The Interaction of Anti-DNA Antibodies with DNA Antigen: Evidence for Hysteresis for High Avidity Binding

Abstract

Antibodies to DNA (anti-DNA) are the serological hallmark of systemic lupus erythematosus. Previous studies have indicated that the phosphodiester backbone is the main antigenic target, with electrostatic interactions important for high avidity. To define further these interactions, the effects of ionic strength on anti-DNA binding of SLE plasmas were assessed in association and dissociation assays by ELISA. As these studies demonstrated, increasing ionic strength to a concentration of 1000 mM NaCl reduced antibody binding although the extent of the reduction varied among samples. In dissociation assays, differences among plasmas were also observed. For one of the plasmas, binding to DNA displayed resistance to dissociation by increasing ionic strength even though these concentrations limited binding in association assays. Time course studies showed a gradual change in binding interactions. These studies indicate that anti-DNA binding can involve both electrostatic and non-electrostatic interactions, with binding in some plasmas showing evidence of hysteresis.

1. Introduction

Antibodies to DNA (anti-DNA) are the serological hallmark of systemic lupus erythematosus (SLE) and markers for classification, diagnosis and disease activity (1–3). These antibodies bind to single stranded (ss) as well as double stranded (ds) DNA and have features consistent with antigen selection. Thus, anti-DNA antibodies demonstrate somatic mutations producing positively charged amino acids such as arginine in the heavy chain complementarity determining regions (CDRs) (4–6). Since anti-DNA antibodies bind to naturally occurring DNA independent of base sequence, these findings have suggested that anti-DNA antibodies target sites along the phosphodiester backbone, with electrostatic interactions with negatively charged phosphate groups key for binding.

In addition to their utility as biomarkers, anti-DNA antibodies play an important role in disease pathogenesis via the formation of immune complexes (ICs) with DNA (7–9). These complexes can deposit in the tissue to fix complement and promote inflammation, with the kidney a major locale for this process. In addition, ICs can enter into cells of the innate immune system to trigger toll-like receptors (TLRs) as well as non-TLR sensors for DNA (10–12). Activation of these internal sensors can stimulate the production of type 1 interferon to drive inflammation. Thus, ICs containing anti-DNA can promote both local and systemic inflammation.

While anti-DNA antibodies can drive pathogenesis, studies of patients as well as animal models suggest that only some antibodies of this specificity can form ICs that either deposit in the tissue or induce cytokine production (13–15). These antibodies are termed pathogenic; the term nephritogenic denotes antibodies that induce renal disease. Among immunochemical properties that can underlie pathogenicity, avidity has been considered important because of its role in determining IC formation. Measurement of antibody avidity for DNA is problematic, however, since DNA is a large macromolecule with a repeating structure (16). As such, the concentration of epitopes that can be bound by antibodies is unknown and it is difficult to calculate precise binding constants.

In the current studies, we have investigated another immunochemical approach to assess the binding properties of anti-DNA antibodies related to IC formation. Thus, we have used an ELISA to assess both antibody association and dissociation. In general, the association of antibody with antigen in solution is considered rapid and diffusion-related while dissociation reflects antibody avidity or functional affinity (17, 18). This relationship also pertains to solid phase assays where antigen is immobilized on a surface although exposed to fluid. As a probe of these interactions, we have used the effects of ionic strength on binding in view of the putative role of electrostatic interactions as major determinants of anti-DNA interaction.

In these studies, we test the hypothesis that, with antibodies binding by charge-charge interactions, ionic strength can influence both association and dissociation via charge shielding. The approach in this study is similar to that of studies using the Crithidia luciliae immunofluorescence assay (CLIFT) (19–21); these studies showed significant effects of ionic strength as well as pH on binding interactions. Crithidia luciliae organisms are hemoflagellates with an organelle called the kinetoplast which contains a pure circle of double stranded DNA; histones are not present (22). While CLIFT is a useful assay for clinical assessments, the structure of DNA in the kinetoplast may differ from that of ordinary mammalian DNA, possibly limiting its utility as a probe for key binding interactions.

To gain further insights into the nature of DNA-anti-DNA interactions, we have, therefore, assessed the effects of ionic strength on both association and dissociation reactions with mammalian DNA in an ELISA using patient plasmas as well as a humanized monoclonal anti-DNA antibody. In addition, we performed kinetic studies to assess the time course of antibody binding. As results presented herein indicate, anti-DNA interactions, while primarily electrostatic, nevertheless, can evolve over time in a slow process taking minutes. As such, these studies suggest that structural changes in either DNA or anti-DNA antibody can determine avidity in a time-dependent process.

