Lupus anti-DNA autoantibodies cross-react with a glomerular structural protein: a case for tissue injury by molecular mimicry.

Anti-DNA autoantibodies are the hallmark of human and murine systemic lupus erythematosus (SLE), an autoimmune rheumatic disease of unknown etiology. Some of these antibodies are believed to be pathogenic for kidney tissue and to initiate immune glomerulonephritis. However, the mechanisms by which anti-DNA antibodies participate in tissue injury remain controversial. We have studied the in vivo pathogenicity of anti-DNA monoclonal antibodies in immune deficient mice, using a panel of murine B cell hybridomas. No consistent genetic or immunochemical differences were found between pathogenic and non-pathogenic anti-DNA antibodies. However, the two antibody populations differed in their cross-reaction with the acidic actin-binding protein, alpha-actinin, that is known to play a major role in the structural integrity of glomerular filtration components. These results suggest that kidney dysfunction in SLE may be facilitated by protein-nucleic acid antigenic mimicry.

Production of high affinity autoantibodies in autoimmune New Zealand Black/New Zealand white F1 mice targeted with an anti-DNA heavy chain.

Lupus-prone, anti-DNA, heavy (H) chain "knock-in" mice were obtained by backcrossing C57BL/6 mice, targeted with a rearranged H chain from a VH11(S107)-encoded anti-DNA hybridoma (D42), onto the autoimmune genetic background of New Zealand Black/New Zealand White (NZB/NZW) F1 mice. The targeted female mice developed typical lupus serologic manifestations, with the appearance of transgenic IgM anti-DNA autoantibodies at a young age (2-3 mo) and high affinity, somatically mutated IgM and IgG anti-DNA Abs at a later age (6-7 mo). However, they did not develop clinical, lupus-associated glomerulonephritis and survived to at least 18 mo of age. L chain analysis of transgenic anti-DNA Abs derived from diseased NZB/NZW mouse hybridomas showed a very restricted repertoire of Vkappa utilization, different from that of nonautoimmune (C57BL/6 x BALB/c)F1 transgenic anti-DNA Abs. Strikingly, a single L chain was repetitively selected by most anti-DNA, transgenic NZB/NZW B cells to pair with the targeted H chain. This L chain had the same Vkappa-Jkappa rearrangement as that expressed by the original anti-DNA D42 hybridoma. These findings indicate that the kinetics of the autoimmune serologic manifestations are similar in wild-type and transgenic lupus-prone NZB/NZW F1 mice and suggest that the breakdown of immunologic tolerance in these mice is associated with the preferential expansion and activation of B cell clones expressing high affinity anti-DNA H/L receptor combinations.

Early anti-nucleosome autoantibodies from a single MRL+/+ mouse: fine specificity, V gene structure and pathogenicity.

In systemic lupus erythematosus, the nucleosome assumes a central role in the autoimmune response to self antigens. To gain insight into the etiology and pathogenesis of anti-nucleosome antibodies (Ab), we analyzed a panel of six IgG-secreting hybridomas derived from a single young MRL +/+ mouse at the onset of the autoimmune response. All monoclonal antibodies (mAb) bound exclusively the native nucleosome, and represented five different clonotypes that recognized diverse nucleosomal epitopes, typical of a polyclonal response. The VH-complementarity-determining region (CDR)3 regions exhibited unique stretches of charged amino acids with different polarity that may be important for the interaction with the nucleosome. These early anti-nucleosome mAb displayed striking structural differences with not only anti-DNA, but also with anti-nucleosome Ab, that appear later in disease. Two of the mAb deposited in kidney glomeruli after in vivo administration to RAG-1-deficient mice, suggesting that diverse B cell clones, possibly selected by the nucleosome itself, may play a role in the initiation of kidney damage.

B cell deletion, anergy, and receptor editing in "knock in" mice targeted with a germline-encoded or somatically mutated anti-DNA heavy chain.

