Research paper
Improved multiplex immunoassay performance in human plasma and synovial fluid following removal of interfering heterophilic antibodies

https://doi.org/10.1016/j.jim.2005.03.009Get rights and content

Abstract

Cytokines, chemokines and soluble adhesion molecules interact in a complex network within the immune system. Fingerprinting of these proteins may allow the use of these proteins as biomarkers for identification of disease, disease subtyping and monitoring therapeutic interventions.

We developed a multiplex immunoassay (MIA) for the detection of 30 proteins in a variety of human body fluids such as plasma and synovial fuid (SF). The measurement of these proteins is hampered by the presence of human (auto-) antibodies, which can cause non-specific binding. We have validated a novel approach for the removal of interfering immunoglobulins using pre-absorption with protein-L.

Interfering (auto-) antibodies, such as rheumatoid factor (RF), were removed using three methods; polyethylene glycol (PEG) precipitation, pre-absorption with human γ-globulin or pre-absorption with protein-L. A significant decrease of RF was observed after a 2 h incubation with protein-L. RF IgM levels were reduced by 89% whereas total IgM, IgG and IgA levels were reduced by 60%. Residual immunoglobulins were blocked with rodent serum and did not interfere with the multiplex immunoassay.

Comparing the MIA with a conventional enzyme-linked immunosorbent assay (ELISA) using a panel of spiked plasma samples resulted in correlation coefficients for all mediators between R2 = 0.88 and R2 = 0.99. Intra-assay variance was less than 10% whereas inter-assay variance ranged between 6% and 16%.

Pathological samples with heterophilic antibodies hamper immunoassays such as ELISA and MIA. We show that pre-absorption with protein-L is a powerful tool for removal of interfering immunoglobulins from human bodily fluids to be used in immunoassays for studying changes in protein patterns.

Introduction

Cytokines and chemotactic cytokines (chemokines) are soluble proteins that have structural similarities and overlapping functions. When secreted during the course of an inflammatory response they interact in a complex network and function as effector molecules for cells of the immune system. During acute and chronic inflammation the dynamic equilibrium of cytokines and chemokines changes. The resulting inflammatory processes can either be systemic or restricted to a local site (Luster, 1998, Davidson and Diamond, 2001, Godessart and Kunkel, 2001, O'Shea et al., 2002). Within the inflamed microenvironment many mediators are secreted that interact directly with the surrounding tissue. The produced cytokines regulate the production of other inflammatory mediators whereas secreted chemokines function as regulatory molecules that attract and direct differentiation of new, potentially inflammatory, cells to the site of inflammation (O'Garra and Murphy, 1994, Luster, 1998, Luther and Cyster, 2001).

During the last years many different therapies have been developed that interfere in the cytokine and chemokine pathways and consequently modulate the inflammatory process. Patient categories that qualify for these forms of therapy are autoimmune diseases such as rheumatoid arthritis (RA). Immune intervention based on blocking the biological effects of tumor necrosis factor alpha (TNFα) and Interleukin 1 alpha (IL1α) is very effective for the treatment of a variety of human diseases, most notable RA and Inflammatory Bowel Disease (Elliott et al., 1994, Fleishmann, 2002).

A variety of techniques are available for detecting cytokines and chemokines, most notably the enzyme-linked immunosorbent assays (ELISA). However this technique has limitations, such as usage of large sample volumes and it is time consuming to obtain a complete protein profile. Recent applications for simultaneous detection of proteins in body fluids has resulted in particle based multiplex immunoassays (MIA). The Bio-Plex system employing the Luminex multi-analyte profiling technology (x-MAP™), allows individual and multiplex analysis of up to a hundred different mediators in a single well containing a sample volume of 50 μl (Vignali, 2000, Kellar et al., 2001, de Jager et al., 2003). However in chronic inflammatory conditions, heterophilic antibodies such as rheumatoid factor (RF) or other auto-antibodies are present (De Rycke et al., 2003, Tiittanen et al., 2004). In addition naturally occurring heterophilic antibodies can be found in up to 40% of the normal population and can either be IgG, IgM, IgA or IgE isotype (Kricka, 1999).

Detection of molecules in plasma or other complex biological fluids containing human antibodies, using a sandwich antibody system, poses problems especially when heterophilic antibodies are present. The antibody pairs used in a MIA are either monoclonal, polyclonal or a combination of both. Mouse and rat are common sources of monoclonal antibodies, whereas polyclonal antibodies are usually produced in rabbit, sheep and goat. Heterophilic- and auto-antibodies react with immunoglobulins from different mammals and thus can crosslink capture and detection antibodies resulting in either false positives or a blockade of the signal and thus critically interfere with immunoassay procedures (Kricka, 1999, Hennig et al., 2000, Kellar et al., 2001, Martins et al., 2004).

In this report we describe the development of a novel method to remove heterophilic antibodies from biological samples using protein-L. In contrast to protein-A and protein-G, protein-L has a high affinity for IgM and to a lesser extent for IgA and IgG (Akerstrom and Bjorck, 1989). In addition we describe the validation of an MIA for the quantification of 30 soluble proteins in plasma and synovial fluid containing heterophilic antibodies. These samples were derived from patients with chronic inflammatory conditions and were used as a model to demonstrate the power of this novel technique when interfering immunoglobulins are removed with protein-L.

Section snippets

Patient material

Heparinized blood samples were collected from patients with rheumatoid arthritis (n = 9) and healthy controls (HC; n = 10). Furthermore, heparinized synovial fluid (SF) samples were obtained from patients with RA (n = 10) and osteoarthritis (OA; n = 5, Table 1). All samples were stored frozen at − 80 °C until analysis. Written informed consent was obtained from the patient during visits to the outpatient clinic.

Before use all samples were centrifuged through a polypropylene centrifuge tube containing a

Immunoglobulin depletion

A multiplex immunoassay for 30 proteins uses 60 different antibodies (coating and developing reagents from mice, rat, chicken, rabbit, sheep and goat). As a consequence the potential interference of heterophillic (auto-) antibodies has to be avoided since this can induce false positive results. This is of particular importance for samples in a complex matrix and those with a high total protein content such as plasma and SF. Three different methods for the removal of immunoglobulins were tested

Discussion

We have described a validation of an MIA for the detection of 30 proteins in complex human body fluids together with a novel method for removal of antibodies from matrices that can interfere with an immunoassay. Previously we and others have validated the x-MAP™ technology for the detection of proteins using either commercial kits or own in house developed assays (Kellar et al., 2001, Prabhakar et al., 2002, de Jager et al., 2003, Prabhakar et al., 2004, Khan et al., 2004).

Since a large number

Acknowledgments

The authors wish to thank Huib de Jong and Joel van Roon (Department of Rheumatology and Clinical Immunology, University Medical Center Utrecht) for providing the patient material. W. de Jager and B.J. Prakken are financially supported by the Dutch Rheumatoid Arthritis Foundation (Nationaal Reumafonds).

References (22)

  • L. De Rycke et al.

    Antinuclear antibodies following infliximab treatment in patients with rheumatoid arthritis or spondylarthropathy

    Arthritis Rheum.

    (2003)
  • Cited by (0)

    View full text