ORIGINAL ARTICLE
Generation of a Highly Specific Monoclonal Anti-Infliximab Antibody for Harmonization of TNF-Coated Infliximab Assays Thomas Van Stappen, PharmD,* Els Brouwers, BSc,* Sophie Tops, BSc,* Nick Geukens, PhD,† Séverine Vermeire, MD, PhD,‡ Paul J. Declerck, PharmD, PhD,* and Ann Gils, PharmD, PhD*
Background: Determination of infliximab (IFX) serum concentrations has been used for treatment optimization of patients with inflammatory bowel disease. A wide range of enzyme-linked immunosorbent assays (ELISA) exists to quantitate IFX. Most of these assays lack specificity and cross-react with other anti-tumor necrosis factor (TNF) agents. The ability of these IFX assays to detect IFX in complex with antidrug antibodies is not known. The objective of our study was to develop an IFX-specific immunoassay to monitor IFX serum concentrations and to evaluate the impact of antidrug antibodies on the assay performance.
Methods: A panel of monoclonal antibodies toward IFX (MA-IFX) was generated by hybridoma technology and evaluated to replace the polyclonal antibody in a TNF-coated IFX assay. The selected monoclonal antibody–based (MA-based) IFX ELISA was benchmarked to a clinically validated, reference polyclonal antibody– based (pAb-based) IFX ELISA using 209 inflammatory bowel disease serum samples. Results: Fifty-five MA-IFX were generated and grouped into 9 clusters. Of the 22 monoclonal antibodies tested, MA-IFX6B7 was selected for use in the IFX ELISA and the assay was further optimized. MA-IFX6B7 is a high-affinity (KD = 1.40E-09 mol/L), noninhibitory IgG1 antibody that binds to the Fab fragment of IFX and exhibits no cross-reactivity with other anti-TNF drugs. The linearity of an IFX dose–response curve was demonstrated in the range of 1.2–37.5 ng/mL (R2 = 0.988). The MA-based assay showed a good Pearson correlation (R = 0.986) and agreement (intraclass correlation Received for publication May 1, 2014; accepted October 28, 2014. From the *Laboratory for Therapeutic and Diagnostic Antibodies, Department of Pharmaceutical and Pharmacological Sciences; †PharmAbs, The KU Leuven Antibody Center, KU Leuven; and ‡Department of Gastroenterology, UZ Leuven, Belgium. Supported in part by a grant G.0617.12 from the Fund for Scientific Research Flanders (FWO, Flanders) and grant IOFKP/12/002 from the KU Leuven. T. Van Stappen is a PhD fellow of the Agency for the Promotion and Innovation through Science and Technology in Flanders (IWT, Flanders). A. Gils has served as a speaker for MSD and Janssen Biologicals, as an advisory board member for Pfizer, and has received an Investigator Initiated Research Grant from Pfizer. S. Vermeire received grant support from Centocor, MSD, Abbvie, and UCB Pharma. The other authors declare no conflict of interest. Correspondence: Ann Gils, PharmD, PhD, Laboratory for Therapeutic and Diagnostic Antibodies, Department of Pharmaceutical and Pharmacological Sciences, KU Leuven, O&N II Herestraat 49, Box 820, B-3000 Leuven, Belgium (e-mail: [email protected]). Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Ther Drug Monit Volume 37, Number 4, August 2015
coefficient = 0.985) with the pAb-based assay. The MA-based assay detects IFX in complex with nonneutralizing anti-IFX antibodies but not when complexed with neutralizing anti-IFX antibodies.
Conclusions: In this study, a highly specific MA-IFX was developed as detection antibody in an ELISA to quantify IFX serum concentrations. The assay was benchmarked to the clinically validated reference pAb-based IFX ELISA. Key Words: infliximab, tumor necrosis factor-alpha, inflammatory bowel diseases, enzyme-linked immunosorbent assay, antibodies, monoclonal (Ther Drug Monit 2015;37:479–485)
INTRODUCTION Inflammatory bowel diseases (IBDs) are chronic autoimmune disorders of the gastrointestinal tract with a relapsing-remitting character. The treatment of IBD varies with severity of diseases and may include 5-aminosalicylic acid derivatives, corticosteroids, immunomodulatory drugs, and more recently, biologicals.1 Infliximab (IFX) and adalimumab (ADM), a chimeric and fully human monoclonal antibody, respectively, toward tumor necrosis factor-alpha (TNF), have proven to be efficacious in inducing and maintaining remission in patients with moderate/severe (fistulizing) Crohn disease and ulcerative colitis.1 However, not all patients benefit equally from these drugs, and 30%–40% of patients lose response within the first year of treatment.2,3 Reasons for this secondary loss of response are predominantly the development of neutralizing antidrug antibodies and increased clearance of anti-TNF–antibody immune complexes.4–6 Previous studies have demonstrated a positive correlation between the IFX trough level (the serum level just before the next infusion) and the clinical response of patients with Crohn disease or ulcerative colitis.7–9 Optimal trough levels are correlated with a long-term response to IFX, whereas patients with low trough levels have more chance to lose response over time. Diverse assay types exist to measure IFX and antidrug antibody serum concentrations [Radio-immunoassay, enzyme-linked immunosorbent assay (ELISA), bridging ELISA, mobility shift assay], each having their own merits and limitations. The most commonly used assays are ELISA techniques. A round robin experiment, performed in the Laboratory for Therapeutic and Diagnostic Antibodies of
479
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Van Stappen et al
KU Leuven (Leuven, Belgium), Sanquin (Amsterdam, the Netherlands), and the University Medical Center Groningen (Groningen, The Netherlands), revealed that a commercially available ELISA (LISA Tracker; Theradiag, Marne La Vallée, France) gives response in samples that contain antibodies toward IFX.10 As a consequence, patients failing IFX therapy may be missed and incorrectly continued on IFX therapy, possibly leading to hypersensitivity reactions and infusion reactions. Additionally, the lack of agreement between all available methods hampers the implementation of universal guidelines for therapeutic drug monitoring of IFX and underlines the need for standardization. In the “first-generation” polyclonal antibody–based (pAb-based) IFX ELISA, TNF was used as capture protein and an IFX monospecific polyclonal antibody preparation conjugated to horseradish peroxidase (pAb-IFX-HRP) as detection antibody. However, the pAb-IFX-HRP preparation can crossreact with other anti-TNF antibodies (ADM, golimumab) and is subject to batch-to-batch variation because of its polyclonal nature, requiring optimization each time a new batch has been produced. The use of a monoclonal antibody toward IFX (MAIFX), being highly specific for its target of interest and highly reproducible, can overcome these limitations. In this perspective, the objective of this study was to develop and validate an IFX-specific immunoassay to monitor IFX serum concentrations and to evaluate the impact of both neutralizing and nonneutralizing antidrug antibodies on the assay performance.
MATERIALS AND METHODS Bovine serum albumin (BSA), horseradish peroxidase (HRP), polyoxyethylene (20) sorbitan monooleate, sulfuric acid (H2SO4), mannitol, sucrose, fluorodinitrobenzene, and citric acid were purchased from Sigma-Aldrich (Steinheim, Germany). VWR International (Haasrode, Belgium) delivered disodium hydrogen phosphate dihydrate (Na2HPO4$2H2O) and sodium chloride (NaCl). Potassium dihydrogen phosphate (KH2PO4) was purchased from Fisher Scientific (Loughborough, United Kingdom). H2O2 and sodium metaperiodate were purchased from Merck (Darmstadt, Germany) and o-phenylenediamine, ethylene glycol, and potassium chloride (KCl) from Acros Organics (Geel, Belgium). TNF was purchased from PeproTech (London, United Kingdom). Complete and incomplete Freund adjuvant was purchased from BD Biosciences (Franklin Lakes, NJ). Dulbecco’s Modified Eagle’s Medium, glutamine, penicillin–streptomycin, and fetal bovine serum were purchased from Invitrogen Life Technologies Europe B.V. (Gent, Belgium). IFX and golimumab were purchased as Remicade and Simponi, respectively, from Janssen Biologics B.V. (Leiden, the Netherlands), ADM as Humira from AbbVie Inc (North Chicago, IL), etanercept as Enbrel from Pfizer Inc (New York, NY), and a human IgG-mixture as Multigam from Biotest AG (Dreieich, Germany).
Ther Drug Monit Volume 37, Number 4, August 2015
Crohn and colitis (B322201213950/S53684). Informed consent was provided by all patients. Generation of monoclonal antibodies in SJL/J mice (Charles River Laboratories, Wilmington, MA) was performed according to local ethical laws and approved by the local ethical committee (P088/2012 KU Leuven, Leuven, Belgium).
