IL-6 B and Κ Foxj1-Mediated Antagonism of NF- Restraint of B

Restraint of B Cell Activation by Foxj1-Mediated Antagonism of NF- κB and IL-6

This information is current as Ling Lin, Steven L. Brody and Stanford L. Peng of September 27, 2021. J Immunol 2005; 175:951-958; ; doi: 10.4049/jimmunol.175.2.951 http://www.jimmunol.org/content/175/2/951 Downloaded from

References This article cites 32 articles, 14 of which you can access for free at: http://www.jimmunol.org/content/175/2/951.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 27, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2005 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Restraint of B Cell Activation by Foxj1-Mediated Antagonism of NF-␬B and IL-61

Ling Lin,* Steven L. Brody,† and Stanford L. Peng2*‡

The forkhead transcription factor Foxj1 inhibits spontaneous autoimmunity, in part by antagonizing NF-␬B activation in T cells. We demonstrate here that Foxj1 also inhibits humoral immune responses intrinsically in B cells; Foxj1 deficiency in B cells results in spontaneous and accentuated germinal center formation, associated with the development of pathogenic autoantibodies and accentuated responses to immunizations—all reflecting excessive activity of NF-␬B and its target gene IL-6, and correlating with a requirement for Foxj1 to regulate the inhibitory NF-␬B component I␬B␤. Thus, Foxj1 restrains B cell activation and the maturation of humoral responses, demonstrating a critical role for at least this forkhead transcription factor in the regulation of B lymphocyte homeostasis. The Journal of Immunology, 2005, 175: 951–958. Downloaded from ystemic autoimmune syndromes like lupus reflect patho- date the B cell-intrinsic function(s) of Foxj1, particularly its rela- genic, hyperactive T and B cell responses that culminate in tionship to Ag-specific Ab responses. We find that Foxj1 modu- S the production of pathogenic autoantibodies, such as anti- lates germinal center B cell formation in vivo via its ability to dsDNA (1). Recent studies suggest that deficient functions in fork- inhibit the NF-␬B-regulated cytokine IL-6 (7, 8). Foxj1 appears to head transcription factors, such as Foxj1 and Foxo3a, predispose to prevent inappropriate B cell responses at least in part by antago- such diseases, because their activities are significantly diminished nizing NF-␬B target genes like IL-6, which would otherwise prop- http://www.jimmunol.org/ in lupus lymphocytes and their deficiency results in spontaneous, agate pathogenic autoimmune responses through dysregulated B multisystem autoimmune syndromes characterized by autoreac- cell hyperactivation. tive, hyperactivated T cells (2, 3). Presumably, such hyperactivated T cells can promote pathogenic autoantibody production by Materials and Methods overexpressing costimulatory molecules and/or effector cytokines, Mice resulting in excessive autoreactive B cell activation. However, a 129, BALB/c, C57BL/6, C57BL/6-IgH , C57BL/6-CD45.1, BXSB, F1 both Foxj1 and Foxo3a are also expressed in B cells (2), and a (NZW ϫ NZB), MRL/ϩ, MRL/lpr, and athymic C57BL/6-nu/nu mice potential immunoregulatory role for the forkhead genes in B cells (The Jackson Laboratory), Rag-2Ϫ/Ϫ (Taconic Farms), and C57BL/6 Foxj1 Ϫ/Ϫ mice (9) were maintained under specific pathogen-free conditions at remains as yet undefined. by guest on September 27, 2021 Pathogenic autoantibody responses require somatic hypermuta- the Washington University School of Medicine. As judged by microsatel- lite markers, mice mutant for Foxj1, which is located on chromosome 11 tion and class switching of natural autoantibodies, such as in the at 78.0 cM, were homozygotic for C57BL/6 loci, including marker context of T-dependent germinal centers, to develop high-affinity, D11Mit333 (66.0 cM; 11qter is ϳ80.0 cM), indicating a Ͻ15 cM residual pathogenic activities, e.g., the production of anti-dsDNA from 129 contribution on chromosome 11. Foxj1 ϩ/ϩ and Ϫ/Ϫ fetal liver chi- 3 germline-encoded anti-ssDNA specificities (1, 4). Most circulating meras (FLCs) were generated in irradiated Rag-2-deficient hosts as previously described (2). For B cell-only chimeras (BOC), fetal livers from natural autoantibodies are produced by natural Ab-secreting B Foxj1 ϩ/ϩ vs Ϫ/Ϫ (CD45.2ϩIgHb) embryos were adoptively transferred cells, such as B1 cells, or by plasma cells that have not undergone into irradiated C57BL/6-nu/nu hosts, and reconstitution of the peripheral B either affinity maturation or class switch recombination (5, 6). As cell lineage was allowed for 8–12 wk. Then, splenic B cell populations a result, the entry of an autoreactive B cell, when activated, into a were purified by negative selection against CD43 (Miltenyi Biotec), and adoptively transferred into C57BL/6-IgHa (for serological studies) or germinal center reaction, therefore, may contribute to and/or un- ϩ C56BL/6-CD45.1 (for flow cytometric studies) animals (one spleen derlie systemic autoimmune diseases like lupus (4, 6). equivalent, ϳ30–40 million B cells, per recipient), and animals were stud- To investigate the role of Foxj1 in B cells, we initiated a series ied 1–2 wk thereafter, as indicated in the text. All experiments were per- of experiments, including adoptive transfers, to isolate and eluci- formed in compliance with the relevant laws and institutional guidelines, as overseen by the Animal Studies Committee of the Washington University School of Medicine.

