Mitigating target interference challenges in bridging immunogenicity assay to detect anti-tocilizumab antibodies

Abstract

Aim: An assay to detect anti-tocilizumab antibodies in the presence of high levels of circulating target and drug is needed for immunogenicity assessment in comparative clinical studies.

Methods: An assay was developed and validated using a combination of blocking agents and dilutions to overcome target interference challenges.

Results: No false-positive signal was detected in serum samples spiked with 350–500 ng/ml of IL-6 receptor. As low as 50 ng/ml of positive control antibodies could be detected in the presence of either 500 ng/ml of IL-6 or 250 μg/ml of the drug product. Assay also demonstrated high sensitivity, selectivity and precision.

Conclusion: A robust, easy to perform immunogenicity assay was developed and validated for detecting anti-tocilizumab antibodies.

Plain language summary

Summary points.

• The presence of soluble drug targets in biological samples can disrupt anti-drug antibody (ADA) assays. This interference occurs when the drug targets bridge the ADA assay reagents, causing higher background values. As a result, false positive results may occur.

• Reducing soluble target interference is crucial to obtain unbiased results in ADA assays. These results are necessary for accurately assessing the safety and efficacy of biopharmaceuticals.

• Previously published reports suggested that IL-6R levels peak around day 12–15 post single dose of tocilizumab and this corresponds to ADA sampling timepoint in clinical studies. However, none of the published reports acknowledge potential interference from IL-6R in bridging format to detect anti-tocilizumab antibodies.

• Meso Scale Discovery-based bridge assay was developed and validated with an aim to detect presence of anti-tocilizumab antibodies in human serum in presence of varying concentrations of IL-6R.

• Target interference was observed with 100 ng/ml of IL-6R in spiked samples.

• Increasing dilution of the samples along with introduction of high concentrations of IL-6 in the assay buffer enabled us to achieve high target (up to 500 ng/ml) and drug tolerance (250 μg/ml of drug in presence of 50 ng/ml of positive control).

• Despite increased dilution, assay retained its sensitivity, specificity and precision.

• This assay is easy to set up with minimal processing steps and ensures no false positives are reported due to presence of IL-6R in the samples.

1. Background

Monoclonal antibodies (mAbs) have emerged as a rapidly expanding category of drugs, demonstrating remarkable therapeutic efficacy across a diverse range of ailments [1]. Tocilizumab (TCZ; brand name Actemra^®/RoActemra^®; Roche) is a monoclonal, humanized antibody that targets the IL-6 receptor (IL-6R). This drug effectively inhibits IL-6-mediated signaling and its associated pro-inflammatory effects by binding to both soluble and membrane-bound forms of IL-6R [2]. TCZ, a pharmacological agent, is frequently employed in the management of certain malignancies and inflammatory and autoimmune disorders, including rheumatoid arthritis [3–5]. Recent research has demonstrated the efficacy of TCZ in managing critical or severe instances of coronavirus disease 2019 (COVID-19) [6,7].

mAbs such as TCZ can elicit an unwanted immunogenic response when administered to humans due to the formation of anti-drug antibodies (ADA), making it crucial to design bioanalytical assays to detect immune responses. The application of ADA assays is typical in detecting antibodies specific to drugs and evaluating the potential for immunogenicity in biotherapeutic compounds [8,9]. ADAs may interfere with therapeutic antibodies efficacy and pharmacokinetic (PK) profile and potentially cause immune-mediated severe reactions such as anaphylaxis [10]. The foremost methods of detection strategies utilized in contemporary research include enzyme-linked immunosorbent assay (ELISA), electrochemiluminescence (ECL), radioimmunoprecipitation assay and surface plasmon resonance [11].

ECL is the most common platform used in bridging assays to detect ADAs due to its low background noise and high sensitivity with multiple orders of dynamic range, superiority over other methods in detecting high and low-affinity antibodies, and increasing assay throughput [12]. The contemporary testing paradigm for detecting and characterizing ADAs involves a multi-tiered strategy encompassing screening, confirmatory and titer approaches. This strategy is widely accepted and recommended by regulatory agencies [13,14]. However, for bridging assays, soluble, shed dimeric or multimeric drug targets can interfere or compete with anti-drug antibodies and can result in false positive results. High levels of circulating target concentrations have the potential to interfere with the detection of ADA [15,16]. This interference can result in either false-positive or, in certain cases, false-negative ADA results [17]. The levels of circulating drug targets in a subject may initially be low and may not cause any interference. However, these levels can significantly increase during treatment. This increase can occur because of the accumulation of drug target complexes. There are a few possible reasons for this accumulation, such as enhanced shedding of drug-engaged cell receptors, increased release of receptors from cellular breakdown, decreased clearance of the drug-target complexes, or feedback mechanisms within the biological pathway. False positive results due to target interference have been reported for patients treated with mepolizumab [18], fulranumab [19] and ofatumumab [20,21].

Some of the well-described strategies to mitigate target and drug interference include pre-treatment of the sample with anti-target proteins [18,22], target specific reagent addition in confirmatory step for removal of the target [17], solid phase separation of the target and polyethylene glycol (PEG) precipitation [17,23]. However, the above-mentioned strategies do have limitations such as; anti-target proteins when added to quench drug target although may effectively reduce target interference but they can also lead to reduction in assay signal if there is an overlap of the complementarity-determining regions (CDRs) between human anti-target protein and therapeutic protein (mAb) of interest. Thus, it is important to ensure that anti-target protein do not have overlapping sequences that may compete with the drug for ADA binding, ensuring availability of such reagents can be time and labor intensive. Target-specific reagent added in confirmatory step can be operationally challenging if there is high incidence of screen-positive samples. Solid phase separation of target and PEG precipitation techniques are laborious and low throughput and have an additional risk of under recovery of ADAs due to multiple processing steps and thereby impacting sensitivity of the assay [17,22–24].

