Abstract
Anti-drug antibody (ADA) positivity is correlated with disease relapse risk when treated with monoclonal antibody (mAb) therapeutics. ADA evaluation can assist with interpreting pharmacokinetic, pharmacological, and toxicology results. Here, we established an ADA assay based on two steps of acid dissociation combined with a bridging immunoassay to provide a comprehensive validation strategy. The three-tiered sample analysis process included screening, confirmation, and titration assays using therapeutic HLX26 (targeting lymphocyte activation gene-3 [LAG-3]) as an example. The cut points were determined by testing 50 individual normal human serum samples, including screening cut point (SCP) (SNR: 1.08), confirmatory cut point (CCP) (% inhibition: 12.65), and titration cut point (TCP) (sample-to-noise ratio [SNR]: 1.17). The assay sensitivity, low positive control (LPC), and high positive control (HPC) titer acceptable range were also set up as 33.0 ng/mL, 41.0 ng/mL, and 320–1280, respectively. After full validation, both the intra-assay and inter-assay precision testing passed with coefficient of variations (CVs) < 20%. The assay enabled excellent drug tolerance up to 768.0 μg/mL at the HPC level and 291.0 μg/mL at the LPC level, while the tolerance of target interference was up to 74.0 ng/mL of soluble LAG3. Moreover, no false-positive results were observed in the presence of 5% hemolyzed serum samples and 150 mg/dL of triglyceride in the serum samples, no hook effect was observed, and the stability performed normally under room temperature for 24 h, 2–8 °C for 7 d, and six freeze/thaw cycles. In summary, this ADA assay is feasible and could be used for evaluating the immunogenicity of HLX26 in clinical trials.
Keywords: Anti-drug antibody, Acid dissociation, LAG-3, Bridging immunoassay, Immunogenicity, Streptavidin magnetic beads
1. Introduction
LAG-3 (lymphocyte activation gene-3) is an immune checkpoint protein expressed on the surface of effector T cells and regulatory T cells (Tregs), which regulates the signaling pathway of T lymphocytes antigen-presenting cells (APCs) and plays an important role in the adaptive immune response. Studies have shown that inhibiting LAG-3 can allow T cells to regain cytotoxic activity and reduce the function of Tregs to suppress immune responses, thereby enhancing the killing effect on tumors [1,2]. LAG-3 belongs to a new generation of tumor immunotherapy targets following PD-1/L1 and CTLA-4 and has huge clinical application prospects [3,4]. Recently, a phase II/III clinical trial of the LAG-3 antibody Relatlimab combined with Nivolumab in patients with metastatic or unresectable melanoma met the primary endpoint of progression-free survival (PFS), representing the world's first Phase III clinical trial to report the efficacy of a LAG-3 antibody [5].
However, there are few reports about the immunogenicity of anti-LAG-3 antibodies. Previous studies have shown that repeated injections of therapeutic antibody drugs could induce an immune response against the drug itself, as well as increased production of anti-drug antibodies (ADAs) capable of neutralizing both the drugs and its endogenous counterparts. ADAs may affect drug exposure, pharmacokinetic characteristics, drug efficacy, and drug toxicity, which may have various clinical consequences [6]. Although both humanized and fully human mAbs greatly reduced immunogenicity, they are not completely non-immunogenic [7]. The immunogenicity of therapeutic proteins can be influenced by many factors, including the patient's genetic background, administration route, dose frequency, and treatment duration [8]. Continuous surveillance for immunogenicity in patients is not only suitable to evaluate treatment options, but can also assist with developing personalized treatment strategies. Therefore, the evaluation of the immunogenicity of antibody drugs is of paramount importance for ensuring the safety and efficacy [9,10].
Various types of assay can be used to detect ADA, including enzyme-linked immunosorbent assay (ELISA), electrochemiluminescence (ECL), radioimmunoprecipitation assay (RIPA), and surface plasmon resonance (SPR), but they often shown poor comparability of results [11]. The aim of this study was to develop and validate a method for the detection of ADA toward anti-LAG-3 in human serum using classical acid-dissociation combined with a bridging immunoassay. This assay type is most commonly used for evaluating ADAs against therapeutical protein products and was designed with reference to previous reports [[12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22]] and validated following the published recommendations and regulatory guidelines [[23], [24], [25], [26]]. In this classical three-tiered ADA assay, the samples were first screened for binding ADA, before HLX26-specific binding ADA confirmation and further analysis of the ADA titers. Neutralizing ADA was not included in this assay. The validation parameters included the screening cut point (SCP), confirmatory cut point (CCP), titration cut point (TCP), drug tolerance, target interference, precision, matrix effect (specificity, hemolysis, lipemia), hook effect, robustness, and stability. The details of the principles and acceptance criteria for method validation are described in the validation plan and report, which remain open to further valuable contributions.
2. Material and methods
2.1. Reagents, consumables, and instruments
Nunc MaxiSorp flat-bottom 96-well plates (F96) and non-binding U-bottom 96-well plates (U96), 1 × phosphate buffered saline (PBS) buffer pH 7.4 and blocking buffers SuperBlock and Blocker Casein were purchased from Thermofisher Scientific, Waltham, Massachusetts, USA. Individual and pooled normal human sera were obtained from ZenBio, Durham, North Carolina, USA. The recombinant proteins FGL1 and LAG3 were supplied by Sino Biological, Beijing, China. Streptavidin (SA)-labeled magnetic beads were purchased from Laizee, Shanghai, China. SA-labeled Horseradish Peroxidase (SA-HRP) was obtained from Abcam, Cambridge, UK. HLX26 and biotin-labeled HLX26 (HLX26-biotin) were prepared by Henlius, Shanghai, China. The TMB was purchased from Surmodics, Eden Prairie, Minnesota, USA. The 1% bovine serum albumin (BSA) and 2% BSA were prepared with 10% BSA (SeraCare, Milford, Massachusetts, USA). The 1 M Tris-HCl buffer pH 9.5 was purchased from Sangon, Shanghai, China, and the 0.3 M glycine solution (pH 2.0) and 2 M H2SO4 were prepared in our lab.
3. ADA assay procedure
On the first day, ELISA plate #1 (F96) was coated with 100 μL/well of FGL1 (2.0 μg/mL in 1 × PBS buffer) overnight, while ELISA plate #2 (F96) was coated with 100 μL/well of HLX26 (0.5 μg/mL in 1 × PBS buffer) overnight. On the second day, 12 μL of samples (including positive controls, negative controls, and samples to be tested), 58 μL of 1% BSA, and 90 μL of 0.3 M glycine were mixed together in one U96 plate before incubating for 30 ± 5 min. In the meantime, 1800 μL SA-labeled beads per plate were prepared by washing twice and blocking with 2% BSA on a vertical mixer (25 rpm) for at least 60 min. HLX26-biotin neutralizing working solution (40.0 μg/mL in 1 M Tris-HCl buffer) was added into the above U96 plate (50 μL/well) and incubated for 60 ± 10 min. In the meantime, ELISA plate #1 was washed and blocked by adding 200 μL/well of SuperBlock for 90 ± 30 min. After neutralization, 30 μL/well of blocked SA-labeled beads were added into the U96 plate for 60 ± 10 min incubation. In the meantime, ELISA plate #2 was washed and blocked by adding 200 μL/well of SuperBlock for 90 ± 30 min. After incubation, the U96 plate was placed on a magnetic plate holder, and then the beads were washed twice with 200 μL/well of 0.05% PBST, followed by adding 150 μL/well of 0.3 M glycine and incubating for 10 ± 2 min. In the meantime, ELISA plate #1 was washed, before adding 46 μL/well of 1 M Tris-HCl. Subsequently, the U96 plate was placed on a magnetic plate and 70 μL of the supernatant was transferred to ELISA plate #1 in duplicate and incubated for 60 ± 10 min. After incubation, 100 μL/well of sample from ELISA plate #1 was transferred to ELISA plate #2 and incubated for 60 ± 10 min. Subsequently, 100 μL/well of HLX26-biotin working solution was added and incubated for 30–35 min. Notably, 1.0 μg/mL HLX26-biotin in Casein Block buffer was used for the screening test, while 1.0 μg/mL HLX26-biotin and 20.0 μg/mL HLX26 mixed in Casein Block buffer were used for the confirmatory test. Finally, the reaction was reflected by the biocatalytic conversion of chromogenic TMB via SA-HRP. After the reaction was stopped by adding 2 M H2SO4, ELISA plate#2 was scanned at 450 nm using a SpectraMax M5 microplate reader. All of the steps were performed at room temperature, and all of the washing steps were performed by adding 300 μL/well of 0.05% PBST, repeated three times, before patting dry on absorbent paper. The schematic diagram of the HLX26 ADA assay process is shown in Fig. 1.
Fig. 1.
Schematic diagram of the HLX26 ADA assay process.
4. Validation parameters and acceptance criteria
Both the determination of validation parameters and the setup of acceptance criteria hereinafter were completely dependent on the regulatory guidelines from the European Medicines Agency [23], the Food and Drug Administration [24], the National Medical Products Administration of China [25], and the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use [26], as well as with reference to previous reports on the establishment and validation of the three-tiered approach to identify and characterize ADAs raised against monoclonal antibodies [[12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22]].
4.1. Determination of cut points
At least 50 individual normal human sera samples were assayed at least six times by at least two analysts over three days to determine the screening cut point (SCP), confirmatory cut point (CCP), and titration cut point (TCP). The sample-to-noise ratio (SNR) was calculated as the mean OD of samples divided by the mean OD of the plate negative control (NC; commercial pooled normal human serum). The inhibition percentage of samples was calculated as 100 × (mean OD without drug – mean OD with drug)/mean OD without drug. Outliers were excluded from the statistics if they were higher than the outlier upper limit (Q3 + N × IQR (Q3-Q1)) or lower than the outlier lower limit (Q1 - N × IQR (Q3-Q1)). Here, the N value was selected from within the range of 1.5–4.0 based on the ratio of outliers. Q1 and Q3 represent the first and third quartiles of the residuals, respectively, and their difference equals the inter-quartile range (IQR). The total number of excluded outliers was <10% of the total sample size. If the data were normally distributed after outlier exclusion, then the SCP, CCP, and TCP were calculated based on the 95th, 99th, and 99.9th percentile of the one-sided confidence interval (CI), respectively; that is SCP = mean (SNR) + 1.645 × SD (SNR), CCP = mean (Inhibition %) + 2.326 × SD (Inhibition %), and TCP = mean (SNR) + 3.090 × SD (SNR), where 1.645, 2.326, and 3.090 are the critical values of the 95%, 99%, and 99.9% confidence level of normal distribution, respectively (SD: standard deviation). If the data were non-normally distributed, then SCP = 95th percentile of SNR, CCP = 99th percentile of Inhibition %, and TCP = 99.9th percentile of SNR.
