The Journal of Immunology

A Unique Hybrid Renal Mononuclear Phagocyte Activation Phenotype in Murine Systemic Lupus Erythematosus Nephritis

Ramalingam Bethunaickan,* Celine C. Berthier,† Meera Ramanujam,* Ranjit Sahu,* Weijia Zhang,‡ Yezou Sun,‡ Erwin P. Bottinger,‡ Lionel Ivashkiv,x Matthias Kretzler,† and Anne Davidson*

Renal infiltration with mononuclear cells is associated with poor prognosis in systemic lupus erythematosus. A renal macrophage/ dendritic cell signature is associated with the onset of nephritis in NZB/W mice, and immune-modulating therapies can reverse this signature and the associated renal damage despite ongoing immune complex deposition. In nephritic NZB/W mice, renal F4/80hi/ CD11cint macrophages are located throughout the interstitium, whereas F4/80lo/CD11chi dendritic cells accumulate in perivas- cular lymphoid aggregates. We show here that F4/80hi/CD11cint renal macrophages have a Gr1lo/Ly6Clo/VLA4lo/MHCIIhi/CD43lo/ CD62Llo phenotype different from that described for inflammatory macrophages. At nephritis onset, F4/80hi/CD11cint cells up- regulate cell surface CD11b, acquire cathepsin and matrix metalloproteinase activity, and accumulate large numbers of auto- phagocytic vacuoles; these changes reverse after the induction of remission. Latex bead labeling of peripheral blood Gr1lo monocytes indicates that these are the source of F4/80hi/CD11cint macrophages. CD11chi/MHCIIlo dendritic cells are found in the kidneys only after proteinuria onset, turnover rapidly, and disappear rapidly after remission induction. expression profiling of the F4/80hi/CD11cint population displays increased expression of proinflammatory, regulatory, and tissue repair/ degradation-associated at nephritis onset that reverses with remission induction. Our findings suggest that mononuclear phagocytes with an aberrant activation profile contribute to tissue damage in lupus nephritis by mediating both local inflamma- tion and excessive tissue remodeling. The Journal of Immunology, 2011, 186: 4994–5003.

enal mononuclear phagocyte infiltration in sys- immune complexes and mononuclear cell renal infiltration are temic lupus erythematosus (SLE) nephritis is associated required for renal damage to occur (4). R with poor disease outcome (1) and is linked intricately Macrophages have distinct activation patterns and functions with the activation of the endothelium, renal chemokine pro- depending on the stimuli to which they are exposed (5–8). Clas- duction, proteinuria, complement activation, progressive hypoxia, sically activated (M1) macrophages, induced by TNF-a/IFN-g or and hypertension (2, 3). In NZB/W SLE-prone mice both Fc by IFN-b, have enhanced Ag presentation capabilities and secrete -mediated interaction of mononuclear cells with renal reactive oxygen intermediates, NO, chemokines, and cytokines that induce a Th1 response. Alternatively activated (M2) macro- phages, activated by IL-4 and IL-13 or induced in response to helminths, secrete protective cytokines such as IL-10 and IL1-RA. *Center for Autoimmunity and Musculoskeletal Diseases, Feinstein Institute for These cells express the mannose receptor, Arginase 1, and the † Medical Research, Manhasset, NY 11030; Department of Medicine, University of transcription factors YM1 and FIZZ-1 and are also capable of Michigan Medical School, Ann Arbor, MI 48109; ‡Department of Medicine, Mount Sinai Medical Center, New York, NY 10029; and xWeill Cornell Graduate School of tissue repair (9). Other macrophage activation patterns represent Medical Sciences, New York, NY 10021 a mixed phenotype likely resulting from simultaneous exposure Received for publication September 13, 2010. Accepted for publication February 13, to inflammatory and homeostatic/suppressive factors in vivo. Of 2011. relevance to SLE, macrophages exposed to immune complexes This work was supported by the Alliance for Lupus Research (to A.D., E.B., and and TLR agonists are characterized by an IL-10hi/IL-12lo pheno- W.Z.), National Institutes of Health Grants R01 AI059636-01A2 (to A.D. and L.I.) and R01 DK085241-01 (to C.C.B., M.K., E.B., W.Z., and A.D.), the New York SLE type termed M2b (10). These cells have Ag presentation capa- Foundation, and Rheuminations (to R.B.). bilities and may continue to produce inflammatory cytokines, but The data presented in this article have been submitted to the Om- they are not involved in tissue repair. A somewhat different phe- nibus under accession number GSE27045. notype is induced by ITAM cross-linking in vitro (11). These cells Address correspondence and reprint requests to Dr. Anne Davidson, Feinstein In- do not produce inflammatory cytokines but markedly upregulate stitute for Medical Research, 350 Community Drive, Manhasset, NY 11030. E-mail IL-10 and CXCL13, a chemokine that facilitates lymphoid neo- address: [email protected] genesis. The online version of this article contains supplemental material. Normal mouse kidneys have a heterogeneous population of mono- Abbreviations used in this article: GBM, glomerular basement membrane; IRF, IFN hi hi hi regulatory ; M1, classically activated; M2, alternatively activated; nuclear phagocytes. The dominant F4/80 /CX3CR1 /MHCII / int int MFI, mean fluorescence intensity; MMP, matrix metalloproteinase; SAM, signifi- CD11b /CD11c renal macrophage population forms a network cance analysis of microarrays; SEM, scanning electron microscopy; SLE, systemic throughout the renal interstitium (12, 13). Normal human kid- lupus erythematosus. neys have a similar interstitial network of CD68+ cells (14). In Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00 mice, these intrinsic renal cells use dendrites to sample their local www.jimmunol.org/cgi/doi/10.4049/jimmunol.1003010 The Journal of Immunology 4995 environment, and they can present Ag, but they are poor NO pro- Materials and Methods ducers and are only weakly phagocytic. This has suggested that Mice they are more like dendritic cells (12, 13, 15, 16), and their no- NZB/NZW F1 females were purchased from The Jackson Laboratory (Bar menclature is now a subject of some debate (15). A minor population Harbor, ME). Urine was tested weekly for proteinuria by dipstick (Fisher hi hi of mononuclear cells in normal kidneys is CD11b /CD62L / Scientific, Pittsburgh, PA). After fixed proteinuria appeared, a single dose Ly6Chi/Gr1int/CCR2hi/F4/80lo/CD11clo/MHCIIlo/CD86lo; during of cyclophosphamide and six doses of CTLA4 Ig and anti-CD154 Ab were acute renal inflammation or ischemia, this macrophage population administered as described previously (19). Mice were sacrificed 3–4 wk increases markedly and secretes proinflammatory cytokines (17). after entering complete remission (#30 mg/dl proteinuria on at least two hi hi lo hi occasions 7 d apart). Nephritic mice were sacrificed 2–6 wk after pro- Small populations of CD11b /CD11c and CD11b /CD11c re- teinuria onset but before the development of terminal renal failure. Young nal dendritic cells have been described (18) but have not been well NZB/W F1 mice were sacrificed at 8–16 wk of age. Kidneys also were characterized. obtained from 129/SvJ mice in which anti-glomerular basement membrane In this study, we analyzed the characteristics and function of the (GBM) disease was induced 14 d previously exactly as described (21). major renal mononuclear phagocyte populations from young and Flow cytometry and cell sorting nephritic NZB/W mice and from mice in which remission of nephritis was induced with a combination of cyclophosphamide Single-cell suspensions were prepared from perfused kidneys and stained and costimulatory blockade (19). We show that CD11b+/CD11cint/ with Abs to CD11c, CD11b, CD4, CD8, B220, PDCA, VLA4, MHCII, hi hi CD80, CD86, CD43, CD44, CD62L, Gr1, Ly6C (BD Pharmingen), and F4/80 (F4/80 ) cells, which are the dominant mononuclear cells F4/80 (Caltag-Invitrogen, Carlsbad CA), as described previously (20, 22). in normal renal interstitium, acquire an activated phenotype dur- CD11b+/CD11chi and CD11b+/CD11cint/F4/80hi cells were isolated using ing active nephritis that reverses upon remission induction. During a BD FACSAria IIu cell sorter. A dump gate for CD49b, CD4, CD8, CD5, active SLE nephritis, F4/80hi cells are a major renal source of and B220 was used to exclude NK cells and lymphocytes, and DAPI was used to exclude dead cells. Isolated cells were .90% pure. Cytospin several proinflammatory cytokines and chemokines (20), but they preparations of isolated cells were stained with Wright-Giemsa stain. also secrete molecules associated with tissue protection and repair The proteolytic activities of cathepsins (B, L, and S) and matrix met- that in excess may contribute to tissue degradation. In contrast, alloproteinases (MMPs; MMP2, MMP3, MMP9, and MMP13) were CD11b+/CD11chi/F4/80lo (CD11chi) cells, which are rare in nor- measured in vivo using activatable fluorescent sensors ProSense 680 and mal kidneys, appear in large numbers in lymphoid aggregates MMPSense 680 (Visen Medical, Bedford, MA; Ref. 23). A total of 5 nmol of probe in 100 ml sterile PBS was injected i.v. 24 h before sacrifice, and during nephritis (20) and disappear upon remission induction. kidney cells were stained with Abs to CD11b, CD11c, and F4/80 as above. These two major populations of infiltrating mononuclear cells are F4/80hi cells from nephritic and young mice were sorted into ProSensehi topographically and functionally distinct from each other and are and ProSenselo populations and subjected to real-time PCR for IL-10, phenotypically different from the proinflammatory Gr1hi macro- IKKε, Mincle, MMP14, and TIMP-1. Results were normalized to b-. phages that infiltrate kidneys during acute inflammatory glomer- Electron microscopy studies ulonephritis or renal ischemia (17). Our data call into question the use of acute inflammatory models of nephritis in the study of For transmission electron microscopy studies, sorted renal cell pellets were chronic SLE and point to a unique hybrid activation phenotype of fixed with 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M sodium cacodylate buffer, postfixed with 1% osmium tetroxide followed by 2% mononuclear phagocytes in chronic SLE nephritis that is associ- uranyl acetate, dehydrated with ethanol, and embedded in LX-112 resin ated closely with proteinuria onset and tissue damage and whose (LADD Research Industries, Burlington VT). Ultrathin sections were cut on reversal is associated with remission induction. a Reichert Ultracut UCT, stained with uranyl acetate followed by lead

