Type II but Not Type I IFN Signaling Is Indispensable for TLR7-Promoted Development of Autoreactive B Cells and Systemic Autoimmunity

Systemic lupus erythematosus (SLE) is a debilitating autoimmune disease mediated by Abs reactive to DNA- and RNA-associated self-antigens. Although extrafollicular Ab-forming cell (AFC) and follicular germinal center (GC) responses are traditionally thought to be involved mainly in anti-pathogen Ab responses, AFCs and GCs can also develop spontaneously in the absence of purposeful immunization or overt infection (1, 2). Dysregulation in spontaneous AFC (Spt-AFC) and Spt-GC responses in SLE patients and SLE-prone mice promotes autoreactive B cell survival and pathogenic autoantibody production (3–8). However, the mechanisms that drive Spt-AFC and spontaneous GC (Spt-GC) responses which can lead to the development of autoreactive B cells, autoantibodies and SLE are incompletely understood.

Genome-wide association studies revealed multiple risk alleles in the TLR pathway that are strongly associated with SLE development (9, 10). Overexpression and overactivity of TLR7 are associated with development of SLE in humans and in mouse models (11–13). Recent collaborative works also have highlighted the significance of TLR7 in promoting extrafollicular double negative B cell populations involved in SLE pathogenesis (8). TLR7 gene escape from X chromosome inactivation enhances the B cell response to TLR7 stimulation (14). Studies in the MRL-lpr model highlight the opposing inflammatory and regulatory roles for TLR7 and TLR9, respectively, in extrafollicular autoreactive B cell responses and SLE development (15). Using a variety of SLE mouse models, we and others discovered that TLR7 signaling specifically in B cells promoted SLE development through formation of Spt-AFCs and Spt-GCs (11, 16, 17). These data highlight the importance of TLR7 signaling in promoting autoimmune AFC and GC responses, leading to the development of autoreactive B cells, autoantibody production and SLE; however, mechanisms of TLR7-mediated autoimmunity and SLE development are not clearly defined.

Type I IFN (T1IFN) and type II (IFN-γ) IFN signaling are both implicated in SLE (9, 18). An IFN-α signature is associated with human SLE (19, 20). Lupus-prone mice deficient in IFN-αR have reduced anti-nuclear Abs (ANAs) and renal disease (21). Exogenous IFN-α treatment accelerates the SLE phenotype in female (NZW × NZB) F1 mice with an early development of anti-dsDNA Abs and nephritis (22). The administration of IFN-α in lupus-prone mice also leads to enhanced short-lived AFC and GC responses (23), and an altered B cell selection in the GCs (24). We and others recently showed a B cell–intrinsic role of T1IFN signaling in loss of tolerance and autoreactive B cell development in the AFC and GC pathways in lupus-prone mice (25, 26). SLE patients also showed an increased production of IFN-γ (27, 28), and blockade of IFN-γ normalizes IFN regulated gene expression in SLE patients (29). MRL/lpr and NZW/NZBF1 lupus-prone mice deficient in IFN-γ have reduced autoantibody production and renal disease (30–35). We and others discovered a B cell–intrinsic requirement of IFN-γ signaling in Spt-AFC and Spt-GC responses and SLE development (4, 36). These data indicate that both T1IFN and IFN-γ signaling contribute to autoreactive B cell development in the AFC and GC pathways. However, the mechanisms and relative contribution of T1IFN and IFN-γ signaling in TLR7-mediated AFC, GC, and autoimmunity development have never been directly investigated using a TLR7-induced model, especially when T1IFN signaling is believed to be strongly associated with TLR signaling and SLE development.

T1IFN and IFN-γ signaling mediate signals through STAT1. Interestingly, we previously reported a B cell–intrinsic requirement of STAT1 in the Spt-GC responses (36) which we also showed dependent on TLR7 signaling in B cells (16). However, how STAT1 downstream of T1IFN and IFN-γ signaling in B cells controls TLR7-mediated AFC and GC responses, promoting development of autoreactive B cells and SLE, remain unanswered. In this study, we used TLR7-induced and -overexpression models of SLE to dissect the relative contribution of T1IFN and IFN-γ signaling in driving autoimmune AFC and GC responses, promoting development of autoreactive B cells, autoantibody production, and SLE. We found that IFN-γ and STAT1 signaling in B cells played an indispensable role in TLR7-promoted AFC and GC responses and SLE development, whereas global T1IFN signaling only moderately contributed to these processes. Using a TLR7-induced SLE model we show that TLR7 not only increased expression of T1IFN genes but also the IFN-γ gene in splenic tissue compared with untreated mice. Our data from IFN-γ reporter mice and RNA sequencing (RNA-seq) analysis indicate that TLR7-driven and T cell–produced IFN-γ activates multiple signaling pathways to regulate TLR7-promoted SLE. Together, our data show that TLR7 drives autoreactive B cell development in the AFC and GC pathways and promotes SLE by activating T1IFN and IFN-γ signaling in which IFN-γ signaling plays an indispensable role, which cannot be overcome by the activation of other downstream pathways of TLR7 signaling.