Monoclonal Antibody Generation Monoclonal antibodies were generated in-house using the method of Galfré and Milstein.11 Briefly, SJL/J mice were immunized by subcutaneous injection of 10 mcg IFX in complete Freund adjuvant, followed 2 weeks later by intraperitoneal injection of 10 mcg IFX in incomplete Freund adjuvant. Minimum 6 weeks later, the mice were boosted intraperitoneally with 10 mcg IFX in saline, 4 and 2 days before the cell fusion. The spleen was then isolated, spleen cells were fused with Sp2/0-Ag14 myeloma cells (ATCC, Manassas, VA) and plated onto 96-well cell-culture plates (Falcon; BD Biosciences). After selection in hypoxanthine–aminopterin–thymidine medium, supernatants were screened for the presence of antibodies against IFX by ELISA, using IFX for capture and HRP-conjugated rabbit anti-mouse antibody for detection. Subsequently, positive clones were selected and grown in Integra CELLine systems (Integra Biosciences AG, Zizers, Switzerland) for production of monoclonal antibody. Monoclonal antibodies were purified on a ProSep-vA Ultra column (Merck Millipore, Darmstadt, Germany). HRP-conjugated monoclonal antibodies were produced as described by Nakane and Kawaoi.12 Briefly, HRP was dissolved in 0.3 mol/L sodium bicarbonate buffer pH 8.1 in a brown glass, and slowly 1% fluorodinitrobenzene in ethanol was added. Then, the mixture was gently mixed for 1 hour at room temperature on a rotator. Subsequently, 8E-02 mol/L sodium metaperiodate was added and allowed to mix for 30 minutes, after which 0.16 mol/L ethylene glycol was added. Next, the (freshly made) “peroxidase-aldehyde” solution was mixed for 1 hour and dialyzed overnight at 48C against conjugate buffer, consisting of 1E-02 mol/L NaHCO3/Na2CO3 pH 9.5. Simultaneously, the monoclonal antibody was dialyzed overnight against conjugate buffer. The next day, the antibody was concentrated using Centricon (30 kDa) centrifugal filters (Merck Millipore) to a final concentration of 20–40 mg/mL and dissolved in the peroxidase-aldehyde solution. The solution was mixed gently for 2–3 hours at room temperature and dialyzed overnight against phosphate-buffered saline (PBS; 140 mmol/L NaCl, 8 mmol/L Na2HPO4$2H2O, 2.7 mmol/L KCl, and 1.5 mmol/L KH2PO4). Finally, the conjugate was sterilized through filtration and stored at 48C.
Evaluation of the Cross-Reactivity of MA-IFX Toward Other Anti-TNF Drugs
This study was approved by the Ethical Committee UZ Leuven in the framework of the Flemish inheritance study for
The cross-reactivity toward other anti-TNF drugs was determined by a sandwich-type ELISA. Briefly, IFX, the Fab fragment of IFX (Fab-IFX), ADM, etanercept, golimumab, and a human IgG-mixture were coated on 96-well plates (Costar; Corning Inc, NY) at a concentration of 4 mcg/mL in PBS. Next, MA-IFX were applied on the plate at 3 different concentrations (1–10–100 ng/mL). Subsequently, plates were
480
Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
Ethical Considerations
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Ther Drug Monit Volume 37, Number 4, August 2015
incubated overnight at 48C, and HRP-conjugated rabbit antimouse antibody (1:10.000 diluted) was used to detect bound MA-IFX. MA-IFX revealing reactivity toward any of the other anti-TNF drugs and/or the human IgG mixture, were considered as cross-reacting antibodies.
Evaluation of the IFX-Neutralizing Activity of MA-IFX The neutralizing activity was determined in a cellbased assay using a human fibrosarcoma cell line (HT1080; ATCC, CCL-121), characterized by a TNF-inducible interleukin-6 (IL-6) expression, according to a procedure described by Gils et al.13 Briefly, HT1080 cells were plated in a 96-well plate at a density of 4E4 cells in Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal bovine serum, 2 mmol/L glutamine, 50 mcg/mL streptomycin, and 50 U/mL penicillin per well and incubated 24 hours at 5% CO2 and 378C humidified atmosphere. Next, Dulbecco’s Modified Eagle’s Medium was replaced by varying concentrations of TNF (0–240 ng/mL) to determine the EC50 and in a subsequent experiment by a fixed concentration of TNF (at a concentration equal to EC50) with varying concentrations of IFX (0–120 ng/mL) to determine the IC80. After 48 hours of incubation, the amount of IL-6 was measured with the “Human IL-6 ELISA MAX Deluxe Kit” (BioLegend, San Diego, CA), according to the manufacturer’s protocol. In the subsequent experiments, a 10-fold excess of MA-IFX over IFX was added to a mixture of TNF (at a concentration equal to EC50) and IFX (at a concentration equal to IC80) and the effect of MA-IFX on IL-6 expression determined. An inhibitory antibody can upregulate IL-6 expression by binding to (and neutralizing) IFX, and thus blocking the inhibitory effect of IFX on TNF. A reference sample, TNF at a concentration equal to EC50 in the absence of IFX and MA-IFX, was set at 100%, to which the IL-6 secretion was compared. Minimum 80% of the IL-6 expression of the reference sample was required to consider a MA-IFX as neutralizing.
Improving the Specificity of Infliximab Assays
Development of the IFX ELISA Selected MA-IFX were subsequently evaluated to replace the polyclonal antibody in the TNF-coated IFX assay. Briefly, in the in-house developed IFX ELISA, 96-well plates were coated overnight with TNF at 48C. Plates were blocked with PBS 1% BSA for 2 hours at room temperature. Samples were diluted 1:150 and 1:300 in PBS 1% BSA, applied on the plate, and incubated for 2 hours at 378C on a shaker. Then, plates were washed and pAb-IFX-HRP or HRP-conjugated MA-IFX was applied for detection of bound IFX and incubated 1 hour at room temperature. Plates were washed and developed using o-phenylenediamine and H2O2 in citrate buffer, and the reaction was stopped with 2 mol/L H2SO4. The absorbance was measured at 490 nm with an ELx808 Absorbance Microplate Reader (BioTek Instruments Inc, Winooski, VT) and the IFX dose–response curve was analyzed by linear and nonlinear regression for the monoclonal antibody–based (MA-based) and pAb-based IFX assay, respectively, with GraphPad Prism 5.0 (GraphPad Software, San Diego, CA).
Determination of the Isotype of MA-IFX6B7
The antibody’s isotype was determined by the “IsoStrip Mouse Monoclonal Antibody Isotyping kit” (Roche Diagnostics, Indianapolis, IN), according to the manufacturer’s protocol.
Determination of the Affinity of MA-IFX6B7 Surface plasmon resonance-based affinity measurements were carried out using a BIAcore 3000 (GE Healthcare Life Sciences, Uppsala, Sweden). CM5 sensor chips were immobilized to 660 resonance units with IFX using 1-ethyl3-(3-dimethylaminopropyl)-carbodiimide hydrochloride/ N-hydroxysuccinimide amine coupling chemistry. Then, MA-IFX6B7 was applied at 3 different concentrations at a flow rate of 30 mL/min. Data were fitted using the 1:1 Langmuir binding global fit isotherm using BIAevaluation software 4.1 (GE Healthcare Life Sciences).
Blood Samples Epitope Binning of MA-IFX Epitope binning was carried out using competition ELISA. Briefly, 4 mcg/mL IFX was coated on the plate, incubated for 72 hours at 48C, and blocked with PBS 1% BSA. Then, MA-IFX were applied at a concentration of 10 mcg/mL and incubated overnight at 48C (“cold” antibody). Next, plates were emptied and HRP-conjugated MA-IFX (“labeled” antibody), 1:64,000 diluted (ie, at a concentration of 25 ng/mL), was applied followed by incubation for 2 hours at room temperature. Then, plates were washed and developed using 400 mcg/mL o-phenylenediamine and 3E-03% H2O2 in 0.1 mol/L sodium citrate, 0.2 mol/L disodium phosphate buffer pH 5 (citrate buffer), and the reaction stopped with 4 mol/L H2SO4. The optical density (OD) obtained for the labeled antibody in the presence and absence of cold antibody was compared and cold antibodies of which the ODþcold antibody OD2cold antibody · 100% gave a percentage lower than 20% were considered competitive with the labeled antibody. Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
Blood samples of patients with IBD were drawn at the Infusion Unit of the Division of Gastroenterology, UZ Leuven, using the “BD Vacutainer SST II Advance tube” (BD Biosciences), containing a gel separator and clot activator. After an incubation time of 30 minutes, serum was prepared by centrifugation (10 minutes, 1960g, RT) and stored at 48C (maximum 1 week) or 2208C (several years).
Validation of the IFX ELISA Lower Limit of Quantification, Accuracy, and Imprecision The assay cutoff value was calculated using 45 antiTNF naive serum samples, measured twice independently. The lower limit of quantification (LLOQ) was defined as the lowest concentration that could be accurately measured with a coefficient of variation (CV) #25%. The accuracy and imprecision were determined by preparing 5 quality control samples in PBS 1% BSA (0.5–1–3–7.5–11.25 mcg/mL). The
481
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Ther Drug Monit Volume 37, Number 4, August 2015
Van Stappen et al
quality control samples were diluted 1:150 and 1:300, and 8 replicates were applied per plate. The IFX dose–response curve was calculated using linear regression (GraphPad Prism 5.0), and the concentrations of the quality control samples were interpolated from the dose–response curve. The experiment was performed 3 times independently by 2 different operators.
