The similarity testing plan is summarized in Supplementary Table 1. A comprehensive panel of over 20 methods was used to generate a detailed, physicochemical, structural and functional comparison of critical and non-critical product quality attributes for 10 lots of BAT1806/BIIB800 and 44 lots of TCZ (16 sourced from China [CN], 16 from the European Union [EU], and 12 from the United States [US]).

BAT1806/BIIB800 is formulated in His buffer 10 mM (containing histidine and histidine monohydrochloride monohydrate), arginine-hydrochloride (HCl) 50 mM, sucrose 20 mg/mL, polysorbate 80 (pH 6.2) 0.5 mg/mL, while TCZ is formulated in phosphate buffer15 mM, sucrose 50 mg/mL, polysorbate 80 (pH 6.5) 0.5 mg/mL.

Analytical similarity acceptance criteria were established based on the criticality of the product quality attributes, method capability, and lot-to-lot variability of the RP. A Tier system (as shown in Supplementary Table 2, see electronic supplementary material [ESM]) was used to rank and classify each quality attribute and associated methods in accordance with regulatory guidance [14, 15]. The attributes that directly relate to the primary mechanism of action are classified as the most critical quality attributes (CQAs) and are designated as Tier 1 (high clinical impact). For product quality attributes placed in Tier 1, an equivalence test was used where the proposed biosimilar is considered equivalent to the RP if the calculated 90% two-sided confidence interval of the mean difference between the proposed biosimilar product and RP is within ± δ. The δ was based on the standard deviation of all the RP lots tested and multiplied by 1.5. Tier 2 classification comprised quantitative assays measuring high-impact CQAs by orthogonal methods. The most sensitive method was used for quantitative characterization test results. Tier 2 group similarity acceptance criteria were categorized into analysis by quality range (QR). The similarity acceptance criterion for a Tier 2 QR-based analysis was > 90% of BAT1806/BIIB800 lots that were to be within the RP mean ± 3 standard deviations. Tier 3 classification covered quantitative assays measuring non-CQAs (with lower clinical risk) and semi-quantitative assays or assays producing a fingerprint or non-quantitative output (e.g., a spectra or chromatographic comparison). For Tier 3 attributes, the similarity assessments used raw data/graphical comparisons.

All analyses were performed at Bio-Thera Solutions, Ltd with the exception of far ultraviolet (UV) and near UV by circular dichroism, performed at the Guangzhou Institutes of Biomedicine and Health, Chinese Academy of Sciences, and sedimentation coefficient by sedimentation velocity analytical ultracentrifugation (SV-AUC), performed by Shanghai ZhenGe Biotech Co. Ltd. Samples were analyzed under identical conditions for both BAT1806/BIIB800 and TCZ.

2.1 Materials

The BAT1806/BIIB800 drug product was manufactured by Bio-Thera Solutions Ltd (Guangzhou, China), while the lots of TCZ were purchased from the Chinese (CN source), EU (EU source), and US markets (US source), and stored according to the manufacturer’s instructions.

2.2 Analytical Methods2.2.1 Intact and Reduced Deglycosylated Mass Analysis by Liquid Chromatography–Mass Spectrometry (LC–MS)

Intact mass analysis was performed by initial desalting using ultracentrifugation (Millipore, Amicon Ultra Centrifugal Filter, 3 kDa MWCO) followed by reverse phase (RPh) LC chromatographic resolution with a BioResolve™ RP (450 Å, 2.1 × 100 mm, 2.7 μm) column (Waters, Milford, MA, USA) on a Waters ACQUITY Bio H-Class ultra performance liquid chromatography (UPLC) system (Waters). Analysis was performed using UV adsorption (ACQUITY UPLC® Tunable UV Detector) and electrospray high-resolution accurate mass spectrometry (HRMS) on a Vion IMS QTof (Waters) using UNIFI software (Waters, v 1.8.2.16).

To determine the similarity of the antibodies at the heavy chain (HC) and light chain subunit level, samples were denatured with 6 M guanidine HCL (Aladdin) and reduced with 0.5 M dithiothreitol (DTT, Sigma) to generate reduced samples. For deglycosylated analysis, samples were diluted with Tris-HCl (pH 7.95, Aladdin) and treated with PNGase F (New England Biolabs); for reduced and deglycosylated analysis, intact antibodies were treated with PNGase F then denatured and reduced as before.

