New examples of antibody-based assay interferences

Changes in medical practice may affect immunoassay-based determinations of drugs and diagnostic markers. In this issue, Dayaldasani et al. ¹ examine the impact on insulin assays of modern insulin analogues that had been synthesized by recombinant DNA technology and designed to alter insulin pharmacodynamics. In essence, cross-reactivity is the issue here. By testing at several concentrations, the profile of interference was established for each analogue with three different insulin assays. In some cases the relationship was non-linear, illustrating the need for this experimental approach and in contrast to the common commercial practice of defining cross-reactivity at a single, typically supraphysiological, concentration. The procedure followed the Clinical Laboratory Standard Institute guideline EP7-A2. ² Knowledge of the cross-reactivity of different analogues enables interpretation of results from treated individuals with diabetes and may also be useful in detecting factitious hypoglycaemia.

In general laboratory use, commercial reagent manufacturers do not provide, and are often unwilling to issue, detailed information on assay formulation and performance. Consequently, clinical laboratories have to generate their own data to justify use in particular clinical circumstances. The biomedical profession is familiar with many areas where cross-reactivity generates results that influence patient care, for better or worse. Steroids are probably the largest group of substances so affected because of the close chemical structures of parent hormones, their metabolites and synthetic analogues.

The repertoire of clinical biochemistry tests is expanding to signalling pathways, an area where specificity problems in quantitative measurements, particularly of receptors, have been exposed with histochemical tests. Laboratories often make use of immunoassay reagents for quantitative assays that were first generated for pharmaceutical industry use in drug trials. In the gastrointestinal tract, for instance, a family of receptors has been identified; the receptors are activated by proteinases – the so-called proteinase-activated receptors (PAR). Thrombin and glucagon are such proteinases in the coagulation process and glucose homeostasis, respectively. Other activators participate in electrolyte transfer across membranes, inflammatory reactions, protection of mucosal surfaces, modification of gut motility and secretory function. There is a strong evidence of the roles of PAR₁ and PAR₂ in irritable bowel syndrome and disease. ³ Modified PARs have been synthesized for pharmacological use as blockers of in vivo activation of receptors, even though their clinical relevance in intestinal disease is not yet clear. The PAR receptors fall within a family of G-protein coupled receptors (GPCR). The activated receptor can stimulate a number of cellular processes, including adenyl cyclase, phosphatidylinositol-phosphate, phosphatidylinositol-3-kinase and cyclic AMP phosphodiesterase activities. PAR receptors can be over-expressed, leading to increased plasma concentrations in certain cancers. ⁴ Commercial assays are available, but assay specificity for PAR has been challenged when commercial kits are compared. The angiotensin II type 1 receptor is a widely studied GPCR. Western blot analysis and immunochemistry showed inconsistencies in tissue specificity. ⁵ Another comparison of angiotensin II type 1 receptor antibodies concluded that results were variable, unpredictable and, importantly, unreliable. ⁶

GPCR are large molecules, with a large hydrophobic membrane spanning region and a short loop exposed on the surface of cells. Antibody specificity can be directed to specific regions of the molecule, the tails or extracellular loops. Recombinant protein will be the source of antigen for antibody generation and calibration. Protein modifications such as glycosylation ⁵ will also influence antigenicity, and the protein may take up a different tertiary structure in vitro compared with the circulating or tissue conformation. Fragments of the protein may also exist in the circulation. The assay manufacturer does not often, and can be reluctant to, provide detailed information on the antigen/immunogen basis for the test or assay interferences. Hundreds of different genes for GPCR have been predicted from knowledge of the genome sequence, so the usual tests for interference would be impossible. Many issues of assay specificity have arisen from immuno-histochemical use of GPCR antibodies. Binding has been seen in histochemical evaluation tests with tissues where no receptor is expressed. Other investigations performed with gene knockout animal tissues, a cell model with gene over-expression and experiments in the presence of antagonists, endorsed the lack of antibody specificity. ⁵

For quantitative measurements of receptors in serum or plasma, the laboratory should satisfy itself whether an assay is fit for purpose and quality. The mechanism for transfer of membrane and intracellular receptor products to blood needs to be clear in the case of normal and disease tissues. Without knowledge of the antigen(s), there is a risk of the assay performance being compromised. Comparison of reagents from two or more sources may help to assure the user of assay performance. Recommendations for adaptation and validation of commercial assays have been reviewed to incorporate discussion of the draft 2013 Food and Drug Administration (FDA) guidance. ⁷ There is a long recognized need for better characterization of antibodies and immunoassay components. Affinity purification of the antibody adds confidence to the reagent quality. ⁵ The binding characteristics, association and dissociation, of an antibody can be tested, for example, with surface plasmon resonance techniques. ⁸ Tandem mass spectrometry used, for example, for steroid analysis has improved assay specificity. As protein assays become more amenable to this technology, the commercial immunoassay-based kits may decline in use in favour of improved specificity: until then, immunoassay will be used. Stable-isotope labelled internal standards with selective reaction monitoring mass spectrometry are becoming suitable for quantitation of biomarkers. ⁹

Changes in gender and sample matrix affect assay performance for many analytes (e.g. cortisol ¹⁰ ); species and reagent batch ¹¹ are also sources of assay interference, leaving the user to validate method performance for the specific clinical requirement. Reference ranges, matrix effects and reference material for quality control in the long term are other issues to be addressed. ¹² ‘Magic peptides’, ¹³ ‘the good, bad and really ugly’ ¹¹ and ‘… reagents of mass distraction’ ¹⁴ are all terms that have been applied to antibodies where use has led to erroneous findings. To this, I suggest could be added ‘buyer beware’.