Development of an antibody-dependent cellular cytotoxicity reporter assay for measuring anti-Middle East Respiratory Syndrome antibody bioactivity

Middle East Respiratory Syndrome coronavirus (MERS-CoV) is a highly virulent pathogen that causes Middle East Respiratory Syndrome (MERS). Anti-MERS-CoV antibodies play an integral role in the prevention and treatment against MERS-CoV infections. Bioactivity is a key quality attribute of therapeutic antibodies, and high accuracy and precision are required. The major methods for evaluating the antiviral effect of antiviral antibodies include neutralization assays using live viruses or pseudoviruses are highly variable. Recent studies have demonstrated that the antibody-dependent cellular cytotoxicity (ADCC) activity of antiviral antibodies is more consistent with the virus clearance effect in vivo than neutralization activity. However, no reports evaluating the ADCC activity of anti-MERS antibodies have been published to date. Here, we describe the development of a robust and reliable cell-based reporter gene assay for the determination of ADCC activity of anti-MERS antibodies using 293T/MERS cells stably expressing the spike protein of MERS-CoV (MERS-S) as target cells and the engineered Jurkat/NFAT-luc/FcγRIIIa stably expressing FcγRIIIA and NFAT reporter gene as effector cells. According to the ICH-Q2 analytical method guidelines, we carefully optimized the experimental conditions and assessed the performance of our assay. In addition, we found that the ADCC activity of afucosylated anti-MERS antibodies is higher than their fucosylated counterparts. The establishment of this ADCC determination system provides a novel method for evaluating the bioactivity of anti-MERS antibodies and improving ADCC activity through modification of N-glycosylation of the Fc segment.


FcγRIIIa binding analysis. The affinity of anti-MERS CoV antibodies to the human FcγRIIIa (CD16a)
was determined using the Octet Red96 system (ForteBio). The biotinylated recombinant human CD16a protein (Sino Biological Inc) was equilibrated in kinetics buffer (phosphate-buffered saline, 0.1% bovine serum albumin, 0.02% Tween-20) at a concentration of 5 μg/mL onto the streptavidin sensors (ForteBio). The anti-MERS antibodies were equilibrated with kinetics buffer at two-fold serial dilution concentrations from 31.3 nM to 4000 nM. The sensors were regenerated by dipping in the regeneration buffer (150 mM NaCl, 300 mM sodium citrate, pH 3.5) and wash solution alternately to remove the antibodies but not the CD16a loaded on the sensors. The Octet assay protocol was performed as follows: baseline (kinetic buffer) for 60 s, loading (rhCD16a proteins) for 150 s, baseline 2 (kinetic buffer) for 60 s, association (anti-MERS antibodies) for 60 s, disassociation (kinetic buffer) for 60 s, regeneration (regeneration buffer) for 5 s and wash (kinetic buffer) for 5 s alternately for three times, then the next recycle was started from the association process at a higher concentration of the same sample. The data of the samples were subtracted from the value of the control sensor and were analyzed using the analysis software provided in the Octet Red96 system.

Generation of the 293T/MERS cell line.
The 293T cells (ATCC, CRL-3216) were transfected with the pcDNA4-MERS-S expression vector using the Lipofectamine transfection system (Invitrogen), followed by zeocin selection (500 μg/mL) 24 h post-transfection. After recovery, the stable pool was subcloned by limited dilution, and the clones with high expression of MERS-S were identified by flow cytometry. For the flow cytometry assay, the cells were incubated with h2E6 anti-MERS antibody against the cell-surface antigen for 30 min at 4 °C and the unconjugated antibodies were washed out with FACS washing buffer (2% FBS in PBS), followed by another 30 min incubation with APC-conjugated anti-human IgG Fc antibody (Biolegend) and washing procedure. Isotype antibodies were included, substituting the anti-MERS antibody as negative controls. The luminescence intensity of the stained cells was analyzed by flow cytometry (FACS Calibur, BD Biosciences). The clones highly expressed the MERS antigen, which were identified by flow cytometry, were further selected by the ADCC reporter gene assay (RGA). The selected 293T/MERS cell line was cultured in high glucose DMEM medium with 10% fetal bovine serum and 500 μg/mL zeocin (Gibco).