The intact mass analysis was summed, and the deconvoluted spectrum of BAT1806/BIIB800 (intact, reduced, and deglycosylated) was compared with lots of TCZ.

2.2.2 Reduced and Non-reduced Peptide Mapping by LC–MS

Reduced peptide mapping was performed by denaturation of samples with 6 M guanidine HCl (Aladdin) and DTT in a 100-mM Tris buffer. Iodoacetamide (0.5 M, Sigma) was used to cap the reduced thiols and the samples were then digested with trypsin (Promega) or Asp-N (Promega).

Non-reduced peptide mapping was performed by denaturation of test samples with 6M guanidine HCl (Aladdin) followed by digestion with either trypsin (Promega) or Asp-N (Promega) in the same manner as for reduced analysis.

Digested samples were chromatographically resolved using an ACQUITY UPLC® Peptide BEH C18 Column (300 Å, 1.7 µm, 2.1 mm × 150 mm, Waters) on an ACQUITY UPLC H-Class system (Waters). Samples were detected by UV absorbance at 214 nm and electrospray HRMS on Xevo G2-S QTof (Waters). Data collection was performed using MassLynx v4.1 software and processed with BiopharmaLynx v1.3.3 software (Waters) with disulphide bond positions identified by comparison of data generated under reducing and non-reducing conditions.

2.2.3 Glycan Mapping by Hydrophilic Interaction Chromatography (HILIC)-High-Performance Liquid Chromatography (HPLC)

Samples were denatured in RapiGest™ (Waters), reduced with DTT then treated with glycopeptidase F (Takara). The glycans were precipitated with acetonitrile and the supernatant was dried. Isolated N-glycans were labelled with procainamide HCl (Sigma) and chromatographically resolved using a HILIC XBridge BEH Amide (2.5 μm 3.0 × 100 mm) column (Waters) on an ACQUITY Arc HPLC (Waters) with fluorescent (FLR) detection (2475FLR, Waters) at an excitation wavelength of 310 nm and an emission wavelength of 370 nm. Data were acquired and processed by the integration algorithm Empower™3 (Waters) software and results reported as relative percentage of glycan (e.g., glycoform percent G0F, percent G1F, and percent G2F).

2.2.4 Total Sialic Acids by RPh HPLC-FLR

N-Acetylneuraminic acid (NANA) and N-glycolylneuraminic acid (NGNA) content were evaluated by hydrolyzing the samples and the NANA and NGNA standard curves (Sigma), and the isolated glycan samples with 4M acetic acid. Samples were chromatographically resolved using a Kromasil 100-5-C18 column (4.6 × 250 mm, 5 µm) on an ACQUITY Arc HPLC instrument (Waters) with FLR detection (2475FLR, Waters) at an excitation wavelength of 373 nm and an emission wavelength of 448 nm. Data based on NANA and NGNA calibration curves were acquired and processed by Empower™3 (Waters), and results reported as molar total sialic acid amount per mole of antibody.

2.2.5 Free Thiol Content by Ellman Method

Free thiol content was evaluated by treating test samples and a calibration line of L-Cys (Sigma) with Ellman's reagent (5,5′-dithiobis-[2-nitrobenzoic acid]). After the reaction, samples were analyzed on a UV-2600 UV spectrophotometer (Shimadzu) at 412 nm. Free thiol concentrations of test samples were calculated from the L-Cys calibration line.

2.2.6 Capillary Isoelectric Focusing

Test samples were mixed with Pharmalyte (Cytiva), isoelectric point (pI) markers (BioCEart, Beijing Biosmart Technical Institute), 4 M urea-capillary isoelectric focusing gel (Aladdin), 0.5 M arginine (Aladdin), and 0.2 M N-2-acetamido iminodiacetic acid (Aladdin). Each mixture was resolved using a μsilFC coated capillary (BioCEart) on a CE 7100 capillary electrophoresis system (Agilent). Data were acquired and processed by OpenLAB software (Agilent).