* Division of Rheumatology and †Division of Pulmonary and Critical Care Medicine, In vivo studies Department of Internal Medicine and ‡Department of Pathology and Immunology, Washington University School of Medicine, St. Louis, MO 63110 Immunizations with 4-hydroxy-3-nitrophenylacetyl (NP)-chicken ␥ globulin and NP-Ficoll were performed with 50 ␮g of Ag in PBS (without Received for publication March 10, 2005. Accepted for publication May 13, 2005. adjuvant) administered i.p. as previously described (4). Where indicated, The costs of publication of this article were defrayed in part by the payment of page animals were treated with neutralizing anti-IL-6 (MP5-20F3; BD Pharm- charges. This article must therefore be hereby marked advertisement in accordance ingen) or control rat IgG Ab, 5 mg i.p. three times per week (10); or with with 18 U.S.C. Section 1734 solely to indicate this fact. RELA/p65 antisense (5Ј-GAAACAGATCGTCCATGGT) or mismatch 1 This work was supported in part by the Rheumatic Diseases, Siteman Cancer, Di- (5Ј-GGAACAGTTCGTCTATGGC) oligonucleotides, 800 ␮g i.v. daily abetes Research and Training, and the Digestive Diseases Research Core (DK52574) (11). ELISA-based serological assessments, including determination of anti- Centers of the Washington University School of Medicine, as well as grants from the ssDNA using sheared salmon sperm DNA, anti-dsDNA using Crithidia National Institutes of Health (AI057471 and AI061478 to S.L.P.), Arthritis Founda- luciliae immunoﬂuorescence, IgK rheumatoid factor (RF) activity using tion, and the Lupus Research Institute. S.L.P. is supported in part by an Arthritis Investigator Award from the Arthritis Foundation. 2 Address correspondence and reprint requests to Dr. Stanford L. Peng, Washington 3 Abbreviations used in this paper: FLC, fetal liver chimera; BOC, B cell-only chi- University School of Medicine, Campus Box 8045, Clinical Sciences Research Building mera; NP, 4-hydroxy-3-nitrophenylacetyl; PNA, peanut agglutinin; RF, rheumatoid 6617, 660 South Euclid Avenue, St. Louis, MO 63110. E-mail address: [email protected] factor.

