Overcoming target interference in bridging anti-drug antibody (ADA) assay by optimizing sample treatment

ABSTRACT

Background

Drug bridging immunoassays are widely employed as the standard approach for detecting anti-drug antibodies (ADAs) in the development of new biological entities. A major challenge in these assays is mitigating target interference, particularly when the soluble target exists in dimeric forms, which can result in false positive signals and compromise assay specificity.

Research design and methods

We developed sensitive and robust ADA assays capable of overcoming target interference to detect antibodies against BI X in both cynomolgus monkey (cyno) plasma and human serum matrices. This was achieved through the implementation of simple sample treatment techniques, specifically, acidification using a panel of different acids, to disrupt dimeric target interactions and minimize the interference.

Results

Optimization of the acid dissociation and subsequent neutralization steps significantly reduced target interference in both cyno and human matrices. These improvements were achieved without the need for additional assay development or complex depletion strategies.

Conclusions

Compared to previously reported methods for mitigating target interference, the acid panel treatment approach is simpler, more time-efficient, and cost-effective. This user-friendly strategy can be readily applied to eliminate soluble dimeric targets during ADA method development, particularly in cases where alternative methodologies are not feasible or applicable.

1. Introduction

Biopharmaceutical advancements have led to the development of novel therapeutic modalities, including biologics, to address a broad spectrum of diseases. Due to their inherent immunogenic potential, new biological entities (NBEs) can elicit immune responses in humans, making anti-drug antibody (ADA) assessments a critical component in evaluating clinical safety, efficacy and pharmacokinetics [1–6].

The industry standard for immunogenicity testing of ADA assessment is the drug bridging immunoassay, performed in an Enzyme Linked Immunosorbent Assay (ELISA) or on the Meso Scale Diagnostics (MSD) Electrochemiluminescence (ECL) platform [7]. This specific method uses conjugated drugs (usually labeled with biotin, SULFO-TAG (MSD) or etc.) as the capture and detection reagents (pair referred as master mix (MM)). In this format, any potential ADA present in a sample will form a bridge between the MM reagents.

A significant advantage of the drug bridging immunoassay is that all isotypes of ADA (IgA, IgG, IgM, etc.) can be detected in this assay setup [8,9]. This method format can also serve as a “one-for-all” ADA approach, as it is applicable across all matrix species, both clinical and non-clinical. On the other hand, a major challenge in developing drug bridging ADA methods is overcoming potential false positive signals due to the presence of soluble multimeric targets in the sample [10,11].

Various articles have been published previously to describe several methodologies for minimizing target interference in ADA assays. Several of the techniques include immunodepletion using anti-target antibodies, target receptors or target specific reagents, solid-phase separation of the target, polyethylene glycol precipitation as well as others [10–16]. While these methods can successfully resolve target interference, several disadvantages have been reported, including a potential decrease in assay sensitivity, high costs of drug-specific reagents, variability in reagent quality (such as stability and affinity), and the labor-intensive nature and low throughput of some techniques [10].

To address or minimize target interference in our case study involving BI X, a single-chain variable fragment (scFv) molecule that binds to a soluble target, we first attempted target immunodepletion using either anti-target antibodies or target receptor as they were the most suitable strategies. However, in this case, no target receptor was available, and we were not successful at identifying a commercially available anti-target antibody that could deplete the target in matrix to reduce the interference. The previously reported high ionic strength dissociation assay, which utilizes magnesium chloride, offers a simple and novel strategy to minimize interference by non-covalently bond dimeric targets in ADA methods [17]. While this approach is particularly advantageous for routine use and effectively addresses target interference issues, it may lead to a reduction in sensitivity in our BI X ADA assay, as evidenced by signal loss (about 25%) observed during method optimization when using salt-based buffers (data not shown). Another straightforward and practical approach to reducing target interference is using low-pH-sample-pretreatment (without assay neutralization) method [18]. For our BI X bridging ADA assay, using low-pH-sample acidification technique without neutralization was not suitable for our assay steps, as low pH may cause protein denaturation or aggregation of the MM during the bridging step.

