Journal of Autoimmunity 61 (2015) 9e16

Contents lists available at ScienceDirect

Journal of Autoimmunity journal homepage: www.elsevier.com/locate/jautimm

Deleting the BAFF receptor TACI protects against systemic lupus erythematosus without extensive reduction of B cell numbers William A. Figgett a, Devy Deliyanti a, Kirsten A. Fairfax a, b, c, Pin Shie Quah a, Jennifer L. Wilkinson-Berka a, Fabienne Mackay a, * a b c

Department of Immunology, Central Clinical School, Monash University, Melbourne 3004, Australia The Walter and Eliza Hall Institute of Medical Research, Molecular Medicine Division, 1G Royal Parade, Parkville 3052, Australia Department of Experimental Medicine, University of Melbourne, Parkville 3052, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 27 March 2015 Received in revised form 21 April 2015 Accepted 26 April 2015 Available online 29 May 2015

B cell-activating factor of the TNF family (BAFF) is an essential B cell survival factor. However, high levels of BAFF promote systemic lupus erythematosus (SLE) in mice and humans. Belimumab (anti-human BAFF) limits B cell survival and is approved for use in patients with SLE. Surprisingly, the efficacy of rituximab (anti-human CD20) in SLE remains controversial, despite depleting B cells more potently than belimumab. This raises the question of whether B cell depletion is really the mechanism of action of belimumab. In BAFF transgenic mice, SLE development is T cell-independent but relies on innate activation of B cells via TLRs, and TLR expression is modulated by the BAFF receptor TACI. Here, we show that loss of TACI on B cells protected against BAFF-mediated autoimmune manifestations while preserving B cells, suggesting that loss of BAFF signaling through TACI rather than loss of B cells may underpin the effect of belimumab in the clinic. Therefore, B cell-sparing blockade of TACI may offer a more specific and safer therapeutic alternative to broad B cell depletion in SLE. © 2015 Elsevier Ltd. All rights reserved.

Keywords: BAFF Systemic lupus erythematosus TACI TLR7 Autoantibodies Lupus nephritis

1. Introduction B cell-activation factor of the tumor necrosis factor superfamily (BAFF; also named BLyS, TNFSF13b) is an essential survival and maturation factor for B cells (reviewed in Ref. [1]). However, excessive BAFF production in BAFF transgenic (Tg) mice leads to the development of an autoimmune disease resembling systemic €gren's syndrome (SS) and lupus erythematosus (SLE) and Sjo

Abbreviations: ANA, anti-nuclear autoantibodies; APRIL, a proliferationinducing ligand; BAFF, B cell-activating factor of the TNF family; BAFFR, BAFF receptor; BCMA, B cell maturation antigen; BM, bone marrow; C3, third component of complement; CSR, class switch recombination; CVID, common variable immunodeficiency; Fo, follicular; GC, germinal center; GSI, glomerulosclerotic index; IC, immune complex; HEL, hen egg lysozyme; MFI, mean fluorescence intensity; MZ, marginal zone; PAS, periodic acid-Schiff; PC, plasma cell; RA, rheumatoid arthritis; RF, rheumatoid factors; RNP, ribonucleoprotein; SLE, systemic lupus erythematosus; € gren's syndrome; TACI, transmembrane activator and Sm, Smith antigen; SS, Sjo cyclophilin ligand interactor; Tg, transgenic; WASp, WiskotteAldrich syndrome protein; WT, wild type. * Corresponding author. Department of Immunology, Central Clinical School, Monash University, 89 Commercial Road, Melbourne, Victoria 3004, Australia. Tel.: þ61 3 99030712; fax: þ61 3 99030031. E-mail address: [email protected] (W.A. Figgett). http://dx.doi.org/10.1016/j.jaut.2015.04.007 0896-8411/© 2015 Elsevier Ltd. All rights reserved.

characterized by the production of a wide range of autoantibodies (reviewed in Ref. [1]). Moreover, elevated BAFF levels have been detected in the serum of patients suffering from various autoimmune conditions, such as SLE, primary SS, and rheumatoid arthritis (RA) [1]. In March 2011, belimumab (trade name Benlysta®), a BAFF neutralizing antibody, was approved for use in a subset of patients with SLE (reviewed in Ref. [2]), highlighting the major importance of BAFF in this pathology. Indeed, the current view supporting B cell number reduction in SLE relies on the notion that the excessive production of BAFF observed in SLE patients is likely linked to the aberrant survival of self-reactive B cells that would normally die throughout the course of B cell tolerance, had they not received excessive survival signals through BAFF receptor (BAFF-R) [2]. Previous studies from our group have shown that excessive BAFF production does not corrupt negative selection of strongly self-reactive B cells, but rather, excess BAFF allows the expansion of low affinity self-reactive B cells that are normally found in the healthy B cell repertoire [3]. As these low affinity self-reactive B cells mostly populate the marginal zone (MZ) and B-1 compartments where self-reactive B cells are known to naturally accumulate, an alternative explanation for disease in BAFF Tg mice has emerged, suggesting that excessive pro-inflammatory

