The path from antibody discovery to approved therapeutic is costly and marked by high attrition. Many promising candidates fail not due to lack of efficacy, but because they cannot be manufactured at scale.
Developability failures—aggregation, poor stability, low yields, unfavorable viscosity—typically surface only in late-stage development, causing enormous financial losses and delays. Shifting developability assessment earlier is a strategic opportunity to de-risk programs and accelerate success.
The Cost of Late-Stage Failures
Advancing a single candidate through Phase III can cost $1–2 billion1, and 50–70% of antibodies entering preclinical development never reach market—many due to manufacturability challenges rather than efficacy or safety issues2. A single late-stage developability failure can represent losses exceeding $500 million, not counting years of lost development time and delayed patient access.
The Case for Early Assessment
Integrating developability evaluation into initial screening, when dozens or hundreds of candidates are still under consideration, offers clear advantages:
- Risk mitigation: Flag problematic candidates before major investment
- Cost savings: Early elimination can save tens to hundreds of millions per program
- Faster timelines: Selecting developable candidates upfront can shorten time-to-clinic by 6–12 months
- Broader screening: Evaluating more candidates in parallel increases the likelihood of finding molecules that combine efficacy and manufacturability
Key Developability Parameters
A robust early assessment covers multiple interconnected properties that together predict how a candidate will behave during manufacturing, formulation, and storage:
- Thermal stability: Low melting temperature indicates susceptibility to unfolding, aggregation, and degradation, which is one of the earliest and most informative filters
- Aggregation propensity: A leading cause of late-stage failure, aggregation compromises yield, product quality, and carries immunogenicity risk
- Expression yield and product quality: Low yields or high levels of misfolded variants signal manufacturing challenges that are costly to resolve later
- Chemical stability: Sequence liabilities such as deamidation and oxidation create charge variants that affect stability and potency
- Hydrophobicity and non-specific binding: Surface hydrophobic patches drive aggregation, increase viscosity, and promote unwanted tissue interactions
- PTMs and Glycosylation profile: Directly influences effector functions, half-life, and aggregation and is highly host-dependent, making representative expression essential
- Viscosity: High-concentration formulations required for subcutaneous delivery are particularly sensitive; problematic viscosity creates manufacturing, fill-finish, and patient tolerability challenges
The Importance of Consistent, High-Quality Data
Developability parameters are only as informative as the data underpinning them. Inconsistent or poor-quality data can lead to misguided go/no-go decisions—advancing flawed candidates or discarding viable ones.
evitria supports this by providing high purity material and critical analytical data for every construct as well as comparable manufacturing across the whole pipeline from small-scale initial screening all the way up to gram scale production.
The Advantage of Starting in CHO
Developability data from non-CHO systems may not predict commercial behavior and leads to costly surprises at the host-switch stage. The expression system used to generate early material significantly impacts data reliability. Antibodies produced in HEK293 or cell-free systems often behave differently than those from Chinese Hamster Ovary (CHO) cells, which dominate commercial manufacturing. These biological discrepancies create the following development risks:
- Glycosylation: CHO cells generate core-fucosylated, complex-type glycans that differ from HEK293 and affect pharmacokinetics3.
- Post-translational modifications (PTMs): Variations in lysine clipping and glycation alter charge heterogeneity and stability.
- Folding and aggregation: Successful expression in HEK293 does not predict the absence of folding challenges or aggregation in a CHO host.
Since CHO cells account for ~70% of approved recombinant protein therapeutics, using CHO-derived material from the outset provides:
- Directly predictive developability data translatable to manufacturing
- Elimination of host-switch variability and repeat characterization
- Glycan profiles representative of the final drug product
- A faster path to GMP readiness without host-dependent caveats
Enabling Early CHO-Based Screening
Historically, CHO was considered too slow for discovery-stage work. Modern approaches have changed this, delivering milligram quantities of purified antibody within weeks—fully compatible with discovery timelines.
This is where evitria’s High-Throughput CHO Expression Platform makes a decisive difference. With the capacity to process dozens to hundreds of variants in parallel in small-scale formats, evitria enables comprehensive screening across large candidate panels without sacrificing manufacturing relevance.
Modern analytical techniques require only microgram to low milligram quantities of material, meaning small-scale CHO productions of just a few mg per variant are more than sufficient for robust profiling of thermal stability, aggregation, charge heterogeneity, and hydrophobicity. Sequence-based computational predictions4 further enrich the assessment before committing to larger-scale production.
By partnering with evitria, discovery teams gain access to high-quality, manufacturing-relevant CHO material without heavy internal infrastructure investment—enabling confident go/no-go decisions that reliably translate to later development stages.