4.2. Assay controls and validation samples
Except for the system suitability NC, which was tested in eight replicates, all of the other controls and validation samples were tested in duplicate and the mean value was reported. The mean OD value of the NC replicates, which met the CV acceptance criteria (≤25%), was defined as the plate NC and used for the calculation of the SNR.
For cut point determination runs, the stock solution of HLX26 ADA was diluted with the pooled human serum to prepare ADA controls at concentrations of 640.0 ng/mL (PC1), 320.0 ng/mL (PC2), 160.0 ng/mL (PC3), 80.0 ng/mL (PC4), 40.0 ng/mL (PC5), 20.0 ng/mL (PC6), 10.0 ng/mL (PC7), 5.0 ng/mL (PC8), and 2.5 ng/mL (PC9). Screening and confirmatory assays were performed in the same analytical assay, and each assay should include two sets of serial dilution PCs (PC1-PC9) and two sets of NC samples in eight replicates. The assay was accepted if at least three of four NCs with a CV of signals of duplicate wells was ≤25% (up to two of the eight replicates can be excluded if proved as outliers by the Dixon test), and at least three of four PCs with a CV of signals of duplicate wells ≤20%, and at least three of four screening PC samples without drug had an SNR value that met the following criteria: PC1 ≥ PC2 ≥ PC3 ≥ PC4 ≥ PC5 ≥ PC6 ≥ PC7 ≥ PC8 ≥ PC9. If the PC samples did not meet the above criteria, the PC at a lower concentration was excluded from the sensitivity calculation.
For other validation assays, each screening/confirmatory assay included two sets of PCs (one HPC with 640.0 ng/mL ADA in duplicate and one LPC with 41.0 ng/mL of ADA in duplicate per set) and one set of NC (each set of NCs was analyzed in eight replicates) as the system suitability control. The assay was accepted if at least three of four NCs with a CV of signals of duplicate wells ≤25% (up to two of the eight replicates can be excluded). The NCs were required to meet the acceptance criteria of the system suitability control as defined in section 3.4, and at least three of four PC (HPC and LPC) with a CV of signals of duplicate wells ≤20%, and at least three of four PC (HPC and LPC) should meet the acceptance criteria of the system suitability control as defined in section 3.4.
Validation samples were tested in duplicate and the mean value was reported. If a validation sample with a CV of signals from duplicate wells was >20%, this sample was excluded from the calculation.
5. Calculation of the assay sensitivity, LPC, and HPC titer control acceptable range
5.1. Calculation of the assay sensitivity
The screening and confirmatory sensitivity of each assay were calculated using the function FORECAST (x, known_y's, known_x's) in Microsoft Office Excel. In the screening assay, x is SCP, known_y's is two concentrations that span SCP, known_x's is the SNR values spanning SCP, and the y value is the individual assay screening sensitivity. In the confirmatory assay, x is the CCP, known_y's is two concentrations that span the CCP, known_x's is the Inhibition % spanning the CCP, and the y value is the individual assay confirmatory sensitivity. Both the screening and confirmatory assay sensitivity = mean + (t0.05, df) × SD, where (t0.05, df) is the critical value for the one-sided 95% confidence level in the T-distribution. The sensitivity of the method refers to the lowest concentration that the positive control sample test result is continuously positive or the reading is equal to the cut point of the method. In order to ensure that the test results remain positive, finally, the sensitivity of the method was reported to be the sensitivity with the higher value in the screening and confirmatory assays.
5.2. Determination of assay LPC
The screening and confirmatory LPC of each assay were calculated using the functions as below. The LPC in screening assay = mean (Individual screening assay sensitivity) + (t0.01, df) × SD (Individual screening assay sensitivity). The LPC in confirmatory assay = mean (Individual confirmatory assay sensitivity) + (t0.01, df) × SD (Individual confirmatory assay sensitivity). (t0.01, df) is the critical value for the one-sided 99% confidence level in the T-distribution. The LPC of the method should be the LPC with the higher value in the screening and confirmatory assays.
5.3. HPC titer control acceptable range
In the cut point assays, PC9 was prepared from HPC (PC1) by two-fold serial dilution, the dilution factor of the last PC that was higher than or equal to TCP was defined as the HPC titer. For example, if the SNR of PC4 is higher than TCP and the SNR of PC5 is below TCP, then the titer of HPC is 160 (including MRD). The HPC titer lower limit and upper limit were defined as 1/2 × median (titer) and 2 × median (titer), respectively. For each titer assay, the HPC was the titer positive control (TPC) and the titer values of all TPCs should be within the validated acceptable range.
6. Establishment of system suitability control acceptance criteria
The acceptance criteria for PC and NC were calculated using all of the system suitability controls and precision samples in accepted validation assays (except long-term stability and cut point validation assays) and were used for the system suitability controls and samples stability determination. The HPC acceptance criteria were defined as mean (OD) ± 3 × SD (OD) in the absence of the drug. The inhibition % lower acceptance range in the presence of the drug was defined as mean (Inhibition %) – 3 × SD (Inhibition %). The NC acceptance criteria were defined as mean (OD) ± 3 × SD (OD). HPC and LPC should meet HPC > LPC ≥ SCP (SNR) and HPC > LPC ≥ CCP (Inhibition %).
7. Method validation
7.1. Precision
For screening assay and confirmatory assay precision evaluation, the precision (%CV) was determined from at least five inter-plate precision assays and one intra-plate precision assay, which were finished by at least two analysts over at least three days. In addition to system suitability controls, each intra-plate precision assay contained five sets of precision samples (each set containing 1 HPC, 1 LPC, and 1 NC), while each inter-plate precision assay contained three sets of precision samples as above. The intra-plate precision was calculated by the intra-plate precision assay result, while the inter-plate precision was calculated by the precision assays. To represent intra-plate precision, the inter-replicate %CVs calculated from each precision sample duplicate wells in each precision assay were used to calculate the overall mean %CV. To represent inter-precision, the SNR values obtained from each precision sample duplicate in each precision assay were used to calculate the overall %CV. Here, all precision samples were analyzed in both screening and confirmatory assays. The intra- and inter-plate precision should be ≤ 20%.
7.2. Drug tolerance
HLX26 ADA was diluted with pooled normal human serum at the concentrations of 2 × HPC and 200.0 ng/mL. HLX26 was diluted with the pooled normal human serum at the concentrations of 2560.0, 1280.0, 640.0, 320.0, 160.0, 80.0, 40.0, and 0.0 μg/mL. Equal volumes of diluted HLX26 ADA and HLX26 were mixed, incubated for at least 1 h at room temperature, and then tested in the screening assay. The drug tolerance at each concentration of HLX26 ADA in each assay was determined by function FORECAST (x, known_y's, known_x's) as in the Calculation of Screening Assay Sensitivity. If the SNR value is equal to the SCP, this SNR and the SNR below the SCP were calculated. The drug tolerance was defined as the y value calculated by this function. If none of the tested samples in the run were affected by drug addition, the drug tolerance of the method was defined as the highest drug concentration added.
7.3. Target interference
HLX26 ADA was diluted with a pooled normal human serum to the concentrations of 2 × HPC, 2 × LPC, and NC. LAG-3 was diluted with pooled normal human serum to the concentrations of 1280.0, 640.0, 320.0, 160.0, 80.0, 40.0, and 0.0 ng/mL. Equal volumes of HLX26 ADA and LAG-3 were mixed, incubated for at least 1 h at room temperature, and then analyzed in at least one assay including the screening and confirmatory assay. The target interference was tested at least once and the median target concentration was defined as the final target interference concentration. The interference level of the target at each concentration of HLX26 ADA was determined by function FORECAST (x, known_y's, known_x's) as in the calculation of drug tolerance. In the confirmatory assay, x is the CCP, known_y's is the two correlated target concentrations higher and lower than the CCP, known_x's is the 2 Inhibition % spanning the CCP (If an Inhibition % was equal to the CCP, this Inhibition % and the Inhibition % that was below the CCP were calculated). The target tolerance was defined by the calculated y value. If none of the tested samples were affected by the added target, the target tolerance was defined by the highest target concentration added.
7.4. Matrix effect
7.4.1. Selectivity
HLX26 ADA was diluted with the pooled normal human serum to the concentrations of 20 × LPC, before diluting with at least ten lots of normal individual human serum at a ratio of 1:19 to prepare spiked samples at the concentrations of LPC. Unspiked normal individual human serum was also treated as the NC level of the selectivity samples. At least one set of LPC and NC level selectivity samples was prepared (1 LPC in duplicate and 1 NC in duplicate per set). At least 80% of individual normal human serum with an SNR value meet the criteria of LPC ≥ SCP > NC.
7.4.2. Hemolysis
HLX26 ADA was diluted with a pooled normal human serum to the concentrations of 20 × LPC, before diluting with 2% and 5% hemolysis samples, respectively, at a ratio of 1:19 (containing 95% hemolysis serum) to prepare spiked hemolyzed samples at the concentrations of LPC. Unspiked 2% and 5% hemolyzed serum samples were also tested at the NC level for selectivity hemolysis samples. At least three sets of hemolyzed LPC and NC level samples were prepared (1 LPC in duplicate and 1 NC in duplicate per set). Interference was concluded if more than 1/3 of the NC samples spiked with the 2% hemolyzed blood screened positive, and/or more than 1/3 of the NC samples spiked with the 5% hemolyzed blood screened positive, and/or more than 1/3 of the LPC samples spiked with the 2% hemolyzed blood screened negative, and/or more than 1/3 of the LPC samples spiked with the 5% hemolyzed blood screened negative.