FIGURE 1. Flow cytometric analysis of renal mononuclear phagocytes. The gating strategy for CD11bhi cells is shown in Supplemental Fig. 1. A, Gating strategy for the F4/80hi (right), F4/80lo (lower left), and CD11chi (upper left) subpopulations of CD11bhi cells. B, The increase in the percentage of CD11bhi cells as a percentage of live cells in the kidneys of nephritic mice is accounted for mainly by an increase in the F4/80hi population and an influx of CD11chi cells. There is an increase in lymphocyte infiltration (percentage of cells in the dump gate) in nephritic mice. Cellular infiltration resolves upon remission induction (young and remission mice compared with nephritic mice; six to eight mice per group; *p , 0.002, xp , 0.05). C, CD43 staining of the CD11chi (black) and F4/80hi (gray) populations from young and nephritic mice. Light gray is an isotype control for CD43. D, VLA4 and MHCII staining of the CD11chi (black) and F4/80hi (gray) populations from young and nephritic mice. Light gray is an isotype control for VLA4. E, Gr1 staining shows a small population of neutrophils in nephritic mice (arrow). Plots are representative of at least five mice per group. Experiments were repeated more than three times. 4996 ACTIVATED RENAL MACROPHAGES IN MURINE SLE

FIGURE 2. Analysis of renal mononuclear phagocyte turnover. Mice were fed BrdU from days 2 to 40 after remission induction (B–F) or for .60 d after remission induction followed by BrdU withdrawal at relapse (G). B, C, F, and G represent cells gated on CD11b and F4/80hi. A, BrdU uptake over time in NZB/W mice (four mice per group). B and C, Downregulation of CD11b expression 21 d after remission (C) compared with the nephritic control (B) occurs in both the BrdU+ (newly arrived) and BrdU2 (resident) population of F4/80hi cells. D, Twenty-one days after remission, all of the CD11chi cells (two shaded histograms) are BrdU+, whereas many F4/80hi cells (two black histograms) are still BrdU2. E, Twenty-one days after remission, CD11b+ cells in the spleens are nearly all BrdU+, compared with only a small proportion of B cells and T cells. F, Forty days after remission induction, one third of F4/80hi cells are still BrdU2. G, Twenty-one days after relapse, upregulation of CD11b is evident in both the BrdU+ (resident) and the BrdU2 (newly arrived) population of F4/80hi cells. Data for B–G are representative of four mice per group. Experiments were repeated once. citrate, and viewed on a JEOL 1200EX transmission electron microscope at the initiation of the remission induction regimen and sacrificed mice 21 80 kV. and 40 d later. In alternate experiments, we fed mice with BrdU for For scanning electron microscopy (SEM) studies, sorted cell pellets were .60 d from the time of the remission until relapse. BrdU then was stop- quick-fixed in 1% osmium tetroxide, 0.1 M sodium cacodylate, 0.2 M ped, and the mice were sacrificed 21 d later. sucrose, and 5 mM MgCl2 (pH 4) (SEM buffer) for 10 s, followed by 2.5% glutaraldehyde in SEM buffer. Secondary fixation was performed with 1% Labeling of peripheral blood monocytes with fluorescent beads osmium tetroxide in SEM buffer for 30 min, and pellets were dehydrated A total of 100 ml per mouse of a 1:10 dilution of yellow-green fluorescent with ethanol. Pellets then were subjected to critical point dry using liquid 1.0-mm latex particles (Polysciences, Warrington, PA) was injected i.v. carbon dioxide in a Tousimis Samdri 795 critical point drier (Rockville, Peripheral blood and organs were harvested 3 d later and subjected to flow MD) followed by sputter coat with gold–palladium in a Denton Vacuum cytometry and immunohistochemical analysis. In alternate experiments, Desk II sputter coater (Cherry Hill, NJ). Samples were examined on a isolated kidney cells were exposed in vitro to latex beads for 30 or 45 min JEOL JSM6400 scanning electron microscope (Peabody, MA), using an and analyzed by flow cytometry. accelerating voltage of 10 kV. RNA amplification, labeling, and GeneChip array BrdU incorporation hybridization Mice were loaded i.p. with 1 mg BrdU (Sigma-Aldrich, St. Louis, MO), We performed Affymetrix microarray analyses of cDNA obtained from followed by feeding with water containing 0.8 mg/ml BrdU for up to 60 d. sorted NZB/W F4/80hi cells from six young kidneys, seven nephritic Groups of four mice were sacrificed at intervals, and kidney cells were kidneys, and four kidneys harvested 3–4 wk after the induction of complete analyzed by flow cytometry as above, with the addition of anti-BrdU Ab remission. RNA was synthesized from sorted cell populations using a (BD Pharmingen) according to the manufacturer’s instructions. Similar PicoPure RNA Isolation kit (Arcturus Molecular Device Corporation, experiments were performed in mice fed with BrdU for 7 or 15 d after Sunnyvale, CA). The quality of total RNA was assessed using a NanoDrop nephritis onset. To determine whether remission was associated with the ND-1000 spectrophotometer (Thermo Scientific, Wilmington, DE) and migration of the cells out of the kidneys, we began BrdU feeding 48 h after verified on a NanoRNA chip (Agilent, Santa Clara, CA).

