Deficiency of the type I interferon receptor protects mice from experimental lupus


  • The Malcolm Randall VA Medical Center (Gainesville, FL) provided resources and the use of their facilities.



Systemic lupus erythematosus (SLE) is diagnosed according to a spectrum of clinical manifestations and autoantibodies associated with abnormal expression of type I interferon (IFN-I)–stimulated genes (ISGs). The role of IFN-I in the pathogenesis of SLE remains uncertain, partly due to the lack of suitable animal models. The objective of this study was to examine the role of IFN-I signaling in the pathogenesis of murine lupus induced by 2,6,10,14-tetramethylpentadecane (TMPD).


IFN-I receptor–deficient (IFNAR−/−) 129Sv mice and wild-type (WT) 129Sv control mice were treated intraperitoneally with TMPD. The expression of ISGs was measured by real-time polymerase chain reaction. Autoantibody production was evaluated by immunofluorescence and enzyme-linked immunosorbent assay. Proteinuria and renal glomerular cellularity were measured and renal immune complexes were examined by immunofluorescence.


Increased ISG expression was observed in the peripheral blood of TMPD-treated WT mice, but not in the peripheral blood of TMPD-treated IFNAR−/− mice. TMPD did not induce lupus-specific autoantibodies (anti-RNP, anti-Sm, anti–double-stranded DNA) in IFNAR−/− mice, whereas 129Sv controls developed these specificities. Although glomerular immune complexes were present in IFNAR−/− mice, proteinuria and glomerular hypercellularity did not develop, whereas these features of glomerulonephritis were found in the TMPD-treated WT controls. The clinical and serologic manifestations observed in TMPD-treated mice were strongly dependent on IFNAR signaling, which is consistent with the association of increased expression of ISGs with lupus-specific autoantibodies and nephritis in humans.


Similar to its proposed role in human SLE, signaling via the IFNAR is central to the pathogenesis of autoantibodies and glomerulonephritis in TMPD-induced lupus. This lupus model is the first animal model shown to recapitulate the “interferon signature” in peripheral blood.

Systemic lupus erythematosus (SLE) is a multiorgan autoimmune disease with protean clinical manifestations, accompanied by a highly characteristic subset of antinuclear autoantibodies (ANAs) reactive with small ribonucleoproteins, such as the U1 small nuclear RNP (snRNP) and double-stranded DNA (dsDNA) (1). Accumulating evidence suggests that dysregulated production of type I interferons (IFN-I), reflected in the overexpression of IFN-I–stimulated genes (ISGs) in peripheral blood mononuclear cells (PBMCs), is associated with SLE (2, 3). This pattern, known as the “interferon signature,” is clinically important, since a correlation has been found with disease activity, renal involvement, and the presence of autoantibodies against components of the U1 snRNP (dsDNA and the Sm/RNP antigens) (4, 5).

Initially recognized for their role in antiviral responses, the IFN-I modulate immunity and autoimmunity by promoting dendritic cell (DC) maturation, T cell survival, and antibody production (6–9). The onset of SLE and other autoimmune disorders has been reported in patients undergoing treatment with recombinant IFNα for hepatitis C infection or neoplastic diseases (10–12), suggesting that IFN-I may have a causal role in the pathogenesis of autoimmune disease. In addition, patients with a chromosomal translocation resulting in trisomy of the short arm of chromosome 9 (which bears the IFN-I gene cluster) have high levels of IFN-I production and develop lupus-like disease (13). However, there is, at present, no direct proof that dysregulation of IFN-I has a causal role in SLE.

Murine lupus induced by 2,6,10,14-tetramethylpentadecane (TMPD) is associated with high levels of IFN-I production (14), as well as with key immunologic and clinical features of SLE, such as the production of anti-dsDNA and anti-Sm autoantibodies and the development of glomerulonephritis, arthritis, and pulmonary vasculitis (15–18). In this study, we investigated whether IFN-I signaling plays a causal role in the pathogenesis of lupus-like disease in this model.



Four-week-old 129Sv/Ev IFN-I receptor α-chain–deficient (IFNAR−/−) mice (19) and wild-type (WT) control (129Sv/Ev) breeding pairs (B&K Universal Limited, Grimston, Aldbrough, UK) were housed in a specific pathogen–free facility in barrier cages. The mice received 0.5 ml of TMPD (Sigma-Aldrich, St. Louis, MO), intraperitoneally, at age 8 weeks. Peritoneal cells, granulomas, the spleen, kidneys, and blood were harvested 6–8 months later. These studies were approved by the University of Florida Institutional Animal Care and Use Committee.

Polymerase chain reaction (PCR).