10

W.A. Figgett et al. / Journal of Autoimmunity 61 (2015) 9e16

autoantibody production by innate B cells, rather than uncontrolled B cell survival, underlies the autoimmune disorders developing in BAFF Tg mice [4,5]. Interestingly, rituximab (an anti-human CD20 B cell depleting antibody) is more effective at reducing B cell numbers than belimumab (reviewed in Ref. [6]), however, rituximab's efficacy in treating SLE remains debated (reviewed in Ref. [7]). This difference suggests the possibility that a function other than inhibition of B cell survival may have contributed to the efficacy of belimumab. One possibility is a role for BAFF inhibition in preventing autoantibody production, which is an amelioration that has been noted in SLE patients responding to belimumab (reviewed in Ref. [2]). Apart from BAFF-R, BAFF binds to at least three other receptors: transmembrane activator and calcium-modulating cyclophilin ligand interactor (TACI, TNFRSF13b) and B cell maturation antigen (BCMA, TNFRSF17), and possibly Nogo-66 receptor [8]. BCMA is mostly expressed on plasma cells (PC) and is important for their survival, essentially in response to a proliferation-inducing ligand (APRIL; also named TNFSF13), which binds to TACI and BCMA (reviewed in Ref. [9]). TACI is expressed on mature B cells, and most highly on MZ and peritoneal B-1 B cells in mice; TACI is expressed most highly on CD19þ/CD27þ memory B cells in humans [10,11]. TACI expression is also upregulated on B cells of SLE patients [12]. Studies using TACI-deficient mice revealed that TACI has a dual role; TACI is essential for the production of antibodies in response to T-independent antigens but is also critical for the maintenance of B cell homeostasis (reviewed in Ref. [1]). TLRs, particularly TLR9 and TLR7, are thought to play a pathogenic role in SLE (reviewed in Ref. [13]); we have previously shown that activation of TLR9 or TLR7 strongly upregulates TACI expression on B cells and that BAFF stimulation of B cells also upregulates TLR9 and TLR7 in these cells [5]. Moreover, we also showed that expression of MyD88 in B cells, a signaling element downstream of TLR7 and TLR9 (as well as TLR2, TLR4, TLR5 and IL-1R), is critical for disease progression in BAFF Tg mice [5]. TACI interacts directly with MyD88, establishing a tight connection between TACI activation and TLR signaling and function [14]. TACI also forms a complex with the activated (cleaved) forms of TLR7 or TLR9, and TACI signals synergize with those of TLR7 and TLR9 in B cells [15]. Considering that BAFF-mediated disease is T cell-independent [5] and that TACI is an essential receptor for T-independent B cell responses with its function closely associated with that of TLRs [1], the possibility emerged that the efficacy behind BAFF inhibition in the clinic may lie in the loss of BAFF signals via TACI rather than loss of a fraction of B cells. Indeed, the work presented here demonstrates that loss of TACI in BAFF Tg mice fully protects the animals against autoantibody production and severe SLE-like pathologies, without having any impact on B cell survival. 2. Materials and methods 2.1. Mice BAFF Tg mice that overexpress murine BAFF behind a liverspecific a1-antitrypsin promoter (MGI: 2386944) [16] and TACI/ mice on a C57BL/6 background have been described previously (MGI: 2182823) [17]. C57BL/6 mice were used as WT controls in all experiments. All mice used in this study were females. Mice were housed under conventional barrier protection and handled in accordance with the guidelines of our institutional ethics committees, in compliance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes.

2.2. Bone marrow chimera mice Recipient mice aged 8e14 weeks were irradiated (850 rad, cumulative from two half doses) using an RS2000 X-ray irradiator (Rad Source), and rescued by i.v. injection of 107 bone marrow (BM) cells, which were harvested aseptically from donor mice by aspirating femurs and tibiae with RPMI 1640 medium (Gibco). Chimera mice were analyzed 12 weeks after irradiation. CD45.1/CD45.2 chimera mice had 99e100% B cell chimerism, and 80e90% T cell chimerism 12 weeks after irradiation (data not shown). BM chimeras are denoted in the format of: donor/recipient. 2.3. Flow cytometry and sterile cell sorting Mouse spleens were filtered to prepare single lymphocyte suspensions in FACS buffer (1% (w/v) BSA, 2 mM EDTA; PBS), followed by treatment with Red Blood Cell Lysis Buffer (eBioscience). Lymphocytes were stained with fluorescent-labeled detection antibodies (BD Biosciences, eBioscience), and analyzed with an LSR Fortessa flow cytometer (BD Biosciences) and FlowJo software (Tree Star). Dead cells were excluded from analysis using a fixable live/ dead aqua marker (Molecular Probes). Viable B220þ B cell subsets were gated as Fo (IgMint CD21int CD23hi), MZ (IgMhi CD21hi CD23lo CD1dhi). Viable PC populations were gated as B220loCD138þ. TLR7 ligand binding was measured by flow cytometry using a TLR7 agonist, CL264-biotin (Invivogen) followed by secondary staining with streptavidineallophycocyanin (SA-APC) (BD Biosciences). The delta mean fluorescence intensity (DMFI) was calculated by subtracting the MFI of cells stained with all reagents except CL264biotin. Cells were fixed for intracellular flow cytometry using Intracellular Fixation and Permeabilization buffers (eBioscience) after staining for surface markers. 2.4. ELISA Total circulating Ig and autoantibodies against dsDNA or rheumatoid factors (RF) were measured as previously described [5]. 384-well high-binding assay plates (Corning) were coated at 4  C overnight with goat anti-mouse Ig capture antibody (Southern Biotech) in coating buffer (50 mM sodium bicarbonate) for total isotype detection, or plates were coated with antigens in coating buffer for autoantibody detection as follows: 5 mg/ml calf thymus DNA (SigmaeAldrich) for anti-DNA autoantibodies; 2 mg/ml of goat g-globulin (Jackson ImmunoResearch) for RF; 5 mg/ml of RNP/Sm antigen mixture (Arotec Diagnostics) for bulk anti-RNP/Sm autoantibodies. Plates were blocked with 4% (w/v) BSA in PBS before incubating with serial 2-fold dilutions of each serum sample for 1 h at 37  C. Antibody isotypes were measured using AP-conjugated antibodies to mouse IgM, IgA, total IgG, IgG2b, and IgG2c (Southern Biotech), and detected using para-nitrophenylphosphate (pNPP) substrate (SigmaeAldrich) with a MultiScan Go spectrophotometric plate reader (Thermo Scientific). Titer was defined as the serum dilution giving an OD405nm 4 times higher than background (where 1 ¼ 1:50). For OD comparisons, 1:100 serum dilutions were used and background absorbance was subtracted using the average of the blank wells (using PBS instead of sample). Urine protein concentrations were measured using a BCA protein assay kit (Pierce). Mouse albumin in the urine was measured by ELISA, using a mouse albumin capture and HRP-detection antibody set (Bethyl). 2.5. Histopathology Impaired kidney function was shown by increased proteinuria, measured using the bicinchonic acid (BCA) total protein assay (Pierce) on urine samples diluted 1:50 in PBS to minimize the effect

