Análisis y comprobación de las hipótesis de estudio

Modelling of the DNA binding site based on the crystal structure of several anti-DNA antibodies suggests that antibodies to ssDNA contain a deep cleft but antibodies to dsDNA have a planar binding (Eilat et al 1994). Herron et al (1991) have described a crystal of an anti-DNA antibody with ssDNA of the murine Fab BV-0401. A large irregular groove between the Vh and Vl regions

seems to comprise the binding site of BY 04-01. Apart from residues from the six CDRs which contributed to the walls of the groove, an arginine residue was also found to form an ionic interaction with the phosphate backbone of ssDNA. The ssDNA used in the crystal was a trinucleotide of deoxythymidylic acid (dT3). The central thymine of this trinucleotide forms hydrogen bonds between the planar rings of a tyrosine residue V lC D R I and a tryptophan residue from

VhCDR3. Yokoyama et al (2000) describe a crystal structure of an antibody Fab fragment (64M-2) in complex with a DNA photoproduct formed by ultraviolet radiation.

On the basis of data from crystal structures of monoclonal antibodies of known sequence, computer programs have been developed that can predict the three dimensional structure of an antibody from its amino acid sequence (Chothia et al 1989, Martin et al 1989). Modelling studies of human monoclonal anti- DNA antibodies were undertaken by Kalsi et al (1995a). Monoclonal anti-DNA

antibody B3 which binds dsDNA appears to do so in a deep cleft bound by arginines from VlCDRI and CDR2 and VhCDR2. The VlCDRI arginine was a

product of somatic mutation. However, in monoclonal antibody WRI176 which binds ssDNA more strongly than dsDNA, no cleft was observed. An exposed tryptophan residue from VhCDR2 was observed to play a role in making

important stacking interactions. This tryptophan is the only replacement mutation in the Vh sequence of WRI176.

In their studies, Hobby et al (1998) have reported the modelling of the three dimensional structure of the V region of AB-88, which is a IgG 1/kappa ssDNA binding antibody from a lupus mouse that bears a crossreactive idiotype (CRI) which is important in the pathogenesis of human and murine disease. The location of the CRI which lies within the framework region of VH is such that it could influence antigen specificity.

1.5.16 Crossreactivity of anti-DNA antibodies

The first molecules to which anti-DNA antibodies were found to be crossreactive were polynucleotides and phospholipids. Monoclonal anti-DNA antibodies from lupus prone mice were reported by Lafer et al (1981) that also bound phospholipids cardiolipin and phosphotidyl glycerol. Phosphodiester phosphate groups that are common to both DNA and crossreactive phospholipids may constitute the shared epitope for these crossreactive antibodies. Phosphodiester groups may also be responsible for crossreactivity between anti- DNA antibodies and bacterial antigens. Carrol et al (1985) have shown that anti- DNA antibodies from lupus prone mice bind endogenous microbial flora and this binding is inhibited by DNA. This binding was found to be due to the

phosphodiester components of the bacterial cell wall. Anti-DNA antibodies have been known to crossreact with mycobacterial cell walls (Shoenfeld et al 1986). Anti-DNA antibodies also crossreact with carbohydrate antigens. Increased levels of the common antiDNA idiotype 16/6 were found to occur in a higher proportion of sera from patients infected with klebsiella pneumoniae as compared to normal healthy controls. A human monoclonal antibody which binds bacterial capsular polysaccharides from group B meningococcus and Escherichia coli Kl, has also been shown to bind DNA as well (Kabat et al 1986). One of the reasons for crossreactivity between anti-DNA and anti-polysaccharide antibodies may be negative charge distribution.

Crossreactive antibodies to DNA and phosphorylcholine occur in mice (Limpanasithikul et al 1995, Ray et al 1996) as well as humans (Putterman et al 1996). PC is an immunodominant component of the pneumococcal cell wall polysaccharide. Antibodies to PC are found to occur naturally in humans and decline in titre with age. A strong association between high titer of antibodies to PC in serum of elderly subjects at 70 years of age and pneumonia related death upto 14 years later has been observed (Nordenstam et al 1990). There is no indication in this study as to why high levels of antibody to PC would be a risk factor for pneumonia. One of the explanations given by the authors is that individuals who are going to die of pneumonia in old age have abnormalities in one or more of their non specific anti-microbial defence mechanisms. This would predispose them to frequent subclinical and clinical infections, which in turn would serve to immunize them with microbial antigens and thus raise their serum levels of anti-PC antibody. Studies by Nordenstam et al (1990) show that most

elderly subjects later dying with pneumonia seemed to maintain their high levels of IgM anti-PC antibody however the anti-PC levels showed a statistically significant decline prior to fatal pneumonia in those still alive at age 79. Also antibodies to PC are less effective in protection as compared to antibodies to pneumococcal capsular antigens (Briles et al 1989). Thus declining levels of anti- PC in elderly individuals with frequent invasive pneumococcal infections may contribute to the susceptibility to pneumococcal infection.

Human anti-PC antibodies are predominantly of the IgG2 isotype presumably reflecting their origin in a T-independent antibody response, although in some individuals IgGl and IgG3 isotypes have been found (Scott et al 1987). Briles et al (1981) have shown that murine anti-PC antibodies, predominantly those expressing the T15 idiotype, are protective against pneumococcal infection. Diamond and Scharff (1984) showed that a single amino acid substitution of alanine for glutamic acid at residue 35 on the CDRl of the heavy chain converts a protective T15+ anti-PC antibody to one with (ds)DNA specificity.