FIGURE 3. In vivo labeling of PBMCs 3 d after the administration of latex beads. A and B, Uptake of fluorescent beads occurs in the peripheral blood CD11bhi/F4/80lo/Gr1lo population. C and D, Bead- laden cells are restricted to the F4/80hi population of renal CD11bhi cells. E–G, No beads are detected in the renal node (E) or thymus (F), but there is some traf- ficking to CD11b+ cells in the spleen (G). H, Immu- nofluorescence analysis of kidneys shows beads within F4/80hi cells (original magnification 310 and 340). Data are representative of five mice per group. Experi- ments were repeated once. The Journal of Immunology 4997

FIGURE 4. Morphology of renal mononuclear phagocytes by Wright-Giemsa staining and SEM. A, Gating strategy for the sorting of F4/80hi (dashed), CD11chi (gray), and F4/80lo (black) populations (see also Supplemental Fig. 1). B, The F4/80lo population consists of round macrophage-like cells and some neutrophils (data not shown). C and D, There is an increase in volume and marked vacuolization of the F4/80hi population in nephritic mice (D) compared with that of young mice (C). E, The CD11chi population has a veiled morphology characteristic of dendritic cells. Original magnification 3100 for Wright-Giemsa stains and 3750 and 310,000 for SEM. E–H, Representative transmission electron microscopy visualization of F4/80hi cells from young (F) and nephritic mice (G, H) show multiple vacuoles in the nephritic mice, many of which have double membranes (inset from H). Data are representative of three to four mice per group. Experiments were repeated once.

A total of 30 ng of total RNA was amplified and labeled with biotin cells, nephritic versus remission F4/80hi cells, and prenephritic versus re- (Ovation Biotin system; NuGEN Technologies, San Carlos, CA), frag- mission F4/80hi cells). Genes regulated between two groups with a q value mented, and hybridized to Affymetrix Mouse Genome 430 Plus 2.0 Gene- (false discovery rate) ,5% were considered significant, and gene lists of Chip arrays (Santa Clara, CA). Arrays were washed, stained, and scanned interest were uploaded for literature-based pathway analyses on Geno- by GeneChip Scanner 3000 7G according to the Affymetrix Expression matix BiblioSphere Pathway Edition software (http://www.genomatix.de). Analysis Technical Manual. The data discussed in this publication have been deposited in the National CEL files were processed using the GenePattern analysis pipeline Center for Biotechnology Information’s Gene Expression Omnibus (http:// (http://www.genepattern.com). Normalization was performed using the www.ncbi.nlm.nih.gov/geo/) and are accessible through accession number robust multichip average method and version 10 of the Mouse Gene GSE27045. custom CDF annotation from Brain Array (http://brainarray.mbni.med. umich.edu/Brainarray/default.asp). The resulting data were log2-trans- Real-time PCR formed. Of the 16,539 mouse genes represented on the Affymetrix Gen- eChip, 14,415 were expressed above the 27 poly-A Affymetrix control Verification of selected differentially expressed genes was performed using expression baseline and used for further analyses. Unpaired significance Custom TaqMan Expression Arrays (Applied Biosystems) as per the manu- analysis of microarrays (SAM) was performed for each statistical com- facturer’s instructions. Five internal controls identified by the microarray parison between the relevant groups (prenephritic versus nephritic F4/80hi analysis were tested, and the data were normalized to the mean of the three

FIGURE 5. Functional studies of renal mononuclear phagocytes. A, In vivo cathepsin labeling of young F4/80hi cells (green), nephritic F4/80hi cells (red), and nephritic CD11chi cells (blue). CD11b2 lymphocytes are shown in gray. B, MFI values for cathepsin activity (**p , 0.02, xp , 0.01). C,Invivo MMP activity of young F4/80hi cells (green), nephritic F4/80hi cells (red), and nephritic CD11chi cells (blue). CD11b2 lymphocytes are shown in gray. D, MFI values for MMP labeling (*p , 0.05). E, In vitro phagocytosis assay shows that the majority of bead uptake is in the renal F4/80hi population (red) compared with the CD11chi population (green). Beads were incubated with renal cells for 45 min in this experiment. F, RNA from sorted F4/80hi ProSensehi (gray, young; black, nephritic) and ProSenselo (white, nephritic) cells was analyzed by real-time PCR. Data are representative of at least four mice per group. Experiments were repeated twice. 4998 ACTIVATED RENAL MACROPHAGES IN MURINE SLE most stable control genes (Gus-B, Oaz1, and b-actin) and then to the mean kidneys evaluated by the quantification of renal cells expressing of the young controls. To identify significantly differentially expressed B220, CD5, CD4, CD8, or CD49b also was observed in the re- genes between groups, SAM (version 2.1) (24) was carried out. Genes mission mice compared with that in the controls (Fig. 1B, Sup- regulated between two groups with a q value (false discovery rate) ,5% were considered significant. plemental Fig. 1). NZB/W kidneys contain several distinct mononuclear Results phagocyte populations Remission induction using cyclophosphamide and Renal CD11b+ and CD11c+ mononuclear phagocytes in the kid- costimulatory blockade neys of young, nephritic, and remission mice were analyzed by We have reported previously that the disappearance of proteinuria flow cytometry using the gating strategy shown in Supplemental after remission induction therapy is associated with decreased Fig. 1. The major (F4/80hi) population in young mice had a phe- glomerular damage and interstitial inflammation scores by light notype similar to that of renal dendritic cells, namely, CD11bhi/ microscopy and with markedly increased longevity (19, 20). Re- CD11cint/F4/80hi/Gr1lo/Ly6Clo/VLA4lo/MHCIIhi/CD43lo/CD62Llo mission of nephritis in the group of mice studied here was char- (17). A smaller population of cells (F4/80lo) was CD11bhi/CD11cint/ acterized by a disappearance of proteinuria from $300 to #30 F4/80lo/Gr1lo. A third, infrequent population (CD11chi) was CD11bhi/ mg/dl on repeated measurements over a 2–4 wk period, whereas CD11chi/F4/80lo/Gr1lo/MHCIIhi/CD43hi/CD62Llo (Fig. 1). A fourth untreated mice remained proteinuric. Blood urea nitrogen levels population, comprising ,2% of all of the cells in the lymphocyte/ remained within the normal range in the treated mice (22.0 6 8.1 monocyte gate, was CD11blo/CD11chi/F4/80lo/Gr1lo and may cor- versus 22.8 6 2.6 in young mice; p = NS), whereas they increased respond to the CD103+ population found in other tissues (25). in controls (49.6 6 25.0; p , 0.001 versus both young and remission Finally, ,0.5% of all of the cells in the lymphocyte/monocyte gate mice). A decrease in interstitial lymphocytic infiltration of the had a plasmacytoid dendritic cell phenotype (B220+/PDCAhi).

