ASSAY DEVELOPMENT

Figure 1. (click to enlarge) SDS-polyacrylamide electrophoresis of purified native autoantigens to demonstrate purity (target specification: greater than 90% pure). Autoantigens are: Lane 1, immunoaffinity-purified Scl-70, 70 kDa; Lane 2, immunoaffinity-purified Ro60 (SSA), 60 kDa; Lane 3, neutrophil proteinase 3, 27 kDa with higher m.wt. glycosylated forms; Lane 4, neutrophil myeloperoxidase, a-subunit 55 kDa, ß-subunit 12 kDa; Lane 5, ß2-glycoprotein 1 from human plasma, 50 kDa.

Autoantigens purified from native source materials are key components for many commercially available autoimmune disease diagnostic tests. Their principal application is activating a solid phase for the binding and subsequent detection of specific autoantibodies in patient fluids. Many diagnostically significant autoantibodies require retention of the authentic autoantigen structure and appropriate posttranslational processing for effective binding. Such attributes are intrinsic to native autoantigens, which have demonstrated binding of autoanti­bodies that remain undetected using equivalent components purified from alternative source materials.

For more than 20 years, purified autoantigens have been used in autoimmune disease diagnostics. During that time, technologies have evolved from radioimmunoassay and hemagglutination platforms to enzyme-linked immunosorbent assay (ELISA), line immunoassay (in which autoantigens are immobilized onto membrane strips), and, more recently, multiplexed microparticle and microarray formats. All of the diagnostic platforms that use purified autoantigens immobilize them onto a solid phase that binds specific auto­antibodies in a patient sample. Auto­antibodies bound to the solid phase are detected using labeled antihuman immunoglobulin conjugates.

In contrast to alternative diagnostic methods, such as immunofluorescence that uses native autoantigens in the form of intact cells (e.g., Hep-2, neutrophils) or tissue sections, technologies using purified autoantigens give rapid, quantitative results that require minimal operator interpretation. Diagnostic kits that employ purified auto­antigens for quantitative or semi­quantitative detection of patient autoantibodies constitute an in­dustry sector that has expanded con­tinuously since its beginning.

Purification of Native Autoantigens

Table I. (click to enlarge) Native autoantigens currently being used for commercial diagnostic applications.

By definition, native autoantigens are purified from native source materials of human or animal origin. Such materials are selected because they are known to contain in sufficient abundance human autoantigens or mammalian homologues thereof that bind effectively to human autoantibodies. The purity and specific activity achieved for the final product depends on the degree of purification expertise applied during the development and subsequent execution of the production process. Until now, purification of native autoantigens has utilized two different methods.

One purification method is conventional protein chromatography from source materials in which the target protein is abundant (purification factors of 10–1000-fold are required). Examples include the following: antineutrophil cytoplasmic antibody antigens (proteinase 3, myeloperoxidase) (purified from human neutrophil granules), β2-glycoprotein 1, C1q, and prothrombin (human plasma), mitochondrial antigen (bovine heart muscle), parietal cell antigen (porcine gastric mucosa), and glomerular basement membrane antigen (bovine kidney basement membrane).

The other method is immuno­affinity purification using human autoantibodies purified and then im­mobilized onto a chromatographic solid phase. This procedure not only is effective at purifying low-abundance target proteins (purification factors in excess of 4000-fold are achievable in one step), but also has the explicit advantage of being highly selective for target proteins that are functional autoantigens. This method has been applied to purifying nuclear antigens from mammalian tissue extracts (e.g., thymus or spleen).

By combining either method with final polishing using high-resolution chromatographic media, a high degree of purity can be attained as assessed by SDS-polyacrylamide gel electrophoresis (see Figure 1). In all cases, the native autoantigens should be tested for not only the quantity but also the quality of contaminating substances. If potential contaminants that are reactive to alternative specific autoantibodies are present, they can be removed by selective immuno­affinity absorption procedures (e.g., removing SmRNP antigen from Sm antigen using anti-RNP immunoaffinity chromatography).

Advantages of Native Autoantigens

Many of the critical properties of purified autoantigens (e.g., purity that can vary among manufacturers using the same source materials) will depend more on the purification technology than on the nature of the source material (i.e., native, recombinant, or synthetic). In many cases, different approaches can produce particular autoantigens that will perform well in an IVD kit. Nevertheless, native autoantigens offer the following advantages that are due to the fact that they are derived from natural source materials.

Table II. (click to enlarge) Authentic structural properties required for effective autoantibody binding to autoantigens commonly used in commercial diagnostic kits. Examples where synthetic or recombinant versions have not been able to achieve the same sensitivity as native autoantigens are listed.