2. Materials and Methods

2.1. Study samples

For this study, plasma samples from eight patients with SLE were purchased from Plasma Services Group; Huntingdon Valley, PA, USA. EDTA was the anticoagulant used. Five additional plasma samples from the donor of plasma #9 collected over a 2.5 year period were also obtained from the same source. The samples were stored long term at −80°C, with an aliquot of the original kept at 4°C to avoid frequent freeze-thaws during the course of the experiments. Approval to use these samples was obtained in accordance with the Department of Veterans Affairs policy.

Before use, the thawed samples were centrifuged at 3000 × g for 2 min at room temperature (RT; 19 to 23°C) to remove debris and the supernatant was removed for storage at 4° C. VAL-1205, a monoclonal anti-DNA antibody, was produced by Catalent Pharma Solutions for Enable Therapeutics, LLC, Framingham, MA, USA. VAL-1205 is a humanized version of the anti-dsDNA antibody 3E10, originally discovered in a mouse model of SLE (23). This antibody has a point mutation (D31N) that improves DNA binding and cell penetration (24).

2.2. Effect of NaCl concentration on association

Anti-DNA activity was analyzed by ELISA. In brief, Immunlon 2HB plates (Thermo-Fisher-Invitrogen, Waltham, MA, USA) were coated overnight with 100 μl/well of 5 μg/ml of calf thymus (CT) dsDNA (Worthington Biochemical Corp., Lakewood, NJ, USA); the DNA was further purified by phenol-chloroform extraction with purity assessed by the OD 260/280 ratio. The DNA was prepared with 1X SSC (saline sodium citrate, pH 7.0; Thermo-Fisher-Invitrogen). Control wells had 1X SSC. The assay plates were incubated overnight at 4°C.

The plates were washed three times with 1X phosphate buffered saline, pH 7.4 (PBS-free; Ca²⁺ free, Mg²⁺ free). The assay wells were then incubated with 200 μl/well of block buffer II (2% bovine serum albumin [BSA], 0.05% Tween-20 [both from Sigma Aldrich, St. Louis, MO, USA] in 1X PBS-free) for 2 h at RT. Plasma/antibody samples were diluted with Tris based ELISA dilution buffer (Tris – EDB 150 mM; 150 mM sodium chloride, 0.1% BSA, 0.05% Tween-20 in 50mM Tris, pH 7.4). In these studies, two to three fold serial dilutions were prepared for each sample. Dilutions were selected to represent the linear portion of antibody binding.

Solutions at 2X the concentration of the desired final antibody dilution values were prepared in Tris – EDB 150 mM. Sodium chloride and stock solutions of urea were prepared at concentrations so that 1:1 dilution with 2X antibody solutions would give sodium chloride concentrations of 250 mM, 500 mM and 1000 mM or 1 M, 2 M or 4 M urea. These sodium chloride and urea solutions also contained 0.1% BSA, 0.05% Tween-20 in 50 mM Tris, pH 7.4. Tris – EDB 150 mM was used for wells where the final NaCl incubation concentration was 150 mM.

Using a multichannel pipette, 50 μl/well of sodium chloride and urea solutions of varying concentrations or Tris – EDB 150mM were distributed to wells of the microtiter plates. Then, 50 μl/well of the 2X antibody solutions were added. The wells were incubated for 1 h at RT. All reactions were performed in duplicate. The plates were then washed 3X with 1X phosphate buffered saline and then incubated for 1 h at RT with 100 μl/well of a 1:1000 dilution of anti-human IgG (γ chain specific)-HRP conjugated (Sigma-Aldrich) in phosphate based ELISA dilution buffer (EDB-P; 1X PBS, 0.1% BSA, 0.05% Tween-20).

After incubation of the conjugate, the plates were washed three times with 1X phosphate buffered saline and then incubated for 30 min at RT in the dark with 100 μl/well of TMB substrate solution (0.01% H₂O₂, 0.015% 3,3′,5,5′- tetramethylbenzidine dihydrochloride [both from Sigma-Aldrich] in 0.1 M citrate buffer, pH 4.0). After addition of 100 μl/well of 2 M H₂SO₄ to terminate the reaction, plates were then analyzed for absorbance at 450 nm using a spectrophotometer (Molecular Devices, San Jose, CA, USA).

In one experiment, longitudinal samples from patient #9 were analyzed.

2.3. Analysis of antibody dissociation with NaCl, urea or CT DNA

Diluted samples were distributed at 100 μl/well and allowed to incubate for 1 h at RT. After this incubation, the plasma/antibody solutions were removed and replaced with Tris-EDB 150 mM containing NaCl (150, 250, 500 or 1000 mM) or CT dsDNA (0, 5, 15, 50 μg/ml). This incubation proceeded for 1 h at RT. The protocol then continued with incubation of secondary antibody and addition of substrate as described above. Data on plasma #6 are not presented for NaCl and CT DNA due to inconsistency in these dissociation assays.