C57BL/6J (B6), B6.129S7-Ifngr1^tm1Agt/J (IFN-γR1^−/−), B6(Cg)-Ifnar1^tm1.2Ees/J (IFN-αR1^−/−), C.129S4(B6)-Ifng^tm3.1Lky/J mice (IFN-γ eYFP), B6.129S(Cg)-Stat1^tm1Dlv/J (STAT1^−/−), B6.129S-Stat1^tm1Mam/Mmjax (STAT1^fl/fl), C57BL/6N-Ifngr1^tm1.1Rds/J (IFN-γR^fl/fl), B6.129-Tbx21^tm2Srnr/J (T-bet^fl/fl), and B6.SB-Yaa/J (B6.yaa) mice were originally purchased from The Jackson Laboratory and bred in house. The B6.Sle1b mice (congenic for the Sle1b sublocus) were described previously (37). B6.Sle1b^Yaa mice were generated by crossing B6.Yaa mice to B6.Sle1b mice. B6.CD23^Cre (B6.Cg-Tg(Fcer2a-cre)5Mbu, provided by Dr. Meinrad Busslinger, were crossed to B6.Sle1b autoimmune mice in house. All animal studies were conducted at Penn State Hershey Medical Center in accordance with the guidelines approved by our Institutional Animal Care and Use Committee. Animals were housed in a specific pathogen-free barrier facility.

For epicutaneous imiquimod (IMQ) treatment, 5% IMQ cream (Glenmark Pharmaceuticals) was applied on the ears of mice 3 times weekly for 4–12 wk, depending on the experiments as previously described (38, 39). For PCR array and RNA-seq sample preparation, mice were treated for 4 wk. To study the systemic autoimmune responses, such as AFC, GC, follicular helper T (Tfh), and Ab responses, mice were treated for 8 wk. To evaluate immune complex deposition and kidney pathology, mice were treated for 12 wk (three times a wk for 8 wk, followed by once a week for another 4 wk).

Flow cytometric analysis of total mouse splenocytes was performed using the following Abs: B220-BV605 (RA3-6B2), CD4-AF700 (RMP4-5), CD44-allophycocyanin (IM7), CD62L-PECy7 (MEK-14), PD1-PE (29F.1A12), NK1.1-biotin (PK136), Ly6G-PB (1A8), and MHCII-PE-Cy7 (M5/114.15.2) (BioLegend); GL7-FITC (GL-7), CD95- PeCy7, CXCR5-biotin (2G8), IFN-γR-biotin (GR20), STAT1-PE (1/STAT1), CD11b-AF700 (M1/70), and CD11c-FITC (HL3) (BD Biosciences); Ly6C-allophycocyanin (HK1.4), PDC-PE (eBio927), and CD8α (clone 53–6.7) (eBioscience). All cells were stained with fixable viability dye-eFluor780 (Invitrogen) prior to surface staining. Stained cells were analyzed using the BD LSR II flow cytometer (BD Biosciences). Data were acquired using FACSDiva software (BD Biosciences) and analyzed using FlowJo software (Tree Star).

Mouse spleens or kidneys were embedded in OCT compound and snap frozen over liquid nitrogen. Five-micrometer-thin sections were cut on a cryostat, mounted on ColorFrost Plus microscope slides (Thermo Fisher Scientific), and fixed in cold acetone for 20 min. The following Abs and reagents were used for immunofluorescence staining of mouse spleen sections for GCs: PE–anti-CD4 (GK1.5; BioLegend), FITC-GL7 (RA3-6B2; BD Biosciences), and allophycocyanin–anti-IgD (11-26c2a; BD Biosciences). Kidney sections were stained for C3 using FITC–anti-C3 (ICL Laboratories) or Biotin–anti-IgG (Jackson ImmunoResearch), followed by streptavidin (SA)–PE. ANA reactivity was detected by indirect immunofluorescence staining of HEp-2 cell slides using sera from indicated mice at a 1:50 dilution and probed with FITC-rat anti-mouse κ (H139-52.1). The images of stained spleen and kidney sections were captured using the Leica DM4000 fluorescence microscope and analyzed using a Leica application suite–advanced fluorescence software (LAS-AF; Leica Microsystems). For the measurement of the GC area, a total of at least 10 randomly selected GCs (GL-7⁺) per mouse spleen section from at least five mice of each genotype were measured for total area (square micrometers) using the LAS-AF quantitation tool.

