Anti-drug antibodies (ADAs) can be induced in vivo in individuals treated with biotherapeutics. Detection of ADAs and evaluation of the effect of ADAs is important for the assessment of drug efficacy, drug safety and the risk of developing immunopathological reactions in treated individuals. In this context, the stability of ADAs in serum samples collected during clinical trials is important as samples may undergo several handling steps from the point of blood sampling to final analysis in the laboratory. The handling of samples may include transportation, separation of serum and other blood components, aliquoting, and further shipment to other laboratories where they may be stored, frozen and thawed several times in-between different rounds of analyses and re-analyses before a study is completed.

It is important that qualitative and quantitative results of antibody analyses are consistent over time and that investigators can trust that measurements made at different points in time are comparable to the values that could be obtained at a time just after blood sampling. As analyses of ADAs are performed as a tiered approach, the investigations of ADA positive samples may have to be performed over periods of time that require intermediate storage of the samples, − in which case knowledge about the stability of the ADAs at different storage conditions is important.

Currently there are no specific guidance on how to perform experimental studies for sample stability regarding ADAs. Although several publications mention sample stability with respect to antibodies, this topic is not described in detail in the guidelines on immunogenicity assessment issued by the leading authorities in the field (EMA, 2017; FDA, 2014; FDA, 2019). The FDA and EMA Guideline on bioanalytical method validation (EMA, 2011; FDA, 2018) provides specific instructions on how to evaluate the stability of drug analytes. As every bioanalytical assay measure a different analyte with different structure and physicochemical properties, it is always important to test analyte stability during assay validation. Contrary, in immunogenicity assays the analyte will have similar structure and physicochemical properties, thus assay-specific stability testing is not always relevant given the amount of stability data already available in the literature. Shankar et al. pointed out that stability of ADAs specific to different drugs is the same (Shankar et al., 2008), which was also accepted by others (Pihl et al., 2014; Hendriks et al., 2014; Boridy and Xing, 2019). Thus, studies on the stability of ADAs provide results that are generally applicable as opposed to results from drug analytes. Stability of antibodies in serum after recurrent freeze-thaw events (FT) have been studied previously (Hendriks et al., 2014; Boridy and Xing, 2019; Michaut et al., 2014; Pinsky et al., 2003; Castejon et al., 2017). Even though it is generally accepted that ADAs are stable after three FT (United States Pharmacopeia and the National Formulary (USP-NF) General chapter 1106, 2013), studies have shown antibody stability after at least 12 FT for ADAs (Michaut et al., 2014), after 10 FT for anti-virus antibodies (Pinsky et al., 2003), after up to 176 FT for anti-treponema antibodies (Castejon et al., 2017) and after eight FT for binding anti-vaccine antibodies (AVAs) and after three FT for functional AVAs (Hendriks et al., 2014). None of the studies on recurrent FT cited here reached a number of FT that had a negative effect on the stability of the antibodies measured.

Human insulin and human insulin-analogues are important biopharmaceuticals in the treatment of diabetes type 1 and type 2 (Mathieu et al., 2021). The occurrence of anti-insulin antibodies in subjects treated with exogenous insulin is well described and the impact of these ADAs to insulin have been studied throughout the modern history of diabetes care (Fineberg et al., 2007; Radermecker et al., 2009; Yang et al., 2022). In the present study we investigate serum samples positive for antibodies to human insulin from individuals with diabetes type 1 or type 2, in order to generate stability data from clinical samples which can be applied to all ADA's.

Stability of anti-insulin antibodies in human serum was studied in experiments addressing the short-term stability during storage at room temperature for up to 72 h and at refrigeration for up to 2 weeks. The stability of ADAs was also investigated by mimicking the “snap freeze” procedure often used when shipping serum or plasma samples from one laboratory to another. Samples were quickly frozen by placing samples that had room temperature directly on dry ice for 4 h followed by storage at −20 °C.

Long-term stability of ADAs to insulin was studied by analysis of samples stored frozen at −20 °C for 2.7 years up to 6.3 years. Additionally, some of the samples included in the long-term stability investigations were subjected to four FT during the storage to examine the possible effect of recurrent FT on the stability of the ADAs.

Testing of serum samples spiked with positive control antibodies, QC samples, are commonly used for documentation of antibody stability (Pihl et al., 2014). However, as QC samples for analysis of samples from clinical trials typically consist of drug-specific polyclonal or monoclonal antibodies of animal origin diluted in human serum, the stability of such QC samples does not necessarily reflect the stability of the study samples. Therefore, it is of interest to gather more data on the stability of ADAs of human origin. It is the intention that this study should provide additional information on the stability of ADAs in human serum that can support knowledge on antibody stability in general and be referenced as part of submissions of validation data for future ADA assays to regulatory authorities.