FIGURE 6. Literature-based analysis using Genomatix BiblioSphere software of the genes that were regulated in F4/80hi cells from nephritic kidneys (n = 7) compared with both young (n = 6) and remission (n = 4) kidneys (with stringent filter criteria, q value ,0.001, fold change $1.5 and #0.7 for the upregulated and downregulated genes, respectively). The picture shows the 101 genes that were cocited in PubMed abstracts in the same sentence linkedby a function word (B3 filter). Yellow to red represents the genes that are upregulated and yellow to blue represents the genes that are downregulated in nephritic compared with young mice. The Journal of Immunology 4999

The latter two populations did not increase in frequency during the our conclusion that the activation of this cell population occurs in disease course and were not studied further. CD11bhi cells in- situ in the kidneys. creased by 3- to 4-fold in nephritic compared with young kidneys The three major renal mononuclear phagocyte populations (Fig. 1A,1B). The F4/80hi subset increased by 2- to 3-fold and have different morphologies upregulated the surface expression of CD11b, F4/80, Ox40L, and CD80 (20). These cells were also MHCIIhi/VLA4lo/CD43lo (Fig. Cells were sorted into the subpopulations shown in Fig. 4A. Iso- lo 1C,1D) The CD11chi subset underwent 10-fold expansion (Fig. 1A, lated F4/80 cells were round, with large nuclei and little cyto- 1B) and was MHCIIlo/VLA4hi/CD43hi (Fig. 1C,1D). The F4/80lo plasm, which is consistent with a macrophage phenotype (Fig. hi subset expanded 3-fold during nephritis but remained a minor 4B). In young and remission mice, F4/80 cells were large and population (Fig. 1A,1B). Few neutrophils (CD11bhi/F4/802/Gr1hi) round with kidney-shaped nuclei and small dendrites (Fig. 4C). In hi were detected either before or during nephritis (Fig. 1E). The renal nephritic mice, the F4/80 cells became swollen and contained CD11b+ population of mice in complete remission resembled that multiple double-membrane vacuoles characteristic of autophago- of young prenephritic mice (Fig. 1A,1B). Importantly, we could not cytic vacuoles (Fig. 4D,4G,4H). Oil Red O stain was negative for hi identify the CD11bhi/CD11clo/F4/80lo/VLA4hi/Ly6Chi population lipid (data not shown). CD11c cells were large with many den- reported by others in acute models of renal inflammation, includ- drites (Fig. 4E) and had a veiled morphology by electron mi- ing anti-GBM disease (Supplemental Fig. 1H,1I). Conversely, the croscopy consistent with a myeloid dendritic cell phenotype (Fig. CD11chi population that we identified in nephritic NZB/W mice 4E). was not increased in frequency in mice with anti-GBM disease Renal mononuclear phagocyte subsets have different (data not shown). functional properties Cell turnover rates of the major mononuclear phagocyte The F4/80hi population in nephritic mice had a significant increase populations in both cathepsin and MMP activity; no increase in either ca- Using BrdU incorporation, we determined that the CD11chi pop- ulation had a renal half-life of 6.5 d in young mice. In contrast, the F4/80hi population had a half-life of 16 d (Fig. 2A); this did not increase in nephritic mice. To analyze the fate of CD11bhi cells during remission, we administered BrdU starting 48 h after the initiation of remission induction. At day 21, 64.6 6 8.7% of the F4/80hi population was still BrdU2 (Fig. 2C,2D), indicating that the cells had not proliferated or been replaced by newly formed cells. In contrast, splenic CD11bhi cells were nearly all BrdU+ (Fig. 2E). Even 40 d after remission induction, 32.7 6 6.7% of the renal F4/80hi population remained BrdU2 (Fig. 2F). Both BrdU+ and BrdU2 populations in remission kidneys downregulated CD11b expression (mean fluorescence intensity [MFI] 41,900 6 15,338 in remission mice versus 74,860 6 16,222 in nephritic controls; p , 0.04, compare Fig. 2B with Fig. 2C,2F). In contrast, at day 21, all of the remaining CD11chi cells were BrdU+, con- firming the rapid turnover of this population (Fig. 2D). In an al- ternate experiment, we administered BrdU for .60 d to mice in remission and stopped it when the mice relapsed. Twenty-one days later, all of the CD11chi cells were BrdU2 (data not shown), indicating that they had been replaced, whereas 34.1 6 11.7% of F4/80hi cells were still BrdU+ (Fig. 2G), indicating that they had been present at the time of relapse. Both BrdU+ and BrdU2 populations had upregulated CD11b (MFI in relapsed mice 72,900 6 8661, p = NS, compared with nephritic controls; com- pare Fig. 2G with Fig. 2C,2F). In sum, these data suggest that the activation of F4/80hi cells, as assessed by the upregulation of CD11b, occurs in situ and that this activation reverses upon re- mission induction. In contrast, the renal CD11chi population turns over rapidly and rapidly clears from the kidneys upon remission induction. Origins of renal F4/80hi macrophages We used latex beads to label PBMCs in nephritic NZB/W mice. Beads were endocytosed into peripheral Gr1lo cells (Fig. 3A,3B) and 3 d later were detected only within the renal F4/80hi pop- ulation (Fig. 3C,3D,3H). No beads were detected in renal lymph nodes (Fig. 3E) or thymus (Fig. 3F), although a small number was FIGURE 7. One-way cluster analysis of genes with significantly altered lo detected in the spleen (Fig. 3G). Thus, peripheral blood Gr1 expression in the PCR confirmation set (see Supplemental Table V). Gene hi monocytes are the precursors of the renal F4/80 population. expression was normalized to the mean of three housekeeping genes as CD11b expression was higher on renal cells than that on matched described in Materials and Methods. Significantly up- or downregulated peripheral blood cells (data not shown), which is consistent with (.2-fold) genes by SAM are shown. 5000 ACTIVATED RENAL MACROPHAGES IN MURINE SLE thepsin or MMP activity was detected in the F4/80lo population, genes were differentially regulated between young and remission and there was only a modest increase in cathepsin activity in kidneys. We examined the 378 coregulated genes in more detail CD11chi cells (Fig. 5A–D). Phagocytic activity was a property of using the natural language processing tool (Genomatix Biblio- the F4/80hi population (3.3 61.3% of the F4/80hi population Sphere) (26, 27). Transcriptional network analysis integrating versus 1.3 6 0.3% of the CD11chi population after 30 min; p , these 378 differentially regulated mRNAs with literature mining 0.05; Fig. 5E) and did not change significantly in nephritic mice (PubMed database) and automated promoter analysis highlighted (data not shown). ITGAM, Stat3, TIMP-1, IL-10, and CCL2 among important reg- Affymetrix microarray analyses were performed using cDNA ulatory nodes (Fig. 6, Supplemental Table IV). from sorted NZB/W F4/80hi cells. We found 694 genes differen- We chose a panel of 75 genes with varying degrees of upreg- tially expressed in nephritic compared with young kidneys (q ulation at nephritis onset and a set of control genes for confirmatory value ,0.01, fold change $1.5 and #0.7 for the upregulated and studies using quantitative real-time PCR. Six additional genes downregulated genes, respectively) (Supplemental Table I) and specific for the M1/M2 macrophage genotype but without sig- 794 genes differentially expressed in remission compared with nificant differences by microarray also were included. Using cut- nephritic kidneys (Supplemental Table II). A total of 378 genes off values of 2-fold overexpression and a q value ,0.05, 49/75 overlapped between the two comparisons and thus were bio- genes were significantly different between young and nephritic markers of nephritis activity (Supplemental Table III). Only 12 mice, and 50/75 were significantly different between nephritic

Table I. Top 10 canonical and signal transduction pathways (Genomatix literature mining) as assessed by Genomatix Pathway System software