For dissociation with urea, initial antibody binding proceeded for 1 h at RT. Buffer was then replaced with 100 μl per well of EDB-P containing varying concentrations of urea (0, 1, 2 and 4 M). The ELISA protocol continued as described. To assess the combined effects of NaCl and urea on the binding of plasma #9, the original incubation mix was removed and replaced with 100 μl per well of either EDB-Tris 150 mM NaCl or EDB-Tris 1 M NaCl plus varying urea concentrations (0, 1 M, 2 M or 4 M). The reaction proceeded for 1 h at RT. After this step, the urea/NaCl solutions were removed and the protocol continued as described.

2.4. Time course of antibody binding before dissociation with NaCl

Diluted plasma or antibody solutions were prepared as outlined above and distributed to the assay plates (100 μl/well) for incubation for 5, 20 or 60 minutes. At the end of the initial incubation time, 100 μl/well of SCAS were added to produce a final concentration of 250, 500 or 1000 mM NaCl per well. Tris – EDB 150 mM was distributed to wells where the final NaCl concentration was 150 mM. After the addition of the SCAS solutions, the incubations continued for a total of 2 h. Anti-DNA activity was then measured.

2.5. Time course of antibody binding before dissociation with DNA

The experimental details are as outlined above except that antibody binding was challenged with ds soluble CT DNA. Plasma/antibody solutions (100 μl/well) were incubated with plated ds CT DNA for 5, 20 or 60 minutes. At the end of the initial incubation time, 100 μl/well of ds CT DNA were added to produce a final concentration of DNA at 0, 5, 15, or 50 μg/ml. Anti-DNA was then measured.

2.6. Effects of NaCl concentration on IgG binding

IgG was purified from 250 μl of undiluted plasma #9 using Nab™ Protein A/G Spin Columns (Thermo-Fisher-Scientific) following the manufacturer’s instructions. Briefly, spin columns were washed with Binding Buffer (0.1 M phosphate, 0.15 M sodium chloride; pH 7.2). Plasma #9 (250 μl) was applied to the column and then mixed by end-over-end tumbling for 10 min at RT. The spin column was then placed into a 2 ml microcentrifuge tube and centrifuged at 5000 × g for 1 min at RT to remove unbound plasma components. The bound IgG fraction was removed in three fractions with elution buffer of 0.1 M glycine, pH 2.5 and subsequently neutralized with 1/10 volume of 1 M Tris, pH 9.0.

The amount of antibody present was determined by measuring the absorbance of each fraction at 280 nm by Nanodrop 1000 (Thermo-Fisher-Scientific). The first elution fraction gave the highest yield (6.7 mg/ml) and was used in subsequent ELISA assays. An initial ELISA experiment was performed to determine concentrations of the purified IgG fraction required to produce levels of antigen binding comparable to those of plasma #9. Plasma #9 and purified IgG from plasma #9 were then analyzed for association and dissociation as described above.

3. Results

3.1. Binding of a monoclonal anti-DNA antibody

In a preliminary experiment to explore association and dissociation assays, we assessed the binding of a humanized monoclonal anti-DNA antibody that is based on a murine product (23). This antibody has been modified by the introduction of a somatic mutation to substantially enhance binding to negatively charged DNA (24). As such, VAL-1205 provides an apt model for binding by electrostatic interactions. As shown in Figure 1A, the extent of binding of VAL-1205 dramatically decreases with increasing ionic strength during the association step; with 500 mM NaCl, binding becomes essentially undetectable.

Figure 1. The effects of ionic strength on the binding of a monoclonal anti-DNA antibody to DNA.

The effect of salt on the binding of VAL-1205, a monoclonal anti-DNA antibody, to double stranded calf thymus DNA (dsDNA; 5 μg/ml) was assessed by ELISA in association and dissociation assays. Panel A shows the binding with increasing concentration of NaCl. Panel B shows the effects of increasing concentrations of NaCl on antibody binding following association in 150 mM NaCl. Results are presented as OD 450 nm values of the ELISA. All assays were performed in duplicate; standard deviations are shown.

In the next experiment, following incubation for 1 h, the salt concentration of the assay buffer was increased following replacement of the initial incubation solution with buffer to produce final concentrations of NaCl of 250 to 1000 mM. As in the case of the association step, the extent of binding was dramatically reduced by high ionic strength although some binding was observed after dissociation with 250 mM NaCl (Figure 1B). These findings are consistent with a predominant role of electrostatic interactions in the binding of VAL-1205. These experiments support the idea that strong DNA binding results from charge-charge interactions that can be inhibited even after equilibrium is established.