ADCC reporter assay using Jurkat/NFAT-luc/FcγRIIIa. To assess the antibody-dependent cytotoxicity of anti-MERS antibodies, an FcγRIIIa reporter assay was performed and optimized. Jurkat/NFAT-luc/FcγRIIIa effector cells were developed as previously described 40,41 . The target 293T/MERS cells and the effector Jurkat/ NFAT-luc/FcγRIIIa cells were counted and diluted in the dilution medium (RPMI 1640 with no phenol red and with 2% low IgG FBS) and were seeded into white 96-well plates (2.5 × 10 4 293T/MERS cells/well in 25 μL of dilution medium and 1 × 10 5 Jurkat/NFAT-luc/FcγRIIIa cells/well in 25 μL of dilution medium, and the ratio of the effector to the target cell was 4:1). Anti-MRES antibodies were serially diluted in a 1:2.5 ratio with dilution medium at an initial concentration of 2 μg/mL, and 50 μL of the serial antibody diluents were added into the 96-well plates. Cultures were incubated at 37 °C with 5% CO 2 for 16 h. By adding of Bright-Glo Luciferase Assay reagent (Promega) and measuring relative luciferase units (RLU) with a plate reader (SpectraMax), the ADCC activity of anti-MERS antibodies can be determined. In this study, the ADCC reporter gene assays that have not been described specifically refer to this method.
ADCC reporter assay using NK92/NFAT-luc/CD16a. For method comparison, NK92/NFAT-luc/ CD16a cells were used to replace Jurkat/NFAT-luc/FcγRIIIa cells as effector cells. The 293T/MERS cells and the effector cells NK92/NFAT-luc/CD16a were counted and mixed at the proportion of 1:2 (5 × 10 5 293T/MERS cells/mL and 1 × 10 6 NK92/NFAT-luc/CD16a cells/mL), and the mixtures were seeded into white 96-well plates in 50 μL of assay medium dilution medium. Anti-MRES antibodies were serially diluted with dilution medium with a 1:5 ratio at the initial concentration of 100 μg/mL, and 50μL of the serial antibody diluents were added into the 96-well plates. The cultures were incubated at 37 °C with 5% CO 2 for 16 h. The luciferase activity was assessed by the addition of Bright-Glo Luciferase Assay reagent.
Accuracy and linearity. AF-h2E6 anti-MERS antibody was diluted to 1, 1.5, 2, 2.5, and 3 μg/mL in dilution medium with the 2 μg/mL concentration as the reference standard. The expected bioactivities of the serial dilutions were 50%, 75%, 100%, 125%, and 150% respectively. Linear analysis was performed using the measured versus the expected potency of each concentration point. The ratios of the measured potency to the expected potency were used to determine the accuracy of the method. All of the samples were subjected to three independent replicate assays.

Specificity.
To test the cell specificity of the RGA assay, we employed the 293T cells instead of the 293T/

Statistical analysis.
A four-parameter logistic model (Sect. 5.3 of European Pharmacopoeia) was employed to determine the 50% concentration of maximal effect (EC 50 ), dose response, as well as linear range of the bioassay. The relative potency of the bioactivity was calculated by the ratio of EC 50 value of the reference to the sample. The signal to noise ratio (SNR) was calculated using the ratio of the upper to the lower asymptotes. Data analysis and graphing were conducted using the GraphPad Prism software.

Effect of glycosylation modification on the binding activity of anti-MERS antibody.
