Treatment of murine lupus with cDNA encoding IFN-γR/Fc (original) (raw)

We demonstrated that: (a) intramuscular injections of plasmids with cDNA encoding IFN-γR/Fc protected MRL-Fas_lpr_ mice from development of lymphoid hyperplasia and lupus-like disease, (b) electroporation at the injection site significantly enhanced the expression of the fusion protein and its protective effects, and (c) this mode of treatment was highly effective in prolonging survival and ameliorating the serologic and histologic parameters of the disease, even when initiated at the advanced stage, thereby making it a good candidate for human application.

Previous studies have attempted to neutralize IFN-γ in mouse lupus models using polyclonal Ab’s (9) or mAb’s (21), as well as soluble recombinant IFN-γR (22). These approaches, however, have limitations: for example, with regard to Ab’s, large quantities may be required, the Ab may not attain sufficient concentrations in secondary lymphoid organs or sites of inflammation, and it may be neutralized by host immune responses. With regard to soluble recombinant receptors, rapid turnover may affect efficacy and necessitate frequent administration. These constraints might explain the negative result with anti-IFNγ mAb treatment of MRL-Fas_lpr_ mice reported previously (21) and the finding that treatment with recombinant soluble IFN-γR in (NZB × W)F1 lupus mice was effective only when initiated early, but not late, when IFN-γ levels are significantly higher (22).

Several of these problems may be overcome by intramuscular injection of nonviral vectors, which can induce sustained production of cytokine agonists or antagonists. This approach was originally implemented in MRL-Fas_lpr_ mice by Raz et al. (23, 24), who showed that injections of a cDNA vector encoding TGF-β commencing early in life significantly reduced disease parameters. We have now used a nonviral vector encoding a secreted IFN-γR/Fc (IFN-γ inhibitory fusion protein) as a means to assess its prophylactic and therapeutic effects in this lupus model. We used IFN-γR fused to IgG1 Fc as the inhibitor instead of the truncated receptor alone, since fusion molecules secreted as homodimers have been reported to have much longer half-lives than truncated IFN-γR (40 vs. 1–3 hours, respectively) (25, 26), and dimeric IFN-γR/Fc fusion proteins exhibit higher ligand avidity than single-chain receptors (27). Finally, IgG1 Fc, commonly used in these cases, was chosen as the partner to create the enhanced half-life–displaying biomolecule because this IgG subclass does not activate complement. In addition, to enhance systemic fusion protein expression, we coupled the intramuscular injections with local electroporation. Among the many methods used to promote naked plasmid DNA cellular transfer and gene expression, application of low field strength, square-wave electric pulses through external or invasive electrodes appears to produce markedly higher transfer efficiency (18, 28–31). This has been attributed to increased numbers of muscle fibers that take up plasmid DNA and probably increased copy number of plasmids introduced into each muscle cell. Indeed, when we applied this procedure, levels of IFN-γR/Fc in the majority of treated animals exceeded 100 ng/mL, and the ligand levels were consequently reduced to approximately 10–25% of those in controls.

Prophylactic treatment initiated at the predisease stage with injections of the VR-1255-IFN-γR/Fc plasmid without local electroporation effectively protected mice from early lethality and reduced serologic and histologic disease markers. These benefits were observed despite low serum levels of the fusion protein, which may, to some extent, be due to the sequestration of the receptor in IFN-γ–producing lymphoid organs and inflammatory sites. This possibility is supported by the fact that decreases in IFN-γ serum levels were still substantial. The degree of prophylaxis afforded by this regimen is not unlike the protection we reported previously in MRL-Fas_lpr_ mice heterozygous for the IFN-γ gene deletion (10) and that reported by others (32) in a subline of long-lived MRL-Fas_lpr_ mice in which IFN-γ levels were one-third that of conventional MRL-Fas_lpr_ mice. In the heterozygous IFN-γ gene–deleted mouse, however, despite reduced GN and prolonged survival, autoantibody levels and IgG kidney deposits were unaffected (10). It appears, therefore, that the active plasmid treatment is more potent than the drop in serum IFN-γ levels suggests, presumably due to the postulated effects of the sequestered receptor in inflammatory sites. Overall, the data indicate that even a fractional inhibition of IFN-γ is sufficient to reduce disease progression and severity. Nevertheless, injections of the active plasmid combined with electroporation significantly improved protection commensurate with several-fold increases in serum levels of the IFN-γ inhibitory fusion protein.