Total RNA from murine peripheral blood was isolated using the PAXgene RNA kit (Qiagen, Valencia, CA). Total RNA from spleen, kidney, and peritoneal cells and ectopic lymphoid tissue (“lipogranulomas”) from TMPD-treated mice were isolated using TRIzol (Invitrogen Life Technologies, Carlsbad, CA). RNA (1 μg) was treated with DNase I (Invitrogen Life Technologies) to remove genomic DNA, and reverse transcribed to complementary DNA using the Superscript First-Strand Synthesis System (Invitrogen Life Technologies) for reverse transcription–PCR. Amplification was carried out in a PTC-100 Programmable Thermal Controller (MJ Research, Waltham, MA) as described previously (14). IFNα consensus and IFNβ primers were used as described previously (20).

Real-time PCR was performed using SYBR Green core reagents (Applied Biosystems, Foster City, CA) with specific primer pairs as described previously (14). The monocyte chemotactic protein 1 (MCP-1) primers used were as follows: AGGTCCCTGTCATGCTTCTG (forward) and GGATCATCTTGCTGGTGAAT (reverse). Transcripts were quantified using the comparative (2math image) method.

Enzyme-linked immunosorbent assay (ELISA).

Antichromatin antibodies were detected in the mouse serum (at a 1:500 dilution) using an ELISA with chicken erythrocyte chromatin, as described previously (21). Antigen-capture ELISAs were performed to detect anti-nRNP/Sm and anti-Su in the mouse serum (at a 1:500 dilution) (21). Levels of anti-dsDNA and anti–single-stranded DNA (anti-ssDNA) antibodies were tested in the sera (at a 1:250 dilution) using S1 nuclease–treated calf thymus DNA as antigen (heat denatured for ssDNA) (21). Total immunoglobulin levels were measured by ELISA as described previously (22).

Cytokine ELISA.

Levels of IFNβ, interleukin-6 (IL-6), tumor necrosis factor α (TNFα), and IL-12 in peritoneal lavage fluid were measured using hamster monoclonal antibodies and rabbit polyclonal antibodies (for TNFα) or rat monoclonal antibody pairs (for IL-12) (BD Biosciences, San Jose, CA). After incubation with biotinylated cytokine-specific antibodies, streptavidin–alkaline phosphatase (1:1,000 dilution; Southern Biotechnology, Birmingham, AL) was added for 30 minutes at 22°C, and the reaction was developed with p-nitrophenyl phosphate substrate in diethanolamine buffer. Results of the cytokine ELISAs were read in a VERSAmax microtiter plate reader (Molecular Devices, Sunnyvale, CA).


Levels of serum ANAs were determined 6 months after TMPD treatment, by indirect immunofluorescence using HEp-2 cells (Immunoconcepts, Sacramento, CA). Sera were screened at a 1:40 dilution and the titers were determined using an Image Titer titration emulation system (Rhigene, Des Plaines, IL).

Immunoprecipitation and Western blot analysis of autoantibodies.

Autoantibodies to cellular proteins in sera (5 μl per sample) were analyzed by immunoprecipitation of 35S-labeled proteins from K562 cell extract, and analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and autoradiography as described previously (23). Autoantibody specificities also were analyzed for reactivity with K562 cell extract by Western blotting (23), using a serum dilution of 1:1,000 followed by 1:1,000 alkaline phosphatase–labeled goat anti-mouse IgG (Southern Biotechnology) for 1 hour. The reaction was developed using Immun-Star AP substrate (Bio-Rad, Hercules, CA).

Flow cytometry.

Lipogranulomas were dissociated using collagenase. A single-cell suspension was stained with anti-B220 and anti-CD4 antibodies (BD Biosciences). B220+ B cells and CD4+ T cells were analyzed by flow cytometry as previously described (14).


Lipogranulomas were excised from the peritoneal wall, fixed in 4% paraformaldehyde, and embedded in paraffin. Immunohistochemistry was carried out at the Molecular Pathology and Immunology Core (University of Florida) using the Autostainer protocol (Dako, Carpinteria, CA). Briefly, 4-μm serial sections were deparaffinized and then blocked with Sniper (Biocare Medical, Walnut Creek, CA). Sections were incubated with rat anti-mouse CD45 receptor (B220; BD Biosciences) or anti-CD3 (Serotec, Raleigh, NC) for 1 hour, followed by incubation with nonbiotinylated rabbit anti-rat immunoglobulin antibodies (Vector, Burlingame, CA) for 30 minutes. Staining was visualized using Mach Gt × Rb horseradish peroxidase polymer (Biocare Medical), the chromagen Cardassian diaminobenzidine (Biocare Medical), and Mayer's hematoxylin counterstain.

Assessment of glomerulonephritis.