The pseudovirus neutralization activity mediated by the Fab segment and the binding affinities with FcγRIIIa mediated by the Fc segment of several anti-MERS-CoV antibodies with the same antigen binding site and different Fc segments were determined. 2E6 is a neutralizing antibody against MERS-CoV targeting the receptor-binding domain of the spike protein, and were generated by mice immunized with recombinant MERS-S and subsequently selection of hybridoma cell lines. 2E6 can block virus entry in vitro with high efficacy 20 . Humanized 2E6 antibodies (h2E6) were generated by preserving the paratope residues and substituting the remaining amino acids of the murine 2E6 antibody (m2E6) with counterparts of the human immunoglobulins. Both m2E6 and h2E6 were belong to the IgG1 subtype. Afucosylated h2E6 antibody (AF-h2E6) were produced by a Chinese hamster ovary (CHO) host cell line lacking the fucose synthesis ability while the h2E6 were produced by a wildtype CHO cell line. The glycosylation of h2E6 and AF-h2E6 were analyzed and there showed nearly no glycans with core fucose in the AF-h2E6, whereas the ratio of glycans with core fucose in the h2E6 was more than 90% (G0F 30.35%, G1F 48.95%, and G2F 14.59%) (Fig. 1A). The pseudovirus neutralization assay was performed to determine the neutralization activity of the anti-MERS CoV antibodies including m2E6, h2E6, AF-h2E6, and PN-h2E6 without N-glycans. No evident differences were observed in the 50% neutralization dose (ND 50

Generation of the 293T/MERS cell line.
To verify whether the anti-MERS antibody has ADCC activity and whether our experiment designed to determine the ADCC activity is reasonable, we utilized the 293T cells which were transiently transfected with pCDNA4-MERS-S plasmid as target cells and the Jurkat/NFAT-luc/ FcγRIIIa cells as effector cells to evaluate the ADCC activity of the AF-h2E6 antibody. The results showed that the relative luciferase units (RLU) was positively associated with the AF-h2E6 concentration, indicating that the AF-h2E6 has ADCC activity and can be evaluated by our RGA detection system ( Fig. 2A). Then, to establish a cell line that stably expresses the MERS antigen, the 293T cells were transfected with pCDNA4-MERS-S, followed by subcloning and pressure screening. The clones showing high expression of the MERS antigen were identified by flow cytometry (Fig. 2B), and the optimized clones were subjected to the ADCC reporter gene assay subsequently (Fig. 2C). Of the six representative antigen-positive clones identified by flow cytometry, clone 9 showed the highest fold induction (FI) of luciferase under the same experimental condition in the ADCC determination assay and was named as 293T/MERS (Fig. 2C).
Scientific RepoRtS | (2020) 10:16615 | https://doi.org/10.1038/s41598-020-73960-x www.nature.com/scientificreports/ Optimization of the ADCC reporter gene assay. To establish an optimal and standardized ADCC reporter gene assay to determine the ADCC activity of the anti-MERS antibodies, we optimized the key experimental parameters of the assay. We performed a simple orthogonal experiment involving the variables of incubation time, cell density, and the ratio of effector cells to target cells, and the results showed that there was no obvious correlation between these experimental factors in this experiment (data not shown), so we optimized these factors separately by changing one factor while keeping the others constant. The cell density of the effector cells was adjusted to 0.25 × 10 5 , 0.5 × 10 5 , 1 × 10 5 , 2 × 10 5 cells/well, and the density of the target cells were correspondingly adjusted at the ratio of the effector to target (E:T) of 1:1, 2:1, 4:1, 8:1. The dose-response curve showed good signal value and high SNR when the effector cells density was 1 × 10 5 cells/well and the ratio of effector cells to target cells was 4:1 (Fig. 3A) (Fig. 3B). According to the fitting curve in the previous optimization process, we optimized the antibody concentration, initiating at a working concentration of 1 μg/mL and serially diluted at a 1:2.5 ratio with dilution medium (Fig. 3C).