7.4.3. Lipemia
HLX26 ADA was diluted with pooled human serum at concentrations of 20 × LPC, before diluting with lipemic serum containing approximately 60.0 mg/dL and 150.0 mg/dL of triglyceride at a ratio of 1:19 (resulting in 95% lipemic matrix content) to prepare lipemic samples at the concentrations of LPC. Unspiked 60.0 mg/dL and 150.0 mg/dL lipemic serum samples were also tested as the NC level of the lipemic validation samples. At least three sets of lipemic LPC and NC level samples were prepared (1 LPC in duplicate and 1 NC in duplicate per set). Interference was concluded if more than 1/3 of NC samples spiked with the 60.0 mg/dL lipemic serum were positive, and/or more than 1/3 of NC samples spiked with the 150.0 mg/dL lipemic serum were positive, and/or more than 1/3 of LPC samples spiked with the 60.0 mg/dL lipemic serum were negative, and/or more than 1/3 of LPC samples spiked with the 150.0 mg/dL lipemic serum were negative.
7.5. Hook effect
HLX26 ADA was prepared for hook effect assessment at a concentration of 400.0 μg/mL (Hook-1) in the pooled human serum, before diluting with pooled human serum at a 5-fold serial dilution to approximately 640.0 ng/mL. All prepared hook effect samples were analyzed in at least one screening assay. If all samples above the HPC concentration were positive and the mean SNR value of these samples was higher than or equal to the mean SNR value of HPC, no hook effect was concluded.
7.6. Assay robustness
Different incubation times (Table 1) were evaluated in the precision determination assays. The assay robustness was demonstrated if the system suitability controls met the acceptance criteria established in 3.5.1.
Table 1.
Assay robustness validation items.
| Test Condition | Sample acidification | Neutralization incubation #1 | Bead incubation | Bead acidification | Neutralization incubation #2 |
|---|---|---|---|---|---|
| Incubation time (min) | 25 | 50 | 50 | 9 | 50 |
| 35 | 70 | 70 | 11 | 70 | |
| Test condition | Transfer incubation | HLX26-biotin incubation | SA-HRP incubation | TMB color | |
| Incubation time (min) | 50 | 30 | 30 | 8 | |
| 70 | 35 | 35 | 12 | ||
7.7. Stability
The stability test was conducted at room temperature (24 h), 2–8 °C (7 d), and six cycles of freezing and thawing. The stability samples included three sets of samples (1 HPC, 1 LPC, and 1 NC, analyzed in duplicate). Under different stability conditions, at least 2/3 of the stability samples should meet the acceptance criteria described in section 3.2.
8. Statistical analysis
SAS JMP Statistical version was used for statistical analysis. The NC outliers in the cut point determination assays and other assays were excluded if the Dixon test determined an α value of 0.05. The Shapiro–Wilk test was performed for normal distribution testing with a p-value of 0.05.
9. Results
9.1. Cut points
Fifty human individual normal sera samples were assayed six times by two analysts over three days to determine the cut points (the details are shown in Table S1). As shown in Table 2, the data of SNRs and Inhibition % in the cut point determination assays were non-normally distributed (P < 0.0001) after removing the outliers. According to the calculation formula mentioned in section 3.1, the SCP, CCP, and TCP were finally set as 1.08, 12.65, and 1.17, respectively. Here, the proportion of outliers was less than 10%, when choosing a value of 4 for N. The details are shown in Table 2.
-
2.
Sensitivity, LPC, and titer
Table 2.
Cut point determination.
| SCP (SNR, N = 4) | CCP (Inhibition %, N = 4) | TCP (SNR, N = 4) | ||||
|---|---|---|---|---|---|---|
| Q1 | 0.97 | −0.88 | 0.97 | |||
| Q3 | 1.01 | 3.65 | 1.01 | |||
| IQR | 0.04 | 4.53 | 0.04 | |||
| Q1–4*IQR | 0.81 | −19.00 | 0.81 | |||
| Q3+4*IQR | 1.17 | 21.77 | 1.17 | |||
| Normal Distribution Test | Shapiro–Wilk W test | |||||
| w = 0.883003 | w = 0.964662 | w = 0.883003 | ||||
| P < 0.0001 | P < 0.0001 | P < 0.0001 | ||||
| Non-normal | Non-normal | Non-normal | ||||
| Cut points | SCP (95%) | 1.08 | CCP (99%) | 12.65 | TCP (99.9%) | 1.17 |
| Outlier | 3/317 (0.95%) | 1/313 (0.32%) | 3/317 (0.95%) | |||
A total of 29 assays were conducted to determine the assay sensitivity, LPC, and HPC titer control acceptable range (the details are shown in Table 3). As calculated according to section 3.3, the screening assay sensitivity and LPC were determined as 16.0 ng/mL and 19.0 ng/mL, while the confirmatory assay sensitivity and LPC were determined as 33.0 ng/mL and 41.0 ng/mL, respectively. Finally, the sensitivity and LPC of this ADA assay were 33.0 ng/mL and 41.0 ng/mL, respectively, which is the higher value in the screening and confirmatory assays. The HPC titer median was determined as 640, with a range of 320 (1/2 × HPC titer median) to1280 (2 × HPC titer median).
Table 3.
Determination of assay sensitivity, LPC, and HPC titer control acceptable range.
| Run No. | Screening assay |
Confirmatory assay |
HPC titer control |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| QC | Conc. (ng/mL) | SNR | S-sensitivity | QC | Conc. (ng/mL) | Inhibition (%) | C-sensitivity | QC | Dilution factor | SNR | Dilution factor | HPC titer | |
| 7 | PC7 | 10.0 | 1.16 | 6 | PC 7 | 10.0 | 13.90 | 9 | PC6 | 32 | 1.30 | 32 | 640 |
| PC8 | 5.0 | 1.07 | PC 8 | 5.0 | 5.72 | PC7 | 64 | 1.16 | |||||
| 8 | PC7 | 10.0 | 1.08 | 10 | PC6 | 20.0 | 19.55 | 12 | PC6 | 32 | 1.19 | 32 | 640 |
| PC8 | 5.0 | 1.02 | PC7 | 10.0 | 10.94 | PC7 | 64 | 1.08 | |||||
| 9 | PC7 | 10.0 | 1.08 | 10 | PC6 | 20.0 | 16.21 | 15 | PC5 | 16 | 1.34 | 16 | 320 |
| PC8 | 5.0 | 1.03 | PC7 | 10.0 | 8.83 | PC6 | 32 | 1.16 | |||||
| 10 | PC6 | 20.0 | 1.14 | 15 | PC 6 | 20.0 | 15.21 | 18 | PC5 | 16 | 1.26 | 16 | 320 |
| PC7 | 10.0 | 1.03 | PC7 | 10.0 | 3.70 | PC6 | 32 | 1.14 | |||||
| 13 | PC7 | 10.0 | 1.08 | 10 | PC6 | 20.0 | 22.61 | 14 | PC6 | 16 | 1.25 | 16 | 320 |
| PC8 | 5.0 | 1.03 | PC7 | 10.0 | 5.65 | PC7 | 64 | 1.08 | |||||
| 14 | PC7 | 10.0 | 1.11 | 8 | PC6 | 20.0 | 14.63 | 14 | PC6 | 32 | 1.20 | 32 | 640 |
| PC8 | 5.0 | 1.05 | PC7 | 10.0 | 11.40 | PC7 | 64 | 1.11 | |||||
| 15 | PC8 | 5.0 | 1.13 | 4 | PC8 | 5.0 | 13.47 | 5 | PC6 | 32 | 1.29 | 32 | 640 |
| PC9 | 2.5 | 0.98 | PC9 | 2.5 | 2.54 | PC7 | 64 | 1.13 | |||||
| 16 | PC6 | 20.0 | 1.17 | 11 | PC7 | 10.0 | 12.67 | 10 | PC6 | 32 | 1.17 | 32 | 640 |
| PC7 | 10.0 | 1.07 | PC8 | 5.0 | 5.49 | PC7 | 64 | 1.07 | |||||
| 17 | PC7 | 10.0 | 1.10 | * | PC8 | 5.0 | 12.74 | 5 | PC6 | 32 | 1.22 | 32 | 640 |
| PC9 | 2.5 | 1.02 | PC9 | 2.5 | 1.00 | PC7 | 64 | 1.10 | |||||
| 18 | PC7 | 10.0 | 1.11 | 8 | PC7 | 10.0 | 13.42 | 9 | PC6 | 32 | 1.21 | 32 | 640 |
| PC8 | 5.0 | 1.05 | PC8 | 5.0 | 5.78 | PC7 | 64 | 1.11 | |||||
| 19 | PC6 | 20.0 | 1.19 | 11 | PC6 | 20.0 | 18.32 | 13 | PC6 | 32 | 1.19 | 32 | 640 |
| PC7 | 10.0 | 1.07 | PC7 | 10.0 | 10.04 | PC7 | 64 | 1.07 | |||||
| 20 | PC7 | 10.0 | 1.14 | 6 | PC7 | 10.0 | 15.95 | 7 | PC6 | 32 | 1.27 | 32 | 640 |