Proteinuria was measured using Albustix (Miles Laboratories, Elkhart, IN), after the mice were killed. Glomerular cellularity was evaluated by counting the number of nuclei per glomerular cross-section (20–30 glomerular cross-sections per mouse) after staining with hematoxylin and eosin (H&E) (24). For assessment of renal immune complex deposition, 4-μm frozen sections were stained with fluorescein isothiocyanate–conjugated goat anti-mouse IgG antibodies or rabbit anti-mouse C3 antiserum, and examined by fluorescence microscopy (16). Glomerular staining intensity was quantified using the Image Titer titration emulation system (Rhigene), which approximates intensity by acquiring series of images obtained at different exposure times. These exposures create an image titer that is equivalent to serial dilutions, from which an end point is selected using a standard curve; the results, expressed in arbitrary units, are designated according to the staining intensity based on the standard curve. Renal tissue from 3 WT mice treated with TMPD and 3 IFNAR−/− mice treated with TMPD was stained in the same manner as described above. Glomerular fluorescence intensity was determined in 3 kidney sections per mouse.

Statistical analysis.

All statistical analyses were carried out using a Mann-Whitney 2-tailed U test.


The interferon signature in peripheral blood cells from TMPD-treated mice.

PBMCs from most patients with SLE show increased expression of a group of ISGs (2, 3). This interferon signature has also been found in the peritoneal cells and ectopic lymphoid tissue of mice treated with TMPD; a previous study showed that TMPD-treated mice express MX1, oligoadenylate synthetase, IFN-inducible 10-kd protein (IP10), IFN regulatory factor 7 (IRF7), and other ISGs at levels significantly higher than those found in the peritoneal cells from mice treated with a mineral oil that does not induce lupus (14). As shown in Figure 1, PBMCs from 129Sv mice treated with TMPD exhibited increased expression of the ISGs MX1 (Figure 1A) and IRF7 (Figure 1B) as compared with untreated control mice. In contrast, expression of IP10 (CXCL10) was comparable in both TMPD-treated and untreated 129Sv mice (Figure 1C). Expression of these genes in IFNAR−/− mice was substantially lower (Figures 1A–C). There was little difference in the expression of IFNβ between IFNAR−/− and 129Sv mice, either with or without TMPD treatment (Figure 1D). IFNα also was expressed comparably in the PBMCs from IFNAR−/− mice and 129Sv controls (Figure 1E).

Figure 1.

The “interferon (IFN) signature” in peripheral blood mononuclear cells of 2,6,10,14-tetramethylpentadecane (TMPD)–treated mice. A–D, Wild-type (WT) 129Sv mice and IFN-I receptor α-chain–deficient (IFNAR−/−) 129Sv mice were treated with TMPD (0.5 ml intraperitoneally) or were left untreated (No Rx). Peripheral blood was collected 6–8 months later for RNA isolation. Expression of the IFN-I–inducible genes MX1 (A), IFN regulatory factor 7 (IRF7) (B), and IFN-inducible 10-kd protein (IP10) (C) and expression of IFNβ (D) were measured by real-time polymerase chain reaction (PCR), normalized to β-actin. Horizontal lines indicate the mean relative gene expression in each group. P values were determined by Mann-Whitney U test. E, IFNα and IFNβ gene expression was analyzed by reverse transcription–PCR analysis of peripheral blood cells from TMPD-treated WT and IFNAR−/− mice; 18S RNA expression is shown as a control.

Expression of IFN-regulated genes in other locations.

Expression of MX1 and the IFN-I–regulated chemokines IP10 (CXCL10; which is also IFNγ regulated) and MCP-1 (CCL2) (25) was reduced ∼100-fold in the peritoneal exudate cells of IFNAR−/− mice treated with TMPD as compared with 129Sv control mice (Figure 2A); comparable results (not shown) were obtained for the expression of IRF7.

Figure 2.

Expression of IFN-I–regulated genes in TMPD-treated mice. Gene expression by cells from different peripheral sites in WT 129Sv mice and IFNAR−/− mice was measured by real-time PCR, normalized to β-actin expression. The levels of the IFN-I–inducible genes MX1, monocyte chemotactic protein 1 (MCP-1), and IP10 were determined in A, peritoneal exudate cells, while expression of MX1, MCP-1, and the non–IFN-regulated gene B cell–attracting chemokine 1 (BCA; CXCL13) was determined in B, ectopic lymphoid tissue (lipogranulomas) and C, the spleens of mice. Horizontal lines indicate the mean relative gene expression in each group. P values were determined by Mann-Whitney U test. See Figure 1 for other definitions.

A similar pattern was evident in the ectopic lymphoid tissue (“lipogranulomas”) after TMPD treatment (Figure 2B). The expression of ISGs MX1 and MCP-1 was substantially higher in WT mice than in IFNAR−/− mice, whereas expression of B cell–attracting chemokine 1 (BCA-1; CXCL13), a chemokine that is not IFN-I regulated, was comparable in the ectopic lymphoid tissue of WT mice and IFNAR−/− mice (Figure 2B). Thus, the high expression of ISGs observed in the peripheral blood (Figure 1) was also apparent at the site of the TMPD-induced inflammatory response in WT mice, but not in IFNAR−/− mice.

