(T1330-01-01) Identification of Vendor-Specific Interactions in Antibody HRP Conjugates that Impact Bioanalytical Assay Performance

Purpose: Consistent, well-performing reagents are critical to ensuring Bioanalytical PK/ADA assay performance in GLP/GCP clinical and non-clinical sample testing.  In an ADA bridging assay, an interaction between drug and the horseradish peroxidase (HRP) labeled anti-digoxigenin (aDIG) antibody led to increased HRP signal in an ELISA, causing a high and variable background signal and inconsistent assay performance.  This work evaluates several commercial sources of aDIG-HRP and identifies aspects of commercially sourced aDIG-HRP quality that impact ADA assay performance.

Methods: A conventional double antigen-bridging format was used to detect ADA in a screening ELISA.  Specifically, a master mix composed of biotinylated and DIG-labeled therapeutic mAb was incubated with sample or positive control in human serum.  After overnight incubation, the mixture was transferred into a streptavidin-coated plate, incubated, and washed.  Then aDIG-HRP was added to each well, incubated, washed, and 3,3′,5,5′-Tetramethylbenzidine (TMB) was added.  After 20 minutes, 1 M phosphoric acid was added to stop the reaction, and optical density (OD) was read at 450 nm in a plate reader. A similar ELISA-based binding assay was also performed to directly test the interaction between the biotinylated drug and aDIG-HRP.  Here, biotin-drug at various concentrations was captured on a streptavidin-coated plate.  After washing, aDIG-HRP or HRP was added to the wells and further incubated.  Following another wash step TMB was added.  After 20 minutes, 1 M phosphoric acid was added to stop the reaction, and OD was read at 450 nm in a plate reader. HRP conjugations were performed in-house as described by the vendor's protocol.  These are based on the Nakane method in which HRP glycans are activated by periodate to create aldehydes which are then reacted with lysines on the protein being labeled. Separation and fractionation of the aDIG-HRP conjugates was performed using size exclusion chromatography (SEC)  on a Thermo Vanquish UHPLC.  Signal was monitored at 280 nm.

Results: While evaluating alternative reagents to decrease background noise and variability troubleshoot  a bridging-ELISA ADA screening assay,  we observed a potential interaction between aDIG-HRP and the biotinylated drug.  This interaction would cause an increased background OD, explaining the assay’s background variability.  A simplified binding ELISA that measures aDIG-HRP activity on material bound to biotinylated drug captured on a strepavidin plate confirms this finding.  In addition, Biacore analysis also showed binding between aDIG-HRP and both biotinylated and non-biotinylated drug. Binding ELISA experiments were also performed on several other commercial sources of aDIG-HRP.  The interactions of aDIG-HRP and the drug appear to be common for three of the commercial sources; one commercial source (“Vendor B”) does not show evidence of drug binding.  A comparison across these commercial sources by SDS-PAGE shows heterogeneous samples with a wide size distribution across all commercial sources tested, including Vendor B.  Furthermore, some SDS-PAGE bands correspond to high molecular weight (MW) species by SDS-PAGE ( > 250 kDa MW marker). HRP was conjugated in-house to the aDIG mAb from Vendor B; this material showed interactions with the drug.  Because the differences between the in-house conjugated aDIG-HRP and vendor-conjugated aDIG-HRP are either the conjugation and purification process or the HRP source, this links the drug binding to the HRP conjugation process.  Similar results were observed using two sources of HRP.  Additionally, a “mock” HRP conjugation in the absence of any mAb also shows interaction with the drug after filtering of the HRP monomer, linking interaction to any potential HRP-HRP conjugations that may occur in the reaction. Comparing vendors with the highest drug binding aDIG-HRP (Vendor A) and Vendor B aDIG-HRP shows differences in the size distribution by SEC; Vendor A aDIG-HRP shows larger-sized species present compared to Vendor B.  Fractionation of Vendor A aDIG-HRP locates the highest binding activity to the largest species fractionated by SEC, linking drug binding to the largest-sized species in the heterogeneous mixture.

Conclusion: HRP conjugation through the glycans potentially activates multiple glycans on HRP, which can react with lysines on either an antibody of interest or another molecule of HRP.  This may lead to formation of extended antibody-HRP assemblies; we call these oligomeric-HRP.  These species are common among several commercial sources of aDIG-HRP, and the largest Oligomeric-HRPs are linked to drug binding and increased background signal in an ADA assay leading to variable assay performance.  Given that one aDIG-HRP source (Vendor B) does not show drug binding and appears to have smaller species by SEC, one can conclude that appropriate conjugation conditions and / or reagent purification can mitigate these effects.  Altogether, this work shows that commercially-sourced reagents show marked differences across vendors, and careful reagent evaluation and characterization are paramount in ensuring consistent assay performance.

References: Nakane, P. K. (1974). Peroxidase-Labeled Antibody A New Method of Conjugation. J. Histochemistry and Cytochemistry., 22(12), 1084-91.

Acknowledgements: All authors are employees of Genentech Inc., a member of the Roche Group.


The buffer signal is observed in several aDIG-HRP vendors in a simplified binding ELISA. Comparison of aDIG-HRP signal in the binding ELISA. Across multiple vendors (VA, VC, and VD) and multiple species, drug 1 shows concentration dependent signal consistent with an interaction with aDIG-HRP, while one vendor (VB) shows low OD signal close to the blank.


Table 1: A drug 1 binding ELISA links interaction signal to aDIG mAb and HRP conjugation process and/or cleanup. Commercially available aDIG-HRP from Vendor B shows low signal in buffer and serum over a drug 1 surface, while aDIG-HRP using Vendor B’s aDIG mAb prepared in house shows high drug 1 binding signal in both buffer and serum.



The fractionation of Vendor A aDIG-HRP links aDIG-HRP interaction with drug 1 to high-MW species eluting around 4-5 minutes by SEC. (A) Representative SEC chromatogram of Vendor A, lot 6 (VA-6) with regions of fractions collected denoted by vertical black lines. (B) The binding ELISA shows high HRP binding activity (green) of Fraction 1 (F1) bound to drug 1 compared to fractions 2-4 (F2, F3, F4). HRP activity observed in bridging ADA ELISA for both the low positive control (LPC) (blue), as well as a negative control in the absence of serum (buffer) (red); with all four fractions low separation is observed between the NC normalized LPC and buffer signals. For comparison, results for bulk (non-fractionated) Vendor B, lot 1 (VB-1) and Vendor A, lot 1 (VA-1) are shown. Signals are normalized by dividing them by the NC OD for the bridging ADA assay.