Background There are five different anti-TNF biologics: three are bivalent full length (FL) antibodies (adalimumab, golimumab, and infliximab), one a bivalent fusion protein (etanercept), and one a univalent PEGylated Fab' (PF) (certolizumab pegol [CZP]). Administration of such protein biologics can induce anti-drug antibodies (ADAbs), of which the majority are anti-idiotypic antibodies (anti-ID).1 The potential cross-linking of bivalent anti-IDs with bivalent biologics can result in the formation of large immune complexes (ICs), which are subsequently cleared by mechanisms using the multiple Fc domains, thereby resulting in a decrease in the efficacy of the biologic. Since univalent biologics, such as CZP, only have one binding Fab' arm, such large cross-linked anti-ID-mediated ICs are unlikely to form. Therefore, anti-IDs may have a different effect on the elimination and bioavailability of univalent and bivalent biologics in vivo.
Objectives To determine if the valency of a biologic will affect the in vitro size and in vivo elimination from the plasma of ICs formed with an anti-ID following intravenous (IV) administration to BALB/c mice.
Methods An anti-ID antibody to CZP was generated and used for subsequent studies to mimic an ADAb response. Univalent PF CZP was reengineered as a bivalent FL humanized IgG1 antibody (similar to adalimumab) to directly compare the effect of valency on IC-mediated clearance. This FL antibody showed very similar TNF neutralization compared to the conventional PF molecule in a bioassay. To generate ICs, the anti-ID was incubated overnight with either PF or FL CZP, and the size of the resultant ICs determined by analytical ultracentrifugation (AU). The anti-ID complexes and PF and FL CZP alone were then administered IV to the mice, and the elimination of the anti-TNFs from the circulation was monitored by quantitative liquid chromatography-mass spectrometry (LC-MS) in serial plasma samples.
Results AU analysis of the immune complexes formed between PF CZP and the anti-ID showed the presence of one peak corresponding in size to one anti-ID molecule bound to two PF molecules (∼3.5x105 Da). In contrast, the FL CZP/anti-ID mixture showed ICs of various sizes up to very large molecular weights (>1x106 Da), with the predominant species corresponding to a complex of two anti-IDs bound to two FL CZP molecules (∼6x105 Da). The in vivo studies showed that the FL CZP/anti-ID ICs were eliminated much faster (t1/2=0.3 hours) than the FL CZP alone (t1/2=44.5 hours), whereas the PF CZP/anti-ID ICs were eliminated more slowly (t1/2=60.5 hours) than the PF CZP alone (t1/2=19.7 hours).
Conclusions The FL CZP molecule formed large ICs with the anti-ID, which led to much faster elimination than the FL molecule alone. This result suggests that an ADAb response to a FL antibody could lead to rapid elimination and loss of efficacy of the drug in patients. In contrast, the PF CZP/anti-ID complex had a longer half-life than the PF CZP alone, presumably because this molecule was not seen as an IC due to the presence of only one Fc. These results showed that an ADAb to a univalent biological reagent may not lead to elimination and instead, may actually increase the in vivo half-life of the molecule.
van Schouwenburg PA. Ann Rheum Dis 2013;72:104–9.
Acknowledgements This study was funded by UCB Pharma. Editorial services were provided by Costello Medical Consulting.
Disclosure of Interest J. Silva Employee of: UCB Pharma, A. Nesbitt Employee of: UCB Pharma