1 Introduction

The measurement of autoantibodies against one or more common host antigens is often used to clinically diagnose autoimmune diseases [1–3]. Normally, autoantibodies to host proteins are IgM isotype and in low abundance in healthy individuals. The low levels of autoantibodies produced by B cells in healthy individuals is due to infrequent interaction with cells that aid autoantibody production; such as antigen-presenting cells (APCs) and T cells, that are eradicated by clonal deletion early in development or silenced by clonal anergy. These processes mean that as T cells survey host peptide fragments present on MHC-II molecules, they remain unresponsive to them, unless they appear “foreign” or altered. SLE patients are known to generate a greater amount of oxidative stress in their tissues, in which excess reactive species, including reactive oxygen species (ROS), reactive nitrogen species (NOS), and reactive chlorine species (RCS), are able to interact with amino acids, lipids, and nucleic acids [4, 5]. A consequence of this posttranslational modification (PTM) of individual amino acids in proteins leads to self-peptide sequences appearing as “nonself” to the immune system and initiation of an adaptive immune response [6, 7]. Indeed, several studies have demonstrated posttranslationally modified antigens to be superior to their native counterparts in autoimmune disease [8–10].

The monitoring of a number of autoantibodies has proved useful in the diagnosis (anti-double-stranded DNA; ds-DNA) of SLE and onset of nephritis (anti-C1q) in SLE patients [11–14]. However, the measurement of autoantibody titers is not relied on for routine monitoring of the clinical course of SLE. One of the reasons for this may be the lack of clinical sensitivity and specificity in test results correlating to disease activity. A number of factors may contribute to the lack of prognostic value of these assays in clinical practice. For example, when one examines the antigens used in enzyme-linked immunosorbent assays (ELISAs) to quantify specific autoantibody levels, they comprise of unmodified antigen from various animal sources or human recombinant protein made in bacteria. The use of PTM antigens may be a more representative antigen source with which to probe for autoantibodies against a specific antigen. Indeed, in some cases, PTM antigens may be the initiating antigen in the immune response, and therefore, autoantibodies to these PTM forms of an antigen may represent a more clinically relevant biomarker [15].

In the inflammatory environment generated during SLE pathology, a greater amount of ROS, RNS, and RCS may be generated, resulting in the formation of a number of PTM to amino acids [16]. For example, exposed cysteine present in SH groups upon exposure to ROS can be oxidized to form disulfides with other SH groups or be oxidized cysteine sulfenic, sulfinic, and sulfonic acid derivatives. The amino acid methionine is also highly susceptible to oxidation resulting in the generation of methionine sulfoxide or methionine sulfone. Chlorination of the side chains of lysine and histidine can occur resulting in the formation of chloramines. Nitrating species (such as peroxynitrite) may mediate the formation of the stable end products of nitration such as 3-nitrotyrosine (3-NT). Peptides bearing 3-NT have been shown to be immunogenic, evoke anti-3NT antibody responses, and induce a prolonged immune response [17, 18]. The result of protein exposure to a variety of reactive species may lead to significant changes in protein structure, the consequence of which may be the generation or exposure of peptides that are more antigenic than unmodified proteins, resulting in an autoimmune response. This may lead to unwanted production and accumulation of autoantibodies to “self-proteins.” However, we can exploit this aberration in immune tolerance to detect specific autoantibodies to PTM host proteins, with the view to correlate their production with disease activity.

2 Materials

All reagents used were analytical grade and solutions should be prepared using ultrapure water (prepared by purifying deionized water to attain a sensitivity of 18 MΩ cm at 25 °C). Treatment of solid-phase proteins requires the use of reactive species prepared in Chelex-200 treated water (to prevent spontaneous Fenton reaction generation of ˙OH from H2O2, ONOO− or HOCl).