Similarly, mAbs present in the clinical sample may bind to ADA due to chronic administration and long half-life. This binding can impede the detection of ADA and potentially lead to false negative results during the analysis of ADA samples [25]. ADA assay development and validation should demonstrate the ability to detect ADA even in the presence of high concentrations of the drug. The reported mean C_max of TCZ in normal healthy volunteer (NHV) studies with intravenous (IV) route of administration is 45.9 μg/ml [26] and with subcutaneous (SC) route of administration is 10.9 μg/ml [27].

Tocilizumab clinical studies in normal healthy volunteers and rheumatoid arthritis (RA) patients have shown a marked increase in it's target serum IL-6R levels post administration [28,29]. The increase in serum IL-6 receptor levels with tocilizumab exposure is believed to be a consequence of the binding of TCZ to those receptors [30].

The pharmacokinetics data from studies in normal healthy volunteers suggest that the serum IL-6 receptor (IL-6R) levels follow different kinetics than TCZ [26,28]. The circulating levels of serum IL-6R (baseline) reported in drug naive subjects is between 25 and 75 ng/ml [31]. Upon administration of single dose of TCZ, C_max reported for IL-6R and TCZ were 270–330 ng/ml and 7.8–15.2 μg/ml, respectively. T_max reported for IL-6R and TCZ were 296 h (246.5–346 h) (approx. 12–14 days) and 89 h (approx. 3–4 days), respectively. This data shows that mean serum IL-6R levels increase around day 15 which is also a common ADA sampling time point, therefore, these high levels of IL-6R do have potential to interfere in bridging assays and may lead to false positive results. This hypothesis further substantiated the need to develop and validate a target tolerant assay.

For biosimilar therapeutics, it is also essential to demonstrate antigenic equivalence between biosimilar and originator drug products, which will enable the use of a single assay format, wherein biosimilar drugs can be used for coating and detection [32].

Here, we report for the first time the results from development and validation of a highly target and drug tolerant immunogenicity assay for the detection of anti-tocilizumab antibodies against originator and proposed tocilizumab biosimilar. Our data demonstrates that high levels of serum IL-6R in clinical samples can interfere in bridging assay format and may lead to generation of false positive results. We overcame this challenge by employing higher dilutions along with addition of high concentrations of IL-6 (natural ligand to IL-6 receptor) in screening and confirmatory assay formats, while retaining sensitivity, specificity and precision of the assay.

2. Materials & methods

2.1. Materials & reagents

Three positive controls (PCs) were used; the first two positive controls were commercially sourced; Human anti-TCZ antibody, monoclonal (Bio-Rad, USA), rabbit anti-TCZ antibody, polyclonal (pAb; Genscript, NJ, USA), and the third polyclonal positive control was custom generated against Dr. Reddy's tocilizumab (DRL_TC) at (Genscript). A negative control (NC) pools/individuals comprising of normal human serum (NHS) and rheumatoid arthritis serum (BioIVT, NY, USA) were used throughout the study. Critical reagents like Actemra^® and RoActemra^® were procured from Jupiter Research Services LLC (Edison, NJ, USA) and Dr. Reddy's tocilizumab (DRL_TC) was manufactured at Dr. Reddy's Laboratories Ltd (Hyderabad, India), Biotinylated-TCZ (Biotin-TC) and Sulfo tag TCZ- were derived from DRL_TC and were produced by Syneos Health (NJ, USA). Streptavidin coated plates, read buffer and all other Meso Scale Discovery (MSD) consumables were sourced from MSD (MD, USA). Recombinant anti-IL-6R antibody (R&D Systems, MN, USA and Merck, MO, USA), anti-IL-6R and IL-6 were sourced from (R&D Systems).

2.2. Methods

The immunoassay used to detect anti-tocilizumab antibodies in human serum samples is a bridging assay based on MSD technology. For the preparation of immune complexes of the drug and positive controls, positive controls sourced either commercially or custom synthesized ADAs (anti-tocilizumab antibodies), were used. The screening assay utilized a bridging assay, which was complemented by an additional competitive displacement step serving as the confirmation assay. The process begins with dissociating immune complexes through acid treatment, high MRD (1/80) with 300 mM acetic acid. This is followed by a pre-incubation step where dissociated anti-tocilizumab antibodies are added to a polypropylene plate. The plate contains Biotin-TC, sulfo-tag-labeled-TC, a titrated volume of neutralization buffer for the screening assay, along with the fixed concentration of IL-6 (master mix). IL-6 at 5 μg/ml is added to quench any excess IL-6 receptor in the samples. For the confirmatory assay, the pre-incubation step involves addition of a neutralization buffer containing 50 μg/ml of the confirmatory drug. The bound complex of biotin-TC-ADA-sulfo-tag-labelled-TC is transferred to MSD gold streptavidin coated plates, where biotin-TC is immobilized on the streptavidin plate. The bound complex, also known as a bridge, generates an ECL signal when the read buffer is added, as represented in Figure 1. The development and validation of the assay method were carried out in accordance with the standard guidance documents [13,14] and Shankar et al. [33].

Figure 1.