Here, we describe an alternative approach to overcoming target interference in drug bridging ADA assays: by optimizing sample treatment using a panel of acids with varying types and concentrations, followed by a neutralization step. This acid treatment – neutralization strategy [17,19–21] has previously been shown to effectively eliminate interference caused by dimeric or multimeric target molecules by disrupting the non-covalent interactions that stabilize these complexes [17]. However, earlier methods typically focused on a limited selection of weak acids and often required additional steps, such as immunodepletion, to achieve sufficient interference reduction.

The objective of our work is to evaluate a broader panel of acids, including strong acids such as hydrochloric acid (HCl), at varying concentrations within a standard bridging ADA assay, applicable to both clinical and non-clinical matrices. We propose that selecting the most optimal acid treatment can eliminate target interference without the need for additional assay optimization or complex depletion strategies.

2. Materials and methods

2.1. Generation of immunoassay reagents

2.1.1. Anti-BI X antibody generation

The rabbit polyclonal positive control (PC) antibody was generated (Maine Biotechnology Services, Inc.) by immunizing New Zealand white rabbits (rabbit source: Charles River Labs) with 250 µg of BI X in Freund’s Complete Adjuvant (FCA, CFA) followed by boosting every 3 weeks with 125 µg of BI X in Freund’s Incomplete Adjuvant (FIA, IFA). Production bleeds started at approximately 10 weeks post-immunization, after confirming high serum titers, and continued until ~1.6 liters of antiserum were obtained (~8 months). Purification of the rabbit serum was performed first by protein A affinity chromatography. This was followed by affinity purification of the protein A purified serum with BI X coupled to agarose beads. Finally, the affinity purified serum was cross-adsorbed against human IgG and cyno IgG to reduce antibodies reactive to the human and cyno IgG backbone.

2.1.2. Labeling of BI X

BI X is a small scFV molecule with limited binding sites, it is important to select the optimal conjugation protocol to ensure the overall final quality [22,23]. The final degree of labeling (DoL) was selected based on small scale tests using reagents with different DoL to evaluate their performance in the assay. The percentage of monomer was monitored by analytical size exclusion chromatography (aSEC) to avoid potential false positive or negative result in the ADA assays due to poor reagent quality.

An automated pressure-based filtration system (Big Tuna, Unchained Labs), Amicon tubes (MilliporeSigma) and PD-10 desalting columns (Cytiva) were used for buffer exchange and purifications. Stunner and Lunatic (Unchained Labs) were used for UV-Vis measurements. Biotin-PEG4-NHS ester (A39259), HEPES (J63578.AP) and Tris (15568) purchased from Thermo Fisher Scientific. The anhydrous DMSO (472301) were purchased from MilliporeSigma. MSD GOLD SULFO-TAG NHS Ester (R91AO–2) was acquired from MSD.

Small scale tests: BI X was buffer exchanged to 20 mM HEPES buffer pH 8.0, using Amicon tube (10 kDa) to the final concentration of 2 mg/mL. Lysine conjugation was performed using NHS ester chemistry. MSD GOLD SULFO-TAG or Biotin-PEG4-NHS were dissolved in anhydrous DMSO to a final concentration of 10 mM (stock solution) and added to the antibody at 4, 7 and 10 label:protein molar ratios (e.g., challenging ratio: 4X-7X-10X). The reaction mixture was mixed thoroughly by repetitive pipetting. The reaction mixture was incubated for 75 min- and then quenched by the addition of excess Tris, followed by further incubation for 20 min. Excess label was removed using PD-10 columns followed by Amicon Tubes (10 kDa), and the solution was buffer exchanged to a formulation buffer (pH 5.0 for SULFO-TAG and pH 6.0 for PEG4-biotin). The DoL was measured by LC-MS and UV-Vis for SULFO-TAG (2.22, 3.41, 4.66 by LC-MS and 2.46, 4.01, 5.45 by UV-Vis for 4X-7X-10X respectively) and by LC-MS for Biotin (2.91, 4.69 and 5.86 for 4X-7X-10X respectively). Purity was checked by aSEC and was 97.52, 98.93, 99.16% monomer for SULFO-TAG 4X-7X-10X and 94.26, 93.11, 93.95% monomer for biotin 4X-7X-10X, respectively. The sample were tested in the assay and DoL ~2 was selected as the optimum DoL for the large scale production.

