Bioanalytical Method Validation (OBX-319)

Document ID: BIOANALYTICAL-001
Version: 1.0
Change History: 1.0 — Initial issue.
Standard(s): ICH M10

Bioanalytical Method Validation — OBX-319

Validation of the bioanalytical assays for OBX-319 in biological matrices (and the anti-drug-antibody assays), covering selectivity, calibration, accuracy and precision, stability, and (for the ligand-binding/immunogenicity assays) the tiered strategy. ICH M10.

OBX-319 is a humanized IgG1 anti-CD19 × anti-CD20 bispecific monoclonal antibody produced in Chinese hamster ovary (CHO) cell culture and administered subcutaneously. Because both binding arms engage membrane-bound B-cell antigens, the molecule exhibits target-mediated drug disposition (TMDD): systemic clearance is time-varying and load-dependent, accelerating while a large circulating and tissue B-cell pool is present and slowing as CD19+ B cells are depleted. The bioanalytical package was therefore designed to (i) quantify OBX-319 across a concentration range wide enough to characterise the nonlinear disposition, (ii) confirm that the circulating analyte retains both binding functionalities, (iii) detect and characterise anti-drug antibodies (ADA) through a tiered immunogenicity strategy appropriate to an autoimmune (SLE) population, and (iv) support the pharmacodynamic (PD) read-outs of B-cell depletion and serological normalisation. All quantitative pharmacokinetic (PK) and immunogenicity methods were validated in accordance with ICH M10; supporting PD biomarker assays were qualified fit-for-purpose against the same framework.

Analytes, Matrices, and Critical Reagents

• Analyte: OBX-319 (intact bispecific IgG1), measured in human serum. The clinical reference standard is the qualified drug-substance lot, cross-referenced to the primary reference standard used in the ICH Q6B analytical control strategy so that bioanalytical calibrators are traceable to the release potency standard.
• Matrix: Serum was selected as the primary PK matrix; K2-EDTA plasma was bridged for exploratory sampling. Individual and pooled lots from SLE donors were included during validation because the intended-use population carries elevated polyclonal IgG, antinuclear antibodies (ANA), rheumatoid factor (RF), and other autoantibodies that can perturb ligand-binding assays.
• Critical reagents. The anti-CD19 arm is supported by soluble recombinant human CD19 extracellular domain (a single-pass type I membrane protein whose ectodomain is readily expressed). CD20 is a four-transmembrane protein with only a small extracellular loop and is not tractable as a soluble antigen; the anti-CD20 arm is therefore supported by an affinity-purified anti-idiotypic antibody raised against the anti-CD20 paratope (and, where a native conformational epitope is required, by CD20-expressing membrane preparations). Each critical reagent (recombinant CD19, anti-idiotypes, labelled conjugates, positive-control ADA) is governed by lot-specific certificates, retest dating, and a documented reagent-bridging protocol so that lot changes do not shift assay performance.

Pharmacokinetic Assay — Serum Quantification of OBX-319

A validated electrochemiluminescence (ECL) ligand-binding assay quantifies OBX-319 in serum. The reportable method captures analyte via recombinant human CD19 ectodomain and detects bound drug with a labelled anti-human IgG Fc (or anti-idiotype) conjugate, yielding a "CD19-arm-competent, Fc-intact" measurement of the intended therapeutic. Method characteristics validated under ICH M10:

• Calibration. A minimum of six non-zero calibrator levels plus anchor, blank, and zero samples were fitted with a weighted four-parameter logistic (4-PL) regression. At least 75% of calibrators (and both LLOQ and ULOQ) met back-calculated accuracy within ±20% (±25% at LLOQ/ULOQ) of nominal.
• Accuracy and precision. Validation QCs were prepared at five levels (LLOQ, low ≤3× LLOQ, mid, high ≥75% ULOQ, ULOQ). Inter- and intra-run mean recovery fell within ±20% of nominal (±25% at LLOQ/ULOQ) with precision (%CV) ≤20% (≤25% at the LLOQ/ULOQ), and total error remained ≤30%, consistent with ligand-binding acceptance criteria.
• Selectivity and specificity. Individual normal and disease-state (SLE) serum lots, including hyperlipidaemic, haemolysed, and high-RF/high-ANA lots, were spiked at the LLOQ and evaluated for interference; the assay met selectivity acceptance in ≥80% of lots. Specificity against endogenous IgG and against the anti-idiotype critical reagents confirmed absence of cross-signal.
• Dilutional linearity and hook effect. High-concentration samples were serially diluted into surrogate matrix to establish dilutional linearity and the minimum required dilution (MRD); the assay was demonstrated free of a high-dose hook (prozone) across the anticipated Cmax range following subcutaneous dosing.
• Parallelism. Incurred (post-dose) samples diluted in parallel to the calibrators, supporting the validity of interpolated concentrations and the absence of matrix-driven non-parallel response.
• Stability. Analyte stability was established for bench-top (room temperature), refrigerated processing, multiple freeze–thaw cycles, and long-term frozen storage bracketing the study sample storage interval, plus stock and working-solution stability of critical reagents. Whole-blood stability supported the sample-handling window at clinical sites.
• Incurred sample reanalysis (ISR). A prespecified proportion of study samples (including peak and elimination-phase time points) was reanalysed; ≥67% agreed within 30% of the original result, meeting the ligand-binding ISR criterion and confirming reproducibility in authentic samples.

