A Generic Approach for Miniaturized Unbiased High-Throughput Screens of Bispecific Antibodies and Biparatopic Antibody-Drug Conjugates

Abstract

The toolbox of modern antibody engineering allows the design of versatile novel functionalities exceeding nature's repertoire. Many bispecific antibodies comprise heterodimeric Fc portions recently validated through the approval of several bispecific biotherapeutics. While heterodimerization methodologies have been established for low-throughput large-scale production, few approaches exist to overcome the bottleneck of large combinatorial screening efforts that are essential for the identification of the best possible bispecific antibody. This report presents a novel, robust and miniaturized heterodimerization process based on controlled Fab-arm exchange (cFAE), which is applicable to a variety of heterodimeric formats and compatible with automated high-throughput screens. Proof of applicability was shown for two therapeutic molecule classes and two relevant functional screening read-outs. First, the miniaturized production of biparatopic anti-c-MET antibody-drug conjugates served as a proof of concept for their applicability in cytotoxic screenings on tumor cells with different target expression levels. Second, the automated workflow enabled a large unbiased combinatorial screening of biparatopic antibodies and the identification of hits mediating potent c-MET degradation. The presented workflow utilizes standard equipment and may serve as a facile, efficient and robust method for the discovery of innovative therapeutic agents in many laboratories worldwide.

Keywords: DuoBody; automation; bispecific antibody; miniaturization; unbiased screening.

Figure 1

Optimization of miniaturized cFAE via TCEP and re-oxidation, applying dhAA or PEG-azide. (a) Scheme of the DuoBody reaction for heterodimerization. (b) Exemplary raw data set from SCX-HPLC monitoring cFAE over time, upon the addition of a 20-fold molar excess of TCEP at 1 mg/mL. Please refer to supplementary Figure S1 indicating a negligible peak caused by PEG-azide. (c) cFAE kinetics for four TCEP concentrations applying a representative parental DuoBody pair to optimize TCEP concentration and incubation time. (d) Re-oxidation of interchain disulfide bonds using 10- or 20-fold dhAA or 20-fold PEG-azide (molar access over TCEP) compared to buffer exchange. PEG-azide yielded the formation of >90% species of the size expected for the re-oxidized antibody similar to buffer exchange, confirmed by CE-SDS, as shown in Figure S2. Numerical data for sections (b,c) are listed in Table S1.

Figure 2

Concentration and time dependency of the miniaturized reaction. (a) HIC HPLC analyses of the cFAE reaction with a 20-fold molar excess of TCEP for 4 h at 1 mg/mL, 0.3 mg/mL and 0.1 mg/mL, revealing a clear concentration dependency. (b) cFAE time course at 0.3 mg/mL and 0.1 mg/mL indicating slower reaction kinetics but completion at >12 h incubation times. The working condition consisting of 0.7–1 mg/mL in 50–100 µL reaction volume and 4 h of cFAE reaction was found optimal. Numerical data for section (b) are listed in Table S2.

Figure 3

Applicability for biparatopic ADC screening. (a) Scheme of the preparative steps including MTG-mediated linker–payload conjugation and heterodimerization to yield “half-ADC” candidates amendable for functional interrogation. (b) Representative CIEX-HPLC chromatograms indicating the successful generation of an REGN5093-based antibody–ADC heterodimer/biparatopic ADC (bpADC) applying the conditions optimized for bispecific antibody generation. (c) Triplicate analyses of biparatopic ADC formation at 1 mg/mL, 40 eq. TCEP and 6 h incubation time (please refer also to Figure S3). (d) Applicability of the generated half ADCs for functional screening such as for cytotoxic potency. Please note that the higher potency of the parental ADC is due to a higher DAR. (e) Potency determination of the samples shown in (d), demonstrating the option to screen for functionality with a DAR2 biparatopic ADC. * Indicates the calculated DAR values.

Figure 4

Heterodimerization and c-MET degradation pilot study. (a) Scheme for automated matrix pipetting and heterodimerization followed by high-throughput functional interrogation and hit confirmation. (b) Representative HIC-HPLC chromatograms for 3 out of 624 c-MET biparatopic combinations. Marked peaks refer to bispecifics (Peak 2) and their respective parental antibodies (Peak 1, Peak 3) and indicate a high heterodimer content. Please, see also Figure S3 for a broader data panel. (c) High-throughput screening for c-MET degradation yielded several potent combinations. Representative data indicate the nanomolar c-MET degradation potency (IC50 values: 0.91 nM, A03xB10v5; 0.80 nM, A03xE03; 1.47 nM, REGN5093) of two novel biparatopic antibodies comprising VHH A03 in combination with Fabs B10v5 or E03, similar to that of the re-produced REGN5093 reference, while the anti-DIG IgG isotype control did not degrade the antigen.