Kidneys from 6-mo-old mice were fixed in 10% neutral buffered formalin and embedded in paraffin. Kidney sections were cut at 3- and 6-μm thickness for periodic acid-Schiff (PAS) and H&E staining, respectively. All images were obtained with an Olympus BX51 microscope and DP71 digital camera using cellSens Standard 1.12 imaging software (Olympus America, Center Valley, PA). Two pathologists blinded to the genotype of the mice examined kidney sections. One kidney section per mouse was evaluated: each glomerulus was examined at 400× magnification and scored from 0 (normal) to 4 (severe) based on glomerular size and lobulation, presence of karyorrhectic nuclear debris, capillary basement membrane thickening, and the degree of mesangial matrix expansion and mesangial cell proliferation (40). H&E sections were scored for overall glomerular damage and interstitial inflammation as described (41).

ELISPOT assays were performed as previously described (25, 36). Briefly, splenocytes in RPMI 1640 containing 10% FBS were plated at a concentration of 1 × 10⁶ cells per well onto multiscreen 96-well filtration plates (Millipore) coated with salmon sperm dsDNA (Invitrogen), nucleosome (histone from Sigma Aldrich plated on a layer of dsDNA), or Smith/ribonucleoprotein (SmRNP; Arotec Diagnostics). Serially diluted (1:2) cells were incubated for 12 h at 37°C. dsDNA-, nucleosome-, and smRNP-specific AFCs were detected by biotinylated anti-κ Ab (Invitrogen) and SA–alkaline phosphatase (Vector Laboratories) or alkaline phosphatase-conjugated anti-mouse IgG (Molecular Probes). Plates were developed using the Vector Blue Alkaline phosphatase Substrate Kit III (Vector Laboratories). ELISPOTs were enumerated and analyzed using a computerized ELISPOT plate imaging/analysis system (Cellular Technology).

Serum autoantibodies were measured using standard ELISA protocols as described (3). Briefly, total IgG autoantibody titers were measured in ELISA plates coated with salmon sperm dsDNA, nucleosome or smRNP and detected with biotinylated secondary Ab followed by SA–alkaline phosphatase (Vector Laboratories). The plates were developed using p-nitrophenyl phosphate, disodium salt (Thermo Fisher Scientific) substrates for alkaline phosphatase and read at λ405 nm on Synergy H1 (BioTek Instruments).

RNA was extracted from spleen tissues by TRIzol as per the manufacturer’s instructions (Ambion). OD values of extracted RNA were measured using NanoDrop to confirm an A260:A280 ratio above 1.9. RNA integrity number was measured using BioAnalyzer (Agilent Technologies). RNA 6000 Pico Kit to confirm RIN above 7. The cDNA libraries were prepared using the Ovation RNA-Seq System V2 (NuGEN Technologies, San Carlos, CA) as per the manufacturer’s instructions. The unique barcode sequences were incorporated in the adaptors for multiplexed high-throughput sequencing. The final product was assessed for its size distribution and concentration using BioAnalyzer High Sensitivity DNA Kit (Agilent Technologies) and then loaded onto TruSeq SR v3 flow cells on an Illumina HiSeq 2500 (Illumina) and run for 100 cycles using a paired-read recipe (TruSeq SBS Kit v3; Illumina) according to the manufacturer’s instructions. Illumina CASAVA pipeline (released version 1.8; Illumina) was used to obtain demultiplexed sequencing reads (fastq files) passed the default purify filter. Additional quality filtering used FASTX-Toolkit (http://hannonlab.cshl.edu/fastx_toolkit) to keep only reads that have at least 80% of bases with a quality score of 20 or more (conducted by fastq_quality_filter function) and reads left with 10 bases or longer after being end trimmed with a base quality score of 20 (conducted by fastq_quality_trimmer function). A bowtie2 index was built for the mouse reference genome (GRCm38) using bowtie version 2.1.0. The RNA-seq reads were mapped using Tophat version 2.0.9 (42) supplied by Ensembl annotation file; GRCm38.78.gtf. Gene expression values were computed using fragments per kilo base per million mapped reads values. Differential gene expression was determined using Cuffdiff tool, which is available in Cufflinks version 2.2.1 supplied by GRCm38.78.gtf. Normalization was performed via the median of the geometric means of fragment counts across all libraries, as described (43). Statistical significance was assessed using a false discovery rate threshold of 0.05 to determine the differentially abundant genes. Genes that were expressed at ≥1.5-fold differences between control and test samples were selected for further analysis. The heatmaps of selected genes were generated via the pheatmap package in R using Euclidean distance and Ward clustering method.