Conclusion
Much of late-stage antibody attrition is avoidable. Early developability assessment using CHO-derived material identifies liabilities before major investments are made. With modern analytics, computational tools, and evitria’s High-Throughput CHO Expression Capabilities, the technical barriers have been largely overcome.
Frequently Asked Questions
EDA is the pre-clinical screening of antibody candidates for physical and chemical stability, expression yield, and solubility—alongside binding affinity—to identify molecules suitable for large-scale manufacturing, formulation, and storage. By evaluating these properties early, teams can deprioritize high-risk candidates before significant resources are committed.
HEK293 and CHO cells differ substantially in post-translational modifications—particularly glycosylation patterns—as well as folding efficiency and product quality. A candidate that expresses well and appears stable in HEK293 may aggregate, under-perform, or require extensive re-optimization when transferred to CHO, the industry-standard manufacturing host. Starting developability screening in CHO eliminates this uncertainty.
Thermal stability and aggregation propensity are the most critical early indicators. These are followed by charge heterogeneity, hydrophobicity, non-specific binding, viscosity, and expression titer. Together, these parameters provide a comprehensive picture of how a candidate is likely to behave during manufacturing, formulation, and long-term storage.
Failures discovered during clinical manufacturing or late-stage trials can cost upwards of $500 million per program. Identifying the same liabilities during pre-clinical screening—when alternatives are still available and pivot costs are minimal—can save tens to hundreds of millions of dollars per program and shorten time-to-clinic by up to 12 months.
Computational design and machine learning models generate large numbers of candidate sequences, but predictions must be validated with real experimental data. High-throughput CHO Expression Platforms, such as evitria’s, can produce and profile dozens to hundreds of variants in parallel, providing standardized, manufacturing-relevant ground truth data that refines and improves predictive models—ensuring that in silico designed antibodies translate reliably into physical, manufacturable molecules.
Modern analytical techniques such as DSF, DLS, and SEC require as little as 50–500 µg per variant, meaning small-scale CHO productions are fully sufficient for robust early profiling. Partnering with a specialized provider like evitria makes sense as soon as candidate panels are ready for screening—gaining access to High-Throughput, manufacturing-relevant CHO material without the cost and time of building internal infrastructure, and generating data that confidently translates to later development stages.
Ideally, developability screening should be integrated into the initial candidate selection phase, while multiple hits or leads are still under consideration. At this stage, High-Throughput Platforms like evitria’s CHO system allow dozens of variants to be profiled in parallel with minimal material, enabling data-driven go/no-go decisions before significant resources are committed.
Yes. Modern transient transfection approaches in CHO cells can deliver purified antibody within a few weeks—fully compatible with discovery timelines. evitria’s High-Throughput Platform is specifically designed to handle large numbers of variants in parallel, making early CHO-based screening a practical and increasingly standard approach.
References
- DiMasi, J. A., Grabowski, H. G. & Hansen, R. W. (2016). Innovation in the pharmaceutical industry: New estimates of R&D costs. Journal Of Health Economics, 47, 20–33. https://doi.org/10.1016/j.jhealeco.2016.01.012 ↩︎
- Jain, T., Sun, T., Durand, S., Hall, A., Houston, N. R., Nett, J. H., Sharkey, B., Bobrowicz, B., Caffry, I., Yu, Y., Cao, Y., Lynaugh, H., Brown, M., Baruah, H., Gray, L. T., Krauland, E. M., Xu, Y., Vásquez, M. & Wittrup, K. D. (2017). Biophysical properties of the clinical-stage antibody landscape. Proceedings Of The National Academy Of Sciences, 114(5), 944–949. https://doi.org/10.1073/pnas.1616408114 ↩︎
- Croset, A., Delafosse, L., Gaudry, J., Arod, C., Glez, L., Losberger, C., Begue, D., Krstanovic, A., Robert, F., Vilbois, F., Chevalet, L. & Antonsson, B. (2012). Differences in the glycosylation of recombinant proteins expressed in HEK and CHO cells. Journal Of Biotechnology, 161(3), 336–348. https://doi.org/10.1016/j.jbiotec.2012.06.038 ↩︎
- Raybould, M. I. J., Marks, C., Krawczyk, K., Taddese, B., Nowak, J., Lewis, A. P., Bujotzek, A., Shi, J. & Deane, C. M. (2019). Five computational developability guidelines for therapeutic antibody profiling. Proceedings Of The National Academy Of Sciences, 116(10), 4025–4030. https://doi.org/10.1073/pnas.1810576116 ↩︎