W.A. Figgett et al. / Journal of Autoimmunity 61 (2015) 9e16

of interfering substances in the urine. Changes in kidney structure were assessed in a blinded protocol in at least 25 randomly selected paraffin-embedded tissue sections from each group studied. Sections were stained with periodic acid-Schiff reagent (PAS) or Masson's modified trichrome. Sections were imaged under a BX51 microscope (Olympus) and glomeruli structures were analyzed using ImageJ software (version 1.48q, National Institutes of Health) and quantitated as described below. 2.6. Glomerulosclerotic index (GSI) and tubular sclerosis quantitation From 5 mm kidney sections with PAS staining, 150e200 glomeruli from mice were examined in a masked protocol. The degree of sclerosis in each glomerulus was subjectively graded on a scale of 0e4 as previously described [18]. Grade 0 is normal; grade 1, sclerotic area up to 25% (minimal); grade 2, sclerotic area 25e50% (moderate); grade 3, sclerotic area 50e75% (moderate to severe); and grade 4, sclerotic area 75e100% (severe). Glomerulosclerosis was defined as glomerular basement membrane thickening, mesangial hypertrophy, and capillary occlusion. A glomerulosclerotic index (GSI) was then calculated using the formula:

GSI ¼

4 X

11

2.9. Real-time PCR Total RNA was isolated from sorted B cell populations using an RNeasy Mini Kit (Qiagen) and 1.5 mg was used for reverse transcription to cDNA using a High capacity cDNA Reverse Transcription Kit (Applied Biosystems). Samples were stored at 20  C until used. Pre-designed TaqMan assays for mouse RN18S rRNA (Mm03928990_g1), TLR7 (Mm00446590_m1) and TLR9 (Mm00446193_m1) mRNA were used with TaqMan Universal Master Mix II (Applied Biosystems). Real-time PCR was performed using an ABI 7900 HT sequence Detection System (Applied Biosystems). Each biological replicate was assayed in triplicate wells. Target mRNA was normalized to 18S rRNA as the endogenous control and the relative fold difference in expression was calculated using the 2DDCt method. 2.10. Statistical analysis Student's t tests (normally distributed data) or ManneWhitney tests (abnormal distribution) were performed where appropriate to assess statistical significance. Error bars represent standard error of the mean (SEM), and statistically significant differences are shown for p < 0.05 (*), p < 0.01 (**), p < 0.001 (***); n.s. (not significant). 3. Results

FiðiÞ

i¼0

where Fi is the percentage of glomeruli with a given score of (i). Similarly, from 5 mm kidney sections stained with Masson's modified trichrome, the degree of tubular sclerosis was graded on a scale of 1e4; grade 1 is normal with sclerotic area less than 25%; grade 2, sclerotic area 25e50%, grade 3, sclerotic area 50e75%; grade 4, more than 75% sclerotic area. Four sections at least 100 mm apart per animal, and 12 non-overlapping fields per section were used. Investigators were masked to the groups. 2.7. Fluorescence microscopy Freshly isolated mouse kidneys were flash-frozen in TissueTek® OCT compound over liquid nitrogen and then stored at 80  C until cryosectioning at 20  C. 7 mm cryosections were cut onto SuperFrost Plus slides (Menzel-Glaser), fixed in acetone for 5 min and then air-dried at RT. Renal Ig deposits and C3 fixation were detected with anti-mouse IgG-Alexa Fluor 488 (Molecular Probes) or anti-mouse C3-FITC (Cederlane), respectively. To detect serum autoantibodies by indirect microscopy, serum was diluted 1:50 or 1:10 in PBS and applied to HEp-2 slides and Crithidia luciliae slides (BIO-RAD), respectively. Slides were washed with PBS and stained with anti-mouse IgG-Alexa Fluor 488 (Molecular Probes) and DAPI (Sigma). All fluorescence slides were viewed with a BX61 fluorescence microscope (Olympus). Microscope images for each experiment were uniformly analyzed using ImageJ software (version 1.48q, National Institutes of Health). 2.8. In vitro cultures B cells were MACS-isolated from splenocyte suspensions using CD43 negative isolation Dynabeads (Invitrogen), and were cultured with or without recombinant mouse BAFF (50 ng/ml) (Enzo Life Sciences) in cell culture media: RPMI 1640 medium supplemented with 10% (v/v) heat-inactivated fetal calf serum (FCS), 2merceptoethanol (50 mM), L-glutamine (2 mM), 100 U/ml penicillin and 100 U/ml streptomycin (Gibco).

3.1. BAFF-mediated upregulation of TLR7 expression in B cells requires TACI signaling Our previous work has shown that TLR7 activation of B cells strongly upregulates the expression of the BAFF receptor TACI [5]. We show here that conversely, BAFF stimulation of B cells upregulates TLR7 expression in a TACI-dependent manner, as TACI/ B cells failed to significantly upregulate TLR7 expression in response to stimulation with BAFF (Fig. 1A). Further supporting these findings, TACI/ MZ and follicular (Fo) B cells express significantly less TLR7 than WT controls, whereas expression of TLR7 is significantly elevated in MZ and Fo B cells from BAFF Tg mice compared to WT controls (Fig. 1B and C, respectively). These results highlight a function of TACI in maintaining normal TLR7 expression levels in resting B cells and increasing TLR7 expression in response to elevated BAFF production. As extensive information demonstrating the pathogenic role of TLR7 in SLE is available (reviewed in Ref. [13]), and considering that TACI expression is highly influenced by TLR7 activation [5,19], we next aimed to test whether TACI is a key pathogenic player in BAFF-mediated autoimmunity. 3.2. Loss of TACI expression prevents class-switched autoantibody production TACI is critical for T cell-independent type 1 and type 2 antibody production [20,21], and IgG autoantibody production in BAFF Tg mice is T cell-independent [5]. Therefore, we hypothesized a possible mechanism whereby TLR7-induced TACI expression combined with excessive BAFF signaling via TACI could lead to the production of large quantities of proinflammatory autoantibodies driving nephritis in BAFF Tg mice. To test this possibility, we reconstituted the immune system of irradiated BAFF Tg mice with either WT or TACI/ BM. The BAFF transgene in these chimeric mice is expressed by hepatocytes to supply high systemic levels of BAFF [5,16], providing an identical environment to test the effects of losing hematopoietic-intrinsic TACI expression in chimera mice. Comparison of these BM chimeras did not reveal a major difference in the reconstitution of splenic B cell subsets (Fig. 2A and B), although numbers of