Schematic representation of anti-drug antibody assay to detect anti-tocilizumab antibodies in human serum. Stepwise representation of principle of the assay is shown (1) Samples containing drug (tocilizumab), anti-drug antibodies and drug target (IL-6R). (2) 300 mM acetic acid is added to dissociate anti-drug antibody and drug complexes along with tocilizumab and IL-6R complexes at MRD of 1/80. (3) Acidified samples containing free drug, target and anti-drug antibodies are neutralized in master mix solution containing biotin-tocilizumab, sulfo-tag labelled-tocilizumab along with fixed concentration of IL-6 to quench IL-6R. (4) Neutralized samples are incubated to allow formation of complexes, followed by (5) washing of plate and addition of read buffer (6) data acquisition on MSD reader and (7) data analysis.

IL-6: Interleukin-6; IL-6R: Interluekin-6 receptor; MRD: Minimum required dilution.

2.3. Data analysis

The screening cut point factor (SCPF) was calculated using drug-naive individuals during assay development and assay validation. Note: Balance design utilizing drug naive matrix samples from 50 individuals generating 300 data points was used during assay validation to generate screening, confirmatory and titer cut point factor values.

2.4. Screening cut point factor analysis

After outlier removal, the normality of the entire dataset was determined in JMP software through distribution analysis with the Shapiro-Wilk test. If the distribution was normal (p ≥ 0.05 and skewness <1), then SCPF was calculated as:

$$\text{SCPF}=\text{Mean}+(1.645\times \text{StdDev})$$

If the distribution was not normal, then data was compiled in the log₁₀ (normalized data). If this distribution was normal, the SCPF was calculated as follows:

$$\begin{array}{c}\text{Value}=\text{Mean}+(1.645\times \text{StdDev})\\ \text{SCPF}={10}^{\text{Value}}\end{array}$$

If the log₁₀ transformed distribution was also not normal, then normalized data was used to calculate the cut point factor non-parametrically in Excel using the 95^th percentile.

Screening plate-based assay cut point (SACP) = mean response of NCs for the Run* SCPF

PC and samples responses ≥ the SACP in screening runs will be reported as putative positive, and responses < the SACP will be reported as negative.

Confirmatory cut point (% signal inhibition cut point) for the competitive inhibition test (confirmatory assay).

The signal inhibition ratio (SIR) was calculated for each sample (n = 50 × 6 sets):

$$\text{SIR}={log}_{10}\left(\frac{\text{sample} \text{with} \text{drug}}{\text{sample} \text{without} \text{drug}}\right)$$

If all the data obtained post-outlier removal was normally distributed, then the confirmatory cut point was derived as:

$$\text{Value}=\text{Mean} \text{SIR}-(2.33\times \text{StdDev})$$

Establish confirmatory cut point as follows:

$$\begin{array}{c}\text{Confirmatory} \text{Cut} \text{Point} \left(\text{CCP}\right)\\ =100\times \left[\right(1-\text{anti}-log\left(\text{value}\right)]\end{array}$$

If the distribution was not normal (p < 0.05 or skewness >1), then the confirmatory cut point (CCP) was calculated as follows:

$$\% \text{inhibition}=\left(1-\left(\frac{\text{with} \text{drug}}{\text{without} \text{drug}}\right)\right)\times 100$$

Identify the value at the 99^th percentile using the entire dataset as a whole.

$$\text{CCP}=\text{Mean} \% \text{inhibition}+(2.33\times \text{StdDev})$$

Report this result as the final CCP, and compute such that a 1% false positive rate is allowed.

2.5. Assay sensitivity

Assay sensitivity for the screening assay is defined as the lowest concentration level at which the assay can detect the positive control antibody.

For determining the assay sensitivity two (2)-fold sequential dilutions of the PCH (500 ng/ml) were performed in 18 independent dilution curves.

Additionally, 9 curves out of 18 curves were performed starting at 125 ng/ml (monoclonal) and 250 ng/ml (polyclonal) since not all curves produced five concentrations above the screening assay cut point and at least one concentration below the screening assay cut point.

The estimated concentration for each dilution curve was based on the last positive sample and the first negative sample according to the formula:

$$\mathrm{Sensitivity}=C_P+\left((V_P-ACP)\times \left(\frac{C_N-C_P}{V_P-V_N}\right)\right)$$

Where:

C_P = Positive Concentration; C_N = Negative Concentration; V_P = Positive Value

V_N = Negative Value; ACP = Assay Cut Point

The screening and confirmatory assay sensitivity were calculated using the equation below:

$$\begin{array}{c}\text{Assay} \text{Sensitivity}=\\ \text{Mean} \text{Concentrations}+{\text{t}}_{0.05, \text{df}}\times \text{StdDev}\end{array}$$

Acceptance criteria for all remaining parameters: Individual NC and PC samples will be considered valid if the %CV of the individual replicate is ≤20.0%. At least 3 out of 4 NC replicates must meet the %CV criteria. At least one positive control at each level sample must meet the %CV criteria.

2.5.1. Drug tolerance

Drug tolerance was performed to assess what impact the level of free tocilizumab has on the ability of the assay to detect anti-tocilizumab antibodies in the assay. Five (5) antibody levels (50, 100, 250, 500 and 1000 ng/ml) were spiked with six different levels of either Dr. Reddy's tocilizumab, Actemra^® or RoActemra^® (250, 100, 50, 10, 5 and 0 μg/ml) to evaluate the drug tolerance of the assay. The concentration of drug that dropped the assay response of the PC sample equal to or above the screening plate-based assay cut point was deemed as the drug tolerance level.

2.5.2. Precision

Positive control samples; PCH (500 ng/ml), DPC (dissociation positive control at 250 ng/ml in presence of 50 μg/ml Dr. Reddy's tocilizumab), PCM (50 ng/ml), PCL-2 (20 ng/ml) and PCL-1 (5 ng/ml) were used to evaluate the inter- and intra-assay precision. The positive control levels were evaluated in six batches (n = 3) and one batch (n = 6) performed by two analysts on two different days in both the screening and confirmatory format. Data from these assays were utilized to establish the inter-assay precision.