A case in point is SSA (Ro60) autoantigen in which autoantibody binding is conformation-dependent, and recombinant Ro60 yields more than 25% false-negative results.^2,3 IVD kit manufacturers that have incorporated recombinant Ro60 have blended it with Ro52, a different autoantigen that the vast majority of SSA sera also react with. Although this strategy may mask false-negative results arising from recombinant Ro60, it has the undesirable effect of reducing the SSA assay’s specificity. Ro52 has a much broader specificity than SSA, and, for example, anti-Ro52 autoantibodies can be detected in many polymyositis and dermatomyositis sera that are Jo-1 positive/SSA negative.^4,5 Although reconstituting recombinant Ro60 with hY-RNA through coexpression to form a ribonucleoprotein complex increases sensitivity, the signal/cutoff ratio was still significantly lower than native SSA.⁶

In addition, some autoantibodies to the multisubunit U1 snRNP complex depend on the authentic ribonucleoprotein quaternary structure. A combination of the recombinant versions of the U1 68K, A, and C proteins failed to detect almost 8% of anti-U1 RNP autoantibodies with ELISA. Only when all subunits were present in a complete U1 RNA–containing ribonucleoprotein complex could such autoantibodies be detected.⁷ Due to its authentic quaternary structure, the native version of U1 snRNP, or SmRNP, remains the most effective component for detecting U1 RNP autoantibodies.

Figure 2. (click to enlarge) Primary sequence homology between human and bovine ENA autoantigens. Amino acid sequences were extracted from the NCBI Entrez Protein database (www.ncbi.nlm.nih.gov) or the ExPASy Proteomics database (www.expasy.org), and aligned using the T-Coffee multiple-sequence alignment program. The total number of positions for which the amino acid identity is conserved between the two species is shown as a percentage of the total number of amino acids for a particular autoantigen.

Authentic Posttranslational Modification. The posttranslational modification of certain amino acids is critical for the recognition of some autoantigens by specific autoanti­bodies. Most notably, citrullination, or deimidation, of arginine residues by peptidylarginine deiminase is required for binding specific rheumatoid arthritis (RA) autoantibodies; symmetrical dimethylation of Sm D arginine residues is required for binding some subpopulations of anti-Sm autoantibodies.^8,9 Although synthetic citrullinated peptide derivatives can detect RA autoantibodies, synthetic or recombinant versions of Sm D antigen have been less effective.⁸8 Expression of Sm D1 in baculovirus-infected insect cells resulted in asymmetrical dimethylarginines that are ineffective in binding autoantibodies.⁹ But short synthetic Sm D1 and D3 peptides containing symmetrical dimethylarginine can detect only some autoantibody subpopulations and do not detect conformation-dependent anti-Sm D autoantibodies.^9,10

By contrast, native Sm D antigens derived from bovine tissue can detect both Sm autoantibodies that require symmetrical dimethylarginine and conformation-dependent autoantibodies. Amino acid sequence alignment reveals that the Sm subunits D1 and D2 are identical for human and bovine, while only two amino acids vary for Sm D3. With the combination of authentic sequence, authentic structure, and correct posttransla­tional processing, native Sm antigen is likely to remain the benchmark component for detecting anti-Sm antibodies.

Homogeneity. After being biosynthesized in situ, native autoantigens are not susceptible to inappropriate and unpredictable aggregation states that occur in recombinant versions. Autoantigen homogeneity is important for most IVD platforms in which effective and reproducibly uniform attachment to a solid phase is required.

Stability. Proteins existing in their correct tertiary and quaternary structures are more stable and less likely to denature via precipitation or aggregation. This factor is significant both during and after application of the component (e.g., coating of immuno­assay plates), and during long-term storage while awaiting application.

Reproducibility of Quality. Although the consistency of purified autoantigen quality can vary among producers using the same source materials, qualified native sources provide a reliable and homogeneous starting material due to the in situ nature of the autoantigen expression. Proteins are expressed at consistent levels with consistently correct structure, regardless of the volume of starting materials. Consequently, en­­suing chromatographic separation procedures are applied with a high degree of predictability, ensuring reproducible end-product quality.

Large Lot Sizes. With homogeneous native starting materials, scale-up of the autoantigen purification process is a straightforward procedure. In this manner, a final product lot size of several hundred milligrams, which is sufficient to coat several tens of thousands of immunoassay plates, can be achieved. IVD manufacturers commonly perform extensive evaluation studies when introducing a new component lot into an established production process. But with the availability of lot sizes for such critical comp­onents that are equivalent to several years of diagnostic kit production, IVD manufacturers can maintain consistent production with minimal component evaluation efforts.

