Systemic lupus erythematosus is an autoimmune disease characterized by increased type I IFNs, autoantibodies, and inflammatory-mediated multiorgan damage. TLR7 activation is an important contributor to systemic lupus erythematosus pathogenesis, but the mechanisms by which type I IFNs participate in TLR7-driven pathologic conditions remain uncertain. In this study, we examined the requirement for type I IFNs in TLR7-stimulated lupus nephritis. Lupus-prone NZM2328, INZM (which lack a functional type I IFN receptor), and NZM2328 IL-1β−/− mice were treated at 10 wk of age on the right ear with R848 (TLR7 agonist) or control (DMSO). Autoantibody production and proteinuria were assessed throughout treatment. Multiorgan inflammation was assessed at the time of decline in health. Renal infiltrates and mRNA expression were also examined after 14 d of treatment. Both NZM2328 and INZM mice exhibited a decline in survival after 3–4 wk of R848 but not vehicle treatment. Development of splenomegaly and liver inflammation were dependent on type I IFN. Interestingly, autoantibody production, early renal infiltration of dendritic cells, upregulation of IL-1β, and lupus nephritis occurred independent of type I IFN signaling. Development of TLR7-driven lupus nephritis was not abolished by the deletion of IL-1β. Thus, although IFN-α is sufficient to induce nephritis acceleration, our data emphasize a critical role for IFN-independent signaling in TLR7-mediated lupus nephritis. Further, despite upregulation of IL-1β after TLR7 stimulation, deletion of IL-1β is not sufficient to reduce lupus nephritis development in this model.
Systemic lupus erythematosus is a devastating autoimmune disease characterized by cycles of disease flare that lead to permanent organ damage (1). Lupus nephritis is a feared complication of systemic lupus erythematosus; 25% of lupus nephritis patients go on to develop end-stage renal disease within 10 y after renal compromise (2) despite current treatment regimens (3). Triggers that lead to renal disease flares still remain unclear; thus, there is a critical need to identify factors that promote lupus nephritis to develop novel targets and therapies.
Polymorphisms in TLR7 are linked with the development of lupus (4, 5). Murine models of lupus have also provided genetic and experimental evidence to support a role for TLR7 activation in the pathogenesis of lupus nephritis (6, 7). Male BXSB mice, which contain an additional copy of TLR7 on the Y chromosome, develop lupus nephritis, whereas female mice are protected (8). Transgenic mice overexpressing TLR7 also develop lupus (9). Inducible murine lupus models, such as the pristane model, are dependent on TLR7 signaling for lupus pathogenesis (10). Interestingly, even repeated TLR7 agonist epicutaneous application to wild-type (WT) mice leads to development of lupus-like characteristics, including mild lupus nephritis (11). Despite these supportive data, the mechanisms by which TLR7 signaling leads to nephritis remain unknown.
TLR7 activation can result in both type I IFN production and NF-κB activation in various cell populations such as dendritic cells (DCs), monocytes, macrophages, and B cells. Type I IFNs, including IFN-α, can promote lupus development (12–14) and are sufficient to induce acceleration of nephritis in lupus-prone mice (15). In addition, various genetic and inducible models of lupus are protected by deletion of type I IFN signaling (12, 16). Given the importance of type I IFN in lupus, we hypothesized that TLR7 induction of type I IFNs was responsible for its effects. To test this, we examined an inducible model of lupus flare via epicutaneous stimulation of TLR7 in lupus-prone mice. The NZM2328 mouse strain was selected for this study because these mice have spontaneous development of lupus characteristics, including autoantibody production around 14 wk of age, splenomegaly, and lupus nephritis around 35 wk of age (17). Surprisingly, we found that TLR7 treatment of young mice accelerated lupus nephritis development, including renal infiltration of innate immune populations, immune complex deposition, and IL-1β upregulation, independently of type I IFNs. Other manifestations, such as liver inflammation and splenomegaly, required type I IFNs to occur. Further, we demonstrated that a novel IL-1β−/− mouse on the NZM2328 background was also susceptible to TLR7-accelerated nephritis. These data support IFN-independent immune activation in the presence of robust TLR7 stimulation as sufficient for acceleration of lupus nephritis.
