New Dogs/Old Tricks: Methods for Assessing the Immunogenicity of Peptide Drugs and Their Impurities
Brian Roberts
Scientific Director, Protein Therapeutics, Epivax
"In order to address the recent draft guidance from the FDA regarding the potential for
immunogenic impurities in generic drug products, we have adapted existing in silico prediction
tools and in vitro laboratory assays to identify potential T cell epitopes (in silico), confirm HLA
binding potential in vitro and determine the ability of peptide drug substances (DS) and their
impurities to generate a T cell response in HLA typed donor PBMC assays. We refer to this three
step approach as PANDA, each step of this analysis is described below.
Step 1. Immunoinformatics assessment: The potential of the DS to stimulate a T cell response
can be rapidly assessed computationally using T cell epitope mapping algorithms, such as
EpiMatrix. The EpiMatrix algorithm is used to screen the primary amino acid sequence of the DS
and its impurities for the presence of HLA DR ligands, which can be considered putative T cell
epitopes. Discrimination between potential inflammatory “T effector†epitopes and regulatory “T
reg†epitopes is performed with a second algorithm known as JanusMatrix which looks for crossconservation with endogenous proteins at the TCR-facing residues. Following the identification of
putative T cell epitopes, the next step is to combine the scores for effector and regulatory T cell
epitope content, providing an overall assessment of immunogenic potential for the DS and
impurities. Treg-adjusted EpiMatrix scores are highly correlated with immune responses in vivo.
Step 2: In vitro - HLA binding: Peptides derived from the DS and impurities that encompass the
predicted T cell epitope content can be evaluated for binding to HLA in assays that measure
binding affinity in dose-ranging studies, in vitro. HLA binding is used to confirm the in silico
analysis and inform the design of additional in vitro assays.
Step 3: In vitro– ex vivo T cell Assay and the measurement of de novo T cell response: In
vitro stimulations of human peripheral blood mononuclear cells using either the drug product or
individual peptides allow for natural antigen processing of the DS and other product components
including impurities. In vitro assays using predicted epitopes derived from the DS product
impurities provide information about the ability of these defined sequences to drive a T cell
response using human T cells. At the end of the culture period, T cell phenotype and/or function
are characterized in assays that measure the magnitude and quality of effector T cells that have
potential to drive ADA development.
Combined, these three steps comprise the orthogonal “PANDA†approach to addressing the
immunogenic impurities within the context of the FDA ANDA draft guidance"