Healthy individuals following vaccination with pneumococcal polysaccharide do not routinely produce anti-DNA antibodies (Grayzel et al

1991). However, Grayzel et al (1991) showed that anti-pneumococcal antibodies from non autoimmune individuals immunized with a polyvalent pneumococcal vaccine showed an anti-dsDNA associated idiotype, 31, which is found on kappa light chains of anti-dsDNA antibodies in patients with SLE. Livneh et al (1994) investigated serum from 14 lupus patients and 28 unaffected family members following vaccination by pneumococcal polysaccaride for the presence of antibodies bearing the 8.12 idiotype, a marker present on lambda light chains of

anti-DNA antibodies in 50% of lupus patients. Elevated titers of the 8.12 idiotype was found in the serum of 57% of SLE patients and in only 9% of family members. Following vaccination with pneumococcal polysaccharide, 8.12 reactive anti-pneumococcal antibodies were produced by 7 of 10 non­ autoimmune individuals and 8.12 reactive anti-DNA antibodies by one. These results suggest that 8.12 reactive antibodies are antigen driven and bind structurally related antigens and that anti-DNA antibodies may be activated in an anti-pneumococcal response. Idiotype measurements may act as useful V region markers for antibodies and may help in distinguishing between pathogenic and non pathogenic subsets of antibodies. Also idiotypes such as the T15+ idiotype found on protective anti-PC antibodies may serve as phenotypic markers for protective immune responses. Sharing of idiotypes between antibodies to different antigens may reflect crossreactivity of antibodies to structurally related antigens. In studies by Ray et al (1996) murine crossreactive anti-PC, anti- dsDNA antibodies were found to arise following PC immunization, and are regulated in the non autoimmune host by apoptosis. It was shown by Limpanasithikul et al (1995) that murine crossreactive anti-dsDNA and anti-PC antibodies not only protect against lethal pneumococcal infection but are also pathogenic and deposit in the kidneys of SCID mice. To investigate the crossreactivity of human polyclonal anti-dsDNA and anti-PC antibodies, studies regarding the same were conducted by me.

In studies by Kowal et al (1999), a high percentage of Fabs reactive with pneumococcal polysaccharide derived from a combinatorial library generated from the spleen of a patient with SLE who had been vaccinated with pneumovax.

were also found to crossreact with dsDNA. This combinatorial library was prepared from the spleen of a lupus patient who had been given a pneumococcal vaccine ten days before splenectomy. In principle, combinatorial libraries are formed by random associations of heavy and light chains of preexisting antibodies and are prepared by cloning the variable regions of inununoglobulin heavy and light chains from the antibodies ( Vh and Vl) into a phagemid.

Phagemids are plasmid based vectors which can replicate as ssDNA and be packaged into bacteriophage (virion which infects bacteria) particles. The recombinant antibody particle can be produced as a fusion protein, connected covalently to a surface protein of the phage particle. This is known as phage display (Johnson and Chiswell 1993). Phage display vectors can be used to produce surface bound scFv or Fab. Various studies comparing antibodies derived from combinatorial libraries to antibodies produced by B cell hybridomas or transformed B cell lines have implicated striking similarities of heavy and light chain pairing (Caton et al 1990, Reason et al 1997).

However due to the procedures used in the construction of the combinatorial libraries and the random combinations of the VH and VL regions, the antibodies obtained from these libraries may not always be representative of the in vivo response. Hence the diverse array of VH/VL combinations from these combinatorial libraries may also result in the formation of antibody structures that are not, or cannot usually be found within a host. Gherardi and Mil stein (1992) in their studies have compared anti-2-phenyloxazolone (phOx) antibodies obtained by hybridomas with single chain Fvs (Fragment variable) from a combinatorial library made with mRNA from the same pool of spleen cells. The

random combinatorial library had been selected on a phOx affinity column and clones were ranked by ELISA. Although both the methods produced good binders, the best binder was isolated from a hybridoma. Also the gene combination of this antibody isolated from the hybridoma was not obtained from the random combinatorial library, and nor were other gene combinations obtained from the combinatorial library typical of the anti-phOx response. Hence the combinatorial library approach may not always lead to the recovery of the original pairs of H and L chains expressed by individual B cells.

Kowal et al (1999) selected eight Fab clones on the basis of expression of a common idiotype 31. The 31 idiotype is expressed on kappa light chains of anti- dsDNA antibodies and is present in elevated titers in about 80% of patients of SLE whose serum is dsDNA reactive. Six out of the eight Fab fragments were encoded by the Vh3 gene family. One Fab was encoded by the VhI family and one by the Vh6 family. Six of the light chains were derived from the V^l gene family. The light chains of the remaining two clones were encoded by A27, a germline gene belonging to the Vk3 family, which is highly represented in the human Vk repertoire and occasionally dominates the anti-pneumococcal response

in humans (Sun et al 1999). Most of the clones showed binding to at least one of the bacterial antigens PC or pneumococcal polysaccharide and four of them showed strong binding to dsDNA. Inhibition studies on three crossreactive antibodies showed that dsDNA and polysaccharide bind in the same or a proximal site on these antibodies. Inhibition assays with a wide range of inhibitor concentration also demonstrated that both dsDNA and pneumococcal

polysaccaride bound with high avidity and in a similar fashion throughout the range of inhibitor concentration.

Similar studies were conducted by me wherein I analysed a combinatorial library for anti-dsDNA antibodies expressing the 8,12 idiotype and crossreactive with the pneumococcal antigen in order to elucidate the potential role of bacterial antigens in triggering the anti-dsDNA response in SLE.