Large scale production: BI X was buffer exchanged to 20 mM HEPES buffer pH 8.0, using Big Tuna (10 kDa filter plate) and concentrated to 2.1 mg/mL. MSD GOLD SULFO-TAG or Biotin-PEG4-NHS ester were dissolved in anhydrous DMSO to a final concentration of 8 mM and added to the antibody at 3.5 and 2.8 label:protein molar ratios, respectively. The reaction mixture was mixed thoroughly by repetitive pipetting. The reaction mixture was incubated for 1 h and quenched by the addition of excess Tris, followed by further incubation for 20 min. Excess label was removed using Big Tuna (10 kDa filter plate) and the solution was buffer exchanged to a formulation buffer (pH 5.0 for SULFO-TAG and pH 6.0 for PEG4-Biotin). The DoL was measured by LC-MS and UV-Vis for SULFO-TAG conjugated BI X (1.7 and 2.3 respectively) and by LC-MS for PEG4-biotin conjugated BI X (1.7). Purity was checked by aSEC and was 98.9% for PEG4-biotin conjugated BI X and 98.6% for SULFO-TAG conjugated BI X. The sample was aliquoted and stored at −80°C. The samples were tested periodically by aSEC and were >98% monomer after 2 years storage at −80°C.

2.2. Anti-BI X drug bridging ADA assay

The drug bridging ADA assays were designed to detect the presence of anti-BI X antibodies in either K₃EDTA cyno plasma or human serum using a multi-tiered approach (screening, confirmation and titration (human method only) according to FDA and EMA guidances [24,25].

In the screening assay, streptavidin-coated MSD plates (Catalog # L15SA, MSD) were blocked with Blocker Casein (Catalog # 37528, Thermo Fisher Scientific) to prevent nonspecific binding. Samples including PC and pooled matrix negative control (NC) were diluted to a minimum required dilution (MRD) of 1:30 in acid for 30 min (±10 min). The MM included biotin-labeled BI X and MSD SULFO-tagged BI X at concentrations of 0.5 µg/mL each in Blocker Casein. For the neutralization step, acid dissociated samples were incubated with the MM containing 1.5 M Tris Base (pH 10.0) (cyno method) or 1.0 M Tris Base (pH 10.0) (human method) for approximately 1 h (total MRD of 90). The neutralized samples were then added to the previously blocked MSD plates which were washed by PBST wash buffer and incubated for approximately 1 h. Plates were read on the MSD sector 600S instrument after the addition of 2X MSD Tripropylamine (TPA) (Catalog # R92TC–23, MSD) read buffer, and the resulting ECL was measured in relative light units (RLUs) which were proportional to the amount of ADA present in the samples. ADA samples were considered positive in the screening assay if they produced a signal greater than or equal to the plate-specific screening cut point.

The confirmatory assay was performed in the same manner, except that the samples were assessed both unspiked and spiked with BI X at a concentration of 10 µg/mL. The samples were confirmed positive if the spiked sample had % signal reduction (% inhibition) greater than or equal to the confirmatory cut point.

The titration assay (human method only) was done in an analogous manner to the screening assay, but the samples were titrated at multiple 2-fold serial dilutions, to assess how far the sample may be diluted before the assay signal drops below the titration cut point. The result of this titration assay was expressed as the highest dilution which still produces a signal above or equal to the plate-specific titration cut point.

ADA statistical analyses were performed in-house (for method development of both ADA methods), by QPS Holdings LLC (cyno ADA method validation) or by B2S Life Sciences (human ADA method validation). Statistical methods used for the analyses are consistent with robust procedures recommended by Devanarayan et al. [26] and Shankar et al. [27] when applied to immunoassay designs described by Mire-Sluis et al. [8].

Both cyno and human ADA assays were successfully developed and validated, meeting all critical method acceptance criteria, including precision, sensitivity, hook effect, selectivity, and drug tolerance. The validation summaries for each ADA method are presented in Tables 1 and 2, respectively.

Table 1.

BI X cyno ADA method validation data Summary of assay parameters and results.

BI X Cyno ADA Method Validation Data Summary
Assay Parameter	Screening	Confirmatory
Sensitivity (ng/mL)	4.73	7.15
NC Precision (%CV)	9.0	8.9
PC Precision (%CV)	10.5 to 17.4	7.6 to 10.2
Cut Point (Factor)	1.05	16.5%
Hook Effect	No Hook observed up to 100 µg/mL of PC	Not Applicable
Selectivitya	97% meet Acceptance Criteria	100% meet Acceptance Criteria
Drug Toleranceb	500 µg/mL of BI X for PC at 100 ng/mL	Not Applicable

^aSelectivity Acceptance Criteria: ≥ 80% blank sample results < their respective cut point(s) and ≥ 80% of spiked matrix samples are expected be ≥ their respective cut point(s).

