Developing anti-drug antibody (ADA) assays for therapeutic peptides can be complex, particularly when the molecule is small, endogenous-like, and when no positive control exists. This article describes the stepwise development of a screening and confirmatory ADA assay for a 47-amino acid peptide targeting a receptor involved in inflammation in a cardiovascular indication, highlighting the challenges encountered and the solutions to overcome them.

Project overview

The sponsor’s request was to develop screening and confirmatory ADA assay methods in human serum. The drug candidate is a 47-amino acid peptide, a fragment of an endogenous protein involved in inflammation in a cardiovascular indication. The drug target is a cell receptor, and the assays were to be implemented to test clinical samples from a Phase 1 study in healthy volunteers after repeated ascending intravenous dose. The method was developed from scratch, meaning no positive control was identified beforehand.

The overall strategy involved

  • Selecting and testing a positive control
  • Producing reagents suitable for the peptide
  • Implementing classical and alternative assay formats
  • Determining screening and confirmatory cut points

Positive control selection

To select a positive control, several commercial antibodies were tested by direct ELISA. Two out of three antibodies produced a signal proportional to drug concentration. The monoclonal antibody was selected because it showed better sensitivity.

Reagent production for a small peptide

Because the drug is a small peptide, conjugation can risk masking recognition sites and preventing antibody binding. Two conjugation strategies were implemented

  • To couple the drug to sulfotag, which is a relatively large molecule, an aminohexanoic acid linker was added to the C-terminal end of the peptide
  • To couple the drug to biotin, which is smaller, conjugation was performed through a C-terminal cysteine

Bridging assay development and troubleshooting

With the positive control and reagents available, a bridging format was developed using the MSD platform. Streptavidin plates were used to coat the biotin drug, and the first configuration tested was a homogeneous format, based on pre-incubation with a master mix containing drug biotin and drug sulfotag.

In this homogeneous format, signals were very low. The maximum signal to noise ratio was below 15, and for the positive control at 100 ng per mL the signal was around one. Troubleshooting was therefore initiated.

Troubleshooting included:

  • Coating plates directly with the positive control and testing detection with conjugated peptides (successful).
  • Adjusting buffer, incubation, and adding protease inhibitors (no significant improvement).
  • Switching to sequential format with a biotinylated spacer to reduce steric hindrance (slight improvement).

Detection Optimization

Although the spacer slightly improved signal, it was not sufficient to justify the increased risk of unspecific interactions. Detection strategies were refined:

  • Positive control detected with anti-rabbit IgG sulfotag.
  • Human antibodies (IgG and IgM) initially tested with protein A/G, which was insufficient.
  • A mixture of anti-species antibodies ultimately provided sufficient sensitivity.

Final assay format

  • The positive control was detected using anti-rabbit IgG sulfotag
  • Human IgG and IgM control antibodies were detected using anti-human IgG and IgM sulfotag
  • In clinical samples, human ADA responses are expected to be detected using anti-human IgG and IgM sulfotag

Screening cut point estimation

Individual human serum samples were collected and screened. Initial testing confirmed:

  • IgG and IgM were recognized without cross-reactivity.
  • High variability in some samples (especially from SST2 advanced tubes in contrast to dry tubes).
  • Because this run was performed by a different operator, interoperator variability was also assessed and confirmed.

After exclusion of analytical and biological outliers, cut points were established:

Confirmatory assay development

Twelve individual serum samples were tested under excess drug conditions:

  • Negative serum showed <50% inhibition.
  • Excess drug (10 μg/mL) achieved >90% inhibition at 16 ng/mL positive control.
  • A hook effect was observed at 1 μg/mL.

The confirmatory assay was thus finalized and ready for clinical implementation.

Pre-screening and preliminary controls

Pre-screening of the 50 individuals collected for validation identified eight analytical outliers based on a high coefficient of variation between replicates and six biological outliers due to extreme responses.

  • Human IgG control at 100 ng per mL produced a signal of approximately 2000
  • Human IgM control at 100 ng per mL produced a signal of approximately 4000
  • Negative control signal was approximately 600
  • Estimated cut point factor was 1.7
  • Low positive control was set at 10 ng per mL

Key Challenges and Lessons

  • Small peptide drug: Conjugation required careful linker design to avoid masking epitopes.
  • Endogenous-like peptide: Increased risk of non-specific interactions demanded an adaptable assay format.
  • Mixed detection antibodies: While necessary for sensitivity, this approach required additional controls to mitigate background signals, potentially addressable via humanization of the positive control.

Conclusion

Thanks to our expertise, iterative optimization, and close communication with the sponsor, we were able to develop a robust ADA assay that addressed the challenges of small peptide conjugation, endogenous-like interference, and mixed detection antibody requirements, ensuring readiness for clinical sample testing.