2.2.7 Extinction Coefficient

Extinction coefficient was calculated using the Edelhoch method [16].

2.2.8 Circular Dichroism Spectroscopy

Formulation buffer was removed from the test samples by ultrafiltration (12,000 rpm) and samples were diluted with phosphate buffer 5 mM (Sigma, 10 and 0.1 mg/mL). Secondary and tertiary structures were assessed with 10 mg/mL for near and 0.1 mg/mL for far UV-circular dichroism (UV-CD), respectively, using Chirascan V100 spectrophotometer UV-CD (Applied Photophysics Ltd) controlled by Chirascan software (Applied Photophysics Ltd). The CD absorption spectra in the far- and near-UV regions were measured at 190–260 nm (to determine secondary structures) and 250–350 nm (to determine tertiary structures), respectively. After scanning and data collection, Chirascan software was used to obtain averaged, baseline subtracted, and smoothed replicates of three spectra per sample.

2.2.9 Fourier-Transform Infrared Spectroscopy

Secondary structure was determined using spectral scanning of freeze-dried, powdered samples on an attenuated total reflection Fourier-transform infrared spectrometer (Thermo Fisher Scientific Inc., Nicolet iS10), running OMNIC Spectra software (Thermo Fisher Scientific Inc.). After background removal and second derivative processing (smoothed twice at level 9), spectra for test samples were compared.

2.2.10 Endogenous Fluorescence

The tertiary structure of test samples was compared by endogenous fluorescence using an Uncle-0383 static light scattering charge-coupled device FLR detector (Unchained Labs) scanning from 250 to 450 nm. After data collection, spectra were analyzed using Uncle Analysis software (Unchained Labs). Comparison of data was performed by visual comparisons.

2.2.11 Differential Scanning Calorimetry (DSC)

DSC was used to measure the thermal unfolding temperatures of specific domains of test samples on a MicroCal VP-DSC (Malvern Panalytical). Samples were diluted to 2 mg/mL with preparation buffer and heated from 40 to 120 °C (heating rate of 110 °C/h) in triplicate. Software was used to process and analyze melting temperature (MicroCal VP-Capillary DSC software, Malvern Panalytical) and normalization for protein concentration; the melting temperature was determined using MicroCal Analysis Launcher software (Malvern Panalytical).

2.2.12 Subvisible Particles by Light Obscuration and FlowCam

Subvisible particles of test samples were assessed by light obscuration using a HIAC 9703+ liquid particle counting system (Beckman Coulter) controlled by PharmSpec software (Beckman Coulter). Particle concentration results were reported as cumulative particle counts per mL for ≥ 2, ≥ 5, ≥ 10, ≥ 25, and ≥ 50 µm size ranges. Samples were also assessed using the FlowCam 8000 (Yokogawa Fluid Imaging Technologies, Inc.) particle imaging system containing a flow cell and a digital camera. Spherical and non-spherical particle counts per mL for particles ≥ 5 µm were reported as those product-related particles that were more likely to be proteinaceous, and potentially have an increased risk to illicit an immunogenic response.

2.2.13 Submicron Particle Size Concentrations by Dynamic Laser Light Scattering (DLS)

The translational diffusion coefficient for particles in solution was determined by DLS using a DynaPro NanoStar DLS (Waters). Samples were prepared at 25 °C and the determined coefficients were used to calculate the equivalent spherical hydrodynamic radius. DLS profiles were performed by visual comparisons.

2.2.14 Size Exclusion Chromatography (SEC) With Multi-Angle Light Scattering (MALS)

SEC with MALS was used to assess the similarity between test samples for aggregates. Samples were chromatographically resolved using a TSK-gel G3000SWXL analytical column (5 µm/7.8 mm × 300 mm, Tosoh) on an e2695 HPLC system (Waters) with a mini-DAWN MALS detector (Wyatt) and a W2414 refractive index detector (Waters). Data were collected and processed using ASTRA software (Wyatt) to generate the absolute molecular weight of monomer and polymer species in kDa.