^bRabbit Polyclonal PC used and reported for the assessment.

Table 2.

BI X human ADA method validation data Summary of assay parameters and results.

BI X Human ADA Method Validation Data Summary
Assay Parameter	Screening	Confirmatory
Sensitivity (ng/mL)	5.31	8.33
NCPrecision (%CV)	4.2	18.9
PC Precision (%CV)	10.7 to 13.4	0.3 to 13.7
Cut Point (Factor)a	1.07	14.4%
Hook Effect	No Hook observed up to 48 µg/mL of PC	No Hook observed up to 48 µg/mL of PC
Selectivityb	95% meet Acceptance Criteria	100% meet Acceptance Criteria
Drug Tolerancec	50 µg/mL of BI X for PC at 100 ng/mL	10 µg/mL of BI X for PC at 100 ng/mL

^aTitration Cut Point Factor: 1.19.

^bSelectivity Acceptance Criteria: ≥ 80% blank sample results < their respective cut point(s) and ≥80% of spiked matrix samples are expected be ≥ their respective cut point(s).

^cMouse Monoclonal PC used and reported for the assessment.

2.3. Measurement of target to BI X in human serum

The ELISA kit (Abcam Inc.) is a competitive immunoassay designed for the quantitative measurement of our target of interest in human plasma and serum. A target-specific antibody has been precoated onto 96-well plates and blocked. Standards or test samples are added to the wells and subsequently a target specific biotinylated detection protein is added and then followed by washing with wash buffer. Streptavidin-Peroxidase Conjugate was added and unbound conjugates were washed away with wash buffer. 3,3′,5,5′-Tetramethylbenzidine (TMB) was then used to visualize Streptavidin-Peroxidase enzymatic reaction. TMB was catalyzed by Streptavidin-Peroxidase to produce a blue color product that changed into yellow after adding acidic stop solution. The optical density (OD) was measured spectrophotometrically at a wavelength of 450 ± 2 nm. The density of yellow coloration is inversely proportional to the amount of target captured in plate.

2.4. Measurement of target to BI X in cyno plasma

The ELISA kit from Novus Biological (Novus Biologicals) uses the Sandwich-ELISA principle. The micro ELISA plate provided in this kit was pre-coated with an antibody specific to our target of interest in cyno. Standards or test samples were added to the micro ELISA plate wells and combined with the specific antibody. Then, a biotinylated detection antibody specific for the target and Avidin-Horseradish Peroxidase (HRP) conjugate were added to each micro plate well and incubated. Free components were washed away. The substrate solution was added to each well. Only those wells that contain target, biotinylated detection antibody and Avidin-HRP conjugate appeared blue in color. The enzyme-substrate reaction was terminated by the addition of stop solution and the color turned yellow. The optical density (OD) was measured spectrophotometrically at a wavelength of 450 ± 2 nm. The OD value is proportional to the concentration of target. The concentration of target in monkey samples can be back-calculated by comparing the OD of the samples to the standard curve.

3. Results

During our early method development of the MSD ECL bridging ADA assay in cyno plasma for BI X, high response of “positive” signals (RLUs) were detected in all of the pooled matrix blanks (matrix NC) and most of the individual blank/naïve matrix lots. The magnitude of signals for the pooled matrix were significantly higher than for the individual matrix (Supplement Figure S1). In addition, the buffer NC in the assay tested negative and showed an acceptable response (RLU), indicating that no false positive signal was observed in the absence of matrix using the bridging method. Based on this observation, we investigated the potential aggregation of critical reagents (capture and detection antibodies), as aggregation is reported to be the most common form of protein alteration in reagents and may lead to nonspecific matrix interferences in immunoassays [23]. aSEC assessment was later conducted to confirm that both labeled capture and detection reagents had a purity of >98% monomers and showed no sign of protein aggregation (data not shown).