Number of Number of Canonical and Signal Transduction Pathways (Number of Genes in the Pathway) p Value Regulated Genes Expected Genes From the 694 genes regulated in macrophages from nephritic compared with young kidneys Canonical pathways 1 Extrinsic prothrombin activation pathway (14) 2.57 3 1024 5 0.63 2 IL-12–mediated signaling events (63) 1.83 3 1023 9 2.86 3 FOXM1 transcription factor network (42) 2.41 3 1023 7 1.90 4 Fibrinolysis pathway (16) 4.81 3 1023 4 0.73 5 Inhibition of matrix metaloproteinases (9) 6.27 3 1023 3 0.41 6 Aurora B signaling (41) 9.45 3 1023 6 1.86 7 Costimulatory signal during T cell activation (20) 1.11 3 1022 4 0.91 8 Phospholipase Cd in phospholipid associated cell signaling (4) 1.15 3 1022 2 0.18 9 PLK1 signaling events (47) 1.81 3 1022 6 2.13 10 Basic mechanism of action of ppara pparb(d) and pparg and 1.86 3 1022 2 0.23 effects on gene expression (5)

Signal transduction pathways 1 IL-10 (265) 1.66 3 1029 40 14.27 2 NF-kB (1341) 3.13 3 1029 118 72.24 3 TLR (398) 8.85 3 1029 50 21.44 4 Integrin (458) 2.00 3 1028 54 24.67 5 IL-1 (623) 3.26 3 1028 66 33.56 6 STAT (1002) 4.90 3 1028 92 53.98 7 IL-1R (103) 8.80 3 1028 21 5.55 8 IL-6 (IFN, b2) (581) 1.68 3 1027 61 31.30 9 TNF (TNF superfamily, member 2) (1035) 2.35 3 1027 92 55.75 10 TNFR superfamily (826) 4.71 3 1027 77 44.49

From the 794 genes regulated in macrophages from remission compared with nephritic kidneys Canonical pathways 1 Aurora B signaling (41) 4.15 3 10210 16 2.46 2 PLK1 signaling events (47) 4.39 3 1029 16 2.82 3 FOXM1 transcription factor network (42) 4.76 3 1027 13 2.52 4 Cdk regulation of DNA replication (18) 4.54 3 1025 7 1.08 5 Aurora A signaling (31) 5.32 3 1025 9 1.86 6 IL-12–mediated signaling events (63) 6.36 3 1025 13 3.78 7 Alternative NF-kB pathway (6) 3.72 3 1023 3 0.36 8 IL-12 signaling mediated by STAT4 (36) 4.73 3 1023 7 2.16 9 TCR signaling in naive CD8+ T cells (56) 5.40 3 1023 9 3.36 10 Costimulatory signal during T cell activation (20) 5.46 3 1023 5 1.20

Signal transduction pathways 1 Aurora kinase (85) 3.67 3 10211 25 5.40 210 2 Cell division cycle 2 (G1 to S and G2 to M) (216) 1.18 3 10 41 13.72 3 Signal transduction and activator of transcription (1002) 4.13 3 10210 111 63.66 4 Polo-like kinase 1 (109) 2.33 3 1029 26 6.93 5 Integrin (458) 2.98 3 1028 60 29.10 6 IL-12 (217) 2.18 3 1027 35 13.79 7 JAK (560) 3.49 3 1027 66 35.58 8 IL-1 (623) 4.36 3 1027 71 39.58 9 IL-10 (265) 5.16 3 1027 39 16.84 10 Guanine nucleotide exchange factor VAV (211) 1.01 3 1026 33 13.41 The Journal of Immunology 5001 mice and treated mice. There was 94% overlap between these two population is derived from circulating Gr1lo monocytes that mi- gene sets (Fig. 7, Supplemental Table V). grate into nephritic kidneys where they become activated in situ Pathway analysis using the Genomatix Pathway System (28) was and form a cuff around the glomeruli and around the edges of performed on the 694- and 794-gene lists. The top 10 canonical large interstitial inflammatory infiltrates (20). Nephritic NZB/W pathways and signal transduction pathways (Genomatix literature kidneys additionally accumulate a population of CD11bhi/CD11chi/ mining) are displayed in Table I. Prothrombotic and tissue deg- F4/80int/MHCIIlo/Gr1lo/VLA4hi/CD43hi/CD62Llo myeloid dendritic radation pathways were among the most significantly regulated cells that are located only within lymphoid aggregates (20) and canonical pathways in nephritic compared with young kidney disappear upon remission induction. Differences in the location of macrophages, and IL-10 was the top significant signal trans- dendritic cells and macrophages similarly have been observed in duction pathway. A total of 28 genes of this pathway were altered human SLE kidney biopsies (14). Notably, the CD11bhi/Ly6Chi/ in the macrophages from remission compared with nephritic VLA4hi/CD43lo/ CD62Lhi inflammatory macrophage population kidneys. The top functions identified using Ingenuity Pathway (17) that is found in most models of acute renal inflammation Analysis software included inflammatory response, cell-to-cell remains a minor population in NZB/W mice and in two other signaling and interaction, hematological system development and murine SLE models that we have studied (30, 31). function, and immune cell trafficking. The two major renal mononuclear phagocyte populations are The gene expression studies revealed a mixed phenotype with functionally distinct. The F4/80hi population is phagocytic and the expression of proinflammatory and anti-inflammatory genes as acquires both MMP and cathepsin activity at nephritis onset. The well as genes involved in tissue repair. We have confirmed the cells also increase their volume due to the accumulation of auto- increased expression of CD11b and MMP14 previously (20, phagocytic vacuoles. Intracellular degradation of cytoplasmic pro- 29). To confirm increased protein expression of CXCL13, IL-10, teins and organelles by autophagy is required for type I IFN re- and IKKε, we performed immunohistochemistry, flow cytometry, sponses to viral infections (32) but is also a physiologic response and Western blot analysis (Fig. 8). To determine if the presence of to inflammation or cellular stress and can be induced in human cathepsin activity could distinguish populations of F4/80hi cells macrophages in vitro by exposure to growth factors and IL-10 with different gene expression profiles, we sorted ProSensehi and (33). Although this process conserves energy, protects the cell ProSenselo subsets from nephritic mice and ProSenselo cells from from death, and helps to clear apoptotic bodies (34), it may en- young mice and performed real-time PCR for the expression of hance the presentation of nuclear and cytosolic Ags (35, 36) and IL-10 (regulatory), IKKε, Mincle (inflammatory), TIMP-1, and facilitate inflammation (37). In contrast, the CD11chi population is MMP14 (repair). Cells from young mice had low expression of all weakly phagocytic, downregulates the expression of MHC class of the markers, whereas increased expression of all of the markers II, and does not acquire MMP activity or accumulate autopha- was observed in both ProSensehi and ProSenselo populations from gocytic vesicles. Downregulation of MHC class II has been ob- nephritic mice with no difference in the expression between the served in dendritic cells exposed to hypoxia (38) or activated by two populations (Fig. 5F). In accord with these data, we were cytokines, including IL-10 (39). unable to distinguish F4/80hi subpopulations on the basis of in- To further understand why infiltrating F4/80hi cells are associ- tracellular IKKε staining (Fig. 8E). ated with tissue damage, we performed gene expression analysis on isolated cell populations from young, nephritic, and remission Discussion mice. One goal of our gene expression analysis was to gain insight Renal mononuclear phagocyte infiltration is associated with poor into the mechanisms by which renal F4/80hi macrophages con- disease outcome in SLE nephritis (1). In this study of NZB/W tribute to the pathogenesis of lupus nephritis by determining lupus-prone mice, we show that there is an expansion of two major whether they exhibit features of classical or alternative activation. populations of mononuclear phagocytes in nephritic kidneys. The In addition to immune complexes and TLR ligands that promote renal F4/80hi population is the major population in normal and an M2b phenotype, lupus nephritic macrophages are exposed to prenephritic kidneys and has a CD11bhi/CD11cint/F4/80hi/Gr1lo/ opposing signals from TNF-a and tissue damage, and we hy- Ly6Clo/VLA4lo/MHCIIhi/CD80hi/CD43lo/CD62Llo phenotype. This pothesized that they may exhibit a mixed or potentially novel