Table I. B cell characteristics of Foxj1 BOCsa

B Cells (B220ϩ) NP-Binding GC-Like (PNAϩB220ϩ)

Splenocytes Percentage of Chimeric Percentage of Chimeric Percentage of Chimeric Donor n (ϫ10Ϫ6) splenocytes (%) splenocytes (%) B cells (%)

WT 5 67.0 Ϯ 2.5 62.2 Ϯ 1.5 2.2 Ϯ 0.16 2.2 Ϯ 0.32 2.6 Ϯ 0.38 3.9 Ϯ 1.1 2.7 Ϯ 0.3 Ϯ 0.13 5.6 Ϯ 0.32b 20.6 Ϯ 2.4b 52 Ϯ 4.3b 2.8 ءKO 5 68.8 Ϯ 3.8 64.6 Ϯ 4.2 9.9 Ϯ 0.76

a BOCs were generated first by chimerization of nu/nu animals by Foxj1 ϩ/ϩ or Ϫ/Ϫ fetal livers (CD45.2), the B cells of which were then adoptively transferred into wild-type C57BL/6-CD45.1 animals. Shown are flow cytometric data (means Ϯ SD) of representative spleens from such BOCs, quantifying the subpopulations indicated, after immunization with NP-Ficoll. Chimeric percentages indicate the percentage of CD45.2-staining cells of the subpopulations indicated. b Significant differences between wild-type (W/T) and knockout (K/O) donor cohorts ( p Ͻ 0.001).

Ig␭ IgGs, and anti-hapten responses using NP-BSA and TNP-BSA, were Table I . The PCR products were cloned into pCR2.1-TOPO (Invitrogen performed as described (4, 12). Germinal centers were identified by stain- Life Technologies), and their sequence was determined by routine ing frozen spleen sections with peanut agglutinin (PNA)-FITC, as previ- sequencing. ously described (4). Results In vitro B cell studies Spontaneous high-affinity autoantibodies in Foxj1 deficiency

B cell purification and stimulation was performed similarly to a previous Downloaded from study (13). Briefly, for FLCs and unmanipulated mice, naive-enriched B Foxj1 is prominently expressed in naive B cells, where, like T cells were purified from spleens by negative selection against CD43 (Milte- cells, it is significantly down-regulated in response to activation nyi), and were further purified over a discontinuous Percoll gradient (70/ (Ref. 2 and Fig. 1A, p Ͻ 0.001 comparing any stimulation condi- 66/60/50%), with resting B cells isolated from the 66–70 interface. For tion to no treatment). In addition, naive, resting B cells from pre- BOCs, CD45.2ϩCD43Ϫ B cells were purified by flow cytometry, followed by Percoll gradient purification. Cells were cultured in RPMI 1640 medium diseased lupus-prone mice expressed significantly lower levels of supplemented with 10% FCS (BioWhittaker) and 100 U of penicillin/strep- Foxj1 (Fig. 1B; p Ͻ 0.001 comparing MRL/ϩ, BXSB, or NZB/W tomycin (Sigma-Aldrich), in the presence or absence of 25 ␮g/ml LPS to 129, BALB, or C57BL/6). Although we could not completely http://www.jimmunol.org/ (LPS; Sigma-Aldrich), 2 ␮g/ml anti-CD40 Ab (BD Pharmingen), 5–10 eliminate the possibility that such results simply reflect a global B ␮ g/ml anti-IgM (Jackson Immunologicals), 3 mM CpG-1 stimulatory cell hyperactivation seen in such autoimmune mouse strains (1), phosphorothioate oligonucleotide 5Ј-TCCATGACGTTCCTGACGTT, 100 ng/ml IFN-␥, 5 ng/ml TGF-␤, and/or 10 ng/ml IL-4 (PeproTech). Where indicated, phosphorothioate decoy or control decoy annealed NF-␬B oligonucleotides, which inhibit the activity of all NF-␬B subunits, were added at 10 ␮M (14). Real-time PCR detection of Foxj1, IL-6, CD80, CD86, ICOS, ICOSL, CD40, CD154, bcl-6, Blimp-1, IRF-4, Mitf, c-myc, Pax5, and Xbp1 were performed as described, with normalization against

