|Home | About | Journals | Submit | Contact Us | Français|
Male NZW/BXSB mice develop anti-phospholipid syndrome (APL) and proliferative glomerulonephritis that is markedly accelerated by the Yaa locus encoding an extra copy of TLR7. Female NZW/BXSB mice with only one active copy of TLR7 develop late onset glomerulonephritis but not APL. Since a major function of TLR7 is to induce Type I interferons, our goal was to determine whether interferon alpha (IFNα) can induce or accelerate SLE manifestations in female NZW/BXSB.
8 week old female NZW/BXSB F1 mice were injected with a single dose of Adenovirus expressing IFNα. Mice were monitored for thrombocytopenia and proteinuria. Sera were tested for anti-cardiolipin and anti-Sm RNP antibodies. Mice were sacrificed at 17 or 22 weeks of age and kidneys and hearts were examined histologically and by immunohistochemistry. Spleen cells were phenotyped and ELISpot assays for autoantibody producing B cells were performed.
IFNα markedly accelerated nephritis and death in female NZW/BXSB F1 mice. A significant increase in spleen cell numbers associated with a striking increase in activated B and T cells were seen. Marginal zone B cells were retained. IFNα increased titers of autoantibodies but thrombocytopenia was not found and cardiac damage was milder than in males.
IFNα accelerates the development of renal inflammatory disease in the female NZW/BXSB mice but induces only mild anti-phospholipid syndrome and does not induce thrombocytopenia. The effect of IFNα on SLE disease manifestations is strain dependent. These findings are relevant to our understanding of the physiologic significance of the interferon signature.
The pathogenic role of IFNα in SLE has been inferred from findings that IFNα can induce lupus-like symptoms, and from the discovery that peripheral blood mononuclear cells from active lupus patients show dysregulated expression of a group of IFNα-induced genes (1). In NZB/W mice administration of exogenous IFNα accelerates SLE-like disease that is similar in its characteristics to that observed in the spontaneous model (2).
IFNα is induced by the ligation of Toll-like receptors that are expressed on B cells and plasmacytoid dendritic cells and are specific for nucleic acid antigens (3). TLR7 overexpression induces SLE in mice and its depletion modulates some of the manifestations of SLE (4). In human SLE, the IFN signature has been associated with active and severe disease (5) and with antibodies to RNA associated antigens (6).
Male NZW/BXSB mice carry two active copies of the TLR7 gene (7) and have an accelerated form of SLE characterized by inflammatory nephritis and anti-phospholipid syndrome (8, 9). Female mice with a single active copy of TLR7 develop nephritis late in life but not the anti-phospholipid syndrome (8). To determine whether exogenous IFNα is sufficient to confer the disease accelerating effects of TLR7 reduplication, we administered an adenovirus expressing IFNα to female NZW/BXSB mice. Although the mice developed accelerated nephritis, IFNα was not sufficient to induce full-blown anti-phospholipid syndrome.
NZW/BXSB mice (Jackson Laboratory, Bar Harbor, ME) were bred in our facility. 35 female F1 mice were treated at 8w of age with a single i.v. injection of 109 particles of IFNα adenovirus (Ad-IFNα, Qbiogene Morgan Irvine, CA). 30 controls received the same dose of β-galactosidase-expressing adenovirus (Ad-LacZ), or no treatment. 15 mice received the TLR7 agonist Imiquimod 25ug i.p. three times weekly for 6 weeks. Mice were tested for proteinuria every two weeks (Multistick, Fisher, Pittsburg, PA) and bled periodically for serologic analysis. Platelets were counted at 8, 17 and 22w of age using a Coulter counter (Beckman Coulter, Fullerton CA). Groups of 6-8 mice were sacrificed at 17 or 22w of age, and the remaining mice observed for proteinuria onset and survival. These experiments were carried out according to protocols approved by the Institutional Animal Care and Use Committees of Columbia University and the Feinstein Institute.
ELISA plates (Falcon Labware, Lincoln Park, NJ) were coated with unlabeled goat anti-mouse IgM, IgG1, IgG2a, IgG2b or IgG3 (Southern Biotechnology, Birmingham, AL) overnight at 4°C. After blocking, the plates were incubated with dilutions of serum for 1hr at 37°C followed by HRP-conjugated goat anti-mouse isotype-specific antibodies and substrate solution (KPL, Gaithersburg, MD). Standard curves were established using serial dilutions of purified antibody of the appropriate isotype (Sigma-Aldrich, St. Louis, MA).
Sm/RNP (Arotec Diagnostics Limited, Wellington, NZ) was coated onto Falcon plates at 1μg/ml in PBS. ELISA was performed according to manufacturers' instructions. Anti-cardiolipin titers were measured as previously described (9). A high titer serum was run in serial dilution on each plate as a quantitation control.
ELISpot assays for total immunoglobulin-secreting cells and for anti-cardiolipin–secreting B cells were performed on isolated spleen and bone marrow cells as previously described (9).
