(M1330-01-06) In Vitro Assessment of Physicochemical Properties of Biologics for the Early Identification of Pharmacokinetic Liabilities in a Developability Workflow

Purpose: Historically, the assessment of pharmacokinetics (PK) of an Fc-asset biotherapeutic has primarily relied onin vivo characterization. Recently, it has emerged in the literature that certain physicochemical properties of biotherapeutics can be assessed in vitro with strong correlative potential to in vivo behavior[1]. Consequently, in vitro screening approaches, comparable to that of Lipinski’s rule of 5 for small molecules, are emerging as an evolving field of science for biotherapeutics. Applying this along a protein engineering developability workflow, it follows that physicochemical attributes relating to target binding domain optimization and framework selection can have negative PK implications, which can be identified using early in vitro assessments in lieu of conducting in vivo PK studies[2-4]. Evaluation of disparate in vitro measured properties, in relation to measured PK parameters, highlight that the most correlative physicochemical properties tend to fall into categorical associations including charge based or non-specific interactions, self-association or aggregation, and biological interaction with the FcRn receptor[1,5]. In this work a series of in vitro assays addressing different categorical physicochemical properties was implemented into the developability workflow to aid in the earlier identification of PK liabilities of biologic modalities. Accelerated identification of PK liabilities would provide numerous benefits to drug development efforts, reducing the time and effort expended on sub-optimal molecular material, thereby reducing the cost, expense, and animal expenditure of non-essential in vivo studies.

Methods: Following a comprehensive literature review and an internal validation process, four in vitro assays measuring varied categorical physicochemical properties were identified as potentially impactful predictors of PK liability: >A baculovirus particle (BVP) binding assay assessing for non-specific binding, a cross interaction chromatography assay (CIC) assessing for self-interaction/association, and an FcRn column assay and FcRn cellular recycling/transcytosis assay assessing for FcRn interaction. These assays were selected based upon a multifaceted scoring motif, selecting for correlative potential to PK, appropriate throughput to be able to inform against the numbers of molecules encountered in early screening campaigns, and the ability to be easily implemented into current developability workflow schema. These assays were internally developed and implemented, then utilized to evaluate a diverse set of 43 monoclonal antibodies (mAbs) selected from an early discovery campaign which exhibited a known range of developability properties that would be representative of an early-stage protein production campaign. The individual assays were validated to ensure differentiation of in vitro output scores could be achieved, and then used to evaluate and rank order the mAbs in terms of developability risk criteria.  Finally, each individual mAb was administered to hFcRn Tg32 mice for PK characterization, and the resulting clearance (CL) was used to establish the in-vitro-to-in vivo-correlation (IVIVC) of the assays and help shape the screening cascade for implementation of these in vitro assays into the existing developability protein engineering workflow.

Results: Each individual assay resulted in a positive correlation to in vivo CL, with a range in ρ of ≥ 0.56-0.72 between all assays as follows: CIC assay: ρ = 0.72, BVP assay: ρ = 0.67, FcRn Column assay: ρ = 0.61,  FcRn recycling: ρ = 0.56. CL values for the test set varied greatly, ranging from 0.093 mL/h/Kg to 2.21 mL/h/kg, ensuring that a broad scope of high, medium, and low clearance molecules for correlation were inclusive in the evaluation dataset for the derivation of the IVIVC relationship of each individual assay. Thresholds for acceptance values were established in each in vitro assay to minimize potential for false positive scorings, while maximizing predictability of true positive and negative scoring. All mAbs with CL values > 0.7 mL/h/kg were successfully identified as PK liabilities in at least one of the in vitro assays (n = 9).  Furthermore, clinically marketed mAbs (n = 5; Eculizumab, Ramucirumab, Rituximab, Fremanezumab, and Obinutuzumab) were included as a test set of “well behaved” controlcompounds, and were identified in all assays as classifying with below threshold scores. Collectively, these results demonstrate the potential for this suite of assays to be utilized in a developability workflow as an accurate early-stage predictor of potential PK liability in lieu of conducting an in vivo PK evaluation.

Conclusion: The positive correlations observed between the in vitro assays and in vivo CL values support the concept of utilizing categorical in vitro assay data to augment and direct the implementation of screening approaches in protein engineering screening cascades. Utilization of this strategic approach will allow for early identification of potential PK liabilities, guiding the lead selection and optimization strategies, as well as informing the early-stage protein engineering efforts to ensure that the PK properties of a targeted drug product profile are achievable.

References: Avery et al., 2018, “Establishing in vitro in vivo correlations to screen monoclonal antibodies for physicochemical properties related to favorable human pharmacokinetics”, MAbs, 10(2):244-255.
Jain et al., 2017, “Biophysical properties of the clinical-stage antibody landscape”, PNAS, 114(5):944-949.
Jarasch et al., 2015, “Developability assessment during the selection of novel therapeutic antibodies”, J Pharm Sci., 104(6):1885-1898.
Wolf Pérez et al., 2022, “Assessment of Therapeutic Antibody Developability by Combinations of In Vitro and In Silico Methods”, Methods Mol Biol., 2313:57-113.
Starr et al., 2019, “Selecting and engineering monoclonal antibodies with drug-like specificity”, Curr Opin Biotechnol., 60:119-127.



Acknowledgements: Large Molecule Research (LMR), Sanofi USAa, Germanyb, Francec, Belgiumd: Katrin Reichelb, Christian Langeb, Sagar Kathuriaa, Alessandro Masiero, Tingwan Suna, Sevim Oezguer Bruederleb, Pep Amengual-Rigob, Lena Wahlb, Norbert Furtmannb, Ailin Wanga, Chemistry Manufacturing and Controls (CMC), Sanofi USA: Jonathan Kingsbury, Charles Starr Adimab, USA: Robert Pejchal, Christin Strong, James Geoghegan, Shu Lin, Rebecca Hurley, Eric Krauland