Due to the abnormal data observed in all tested blank cyno plasma lots (both pooled and individual), we hypothesized that the “positive” signals were caused by dimeric target interference in the bridging ADA assay Figure 1(a). To test this hypothesis, we conducted experiments to quantify target levels in both cyno plasma and human serum using specific ELISA kits. Based on our results, the average target concentration in cyno plasma was 723 ng/mL (N = 8), while the average concentration in human serum was 1126 µg/mL (N = 24). These findings are consistent with previously reported data in the literature, particularly the target concentration in human serum (~1000 µg/mL), supporting the possibility of target interference in the ADA assay.

Figure 1a.

Target interference in BI X drug bridging anti-drug antibody (ADA) assay on the meso scale diagnostics (MSD) electrochemiluminescence (ECL) platform. BI X is a single-chain variable fragment (scFv) molecule. Its target exists as a dimer, which can bridge within the bridging ADA assay format and generate a false positive signal in ADA negative samples.

In addition to using specific ELISA kits to measure target levels, we aimed to confirm the possibility of target interference by creating our own interference samples. This was done by spiking recombinant human and cyno dimeric targets into NC and PC (rabbit polyclonal) samples prepared in buffer. These spiked samples were then tested and compared to unspiked controls in the ADA method. No matrix or acid treatment was used in this assessment. The data, presented in Supplementary Figure S2, confirm that target interference, caused by both human and cyno dimeric target forms, is observed in the ADA assay, particularly at the NC and PC 100 ng/mL levels, where approximately a 10-fold increase in RLU signal was detected. This evaluation clearly demonstrates that the dimeric target can bridge in the ADA method and cause false positive signals.

To overcome target interference in the ADA assay, we initially employed an immunodepletion strategy by screening multiple commercially available anti-target antibodies, with the goal of depleting the target from the matrix. However, we were unable to identify an antibody that could effectively bind to the target and mitigate the interference. Additional assay optimizations using detergent- or salt-based buffers were also evaluated, but these strategies resulted in minimal to no improvement. In some cases, the use of these buffers even led to a reduction in assay sensitivity (data not shown).

Various previously published studies have reported that acid dissociation can effectively eliminate target interference caused by dimeric or multimeric target proteins. However, most of these methods focused on weak acids, such as acetic acid, and often required additional steps beyond the standard bridging ADA method to achieve sufficient target depletion.

We adopted a similar approach using acid dissociation but evaluated a broader selection of acids at varying concentrations within a standard bridging ADA assay in both cyno plasma and human serum. Both methods were analyzed using drug bridging ECL platform with acid dissociation Figure 1(b). The assays employed the same blocking buffer, assay buffer, MM concentrations, and sample dilution factor. The only variable parameter, aside from the sample matrix species, was the sample treatment step, in which a panel of different acids was evaluated to optimize the method for overcoming target interference.

Figure 1b.

BI X drug bridging ADA assay with acid dissociation on the MSD ECL platform. To overcome target interference, samples are initially subjected to acid treatment. During this acid dissociation step, the dimeric target is disrupted into its monomeric form. Following dissociation, the treated samples are incubated with master mix (SULFO-TAG-BI X and Biotin-BI X) during the neutralization step. The resulting neutralized mixture is then transferred to a streptavidin-coated MSD plate for ADA detection.

The panel of acids and their corresponding pH values evaluated in this study is presented in Table 3. For the purposes of this paper, we focus primarily on four specific conditions: (1) Control (No Acid), (2) 300 mM Acetic Acid, (3) 300 mM Glycine HCl, and (4) 125 mM HCl, as these demonstrated the greatest impact in reducing target interference in our case studies.

Table 3.

List of all acids used in the acid panel evaluation.

Acid Panel List
Acid	pH
Control (No Acid)	Not Applicable
300 mM Acetic Acid	2.60
600 mM Acetic Acid	2.45
300 mM Citric Acid	3.20
300 mM Glycine HCl	2.25
25 mM HCl	2.00
50 mM HCl	1.65
125 mM HCl	1.01

Figure 2 presents the initial results of different acid treatments in the cyno plasma ADA method, while Figure 3 shows the corresponding results for the human serum ADA method. In both assays, all three acid treatment conditions resulted in decreased assay signal compared to the no acid control. The signal threshold for optimal performance was estimated to be approximately 100 RLU or lower for both pooled and individual matrix lots following acid dissociation and subsequent neutralization.