Spleens were analyzed for B and T cell markers as previously described (10) using antibodies to CD4, CD8 (Caltag, Burlingham, CA) and CD19. Splenic dendritic cells were identified using PE anti-CD11b and FITC anti-CD11c. B cell subsets were identified as follows: follicular (CD19+/IgM+/IgD+); marginal zone (CD19+/CD21hi/CD23lo); class-switched (CD19+/IgMlo/IgDlo); immature (CD19+/IgMhi/IgDlo). CD4+ T cells were classified as naive (CD62Lhi/CD44lo) or memory (CD62Llo/CD44hi). Activated T and B cells were defined as CD69hi. Unless otherwise stated all antibodies were obtained from BD Pharmingen, San Diego, CA.
Scoring of H&E sections of the kidneys and hearts was performed using a 1 to 4 scale for glomerular damage and interstitial inflammation as previously described (9). Histologic analyses were performed by observers blinded to the treatment group of the mice.
Cryosections of kidney were stained with FITC-conjugated anti-mouse IgG2a, PE-conjugated anti-mouse IgD (Southern Biotech, Birmingham, Alabama), F4/80 (Invitrogen, Carlsbad, California), CD4, CD19, or CD11c (BD Pharmingen) in 2%BSA/PBS containing 0.5% anti-mouse CD16/CD32 (BD Pharmingen) for 1hr at room temperature. Slides were counterstained with DAPI (Invitrogen) and images were captured using a digital CCD-camera system connected to a Zeiss microscope (Zeiss, Thornwood NY).
Proteinuria and survival data for the entire group are shown in Figure 1 and were analyzed using Kaplan-Meier curves and log-rank test. Data in bar graphs are shown as mean + 1 SD. Comparisons in Figure 2 and Table 1 were performed using Mann-Whitney test. p values ≤ 0.05 were considered significant.
Administration of a small dose of Ad-IFNα markedly accelerated disease onset in female NZW/BXSB mice and proteinuria occurred 4-12 weeks after virus injection, followed rapidly by death (IFNα vs. controls p<0.0001 for proteinuria and death - Figure 1A). Serum IFNα levels were below detectable levels in treated mice. However two weeks after virus injection there was a significant increase in the expression of several IFN inducible genes in the spleens of treated mice to levels comparable with those of proteinuric male mice (data not shown). Imiquimod had no effect on autoantibody production, proteinuria onset or mortality and these mice were not studied further.
The kidneys of IFNα treated mice contained intense IgG deposits. Infiltrating F4/80 positive macrophages were located in the interstitium and around glomeruli, whereas CD11c positive dendritic cells were within the glomerular tuft (Figure 1B, 1D). This pattern is similar to that observed in male NZW/BXSB mice (11). In contrast, few glomerular or interstitial infitrates were present in 22 week control mice (Figure 1C, 1E). 17-22w IFNα treated mice had significant renal damage compared with untreated controls (Figure 1F - p<0.001 for glomerular score and p<0.01 for interstitial score). Mild cardiac damage was also observed in the IFNα treated mice but not in the untreated controls (Figure 1G). In contrast, thrombocytopenia was not found (Figure 1H), despite the presence of high titers of anti-phospholipid autoantibodies in the serum.
Serum IgG2a levels increased significantly in the IFNα treated mice compared with controls (1036.2 +/- 452.9 vs. 433.4 +/- 258.9 ug/ml; p < 0.0072 at 12w). The difference was no longer significant at 17w. In contrast, serum IgG1 levels were significantly lower in the IFNα group than in the controls at both 12 and 17w (59.8 +/- 31.5 vs. 142.3 +/-71.5 ug/ml; p< 0.001 and 116.1 +/- 91.0 vs. 209.6 +/- 95.5ug/ml; p < 0.02 respectively). Autoantibodies to both cardiolipin (p < 0.004 at 14w and p < 0.03 at 18-20w) and Sm/RNP (p < 0.02 at 12w, p < 0.04 at 16w) arose earlier in the IFNα treated mice than in age matched untreated controls; the maximal autoantibody titers reached were comparable to male mice (Figure 2A, 2B). Low titer anti-dsDNA antibodies arose late in both treated and untreated mice (not shown). In accordance with these data, ELISpot analysis of spleens revealed an increase in both the frequency (p < 0.04 at 17w and p < 0.0001 at 22w) and total number of IgG (p < 0.003 at 17w and p < 0.0001 at 22w) and IgG anti-cardiolipin antibody producing B cells (p < 0.0001 at 22w) in IFNα treated mice compared with controls (Figure 2C, 2D). Anti-cardiolipin producing B cells also increased in the bone marrows (3.93 +/-.99 per 105 cells in IFNα treated vs. 1.37 +/- 0.87 in controls at 17w; p < 0.02; 9.93 +/-4.48 in IFNα treated vs.3.52 +/- 2.92 in controls at 22w; p < 0.03)
Phenotypic analysis of spleen cells revealed a marked increase in spleen cell number in the treated mice compared with controls, with a significant increase in CD11b positive cells and activated CD4 T cells and B cells (Table 1). Circulating monocytes in the blood were also twofold increased (not shown). All B cell subsets were expanded in the treated mice. This is in contrast to male NZW/BXSB mice in which the overexpression of TLR7 results in loss of the marginal zone subset and a marked increase in the follicular to marginal zone B cell ratio (Kahn and Davidson, unpublished). There was also a significant increase in the number of class switched B cells and plasma cells in the spleens of IFNα treated mice (Table 1). Germinal centers appeared earlier in the spleens of IFNα treated mice compared with untreated controls, but the spleens became disorganized with loss of follicular architecture and germinal centers as the mice aged, similar to males (Figure 2E-J).