2.1 Reactive Species Reagents

Posttranslation modification of proteins may be achieved by exposure to a number of reactive species. We routinely expose proteins to peroxynitrite (ONOO−), hydrogen peroxide (H2O2), hypochlorus acid (HOCl), and hydroxyl radical (˙OH).

  1. 1.

    Peroxynitrite ONOO− synthesis.

    Peroxynitrite can be purchased commercially in small quantities, but can be made in the laboratory (see Note 1).

    1. (a)

      In a fume hood, 50 mL of 600 mM HCl, and 700 mM H2O2 is added to 50 mL 600 mM NaNO2, on a magnetic stirring plate and then immediately added to 50 mL 1.2 M NaOH. A yellow ONOO− solution will form. To neutralize excess H2O2, add MnO2 powder (~10 g) and mix for 30 min. Next, remove the MnO2 by filtering the peroxynitrite solution through general use filter paper. The resulting supernatant is concentrated by overnight freezing at −20 °C, then the yellow-colored ONOO− solution is decanted during thawing.

    2. (b)

      The concentration of the ONOO− solution is determined spectrophotometrically at an absorbance setting of 302 nm, using the Beer-Lambert equation (ε = 1,670 M−1 cm−1) to estimate the concentration. A cuvette containing 0.4 M NaOH should act as a blank. The ONOO− aliquots can be stably stored at −80 °C for up to 6 months.

  2. 2.

    Both H2O2 and HOCl can be purchased relatively cheaply from general lab suppliers.

  3. 3.

    The use of 1 M H2O2 with and without 0.1 mM CuCl2 enables the generation of ˙OH to be generated through the Fenton reaction with the Cu2+ acting catalytically.

2.2 Modification of Solid-Phase Proteins by Reactive Species

2.2.1 Binding of Proteins of Interest to 96-Well Plates

  1. 1.

    Prepare sufficient sodium carbonate buffer to dilute a protein to bind to a 96-well MaxiSorp plate (Nunc). For 100 mL of buffer, add 0.1575 g Na2CO3 and 0.294 g NaHCO3 to 100 mL of double-distilled water. The pH of the sodium carbonate buffer should be pH 9.6, adjust with 1 M HCl, if necessary. Prepare 2 μg of protein of interest in sodium carbonate buffer, pH 9.6 to a final volume of 100 μL (see Note 2).

  2. 2.

    Allow the protein to bind to the plates overnight at 4 °C, then wash each well twice with 150 μL phosphate-buffered saline pH 7.5 containing 0.075 % v/v Tween-20 (PBST).

2.2.2 Reactive Species Treatment of Bound Proteins of Interest

  1. 3.

    The solid-phase proteins can now be modified by addition of various reactive species (Fig. 1). To generate a large proportion of PTM protein, a nonphysiological concentration of reactive species will be more effective than a physiological concentration (see Note 3). Initially, try exposing bound protein to 50 mM H2O2 (± 0.1 mM CuCl2), 0.1 mM HOCl, or 0.5 mM ONOO− prepared in 200 mM phosphate buffer (see Note 4), for 1 h at room temperature.

    Fig. 1
    figure 1

    Schematic outline of suggested modifications of solid-phase bound host protein by reactive species and testing for autoantibodies against PTM forms of antigen. (a) Proteins of interest are first bound to a 96-well plate. (b) The target protein is then modified. (c) Patient or control sera are exposed to the unmodified and modified forms of antigen and autoantibodies that recognize various form of the protein bind. (d) The amount of autoantibody recognizing various PTM forms of the protein is quantified spectrophotometrically

    Full size image
  2. 4.

    Wash the plates containing unmodified and modified proteins three times with PBST, and then block each well with 5 % (w/v) dried milk powder in PBS for 2 h at 37 °C. Next, wash the plates twice more with PBST. The plates are now ready to test with autoimmune and control serum samples.

3 Methods

3.1 Enzyme-Linked Immunosorbent Assay (ELISA) for Antibodies Against PTM Proteins in SLE Serum

  1. 1.

