Bispecific antibodies (bsAbs) are one of biotech’s most exciting therapeutic classes, promising the ability to bind two different targets simultaneously — unlocking new treatment possibilities in oncology, immunology, and beyond. But despite the buzz, the sobering truth from industry insiders is this: most bispecific programs collapse long before they ever hit GMP manufacturing at a CDMO.
And when they do fail, the CDMO often takes the blame. But according to manufacturing leads and process scientists at multiple contract organizations, the harsh reality is different:
“By the time it gets to us, it’s already too late.”
In other words, the CDMO didn’t fail the molecule. The molecule was set up to fail — months or years earlier — by upstream design, candidate selection, and early development choices that baked in manufacturing liabilities from day one.
This post breaks down the most common failure points, why they happen, and how to design bispecific programs for manufacturability from the very start.
The Harsh Reality: When “Manufacturability” Is an Afterthought
Bispecifics bring unique engineering and production challenges that go beyond standard monoclonal antibody (mAb) manufacturing. Two binding arms, two different antigen specificities, and complex architectures (e.g., knobs-into-holes, CrossMab, DVD-Ig) all mean more ways for the molecule to misbehave.
If manufacturability isn’t evaluated alongside potency and binding during early R&D, programs can advance into cell line development (CLD) and process optimization with hidden flaws that no amount of downstream engineering can fix. By the time the CDMO runs into these issues, timelines are blown, budgets are drained, and the molecule’s fate is sealed.
The Top Four Bispecific Manufacturing Failure Points
Drawing from CDMO feedback, tech transfer case studies, and failed project post-mortems, four failure points account for the majority of bsAb collapses:
1. Mispairing and Aggregation Baked in at Candidate Selection
Bispecific formats rely on complex heavy/light chain pairing, and if that pairing is not well-controlled by design, the result is product heterogeneity and aggregation.
Why it happens:
- Inadequate in silico modeling of chain pairing during sequence design.
- Over-reliance on binding data without evaluating structural stability.
- Using a format unsuited to the targets or desired valency.
Downstream impact:
- High aggregate levels detected in purification and stability studies.
- Low yield of correctly paired product even under optimized process conditions.
- Increased risk of immunogenicity from mispaired species.
2. Cell Lines That Collapse Under Scale-Up
A bispecific may express reasonably well in small-scale CHO culture during CLD, but crumble during large-scale bioreactor runs.
Why it happens:
- Over-engineered expression constructs that place metabolic stress on host cells.
- Lack of stability testing under fed-batch or perfusion conditions.
- Ignoring subtle productivity declines during extended culture.
Downstream impact:
- Dramatic titer drops during GMP scale-up.
- Batch-to-batch inconsistency and failed manufacturing campaigns.
- Long troubleshooting cycles that burn program runway.
3. CDMOs Blamed for Problems They Didn’t Create
When a bsAb fails at GMP scale, the CDMO is often in the firing line — even when they inherited an unstable molecule and a problematic process.
Why it happens:
- Incomplete tech transfer packages from the sponsor.
- Minimal process characterization before transfer.
- Unrealistic sponsor expectations about CDMO “rescue” capabilities.
Downstream impact:
- Strained sponsor–CDMO relationships.
- Delayed program timelines and cost overruns.
- Molecule termination despite potential therapeutic value.
4. Low Titers That No Tech Transfer Can Fix
Some bispecifics simply never produce enough product, even under optimal expression and purification conditions.
Why it happens:
- Poor codon optimization for the host expression system.
- Intrinsically unstable protein domains.
- Designs that require extensive post-translational modifications, slowing production.
Downstream impact:
- COGs (cost of goods) too high for commercial viability.
- Inability to produce enough drug substance for clinical trials.
- Early project kill despite strong biological activity.
The Common Thread: Problems Are Locked in Upstream
Across all four failure points, the pattern is the same: manufacturability flaws are embedded before the CDMO ever sees the molecule. This means traditional late-stage fixes — process optimization, chromatography tweaks, formulation changes — have minimal impact.
By the time a CDMO runs into these problems during GMP manufacturing, the physics and biology of the molecule are working against them.
Manufacturability by Design: The New Imperative for bsAbs
For bispecific programs to survive long enough to make an impact, manufacturability must be engineered in from day one. This requires shifting some of the decision-making and evaluation typically reserved for late-stage CMC into the earliest R&D phases.
Key components of “manufacturability by design” include:
- Sequence-level design tools to predict mispairing risk, aggregation propensity, and expression burden before synthesis.
- In vitro stress testing during candidate selection to identify stability liabilities early.
- Mini-scale process simulations to approximate CLD and DSP conditions before committing to scale-up.
- Cross-functional evaluation between discovery, process development, and manufacturing teams.
Practical Steps to Catch Failures Before They Scale
Here’s how biotech teams can build manufacturability into their bsAb programs:
1. Include CDMO manufacturing input early.
Engage with a manufacturing group — internal or external — at the pre-candidate selection stage. They can flag risky designs based on past production experience.
