P055 Immunological cross-reactivity of anti-drug antibodies to adalimumab and ABP 501

Abstract

Background: ABP 501 AMJEVITA®; adalimumab) is an approved biosimilar to adalimumab (Humira®), a fully human recombinant monoclonal antibody. Amgen validated separate assays to independently detect ADA against ABP 501 and adalimumab; this testing strategy was applied to both binding and neutralising ADA detection to evaluate immunogenicity of ABP 501 compared with adalimumab. The goal of the present analyses is to assess the extent of ADA crossreactivity between the biosimilar and reference product among patients with rheumatoid arthritis and psoriasis in two pivotal Phase 3 studies. Methods: All protocol-defined antibody samples collected from drug-exposed subjects, irrespective of treatment group, were tested for binding antibodies against ABP 501 and adalimumab using a validated electrochemiluminescence-based assay. Binding ADA magnitude was expressed as signal-to-noise (S/N), defined as the mean signal of the study sample divided by the mean signal of the negative control analysed on the same plate. Samples positive for binding ADAs were then tested in a TNFα-target binding assay for neutralising activity. Neutralising antibody titer was reported as the highest dilution that tested positive in the assay. Validated assay parameters for both binding and neutralising assays such as assay cut point, sensitivity, precision, and drug tolerance for each assay were highly similar. Correlation of binding and neutralising antibody results (S/N or titer) between ABP 501 or and adalimumab assays was evaluated using the Pearson's correlation coefficient. Concordance of antibody results (positive/negative) was evaluated using the kappa statistic. Results: Irrespective of treatment group, the binding ADA results for ABP 501 and adalimumab assays correlated well, with a Pearson correlation coefficient >0.950 and concordance measure of (kappa > 0.860). In a small subset of samples that were not concordant, the distribution of sample reactivity was detected equally across all assays. The magnitude of ADAs was at or near the assay cut point, indicating low probability of clinical impact. Furthermore, neutralising antibody titer results correlated well between ABP 501 and adalimumab assays with a Pearson correlation coefficient > 0.780 and strong concordance (kappa > 0.884). Conclusions: No meaningful differences were observed in the detection of binding and neutralising ADAs in the different ADA assays, providing evidence of high similarity between ABP 501 and adalimumab.