2.2.15 Aggregate Profile by SV-AUC

An AUC consists of an optical detection system integrated into an ultracentrifuge allowing for real-time detection of the evolution of the concentration distribution of particles subjected to centrifugation. Aggregate profiles of test samples were determined using SV-AUC and performed on an Optima AUC analytical ultracentrifuge (Beckman Coulter) at 4000 rpm and the detection wavelength of 280 nm. AUC-SV was used to characterize aggregates and fragments with respect to heterogeneity and particle size. Samples were analyzed in triplicate and process by SEDFIT software (https://sedfitsedphat.nibib.nih.gov/software/default.aspx).

2.2.16 Size Variants by SEC-HPLC

Size variants of each sample were determined by SEC-HPLC with UV absorbance. Chromatographic resolution was achieved with a Tosoh Bioscience TSK Gel G3000 SWXL column (7.8 × 300 mm, 5 µm, Tosoh) on an e2695 HPLC system (Waters) using a 2489 UV/Vis Detector scanning at 280 nm. Data were acquired and processed by Empower™3 (Waters) software.

2.2.17 Low Molecular Weight Fragments by Capillary Electrophoresis (CE)-Sodium Dodecyl Sulfate (SDS)

Low molecular weight fragments present in samples were determined using CE-SDS either on reduced or non-reduced samples. For reduced analysis, samples were mixed with 70 µL of CE-SDS sample buffer (pH 9.4, BioCEart), and 5 µL of 2-Mercaptoethanol (Sigma) and boiled at 70 °C for 10 min. For non-reduced analysis, samples were mixed with 70 µL of CE-SDS sample buffer (pH 6.5) and 5 µL of iodoacetamide (Sigma) and heated at 65 °C for 4 min. Both sets of samples were electrophoretically separated using a bare fused-silica capillary (50 µm, 33 cm, Beckman Coulter) on an Agilent CE 7100 (Agilent Technologies) using UV detection at 220 nm. Data were acquired and processed by OpenLAB software with integration capabilities (Agilent).

2.2.18 Charge Variants by Ion Exchange (IEC)-HPLC

IEC-HPLC was used for the assessment of charge variants of test samples with and without carboxypeptidase B (CpB) treatment. For CpB treatment, samples were diluted with 20 mM of ACES buffer (pH 7.5) to a concentration of 5 mg/mL, then CpB (YAXINBIO, CAS #9025-24-5) at 5 mg/mL was added at a 50:1 (sample:CpB) ratio, mixed and incubated (30 min, 37 °C). Both sets of samples were then chromatographically separated on a Thermo MAbPac SCX-10 (10 μm, 4.0 × 250 mm) column (Thermo Fisher Scientific) using a 1260 HPLC system (Agilent) equipped with an e2695 UV detector (Waters). Data were acquired and processed by Empower™3 (Waters) software, using the ApexTrack integration algorithm. Results were reported as a relative percentage of the charge variants (e.g., percent of acidic region, percent of main peak, and percent of basic region).

2.2.19 Purity Hydrophobic Interaction (HIC)-HPLC and RPh-UPLC

HIC-HPLC was used to separate intact antibodies from their aggregates and variants under non-denaturing conditions. Chromatographic separation was performed with an Agilent Bio HIC (4.6 × 100 mm, 3.5 µm) column (Agilent) on an e2695 HPLC system (Waters) equipped with a 2489 UV/Vis Detector and data were collected at 280 nm. Data collection was performed using Empower™3 (Waters) software with the percentage of main peak, percentage of pre-peaks (eluting earlier that main peak), and percentage of post-peaks (eluting after main peak) evaluated.

RPh-UPLC was used to compare different proteins by their relative hydrophobicity. RPh is more denaturing than HIC and therefore offers complementary analysis. Chromatographic separation was performed on a BioResolve RPh mAb Polyphenyl (450 Å, 2.7 μm, 2.1 mm × 100 mm) column (Waters) and an ACQUITY UPLC H-Class Bio UPLC system (Waters) equipped with a Tunable UV detector (Waters), collecting data at 280 nm. Data collection was performed as described for HIC-HPLC.

