Erythematosus Human and Murine Systemic Lupus Bone Marrow Alters B Cell Development in Neutrophil-Mediated IFN Activation In

The Journal of Immunology

Neutrophil-Mediated IFN Activation in the Bone Marrow Alters B Cell Development in Human and Murine Systemic Lupus Erythematosus

Arumugam Palanichamy,* Jason W. Bauer,†,1 Srilakshmi Yalavarthi,‡ Nida Meednu,* Jennifer Barnard,* Teresa Owen,* Christopher Cistrone,* Anna Bird,* Alfred Rabinovich,* Sarah Nevarez,* Jason S. Knight,‡ Russell Dedrick,x Alexander Rosenberg,* Chungwen Wei,{ Javier Rangel-Moreno,* Jane Liesveld,‖ Inaki Sanz,{ Emily Baechler,† Mariana J. Kaplan,‡,2 and Jennifer H. Anolik*

Inappropriate activation of type I IFN plays a key role in the pathogenesis of autoimmune disease, including systemic lupus erythematosus (SLE). In this study, we report the presence of IFN activation in SLE bone marrow (BM), as measured by an IFN gene signature, increased IFN regulated chemokines, and direct production of IFN by BM-resident cells, associated with profound changes in B cell development. The majority of SLE patients had an IFN signature in the BM that was more pronounced than the paired peripheral blood and correlated with both higher autoantibodies and disease activity. Pronounced alterations in B cell development were noted in SLE in the presence of an IFN signature with a reduction in the fraction of pro/pre-B cells, suggesting an inhibition in early B cell development and an expansion of B cells at the transitional stage. These B cell changes strongly correlated with an increase in BAFF and APRIL expression in the IFN-high BM. Furthermore, we found that BM neutrophils in SLE were prime producers of IFN-a and B cell factors. In NZM lupus-prone mice, similar changes in B cell development were observed and mediated by IFN, given abrogation in NZM mice lacking type-I IFNR. BM neutrophils were abundant, responsive to, and producers of IFN, in close proximity to B cells. These results indicate that the BM is an important but previously unrecognized target organ in SLE with neutrophil-mediated IFN activation and alterations in B cell ontogeny and selection. The Journal of Immunology, 2014, 192: 906–918.

ystemic lupus erythematosus (SLE) is a complex auto- CpG, respectively, leading to the initiation of JAK/STAT signaling immune disease that affects multiple target organs. Both cascade, resulting in abundant secretion of type-I IFN (7). Several S the innate and adaptive arms of the immune system con- lines of evidence indicate the connection between type-I IFN and tribute to the pathogenesis of this autoimmune disorder (1, 2). With development of SLE in murine and human studies. Administration respect to innate immune system dysregulation, inappropriate acti- of type-I IFN to mice accelerates the development of autoimmu- vation of type-I IFN plays a critical role in the pathophysiology nity associated with glomerulonephritis (8). In humans, elevated of SLE (3, 4). IFN, a key mediator molecule capable of mounting levels of IFN in the serum of lupus patients were reported almost a ﬁrst line of antiviral response also possesses multiple immune- three decades ago (9). An important link between IFN and SLE modulatory properties that include differentiation of monocytes was revealed by studies of patients receiving IFN-a as a thera- into APCs, activation of T lymphocytes, and differentiation of peutic agent against malignant carcinoid tumors or viral hepatitis, B lymphocytes into Ab-producing plasma cells (5, 6). with a subset developing autoimmune phenomena, including Abs Plasmacytoid dendritic cells (pDCs) are the major producers of against dsDNA and clinical lupus (10). The role of IFN activation type-I IFN in response to infection by a wide array of viruses. pDCs in the initiation and propagation of the disease has been further express TLR7 and -9, which recognize ssRNA and demethylated highlighted by the seminal ﬁnding of upregulation of IFN-inducible

*Division of Allergy, Immunology and Rheumatology, Department of Medicine, Uni- The opinions expressed in this article are the authors’ own and do not reflect the versity of Rochester Medical Center, Rochester, NY 14642; †Department of Medicine, views of the National Institutes of Health, the Department of Health and Human University of Minnesota, Minneapolis, MN 55455; ‡Division of Rheumatology, Depart- Services, or the U.S. Government. ment of Internal Medicine, University of Michigan Medical School, Ann Arbor, MI x { Address correspondence and reprint requests to Dr. Jennifer H. Anolik, Division of 48109; XDx, Inc., Brisbane, CA 94005; Department of Medicine, Emory University, ‖ Allergy, Immunology and Rheumatology, Department of Medicine, University of Atlanta, GA 30332; and Division of Hematology and Oncology, Department of Med- Rochester, 601 Elmwood Avenue, Box 695, Rochester, NY 14642. E-mail address: icine, University of Rochester Medical Center, Rochester, NY 14642 [email protected] 1Current address: UCB, Inc., Atlanta, GA. The online version of this article contains supplemental material. 2Current address: Systemic Autoimmunity Branch, Intramural Research Program, Abbreviations used in this article: BM, bone marrow; BMMC, bone marrow mono- National Institute of Arthritis and Musculoskeletal and Skin Diseases/National Insti- nuclear cell; Grans, granulocytes; MTG, MitoTracker G; NC, normal control; NET, tutes of Health, Bethesda, MD. neutrophil extracellular trap; NZM, New Zealand Mixed 2328; PB, peripheral blood; Received for publication August 9, 2013. Accepted for publication November 20, pDC, plasmacytoid dendritic cell; qPCR, quantitative PCR; SLE, systemic lupus 2013. erythematosus; SLEDAI, SLE Disease Activity Index; SM, switched memory; USM, unswitched memory; WB, whole blood. This work was supported in part by Grants R01 AI077674 (to J.H.A.), P01 AI078907 (to I.S. and J.H.A.), and U19 AI56390 from the Autoimmunity Center of Excellence Ó (to I.S. and J.H.A.), Grant R01 HL088419 (to M.J.K.), and a Rheumatology Research Copyright 2014 by The American Association of Immunologists, Inc. 0022-1767/14/$16.00 Foundation Rheumatology Scientist Development Award (to J.S.K.). www.jimmunol.org/cgi/doi/10.4049/jimmunol.1302112 The Journal of Immunology 907 genes in the peripheral blood (PB) of SLE patients (11, 12). Both mean 6 SEM) in the last draw, suggesting minimal change in PB admixture pDCs and, more recently, neutrophils (13) have been implicated as through the procedure. Preparation of BM smears and flow cytometry as- drivers of IFN activation in SLE. sessment also demonstrated the expected high frequency of precursor populations in the BM aspirates. Within the adaptive compartment of the immune system, dysregulation of B cells has been shown to play a critical role in SLE RNA preparation and quantitative PCR (14). Because the disease is characterized by the generation of For RNA isolation, PB or BM aspirate was placed in PAXgene tubes large amounts of autoantibodies directed against chromatin and (PreAnalytix, Franklin Lakes, NJ) and total RNA isolated according to the other self-Ags, the loss of B cell tolerance clearly plays a key role manufacturer’s protocol with on-column DNase treatment. RNA yield and (15). B cells contribute to the immune pathogenesis and end-organ integrity were assessed using an Agilent Lab-on-a-Chip Bioanalyzer (Agilent Technologies, Palo Alto, CA). In some experiments, CD10+ cells were first damage in SLE via both Ab-dependent and -independent path- purified from the BM aspirate. Total cellular mRNA was reverse-transcribed ways. In an autoimmune setting, B cells can present self-Ag, acti- to cDNA immediately following purification using the High Capacity cDNA vate T cells, and produce proinflammatory cytokines including Reverse Transcription Kit (Applied Biosystems, Foster City, CA). Quanti- TNF-a and IL-6, in addition to secreting autoantibodies (16–18). tative real-time PCR was used to quantify specific cDNA using TaqMan Gene Expression Assays (Applied Biosystems, Foster City, CA). Autoantibodies produced by B cells and RNA- and DNA-containing The expression of five IFN-regulated genes was measured (C1orf29, immune complexes in SLE stimulate pDCs to produce large quan- IFIT1, IRF7, G1P2,andCEB1) by quantitative PCR (qPCR) (12). In a subset tities of IFN-a (19–22) and also contribute to the more recently of patients, examination of three IFN-inducible genes—IFIT1, IRF7,and identified neutrophil activation characteristic of the disease, thereby G1P2—was performed, as this was sufficient to assess IFN activation. establishing a critical link between the adaptive and innate com- These three to five genes were selected because they exhibited the highest correlation with a full 82-gene IFN signature generated by microarray (R2 = partments of the immune system (13). 0.988) (J.W. Bauer and E. Baechler, unpublished observations). Data were Interestingly, it has been demonstrated previously that IFN-a normalized to a housekeeping gene (GAPDH), calibrated using a healthy impairs B cell lymphopoiesis in the bone marrow (BM) of young control blood sample that was included in each run, and (for BM) recali- normal mice (23). Moreover, lupus-prone mice exhibit an age- and brated against a healthy BM. Data are either expressed as a normalized fold change for each gene or an IFN score calculated as follows. Each gene was autoantibody-related decline in B cell lymphopoiesis at the same scaled to the 95th percentile to minimize the effects of outliers, the scaled stage as the inhibition mediated by IFN and an expansion of IFN- values averaged across three genes (IFIT1, GIP2,andIRF7), and averages producing, TLR9-expressing pDCs in the BM (24). Overall, these transformed to a 100-point scale (the final score). The expression of BAFF, results raise the intriguing possibility that B cell lymphopoiesis may APRIL, IFNa ,andIFNb mRNA was also measured in total BM aspirate or be altered in human lupus due to the presence of TLR-stimulating, purified BM neutrophil fractions using similar methods. For murine studies, one IFN-regulated gene (Mx-1), IFN-a, IFN-b, BAFF,andAPRIL,andone interferogenic autoantibodies that have direct BM-mediated effects. housekeeping gene (b-actin) were quantified by real-time PCR as described. Although limited studies have found impaired BM stromal cell function in human SLE (25), the idea that the BM may be an Measurement of serum and BM chemokine and cytokine levels important target organ in SLE is largely unexplored. Given ab- BM supernatant and PB serum were collected from heparinized BM aspirates errant type-I IFN activation in the PB compartment in human and PB drawn in serum tubes, respectively, following a centrifugation. SLE, we examined IFN-regulated gene expression in the BM and SearchLight (Pierce, Woburn, MA) chemiluminescence sandwich-based immunoassays were used to measure serum levels of CCL19 (MIP-3b). changes in B cell development. In the current study, we report for CXCL10 (IP-10) and CCL2 (MCP-1) were measured by multiplex (Milliplex the first time, to our knowledge, the presence of type-I IFN acti- MAG 29; Millipore). These chemokines are IFN regulated, and they exhibited vation in the BM microenvironment in human SLE in association the strongest correlations with disease activity in our prior publications (27). with a reduction in the commitment of precursor cells into the B cell Additional cytokines and chemokines were measured on the MAG 29. Concentrations (picograms per milliliter) were determined using seven-point lineage and increased differentiation of B cells into the transitional standard curves that were generated with recombinant protein. Each sample compartment. We also find that neutrophils are a prime producer of was run in duplicate. BAFF levels were measured by ELISA per the man- type-I IFN in the SLE BM and produce key mediators that affect ufacturer’s instructions (Quantikine; R&D Systems, Minneapolis, MN). B cell development, including APRIL and BAFF. Data in murine Analysis of IFN activation in purified cell populations lupus further implicate neutrophils as a key BM population both stimulated by and producing type-I IFN and mediating altered PB from five SLE patients and three NC was collected, and gene expression patterns were compared between four different cell preparations. Blood was B cell lymphopoeisis. These findings provide new insights into collected in Vacutainer CPT Cell Preparation Tubes with sodium citrate, and the role of IFN in SLE and reveal a novel link between innate cell fractions were prepared as follows. 1) Whole blood (WB): RNA was immune activation and dysregulation of early B cell homeostasis, purified from 1.5 ml WB after lysis in Tri Reagent RT (Molecular Research with critical implications for autoimmunity. Center, Cincinnati, OH) according to the manufacturer’s instructions. 2) RBCs: WB was centrifuged for 15 min at 2000 3 g. The RBC pellet was suspended in an equal volume of distilled PBS plus 2% FBS after removal Materials and Methods of the plasma and buffy coat. The RBC suspension was passed through Study population and sample procurement a LeukoLOCK filter (Ambion Life Technologies) to remove residual WBCs, and RNA was prepared from 1.5 ml filtrate using Tri Reagent RT Detailed written informed consent was obtained from all patients and (Molecular Research Center). 3) PBMCs: the CPT tube was centrifuged for healthy donors, in accordance with protocols approved by the Human 15 min at 2000 3 g. The cells and plasma in the upper fraction were mixed Subjects Institutional Review Board of the University of Rochester Medical by several inversions of the tube and then diluted with 5 ml distilled PBS. Center. Clinical data included a comprehensive medical history, med- Cells were pelleted by centrifugation at 1750 3 g for 5 min. 4) Gran- ications profile, clinical laboratory tests, and assessment of SLE disease ulocytes (Grans): Grans were isolated from the buffy coat fraction using activity by the SLE Disease Activity Index (SLEDAI) (Table I) (26). a CD66b Positive Selection kit (StemCell Technologies, Vancouver, BC, Autoantibodies were measured by clinical multiplex assay (Bio-Rad). Canada). RNA was purified as above. The PBMCs and Grans cell pellets Research BM aspirates and paired PB samples were drawn from 28 SLE were lysed in 1 ml RLT plus 2-ME from an RNeasy kit (Qiagen, Valencia, patients and .20 normal controls (NC). BM mononuclear cells (BMMCs) CA). RNAs from all cell preparations were purified using RNeasy Mini and PBMCs were isolated from heparinized BM aspirate and heparinized columns according to the manufacturer’s instructions (Qiagen). Gene ex- PB by Ficoll-Hypaque density-gradient centrifugation (Pharmacia Biotech, pression evaluated by quantitative RT-PCR was normalized to 18s and Uppsala, Sweden). To avoid PB contamination of marrow aspirates, as- 28s rRNA and relative gene expression levels calculated as a normalized pirate volumes per draw were limited with careful needle tip position fold-change compared with the average expression level in WB from NC. adjustments during the procedure. When nucleated cell differentials were IFI27 and IFI44 were used as type-I IFN–regulated genes in this assay performed between the first and last marrow draws, the percentage of because they are well described to be downstream of IFN-a (12) and were nucleated RBCs was 33 6 4.7% in the first draw and 27 6 4.5% (n =4; expressed at high levels in RBCs in prior studies. 908 BONE MARROW IFN ACTIVATION IN SLE