Materials and Methods
All mice were bred and housed at the University of Michigan in specific pathogen-free housing. All mice were treated according to our University of Michigan Institutional Animal Care and Use Committee–approved protocol. NZM2328 mice and INZM (lacking the α-chain of the IFN-α/β receptor) mice were a kind gift of Dr. C. Jacob, University of Southern California (12). NZM2328 IL-1β−/− mice were generated through the University of Michigan Transgenic Animal Model Core via CRISPR/Cas9 technology using IL-1β CRISPR/Cas9 knockout (KO) Plasmid sc-421097 from Santa Cruz Biotechnology (Santa Cruz, CA). An 8-bp deletion resulting in a frame shift mutation was confirmed via Sanger sequencing through the University of Michigan DNA Sequencing Core (Supplemental Fig. 1). NZM2328 IL-1β−/− mice were backcrossed onto the NZM2328 background for three generations to eliminate off-target effects of using the CRISPR system, followed by heterozygote crossing to develop mice homozygous for the IL-1β deletion. Female 10-wk-old mice were used for all experiments. The NZM mouse model has minimal autoantibody production and no nephritis at this age (12). Males were not used, as they do not achieve a lupus phenotype in this model, and lupus is predominantly a female disease (18).
TLR7 cutaneous stimulation
Female 10-wk-old mice were treated via epicutaneous application of 100 μg of the TLR7 agonist R848 (Enzo Life Sciences) dissolved in 8 μl of DMSO, or DMSO alone as a control, to the right ear three times weekly until euthanasia. DMSO and R848-treated mice were housed in separate cages to avoid cross-contamination. For some studies, mice were treated for only 2 wk before euthanasia (for evaluation of manifestations preproteinuria); others were treated until they developed proteinuria or became moribund (for survival studies), followed by euthanasia. Control mice were harvested with their paired R848-treated littermates.
Analysis of anti-dsDNA and IgG serum levels
Serum was collected every 2 wk. Anti-dsDNA and IgG levels were analyzed with the use of ELISA kits (Alpha Diagnostic, San Antonio, TX, and Innovative Research, Novi, MI).
Analysis of proteinuria
Urine was collected weekly, and protein was screened via dipstick followed by albumin (alb) measurements via Albuwell kits (Exocell, Philadelphia, PA) and total creatinine (Cr) via commercial kit (BioAssay Systems, Hayward, CA). Urinary protein excretion was represented by the alb/Cr ratio.
Renal histopathology and immune complex deposition scoring
Glomerular inflammation (activity index) and scarring (chronicity index) of murine kidneys were quantified in a blinded fashion (by J.B. Hodgin) on perfused kidneys fixed in 10% formalin followed by 3-μm sectioning and periodic acid–Schiff (PAS) staining as previously described by us and others (19, 20). In brief, a semiquantitative scoring system (0: no involvement; 0.5: minimal involvement [10% of section]; 1: mild involvement [10–30% section]; 2: moderate involvement [31–60% of section]; and 3: severe involvement [60% of section]) was used to assess 13 different parameters of activity and chronicity (mesangial hypercellularity, mesangial deposits, mesangial sclerosis, endocapillary cellular infiltrate, subepithelial and subendothelial deposits, capillary thrombi, capillary sclerosis, cellular or organized crescents, synechiae, tubular atrophy, and interstitial fibrosis). The chronicity and activity index were generated by compiling the scores from groups of related parameters (for activity: mesangial hypercellularity, mesangial deposits, and endocapillary cellular infiltrate; for chronicity: interstitial fibrosis, tubular atrophy, synechiae, organized crescents, and capillary sclerosis). Glomerular immune complexes were quantified by immunofluorescence microscopy as previously described (19). Briefly, 6-μm frozen kidney sections were stained for 1 h at 4°C with Texas Red–conjugated anti-mouse IgG (Sigma-Aldrich) and FITC-conjugated anti-C3 (Immunology Consultants Laboratory, Portland, OR) followed by Hoechst (Invitrogen, Eugene, OR) counterstain to stain DNA. Quantification of immune complex staining in eight glomeruli per mouse was performed at the Center for Live-Cell Imaging at the University of Michigan using MetaMorph version 7.0.6 to calculate the mean fluorescence in a defined area for each stain. Glomeruli were identified based on DAPI staining and outlined to define the area for analysis. Both FITC and Texas Red staining were calculated and shown as staining per glomerulus.