2.3 Functional Assays

In addition to structural and physicochemical characterization, functional similarity is an essential part of biosimilarity. Given that the mechanism of action of tocilizumab is the inhibition of IL-6 binding to both sIL-6R and mIL-6R and, therefore, receptor signalling, multiple assays targeting this pathway were used to demonstrate that BAT1806/BIIB800 is similar to TCZ in terms of biological activity.

2.3.1 Binding to sIL-6R by Enzyme-Linked Immunosorbent Assay (ELISA)

Direct binding of test samples to sIL-6R was determined via an ELISA where recombinant human sIL-6R (Sino Biological) was coated onto ELISA plates. In brief, serially diluted samples were added to the sIL-6R immobilized plate, washed, then detected by goat anti-human IgG (fragment crystallisable [Fc] fragment) conjugated to horseradish peroxidase (HRP, Jackson ImmunoResearch) followed by 3,3′,5,5′ tetramethylbenzidine (TMB; Sigma-Aldrich) substrate. The reaction was terminated with 0.2 M sulfuric acid and absorbance measured with a SpectraMax M4 microplate reader (Molecular Devices) at 450 nm. After assessing parallelism of the dose–response curves, the binding activity of test samples relative to the qualified BAT1806/BIIB800 internal reference standard, B0520180301STD, derived from the commercial process lot (B0520180301; referred to as ‘reference standard’ hereafter) was determined using a four-parameter logistic model fit by SoftMax Pro Software (Molecular Devices). Results were reported as a percentage of the relative binding values.

2.3.2 Competitive Inhibition of IL-6 Binding to sIL-6R by ELISA

Competitive inhibition of test samples was determined by immobilization of recombinant human sIL-6 (Bio-Thera Solutions) onto ELISA plates. Serial dilutions of test samples and reference standard were mixed with IL-6R-His (Bio-Thera Solutions) at a fixed concentration of 1 µg/mL. After washing, rabbit anti-His antibody (Bio-Thera Solutions) was added, followed by anti-rabbit IgG conjugated to HRP (Wuhan Boster Biological Technology Co., Ltd.) and TMB (Sigma-Aldrich) substrate. The reaction was terminated with 0.2 M sulfuric acid, absorbance measured, and the binding activity of samples relative to the reference standard was determined as before. Results were reported as a percentage of the relative binding value.

2.3.3 Binding Kinetics to sIL-6R by Surface Plasmon Resonance (SPR)

Binding kinetics of test samples to sIL-6R (Sino Biological) were assessed by SPR using a Biacore™ T200 (GE HealthCare) running multiple-cycle kinetics. sIL-6R serial concentrations were added to a carboxy-methylated sensor chip (GE HealthCare) with pre-immobilized Protein A, which was used as the Fc region capture reagent of each sample. On-rates (ka), off-rates (kd), and equilibrium binding were calculated. Data collection was performed by Biacore T200 Evaluation software (GE HealthCare).

2.3.4 Inhibition of Signal Transducer and Activator of Transcription (STAT) 3 Phosphorylation by Homogeneous Time Resolved Fluorescence (HTRF)

IL-6 binding to mIL-6R results in a complex that recruits glycoprotein (gp)130 to activate intracellular signalling [17]. gp130 dimerization activates Janus kinase (JAK), a tyrosine receptor kinase, which phosphorylates tyrosine residues on gp130 to recruit and phosphorylate signal transducers and STAT3 [17]. Phosphorylated STAT3 then relocates into the nucleus where it regulates transcription and expression of target genes, activating a variety of immune responses [18]. The ability of test samples to inhibit STAT3 phosphorylation was determined by an HTRF assay (Cisbio, 62AT3PEG) on a SpectraMax M5e FLR microplate reader (Molecular Devices). TF-1 cells (human erythroleukemia cell line expressing mIL-6R) were incubated with serial test sample concentrations, then stimulated with a constant concentration of IL-6 (Sino Biological, 8 ng/mL). Following lysis, phosphorylated STAT3 was detected using the Phospho-STAT3 (Tyr705) kit (Cisbio) and a SpectraMax M5e FLR microplate reader (Molecular Devices) equipped with the HTRF detection model. The percentage of the relative potency of samples was calculated against a dose–response curve of the reference standard. Data collection and processing was achieved by SoftMax Pro Software (Molecular Devices).

