Assay Formats

The assay of anti-dsDNA has shown continuous refinement, resulting from innovations in the form of the DNA used as substrate as well as the method for antibody detection. The following assays have been used in the clinical setting: complement fixation, Farr-type immunoprecipitation, Crithidia luciliae immunofluorescence, filter binding, solid-phase radiobinding assay, and enzyme-linked immunoabsorbent assay (ELISA). These assays differ in the source of DNA used as antigen as well as the spectrum of antibodies that can be detected.^4,6 For example, a Farr-type immunoprecipitation assay involves the formation of an immune complex with a radiolabeled DNA that can be precipitated by ammonium sulfate. This complex must have sufficient avidity to remain intact in high salt. As such, a Farr assay likely detects a more limited subset of high-avidity antibodies than an ELISA (which can detect lower-avidity antibodies due to high antigen density at the solid-phase surface and the potential for cross-linking). (See Table 22A.1.)

TABLE 22A.1 PROPERTIES OF ANTI-DNA ANTIBODIES IN NORMAL SUBJECTS AND PATIENTS WITH SLE

In general, anti-DNA assays provide useful information for diagnosis as well as prognosis, although results of various assays of individual sera can differ based on the immunochemical properties of the antibodies present. Among assays, an ELISA likely detects the broadest number of specificities because it measures low- as well as high-avidity antibodies. In a mature antigen-driven response, the significance of low-avidity specificities is uncertain. An ELISA does not entail the use of radioactivity, however, and facilitates high throughput screening because of the use of a multi-well plate platform. These features make an ELISA an attractive choice for the clinical laboratory despite the detection of lower-avidity specificities.^7–11

There are several features of the routine testing of anti-DNA antibodies that bear note. The first concerns the dichotomy of anti-ss versus anti-dsDNA antibodies. Although antibodies to ssDNA can occur in patients with diagnoses other than SLE, sera from patients with SLE in general bind both antigenic forms. Indeed, as shown by cross-inhibition studies as well as the characterization of monoclonal antibodies, many antibodies bind ss and dsDNA. This pattern of specificity likely reflects antibody interaction with a determinant on the phosphodiester backbone that can be presented irrespective of strandedness.¹² Sera with exclusive specificity for dsDNA are in fact uncommon among SLE patients, and in patient sera anti-ssDNA occur more frequently than anti-dsDNA. Anti-ssDNA antibodies are also technically easier to measure and provide more sensitive assays.

Antibody Avidity and Specificity

Another issue regarding anti-DNA assays centers on avidity. As an antigen, DNA presents a repeating structure that allows a single antibody to contact epitopes on an extended polynucleotide chain via each Fab site of an IgG molecule. This type of binding, termed monogamous (or bivalent) interaction, leads to a dramatic increase in antibody avidity because of cross-linking. Thus, although each Fab site can contact only a few nucleotides, most sera require much larger pieces of DNA for binding. These pieces are generally at least 35 to 40 nucleotides in length, a span that covers the distance between each Fab site. Furthermore, some antibodies require DNA pieces hundreds of bases long for binding, likely because of low concentration of each epitope on a DNA molecule or the need for a conformational change in the DNA chain for the juxtaposition of each epitope.^13–15 For DNA, like other multivalent antigens, the term avidity is relative. (See Box 22A.1.)

Finally, although antibodies to DNA can be readily detected in patient sera, DNA (both inside and outside the cell) exists in the form of nucleosomes, the basic structure of DNA in chromatin. In this structure, DNA is wrapped around a histone core and binds tightly to proteins. As an antigen, therefore, DNA can be considered a component or epitope of chromatin (with many antibodies to chromatin showing sufficient interaction with free DNA to allow detection in the absence of the protein components).^16,17 Although chromatin preparations may mimic more closely the antigenic form of DNA in vivo, they are less well defined antigenically, leading to preference in the use of a purified DNA for serologic assays.

Anti-DNA Expression in SLE

In the context of SLE, dsDNA in the B conformation is the relevant antigenic determinant. This structure is widely expressed on DNA independently of species origin and synthetic dsDNA molecules (depending on sequence). With dsDNA as an antigen, anti-DNA expression is highly specific for SLE and occurs rarely in patients with other clinical diagnoses. These antibodies are expressed in approximately 50% of patients at some time during the course of their illness.^4,5 Frequently, antibodies to DNA are expressed concomitantly with antibodies to histones and other components of chromatin. Such expression, called linkage, likely reflects the role of nucleosomes in driving autoantibody production in SLE. The association is not invariable, however, because antibodies to histones can occur in the absence of antibodies to DNA in drug-induced lupus.^5,16,17