The accession number for the RNA-seq data reported in this paper is in the Gene Expression Omnibus: GSE141433 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE141433).

The p values were calculated using unpaired, nonparametric, Mann–Whitney, Student t test (for comparison of two groups), or one-way or two-way ANOVA, with a follow-up Tukey multiple-comparison test (for comparison of more than two groups) or two-way ANOVA, with a follow-up Sidak multiple-comparison test (for comparison of two groups). The p values are as follows: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. GraphPad Prism 6 software was used.

To identify the mechanisms and determine the role of T1IFN and IFN-γ signaling in differential regulation of TLR7-mediated autoimmunity, we have used a TLR7-induced SLE model in which we treated 8 wk old female mice with TLR7 agonist IMQ as described (38). IMQ-treated B6 (B6-IMQ) and autoimmune-prone B6.Sle1b mice (Sle1b-IMQ) displayed more splenocytes compared with untreated control counterparts (Supplemental Fig. 1A). Overall percentage of B220⁺ B cells and CD4⁺ T cells was not significantly different in untreated and treated B6 and B6.Sle1b mice (Supplemental Fig. 1B, 1C), but the numbers of B cells and CD4⁺ T cells were higher in Sle1b-IMQ mice than three other groups (Supplemental Fig. 1D, 1E). In addition, Sle1b-IMQ mice had much higher percentage of splenic B220⁺GL7^hiCD95^hi GC B cells and CD4⁺CD44^hiPD-1^hiCXCR5^hi Tfh cells than untreated B6.Sle1b mice (Fig. 1A, 1B). Flow cytometry gating strategy for GC B cells and Tfh cells is shown (Supplemental Fig. 1F). Elevated GC and Tfh responses in Sle1b-IMQ mice were strongly associated with increased numbers of dsDNA- and nucleosome-specific splenic AFCs, serum ANA reactivity, and autoantibody titers (Fig. 1C–F). Treated B6 mice (B6-IMQ) had a 2-fold higher GC and Tfh responses than untreated B6 mice but no autoantibody production. We then treated 8-wk-old mice with IMQ for 12 wk and found that Sle1b-IMQ mice had more severe nephritis than B6.Sle1b untreated mice and B6-treated mice (Fig. 1G, 1H). These elevated responses in Sle1b-IMQ mice were accompanied by an increased expression of IFN gene signatures in the splenic tissue (Fig. 1I), and elevated IFN-αR and IFN-γR expression on B cells upon TLR7 stimulation in vitro (Fig. 1J). Thus, Sle1b-IMQ mice represent a strong TLR7-driven systemic autoimmunity model with autoimmune phenotypes (elevated AFC, GC, Tfh, ANA, and nephritis) closely resembling human SLE.

