With the biotherapeutic market gearing towards the production of a new era of modalities, novel challenges await Chinese hamster ovary (CHO)-based process development. Multi-specific antibodies present complications due to the requirement of efficient multi-chain expression. This poses challenges for the expression system, cell line selection strategy, analytical characterisation and downstream purification. High levels of productivity are often driven by optimised vector designs that achieve delicately balanced chain expression. Additionally, consideration of product purity earlier in the upstream process is paramount since product related impurities are common with these complex proteins. In this study, we investigated a holistic approach to bespoke process development for bispecific antibodies (BsAbs). Early in the ApolloX™ cell line development process, we evaluated the use of a vector and transfection design toolbox to rapidly evaluate various multi-gene vector constructs and identify optimal DNA amounts/selection pressure stringency for balanced chain expression. Selected pools were taken forward into single cell clone generation based on their growth and heterodimer production performance. Screening stages in cell line development were complemented by advanced analytics such as high-throughput product purity analysis and mass spectrometric evaluation of product related impurities. Targeted downstream purification development approaches based on molecule bioinformatics were also adopted. Application of this toolbox-style approach boosted target BsAb titres for a 4-chain BsAb by 3-fold in transfectant pools. Additionally, average heterodimer purity of over 80% was achieved across 3 different model BsAb formats. By adopting a testing toolbox strategy catered to the variability of multi-specific antibody structures, we were able to successfully navigate the complexity of BsAb production.

Effective T-cell activation is driven by three signals: Signal 1 delivered through antigen recognition via the T-cell receptor, signal 2 through co-stimulatory receptors, and signal 3 mediated by cytokines. CD3-targeting T-cell engagers provide signal 1 to T cells and have shown significant clinical benefit in hematological malignancies but have faced challenges in solid tumors due to on target off tumor toxicities. Emerging clinical data supports the hypothesis that for efficient and sustained activity in the presence of signal 1, engagement of co-stimulatory molecules like CD28 could be important for effective anti-tumor activity in solid tumors. We present the development of a panel of bispecific antibodies (bsAbs), targeting the costimulatory molecule CD28 and the tumor associated antigen (TAA) Nectin-4, a cell adhesion molecule overexpressed in bladder cancer and other malignancies. To select our development candidate RNDO-564, we screened a panel of bsAbs with varying CD28-potencies for both functional activity and biophysical stability to assess and reduce manufacturability risks. Functional characterization of our bsAbs showed robust tumor cytotoxicity and IL-2 secretion while posing a lower safety risk providing that the activity is dependent on presence of signal 1 and expression of the TAA. Additionally, we demonstrated that the panel of CD28 binders can be paired with various TAA-binding arms, enabling the design of new bispecific antibodies for different indications. In addition to the desired biological activity, the CD28 x Nectin-4 bsAbs also exhibited favorable developability profiles as demonstrated by subjecting the panel to accelerated stress conditions like thermal stress at elevated temperatures for extended periods of time or low pH hold. This extensive functional and biophysical characterization enabled the selection and successful manufacture of our clinical candidate RNDO-564.

DNA, RNA, and proteins are synthesized using template molecules, but glycosylation is not believed to be constrained by a template. However, if cellular environment is the only determinant of glycosylation, all sites should receive the same glycans on average. This template-free assertion is inconsistent with observations of microheterogeneity—wherein each site receives distinct and reproducible glycan structures. Here, we test the assumption of template-free glycan biosynthesis. Through structural analysis of site-specific glycosylation data, we find protein-sequence and structural features that predict specific glycan features. To quantify these relationships, we present a new amino acid substitution matrix that describes “glycoimpact” -- how glycosylation varies with protein structure. High-glycoimpact amino acids co-evolve with glycosites, and glycoimpact is high when estimates of amino acid conservation and variant pathogenicity diverge. We report hundreds of disease variants near glycosites with high-glycoimpact, including several with known links to aberrant glycosylation (e.g., Oculocutaneous Albinism, Jakob-Creutzfeldt disease, Gerstmann-Straussler-Scheinker, and Gaucher’s Disease). Finally, we validate glycoimpact quantification by studying oligomannose-complex glycan ratios on HIV ENV, differential sialylation on IgG3 Fc, differential glycosylation on SARS-CoV-2 Spike, and fucose-modulated function of a tuberculosis monoclonal antibody. In all, we show glycan biosynthesis is accurately guided by specific, genetically-encoded rules, and this presents a plausible refutation to the assumption of template-free glycosylation.