␤-tubulin (13, 15). Western blot analyses of NF-␬B activities and proteins by guest on September 27, 2021 were performed as previously described (2). Reporter assays A promoter-reporter construct for IL-6 was constructed by PCR from C57BL/6 genomic DNA, using primers 5Ј-GGGGTACCATTCAAATC CTGTCATCCAGTAGAAGGGAG and 5Ј-GAAGATCTGAAAACCG GCAAGTGAGCAGATAGCACAGT, producing an ϳ1257-bp fragment corresponding to the putative promoter (Ϫ1058/ϩ199), flanked by KpnI and BglII restriction sites. The amplicon was cloned into the KpnI-BglII sites of TK-luc and then confirmed by routine sequencing, generating IL- 6-luc. Reporter assays involved Dual-Luciferase (Promega) assays using M12 murine B cell lymphoma cells, electroporated in the presence of 10 ␮g of IL-6-luc or 20 ␮gofNF-␬B-luc, 400 ng of pRL-CMV (Renilla luciferase control reporter; Promega), and 10 ␮g of pcDNA3 (Invitrogen Life Technologies) or pcDNA3-Foxj1, as described (2). Primary B cell transfection was performed by a modification of a previously described protocol (16): 107 cells, purified by negative selection and Percoll gradient centrifugation as described above, were incubated for 10 min at room temperature in 0.4-cm cuvettes in 400 ␮l of RPMI 1640 medium supplemented with 10% FCS, 100 ␮gofNF-␬B-luc, and 1 ␮g of pRL-CMV. The cells were then electroporated with a Bio-Rad electroporation system at 280 V, 975 ␮F, incubated at room temperature for 5 min, and then cultured in 5–10 ml of RPMI 1640 medium supplemented with 10% FCS and 25 ␮g/ml LPS. Dual-Luciferase assays were then performed after 4–5 h of incubation at 37°C. Somatic hypermutation assay To assess for somatic hypermutation, NP-binding B cells from immunized FLCs were purified by flow cytometry using NP-FITC-BSA (Biosearch