Absence of Host Cell Protein Contamination. Contamination by host cell proteins is a particular concern for autoantigens purified from source materials generated by heterologous expression. Many healthy individuals possess antibodies against proteins from the most commonly used host cells. Such antibodies will lead to false-positive results when host cell protein contamination is present even at trace amounts. Although this is particularly evident for prokaryotic expression systems (e.g., bacteria, yeast) in general, the closer the source material is phylogenetically to human, the less likely that host cell protein contamination will react with antibodies in healthy individuals. Since native autoantigens are purified from human or mammalian source materials (i.e., phylogenetically close to human), the risk of any contaminants causing false-positive results is minimal.

Intellectual Property Issues. Investigators using autoantigens derived from nonnative sources should realize that many stages of synthetic or recombinant autoantigen protein production and application are subject to intellectual property or patent protection. Many of those patents are recent, such as the following: citrullinated synthetic peptides for the detection of RA autoantibodies (2005; U.S. Patent No. 6,858,438), recombinant thyroid peroxidase (2007; U.S. Patent No. 7,196,181), and reconstitution of SSA antigen with hY-RNA through coexpression (2006; U.S. Patent No. 7,122,346). Native autoantigens are prepared using proprietary technology that does not expose the end-user to patent issues.

Limitations of Native Autoantigens

In autoimmune diseases in which the primary autoantigen amino acid sequence varies significantly between species and no adequate human source material is available, native auto­antigens have limited use to IVD kit manufacturers. This is indeed the case for the autoantigen Ro52, in which the human amino acid sequence of a nuclear autoantigen varies considerably from other mammalian species, and some tissue-­specific autoantigens (e.g., thyroid peroxidase). In general, there appears to be less interspecies sequence homology for tissue-specific autoantigens than nuclear autoantigens.

Other autoantigens (e.g., centro­mere proteins) are prominent only during cell division. Accord­ingly, they are only present at very low abundance in native source materials, other than cultured mammalian cells, which is prohibitively expensive. In such cases, recombinant technology has been able to provide acceptable autoantigen versions.

For novel or esoteric autoantigens of known sequence, it is also easier to generate a recombinant product with a purification affinity tag rather than establish a purification protocol from scratch for a native source material. In any such cases, it remains to be seen whether the recombinant version fails to detect autoantibodies dependent on authentic structure. In other cases, synthetic peptides, or derivatives thereof, are the most effective versions. For example, citrullinated peptides are more effective at binding RA autoantibodies than the full-length protein equivalent, due to the greater density of citrullinated arginine residues that can be achieved.⁸

Conclusion

During the more than 20 years that autoimmune diagnostic kits have been available commercially, native autoantigens that retain authentic structure have become benchmark components in their field. Reservations that autoantigens purified from native source materials may yield false-positive results due to non­specific protein content have been allayed by the availability of highly pure versions resulting from applying high-resolution final polishing chromatographic procedures. False­negative results for animal-derived autoantigens from differing primary sequences have also not been apparent due to the high degree of homology with human autoantigens. Furthermore, in many cases, authentic tertiary or quaternary structure, or correct posttranslational modification of specific amino acids, appear to be more critical for autoantibody recognition than any effects that minor variations in primary amino acid sequence would have.

However, the diversity of auto­immune diseases has meant that native autoantigen versions for certain disease specificities are not always available. In such cases, synthetic or recombinant autoantigens have been able to provide effective alternatives. Rather than a dogmatic approach (favoring only native or only recombinant/synthetic autoantigen versions), most manufacturers of autoimmune diagnostics for any particular disease specificity have therefore taken a pragmatic “what works best” approach to autoantigen component selection for their particular application. Consequently, manufacturers of diagnostic kits that cover a reasonable range of autoimmune disease specificities use at least some native autoantigen components in their product range. The irreplace­ability of native autoantigens for many autoimmune disease specificities, along with their improving quality and availability due to the application of continual improvement and scale-up procedures to their production processes, ensure that these components will continue to play a vital role in autoimmune diagnostics.

Neil J. Cook, PhD, is managing director at Arotec Diagnostics Ltd. (Wellington, New Zealand). He can be reached at neil.cook@arodia.com.	Richard H. Steele, MD, is a clinical immunologist and immunopathologist at Wellington Hospital (Wellington, New Zealand). He can be reached at richard.steele@ccdhb.org.nz.