Although the induction of T1IFNs by TLR signaling and their roles in SLE pathogenesis have been well described (44–46), we know little about how IFN-γ signaling may contribute to TLR-mediated SLE autoimmunity. Therefore, using the TLR7-induced Sle1b-IMQ model, in this study we determined the relative contribution of T1IFN and IFN-γ signaling in TLR7-mediated SLE autoimmune responses. Although IMQ-treated B6.Sle1b mice deficient in IFN-αR (Sle1b.IFN-αR^−/−-IMQ) had lower GC B cell and Tfh cell responses than Sle1b-IMQ mice, these responses were diminished in treated B6.Sle1b mice deficient in IFN-γR (Sle1b.IFN-γR^−/−-IMQ) (Fig. 2A, 2B). The frequency of GCs between Sle1b-IMQ and Sle1b.IFN-αR^−/−-IMQ mice was similar, but GC sizes were smaller in Sle1b.IFN-αR^−/−-IMQ mice than Sle1b-IMQ control mice, whereas a negligible number of very small GCs was observed in Sle1b.IFN-γR^−/−-IMQ mice (Fig. 2C, 2D). In addition, the number of dsDNA- and nucleosome-specific splenic AFCs (Fig. 2E), serum dsDNA-, nucleosome- and SmRNP-specific Abs (Fig. 2F), and serum ANA reactivity (Fig. 2G) were moderately reduced in Sle1b.IFN-αR^−/−-IMQ mice, whereas they were diminished in Sle1b.IFN-γR^−/−-IMQ mice compared with Sle1b-IMQ control mice. Consistent with systemic autoimmune B and T cell responses and serum ANA reactivity, immune complex (IC) deposition (IgG and C3) and severe glomerulonephritis (GN) in B6.Sle1b-IMQ mice were moderately reduced in Sle1b.IFN-αR^−/−-IMQ mice, whereas they were completely absent in Sle1b.IFN-γR^−/−-IMQ mice (Fig. 2H–J).

Next, we determined the role of IFN-γ signaling in SLE autoimmunity development in B6.Sle1b mice overexpressing TLR7 by breeding Sle1b.IFN-γR^−/− mice to B6.Yaa mice that carry an extra copy of TLR7 to generate Sle1b^YaaIFN-γR^−/− mice. B6.Sle1b male mice expressing Yaa (Sle1b^Yaa) developed splenomegaly, which was reversed to a level observed in B6.Sle1b males in the absence of IFN-γR (Fig. 3A). Elevated GC B cell (Fig. 3B, 3D, 3E) and Tfh cell (Fig. 3C) responses in Sle1b^Yaa mice were significantly reduced in Sle1b^YaaIFN-γR^−/− mice. Similarly, the number of dsDNA- and nucleosome-specific splenic AFCs (Fig. 3F), serum dsDNA-, nucleosome- and SmRNP-specific Abs (Fig. 3G), and serum ANA reactivity (Fig. 3H) were significantly reduced in Sle1b^YaaIFN-γR^−/− mice compared with Sle1b^Yaa control mice. Finally, IC deposition in the kidney and the developed GN in Sle1b^Yaa mice were also diminished in Sle1b^YaaIFN-γR^−/− mice (Fig. 3I–K). Our data from TLR7-induced and -overexpression SLE models indicate that IFN-γ signaling plays a nonredundant role in TLR7-promoted AFC and GC responses and subsequent autoantibody production and SLE pathogenesis, whereas T1IFN signaling only moderately contributes to these responses.

To investigate how IFN-γ may promote autoreactive B cell development into the AFC and GC pathways in TLR7-accelerated SLE, we analyzed the gene transcripts in spleens of IMQ-treated B6.Sle1b and Sle1b.IFN-γR^−/− mice by RNA-seq. Clustering of differentially expressed genes showed a distinct transcriptional profile in Sle1b.IFN-γR^−/−-IMQ mice compared with B6.Sle1b-IMQ control mice. We found that 130 and 213 genes were up and downregulated, respectively, in Sle1b.IFN-γR^−/−-IMQ mice (p < 0.05; fold change ≥1.5; Fig. 4A). Functional classification of the differentially expressed genes with gene ontology analysis revealed many biological processes or the pathways that were downregulated (Fig. 4B) or upregulated (Fig. 4C) in the absence of IFN-γR signaling. Genes differentially regulated between B6.Sle1b-IMQ and B6.Sle1b.IFN-γR^−/−-IMQ splenic tissues were associated with gene transcription regulation (Fig. 4D), ubiquitination/proteasome complex (Fig. 4E), cell proliferation and migration (Fig. 4F), autoimmunity (Fig. 4G), inflammation (Fig. 4H), cellular and oxidative stress (Fig. 4I), and cellular metabolism (Fig. 4J). These results indicate that IFN-γ signaling promotes TLR7-mediated autoimmunity by regulating multiple biological processes.