3.2. Effects of ionic strength on SLE anti-DNA association

Using this approach, we next tested a series of SLE plasmas that had been selected on the basis of high anti-DNA binding; plasmas with high binding were used to allow better assessment of any changes in activity. Figure 2 presents the results of titrations of these plasmas in solutions of NaCl ranging from 150 mM to 1000 mM. As these data indicate, plasmas differed with respect to the effects of ionic strength on association, with some plasmas, nevertheless, able to bind DNA in NaCl concentrations as high as 500 mM. Binding in 1000 mM was at most very limited. These findings differ from those of the VAL-1205 monoclonal which showed essentially no binding at comparable salt concentrations.

Figure 2. The effects of ionic strength on the binding of SLE plasmas to DNA.

Plasmas from six different SLE donors were tested by ELISA to investigate the effects of increasing NaCl concentrations (150, 250, 500 and 1000 mM) on the association step of antibody binding to dsDNA (5 μg/ml). Plasma dilutions ranged from 1:200 to 1:5400. Panels A – F represent the results of the different plasmas. Results are presented as OD 450 nm values of the ELISA. All assays were performed in duplicate; standard deviations are shown.

3.3. Effects of ionic strength on SLE anti-DNA dissociation

In the next experiments, we assessed the effects of ionic strength on dissociation by changing the buffer after one hour and allowing dissociation to occur over an additional hour. One hour was used for the initial incubation on the basis of preliminary experiments to show that the extent of binding had plateaued (data not shown). The studies assessing dissociation demonstrated that some plasmas maintained DNA interaction even in the presence of 500 mM to 1000 mM NaCl. In the case of one of the plasmas tested (Figure 3B), almost complete antibody binding was observed despite the presence of a concentration of NaCl sufficient to promote dissociation of VAL-1205 as well as other plasmas in this panel. This finding suggests that the binding of anti-DNA antibodies in some plasmas is less dependent on electrostatic interactions than others and that association and dissociation assays can show differences in the effects of increasing ionic strength.

Figure 3. Effects of ionic strength on dissociation of anti-DNA antibodies from DNA in SLE plasmas.

The effect of NaCl concentrations on the dissociation from DNA of anti-DNA from five SLE plasmas was analyzed. With plates coated with DNA at a concentration of 5 μg/ml, antibody binding to DNA was allowed to proceed for 1 h in 150 mM NaCl; after this incubation, incubation buffer (150 mM NaCl) was changed to solutions of varying NaCl concentrations (150, 250, 500 and 1000 mM). The reaction continued for another hour. Panels A – E represent the results of the different plasmas presented as OD 450 nm values. Assays were performed in duplicate; standard deviations are shown.

The results presented thus far are consistent with the findings reported using the CLIFT assay, indicating that predominant anti-DNA binding mechanisms may vary (19–21). Since anti-DNA in plasmas are probably heterogeneous in composition, it is likely that, while most anti-DNA bind by electrostatic interactions, there can be antibody subpopulations which show a greater contribution of other binding mechanisms such as hydrogen bonds. Furthermore, it is possible that, during the incubation period in the ELISA, the predominant mode of binding of an individual antibody preparation may change so that the effects of ionic strength on association and dissociation differ. According to this idea, an antibody-antigen interaction can change such that other bond types (e.g., hydrogen bonds) develop over time.

3.4. Time course of SLE anti-DNA binding

One way to explore the changing profile of anti-DNA is a kinetic or time-course experiment. For this purpose, we tested the effects of ionic strength on dissociation after a variable time of incubation of antibodies with DNA at 150 mM. In these experiments, we tested plasmas that retained observable binding after incubation with 1000 mM NaCl. As these data indicate (Figure 4), for one of the plasmas (plasma #9), the amount of binding that could resist the effects of high ionic strength increased over time so that eventually almost all binding was salt resistant and thus non-ionic. For other antibodies, the extent of binding did not change, suggesting that non-ionic interactions did not develop.

Figure 4. The effects of incubation time of plasmas with DNA on subsequent dissociation of anti-DNA antibodies at high ionic strength.

Six SLE plasmas were incubated with plates coated with dsDNA for 5, 20 or 60 minutes before the incubation buffer (150 mM NaCl) was changed to a solution of 1000 mM NaCl. For each assay condition, the assay proceeded for a total of 120 min. The percent antibody binding was calculated as (OD 450 nm 1000 mM/OD 450 nm 150 mM)*100 for a given time point.