Inflammation of murine livers were quantified blindly by J.B. Hodgin on livers fixed in 10% formalin followed by 3-μm sectioning and PAS staining as described by others (21). In brief, a scoring system (0: <0.5 inflammatory foci/field; 1: 0.5–1.0 foci/field; 2: 1.0–2.0 foci/field; 3: >2.0 foci/field) was used.
Images of H&E-stained skin sections and PAS-stained kidney and liver sections were captured using an Olympus BX41 microscope with a ×100 objective (total magnification = ×1000). Images of kidney immune complex staining were captured at the Center for Live-Cell Imaging at the University of Michigan Medical School using an Olympus IX70 inverted microscope (Olympus, Center Valley, PA) with a ×40 objective.
Following euthanasia, the lymph nodes (LNs) (draining LN [dLN] from the cervical chain on the treated side and non-dLN from the inguinal chain) and spleen were removed, teased apart, and passed through a 70-μm filter to generate single-cell suspensions. One of the kidneys was also removed, minced, and digested as previously described (19) with 0.1 mg/ml Liberase (Roche), 200 U/ml DNAse (Roche), and 2.4 mM CaCl2 in DMEM (Invitrogen) at 37°C in a humidified incubator for 1 h. The tissue was then passed through a 70-μm cell strainer, and RBCs were lysed with multispecies RBC Lysis Buffer (eBioscience). Live cells were counted via trypan blue exclusion. The cells were incubated in flow block (1% horse serum and 1% BSA in PBS) for 1 h, then stained for 1 h on ice using the following Abs: CD3-allophycocyanin, F4/80-PE, CD11b-allophycocyanin, CD11c-PE, CD4 allophycocyanin/Cy7, CD8 allophycocyanin/Cy7, B220 allophycocyanin (BioLegend, San Diego, CA), and Qdot 506 Streptavidin Conjugate (Invitrogen), followed in some cases by permeabilization and intracellular staining for IgH+L A488 (Southern Biotech, Birmingham, AL). The flow data were collected using a BD LSR II flow cytometer and analyzed via FlowJo version 10.0.7 (Tree Star). The following gating strategy was used: live cells were gated for T cells: CD3+; macrophages: CDllcintCDllb+F480+; DCs: CDllc+CDllb+F480−; B cells: CD4−CD8−IgH+L+B220+; and plasma cells: CD4−CD8−IgH+LhiB220int-low.
Bone marrow–derived macrophages
Bone marrow–derived macrophages (BMDM) were generated as previously described (22). Bone marrow was flushed from the tibiae and femurs of 10-wk-old NZM2328 mice and NZM IL1β−/− mice and plated in macrophage differentiation medium (59% IMDM, 10% FBS, 30% l-cell supernatant, and 1% penicillin/streptomycin) for 7 d at 37°C in 5% CO2. BMDM were then plated 2 × 106 cells per well of a 6-well plate for detection of IL-1β activation.
BMDM were incubated with or without 1 μg of LPS for 4 h followed by 1 h incubation with or without 5mM of ATP to activate the inflammasome. Secreted IL-1β was measured via ELISA (DuoSet ELISA kit; R&D, Abingdon, U.K.). Cells were lysed with RIPA plus protease inhibitors for 10 min on ice. Ten micrograms of each lysate was then run on a 10% SDS–polyacrylamide gel and blotted on nitrocellulose membranes (GE Healthcare). The membranes were blocked in 4% milk/TBST, followed by probing with anti-murine IL-1β Ab 1:500 (Cell Signaling Technology) overnight. Then, the membrane was incubated with anti-rabbit IgG-HRP 1:1000 (Abcam). The pro–IL-1β band was detected using WesternBright Quantum (Advansta) and imaged using Ω Lum C system (Aplegen).