In addition, the expression of MX1 and MCP-1, but not BCA-1, was decreased ∼100-fold in the spleens of IFNAR−/− mice as compared with WT control mice (Figure 2C). Together with previous evidence indicating that IFNα/β production by peritoneal and lipogranuloma cells is greatly increased by TMPD treatment (14), the present results suggest that the overexpression of a variety of ISGs, including chemokines implicated in the recruitment of inflammatory cells in TMPD-treated mice, requires signaling via the IFNAR. Thus, TMPD-induced lupus recapitulates the interferon signature observed in human SLE.

Ectopic lymphoid tissue in TMPD-treated IFNAR−/− mice.

In WT mice, TMPD causes chronic peritoneal inflammation culminating in the development of ectopic lymphoid tissue, the site of a germinal center–like reaction giving rise to autoantibody-secreting cells (Nacionales D, et al: unpublished observations). Therefore, it was of interest to examine whether peritoneal inflammation and ectopic lymphoid tissue develop in IFNAR−/− mice. Absence of the IFNAR did not diminish the peritoneal inflammatory response substantially at 6 months after TMPD treatment, since the total number of peritoneal inflammatory cells was comparable in TMPD-treated WT mice and IFNAR−/− mice (Figure 3A). Similarly, spleen weight was not significantly different between the WT and IFNAR−/− mice (Figure 3B).

Figure 3.

Inflammatory response to TMPD in IFNAR−/− mice compared with WT 129Sv mice. A, Peritoneal cell counts. Peritoneal lavage was performed at 6–8 months after TMPD treatment in WT or IFNAR−/− mice. Total cells were counted using a hemocytometer. B, Spleen weight. Spleens of WT or IFNAR−/− mice were weighed at 6–8 months after TMPD treatment. Horizontal lines indicate the mean in each group. P values were determined by Mann-Whitney U test. C, Flow cytometric analysis of lipogranuloma cells. T and B cells were stained using anti-CD4 and anti-B220, respectively. Staining results are expressed as the mean and SD percentage of total isolated lipogranuloma cells. D, Hematoxylin and eosin (H&E) staining (top) and immunoperoxidase staining with anti-B220 (middle) and anti-CD3 (bottom) of ectopic lymphoid tissue from WT and IFNAR−/− mice treated 6–8 months earlier with TMPD. Arrows indicate groups of B cells within more diffuse T cell infiltrates. (Original magnification × 100.) NS = not significant (see Figure 1 for other definitions).

The peritoneal cavities of WT and IFNAR−/− mice both contained lipogranulomas (ectopic lymphoid tissue [14]). Analysis of the isolated lipogranuloma cells by flow cytometry confirmed the presence of a similar percentage of B lymphocytes (B220+ B cells) and T lymphocytes (CD4+ T cells) in the ectopic lymphoid tissue from TMPD-treated WT mice and TMPD-treated IFNAR−/− mice (Figure 3C). Staining of paraffin sections with H&E revealed that the lipogranulomas from IFNAR−/− and WT mice had a similar histologic appearance, with numerous oil droplets surrounded by mononuclear cell infiltrates (Figure 3D). Consistent with the findings in BALB/c mice (14), ectopic lymphoid tissue from TMPD-treated 129Sv mice contained B220+ B cells and CD3+ T cells (Figure 3D). Evaluation of the serial sections from these mice revealed groups of B cells within more diffuse T cell infiltrates (Figure 3D, arrows). Ectopic lymphoid tissue from TMPD-treated IFNAR−/− mice displayed a similar immunohistologic pattern (Figure 3D).

Lack of lupus-associated autoantibodies in IFNAR−/− mice.

ANAs, one of the hallmarks of SLE, are produced at high levels in TMPD-treated BALB/c mice. Not surprisingly, sera from TMPD-treated 129Sv mice also were strongly ANA positive (Figure 4A). In addition, ANAs were readily detectable in the sera from TMPD-treated IFNAR−/− mice, but the mean titer was lower than in 129Sv controls (Figure 4A). Untreated mice were ANA negative or had only low titers.

Figure 4.

