Partial Protection From Lupus-Like Disease by B-Cell Specific Type I Interferon Receptor Deficiency

Systemic lupus erythematosus (SLE) is an autoimmune disease that can present with many different permutations of symptom presentation. A large subset of SLE patients have been shown to present with elevated interferon stimulated gene (ISG) expression, and Type I IFNs (IFNαβ) have been shown to drive disease in murine models through global IFNα Receptor (IFNAR) knockouts. However, the disease contribution of distinct immune cell subsets in response to constitutively increased levels of IFNαβ is not fully understood. We utilized a B-cell specific IFNAR knockout (BΔIFNAR) on the B6.Nba2 spontaneous-lupus background to determine the contribution of IFNαβ stimulated B cells in disease. We found that IFNαβ signaling in B cells is driving increased splenomegaly, increased populations of activated B cells, and increased populations of germinal center (GC) B cells, memory B cells, and plasma blasts/cells, but did not affect the development of glomerulonephritis and immune-complex deposition. IFNAR expression by B cells also drove production of anti-chromatin IgG, and anti-dsDNA and -nRNP IgG and IgG2C auto-antibody levels, as well as increased Bcl2 expression, affecting GC B cell survival in B6.Nba2 mice.


INTRODUCTION
Systemic Lupus Erythematosus (SLE) is a chronic autoimmune disease. SLE patients have a lower quality of life when compared to that of patients with other chronic illnesses (1). Treatment of SLE has advanced in the recent past with the FDA approval of Belimumab in 2011, the first SLE drug approved in 50 years (2). However, one remaining critical obstacle in the field is the heterogeneity of symptom presentation between SLE patients. For physicians using the American College of Rheumatology SLE diagnosis criteria, there are 330 different permutations of symptom presentations that can be diagnosed as SLE (3). As the field moves toward more targeted treatment strategies, it is imperative to understand which presentations of SLE could be treated with a particular therapy, and which presentations are not likely to respond to the therapy.
It has been known for decades that many SLE patients present with elevated levels of type I interferons (IFNab) (4)(5)(6). In 2003, multiple publications showed that SLE patient PBMCs had a significant increase in IFNab-stimulated gene expression (7)(8)(9)(10). Combined with the knowledge that interferon treatment for non-SLE conditions may lead to the development of autoimmune responses and SLE symptoms (11,12), interferons quickly became an area of study in SLE. As such, blocking the type I IFN receptor (IFNAR) was shown to inhibit IFNab production and decrease the development of plasma cells in patients (13,14). Furthermore, according to data presented at the 2020 EULAR E-Congress of Rheumatology, phase III trials of Anifrolumab, a monoclonal antibody that blocks type IFNAR, have been successful in reducing flares and BICLA (BILAGbased Combined Lupus Assessment) scores in patients (14)(15)(16).
The role of IFNab in SLE has been confirmed by lupus-like disease amelioration in global IFNAR knockouts on a multitude of murine lupus backgrounds (17)(18)(19)(20)(21). In the B6.Nba2 spontaneous lupus mouse model, global IFNAR deficiency leads to decreased autoantibody production, a reduction in serum IgG, reduced splenic Ifi202 expression and decreased splenomegaly (17). Similarly, IFNAR-deficient New Zealand Black (NZB) mice displayed significantly reduced auto-antibodies and a significant reduction in splenomegaly (19,22). However, despite the overwhelming evidence supporting blocking IFNAR in SLE, the fact that IFNAR is expressed by most cells of the body and stimulation contributes a wide array of anti-viral responses, suggests that IFNAR blockade could have unwanted side effects. Moreover, as not all patients present with elevated IFNab levels, it is unlikely that such therapy will benefit all patients. To determine which aspects of disease and symptom presentation are IFNabdriven through B cells, we generated B cell-specific IFNAR-/-lupusprone B6.Nba2 mice (BDIFNAR). Mice were followed for four months to determine the primary effects of B cell-specific IFNARdeficiency on cellular and serological measures of lupus-like disease observed in B6.Nba2 mice.

Organ Harvest and Processing
B6, B6.BDIFNAR, B6.Nba2.BDIFNAR and B6.Nba2 female mice were harvested at 4 months of age in several separate cohorts. Spleens were harvested and weighed. A middle portion of the spleen was frozen in OCT for immunohistochemistry. A small portion of the spleen was made into a single cell suspension in 1x phosphate buffered saline (PBS) pH7.4 and incubated at 37°C for 10 mins for isolation of spleen supernatant and single cells for RNA extraction. The remaining spleen was similarly processed into single cells after which red blood cells (RBCs) were lysed using ACK (Ammonium-Chloride-Potassium) lysis buffer (0.15 M NH4Cl, 0.01 M KHCO3, 0.2 mM EDTA, pH 7.3) and the cells were used for flow cytometry. Bone marrow from the left femur and left tibia was flushed out with 1x PBS and used for flow cytometry after RBC lysis as described above. One kidney was removed and dissected into two coronal sections. One half was preserved in OCT, the other half was fixed in 10% formalin.
Where indicated, specific cell subsets were isolated by cell sorting on a BD FACSAriaII. All cell subsets were gated as shown in Supplemental Figure 1.