Isolation and analysis of BM neutrophils bred at the University of Michigan animal facilities. Only female mice were used in the experiments described in this study, at age 12 wk (prior to BM neutrophils were isolated using previous published protocols (28, 29). proteinuria onset). BALB/c mice were purchased from The Jackson Labo- In brief, Percoll gradient solutions with various densities were prepared ratory. Single-cell suspensions from spleen and BM were prepared, and 3 from Percoll (Sigma-Aldrich) using 10 PBS (Invitrogen). Cells isolated multiparameter flow cytometry for B cell analysis was performed as previ- in the top gradient (called M1) contained mostly immature neutrophils, ously described by our group (subset definitions in figure legend) (39). For whereas cells isolated in the lower gradient (M2) consisted primarily of immunohistochemistry, mouse tibias were fixed in neutral buffered formalin mature neutrophils. Phenotype was confirmed by flow cytometry as pre- for 48 h. Bones were immersed in 14% EDTA for 2 wk and transferred viously described (30). In some instances, CD66b magnetic isolation was briefly to PBS. Fixed tibias were progressively dehydrated by transferring used for neutrophil purification. Morphology was confirmed by Giemsa them through graded alcohols (70%, 95%, and absolute alcohol), immersed staining and BAFF, APRIL, and IFN protein expression defined by im- in xylenes, and finally embedded in paraffin. The 5-mm formalin-fixed munofluorescence or flow cytometry. Murine BM neutrophils were isolated paraffin-embedded sections were incubated at 60˚C to melt paraffin and as described previously (31). Briefly, BM was flushed from femurs and transferred quickly to xylenes. Tissues were hydrated by immersing the tis- tibias with HBSS with 15 mM EDTA. Cells were then spun on a discon- sues in absolute alcohol, 95% alcohol, 70% alcohol, and finally in water. Ags 3 tinuous Percoll gradient (52, 69, and 78%) at 1500 g for 30 min. Cells were unmasked by boiling tibias in DakoCytomation Ag retrieval solution from the 69–78% interface were isolated, and RBCs were lysed. (S1699; DakoCytomation) for 30 min and then blocking with 5% normal Analysis of BM pDCs donkey serum (017-000-121; Jackson ImmunoResearch Laboratories) for 30 min at room temperature. Primary Abs were added to the slides and in- BMMCs and PBMCs were stimulated with CpG 2216 (type A: potent pDC cubated at room temperature, in a humid chamber overnight, washed in PBS, stimulator) or CpG 2006 (type B: potent B cell stimulator) or TLR 7/8 ligand and secondary Abs were added to the slides and incubated for 2 h at room R848 (5 mg) for 2 h at 37˚C, followed by protein transport inhibition (bre- temperature. Tissues were washed in PBS and mounted with Prolong Gold feldin A) for 2 h and staining for surface Lin1 (CD3, CD14, CD16, CD19, Anti-fade with DAPI (Invitrogen). Pictures were taken with a Zeiss Axioplan 2 CD20, and CD56), CD304, CD123 markers, and intracellular IFN-a. Microscope and recorder with a Zeiss AxioCam digital camera (Carl Zeiss). Flow cytometry analysis of human B cell populations Bioinformatics analysis of phenotyping data Flow cytometry was performed on a subset of patient samples. Single-cell Samples were hierarchically clustered based on B cell subset proportions suspensions of Ficoll-isolated BMMCs or PBMCs (106/sample) were la- using simplicial distance [a metric useful for compositional data described beled at 4˚C with predetermined optimal concentrations of fluorophore- by Aitchison (40)] and complete linkage. Matlab (Mathworks, Natick, conjugated mAbs against CD19, IgD, CD27, IgM, CD24, and CD38 sur- MA) was used for bioinformatics analyses and visualization. face markers. Pair-matched isotype controls were also used. Previously, the cells were resuspended in RPMI 1640 with 10% FBS and 1% penicillin, Statistical analysis streptomycin, and L-Glutamine, pulse loaded with MitoTracker G (MTG; The nonparametric Mann–Whitney U test was used to compare the dis- 20 nM) (Invitrogen), and chased for 30 min at 37˚C. Mature naive B cells tribution of continuous variables between pairs of groups. Three group express the ABCB1 transporter and effectively extrude the loaded MTG, comparisons were conducted using nonparametric ANOVA (Kruskal– whereas other B cell subsets (including pro-, pre-, immature, transitional, Wallis) followed by Dunn post hoc tests. For comparison of frequency memory, and plasma cell) retain MTG. Analyses were performed on an within groups, Fisher exact test was used. Spearman nonparametric cor- LSRII (BD Biosciences) with FACSDiva software (BD Biosciences) and relation coefficient and associated p values were computed to assess the CD19 gating for identification of B cells (32). B cells were classified along association between pairs of continuous variables. All tests were two- a developmental pathway based on the expression of defined surface +++ +++ 2 2 sided, and p values #0.05 were considered statistically significant. markers as follows: pro/pre- (CD38 , CD24 , IgD , IgM ), immature Statistical analyses were performed using Prism software (GraphPad). (CD38+++, CD24+++, IgD2, IgM+), T1 (IgD+, CD272, MTG+, CD24+++, CD38+++), T2 (IgD+, CD272, MTG+, CD24++, CD38++), T3 (IgD+, CD272, MTG+, CD24low, CD38low), mature naive (IgD+, CD272, MTG2, Results CD24low, CD38low), switched memory (IgD2, CD27+), unswitched mem- SLE BM displays an IFN signature and elevated IFN-regulated 2 2 ory (USM) (IgD+, CD27+), and double-negative memory (IgD , CD27 , chemokines CD102) (33–38). An IFN signature has previously been reported in the PB of SLE Murine studies patients, but the activation of this cytokine in the BM has not been Breeding pairs of lupus-prone New Zealand Mixed 2328 (NZM) and well studied. Therefore, the IFN signature was assessed in BM and NZM2328 IFNabR2/2 mice were a gift from Dr. Chaim Jacob. Mice were paired PB from SLE patients (n = 28) and NC (n = 20) by ex-