2.3.5 Inhibition of IL-6/sIL-6R–Induced Vascular Endothelial Growth Factor (VEGF) Release in Human Fibroblast-Like Synovial Cells (HFLS)

IL-6 and sIL-6R in the synovia of patients with RA have been shown to correlate with local joint inflammation and the severity of joint damage in chronic synovitis [19, 20]. VEGF is upregulated by proinflammatory cytokines, including IL-6 [21]. To determine the ability of test samples to inhibit VEGF secretion, a cell-based ELISA assay was developed using HFLS cells from patients with RA (HFLS-RA). HFLS cells, which do not express mIL-6R but do express gp130, respond to IL-6 and sIL-6R stimulation through the trans-signalling pathway. HFLS-RA cells (Cell Applications, Inc.) were incubated with serial test sample concentrations, the reference standard, and a fixed dose of IL-6 and sIL-6R (both from Sino Biological). After incubation, secreted VEGF was detected according to the manufacturers’ instructions (human VEGF ELISA kit, Wuhan Boster Biological Technology Co., Ltd.). The relative potency of test samples was calculated against the reference standard at each concentration. Data collection and processing was achieved by SoftMax Pro Software (Molecular Devices).

2.3.6 Binding to mIL-6R by Flow Cytometry

Direct binding of test samples to mIL-6R was performed by incubating various concentrations of samples and reference standard with mIL-6R-overexpressing HEK-Blue™ IL-6 cells (InvivoGen). After washing, Alexa Fluor 488 goat anti-human IgG (Thermo Fisher Scientific Inc.) was added and mean fluorescence of FITC-A detected by flow cytometry on a CytoFLEX (Beckman Coulter) was evaluated. CytExpert software (Beckman Coulter) was used for data collection. Dose–response curves of the reference standard and samples were fitted using a four-parameter logistic model by GraphPad Prism 8 (GraphPad) and the binding activity of samples relative to the reference standard was determined.

2.3.7 Inhibition of IL-6–Mediated Proliferation in TF-1 Cells

Inhibition of IL-6–mediated proliferation by test samples was determined using TF-1 cells engineered to express mIL-6R (National Institutes for Food and Drug Control China). TF-1 cells were incubated with varying concentrations of reference standard and test samples in the presence of IL-6 (4 ng/mL, Sino Biological). After incubation, CellTiter-Glo® Luminescent Cell Viability Assay (Promega) solution was added and luminescence measured with a SpectraMax M5e microplate reader (Molecular Devices). The percentage of the inhibition of TF-1 cell proliferation relative to the reference standard was determined using a four-parameter logistic model fit by SoftMax Pro Software (Molecular Devices).

2.3.8 IL-6 Blockage Activity by Secreted Embryonic Alkaline Phosphatase (SEAP) Reporter Gene Assay

The SEAP reporter gene assay was developed based on the JAK/STAT3 signalling pathway to evaluate the blocking activity of test samples against IL-6. HEK-Blue™ IL-6 cells expressing mIL-6R, STAT3 cDNA, and interferon-β response elements–driven SEAP reporter genes were incubated with varying concentrations of reference standard and test samples in the presence of IL-6 (1.5 ng/mL, Sino Biological). IL-6–mediated activation of STAT3 and downstream SEAP was measured by the addition of QUANTI-Blue™ (InvivoGen) solution and quantified with a SpectraMax M5e microplate reader (Molecular Devices). Blockage activity was calculated against a dose–response curve of the internal reference standard by SoftMax Pro Software (Molecular Devices).