The association of the IFN signature with SLE in humans and the ability to induce accelerated SLE in some murine SLE models with IFNα has led to the hypothesis that IFNα is an important pathogenic cytokine in SLE and a target for therapy (3). IFNα accelerates SLE in female NZB/W mice in which it induces early development of anti-dsDNA antibodies and nephritis (2). We have shown in NZB/W mice that IFNα rapidly induces germinal centers that generate short-lived plasma cells producing pathogenic IgG2a autoantibodies in a T cell-dependent fashion, whereas pathogenic IgG3 autoantibodies are generated in a T cell-independent fashion (Liu, submitted). In other mouse models however, Type I interferons have protective effects with respect to SLE initiation, indicating strain heterogeneity (12).
NZW/BXSB mice develop autoantibodies to cardiolipin and to RNA-associated antigens. Together with immune mediated thrombocytopenia, inflammatory glomerulonephritis and a thrombotic vasculopathy that affects the small coronary arteries leading to myocardial infarcts, myocardial fibrosis and a dilated cardiomyopathy (8, 9, 13). Male mice have markedly accelerated disease because they carry the Yaa gene, a reduplication of a portion of the X chromosome that includes the TLR7 gene (7). TLR7 recognizes ssRNA, is required for the production of autoantibodies to RNA associated autoantigens (14), and appears to be responsible for much of the phenotype associated with the Yaa locus (4).
Ligation of TLR7 on plasmacytoid dendritic cells is a potent stimulus for secretion of IFNα as well as other cytokines including TNFα, IL-12 and IL-6 (14). We therefore wished to know whether the accelerating effects of TLR7 overexpression could be mimicked by administration of IFNα. We first showed that the TLR7 agonist Imiquimod did not accelerate disease in female NZW/BXSB mice, consistent with the finding that optimal TLR7 agonism requires the presence of IFNα (15). In contrast, IFNα accelerated the onset of anti-cardiolipin and anti-Sm/RNP autoantibodies and caused early mortality due to nephritis. However, the IFNα treated female mice developed only mild cardiac disease and they did not become thrombocytopenic. Like male NZW/BXSB mice, the IFNα treated females had marked splenomegaly, an increase in activated B and T cells and an increase in plasma cells; however they did not lose their marginal zone B cells. Thus the loss of marginal zone B cells in male NZW/BXSB mice and in female TLR7-overexpressing mice (4) is not mediated through IFNα and may be due to a TLR7 mediated intrinsic defect in B cell selection. The differences in B cell selection between male and female mice may be one reason for the failure to develop anti-platelet antibodies and the milder anti-phospholipid syndrome in the treated females. Alternatively, TLR7 ligation induces many cytokines in addition to IFNα that may cooperatively contribute to disease phenotype and severity. Finally it is possible that other genes in the Yaa locus could contribute to a more severe disease phenotype in males.
The accelerating effects of IFNα are clearly different between NZB/W and NZW/BXSB mice. In both strains IFNα recapitulates the autoantibody profile of the spontaneous disease (2) with early onset of B and T cell activation. In NZB/W mice IFNα induces abundant germinal centers, large numbers of short-lived plasma cells in the spleens with lack of expansion of these cells in the bone marrow, and high serum levels of IgG2a and IgG3 that deposit in the kidneys (Liu, submitted). In NZW/BXSB mice it accelerates the development of large germinal centers followed by splenic disorganization, similar to what is seen in males, an increase in anti-cardiolipin producing B cells in the bone marrow and an increase only in serum levels of IgG2a.
These data show in sum that the effects of IFNα do not fully recapitulate the B cell or disease phenotype associated with TLR7 overexpression and that its effects on disease phenotype are strain dependent. Further characterization of effects of excess IFNα in different lupus-prone mice models and on responses of these mice to therapy will increase our understanding of the physiologic significance of the interferon signature in humans.
This work was supported by funds from the New York SLE Foundation (MR and PK) and from NIAMS (MR) and NIAID (AD).