FIGURE 8. Protein expression of analytes of interest. A and B, CXCL13 staining was observed in periglomerular infiltrates of nephritic mice (A) but not young controls (B). Original magnifica- tion 310. C, Periglomerular F4/80hi macro- phages were preset only in nephritic mice (same mouse as A). Original magnification 310. D, Intracellular IKKε staining was significantly in- creased in F4/80hi cells of nephritic mice (blue) compared with young mice (red) and isotype controls (purple). E, IKKε staining was uniform in both ProSensehi and ProSenselo populations from nephritic mice. F and G, Western blot analysis for IL-10 in sorted F4/80hi renal cells from young NZB/W mice (lane 1), mice with new onset proteinuria (lane 2), and mice with established proteinuria (lanes 3 and 4). Red bands show the actin control. H, IL-10/actin ra- tio. 5002 ACTIVATED RENAL MACROPHAGES IN MURINE SLE phenotype. Our data demonstrate that extensive changes in gene mixed phenotype reflects an evolution of one type of macrophage expression occur at proteinuria onset and reverse with remission from another. induction. Surprisingly, we found that the expression profile has Our data, in sum, point to a unique activation profile of mac- features that are consistent with a regulatory phenotype (7, 40), rophages and dendritic cells in SLE kidneys quite different from including the expression of IL-10, PPAR-g, antiapoptotic genes, that found in acute renal inflammation or ischemia. We surmise that cytokine inhibitors such as IL-1RA and IL-18 binding protein, this phenotype results from chronic exposure of SLE kidney efferocytosis receptors such as Mer kinase and transglutaminase 2, macrophages to immune complexes, cytokines such as TNF-a, and anti-inflammatory receptors such as the PGE2 receptor. In- TLR signals, fibrinogen, dead and dying cells, hypoxia, and other duction of SOCS3 is associated with a failure to upregulate IL-6 danger signals. The resultant mixed phenotype is associated with and TNF-a. There is an increase in CD11b, an integrin required chronic and progressive renal injury. The M1-like characteristics for cell activation, adhesion, and phagocytosis but whose en- of kidney macrophages likely contribute to the local inflammatory gagement also can mediate negative feedback on the TLR path- process, but it is less clear whether the M2 components of the ways and induce the production of IL-10 (11, 41). phenotype are protective or pathogenic. It is possible that the The expression profile also displays a proinflammatory pheno- expression of M2 genes that are used to resolve inflammation or type. Proinflammatory genes include Trem-1, an ITAM-containing repair tissues in other settings “inadvertently” contributes to patho- cell surface molecule that amplifies TLR-mediated responses (42), genesis by promoting excessive tissue remodeling and linked pro- activating Fc receptors and formyl peptidyl receptors, C-type lec- liferation. tins associated with tissue injury such as Mincle (43) and oxidized Another striking finding is that pathway analysis identified low-density lipoprotein receptor 1 (44, 45), chemokines such as central roles for genes with allelic polymorphisms that have been CCL2, CCL5, CCL7, and CCL8, and costimulatory molecules such linked to SLE, including Il10, Itgam, Ptpn22, Atg5, and IRF7. The as CD80 and CD40. We also detected the increased expression of pathogenic role of IL-10–secreting macrophages in chronic in- C3, complement factor B, and properdin, suggesting a role for the flammation has been suggested by our previous studies in NZB/W activated F4/80hi population in the amplification of the alternative mice with proliferative glomerulonephritis accelerated by type I complement pathway (46); this, together with the increased pro- IFNs. In these studies, the depletion of IL-10–producing phago- duction of coagulation factors, could propagate endothelial damage cytic renal macrophages prevented ongoing crescent formation and thrombosis. The cells also upregulate CXCL13, a chemokine and renal damage (29). IL-10 also may augment mesangial pro- that can be induced by IL-10 as well as by IL-1 and TNF-a (11, 47, liferation in inflamed kidneys (29, 55). IL-10 antagonism is 48) and that helps to orchestrate lymphoid neogenesis in inflamed therapeutic in NZB/W mice (56), and a small clinical study of tissues. Finally, we observed the expression of MMPs and cathepsin anti–IL-10 Ab in showed salutary effects (57). It will be proteases that are associated with tissue repair during the resolution worthwhile in future studies to examine the function of the other of acute inflammation but could mediate tissue degradation if molecules identified in genetic studies, specifically in macro- chronically expressed. phages and dendritic cells. Many of the proinflammatory genes expressed by activated renal There has been much recent interest in manipulating macro- macrophages in nephritic NZB/W kidneys are induced in human phage programming and migration in acute renal diseases. Ther- macrophages by TNF-a (L. Ivashkiv, unpublished observations) apeutic targeting of key molecules identified here in the murine but not by the ITAM cross-linking that induces a regulatory models should help to define the function of the major renal macrophage phenotype. TNF-a, made by multiple cell types in mononuclear phagocyte cell types found in SLE kidneys. An inflamed kidneys including nonlymphoid renal cells (20), is enhanced understanding of macrophage and dendritic cell phe- pathogenic in the effector phase of SLE nephritis (49). One gene notypes and function in patients with SLE nephritis will be needed induced by TNF-a and highly upregulated in F4/80hi cells from to appropriately target these cell populations in humans. nephritic kidneys is IKKε (IKKi), a kinase that enhances the in- duction of a subset of NF-kB–regulated genes and phosphorylates Acknowledgments the IFN regulatory transcription factors (IRFs) IRF-3 and IRF-7, We thank Haiou Tao for technical assistance, Stella Stefanova for assistance ε major inducers of type I IFNs. IKK phosphorylates STAT1 and with cell sorting, and Dr Yong Du and Dr. Chandra Mohan for the gift of enhances the formation of the ISGF3 transcriptional complex (50) kidneys from anti-GBM mice. that helps to mediate some of the inhibitory and antiproliferative ε effects of type I IFNs (51, 52). IKK is dispensable for acute in- Disclosures flammatory responses (53). In contrast, its expression is markedly The authors have no financial conflicts of interest. upregulated in F4/80hi macrophages derived from chronically inflamed adipose tissue of mice fed a high-fat diet, and it is es- sential for maintaining low-grade inflammation and propagating References ε obesity in these mice (53). IKK -deficient mice are less sensitive 1. Hill, G. S., M. Delahousse, D. Nochy, P. Re´my, F. Mignon, J. P. Me´ry, and to arthritis induction in a passive transfer model; the inhibition of J. Barie´ty. 2001. Predictive power of the second renal biopsy in lupus nephritis: type I IFNs and IKKε is therapeutic in this disease (54). These significance of macrophages. Kidney Int. 59: 304–316. observations suggest a role for IKKε in chronic inflammation. 2. Sean Eardley, K., and P. Cockwell. 2005. Macrophages and progressive tubu- lointerstitial disease. Kidney Int. 68: 437–455. We have confirmed the increased protein expression of CD11b, 3. Knight, S. F., J. E. Quigley, J. Yuan, S. S. Roy, A. Elmarakby, and J. D. Imig. CD80, CD86, MMP2, MMP14, IKKε, CXCL13, and IL-10 and the 2008. Endothelial dysfunction and the development of renal injury in sponta- presence of functional cathepsin and MMP activity in the F4/80hi neously hypertensive rats fed a high-fat diet. Hypertension 51: 352–359. 4. Bergtold, A., A. Gavhane, V. D’Agati, M. Madaio, and R. Clynes. 2006. FcR- macrophage population of nephritic mice in this and previous bearing myeloid cells are responsible for triggering murine lupus nephritis. J. studies (Figs. 5, 8, Refs. 20, 29). Further work will be needed to Immunol. 177: 7287–7295. confirm the protein expression of other genes in the expression 5. Mosser, D. M. 2003. The many faces of macrophage activation. J. Leukoc. Biol. profile and to determine whether the macrophages express the 73: 209–212. 6. Mantovani, A., A. Sica, S. Sozzani, P. Allavena, A. Vecchi, and M. Locati. 2004. genes comprising the hybrid profile simultaneously, there are The chemokine system in diverse forms of macrophage activation and polari- different subsets expressing different aspects of the profile, or the zation. Trends Immunol. 25: 677–686. The Journal of Immunology 5003