| PC8 | 5.0 | 1.07 | PC8 | 5.0 | 10.51 | PC7 | 64 | 1.14 | |||||
| 21 | PC7 | 10.0 | 1.11 | 6 | PC7 | 10.0 | 13.11 | 9 | PC6 | 32 | 1.24 | 32 | 640 |
| PC8 | 5.0 | 1.07 | PC8 | 5.0 | 8.67 | PC7 | 64 | 1.11 | |||||
| 22 | PC8 | 5.0 | 1.09 | 5 | PC8 | 5.0 | 13.64 | 5 | PC6 | 32 | 1.32 | 32 | 640 |
| PC9 | 2.5 | 1.01 | PC9 | 2.5 | 8.59 | PC7 | 64 | 1.16 | |||||
| 23 | PC7 | 10.0 | 1.10 | 9 | PC7 | 10.0 | 15.32 | 8 | PC6 | 32 | 1.21 | 32 | 640 |
| PC8 | 5.0 | 1.02 | PC8 | 5.0 | 7.18 | PC7 | 64 | 1.10 | |||||
| 24 | PC7 | 10.0 | 1.08 | 10 | PC6 | 20.0 | 16.72 | 13 | PC5 | 16 | 1.35 | 16 | 320 |
| PC8 | 5.0 | 1.05 | PC7 | 10.0 | 11.22 | PC6 | 32 | 1.16 | |||||
| 25 | PC8 | 5.0 | 1.16 | 4 | PC8 | 5.0 | 15.85 | 4 | PC6 | 32 | 1.21 | 32 | 640 |
| PC9 | 2.5 | 1.01 | PC9 | 2.5 | 3.21 | PC7 | 64 | 1.16 | |||||
| 26 | PC8 | 5.0 | 1.09 | 4 | PC7 | 10.0 | 13.83 | 7 | PC6 | 32 | 1.26 | 32 | 640 |
| PC9 | 2.5 | 1.05 | PC8 | 5.0 | 11.79 | PC7 | 64 | 1.13 | |||||
| 27 | PC7 | 10.0 | 1.08 | 10 | PC4 | 80.0 | 21.23 | 57 | PC6 | 32 | 1.17 | 32 | 640 |
| PC8 | 5.0 | 1.06 | PC5 | 40.0 | 6.15 | PC7 | 64 | 1.08 | |||||
| 28 | PC6 | 20.0 | 1.13 | 14 | PC5 | 40.0 | 16.22 | 23 | PC5 | 16 | 1.21 | 16 | 320 |
| PC7 | 10.0 | 1.05 | PC6 | 20.0 | 11.96 | PC6 | 32 | 1.13 | |||||
| 29 | PC8 | 5.0 | 1.09 | 4 | PC6 | 20.0 | 14.72 | 14 | PC6 | 32 | 1.18 | 32 | 640 |
| PC9 | 2.5 | 1.07 | PC7 | 10.0 | 11.07 | PC7 | 64 | 1.10 | |||||
| 30 | PC8 | 5.0 | 1.10 | 4 | PC5 | 40.0 | 25.55 | * | PC5 | 16 | 1.35 | / | * |
| PC9 | 2.5 | 1.04 | PC7 | 10.0 | 10.17 | PC7 | 64 | 1.10 | |||||
| 31 | PC7 | 10.0 | 1.10 | 7 | PC5 | 40.0 | 16.42 | 25 | PC5 | 16 | 1.23 | 16 | 320 |
| PC8 | 5.0 | 1.07 | PC6 | 20.0 | 11.26 | PC6 | 32 | 1.12 | |||||
| 32 | PC6 | 20.0 | 1.11 | 14 | PC5 | 40.0 | 15.47 | 31 | PC5 | 16 | 1.18 | 16 | 320 |
| PC7 | 10.0 | 1.06 | PC6 | 20.0 | 8.84 | PC6 | 32 | 1.11 | |||||
| 33 | PC6 | 20.0 | 1.09 | 15 | PC5 | 40.0 | 19.98 | 28 | PC5 | 16 | 1.30 | 16 | 320 |
| PC7 | 10.0 | 1.07 | PC6 | 20.0 | 7.59 | PC6 | 32 | 1.09 | |||||
| 34 | PC6 | 20.0 | 1.09 | 19 | PC6 | 20.0 | 14.69 | 17 | PC5 | 16 | 1.21 | 16 | 320 |
| PC7 | 10.0 | 1.02 | PC7 | 10.0 | 7.25 | PC6 | 32 | 1.09 | |||||
| 35 | PC6 | 20.0 | 1.15 | 12 | PC6 | 20.0 | 15.60 | 17 | PC5 | 16 | 1.32 | 16 | 320 |
| PC7 | 10.0 | 1.06 | PC7 | 10.0 | 5.27 | PC6 | 32 | 1.15 | |||||
| 36 | PC6 | 20.0 | 1.14 | 11 | PC7 | 10.0 | 14.79 | 9 | PC5 | 16 | 1.31 | 16 | 320 |
| PC7 | 10.0 | 1.07 | PC8 | 5.0 | 5.50 | PC6 | 32 | 1.14 | |||||
| 37 | PC8 | 5.0 | 1.11 | 4 | PC7 | 10.0 | 13.23 | 7 | PC6 | 32 | 1.17 | 32 | 640 |
| PC9 | 2.5 | 0.98 | PC8 | 5.0 | 12.23 | PC7 | 64 | 1.11 | |||||
| Screening sensitivity | 16 | Confirmatory sensitivity | 33 | Median | 640 | ||||||||
| Screening LPC | 19 | Confirmatory LPC | 41 | HPC titer control acceptable range: 320–1280 | |||||||||
Note: *As the CV value of duplicate wells exceeded 20%, the two dilutions spanning the cut points are discontinuous and the assay data were removed from the result of the S-sensitivity, C-sensitivity, and HPC titer columns. LPC: low positive control; HPC: high positive control. SNR: sample-to-noise ratio.
10. Determination of system suitability control acceptance criteria
Fifty-four sets of HPC, LPC, and NC data from 12 assays were collected to determine the system suitability control acceptance criteria (the details are shown in Table 4). The mean OD values ± 3 × SD were used for the quality control of HPC (0.7152 ± 0.4131), LPC (0.1209 ± 0.0315), and NC (0.0787 ± 0.0093), respectively in the screening assay, while the mean Inhibition % ± 3 × SD was used for the quality control of HPC (84.76% ± 4.9060%) and LPC (36.01% ± 10.4856%), respectively, in the confirmatory assay.
Table 4.
System suitability control acceptance criteria.
| Run No. | QC | Screening assay (OD values) |
Confirmatory assay (%Inhibition) |
|||
|---|---|---|---|---|---|---|
| HPC | LPC | NC | HPC | LPC | ||
| 38 | 1 | 0.9400 | 0.1403 | 0.0819 | 86.52 | 44.90 |
| 2 | 0.7903 | 0.1258 | 84.84 | 29.33 | ||
| 39 | 1 | 0.8968 | 0.1350 | 0.0803 | 86.35 | 40.81 |
| 2 | 0.6995 | 0.1254 | 82.59 | 33.81 | ||
| 40 | 1 | 0.7327 | 0.1298 | 0.0803 | 84.63 | 35.13 |
| 2 | 0.5581 | 0.1165 | 80.58 | 34.08 | ||
| 41 | 1 | 0.7493 | 0.1166 | 0.0777 | 85.97 | 35.68 |
| 2 | 0.6220 | 0.1081 | 85.37 | 26.64 | ||
| 42 | 1 | 0.8028 | 0.1292 | 0.0759 | 88.19 | 41.64 |
| 2 | 0.7029 | 0.1151 | 87.01 | 36.40 | ||
| 43 | 1 | 0.5591 | 0.1121 | 0.0778 | 84.15 | 32.29 |
| 2 | 0.4999 | 0.1051 | 80.08 | 26.45 | ||
| 44 | 1 | 0.4222 | 0.1131 | 0.0853 | * | * |
| 2 | 0.3635 | 0.1028 | ||||
| 45 | 1 | 0.7156 | 0.1194 | 0.0753 | 87.06 | 35.18 |
| 2 | 0.7004 | 0.1114 | 86.89 | 34.38 | ||
| 46 | 1 | 0.7291 | 0.1194 | 0.0757 | 87.57 | 39.53 |
| 2 | 0.7699 | 0.1233 | 88.58 | 40.79 | ||
| 47 | 1 | 0.7832 | 0.1202 | 0.0752 | 85.41 | 31.86 |
| 2 | 0.7248 | 0.1189 | 85.35 | 36.00 | ||
| 48 | 1 | 0.4987 | 0.1029 | 0.0788 | * | * |
| 2 | 0.4486 | 0.1059 | ||||
| 49 | 1 | 0.8124 | 0.1358 | 0.0767 | 81.81 | 44.18 |
| 2 | 0.7036 | 0.1444 | 81.47 | 45.15 | ||
| 38 | 1 | 1.0571 | 0.1429 | 0.0835 | 87.91 | 41.36 |
| 2 | 0.9953 | 0.1259 | 0.0796 | 87.20 | 36.93 | |
| 3 | 0.9125 | 0.1107 | 0.0812 | 86.64 | 27.82 | |
| 4 | 0.6594 | 0.1164 | 0.0817 | 81.98 | 31.53 | |
| 5 | 0.8148 | 0.1305 | 0.0828 | 84.19 | 37.09 | |
| 39 | 1 | 0.8849 | 0.1324 | 0.0814 | 87.76 | 40.56 |
| 2 | 0.8363 | 0.1272 | 0.0805 | 86.32 | 37.66 | |
| 3 | 0.7583 | 0.1279 | 0.0811 | 83.85 | 39.56 | |
| 4 | 0.7822 | 0.1253 | 0.0826 | 84.75 | 36.79 | |
| 5 | 0.8189 | 0.1291 | 0.0791 | 85.58 | 38.26 | |
| 40 | 1 | 0.7298 | 0.1267 | 0.0804 | 84.98 | 37.81 |
| 2 | 0.7394 | 0.1252 | 0.0830 | 83.78 | 37.14 | |
| 3 | 0.6122 | 0.1089 | 0.0762 | 81.25 | 29.94 | |
| 4 | 0.6139 | 0.1188 | 0.0810 | 82.54 | 32.07 | |
| 5 | 0.6228 | 0.1125 | 0.0814 | 82.10 | 30.93 | |
| 41 | 1 | 0.7332 | 0.1147 | 0.0777 | 85.90 | 29.21 |
| 2 | 0.6711 | 0.1201 | 0.0806 | 85.96 | 33.97 | |
| 3 | 0.6356 | 0.1173 | 0.0787 | 84.47 | 34.44 | |
| 42 | 1 | 0.8711 | 0.1212 | 0.0733 | 89.15 | 40.92 |
| 2 | 0.8172 | 0.1194 | 0.0738 | 88.18 | 34.92 | |
| 3 | 0.7505 | 0.1158 | 0.0767 | 87.37 | 36.44 | |
| 43 | 1 | 0.5810 | 0.1173 | 0.0788 | 82.25 | 33.59 |
| 2 | 0.6089 | 0.1080 | 0.0775 | 84.45 | 29.44 | |
| 3 | 0.5476 | 0.1052 | 0.0807 | 82.56 | 27.19 | |
| 47 | 1 | 0.6804 | 0.1200 | 0.0781 | 84.00 | 34.10 |
| 2 | 0.7081 | 0.1204 | 0.0756 | 84.40 | 35.80 | |
| 3 | 0.6515 | 0.1095 | 0.0739 | 84.30 | 33.80 | |
| 49 | 1 | 0.7739 | 0.1328 | 0.0742 | 82.30 | 45.00 |
| 2 | 0.8089 | 0.1287 | 0.0733 | 83.40 | 42.20 | |
| 3 | 0.7204 | 0.1433 | 0.0745 | 77.90 | 49.60 | |
| Mean | 0.7152 | 0.1209 | 0.0787 | 84.76 | 36.01 | |
| SD | 0.1377 | 0.0105 | 0.0031 | 2.4530 | 5.2428 | |
| Mean – 3*SD | 0.3021 | 0.0894 | 0.0694 | 77.40 | 20.28 | |
| Mean + 3*SD | 1.1283 | 0.1524 | 0.0880 | 92.12 | 51.74 | |
Note: *No confirmatory assay was performed in runs 44 and 48.