    Following binding and PTM of proteins of interest to a 96-well plate, add control and patient serum (Fig. 2). Initially, make a 1:50 dilution of patient serum in PBST containing 5 % (w/v) dried milk powder. For example, dilute 20 μL of serum in 980 μL of PBST containing 3 % (w/v) dried milk powder, and add 100 μL of the diluted serum to each well for 1 h at 37 °C. As a negative control add 100 μL of 40 μg/mL human IgG prepared in PBST containing 5 % (w/v) milk powder to separate wells containing test proteins.

    Fig. 2
    figure 2

    Suggested layout of samples and controls for screening patient sera for autoantibodies against PTM self-antigens. (a) Unmodified protein and two PTM-proteins can be prepared on each plate. (b) Additional PTM forms of a protein can be placed on a second 96-well plate along with controls and reference samples for comparison with the first plate. Shaded wells represent a known positive reference sample run on each plate in sextruplicate used to standardize the ELISA

    Full size image
  2. 2.

    Wash the wells twice with PBST and add 100 μL alkaline phosphatase-conjugated goat antihuman IgG antibody (see Note 5), diluted 1:2,000 in PBST containing 5 % (w/v) milk powder for 1 h at 37 °C.

  3. 3.

    Wash each well twice with PBST, then add 100 μL p-nitrophenyl phosphate (PNPP) substrate (dilute in accordance with manufacturer’s instructions) and incubate in the dark for 30 min. The reaction is terminated by addition of 11 μL 500 mM NaOH (50 mM NaOH final concentration).

  4. 4.

    Read the yellow-colored product at an absorbance maximum of 405 nm in a 96-well plate reader (see Note 6).

3.2 Standardization of the ELISA

  1. 1.

    Unless all the test and control samples are run on a single plate in replicate on a single day, it is important to standardize results between plates and performed on different days. A reference serum sample known to contain a relatively high antibody titer to an unmodified antigen can act as a standardization control, with which to compare the data with PTM of the antigen. Negative controls do not receive serum samples, but do receive wash and blocking buffers and secondary antibody and substrate and demonstrate the level of background, due to nonspecific antibody interactions.

  2. 2.

    Take the reference sample (see Note 7) and prepare twofold serial dilutions from 1:25 to 6,400 (eight different dilutions). Perform ELISA above in triplicate and construct a standard curve of the OD405 (Y-axis) vs. the dilution (X-axis). The correlation coefficient of the curve should be greater than 0.95 and can be calculated plotting the data using Prism 5 software or other statistical package. Chose a serum dilution that gives an approximate OD405 = 1 AU. This reference sample at the chosen dilution (in sextruplicate) should thereafter be added to each plate being used for all sample testing by ELISA (see Note 8).

  3. 3.

    A simple ELISA unit (EU) can now be defined as the percentage of the OD405nm of the reference sample using the equation:

    EU = mean OD405nm (sample)/mean OD405nm (reference) × 100

  4. 4.

    The inter- (between plates) and intra- (within plate) assay variation of the ELISA should also be validated. The inter-assay variation provides information of the reproducibility of the assay when performed on different plates, at different times, using different batches of buffers and the influence of different operators. The intra-assay variation provides information on pipetting error, mixing, timing, and influence of well positions. To assess the inter-assay variation, run a positive and negative control standard and zero analyte (buffers only) at least 20 times on 5 separate plates in five consecutive days. For intra-assay variation, run the above in a single plate. Read the plates and calculate the mean OD ± standard deviation (SD) of the positive, negative, and blank control samples. The SD is an estimate of reproducibility of the replicate data points and is a useful tool for estimating the variability of replicate results normally calculated as the coefficient of variation (CV). The CV is calculated as %CV = SD/mean × 100 (see Note 9).

3.3 Statistics

To test the comparison of autoantibody titers to PTM antigen compared to untreated antigen, use an ANOVA for nonparametric data (Friedman test) with Dunn posttest (if multiple groups are being compared).