Autoantibody production in 2,6,10,14-tetramethylpentadecane (TMPD)–treated type I interferon receptor α-chain–deficient (IFNAR−/−) mice compared with wild-type (WT) 129Sv mice. A, Sera from WT or IFNAR−/− mice treated 6–8 months earlier with TMPD, as well as sera from untreated (No Rx) mice, were tested for antinuclear antibodies by immunofluorescence (1:40 dilution). P values were determined by Mann-Whitney U test. B, To examine the immunofluorescence pattern, HEp-2 cells were incubated with sera (1:40 dilution) from representative TMPD-treated WT or IFNAR−/− mice. Arrows indicate staining of mitotic chromosomes in the sera from IFNAR−/− mice. (Original magnification × 200.) C, Immunofluorescence staining was used to analyze the frequency of nuclear (Nuc) or cytoplasmic (Cyt) staining, or both staining patterns (Nuc + Cyt) in WT and IFNAR−/− sera. D, Sera from TMPD-treated or untreated IFNAR−/− mice or 129Sv controls were tested for antichromatin antibodies using an enzyme-linked immunosorbent assay (ELISA) for detection of reactivity with chicken erythrocyte chromatin. E, Serum levels of the lupus autoantibodies IgG anti–nuclear RNP/Sm (anti-nRNP/Sm), anti-Su, anti–double-stranded DNA (anti-dsDNA), and anti–single-stranded DNA (anti-ssDNA) were measured by ELISA in IFNAR−/− mice or 129Sv controls, 6–8 months after treatment with TMPD (15 per group) or without treatment (No Rx) (n = 12 per group). Horizontal lines indicate the mean in each group. P values were determined by Mann-Whitney U test.

All sera from TMPD-treated WT mice exhibited a nuclear staining pattern that was sometimes accompanied by cytoplasmic staining (Figures 4B and C). In contrast, half of the serum samples from TMPD-treated IFNAR−/− mice showed cytoplasmic staining, and the remaining serum samples displayed nuclear staining or both nuclear and cytoplasmic staining (Figures 4B and C). Although some of the sera from both the WT and IFNAR-deficient mice showed staining for mitotic chromosomes, which is consistent with the presence of antichromatin antibodies (Figure 4B, arrows), only the sera from TMPD-treated WT mice were positive for antichromatin antibodies by ELISA (Figure 4D).

In striking contrast to the immunofluorescence results, lupus-associated autoantibodies (IgG anti-dsDNA, anti-ssDNA, anti-nRNP/Sm, and anti-Su) were nearly undetectable by ELISA in the sera of TMPD-treated IFNAR−/− mice but were present at high concentrations in the sera from 129Sv controls (Figure 4E). Longitudinal studies, with followup to 8 months after treatment, indicated that the absence of this subset of autoantibodies in IFNAR−/− mice was not explained by delayed onset of autoantibody production (results not shown). Moreover, these lupus autoantibodies were absent in IFNAR−/− mice with high-titer ANAs (Figure 4E), suggesting that the absence of IFN signaling affected the production of only a subset of ANAs while having little effect on other autoantibody subsets.

The specificities of the ANAs in IFNAR−/− mice could not be determined by Western blotting or immunoprecipitation, whereas anti-Sm/RNP autoantibodies were readily detectable by both techniques in the sera from WT mice (Figures 5A and B). Interestingly, the levels of total serum IgG2a (the primary isotype of anti-nRNP/Sm, anti-dsDNA, and other autoantibodies induced by TMPD) were comparable in TMPD-treated IFNAR−/− mice and WT mice, whereas levels of total IgM were higher in IFNAR−/− mice (Figure 5C), indicating that the effect of IFNAR deficiency on autoantibody production was not mediated at the level of IgG2a production or isotype switching.

Figure 5.

TMPD-induced autoantibodies in IFNAR−/− mice compared with WT mice. A, Western blotting. Total proteins from K562 cell extract were probed with sera from IFNAR−/− or WT mice 6 months after TMPD treatment. Sera from the IFNAR−/− mice were negative, whereas sera from 2 of 4 WT mice exhibited reactivity with the U1-70K (anti-RNP) antigen and sera from 1 of 4 WT mice reacted with the U1-A protein, as determined by comparison with human reference sera and monoclonal antibody 2.73 (anti–U1-70K) (not shown). B, Immunoprecipitation. Sera from TMPD-treated IFNAR−/− or WT mice were tested for reactivity with radiolabeled K562 cell extract. Positions of the U small nuclear RNP proteins U5-200K (indicative of anti-Sm reactivity) and U1-A, B′/B, C, D, E/F, and G are indicated. C, Expression of total IgG2a and IgM. Levels of total IgG2a (the predominant isotype of TMPD-induced autoantibodies) and IgM in sera from TMPD-treated or untreated IFNAR−/− mice or WT controls were measured by enzyme-linked immunosorbent assay. IgM levels were significantly different between the IFNAR−/− and control groups, as determined by Mann-Whitney U test. Horizontal lines indicate the mean in each group. See Figure 4 for definitions.

To further explore how autoantibodies in TMPD-treated mice were generated, we measured the expression of several cytokines implicated in the expansion and differentiation of germinal center B cells, including the cytokines IL-12 (26), IFNβ, IL-6 (9), TNFα (27), and B lymphocyte stimulator (BLyS; trademark of Human Genome Sciences, Rockville, MD)/BAFF (28–30). TMPD-induced lupus is milder (decreased autoantibody production and/or less severe nephritis) in IL-6– or IL-12–deficient mice (21, 31), whereas the effect of a deficiency of TNFα or BLyS/BAFF has not been investigated. Consistent with the known effect of IFN-I on maturation of myeloid DCs (major producers of IL-12), levels of IL-12 in the peritoneal lavage fluid were reduced in TMPD-treated IFNAR−/− mice as compared with control mice. In contrast, there was no difference in the levels of IFNβ, IL-6, or TNFα in the peritoneal lavage fluid from these mice (Figure 6A).