PCH, DPC, PCM, PCL-2 and PCL-1 were also analyzed in 1 run (n = 6) to determine the intra-assay precision.

2.5.3. Hook effect

Prozone (hook effect) was performed to assess the shape of the dose response curve of an ultra-high positive control. An ultra-high concentration of PC stock (20,000 ng/ml) was prepared in normal human serum then diluted threefold and all dilutions were tested. Samples were analyzed at n = 3 and the data were graphed as the theoretical PC concentration versus the S/N ratio. For the results to be acceptable all PC concentrations above the PCL must have S/N values above the screening cut point factor for the assay.

2.5.4. Stability

To assess stability of PC samples for their intended use, short-term (bench top), freeze/thaw and long-term stability were evaluated [34]. Three (3) replicates at the PCH (500 ng/ml), DPC (dissociation positive control at 250 ng/ml in presence of 50 μg/ml Dr. Reddy's tocilizumab), PCL (20 and 5 ng/ml) levels were evaluated for each stability experiment. The acceptance criteria included each PC must have a %CV ≤25.0% and a minimum of 2 of 3 PCs at each level must meet acceptance criteria for the stability experiment to be deemed acceptable.

3. Results & discussion

To ascertain the immunogenicity of a therapeutic protein, it is imperative to devise screening and confirmatory assays that can effectively measure immune responses. The assay development process entails optimizing the procedures and parameters utilized in processing and detecting the ADAs. The assay development experiments included optimization of MRD, overcoming serum IL-6 receptor and serum IL-6 interference, drug interference and evaluation and demonstration of antigenic equivalence, including other parameters.

3.1. Initial minimum required dilution optimization

US FDA recommends evaluating minimum required dilution (MRD) by testing against assay buffer that has been spiked with positive controls. It is generally considered more appropriate to conduct this comparative assessment during the process of assay development. If there are significant differences between the buffer and matrix samples, it suggests that the assay may need additional optimization, such as selecting a higher MRD, before proceeding with validation [35].

MRD is the minimum required dilution necessary for the detection of ADA in a biological matrix with the least interference. Focus was placed on optimizing MRD to develop highly sensitive and drug-tolerant assay. Initially, 1/20 MRD was optimized wherein the sensitivity of the assay obtained was 8.688 ng/ml and 7.212 ng/ml for polyclonal antibody in screening and confirmatory assays, respectively and 2.898 ng/ml and 3.446 ng/ml for monoclonal antibody in screening and confirmatory assays, respectively and drug tolerance achieved was up to 50 μg/ml of drug at 100 ng/ml of positive control (both monoclonal and polyclonal).

As published previously, the pharmacokinetics of TCZ and IL-6 receptors are different, and it is well documented that TCZ undergoes target-mediated drug disposition [26]. Therefore, as tocilizumab concentrations decrease, IL-6 receptor levels increase in the serum [26,28]. And in a bridging assay higher circulating target can bind can lead to false positive results. Thus, the target tolerance of the assay was evaluated at 1/20 MRD.

As shown in Figure 2A & B, at 1/20 MRD, it was observed that all negative control unspiked (0 ng/ml) samples either pooled or from two individuals (individual-1 and individual-2; unspiked), were below screening and confirmatory cut points. High positive control (500 ng/ml) was positive and above cut points in screening and confirmatory assays. All the samples (pooled and or individual) spiked with varying concentrations of IL-6 receptor (50–300 ng/ml) appeared screen positive in screening and confirmatory assays, only the (pooled and or individual) samples spiked with 50 ng/ml of IL-6 receptor were below confirmatory cut point, indicating the assay was not target tolerant and could tolerate up to only 50 ng/ml of IL-6R in the confirmatory assay at 1/20 MRD. Therefore, the next set of experiments were designed and executed to reduce the target interference, as target interference could lead to an increase in the reporting of false positives.

Figure 2.

Evaluation of IL-6R target tolerance at 1/20 minimum required dilution in screening and confirmatory assays. (A) Mean RLUs of the analysis of unspiked (0 ng/ml) and spiked with 50–300 ng/ml of IL-6R, NC pooled (purple), individual-1 (blue), Individual-2 (orange) and PCH (500 ng/ml) (green) obtained in screening assay. (B) % inhibition values of the analysis of unspiked (0 ng/ml) and spiked with 50–300 ng/ml of IL-6R; NC pooled (purple), individual-1 (blue), Individual-2 (orange) and PCH (500 ng/ml) (green) obtained in confirmatory assay. Plate Specific Screening cut point and Confirmatory cut points are represented in black dotted lines.

IL-6R: Interleukin-6 receptor; MRD: Minimum required dilution; NC: negative control; PCH: Positive control high; PCL: Positive control low; RLU: Relative light unit.

3.2. Strategy 1 for establishing target tolerance of the assay

One of the disadvantages of bridging immunoassay is interference from the circulating target in the matrix, which can lead to false positive results. It is reported that the serum IL-6 receptor attains C_max (approx. 270–330 ng/ml) after TCZ attains C_max (7.8–15.2 μg/ml) in a counterclockwise hysteresis model [28]. The circulating levels of serum IL-6 receptor (baseline) in drug naive subjects is between 25 and 75 ng/ml. Therefore, it was necessary to develop a method that tolerates up to 350 ng/ml of serum IL-6 receptor.

3.2.1. Anti-IL-6R antibody

As a first step to quench excess circulating IL-6 receptor. Anti-IL-6R antibody was added to the master mix to remove any interfering IL-6R in the samples.