Considering the increased expression of the Ifn-γ gene in splenic tissue (Fig. 1I) and regulation of multiple signaling pathways by IFN-γ signaling (Fig. 4) in IMQ-treated mice, we next determined the cellular source of IFN-γ in TLR7-mediated autoimmunity. We treated IFN-γ–eYFP reporter mice with IMQ for 6 wk and found that CD4⁺ T cells were the major producers of IFN-γ, although CD8⁺ T cells and NK cells produced this cytokine to a certain extent (Fig. 5A). Of the CD4⁺CD44^hi effector T cells, a higher frequency of pre-Tfh and Tfh cells produced IFN-γ than non-Tfh cells upon IMQ treatment (Fig. 5B–D). Dendritic cells, macrophages, eosinophils, monocytes, and neutrophils did not produce IFN-γ in these mice (Fig. 5E). These data together with increased IFN-γR expression on B cells upon TLR7 stimulation ex vivo (Fig. 1J) point to an important role for TLR7 in regulating the IFN-γ–IFN-γR pathway in B cells by inducing T cell IFN-γ production and upregulating IFN-γR on B cells.

To determine how TLR7-induced and T cell–produced IFN-γ may signal in B cells for regulating autoimmune AFC, GC, and Tfh responses and promoting SLE development, we conditionally deleted the Ifngr1 gene in peripheral B cells in B6.Sle1b mice (designated Sle1bIFN-γR^fl/fl.CD23^Cre+). IMQ-treated Sle1bIFN-γR^fl/fl.CD23^Cre+ mice did not show splenomegaly that was observed in IMQ-treated Sle1bIFN-γR^fl/fl control mice (Fig. 6A, 6B). Sle1bIFN-γR^fl/fl.CD23^Cre+ mice also had markedly reduced B220⁺GL-7^hiCD95^hi GC B cell and CD4⁺CD44^hiCXCR5^hiPD-1^hi Tfh cell responses, including reduced frequency and size of GCs, compared with Sle1bIFN-γR^fl/fl mice (Fig. 6C–G). Sle1bIFN-γR^fl/fl.CD23^Cre+ mice also had diminished serum ANA reactivity (Fig. 6H), reduced serum autoantibody titers (Fig. 6I), and splenic autoantibody-producing AFCs (Fig. 6J). Dampened AFC, GC, and Tfh responses in Sle1bIFN-γR^fl/fl.CD23^Cre+ mice were strongly associated with diminished IC deposition in the kidney and with significantly improved kidney inflammation such as GN, interstitial nephritis, and vasculitis score (Fig. 6K–M). These data highlight the nonredundant role of IFN-γ signaling in B cells in TLR7-mediated AFC, GC, and Tfh responses and the development of autoreactive B cells, autoantibodies, and lupus nephritis.

We next asked how much of a role STAT1 that functions downstream of IFN-γ signaling played in TLR7-driven GC, Tfh, and IgG Ab responses under a scenario in which several pathways (NF-κB, IRF5/7, and P38/MAPK) can be potentially activated by the strong in vivo TLR7 stimulation. We first treated B6 and B6.STAT1^−/− mice with IMQ and found that STAT1 is absolutely required for the development of TLR7-driven GC B cell, Tfh cell, and plasma cell responses (Fig. 7A–C). Next, by using conditional knockout mice with peripheral B cells deficient in STAT1 (STAT1^fl/flCD23^Cre+), we found the requirement of B cell–intrinsic STAT1 in TLR7-mediated GC, Tfh, plasma cell, and IgG Ab responses (Fig. 7D–F). Consistent with these data, B cells stimulated with TLR7 ligand had much higher STAT1 expression than unstimulated B cells (Fig. 7G). Similar to the increased STAT1 expression, T-bet downstream of STAT1 was also upregulated in B cells stimulated with TLR7 ligand (Fig. 7H). To solidify the importance of B cell–intrinsic T-bet in TLR7 driven GC, Tfh, and Ab responses, we treated T-bet^fl/flCD23^Cre+ mice where peripheral B cells are deficient in T-bet. We observed a crucial B cell–intrinsic role of T-bet in TLR7-mediated GC, Tfh, and IgG Ab responses (Fig. 7I–K). These data indicate that STAT1 and T-bet downstream of TLR7 and IFN-γ signaling play a significant role in TLR7-driven GC, Tfh, and IgG Ab responses, which cannot be overcome by other pathways potentially activated by TLR7 stimulation.

Cellular and molecular mechanisms by which TLR7 overexpression or overactivity promotes autoreactive B cell development in the extrafollicular AFC and follicular GC pathways that are activated in SLE remain largely unknown. In this study, using the TLR7-induced and TLR7-overexpression models of SLE, as well as an IFN-γ reporter mouse model, we demonstrate that TLR7 overactivation drives development of autoreactive B cells in both the AFC and GC pathways and promotes SLE by driving predominantly IFN-γ production by T cells and mediating IFN-γ signaling in B cells. TLR7-driven and T cell–produced IFN-γ controls gene transcription in multiple cellular pathways important for immune cell survival, proliferation, metabolism, and autoimmunity.