Figure 6.

Cytokine expression in IFNAR−/− mice compared with WT mice. A, Interleukin-12 (IL-12), IFNβ, IL-6, and tumor necrosis factor α (TNFα) were measured by ELISA in the peritoneal lavage fluid of WT and IFNAR−/− mice. B, B lymphocyte stimulator (BLyS)/BAFF mRNA was quantified in lipogranulomas and peritoneal cells from WT and IFNAR−/− mice by real-time polymerase chain reaction, normalized to β-actin expression. Horizontal lines indicate the mean in each group. P values were determined by Mann-Whitney U test. NS = not significant (see Figure 4 for other definitions).

BLyS/BAFF overexpression in mice increases the production of polyclonal IgG, IgA, and IgE, leading to autoantibody production and a lupus-like syndrome (32). Although IFN-I regulates BLyS/BAFF expression (33), the levels in both peritoneal exudate cells and lipogranuloma cells were comparable in IFNAR−/− mice and WT mice (Figure 6B), suggesting that BLyS/BAFF expression is regulated, at least partly, through IFN-I–independent mechanisms, and that its expression is not sufficient to induce production of the lupus autoantibodies anti-nRNP/Sm, anti-Su, and anti-dsDNA.

Abolishment of glomerulonephritis in IFNAR−/− mice.

We next examined the effect of IFNAR deficiency on the induction of lupus nephritis. As shown in Figure 7A, ≥2+ proteinuria was detected 6 months after TMPD treatment in 5 of 12 female 129Sv control mice, but not in any of the TMPD-treated female IFNAR−/− mice. The number of nuclei per glomerular cross-section, a measure of glomerular cellularity, was increased in TMPD-treated 129Sv mice as compared with untreated 129Sv controls (Figure 7B). In contrast, there was no difference in glomerular cellularity between TMPD-treated IFNAR−/− mice and untreated IFNAR−/− controls; moreover, the number of nuclei/glomerulus was comparable with that in untreated WT mice. The number of nuclei/glomerulus was significantly lower in TMPD-treated IFNAR−/− mice than in TMPD-treated 129Sv controls (P = 0.0007 by Mann-Whitney U test).

Figure 7.

Absence of renal disease in IFNAR−/− mice. A, Presence of proteinuria, defined as ≥2+ on dipstick analysis, was identified in TMPD-treated WT 129Sv controls, but not in TMPD-treated IFNAR−/− mice or in untreated (No Rx) mice. B, Glomerular cellularity was determined in WT and IFNAR−/− mice as the number of nuclei per glomerular cross-section; results are representative of 3 experiments. Horizontal lines indicate the mean in each group. P values were determined by Mann-Whitney U test. C, Immune complexes were determined in WT and IFNAR−/− mice by direct immunofluorescence of glomeruli for IgG and complement component C3 (original magnification × 200). D, IgG and C3 immunofluorescence staining intensity was measured by titration emulation (3 mice/group) in IFNAR−/− and WT mice. Bars show the mean and SD. NS = not significant (see Figure 4 for other definitions).

Interestingly, direct immunofluorescence revealed glomerular immune complex deposits, consisting of IgG and C3, in TMPD-treated IFNAR−/− mice and 129Sv control mice (Figure 7C). Quantification of the glomerular staining for IgG and C3 confirmed that immune complex deposition was not significantly decreased in IFNAR−/− mice as compared with that in the control mice (Figure 7D), suggesting that signaling through the IFNAR is not required for immune complex formation and renal deposition. In contrast, IFNAR signaling was necessary for the development of an inflammatory response to renal immune complexes, manifested as glomerular hypercellularity and proteinuria.


Most immunocompetent mouse strains develop ANAs, as well as lupus-specific autoantibodies such as anti-Sm, anti-dsDNA, and anti–ribosomal P, following intraperitoneal injection of TMPD (15, 16, 34). Many strains, such as BALB/c and SJL, develop glomerulonephritis, while some strains, such as BALB/c, develop arthritis and others, such as B6, develop pulmonary vasculitis (17, 18). Thus, TMPD induces an autoimmune disease that meets 4 of the American College of Rheumatology classification criteria for SLE (35). The present study demonstrates that PBMCs from mice with TMPD-induced lupus overexpress ISGs, an abnormality exhibited by most patients with SLE (2, 3), and that signaling via the IFNAR is central to the pathogenesis of lupus in this model.