Assessment of the accuracy, specificity, precision, and linearity of the ADCC reporter gene assay. The ADCC reporter gene assay was validated according to the International Conference on Harmonization Q2 (ICH-Q2) guidelines 42 by assessing the accuracy, specificity, linearity, and precision. For the precision assessment, the AF-h2E6 was diluted to three copies, one of which served as the reference, and the relative activity was the ratio of EC 50 value of the reference to sample. Repeatability tests were performed for three times independently on different days, and the results indicated good precision ( Table 1). The accuracy of an analytical procedure is an indication of agreement between the measured and expected values. The linearity of an analytical procedure indicated its ability to obtain test results that are directly proportional to the concentration of the sample analyte within a given range 43 . The AF-h2E6 anti-MERS antibody were diluted at the starting concentrations of 50%, 75%, 100%, 125%, and 150%, respectively, relative to the established antibody concentration in the RGA, using another 100% established antibody concentration as the reference, and the measured relative www.nature.com/scientificreports/ activity was the ratio of EC 50 value of the reference to each sample. The relative bioactivity of each sample was determined and showed high closeness of agreement between the measured and expected biological activity, indicating excellent accuracy of the ADCC reporter gene assay ( Table 2). The measured data of activity at the concentrations of 50%, 75%, 100%, 125%, and 150% were linearization analyzed and showed good linearity in the established bioassay (Fig. 4A). Specificity means that the result of the experiment is not caused by nonspecific interactions between cells in the system and the assay could unequivocally evaluate the analyte of interest in the presence of other components. When we replaced the 293T/MERS with the parental 293T cells or replaced the Jurkat/NFAT-luc/FcγRIIIa with the Jurkat/NFAT-luc cells, no luminescence signals mediated by the AF-h2E6 antibody were observed (Fig. 4B). Bevacizumab (anti-VEGF mAb), even rituximab (anti-CD20 mAb), trastuzumab (anti-HER2 mAb), and cetuximab (anti-EGFR mAb) which have the ability to activate ADCC, all failed to produce a dose-dependent signal curve in our established ADCC activity determination assay for anti-MERS antibodies (Fig. 4C). These results confirmed the specificity of our assay.

comparison of antibody biological activity determined by ADcc reporter assay and cytotoxicity assay. ADCC is traditionally mediated by NK cells, so we evaluated the level of agreement between
ADCC reporter assay using Jurkat/NFAT-luc/FcγRIIIa cells and using NK92/NFAT-luc/CD16 cells. The ADCC activity of m2E6, h2E6, AF-h2E6 and PN-h2E6 were determined. The m2E6 antibody failed to produce a dosedependent signal curve in the ADCC detection systems based on the human Fc receptor while the h2E6 showed obvious ADCC activity. The ADCC activity enhanced significantly when the core fucose expression lacked in the AF-h2E6 antibody, whereas the ADCC activity disappeared when all of the N-glycan were removed in the PN-h2E6 (Fig. 5A,B). These results demonstrated that the ADCC activity which mediated by anti-MERS antibodies with the same Fab segment but having different glycosylated Fc segment were different. In addition, the results also proved the specificity of the determination of ADCC activity, as the luciferase activity was not caused by nonspecific cross-linking between cells or between antibodies and cells since m2E6 and PN-h2E6 failed to activate the luciferase activity in the same culture system. No significant differences were found between ADCC procedure based on the effector cell of Jurkat/NFAT-luc/FcγRIIIa (Fig. 5A) and NK92/NFAT-luc/CD16 (Fig. 5B), which demonstrated high consistency for the two assays. Furthermore, a traditional cytotoxicity assay was performed to demonstrate that the 293T/MERS target cells were killed following incubation with anti-MERS antibodies and NK92/CD16a cells. The 293T/MERS target cells were loaded with an acetoxymethyl ester of a fluorescence enhancing ligand. After the ligand had penetrated the cell membrane, the ester bonds were hydrolyzed within the cell to form a hydrophilic ligand (TDA), which no longer passes through the membrane. The labeled target cells were incubated together with the serially diluted  www.nature.com/scientificreports/ anti-MERS antibodies and the NK92/CD16a effector cells. When the NK92/CD16a effector cells were activated by the anti-MERS antibodies bound on the target cells, the TDA-labeled target cells were lysed and the TDA was released in the cell culture supernatant. The subsequently measured fluorescence signal, emitted by the chelate between the added europium and the released TDA, was correlated to the number of lysed target cells. The results showed that the death of target cells mediated by ADCC was similar with the RLU signal in the reporter assay (Fig. 5C). These results demonstrate that the luciferase activity detected in the reporter assay can be an indirect measure of ADCC. The results showed that when the passage number of the two cell lines was between 7 and 37, there were no significant differences of the overall fluorescence intensity and the signal-noise ratio of the dose-response curves ( Fig. 6A and 6B), demonstrating excellent stability of the 293T/MERS and Jurkat/NFAT-luc/FcγRIIIa cell lines and robustness and consistency of the established bioassay.

