Drug toxicity to T-antigen-immortalized human liver epithelial (THLE) cells stably transfected with plasmid vectors that encoded human cytochrome P450s 1A2, 2C9, 2C19, 2D6, or 3A4, or an empty plasmid vector (THLE-Null), was investigated. An automated screening platform, which included 1% dimethyl sulfoxide (DMSO) vehicle, 2.7% bovine serum in the culture medium, and assessed 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium reduction, was used to evaluate the cytotoxicity of 103 drugs after 24h. Twenty-two drugs caused cytotoxicity to THLE-Null cells, with EC₅₀ ≤ 200 μM; 21 of these drugs (95%) have been reported to cause human liver injury. Eleven drugs exhibited lower EC₅₀ values in cells transfected with CYP3A4 (THLE-3A4 cells) than in THLE-Null cells; 10 of these drugs (91%) caused human liver injury. An additional 8 drugs, all of which caused human liver injury, exhibited potentiated THLE-3A4 cell toxicity when evaluated using a manual protocol that included 0.2% or 1% DMSO, but not bovine serum. Fourteen of the drugs that exhibited potentiated THLE-3A4 cell toxicity are known to be metabolized by P450s to reactive intermediates. These drugs included troglitazone, which was shown to undergo metabolic bioactivation and covalent binding to proteins in THLE-3A4 cells. A single drug (rimonabant) exhibited marked THLE cell toxicity but did not cause human liver injury; this drug had very low reported plasma exposure. These results indicate that evaluation of toxicity to THLE-Null and THLE-3A4 cell lines during drug discovery may aid selection of drugs with reduced propensity to cause drug-induced liver injury and that consideration of human exposure is required to enhance data interpretation.

The carboxylic acid NSAID fenclozic acid exhibited an excellent preclinical safety profile and promising clinical efficacy, yet was withdrawn from clinical development in 1971 due to hepatotoxicity observed in clinical trials. A variety of modern in vitro approaches have been used to explore potential underlying mechanisms. Covalent binding studies were undertaken with [(14)C]-fenclozic acid to investigate the possible role of reactive metabolites. Time-dependent covalent binding to protein was observed in NADPH-supplemented liver microsomes, although no metabolites were detected in these incubations or in reactive metabolite trapping experiments. In human hepatocytes, covalent binding was observed at lower levels than in microsomes and a minor uncharacterizable metabolite was also observed. In addition, covalent binding was observed in incubations undertaken with dog and rat hepatocytes, where a taurine conjugate of the drug was detected. Although an acyl glucuronide metabolite was detected when liver microsomes from human, rat and dog were supplemented with UDPGA, there was no detectable UDPGA-dependent covalent binding. No effects were observed when fenclozic acid was assessed for P450-dependent and P450-independent cytotoxicity to THLE cell lines, time-dependent inhibition of five major human cytochrome P450 enzymes, inhibition of the biliary efflux transporters BSEP and MRP2 or mitochondrial toxicity to THLE or HepG2 cells. These data suggest that Phase 1 bioactivation plays a role in the hepatotoxicity of fenclozic acid and highlight the unique insight into mechanisms of human drug toxicity that can be provided by investigations of biotransformation and covalent binding to proteins.

The oral dipeptidyl peptidase 1 (DPP1) inhibitor AZD5248 showed aortic binding in a rat quantitative whole-body autoradiography (QWBA) study, and its development was terminated prior to human dosing. A mechanistic hypothesis for this finding was established invoking reactivity with aldehydes involved in the cross-linking of elastin, a major component of aortic tissue. This was tested by developing a simple aldehyde chemical reactivity assay and a novel in vitro competitive covalent binding assay. Results obtained with AZD5248, literature compounds, and close analogues of AZD5248 support the mechanistic hypothesis and provide validation for the use of these assays in a two tier screening approach to support lead optimization. The strengths and limitations of these assays are discussed.