12

W.A. Figgett et al. / Journal of Autoimmunity 61 (2015) 9e16

Fig. 1. BAFF signaling through TACI upregulates TLR7 expression in B cells. Whole splenic B cell fractions were MACS-isolated from WT or TACI/ spleens and cultured ±BAFF for 5 h, then RNA was extracted and analyzed for TLR7 mRNA by real-time PCR (A). Results are expressed as fold change compared to unstimulated control cells. Gated MZ (B) and Fo (C) B cells from splenocytes isolated from WT, TACI/ and BAFF Tg mice were analyzed for TLR7 ligand binding by flow cytometry. Symbols represent individual mice. Pooled data from three independent experiments are shown and at least n ¼ 3 mice per group are plotted; fold-changes are relative to the mean of the corresponding WT control group of each experiment. Representative histograms of TLR7 ligand binding are shown in Appendix A, Figure S1.

CD138þB220int plasma cells (PC) in the spleen were significantly reduced in TACI//BAFF Tg chimera mice when compared to WT/BAFF Tg chimera controls (Fig. 2C). This difference suggests that in a high BAFF environment, TACI signaling may promote additional PC differentiation in a B cell-intrinsic manner, and this likely involves the previously described TACI-mediated upregulation of Blimp-1 expression in activated B cells [22]. Although TACI is a negative regulator of B cells and numbers of B cells are elevated in TACI/ mice [10,23], no difference was noted between MZ and Fo numbers in TACI//BAFF Tg and WT/BAFF Tg chimeras (Fig. 2A and B). This suggested that BAFF-R signaling rather than loss of TACI expression controlled B cell expansion in the context of a prolonged exposure to high levels of BAFF in vivo, and indeed BAFF-R is the predominant pro-survival receptor for BAFF (reviewed in Ref. [1]). As expected, B cell survival was not compromised in the absence of TACI on B cells. Consistent with reduced numbers of PC in TACI//BAFF Tg mice (Fig. 1C), levels of total serum Ig isotypes in these animals were reduced compared to that of WT/BAFF Tg mice (Fig. 2DeH). Previous studies have shown that APRIL-induced production of IgA is dependent on TACI expression on B cells [14,24]. As a result, total IgA levels were significantly reduced in TACI//BAFF Tg mice compared to WT/BAFF Tg controls (Fig. 2E). Similar observations were made for total IgM, IgG, IgG2b and IgG2c, of which levels in TACI//BAFF Tg controls were comparable to those of WT/WT healthy control chimera mice (Fig. 2DeH). Strikingly, TACI//BAFF Tg chimera mice had significantly reduced levels of circulating autoantibodies compared to WT/BAFF Tg controls, including the complete absence of a range of autoantibody types such as rheumatoid factors (RF, anti-IgG), anti-dsDNA, anti-ribonucleoprotein (RNP), and anti-Smith antigen (Sm) (Fig. 2IeM, and Appendix A Fig. S2). Anti-dsDNA IgA autoantibodies are produced in a T cell-dependent manner in BAFF Tg mice [5], and IgA autoantibodies contribute to disease symptoms similar to IgA nephropathy [25]. Therefore, our results confirm that TACI signaling entirely controls autoantibody production in BAFF Tg mice whether being T cell-dependent (IgA) or T cell-independent. Indirect fluorescence microscopy using C. luciliae and HEp-2 diagnostic slides confirmed that TACI//BAFF Tg lacked IgG anti-dsDNA and anti-nuclear autoantibodies (ANA) (Fig. 3A and B, respectively). In particular, without TACI expression on B cells these mice had significantly reduced levels of IgG2b autoantibodies against dsDNA (Fig. 2J), and this isotype is implicated in complement fixation, inflammation and recruitment of FcgRIV macrophages [5,26]. Consistent with these findings, IgG deposits were

observed directly in the kidneys of WT/BAFF Tg chimeras and BAFF Tg mice, but these deposits were substantially reduced in the kidneys of TACI//BAFF Tg chimera mice (Fig. 3C). As a consequence of reduced Ig deposition in the glomeruli, substantially less C3 complement fixation was observed in the kidneys of TACI//BAFF Tg chimera mice (Fig. 3D). In conclusion, production of autoantibodies in BAFF Tg mice is dependent on TACI expression by B cells. 3.3. Loss of TACI expression protects against BAFF-driven autoimmune tissue damage TACI//BAFF Tg chimeric mice failed to produce pathogenic autoantibody isotypes, and this lack of autoantibodies corresponded with greatly reduced organ pathologies. The kidneys of WT/BAFF Tg or BAFF Tg mice exhibited diffuse glomerulosclerosis, which was characterized by basement membrane thickening, mesangial expansion, and capillary occlusion, but in contrast, TACI//BAFF Tg chimera mice were protected (Fig. 4AeD) and had a significantly reduced glomerulosclerotic disease index (GSI) similar to healthy controls (Fig. 4C). Tubulointerstitial fibrosis, which consists of accumulated extracellular matrix, prominent inflammatory infiltration as well as tubular dilation and atrophy, was observed in WT/BAFF Tg chimera and BAFF Tg kidneys and these pathological features were absent in TACI//BAFF Tg kidney tissues (Fig. 4AeD). Furthermore, increased total protein, and increased albumin concentrations were detected in the urine of WT/BAFF Tg chimera mice, in contrast to TACI//BAFF Tg mice, which lacked these signs of nephritis (Fig. 4E and F). In conclusion, TACI signaling is critical for autoimmune nephritis development in BAFF Tg mice. Moreover, loss of TACI expression totally protects mice against BAFF-mediated autoimmunity. 4. Discussion Our work demonstrates a central and unappreciated role for the BAFF receptor TACI in driving autoantibody production and autoimmune tissue damage in response to excessive BAFF signaling. Our results challenge previous views that excessive B cell survival is the major event underpinning loss of B cell homeostasis and resulting autoimmunity in BAFF Tg mice, and in SLE patients with high BAFF levels [3]. Historically, this original notion has supported strategies aimed at reducing B cell survival/numbers in autoimmunity using B cell-depleting treatments such as belimumab or rituximab, and