It has been observed from published immunogenicity assays that blocking antibodies (anti-target antibodies) enhance the target tolerance of the assay [17]. Therefore, different concentrations of anti-IL-6R antibody were evaluated to enhance the target tolerance of the assay. Four sets of samples were prepared [a] negative control (NC), [b] NC spiked with 750 ng/ml of IL-6R, [c] NC spiked with 500 ng/ml of IL-6R, and [d] NC spiked with 100 ng/ml of IL-6R. These four samples were further processed with the addition of varying concentrations of anti-IL-6R antibody (0, 10, 50, 500, 1000 and 4000 ng/ml). From the results (Figure 3A), it is evident that except sample [a] NC, all samples in the presence and absence of anti-IL-6R antibody showed signal above plate specific screening cut point. Based on the findings, it can be concluded that the addition of anti-IL-6R antibody to the screening assay did not enhance the assay's target tolerance. In contrast, it enhanced the background and overall signal of the assay.

Figure 3.

Strategies employed to mitigate target IL-6R interference in the assay. (A) RLUs of NC pooled unspiked 0 ng/ml (orange), spiked with 750 ng/ml of IL-6R (green), 500 ng/ml of IL-6R (yellow) and with 100 ng/ml of IL-6R (blue) with varying concentrations of anti-IL-6R antibody (0–4000 ng/ml). Plate Specific Screening cut point (black dotted line). (B) Analysis of PCH 500 ng/ml (green), PCL 50 ng/ml (yellow), NC pooled spiked with 350 ng/ml of IL-6R and 50 μg/ml of anti-IL-6R antibody (blue), unspiked NC (purple) and plate specific screening cut point (black dotted line) with varying conditions. All samples were acidified 20-fold with 300 mM acetic acid. Condition-1 acidified samples were incubated to allow acid dissociation followed by neutralized in master mix containing 1M tris and 50 μg/ml of anti-IL-6R antibody. Condition-2 acidified samples were incubated to allow acid dissociation followed by neutralized in master mix containing 60 mM tris and 50 μg/ml of anti-IL-6R antibody. Condition-3 acidified samples were incubated post acid dissociation followed by addition of 50 μg/ml of anti-IL-6R diluted in 60 mM tris, followed by incubation and addition of samples to master mix containing 60 mM tris. Condition-4 post acid dissociation of samples, 50 μg/ml of anti-IL-6R antibody was added followed by incubation and addition of master mix containing 1M tris. (C) Mean RLUs of the analysis of PCH (500 ng/ml), PCL (50 ng/ml), NC-unspiked, NC spiked with 50, 100 and 350 ng/ml of IL-6R with acid dissociation (blue) and without acid dissociation (orange). Plate Specific Screening cut point without acid dissociation is represented in red dotted lines and with acid dissociation in black dotted lines (D) Mean RLUs of analysis of PCH 500 ng/ml (green), PCL 50 ng/ml(yellow)), NC (purple), NC spiked with 100 ng/ml of IL-6R (orange), NC spiked with 350 ng/ml of IL-6R (blue), NC spiked with 750 ng/ml of IL-6R (pink) at three different MRD's 1/20, 1/40 and 1/80 in presence of 5 μg/ml of IL-6 in master mix. Plate Specific Screening cut point is represented in black dotted lines. (E) % inhibition values of analysis of PCH 500 ng/ml (green), PCL 50 ng/ml (yellow), NC (purple), NC spiked with 100 ng/ml of IL-6R (orange), NC spiked with 350 ng/ml of IL-6R (blue), NC spiked with 750 ng/ml of IL-6R (pink) at three different MRD's 1/20, 1/40 and 1/80 in presence of 5 μg/ml of IL-6 in master mix. Confirmatory cut point is represented in black dotted lines.

IL-6R: Interleukin-6 receptor; MRD: Minimum required dilution; NC: negative control; PCH: Positive control high; PCL Positive control low; RLU: Relative light unit.

3.2.2. Anti-IL-6R antibody with mild basic pH

Chen et al. [21] have reported a reduction in interference from target in bridging immunoassay by using a combination of mild basic pH and anti-target antibodies [21]. A similar strategy with four different conditions with varying pH, sample processing conditions along with 50 μg/ml of anti-target antibody combinations were tested. Samples tested in this experiment were negative control (NC), PCH (500 ng/ml), PCL (50 ng/ml) and NC (negative control spiked with 350 ng/ml IL-6R with 50 μg/ml of anti-IL-6R antibody). As shown in (Figure 3B), except negative control (NC) sample. All samples showed a signal above plate specific cut point, indicating target interference at different conditions of mild basic pH, sample processing in the presence of anti-IL-6R antibody did not reduce target interference, and samples incubated with IL-6R at 350 ng/ml were still above the plate specific screening cut point, i.e., ADA positive.

3.2.3. With & without acid dissociation

Acid dissociation is a routinely employed technique in ligand binding assays to dissociate ADA and target complexes, improving or reducing target interference [36]. However, in some cases, the acidic conditions can also aggravate target interference by disrupting the drug–target complex and by releasing the accumulated target, thereby increasing the availability of free targets. To test this hypothesis, target interference experiments were performed by spiking negative control samples with varying concentrations of IL-6R (50, 100 and 350 ng/ml), and these samples were evaluated for target interference in the presence and absence of acid dissociation. As shown in Figure 3C, response values were slightly higher in samples without acid dissociation for negative control spiked and unspiked samples. Negative control samples spiked with 350 and 100 ng/ml of IL-6R showed response above plate-specific screening cut point. This confirms presence of target interference at concentrations greater than 100 ng/ml of IL-6R when tested at MRD 1/20. Based on the results, we concluded that at MRD 1/20, there was no reduction in target interference, regardless of the presence or absence of acid dissociation.