11. Precision and robustness
Five sets of HPC, LPC, and NC were included in one assay to evaluate the intra-assay precision. In the screening assay, the CVs of the SNR data were detected as 11.09%, 9.80%, and 2.00% for HPC, LPC, and NC, respectively; the CVs of the inhibition % data were detected as 2.86% and 15.14% for HPC and LPC, respectively, in the confirmatory assay (Table 5). Theoretically and in fact, NC samples have a wide range of Inhibition % values and a large CV, and therefore are not considered in the precision evaluation. Based on the above data (details shown in Table 5), all of the CVs (except the CV of the NC in the confirmatory assay) were less than 20%, which demonstrated acceptable intra-assay precision of this ADA assay.
Table 5.
Intra-plate and inter-plates precision calculation.
| Intra-plate precision | |||||||
|---|---|---|---|---|---|---|---|
| Lot | Sample | Screening assay (SNR values) |
Confirmatory assay (%Inhibition) |
||||
| HPC | LPC | NC | HPC | LPC | NC | ||
| 38 |
1 | 12.91 | 1.74 | 1.02 | 87.91 | 41.36 | 6.47 |
| 2 | 12.15 | 1.54 | 0.97 | 87.20 | 36.93 | 2.76 | |
| 3 | 11.14 | 1.35 | 0.99 | 86.64 | 27.82 | −3.20 | |
| 4 | * | 1.42 | 1.00 | 81.98 | 31.53 | 2.69 | |
| 5 | 9.95 | 1.59 | 1.01 | 84.19 | 37.09 | 8.70 | |
| Mean | 11.54 | 1.53 | 1.00 | 85.58 | 34.95 | 3.48 | |
| STD | 1.28 | 0.15 | 0.02 | 2.45 | 5.29 | 4.53 | |
| %CV |
11.09 |
9.80 |
2.00 |
2.86 |
15.14 |
/ |
|
|
Robustness | |||||||
| 39 | 1 | 11.02 | 1.65 | 1.01 | 87.76 | 40.56 | 6.39 |
| 2 | 10.41 | 1.58 | 1.00 | 86.32 | 37.66 | 0.99 | |
| 3 | 9.44 | 1.59 | 1.01 | 83.85 | 39.56 | 6.78 | |
| 4 | 9.74 | 1.56 | 1.03 | 84.75 | 36.79 | −0.48 | |
| 5 | 10.20 | 1.61 | 0.99 | 85.58 | 38.26 | 2.91 | |
| Mean | 10.09 | 1.60 | 1.01 | 85.65 | 38.57 | 3.32 | |
| STD | 0.65 | 0.03 | 0.01 | 1.50 | 1.50 | 3.22 | |
| %CV | 6.44 | 1.88 | 0.99 | 1.75 | 3.89 | / | |
| 40 | 1 | 9.09 | 1.58 | 1.00 | 84.98 | 37.81 | −6.97 |
| 2 | 9.21 | 1.56 | 1.03 | 83.78 | 37.14 | 6.14 | |
| 3 | 7.62 | 1.36 | 0.95 | 81.25 | 29.94 | 4.72 | |
| 4 | 7.65 | 1.48 | 1.01 | 82.54 | 32.07 | 4.44 | |
| 5 | 7.76 | 1.40 | 1.01 | 82.10 | 30.93 | 8.97 | |
| Mean | 8.16 | 1.48 | 1.00 | 82.93 | 33.58 | 3.46 | |
| STD | 0.85 | 0.10 | 0.03 | 1.47 | 3.64 | 6.10 | |
| %CV | 10.42 | 6.76 | 3.00 | 1.77 | 10.84 | / | |
| 38–40 |
Mean | 9.88 | 1.53 | 1.00 | 84.72 | 35.70 | 3.42 |
| STD | 1.60 | 0.11 | 0.02 | 2.17 | 4.15 | 4.41 | |
| %CV |
16.19 |
7.19 |
2.00 |
2.56 |
11.62 |
/ |
|
|
Inter-plates precision | |||||||
| 41 | 1 | 9.44 | 1.48 | 1.00 | 85.90 | 29.21 | −1.29 |
| 2 | 8.64 | 1.55 | 1.04 | 85.96 | 33.97 | 0.62 | |
| 3 | 8.18 | 1.51 | 1.01 | 84.47 | 34.44 | 4.96 | |
| 42 | 1 | 11.48 | 1.60 | 0.97 | 89.15 | 40.92 | 3.82 |
| 2 | 10.77 | 1.57 | 0.97 | 88.18 | 34.92 | 2.85 | |
| 3 | 9.89 | 1.53 | 1.01 | 87.37 | 36.44 | 6.26 | |
| 43 | 1 | 7.47 | 1.51 | 1.01 | 82.25 | 33.59 | 6.73 |
| 2 | 7.83 | 1.39 | 1.00 | 84.45 | 29.44 | 2.97 | |
| 3 | 7.04 | 1.35 | 1.04 | 82.56 | 27.19 | 3.59 | |
| 47 | 1 | 9.05 | 1.60 | 1.04 | 83.95 | 34.08 | 6.40 |
| 2 | 9.42 | 1.60 | 1.01 | 84.38 | 35.80 | 2.65 | |
| 3 | 8.66 | 1.46 | 0.98 | 84.27 | 33.79 | 0.68 | |
| 49 | 1 | 10.09 | 1.73 | 0.97 | 82.31 | 44.95 | 2.70 |
| 2 | 10.55 | 1.68 | 0.96 | 83.37 | 42.19 | 0.68 | |
| 3 | 9.39 | 1.87 | 0.97 | 77.92 | 49.55 | −0.13 | |
| 38–43, 47 & 49 | Mean | 9.52 | 1.55 | 1.00 | 84.58 | 35.86 | 3.16 |
| STD | 1.45 | 0.12 | 0.02 | 2.44 | 5.11 | 3.52 | |
| %CV | 15.23 | 7.74 | 2.00 | 2.88 | 14.25 | / | |
Note: *As the CV value of duplicate wells exceeded 20%, the assay data were removed from the precision calculation.
Next, the robustness was investigated in intra-assay precision assays. In the robustness assay, both the lower and upper limits of the incubation time range (Table 1). Similarly, all of the CVs (except the CV of the NC in the confirmatory assay) in each robustness assay were less than 20%, and all of the CVs (except the CV of the NC in the confirmatory assay) among the above intra-assay precision run and two robustness runs were also less than 20%, demonstrating acceptable robustness of the assay (Table 5).
Another five inter-assay precision runs were conducted, with each assay containing three sets of precision samples as above. Combined with the data collected from intra-assay precision and robustness runs, the inter-assay precision was calculated and also passed with CVs ≤20% (Table 5).
12. Drug tolerance
As shown in Table 6, for the two SNRs spanning SCP, the final concentrations of HLX26 were 320.0 and 160.0 μg/mL in the serum with 100.0 ng/mL HLX26 ADA (LPC), and 1280.0 and 640.0 μg/mL in the serum containing 640.0 ng/mL of HLX26 ADA (HPC). Calculated by the function, the drug tolerance was up to 768.0 μg/mL at HPC level and 291.0 μg/mL in LPC level.
Table 6.
Drug tolerance.
| Samples | HLX26 Conc. (μg/mL) | Mean OD | SNR |
|---|---|---|---|
| LPC-HLX26-1 | 2560.0 | 0.0792 | 0.93 |
| LPC-HLX26-2 | 1280.0 | 0.0813 | 0.95 |
| LPC-HLX26-3 | 640.0 | 0.0907 | 1.06 |
| LPC-HLX26-4 | 320.0 | 0.0999 | 1.17 |
| LPC-HLX26-5 | 160.0 | 0.1034 | 1.21 |
| LPC-HLX26-6 | 80.0 | 0.1161 | 1.36 |
| LPC-HLX26-7 | 40.0 | 0.1240 | 1.45 |
| LPC-HLX26-8 | 0.0 | 0.1251 | 1.47 |
| HPC-HLX26-1 | 2560.0 | 0.0815 | 0.96 |
| HPC-HLX26-2 | 1280.0 | 0.0944 | 1.11 |
| HPC-HLX26-3 | 640.0 | 0.1159 | 1.36 |
| HPC-HLX26-4 | 320.0 | 0.1757 | 2.06 |
| HPC-HLX26-5 | 160.0 | 0.2470 | 2.90 |
| HPC-HLX26-6 | 80.0 | 0.3142 | 3.68 |
| HPC-HLX26-7 | 40.0 | 0.4216 | 4.94 |
| HPC-HLX26-8 | 0.0 | 0.5057 | 5.93 |
| HLX26 ADA (ng/mL) | HLX26 Conc. (μg/mL) | SNR | Drug tolerance (μg/mL) | |
|---|---|---|---|---|
| 100.0 | 640.0 | 320.0 | 1.06 | 291.0 |
| 320.0 | 160.0 | 1.17 | ||
| HPC | 2560.0 | 1280.0 | 0.96 | 768.0 |
| 1280.0 | 640.0 | 1.11 | ||
13. Target interference
The SNRs of HPC and LPC were still higher than SCP, even spiked with up to 640.0 ng/mL of LAG-3 Table 7. It indicates that no false negative results and no interference occurs up to this level of concentration could be concluded. However, the addition of LAG-3 in the NCs could generate a false positive result, indicating interference, and the highest tolerance concentration was 74.0 ng/mL.
Table 7.