Table I. Patient characteristics

IFN-high IFN-low Category Feature (n =16) (n = 12) p Value Demographic Agea 42.3 43.7 NS Femaleb 16 10 NS Whiteb 11 8 NS African Americanb 54NS Disease Disease duration (y)a 10.5 10.8 NS SLEDAIa 3.6 2.6 NS + Anti-DNAb 93NS + Anti-RBPb 12 3 0.02 No. of autoantibodiesc 3.4 1.5 0.003 Low complementsb 54NS WBCa 5.2 5.0 NS Lymphocyte counta 1.5 1.5 NS Medications Prednisone (mg) a 5.7 4.6 NS ISb 11 7 NS Steroids and/or ISb 13 10 NS aContinuous variables (mean) were compared using Mann–Whitney. bData indicate the number of individuals positive for the feature within each group. These variables were compared using Fisher exact test. cAutoantibodies include anti-nuclear Abs, anti-DNA, anti-RNP, anti-Sm, anti-Ro, and anti-La. IS, Immunosuppressive; RBP, RNA binding proteins. The Journal of Immunology 909

amining the relative expression of previously defined three IFN- inducible genes, IFIT1, IRF7, and G1P2, by quantitative RT-PCR. An IFN score was calculated for each sample as the summation of the relative expression of these three genes normalized to a housekeeping gene (GAPDH). Next, we divided the SLE patients into two groups based on the BM IFN signature: 1) IFN high (IFN score $ mean + 2 SD NC); or 2) IFN low (IFN score , mean + 2 SD NC). Notably, an IFN-high signature was demonstrated in the BM of more than half of the SLE patients (57%) (Table I) and correlated with that in the PB (R2 = 0.78; p , 0.0001, Spearman correlation) (Fig. 1A). Examination of the fold change of individual genes (G1P2, IFIT1, and IRF7) in IFN-high and IFN-low SLE (expressed relative to a normal BM) revealed interesting differences, with the highest expression in IFIT1 followed by G1P2 and IRF7 (Fig. 1B, 1C). IFN-high SLE patients expressed significantly higher levels of all three IFN-inducible genes in the BM compared with NC and IFN-low SLE patients (p , 0.0001, Krus- kal–Wallis followed by Dunn post hoc tests). The relationship between the BM and PB IFN signature was further reflected at the individual gene level (BM versus PB expression: R2 = 0.80, 0.82, and 0.82 for G1P2, IFIT1,andIRF7,respectively;p , 0.0001, Spearman correlation). Remarkably, the BM demonstrated higher expression of IFN-inducible genes compared with PB (Fig. 1B, 1C). The magnitude of this difference varied by gene, suggesting interesting tissue-specific differences in IFN activation and gene expression (BM versus PB for IFN high group: 45.7- versus 18.5-fold, p = 0.005 for G1P2; 108.5- versus 54-fold, p = 0.013 for IFIT1;6.3- versus 5.4-fold, NS for IRF7, Wilcoxon signed-rank test). Table I reveals that the clinical, demographic, and immunologic features of the BM IFN-high and -low SLE groups as a whole were similar. However, a notable difference was the enrichment for more autoantibody specificities in the IFN-high group. Thus, the mean number of autoantibodies present was significantly higher in the IFN-high group (p = 0.003) (Table I). There was also a higher frequency of autoantibodies against RNA binding proteins (RNP, Sm, Ro, and La) in the IFN-high group (12 of 16 positive for at least one) versus the IFN low group (3 of 12) (p = 0.02), but no significant differences for anti-DNA (9 of 16 versus 3 of 12; p = 0.13, Fisher exact test). Additionally, on specific analysis of the IFN-high group, significant correlations were observed between the degree of IFN activation in the BM and peripheral lymphopenia (correlation coefficient R2 = 20.58 for IFN-high group total BM score and absolute lymphocyte counts; p = 0.019) and disease activity (correlation coefficient R2 = 0.48 for IFN-high group IFIT1 in the BM and SLEDAI, p = 0.05, Spearman correlation). Given an increased expression of IFN-inducible genes in the BM and paired PB of SLE patients, we next analyzed the levels of IFN-regulated chemokines in the BM supernatants and PB serum FIGURE 1. IFN signature in the BM of SLE patients. IFN signature was measured in terms of relative expression of the three IFN-inducible genes in SLE patients with or without an IFN signature. In particular, the GIP2, IFIT1, and IRF7 (which was shown to reflect the microarray gene levels of three IFN-regulated chemokines were examined—CCL2 expression technology) normalized to a housekeeping gene (GAPDH), (MCP-1), CCL19 (MIP-3b), and CXCL10 (IP-10)—that have using a healthy control blood sample that was included in each run, and previously been shown to be expressed in elevated levels in the (for BM) recalibrated against a healthy BM. IFN score (summation of sera of SLE patients (27). The expression levels of MIP-3b and IP- scores of all three genes) in the SLE PB correlated significantly with the 10 were significantly different on three-group comparison (NC, 2 BM with an R = 0.78; p , 0.0001 (n = 26) (Spearman correlation). SLE IFN-low SLE, and IFN-high SLE; Kruskal–Wallis) (p =0.004 patients were divided into IFN-high (n = 16) and IFN-low (n = 12) groups and p = 0.012, respectively), with higher levels of MIP-3b in the based on the BM IFN scores (IFN-high group: IFN score .2 SD above the normal mean) (n = 20 NC). Dotted lines show the cutoff between IFN-high and IFN-low groups in BM and PBL as determined from the IFN score (A). Fold change of individual IFN-inducible genes in IFN-high and IFN-low IFN high compared with NC and IFN-low (BM) (Dunn post hoc test) and SLE groups are shown in (B) (BM) and (C) (PB), respectively (6 SEM). p = 0.0003 for three-group comparison with significance on IFN-high IFN-high SLE patients expressed significantly higher levels of all three compared with NC and IFN-low (PB), #p = 0.0008, **p = 0.0004 (Kruskal– IFN-inducible genes compared with NC and IFN-low SLE patients. *p , Wallis three-group comparison with significance maintained in high versus 0.0001 for three-group comparison (Kruskal–Wallis) with significance on NC and high versus low on Dunn multiple comparison test). 910 BONE MARROW IFN ACTIVATION IN SLE

IFN-high SLE BM supernatant compared with NC (p , 0.005, NC donors (Mann–Whitney) (Fig. 3). In the WB RNA, the aver- Dunn post hoc) and higher levels of IP-10 in the IFN-high and age IFI27 expression was 270-fold higher in SLE compared with -low SLE compared with NC (p , 0.05, Dunn post hoc) (Fig. NC, with the average IFI44 expression 18-fold higher. Interest- 2A). In the PB, MIP-3b (p = 0.008) and IP-10 (p =0.042)were ingly, these genes showed different levels of expression in dif- also significantly different on three-group comparison (p values ferent cell fractions. For example, the average expression level by Kruskall–Wallis in Fig. 2B), with significantly higher levels increases in SLE of IFI27 were higher in RBC (340-fold) relative in the IFN-high SLE, as has been previously reported (27). Thus, to WB (250-fold) and somewhat lower in PBMC (76-fold) and elevated IFN-inducible chemokines provide further evidence for Grans (55-fold). Because erythrocytes extrude their nuclei before IFN activation in lupus BM. entering the circulation from the BM, this suggests that erythropoiesis occurred in a high IFN environment. As an additional IFN signature in BM-derived cell populations verification of the BM IFN signature, we also sorted CD10+ To provide further evidence that IFN is actually produced in SLE precursor B cells from an SLE BM sample and demonstrated the BM, we purified specific cell fractions from the PB including presence of the IFN signature (data not shown). RBCs. This cell population lacks nuclei, and thus the presence of an IFN signature would suggest erythropoiesis in a high IFN BM Association of IFN activation in SLE BM with changes in environment. Purified populations of RBCs, PBMCs, and Grans B cell development from SLE patients with a high IFN signature (n = 5) and NC (n =3) It has previously been reported that type-I IFN in murine BM has were compared. The expression levels of cell-type specific genes an impact on B cell lymphopoiesis (23). Thus, we next explored (HBA in RBCs, TRBC1 in lymphocytes, and PGLYRP1 in Grans; whether IFN activation in human SLE affects the early B cell Supplemental Fig. 1) verified the purity of the cell populations. developmental pathway in the BM. Although the BM is a micro- Relative expression levels of two IFN-regulated genes, IFI27 and environment for the generation of early B cells, it also serves as IFI44, were elevated in all cell fractions in SLE compared with a niche for mature B cells as memory and long-lived plasma cells home to the BM compartment. Thus, it is critical to separate B cell precursors from mature B cells when studying the BM cell com- position. B cell ontogeny initiates in the BM and follows a se- quential pathway of hematopoiesis: stem cell → pro-B → pre-B → immature B → early transitional B cell followed by migration to the periphery for further differentiation into a mature B cell phenotype. As expected, a significant fraction of early B cells identified by CD24+++CD38+++ surface expression was found in the BM (Fig. 4A). In parallel, these cells were also positive for CD10 expression consistent with a B cell precursor phenotype (Fig. 4B). The fraction of early B cells overall did not vary significantly across NC, IFN-low, and IFN-high BM. However, further classification of early B cells into pro/pre-, immature, and transitional (type 1 and 2) B cells based on the expression of IgD and IgM revealed interesting differences with a reduction of CD19+, CD24+++, CD38+++, IgD2, IgM2 pro/pre-B cells in the early B cell compartment in IFN-high SLE BM (mean 6 SD, high SLE BM versus low SLE BM versus NC BM: 60.3 6 24.1 versus 83.6 6 2.9 versus 79.3 6 5.6, respectively; p = 0.019 by nonparametric ANOVA Kruskal–Wallis with significant differences