Reactivity of the ELISA Toward Adalimumab The cross-reactivity of the ELISA toward ADM was evaluated using 3 different concentrations of ADM (3–7– 11.25 mcg/mL) and compared with the ADM assay9 (TNF-coated plate in which ADM bound to TNF is detected by a monospecific polyclonal antibody toward ADM). In addition, 20 patient samples who received ADM treatment were analyzed with the MA-based IFX assay and compared with the results obtained with the ADM ELISA.
Comparative Evaluation of IFX Values The monoclonal antibody–based ELISA was benchmarked to the clinically validated polyclonal antibody–based ELISA.10 In total, 209 IBD serum samples were analyzed in parallel by the MA-based and pAb-based IFX assay. The Pearson correlation and the relative Bland–Altman comparison14 were calculated using GraphPad Prism 5.0 and the intraclass correlation coefficient (ICC) using the “2-way mixed single measure test” from SPSS Statistics version 22 (IBM, New York, NY).
Impact of Neutralizing and Nonneutralizing Anti-IFX Antibodies Six monoclonal antibodies with different properties were selected to evaluate the ability of the IFX ELISA to detect IFX in the presence of other monoclonal antibodies. MA-IFX13E7 and MA-IFX4B6 do not inhibit IFX and do not bind to Fab-IFX, while cross-reacting with other anti-TNF drugs (nonneutralizing anti-IFX antibodies). MA-IFX20G2, MA-IFX3D5, MA-IFX10F9 do inhibit IFX and bind to FabIFX, while not cross-reacting with other anti-TNF drugs (neutralizing anti-IFX antibodies). MA-33H1F715 is a monoclonal antibody raised toward plasminogen activator inhibitor-1 and does not bind to IFX. IFX was spiked to buffer in a concentration of 3 mcg/mL in the absence (-MA) and presence of a 6-mcg/mL monoclonal antibody. Samples were preincubated for 10 minutes at 378C on a shaker, diluted 1:150 and 1:300, and applied on the IFX ELISA. In addition, 41 serum samples with antibodies toward IFX (detected with a bridging ELISA10) were analyzed in parallel on the MA-based and pAb-based IFX assay.
RESULTS
covering all clusters were evaluated for replacement of pAb-IFX-HRP in the IFX assay.
Selection of MA-IFX6B7 for Use in the IFX ELISA The ability of 22 monoclonal antibodies to detect IFX in spiked buffers and serum samples in the IFX ELISA setup was evaluated. Based on affinity, nonneutralizing properties, high specificity for IFX, and minimal serum influence, MA-IFX6B7 was selected for development of the assay. MA-IFX6B7 is a high-affinity (KD = 1.40E-09 mol/L), noninhibitory IgG1 antibody with k light chains. MA-IFX6B7 does not cross-react with other anti-TNF drugs (ADM, etanercept, golimumab) and binds the Fab fragment of IFX.
Lower Limit of Quantification, Accuracy, and Imprecision The use of MA-IFX6B7-HRP in the TNF-coated IFX assay yielded a linearity for the IFX dose–response curve in the range of 1.2–37.5 ng/mL (R2 = 0.988). Taking the 1:150 serum dilution into account, the limit of detection (LOD) was calculated to be 0.2 mcg/mL. The cutoff value of the assay was determined based on the mean OD obtained with 45 antiTNF naive serum samples plus 3 SDs compared with the mean OD of the IFX concentration at 1.2 ng/mL. The mean OD of the samples was 7.3E-03 with a standard deviation of 4.1E-03, giving a value of 2.0E-02, which was lower than the mean OD (6.0E-02) of 1.2 ng/mL IFX. Therefore, the cutoff for IFX determination was set at 0.2 mcg/mL. The assay complied with the requirements for accuracy: (1) the mean concentration was within 20% of the nominal value at each concentration level (and lower than 25% of the nominal value at the LLOQ/upper limit of quantification), (2) the imprecision (CV) was lower than 20% for all quality control samples (and lower than 25% for the LLOQ/upper limit of quantification) (Table 1). As 0.5 mcg/mL was still accurately detected (CV = 21%), the LLOQ was set at 0.5 mcg/mL.
Evaluation of Reactivity Toward ADM and Application of the ELISA on Samples From Patients Treated With ADM Applying different concentrations of ADM (3–7–11.25 mcg/mL) to the assay revealed no value above the LOD (Fig. 1). In 20 patient samples with ADM values ranging from 1 to 15 mcg/mL, no apparent IFX levels were detected (all
TABLE 1. Accuracy and Imprecision of 5 Quality Control Samples Within the Range of the Assay 0.5 mcg/mL, %
1 mcg/mL, %
3 mcg/mL, %
7.5 mcg/mL, %
11.25 mcg/mL, %
91 21
85 15
96 13
107 7
106 7
A panel of 55 anti-IFX monoclonal antibodies was evaluated: (1) for differential cross-reactivity with the Fab fragment of IFX, ADM, etanercept, golimumab, and a human IgG preparation, (2) for their ability to inhibit the TNF-IFX interaction, and (3) through epitope binning. Based on this methodology, the anti-IFX antibodies were grouped into 9 clusters. Subsequently, 22 HRP-conjugated MA-IFX
Accuracy Imprecision (CV)
482
Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
QC Samples
QC, quality control.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Ther Drug Monit Volume 37, Number 4, August 2015
FIGURE 1. Detection of ADM in the pAb-based ADM ELISA (empty bar) compared with the MA-based IFX ELISA (filled bar). ADM, tested in 3 different concentrations, was not detected in the MA-based IFX ELISA. Four patients (patients 1–4) on ADM therapy with detectable ADM levels did not reveal detectable values in the MA-based IFX ELISA.
values below LOD). Data of 4 representative samples are presented in Figure 1.
Comparative Evaluation of IFX Values Comparison of the IFX values (n = 209) obtained with the pAb-based IFX ELISA and with the currently developed MA-based IFX ELISA revealed a correlation (R) of 0.986 (Fig. 2). The ICC was 0.985. Results were graphically illustrated with a Bland–Altman plot, showing the relative difference between the 2 measurements to the average measured value (Fig. 3). The 95% limits of agreement were 226% and 30%. Based on these criteria, the methods can be considered interchangeable. One sample (Figs. 2, 3: open dot) revealed a much higher signal in the pAb-based IFX assay compared with that in the MA-based IFX assay. Further analysis showed that this
Improving the Specificity of Infliximab Assays
FIGURE 3. The Bland–Altman comparison of IFX values determined with the MA-based and pAb-based IFX ELISA. The relative (%) difference between IFX values determined with the MA-based and pAb-based IFX ELISA is plotted against the average measured IFX value (mcg/mL). The dotted lines represent the upper (30%) and lower (226%) 95% limits of agreement. One sample (open dot) contained ADM from previous ADM treatment and was therefore overestimated in the pAb-based IFX assay.
patient had received ADM 3 days before IFX infusion. The discrepancy between both assays for this sample is due to the ADM present in the sample. Subsequently, this value was excluded from the Pearson correlation, ICC, and 95% limits of agreement.
Impact of Neutralizing and Nonneutralizing Anti-IFX Antibodies Buffer was spiked with IFX (3 mcg/mL) in the absence (-MA) and presence of a 2-fold molar excess of antibody over IFX. On addition of a monoclonal antibody not binding to IFX (MA-33H1F7) to the spiked IFX sample, full recovery of IFX was achieved (97%). Addition of nonneutralizing antiIFX antibodies, MA-IFX4B6 and MA-IFX13E7, resulted in a detection of 78%–87% of the spiked IFX. However, when neutralizing anti-IFX antibodies, MA-IFX20G2, MAIFX3D5, and MA-IFX10F9, were preincubated with IFX, less than 20% of the spiked IFX was detected (Fig. 4). Addition of 6 mcg/mL of any MA to the buffer revealed a response below the LLOQ. IFX levels of 41 antibodies toward IFX-positive samples, ranging from 1 to 30 mcg/mL pAb-IFX equivalents, were analyzed with the MA-based and pAb-based IFX ELISA. All serum samples were below the detection limit (,0.30 mcg/mL) in the pAb-based IFX ELISA, whereas the MA-based IFX ELISA revealed 36 serum samples with undetectable levels (,0.2 mcg/mL) and 5 serum samples with detectable but not quantifiable levels (.0.2 mcg/mL, but ,0.5 mcg/mL). Overall, a good agreement was demonstrated between the MA-based and pAb-based IFX assay.
DISCUSSION
FIGURE 2. Comparison of IFX values determined with the MA-based and pAb-based IFX ELISA. The MA-based IFX ELISA correlated well with the pAb-based IFX ELISA (R = 0.986). One sample (open dot) contained ADM from previous ADM treatment.