Figure 2.

Acids panel evaluation in BI X cynomolgus monkey (cyno) ADA assay. Two blank pooled plasma lots (average relative light units (RLUs) values shown in the figure) and eight blank individual (Ind) lots were evaluated. All pooled and Ind plasma lots used in this figure are different from those used in Supplement Figure S1. The no acid condition (blue bar) showed the highest RLU values for both the pooled (average) lot and all Ind lots. Among the three acid conditions, both 300 mM Glycine HCl (green bar) and 125 mM HCl (purple bar) achieved optimal assay results: all tested plasma lots had RLU values below the upper limit of assay signals.

Figure 3.

Acids panel evaluation in BI X human ADA assay. Eight blank pooled serum lots (average RLU values shown in the figure) and four blank individual (Ind) lots were tested. The no acid condition (blue bar) showed the highest RLU for both the pooled (average) lot and all of the Ind lots. Of the three acid conditions only 125 mM HCl (purple bar) achieved the optimal result: all tested serum lots have RLU values lower than the upper limit of assay signals; both the 300 mM Acetic acid (orange bar) and Glycine HCl (green bar) showed improvement in reducing target interference (RLU significantly lowered compared to no acid condition), however, the majority of the lots still have RLUs that are above the upper limit of assay signals.

In the cyno ADA method (Figure 2), two pooled blank plasma lots (average values shown) and eight individual blank lots were evaluated. Among the three acid conditions tested, both 300 mM Glycine HCl and 125 mM HCl achieved optimal signal reduction. We proceed with 300 mM glycine HCl as the final acid for the method as it is a weaker acid compared to HCl, making it gentler on the samples while still being effective.

In the human serum ADA method (Figure 3), eight pooled blank serum lots (average values shown) and four individual blank lots were assessed. While 300 mM Acetic Acid and 300 mM Glycine HCl reduced target interference to some extent, their effects were inconsistent between pooled and individual samples. Of the three acid treatments, only 125 mM HCl consistently achieved the desired signal reduction across both pooled and individual lots.

Following the identification of the most suitable acid for both cyno and human ADA assays through initial acid panel screening, we conducted further evaluations using additional individual lots. This was performed to ensure that our selected acids could effectively overcome target interference in the matrix across a large number of individuals. The corresponding findings are presented in Figures 4 and 5.

Figure 4.

Individual cyno plasma lots with 300 mM Glycine HCl acid treatment. A total of 40 individual cyno plasma lots were tested in the cyno ADA assay under the 300 mM Glycine HCl acid treatment condition. Of these, 37 lots (92.5%) produced assay signals below the plate cut point (assay signal at 55 RLU; mean value across four different plates, with a standard deviation of 3.5 and a %CV of 6.4%). Only three lots exceeded the cut point, and these outliers were within two-fold of the threshold.

Figure 5.

Individual human serum lots with 125 mM HCl acid treatment. A total of 100 individual human serum lots were tested in the human ADA assay under the 125 mM HCl acid treatment condition. Of these, 93 lots (93%) produced assay signals below the plate cut point (76 RLU; mean value across eight different plates, with a standard deviation of 5.7 and a %CV of 7.5%). Seven lots exceeded the cut point: five were within two-fold of the threshold, one was within three-fold, and one was within four-fold.

Figure 4 presents data from 40 individual cyno plasma lots tested in the cyno ADA assay under the 300 mM Glycine HCl acid treatment condition, while Figure 5 shows results from 100 individual human serum lots tested in the human ADA assay under the 125 mM HCl acid treatment condition. The overall outcomes demonstrate that the specific acid conditions used in both ADA methods are sufficient and acceptable for minimizing target interference, achieving a success rate of over 90% in each case.

4. Discussion

In the BI X bridging ADA case study, abnormally high false positive signals were observed in both cyno and human matrices. A series of investigative experiments confirmed that these false positive results were caused by interference from dimeric forms of the target protein. This interference was effectively mitigated by optimizing the acid dissociation step during sample dilution, using a panel of different acids. The acid treatment, employing a low pH range (~1.0–3.2), facilitated the dissociation of dimeric targets into their monomeric forms by disrupting the non-covalent interactions that hold the subunits together. As a result, target interference was significantly reduced, leading to accurate ADA measurement and preventing false positive outcomes.