Discussion
Bioactivity is a key quality attribute of therapeutic antibodies. In recent years, the therapeutic antibodies have been rapidly developed, raising the high requirement for bioactivity determination assay. ADCC activity serves as a critical mechanism of action of many therapeutic antibodies, including antiviral antibodies [31][32][33][34][35][36] . The effector Table 2. The closeness of agreement between the measured and the expected biological activity.  www.nature.com/scientificreports/ cells used in the classic ADCC assay are primary human peripheral blood mononuclear cells (PBMCs) and the subset natural killer (NK) cells that are isolated from human blood. The target cells are pre-loaded with radioactive tracer such as radioactive chromium (51Cr) 44 or fluorescent dyes such as calcein-AM 45,46 , and the release of these substances are detected to determine the cytotoxicity. In addition, ADCC activity can also be measured by detection of cell death and proliferation such as the measurement of released LDH 47 . All of these assays have poor reproducibility and low sensitivity, due to the inherent variability of the primary effector cells and the high levels of spontaneous release independent on the antibody-dependent killing of target cells. These highly variable traditional methods are thus not suitable for routine quantitative measurement of ADCC activity of therapeutic antibodies 48 49 . The ADCC reporter assay showed improved precision and overall consistence compared to the classic ADCC assay. In our previous work, we have established several novel bioactivity determination methods for different kinds of antibodies based on reporter gene assays, and these methods have obvious advantages compared to the classic bioactivity determination methods if there ever were 43,50,51 . The documented reporter-based surrogate ADCC assay were all designed for antitumor antibodies targeted to hematopoietic or solid tumor cells 41,49 , and there have been no reports about surrogate ADCC assay for antivirus antibodies. Different from methods for antitumor antibodies, there are no natural target cells for antiviral antibodies, so it would be of great value to design an ADCC reporter gene assay for antiviral antibody.
The target cells used in the ADCC activity determination assay for the antiviral antibodies, unlike the readyto-use tumor cell lines used for antitumor antibodies, were prepared by transient infection of the susceptible cells with live viruses or pseudoviruses [52][53][54] , so the preparation procedure is complex, and the target cells are unstable and highly variable. In this study, we generated a 293T/MERS cell line stably expressing the MERS-S as the target cell in the ADCC assay (Fig. 2). The effector cells used in the reporter assay are engineered Jurkat cells expressing the human Fc receptor FcγRIIIa (CD16a) and a luciferase reporter gene driven by NFAT response elements. Jurkat cells are CD4 + T cells, and do not have cytotoxic capability as effector cells. We selected the engineered Jurkat cells rather than NK cells, because Jurkat cells share the same ability to activate the appropriate pathways for cell activation after the crosslinking of FcγRIIIa as that of NK cells 55,56 and Jurkat cells are much more simple to handle than NK cell line. The ADCC reporter assay using Jurkat/NFAT-luc/FcγRIIIa is more suitable for release control, batch-to batch consistency and stability tests for mAbs with ADCC activity where accuracy and precision are important, and engineered Jurkat cells are commonly employed in the surrogate ADCC measurement of mAbs 41,49 and cytotoxicity measurement of CD3 based bispecific antibodies (BiTEs) 57

in industry.