FIGURE 2. IFN-regulated chemokines are upregulated in SLE BM with an IFN signature. IFN-regulated chemokines IP-10 (CXCL10), MCP-1 FIGURE 3. IFN signature in SLE RBCs reflects erythropoiesis in an (CCL2), and MIP-3b (CCL19) were assessed in BM supernatant (A) and IFN-rich BM environment. RBCs, PBMCs, and Grans were purified as PB serum or plasma (B)inNC(n = 23) and SLE patients with low (n = 12) described in the Materials and Methods, RNA prepared, and the expression or high (n = 16) IFN signatures as defined in Fig. 1 (mean 6 SEM). *p = of two IFN-regulated genes, IFI27 and IFI44, measured. Gene expression 0.012 for three-group comparison (Kruskal–Wallis) (BM) and 0.042 (PB) in IFN-high SLE patients (n = 5) and NC (n = 3) is represented relative to with significance on IFN-high and IFN-low compared with NC (Dunn post the average expression level in WB from NC after normalization to house- hoc test), **p = 0.004 (BM) and p = 0.008 (PB) for three-group com- keeping genes. All SLE values are significantly different: p = 0.03 compared parison with significance on IFN-high compared with NC. with NC (mean 6 SEM) (except for IFI27 in Grans) (Mann–Whitney). The Journal of Immunology 911

FIGURE 4. Abnormalities in early B cell development in SLE BM with an IFN signature. (A) Flow cytometric analysis of CD19+ BM cells separates B cells into early (CD24+++, CD38+++ expression) versus mature (excluding CD24+++, CD38+++) populations. Early B cells were further divided into pro/pre-, immature, and early transitional (T1 and T2) B cells based on the expression of IgD and IgM. IFN-high lupus patients exhibit reduced pro/pre-B cells and elevated early transitional T1 and T2 cells compared with NC and IFN-low SLE. Representative examples of this phenomenon are shown in (A). (B) B cell precursors express high levels of CD10. (C) Pro/pre-B cells are expressed as a fraction of CD24/CD38 high early B cells (left panel) or total B cells (right panel) and are significantly decreased in the IFN-high group. Immature B cell (D) and early transitional (T1/T2) (E) fractions are shown. Transitional B cells are increased in the IFN-high BMs. *p = 0.012 for three-group comparison (Kruskal–Wallis) with significance on IFN-high compared with IFN-low (Dunn post hoc test). between IFN-high and IFN-low SLE BM on Dunn post hoc SLE, and NC, respectively; p = 0.011 by Kruskal–Wallis with analysis) (Fig. 4C). Immature B cells (CD19+, CD24+++, CD38+++, significant differences between IFN-high and IFN-low SLE BM IgD2,IgM+) represented a small fraction of the early B cell com- on Dunn post hoc analysis; Fig. 4E). With the reduction in the partment across all BM samples, without any significant group fraction of pro/pre-B cells and expansion of early transitional (T1 differences (Fig. 4D). and T2) B cells in the IFN-high SLE BM, we wanted to further Strikingly, transitional B cells (T1 and T2, identified by CD19+, examine the transitional B cell compartment in detail. We have CD24hi, CD38hi, IgD+, and IgM+ expression) were significantly previously shown that, similar to mouse, transitional B cells exist overrepresented in the early B cell compartment in IFN-high SLE in a phenotypic continuum in humans, namely T1 (CD19+, IgD+, BM, whereas IFN-low SLE and NC BM displayed comparable CD272,MTG+,CD24+++, and CD38+++), T2 (CD19+,IgD+,CD272, levels of T1/T2 B cells (27, 4.6, and 7.5% in IFN-high SLE, IFN-low MTG+,CD24++,andCD38++),andT3(CD19+,IgD+,CD272, 912 BONE MARROW IFN ACTIVATION IN SLE

FIGURE 5. Modulation of transitional B cell compartment in SLE BM with an IFN signature. CD19+, IgD+, CD272 naive B cells were examined for MTG-retaining transitional B cells. These cells were further divided into transitional B cell subsets T1, T2, and T3 based on the relative expression of CD24 and CD38 markers. Representative examples of distribution of transitional B cells in the NC BM, IFN-low SLE BM, and IFN-high SLE BM are shown in (A). Representation of transitional subsets as fraction of total transitional B cells is shown in (B). SLE BM with an IFN signature consisted of significantly higher fractions of transitional T2 B cells in the transitional compartment. (C) Cluster analysis reveals a correlation between pro/precontraction and T1/T2 expansion in the BM (Spearman coefficient 20.93; p , 0.0001). The stacked-bar plot represents the relative proportions of the early B cell subsets (CD24CD38 high). Samples were clustered (by flow data) based on simplicial distance and complete linkage. Sample group is (Figure legend continues) The Journal of Immunology 913

and the early transitional (T1 and T2) B cells. Because IFN has previously been demonstrated to inhibit early B cell lymphopoiesis at the pro/pre-B cell stage in murine BM (23, 42), transitional B cells may be effectively selected to maintain B cell homeostasis. Indeed, SLE BM with IFN activation demonstrated a reduced pro/ pre-/immature to T1/T2 ratio (3.3 versus 16.6 versus 13.2; IFN- high SLE versus IFN-low SLE versus NC). This suggests an enhanced selection of B cells into the transitional compartment in IFN-high SLE BM. Clustered, stacked bar charts revealed an association between the pro/pre-contraction and T1/T2 expansion in the B cell compartment in individual patients (Fig. 5) (Spearman coefficient 20.93; p , 0.0001). As early B cells generated in the BM were significantly influ- enced by IFN activation, we wanted to examine the relationship between IFN activation and mature B cell subsets in the BM that are believed to be generated in the periphery and home to the BM niche. Interestingly, we found significant differences in the CD27+- expressing memory cells in SLE high IFN BM, with a significant reduction of CD27+/IgD+ USM and a trend toward expansion of CD27+/IgD2 switched memory (SM) B cells (p = 0.05 for three- group uniform sampling method comparison Kruskal–Wallis) (Supplemental Fig. 2). Clustered mature B cell compositions show that these B cell changes are related to number of autoantibodies as well as gene and chemokine expression (Supplemental Fig. 2) Thus, IFN-high samples cluster together and correlate with relative expansion of SM (Spearman coefficient between BM IFN score and SM, 0.42; p = 0.031) and more autoantibodies (Spearman coefficient between number of autoantibodies and SM, 0.70; p , 0.0001; Spearman coefficient between number of autoantibodies and double negative memory, 0.50; p = 0.02). In the PB compartment, there were similar disturbances in the memory B cell subsets in the IFN- high SLE (p = 0.02 and 0.008 for USM and SM, respectively, FIGURE 6. BAFF and APRIL are upregulated in SLE IFN-high BM. compared with NC PB, Mann–Whitney; data not shown). Thus, IFN BAFF (top panel) and APRIL (bottom panel) mRNA was measured by may have effects on PB B cell activation and differentiation in qPCR in the total BM aspirate in NC (n = 24) and SLE (n = 26). Ex- addition to direct BM effects. pression was normalized to the housekeeping gene GAPDH and then to the NC samples (mean 6 SEM). SLE patients were divided into IFN-high and BAFF and APRIL are increased in the SLE IFN-rich BM IFN-low groups based on the BM IFN scores as in Fig. 1. *p , 0.0001 for microenvironment three-group comparison (Kruskal–Wallis) with significance on IFN-high It has been reported that type I IFN counteracts the IL-7–dependent compared with NC and IFN-low (BM) (p , 0.005 and 0.0005, respec- growth and survival of murine BM B lineage cells at the pro-B cell tively; Dunn post hoc test), **p = 0.0009 for three-group comparison (Kruskal–Wallis) with significance on IFN-high compared with NC and stage (43) and is also required for the earliest recognizable stages IFN-low (BM) (p , 0.005 in both cases, Dunn post hoc test). of human B cell lymphopoiesis (44). However, IL-7 was not significantly different in the SLE BM supernatants compared with healthy controls (Fig. 5, Supplemental Fig. 3), not necessarily MTG+, CD24+/low and CD38+/low) (41). Analysis of these distinct surprising given that IFN effects may be downstream of IL-7. transitional subsets revealed that the IFN-high SLE BM exhibit There were significantly higher levels of TNF-a in the SLE BM a marked increase of T2 B cells compared with NC and IFN-low supernatants (21.3 6 4.7 pg/ml for SLE versus 10.2 6 1.5 pg/ml BM (21.1 6 10.4, 8.8 6 4.3, and 11.4 6 5.2%, as a fraction of for NC; p = 0.003; Supplemental Fig. 3), which is notable because total transitional B cells in IFN-high SLE, IFN-low SLE, and NC proinflammatory cytokines such as IL-1 and TNF-a have also BM, respectively; p = 0.007 by Kruskal–Wallis with significant been shown to inhibit human lymphoid progenitors (44) (IL-1 and differences between IFN-high and IFN-low SLE BM and between IFN-g were not different, data not shown). Additionally, levels of IFN-high SLE and NC BM; Fig. 5). Notably, this expansion was G-CSF, a stimulator of granulopoeisis, were significantly in- also significant when expressed as a fraction of total B cells. creased in SLE BM (68.4 6 11.8 pg/ml for SLE versus 32.0 6 5.6 As a further assessment of B lymphopoiesis, we measured the pg/ml for NC; p = 0.005). Higher levels of G-CSF and TNF-a ratio between earlier B cell precursors (pro-, pre-, and immature) appear to be related to IFN activation (Fig. 5B) (Spearman cor- shown by symbols below the dendrogram. High SLEDAI (.4) is indicated by a circled group label. IFN signature gene expression for BM is shown as heat maps below the CD19 plot. Responses for each gene (shown for BM) were scaled based on its maximum. X indicates missing data. Similarly, log10 chemokine expression in BM is displayed, also scaled to the maximum. Presence or absence of autoantibodies is indicated with closed or open circles, respectively, underneath the gene expression data. IFN-high SLE samples cluster together to the right with high transitional (Spearman coefficient 0.40; p = 0.043) and contracted pro/pre-B cells (Spearman coefficient 20.44; p = 0.025) and more autoantibodies (Spearman coefficient 0.61; p = 0.001). There is also a correlation between BAFF and contracted pro/pre- (p = 0.016), expanded T1/T2 (p = 0.05), and IFN signature (p , 0.0001). *p # 0.01 compared with NC and IFN-low SLE BM, *p = 0.007 for three-group comparison (Kruskal–Wallis) with significance on IFN-high compared with IFN-low and NC (Dunn post hoc test). 914 BONE MARROW IFN ACTIVATION IN SLE