3.3. Strategy 2: change in minimum required dilution

The required target tolerance for ADA assay of TCZ is approx. 350 ng/ml. It is assumed that with the increase in MRD of the assay, the target tolerance of the assay will increase because, with the dilution of the sample, the concentration of the IL-6R will decrease.

In order, to examine the hypothesis, MRDs 1/20, 1/40 and 1/80 were evaluated, by testing PCH (500 ng/ml), PCL (50 ng/ml), negative control, NC spiked with 350 ng/ml of IL-6R, 750 ng/ml and 100 ng/ml of IL-6R. Further, along with an increase in MRD, IL-6 was also added to the master mix to purge the samples of any surplus soluble IL-6R. Figure 3D & E demonstrate that no target interference was detected up to 750 ng/ml at 1/80 MRD with IL-6 at 5 μg/ml in screening and confirmatory assays.

Thus, the finalized conditions for the assay were as follows: an MRD of 1/80 with a Biotin-TC concentration of 2 μg/ml and a Sulfo-Tag-TC concentration of 2 μg/ml with 5 μg/ml of IL-6.

3.4. Confirmation of target tolerance at 1/80 MRD

3.4.1. Target interference (IL-6R interference)

Post-establishing conditions of 1/80 MRD with 5 μg/ml of IL-6 assay target tolerance was tested with spiking pooled and individual normal healthy volunteer matrices with varying concentrations of soluble IL-6 receptor. As shown in Figure 4, a full evaluation of target tolerance of IL-6R at 1/80 MRD was performed with five individuals and pooled NC samples. Serum samples from five drug naive individuals were spiked with 250, 350 and 500 ng/ml of IL-6R. All five individuals showed a target tolerance of up to 350 ng/ml in confirmatory assay and 4/5 individuals showed target tolerance up to 500 ng/ml of IL-6R. Pooled NC samples spiked with 250 and 350 ng/ml also showed target tolerance and data with NC sample spiked with 500 ng/ml could not be obtained due to execution error. However, based on the data obtained from Figures 3D-E and Figure 4, we conclude the assay achieved target interference up to 500 ng/ml at 1/80 MRD.

Figure 4.

Evaluation of IL-6R target tolerance in individual serum samples at 1/80 minimum required dilution. % inhibition values of analysis of PCH (500 ng/ml) (green), PCH (500 ng/ml spiked with 500 ng/ml of IL-6R) (red), NC/Individuals + 250 ng/ml of IL-6R (blue), NC/individual + 350 ng/ml of IL-6R (yellow) or NC/individual + 500 ng/ml of IL-6R (orange). Confirmatory cut point is represented in red-dotted lines.

IL-6R: Interleukin-6 receptor; NC: Negative control; PCH: Positive control high.

3.4.2. Sensitivity at 1/80 MRD

While the desired target tolerance was achieved at 1/80, it was vital to ensure assay sensitivity was not impacted at this MRD. Sensitivity evaluation was performed in assay development and assay validation using monoclonal and polyclonal positive controls. Here, results of assay sensitivity achieved at 1/80 during method validation using two positive controls are presented. Sensitivity for monoclonal and polyclonal PC's were 3.140 and 10.838 ng/ml, respectively in screening assay. Confirmatory sensitivity achieved for monoclonal PC was 3.402 ng/ml and for polyclonal PC was 10.338 ng/ml in normal healthy volunteer matrix and further sensitivity of the assay in the rheumatoid arthritis matrix was also comparable with screening assay sensitivity being 3.872 ng/ml (monoclonal) and 16.913 ng/ml (polyclonal) and confirmatory assay sensitivity being 5.246 ng/ml (monoclonal) and 20.389 ng/ml (polyclonal).

3.4.3. Drug tolerance at 1/80 MRD

It is well established that high sensitivity at times can lead to poor drug tolerance [37]. To test this hypothesis, drug and positive control complexes were prepared. Varying concentrations of three drugs, DRL_TC, Actemra^® and RoActemra^® and positive control against tocilizumab were complexed. Concentrations of drugs used were 250, 100, 50, 10, 5 and 0 μg/ml and positive control concentrations were 1000, 500, 250, 100, 50 and 0 ng/ml. The concentration of the drug that dropped the assay response of the PC sample equal to or above the plate-specific screening cut point was deemed as the drug tolerance level. Based on data obtained during assay validation, the drug tolerance of the assay was determined as 250 μg/ml of drug can be tolerated at 50 ng/ml of positive control at 1/80 MRD (Figure 5). Note: Data with only 50 ng/ml of positive control is shown in Figure 5.

Figure 5.

Evaluation of drug tolerance of the assay at 1/80 minimum required dilution. Mean RLUs of the analysis of NC (purple), PC 50 ng/ml +250 μg/ml of drug (green), PC 50 ng/ml +100 μg/ml of drug (yellow), PC 50 ng/ml +50 μg/ml of drug (grey), PC 50 ng/ml +10 μg/ml of drug (blue), PC 50 ng/ml +5 μg/ml of drug (pink), PC 50 ng/ml + 0 μg/ml of drug (orange). Plate-specific screening cut point is represented in (red dotted line). Drug here denotes data obtained with all the three drug products DRL_TC, Actemra^® and RoActemra^®.

DRL_TC: Dr. Reddy's tocilizumab; NC: Negative control; PC: Positive control; RLU: Relative light unit.

