Elsevier

Clinica Chimica Acta

Volume 411, Issues 17–18, 6 September 2010, Pages 1367-1370
Clinica Chimica Acta

Poor agreement between commercial ELISAs for plasma fetuin-A: An effect of protein glycosylation?

https://doi.org/10.1016/j.cca.2010.04.020Get rights and content

Abstract

Background

Fetuin-A is a circulating inhibitor of ectopic calcification. Low plasma levels have been associated in some studies with increased vascular calcification, aortic stiffness and mortality in patients with Chronic Kidney Disease (CKD). However, there are other studies examining the association of fetuin-A with vascular parameters and mortality, which do not show these associations. These conflicting data may be explained by methodological differences.

Methods

We compared plasma fetuin-A measurements made with two widely-used commercial fetuin-A ELISA kits (Biovendor, Modrice, Czech Republic; Epitope Diagnostics Inc., San Diego, US) in samples from patients with and without CKD. We evaluated the effect of differences in fetuin-A glycosylation status on assay specificity.

Results

Deming regression analysis showed poor agreement between methods (for CKD cohort: y = −0.05 + 2.52x, Sy|x = 0.099 g/L, R2 = 0.694). The Epitope Diagnostics kit demonstrated significant positive bias and greater specificity for deglycosylated fetuin-A relative to the Biovendor assay.

Conclusion

The apparently contradictory nature of reports of the association of fetuin-A with biological variables may reflect differences in the specificity of different ELISA methods for glycosylated plasma fetuin-A.

Introduction

Human fetuin-A (also known as α2-Heremans–Schmidt glycoprotein) is a 60 kDa glycoprotein which is secreted by the liver into blood plasma [1]. Fetuin-A is a member of the cystatin superfamily of cysteine protease inhibitors and like other Type 3 cystatins has extensive post-translational modifications [2]. Fetuin-A has very complex biology and is purported to function in a diverse array of physiological process. Recent work has focussed on its role as a systemic inhibitor of ectopic mineral deposition, particularly in the setting of chronic kidney disease (CKD) where patients suffer from a heavy burden of early and progressive arterial calcification. Widespread soft-tissue calcification is seen in fetuin-A knockout mice with renal impairment or fed a mineral rich feed [3] and in fetuin-A/apoplipoprotein E-deficient mice, fetuin-A has also been found to protect against atherosclerotic calcification [4].

We recently reported that lower plasma fetuin-A concentrations were a strong and independent predictor of progressive aortic stiffness in cohort with stage 3 and 4 chronic kidney disease (CKD) [5]. However, reports in similar cohorts are not congruent with our results [6], [7]. We sought to explain these discordant data, with the hypothesis that gycosylation of the native protein may be important to the biological action of the protein and that available ELISA assays may differ in their specificity for different glycosylated forms. In the setting of stage 3 and 4 CKD, reported concentrations differ by over 0.5 g/L [5], [6], [8]. In-house methods, based on indirect ELISA and nephelometry have been used extensively by some groups [9], [10], but the majority of publications report the use of one of two commercial ELISA kits manufactured by Biovendor (Modrice, Czech Republic) and Epitope Diagnostics Inc (San Diego, USA). We compared the analytical performance of these two assays, looking at patient and normal plasma and examined the effects of glycosylation of the protein, with reference to potential biological effects.

Section snippets

Study participants

We used 178 lithium heparin plasma samples collected at the baseline visit of patients enrolled in a prospective cohort study of cardiovascular risk in patients with CKD stages 3 and 4. The details of this study have been previously described [5]. In order to evaluate the effect of CKD status, samples were also obtained from a cohort of 78 otherwise unselected patients (50 male; mean age 58 ± 10.4 yr) in whom there was no history of type 2 diabetes mellitus nor biochemical evidence of renal

Results

The analytical characteristics of each ELISA kit are summarised in Table 1. We evaluated the agreement between results in two cohorts: firstly in 178 patients with stage 3 and 4 CKD (Fig. 1A and C) and secondly in a group of 78 patients without biochemical evidence of renal dysfunction or history of type 2 diabetes mellitus but whom were otherwise unselected (Fig. 1B and D). The mean plasma fetuin-A concentration in CKD cohort (Biovendor, 0.236 ± 0.076 g/L; Epitope Diagnostics, 0.526 ± 0.147 g/L) was

Discussion

There is poor agreement between plasma fetuin-A measurements made using different ELISA methods. Overall, the Biovendor assay demonstrated better analytical performance; lower limit of detection, lower imprecision and better recovery of native plasma fetuin-A at the levels tested. The apparent over-recovery of both native and recombinant plasma fetuin-A using the Epitope Diagnostics assays suggests a degree of inaccuracy.

Measurements of plasma fetuin-A concentration in a population of patients

Acknowledgements

We thank Mrs Elizabeth Stainer and Mrs Joanna Lavender for technical assistance with the Grifols Triturus ELISA analyser.

References (13)

There are more references available in the full text version of this article.

Cited by (25)

  • From infancy to aging: Biological and behavioral modifiers of Fetuin-A

    2016, Biochimie
    Citation Excerpt :

    The reported Fetuin-A values were 2.2 higher when using the Epitope™ ELISA kit than when using the Biovendor™ ELISA kit. Smith and collaborators demonstrated that glycosylation status may explain, in part, the difference between the two commercial kits [53]. Therefore, it is possible that the degree of post-translational modification of Fetuin-A may influence measurement accuracy.

  • Serum fetuin-A levels and abdominal aortic calcification in healthy men - The STRAMBO study

    2015, Bone
    Citation Excerpt :

    Blood was collected in the standardized manner, but in the early afternoon and after a non-standardized lunch. There are substantial differences between kits used for the fetuin-A measurements which may be related to differences in specificity of antibodies for glycosylated and phosphorylated forms of fetuin-A [47]. Therefore, our results need to be checked in other cohorts and using other kits.

  • HSP70, Peroxiredoxin-3 and -6 are upregulated during renal warm ischaemia in a donation after circulatory death model

    2014, Journal of Proteomics
    Citation Excerpt :

    It can only be speculated that these multiple bands of A2HSG represent post-translational modified species. Variable expression of A2HSG during different antibody studies has previously been reported [56]. Smith et al. concluded that A2HSG exhibits variable biological detection by targeted antibody probes depending on the methodology [56].

  • Fetuin-A and the cardiovascular system

    2012, Advances in Clinical Chemistry
    Citation Excerpt :

    Nephelometry has also been used to measure fetuin-A. Differences among the various assays have not been clarified completely. Thus, standardization of fetuin-A assays is urgently needed [89,90]. In this regard, the concept of measuring FMC, but not serum fetuin-A, may substantially help our understanding of seemingly contradictory findings, although research in this area is still quite primitive.

View all citing articles on Scopus
View full text