Target interference.
| Sample | LAG-3 Conc (ng/mL) | Screening assay |
Confirmatory assay |
||
|---|---|---|---|---|---|
| Mean OD | SNR | Mean OD | Inhibition (%) | ||
| HPC-LAG3-1 | 640.0 | 0.8093 | 10.69 | 0.0942 | 88.36 |
| HPC-LAG3-2 | 320.0 | 0.802 | 10.59 | 0.0935 | 88.34 |
| HPC-LAG3-3 | 160.0 | 0.7904 | 10.44 | 0.091 | 88.49 |
| HPC-LAG3-4 | 80.0 | 0.7647 | 10.1 | 0.0897 | 88.27 |
| HPC-LAG3-5 | 40.0 | 0.7488 | 9.89 | 0.0914 | 87.79 |
| HPC-LAG3-6 | 20.0 | 0.7822 | 10.33 | 0.0929 | 88.12 |
| HPC-LAG3-7 |
0.0 |
0.7618 |
10.06 |
0.0901 |
88.17 |
| LPC-LAG3-1 | 640.0 | 0.1546 | 2.05 | 0.077 | 50.19 |
| LPC-LAG3-2 | 320.0 | 0.1354 | 1.8 | 0.0759 | 43.94 |
| LPC-LAG3-3 | 160.0 | 0.124 | 1.65 | 0.0726 | 41.45 |
| LPC-LAG3-4 | 80.0 | 0.1234 | 1.64 | 0.0727 | 41.09 |
| LPC-LAG3-5 | 40.0 | 0.1216 | 1.61 | 0.0723 | 40.54 |
| LPC-LAG3-6 | 20.0 | 0.1211 | 1.61 | 0.0726 | 40.05 |
| LPC-LAG3-7 |
0.0 |
0.3781 |
5.02 |
0.0758 |
79.95 |
| NC-LAG3-1 | 640.0 | 0.1085 | 1.44 | 0.0764 | 29.59 |
| NC-LAG3-2 | 320.0 | 0.0916 | 1.22 | 0.0763 | 16.7 |
| NC-LAG3-3 | 160.0 | 0.082 | 1.09 | 0.0729 | 11.1 |
| NC-LAG3-4 | 80.0 | 0.0769 | 1.02 | 0.073 | 5.07 |
| NC-LAG3-5 | 40.0 | 0.0757 | 1.01 | 0.072 | 4.89 |
| NC-LAG3-6 | 20.0 | 0.0761 | 1.01 | 0.0781 | −2.63 |
| NC-LAG3-7 | 0.0 | 0.0734 | 0.97 | 0.081 | −10.35 |
| HLX26 ADA (ng/mL) | Screening assay |
Confirmatory assay |
||||||
|---|---|---|---|---|---|---|---|---|
| LAG-3 (ng/mL) | SNR | Interference (ng/mL) | LAG-3 (ng/mL) | Inhibition (%) | Interference (ng/mL) | |||
| 0.0 rowhead | 160.0 | 80.0 | 1.09 | 74.0 | 320.0 | 160.0 | 16.7 | 102.0 |
| 80.0 | 40.0 | 1.02 | 160.0 | 80.0 | 11.1 | |||
14. Matrix effect
The selectivity was validated using ten lots of normal individual human serum diluted with 2000.0 ng/mL of HLX26 ADA (20 × LPC). These spiked samples were assayed together with unspiked samples. The results (Table 8) showed that 100% (10/10) of the spiked samples were positive and 80% (8/10) of the unspiked samples tested negative. Even if 20% of the unspiked samples tested false positive, they still met the preset criteria: at least 80% of individual normal human serum with an SNR value meet the need of LPC ≥ SCP > NC. Similarly, the interference of hemolysis and triglyceride on the assay were also validated. The spiked and unspiked 2% and 5% hemolyzed serum samples, together with the spiked and unspiked 60.0 mg/dL and 150.0 mg/dL lipemic serum samples, were tested. As expected, all of the spiked samples tested positive and all of the unspiked samples tested negative, demonstrating that there was no interference from hemolysis and triglycerides on the assay.
Table 8.
Matrix effect.
| Hemolysis |
Lipemia |
Selectivity |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Samples | Mean OD | SNR | Samples | Mean OD | SNR | Samples | Mean OD | SNR | Samples | Mean OD | SNR |
| Hemo-1-LPC-1 | 0.1236 | 1.63 | Lipe-1-LPC-1 | 0.1288 | 1.66 | Serum1-LPC | 0.101 | 1.18 | Serum1-NC | 0.0773 | 0.91 |
| Hemo-1-LPC-2 | 0.1162 | 1.53 | Lipe-1-LPC-2 | 0.1051 | 1.35 | Serum2-LPC | 0.103 | 1.21 | Serum2-NC | 0.0788 | 0.92 |
| Hemo-1-LPC-3 | 0.1215 | 1.60 | Lipe-1-LPC-3 | 0.105 | 1.35 | Serum3-LPC | 0.1019 | 1.19 | Serum3-NC | 0.0759 | 0.89 |
| Hemo-1-NC-1 | 0.0816 | 1.08 | Lipe-1-NC-1 | 0.0824 | 1.06 | Serum4-LPC | 0.1129 | 1.32 | Serum4-NC | 0.0766 | 0.90 |
| Hemo-1-NC-2 | 0.0822 | 1.08 | Lipe-1-NC-2 | 0.0777 | 1.00 | Serum5-LPC | 0.1012 | 1.19 | Serum5-NC | 0.0761 | 0.89 |
| Hemo-1-NC-3 | 0.0855 | 1.13 | Lipe-1-NC-3 | 0.0773 | 0.99 | Serum6-LPC | 0.1041 | 1.22 | Serum6-NC | 0.081 | 0.95 |
| Hemo-2-LPC-1 | 0.118 | 1.55 | Lipe-2-LPC-1 | 0.1176 | 1.51 | Serum7-LPC | 0.0974 | 1.14 | Serum7-NC | 0.0785 | 0.92 |
| Hemo-2-LPC-2 | 0.1154 | 1.52 | Lipe-2-LPC-2 | 0.106 | 1.36 | Serum8-LPC | 0.1036 | 1.21 | Serum8-NC | 0.0778 | 0.91 |
| Hemo-2-LPC-3 | 0.1154 | 1.52 | Lipe-2-LPC-3 | 0.1177 | 1.51 | Serum9-LPC | 0.1387 | 1.63 | Serum9-NC | 0.1029 | 1.21 |
| Hemo-2-NC-1 | 0.0813 | 1.07 | Lipe-2-NC-1 | 0.0788 | 1.01 | Serum10-LPC | 0.1589 | 1.86 | Serum10-NC | 0.0946 | 1.11 |
| Hemo-2-NC-2 | 0.0788 | 1.04 | Lipe-2-NC-2 | 0.0803 | 1.03 | NA | NA | ||||
| Hemo-2-NC-3 | 0.0757 | 1.00 | Lipe-2-NC-3 | 0.0809 | 1.04 | ||||||
Note: Serum 1–10: Human normal individual sera, Hemo-1: 5% Hemolytic serum, Hemo-2: 2% Hemolytic serum, Lipe-1: 150.0 mg/dL Triglyceride containing serum, Lipe-2: 150.0 mg/dL Triglyceride containing serum.
15. Hook effect
The hook effect refers to the phenomenon of false negatives due to inappropriate antigen-antibody ratios. In this study, HLX26 ADA was prepared for assessment at the starting concentration of 400.0 μg/mL (Hook-1) in 5-fold serial dilutions to approximately 640.0 ng/mL. The results (Table 9) showed that all hook effect samples above the HPC concentration were positive and the mean SNR value of these samples was higher than that of HPC, demonstrating the lack of hook effect.
Table 9.
Hook effect.
| Samples | HLX26 ADA Conc. (ng/mL) | Mean OD | SNR |
|---|---|---|---|
| Hook-1 | 400000.0 | 2.5463 | 32.73 |
| Hook-2 | 80000.0 | 2.5607 | 32.91 |
| Hook-3 | 16000.0 | 2.6552 | 34.13 |
| Hook-4 | 3200.0 | 1.6843 | 21.65 |
| Hook-5 | 640.0 | 0.3890 | 5.00 |
16. Stability
Based on the requirements, the short-term stability under conditions of room temperature (24 h), 2–8 °C (7 d), and six freeze/thaw cycles were investigated. As shown in Table 10, all of the results meet the acceptance criteria described in sections 3.2 and 3.4.
Table 10.
Stability.
| Room temperature | 2–8 °C | Freeze-thaw | ||||||
|---|---|---|---|---|---|---|---|---|
| Sample | Mean OD | SNR | Sample | Mean OD | SNR | Sample | Mean OD | SNR |
| HPC-1 | 0.4276 | 5.43 | HPC-1 | 0.3911 | 4.96 | HPC-1 | 0.4106 | 5.21 |
| HPC-2 | 0.4291 | 5.45 | HPC-2 | 0.5195 | 6.59 | HPC-2 | 0.4310 | 5.47 |
| HPC-3 | 0.4098 | 5.20 | HPC-3 | 0.3977 | 5.05 | HPC-3 | 0.4033 | 5.12 |
| LPC-1 | 0.0990 | 1.26 | LPC-1 | 0.0943 | 1.20 | LPC-1 | 0.0983 | 1.25 |
| LPC-2 | 0.0974 | 1.24 | LPC-2 | 0.0965 | 1.22 | LPC-2 | 0.0959 | 1.22 |
| LPC-3 | 0.0938 | 1.19 | LPC-3 | 0.0930 | 1.18 | LPC-3 | 0.0956 | 1.21 |
| NC-1 | 0.0747 | 0.95 | NC-1 | 0.0758 | 0.96 | NC-1 | 0.0787 | 1.00 |
| NC-2 | 0.0761 | 0.97 | NC-2 | 0.0752 | 0.95 | NC-2 | 0.0768 | 0.97 |
| NC-3 | 0.0751 | 0.95 | NC-3 | 0.0750 | 0.95 | NC-3 | 0.0752 | 0.95 |
17. Discussion and conclusions
Regulatory guidelines and recommendations exist for the validation method for the immunogenicity assay of therapeutic protein products [[23], [24], [25], [26]]. Although the validation strategies and plans are roughly the same, the selection of validation parameters and process details cannot be completely consistent for each method and each product [27]. Instead, this depends on the characteristics of the method and the different properties of the product. In this study, we provided a detailed case about assay validation, including the assay design, the selection of validation parameters, the acceptance criteria, and the interpretation of the results.
First, a three-tiered ADA assay based on the classical bridging immunoassay method [28] was established considering its strong operability and high cost-effectiveness. Before the assay, it is crucial to isolate and enrich ADAs from the serum to be tested as much as possible [14]. The ADAs in serum usually exist in two forms: free ADA and antibody-bound ADA, the latter is most likely to be missed. As documented, acid dissociation is the commonly used strategy for separating drug-ADA [12,14,16,19].