Real-time quantitative PCR analysis
Kidney tissue was homogenized in TriPure (Roche), and RNA was purified via Direct-zol RNA MiniPrep (Zymo Research). RNA (100 ng) was reverse-transcribed into cDNA, and quantitative real-time PCR analysis was completed on an ABI PRISM 7900HT (Applied Biosystems) by the DNA Sequencing Core at the University of Michigan. The primers used were as follows (all listed 5′→3′): C-C Motif Chemokine Ligand 2 (ccl2) 5′-AGGTCCCTGTCATGCTTCTG-3′ (forward), 5′-GGATCATCTTGCTGGTGAAT-3′ (reverse); Myxovirus (influenza virus) resistance 1 (mx1) 5′-GATCCGACTTCACTTCCAGATGG-3′ (forward), 5′-CATCTCAGTGGTAGTCAACCC-3′ (reverse); C-X-C motif chemokine ligand 10(cxcl10) 5′-ATCATCCCTGCGAGCCTAT-3′ (forward), 5′-ATTCTTGCTTCGGCAGTTAC-3′ (reverse) chemokine (C–C motif) ligand 5 (ccl5) 5′-CAATCTTGCAGTCGTGTTTG-3′ (forward), 5′-GGAGTGGGAGTAGGGGATTA-3′ (reverse); chemokine (C–C motif) ligand 4 (ccl4) 5′-AGCAACACCATGAAGCTCTG-3′ (forward), 5′-CTGTCTGCCTCTTTTGGTCA-3′ (reverse); C-X-C motif chemokine 13 (cxcl13) 5′-AGAGGTTTGCGAGATGGACT-3′ (forward), 5′-GAGCCTGGACCTTTAAGCTG-3′ (reverse); β-Actin 5′-TGGAATCCTGTGGCATCCTGAAAC-3′ (forward), 5′-TAAAACGCAGCTCAGTAACAGTCCG-3′ (reverse); TNF-α (tnf) 5′-CCCACTCTGACCCCTTTACT-3′ (forward), 5′-TTTGAGTCCTTGATGGTGGT-3′ (reverse); il1b 5′-CCCTGCAGCTGGAGAGTGTGGA-3′ (forward), 5′-CTGAGCGACCTGT-CTTGGCCG-3′ (reverse); il6 5′-TAGTCCTTCCTACCCCAATTTCC-3′ (forward), 5′-TTGGTCCTTAGCCACTCCTTC-3′ (reverse). Gene expression was calculated relative to β-actin (2^β-actin/2^gene).
All statistics were completed using GraphPad Prism 6.0. For figures in which more than two comparisons were made, ANOVA testing was used. A Tukey multiple comparison test was used for normally distributed data, and a Kruskal–Wallace/Dunn multiple comparison test was used for nonparametric data. For figures in which comparisons between two groups were completed, a two-tailed Student t test for normally distributed data was used. Welch’s correction was applied when required for significant differences in variances. For nonnormally distributed data, a Mann–Whitney U test was used. Survival analysis following R848 application was completed via log-rank test. Pearson correlation was used for comparison of renal activity scores and alb/Cr values.
TLR7 epicutaneous stimulation leads to an IFN-independent decline in survival
TLR7 stimulation in the skin may serve an important role in lupus flare induction (11, 23); thus, we first examined the effects of cutaneous TLR7 stimulation on young, preautoimmune NZM2328 lupus-prone mice. These mice typically develop lupus nephritis around 35 wk of age, followed by a decline in survival (24). Intriguingly, R848 topical treatment of 10-wk-old NZM2328 mice led to a rapid decline in survival when compared with DMSO-treated controls (Fig. 1A). This decline was marked by a nephrotic appearance, including ascites, edema, and lethargy. Given that TLR7 stimulation is associated with increased IFN signaling, we hypothesized that this accelerated death was driven by type I IFNs. Thus, we next examined the effects of R848 on NZM2328 mice that are deficient in the type I IFN receptor (INZM). These mice are protected from naturally occurring development of lupus for up to 2 y of life (12). To our surprise, treatment of these mice with R848 led to edema, ascites, and a rapid decline in survival in a time frame similar to NZM2328 mice (Fig. 1). No difference in the survival curves between NZM and INZM mice was noted via log-rank test (p = 0.723). Median survival was 30 d for both strains. These data suggest that despite absent type I IFN signaling, both INZM and NZM2328 mice exhibit a rapid decline in survival following epidermal TLR7 activation.
