Immunogenicity Testing for Biologics

Every therapeutic protein carries a quiet question: will the patient’s immune system decide it is foreign? When it does, the result is anti-drug antibodies (ADAs). Some are harmless. Others clear the drug faster, block its activity or, in rare cases, attack a protein the body cannot live without. That is why regulators expect a planned, risk-based immunogenicity program for every biologic, and why the assays behind it deserve as much care as any potency or purity method.

Why Anti-Drug Antibodies Matter

FDA’s 2014 guidance, Immunogenicity Assessment for Therapeutic Protein Products, frames the stakes clearly (FDA, 2014). Neutralizing antibodies can block the active site of a therapeutic enzyme or protein and reduce efficacy. Non-neutralizing antibodies can still change clearance and exposure. Anaphylaxis is a separate concern, and the guidance notes that ADA status alone does not predict it. The most serious outcome is cross-reactivity: antibodies that neutralize both the drug and a non-redundant endogenous protein.

The best-known example is epoetin-associated pure red cell aplasia (PRCA). Cases linked to Eprex (epoetin alfa) were first identified in 1998 and 1999, and European case counts peaked in 2002. A RADAR review reported 191 epoetin-associated PRCA patients between 1999 and 2004, about 95% of whom had received subcutaneous Eprex (McKoy et al., Transfusion, 2008). Patients developed antibodies that neutralized their own erythropoietin as well as the drug.

The Regulatory Framework

Three documents anchor most programs. FDA’s August 2014 guidance sets out the risk-based approach to assessing and mitigating immune responses (FDA, 2014). FDA’s January 2019 guidance, Immunogenicity Testing of Therapeutic Protein Products: Developing and Validating Assays for Anti-Drug Antibody Detection, covers assay design and validation (FDA, 2019). In Europe, the EMA guideline on immunogenicity assessment of biotechnology-derived therapeutic proteins (EMEA/CHMP/BMWP/14327/2006 Rev 1) took effect on 1 December 2017. It adds detail on assays and asks for a multidisciplinary immunogenicity summary, including a risk assessment, in marketing authorization applications (EMA, 2017).

The Tiered Testing Approach

FDA recommends a multi-tiered approach (FDA, 2019). A sensitive screening assay runs first. Positive samples move to a confirmatory assay, then to titration and neutralization assays that characterize the response.

• Screening. A binding assay tuned for sensitivity. A low but defined false-positive rate of approximately 5% is desirable, so the cut point errs toward catching true positives.
• Confirmatory. Usually a competitive inhibition step with excess drug to show specificity. FDA considers a 1% false-positive rate adequate for this cut point.
• Titer. The reciprocal of the highest dilution that reads at or just above the cut point. Titer helps link magnitude of response to PK, safety and efficacy.
• Neutralizing antibody (NAb). FDA generally recommends a cell-based bioassay. A competitive ligand binding format may suit some products, but FDA asks sponsors to discuss alternatives with the Agency first.

Cut points are set statistically from drug-naive samples from the relevant patient population, not from healthy volunteers alone where disease state changes the matrix. Getting this wrong is costly. A cut point that is too high hides true positives; one that is too low floods the confirmatory tier.

“An ADA assay that cannot see antibodies in the presence of drug is not a safety tool. It is a false reassurance.”

Sensitivity, Drug Tolerance and Sampling

FDA recommends that screening and confirmatory IgG and IgM ADA assays achieve a sensitivity of at least 100 ng/mL, with a higher limit possibly acceptable depending on risk and prior knowledge (FDA, 2019). Earlier practice often accepted 250 to 500 ng/mL, but data suggesting clinical events at concentrations as low as 100 ng/mL drove the tighter target (Center for Biosimilars, 2019).

Sensitivity on paper means little if circulating drug masks the antibodies. FDA calls the assessment of sensitivity in the presence of expected drug levels, the assay’s drug tolerance, critical to judging whether a method is fit for purpose. Practical levers include acid dissociation, sample dilution and collecting samples at trough. The guidance also notes that samples taken 3 to 6 weeks after first exposure are generally optimal for IgG responses, with a sample about 30 days after the last dose in multiple-dose studies.

Positive controls matter too. FDA recommends affinity-purifying animal-derived positive control antibodies against the therapeutic protein. Remember that a surrogate control cannot fully mimic a polyclonal human response, so sensitivity values are estimates, not absolutes.

Risk-Based Assessment: Product and Patient Factors

The 2014 guidance asks sponsors to weigh factors on both sides (FDA, 2014). The risk assessment then sets how intensive sampling, assay development and clinical monitoring need to be.

Factor type	Examples from FDA guidance
Patient	Immune status, prior sensitization, HLA genotype, tolerance to the endogenous protein
Treatment	Route (subcutaneous is more immunogenic than IV), dose, intermittent dosing
Product structure	Non-human sequences, novel formats, glycosylation, PEGylation, deamidation, oxidation
Process and impurities	Aggregates, host cell proteins, endotoxin, nucleic acids
Formulation and container	Excipient degradation, leachables, silicone oil, prefilled syringes

Many of these factors are set long before the clinic. Host cell choice, glycosylation profile and aggregation propensity trace back to cell line development and process design, so immunogenicity risk is a CMC topic as much as a clinical one.

In Silico and In Vitro Prediction

Early in development, in silico tools screen sequences for peptides likely to bind HLA class II. In vitro tools follow: PBMC or dendritic cell–T cell assays, MHC-associated peptide proteomics (MAPPs) and dendritic cell maturation assays. A 2021 European Immunogenicity Platform review warns that affinity-only algorithms over-predict and that applying these methods blindly to predict immunogenicity is not recommended (Ducret et al., mAbs, 2021). It suggests 10 to 25 donors early in development and about 50 later for T-cell assays. Use these tools to rank candidates and guide de-immunization, not to replace clinical data.

Biosimilars and Interchangeability

For biosimilars, immunogenicity is a comparative question: is the ADA profile meaningfully different from the reference product? FDA’s October 2025 draft guidance says that when comparative analytics support high similarity, a well-designed human PK similarity study and an immunogenicity assessment may be enough, without a comparative efficacy study (FDA, 2025). Separately, FDA’s June 2024 draft update on interchangeability removed the default expectation for switching studies, citing experience that switching poses minimal added risk (Center for Biosimilars, 2024). Fewer clinical studies put more weight on assay quality. When immunogenicity data carry the clinical comparison, a weak ADA method weakens the whole package.

What This Means for Biologics Developers

• Start the risk assessment early. Document product and patient factors before first-in-human and update them as data arrive.
• Build assays for the matrix you will test. Set cut points with disease-state samples and prove drug tolerance at expected trough levels.
• Plan the NAb assay in parallel. Cell-based formats take time to develop and validate.
• Link immunogenicity to CMC changes. Formulation, container closure and process changes can shift risk, as Eprex showed.
• Integrate the data. Regulators want ADA results tied to PK, safety and efficacy, not reported in isolation.

Terraforme Biosciences helps sponsors build immunogenicity strategies, review ADA assay validation packages and prepare the integrated summaries regulators expect. Learn more about our regulatory guidance services.

Conclusion

Immunogenicity testing is not a box to tick after the pivotal trial. It is a risk-based program that starts with molecule design and runs through life cycle changes. Sound cut points, real drug tolerance and a well-chosen neutralizing assay turn ADA data into evidence that protects patients and supports approval.