W.A. Figgett et al. / Journal of Autoimmunity 61 (2015) 9e16

13

Fig. 2. Loss of TACI prevents autoantibody production and reduces PC numbers in BAFF Tg bone marrow chimera mice. BAFF Tg chimera mice were generated using BM from WT or TACI/ donors. After 12 weeks of reconstitution, tissues (spleen and serum, as indicated) were harvested for analysis. Reconstitution of MZ B (A), Fo B (B), and plasma cell (PC) (C) populations in the spleen were analyzed by flow cytometry. Results are shown as absolute numbers of cells per spleen. Total levels of Ig isotypes (DeH) and titers of anti-dsDNA autoantibodies (IeK), or RF (L), or anti-RNP/Sm (M), as indicated, were measured by ELISA. Data are representative of 4 independent chimera cohorts and at least n ¼ 3 chimera mice per group are plotted. Serum samples from 4 experiments were combined in ELISA experiments.

substantial ongoing investment is being placed in the development of additional biologics based on the same rationale (reviewed in Ref. [27,28]). However, our results entirely change this notion and suggest that targeting TACI, which does not impact on B cell survival/numbers, is sufficient to suppress all pathogenic manifestations resulting from excessive BAFF production. The ability for TACI signaling to upregulate TLR7 expression shown in our study is significant since TLR7 signaling has emerged as a critical pathogenic pathway in SLE, in contrast to TLR9, which may actually play a regulatory or protective role in these circumstances [29]. The different roles of TLR7 and TLR9 in autoimmunity

have been demonstrated in studies using B6.NBa2 mice, in which TLR9-deficiency increased autoantibody levels, yet TLR7/TLR9/ double deficiency negated autoantibody production in this spontaneous mouse model of SLE [30]. Furthermore, in a chimeric murine autoimmunity model where SLE-like disease is driven by the absence of the WiskotteAldrich syndrome protein (WASp) in B cells, TLR7 deletion in B cells was protective against pathogenic autoantibody isotype production and immune complex (IC) glomerulonephritis, whereas B cell-specific TLR9 deletion caused increased production of autoantibodies targeting systemic autoantigens and led to increased disease severity [29]. In the BAFF Tg

14

W.A. Figgett et al. / Journal of Autoimmunity 61 (2015) 9e16

Fig. 3. Pathogenic autoantibody production and kidney Ig deposition in BAFF Tg mice require TACI expression. BAFF Tg or WT control chimeras were generated using WT or TACI/ BM, and analyzed after 12 weeks. Serum autoantibodies were visualized by indirect fluorescence microscopy, using HEp-2 (A) and C. luciliae (B) slides to capture IgG against ANA and anti-dsDNA, respectively. IgG deposition (C) and C3 (D) were observed directly in kidney cryosections by fluorescence microscopy. Untreated WT and BAFF Tg chimera mice were used as negative and positive controls for autoantibody production, respectively. Images are representative of 4 independent experiments including at least n ¼ 5 mice per group. Scale bars represent 50 mm (A, C, D) or 12.5 mm (B), as indicated.

Fig. 4. TACI expression is critical for the development of nephritis in BAFF Tg chimera mice. BAFF Tg chimera mice were generated using BM from WT or TACI/ donors. After 12 weeks, tissues were harvested for analysis of autoimmune organ pathology. Kidney paraffin sections were analyzed by PAS (A), and Masson's Trichrome staining (B). PAS micrographs are annotated for glomeruli: G, and mesangial thickening: arrows. Kidney pathology was assessed by scoring the glomerular sclerotic index (GSI) (C) and tubular fibrosis (D). Proteinuria was quantified by the BCA total protein assay (E), and urine albumin ELISA (F). The fold-change in proteinuria is relative to the WT control group. Data are representative of 4 independent experiments and at least n ¼ 4 mice per group are plotted. The scale bars shown in the WT images (AeB) represent 75 mm.

