Pharmacodynamics-based approach for efficacious human dose projection of BMS-986260, a small molecule transforming growth factor beta receptor 1 inhibitor.

Transforming growth factor beta (TGF-ß) is a pleiotropic cytokine that has a wide array of biological effects. For decades, tumor biology implicated TGF-ß as an attractive therapeutic target due to its immunosuppressive effects. Toward this end, multiple pharmaceutical companies developed a number of drug modalities that specifically target the TGF-ß pathway. BMS-986260 is a small molecule, selective TGF-ßR1 kinase inhibitor that was under preclinical development for oncology. In vivo studies across mouse, rat, dog, and monkey and cryopreserved hepatocytes predicted human pharmacokinetics (PK) and distribution of BMS-986260. Efficacy studies of BMS-986260 were undertaken in the MC38 murine colon cancer model, and target engagement, as measured by phosphorylation of SMAD2/3, was assessed in whole blood to predict the clinical efficacious dose. The human clearance is predicted to be low, 4.25 ml/min/kg. BMS-986260 provided a durable and robust antitumor response at 3.75 mg/kg daily and 1.88 mg/kg twice-daily dosing regimens. Phosphorylation of SMAD2/3 was 3.5-fold less potent in human monocytes than other preclinical species. Taken together, the projected clinical efficacious dose was 600 mg QD or 210 mg BID for 3 days followed by a 4-day drug holiday. Mechanism-based cardiovascular findings in the rat ultimately led to the termination of BMS-986260. This study describes the preclinical PK characterization and pharmacodynamics-based efficacious dose projection of a novel small molecule TGF-ßR1 inhibitor.

Investigation of a Degradant in a Biologics Formulation Buffer Containing L-Histidine.

PURPOSE: An unknown UV 280 nm absorbing peak was observed by SEC for protein stability samples formulated in L-histidine during a stress stability study. Understanding the source would enhance the confidence in the SEC results. We identified the unknown peak, studied the cause, and evaluated ways to eliminate it. METHODS: The unknown peak was fractionated by preparative size exclusion chromatography separations, and subsequently analyzed by Hydrophilic Interaction Chromatography (HILIC) coupled with Time-of-Flight (TOF) high resolution mass spectrometry. The possible degradation was also studied with the presence of different excipients, including metal cations, chelating agents, and amino acids. RESULTS: The unknown peak was identified to be trans-urocanic acid, a degradant of histidine, based on evidences from HILIC retention time, UV profile, accurate mass measurement, trans-cis isomerization, and pI measurement. The degradation from histidine to urocanic acids was not affected by the presence of Fe(2+), but slightly activated by Mn(2+). The chelating agents, EDTA and DTPA, counteracted the Mn(2+) effects. This degradation was evidenced to be caused by contamination. Adding alanine or cysteine as an excipient was found to reduce this degradation by 97 and 98%, respectively. CONCLUSIONS: L-histidine formulation buffer can be contaminated to induce histidine degradation to trans-urocanic acid, which shows a large UV 280 nm absorbing peak at the total permeation volume under SEC conditions. Amino acids alanine and cysteine effectively inhibit this histidine degradation.

Non-basic azolotriazinone MCHR1 antagonists for the treatment of obesity: An empirical brain-exposures-driven candidate selection for in vivo efficacy studies.

Non-basic azolotriazinones were explored using an empirical free brain exposures-driven approach to identify potent MCHR1 antagonists for evaluation in in vivo efficacy studies. An optimized lead from this series, 1j (rMCHR1 Ki=1.8 nM), demonstrated a 6.9% reduction in weight gain relative to vehicle in a rat model at 30 mg/kg after 4 days of once-daily oral treatment as a glycine prodrug. Despite a promising efficacy profile, an assessment of the biliary toxicity risk of this compound rendered this compound non-progressible.

Dihydropyrrolopyrazol-6-one MCHR1 antagonists for the treatment of obesity: Insights on in vivo efficacy from a novel FLIPR assay setup.