To study the relative contributions of clonal deletion, clonal anergy, and receptor editing to tolerance induction in autoreactive B cells and their dependence on B cell receptor affinity, we have constructed "knock in" mice in which germline encoded or somatically mutated, rearranged anti-DNA heavy (H) chains were targeted to the H chain locus of the mouse. The targeted H chains were expressed on the vast majority of bone marrow (BM) and splenic B cells and were capable of Ig class switching and the acquisition of somatic mutations. A quantitative analysis of B cell populations in the BM as well as of Jkappa utilization and DNA binding of hybridoma Abs suggested that immature B cell deletion and light (L) chain editing were the major mechanisms affecting tolerance. Unexpectedly, these mechanisms were less effective in targeted mice expressing the somatically mutated, anti-DNA H chain than in mice expressing the germline-encoded H chain, possibly due to the greater abundance of high affinity, anti-DNA immature B cells in the BM. Consequently, autoreactive B cells that showed features of clonal anergy could be recovered in the periphery of these mice. Our results suggest that clonal deletion and receptor editing are interrelated mechanisms that act in concert to eliminate autoreactive B cells from the immune system. Clonal anergy may serve as a back-up mechanism for central tolerance, or it may represent an intermediate step in clonal deletion.

Expression of an anti-DNA-associated VH gene in immunized and autoimmune mice.

The J558 family BW-16 VH gene is closely associated with the autoimmune response to DNA of lupus-prone mice. We have followed the expression of VHBW-16-encoded heavy chains in cDNA libraries prepared from unmanipulated normal (C3H, AKR) and autoimmune (New Zealand Black/New Zealand White F1) mice, and from mice immunized with a highly immunogenic peptide/DNA complex. The prevalence, clonal heterogeneity, and structural properties (somatic mutation, complementarity-determining region 3 composition) of these H chains were investigated, and the DNA affinity of VHBW-16-encoded hybridoma mAb was measured in solution. We find that H chains encoded by VHBW-16 are very rare in Igs of normal mice, but increase significantly in peptide/DNA-immunized mice, and dramatically in diseased mice. The experimentally induced VHBW-16-encoded H chains are clonally restricted, somatically mutated, partly switched from IgM to IgG, and give rise to anti-DNA Abs with low affinity. In contrast, the VHBW-16-encoded H chains from diseased New Zealand Black/New Zealand White mice are clonally heterogeneous, exclusively of the IgG isotype, and produce high affinity anti-DNA autoantibodies. We conclude that the experimentally induced and spontaneous immune responses to DNA are qualitatively similar, but quantitatively different, and may truly reflect the principles of self immunologic tolerance.

B cell development under the condition of allelic inclusion.

Mice whose IgH alleles are engineered to encode two distinct antibody heavy (H) chains generate a normal-sized B cell compartment in which most cells stably express the two heavy chains. This demonstrates that "toxicity" of bi-allelic H chain expression and cell-autonomous mechanisms of silencing in-frame IgH gene rearrangements do not significantly contribute to allelic exclusion at the IgH locus. Notwithstanding, the stability of the various engineered IgH loci during B cell development in the bone marrow differed substantially from each other.

Structural elements controlling anti-DNA antibody affinity and their relationship to anti-phosphorylcholine activity.

The heavy chains of the anti-DNA autoantibody D42, derived from an autoimmune NZB/NZW F1 mouse, and the anti-phosphorylcholine (PC) Ab 6G6, obtained from an immunized CBA/J mouse, are both encoded by the same VH11 (S107) germ-line gene. We have combined VH CDR3 exchange, site-directed mutagenesis, transfection of mutant hybridoma cells, and equilibrium affinity measurements of the expressed Abs to analyze the contribution of germ line-encoded genetic elements and somatic mutations to DNA binding affinity as well as to test the possibility that an anti-PC Ab can be converted to an anti-DNA Ab by somatic mutation. Our results indicate that the major contribution to DNA binding affinity is provided by the nonmutated rearranged configuration of the autoantibody and depends on the particular H/L chain pairing and the structure of the VH CDR3 gene segment. Two heavy chain somatic mutations increased the DNA binding affinity of the D42 anti-DNA Ab by about 10-fold. One of these mutations abolished PC binding of the 6G6 Ab, but did not convert it to an anti-DNA. These results suggest that activation of autoreactive B cells bearing germ-line-encoded Ag receptors may be required to initiate the affinity maturation of an anti-DNA response, and that B cells responding by somatic mutation to external Ags, such as PC, may be unable to contribute significantly to the high affinity immune response that is typical of disease-related anti-DNA autoantibodies.