7. Mosser, D. M., and J. P. Edwards. 2008. Exploring the full spectrum of mac- of systemic lupus erythematosus nephritis in NZM2410 mice. Arthritis Rheum. rophage activation. Nat. Rev. Immunol. 8: 958–969. 62: 1457–1468. 8. Hume, D. A. 2008. Differentiation and heterogeneity in the mononuclear 32. Orvedahl, A., and B. Levine. 2009. Autophagy in Mammalian antiviral immu- phagocyte system. Mucosal Immunol. 1: 432–441. nity. Curr. Top. Microbiol. Immunol. 335: 267–285. 9. Gordon, S., and F. O. Martinez. 2010. Alternative activation of macrophages: 33. Santiago-Schwarz, F., A. Valentino, C. Martin, and C. DiMaio. 2010. IL-10 mechanism and functions. Immunity 32: 593–604. interruption of dendritic cell (DC) differentiation during the monocyte-to-DC 10. Martinez, F. O., A. Sica, A. Mantovani, and M. Locati. 2008. Macrophage ac- transition is associated with increased levels of LC3, mature autophagosome tivation and polarization. Front. Biosci. 13: 453–461. formation, and survival of monocyte-macrophage-like cells. J. Immunol. 184: 11. Wang, L., R. A. Gordon, L. Huynh, X. Su, K. H. Park Min, J. Han, J. S. Arthur, 134.121. G. D. Kalliolias, and L. B. Ivashkiv. 2010. Indirect inhibition of Toll-like re- 34. Qu, X., Z. Zou, Q. Sun, K. Luby-Phelps, P. Cheng, R. N. Hogan, C. Gilpin, and ceptor and type I interferon responses by ITAM-coupled receptors and integrins. B. Levine. 2007. Autophagy gene-dependent clearance of apoptotic cells during Immunity 32: 518–530. embryonic development. Cell 128: 931–946. 12. Kru¨ger, T., D. Benke, F. Eitner, A. Lang, M. Wirtz, E. E. Hamilton-Williams, 35. Crotzer, V. L., and J. S. Blum. 2009. Autophagy and its role in MHC-mediated D. Engel, B. Giese, G. Mu¨ller-Newen, J. Floege, and C. Kurts. 2004. Identifi- antigen presentation. J. Immunol. 182: 3335–3341. cation and functional characterization of dendritic cells in the healthy murine 36. Mu¨nz, C. 2010. Selective macroautophagy for immunity. Immunity 32: 298– kidney and in experimental glomerulonephritis. J. Am. Soc. Nephrol. 15: 613– 299. 621. 37. Chang, Y. P., C. C. Tsai, W. C. Huang, C. Y. Wang, C. L. Chen, Y. S. Lin, 13. Soos, T. J., T. N. Sims, L. Barisoni, K. Lin, D. R. Littman, M. L. Dustin, and J. I. Kai, C. Y. Hsieh, Y. L. Cheng, P. C. Choi, et al. 2010. Autophagy facilitates P. J. Nelson. 2006. CX3CR1+ interstitial dendritic cells form a contiguous IFN-gamma-induced Jak2-STAT1 activation and cellular inflammation. J. Biol. network throughout the entire kidney. Kidney Int. 70: 591–596. Chem. 285: 28715–28722. 14. Segerer, S., F. Heller, M. T. Lindenmeyer, H. Schmid, C. D. Cohen, 38. Mancino, A., T. Schioppa, P. Larghi, F. Pasqualini, M. Nebuloni, I. H. Chen, D. Draganovici, J. Mandelbaum, P. J. Nelson, H. J. Gro¨ne, E. F. Gro¨ne, et al. S. Sozzani, J. M. Austyn, A. Mantovani, and A. Sica. 2008. Divergent effects of 2008. Compartment specific expression of dendritic cell markers in human hypoxia on dendritic cell functions. Blood 112: 3723–3734. glomerulonephritis. Kidney Int. 74: 37–46. 39. Choi, Y. E., H. N. Yu, C. H. Yoon, and Y. S. Bae. 2009. Tumor-mediated down- 15. Ferenbach, D., and J. Hughes. 2008. Macrophages and dendritic cells: what is regulation of MHC class II in DC development is attributable to the epigenetic the difference? Kidney Int. 74: 5–7. control of the CIITA type I promoter. Eur. J. Immunol. 39: 858–868. 16. Geissmann, F., C. Auffray, R. Palframan, C. Wirrig, A. Ciocca, L. Campisi, 40. Tabas, I. 2010. Macrophage death and defective inflammation resolution in E. Narni-Mancinelli, and G. Lauvau. 2008. Blood monocytes: distinct subsets, atherosclerosis. Nat. Rev. Immunol. 10: 36–46. how they relate to dendritic cells, and their possible roles in the regulation of 41. Han, C., J. Jin, S. Xu, H. Liu, N. Li, and X. Cao. 2010. Integrin CD11b nega- T-cell responses. Immunol. Cell Biol. 86: 398–408. tively regulates TLR-triggered inflammatory responses by activating Syk and 17. Li, L., L. Huang, S. S. Sung, A. L. Vergis, D. L. Rosin, C. E. Rose, Jr., P. I. Lobo, promoting degradation of MyD88 and TRIF via Cbl-b. Nat. Immunol. 11: 734– and M. D. Okusa. 2008. The chemokine receptors CCR2 and CX3CR1 mediate 742. monocyte/macrophage trafficking in kidney ischemia-reperfusion injury. Kidney 42. Bouchon, A., F. Facchetti, M. A. Weigand, and M. Colonna. 2001. TREM-1 Int. 74: 1526–1537. amplifies inflammation and is a crucial mediator of septic shock. Nature 410: 18. Kurts, C. 2006. Dendritic cells: not just another cell type in the kidney, but 1103–1107. a complex immune sentinel network. Kidney Int. 70: 412–414. 43. Yamasaki, S., E. Ishikawa, M. Sakuma, H. Hara, K. Ogata, and T. Saito. 2008. 19. Schiffer, L., J. Sinha, X. Wang, W. Huang, G. von Gersdorff, M. Schiffer, Mincle is an ITAM-coupled activating receptor that senses damaged cells. Nat. M. P. Madaio, and A. Davidson. 2003. Short term administration of costimula- Immunol. 9: 1179–1188. tory blockade and cyclophosphamide induces remission of systemic lupus 44. Pyz, E., A. S. Marshall, S. Gordon, and G. D. Brown. 2006. C-type lectin-like erythematosus nephritis in NZB/W F1 mice by a mechanism downstream of receptors on myeloid cells. Ann. Med. 38: 242–251. renal immune complex deposition. J. Immunol. 171: 489–497. 45. Hu, C., B. Y. Kang, J. Megyesi, G. P. Kaushal, R. L. Safirstein, and J. L. Mehta. 20. Schiffer, L., R. Bethunaickan, M. Ramanujam, W. Huang, M. Schiffer, H. Tao, 2009. Deletion of LOX-1 attenuates renal injury following angiotensin II in- M. P. Madaio, E. P. Bottinger, and A. Davidson. 