3.4.4. Precision at 1/80 MRD

In normal and rheumatoid arthritis matrix, the inter and intra-assay precision of the assay was tested at 1/80 MRD in both screening and confirmatory assays. The quality control samples tested were PCH (positive control high), PCL (positive control low), PCM (positive control mid) and DPC (dissociation positive control), all the runs met set acceptance criteria and precision of the assay was <20% at 1/80 MRD in both the matrices tested (Supplementary Table S1).

In a normal healthy volunteer matrix, the positive control sample concentrations tested were PCH (500 ng/ml), PCM (250 and 50 ng/ml) and PCL (5 and 20 ng/ml). In the rheumatoid arthritis matrix, the positive control sample concentrations tested were PCH (500 ng/ml), PCM (100 ng/ml) PCL (8 and 12 ng/ml).

3.4.5. Selectivity at 1/80 MRD

Selectivity of the assay was evaluated in normal and rheumatoid arthritis matrix and selectivity samples tested were; unspiked, spiked with PCL and PCH. All the samples met the acceptance criteria, and the method was deemed selective. Concentrations of PC's tested in normal and rheumatoid arthritis matrix were:

In a normal healthy volunteer matrix, the positive control sample concentrations tested were PCH (500 ng/ml), PCM (50 ng/ml), and PCL (5 and 20 ng/ml). In the rheumatoid arthritis matrix, the positive control sample concentrations tested were PCH (500 ng/ml), PCM (100 ng/ml), PCL (8 and 12 ng/ml).

3.4.6. Specificity at 1/80 MRD

As IL-6 is a natural ligand for IL-6 receptor and it is well-documented post tocilizumab administration, serum IL-6 levels increase [28]. To evaluate specificity of the assay, the following samples were prepared by spiking 500 ng/ml of IL-6, samples were NC, PCH (500 ng/ml), and serum samples from five individuals and responses were analysed. Based on the data obtained in Figure 6, no interference from IL-6 (up to 500 ng/ml) was observed in confirmatory assay, thus confirming the specificity of the assay.

Figure 6.

Evaluation of specificity of the assay in presence of higher concentrations of IL-6 at 1/80 minimum required dilution. % inhibition values of the analysis of PCH (500 ng/ml) (green), PCH spiked with 500 ng/ml of IL-6 (yellow), NC spiked with 500 ng/ml of IL-6 (blue), individuals (1–5) spiked with 500 ng/ml of IL-6 (orange). NC spiked with 500 ng/ml was below the confirmatory cut point. Confirmatory cut point is represented in red dotted line.

IL-6: Interleukin-6; NC: Negative control; PCH: Positive control high.

3.4.7. Prozone (Hook) effect at 1/80 MRD

The phenomenon of prozone or hook effect was evaluated to ensure that there was no false negative result or change in shape of dose response curve in presence of ultra-high positive control. An ultra-high concentration of PC stock (20000 ng/ml) was prepared in normal human serum and diluted to three-fold and all dilutions were tested. Samples were analyzed at n = 3 and the data were graphed as the theoretical PC concentration versus the S/N ratio. For the results to be acceptable all PC concentrations above the PCL must have S/N values above the screening cut point factor for the assay for normal human serum. The experiment met acceptance criteria thus demonstrating that no prozone or hook effect occur up to 20,000 ng/ml of PC concentration, data obtained is shown in Supplementary Figure S1.

3.4.8. Stability of positive control at 1/80 MRD

Freeze–thaw, short term or bench-top and long-term stability were evaluated for positive control samples PCH (500 ng/ml), PCL (5 and 20 ng/ml), DPC (250 ng/ml of PC with 50 μg/ml of drug) (n = 3). Freeze–thaw stability PC samples subjected to eight cycles of freezing and thawing and were stored at -80°C (nominal) and -20°C (nominal) prior to testing. Bench-top or short-term stability was tested by removing PC's from frozen storage of -80°C and -20°C and then were placed at ambient conditions (bench top) for at least 48 before testing. Long-term storage stability PC samples were stored at -20 and -80°C and were analyzed alongside fresh run-qualifying PCs. Stability testing data confirmed; eight freeze–thaw cycles stability, short term stability up to 29 h and long-term stability up to 603 days when stored at -20 and -80°C data is shown in Supplementary Table S2.

3.4.9. Antigenic equivalence

One important aspect of biosimilar development is to demonstrate the similarity between the biosimilar and originator products. This can be achieved through physicochemical and biological testing, as well as pharmacokinetics and immunogenicity assays. These assays should be able to detect both the originator and biosimilar products in a comparable manner. If the assays can detect the originator and biosimilar products indistinguishably, a ‘one-assay’ approach using the biosimilar can be employed. This testing is commonly known as drug comparability, ADA binding comparability, or antigenic equivalence.

To demonstrate antigenic equivalence between Dr. Reddy's tocilizumab (DRL_TC) and reference products RoActemra^® and Actemra^®. Sensitivity, drug tolerance, accuracy and precision experiments were performed in a normal healthy volunteer matrix at 1/80 MRD. Polyclonal antibody (generated against DRL_TC) was used as a positive control to demonstrate antigenic equivalence in screening and confirmatory format.

3.5. Evaluation of other parameters

3.5.1. Antigenic equivalence precision

The precision of the assay was evaluated as part of antigenic equivalence with the positive control samples PCH (500 ng/ml), PCM (50 ng/ml), PCL (20 ng/ml) and DPC (250 ng/ml of positive control and 50 μg/ml of drug), PC's samples were evaluated with all three drugs; DRL_TC, Actemra^® and RoActemra^® in the confirmatory assay. As shown in Figure 7A, the response of the positive controls was similar across all three drugs when tested in a confirmatory assay wherein three different drugs were used as confirmatory drugs and precision of all the PCs tested in confirmatory assay was <15%.