The positive results with early-life treatment provided the impetus to apply this mode of therapy to mice with established disease. Despite the severity of the underlying illness, survival was extended beyond expectations, with 100% of the mice alive at 14 months of age, the latest point of observation. The effectiveness of this treatment in late disease is, to our knowledge, unprecedented, and it is remarkable that inhibition of a single molecule has such a profound effect in the course of this multifactorial disease. Only a few other attempts have been made to modify late-stage lupus disease in animal models, and only those using reagents that block costimulation (CTLA-4Ig, anti-CD40L, or combination) have shown any success (6, 33, 34). These treatments, however, may severely compromise immune responsiveness in general, and early trials in humans with anti-CD40L had to be discontinued due to complications from thromboembolic episodes (35).

Regardless of whether the VR-1255 IFN-γR/Fc plasmid treatment was initiated early or late, there was, as expected, a drop in serum levels of polyclonal IgG2a (Th1 dependent) and a concomitant increase in the levels of polyclonal IgG1 (Th2 dependent). Similarly, there was a significant decline in the dominant IgG2a anti-chromatin subclass, even when treatment was initiated late. Unlike the situation with polyclonal IgG subclasses, however, there was no compensatory switch of anti-chromatin from IgG2a to IgG1. These results are, again, compatible with our previous observations with the IFN-γ gene–deleted MRL-Fas_lpr_ mice (10) and strongly indicate that an autoimmune response against chromatin is highly dependent on the presence of IFN-γ.

Because of the highly pleiotropic properties of IFN-γ (36), it is very difficult to define the exact mechanism(s) by which this cytokine promotes autoimmunity. Nevertheless, based on the elevated MHC class I and II expression on splenic and peritoneal monocytes of IFN-γ–hyperproducing MRL-Fas_lpr_ mice (10, 37, 38), we postulate that a major effect of IFN-γ blockade would be downregulation of MHC expression, as we documented previously in IFN-γ gene–deleted MRL-Fas_lpr_ mice (10), and consequently reduced autoreactivity. This reduction in MHC expression may encompass not only professional antigen-presenting cells, but also nonprofessional antigen-presenting cells in the afflicted tissues, such as tubular epithelial cells in the kidney (ref. 10, and the present study). These cells have been shown to hyperexpress class II MHC in these mice and to function as antigen-presenting cells (39, 40). The postulated inhibition of autoreactivity was, in fact, reflected in the reduced lymphoid hyperplasia and the lower frequency of both cycling (BrdUhi) and activated (CD44hi) DN cells, presumably because these cells, lacking coreceptors, would require upregulation of MHC and increased Ag presentation to be engaged. In contrast, the frequency of BrdUhi CD8+ cells was unaltered by IFN-γ inhibition. Curiously, however, the frequency of BrdUhi CD4+ cells was increased, an unexplained finding that may be attributed to a regulatory or compensatory mechanism, since a high proportion of these cells was also positive for intracellular IFN-γ.

Additional mechanisms by which kidney disease may be reduced in mice treated with the VR-1255-IFN-γR/Fc plasmid are decreased expression of inflammatory response-promoting molecules, such as ICAM-1 and MCP-1, as documented herein. ICAM-1, a cell-surface protein that regulates immune cell interactions, and MCP-1, a macrophage-attracting chemokine, have been shown previously to be highly expressed in diseased kidneys of MRL-Fas_lpr_ mice (41, 42). In addition, congenic MRL-Fas_lpr_ mice with deleted MCP-1 (42) or ICAM-1 (43) genes showed reduced GN.

Gene therapy for autoimmune diseases continues to receive considerable attention (44, 45). The delivery of IFN-γ inhibitory molecules by intramuscular injection of plasmid vectors is simple and appears to be nontoxic and safe. The use of this approach for gene therapy of autoimmune diseases circumvents several problems encountered with viral vectors (46) because the plasmid will not reactivate to a pathogenic state, is unlikely to be incorporated in genomic DNA or to be neutralized by the host’s immune response, and does not stimulate a local inflammatory response. This approach may also be superior to recombinant soluble molecules in that it provides a depot of genetic material for long-term expression of the active biomolecule. In addition, desirable effects on affected microenvironments may be more potent, since recent studies (47) have shown that naked DNA is transported from the injection site to not only the regional lymph nodes, but also to distant organs such as the spleen and very likely to the afflicted tissues. Finally, since inhibition of IFN-γ has been shown to be of benefit in other autoimmune diseases, such as myasthenia gravis (48, 49) and insulin-dependent diabetes mellitus (50), the mode of therapy outlined here may have broader application.