Several retrospective studies (Maser et al7 in Crohn disease, Seow et al8 in ulcerative colitis) have demonstrated the relevance of therapeutic drug monitoring of IFX and ADM as an indicator of therapy effectiveness in IBD patients.
Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
483
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Van Stappen et al
FIGURE 4. Evaluation of the impact of (non-) neutralizing anti-IFX monoclonal antibodies on the response of IFX in the assay. IFX was spiked to buffer in a concentration of 3 mcg/mL in the absence (-MA) and presence of 6 mcg/mL monoclonal antibody. In the presence of 6 mcg/mL monoclonal antibody not binding to IFX (MA-33H1F7) or 6 mcg/mL nonneutralizing anti-IFX antibody (MA-IFX13E7 and MA-IFX4B6), detection of 3 mcg/mL IFX was not hampered. The addition of 6 mcg/mL neutralizing anti-IFX antibody (MA-IFX20G2, MA-IFX3D5, and MA-IFX10F9) to 3 mcg/mL IFX resulted in sharply reduced detection of IFX.
High trough levels are correlated with a sustained clinical response, whereas low trough levels (frequently accompanied with antidrug antibody formation) are linked with a higher chance of loss of response.7,8 Prospective studies (TAXIT16 and TAILORIX17) are ongoing, but results of the optimization phase of the TAXIT study revealed a significant decrease in C-reactive protein and Harvey–Bradshaw index after dose intensification in patients with Crohn disease with low trough levels (,3 mcg/mL), whereas dose de-escalation did not have any effect on C-reactive protein and Harvey–Bradshaw index in patients with Crohn disease nor on partial Mayo score of patients with ulcerative colitis.18 The results of these studies will help us refine optimal therapy guidelines and algorithms to further improve the management of IBD patients on IFX therapy. In addition, Steenholdt et al19 has recently shown, based on a prospective, randomized controlled single-blind multicenter study with 69 patients (Crohn disease) who lost response to IFX, that therapeutic drug monitoring of IFX is more cost-effective compared with dose intensification based on empirical clinical decisions. To date, different assay types are being used to measure IFX levels, and comparative studies have shown a good correlation between some (commercially) available immunoassays.10,20 The high variety in assay types creates, however, a wide range of treatment algorithms, making it impossible to generalize strategies, even within 1 country.20 Randomly switching from 1 assay to another is therefore not recommended.21 In this article, the characterization of a newly generated monoclonal antibody, MA-IFX6B7, is reported, which can be used as a conjugate in a TNF-coated IFX assay. A
484
Ther Drug Monit Volume 37, Number 4, August 2015
monoclonal antibody has the unique advantage that it can be constantly reproduced once the hybridoma has been generated. MA-IFX6B7 is a high-affinity antibody with very good specificity toward IFX. The ELISA does not crossreact with other anti-TNF drugs (ADM, golimumab, etanercept) or with antibodies toward IFX. Importantly, the performance of this newly described assay is in very good agreement with the clinically validated pAb-based ELISA, which was demonstrated using a Bland–Altman plot and the ICC. The currently described MA-based IFX assay is able to measure IFX in the presence of antibodies not binding to IFX and nonneutralizing anti-IFX antibodies. However, intrinsically related to TNF-coated assays, when IFX is in complex with a neutralizing anti-IFX antibody, the paratope of IFX that binds TNF is not accessible, and therefore the IFX/ antibody complex cannot bind to the TNF-coated plate and subsequently no IFX is detected. However, detection of IFX in complex with neutralizing anti-IFX antibodies depends also on the nature of the antibody because complex formation is an equilibrium between association and dissociation. When the antibody has a high off-rate, IFX will dissociate from the complex and bind to the TNF-coated plate, what might explain the determination of small amounts of IFX in some samples that contain antibodies toward IFX. However, the IFX concentration measured was never above the limit of quantification. In addition to its use in the IFX assay, the full panel of in-house generated MA-IFX may serve as a toolbox to develop other IFX and/or antibodies toward IFX assays, which are not limited to solid-phase ELISA.
CONCLUSIONS In conclusion, a highly specific MA-based ELISA has been developed to quantify IFX serum concentrations. The assay was benchmarked to the clinically validated reference pAb-based IFX ELISA, revealing a very good agreement. The improved specificity of the assay will rule out any clinical misinterpretation owing to ADM (or other anti-TNF drugs) still present in the serum sample. Additionally, the assay is able to measure IFX in the presence of antibodies not binding to IFX and nonneutralizing anti-IFX antibodies. REFERENCES 1. Rutgeerts P, Vermeire S, Van Assche G. Biological therapies for inflammatory bowel diseases. Gastroenterology. 2009;136:1182–1197. 2. Yanai H, Hanauer SB. Assessing response and loss of response to biological therapies in IBD. Am J Gastroenterol. 2011;106:685–698. 3. Allez M, Karmiris K, Louis E, et al. Report of the ECCO pathogenesis workshop on anti-TNF therapy failures in inflammatory bowel diseases: definitions, frequency and pharmacological aspects. J Crohns Colitis. 2010;4:355–366. 4. Ben-Horin S, Chowers Y. Review article: loss of response to anti-TNF treatments in Crohn’s disease. Aliment Pharmacol Ther. 2011;33: 987–995. 5. Nanda KS, Cheifetz AS, Moss AC. Impact of antibodies to infliximab on clinical outcomes and serum infliximab levels in patients with inflammatory bowel disease (IBD): a meta-analysis. Am J Gastroenterol. 2013;1: 40–47.
Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Ther Drug Monit Volume 37, Number 4, August 2015
Improving the Specificity of Infliximab Assays
6. Rojas JR, Taylor RP, Cunningham MR, et al. Formation, distribution, and elimination of infliximab and anti-infliximab immune complexes in cynomolgus monkeys. J Pharmacol Exp Ther. 2005;313:578–585. 7. Maser EA, Villela R, Silverberg MS, et al. Association of trough serum infliximab to clinical outcome after scheduled maintenance treatment for Crohn’s disease. Clin Gastroenterol Hepatol. 2006;4:1248–1254. 8. Seow CH, Newman A, Irwin SP, et al. Trough serum infliximab: a predictive factor of clinical outcome for infliximab treatment in acute ulcerative colitis. Gut. 2010;59:49–54. 9. Vande Casteele N, Ballet V, Van Assche G, et al. Early serial trough and antidrug antibody level measurements predict clinical outcome of infliximab and adalimumab treatment. Gut. 2012;61:321. 10. Vande Casteele N, Buurman DJ, Sturkenboom MG, et al. Detection of infliximab levels and anti-infliximab antibodies: a comparison of three different assays. Aliment Pharmacol Ther. 2012;36:765–771. 11. Galfre G, Milstein C. Preparation of monoclonal antibodies: strategies and procedures. Methods Enzymol. 1981;73:3–46. 12. Nakane PK, Kawaoi A. Peroxidase-labeled antibody. A new method of conjugation. J Histochem Cytochem. 1974;22:1084–1091. 13. Gils A, Vande Casteele N, Poppe R, et al. Development of a universal anti-adalimumab antibody standard for interlaboratory harmonization of antibody titers in patients who develop anti-adalimumab antibodies. Ther Drug Monit. 2014;36:669–673. 14. Dewitte K, Fierens C, Stockl D, et al. Application of the Bland-Altman plot for interpretation of method-comparison studies: a critical investigation of its practice. Clin Chem. 2002;48:799–801.
15. Bijnens AP, Gils A, Knockaert I, et al. Importance of the hinge region between alpha-helix F and the main part of serpins, based upon identification of the epitope of plasminogen activator inhibitor type 1 neutralizing antibodies. J Biol Chem. 2000;275:6375–6380. 16. Vande Casteele N, Compernolle G, Ballet V, et al. OP11 Individualised infliximab treatment using therapeutic drug monitoring: a prospective controlled Trough level Adapted infliXImab Treatment (TAXIT) trial. J Crohns Colitis 2012;6:S6. 17. Study Investigating Tailored treatment with infliximab for Active Crohn’s disease (TAILORIX). Available at: http://clinicaltrials.gov/ct2/show/ study/NCT01442025. Accessed August 21, 2013. 18. Vande Casteele N, Compernolle G, Ballet V, et al. Results on the optimisation phase of the prospective controlled trough level adapted infliximab treatment (TAXIT) trial. Gastroenterology. 2012;142: S211–S212. 19. Steenholdt C, Brynskov J, Thomsen OO, et al. Individualised therapy is more cost-effective than dose intensification in patients with Crohn’s disease who lose response to anti-TNF treatment: a randomised, controlled trial. Gut. 2014;63:919–927. 20. Steenholdt C, Ainsworth MA, Tovey M, et al. Comparison of techniques for monitoring infliximab and antibodies against infliximab in Crohn’s disease. Ther Drug Monit. 2013;35:530–538. 21. Van Assche G, Vermeire S, Ballet V, et al. Switch to adalimumab in patients with Crohn’s disease controlled by maintenance infliximab: prospective randomised SWITCH trial. Gut. 2012;61: 229–234.
Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
485
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Generation of a Highly Specific Monoclonal Anti-Infliximab Antibody for Harmonization of TNF-Coated Infliximab Assays Thomas Van Stappen, PharmD,* Els Brouwers, BSc,* Sophie Tops, BSc,* Nick Geukens, PhD,† Séverine Vermeire, MD, PhD,‡ Paul J. Declerck, PharmD, PhD,* and Ann Gils, PharmD, PhD*
Background: Determination of infliximab (IFX) serum concentrations has been used for treatment optimization of patients with inflammatory bowel disease. A wide range of enzyme-linked immunosorbent assays (ELISA) exists to quantitate IFX. Most of these assays lack specificity and cross-react with other anti-tumor necrosis factor (TNF) agents. The ability of these IFX assays to detect IFX in complex with antidrug antibodies is not known. The objective of our study was to develop an IFX-specific immunoassay to monitor IFX serum concentrations and to evaluate the impact of antidrug antibodies on the assay performance.
Methods: A panel of monoclonal antibodies toward IFX (MA-IFX) was generated by hybridoma technology and evaluated to replace the polyclonal antibody in a TNF-coated IFX assay. The selected monoclonal antibody–based (MA-based) IFX ELISA was benchmarked to a clinically validated, reference polyclonal antibody– based (pAb-based) IFX ELISA using 209 inflammatory bowel disease serum samples. Results: Fifty-five MA-IFX were generated and grouped into 9 clusters. Of the 22 monoclonal antibodies tested, MA-IFX6B7 was selected for use in the IFX ELISA and the assay was further optimized. MA-IFX6B7 is a high-affinity (KD = 1.40E-09 mol/L), noninhibitory IgG1 antibody that binds to the Fab fragment of IFX and exhibits no cross-reactivity with other anti-TNF drugs. The linearity of an IFX dose–response curve was demonstrated in the range of 1.2–37.5 ng/mL (R2 = 0.988). The MA-based assay showed a good Pearson correlation (R = 0.986) and agreement (intraclass correlation Received for publication May 1, 2014; accepted October 28, 2014. From the *Laboratory for Therapeutic and Diagnostic Antibodies, Department of Pharmaceutical and Pharmacological Sciences; †PharmAbs, The KU Leuven Antibody Center, KU Leuven; and ‡Department of Gastroenterology, UZ Leuven, Belgium. Supported in part by a grant G.0617.12 from the Fund for Scientific Research Flanders (FWO, Flanders) and grant IOFKP/12/002 from the KU Leuven. T. Van Stappen is a PhD fellow of the Agency for the Promotion and Innovation through Science and Technology in Flanders (IWT, Flanders). A. Gils has served as a speaker for MSD and Janssen Biologicals, as an advisory board member for Pfizer, and has received an Investigator Initiated Research Grant from Pfizer. S. Vermeire received grant support from Centocor, MSD, Abbvie, and UCB Pharma. The other authors declare no conflict of interest. Correspondence: Ann Gils, PharmD, PhD, Laboratory for Therapeutic and Diagnostic Antibodies, Department of Pharmaceutical and Pharmacological Sciences, KU Leuven, O&N II Herestraat 49, Box 820, B-3000 Leuven, Belgium (e-mail: [email protected]). Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Ther Drug Monit Volume 37, Number 4, August 2015
coefficient = 0.985) with the pAb-based assay. The MA-based assay detects IFX in complex with nonneutralizing anti-IFX antibodies but not when complexed with neutralizing anti-IFX antibodies.
Conclusions: In this study, a highly specific MA-IFX was developed as detection antibody in an ELISA to quantify IFX serum concentrations. The assay was benchmarked to the clinically validated reference pAb-based IFX ELISA. Key Words: infliximab, tumor necrosis factor-alpha, inflammatory bowel diseases, enzyme-linked immunosorbent assay, antibodies, monoclonal (Ther Drug Monit 2015;37:479–485)
INTRODUCTION Inflammatory bowel diseases (IBDs) are chronic autoimmune disorders of the gastrointestinal tract with a relapsing-remitting character. The treatment of IBD varies with severity of diseases and may include 5-aminosalicylic acid derivatives, corticosteroids, immunomodulatory drugs, and more recently, biologicals.1 Infliximab (IFX) and adalimumab (ADM), a chimeric and fully human monoclonal antibody, respectively, toward tumor necrosis factor-alpha (TNF), have proven to be efficacious in inducing and maintaining remission in patients with moderate/severe (fistulizing) Crohn disease and ulcerative colitis.1 However, not all patients benefit equally from these drugs, and 30%–40% of patients lose response within the first year of treatment.2,3 Reasons for this secondary loss of response are predominantly the development of neutralizing antidrug antibodies and increased clearance of anti-TNF–antibody immune complexes.4–6 Previous studies have demonstrated a positive correlation between the IFX trough level (the serum level just before the next infusion) and the clinical response of patients with Crohn disease or ulcerative colitis.7–9 Optimal trough levels are correlated with a long-term response to IFX, whereas patients with low trough levels have more chance to lose response over time. Diverse assay types exist to measure IFX and antidrug antibody serum concentrations [Radio-immunoassay, enzyme-linked immunosorbent assay (ELISA), bridging ELISA, mobility shift assay], each having their own merits and limitations. The most commonly used assays are ELISA techniques. A round robin experiment, performed in the Laboratory for Therapeutic and Diagnostic Antibodies of
479
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Van Stappen et al
KU Leuven (Leuven, Belgium), Sanquin (Amsterdam, the Netherlands), and the University Medical Center Groningen (Groningen, The Netherlands), revealed that a commercially available ELISA (LISA Tracker; Theradiag, Marne La Vallée, France) gives response in samples that contain antibodies toward IFX.10 As a consequence, patients failing IFX therapy may be missed and incorrectly continued on IFX therapy, possibly leading to hypersensitivity reactions and infusion reactions. Additionally, the lack of agreement between all available methods hampers the implementation of universal guidelines for therapeutic drug monitoring of IFX and underlines the need for standardization. In the “first-generation” polyclonal antibody–based (pAb-based) IFX ELISA, TNF was used as capture protein and an IFX monospecific polyclonal antibody preparation conjugated to horseradish peroxidase (pAb-IFX-HRP) as detection antibody. However, the pAb-IFX-HRP preparation can crossreact with other anti-TNF antibodies (ADM, golimumab) and is subject to batch-to-batch variation because of its polyclonal nature, requiring optimization each time a new batch has been produced. The use of a monoclonal antibody toward IFX (MAIFX), being highly specific for its target of interest and highly reproducible, can overcome these limitations. In this perspective, the objective of this study was to develop and validate an IFX-specific immunoassay to monitor IFX serum concentrations and to evaluate the impact of both neutralizing and nonneutralizing antidrug antibodies on the assay performance.
MATERIALS AND METHODS Bovine serum albumin (BSA), horseradish peroxidase (HRP), polyoxyethylene (20) sorbitan monooleate, sulfuric acid (H2SO4), mannitol, sucrose, fluorodinitrobenzene, and citric acid were purchased from Sigma-Aldrich (Steinheim, Germany). VWR International (Haasrode, Belgium) delivered disodium hydrogen phosphate dihydrate (Na2HPO4$2H2O) and sodium chloride (NaCl). Potassium dihydrogen phosphate (KH2PO4) was purchased from Fisher Scientific (Loughborough, United Kingdom). H2O2 and sodium metaperiodate were purchased from Merck (Darmstadt, Germany) and o-phenylenediamine, ethylene glycol, and potassium chloride (KCl) from Acros Organics (Geel, Belgium). TNF was purchased from PeproTech (London, United Kingdom). Complete and incomplete Freund adjuvant was purchased from BD Biosciences (Franklin Lakes, NJ). Dulbecco’s Modified Eagle’s Medium, glutamine, penicillin–streptomycin, and fetal bovine serum were purchased from Invitrogen Life Technologies Europe B.V. (Gent, Belgium). IFX and golimumab were purchased as Remicade and Simponi, respectively, from Janssen Biologics B.V. (Leiden, the Netherlands), ADM as Humira from AbbVie Inc (North Chicago, IL), etanercept as Enbrel from Pfizer Inc (New York, NY), and a human IgG-mixture as Multigam from Biotest AG (Dreieich, Germany).
Ther Drug Monit Volume 37, Number 4, August 2015
Crohn and colitis (B322201213950/S53684). Informed consent was provided by all patients. Generation of monoclonal antibodies in SJL/J mice (Charles River Laboratories, Wilmington, MA) was performed according to local ethical laws and approved by the local ethical committee (P088/2012 KU Leuven, Leuven, Belgium).