Strikingly, although IFNAR-deficient mice developed ANAs, they did not develop lupus-specific autoantibodies, such as anti-dsDNA or anti-Sm, following TMPD treatment. This finding strongly suggests that the dysregulation of IFN-I promotes production of autoantibodies against a subset of nucleic acid–protein autoantigens. Similarly, although the IFNAR−/− mice did not develop nephritis (defined by the presence of proteinuria and/or hematuria), glomerular immune complex deposition was apparent. Thus, the results from this model provide new evidence that IFN-I is involved in the pathogenesis of lupus.

PBMCs from most SLE patients express high levels of a multitude of ISGs (2, 3). However, there is, at present, no direct proof that abnormal IFN-I production causes SLE in humans, and also no animal model that recapitulates the interferon signature. The peritoneal inflammatory response to TMPD is associated with increased ISG expression (14). In the present study, PBMCs from TMPD-treated mice, similar to the pattern in lupus patients, exhibited increased expression of ISGs (Figure 1), which was abolished in the absence of IFNAR signaling (Figure 2). Increased ISG expression has not been reported in other lupus models, with the exception of increased levels of Ifi-202 in NZB/NZW mice, attributable to a polymorphism of the promoter region (36). To our knowledge, TMPD-induced lupus is the first animal model of lupus that has been shown to display the interferon signature.

There is other evidence implicating IFN-I in the pathogenesis of autoimmune disease. In NZB mice, although IFN-I expression is not elevated spontaneously, autoimmune hemolytic anemia and glomerular immune deposits are ameliorated in IFNAR-deficient animals (37). However, functional improvement in renal disease could not be assessed, because NZB mice do not generally develop proteinuria or hematuria, nor do they develop arthritis, serositis, skin rashes, or central nervous system disease (37–39). The frequency of ANA positivity in NZB mice is lower than 10%, and this strain does not generally develop lupus-specific autoantibodies such as anti-Sm and anti-dsDNA (38). In contrast to NZB mice, which is perhaps best regarded as a model for IFN-mediated autoimmune hemolytic anemia (39), (NZB × NZW)F1 mice develop a florid lupus-like disease with glomerulonephritis and positive findings of ANAs and anti-dsDNA antibodies (39), which can be exacerbated by exogenous IFNα (40). However, it is not clear whether IFNα plays a causal role or merely accelerates preexisting subclinical inflammatory disease.

In MRL-lpr/lpr mice, IFN-I ameliorates lupus (41). Male BXSB mice have a duplication of the TLR7 gene, which enhances responsiveness to RNA ligands capable of stimulating IFN-I production (42); however, there is no information on ISG expression in the PBMCs of these mice. Thus, to date, none of the mouse models of spontaneous lupus have been unequivocally shown to exhibit the interferon signature, nor is there direct proof that increased IFN-I production is required for the development of lupus.

Patients with SLE who produce anti-dsDNA or anti-RNP (Sm, RNP, Ro, or La) autoantibodies have higher levels of ISGs than do patients without this autoantibody profile, suggesting that IFN-I is linked to the pathogenesis of these autoantibodies (5). Although it has been suggested that anti-RNP antibodies in human SLE may drive IFN-I production (43), the induction of anti-dsDNA, anti-Sm, anti-RNP, and antichromatin autoantibodies by TMPD was abrogated in IFNAR−/− mice, strongly suggesting that IFN-I signaling drives the production of anti-RNP autoantibodies (and not vice versa). Although the relationship between anti-RNP antibodies and IFN-I might be intrinsically different in human lupus as compared with the murine disease, this is unlikely, since studies have shown the production of anti-RNP autoantibodies in patients in whom a chromosomal translocation causes overproduction of IFN-I (13) as well as in individuals who have received exogenous IFNα (12).

Interestingly, TMPD-treated IFNAR−/− mice still produced ANAs, although at a lower mean titer than that in WT mice (Figure 4A). However, these were not the typical “lupus” autoantibodies (Figures 4 and 5). In many cases, the immunofluorescence pattern displayed by sera from TMPD-treated IFNAR−/− mice suggested reactivity with cytoplasmic filaments (Figure 4B). In other cases, there was both nuclear and chromosomal staining, suggestive of reactivity with chromatin antisense antibodies (Figure 4B), but the reactivity with chromatin, ssDNA, or dsDNA could not be verified (Figures 4D and E). Thus, the autoantibody specificities in IFNAR−/− mice remain to be determined. There may be parallels between the generation of ANAs in IFNAR−/− mice and ANA positivity in healthy individuals who do not have manifestations of SLE. This subset does not exhibit the interferon signature (5), further suggesting that IFN-I has a unique role in the pathogenesis of lupus-specific autoantibodies in humans as well as mice.