Four poly(ADP-ribose) polymerase (PARP) inhibitors have now presented phase 3 monotherapy data showing compelling benefit of targeting tumours enriched with DNA damage response (DDR) pathway deficiencies, including BRCA gene mutations. Indirect treatment comparisons using the published clinical data from these late stage trials suggest similar levels of monotherapy efficacy are observed in spite of reported differences in PARP trapping potency. However, there is greater diversity in the observed safety profiles. To try and understand these observations, we have carried out a head-to-head comparison of these four PARP inhibitors (olaparib, niraparib, rucaparib and talazoparib) as well as veliparib, which recently reported phase 3 chemotherapy combination data. In our studies, we included an assessment of molecular mechanism of action that included PAR inhibition, PARP trapping and synthetic lethality in isogenic BRCA mutant and wild type models. In addition, an assessment of selectivity in terms of both inhibition of PARP family members using a novel chemoproteomic approach, as well as secondary (off-target) activities was performed. Finally, effects on human haematopoietic stem cell viability and bio-distribution to bone marrow in the rat were tested and compared. A detailed correlation of our datasets with the observed clinical results, including adverse events, suggests these preclinical experiments provide an excellent predictor of clinical response and could be used to assess emerging as well as novel PARP inhibitors. OlaparibVeliparibRucaparibNiraparibTalazoparibCompanyAZAbbVieClovisTesaroPfizerPhaseApprovedIIIApprovedApprovedIIIPARP1 SPR Kd (µM)0.0010.0070.0010.0130.002PARP2 SPR Kd (µM)0.0010.0140.0230.0430.005PARPs with Proteomic Kd <1 µM1,2,3,4,131,2,3,4,131,2,3,4,10,131,2,131,2,3,4,5a,13, 16Sec. Pharm. #, top hit µM0/855/85 5HT7, 0.513/85 5HT4, 0.517/84 DAT, 0.040/85Monotherapy dose (mg)300 bd (tablet)500 bd600 bd300 od1 od Citation Format: Elisabetta Leo, Jeffrey Johannes, Giuditta Illuzzi, Andrew Zhang, Paul Hemsley, Michal J. Bista, Jonathan P. Orme, Verity A. Talbot, Ana J. Narvaez, Elizabeth Underwood, Andrew Pike, Jenni K. Nikkila, Lucy Riches, Sinbad Sweeney, Frida Gustafsson, Anna Cronin, Piero Ricchiuto, Debora A. Roaquin, Fiona Pachl, Eric Miele, Ruth MacDonald, Glen Hawthorne, Andrew N. Mead, Mark J. O'Connor. A head-to-head comparison of the properties of five clinical PARP inhibitors identifies new insights that can explain both the observed clinical efficacy and safety profiles [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr LB-273.

Human methionine synthase reductase (MSR), a diflavin oxidoreductase, plays a vital role in methionine and folate metabolism by sustaining methionine synthase (MS) activity. MSR catalyzes the oxidation of NADPH and shuttles electrons via its FAD and FMN cofactors to inactive MS-cob(II)alamin. A conserved aromatic residue (Trp697) positioned next to the FAD isoalloxazine ring controls nicotinamide binding and catalysis in related flavoproteins. We created four MSR mutants (W697S, W697H, S698Δ, and S698A) and studied their associated kinetic behavior. Multiwavelength stopped-flow analysis reveals that NADPH reduction of the C-terminal Ser698 mutants occurs in three resolvable kinetic steps encompassing transfer of a hydride ion to FAD, semiquinone formation (indicating FAD to FMN electron transfer), and slow flavin reduction by a second molecule of NADPH. Corresponding experiments with the W697 mutants show a two-step flavin reduction without an observable semiquinone intermediate, indicating that W697 supports FAD to FMN electron transfer. Accelerated rates of FAD reduction, steady-state cytochrome c(3+) turnover, and uncoupled NADPH oxidation in the S698Δ and W697H mutants may be attributed to a decrease in the energy barrier for displacement of W697 by NADPH. Binding of NADP(+), but not 2',5'-ADP, is tighter for all mutants than for native MSR. The combined studies demonstrate that while W697 attenuates hydride transfer, it ensures coenzyme selectivity and accelerates FAD to FMN electron transfer. Moreover, analysis of analogous cytochrome P450 reductase (CPR) variants points to key differences in the driving force for flavin reduction and suggests that the conserved FAD stacking tryptophan residue in CPR also promotes interflavin electron transfer.