Structure-function correlates of autoantibodies to nucleic acids. Lessons from immunochemical, genetic and structural studies.

Nucleic acid binding autoantibodies are the hallmark of the human autoimmune disease, systemic lupus erythematosus (SLE) and are also prevalent in mouse models of this disease. The immunologic stimuli for the production of these antibodies as well as their pathogenic mechanisms are not well understood. However, extensive immunochemical and genetic studies, together with initial crystallographic analysis and computer modeling, have suggested several structure-function correlates which will form the basis for future research. The anti-DNA and anti-RNA autoantibodies comprise a continuous spectrum of specificities in which a delicate balance exists between the binding to the sugar-phosphate backbone and the interactions with the heterocyclic bases of the nucleic acid. Prominent in these interactions are the products of specific V-region immunoglobulin genes, some of which appear to be uniquely suitable for nucleic acid binding. Other structural elements encoded by D minigenes, N sequences and somatic mutations, help to increase the affinity of the binding interaction, and may also increase the repertoire of nucleic acid binding antibodies by combining with a relatively large number of additional V-gene products. Initial crystallographic analyses of anti-DNA antibodies indicate some fundamental differences in the structure and shape of ssDNA and dsDNA antibody combining sites. However, they also suggest a considerable degree of flexibility of both antibody and antigen, which is induced by their binding interaction.

Primary sequence and location of the idiotopes of V-88, a DNA-binding monoclonal autoantibody, determined by idiotope scanning with synthetic peptides on pins.

In this study, the primary sequence and location of the idiotopes of monoclonal antibody (mAb) V-88 have been examined. V-88 was derived from an adult (NZB x NZW)F1 mouse, has been partially defined previously with polyclonal anti-idiotype antisera, and is a member of the 16/6 idiotype (Id) family. From the inferred primary amino acid sequence of the antibody, sets of hexapeptides, overlapping by five residues, were synthesized on pins and used to scan the expression of epitopes (idiotopes) in the V regions of the light and heavy chains. A heterologous rabbit antiserum raised against the native antibody V-88, and absorbed to make it idiotype specific, was found to react with eight major epitopes distributed between the VH and VL regions. Half of these determinants mapped to the complementarity determining regions, with the others in framework sequences. Thus, the idiotype of antibody V-88 comprises, at least in part, continuous linear idiotopes in both hypervariable and framework areas. The process of absorbing the anti-idiotype antiserum on normal mouse immunoglobulin removed much of the background antibody activity against V region peptides, but left the activity against the dominant idiotopes. The sequence of a major idiotope, VATISG, in the FW2/CDR2 VH region is homologous to sequences of human antibodies that express the 16/6 idiotype, suggesting that Id.16/6 is at least in part defined by this region of the antibody. The same VH area is also homologous to sequences in bacterial and mammalian heat-shock proteins (hsp60-65). Thus there may be a functional link through idiotype connections, especially those involving Id.16/6, between anti-bacterial responses and production of autoantibodies, and some bacterial antigens may function indirectly as superantigens for B cells.

Cross-reactions of anti-DNA autoantibodies with cell surface proteins.

Anti-DNA autoantibodies are thought to play a major role in the pathogenesis of systemic lupus erythematosus. However, the mechanism(s) by which they participate in tissue and organ damage is not well understood. It has been suggested that these antibodies combine with DNA or DNA-histone complexes to produce circulating immune complexes which may deposit in various tissues. Alternatively, anti-DNA autoantibodies could interact directly with tissue components by way of immunological cross-reaction. In this study we have used a panel of mouse monoclonal autoantibodies with anti-nuclear specificity and measured their binding to membrane proteins of several tissues and cell lines. We show that the anti-DNA antibodies, but not anti-RNA or anti-histone antibodies bind to membrane proteins of molecular weights 102, 80, 42, 35 and 31 kDa, which are expressed in different combinations on several cell types. The binding of anti-DNA antibodies to these cell surface proteins was not affected by DNase treatment of the target cells, was increased by DNase treatment of the antibody preparations and was completely inhibited by DNA, indicating a true cross-reaction and not an indirect interaction of antibody and membrane proteins through a DNA bridge. Our results suggest that direct binding of anti-DNA autoantibodies to cell surface membrane proteins may play an important role in the induction of the pleomorphic tissue damage in systemic lupus erythematosus.