Technologies), and their genomic DNA amplified by PCR using VH186.2- ϩ ϩ Ϫ Ϫ and JH4-specific primers (17, 18). For BOCs, Foxj1 / or / fetal ϩ FIGURE 1. Expression pattern of Foxj1 in B cells. Foxj1 expression livers (CD45.2 ) were used first to chimerize nu/nu recipient mice, the splenic B cells of which were then adoptively transferred into wild-type was assessed by real-time PCR on cDNA from C57BL/6 naive B cells C57BL/6-CD45.1ϩ recipient mice. After immunization, NP-binding, treated with the indicated stimuli in vitro for 24 h (A) or naive B cells CD45.2ϩ B cells were purified by flow cytometry; representative rates of purified from 6-wk-old mice of the indicated strains (B). Error bars indicate chimerization, subdivided by representative B cell subsets, are shown in SDs for individually tested cells from three animals. The Journal of Immunology 953 Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 2. Ab responses in the absence of Foxj1. “Standard” FLCs (A–C) were generated by chimerization of Rag-deficient animals by Foxj1 ϩ/ϩ or Ϫ/Ϫ fetal livers; BOCs (D–F) were generated first by chimerization of nu/nu animals by Foxj1 ϩ/ϩ or Ϫ/Ϫ fetal livers (IgHb), the B cells of which were then adoptively transferred into wild-type C57BL/6-IgHa animals. By ELISA, sera from these chimeras were assessed for the presence of spontaneous IgG (FLC, A)orIgGb (BOC, D) anti-DNA or IgK RF autoantibodies at the indicated dilutions; hapten-specific IgG (FLC, B)orIgGb (BOC, E) isotype Abs 35 days after immunization at the indicated dilutions, or affinity maturation of IgG (FLC, C)orIgGb (BOC, F) anti-hapten responses at the times indicated after immunization by relative reactivity ϭ against NP3-BSA vs NP34-BSA at 1/100 dilution. In A and D, B and E, and C and F, n 4, 3, and 5, respectively, of each genotype, representative of at least three trial experiments containing the same number of animals. For reference, sera from 12-wk-old lupus-prone MRL/lpr mice developed anti-DNA and RF activities of 1.8–2.0 OD405 at 1/100 dilution, with titers of at least 1/2000 (data not shown). Dashed lines (A, B, D, and E) indicate the threshold for positivity, as determined by three SDs above the average OD405 generated by sera at 1/100 dilution from nonautoimmune BALB/c (A and D, which have lower spontaneous autoantibody activity than C57BL/6 animals) or unimmunized Foxj1 ϩ/ϩ FLC mice (B and E). these findings in highly purified, resting B cells raised the possi- (Fig. 2A), prompting our prior conclusion that Foxj1 plays a bility that loss-of-function in Foxj1 might contribute to B cell hy- dispensable role in B cell homeostasis (2). Nonetheless, close fur- peractivation and/or tolerance loss in autoimmunity, as it appears ther inspection of the data in repeated experiments revealed that to do for T cells (2). Foxj1 Ϫ/Ϫ sera consistently exhibited higher spontaneous activities However, analysis of Foxj1Ϫ/Ϫ FLCs revealed normal numbers of both anti-DNA and RF, at least as judged by OD values (Fig. of B1, marginal zone, T1, T2, and follicular B cells, and also 1A, comparing OD values of Ϫ/Ϫ vs ϩ/ϩ sera at 1/100–1/300 demonstrated that overall titers of spontaneous anti-DNA and RF dilution). This consistent, albeit subtle, dissociation between titers autoantibodies are generally comparable to their ϩ/ϩ counterparts and OD activities suggested that the Foxj1 Ϫ/Ϫ autoantibodies 954 ROLE OF Foxj1 IN B CELLS Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 3. Somatic hypermutation in the absence of Foxj1. NP-binding B cells were purified by flow cytometry from Foxj1 Ϫ/Ϫ vs Foxj1 ϩ/ϩ FLCs. A, Sequences of 10 individual clones obtained 35 days after immunization with NP-Ficoll from Foxj1 Ϫ/Ϫ vs ϩ/ϩ FLCs are shown, representative of 20 total clones sequenced for each genotype. Shaded regions are marked as corresponding to CDR1 and CDR2. The germline VH186.2 sequence is shown. B, Mutation frequency was calculated as the percentage of point mutations observed per base pair. Similar results were obtained with BOCs, analyzing NP-binding CD45.2ϩ B cells (BOC in B). were, in fact, of higher affinity, as opposed to higher quantity. anti-ssDNA, Abs are capable of exhibiting Crithidia immunore- Indeed, all Foxj1Ϫ/Ϫ sera tested positive for anti-dsDNA activity activity (19) and of mediating immune-complex disease (20), these by Crithidia luciliae immunofluorescence, in contrast to Foxj1ϩ/ϩ findings together suggested the development of only natural, low- sera (14 of 14 vs 0 of 12 sera, respectively; p Ͻ 0.0001). Because, affinity autoimmunity in Foxj1ϩ/ϩ FLCs, but the maturation of this in general, only high-affinity anti-dsDNA, but not lower affinity response to a high-affinity, potentially pathogenic autoimmunity in The Journal of Immunology 955

FIGURE 4. Germinal center formation in the absence of Foxj1 requires the RELA component of NF-␬B ϩ ϩ Ϫ Ϫ

and IL-6. Foxj1 / vs / FLCs were immunized Downloaded from with vehicle (PBS) vs TNP-Ficoll, and their spleens were assessed 21 days later for the presence of germinal centers by PNA staining: A shows representative staining patterns, summarized in B as the average number of germinal centers observed per spleen section (at least 20 sections per spleen were analyzed). C, Similar studies

were performed in BOCs (see Fig. 2). Similar results http://www.jimmunol.org/ were obtained with NP-Ficoll (data not shown). D, Foxj1 Ϫ/Ϫ BOCs were immunized with NP-Ficoll, and concomitantly treated with antisense or mismatch control oligonucleotides against the NF-␬B RELA component (p65), or a neutralizing Ab against murine IL-6 or isotype control. Spleens were assessed 21 days later for the presence of germinal centers by PNA staining. At the end of the experiment, (E) lysates of whole spleens were analyzed by Western blot for RELA and p50 expression levels. by guest on September 27, 2021