W.A. Figgett et al. / Journal of Autoimmunity 61 (2015) 9e16

mouse model, our group also found that BAFF Tg chimeras reconstituted with TLR9/ bone marrow (BM) (TLR9//BAFF Tg mice) were not protected from autoimmunity, although MyD88 expression in B cells was critical for disease [5], suggesting that TLR7 was more likely to play an important role in this model. Our results above showed that TACI expression is required for optimal induction of TLR7 expression in response to BAFF. Conversely, activation of TLR7 is very effective at augmenting TACI expression in B cells [5], and in support of this notion, TACI expression is elevated in B Cells from patients with SLE [12]. Together these results suggest that a deleterious signaling loop leading to excessive TACI expression linked to elevated TLR7 levels may be key for disease pathogenesis in BAFF Tg mice via MyD88-dependent collaborative signals. Indeed, previous studies have shown that overexpression of TLR7 in transgenic mice promotes spontaneous autoimmunity [31]. This is not unique to the BAFF Tg model of SLE, as TLR7/MyD88 dependency and/or T cell independency has been observed in other animal models of SLE [32,33]. Our work unambiguously demonstrates that BAFF-mediated autoantibody production followed by autoantibody-mediated inflammation and tissue destruction, is mediated via signaling through its receptor TACI. This information is of value for understanding the mechanism of action of belimumab in patients with SLE. Indeed, belimumab has always been compared to rituximab, as both agents reduce B cell numbers, with rituximab being more effective than belimumab at depleting B cells (reviewed in Ref. [6]). Belimumab leaves memory B cells intact, as they are not dependent on BAFF for survival, whereas rituximab robustly depletes B cells but leaves many PC intact due to low CD20 expression (reviewed in Ref. [2]). Despite this difference, belimumab met the primary endpoint when tested in clinical trials for SLE whereas the exact efficacy of rituximab in SLE remains controversial (reviewed in Ref. [7]). The discrepancy between a partial B cell depleting potency and, yet, a greater efficacy in treating SLE suggested that perhaps the impact on B cell survival was not the therapeutic mechanism of action of belimumab. Indeed, as mentioned above, belimumab does not affect the survival of human memory B cells [34]. In fact, numbers of memory B cells are transiently increased in belimumab-treated patients before returning to baseline levels [34]. Considering that human memory B cells expressed the highest levels of TACI compared to other human B cell subsets [10,11], and the role of TACI in negatively controlling innate-activated B cells, it is conceivable that decreased BAFF signaling via TACI may have initially affected homeostasis of human memory B cells in belimumab-treated patients. Our results clearly show that TACI signaling is central to the autoimmune disease triggered by excessive BAFF production in BAFF Tg mice, therefore, it is conceivable that the real mechanism of action of belimumab lies in its ability to sequester BAFF away from TACI and prevent TACI-mediated pro-inflammatory autoantibody production and subsequent tissue damage driven by elevated serum BAFF detected in patients with SLE. Of note, the therapeutic efficacy of belimumab in patients with SLE has been characterized in part by the observation of significantly reduced levels of autoantibodies and total Ig isotypes [35], consistent with our observation in TACI//BAFF Tg chimera mice. APRIL can also signal through TACI, however it is unlikely to control TACI-mediated autoantibody production as APRIL Tg mice, unlike BAFF Tg mice, do not develop autoimmunity (reviewed in Ref. [1]). Therefore, therapies that block both BAFF and APRIL, such as atacicept (a TACIFc decoy receptor), may not offer additional protection. However, it remains to be determined whether similar mechanisms of action apply in mice and patients. In that regard, future studies including alternative models of SLE and the ability to manipulate TACI signaling at various stages of disease will be very informative.

15

Our observations are of primary importance in the clinic. B cell depletion in either case carries some significant health risk of a greater vulnerability to severe infections. Our work showed that targeting TACI rather than BAFF is likely to offer an alternative therapeutic approach, which would be expected to have an improved safety profile. Indeed, loss of TACI did not compromise B cell survival; rather, deletion of TACI increased B cell numbers whilst decreasing autoantibody levels (Fig. 2). Circulating B cell lymphopenia is commonly observed in human SLE and this may involve TLR7-mediated B cell necroptosis [36], therefore a TACIblocking treatment, expected to reduce TLR7 expression in B cells (as seen in Fig. 1), may ameliorate B cell lymphopenia in addition to stopping TACI-mediated autoantibody production. BAFF-driven expansion of PC numbers in WT/BAFF Tg chimeric mice was not seen in TACI//BAFF Tg chimeras, suggesting that a TACI-specific therapy may better target autoreactive PC in human SLE, as PC are less well targeted with rituximab. Moreover, T cell-dependent B cell responses were not affected in mice lacking TACI except for the production of autoreactive IgA [5]. Moreover, in some settings, TACI deficiency resulted in improved T-dependent responses, which were demonstrated in the cases of intestinal pathogens [37], and anti-NP-OVA IgG responses [22], albeit with impaired responses to some T-independent antigens [20,38]. This suggests that inhibition of TACI in humans will maintain T cell-dependent protective B cell functions, providing scope for a significantly improved safety profile over broad B cell-reducing strategies. One report has described the development of autoimmunity and lymphoma in TACI/ mice [23]. However, our group and others have not noted this problem in our respective independently-generated colonies of TACI/ mice over the past ten years [17]. Mutations in the human TNFRSF13B (TACI) gene have been reported and strongly associate with patients suffering from common variable immunodeficiency (CVID) [15]. Some mutations may affect clustering of TACI receptors and prevent some aspects of its signaling such as antibody responses, leading to symptoms of immunodeficiency [39]. However, similar mutations can be detected in asymptomatic family members, suggesting that multiple factors in addition to a mutation in the TACI gene contribute to CVID [40]. Together these findings suggest that blocking TACI alone will not cause specific safety concerns, however, this is certainly an aspect that would require further careful assessment in humans. Interestingly, BAFF is the only ligand in the TNF superfamily, which at high concentration, can oligomerise into 60mers [41]. These 60mers have been detected in the serum of mice developing SLE [42]. Moreover, some signals via TACI are only triggered by the 60-mer form of BAFF but not the trimeric soluble form [42]. Whether 60-mers are responsible for TACI-mediated disease in human SLE remains to be assessed but this aspect would further explain the specific involvement of TACI in BAFF-mediated autoimmunity. 5. Conclusions Our work provides critical new information on the mechanism of action of BAFF in SLE. Hematopoietic-specific deletion of TACI expression in a mouse model of SLE resulted in complete protection from autoantibody production and organ pathology without extensive loss of B cells; therefore, we suggest that limiting TACI signaling in SLE may specifically target autoantibody generation, which unlike B cell-depleting agents is likely to preserve important protective functions of B cells. While B cell-depleting treatments such as rituximab are generally well-tolerated by many patients, an increasing number of serious adverse events associated with the use of rituximab are being reported [43], indicating the important role of B cells in immunity and a need for more specific and less