3.5. Binding properties of purified IgG

Since these titrations involved plasmas, it is possible that a component of plasma other than IgG somehow was affecting the interaction of anti-DNA with DNA, perhaps binding to DNA. A previous study on the Farr assay demonstrated that antibody binding may be increased by the presence of a plasma component that interacts with DNA (25). Therefore, we purified IgG from plasma #9 and assessed both association and dissociation. As shown in Figure 5, the purified IgG showed the same pattern of binding as did the plasma, indicating that DNA interaction results from the IgG itself. Thus, these data indicate that IgG is the only protein species involved with anti-DNA binding.

Figure 5. Comparison of effects of ionic strength on the anti-DNA activity of plasma and purified IgG.

The effects of ionic strength (150, 250, 500 and 1000 mM NaCl) on association (A and B) and dissociation (C and D) assays for SLE plasma #9 (panels A and C) and IgG fraction from SLE plasma #9 (B and D) were determined by ELISA. Dilutions of plasma and IgG fraction were selected based on linear portions of binding reactivity for the given antibody source. Results are reported as OD 450 nm values. All assays were performed in duplicate; standard deviations are shown.

3.6. Binding properties of longitudinal SLE samples

For plasma #9, we had a set of longitudinal plasmas whose anti-DNA levels varied. We, therefore, investigated whether the pattern of DNA binding was stable over time. As shown in Figure 6, for the two samples that had high levels of anti-DNA activity (Samples 2 and 5), antibody binding showed similar time-dependent increases in resistance to the effects of ionic strength. Sample 4 had very little anti-DNA activity and resistance to NaCl did not change over time. These findings suggest that, in samples with high anti-DNA activity, the profile with respect to the role of ionic and non-ionic interactions can be similar although this pattern may change when overall levels of anti-DNA activity are low.

Figure 6. The properties of anti-DNA antibodies in longitudinal plasma samples.

Panel A shows anti-DNA antibody binding to plates coated with dsDNA (5 μg/ml) for six SLE plasmas collected from the same donor. Samples are labeled as bleeds 1 – 6, with number 1 being the first sample collected; increasing bleed number represents sequentially collected samples. Samples 2, 4, and 5 were assessed for the effects of increasing ionic strength on anti-DNA binding after initial incubation times (150 mM NaCl; 5, 20 and 60 min) at 150 mM. Panel B shows results for dissociation at 1000 mM NaCl. Results are presented in terms of either OD 450 nm values or percent anti-DNA activity calculated as (OD 450 nm 1000 mM/OD 450 nm 150 mM)*100 for a given time point.

3.7. Effects of urea on SLE anti-DNA association and dissociation

As another approach to investigate the binding of antibodies to DNA, we determined the effects of urea on antibody association and dissociation. Urea is a chaotropic agent that can be used to determine the role of non-ionic interactions as well as assess antibody avidity by its effect of antigen-antibody interactions (26, 27). We, therefore, determined the effects of concentrations of urea as high as 4 M on antibody association (Figure 7). As these data indicate, urea causes a variable inhibition of antibody binding, with limited reduction in antibody binding at 1–2 M, with more significant inhibition at 4 M.

Figure 7. The effects of urea on anti-DNA association.

The panels show the results of performing initial binding of anti-DNA in six SLE plasmas (#5 [A], #6 [B], #9 [C], #10 [D], #11 [E] and #19 [F]) to calf thymus DNA (CT DNA) in the presence of various concentrations of urea (no urea [circle], 1 M [square], 2 M [diamond] and 4 M [triangle]). Control conditions included no DNA without urea (data not shown; binding was negligible) and no DNA with 4 M urea (open triangle). The results are presented as the average OD 450 nm values for duplicate wells with the standard deviation indicated by error bars.

As in the case of the effects of ionic strength, we also assessed the effects of urea on antibody dissociation. This approach has been frequently used to assess antibody avidity since it involves the activity of the chaotropic agent after the bonds between antibody and antigen have formed (28–31). Figure 8 presents these results. As these data indicate, the plasmas differed with respect to the effects of increasing concentrations of urea although, in general, they showed significant binding in urea at a concentration of 1–2 M although binding was observed in urea concentrations of 4 M.

Figure 8. The effects of urea on anti-DNA dissociation.

The experiments were performed as described in Materials and Methods. Data shown represent results of six SLE plasmas (#5 [A], #6 [B], #9 [C], #10 [D], #11 [E] and #19 [F]) tested for the effects of urea on the dissociation of anti-DNA binding (no urea [circle], 1 M [square], 2 M [diamond] and 4 M [triangle]. The results are presented as the average OD 450 nm values for duplicate wells with the standard deviation indicated by error bars.