Structure and binding properties of a monoclonal anti-idiotypic autoantibody to anti-DNA with epibody activity.

We report the isolation and characterization of a mouse autoanti-idiotypic mAb (D7.4 IgG2a), which is directed against a major public Id (A52 IgG2b) in the murine and human autoimmune response to DNA. The natural anti-Id mAb has been produced in the course of the SLE-like disease in a female NZB/NZW F1 mouse and showed a dual specificity (epibody activity) for the public Id (A52) and for the autoantigen (DNA). The two binding activities were shown to reside in the Fab portion of the epibody and were highly specific for their respective Ag. A complete nucleotide sequence analysis of the D7.4 H and L chain V-region genes combined with computer comparisons to available Ig sequences may suggest a charge interaction between the H chain CDR3 segments of the Id and anti-Id antibodies. The D7.4 epibody may be a component of the self-binding, idiotypically connected network of natural antibodies. Alternatively, it could be elicited against the potentially pathogenic, DNA-containing immune complexes in order to facilitate their removal from the circulation of diseased individuals.

Secretion of a soluble, chimeric gamma delta T-cell receptor-immunoglobulin heterodimer.

Soluble derivatives of T-cell antigen receptors (TCRs) should prove invaluable for studying the interaction of these receptors with antigens and major histocompatibility complex molecules, for structural studies, and for the identification of unknown ligands. We have engineered chimeric proteins, containing the extracellular domains of the mouse V gamma 1.1-C gamma 4 and V delta 6.2-C delta (V, variable; C, constant) TCR chains fused to the hinge region, CH2 (H, heavy), and CH3 domains of human IgG1 heavy chain, and expressed them by transient transfection in COS cells. We show here that TCR gamma-IgH and TCR delta-IgH chimeric chains are produced intracellularly in significant amounts, that the two chains can assemble correctly to form disulfide-linked, glycosylated heterodimers, and that a selective mechanism allows secretion of correctly paired receptor chains into the medium. Identity of the chimeric secreted TCR gamma delta-IgH heterodimer was confirmed by immunoblot analysis using V gamma 1-specific anti-peptide antiserum and immunoprecipitation analysis using the monoclonal antibody UC7, which is shown to be specific for the TCR delta chain. In addition, the soluble TCR gamma delta-IgH heterodimer can be immunoprecipitated with the anti-clonotypic monoclonal antibody F10/56, which suggests that the fusion protein likely has a structural conformation similar to that of the native TCR. The COS cell expression system may prove useful for the production of additional TCR-IgH fusion proteins.

Recurrent utilization of genetic elements in V regions of antinucleic acid antibodies from autoimmune mice.

Two IgG anti-DNA and two IgG anti-RNA autoantibodies derived from lupus prone NZB/NZW F1 mice have been analyzed for their Ag fine specificities and for their H and L chain V-region sequences. A remarkable similarity of VH gene sequences with previously sequenced antinucleic acid autoantibodies (Eilat, D., D. M. Webster and A. R. Rees. J. Immunol. 141:1745, 1988) was noted. This finding indicates that a small number of unique VH genes is involved in this autoimmune response and that the sequences of these genes are correlated with the different specificities for the autoantigen. The VK sequences appeared, by computer search, to be selected nonrandomly, but their use was not restricted to autoantibodies. An additional striking feature was evident in the construction of the D region elements, giving rise to CDR3 peptides that can interact with DNA and RNA. These constructs probably include D-D fusion products, which are relatively rare in Ig rearrangements.