Our investigation of the structure-activity and structure-liability relationships for dihydropyrrolopyrazol-6-one MCHR1 antagonists revealed that off-rate characteristics, inferred from potencies in a FLIPR assay following a 2 h incubation, can impact in vivo efficacy. The in vitro and exposure profiles of dihydropyrrolopyrazol-6-ones 1b and 1e were comparable to that of the thienopyrimidinone counterparts 41 and 43 except for a much faster MCHR1 apparent off-rate. The greatly diminished dihydropyrrolopyrazol-6-one anti-obesity response may be the consequence of this rapid off-rate.

Estimation of the extent of in vivo formation of a mutagenic aromatic amine from a potent thyromimetic compound: correlation of in vitro and in vivo findings.

The development of compounds with the potential for genotoxicity poses significant safety risks as well as risks of attrition. Although genotoxicity evaluation of the parent molecule is routine and reasonably predictive, assessing the risk of commercialization when release of a genotoxic degradant and/or metabolite from a nongenotoxic parent molecule is suspected is much more challenging and resource intensive. Much of the risk of the formation of a genotoxic degradant/metabolite can be discharged with the conduct of carcinogenicity studies in models where the compound is formed, but this approach requires a great deal of time and resources. In this manuscript, we investigated the contribution of various factors (pH, serum instability, and hepatic metabolism) to the formation of a mutagenic aromatic amine from a potent and highly selective thyromimetic compound ([3-(3,5-dibromo-4-(4-hydroxy-3-isopropyl-5-methylphenoxy)-2-methylphenylamino)-3-oxopropanoic acid], compound 1), under in vitro conditions. The kinetic parameters obtained from in vitro experiments combined with the pharmacokinetics of 1in vivo (e.g., plasma concentration-time profile and clearance) were used to estimate the extent of in vivo formation of [4-(4-amino-2,6-dibromo-3-methylphenoxy)-2-isopropyl-6-methylphenol] (compound 2), in rats upon administration of a single oral dose of 1. The agreement between the predicted values (1.9% conversion of total administered dose) with the observed levels of 2 in rats (0.2%-2.2% of the 10 mg/kg dose, 10 mg/kg) further prompted the utilization of this approach to predict the extent of release of this mutagen in humans upon administration of 1. The projection of 0.13% conversion to 2 from an efficacious daily dose of 15 mg of 1 translated to the generation of 20 µg of 2 and provided the basis for the decision to terminate the development of 1.

Discovery of a Parenteral Small Molecule Coagulation Factor XIa Inhibitor Clinical Candidate (BMS-962212).

Factor XIa (FXIa) is a blood coagulation enzyme that is involved in the amplification of thrombin generation. Mounting evidence suggests that direct inhibition of FXIa can block pathologic thrombus formation while preserving normal hemostasis. Preclinical studies using a variety of approaches to reduce FXIa activity, including direct inhibitors of FXIa, have demonstrated good antithrombotic efficacy without increasing bleeding. On the basis of this potential, we targeted our efforts at identifying potent inhibitors of FXIa with a focus on discovering an acute antithrombotic agent for use in a hospital setting. Herein we describe the discovery of a potent FXIa clinical candidate, 55 (FXIa Ki = 0.7 nM), with excellent preclinical efficacy in thrombosis models and aqueous solubility suitable for intravenous administration. BMS-962212 is a reversible, direct, and highly selective small molecule inhibitor of FXIa.

A rabbit model for sublingual drug delivery: comparison with human pharmacokinetic studies of propranolol, verapamil and captopril.

A rabbit model for investigating sublingual drug absorption was established yielding results consistent with clinical data reported in the literature. Using propranolol as a model compound the effect of formulation and dosing variables was explored as a means to characterize the limiting parameters of this model. In addition, verapamil and captopril were selected as reference compounds to compare this model to sublingual absorption in humans. Rabbits were dosed sublingually and systemic absorption was measured over time. Sublingual absorption of propranolol was dependent on dosing solution pH and volume. Intra-oral spray device did not affect the overall exposure compared to instillation using a syringe. Despite species and dosing regimen differences the relative bioavailabilities of propranolol and verapamil were very similar in rabbits and humans. In contrast, captopril absorption from the sublingual cavity of rabbits was low and did not agree with that observed in man. Here we report a sublingual rabbit model of drug delivery and its potential utility in preclinical development of intra-oral dosage forms.