Plasma #9, which showed resistance to the effects of high salt, also showed significant binding in urea at 4 M. For this plasma, we could investigate the combined effects of both a chaotropic agent and ionic strength; this experiment could not be performed with the other plasmas for which 1000 mM NaCl caused virtually complete dissociation of antibody binding. Figure 9 presents the results of the combined effects of urea and 1000 mM NaCl on antibody dissociation. As these data indicate, plasma #9, under conditions of dissociation with the combination of 1000 mM NaCl and increasing urea concentrations, nevertheless, had measureable binding even in the highest concentration of urea tested. Together, these studies indicate that the binding to DNA of most plasmas shows less effects on association and dissociation with a chaotropic agent than with increasing ionic strength, likely related to the greater role of ionic interactions. For plasma #9, which showed resistance to dissociation with 1 M NaCl, binding was present with both urea and high salt, indicative of the role of different non-covalent interactions.

Figure 9. The combined effects of NaCl and urea on anti-DNA dissociation.

The effects of NaCl (150 mM [panel A] and 1 M [panel B]) in combination with varying doses of urea (no urea, 1 M, 2 M, and 4 M) on anti-DNA dissociation were assessed for SLE plasma #9. The experiments were performed by ELISA as described in Materials and Methods. Results are presented as the average OD 450 nm values for duplicate wells with the standard deviation indicated by error bars.

3.8. Effects of DNA on SLE anti-DNA dissociation

Both high salt and urea have effects on protein and DNA structure as well as binding interactions. Especially at high concentration of a chaotropic agent like urea, protein denaturation can occur; urea can also interact with DNA, affecting interpretation of these experiments (32–34). As another approach to explore antibody interactions with DNA at ordinary salt conditions, we next performed dissociation assays in which we added high concentrations of DNA after an initial period of incubation with plasma at 150 mM. In this assay format, once an antibody dissociates from the solid phase antigen, it could bind to fluid phase DNA and, therefore, be undetectable in the ELISA. This approach differs from the use of high ionic strength and urea since these treatments could alter the structure of the antibody to promote dissociation or prevent re-association. As results in Figure 10 indicate, with high concentrations of fluid phase DNA, the level of anti-DNA activity measured decreases depending on the plasma.

Figure 10. The effects of soluble DNA on dissociation of anti-DNA antibodies from DNA.

The ability of soluble DNA to cause dissociation of anti-DNA from SLE plasma was analyzed by ELISA assay after establishment of equilibrium. For these determinations, dilutions of five SLE plasmas were first incubated with solid phase bound dsDNA (5μg/ml) for 1 h; the starting plasma dilutions were determined by prior titration. After the initial 1 h incubation, buffer was removed and the samples were then incubated with varying concentrations of soluble dsDNA (0, 5, 15 and 50 μg/ml) for 1 h. Panels A – E represent the results of the different plasmas. All treatments were performed in duplicate; standard deviations are shown.

To determine whether the nature of the bonds involved changed over time, we performed a kinetic experiment in which high concentrations of soluble DNA were added over time. Figure 11 presents these results. As these data indicate, the extent of dissociation varies with the length of incubation. While this effect was most evident with plasma #9, it could also be observed to a limited extent with another plasma tested, plasma #10. Together, these studies indicate that anti-DNA antibody binding mechanisms can evolve over time, with some antibodies becoming increasingly resistant to dissociation with either high concentrations of salt or DNA.

Figure 11. Effects of the duration of binding reactions of antibodies with DNA on subsequent dissociation.

Two plasmas (SLE #9 and #10) were incubated for 5, 20 or 60 min with plates coated with dsDNA (5μg/ml). After the initial binding reaction, incubation solution without DNA was replaced with solutions containing soluble dsDNA (0, 5, 15 or 50 μg/ml). The incubations had a total time of 120 min. Results for 50 μg/ml dsDNA are shown. Percent of anti-DNA activity was calculated as (OD 450 nm 1000 mM/OD 450 nm 150 mM)*100 for a given time point.

4. Discussion

The results presented herein provide new insights into the role of electrostatic or ionic interactions as determinants of anti-DNA avidity. The focus on electrostatic interactions derives from the physical chemical nature of DNA as antigen, with the phosphate groups in the phosphodiester backbone presently an extended array of negative charge (35). Consistent with the importance of negatively charged phosphate groups as antigenic targets, many anti-DNA antibodies contain positively charged amino acid residues in the complementarity determining regions (CDR) (4–6). Since anti-DNA antibodies bind widely to DNA antigens independent of species origin, the phosphodiester backbone, rather than sequence, has appeared to be the major determinant recognized (1).

In our studies, we have used the effects of ionic strength as a probe for bond types based on the idea that high salt conditions can influence electrostatic or coulombic interactions by charge shielding. The effect of high ionic strength on anti-DNA interactions is the principle that underlies the Farr binding assay, long considered a standard measure of high avidity binding (36). In the Farr assay, a source of radioactive DNA (usually a closed circle plasmid) is incubated with serum; at equilibrium, saturated ammonium sulfate is added to produce 50% saturation. At 50% saturation, antibody is no longer soluble and precipitates along with any DNA which is bound, with just a single antibody sufficient to allow precipitation.