FIGURE 7. Neutrophils in SLE BM produce increased levels of IFN-a and APRIL. (A) Neutrophils were isolated from BM as described in the Materials and Methods for SLE and healthy controls (n = 8). IFNa was measured by qPCR and compared with IFNb. APRIL and BAFF were also measured. Expres- sion was normalized to the housekeeping gene GAPDH and then to the control PB neutrophils (mean 6 SEM). (B) Linear regression analysis of the relationship of IFNa RNA levels and APRIL or BAFF RNA levels in the BM cell populations in (A) plotted on a log scale. (C) Giemsa staining (left panel; original magnification 340, cropped for higher magnification) demonstrating purity and morphology of isolated neutrophils from the BM. By immunofluorescence (original magnification 320, cropped for higher magnification), BM neutrophils (green elastase +) express APRIL (red) protein (overlay yellow) (representative of two experiments). *p = 0.036 compared with matched healthy control population (Mann–Whitney).

relation coefficient relative to the BM IFN score of 0.58 and 0.46, play a critical role in the induction and propagation of autoimmune respectively; p = 0.001 and p = 0.008). responses in SLE by providing autoantigen and immunostimu- We next sought to understand why transitional B cells are in- latory molecules that activate pDCs (30, 45, 46). In addition to creased in the SLE IFN-high BM in contrast to the inhibition seen potentiating type-I IFN production by pDCs, SLE neutrophils in early B cell development. One mechanism could be increased may produce IFN as well (30). Remarkably, we found that the BAFF, a key B cell survival cytokine that is known to be regulated Percoll-isolated M2 (mature) neutrophil fraction in SLE BM had by type-I IFN (8). We did not observe a difference in BAFF protein a significantly higher expression of IFN-a when compared with levels in BM supernatants from SLE patients and controls or be- healthy control M2 BM neutrophils (Fig. 7; p = 0.036, Mann– tween IFN-high and IFN-low SLE (522 6 56 pg/ml for IFN-high Whitney U test). APRIL expression was also higher in SLE BM SLE versus 488 6 45 pg/ml for IFN-low SLE; Supplemental Fig. neutrophils. In contrast, IFN-b was not significantly elevated. 3). However, the cellular expression of BAFF in the BM aspirates Notably, there was a significant correlation (Spearman) between may be a more accurate reflection of BAFF activation. Indeed, neutrophil IFN-a expression and both APRIL (p = 0.0004) and qPCR of BAFF mRNA levels showed significantly higher ex- BAFF expression (p = 0.0001) (Fig. 7B), suggesting the IFN-a pression in the IFN-high SLE group. On cluster analysis (Fig. 5B), may be driving APRIL and BAFF. The expression of APRIL there was a correlation between BAFF and the contracted pre-/pro- (Fig. 7C) and BAFF at the protein level was confirmed by im- (p = 0.016) and expanded T1/T2 (p = 0.05) B cells in the BM. munofluorescence. BAFF also strongly correlated with the IFN signature (Spearman Lupus-prone mice display disturbances in BM B cell coefficient 0.65, p = 3.8E-6). Similarly, the gene expression for the development and neutrophils related cytokine APRIL was significantly increased in the IFN- high SLE BM (Fig. 6). To directly address the putative role of type-I IFNs in altered BM B cell development in lupus, we examined lupus-prone female a Neutrophils are a prime producer of IFN- in SLE BM NZM2328 mice (predisease onset, age 12 wk). Compared to age- Given that pDCs are major producers of IFN-a, we postulated that and sex-matched BALB/c mice, NZM mice had profound alter- apoptotic fragments and immune complexes in the BM may serve ations in the BM B cell compartment with significant reductions in as ligands for TLR9 and TLR7 on pDCs, leading to abundant IFN precursor B cells (particularly at the pre- stage) and subsequent production. Notably, pDC fractions were not significantly different B cell progeny (immature and mature) (Fig. 8A). Interestingly, between SLE and NC BM, and pDC production of IFN in re- similar to human SLE, transitional B cells were in contrast in- sponse to TLR9 or TLR7 stimulation was not significantly dif- creased. These B cell developmental differences were abrogated ferent (data not shown). Recent studies indicate that neutrophils in NZM mice lacking type-I IFN receptor (Fig. 8A). Notably, total The Journal of Immunology 915

FIGURE 8. Murine SLE is associated with disturbances in B cell lymphopoiesis related to BM IFN activation and driven by neutrophils. (A) NZM mice age 12 wk were compared with age-matched NZM mice lacking type-I IFNR (I-NZM) and BALB/C mice (n = 5/group). B cell subsets in the BM were deﬁned as follows: pro- (B220lowIgM2CD43+), pre- (B220+IgM2CD43low), immature (B220+IgM+CD232), T1 (B220+AA4.1+IgMhiCD232), T2 (B220+ AA4.1+IgM+CD23+CD24hi), and mature (B220+IgM+CD23+AA4.12CD24low). Data are shown for B cell subsets expressed relative to the total live lymphocyte gate, except for T1 and T2, which are expressed relative to the early B cells (AA4.1+). Similar trends were observed when expressed relative to B220+ B cells for all subsets including the T1 and T2. p values were calculated by nonparametric ANOVA (three-group comparison Kruskal–Wallis with Dunn post hoc test). (B) BM and spleen cells (data not shown) were isolated from the above groups of mice (n = 7/group). Neutrophils were isolated from BM as described in the Materials and Methods. IFNa, IFNb, APRIL, and BAFF were measured by qPCR and expression normalized to the housekeeping gene actin and then to BALB/C controls (mean 6 SEM). (C) Neutrophils were isolated from a lupus-prone mouse BM (n = 2) and stimulated in vitro with IFN-a (1000 U/ml) for 4 h prior to qPCR. Expression is normalized to housekeeping gene and then to media control cultured (Figure legend continues) 916 BONE MARROW IFN ACTIVATION IN SLE

