What should you consider when thinking about your source and type of antigen?

Native autoantigens – the benchmark
Native autoantigens have a proven track record in autoimmune disease diagnostics. They have been used extensively by benchmark manufacturers of autoimmune diagnostic kits for over 20 years and have proven their reliability many times over. At AROTEC Diagnostics, autoantigen production is a result of extensive process development and control combined with proven scale-up procedures. We are therefore able to produce large batches of quality native autoantigens that are most appropriate to the diagnostic industry, and remain excellent value for money. The use of recombinant versions of autoantigens where the authentic human sequence is required (Ro52 and tissue specific autoantigens) or where there is no viable native source available (CENP proteins) is however a valid strategy. Indeed both Ro52 and CENP B proteins are produced at AROTEC Diagnostics using recombinant technology. The continuing improvement in native autoantigen quality due to the application of advanced chromatographic process technology will ensure that these reagents retain their benchmark status for many years to come.

Recombinant autoantigens often do not have the correct conformation
It is now well documented that many autoantibodies react with their corresponding autoantigen in a conformationally-dependent manner. Although this is especially evident for tissue-specific autoantibodies (e.g. thyroid), it is also becoming apparent that tertiary and quaternary autoantigen structures contribute to the binding of many systemic (ENA or ANCA) autoantibodies. The scientific literature is littered with cases of recombinant autoantigens, even though they have the correct primary sequence and may have been expressed as “soluble” proteins, being unable to effectively bind their corresponding autoantibodies. Sm, Ro60 (SSA) antigen and Proteinase 3 (cANCA) antigen are important examples where native proteins are clearly superior to their recombinant versions because autoantibody binding is conformation-dependent.

Recombinant autoantigens have aggregation and solubility issues
Overexpression of a foreign protein is usually a destabilizing process for a host cell. The expressed protein will therefore often be ejected from the cell or be compartmentalized into inclusion bodies. Host cell efforts to neutralize the foreign protein will usually result in aggregation and reduced solubility. In our experience even recombinant proteins that are extractable using buffers appropriate for “soluble” proteins often display degrees of aggregation that can vary considerably between fermentation runs. Aggregation and solubility issues have marked effects on purified autoantigen application procedures (e.g. immunoassay plate coating) and can be a significant source of assay variability.

Association with non-protein entities may be different or lacking in recombinant autoantigens
Many native autoantigens associate with non-protein entities which in many cases are thought to be involved in autoantigen binding, e.g. the ENA autoantigens Sm, RNP, SSA and SSB all associate with specific RNA molecules. It is not known if recombinant autoantigens associate with such entities and if so, it could only be with those of the host cell. It is unlikely that an autoantigen associated with, say, bacterial RNA possesses the same autoantibody reactivity as the native mammalian complex

Post-translational processing may be deficient in recombinant autoantigens
Many expression systems, while being able to reproduce the primary sequence of the protein of interest, are unable to post-translationally process the expressed autoantigen as would be the case in the native human or mammalian cell (examples of post-translational processing include phosphorylation, glycosylation, acylation, dimethylation and citrullination). Sm antigen is a well known example where native autoantigen effectively binds patient autoantibodies even in a denatured stated (e.g. in Western blot) whereas the recombinant version is ineffective. In host cells which are capable of some post-translational modifications, the modification carried out may not be identical to what would be the case in human mammalian cells e.g. glycosylation of thyroid peroxidase expressed in insect Sf9 cells is different from that of the native human protein. In general it is believed that only mammalian cells can post-translationally modify an over-expressed mammalian protein in the appropriate manner.

Contamination by host cell proteins
Although many recombinant autoantigens exhibit a high degree of purity, the type (as opposed to the amount) of any impurity present is of great importance. Many healthy individuals possess antibodies to proteins of host cells (e.g. bacteria, yeast) in which recombinant proteins are often expressed. Testing the sera of such individuals for the presence of antibodies against a recombinant autoantigen will often give a false-positive result due to antibody reactivity with contaminating host cell proteins. As a general rule of thumb, it can be said that the further the host cell is phylogenically from the individual being tested (i.e. human), the greater the likelihood of a false positive result. Thus the incidence of false positive results will be as follows for autoantigens purified from the following sources:

Incidence of false positives: bacteria ≈ yeast >> insect cell > mammalian ≈ human

Fusion partners may cause interference
Most recombinant proteins are expressed as fusion partners with peptides or proteins which simplify the purification process. In many cases (e.g. ß-galactosidase, glutathione-S-transferase) healthy individuals have been found to have antibodies against the bacterial version of the fusion protein itself, thereby leading to false positive results. Other fusion partners (e.g. peptide epitopes, hexahistidine) are more innocuous, although there have been some reports in the literature stating that they interfere with protein structure.

Shouldn't recombinant autoantigens be more affordable?
Not necessarily, some reasons are:
Expression of autoantigens in economical expression systems (e.g. E. coli, yeast) has been found to yield antigens of deficient quality (false positives, incorrect conformation etc.) therefore more expensive eukaryotic expression systems (baculovirus, mammalian cells) need to be used.
The market for purified autoantigens is not so great to allow significant economies of scale. Therefore the high developmental costs for recombinant autoantigens need to be recuperated on a relatively small market volume, making the development premium charged per unit sold very high.