2023 FDA Science Forum
Proteomics identifies multiple potential pharmacodynamic biomarkers for interferon β-1a biologics
- Authors:
- Center:
-
Contributing OfficeCenter for Drug Evaluation and Research
Abstract
Background: Pharmacodynamic (PD) biomarkers may be used for similarity assessment of proposed biosimilars without relying on costly and time-consuming clinical studies with efficacy endpoints. Proteomics has the potential to identify novel PD biomarkers for biologics, especially those with complex mechanism of action and previously uncharacterized PD biomarkers.
Purpose: To evaluate the utility of proteomics for identifying PD biomarkers for interferon beta-1a biologics (IFNβ-1a and pegylated IFNβ-1a), that are approved to treat multiple sclerosis.
Methodology: In this study, 36 healthy subjects randomized to therapeutic doses of IFNβ-1a (n=12 [30µg IM]), pegIFNβ-1a (n=12 [125µg SC]) and placebo (n=6 each) were profiled for >7000 plasma proteins at baseline/pre-treatment and at 9 timepoints, over 6 days in the IFNβ-1a group, and at 11 timepoints, over 13 days in the pegIFNβ-1a and placebo-specific groups, using the SOMAscan® assay v4.1 (SomaLogic). We identified differentially expressed proteins (DEPs) using linear-mixed effect models and ANOVA, as analytes with treatment*time interaction p-values<6.8x10E-6 (Bonferroni [BF]) adjusted p<0.05). We further prioritized signals based on biological relevance, maximal fold change ≥2 from baseline, and a significant difference (t-test BF adjusted p<0.05)) in baseline adjusted area under the effect curve (AUEC) from placebo, with both products. Analysis was conducted in R (v4.1.2).
Results: We identified 248 and 528 DEPs by IFNβ-1a and pegIFNβ-1a respectively, of which, 31 analytes were prioritized based on maximal fold change ≥2 from baseline and overlap between the two products. Of these, the majority had a significant AUEC compared with placebo in response to either product; 8 proteins showed > 4-fold maximal change from baseline. We identified previously reported candidates, beta-2microglobulin and interferon-induced GTP-binding protein (Mx1) with ~ 50% coefficient of variation (CV) for AUEC, and many new candidates (including I-TAC, C1QC, and IP-10) with CVs ranging from 26%-129%. Upstream regulator analysis of DEPs predicted activation of IFNB1 signaling as well as other cytokine, enzyme and transcription signaling networks by both products.
Conclusion: Using proteomics, we identified several plasma proteins as potential PD biomarkers of IFN-β1a biologics for further investigation to support biosimilar development programs.