Kidney Immunohistochemistry
Formalin-fixed kidneys were transferred to 80% ethanol and embedded in paraffin. Three µm sections were cut and stained with hematoxylin/eosin (Histology Core, Lerner Research Institute) or Periodic acid shiff (PAS) (Thermo Scientific, Waltham, MA, USA) for detection of kidney morphology, damage and cellular infiltrates. Sections were scored in a blinded fashion by a renal pathologist (JN) at the Cleveland Clinic. Kidneys were evaluated on a scale of 0-5 for mesangial and endothelium hypercellularity. Glomerular area was calculated by measuring the area of 5-15 individual glomeruli per section for each mouse. Immunofluorescence was performed on 5µm sections of kidneys preserved in OCT ™ (Scigen, Paramount, CA, USA). Sections were thawed, fixed in acetone, blocked with 10% normal goat serum, and stained using Texas-Red conjugated anti-IgG or anti-IgM antibodies (SouthernBiotech, Birmingham, AL, USA) and FITC-conjugated anti-Complement factor C′3 antibodies (Immunology Consultants Laboratory Inc., Portland, OR, USA). Imaging was done on a Keyence BZ-X700 All-in-one microscope (Keyence, Osaka, Osaka, Japan) and images were quantified using the Keyence BZ-X analysis software (Keyence, Osaka, Osaka, Japan). A total of 5-15 glomeruli were quantified for each section and the average calculated per mouse.

Spleen Immunohistochemistry
For immunofluorescent staining of spleens, 5µm sections were cut of spleens preserved in OCT ™ compound. On day one of staining, sections were thawed, fixed in acetone, blocked with 10% normal goat serum, and stained using FITC conjugated anti-B220 IgG2a (eBioscience, San Diego, CA, USA) and biotin conjugated anti-GL-7 IgM antibodies (eBioscience, San Diego, CA, USA) followed by Texas-Red conjugated streptavidin (Southern Biotech, Birmingham, AL, USA). Imaging and quantification was done using the same process as for kidney immunofluorescence.
Quantitative Reverse-Transcriptase-PCR (qRT-PCR) RNA was isolated from frozen splenocytes using the RNeasy mini kit (Qiagen, Valencia, CA, USA) and quantified using nanodrop technology (Nanodrop ND-1000 Spectrophotometer, Thermo Fisher). Complementary DNA (cDNA) was made using qScript ™ DNA supermix (Quanta BioSciences, Gaithersburg, MD) and quantified using nanodrop technology and diluted for qPCR. qPCR was performed with 100ng cDNA using SYBR ™ green PerfeCTa ® SYBR ® Green FastMix ® ROX (Quanta BioSciences) and run on a Step One Plus real time PCR system (Applied Biosciences, Foster City, CA, USA). All transcripts were analyzed using b-actin expression as a control. Primers

Anti-Chromatin IgG Enzyme-Linked Immunosorbent Assay
2HB plates (Immunlon, Thermo Fisher Scientific, Waltham, MA, USA) were coated with purified chromatin overnight as previously described (17). The following day plates were blocked with 5% gelatin in PBS for ≥2 h. Serum samples were diluted 1:300 in serum diluent (0.5% g-globulins and 5% gelatin in 1x PBS tween). HRPconjugated anti-IgG antibody (Southern Biotech, Birmingham, AL, USA) was added and plates were developed with TMB substrate kit (Thermo Scientific, Waltham, MA, USA) and OD 450nm was read using a Victor 3 plate reader (Perkin Elmer, Waltham, MA, USA).

Statistics
Statistical analyses were performed in GraphPad Prism (La Jolla, CA, USA). All comparisons between two groups were done by Student's T test with Welch's correction. Statistical significance is defined as p < 0.05.