TLR7 epicutaneous stimulation leads to accelerated development of murine lupus
We next examined the development of systemic lupus characteristics following R848 stimulation. Using our assay, dsDNA Abs typically rise to ∼100,000 U/ml during nephritis onset in NZM2328 mice (19). Importantly, although only NZM2328 mice treated with R848 demonstrated significant acceleration of total IgG production, both NZM2328 and INZM mice demonstrated increased dsDNA Abs in the serum (Fig. 2A, 2B); however, the rise of dsDNA Abs in the serum was greater in NZM versus INZM mice (p = 0.0167 NZM versus INZM at 2 wk). This suggests that autoantibody production following TLR7 epicutaneous stimulation is enhanced by but not dependent on type I IFN signaling. Consistent with previous literature (12, 25), splenomegaly was detected following R848 stimulation in NZM2328 but not INZM mice, supporting a role for IFNs in TLR7-driven splenomegaly (Fig. 2C). An increase in total splenic cells following R848 stimulation was detected in only NZM2328 mice; analysis of cell subsets identified no significant changes in T cells, DCs, or macrophages (Fig. 2D, 2E). Further, NZM2328 mice treated with R848 demonstrated accelerated development of liver inflammation as indicated by necrotic hepatocytes and focal portal inflammation. INZM mice treated with R848 did not develop liver inflammation, indicating that type I IFNs are required for this manifestation following TLR7 activation (Fig. 2F, 2G).
We next examined changes in B cells and Ab-secreting cells in the spleen and LNs. As shown in Fig. 3, both total B cells (B220+) and Ab-secreting cells (IgH+Lhigh) increased in the spleen of NZM2328 mice, but only Ab-secreting cells increased in INZM mice after R848 stimulation (Fig. 3A–C), suggesting that amplification of splenic Ab production is independent of type I IFNs. We then examined the LN populations to determine if IFN-independent B cell activation was also occurring there. As shown in Fig. 3D, stimulation with R848 resulted in increased total number of cells in the cervical (draining) LN in both NZM2328 and INZM mice. In contrast, increased cell numbers were noted in inguinal (nondraining) LN in only NZM mice after R848 treatment. The total number of B cells (B220+) in both NZM2328 and INZM mice was increased in dLN but not the non-dLN following R848 stimulation (Fig. 3E). NZM2328 mice showed an increase in the Ab-secreting cell population in the dLN and non-dLN, but INZM mice only demonstrated an increase in the dLN (Fig. 3F). Of note, the overall numbers of B cells and Ab-secreting cells were significantly smaller in the INZM mice. These data indicate that TLR7 stimulation amplifies local B cell and Ab production responses in the absence of type I IFN signaling, but the responses may be diminished. However, systemic amplification of Ab-secreting cells in the non-dLN required the presence of type I IFN signaling. Together, this suggests that the accelerated autoantibody production seen after TLR7 cutaneous stimulation is partially type I IFN–independent and that the Ab-secreting cells in the dLN and spleen may serve as a site for autoantibody production in this model.
TLR7 epicutaneous stimulation leads to accelerated development of lupus nephritis in a type I IFN–independent manner
Past work has shown type I IFNs are sufficient to induce renal flares in lupus-prone mice and are required for lupus nephritis development in several murine models (12, 15, 16, 26). We next examined whether TLR7 epicutaneous stimulation led to accelerated lupus nephritis development and whether it was dependent on type I IFNs. Surprisingly, both NZM2328 and INZM mice treated with R848 demonstrated a rise in urinary alb/Cr ratio, supportive of glomerular damage (Fig. 4A, 4B). Histopathologic scoring demonstrated a significant increase in renal activity score when mice were treated with R848 until moribund (Fig. 4C, 4D). A strong positive correlation was detected between renal activity score and urinary alb/Cr ratio for both NZM2328 and INZM mice (Fig. 4E), which supports renal inflammation in R848-treated mice. No significant increase in renal chronicity index score for prolonged exposure in NZM2328 or INZM mice was detected (Fig. 4F). As immune complex deposition is a hallmark of lupus nephritis, this was also assessed. R848 treatment of the NZM2328 and INZM mice led to a significant increase in both IgG and C3 deposition within the kidney (Fig. 5A–C). To examine whether TLR7-induced lupus created similar transcriptional changes to naturally occurring lupus nephritis found in older, untreated NZM2328 mice, we examined transcriptional signatures of the kidneys of both. As shown in Supplemental Fig. 2, identical upregulation of various inflammatory and type I IFN–associated genes was noted in both R848-induced and naturally occurring nephritis. Together, these parameters support development of accelerated lupus nephritis following cutaneous stimulation with a TLR7 agonist in a type I IFN–independent manner.