Several groups, including ours, previously identified a B cell–intrinsic role of TLR7-MyD88 signaling in promoting autoimmune AFC and GC responses, leading to SLE pathogenesis in the mouse models (5, 11, 16, 17, 47). We and others also reported B cell–intrinsic roles of T1IFN and IFN-γ signaling in regulating SLE-associated AFC, GC, and autoantibody responses in SLE-prone mice (4, 25, 36). Although TLR7 signaling is believed to drive T1IFN production and activation of T1IFN signaling in SLE (19, 45, 48, 49), how TLR7-driven IFN-γ production and activation of IFN-γ signaling in B cells may promote TLR7-mediated autoimmune AFC and GC responses and SLE development has previously not been addressed. Using the TLR7-induced SLE model and B cell–specific conditional deletion of the Ifngr1 gene, we found that IFN-γ signaling in B cells was essential for TLR7-mediated AFC, GC, and Tfh responses and GN development, whereas global T1IFN signaling deficiency moderately reduced these responses and was associated with less kidney pathology. Using IFN-γ reporter mice, we further established that TLR7 induced IFN-γ production primarily by CD4⁺ T cells, including pre-Tfh and Tfh. Together, our data demonstrate that TLR7-driven and T cell–produced IFN-γ act nonredundantly to activate IFN-γ signaling in B cells, promoting development of autoreactive B cells and SLE pathogenesis.

Although MRL/lpr and NZW/NZBF1 lupus-prone mice deficient in IFN-γ or treated with anti–IFN-γ blocking therapy had reduced autoantibody production and renal disease (30–35), the role of IFN-γ signaling in autoreactive B cell development in either the extrafollicular or the GC pathways in these mice was not tested. David Rawlings’ and our groups previously discovered the role of IFN-γ signaling in regulating SLE-associated GC responses; however, the link between TLR7 and IFN-γ signaling in autoreactive B cell expansion in the GC pathway was not directly tested in these studies using a TLR7-induced SLE mouse model (4, 36). In this study, we find a significantly reduced number of autoantibody-producing AFCs and dampened GC responses in the spleens of TLR7-induced SLE-prone mice. Our current data highlight the importance of B cell–intrinsic IFN-γ signaling downstream of TLR7 signaling in regulating autoreactive B cell development in both the extrafollicular and GC pathways, leading to TLR7-promoted autoantibody production and SLE development.

Synergy between TLR-MyD88 and IFN-γ–STAT1 signaling in potentiating inflammatory responses in macrophages has previously been described (50). IFN-γ–STAT1 signaling was found to aid TLR-mediated inflammatory cytokine production by regulating cellular metabolism (51) and gene transcription through chromatin remodeling of the inflammatory gene promoter and enhancer regions (52). MyD88 was also shown to directly interact with IFN-γR to stabilize the IFN-γ–induced cytokine and chemokine mRNA (53). In addition, STAT1 was shown to interact with cell cycle regulators cyclin D1 and Cdk4 (54). Consistent with these previous reports, we found that global and B cell–specific deficiency of STAT1 resulted in reduced GC B cell and Tfh responses upon treatment with a TLR7 ligand, highlighting the importance of STAT1 in B cells in TLR7-mediated GC and Tfh responses. Published data and our current findings collectively suggest that TLR7-MyD88 and IFN-γ–STAT1 signaling may synergize to activate multiple biological processes in B cells that are important for mediating SLE-associated AFC, GC, and Tfh responses in TLR7-accelerated SLE autoimmunity.