To ensure that the resulting target monomers could no longer bind to the MM, a factor that could lead to false negative results or reduced assay sensitivity, an additional target interference assessment was conducted using the bridging ADA method. In this evaluation, recombinant cyno and human dimeric targets were spiked into NC and PC (rabbit polyclonal) samples prepared in buffer and compared to unspiked controls under the 125 mM HCl acid treatment condition to assess any impact on assay performance. The results are presented in Supplementary Figure S3. The data showed that no significant differences in assay signal were observed across all samples at both NC and PC levels. This indicates that target interference was not detected in either negative or positive samples in the presence of the dimeric target under the acid treatment condition. Additionally, the monomeric forms of the target protein did not lead to false negative results or a decrease in assay sensitivity, as evidenced by the consistent signal observed in the PC samples.

In the case study, different acid treatments demonstrated varying degrees of effectiveness in mitigating target interference. Strong acids, such as HCl, were more effective at denaturing target proteins, thereby reducing interference in the ADA assay. However, it is important to consider that acids may potentially impact assay sensitivity [28,29]. For this analysis, two different PCs, a rabbit polyclonal ADA and a mouse monoclonal ADA, were evaluated in the human ADA assay. Sensitivity of each PC was assessed at 100 ng/mL under the 125 mM HCl acid condition, following neutralization. As shown in Supplement Figure S4, both PCs achieved the desired sensitivity, with signal-to-NC ratios exceeding the estimated cut point factor of 1.1 at the tested concentration. Although HCl did not negatively impact method performance in this case, selecting an appropriate acid treatment for a specific program should be determined on a case-by-case basis.

To ensure optimal assay conditions, evaluating a range of acid concentrations should be considered. In our cyno ADA method, both 300 mM glycine HCl and 125 mM HCl effectively mitigated target interference. Among these, 300 mM glycine HCl was selected as the optimal condition, as it sufficiently met assay requirements. In contrast, the human ADA method required 125 mM HCl to achieve acceptable performance. Lower concentrations of HCl (50 mM and 25 mM) were also tested but were insufficient to fully eliminate target interference in human serum. Furthermore, 600 mM acetic acid and 300 mM citric acid were evaluated in both ADA methods; however, neither condition achieved the desired reduction in target interference.

In addition to testing a panel of different acids at various concentrations, incubation time during acid dissociation is a critical parameter to monitor [28]. In our ADA assays, an optimal incubation time of approximately 30 ± 10 min has been shown to effectively deplete the target while maintaining assay sensitivity. Extending the dissociation period beyond 1 h has been observed to negatively impact assay sensitivity, regardless of the extent of target reduction achieved. Therefore, defining an optimal incubation time range is essential for identifying the most suitable acid treatment conditions.

Moreover, neutralization of the final reaction mixture, ideally to a pH range of 7.0 to 7.5, is an important step for maintaining assay performance when implementing acid dissociation conditions. In both BI X ADA methods, Tris base at either 1.0 M or 1.5 M (pH 10.0) was used for neutralization. To confirm complete neutralization and ensure robust assay performance, PC recovery was evaluated by comparing the sensitivity curves of the polyclonal PC under acid-treated conditions (PC curve prepared in matrix) and untreated conditions (PC curve prepared in buffer). The results demonstrated comparable performance between the two conditions, indicating minimal impact of acid treatment on PC recovery in both cyno and human ADA methods (Supplementary Figures S5 and S6).

Beyond overcoming the target interference, acid treatment has also been reported to improve drug tolerance and reduce other matrix interferences in ADA assays [11,29–31]. In our case, drug tolerance was not a primary concern for either the cyno or human ADA methods. However, incorporating acid dissociation during the sample dilution step resulted in a drug tolerance limit of 500 µg/mL of BI X with the rabbit polyclonal PC at a concentration of 100 ng/mL, and 50 µg/mL of BI X with the mouse monoclonal PC at the same concentration, with minimal to no additional assay optimization. Drug tolerance using the mouse monoclonal PC was reported for the clinical ADA assay. This represents a 50-fold improvement compared to the drug tolerance observed without acid treatment and exceeds the requirement for this specific clinical program by approximately 100-fold.

