We have examined the ex vivo costimulatory phenotype of DCs from NZM2410 and NZB-W/F1 lupus-prone mice, compared with DCs from age-matched non-autoimmune mice. We found that in lupus-prone mice, after the onset of the disease, DCs expressed an abnormal state of activation with decreased levels of CD80 and CD54, normal levels of CD86, and a specific increase in cells expressing CD40. Importantly, the CD40 increase was already present months before the onset of the disease. The overexpression of CD40 may be important in SLE development because CD40 triggering could be responsible for the inappropriate stimulation of DCs to live longer, produce an excess of pro-inflammatory cytokines, and deliver abnormal activation signals to autoreactive T and B cells.
In the pre-disease stage, the increase in CD40 positivity was mostly present in myeloid DCs and, to a smaller extent, in CD8α+
DCs. Both DC subsets have been proposed to induce peripheral tolerance [25
] unless they receive an activation signal, such as CD40 triggering, that induces the production by DCs of large amounts of pro-inflammatory cytokines and stimulates T cells [26
]. Our results therefore suggest that in young lupus-prone mice, months before the appearance of any autoantibodies, DCs are prone to escape from a tolerogenic status.
The small pDC subset has been proposed to be important in lupus pathogenesis [27
] because pDCs can produce large amounts of type I IFN, which is considered altered in this disease [5
]. CD40 triggering was one of the first stimuli shown to induce the production of type I IFN by pDCs [28
]. Although the increase in CD40 expression that we found in pDCs, after the onset of lupus, did not reach statistical significance with non-parametric tests, we cannot exclude CD40-CD40L interactions as a mechanism by which pDCs are induced to overproduce type I IFN and sustain autoimmunity.
We have found that resting lupus BM-DCs, differentiated in the artificial environment of culture in vitro, did not express increased levels of CD40. Furthermore, splenic DCs from mice with lupus showed a normal upregulation of CD40 expression on activation in vitro. These results suggest that lupus DCs do not have a primary and constitutive alteration in their regulation of CD40, but they may be induced to overexpress CD40 by a chronic stimulus.
The decreased CD80/CD86 ratio in DCs from NZM2410 mice after the onset of the disease resembles the defective costimulatory profile found in DCs from patients with SLE [2
] and validates the use of NZM2410 and NZB-W/F1 strains as murine models for the study of human SLE. CD80 and CD86 are ligands of CD28/CTLA4 and may have different functions [29
]: although both CD80 and CD86 activate effector T cells, CD80 seems to be especially important for the stimulation of T regulatory cells and, therefore, the inhibition of the immune response. Indeed, DNA immunization with plasmids encoding CD80 induces weaker responses than with plasmids encoding CD86 [30
]. In diabetic mice and in lupus-prone MRL/lpr
mice, the blockade of CD80 worsens the severity of both diseases, whereas blockade of CD86 prevents diabetes and has mild effects on lupus [31
]. These data suggest the fascinating hypothesis that the unbalanced expression of CD80 and CD86 in lupus DCs may impair the capacity of DCs to engage regulatory T cells, which would lead to the inappropriate stimulation of autoreactive B cells and the maintenance of the autoimmune disorder.
Recently, Zhu and colleagues [14
] reported that DCs from NZM2410 single-locus derivative B6.Sle3
mice are hyperactivated and hyperstimulate T cells. Some of our results agree with their data, but some differ. Indeed, they found, as we did, that DCs from the spleens of mice 9 to 12 months old hyperstimulate T cells (data not shown). However, they also demonstrated increased expression of CD40, CD54, CD80, and CD86, with a normal CD80/CD86 ratio, whereas we found an increased expression of CD40 with decreased levels of CD80 and CD54. Although we cannot exclude technical reasons to explain the dissimilarities, we think that genetic and disease differences between the two strains of mice can account for the discrepancies. Indeed, NZM2410 mice develop an overt form of lupus, with the full repertoire of autoimmune features and severe glomerulonephritis, whereas B6.Sle3
mice display some autoimmune features but do not develop glomerulonephritis. We therefore propose that NZM2410 mice and their single-locus derivative B6.Sle3
mice have a genetic abnormality, still to be determined and not intrinsic to DCs, that induces upregulation of CD40 in DCs in vivo
before and after the onset of the disease. In NZM2410 mice, the involvement of other immune abnormalities would lead to a decreased CD80/CD86 ratio on DCs, and ultimately to severe overt disease, hypothetically because of impaired activation of regulatory T cells.
CD40-CD40L interaction has been found to be involved in the pathogenesis of lupus in four murine models [33
], and ectopic expression of CD40L on B cells induces a lupus-like autoimmune disease [38
]. In patients with SLE, T cells overexpress CD40L for extended periods [39
], and activated B cells express CD40L ectopically [41
]. Analysis of the literature suggests that the overexpression of CD40 and CD40L may be independent phenomena and may both be required for full lupus development.
Long-term administration of anti-CD40L antibody prevents the production of high-titer autoantibodies and disease onset in mice [34
]. In (SWRxNZB)F1 mice, it also prevented DC accumulation, suggesting that it may be CD40-CD40L dependent [11
]. Our finding of the increase in CD40-expressing DCs in lupus-prone mice before the onset of the disease suggests that CD40 on DCs may have a role in the pathogenesis of this autoimmune disease. We therefore propose that the marked effects of the blockade of CD40L in patients and murine model of lupus are also due to the inhibition of CD40 triggering on DCs. Because anti-CD40L blockade has raised safety concerns in patients with SLE [43
], we propose that the inhibition of the expression and function of CD40 in DCs may be an alternative therapeutic strategy.