16

W.A. Figgett et al. / Journal of Autoimmunity 61 (2015) 9e16

toxic B cell-modulating therapies. Specifically targeting TACI in SLE may offer great efficacy without some of the most concerning adverse events resulting from broad B cell depletion. Disclosures The authors have no financial conflicts of interest. Acknowledgments This work was supported by the National Health and Medical Research Council and Victorian State Government Operational Infrastructure Support. JW-B and FM are Senior Research Fellows of the NHMRC. We thank M. Hibbs, F. Vincent and E. McAllister for review of the manuscript. We thank M. Le Page, Z. Nasa, M. Fuchsberger, M. Bijker, H. McGuire, R. Stokes, A. Nguyen, T. Nguyen, and J. Kang for technical assistance. The authors acknowledge the facilities and scientific and technical assistance of the Monash Animal Research Platform (MARP), Monash Histology Platform, Monash Micro Imaging, and the Monash AMREP Flow Cytometry facility. Appendix A. Supplementary material Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.jaut.2015.04.007. References [1] F. Mackay, P. Schneider, Cracking the BAFF code, Nat. Rev. Immunol. 9 (2009) 491e502. [2] F.B. Vincent, E.F. Morand, P. Schneider, F. Mackay, The BAFF/APRIL system in SLE pathogenesis, Nat. Rev. Rheumatol. 10 (2014) 365e373. [3] M. Thien, T.G. Phan, S. Gardam, M. Amesbury, A. Basten, F. Mackay, et al., Excess BAFF rescues self-reactive B cells from peripheral deletion and allows them to enter forbidden follicular and marginal zone niches, Immunity 20 (2004) 785e798. [4] C.A. Fletcher, J.R. Groom, B. Woehl, H. Leung, C. Mackay, F. Mackay, Development of autoimmune nephritis in genetically asplenic and splenectomized BAFF transgenic mice, J. Autoimmun. 36 (2011) 125e134. [5] J.R. Groom, C.A. Fletcher, S.N. Walters, S.T. Grey, S.V. Watt, M.J. Sweet, et al., BAFF and MyD88 signals promote a lupuslike disease independent of T cells, J. Exp. Med. 204 (2007) 1959e1971. [6] M. Ramos-Casals, I. Sanz, X. Bosch, J.H. Stone, M.A. Khamashta, B-celldepleting therapy in systemic lupus erythematosus, Am. J. Med. 125 (2012) 327e336. [7] A. Kamal, M. Khamashta, The efficacy of novel B cell biologics as the future of SLE treatment: a review, Autoimmun. Rev. 13 (2014) 1094e1101. [8] L. Zhang, S. Zheng, H. Wu, Y. Wu, S. Liu, M. Fan, et al., Identification of BLyS (B lymphocyte stimulator), a non-myelin-associated protein, as a functional ligand for Nogo-66 receptor, J. Neurosci. 29 (2009) 6348e6352. [9] W.A. Figgett, F.B. Vincent, D. Saulep-Easton, F. Mackay, Roles of ligands from the TNF superfamily in B cell development, function, and regulation, Semin. Immunol. 26 (2014) 191e202. [10] W.A. Figgett, K. Fairfax, F.B. Vincent, M.A. Le Page, I. Katik, D. Deliyanti, et al., The TACI receptor regulates T-cell-independent marginal zone B cell responses through innate activation-induced cell death, Immunity 39 (2013) 573e583. [11] A.J. Novak, J.R. Darce, B.K. Arendt, B. Harder, K. Henderson, W. Kindsvogel, et al., Expression of BCMA, TACI, and BAFF-R in multiple myeloma: a mechanism for growth and survival, Blood 103 (2004) 689e694. [12] L.D. Zhao, Y. Li, M.F. Smith Jr., J.S. Wang, W. Zhang, F.L. Tang, et al., Expressions of BAFF/BAFF receptors and their correlation with disease activity in Chinese SLE patients, Lupus 19 (2010) 1534e1549. [13] T. Celhar, R. Magalhaes, A.M. Fairhurst, TLR7 and TLR9 in SLE: when sensing self goes wrong, Immunol. Res. 53 (2012) 58e77. [14] B. He, R. Santamaria, W. Xu, M. Cols, K. Chen, I. Puga, et al., The transmembrane activator TACI triggers immunoglobulin class switching by activating B cells through the adaptor MyD88, Nat. Immunol. 11 (2010) 836e845. [15] N. Romberg, N. Chamberlain, D. Saadoun, M. Gentile, T. Kinnunen, Y.S. Ng, et al., CVID-associated TACI mutations affect autoreactive B cell selection and activation, J. Clin. Invest. 123 (2013) 4283e4293. [16] F. Mackay, S.A. Woodcock, P. Lawton, C. Ambrose, M. Baetscher, P. Schneider, et al., Mice transgenic for BAFF develop lymphocytic disorders along with autoimmune manifestations, J. Exp. Med. 190 (1999) 1697e1710. [17] M. Yan, S.A. Marsters, I.S. Grewal, H. Wang, A. Ashkenazi, V.M. Dixit, Identification of a receptor for BLyS demonstrates a crucial role in humoral immunity, Nat. Immunol. 1 (2000) 37e41.