Here, we introduced an ADA purification procedure with two steps of acid-dissociation and accumulation with magnetic beads. In the first step of acid-dissociation, the original ADA-antibody complex was dissociated. The released ADAs and the free ADAs combined with biotin-labeled antibody and SA-labeled magnetic beads form a trimer complex. In the second step of acid-dissociation, the ADAs are released from the complex. The ADAs in both forms were purified and accumulated. Notably, the requirement for incubation time in the second step of acid dissociation is strict (described in Materials and Methods), otherwise, the biotin-labeled antibody is also dissociated from SA-labeled magnetic beads, resulting in drug interference.
Second, in pilot experiments, we found that the target LAG-3 in serum samples could induce false-positive results. In previous studies, solid-phase immunodepletion and in-solution immunocompetition of target protein have been used to remove target protein in serum samples by binding to a polyclonal antibody [[29], [30], [31]]. However, some attempts failed. In this study, we attempted to eliminate target interference by the formation of a target and ligand complex in the solid phase. Fibrinogen-like Protein 1 (FGL-1) is a major immune inhibitory ligand of LAG-3 [32]. In theory, LAG-3 in serum can be removed after binding with FGL-1 coated on an ELISA plate. Indeed, our results demonstrated that the attempt was feasible when the coated ligand FGL-1 was in excess relative to the target protein LAG-3.
Finally, the method was fully validated to meet regulatory requirements (Table 11). Due to the large dose and long half-life of monoclonal antibody drugs, the high concentration of drugs, as well as the soluble target in the body circulation brings great challenge to the testing of immunogenicity, especially anti-drug antibodies [[33], [34], [35]]. This method should first be able to withstand the test of target interference and drug tolerance. As expected, the assay could tolerate up to 768.0 μg/mL and 291.0 μg/mL of anti-LAG-3 mAb at the ADA concentrations of 640.0 ng/mL and 100.0 ng/mL, respectively; this meets the FDA drug tolerance guideline of detecting 100 ng/ml of ADA in the presence of 500-fold excess of free drug and has a wide range of sensitivity. The assay can also afford target interference up to 74.0 ng/mL of LAG-3, which is higher than the reported 5.0 ng/mL of LAG-3 in normal human serum [[36], [37], [38]].
Table 11.
Summary of validation.
| No. | Validation items | Results | ||||
|---|---|---|---|---|---|---|
| 1 | Cut point (screen, confirmation, titer) | SCP: 1.08 | ||||
| CCP: 12.65 | ||||||
| TCP: 1.17 | ||||||
| 2 | Sensitivity, LPC, titer | Sensitivity: 33.0 ng/mL | ||||
| LPC: 41.0 ng/mL | ||||||
| HPC titer acceptance range: 320–1280 | ||||||
| 3 | Establish system suitability control acceptance criteria | |||||
| 4 | Precision | Samples | Mean | STD | RSD% | |
| Intra-plate SNR | HPC | 10.84 | 1.91 | 17.62 | ||
| LPC | 1.53 | 0.15 | 9.80 | |||
| NC | 1.00 | 0.02 | 2.00 | |||
| Inter-plates SNR | HPC | 8.83 | 1.19 | 13.48 | ||
| LPC | 1.58 | 0.17 | 10.76 | |||
| NC | 1.00 | 0.03 | 3.00 | |||
| 5 | Drug tolerance | HPC level: up to 768.0 μg/mL 100.0 ng/mL HLX26 ADA level: up to 291.0 μg/mL |
||||
| 6 | Target interference | HPC & LPC: no false negative detected NC: 74.0 ng/mL (a false positive is detected if LAG-3 > 74.0 ng/mL) |
||||
| 7 | Matrix effect | |||||
| Selectivity | Pass | |||||
| Hemolysis | Pass | |||||
| Lipemia | Pass | |||||
| 8 | Hook effect | Pass | ||||
| 9 | Assay robustness | Pass | ||||
| 10 | Stability | |||||
| Room temperature | Pass (up to 24 h) | |||||
| 2–8 °C | Pass (up to 7 d) | |||||
| Freeze-thaw | Pass (up to 6 cycles) | |||||
Furthermore, the effects of other components (such as hemolysis and hyperlipidemia) in serum samples on the assay results were also evaluated. Besides free drugs, hemolysis and high concentration of triglycerides are also frequently encountered in clinical serum sample collection. Here, we examined both high and low concentrations of hemolysis and triglycerides separately. The results proved that this assay can withstand up to 5% hemolysis and 150.0 mg/dL of triglycerides. Notably, it is possible to add or adjust the concentrations based on the actual situation, scientific justification, and study requirements, as well as specific requirements for the evaluation of stability. Other routine verification items can also meet the requirements of method application and regulations, including the hook effect, short-term stability, robustness, and precision.
In summary, we have established and validated an ADA assay based on bridging immunoassay with two steps of acid dissociation and magnetic bead accumulation to ensure complete detection of ADAs in two forms, as well as the solid-phase immunodepletion procedure for eliminating the target protein interference. Considering the R&D of other anti-LAG-3 biosimilars in the future, the comparable results for ADA formation facilitate rational interpretation of clinical efficacy, which can be used in good routine clinical practice.
However, this method also has some limitations that warrant discussion. The cut points were developed with commercially healthy volunteers as outlined in previous reports [39,40], and whether the same cut points still work for the study population from the clinical trials should be reevaluated [40] according to the above guidance [[23], [24], [25], [26]].
Declarations
Author contribution statement
Jialiang Du, Yalan Yang, Lingling Zhu and Shaoyi Wang: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper. Chuanfei Yu: Conceived and designed the experiments; Wrote the paper. Chunyu Liu and Caifeng Long: Performed the experiments. Baowen Chen: Analyzed and interpreted the data. Gangling Xu: Contributed reagents, materials, analysis tools or data. Linglong Zou and Lan Wang: Conceived and designed the experiments; analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.
Funding statement
Lan Wang was supported by Ministry of Science and Technology of China [2019ZX09732-002].
Yalan Yang was supported by Zhongshan Science and Technology Bureau, China [210204163866513].
Data availability statement
Data included in article/supp. material/referenced in article.
Declaration of interest’s statement
The authors declare no competing interests.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.heliyon.2023.e13999.
Contributor Information
Linglong Zou, Email: linglong_zou@henlius.com.
Lan Wang, Email: wanglan@nifdc.org.cn.
Appendix A. Supplementary data
The following is the Supplementary data to this article.
References
- 1.Lythgoe M.P., Liu D.S.K., Annels N.E., Krell J., Frampton A.E. Gene of the month: lymphocyte-activation gene 3 (LAG-3) J. Clin. Pathol. 2021;74:543–547. doi: 10.1136/jclinpath-2021-207517. [DOI] [PubMed] [Google Scholar]
- 2.Shan C., Li X., Zhang J. Progress of immune checkpoint LAG-3 in immunotherapy. Oncol. Lett. 2020;20:207. doi: 10.3892/ol.2020.12070. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Maruhashi T., Sugiura D., Okazaki I.M., Okazaki T. LAG-3: from molecular functions to clinical applications. J. Immunother. Cancer. 8. 2020 doi: 10.1136/jitc-2020-001014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ruffo E., Wu R.C., Bruno T.C., Workman C.J., Vignali D.A.A. Lymphocyte-activation gene 3 (LAG3): the next immune checkpoint receptor. Semin. Immunol. 2019;42 doi: 10.1016/j.smim.2019.101305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Tawbi H.A., Schadendorf D., Lipson E.J., Ascierto P.A., Matamala L., Castillo G.E., Rutkowski P., Gogas H.J., Lao C.D., De Menezes J.J., Dalle S., Arance A., Grob J.J., Srivastava S., Abaskharoun M., Hamilton M., Keidel S., Simonsen K.L., Sobiesk A.M., Li B., Hodi F.S., Long G.V. RELATIVITY-047 Investigators. Relatlimab and nivolumab versus nivolumab in untreated advanced melanoma. N. Engl. J. Med. 2022;386:24–34. doi: 10.1056/NEJMoa2109970. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Vaisman-Mentesh A., Gutierrez-Gonzalez M., DeKosky B.J., Wine Y. The molecular mechanisms that underlie the immune biology of anti-drug antibody formation following treatment with monoclonal antibodies. Front. Immunol. 2020;11:1951. doi: 10.3389/fimmu.2020.01951. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Baker M.P., Reynolds H.M., Lumicisi B., Bryson C.J. Immunogenicity of protein therapeutics: the key causes, consequences and challenges. Self. Nonself. 2010;1:314–322. doi: 10.4161/self.1.4.13904. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Schellekens H. Immunogenicity of therapeutic proteins: clinical implications and future prospects. Clin. Ther. 2002;24:1720–1740. doi: 10.1016/s0149-2918(02)80075-3. discussion 1719. [DOI] [PubMed] [Google Scholar]
- 9.Nechansky A., Kircheis R. Immunogenicity of therapeutics: a matter of efficacy and safety. Expert Opin. Drug Discov. 2010;5:1067–1079. doi: 10.1517/17460441.2010.514326. [DOI] [PubMed] [Google Scholar]
- 10.Bray-French K., Hartman K., Steiner G., Marban-Doran C., Bessa J., Campbell N., Martin-Facklam M., Stubenrauch K.G., Solier C., Singer T., Ducret A. Managing the Impact of immunogenicity in an Era of immunotherapy: from Bench to Bedside. J. Pharm. Sci. 2021;110:2575–2584. doi: 10.1016/j.xphs.2021.03.027. [DOI] [PubMed] [Google Scholar]