Anti-DNA (in contrast to other antinuclear antibodies in SLE) shows highly variable levels of expression, leading to its utility in assessing prognosis and disease activity as well as diagnosis. In longitudinal studies, anti-DNA expression in individual sera can range from undetectable levels to striking amounts in terms of titers. Frequently, high levels of anti-DNA expression are associated with an intensification of disease activity, in particular of glomerulonephritis. In the clinical setting, a depression in complement levels often accompanies increased anti-DNA levels, pointing to a role of immune complexes that deposit in the kidney in the immunopathogenesis of nephritis.^4–6

Whereas an elevation of anti-DNA levels can mark the worsening of renal disease, serologic and clinical disease activity can be discordant. Thus, patients with serologic activity (i.e., increased anti-DNA levels) may lack nephritis and patients with active nephritis may have only low levels of anti-DNA. Both situations can be readily explained. Thus, although anti-DNA production may be pathologic, only a subset of these antibodies may be pathogenic or nephritogenic. The properties conferring nephritogenicity are not well defined, although they likely depend on avidity, charge, and fine specificity for DNA. Existing assays do not distinguish those specificities that cause renal disease. The reverse situation (i.e., active nephritis without anti-DNA) may result from a failure of a particular assay to measure anti-DNA or from the role played by another autoantibody system in nephritis.^18–22

The utility of anti-DNA measurements in staging other disease manifestations is much less clear. Thus, despite the value of anti-DNA determinations in nephritis, this antibody should not be viewed as a general measure of disease activity. Other issues concerning the role of anti-DNA as a marker involve the magnitude of change viewed as clinically significant and the timing with respect to flares. These issues have increased in relevance in the context of drug development with agents that can specifically lower anti-DNA production. At present, anti-DNA cannot be considered a surrogate marker for disease that can guide the development of new agents or the utilization of existing agents in routine care. For those patients in whom anti-DNA expression correlates with activity of nephritis this antibody system is nevertheless of a very useful laboratory test.

Anti-DNA Expression in Normal Immunity

As a molecule, DNA is structurally enormously diverse because of sequence microheterogeneity. Although this molecular diversity has been extensively characterized in the context of gene regulation, its potential role in immunology was long neglected. This neglect resulted from several factors: (1) the focus on clinically useful assays that facilitate the detection of antibodies to dsDNA in the B conformation, (2) the seemingly exclusive expression of anti-DNA antibodies in SLE, implying responses to DNA only in conditions of aberrant immunity, and (3) the prevailing belief that with respect to the immune system DNA is structurally simple and uniform, with single- and double-stranded conformations the only relevant antigenic forms. Because anti-DNA assays uniformly support the high association of anti-DNA with SLE, there was little reason to question the prevailing dogma on the antigenicity of DNA.

Studies begun in the 1980s have redefined the antigenic properties of DNA, and the expression of anti-DNA in normal and aberrant immunity. These studies, which have explored the antigenicity of base sequence as well as conformation, originated in a survey of the binding of sera from patients with SLE and normal human subjects (NHS) with a panel of naturally occurring DNA antigens that differed in species origin and included both bacterial and mammalian molecules. The rationale for assaying DNA from various species was to enlarge the spectrum of sequential determinants for testing.²³

The results of these studies were remarkable and refuted the notion that anti-DNA expression is specific for SLE. Thus, these studies showed that sera from NHS contain antibodies to dsDNA from certain bacterial species.^23–28 As shown by various immunochemical approaches, these antibodies differ significantly in immunochemical properties from antibodies found in SLE patients in their specificity, avidity, and pattern of light-chain and isotype expression (Table 22A.1). Importantly, these antibodies are highly specific in their binding to DNA from a given bacteria and do not cross-react with either mammalian DNA or DNA from another bacteria. This reactivity indicates interaction with a nonconserved determinant. In contrast, antibodies in SLE sera show broad binding to DNA, and as expected for antibodies binding to a conserved determinant (i.e., B DNA), show cross-reactivity with both bacterial and mammalian DNA. In general, levels of antibodies to bacterial DNA in NHS antibodies are similar to those of patients with SLE when measured with the same antigen.

In their properties and expression, antibodies to bacterial DNA in normal human sera resemble antibodies to bacterial carbohydrates. These findings suggest that bacterial DNA can serve as an immunogen in normal immunity and drive the production of specific antibodies in ordinary encounters with microorganisms during infection or colonization. Because bacterial DNA differs from mammalian DNA in sequence, an obvious basis for antigenic recognition is present. The generation of antibodies to bacterial DNA is selective, however, and sera from NHS lack appreciable amounts of antibodies to DNA from many common sources, including E. coli. The reason antibodies are generated against only certain bacterial DNA is unknown.