Antibody interference in immunoassay: know your enemy

Antibody interference in clinical immunoassay continues to alarm clinicians, confound immunoassay manufacturers and challenge clinical biochemists. Every year sees the publication of case studies and articles describing situations where immunoassays have returned incorrect results and the adverse effect this can have on patient care. This month’s edition of the Annals contains three such articles.^1–3 Essentially, antibody-related immunoassay interference is caused by circulating antibodies that can bind to either the analyte or the immunoassay reagents potentially generating incorrect results (see ref. ⁴ for a comprehensive review). The incidence of assay interference is difficult to establish and is both assay and analyte dependent. Thyroglobulin immunoassay is notoriously susceptible to interference; results from the method comparison study cited here suggest an incidence of 12%. ¹ For thyroid-stimulating hormone (TSH) immunoassay, the incidence of 0.6% using polyethylene glycol (PEG) precipitation and gel filtration chromatography (GFC) studies ² is equally disturbing, given the widespread use of this assay as a ‘front-line’ test for thyroid dysfunction. This is not dissimilar to published values. ⁵

While the mechanism of antibody interference is often considered academic, once established it can be used as a basis to improve assay design or guide the selection of alternative methods. Interfering antibodies can be specific to the assay reagents (anti-animal antibodies) or the analyte (autoantibodies if to an endogenous molecule). Weak polyspecific antibodies (heterophilic antibodies) are also a common cause of antibody interference as, in accordance with Murphy’s law, they too can have affinity for analytes or assay reagents. Falsely increased results can be generated with competitive immunoassays by antibodies which sequester tracer or block the assay antibody. Cross-linking of capture and detection antibody in two-site assays also leads to false-positive results, and this is the likely mechanism for the false-positive adrenocorticotrophic hormone (ACTH) result described here. ³ Negative interference is usually caused by blocking the binding of either the capture or detection antibody in two-site assays, which is a likely cause of thyroglobulin assay interference. ¹ Falsely low results in competitive assays are rare, but have been described; in this case the interfering antibody must prevent displacement of the tracer from the detection antibody. Anti-analyte antibodies usually present as falsely elevated results as the analyte:antibody complex can be immunoreactive, but not biologically active. The clearance of the analyte:antibody complex is often delayed compounding the effect. These are widely referred to as ‘macro-analytes’; this is a plausible mechanism for the falsely elevated TSH cases described here. ²

Considerable effort has gone into minimizing the risk of antibody interference and most commercial immunoassay kits contain reagents to address this. Anti-animal antibodies can be specifically targeted using excess animal immunoglobulin without affinity for the analyte in question. Heterophilic blocking agents are more mysterious, as their composition is commercially sensitive, but these can be antibody fragments designed to specifically block the invariant tail region (Fc) region of antibody reagents, where heterophiles usually bind. The use of antigen binding (FAb) fragments, rather than intact antibodies, also protects against heterophile interference using similar logic.

There are several tools that laboratories can use to confirm the presence of assay interference. Unfortunately these lack sensitivity, so a variety of methods is often required. Method comparison, immunosubtraction and GFC are used in these three articles.^1–3 When considering method comparison it is best to use a genuinely independent method. Unfortunately, many laboratories currently use rather similar methods for comparison studies. For example, the use of a second one-step method (competitive rather than back titration) to check free thyroxine (fT4) results will fail as both assays will be similarly affected by anti-T4 antibodies. Mass spectrometric (MS) methods, where applicable, are an obvious choice when checking small molecule immunoassays: indeed these approaches have started to replace immunoassay as front-line methods. Peptide MS assays, while more challenging, are now becoming available and these can be used to check immunoassay results. These have been described for the assay of thyroglobulin. ⁶ If immunoassay is to be used as a comparator then a different architecture is preferable. Crane et al. ¹ use a competitive assay to check results from a two-site assay. The pattern of results from different assays can give insight into the nature of the interference. ‘Macro-analytes’ typically give higher results in assays with higher affinity capture antibodies, longer incubation times and greater sample dilution as the ‘macrocomplex’ is more likely to dissociate during assay. PEG precipitation is the most commonly used immunosubtraction method and while not without problems, ⁷ it is widely accepted for the detection of macroprolactin but also a variety of other analytes, including TSH. ² More specific methods such as anti-immunoglobulin agaroses are available but are unfortunately more expensive. Proprietary heterophile blocking reagents are available (e.g. http://scantibodies.com/ scantibodies – accessed 1 July 2013) and these can be useful in the detection of heterophiles.. However results are not always predictable ³ and the effect of any dilution incurred needs to be considered. This type of reagent is included by many assay manufacturers as a matter of course. As this method is insensitive it should not be used in isolation. For example, macrohormone interference is unlikely to be detected.

Linearity and recovery studies were not used in these three publications^1–3; however, they are simple and often enlightening. In the presence of heterophilic antibodies a dramatic reduction in apparent concentration is often observed upon dilution; this is due to an overwhelmed blocking agent becoming effective as the interfering heterophile is diluted. Dilution of macro-analyte complexes can give spectacular increases in apparent concentration as large amounts of analyte can be released from low affinity high capacity macrocomplexes. Cross-linking heterophiles can also show a more subtle increase in apparent concentration on dilution due to a complex mechanism akin to pH buffering.

The three cases presented here highlight the need for ‘total’ and ‘bioactive’ analyte assays. For TSH and ACTH the clinician is typically interested in the biologically active hormone, as this is likely to correlate with the clinical presentation. For tumour markers such as thyroglobulin total analyte concentration is more relevant. Thus harsh physical methods such as tryptic digestion followed by MS analysis can be used for the quantitation of thyroglobulin, as antibody interference is negated by the proteolytic digestion ⁶ and ‘total’ analyte concentration is returned. More challenging is the estimation of biologically active (or ‘free’) hormones. Even GFC, which is widely touted as the ‘gold standard’ method for macro-analyte analysis, is not fool-proof as the dilution incurred during filtration can affect the ratio of antibody-bound to ‘free’ analyte. GFC, as stated by Mills et al., ² can also be confounded by cross-linking antibodies, which can also present as a high molecular mass immunoreactivity.

There is clear variation in the approach to antibody interference between clinical laboratories. As the problem is complex it is difficult to give prescriptive guidance but a genuinely independent method comparison (where possible), immunosubtraction plus linearity and recovery studies is a reasonably comprehensive approach. Otherwise the consolidation of assay platforms across wide geographical regions, commercialization of laboratories and reduction in the amount of time spent on clinical validation may herald the perfect storm for antibody interference in immunoassay.