- 11.Ducret A., Ackaert C., Bessa J., Bunce C., Hickling T., Jawa V., Kroenke M.A., Lamberth K., Manin A., Penny H.L., Smith N., Terszowski G., Tourdot S., Spindeldreher S. Assay format diversity in pre-clinical immunogenicity risk assessment: toward a possible harmonization of antigenicity assays. mAbs. 2022;14 doi: 10.1080/19420862.2021.1993522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hoffmann E., Jordan G., Lauer M., Ringler P., Kusznir E.A., Rufer A.C., Huber S., Jochner A., Winter G., Staack R.F. Generation, Characterization, and Quantitative Bioanalysis of drug/anti-drug antibody immune Complexes to facilitate Dedicated in Vivo studies. Pharm. Res. (N. Y.) 2019;36:129. doi: 10.1007/s11095-019-2661-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Jani D., Marsden R., Gunsior M., Hay L.S., Ward B., Cowan K.J., Azadeh M., Barker B., Cao L., Closson K.R., Coble K., Dholakiya S.L., Dusseault J., Hays A., Herl C., Hodsdon M.E., Irvin S.C., Kirshner S., Kolaitis G., Kulagina N.…Zhong Z.D. Anti-drug antibody sample testing and reporting harmonization. AAPS J. 2022;24:113. doi: 10.1208/s12248-022-00762-6. [DOI] [PubMed] [Google Scholar]
- 14.Kavita U., Duo J., Crawford S.M., Liu R., Valcin J., Gleason C., Dong H., Gadkari S., Dodge R.W., Pillutla R.C., DeSilva B.S. A systematic study of the effect of low pH acid treatment on anti-drug antibodies specific for a domain antibody therapeutic: Impact on drug tolerance, assay sensitivity and post-validation method assessment of ADA in clinical serum samples. J. Immunol. Methods. 2017;448:91–104. doi: 10.1016/j.jim.2017.06.002. [DOI] [PubMed] [Google Scholar]
- 15.Myler H., Pedras-Vasconcelos J., Phillips K., Hottenstein C.S., Chamberlain P., Devanaryan V., Gleason C., Goodman J., Manning M.S., Purushothama S., Richards S., Shen H., Zoghbi J., Amaravadi L., Barger T., Bowen S., Bowsher R.R., Clements-Egan A., Geng D., Goletz T.J., Gunn G.R., Hallett W., Hodsdon M.E., Janelsins B.M., Jawa V., Kamondi S., Kirshner S., Kramer D., Liang M., Lindley K., Liu S., Liu Z., McNally J., Mikulskis A., Nelson R., Ahbari M.R., Qu Q., Ruppel J., Snoeck V., Song A., Yan H., Ware M. Anti-drug antibody validation testing and reporting harmonization. AAPS J. 2021:4. doi: 10.1208/s12248-021-00649-y. 24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Niu H., Klem T., Yang J., Qiu Y., Pan L. A biotin-drug extraction and acid dissociation (BEAD) procedure to eliminate matrix and drug interference in a protein complex anti-drug antibody (ADA) isotype specific assay. J. Immunol. Methods. 2017;446:30–36. doi: 10.1016/j.jim.2017.04.002. [DOI] [PubMed] [Google Scholar]
- 17.Schlain B., Amaravadi L., Donley J., Wickramasekera A., Bennett D., Subramanyam M. A novel gamma-fitting statistical method for anti-drug antibody assays to establish assay cut points for data with non-normal distribution. J. Immunol. Methods. 2010;352:161–168. doi: 10.1016/j.jim.2009.10.012. [DOI] [PubMed] [Google Scholar]
- 18.Shen M., Dong X., Tsong Y. Statistical evaluation of several methods for cut-point determination of immunogenicity screening assay. J. Biopharm. Stat. 2015;25:269–279. doi: 10.1080/10543406.2014.979196. [DOI] [PubMed] [Google Scholar]
- 19.Yuan L., Gleason C.R., Stocker D., Li L., Shen J.X., Ji Q.C. A bridging immunogenicity assay for anti-cabiralizumab antibodies: overcoming the low assay cut point and drug tolerance challenges. Bioanal. 2021;13:395–407. doi: 10.4155/bio-2020-0300. [DOI] [PubMed] [Google Scholar]
- 20.Zhang J., Yu B., Zhang L., Roskos L., Richman L., Yang H. Non-normal random effects models for immunogenicity assay cut point determination. J. Biopharm. Stat. 2015;25:295–306. doi: 10.1080/10543406.2014.972515. [DOI] [PubMed] [Google Scholar]
- 21.Zhang J., Li W., Roskos L.K., Yang H. Immunogenicity assay cut point determination using nonparametric tolerance limit. J. Immunol. Methods. 2017;442:29–34. doi: 10.1016/j.jim.2017.01.001. [DOI] [PubMed] [Google Scholar]
- 22.Zhang J., Arends R.H., Kubiak R.J., Roskos L.K., Liang M., Lee N., Chen C.C., Yang H. A new method for identification of outliers in immunogenicity assay cut point data. J. Immunol. Methods. 2020;112817:484–485. doi: 10.1016/j.jim.2020.112817. [DOI] [PubMed] [Google Scholar]
- 23.FDA . 2019. Guidance on immunogenicity testing of therapeutic protein products – Developing and validating assays for anti-drug antibody detection; p. 1.https://www.fda.gov/ucm/groups/fdagovpublic/@fdagov-drugsen/documents/document/ucm629728.pdf [Google Scholar]
- 24.Ema . 2017. Guideline on Immunogenicity Assessment of Therapeutic Proteins (Rev1)www.ema.europa.eu/docs/en GB/documentlibrary/Scientificguideline/2017/06/WC500228861.pdf [Google Scholar]
- 25.ICH M10 Guideline on bioanalytical method validation and study sample analysis. https://database.ich.org/sites/default/files/M10_Guideline_Step4_2022_0524.pdf
- 26.CDE https://www.cde.org.cn/main/news/viewInfoCommon/a0908879d6c54c7318f0881611b51122 Technical guidelines for drug immunogenicity research (Mar 5, 2021)
- 27.Ruwaard J., Marsman A.F., Nurmohamed M.T., van der Horst-Bruinsma I.E., Te Velthuis H., Bloem K., de Vries A., Rispens T., Wolbink G. Antidrug antibody detection for adalimumab depends on the type of assay used: an experimental approach to help clinicians interpret diagnostic data. Clin. Exp. Rheumatol. 2019;37:756–761. [PubMed] [Google Scholar]
- 28.Imbrechts M., Van Stappen T., Compernolle G., Tops S., Gils A. Anti-infliximab antibodies: How to compare old and new data? J. Pharm. Biomed. Anal. 2020;177 doi: 10.1016/j.jpba.2019.112842. [DOI] [PubMed] [Google Scholar]
- 29.Dai S., Schantz A., Clements-Egan A., Cannon M., Shankar G. Development of a method that eliminates false-positive results due to nerve growth factor interference in the assessment of fulranumab immunogenicity. AAPS J. 2014;16:464–477. doi: 10.1208/s12248-014-9581-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Liao K., Meyer E., Lee T.N., Loercher A., Sikkema D. Inhibition of interleukin-5 induced false positive anti-drug antibody responses against mepolizumab through the use of a competitive blocking antibody. J. Immunol. Methods. 2017;441:15–23. doi: 10.1016/j.jim.2016.11.010. [DOI] [PubMed] [Google Scholar]
- 31.Wang Y., Luong M., Guadiz C., Zhang M., Gorovits B. Addressing soluble target interference in the development of a functional assay for the detection of neutralizing antibodies against a BCMA-CD3 bispecific antibody. J. Immunol. Methods. 2019;474 doi: 10.1016/j.jim.2019.112642. [DOI] [PubMed] [Google Scholar]
- 32.Shi A.P., Tang X.Y., Xiong Y.L., Zheng K.F., Liu Y.J., Shi X.G., Lv Y., Jiang T., Ma N., Zhao J.B. Immune checkpoint LAG3 and its ligand FGL1 in cancer. Front. Immunol. 2022;12 doi: 10.3389/fimmu.2021.785091. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Zoghbi J., Xu Y., Grabert R., Theobald V., Richards S. A breakthrough novel method to resolve the drug and target interference problem in immunogenicity assays. J. Immunol. Methods. 2015;426:62–69. doi: 10.1016/j.jim.2015.08.002. [DOI] [PubMed] [Google Scholar]
- 34.Chen J., Kendra K., Torri A., Sumner G. Overcoming multimeric target interference in a bridging immunogenicity assay with soluble target receptor, target immunodepletion and mild acidic assay pH. Bioanalysis. 12. 2020:1071–1085. doi: 10.4155/bio-2020-0110. [DOI] [PubMed] [Google Scholar]
- 35.Zhong Z.D., Clements-Egan A., Gorovits B., Maia M., Sumner G., Theobald V., Wu Y., Rajadhyaksha M. Drug target interference in immunogenicity assays: recommendations and mitigation mtrategies. AAPS J. 2017;19:1564–1575. doi: 10.1208/s12248-017-0148-7. [DOI] [PubMed] [Google Scholar]
- 36.Yoshida J., Ishikawa T., Doi T., Ota T., Yasuda T., Okayama T., Sakamoto N., Inoue K., Dohi O., Yoshida N., Kamada K., Uchiyama K., Takagi T., Konishi H., Naito Y., Itoh Y. Clinical significance of soluble forms of immune checkpoint molecules in advanced esophageal cancer. Med. Oncol. 2019;36:60. doi: 10.1007/s12032-019-1285-x. [DOI] [PubMed] [Google Scholar]
- 37.Shapiro M., Herishanu Y., Katz B.Z., Dezorella N., Sun C., Kay S., Polliack A., Avivi I., Wiestner A., Perry C. Lymphocyte activation gene 3: a novel therapeutic target in chronic lymphocytic leukemia. Haematologica. 2017;102:874–882. doi: 10.3324/haematol.2016.148965. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Guo W., Zhou M., Qiu J., Lin Y., Chen X., Huang S., Mo M., Liu H., Peng G., Zhu X., Xu P. Association of LAG3 genetic variation with an increased risk of PD in Chinese female population. J. Neuroinflammation. 2019;16:270. doi: 10.1186/s12974-019-1654-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Devanarayan V., Smith W.C., Brunelle R.L., Seger M.E., Krug K., Bowsher R.R. Recommendations for systematic statistical computation of immunogenicity cut points. AAPS J. 2017;19:1487–1498. doi: 10.1208/s12248-017-0107-3. [DOI] [PubMed] [Google Scholar]
- 40.Tan C.Y., Steeno G.S., You Z., Gaitonde P., Cai C.H., Kamerud J., Gorovits B., Baltrukonis D.J. Criteria to reevaluate anti-drug antibody assay cut point suitability in the target population. AAPS. J. 2020:19. doi: 10.1208/s12248-019-0400-4. 22. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data included in article/supp. material/referenced in article.