BM and spleen cells from these NZM mice displayed significantly Interesting recent data have highlighted the role of neutrophils higher in vivo expression of IFN-a expression by qPCR compared in amplifying IFN activation (45) both by directly producing IFN with BALB/c controls (p = 0.03 for BM and p = 0.019 for spleen, (30) and indirectly by stimulating pDCs. This may be particularly shown for BM, Mann–Whitney U test) (Fig. 8B). Changes in type- relevant in the BM microenvironment where neutrophils develop. I IFN–regulated genes (Mx-1) were more variable. However, BM Neutrophil extracellular traps (NETs; weblike structures released neutrophils displayed significant upregulation of Mx-1 (10-fold; by neutrophils as a means of immobilizing and killing invading p = 0.01), as well as increases in IFN-a, IFN-b,andBAFF microbes), which contain nuclear material (DNA and histones) (p = 0.03, 0.03, and 0.04, respectively; p = 0.08 for APRIL). and neutrophil proteins, are increased by IFN in SLE and in turn Moreover, stimulation of BM neutrophils with IFN-a in vitro in- activate pDCs to produce high levels of IFN in a DNA- and TLR9- duced further upregulation of both IFNa and BAFF (Fig. 8C). In dependent manner (45, 46, 51). Additionally, in one of the few bone sections, we also detected numerous Gr-1+myeloperoxidase+ studies examining SLE BM, Nakou et al. (52) have demonstrated neutrophils in NZM mice in close contact with B220+ B cells (Fig. a cell death and granulopoiesis signature but not an IFN signature. 8D) and also expressing APRIL and BAFF. In combination, these The authors speculated that the absence of an IFN signature might data confirm that neutrophils are a major cell population mediat- be related to differences in the patient populations examined. Our ing IFN production and the secretion of cytokines that alter B cell data provide new evidence that BM neutrophils contribute to local development in the BM and further highlight the IFN-driven na- IFN production in SLE with potentially critical effects on devel- ture of the process. oping B cells. IFN-a acts at the center of an amplification loop in SLE, con- Discussion tributing to B cell abnormalities, promoting the differentiation of Although SLE is known to be associated with a type I IFN signature activated B cells into plasmablasts (5), and, in conjunction with in the PB (11, 12), the precise site and mechanism of IFN pathway TLR stimulation, triggering B cell expansion (53) and a lowered activation remain unclear. In this study, we demonstrate for the activation threshold for autoreactive B cells (54). Our data suggest first time, to our knowledge, that the BM in human SLE is a novel role for IFN-a activation in the BM of SLE patients by a primary site of IFN activation based on the expression of IFN- decreasing B cell lymphopoiesis, contributing to B cell lympho- inducible genes and chemokines. Notably, IFN activation in SLE penia, and potentially decreasing the stringency of B cell–negative BM was strongly associated with changes in B cell lympho- selection. This is in accord with data from Lin et al. (23) and poeisis, initiated in the early stages of immune cell develop- Wang et al. (43) demonstrating that type-I IFN–producing mac- ment in the BM microenvironment, and likely mediated at least rophages in murine BM oppose IL-7–induced survival of B cell in part by IFN-induced BAFF upregulation. Further, we dem- precursors at the pro-B stage, possibly by Bcl-2–driven apoptosis onstrate that neutrophils in SLE BM are a prime producer of of pro-B cells mediated by type-I IFN (55). Similarly, in NZM IFN, stimulating APRIL, BAFF, and further IFN expression, lupus-prone mice with elevated levels of IFN in the BM, we find thus providing multiple direct and indirect links between neu- a loss of precursor B cells, most striking at the pre-B cell stage. trophils and B cell development and survival. Overall, our results These findings are consistent with published literature that NZB highlight the importance of the BM as a target organ in SLE and mice have a loss of pre-B cells in the BM that develops with age provide a novel connection between IFN activation and B cell se- and autoantibody onset (24, 42). However, the specific association lection. between an IFN signature and B cell development in murine lupus Genomic approaches have demonstrated that the majority of and human SLE has not been previously reported. SLE patients display a type-I IFN signature based on WB or PBMC Our findings indicate that, similar to mouse models, perturba- gene expression profiling. The signals generating IFN activation tions in early B cell development are observed in human SLE in the in SLE are relatively well characterized. Thus, RNA- and DNA- BM compartment and are likely at least partly mediated by IFN. We containing immune complexes that are abundant in SLE activate also found disturbances in several cytokines and chemokines in the pDCs via TLRs (20–22) to produce large quantities of IFN-a. SLE BM. An increase in TNF-a and G-CSF in the SLE BMs, Although pDCs are paradoxically decreased in the PB of SLE cytokines known to impair B lymphopoiesis, suggests additional patients (6), it is speculated that they may accumulate in sites of IFN-independent mechanisms for inhibition of B cell development inflammation such as the skin (47, 48) or kidney (49). Other than (56). Other studies have demonstrated increased apoptosis of these few publications, however, there is surprisingly little liter- CD34+ BM stem cells in SLE and impaired function of stromal ature addressing the primary site of IFN activation in SLE and no cells (25). It has also been appreciated recently that hematopoietic consideration of the BM as an important location. We hypothe- progenitors express functional TLRs, with TLR ligands blocking sized in this paper that the BM might be an ideal organ for the B lineage differentiation (57), providing another potential IFN- production of IFN given the presence of long-lived plasma cell independent mechanism for inhibition of B cell lymphopoeisis populations as a putative source of interferogenic autoantibodies, in SLE. Ag in the form of apoptotic/dying cells, and BM pDCs as IFN Although we noted an inhibition at the pro/pre-B cell stage, an producers. Indeed, BM DCs from lupus-prone mice have been increase of early transitional B cells into the T2 compartment was demonstrated to produce IFN in a TLR9-dependent fashion (24). observed in the IFN-activated BM. The precise molecular mech- Other reports indicate that distinct cell types, including BM- anisms involved in this process are not fully understood, but this is resident macrophages (43), may also produce IFN. Similarly, likely to have critical implications for the breach in B cell tolerance a mouse model of viral lung infection demonstrated alveolar in SLE. The selection of newly formed B cells into the mature macrophages as prime producers of IFN-a (50). follicular compartment occurs via these intermediate subsets or