B Cells From B6.Nba2.BDIFNAR Mice Fail to Respond to IFNab
To study how IFNab stimulation of B cells specifically alters lupus-like disease and contributes to symptom presentation, we created B cell-specific IFNAR-deficient mice (BDIFNAR) on the lupus-prone B6.Nba2 background ( Figure 1A). To functionally verify our model we stimulated primary splenocytes with murine recombinant IFN-aA overnight, and measured cell surface markers and intracellular Ifit2, an IFNab-induced protein, using flow cytometry (28). B cells from B6.Nba2.BDIFNAR mice did not show production of Ifit2 in response to IFN-aA stimulation, whereas B cells from WT B6.Nba2 lupus-prone control mice upregulated Ifit2 levels by~25 fold ( Figure 1B). This effect was specific to B cells, as both T cells and dendritic cells from B6.Nba2.BDIFNAR mice upregulated Ifit2 expression after stimulation ( Figure 1B and data not shown).

B6.Nba2.BDIFNAR Mice Express Normal Frequencies of Developing B Cells
To confirm that B6.Nba2.BDIFNAR mice were not displaying B cell developmental defects, frequencies of early B progenitor cells, pro-B cells, pre-B cells and immature B cells were determined. No d iff er e nc es we re id e nt ifi ed b et wee n B 6.N ba2 an d B6.Nba2.BDIFNAR mice (Supplemental Figure 2). Interestingly, we observed significantly reduced levels of total splenic pDCs in B6.Nba2.BDIFNAR mice, however levels of interferon-producing SiglecH+ pDCs were unchanged ( Figures 1C, D), suggesting that IFNab levels would be unaffected. Studies of the expression of interferon stimulated genes (ISGs) in spleen and peripheral blood m o n o n u c l e a r c e l l s ( P B M C s ) f r o m B 6 . N b a 2 a n d B6.Nba2.BDIFNAR mice confirmed that IFNab levels were indeed unchanged in these mice ( Figures 1E-G).
Serum Anti-Nuclear Autoantibodies Are Reduced in B6.Nba2.BDIFNAR Mice Hypergammaglobulinemia and elevated levels of serum ANA are hallmarks of B6.Nba2 lupus-like disease (22,25). We compared serum antibodies in 4 months old B6.Nba2 and B6.Nba2.BDIFNAR mice and found no significant differences in the levels of serum IgM  Figure 2E). There were no differences in total serum IgG subclasses ( Figure 2F), however, anti-dsDNA and anti-nRNP IgG 2c antibodies, but not other IgG subclasses, were decreased in serum of B6.Nba2.BDIFNAR mice ( Figures 2G, H).   (29). We found no difference in mesangial hypercellularity and glomerular area between B6.Nba2.BDIFNAR mice and B6.Nba2 controls at 4 months of age (Supplemental Figure 3). Similarly, we found no difference in the glomerular deposition of IgG-or IgM-IC and no difference in complement C′3 fixation (Supplemental Figure 3 and data not shown). Surprisingly, we did not also see differences in total IgG2c-IC between B6.Nba2 and B6.Nba2.BDIFNAR mice (Supplemental Figure 3).

B6.Nba2.BDIFNAR Mice Display Reduced Splenomegaly and Reduced Activation of B Cells
Splenomegaly, a pronounced presentation of B6.Nba2 lupus-like disease, was quantified and found to be significantly reduced in B6.Nba2.BDIFNAR mice as compared with B6.Nba2 age-matched mice ( Figure 3A). Further analyses of the spleen composition, revealed no significant differences in the frequency of total B cells, total T cells, macrophages, conventional DCs (cDC) and neutrophils ( Figures 3B-D and data not shown). As previously mentioned, total pDCs were significantly reduced in BDIFNAR mice while the subset of IFNab producing pDCs was unchanged (see above, Figures 1C, D). Levels of transitional B cells ( Figure  3E) were similarly unchanged, as were transcriptional levels of Baff and splenocyte secreted BAFF ( Figures 3F, G). Populations of follicular mature B cells and marginal zone B cells also remained unchanged (Supplemental Figures 4A, B). However, frequencies of CD69+ and CD40+ B cells were significantly reduced in B6.Nba2.BDIFNAR mice (Figures 3H, I). This was not due to a general reduction in the level of cellular activation, as the CD69+ activated T cell population was unaffected ( Figure 3J) Figures 4A, B). This reduction was specific to the B6.Nba2 model as levels of PB/ PC and memory B cells were unchanged between B6 and B6.BDIFNAR mice (Figures 4A, B). It should be noted that the decrease in splenic PB/PC in B6.Nba2.BDIFNAR mice was not due to an increase in homing to the bone marrow, as bone marrow PB/PC levels were unchanged (Supplemental Figure  1E). Interestingly, we found no significant differences in splenic transcript levels of Blimp ( Figure 4C) and Irf4 ( Figure 4D) both of which are required for plasma cell differentiation (30)(31)(32), suggesting that the differentiation process to PB/PC and memory B cells was unaffected by IFNAR-deficiency.