Because disposition is target-mediated, a total-drug format (Fc capture/Fc detection) was additionally characterised so that TMDD modelling could be informed by both target-competent and total exposure; the difference between formats is interpreted against the declining B-cell target sink over the 52-week dosing period.

Confirmation of Bispecific Integrity — Dual-Binding Assay

To confirm that circulating drug retains both functionalities (and to distinguish intact bispecific from any single-arm degradant), a dual-binding bridging assay was qualified: capture on recombinant human CD19 ectodomain with detection via the anti-idiotype to the anti-CD20 arm. Only molecules that simultaneously present a functional anti-CD19 paratope and a functional anti-CD20 paratope generate signal. This assay supports interpretation of the PK data and links the clinical exposure measurement to the Q6B potency and identity attributes of the bispecific.

Immunogenicity — Tiered Anti-Drug Antibody Strategy

ADA against OBX-319 were assessed with a validated, multi-tier ECL bridging strategy (screen → confirm → titre → neutralising), consistent with ICH M10 and prevailing health-authority immunogenicity expectations. The autoimmune indication was explicitly addressed because SLE sera frequently contain RF, ANA, heterophilic antibodies, and pre-existing reactivity that can produce false-positive bridging signal.

• Screening. A bridging format (biotinylated OBX-319 capture / ruthenylated OBX-319 detection on a streptavidin plate) detects ADA of any isotype. A statistically derived screening cut point was set to yield a nominal 5% false-positive rate using drug-naïve individual lots (including SLE lots), with fixed- or floating-cut-point handling justified by the validation data distribution.
• Confirmatory. Reactive samples were competed with excess unlabelled OBX-319; specific reduction in signal above a confirmatory cut point (nominal 1% false-positive rate) confirms drug-specific antibody and screens out RF/heterophilic and matrix artefacts. Where warranted, arm-specific competition (soluble CD19 versus anti-CD20 anti-idiotype) supports domain-level characterisation of the antibody response.
• Titre. Confirmed-positive samples were titrated to an end-point titre for magnitude, enabling treatment-emergent versus boosted classification against baseline.
• Neutralising antibody (NAb). A competitive ligand-binding NAb assay assesses interference with target engagement; because OBX-319 has two independent paratopes, neutralisation is characterised for each binding arm so that a NAb blocking only one arm is not misreported as fully neutralising.
• Assay performance characteristics. Sensitivity was established with an affinity-purified polyclonal positive control to a target of ≤100 ng/mL (with 500 ng/mL as an acceptable threshold), and drug tolerance was quantified at low and high positive-control levels so that ADA detection in the presence of persistent circulating drug is understood across the dosing interval. Selectivity, precision (intra-/inter-run), and robustness (including RF-spiked lots) were validated; pre-existing (baseline) reactivity and its relationship to CHO-derived product were characterised. The tiered results integrate with the PK and clinical safety review of injection/infusion reactions.

Pharmacodynamic Biomarker Assays

• CD19+ B-cell enumeration (flow cytometry). Circulating B-cell depletion is the pivotal PD marker; validation supported a measuring interval spanning the observed on-treatment change from an approximate baseline of ~210 cells/µL to nadir values of ~7 cells/µL on active arms. A critical, modality-specific control was applied: because the therapeutic occupies CD19 and CD20 epitopes, detection antibodies were selected as non-competing clones binding CD19 (and CD20) epitopes distinct from those of OBX-319, and B-cell identity was corroborated with additional lineage markers (e.g., CD45, HLA-DR, CD22) to prevent drug-induced epitope masking from being misread as depletion. The panel was qualified fit-for-purpose for lineage gating, absolute counting, lower limit of quantification, and specimen-stability window.
• Serological markers. Anti-double-stranded DNA (anti-dsDNA) and complement C3/C4 were measured by validated clinical immunoassays supporting the observed pattern of falling anti-dsDNA and normalising C3/C4 in responders; these methods followed the same accuracy/precision and stability discipline and were bridged to the central-laboratory reference ranges used in the efficacy analysis.

Cross-Study Consistency and Documentation

Method transfers between the nonclinical program (cynomolgus monkey, the sole pharmacologically relevant species owing to absence of rodent target cross-reactivity) and the clinical program were controlled by species-specific validation of the PK and ADA assays, with shared critical reagents cross-qualified where applicable. Reagent lot changes, cut-point re-establishment, and any partial validations were documented under change control. All PK and immunogenicity study samples were analysed under GLP/GCLP-aligned procedures with SOP-governed acceptance, run-acceptance, and reanalysis rules, and the validation reports are cross-referenced to the assay control strategy (ICH Q6B) and to the integrated PK, immunogenicity, and safety summaries in this submission. ICH M10.