2. Run forced-degradation studies on early candidates.
Test under pH, temperature, and mechanical stress conditions that mimic process environments. If aggregation or mispairing spikes, the design needs refinement.
3. Model expression and folding in silico.
Use computational tools to assess whether heavy and light chains are likely to pair correctly, and if the final protein will fold stably in CHO or other expression hosts.
4. Treat CLD stability as a gatekeeper.
Monitor productivity, cell health, and product quality over extended culture times at bench scale. Reject cell lines showing instability trends before scale-up.
5. Demand robust tech transfer packages.
When working with a CDMO, ensure the transfer includes complete process descriptions, raw data, and quality results — not just SOPs.
Why This Matters for the Industry
The current bsAb failure rate is not just a sponsor problem or a CDMO problem — it’s an ecosystem problem. As the market for bispecifics grows, CDMOs are under pressure to deliver faster and at higher quality, but they can only work with what they’re given.
If sponsors continue to advance candidates without manufacturability screening, we’ll see the same cycle of:
- Exciting preclinical data →
- Rush into scale-up →
- Process collapse at GMP →
- Costly program termination.
Breaking that cycle means making manufacturability as much a part of candidate selection as potency or target specificity.
Beyond Pairing and Stability: The Hidden Liabilities That Emerge Downstream
For bispecific antibodies, manufacturability challenges don’t end with correct chain pairing and a stable cell line. Even molecules that look promising in expression and early stability assays can run into severe problems once they enter purification, storage, and formulation.
This is where “late-stage liabilities” show up — issues that may not be obvious during candidate selection but can destroy timelines and inflate costs during GMP manufacturing. Common examples include:
- pI heterogeneity – Multiple charge variants complicate ion-exchange chromatography, often requiring extra purification steps or causing yield loss.
- Hydrophobic surface patches – “Sticky” regions on the protein surface can cause non-specific binding to chromatography resins, filters, and even container walls, leading to recovery losses and batch variability.
- Aggregation under stress – A molecule may be stable in short-term storage but aggregate during freeze–thaw cycles or under long-term refrigerated conditions, affecting potency and shelf life.
- Formulation incompatibility – Certain bsAb formats have solubility limits or excipient incompatibilities that emerge only during high-concentration formulation, needed for subcutaneous delivery or small-volume IV dosing.
The cost of discovering these problems late is measured in lost months and millions: reformulating, re-validating purification steps, or redesigning the molecule entirely. That’s why developability assessments — evaluating a candidate’s biophysical and chemical stability under realistic purification and formulation conditions — should be part of early-stage bispecific design.
Molecules that clear only the “pairing and expression” hurdle are still at risk. To truly survive the journey to the clinic, they must also be engineered to withstand the real-world stresses of downstream manufacturing and distribution.
Conclusion: Build It to Make It!!
The bispecific pipeline is rich with therapeutic promise, but promise alone doesn’t make a drug. Molecules must not only work in vitro and in vivo — they must be buildable at scale, reproducibly, and within a viable cost structure. And they must survive the entire journey through purification, storage, and formulation without falling apart.
When manufacturability is ignored until late development, the result is almost always the same: expensive, avoidable failure. The same is true when downstream developability is left unchecked — hidden liabilities like pI heterogeneity, hydrophobic patches, or freeze–thaw instability will surface late, costing months and millions to fix.
Applying “manufacturability by design” principles from the very start — from sequence design to early cell line work, and from purification resilience to formulation stability — dramatically improves the odds that a bispecific program will make it to the clinic and beyond.
The CDMO didn’t fail the molecule. The molecule was set up to fail — sometimes upstream, sometimes downstream.
Let’s stop doing that!!
Top 5 FAQ: Bispecific Antibody Manufacturing Failures
1. Why do so many bispecific antibody (bsAb) programs fail before GMP manufacturing?
Most failures stem from upstream issues such as chain mispairing, aggregation propensity, unstable cell lines, and inherently low titers. These problems are often locked in during candidate design and early development, making them extremely difficult to fix once the molecule reaches a CDMO.
2. What is mispairing and why is it so damaging to bsAb production?
Mispairing occurs when heavy and light chains assemble incorrectly, creating unwanted species. This reduces the proportion of correctly formed bsAb, increases aggregates, and can cause immunogenicity risks — all of which hurt yield, quality, and regulatory acceptance.
3. How can cell line instability derail a bispecific program?
Even if small-scale expression looks promising, some bsAb-producing cell lines collapse at large scale due to metabolic stress or genetic instability. This can lead to dramatic titer drops and inconsistent quality during GMP manufacturing.
4. Can a CDMO “rescue” a bsAb with poor manufacturability?
Generally, no. If the molecule’s issues are inherent to its sequence or format, process optimization has limited impact. CDMOs can improve yields and robustness, but they cannot fix fundamental stability or expression problems.
5. How can manufacturability be built into a bispecific program from day one?
Sponsors should integrate manufacturability assessments into early R&D — including in silico mispairing prediction, forced-degradation testing, early stability studies, and pilot-scale expression trials. Cross-functional review with manufacturing experts before candidate selection is critical.
