilities, guided by metabolic mechanism hypotheses. By adding a methyl group onto the α carbon of a basic amine, TDI activities of both the truncated and full molecules (7a and 11) were completely eliminated. We propose that truncated molecules, albeit with caveats, may be used as surrogates for full molecules to investigate TDI.

Discovery and Biological Profiling of Potent and Selective mTOR Inhibitor GDC-0349.

Aberrant activation of the PI3K-Akt-mTOR signaling pathway has been observed in human tumors and tumor cell lines, indicating that these protein kinases may be attractive therapeutic targets for treating cancer. Optimization of advanced lead 1 culminated in the discovery of clinical development candidate 8h, GDC-0349, a potent and selective ATP-competitive inhibitor of mTOR. GDC-0349 demonstrates pathway modulation and dose-dependent efficacy in mouse xenograft cancer models.

Evaluation of time-dependent cytochrome p450 inhibition in a high-throughput, automated assay: introducing a novel area under the curve shift approach.

Early in the drug discovery process, the identification of cytochrome P450 (CYP) time-dependent inhibition (TDI) is an important step for compound optimization. Here we describe a high-throughput, automated method for the evaluation of TDI utilizing human liver microsomes and conventional CYP-specific mass spectrometer-based probes in a 384-well format. One of the key differences from other published TDI assays is the use of a shift in area the under curve of the percent activity remaining versus inhibitor concentration plot (AUC shift) rather than the traditional fold-shift in IC50, to determine the magnitude of TDI. An AUC shift of < 15% suggests negative TDI and > 15% suggests potential TDI. This AUC shift was used to achieve quantitative data reporting, even in the case of weak inhibitors for which IC50 values cannot be quantified. An Agilent Technologies BioCel 1200 System was programmed such that the TDI liability of up to 77 test compounds, incubated at four test concentrations, with and without NADPH in the pre-incubation, can be analyzed in a single run. The detailed automated methodology, assay validation, data reporting and the novel TDI AUC shift approach to describe magnitude of TDI are presented.

Comparison of the function and expression of CYP26A1 and CYP26B1, the two retinoic acid hydroxylases.

All-trans-retinoic acid (atRA) is an important signaling molecule in all chordates. The cytochrome P450 enzymes CYP26 are believed to partially regulate cellular concentrations of atRA via oxidative metabolism and hence affect retinoid homeostasis and signaling. CYP26A1 and CYP26B1 are atRA hydroxylases that catalyze formation of similar metabolites in cell systems. However, they have only 40% sequence similarity suggesting differences between the two enzymes. The aim of this study was to determine whether CYP26A1 and CYP26B1 have similar catalytic activity, form different metabolites from atRA and are expressed in different tissues in adults. The mRNA expression of CYP26A1 and CYP26B1 correlated between human tissues except for human cerebellum in which CYP26B1 was the predominant CYP26 and liver in which CYP26A1 dominated. Quantification of CYP26A1 and CYP26B1 protein in human tissues was in agreement with the mRNA expression and showed correlation between the two isoforms. Qualitatively, recombinant CYP26A1 and CYP26B1 formed the same primary and sequential metabolites from atRA. Quantitatively, CYP26B1 had a lower K(m) (19nM) and V(max) (0.8 pmol/min/pmol) than CYP26A1 (K(m)=50 nM and V(max)=10 pmol/min/pmol) for formation of 4-OH-RA. The major atRA metabolites 4-OH-RA, 18-OH-RA and 4-oxo-RA were all substrates of CYP26A1 and CYP26B1, and CYP26A1 had a 2-10-fold higher catalytic activity towards all substrates tested. This study shows that CYP26A1 and CYP26B1 are qualitatively similar RA hydroxylases with overlapping expression profiles. CYP26A1 has higher catalytic activity than CYP26B1 and seems to be responsible for metabolism of atRA in tissues that function as a barrier for atRA exposure.

Discovery of novel PI3-kinase δ specific inhibitors for the treatment of rheumatoid arthritis: taming CYP3A4 time-dependent inhibition.

PI3Kδ is a lipid kinase and a member of a larger family of enzymes, PI3K class IA(α, ß, δ) and IB (γ), which catalyze the phosphorylation of PIP2 to PIP3. PI3Kδ is mainly expressed in leukocytes, where it plays a critical, nonredundant role in B cell receptor mediated signaling and provides an attractive opportunity to treat diseases where B cell activity is essential, e.g., rheumatoid arthritis. We report the discovery of novel, potent, and selective PI3Kδ inhibitors and describe a structural hypothesis for isoform (α, ß, γ) selectivity gained from interactions in the affinity pocket. The critical component of our initial pharmacophore for isoform selectivity was strongly associated with CYP3A4 time-dependent inhibition (TDI). We describe a variety of strategies and methods for monitoring and attenuating TDI. Ultimately, a structure-based design approach was employed to identify a suitable structural replacement for further optimization.

Changes in maternal liver Cyp2c and Cyp2d expression and activity during rat pregnancy.

During human pregnancy, CYP2C9, CYP2C19, and CYP2D6 activities are altered. The aim of the current study was to determine if this phenomenon can be replicated in the rat, and to evaluate the mechanisms that contribute to the changes in Cyp2c and Cyp2d activity during pregnancy. The intrinsic clearance of dextromethorphan O-demethylation, a measure of Cyp2d2 activity, was decreased 80% at both days 9 and 19 of gestation when compared to non-pregnant controls. The decreased intrinsic clearance was a result of both decreased V(max) and increased K(m)-values at both days of gestation. Quantitative RT-PCR revealed that transcripts of Cyp2d2 and Cyp2d4 were significantly decreased at day 19 of pregnancy (p<0.05) when compared to day 9 and non-pregnant controls. The decrease in Cyp2d mRNA levels correlated with a decrease in several nuclear receptor mRNA levels (RARalpha, RXRalpha, HNF1 and HNF3beta) but not with the mRNA levels of nuclear receptors usually associated with regulation of P450 enzymes (PXR, CAR and HNF4alpha). In contrast, Cyp2c12 and Cyp2c6 transcription and protein expression were not significantly altered during rat pregnancy although the intrinsic clearance of Cyp2c6 mediated diclofenac 4'-hydroxylation was increased 2-fold on day 19 of gestation when compared to non-pregnant controls. The increase in intrinsic clearance was due to a decrease in the K(m)-value for 4'-hydroxydiclofenac formation. These data show that pregnancy significantly alters the expression and activity of drug metabolizing enzymes in an enzyme and gestational stage specific manner. These changes are likely to have toxicological and therapeutic implications.