Zinpentraxin alfa is a recombinant form of a native protein called pentraxin-2 (PTX-2), also known as recombinant human serum amyloid P (Raghu et al., 2022). Zinpentraxin alfa is being investigated as a potential treatment for fibrotic diseases such as idiopathic pulmonary fibrosis (IPF) and myelofibrosis (MF) (Duffield and Lupher Jr., 2010; Dillingh et al., 2013; Verstovsek et al., 2018; Raghu et al., 2019), and was granted FDA breakthrough therapy designation based on the data from a double-blind, multicenter phase 2 clinical trial that compared zinpentraxin alfa treatment with placebo in 111 IPF patients (Raghu et al., 2018). Like the native protein (Lupher Jr, 2012), zinpentraxin alfa is expressed and purified as a non-covalent, homo-pentameric glycoprotein. In native PTX-2, each glycan exhibits a typical complex biantennary structure with α2,6-linked bi-sialylated species being the predominant form. Zinpentraxin alfa also exhibits a typical complex biantennary structure as the predominant glycoform similar to the native protein, while additionally exhibiting α2,3 sialic acid linkages, which are typical for recombinant proteins that are expressed in CHO cell lines. Because the endogenous PTX-2 plays a significant role in modulating fibrosis, inflammation, and proliferation (MacDonald and Kilpatrick, 2006; Pilling et al., 2007; Castano et al., 2009), the consequences of neutralizing antibodies induced by zinpentraxin alfa treatment could be severe if ADAs cross-react to the endogenous counterpart.

Immunogenicity monitoring is essential in clinical trials as immunogenic responses to therapeutics have been found to result in altered drug efficacy and to trigger unwanted safety events that can differ among patients (Koren et al., 2002; Schellekens, 2002; Schellekens and Casadevall, 2004; Gunn 3rd et al., 2016). Therefore, immunogenicity analysis is mandated by regulatory agencies throughout the drug development, clinical studies, and post-launch phases (Agency, 2015; 2019). A variety of different assay formats have been developed and validated for ADA detection which include direct binding assays (Mire-Sluis et al., 2004), bridging immunoassays (Patton et al., 2005), affinity capture elution (ACE) assays (Bourdage et al., 2007), and precipitation and acid dissociation (PandA) assays (Zoghbi et al., 2015). Among these, the bridging immunoassay is the most commonly conducted format in which drugs conjugated with different tags are used as capture and detection reagents to form a sandwich with the ADAs. Advantages of the bridging assay format include good assay sensitivity, ability to detect a majority of ADA immunoglobulin isotypes (Aalberse et al., 1999), and operational simplicity for high throughput analysis (Hafeez et al., 2018; Singh et al., 2018; Lu et al., 2020). A disadvantage of the bridging assay format is its vulnerability to interference from therapeutic drugs, which can bind to ADAs and hinder the formation of a complex with the capture and detection reagents, potentially leading to false negative results. In addition, drug targets can also interfere, leading to either false positive results when multimeric soluble targets are present, or false negative results in the presence of monomeric soluble targets.

A direct electrochemiluminescent (ECL) assay using solid phase with extraction acid dissociation (SPEAD) sample treatment (Patton et al., 2005; Smith et al., 2007) was developed by Promedior (part of Roche, Basel, Switzerland) for the detection of anti-zinpentraxin alfa antibodies in human serum and implemented in a phase 2 study (Tremblay and Mascarenhas, 2021; Verstovsek et al., 2023) in subjects with primary myelofibrosis (PMF). To allow ADA binding, this assay requires overnight incubation of sample and biotinylated zinpentraxin alfa (Butterfield et al., 2010). The immune complexes are then captured on a streptavidin plate, followed by washing and acid treatment to break apart the ADA binding. Dissociated ADAs are transferred to a second plate for detection using Sulfo-tagged zinpentraxin alfa in a direct ECL assay. While this assay format exhibits good drug tolerance, the SPEAD treatment requires significant manipulation of the samples, which is time-consuming and potentially increases the assay variability and loss of low binding antibodies. In this regard, greater variability was observed in positive controls. Interference from proteins bound to the drug is another potential limitation of the SPEAD assay method due to the fact that this format does not confirm the existence of immunoglobulins directed to the drug, which could lead to a false positive signal. Considering the aforementioned limitations of the SPEAD method, the first format evaluated was the most commonly used homogeneous bridging immunoassay. However, this format had high background signal which led to poor assay sensitivity. This led us to evaluate alternative assay formats that could provide dependable ADA detection while meeting the analytical performance criteria laid out in regulatory guidance documents. Fig. 1 depicts a diagrammatic representation of each assay format.

The step-wise bridging format (Fig. 1a) is similar to the homogeneous bridging assay format with ADA and drug binding taking place on a solid surface. In this format, drug is coated on the plate, followed by sample incubation. Captured ADAs are detected using biotinylated drug.

The direct binding format (Fig. 1b) employs a similar protocol as the step-wise bridging format while using Protein A/G as the detection reagent, which binds IgG, IgA, and IgM with different affinities (Intaramat et al., 2016). In this assay format, an assay blocker such as ChonBlock is used at both plate blocking and sample dilution steps to reduce the non-specific background signal.

The total ADA format (Fig. 1c) employs the addition of drug to bind and capture ADAs before transferring to an ELISA plate pre-coated with a monoclonal antibody against the drug. This method makes it possible to detect both free and bound ADAs in samples regardless of the level of drug present in the sample. In this format, anti-human IgG monoclonal antibody (Clone R10Z8E9, Genentech) is used as the detection reagent.

The semi-homogenous format (Fig. 1d) employs incubating samples overnight with biotinylated drug to allow the binding of ADAs. The immune complexes are then captured on a streptavidin plate followed by ADA detection using HRP-conjugated Protein A/G.

In this paper, we present results from evaluation of the sensitivity, precision, and drug tolerance of the four ADA assay formats (Fig. 1). A subset of samples from a phase 2 study of zinpentraxin alfa in subjects with primary myelofibrosis was selected and analyzed using different assay formats to further evaluate and compare their performance. The objective of this study was to assess the critical parameters of each format to identify the one most suitable for subsequent validation and analysis of clinical samples in the future.