Purpose: Eliminating drug interference remains a significant obstacle to immunogenicity assessment for studies involving a therapeutic protein product with a multiple dosing regimen or long drug half-life. Many approaches to reduce drug interference have been developed including disruption of ADA-drug complex through acid dissociation, or isolation of the ADA through ACE, SPEAD, or PandA. Each approach is associated with individual strengths and limitations related to degree of drug tolerance, degree of imprecision, differential sensitivity and drug tolerance across testing tiers, persistence of drug target interference, and drug and drug target interference that increasingly impacts detection of lower ADA concentrations in samples. Of these, only a less-common approach, PandA (PEG Precipitation and Acid dissociation) claims complete elimination of drug interference across the full range of sensitivity. However, with the novelty of the format, concerns of imprecision, technique-dependent wash steps, and other unique PandA-specific optimization considerations, the industry has been slow to adopt this format despite strong claims of interference-related advantages. Two significant considerations, insufficient optimization of assay drug concentration and PEG percentage in the procedure, have shown the potential concern of limiting ADA recovery and even differential ADA recovery within subjects over time. Here we characterize the unique conditions contributing to this phenomenon. We also introduce a PandA optimization strategy to avoid these issues, resulting in ADA detection largely unaffected by the amount of drug contained in a sample.
Methods: The PandA method, as outlined by Zoghbi et al, is dependent on four principles: First, the sample is saturated with excess (assay) drug to bind all ADA into drug-antibody complexes. Second, the total ADA is precipitated with the use of PEG and any non-precipitated interferences are washed from the sample. Next, the sample is acidified to disrupt the drug-antibody complexes and coated to a bare high-bind MSD plate under acidic conditions. Lastly, the ADA are specifically detected with Sulfo-TAG conjugated drug.
Sample combinations were prepared at various levels of positive control anti-drug antibody in the presence (+drug) and absence(-drug) of drug. To optimize the concentration of excess assay drug needed to bind all ADA into immune complexes for precipitation, the method MRD was performed in buffer containing 0, 25, 50, and 100µg/mL of drug. Subsequently, to optimize the PEG percentage needed for precipitation, the samples were incubated overnight with 3%, 4.5%, and 6% PEG solutions in the presence and absence of excess drug.
Results: With sub-optimal assay drug conditions (< 100µg/mL), the presence of up to 500µg/mL of drug in +Drug samples increases recovery of 1000ng/mL ADA up to 403% in comparison to -Drug samples. With sub-optimal assay drug conditions (< 100µg/mL), the presence of up to 500µg/mL of drug in +Drug samples increases recovery of 100ng/mL ADA up to 200% in comparison to -Drug samples. As assay drug conditions increase from 0 to 100µg/mL the recovery of 1000ng/mL and 100ng/mL ADA in -Drug samples increase 391% and 148% respectively. As assay drug conditions increase from 0 to 100µg/mL the percent difference of recovery between 500µg/mL +Drug samples and -Drug samples with 1000ng/mL and 100ng/mL ADA decrease from 403% to 130% and 200% to 110% respectively. With optimal assay drug conditions (100µg/mL), increase of PEG percentage from 3% to 6% increases recovery of 1000ng/mL and 100ng/mL ADA an additional 161% and 144% respectively. With optimal assay drug conditions (100µg/mL), increase of PEG percentage to 6% further reduces the difference of recovery between 500µg/mL +Drug samples and -Drug samples with 1000ng/mL and 100ng/mL ADA to 96% and 93% respectively.
Conclusion: Careful optimization of both assay drug concentration and PEG percentage are necessary for complete ADA recovery and sensitivity. Sub-optimal conditions could contribute to perceptions of variable method performance during method development and hesitancy to use the highly drug tolerant PandA format. Sub-optimal assay drug concentrations result in reduced ADA-drug complex formation and, subsequently, reduced ADA detection. Sub-optimal PEG percentages result in reduced precipitation of ADA-drug complexes and, consequently, also reduced ADA detection. These results indicate that sub-optimal assay drug concentrations can result in variable recovery depending on sample drug concentration. This could present as variable ADA sensitivity and titer across timepoints according to drug concentration increase or decrease within a given subject over time. For the same reason it could also result in difficulty comparing ADA results across subjects. As described herein, we recommend optimization of assay drug concentrations followed by optimization of PEG percentage using comparison of experiment samples without drug and with drug at the highest concentration expected in study samples. As demonstrated, this can enable use of the highly drug tolerant PandA format without concern over these issues.