[18] J.L. Wilkinson-Berka, D.J. Kelly, S.M. Koerner, K. Jaworski, B. Davis, V. Thallas, et al., ALT-946 and aminoguanidine, inhibitors of advanced glycation, improve severe nephropathy in the diabetic transgenic (mREN-2)27 rat, Diabetes 51 (2002) 3283e3289. [19] L.S. Treml, G. Carlesso, K.L. Hoek, J.E. Stadanlick, T. Kambayashi, R.J. Bram, et al., TLR stimulation modifies BLyS receptor expression in follicular and marginal zone B cells, J. Immunol. 178 (2007) 7531e7539. [20] G.U. von Bulow, J.M. van Deursen, R.J. Bram, Regulation of the T-independent humoral response by TACI, Immunity 14 (2001) 573e582. [21] E. Ozcan, L. Garibyan, J.J. Lee, R.J. Bram, K.P. Lam, R.S. Geha, Transmembrane activator, calcium modulator, and cyclophilin ligand interactor drives plasma cell differentiation in LPS-activated B cells, J. Allergy Clin. Immunol. 123 (2009) 1277e1286 e5. [22] S. Tsuji, C. Cortesao, R.J. Bram, J.L. Platt, M. Cascalho, TACI deficiency impairs sustained Blimp-1 expression in B cells decreasing long-lived plasma cells in the bone marrow, Blood 118 (2011) 5832e5839. [23] D. Seshasayee, P. Valdez, M. Yan, V.M. Dixit, D. Tumas, I.S. Grewal, Loss of TACI causes fatal lymphoproliferation and autoimmunity, establishing TACI as an inhibitory BLyS receptor, Immunity 18 (2003) 279e288. [24] D. Sakurai, H. Hase, Y. Kanno, H. Kojima, K. Okumura, T. Kobata, TACI regulates IgA production by APRIL in collaboration with HSPG, Blood 109 (2007) 2961e2967. [25] D.D. McCarthy, J. Kujawa, C. Wilson, A. Papandile, U. Poreci, E.A. Porfilio, et al., Mice overexpressing BAFF develop a commensal flora-dependent, IgA-associated nephropathy, J. Clin. Invest. 121 (2011) 3991e4002. [26] M. Ehlers, H. Fukuyama, T.L. McGaha, A. Aderem, J.V. Ravetch, TLR9/MyD88 signaling is required for class switching to pathogenic IgG2a and 2b autoantibodies in SLE, J. Exp. Med. 203 (2006) 553e561. [27] Z. Liu, A. Davidson, BAFF and selection of autoreactive B cells, Trends Immunol. 32 (2011) 388e394. [28] F.B. Vincent, D. Saulep-Easton, W.A. Figgett, K.A. Fairfax, F. Mackay, The BAFF/ APRIL system: emerging functions beyond B cell biology and autoimmunity, Cytokine Growth Factor Rev. 24 (2013) 203e215. [29] S.W. Jackson, N.E. Scharping, N.S. Kolhatkar, S. Khim, M.A. Schwartz, Q.Z. Li, et al., Opposing impact of B cell-intrinsic TLR7 and TLR9 signals on autoantibody repertoire and systemic inflammation, J. Immunol. 192 (2014) 4525e4532. [30] M.L. Santiago-Raber, I. Dunand-Sauthier, T. Wu, Q.Z. Li, S. Uematsu, S. Akira, et al., Critical role of TLR7 in the acceleration of systemic lupus erythematosus in TLR9-deficient mice, J. Autoimmun. 34 (2010) 339e348. [31] J.A. Deane, P. Pisitkun, R.S. Barrett, L. Feigenbaum, T. Town, J.M. Ward, et al., Control of toll-like receptor 7 expression is essential to restrict autoimmunity and dendritic cell proliferation, Immunity 27 (2007) 801e810. [32] S. Chevrier, T. Kratina, D. Emslie, A. Karnowski, L.M. Corcoran, Germinal center-independent, IgM-mediated autoimmunity in sanroque mice lacking Obf1, Immunol. Cell Biol. 92 (2014) 12e19. [33] R.A. Herlands, S.R. Christensen, R.A. Sweet, U. Hershberg, M.J. Shlomchik, T cell-independent and toll-like receptor-dependent antigen-driven activation of autoreactive B cells, Immunity 29 (2008) 249e260. [34] A.M. Jacobi, W. Huang, T. Wang, W. Freimuth, I. Sanz, R. Furie, et al., Effect of long-term belimumab treatment on B cells in systemic lupus erythematosus: extension of a phase II, double-blind, placebo-controlled, dose-ranging study, Arthritis Rheum 62 (2010) 201e210. [35] R. Furie, M. Petri, O. Zamani, R. Cervera, D.J. Wallace, D. Tegzova, et al., A phase III, randomized, placebo-controlled study of belimumab, a monoclonal antibody that inhibits B lymphocyte stimulator, in patients with systemic lupus erythematosus, Arthritis Rheum 63 (2011) 3918e3930. [36] H. Fan, F. Liu, G. Dong, D. Ren, Y. Xu, J. Dou, et al., Activation-induced necroptosis contributes to B-cell lymphopenia in active systemic lupus erythematosus, Cell Death Dis. 5 (2014) e1416. [37] S. Tsuji, L. Stein, N. Kamada, G. Nunez, R. Bram, B.A. Vallance, et al., TACI deficiency enhances antibody avidity and clearance of an intestinal pathogen, J. Clin. Invest. 124 (2014) 4857e4866. [38] G.T. Mantchev, C.S. Cortesao, M. Rebrovich, M. Cascalho, R.J. Bram, TACI is required for efficient plasma cell differentiation in response to T-independent type 2 antigens, J. Immunol. 179 (2007) 2282e2288. [39] J.J. Lee, I. Rauter, L. Garibyan, E. Ozcan, T. Sannikova, S.R. Dillon, et al., The murine equivalent of the A181E TACI mutation associated with common variable immunodeficiency severely impairs B-cell function, Blood 114 (2009) 2254e2262. [40] W. Koopmans, S.T. Woon, A.E. Brooks, P.R. Dunbar, P. Browett, R. Ameratunga, Clinical variability of family members with the C104R mutation in transmembrane activator and calcium modulator and cyclophilin ligand interactor (TACI), J. Clin. Immunol. 33 (2013) 68e73. [41] T.G. Cachero, I.M. Schwartz, F. Qian, E.S. Day, C. Bossen, K. Ingold, et al., Formation of virus-like clusters is an intrinsic property of the tumor necrosis factor family member BAFF (B cell activating factor), Biochemistry 45 (2006) 2006e2013. [42] C. Bossen, T.G. Cachero, A. Tardivel, K. Ingold, L. Willen, M. Dobles, et al., TACI, unlike BAFF-R, is solely activated by oligomeric BAFF and APRIL to support survival of activated B cells and plasmablasts, Blood 111 (2008) 1004e1012. [43] P.M. Kasi, H.A. Tawbi, C.V. Oddis, H.S. Kulkarni, Clinical review: Serious adverse events associated with the use of rituximab e a critical care perspective, Crit. Care 16 (2012) 231.