The role of germline gene expression and somatic mutation in the generation of autoantibodies to DNA.

Several distinctive features of anti-DNA autoantibodies have been identified by a detailed analysis of the available heavy and light chain sequences. They include unique VH gene segments that are not normally expressed in antibodies to external antigens, somatic mutations which may serve to change the antigenic specificity as well as to increase affinity, a less stringent choice of light chains, and a unique basic peptide in the heavy chain CDR3. It is proposed that in the majority of cases, the regulatory mechanism of self-tolerance in the healthy animal operates via VH gene expression to prevent the synthesis of potentially high affinity anti-DNA autoantibodies.

[Genetic origins of anti-DNA autoantibodies]. / Origines génétiques des auto-anticorps anti-ADN.

A detailed analysis of variable region sequences, derived from monoclonal lupus anti-DNA autoantibodies has led to several conclusions with respect to the etiology and pathogenesis of systemic lupus erythematosus (SLE). It was found that the gene elements which are expressed in anti-DNA heavy and light chains are present in the germ lines of normal and diseased animals. However, several VH gene segments are not normally expressed in antibodies to external antigens and may, therefore, be excluded by the regulatory mechanism of self-tolerance. The presence of somatic point mutations, which may serve to change antigenic specificity as well as to increase affinity, indicates that an antigen (DNA)-driven mechanism rather than a polyclonal activation is responsible for the autoimmune response in SLE. An additional unique motif of anti-DNA heavy chains was found to consist of an arginine-rich basic peptide in the third complementarity-determining (hypervariable) region. These data have enabled us to build a computer model of the anti-DNA binding site and to design specific inhibitors of the autoimmune reaction.

The VH gene sequences of anti-DNA antibodies in two different strains of lupus-prone mice are highly related.

The heavy and light chain V region sequences of an IgG anti-DNA autoantibody (PME77), derived from a lupus-prone (NZB x NZW)F1 mouse have been determined by mRNA sequencing. The V kappa gene segment belongs to the V kappa 1A gene sub-group and is found in several (NZB x NZW)F1 and MRL lpr/lpr anti-DNA antibodies, as well as in other antibodies of unrelated specificities. The VH gene segment appears to represent a unique gene or a subfamily of the large J558 VH gene family of the mouse, and is highly related to a germ-line sequence of a major anti-DNA idiotype (H130, IgM) of MRL mice. This anti-DNA-related VH segment has not been found, so far, to be expressed in antibodies with specificities for external or synthetic antigens; therefore, expression of such specificities may be regulated by powerful mechanisms of self tolerance in the healthy animal. In addition, both the heavy and light chain of the PME77 IgG antibody were found to contain somatic point mutations with a high ratio of replacement to silent mutations in complementarity determining regions. This IgM to IgG sequence relationship suggests an affinity maturation process, which is driven by the autoantigen.

IgG rheumatoid factor in purified IgG fractions and whole sera from patients with rheumatoid arthritis.

There are inherent technical difficulties in measuring IgG rheumatoid factor (IgG-RF) in the serum of patients with rheumatoid arthritis (RA). These arise from measuring a reaction between two IgG molecules and the interference of IgM-RF in the reaction. We compared the prevalence of IgG-RF in whole sera and purified IgG fractions from 58 RA patients (43 of whom were latex or sheep cell agglutination positive). Methods of purification were: ammonium sulphate precipitation and DEAE cellulose or protein A-Sepharose chromatography. IgG-RF was measured by two methods: (1) radioimmunoassay and ELISA with a monoclonal myeloma IgG (IgG4,K) as the antigen and radiolabelled rabbit anti-human IgG (previously absorbed on a column with IgG4,K) as the second antibody; (2) ELISA using rabbit IgG as the antigen and a peroxidase conjugated goat anti-human IgG as the second antibody. When whole sera were assayed, 18 (31%) contained IgG-RF. In contrast, only three of the IgG fractions (5%) were positive for IgG-RF by all methods, while the remainder were uniformly negative. These results suggest that IgG-RF determination in whole sera does not accurately reflect IgG-RF activity.