3.4 Interpretation

The objective of this method is to make comparative evaluations of antibody levels against various forms of host antigens that may or may not have undergone PTM during the course of the disease. By measuring a PTM form of antigen, it may increase the specificity or sensitivity of the ELISA with respect to (a) predicting onset of disease activity, e.g., nephritis, and (b) monitoring efficacy of treatment, e.g., reduction in acute-phase response or efficiency of B cell depletion.

  1. 1.

    Sensitivity and specificity. Determination of a positive or negative antibody titer by ELISA requires the establishment of an arbitrary “cutoff” titer. Often it is only possible retrospectively to calculate the clinical sensitivity and specificity of an assay. The actual proportions of positives that are correctly identified as positive define the clinical sensitivity of the assay, while the actual number of negatives that are identified as truly negative provide the basis of the clinical specificity.

  2. 2.

    Application of receiver operating characteristics (ROC) curves. ROC curves are a way of plotting specificity against sensitivity to determine optimal cutoff values for an assay [19]. Various graphical software programs (e.g., GraphPad Prism V5) can be used to construct such curves. In addition, the area under ROC curves may be used to assess if PTM of an antigen increases its diagnostic utility through increased clinical sensitivity or specificity, when compared to the unmodified antigen. Therefore, ROC curves provide a useful statistical tool to retain or discard assays that improve the accuracy of the method (in terms of increased specificity or sensitivity) to monitor disease activity.

4 Notes

  1. 1.

    A number of manufacturers sell peroxynitrite and these may be easily found on the Internet.

  2. 2.

    A 96-well can only absorb a limited amount of protein. Proteins prepared in sodium carbonate buffer at 2 μg/100 μL will likely be in excess of the maximum amount of protein that can bind to a well, allowing saturated binding of protein to each well.

  3. 3.

    The concentrations of reactive species cited in our hands are sufficient to cause PTM of proteins that are detectable by mass spectrometry, without removing the bound protein from the wells of the plate, although these concentrations need to be validated for each protein of interest.

  4. 4.

    It is important to prepare the reactive species in 200 mM phosphate buffer (pH 7.4) in Chelex-200-treated water. Pass Milli-Q double-distilled water through a prepacked column containing Chelex-200 resin into a container that has been washed with 5 mM EDTA (pH 7.0). To prepare the sodium phosphate buffer, prepare two solutions: A and B. Solution A is comprised of 5.52 NaH2PO4·H2O/100 mL of Chelex-200-treated H2O (400 mM). Solution B is prepared by adding 10.73 g Na2HPO4·7H2O/100 mL Chelex-200-treated H2O (400 mM). To obtain the desired pH of 7.4, mix 3.8 mL of solution A to 10.1 mL of solution B, then dilute to 200 mL with Chelex-200-treated water. Check the pH of the buffer, and adjust volumes of solutions A and B.

  5. 5.

    Do not use Chelex-treated water-based buffers to dilute secondary antibodies conjugated to alkaline phosphatase (AP). AP is a zinc metaloenzyme that requires the presence of zinc to activate the active site of the enzyme when it interacts with substrate.

  6. 6.

    Ensure the contents of each well are completely mixed, but ensure there are no air bubbles in the wells, as this will affect the spectrophotometric readout. Do not vary the time of the substrate incubation, as this will increase variation in the data.

  7. 7.

    The reference sample could be a large batch of a single serum sample or a pooled sample of several patients who have relatively high antibody titers against the antigen of interest.

  8. 8.

    If the OD results of the reference sample are not consistent between plates (within the inter-assay variation), repeat the assay until consistent results are obtained.

  9. 9.

    As a rough estimate if the intra- or inter-assay individual control results lie outside a mean value ± 2 SD, then you can be confident 95 % of the time the result is unreliable. If an individual result lies outside a mean value ± 3 SD, then you can be confident 99 % of the time the result is unreliable.