Since high avidity antibody can withstand high concentrations of salt (saturated ammonium sulfate is 4.2 M), the Farr assay detects binding of only certain antibody subpopulations, presumably those that are high avidity. In contrast, use of polyethylene glycol (PEG) to precipitate immune complexes allows detection of a broader array of antibodies including low avidity specificities (37–39). Indeed, the ratio of antibody bound in ammonium sulfate compared to the amount bond in PEG provides a measure of the amount of high avidity antibodies (40, 41). Values of anti-DNA activity in the Farr assay can also differ from those in other assay approaches such as an ELISA, suggesting that the binding properties of anti-DNA are diverse, with the conditions of the assay determining which properties can be measured (39).

In the Farr assay, high salt conditions occur only after equilibrium is presumably reached. As such, the assay assesses the effects of ionic strength on dissociation, inferring avidity from resistance to dissociation. In contrast to studies using the Farr assay, the studies on the CLIFT assay, as well as ours, involve the effects of ionic strength on both association and dissociation. The studies on CLIFT indicated that certain sera showed significant binding with salt in the range of 1 to 5 M, implying substantial contribution to binding energy from bond types other than electrostatic or coulombic (20, 21). Importantly, these studies showed that certain sera, while sensitive to ionic strength in association assays, become resistant to increased concentrations of salt in dissociation assays. These findings are striking and suggest that bond types are not static during the interaction of DNA and anti-DNA and that non-electrostatic interactions can be increasingly favored over time.

In our study, we used an ELISA to compare the effects of ionic strength on both association and dissociation. An ELISA differs significantly from the CLIFT since it can assess the behavior of DNA in a linear conformation; the CLIFT is a closed circle and may have conformational restraints that differ from linear DNA since it can be supercoiled. Although the ELISA is characterized as a solid phase assay, we have presented evidence that at least some of the DNA in the ELISA exists in the fluid phase where it is highly flexible, adopting “soluble”-like structure, albeit attached to the polystyrene surface (42). The presence of “soluble” DNA results from the very high molecular weight of the DNA preparation used in the assay, with only some regions of DNA tightly adherent to the solid phase support.

In contrast to the situation with the ELISA, the actual disposition of the DNA in the CLIFT is not known since the organisms are fixed to a glass slide. Despite these differences, our findings, nevertheless, confirm that the binding of anti-DNA antibodies to DNA involves both electrostatic and non-electrostatic interactions and that the relative contribution of different interactions can change over time. These findings contrast with the usual model of anti-DNA binding based on the interaction of positively charge amino acids in the antibody combining site with negatively charged phosphodiester backbone. Therefore, anti-DNA binding is more complex than previously thought and a variety of bond types may contribute binding energy to the overall interaction.

The studies of Smeenk et al and Van Oss et al provided several examples of sera, considered to have high avidity, whose binding to DNA (association) was prevented at NaCl concentrations of 1.0 and 5.0 M but whose dissociation was, nevertheless, resistant to NaCl at the same concentrations (20, 21). These authors suggested that, over time, hydrogen bonding becomes an important mechanism for anti-DNA binding. The authors of these studies termed these binding properties as hysteresis, a term originally used in physics in the context of magnetization. Hysteresis signifies that the properties of a system depend on its history and that the current state depends on past events. While the studies on CLIFT indicated hysteresis with antibody bonding, they did not determine the time for the creation of new bonds.

In our studies, we identified one plasma that showed clear evidence of hysteresis. Thus, we demonstrated that the binding of plasma #9 could be prevented by ionic strengths in the 500 mM to 1000 mM range; after equilibrium was established at 150 mM, however, the binding became resistant to the same ionic strengths. In view of these findings, we performed experiments to determine the time required for the formation of new bonds. For this purpose, we incubated plasma with DNA for varying lengths of time (5, 20 and 60 minutes) at 150 mM NaCl followed by a second incubation in buffers of high ionic strength. These studies demonstrated that, while a component of anti-DNA was resistant to high ionic strength at 5 minutes of incubation with DNA at 150 mM NaCl, maximal resistance to the effects of increasing ionic strength was observed at 60 minutes. Together, these findings suggest that the establishment of new bonds is a relatively slow process, operating at a different time frame than the initial association of antibody and antigen which should be very rapid as it is limited by diffusion.