TLR7 epicutaneous stimulation leads to upregulation of IL-1β, but nephritis is not dependent on IL-1β
Because TLR7-mediated lupus nephritis developed in the absence of type I IFN signaling, we next assessed what other pathways may play a role in TLR7-mediated lupus nephritis. We chose to examine NF-κB–regulated cytokines, as this transcription factor is also activated downstream of TLR7, with the hypothesis that cytokines upregulated in the kidney in both NZM and INZM mice may be important in TLR7-mediated nephritis. We examined mRNA changes in the kidney of INZM and NZM mice treated for 2 wk with R848 or DMSO control. Following stimulation, there were significant changes in the expression of NF-κB–regulated cytokines tnf or il1b (Fig. 6A–C) in the NZM mice treated with R848 versus control, but there was only a significant upregulation of il1b in INZM mice treated with R848 versus DMSO (Fig. 6B). These data support a potential common role for IL-1β in TLR7-mediated lupus nephritis flare in NZM and INZM mice.
To examine the role of IL-1β in TLR7-mediated nephritis, we generated NZM2328 mice that lack IL-1β (Supplemental Fig. 1) and treated them epicutaneously with R848 or DMSO. These mice also exhibited a decline in survival (Fig. 6D), increase in dsDNA IgG Abs (Fig. 6E), and development of lupus nephritis (Fig. 6F–H). These data indicate that like type I IFNs, IL-1β is not required for TLR7-mediated nephritis.
TLR7 epicutaneous stimulation leads to early infiltration of DCs and upregulation of ccl2
Previous work has identified important roles for macrophage and DC infiltration into the kidney prior to nephritis onset (19, 26). We thus examined inflammatory cell recruitment into the kidney prior to lupus nephritis onset to determine the earliest cellular contributors to TLR7-induced nephritis. Two weeks after R848 or DMSO treatment, prior to proteinuria onset, kidneys were removed and digested, and pre/early nephritic renal immune cell populations were examined by flow cytometry. (T cell, macrophage, and DC population gating strategies are shown in Fig. 7A). Consistent with previous data (19), we did not see a change in the T cell (CD3+) or B cell (CD19+) population prior to proteinuria onset (Fig. 7B, 7C). We also did not see changes in the macrophage (CDllb+CDllcint F4/80+) population at this time point (Fig. 7D). However, we detected a significant infiltration of the DC population (CDllb+ CDllc+ F4/80−) in both the NZM and INZM mice treated with R848 but not DMSO (Fig. 7E). The number of infiltrating cells was overall smaller in INZM mice, but this was not statistically different between NZM and INZM mice (p = 0.2391 by unpaired Student t test). This suggests that TLR7 activation promotes early infiltration of DCs into the kidney, similar to what has been observed in other genetic models of murine lupus (26) but in an IFN-independent fashion.