In humans, the interferon signature is associated with lupus nephritis and increased disease severity (2, 5). In the TMPD-induced lupus model, development of nephritis was also strongly dependent on signaling through the IFN-I receptor (Figure 7). Interestingly, although there was little difference in the immune complex deposition between the IFNAR-deficient and control mice (Figures 7C and D), IFNAR−/− mice failed to develop proteinuria or glomerular hypercellularity (Figures 7A and B). The uncoupling of glomerular immune complex deposition from functionally significant nephritis (proteinuria and/or hematuria) is reminiscent of the pattern in NZB/NZW mice, in which the common γ-chain of Fcγ receptor type I (FcγRI)/FcγRIII is lacking (44), and is also similar to that in TMPD-treated IL-12–deficient mice, which develop anti-dsDNA antibodies and renal immune complexes but not nephritis (31). At present, the mechanism of protection is not known.

Renal expression of the IFN-inducible chemokine MCP-1 (25) was lower in IFNAR−/− mice (results not shown), suggesting that recruitment of CCR2+ inflammatory cells to the glomeruli may be decreased. This would be consistent with the decreased glomerular cellularity seen in the IFNAR−/− mice (Figure 7), the importance of monocyte/macrophages in the pathogenesis of lupus nephritis (45), and the increased urinary MCP-1 levels in patients with active lupus nephritis (46). Alternatively, since FcγRI is IFNα inducible (47), there could be a preponderance of renal FcγRIIb (antiinflammatory) expression over FcγRI (proinflammatory) expression in IFNAR−/− mice. Furthermore, by promoting the generation of autoantibodies against nucleic acid–protein autoantigens, IFNAR signaling might enhance the formation of immunostimulatory immune complexes that become trapped in the glomeruli, causing inflammation (48). Further studies will be necessary to distinguish these possibilities and confirm their importance.

The precise mechanism by which IFN-I promotes autoantibody production remains to be elucidated. The results of our study suggest that abnormal IFN-I signaling is an early event in lupus pathogenesis. The interferon signature appeared in the murine PBMCs 2 weeks after TMPD treatment (Figure 1), long before the onset of autoantibody production (at 3–4 months) or renal disease (at 4–6 months) (16). IFN-I could promote autoimmunity through its effects on downstream cytokines such as IL-12 and IFNγ (31). Although IFNγ−/− mice have milder disease, we have not found increased levels of either IFNγ protein or IFNγ messenger RNA (mRNA) following TMPD treatment (14). In contrast, levels of IL-12 mRNA and protein are elevated. The stimulation of IL-12 production by TMPD was abrogated in IFNAR−/− mice (Figure 6), suggesting that IL-12 plays a role in SLE downstream of IFNAR signaling. IFN-I acts synergistically with the NF-κB pathway and is required for the production of bioactive IL-12p70 following the engagement of Toll-like receptors (TLRs) on myeloid DCs (49). The decreased production of IL-12 in TMPD-treated IFNAR−/− mice may reflect reduced maturation of myeloid DCs, which is promoted by TLR ligation as well as IFN-I (50). This could, in turn, greatly diminish T cell–dependent antibody (or autoantibody) responses (7).

Furthermore, IFN-I could also promote the differentiation of autoreactive B cells into plasma cells (9). However, a subset of B cells can differentiate into autoantibody-secreting cells in the absence of IFNAR signaling, since ANAs reactive with targets other than the classic lupus autoantigens were produced nearly as efficiently in IFNAR−/− mice as in 129Sv controls (Figure 4A). Unexpectedly, BLyS/BAFF, an IFN-inducible cytokine implicated in the pathogenesis of autoantibodies and SLE (32, 33) and in the development of plasma cells (28), was expressed at comparable levels in WT and IFNAR−/− mice (Figure 6). Although it is unlikely that a lack of BLyS/BAFF expression explains the absence of anti-DNA, anti-RNP, anti-Sm, and anti-Su autoantibodies in the IFNAR−/− mice, this pathway could be involved in the production of ANAs in IFNAR−/− mice (Figure 4A).

In summary, TMPD-induced lupus is a model of interferon signature–associated SLE. IFN-I signaling is critical for the development of proliferative nephritis and lupus autoantibodies in this model. As the first animal model in which high IFN-I production appears prior to the onset of disease, TMPD-induced lupus will be a useful tool for investigating both the origins of the interferon signature and its relationship to disease pathogenesis.


We thank Mr. Dustin S. Vale-Cruz (Department of Anatomy and Cell Biology, University of Florida) for assisting with the fluorescence microscopy.


Dr. Reeves had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Nacionales, Reeves.

Acquisition of data. Nacionales, Kelly-Scumpia, Lee, Weinstein, Lyons.

Analysis and interpretation of data. Nacionales, Kelly-Scumpia, Lee, Sobel, Satoh, Reeves.

Manuscript preparation. Nacionales, Kelly-Scumpia, Lee, Reeves.

Statistical analysis. Nacionales, Kelly-Scumpia.