neutrophils. (D) Decalcified BM sections were immunostained with Abs to Gr-1 and myeloperoxidase (MPO; neutrophils), B220 (B cells), proliferating cell nuclear Ag (PCNA), APRIL, and BAFF (original magnification 3200; images are cropped for size, insets on the right). Neutrophils are prominent in the SLE BM and adjacent to B cells (white arrow). Many BM neutrophils express APRIL, with a smaller fraction coexpressing BAFF. The pictures are representative of 10 different samples from BALB/C and NZM mice. The Journal of Immunology 917 transitional B cells (58), a key target of negative selection of References autoreactivity (59, 60). Among the numerous signals regulating 1. Crampton, S. P., E. Voynova, and S. Bolland. 2010. Innate pathways to B-cell the development from one transitional stage to another and ulti- activation and tolerance. Ann. N. Y. Acad. Sci. 1183: 58–68. 2. Rahman, A., and D. A. Isenberg. 2008. Systemic lupus erythematosus. N. Engl. mately selection into the mature repertoire, BAFF is key (61) and J. Med. 358: 929–939. was notably increased in IFN-high SLE BM in the current study. 3. Baechler, E. C., P. K. Gregersen, and T. W. Behrens. 2004. The emerging role of BAFF knockout mice have a block in B cell development at the T1 interferon in human systemic lupus erythematosus. Curr. Opin. Immunol. 16: 801–807. stage (62), and transgenic expression of BAFF results in dramatic 4. Ro¨nnblom, L., and V. Pascual. 2008. The innate immune system in SLE: type I increases in the late transitional compartment (63), loss of B cell interferons and dendritic cells. Lupus 17: 394–399. 5. Jego, G., A. K. Palucka, J. P. Blanck, C. Chalouni, V. Pascual, and J. Banchereau. tolerance, and development of an SLE-like phenotype (64). Al- 2003. Plasmacytoid dendritic cells induce plasma cell differentiation through though numerous studies have demonstrated increased levels of type I interferon and interleukin 6. Immunity 19: 225–234. BAFF in the human SLE PB, none have examined this phenom- 6. Blanco, P., A. K. Palucka, M. Gill, V. Pascual, and J. Banchereau. 2001. In- duction of dendritic cell differentiation by IFN-alpha in systemic lupus eryth- enon in the BM, where critically, B cells are developing. In- ematosus. Science 294: 1540–1543. triguingly, IFN has been reported to increase the expression of 7. Haeryfar, S. M. 2005. The importance of being a pDC in antiviral immunity: the BAFF by human neutrophils (65), providing a mechanistic link IFN mission versus Ag presentation? Trends Immunol. 26: 311–317. 8. Mathian, A., A. Weinberg, M. Gallegos, J. Banchereau, and S. Koutouzov. 2005. between our observations of BM IFN activation and enhanced IFN-alpha induces early lethal lupus in preautoimmune (New Zealand Black x B cell differentiation into the T2 compartment. The demonstration New Zealand White) F1 but not in BALB/c mice. J. Immunol. 174: 2499–2506. 9. Hooks, J. J., H. M. Moutsopoulos, S. A. Geis, N. I. Stahl, J. L. Decker, and in this study that lupus-prone NZM mice have elevated IFN and A. L. Notkins. 1979. Immune interferon in the circulation of patients with au- BAFF in the BM and a notable increase in the transitional com- toimmune disease. N. Engl. J. Med. 301: 5–8. partment supports this mechanism. In turn, it has previously been 10. Ro¨nnblom, L. E., G. V. Alm, and K. Oberg. 1991. Autoimmune phenomena in patients with malignant carcinoid tumors during interferon-alpha treatment. Acta demonstrated that APRIL produced by neutrophils promotes the Oncol. 30: 537–540. survival of human plasma cells in mucosa (66) and BM (67) and 11. Bennett, L., A. K. Palucka, E. Arce, V. Cantrell, J. Borvak, J. Banchereau, and possibly the retention of memory B cells (68, 69). Further, IFN is V. Pascual. 2003. Interferon and granulopoiesis signatures in systemic lupus erythematosus blood. J. Exp. Med. 197: 711–723. known to stimulate APRIL production by DCs (67). We now 12. Baechler, E. C., F. M. Batliwalla, G. Karypis, P. M. Gaffney, W. A. Ortmann, provide evidence that BM neutrophils in human and murine lupus K. J. Espe, K. B. Shark, W. J. Grande, K. M. Hughes, V. Kapur, et al. 2003. respond to IFN, upregulating IFN target genes APRIL and BAFF. Interferon-inducible gene expression signature in peripheral blood cells of patients with severe lupus. Proc. Natl. Acad. Sci. USA 100: 2610–2615. The relevance of these findings for the human BM microenvi- 13. Knight, J. S., and M. J. Kaplan. 2012. Lupus neutrophils: ‘NET’ gain in un- ronment in SLE merit further study but are in accord with the derstanding lupus pathogenesis. Curr. Opin. Rheumatol. 24: 441–450. 14. Ahmed, S., and J. H. Anolik. 2010. B-cell biology and related therapies in higher APRIL expression by neutrophils in the human SLE BM systemic lupus erythematosus. Rheum. Dis. Clin. North Am. 36: 109–130, viii– observed in this study and the increase in BM memory B cells. ix. Additionally, G-CSF and TNF-a were both noted to be elevated in 15. Arbuckle, M. R., M. T. McClain, M. V. Rubertone, R. H. Scofield, G. J. Dennis, J. A. James, and J. B. Harley. 2003. Development of autoantibodies before the the SLE BM and have also been described to stimulate BAFF clinical onset of systemic lupus erythematosus. N. Engl. J. Med. 349: 1526– production by neutrophils (70). These cytokines also impair B 1533. lymphopoiesis (56), providing a potential explanation for both the 16. Lipsky, P. E. 2001. Systemic lupus erythematosus: an autoimmune disease of B cell hyperactivity. Nat. Immunol. 2: 764–766. inhibition of early B cell development and enhanced transitional 17. Browning, J. L. 2006. B cells move to centre stage: novel opportunities for B cell selection observed in this study. Intriguingly, it was recently autoimmune disease treatment. Nat. Rev. Drug Discov. 5: 564–576. 18. Anolik, J. H., R. J. Looney, F. E. Lund, T. D. Randall, and I. Sanz. 2009. Insights published that BM neutrophils from NZM mice have enhanced into the heterogeneity of human B cells: diverse functions, roles in autoimmu- NET formation (71). Thus, it is tempting to speculate that some of nity, and use as therapeutic targets. Immunol. Res. 45: 144–158. the effects of neutrophils on the BM B cell compartment are 19. Boule´, M. W., C. Broughton, F. Mackay, S. Akira, A. Marshak-Rothstein, and I. R. Rifkin. 2004. Toll-like receptor 9-dependent and -independent dendritic cell mediated through NET formation. activation by chromatin-immunoglobulin G complexes. J. Exp. Med. 199: 1631– In conclusion, our results indicate that the majority of lupus patients 1640. have local production of IFN-a in the BM by resident neutrophils 20. Ba˚ve, U., G. V. Alm, and L. Ro¨nnblom. 2000. The combination of apoptotic U937 cells and lupus IgG is a potent IFN-alpha inducer. J. Immunol. 165: 3519– that also contribute to elevated levels of the B cell survival factors 3526. BAFF and APRIL in an IFN-dependent manner. We suggest that the 21. Ba˚ve, U., M. Magnusson, M. L. Eloranta, A. Perers, G. V. Alm, and L. Ro¨nnblom. 2003. Fc gamma RIIa is expressed on natural IFN-alpha- BM may be an ideal organ for the production of IFN given the producing cells (plasmacytoid dendritic cells) and is required for the IFN- presence of long-lived plasma cell populations producing autoanti- alpha production induced by apoptotic cells combined with lupus IgG. J. bodies, apoptotic B cells providing nuclear Ags, and neutrophils as Immunol. 171: 3296–3302. 22. Means, T. K., E. Latz, F. Hayashi, M. R. Murali, D. T. Golenbock, and both direct IFN producers and potentiators of pDC and B cell acti- A. D. Luster. 2005. Human lupus autoantibody-DNA complexes activate DCs vation. In the setting of IFN activation, an inhibition of early B through cooperation of CD32 and TLR9. J. Clin. Invest. 115: 407–417. lymphopoiesis may occur, and with low numbers of precursor B cells 23. Lin, Q., C. Dong, and M. D. Cooper. 1998. Impairment of T and B cell development by treatment with a type I interferon. J. Exp. Med. 187: 79–87. entering the transitional B cell compartment, the stringency of neg- 24. Lian, Z. X., K. Kikuchi, G. X. Yang, A. A. Ansari, S. Ikehara, and ative selection may be compromised leading to enhanced recruitment M. E. Gershwin. 2004. Expansion of bone marrow IFN-alpha-producing dendritic cells in New Zealand Black (NZB) mice: high level expression of TLR9 of autoreactive clones into the mature B cell repertoire. and secretion of IFN-alpha in NZB bone marrow. J. Immunol. 173: 5283–5289. 25. Papadaki, H. A., D. T. Boumpas, F. M. Gibson, D. R. Jayne, J. S. Axford, E. C. Gordon-Smith, J. C. Marsh, and G. D. Eliopoulos. 2001. Increased apo- Acknowledgments ptosis of bone marrow CD34(+) cells and impaired function of bone marrow We thank the patients for participation and the clinical coordinators, stromal cells in patients with systemic lupus erythematosus. Br. J. Haematol. Debbie Campbell, Maria Allen, Kelly Callahan, and Natasha Nedelkova. 115: 167–174. We also thank Hengwei Zhang and Lianping Xing for assisting with the 26. Touma, Z., M. B. Urowitz, D. Iban˜ez, and D. D. Gladman. 2011. SLEDAI-2K 10 days versus SLEDAI-2K 30 days in a longitudinal evaluation. Lupus 20: 67–70. preparation of mouse bone samples for immunofluorescence, Dr. Chaim Jacob 27. Bauer, J. W., M. Petri, F. M. Batliwalla, T. Koeuth, J. Wilson, C. Slattery, 2 2 for providing breeding pairs of NZM2328 and NZM2328 IFNabR / A. Panoskaltsis-Mortari, P. K. Gregersen, T. W. Behrens, and E. C. Baechler. mice, biostatistician Ollivier Hyrien for input, and John Looney for enlight- 2009. Interferon-regulated chemokines as biomarkers of systemic lupus eryth- ening discussions. ematosus disease activity: a validation study. Arthritis Rheum. 60: 3098–3107. 28. Hu, Y. 2012. Isolation of human and mouse neutrophils ex vivo and in vitro. Methods Mol. Biol. 844: 101–113. 29. Cowand, J. B., and N. Borregaard. 1999. Isolation of neutrophil precursors from Disclosures bone marrow for biochemical and transcriptional analysis. J. Immunol. Methods The authors have no financial conflicts of interest. 232: 191–200. 918 BONE MARROW IFN ACTIVATION IN SLE

