Cutting-edge multi-level analytical and structural characterization of antibody-drug conjugates: present and future

ABSTRACT

Introduction: The development and optimization of antibody drug conjugates (ADCs) rely on improving their analytical and bioanalytical characterization, by assessing critical quality attributes (CQAs). Among the CQAs, the glycoprofile, drug load distribution (DLD), the amount of unconjugated antibody (D0), the average drug-to-antibody ratio (DAR), the drug conjugation sites and the residual drug-linker and related product proportions (SMDs) in addition to high and low molecular weight species (H/LMWS), and charge variants are the most important ones.

Areas covered: The analytical and structural toolbox for the characterization of 1^st, 2^d and 3^d generation ADCs was significantly extended in the last 3 years. Here, we reviewed state-of-the-art techniques, such as liquid chromatography, high resolution native and ion mobility mass spectrometry, multidimensional liquid chromatography and capillary electrophoresis hyphenated to mass spectrometry, reported mainly since 2016.

Expert commentary: These emerging techniques allow a deep insight into important CQAs that are related to ADC Chemistry Manufacturing and Control (CMC) as well as an improved understanding of in vitro and in vivo ADC biotransformations. This knowledge and the development of quantitative bioanalytical assays will continue to contribute to early-developability assessment for the optimization of all the ADC components (i.e. antibody, drug, and linker) and help to bring next-generation ADCs into late clinical development and to the market.

KEYWORDS:

• Brentuximab vedotin
• trastuzumab emtansine
• gemtuzumab ozogamicin
• inotuzumab ozogamicin
• sacituzumab govitecan
• 3G-ADCs
• small-protein fragment-DCs

Article highlights

• The inherent complexity of ADCs, which results from the heterogeneity and PTMs, the linker-payload conjugation, combined with changes during manufacturing processes, purification and storage necessitate advanced methodology for their characterization.

• High-resolution native MS provides accurate mass measurements (within 30 ppm) of intact ADCs and can also yield the drug load distribution (DLD) and average DAR. Native MS is furthermore unique in its ability to simultaneously detect covalent and non-covalent species within a mixture.

• Native IM-MS reveals the drug-loading profile of ADCs, the CCCs of each payload species, highlighting slight conformational differences.

• Additional developments are needed for highly hydrophobic payloads such as PBDs [231,232] or highly loaded payloads.

• As an orthogonal method, LC-MS following IdeS digestion of ADCs can be used to measure the drug-load distribution on light chain and Fd fragments, as well as the average DAR for both monomeric and multimeric species. In addition, the Fc fragment can be analyzed in the same run, providing a complete glycoprofile and demonstrating the presence or absence of additional conjugation.

• The linker-payload and glycan modifications of Cys linked ADCs can be simultaneously characterized using a HILIC-MS middle-up analysis, which is orthogonal to RPLC-MS.

• Because the drug molecules are frequently hydrophobic, all enzymatic digestion steps used to prepare the sample for peptide mapping have to be adapted to keep the drug-loaded peptides soluble. When the procedures are optimized, unambiguous maps are obtained by LC-MS, while the positional isomers of the cytotoxic drug can be determined by RP-HPLC after digestion with IdeS and reduction.

• Multiple Attributes Method (MAM) based on HRMS data with automated identification and quantitation to monitor CQAs in one single analysis for mAbs may also be used for ADCs.

• All these methods are also useful for mAbs and ADCs comparability [233], forced degradation [234] and developability studies [235].On the bioanalytic side, LC-MS quantitation of larg biotherapeutic proteins such as mAbs and ADCs at the intact level presents many challenges in both LC and MS due to their high molecular weight, large size, structural complexity, and heterogeneity.

• ADCs can be quantitated at the intact level with highly selective sample preparation combined with LC-HRMS analysis. Technological advances in processing software, protein chromatographic columns, ionization techniques, and high-resolution mass spectrometry are still required to successfully quantify large intact proteins at much lower levels.

• Due to the implementation of an electrophoretic separation, CE has demonstrated to be particularly relevant for the characterization of ADCs. Depending on the separation mode applied, CE can provide, on the intact or/and reduced ADCs level, the analysis of a substantial number of micro-heterogeneities which includes size or charge variants, the number of conjugated drugs, glycoforms and conventional PTMs.

• CE-MS/MS represents an emerging analytical technique able to provide a comprehensive characterization of the primary structure of ADCs. Due to the intrinsic properties of CE-MS/MS, the analysis of an ADC peptide mixture allows the characterization of the amino acid sequence and PTMs with an outstanding robustness, in addition CE-MS/MS analysis enables to locate the conjugated drug on the peptide backbone and estimate the conjugation level independently. As such, CE-MS/MS appears as an orthogonal methodology compared to LC-MS/MS peptide mapping for the characterization of ADCs.

• Liquid chromatography coupled online to surface plasmon resonance (LC-SPR) to allow label-free affinity evaluation of mAb and ADC sample constituents (size and charge variants), under near-native conditions is another emerging analytical method and recently reported by Lakayan et al. [236]

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was supported in part by the Swiss National Science Foundation (to D. Guillarme and S. Fekete: 31003A 159494); the ‘Association Nationale de la Recherche et de la Technologie’ (ANRT) and Syndivia (to A. Ehkirch); and the IdeX program of the University of Strasbourg (to O. Hernandez-Alba). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.