To identify potential mechanisms of chemotaxis into the kidney, we examined chemokine expression in the kidney following DMSO or R848 treatment with the hypothesis that relevant chemokines should rise in both NZM and INZM mice with R848 treatment. As shown in Fig. 8, there were no changes in ccl4 (MIP-2) or cxcl13 expression in either strain following R848 treatment (Fig. 8A, 8B). ccl5 expression increased in NZM2328 but not INZM mice after R848 exposure (Fig. 8C). IP-10 (cxcl10) expression also increased in only NZM2328 mice after R848 exposure (Fig. 8D). Importantly, upregulation of ccl2 (MCP-1), a chemokine that rises during lupus nephritis (27) and whose receptor (CCR2) is expressed on monocytes, macrophages, and DCs (28, 29), was demonstrated to increase in both NZM and INZM mice after R848 treatment (Fig. 8E). These data suggest TLR7-mediated upregulation of ccl2 may serve as a potential trigger for induction of inflammatory cell infiltration in an IFN-dependent manner.
In this paper, we examine a novel model of lupus flare in which accelerated lupus nephritis is induced in genetically prone mice after 3–4 wk of epicutaneous application of a TLR7 agonist, R848. R848 treatment led to accelerated development of autoantibodies, splenomegaly, liver inflammation, and lupus nephritis. With the use of NZM mice lacking the type I IFN receptor, we demonstrated that TLR7-mediated splenomegaly and liver inflammation were dependent on type I IFN signaling. Surprisingly, however, renal injury was independent of type I IFNs. Indeed, INZM mice demonstrated proteinuria, increased renal immune complex formation, upregulation of NF-κB–regulated cytokines, and infiltration of CDllb+CDllc+F4/80low DCs in the kidney, similar to NZM mice.
Human genetics support a role for TLR7 in lupus (4, 5), and murine data has further supported this (9, 10, 11). BXSB male mice develop lupus secondary to the Yaa locus that contains a duplication of the TLR7 gene (23). The role of TLR7 has also been demonstrated in transgenic mice overexpressing TLR7 (9) and in lupus inducible models such as the pristane model (10). A recent paper also demonstrated that TLR7 epicutaneous stimulation of WT mice led to development of mild lupus characteristics after long-term (13–15 wk) treatment (11). Our model furthers these observations and demonstrates that TLR7 stimulation rapidly (in 3–4 wk) accelerates lupus development in young, lupus-prone mice. Further, we show that mice that are otherwise protected from lupus development in the absence of functional IFN signaling (12) are similarly susceptible to the nephritis-inducing effects of TLR7. Others have shown that overexpression of IFN-α is sufficient to stimulate lupus nephritis in a similar time frame to TLR7 treatment (15). However, in the presence of TLR7 agonist, type I IFN signaling is not required. This observation may reflect duplicative roles of other inflammatory cytokines induced by TLR7.
Recently, the role of type I IFN has been explored in other murine models of lupus. Treatment of mice with HgCl2 is able to induce autoantibody production in a lysosomal TLR–dependent but type I IFN–independent fashion (30), similar to the data we see in the R848 NZM model. Development of renal disease and direct contribution of TLR7 in this xenobiotic model were not assessed (30). In a different model, exacerbation of autoimmunity via deletion of TLR9 in the MRL/lpr mouse was found to be type I IFN dependent (31). In this model, IFNs are required for development of renal disease but not anti-nucleosome Ab production. HepG-2–positive anti-nuclear Abs and anti-RNA Abs were dependent on type I IFN signaling (31). Combined with our findings, these data suggest that the role for IFNs in lupus development may vary depending on the stimuli. Autoantibody production may be independent of IFNs when the trigger is strong environmental exposure (such as R848 or HgCl2). Conversely, lupus that develops based on genetic factors without the need for external triggers (such as TLR9-deficient MRL/lpr or NZM2328 mice) may require IFN signaling for autoantibody and nephritis development.
NF-κB activation occurs downstream of TLR7 stimulation, and activation of this pathway is important for stimulation of lupus nephritis (32). Thus, TLR7 activation of NF-κB pathways may be a critical step in driving lupus flares. Mutations in A20 that drive NF-κB activation in WT mice led to lupus nephritis (33, 34), possibly through decreased regulation of inflammasome activity (35). Although roles for TNF-α and IL-6 in lupus development have been suggested (36), and IL-6 can exacerbate TLR7-driven lupus (37), we do not see these cytokines significantly upregulated in the kidney in an IFN-independent manner in our TLR7 stimulation model. Intriguingly, we see an upregulation of IL-1β in the kidney prior to the onset of lupus nephritis in both NZM2328 and INZM mice. This finding coincides with data that support a potential role for IL-1β and inflammasome activation in lupus development (reviewed in Ref. 38). However, deletion of IL-1β in NZM2328 mice did not reduce the development of lupus nephritis after TLR7 exposure. Given that TLR7 activates pleotropic pathways, generation of double or triple KOs may be required to hinder lupus nephritis development in the presence of such a strong inflammatory activator.