B6.Nba2.BDIFNAR Mice Display Reduced Germinal Center (GC) B Cells With No Change in T Follicular Helper (TFH) Cells
We quantified the GC B cells to determine if the significant decrease in PB/PC and memory B cells was due to reduced GC B cell population. Flow cytometric analyses identified a significant   Figure 4E), although immunohistochemistry analyses did not support differences in GC area ( Figures 4F, G). The reduction in GC B cell numbers, was not due to changes in the TFH cell population ( Figure 5A) or in levels of IL-6, IL-1b, and IL-21, known to be involved in the GC reaction ( Figures 5B-D). Similarly, expression levels of Bcl6, a crucial transcription factor for TFH and GC B cells (33)(34)(35), was unchanged ( Figure 5E).

IFNAR-Deficiency Results in Increased Apoptosis of GC B Cells
IFNa signaling has been found to drive Bcl2 expression in lymphocytes, suggesting that a lack of IFNa signaling may affect B cell survival (36). We evaluated expression of the anti-apoptotic factors Bcl2 and Bclxl in total splenocytes and in sorted GC B cells and PB/PC. While there was no change in expression in total splenocytes ( Figure 6A), GC B cells from B6.Nba2.BDIFNAR mice showed decreased expression of Bcl2 and Bclxl (p < 0.05 and p = 0.1428, respectively), and a trend towards elevated expression of Bim (p = 0.1202) ( Figure 6B). In contrast, there were no differences in the expression level of either factor in PB/PC cells ( Figure 6C). In support of a role for IFNa in regulating Bcl2, splenic B cells from the B6.Nba2 mice showed increased intracellular Bcl2 protein ex vivo in response to recombinant IFN-aA stimulation, while B cells from B6.Nba2.BDIFNAR mice were unable to upregulate intracellular Bcl2 in response to the same conditions (data not shown).