Monoclonal Antibody Generation Monoclonal antibodies were generated in-house using the method of Galfré and Milstein.11 Briefly, SJL/J mice were immunized by subcutaneous injection of 10 mcg IFX in complete Freund adjuvant, followed 2 weeks later by intraperitoneal injection of 10 mcg IFX in incomplete Freund adjuvant. Minimum 6 weeks later, the mice were boosted intraperitoneally with 10 mcg IFX in saline, 4 and 2 days before the cell fusion. The spleen was then isolated, spleen cells were fused with Sp2/0-Ag14 myeloma cells (ATCC, Manassas, VA) and plated onto 96-well cell-culture plates (Falcon; BD Biosciences). After selection in hypoxanthine–aminopterin–thymidine medium, supernatants were screened for the presence of antibodies against IFX by ELISA, using IFX for capture and HRP-conjugated rabbit anti-mouse antibody for detection. Subsequently, positive clones were selected and grown in Integra CELLine systems (Integra Biosciences AG, Zizers, Switzerland) for production of monoclonal antibody. Monoclonal antibodies were purified on a ProSep-vA Ultra column (Merck Millipore, Darmstadt, Germany). HRP-conjugated monoclonal antibodies were produced as described by Nakane and Kawaoi.12 Briefly, HRP was dissolved in 0.3 mol/L sodium bicarbonate buffer pH 8.1 in a brown glass, and slowly 1% fluorodinitrobenzene in ethanol was added. Then, the mixture was gently mixed for 1 hour at room temperature on a rotator. Subsequently, 8E-02 mol/L sodium metaperiodate was added and allowed to mix for 30 minutes, after which 0.16 mol/L ethylene glycol was added. Next, the (freshly made) “peroxidase-aldehyde” solution was mixed for 1 hour and dialyzed overnight at 48C against conjugate buffer, consisting of 1E-02 mol/L NaHCO3/Na2CO3 pH 9.5. Simultaneously, the monoclonal antibody was dialyzed overnight against conjugate buffer. The next day, the antibody was concentrated using Centricon (30 kDa) centrifugal filters (Merck Millipore) to a final concentration of 20–40 mg/mL and dissolved in the peroxidase-aldehyde solution. The solution was mixed gently for 2–3 hours at room temperature and dialyzed overnight against phosphate-buffered saline (PBS; 140 mmol/L NaCl, 8 mmol/L Na2HPO4$2H2O, 2.7 mmol/L KCl, and 1.5 mmol/L KH2PO4). Finally, the conjugate was sterilized through filtration and stored at 48C.
Evaluation of the Cross-Reactivity of MA-IFX Toward Other Anti-TNF Drugs
This study was approved by the Ethical Committee UZ Leuven in the framework of the Flemish inheritance study for
The cross-reactivity toward other anti-TNF drugs was determined by a sandwich-type ELISA. Briefly, IFX, the Fab fragment of IFX (Fab-IFX), ADM, etanercept, golimumab, and a human IgG-mixture were coated on 96-well plates (Costar; Corning Inc, NY) at a concentration of 4 mcg/mL in PBS. Next, MA-IFX were applied on the plate at 3 different concentrations (1–10–100 ng/mL). Subsequently, plates were
480
Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
Ethical Considerations
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Ther Drug Monit Volume 37, Number 4, August 2015
incubated overnight at 48C, and HRP-conjugated rabbit antimouse antibody (1:10.000 diluted) was used to detect bound MA-IFX. MA-IFX revealing reactivity toward any of the other anti-TNF drugs and/or the human IgG mixture, were considered as cross-reacting antibodies.
Evaluation of the IFX-Neutralizing Activity of MA-IFX The neutralizing activity was determined in a cellbased assay using a human fibrosarcoma cell line (HT1080; ATCC, CCL-121), characterized by a TNF-inducible interleukin-6 (IL-6) expression, according to a procedure described by Gils et al.13 Briefly, HT1080 cells were plated in a 96-well plate at a density of 4E4 cells in Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal bovine serum, 2 mmol/L glutamine, 50 mcg/mL streptomycin, and 50 U/mL penicillin per well and incubated 24 hours at 5% CO2 and 378C humidified atmosphere. Next, Dulbecco’s Modified Eagle’s Medium was replaced by varying concentrations of TNF (0–240 ng/mL) to determine the EC50 and in a subsequent experiment by a fixed concentration of TNF (at a concentration equal to EC50) with varying concentrations of IFX (0–120 ng/mL) to determine the IC80. After 48 hours of incubation, the amount of IL-6 was measured with the “Human IL-6 ELISA MAX Deluxe Kit” (BioLegend, San Diego, CA), according to the manufacturer’s protocol. In the subsequent experiments, a 10-fold excess of MA-IFX over IFX was added to a mixture of TNF (at a concentration equal to EC50) and IFX (at a concentration equal to IC80) and the effect of MA-IFX on IL-6 expression determined. An inhibitory antibody can upregulate IL-6 expression by binding to (and neutralizing) IFX, and thus blocking the inhibitory effect of IFX on TNF. A reference sample, TNF at a concentration equal to EC50 in the absence of IFX and MA-IFX, was set at 100%, to which the IL-6 secretion was compared. Minimum 80% of the IL-6 expression of the reference sample was required to consider a MA-IFX as neutralizing.
Improving the Specificity of Infliximab Assays
Development of the IFX ELISA Selected MA-IFX were subsequently evaluated to replace the polyclonal antibody in the TNF-coated IFX assay. Briefly, in the in-house developed IFX ELISA, 96-well plates were coated overnight with TNF at 48C. Plates were blocked with PBS 1% BSA for 2 hours at room temperature. Samples were diluted 1:150 and 1:300 in PBS 1% BSA, applied on the plate, and incubated for 2 hours at 378C on a shaker. Then, plates were washed and pAb-IFX-HRP or HRP-conjugated MA-IFX was applied for detection of bound IFX and incubated 1 hour at room temperature. Plates were washed and developed using o-phenylenediamine and H2O2 in citrate buffer, and the reaction was stopped with 2 mol/L H2SO4. The absorbance was measured at 490 nm with an ELx808 Absorbance Microplate Reader (BioTek Instruments Inc, Winooski, VT) and the IFX dose–response curve was analyzed by linear and nonlinear regression for the monoclonal antibody–based (MA-based) and pAb-based IFX assay, respectively, with GraphPad Prism 5.0 (GraphPad Software, San Diego, CA).
Determination of the Isotype of MA-IFX6B7
The antibody’s isotype was determined by the “IsoStrip Mouse Monoclonal Antibody Isotyping kit” (Roche Diagnostics, Indianapolis, IN), according to the manufacturer’s protocol.
Determination of the Affinity of MA-IFX6B7 Surface plasmon resonance-based affinity measurements were carried out using a BIAcore 3000 (GE Healthcare Life Sciences, Uppsala, Sweden). CM5 sensor chips were immobilized to 660 resonance units with IFX using 1-ethyl3-(3-dimethylaminopropyl)-carbodiimide hydrochloride/ N-hydroxysuccinimide amine coupling chemistry. Then, MA-IFX6B7 was applied at 3 different concentrations at a flow rate of 30 mL/min. Data were fitted using the 1:1 Langmuir binding global fit isotherm using BIAevaluation software 4.1 (GE Healthcare Life Sciences).
Blood Samples Epitope Binning of MA-IFX Epitope binning was carried out using competition ELISA. Briefly, 4 mcg/mL IFX was coated on the plate, incubated for 72 hours at 48C, and blocked with PBS 1% BSA. Then, MA-IFX were applied at a concentration of 10 mcg/mL and incubated overnight at 48C (“cold” antibody). Next, plates were emptied and HRP-conjugated MA-IFX (“labeled” antibody), 1:64,000 diluted (ie, at a concentration of 25 ng/mL), was applied followed by incubation for 2 hours at room temperature. Then, plates were washed and developed using 400 mcg/mL o-phenylenediamine and 3E-03% H2O2 in 0.1 mol/L sodium citrate, 0.2 mol/L disodium phosphate buffer pH 5 (citrate buffer), and the reaction stopped with 4 mol/L H2SO4. The optical density (OD) obtained for the labeled antibody in the presence and absence of cold antibody was compared and cold antibodies of which the ODþcold antibody OD2cold antibody · 100% gave a percentage lower than 20% were considered competitive with the labeled antibody. Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
Blood samples of patients with IBD were drawn at the Infusion Unit of the Division of Gastroenterology, UZ Leuven, using the “BD Vacutainer SST II Advance tube” (BD Biosciences), containing a gel separator and clot activator. After an incubation time of 30 minutes, serum was prepared by centrifugation (10 minutes, 1960g, RT) and stored at 48C (maximum 1 week) or 2208C (several years).
Validation of the IFX ELISA Lower Limit of Quantification, Accuracy, and Imprecision The assay cutoff value was calculated using 45 antiTNF naive serum samples, measured twice independently. The lower limit of quantification (LLOQ) was defined as the lowest concentration that could be accurately measured with a coefficient of variation (CV) #25%. The accuracy and imprecision were determined by preparing 5 quality control samples in PBS 1% BSA (0.5–1–3–7.5–11.25 mcg/mL). The
481
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Ther Drug Monit Volume 37, Number 4, August 2015
Van Stappen et al
quality control samples were diluted 1:150 and 1:300, and 8 replicates were applied per plate. The IFX dose–response curve was calculated using linear regression (GraphPad Prism 5.0), and the concentrations of the quality control samples were interpolated from the dose–response curve. The experiment was performed 3 times independently by 2 different operators.