2.3.9 Binding Affinity to Fc Receptors (neonatal fragment crystallisable receptor [FcRn] and fragment crystallisable gamma receptors [FcγR] IIa, IIb, IIIa) by Bio-Layer Interferometry (BLI)

The FcRn, FcγRIIa-131H/131R, FcγRIIb, FcγRIIIa-158V/158F binding affinity of test samples was determined by BLI using an Octet QKe platform (Sartorius) with multiple-cycle kinetics. Briefly, streptavidin sensors (Sartorius) were used to load the biotinylated FcRn (ACRO Biosystems), FcγRIIa-131H/131R (ACRO Biosystems), FcγRIIb (ACRO Biosystems), and FcγRIIIa-158V/158F (ACRO Biosystems), respectively. Reference standard and test samples were tested at varying concentrations with the affinity constant KD obtained using the 1:1 kinetic or steady-state fitting model. Data were collected and processed with Octet Data Analysis software (Sartorius). Results were reported as a percentage of affinity relative to reference standard.

2.3.10 Binding Affinity to Complement Component 1q (C1q) by BLI

To determine the ability of test samples to bind to C1q, BLI with multiple-cycle kinetics was performed on an Octet QKe platform using a sensor chip with pre-immobilized Protein L sensors (Sartorius) to capture samples via their light chain. The percentage of relative binding was reported by comparing the affinity constant KD (which was obtained by the 1:1 kinetic model fitting) from test samples against the reference standard. Data were collected and processed with Octet Data Analysis software (Sartorius).

2.3.11 Binding Affinity to FcγRIa by SPR

FcγRIa binding was assessed by SPR with multiple-cycle kinetics on a Biacore™ T200 (GE HealthCare), using a sensor chip with pre-immobilized Protein A (GE HealthCare). Reference standard and test samples were assessed at varying concentrations with the affinity constant KD obtained using the 1:1 kinetic or steady-state fitting model. Results were reported as a percentage of affinity relative to reference standard. Data were collected and processed with Biacore T200 Evaluation software (GE HealthCare).

2.3.12 Binding Affinity to FcγRIIIb by SPR

FcγRIIIb binding was assessed by SPR with multiple-cycle kinetics on a Biacore™ T200 (GE Healthcare) using a CM5 Chip (GE Healthcare) with pre-immobilized anti-His Ab (GE HealthCare) to capture human FcγRIIIb conjugated with His tag (R&D Systems). Samples were tested at various concentrations in parallel and reported as affinity constant KD by the steady-state model fitting. Data were collected and processed with Biacore T200 Evaluation software (GE HealthCare).

2.3.13 Antibody-Dependent Cell-Mediated Cytotoxicity (ADCC)

ADCC activity of test samples was evaluated by two reporter gene assays and one peripheral blood mononuclear cell (PBMC)-based lactate dehydrogenase (LDH) cytotoxicity assay. For the reporter gene assay, mIL-6R-expressing TF-1 and HEK-Blue™ IL-6 cells were used as target cells, and Jurkat cells overexpressing FcγRIIIa (158V) and nuclear factor of activated T cells (NFAT) response elements-driven luciferase (Jurkat/NFAT-luc+ FcRγRIIIa [158V]) were used as effector cells. For the PBMC-based LDH cytotoxicity assay, TF-1 cells and PBMC were used as the target and effector cells, respectively. Raji cells (CD20 expressing cell line) with ofatumumab (an anti-CD20 monoclonal antibody, Novartis) were used as a positive control, while pertuzumab (an anti-Her2 monoclonal antibody, Novartis) was a negative control. Briefly, target and effector cells were seeded in 96-well plates and incubated with a serial dilution of test samples. Luciferase or cell cytotoxicity were detected using a Bio-Lite™ Luciferase Assay System (Vazyme) or CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega) and quantified with a SpectraMax M5e microplate reader (Molecular Devices). Data were collected and processed with SoftMax Pro software (Molecular Devices).

2.3.14 Complement-Dependent Cytotoxicity (CDC)

The CDC activity was evaluated using TF-1 and HEK-Blue™ IL-6 cells as target cells and human serum complement. Raji cells with ofatumumab were used as a positive control and pertuzumab as a negative control. Briefly, target cells were seeded in 96-well plates and incubated with a serial dilution of test samples, followed by human complement (Quidel). Cell viability was detected using Cell Counting-Lite 2.0 Luminescent Cell Viability Assay (Vazyme) reagent and quantified with a SpectraMax M5e microplate reader (Molecular Devices). Data were collected and processed with SoftMax Pro software (Molecular Devices).