Permeability characteristics of calu-3 human bronchial epithelial cells: in vitro-in vivo correlation to predict lung absorption in rats.

The goal of this study was to evaluate the permeability characteristics of Calu-3, human bronchial epithelial cells to passive and actively transported drugs and to correlate the data with other in vitro models and rat lung absorption in vivo. Air-interface cultured Calu-3 cells grown on collagen-coated permeable filter supports formed "tight" polarized and well differentiated cell monolayers with apical microvilli and tight-junctional complexes. Within 8-10 days, cell monolayers developed trans-epithelial electrical resistance (TEER) > 1000 ohm cm2 and potential difference about 11-16 mV. Solute permeability was dependent on lipophilicity, and inversely related to molecular size. Calu-3 cells actively transported amino acids, nucleosides and dipeptide analogs, but not organic anions, organic cations or efflux pump substrates. The permeability characteristics of Calu-3 cells correlated well with primary cultured rabbit tracheal epithelial cells in vitro (r2 = 0.91), and the rate of drug absorption from the rat lung in vivo (r2 = 0.94). The absorption predicted from the regression equation correlated well with observed values. In conclusion, in vitro-in vivo correlation studies indicate that the Calu-3 cell culture model is a potentially useful model to predict absorption of inhalation delivery drug candidates.

Identification of a nonbasic melanin hormone receptor 1 antagonist as an antiobesity clinical candidate.

Identification of MCHR1 antagonists with a preclinical safety profile to support clinical evaluation as antiobesity agents has been a challenge. Our finding that a basic moiety is not required for MCHR1 antagonists to achieve high affinity allowed us to explore structures less prone to off-target activities such as hERG inhibition. We report the SAR evolution of hydroxylated thienopyrimidinone ethers culminating in the identification of 27 (BMS-819881), which entered obesity clinical trials as the phosphate ester prodrug 35 (BMS-830216).

Application of a kosmotrope-based solubility assay to multiple protein therapeutic classes indicates broad use as a high-throughput screen for protein therapeutic aggregation propensity.

Aggregation propensity is a critical attribute of protein therapeutics that can influence production, manufacturing, delivery, and potential activity and safety (immunogenicity). It is therefore imperative to select molecules with low aggregation propensity in the early stages of drug discovery to mitigate the risk of delays or failure in clinical development. Although many biophysical methods have been developed to characterize protein aggregation, most established methods are low-throughput, requiring large quantities of protein, lengthy assay times, and/or significant upstream sample preparation, which can limit application in early candidate screening. To avoid these limitations, we developed a reliable method to characterize aggregation propensity, by measuring the relative solubility of protein therapeutic candidates in the presence of the kosmotropic salt ammonium sulfate. Manual bench-scale and automated plate-based methods were applied to different protein therapeutic formats including Adnectins, domain antibodies, PEGylated Adnectins, Fc fusion proteins, and monoclonal antibodies. The kosmotrope solubility data agreed well with the aggregation propensity observed by established methods, while being amenable to high-throughput screening because of speed, simplicity, versatility and low protein material requirements. The results suggest that kosmotrope-based solubility assessment has broad applicability to selecting protein therapeutic candidates with low aggregation propensity and high "developability" to progress into development.

Reductions in log P improved protein binding and clearance predictions enabling the prospective design of cannabinoid receptor (CB1) antagonists with desired pharmacokinetic properties.

Several strategies have been employed to reduce the long in vivo half-life of our lead CB1 antagonist, triazolopyridazinone 3, to differentiate the pharmacokinetic profile versus the lead clinical compounds. An in vitro and in vivo clearance data set revealed a lack of correlation; however, when compounds with <5% free fraction were excluded, a more predictable correlation was observed. Compounds with log P between 3 and 4 were likely to have significant free fraction, so we designed compounds in this range to give more predictable clearance values. This strategy produced compounds with desirable in vivo half-lives, ultimately leading to the discovery of compound 46. The progression of compound 46 was halted due to the contemporaneous marketing and clinical withdrawal of other centrally acting CB1 antagonists; however, the design strategy successfully delivered a potent CB1 antagonist with the desired pharmacokinetic properties and a clean off-target profile.