Reactivity of the ELISA Toward Adalimumab The cross-reactivity of the ELISA toward ADM was evaluated using 3 different concentrations of ADM (3–7– 11.25 mcg/mL) and compared with the ADM assay9 (TNF-coated plate in which ADM bound to TNF is detected by a monospecific polyclonal antibody toward ADM). In addition, 20 patient samples who received ADM treatment were analyzed with the MA-based IFX assay and compared with the results obtained with the ADM ELISA.
Comparative Evaluation of IFX Values The monoclonal antibody–based ELISA was benchmarked to the clinically validated polyclonal antibody–based ELISA.10 In total, 209 IBD serum samples were analyzed in parallel by the MA-based and pAb-based IFX assay. The Pearson correlation and the relative Bland–Altman comparison14 were calculated using GraphPad Prism 5.0 and the intraclass correlation coefficient (ICC) using the “2-way mixed single measure test” from SPSS Statistics version 22 (IBM, New York, NY).
Impact of Neutralizing and Nonneutralizing Anti-IFX Antibodies Six monoclonal antibodies with different properties were selected to evaluate the ability of the IFX ELISA to detect IFX in the presence of other monoclonal antibodies. MA-IFX13E7 and MA-IFX4B6 do not inhibit IFX and do not bind to Fab-IFX, while cross-reacting with other anti-TNF drugs (nonneutralizing anti-IFX antibodies). MA-IFX20G2, MA-IFX3D5, MA-IFX10F9 do inhibit IFX and bind to FabIFX, while not cross-reacting with other anti-TNF drugs (neutralizing anti-IFX antibodies). MA-33H1F715 is a monoclonal antibody raised toward plasminogen activator inhibitor-1 and does not bind to IFX. IFX was spiked to buffer in a concentration of 3 mcg/mL in the absence (-MA) and presence of a 6-mcg/mL monoclonal antibody. Samples were preincubated for 10 minutes at 378C on a shaker, diluted 1:150 and 1:300, and applied on the IFX ELISA. In addition, 41 serum samples with antibodies toward IFX (detected with a bridging ELISA10) were analyzed in parallel on the MA-based and pAb-based IFX assay.
RESULTS
covering all clusters were evaluated for replacement of pAb-IFX-HRP in the IFX assay.
Selection of MA-IFX6B7 for Use in the IFX ELISA The ability of 22 monoclonal antibodies to detect IFX in spiked buffers and serum samples in the IFX ELISA setup was evaluated. Based on affinity, nonneutralizing properties, high specificity for IFX, and minimal serum influence, MA-IFX6B7 was selected for development of the assay. MA-IFX6B7 is a high-affinity (KD = 1.40E-09 mol/L), noninhibitory IgG1 antibody with k light chains. MA-IFX6B7 does not cross-react with other anti-TNF drugs (ADM, etanercept, golimumab) and binds the Fab fragment of IFX.
Lower Limit of Quantification, Accuracy, and Imprecision The use of MA-IFX6B7-HRP in the TNF-coated IFX assay yielded a linearity for the IFX dose–response curve in the range of 1.2–37.5 ng/mL (R2 = 0.988). Taking the 1:150 serum dilution into account, the limit of detection (LOD) was calculated to be 0.2 mcg/mL. The cutoff value of the assay was determined based on the mean OD obtained with 45 antiTNF naive serum samples plus 3 SDs compared with the mean OD of the IFX concentration at 1.2 ng/mL. The mean OD of the samples was 7.3E-03 with a standard deviation of 4.1E-03, giving a value of 2.0E-02, which was lower than the mean OD (6.0E-02) of 1.2 ng/mL IFX. Therefore, the cutoff for IFX determination was set at 0.2 mcg/mL. The assay complied with the requirements for accuracy: (1) the mean concentration was within 20% of the nominal value at each concentration level (and lower than 25% of the nominal value at the LLOQ/upper limit of quantification), (2) the imprecision (CV) was lower than 20% for all quality control samples (and lower than 25% for the LLOQ/upper limit of quantification) (Table 1). As 0.5 mcg/mL was still accurately detected (CV = 21%), the LLOQ was set at 0.5 mcg/mL.
Evaluation of Reactivity Toward ADM and Application of the ELISA on Samples From Patients Treated With ADM Applying different concentrations of ADM (3–7–11.25 mcg/mL) to the assay revealed no value above the LOD (Fig. 1). In 20 patient samples with ADM values ranging from 1 to 15 mcg/mL, no apparent IFX levels were detected (all
TABLE 1. Accuracy and Imprecision of 5 Quality Control Samples Within the Range of the Assay 0.5 mcg/mL, %
1 mcg/mL, %
3 mcg/mL, %
7.5 mcg/mL, %
11.25 mcg/mL, %
91 21
85 15
96 13
107 7
106 7
A panel of 55 anti-IFX monoclonal antibodies was evaluated: (1) for differential cross-reactivity with the Fab fragment of IFX, ADM, etanercept, golimumab, and a human IgG preparation, (2) for their ability to inhibit the TNF-IFX interaction, and (3) through epitope binning. Based on this methodology, the anti-IFX antibodies were grouped into 9 clusters. Subsequently, 22 HRP-conjugated MA-IFX
Accuracy Imprecision (CV)
482
Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
QC Samples
QC, quality control.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Ther Drug Monit Volume 37, Number 4, August 2015
FIGURE 1. Detection of ADM in the pAb-based ADM ELISA (empty bar) compared with the MA-based IFX ELISA (filled bar). ADM, tested in 3 different concentrations, was not detected in the MA-based IFX ELISA. Four patients (patients 1–4) on ADM therapy with detectable ADM levels did not reveal detectable values in the MA-based IFX ELISA.
values below LOD). Data of 4 representative samples are presented in Figure 1.
Comparative Evaluation of IFX Values Comparison of the IFX values (n = 209) obtained with the pAb-based IFX ELISA and with the currently developed MA-based IFX ELISA revealed a correlation (R) of 0.986 (Fig. 2). The ICC was 0.985. Results were graphically illustrated with a Bland–Altman plot, showing the relative difference between the 2 measurements to the average measured value (Fig. 3). The 95% limits of agreement were 226% and 30%. Based on these criteria, the methods can be considered interchangeable. One sample (Figs. 2, 3: open dot) revealed a much higher signal in the pAb-based IFX assay compared with that in the MA-based IFX assay. Further analysis showed that this
Improving the Specificity of Infliximab Assays
FIGURE 3. The Bland–Altman comparison of IFX values determined with the MA-based and pAb-based IFX ELISA. The relative (%) difference between IFX values determined with the MA-based and pAb-based IFX ELISA is plotted against the average measured IFX value (mcg/mL). The dotted lines represent the upper (30%) and lower (226%) 95% limits of agreement. One sample (open dot) contained ADM from previous ADM treatment and was therefore overestimated in the pAb-based IFX assay.
patient had received ADM 3 days before IFX infusion. The discrepancy between both assays for this sample is due to the ADM present in the sample. Subsequently, this value was excluded from the Pearson correlation, ICC, and 95% limits of agreement.
Impact of Neutralizing and Nonneutralizing Anti-IFX Antibodies Buffer was spiked with IFX (3 mcg/mL) in the absence (-MA) and presence of a 2-fold molar excess of antibody over IFX. On addition of a monoclonal antibody not binding to IFX (MA-33H1F7) to the spiked IFX sample, full recovery of IFX was achieved (97%). Addition of nonneutralizing antiIFX antibodies, MA-IFX4B6 and MA-IFX13E7, resulted in a detection of 78%–87% of the spiked IFX. However, when neutralizing anti-IFX antibodies, MA-IFX20G2, MAIFX3D5, and MA-IFX10F9, were preincubated with IFX, less than 20% of the spiked IFX was detected (Fig. 4). Addition of 6 mcg/mL of any MA to the buffer revealed a response below the LLOQ. IFX levels of 41 antibodies toward IFX-positive samples, ranging from 1 to 30 mcg/mL pAb-IFX equivalents, were analyzed with the MA-based and pAb-based IFX ELISA. All serum samples were below the detection limit (,0.30 mcg/mL) in the pAb-based IFX ELISA, whereas the MA-based IFX ELISA revealed 36 serum samples with undetectable levels (,0.2 mcg/mL) and 5 serum samples with detectable but not quantifiable levels (.0.2 mcg/mL, but ,0.5 mcg/mL). Overall, a good agreement was demonstrated between the MA-based and pAb-based IFX assay.
DISCUSSION
FIGURE 2. Comparison of IFX values determined with the MA-based and pAb-based IFX ELISA. The MA-based IFX ELISA correlated well with the pAb-based IFX ELISA (R = 0.986). One sample (open dot) contained ADM from previous ADM treatment.
Several retrospective studies (Maser et al7 in Crohn disease, Seow et al8 in ulcerative colitis) have demonstrated the relevance of therapeutic drug monitoring of IFX and ADM as an indicator of therapy effectiveness in IBD patients.
Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
483
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Van Stappen et al
FIGURE 4. Evaluation of the impact of (non-) neutralizing anti-IFX monoclonal antibodies on the response of IFX in the assay. IFX was spiked to buffer in a concentration of 3 mcg/mL in the absence (-MA) and presence of 6 mcg/mL monoclonal antibody. In the presence of 6 mcg/mL monoclonal antibody not binding to IFX (MA-33H1F7) or 6 mcg/mL nonneutralizing anti-IFX antibody (MA-IFX13E7 and MA-IFX4B6), detection of 3 mcg/mL IFX was not hampered. The addition of 6 mcg/mL neutralizing anti-IFX antibody (MA-IFX20G2, MA-IFX3D5, and MA-IFX10F9) to 3 mcg/mL IFX resulted in sharply reduced detection of IFX.
High trough levels are correlated with a sustained clinical response, whereas low trough levels (frequently accompanied with antidrug antibody formation) are linked with a higher chance of loss of response.7,8 Prospective studies (TAXIT16 and TAILORIX17) are ongoing, but results of the optimization phase of the TAXIT study revealed a significant decrease in C-reactive protein and Harvey–Bradshaw index after dose intensification in patients with Crohn disease with low trough levels (,3 mcg/mL), whereas dose de-escalation did not have any effect on C-reactive protein and Harvey–Bradshaw index in patients with Crohn disease nor on partial Mayo score of patients with ulcerative colitis.18 The results of these studies will help us refine optimal therapy guidelines and algorithms to further improve the management of IBD patients on IFX therapy. In addition, Steenholdt et al19 has recently shown, based on a prospective, randomized controlled single-blind multicenter study with 69 patients (Crohn disease) who lost response to IFX, that therapeutic drug monitoring of IFX is more cost-effective compared with dose intensification based on empirical clinical decisions. To date, different assay types are being used to measure IFX levels, and comparative studies have shown a good correlation between some (commercially) available immunoassays.10,20 The high variety in assay types creates, however, a wide range of treatment algorithms, making it impossible to generalize strategies, even within 1 country.20 Randomly switching from 1 assay to another is therefore not recommended.21 In this article, the characterization of a newly generated monoclonal antibody, MA-IFX6B7, is reported, which can be used as a conjugate in a TNF-coated IFX assay. A
484
Ther Drug Monit Volume 37, Number 4, August 2015
monoclonal antibody has the unique advantage that it can be constantly reproduced once the hybridoma has been generated. MA-IFX6B7 is a high-affinity antibody with very good specificity toward IFX. The ELISA does not crossreact with other anti-TNF drugs (ADM, golimumab, etanercept) or with antibodies toward IFX. Importantly, the performance of this newly described assay is in very good agreement with the clinically validated pAb-based ELISA, which was demonstrated using a Bland–Altman plot and the ICC. The currently described MA-based IFX assay is able to measure IFX in the presence of antibodies not binding to IFX and nonneutralizing anti-IFX antibodies. However, intrinsically related to TNF-coated assays, when IFX is in complex with a neutralizing anti-IFX antibody, the paratope of IFX that binds TNF is not accessible, and therefore the IFX/ antibody complex cannot bind to the TNF-coated plate and subsequently no IFX is detected. However, detection of IFX in complex with neutralizing anti-IFX antibodies depends also on the nature of the antibody because complex formation is an equilibrium between association and dissociation. When the antibody has a high off-rate, IFX will dissociate from the complex and bind to the TNF-coated plate, what might explain the determination of small amounts of IFX in some samples that contain antibodies toward IFX. However, the IFX concentration measured was never above the limit of quantification. In addition to its use in the IFX assay, the full panel of in-house generated MA-IFX may serve as a toolbox to develop other IFX and/or antibodies toward IFX assays, which are not limited to solid-phase ELISA.
CONCLUSIONS In conclusion, a highly specific MA-based ELISA has been developed to quantify IFX serum concentrations. The assay was benchmarked to the clinically validated reference pAb-based IFX ELISA, revealing a very good agreement. The improved specificity of the assay will rule out any clinical misinterpretation owing to ADM (or other anti-TNF drugs) still present in the serum sample. Additionally, the assay is able to measure IFX in the presence of antibodies not binding to IFX and nonneutralizing anti-IFX antibodies. REFERENCES 1. Rutgeerts P, Vermeire S, Van Assche G. Biological therapies for inflammatory bowel diseases. Gastroenterology. 2009;136:1182–1197. 2. Yanai H, Hanauer SB. Assessing response and loss of response to biological therapies in IBD. Am J Gastroenterol. 2011;106:685–698. 3. Allez M, Karmiris K, Louis E, et al. Report of the ECCO pathogenesis workshop on anti-TNF therapy failures in inflammatory bowel diseases: definitions, frequency and pharmacological aspects. J Crohns Colitis. 2010;4:355–366. 4. Ben-Horin S, Chowers Y. Review article: loss of response to anti-TNF treatments in Crohn’s disease. Aliment Pharmacol Ther. 2011;33: 987–995. 5. Nanda KS, Cheifetz AS, Moss AC. Impact of antibodies to infliximab on clinical outcomes and serum infliximab levels in patients with inflammatory bowel disease (IBD): a meta-analysis. Am J Gastroenterol. 2013;1: 40–47.
Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Ther Drug Monit Volume 37, Number 4, August 2015
Improving the Specificity of Infliximab Assays
6. Rojas JR, Taylor RP, Cunningham MR, et al. Formation, distribution, and elimination of infliximab and anti-infliximab immune complexes in cynomolgus monkeys. J Pharmacol Exp Ther. 2005;313:578–585. 7. Maser EA, Villela R, Silverberg MS, et al. Association of trough serum infliximab to clinical outcome after scheduled maintenance treatment for Crohn’s disease. Clin Gastroenterol Hepatol. 2006;4:1248–1254. 8. Seow CH, Newman A, Irwin SP, et al. Trough serum infliximab: a predictive factor of clinical outcome for infliximab treatment in acute ulcerative colitis. Gut. 2010;59:49–54. 9. Vande Casteele N, Ballet V, Van Assche G, et al. Early serial trough and antidrug antibody level measurements predict clinical outcome of infliximab and adalimumab treatment. Gut. 2012;61:321. 10. Vande Casteele N, Buurman DJ, Sturkenboom MG, et al. Detection of infliximab levels and anti-infliximab antibodies: a comparison of three different assays. Aliment Pharmacol Ther. 2012;36:765–771. 11. Galfre G, Milstein C. Preparation of monoclonal antibodies: strategies and procedures. Methods Enzymol. 1981;73:3–46. 12. Nakane PK, Kawaoi A. Peroxidase-labeled antibody. A new method of conjugation. J Histochem Cytochem. 1974;22:1084–1091. 13. Gils A, Vande Casteele N, Poppe R, et al. Development of a universal anti-adalimumab antibody standard for interlaboratory harmonization of antibody titers in patients who develop anti-adalimumab antibodies. Ther Drug Monit. 2014;36:669–673. 14. Dewitte K, Fierens C, Stockl D, et al. Application of the Bland-Altman plot for interpretation of method-comparison studies: a critical investigation of its practice. Clin Chem. 2002;48:799–801.
15. Bijnens AP, Gils A, Knockaert I, et al. Importance of the hinge region between alpha-helix F and the main part of serpins, based upon identification of the epitope of plasminogen activator inhibitor type 1 neutralizing antibodies. J Biol Chem. 2000;275:6375–6380. 16. Vande Casteele N, Compernolle G, Ballet V, et al. OP11 Individualised infliximab treatment using therapeutic drug monitoring: a prospective controlled Trough level Adapted infliXImab Treatment (TAXIT) trial. J Crohns Colitis 2012;6:S6. 17. Study Investigating Tailored treatment with infliximab for Active Crohn’s disease (TAILORIX). Available at: http://clinicaltrials.gov/ct2/show/ study/NCT01442025. Accessed August 21, 2013. 18. Vande Casteele N, Compernolle G, Ballet V, et al. Results on the optimisation phase of the prospective controlled trough level adapted infliximab treatment (TAXIT) trial. Gastroenterology. 2012;142: S211–S212. 19. Steenholdt C, Brynskov J, Thomsen OO, et al. Individualised therapy is more cost-effective than dose intensification in patients with Crohn’s disease who lose response to anti-TNF treatment: a randomised, controlled trial. Gut. 2014;63:919–927. 20. Steenholdt C, Ainsworth MA, Tovey M, et al. Comparison of techniques for monitoring infliximab and antibodies against infliximab in Crohn’s disease. Ther Drug Monit. 2013;35:530–538. 21. Van Assche G, Vermeire S, Ballet V, et al. Switch to adalimumab in patients with Crohn’s disease controlled by maintenance infliximab: prospective randomised SWITCH trial. Gut. 2012;61: 229–234.
Copyright 2014 Wolters Kluwer Health, Inc. All rights reserved.
485
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