Similar to others (12), we found that splenomegaly following TLR7 stimulation was type I IFN dependent. Interestingly, the massive R848-induced splenomegaly was out of proportion to the small increases in B and T cell populations in the spleen that were identified by flow. This supports a possible role for extra medullary hematopoiesis resulting in R848-mediated splenomegaly, consistent with findings by other groups (39, 40).
TLR7 epicutaneous stimulation can lead to development of autoantibody production in the absence of type I IFN signaling. We were able to see an increase in dsDNA Abs in NZM2328 and INZM mice following R848 stimulation, although the increase in anti-dsDNA levels was less robust in INZM mice. Increases in B (B220+) cell numbers in the spleen were noted in only the NZM mice, which suggests that expansion of B cell populations in the spleen may require type I IFNs following TLR7 exposure. This would be consistent with a recently described role for splenic follicular DC production of type I IFN in promotion of autoreactive B cell populations (41, 42). Type I IFNs also enhance the TLR7 signaling response, allowing for autoantibody production (43), and they support survival of transitional stage B cells in the spleen (44). Thus, the more extensive systemic immune activation seen in NZM mice may be secondary to these effects of type I IFN. Overexpression of TLR7 stimulation has also been shown to lead to expansion of transitional stage B cell populations in a type I IFN–independent manner (45). Given that we see a rise in Ab-secreting cells in the spleen and dLN but not in the non-dLN of INZM mice, this may indicate that a certain threshold of TLR7 stimulation, and possibly other cytokines, is needed to generate Ab-secreting cell expansion in a type I IFN independent manner.
TLR7 stimulation results in the recruitment of CD11b+CD11c+F4/80− DCs in the kidney prior to proteinuria onset without significant changes in the T cell or B cell population. Interestingly, this recruitment occurs in a type I IFN–independent manner. The DC chemoattractant MCP-1/CCL2 was increased in the kidney in both NZM and INZM mice, which supports its potential role for instigating nephritis downstream of TLR7 activation. Intriguingly, CCL2 has been shown to rise during human lupus nephritis development (27) and is proposed as a urinary biomarker for disease (46). Blockade of CCL2 demonstrates some efficacy in murine lupus models (47). Macrophages, which also are recruited by CCL2, were not increased in the kidney after R848 treatment, which may reflect differential upregulation of CCL2’s receptor, CCR2, on DCs and macrophages in this model. Overall, consideration of the role of CCL2 downstream of TLR7-driven lupus should be made in future studies.
In summary, we have demonstrated a novel role for TLR7 epicutaneous stimulation in mediating lupus flare in lupus-prone mice. Following stimulation, splenomegaly and liver inflammation occur in a type I IFN–dependent manner. Importantly, autoantibody production and lupus nephritis occur independent of IFN signaling. We demonstrate that IL-1β is not required for development of TLR7-activated lupus nephritis. Future studies should address the role of CCL2 in TLR7-mediated lupus nephritis. In addition, our data lend a note of caution to ongoing trials using type I IFN blockade in lupus nephritis; consideration of the upstream drivers of nephritis (which may vary in individual patients) may be important for identifying effective treatment modalities. In particular, our data would suggest that blockade of type I IFN signaling my not be effective if TLR7 is driving the phenotype.
This work was supported by the Stone Family Foundation and the National Institutes of Health via the National Institute of Arthritis and Musculoskeletal and Skin Diseases under Awards K08AR063668 and R01AR071384 (to J.M.K.) and National Institute of Allergy and Infectious Diseases Research Training in Experimental Immunology Training Grant T32AI007413.
The online version of this article contains supplemental material.
Abbreviations used in this article:
bone marrow–derived macrophage
The authors have no financial conflicts of interest.