DISCUSSION
IFNab is best known as an anti-viral response (37,38), but it also plays a significant role in murine and human SLE (4)(5)(6)(7)(17)(18)(19)(20). This has been shown through increased ISG expression in patients (7)(8)(9)(10) and protection from murine lupus-like symptom development in global IFNAR knockouts (17)(18)(19)(20)25). Still, the specific response to IFNab by individual immune cells and their relative contribution to disease characteristics during lupus  by B cells contributes to increased levels of anti-chromatin IgG, anti-dsDNA and anti-nRNP IgG and IgG2c autoantibodies in lupus-prone B6.Nba2 mice, without significantly increasing total levels of IgG or IgM and without skewing total IgG subtype presentation. Similar to the lack of changes in glomerular IgG2c deposition observed in B6.Nba2 global IFNAR knockout mice (17), the reduced levels of anti-dsDNA and anti-nRNP specific IgG2c autoantibodies in the serum of B6.Nba2.BDIFNAR mice was not accompanied by a decrease in IgG2c deposition in the kidney glomeruli. A few studies have previously described a specific effect of IFNAR signaling in driving immunoglobulin class switching to IgG2c (or IgG2a) (39,40), however the implication of specific IgG subclass presentation remains unclear in lupus-like disease settings.
Corresponding with reduced levels of anti-chromatin and anti-dsDNA IgG and IgG2c specific autoantibodies, B6.Nba2.BDIFNAR mice also displayed significantly decreased populations of germinal center B cells, memory B cells and PB/PCs as compared to B6.Nba2 mice, highlighting a role for IFNab stimulation to positively affect the germinal center reaction. Surprisingly, we did not observe any difference in GC areas within the spleen. The reason for this is currently unknown, but the fact that other GC cell types, including TFH cells, were unchanged suggest that the process of GC formation was unaltered and the observed effect specific to GC B cells. It should also be noted that levels of GC B cells in B6.Nba2.BDIFNAR mice remained significantly higher than levels in B6.BDIFNAR mice, while memory B cells and PB/PCs were actually lower in B6.Nba2.BDIFNAR mice as compared with B6.BDIFNAR mice. This suggests that both IFNab-dependent and IFNab-independent factors drive elevated numbers of GC B cells in B6.Nba2 mice, while IFNab is specifically affecting the accumulation of post-GC cells.
To determine if removal of IFNAR on B cells affected other known mechanisms involved in driving a GC reaction, we analyzed intra-splenic levels of several factors, including IL-6, IL-21, and IL-1b known to be involved in the GC reaction. Of these cytokines, IL-6 and IL-21 were of particular interest, as IL-6 has been associated with increased cell growth and survival in a number of in vitro studies (41)(42)(43)(44) and affect GC B cells indirectly via differentiation of TFH cells in vivo (45) while IL-21 is known to drive both TFH and GC B cell populations in a virus-infection model (45). Neither IL-6 nor IL-21 were differentially present in spleens of B6.Nba2 and B6.Nba2.BDIFNAR mice. Furthermore, we did not observe differences in the levels of TFH between the mice, suggesting that the observed changes were due to intrinsic effects in the GC B cells themselves subsequently affecting post-GC B cell populations including memory B cells and PB/PCs. While IFNab alone can directly increase Bcl2 expression (36,46), the increased expression of Bcl2 in GC B cells in the B6.Nba2 model may be compounded by the increased expression of CD40 in the model. CD40 expression by B cells is required for GC formation, Ig isotype class-switching, and sustained production of antibodies, even in the presence of other CD40 expressing APCs (47). Interestingly, ex vivo cross-linking of CD40 on human B cells by anti-CD40 specific antibodies also upregulated expression of the anti-apoptotic factor BCL2 in the presence of IL-21 (48), thus directly affecting B cell survival. Similarly, concomitant stimulation of CD40 and the B cell receptor induced increased expression of anti-apoptotic BCL2 family members in human tonsil GC B cells (49), further supporting a potential survival contribution from CD40 expression in B cells. We observed a significant decrease in CD40+ B cells and reduced expression of Bcl2 and Bclxl mRNA in GC B cells from B6.Nba2.BDIFNAR. Based on our data presented here, it is reasonable to hypothesize that in the B6.Nba2 lupus-like model IFNab may be increasing B cell activation, leading to increases in CD40+ B cells. CD40 expression and IFNab itself may in combination lead to increased survival of GC B cells by increasing pro-apoptotic factor Bcl2. This survival advantage could subsequently contribute to increased differentiation into memory B cells and PB/PCs and elevated production of auto-antibodies. Further studies are needed to establish if altered GC B cell survival is the sole effect of IFNab-induced signaling in B cells in this model of lupus. In summary, in a disease where symptom presentation is extremely heterogeneous, providing effective treatment for each patient means understanding the disease and drivers of symptoms. While we know that blocking IFNAR globally drastically reduces murine lupus-like disease, this approach will also affect IFNab responses elicited by viral and/or bacterial infections. In support thereof, Anifrolumab (anti-IFNAR antibody) treatment, while successfully reducing levels of plasma cells (13,14) and anti-dsDNA antibody levels in a Phase IIb trial (50), resulted in reactivation of herpes simplex virus in a significant number of patients (50). Finally, our studies suggests that therapies that target IFNAR, such as Anifrolumab, or IFNab may be more successful in patients who have higher levels of autoantibodies, such as anti-dsDNA antibody levels, or increased B cell activation. Understanding how IFNab affect additional cell subsets as these contribute to disease and symptom presentation, may allow us to identify better therapeutic targets, combinatorial treatments or cellspecific therapies to more specifically treat individual patients.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
All mice were maintained in the Biological Research Unit at the Lerner Research Institute, in accordance with Cleveland Clinic Foundation Animal Research Committee guidelines. Animal studies were approved by the Institutional Animal Care and Use Committee of the Lerner Research Institute of the Cleveland Clinic Foundation and conducted in compliance with guidelines issued by the National Institutes of Health.

AUTHOR CONTRIBUTIONS
EK harvested the mice, and performed all in vitro studies, all ELISAs, and all flow cytometry. NP contributed to staining, imaging, and quantification of the kidneys and spleen, as well as genotyping and mouse care. JN blindly scored pathology for H&E and PAS kidney samples. MP contributed to the breeding and genotyping of the mice and to sectioning frozen tissues for immunohistochemistry. TJ designed and guided the study. The manuscript was written by EK and TJ. All authors contributed to the article and approved the submitted version.

FUNDING
Funding for this study was provided by NIH R01AI118774 (TNJ).  Stains were used for detection of renal morphology, glomerulonephritis, immune complex deposition and complement fixation. (E) Kidneys were scored using both H&E and PAS stains by a blinded pathologist for mesangial hypercellularity. (F) Glomerular area was measured and calculated using the H&E staining. (G, H) IgG and IgG2c deposition were measured using Image Pro Software and Keyence software respectively. Two sections, representing > 10 glomeruli were analyzed per mouse for each evaluation. Each symbol represents the average score per mouse and data are shown as Mean ± SEM. n = 11 (B6.Nba2); n = 10 (B6.Nba2.BDIFNAR). * p < 0.05, Student's unpaired t-test with Welch's correction.