(T-127) Surface plasmon resonance enabled mechanistic pharmacokinetic/pharmacodynamic modeling to support covalent inhibitor drug development: Bruton’s tyrosine kinase inhibitor case study

John Kalvass, Ph.D. – Senior Research Fellow, AbbVie Inc.; Sanjay Panchal, Ph.D. – Principal Research Scientist, AbbVie Inc; Colin Phipps, Ph.D. – Principle Research Scientist, AbbVie Inc; John Wan, Ph.D. – Senior Scientist, AbbVie Inc.

Objectives: Covalent inhibitors are proven to be an effective and versatile modality with approved drugs across multiple therapeutic areas. Predicting the extent and duration of clinical target inhibition by pharmacokinetic/pharmacodynamic (PK/PD) modeling at the drug discovery phase would greatly benefit the candidate drug selection and accelerate project timelines. The quality of modeling results heavily relies on the reliability of input data. However, accurately characterizing the binding kinetic parameters of covalent drugs has been challenging due to the rapid and irreversible generation of inactivated drug-receptor complexes.

Methods: A surface plasmon resonance (SPR) method was developed using His-tagged human Bruton's tyrosine kinase (BTK) protein to determine the binding association rate constant k_on, dissociation rate constant k_off, and inactivation rate constant k_inact for multiple covalent BTK inhibitors including ibrutinib, acalabrutinib, zanubrutinib and ABBV-992. These parameters were incorporated into a mechanistic PK/PD model, together with the published human PK for ibrutinib and zanubrutinib, plus the BTK turn-over rate derived from literature, to simulate the receptor occupancy (RO) time-course in peripheral blood at the clinically tested dose levels. Clinically observed RO data of the approved drugs was used to verify the model predictions.

Results: According to the SPR assay results, ibrutinib showed the most rapid kon among the studied cBTKi, followed by zanubrutinib, acalabrutinib and ABBV-992. Acalabrutinib and zanubrutinib were found to have similar koff values and both dissociated faster than ibrutinib from the target. ABBV-992 had the slowest koff among the studied cBTKi. Zanubrutinib was determined to have the highest kinact, while kinact values for the other cBTKi were within 3-fold. The indicator of intrinsic potency for covalent inhibitors, kinact/KI, was correlated with kon, suggesting superior commitments of covalency for these drugs in general. By integrating the measured binding kinetics, reported clinical PK and target turn-over rate, RO levels predicted by the translational PK/PD model were consistent with the clinical findings in B-cell lymphoma patients for ibrutinib and zanubrutinib. For ABBV-992, the unbound human efficacious exposure predicted to maintain >95% RO with daily dosing was as low as 0.1 µg•h/mL. This implies a desirable drug-like property for this molecule to advance into the clinic.

Conclusions: This work demonstrates a translational PK/PD modeling approach for cBTKi that is enabled by robust binding kinetic measurements through innovative SPR technology. We have illustrated the value of utilizing such a platform, which can be potentially applied to covalent inhibitors for other targets. This framework efficiently facilitates the candidate drug prioritization by intrinsic potency and predicts human efficacious exposure at the preclinical stage with a high level of confidence.

Citations: [1]. Nat Rev Drug Discov 21, 881–898 (2022). https://doi.org/10.1038/s41573-022-00542-z
[2]. J of Med Chem 2019 62 (7), 3228-3250. DOI: 10.1021/acs.jmedchem.9b00167
[3]. J Clin Oncol. 2013;31(1):88-94.
[4]. N Engl J Med. 2016;374(4):323-332.
[5]. Blood. 2019;134(11):851-859.
[6]. Clin Transl Sci. 2020;13(2):325-336.