As previously stated, the acid treatment technique used to overcome target interference is applicable only to non-covalently bonded target proteins [17,31,32]. The case study presented here is based on a specific ADA program involving a soluble dimeric target. This same treatment approach may also be relevant for minimizing multimeric targets in study samples.

Lastly, for future investigations, the acid panel strategy will be expanded to support a range of assessments, including diverse drug modalities (from monoclonal antibodies to bispecifics), target forms, and PCs with a broad range of binding affinities to the drug of interest.

5. Conclusion

Immunogenicity assessment remains a critical aspect of NBE development, especially in the context of endogenous dimeric targets that can compromise ADA assay specificity and lead to false positives. This study demonstrates that optimized acid treatment during sample dilution effectively mitigates such interference in both cyno and human ADA assays. Acidification at low pH facilitates the dissociation of dimeric targets into monomeric subunits by disrupting non-covalent interactions. In our study, the monomeric form of the target protein, following acid dissociation, did not adversely affect assay performance after neutralization, as evidenced by maintained sensitivity and the absence of false negative responses of the PC.

Our findings highlight the importance of tailoring acid treatment conditions, including acid type, concentration, pH, and incubation time. Additionally, precise pH control during neutralization is essential to maintain assay integrity, with an optimal range of 7.0 to 7.5 ensuring complete neutralization and reliable signal recovery. When feasible, evaluating multiple types of PCs (e.g., monoclonal and polyclonal ADAs) following acid treatment and assessing signal recovery after neutralization should be performed to ensure assay robustness.

The acid panel treatment approach offers several practical advantages: it eliminates the need for additional depletion steps or supplementary reagents, preserves time efficiency comparable to standard bridging ADA methodologies, and may improve drug tolerance in ADA assays. This streamlined and cost-effective strategy is particularly beneficial when other reported target mitigation techniques are impractical or unavailable. Future investigations will extend the acid panel strategy to support broader testing across diverse therapeutic modalities, target forms, and PCs with varying affinities, further enhancing its utility in ADA method development.

Funding Statement

All work described in this paper was funded by Boehringer Ingelheim Pharmaceuticals, Inc.

Article highlights

• In the BI X bridging ADA case study, false positive signals in cyno and human matrices, caused by dimeric target protein interference, were effectively mitigated through the use of a panel of acids with varying types and concentrations, followed by a neutralization step.

• In both BI X cyno and human ADA methods, the optimization of the acid dissociation and subsequent neutralization steps significantly reduced target interference without the need for additional assay development.

• The overall results demonstrate that the specific acid conditions used in both ADA methods are sufficient and acceptable for minimizing target interference, achieving a success rate of over 90% in each case.

• The acid panel treatment approach offers several practical advantages: it eliminates the need for depletion steps, does not require supplementary reagents such as anti-target antibodies or receptors, and maintains time efficiency comparable to standard bridging ADA methodologies.

• This user-friendly strategy can be readily applied to eliminate soluble dimeric targets during ADA method development, particularly in cases where alternative methodologies are not feasible or applicable.

Author contributions

Sally Ye conceived and planned the experiments. Sally Ye, Janice Gambardella, Lioudmila Zaslavskaia, and Daniel Kim conducted the experiments. Andrey Konovalov, Stephanie Kostuk, and Hamid Samareh Afsari contributed to the generation and characterization of critical reagents. Sally Ye, Corina Place, Kelly Coble, and Alison J. Johnson contributed to the interpretation of the results. Sally Ye drafted the original manuscript in collaboration with Stephanie Kostuk and Alison J. Johnson. Sally Ye and Hamid Samareh Afsari prepared the revised version of the manuscript. Sally Ye and Alison J. Johnson contributed to the creation of the manuscript’s figures. Sally Ye, Hamid Samareh Afsari, Corina Place, and Kelly Coble participated in the final review of the manuscript. All authors provided critical feedback and helped shape the final manuscript.

Disclosure statement

All authors are/were employed at Boehringer Ingelheim Pharmaceuticals Inc. in Ridgefield, CT, at the time of the experiments. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Ethical conduct of research

All (pooled and individual) cynomolgus monkey plasma and human serum lots were obtained from BioIVT, LLC (NY, USA). The authors state that they have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations.

Data availability statement

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials.

Supplementary Information

Supplemental data for this article can be accessed online at https://doi.org/10.1080/17576180.2025.2546709