When the anti-MERS antibody recognize the MERS-S antigen on the surface of 293T/MERS cells, the Fc segment of the antibody bind to the FcγRIIIa on the Jurkat/NFAT-luc/FcγRIIIa cells and initiate the downstream signaling pathways, resulting in the expression of the luciferase reporter gene driven by NFAT elements. The luciferase expression was positively associated with the concentration of anti-MERS antibodies, and the bioactivity of anti-MERS antibodies can be determined by the measurement of RLUs. For the application of this new ADCC reporter assay in the research and development of vaccines and antibodies, we optimized the experimental conditions of this method, and obtained a typical dose-dependent curve with suitable SNR and uniform distributed data points (Fig. 3).
In addition, the precision, accuracy, linearity, and specificity of the analytical method were evaluated according to the ICH-Q2 analytical method guidelines. The small value of coefficient of variation (CV) demonstrated the excellent precision ( Table 1). The high consistency between the measured and the expected biological activity demonstrated good accuracy (Table 2). Besides, the method showed good linearity in the concentration ranges of 50%-150% (Fig. 4A). By replacing the target cells or the effector cells with the corresponding parental cells, no non-specific signals in the cells caused by stimulation of the anti-MERS antibody were observed (Fig. 4B). Furthermore, only anti-MERS antibodies but not non-specific antibodies could induce luciferase expression in the ADCC assay (Fig. 4C), proving the high specificity of this ADCC assay. The ADCC assay also showed sufficient sensitivity in detecting the different bioactivity of different anti-MERS antibodies derived from different species or modified with different glycans (Fig. 5A). Cell passage number may affect the stability of transgenic reporter gene cells, and our results proved the excellent stability of the 293T/MERS and Jurkat/NFAT-luc/FcγRIIIa cell lines (Fig. 6).
NK cells are the key mediators of ADCC activity. The bioactivity of anti-MERS antibodies were determined in two ADCC reporter assays using Jurkat/NFAT-luc/FcγRIIIa effector cells (Fig. 5A) and NK92/NFAT-luc/CD16 effector cells (Fig. 5B) separately, and the results showed highly consistency. In addition, a traditional cytotoxicity assay mediated by NK92/CD16a effector cells was performed to prove that the results were concordance with those of the established RGA, and the similar correlation of antibody glycosylation with ADCC biological activity of different anti-MERS antibodies demonstrated that the luciferase activity driven by FcγRIIIa-NFAT signaling pathway in the effector cells can be a convincing indication of ADCC activity (Fig. 5C). The ADCC bioactivity of the anti-MERS antibodies is closely related to the affinity of Fc segment of antibody to the Fc receptor FcγRIIIa (Fig. 1C), whereas differ from the neutralization activity mediated by the Fab segment of these antibodies (Fig. 1B). Consistent with published reports 58,59 , our results showed that the afucosylated anti-MERS antibody AF-h2E6 exhibits higher binding affinity to CD16a (Fig. 1C) and increased ADCC potency than their fucosylated counterparts h2E6 (Fig. 5).
In the current study, we developed a robust reporter-based ADCC assay for the determination of bioactivity of anti-MERS antibodies targeting MERS spike protein for the first time. The ADCC assay will be a useful tool in the antibody characterization and lot release of therapeutic antibodies. Furthermore, our study provides a method for improving the ADCC bioactivity of anti-MERS antibodies by N-glycosylation modification of the Fc segment, which has been reported in other antiviral antibodies such as Ebola 60 , while there are no reports on the ADCC activity of anti-MERS antibodies nor corresponding ADCC determination methods to date. As for now, this ADCC activity determination method is validated for mAbs targeting MERS spike protein but not for the corresponding spike-specific antibodies in human serum. We believe that the MERS-CoV-specific ADCC assay is highly applicable in the screening, characterization, release control, and stability study of MERS-CoV neutralizing antibodies which are mostly targeting RBD region of spike protein, and may provide good technical rational for other anti-coronaviruses research.