30. Denny, M. F., S. Yalavarthi, W. Zhao, S. G. Thacker, M. Anderson, A. R. Sandy, 51. Villanueva, E., S. Yalavarthi, C. C. Berthier, J. B. Hodgin, R. Khandpur, W. J. McCune, and M. J. Kaplan. 2010. A distinct subset of proinflammatory A. M. Lin, C. J. Rubin, W. Zhao, S. H. Olsen, M. Klinker, et al. 2011. Netting neutrophils isolated from patients with systemic lupus erythematosus induces neutrophils induce endothelial damage, infiltrate tissues, and expose immunos- vascular damage and synthesizes type I IFNs. J. Immunol. 184: 3284–3297. timulatory molecules in systemic lupus erythematosus. J. Immunol. 187: 538– 31. Boxio, R., C. Bossenmeyer-Pourie´, N. Steinckwich, C. Dournon, and O. Nu¨sse. 552. 2004. Mouse bone marrow contains large numbers of functionally competent 52. Nakou, M., N. Knowlton, M. B. Frank, G. Bertsias, J. Osban, C. E. Sandel, neutrophils. J. Leukoc. Biol. 75: 604–611. H. Papadaki, A. Raptopoulou, P. Sidiropoulos, I. Kritikos, et al. 2008. Gene 32. Anolik, J. H., J. W. Friedberg, B. Zheng, J. Barnard, T. Owen, E. Cushing, expression in systemic lupus erythematosus: bone marrow analysis differentiates J. Kelly, E. C. Milner, R. I. Fisher, and I. Sanz. 2007. B cell reconstitution after active from inactive disease and reveals apoptosis and granulopoiesis signatures. rituximab treatment of lymphoma recapitulates B cell ontogeny. Clin. Immunol. Arthritis Rheum. 58: 3541–549. 122: 139–145. 53. Bekeredjian-Ding, I. B., M. Wagner, V. Hornung, T. Giese, M. Schnurr, 33. Anolik, J. H., J. Barnard, A. Cappione, A. E. Pugh-Bernard, R. E. Felgar, S. Endres, and G. Hartmann. 2005. Plasmacytoid dendritic cells control TLR7 R. J. Looney, and I. Sanz. 2004. Rituximab improves peripheral B cell abnor- sensitivity of naive B cells via type I IFN. [Published erratum appears in 2005 J. malities in human systemic lupus erythematosus. Arthritis Rheum. 50: 3580– Immunol. 174: 5884.] J. Immunol. 174: 4043–4050. 3590. 54. Uccellini, M. B., L. Busconi, N. M. Green, P. Busto, S. R. Christensen, 34. Bohnhorst, J. Ø., M. B. Bjørgan, J. E. Thoen, J. B. Natvig, and K. M. Thompson. M. J. Shlomchik, A. Marshak-Rothstein, and G. A. Viglianti. 2008. Autoreactive 2001. Bm1-Bm5 classification of peripheral blood B cells reveals circulating B cells discriminate CpG-rich and CpG-poor DNA and this response is modu- germinal center founder cells in healthy individuals and disturbance in the B cell lated by IFN-alpha. J. Immunol. 181: 5875–5884. subpopulations in patients with primary Sjo¨gren’s syndrome. J. Immunol. 167: 55. Gamero, A. M., R. Potla, J. Wegrzyn, M. Szelag, A. E. Edling, K. Shimoda, 3610–3618. D. C. Link, J. Dulak, D. P. Baker, Y. Tanabe, et al. 2006. Activation of Tyk2 and 35. Pascual, V., Y. J. Liu, A. Magalski, O. de Bouteiller, J. Banchereau, and Stat3 is required for the apoptotic actions of interferon-beta in primary pro- J. D. Capra. 1994. Analysis of somatic mutation in five B cell subsets of human B cells. J. Biol. Chem. 281: 16238–16244. tonsil. J. Exp. Med. 180: 329–339. 56. Ueda, Y., K. Yang, S. J. Foster, M. Kondo, and G. Kelsoe. 2004. Inflammation 36. Anolik, J. H., J. Barnard, T. Owen, B. Zheng, S. Kemshetti, R. J. Looney, and controls B lymphopoiesis by regulating chemokine CXCL12 expression. J. Exp. I. Sanz. 2007. Delayed memory B cell recovery in peripheral blood and lym- Med. 199: 47–58. phoid tissue in systemic lupus erythematosus after B cell depletion therapy. 57. Sioud, M., Y. Fløisand, L. Forfang, and F. Lund-Johansen. 2006. Signaling Arthritis Rheum. 56: 3044–3056. through toll-like receptor 7/8 induces the differentiation of human bone marrow 37. Cuss, A. K., D. T. Avery, J. L. Cannons, L. J. Yu, K. E. Nichols, P. J. Shaw, and CD34+ progenitor cells along the myeloid lineage. J. Mol. Biol. 364: 945–954. S. G. Tangye. 2006. Expansion of functionally immature transitional B cells is 58. Loder, F., B. Mutschler, R. J. Ray, C. J. Paige, P. Sideras, R. Torres, associated with human-immunodeficient states characterized by impaired hu- M. C. Lamers, and R. Carsetti. 1999. B cell development in the spleen takes moral immunity. J. Immunol. 176: 1506–1516. place in discrete steps and is determined by the quality of B cell receptor-derived 38. Sims, G. P., R. Ettinger, Y. Shirota, C. H. Yarboro, G. G. Illei, and P. E. Lipsky. signals. J. Exp. Med. 190: 75–89. 2005. Identification and characterization of circulating human transitional 59. Carsetti, R., G. Ko¨hler, and M. C. Lamers. 1995. Transitional B cells are the B cells. Blood 105: 4390–4398. target of negative selection in the B cell compartment. J. Exp. Med. 181: 2129– 39. Bekar, K. W., T. Owen, R. Dunn, T. Ichikawa, W. Wang, R. Wang, J. Barnard, 2140. S. Brady, S. Nevarez, B. I. Goldman, et al. 2010. Prolonged effects of short-term 60. Wardemann, H., S. Yurasov, A. Schaefer, J. W. Young, E. Meffre, and anti-CD20 B cell depletion therapy in murine systemic lupus erythematosus. M. C. Nussenzweig. 2003. Predominant autoantibody production by early human Arthritis Rheum. 62: 2443–2457. B cell precursors. Science 301: 1374–1377. 40. Aitchison, J. 2001. Simplicial inference. In Algebraic Methods in Statistics and 61. Miller, J. P., J. E. Stadanlick, and M. P. Cancro. 2006. Space, selection, and Probability: Contemporary Mathematics Series. A. G. V. Marlos, and D. S. surveillance: setting boundaries with BLyS. J. Immunol. 176: 6405–6410. P. Richards, eds. American Mathematical Society, Providence, RI, p. 1–22. 62. Sasaki, Y., S. Casola, J. L. Kutok, K. Rajewsky, and M. Schmidt-Supprian. 2004. 41. Palanichamy, A., J. Barnard, B. Zheng, T. Owen, T. Quach, C. Wei, R. J. Looney, TNF family member B cell-activating factor (BAFF) receptor-dependent and I. Sanz, and J. H. Anolik. 2009. Novel human transitional B cell populations -independent roles for BAFF in B cell physiology. J. Immunol. 173: 2245–2252. revealed by B cell depletion therapy. J. Immunol. 182: 5982–5993. 63. Batten, M., J. Groom, T. G. Cachero, F. Qian, P. Schneider, J. Tschopp, 42. Merchant, M. S., B. A. Garvy, and R. L. Riley. 1996. Autoantibodies inhibit J. L. Browning, and F. Mackay. 2000. BAFF mediates survival of peripheral interleukin-7-mediated proliferation and are associated with the age-dependent immature B lymphocytes. J. Exp. Med. 192: 1453–1466. loss of pre-B cells in autoimmune New Zealand Black Mice. Blood 87: 3289– 64. Mackay, F., S. A. Woodcock, P. Lawton, C. Ambrose, M. Baetscher, 3296. P. Schneider, J. Tschopp, and J. L. Browning. 1999. Mice transgenic for BAFF 43. Wang, J., Q. Lin, H. Langston, and M. D. Cooper. 1995. Resident bone marrow develop lymphocytic disorders along with autoimmune manifestations. J. Exp. macrophages produce type 1 interferons that can selectively inhibit interleukin- Med. 190: 1697–1710. 7-driven growth of B lineage cells. Immunity 3: 475–484. 65. Scapini, P., F. Bazzoni, and M. A. Cassatella. 2008. Regulation of B-cell- 44. Ryan, D. H., B. L. Nuccie, I. Ritterman, J. L. Liesveld, and C. N. Abboud. 1994. activating factor (BAFF)/B lymphocyte stimulator (BLyS) expression in hu- Cytokine regulation of early human lymphopoiesis. J. Immunol. 152: 5250– man neutrophils. Immunol. Lett. 116: 1–6. 5258. 66. Huard, B., T. McKee, C. Bosshard, S. Durual, T. Matthes, S. Myit, O. Donze, 45. Garcia-Romo, G. S., S. Caielli, B. Vega, J. Connolly, F. Allantaz, Z. Xu, C. Frossard, C. Chizzolini, C. Favre, et al. 2008. APRIL secreted by neutrophils M. Punaro, J. Baisch, C. Guiducci, R. L. Coffman, et al. 2011. Netting neu- binds to heparan sulfate proteoglycans to create plasma cell niches in human trophils are major inducers of type I IFN production in pediatric systemic lupus mucosa. J. Clin. Invest. 118: 2887–2895. erythematosus. Sci. Transl. Med. 3: 73ra20. 67. Matthes, T., I. Dunand-Sauthier, M. L. Santiago-Raber, K. H. Krause, O. Donze, 46. Lande, R., D. Ganguly, V. Facchinetti, L. Frasca, C. Conrad, J. Gregorio, J. Passweg, T. McKee, and B. Huard. 2011. Production of the plasma-cell sur- S. Meller, G. Chamilos, R. Sebasigari, V. Riccieri, et al. 2011. Neutrophils ac- vival factor a proliferation-inducing ligand (APRIL) peaks in myeloid precursor tivate plasmacytoid dendritic cells by releasing self-DNA-peptide complexes in cells from human bone marrow. Blood 118: 1838–1844. systemic lupus erythematosus. Sci. Transl. Med. 3: 73ra19. 68. Liu, Y. J., C. Barthe´le´my, O. de Bouteiller, C. Arpin, I. Durand, and 47. Blomberg, S., M. L. Eloranta, B. Cederblad, K. Nordlin, G. V. Alm, and J. Banchereau. 1995. Memory B cells from human tonsils colonize mucosal L. Ro¨nnblom. 2001. Presence of cutaneous interferon-alpha producing cells in epithelium and directly present antigen to T cells by rapid up-regulation of B7-1 patients with systemic lupus erythematosus. Lupus 10: 484–490. and B7-2. Immunity 2: 239–248. 48. Farkas, L., K. Beiske, F. Lund-Johansen, P. Brandtzaeg, and F. L. Jahnsen. 2001. 69. Moser, K., K. Tokoyoda, A. Radbruch, I. MacLennan, and R. A. Manz. 2006. Plasmacytoid dendritic cells (natural interferon- alpha/beta-producing cells) Stromal niches, plasma cell differentiation and survival. Curr. Opin. Immunol. accumulate in cutaneous lupus erythematosus lesions. Am. J. Pathol. 159: 237– 18: 265–270. 243. 70. Scapini, P., B. Nardelli, G. Nadali, F. Calzetti, G. Pizzolo, C. Montecucco, and 49. Tucci, M., C. Quatraro, L. Lombardi, C. Pellegrino, F. Dammacco, and M. A. Cassatella. 2003. G-CSF-stimulated neutrophils are a prominent source of F. Silvestris. 2008. Glomerular accumulation of plasmacytoid dendritic cells in functional BLyS. J. Exp. Med. 197: 297–302. active lupus nephritis: role of interleukin-18. Arthritis Rheum. 58: 251–262. 71. Knight, J. S., W. Zhao, W. Luo, V. Subramanian, A. A. O’Dell, S. Yalavarthi, 50. Kumagai, Y., O. Takeuchi, H. Kato, H. Kumar, K. Matsui, E. Morii, K. Aozasa, J. B. Hodgin, D. T. Eitzman, P. R. Thompson, and M. J. Kaplan. 2013. Pepti- T. Kawai, and S. Akira. 2007. Alveolar macrophages are the primary interferon- dylarginine deiminase inhibition is immunomodulatory and vasculoprotective in alpha producer in pulmonary infection with RNA viruses. Immunity 27: 240–252. murine lupus. J. Clin. Invest. 123: 2981–2993.