2008. Activated renal macro- fusion. Kidney Int. 76: 521–527. phages are markers of disease onset and disease remission in lupus nephritis. J. 46. Holers, V. M. 2008. The spectrum of complement alternative pathway-mediated Immunol. 180: 1938–1947. diseases. Immunol. Rev. 223: 300–316. 21. Xie, C., R. Sharma, H. Wang, X. J. Zhou, and C. Mohan. 2004. Strain distri- 47. Ishikawa, S., S. Nagai, T. Sato, K. Akadegawa, H. Yoneyama, Y. Y. Zhang, bution pattern of susceptibility to immune-mediated nephritis. J. Immunol. 172: N. Onai, and K. Matsushima. 2002. Increased circulating CD11b+CD11c+ 5047–5055. dendritic cells (DC) in aged BWF1 mice which can be matured by TNF-alpha 22. Ramanujam, M., X. Wang, W. Huang, Z. Liu, L. Schiffer, H. Tao, D. Frank, J. Rice, into BLC/CXCL13-producing DC. Eur. J. Immunol. 32: 1881–1887. B. Diamond, K. O. Yu, et al. 2006. Similarities and differences between selective 48. Perrier, P., F. O. Martinez, M. Locati, G. Bianchi, M. Nebuloni, G. Vago, and nonselective BAFF blockade in murine SLE. J. Clin. Invest. 116: 724–734. F. Bazzoni, S. Sozzani, P. Allavena, and A. Mantovani. 2004. Distinct tran- 23. Nahrendorf, M., F. K. Swirski, E. Aikawa, L. Stangenberg, T. Wurdinger, scriptional programs activated by interleukin-10 with or without lipopolysac- J. L. Figueiredo, P. Libby, R. Weissleder, and M. J. Pittet. 2007. The healing charide in dendritic cells: induction of the B cell-activating chemokine, CXC myocardium sequentially mobilizes two monocyte subsets with divergent and chemokine ligand 13. J. Immunol. 172: 7031–7042. complementary functions. J. Exp. Med. 204: 3037–3047. 49. Aringer, M., G. Steiner, W. B. Graninger, E. Ho¨fler, C. W. Steiner, and 24. Tusher, V. G., R. Tibshirani, and G. Chu. 2001. Significance analysis of J. S. Smolen. 2007. Effects of short-term infliximab therapy on autoantibodies in microarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. systemic lupus erythematosus. Arthritis Rheum. 56: 274–279. USA 98: 5116–5121. 50. Tenoever, B. R., S. L. Ng, M. A. Chua, S. M. McWhirter, A. Garcı´a-Sastre, and 25. Ginhoux, F., K. Liu, J. Helft, M. Bogunovic, M. Greter, D. Hashimoto, J. Price, T. Maniatis. 2007. Multiple functions of the IKK-related kinase IKKepsilon in N. Yin, J. Bromberg, S. A. Lira, et al. 2009. The origin and development of interferon-mediated antiviral immunity. Science 315: 1274–1278. nonlymphoid tissue CD103+ DCs. J. Exp. Med. 206: 3115–3130. 51. Laver, T., S. E. Nozell, and E. N. Benveniste. 2008. IFN-beta-mediated in- 26. Wiggins, J. E., S. R. Patel, K. A. Shedden, M. Goyal, B. L. Wharram, S. Martini, hibition of IL-8 expression requires the ISGF3 components Stat1, Stat2, and M. Kretzler, and R. C. Wiggins. 2010. NFkappaB promotes inflammation, co- IRF-9. J. Interferon Cytokine Res. 28: 13–23. agulation, and fibrosis in the aging glomerulus. J. Am. Soc. Nephrol. 21: 587– 52. Tsuno, T., J. Mejido, T. Zhao, H. Schmeisser, A. Morrow, and K. C. Zoon. 2009. 597. IRF9 is a key factor for eliciting the antiproliferative activity of IFN-alpha. J. 27. Schmid, H., A. Boucherot, Y. Yasuda, A. Henger, B. Brunner, F. Eichinger, Immunother. 32: 803–816. A. Nitsche, E. Kiss, M. Bleich, H. J. Gro¨ne, et al; European Renal cDNA Bank 53. Chiang, S. H., M. Bazuine, C. N. Lumeng, L. M. Geletka, J. Mowers, (ERCB) Consortium. 2006. Modular activation of nuclear factor-kappaB tran- N. M. White, J. T. Ma, J. Zhou, N. Qi, D. Westcott, et al. 2009. The protein scriptional programs in human diabetic nephropathy. Diabetes 55: 2993–3003. kinase IKKepsilon regulates energy balance in obese mice. Cell 138: 961–975. 28. Frisch, M., B. Klocke, M. Haltmeier, and K. Frech. 2009. LitInspector: literature 54. Corr, M., D. L. Boyle, L. Ronacher, N. Flores, and G. S. Firestein. 2009. Syn- and signal transduction pathway mining in PubMed abstracts. Nucleic Acids Res. ergistic benefit in inflammatory arthritis by targeting I kappaB kinase epsilon and 37(Web Server issue): W135–W140. interferon beta. Ann. Rheum. Dis. 68: 257–263. 29. Triantafyllopoulou, A., C. W. Franzke, S. V. Seshan, G. Perino, G. D. Kalliolias, 55. Kalechman, Y., U. Gafter, T. Weinstein, A. Chagnac, I. Freidkin, A. Tobar, M. Ramanujam, N. van Rooijen, A. Davidson, and L. B. Ivashkiv. 2010. Pro- M. Albeck, and B. Sredni. 2004. Inhibition of interleukin-10 by the immuno- liferative lesions and metalloproteinase activity in murine lupus nephritis me- modulator AS101 reduces mesangial cell proliferation in experimental mesan- diated by type I interferons and macrophages. Proc. Natl. Acad. Sci. USA 107: gioproliferative glomerulonephritis: association with dephosphorylation of 3012–3017. STAT3. J. Biol. Chem. 279: 24724–24732. 30. Ramanujam, M., P. Kahn, W. Huang, H. Tao, M. P. Madaio, S. M. Factor, and 56. Ishida, H., T. Muchamuel, S. Sakaguchi, S. Andrade, S. Menon, and M. Howard. A. Davidson. 2009. Interferon-alpha treatment of female (NZW x BXSB)F(1) 1994. Continuous administration of anti-interleukin 10 antibodies delays onset of mice mimics some but not all features associated with the Yaa . Arthritis autoimmunity in NZB/W F1 mice. J. Exp. Med. 179: 305–310. Rheum. 60: 1096–1101. 57. Lauwerys, B. R., N. Garot, J. C. Renauld, and F. A. Houssiau. 2000. Interleukin- 31. Ramanujam, M., R. Bethunaickan, W. Huang, H. Tao, M. P. Madaio, and 10 blockade corrects impaired in vitro cellular immune responses of systemic A. Davidson. 2010. Selective blockade of BAFF for the prevention and treatment lupus erythematosus patients. Arthritis Rheum. 43: 1976–1981.