Although T-bet has been associated with age-associated B cells and class switching to the IgG2a/c isotype (55–57), we found significantly reduced TLR7-driven GC and Tfh responses in B cell conditional T-bet–deficient mice. Rubtsova et al. (58) recently showed a B cell–intrinsic role of T-bet in B cell differentiation into age-associated B cells, Spt-GC B cells, and plasma cells and which subsequently increased kidney pathology. In addition, CD11c^hiT-bet⁺ B cells have recently been described in SLE patients (59). An earlier report showed that TLR7 stimulation of B cells upregulated T-bet and that T-bet expression was elevated in 564Igi autoreactive B cells (60). Rubtsova et al. (61) similarly found that BCR and TLR7 signaling synergize with IFN-γ signaling to drive T-bet expression in B cells. The Cancro group recently reported that T-bet expression is induced in B cells by TLR signaling but not to BCR and/or anti-CD40 stimulation alone (62). Published data and our current observation together suggest that TLR and IFN-γ–STAT1 signaling cooperate to drive T-bet expression in B cells, which is important for GC, Tfh, and Th1-biased IgG2c Ab responses. Consistent with this idea, we found upregulation of STAT1 and T-bet in TLR7 stimulated B cells, implicating the involvement of both T-bet and STAT1 in TLR7-mediated AFC, GC, and SLE autoimmune responses.

TLR7 can potentially activate multiple signaling pathways (e.g., NF-κB, IRF5/7, and p38/MAPK) to promote autoimmunity (63). Both canonical and noncanonical NF-κB pathways have been shown to be important for foreign Ag-driven AFC and GC responses (64, 65), although the role of NF-κB in spontaneous or TLR7-mediated AFC and GC responses in SLE is unclear. Altered NF-κB activity was previously reported in B cells and T cells from SLE patients (66). IRF5 deficiency abrogates disease development in the FcγRIIB^{−/− Yaa} model of SLE (67). Despite the potential activation of these pathways in TLR7-induced and TLR7-overexpression models of SLE, we found that IFN-γR and STAT1 signaling in B cells played an important role in TLR7-mediated AFC and GC responses. Given the importance of both TLR7 and IFN-γ–STAT1 signaling in autoimmune AFC and GC responses, we postulate that IFN-γ–STAT1 signaling in B cells may co-operate with the NF-κB or IRF5/7 pathway to orchestrate a genetic program required for autoimmune AFC and GC responses in TLR7-accelerated SLE. Cooperation of NF-κB and STAT1 in the transcriptional control of anti-microbial genes in response to in vitro Listeria monocytogenes infection has previously been described (68).

Our RNA-seq data derived from splenic tissues of IMQ-treated Sle1b.IFN-γR^−/− and B6.Sle1b mice revealed differential expression of multiple genes and pathways involved in the autoimmune response. Specifically, multiple biological processes are disrupted in the absence of IFN-γR signaling such as cell proliferation and migration, inflammation, and cellular metabolism that can contribute to the development of autoimmunity. In support of the overall reduction in the number of autoantibody-producing AFCs in Sle1b.IFN-γR₁^−/− mice, the xbp-1 gene that encodes for the transcription factor (XBP-1) and is required for plasma cell differentiation and autoimmune responses (69, 70), is downregulated in these mice. IL-6 signaling is involved in autoimmune GC formation and autoimmune disease progression likely by increasing the production of IFN-γ (71, 72). Interestingly, il6r gene expression is downregulated in Sle1b.IFN-γR₁^−/− mice. In addition, the relationship between IL-6 and IFN-γ signaling in autoimmunity is further supported by the upregulation of the Smyd2_TSS3427 gene in Sle1b.IFN-γR^−/− mice, a gene that encodes a histone methyltransferase that negatively regulates IL-6 production (73). Finally, genes encoding chemokine receptor Cxcr3 and its ligand Cxcl9 that play an important role in autoimmunity development (74–76) are downregulated in the absence of IFN-γR signaling. Overall, our data indicate that TLR7 stimulation induces expression of genes that are associated with autoimmune responses in an IFN-γR signaling dependent manner.

In summary, our data define a previously unrecognized indispensable role for IFN-γ signaling and its downstream transcription factors STAT1 and T-bet in B cells in promoting TLR7-accelerated AFC and GC responses and SLE autoimmunity. In contrast, we find moderate effects of T1IFN signaling on these processes. Our current T1IFN-signaling data reflect the moderate efficacy of the clinical trials for anti–IFN-α or anti–IFN-αR therapy in SLE patients. Future treatment approaches could aim to reduce TLR7 signaling in combination with T1IFN or IFN-γ signaling to a threshold that can prevent or treat SLE patients and maintain protection against pathogenic infection.

We thank the Pennsylvania State University Hershey Medical Center flow cytometry core facility for assistance. We thank the Pennsylvania State University Hershey Medical Center Department of Comparative Medicine for animal housing and care. We thank Dr. Meinrad Busslinger for providing CD23^Cre mice.

The authors have no financial conflicts of interest.