It is of interest that studies with both the CLIFT and the ELISA demonstrate similar phenomena despite the differences in the assays, including the nature of the DNA antigen, as well as the range of the NaCl concentrations used. In our study, the highest concentration of NaCl used for association was 1000 mM; at this ionic strength, we showed the most limited antibody binding. In contrast, Smeenk et al and Van Oss et al showed that some sera could bind at 2.0 and 5.0 M NaCl (20, 21). We would suggest that these differences reflect the form of the DNA antigen in the respective assays and any potential structural changes resulting from fixation. In addition, the other investigators had more examples of sera whose binding became resistant to high ionic strength. We are, therefore, assaying plasmas from different cohorts to better determine the frequency of different binding types.

In these experiments, we also investigated the effects of urea, a chaotropic agent, on antibody association and dissociation. The effects of chaotropic agents have been used widely to assess the avidity of serum or plasma antibody where the concentration of antibodies is unknown; this approach has been especially popular with the assessment of vaccine responses (26–31). Chaotropic agents can affect protein-protein interactions and bonds such as hydrogen bonds because of the effects on water structure around non-polar residues. Some chaotropic agents such as KSCN can also affect ionic interactions but, in our experiments, we used urea to assess better the contribution of non-ionic interactions.

The chaotropic agents have been used much more extensively to evaluate the antibody response to protein antigens where bond types differ from those of anti-DNA antibodies (26–31). Chaotropic agents, however, can lead to unfolding of proteins, a kind of denaturing effect that complicates the interpretation of avidity assays based on antibody dissociation with these agents. Interestingly, the anti-DNA antibody in the plasmas we studied showed significant binding activity in 1–2 M urea in both association and dissociation assays although, in general, binding activity was diminished in 4 M urea. Urea, however, can interact with DNA, suggesting that this agent can affect both antigen and antibody (32–34).

For plasma #9, which showed evidence of hysteresis, the effects of increasing urea concentrations were less marked than those of other plasmas. Furthermore, because plasma #9 showed resistance to dissociation at high salt, we could study the combined effects of high salt and urea. As these data showed, plasma #9 showed significant binding after dissociation with the combination of high salt and urea, further evidence for the nature of bond interactions that anti-DNA antibodies can form. Together, these studies point to differences between the effects of ionic strength and chaotropic agents in association and dissociation assays; they also suggest that the nature of the predominant bond type involved in antigen binding will affect the utility of using chaotropic agents to determine avidity for anti-DNA-DNA interactions.

To show further that anti-DNA binding can strengthen over time, we performed studies using high concentrations of DNA rather than high salt for dissociation. This approach has an advantage since high salt could potentially alter the conformation of DNA and perhaps prevent re-association of the antibody. Despite these differences in the assay conditions, nevertheless, we could demonstrate a similar time-dependent increase of antibody binding, showing less dissociation in the presence of high concentration of DNA. In this regard, the slow acquisition of a high avidity state can limit the use of antigen competition assays to determine affinity using Scatchard analysis.

While the acquisition of a high avidity state appears well established with two different assays, the basis of this finding remains unclear. Presumably, the slow kinetics suggests that DNA undergoes some type of rare or unusual structural change that exposes sites for which hydrogen bonding (or other non-electrostatic interactions) stabilizes binding. DNA can exist in a wide variety of backbone structures and it is possible that one of these structures is the preferred antigenic form; exposure of bases could also occur (43). It is also possible that anti-DNA antibodies induce changes in the DNA to generate a preferred structural site. Some anti-DNA antibodies are nucleolytic, suggesting that a gradual degradation of the DNA could lead to a new epitope that can be more tightly bound (44, 45).

Together, these findings provide further evidence for the unique nature of DNA-anti-DNA interactions, in particular, the slow acquisition of high avidity binding by a process that can involve hysteresis. In addition to providing information relevant to the design of assays with respect to salt conditions as well as assay duration, the data presented herein raise questions concerning the in vivo formation of immune complexes. Since avidity may be slow to develop, the levels of DNA and anti-DNA in the blood as well as hydrodynamic effects of blood flow may limit immune complex formation. Indeed, some studies suggest that levels of DNA-containing immune complexes in the blood may be low (46). In contrast, the formation of complexes of planted DNA antigen in the kidney may provide an environment where avidity can slowly develop. Future studies will address the key interactions involved in the establishment of high avidity and the impact on immune complex-mediated renal disease. A better understanding of these molecular interactions may also be valuable to the development of new therapeutic approaches targeting these complexes.

Key Points.

• The effects of ionic strength can indicate the bond types mediating anti-DNA interactions

• Differing effects on ionic strength in association and dissociation assays can indicate changes in anti-DNA interactions

• Increasing ionic strength can reduce anti-DNA binding in association assays

• Resistance to increasing ionic strength in dissociation assays can indicate development of non-ionic bonds

• High avidity anti-DNA binding can develop slowly over minutes and show evidence of hysteresis