Figure 7.

Evaluation of antigenic equivalence of the three drugs in precision, sensitivity and drug tolerance assay. (A) % inhibition values of analysis of PCH (500 ng/ml), PCM (50 ng/ml), PCL (20 ng/ml) and DPC (250 ng/ml + 50 μg/ml of drug). Three drugs evaluated in confirmatory assay were Actemra^® (blue), RoActemra^®(red) and DRL_TC (green). Confirmatory cut point (purple dotted line). (B) % inhibition values of analysis of sensitivity of the assay. Three drugs evaluated to assess sensitivity in confirmatory assay were Actemra^® (blue), RoActemra^®(red) and DRL_TC (green). Confirmatory cut point (purple dotted line). (C) Mean RLUs values of analysis of drug tolerance samples, samples were generated by adding varying concentrations (0–1000 μg/ml) of three drugs; Actemra^® (blue), RoActemra^®(Red) and DRL_TC (green) along with 100 ng/ml of positive control. Plate-specific screening cut point; RoActemra^® (red dotted line), DRL_TC (purple dotted line), Actemra^® (solid black line).

DRL_TC: Dr. Reddy's tocilizumab; PCH: Positive control high; PCL: Positive control low; PCM: Positive control mid; RLU: Relative light units.

3.5.2. Antigenic equivalence sensitivity

The assay's sensitivity was investigated as part of antigenic equivalence; serially diluted positive control samples (positive control was generated against DRL_TC) were tested in screening and confirmatory assays. Three drugs; DRL_TC, Actemra^®, and RoActemra^®, were tested in confirmatory assay to assess sensitivity of the assay. The data obtained was similar across all three drugs when evaluated in a confirmatory assay where sensitivity achieved by all the three different drugs was between 24 and 12 ng/ml, considering CCP of 14.035% approximately, as shown in Figure 7B.

3.5.3. Antigenic equivalence drug tolerance

Antigenic equivalence through drug tolerance experiment was demonstrated by preparing drug and positive control complexes. Note: for drugs, all three drugs, DRL_TC, Actemra^® and RoActemra^®, were incubated with varying concentrations (50–2000 ng/ml) of polyclonal positive control generated against DRL_TC. As shown in Figure 7C, drug tolerance achieved for this assay was similar across all the three drugs, and the assay could tolerate up to 1000 μg/ml of drug at 100 ng/ml of polyclonal positive control. Note: Data with only combination of 100 ng/ml of PC with 0–1000 μg/ml of drug is shown.

Thus, based on data from Figures 7A-C, it is concluded that antigenic equivalence is established between DRL_TC, Actemra^®, and RoActemra^®, which justifies single assay suitability where DRL_TC can be used as confirmatory drug for all confirmatory assays at 1/80 MRD.

In summary, target interference from IL-6R was observed in bridging immunogenicity assay developed to evaluate incidence of anti-tocilizumab antibodies in clinical trial samples. By increasing MRD and introducing high concentrations of IL-6 natural ligand of IL-6R we were able to enhance target tolerance of the assay. Further, high target tolerance was achieved by utilizing an easy to perform protocol with minimal processing steps, which makes the assay robust and rugged, less prone to execution errors and increases throughput of the assay.

4. Conclusion

TCZ binds to both soluble and membrane-bound IL-6R [5]. Binding of TCZ to IL-6R leads to increase in serum levels of IL-6R. IL-6R levels are highest around day 12–15 when a single dose of tocilizumab is administered to normal healthy volunteers [26,28] or continue to be high when repeated dose of tocilizumab is administered in patients [29]. These high levels of serum IL-6R can interfere in bridging assays and may lead to false positive signal.

Immunogenicity assays to evaluate incidence of anti-tocilizumab antibodies have been published by many groups [38–44]. All of these groups have employed bridging assay to detect incidence of anti-tocilizumab antibodies in clinical trial samples. However, to best of our knowledge none of the published reports acknowledge or investigate impact of increase in serum IL-6 receptor levels on specificity of immunogenicity assays employed during sample analysis of the clinical samples.

We hypothesized high levels of serum IL-6R may interfere with reporting of ADA results in a bridging assay format and may lead to increase in reporting of false positive results. To test this hypothesis, we spiked varying levels of IL-6R in negative control and in individual matrix samples and observed as low as 100 ng/ml of IL-6R in samples led to false positive result. This observation prompted us to develop and validate an easy to use, robust, rugged, high throughput assay with minimal processing steps. Different strategies were employed to enhance target tolerance of the assay, addition of anti-IL-6R antibody to quench IL-6R, addition of anti-IL-6R antibody with varying pH to break non-covalent bonds between target and drug, evaluate target tolerance of the assay in presence and absence of acid dissociation. Finally, with combination of high MRD (dilution) along with high concentrations of IL-6 a natural ligand to IL-6R we could achieve target tolerance up to 500 ng/ml of IL-6R.

Despite high dilutions introduced to reduce target interference assay retained its sensitivity when tested with monoclonal and polyclonal positive controls, assay could tolerate 500 ng/ml of IL-6 and up to 250 μg/ml of drug in presence of 50 ng/ml of positive control. Assay also demonstrated precision, selectivity and sensitivity in both healthy and rheumatoid arthritis matrix.

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/17576180.2024.2349417

Financial disclosure

The authors have no 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. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Competing interests disclosure

The authors are employees of Dr. Reddy's Laboratories Ltd or Syneos Health. The authors have no other competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript apart from those disclosed.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.