Foxj1Ϫ/Ϫ FLCs; thus, Foxj1 might also play a negative regulatory peared to be of higher affinity than Foxj1ϩ/ϩ responses, as judged role in B cells, as it does in T cells (2). by the relative activity of their sera against TNP3-BSA vs TNP34- BSA (Fig. 2C; p Ͻ 0.0001 comparing Foxj1 ϩ/ϩ to Ϫ/Ϫ affinities Accentuated humoral immunity in Foxj1 deficiency at days 28 and 35). Analogous findings were observed during Td To gain insight into the mechanisms by which Foxj1 might regu- immunization, which also elicited higher titer anti-hapten relate B cells, we turned to model immunogens and assessed the sponses in Foxj1Ϫ/Ϫ than Foxj1ϩ/ϩ FLCs (Fig. 2B and data not Ϫ/Ϫ response of Foxj1 FLCs to immunization with the T-indepen- shown; median titers 1/8100 vs 1/900 on TNP3-BSA, respectively; dent (Ti) Ag, NP-Ficoll, and the T-dependent (Td) Ag, NP-chicken p Ͻ 0.001) and appeared to be associated with a more rapid onset ␥ globulin. In response to NP-Ficoll, Foxj1 Ϫ/Ϫ FLCs mounted of affinity maturation in Foxj1Ϫ/Ϫ FLCs (Fig. 2C; p Ͻ 0.001 com- exaggerated anti-hapten responses, with higher titers (Fig. 2B; me- paring Foxj1 ϩ/ϩ to Ϫ/Ϫ affinities at day 14, p Ͻ 0.01 at day 28). Ͻ dian titers 1/400 vs 1/1600 on TNP3-BSA, respectively; p Therefore, we conclude that Ag-specific humoral immune re- 0.001). Interestingly, Foxj1 Ϫ/Ϫ Ti anti-NP responses also ap- sponses are accentuated in the absence of Foxj1. 956 ROLE OF Foxj1 IN B CELLS

FIGURE 5. Foxj1 regulates NF-␬B activity via I␬B␤. A, Spontaneous NF-␬B and NFAT activity was assessed in resting B cells from Foxj1Ϫ/Ϫ vs Foxj1ϩ/ϩ BOCs via NF-␬B- and NFAT-luciferase reporters. B, Western blot analysis of various NF-␬B pathway components was performed on resting B cells from Foxj1Ϫ/Ϫ vs Foxj1ϩ/ϩ BOCs. C, Real-time PCR analysis of mRNA of the indicated I␬B subunits was performed on resting B cells from Foxj1Ϫ/Ϫ vs Foxj1ϩ/ϩ Downloaded from BOCs. For A and C, SDs reﬂect triplicate samples, representative of at least three separate experiments. For B, results reﬂect combined B cell extracts from at least three animals of each genotype, representative of at least three separate experiments. http://www.jimmunol.org/ by guest on September 27, 2021

For evidence that somatic hypermutation, in fact, could be re- B cell hyperactivation in the absence of Foxj1 is B cell intrinsic sponsible for the apparently increased affinity maturation in Because the humoral responses of Foxj1Ϫ/Ϫ FLCs might simply be Ϫ/Ϫ Foxj1 FLCs, we analyzed IgH sequences. NP-binding B cells a secondary effect of the hyperactivated Foxj1Ϫ/Ϫ Th cells (2), we were purified by flow cytometry from Foxj1 ϩ/ϩ and Ϫ/Ϫ FLCs developed BOCs by performing fetal liver chimerization in athy- 35 days after immunization with NP-Ficoll, their genomic DNA mic nu/nu host mice, creating animals containing Foxj1 ϩ/ϩ vs b was amplified by PCR using VH182.6-specific primers, and se- Ϫ/Ϫ B cells (IgH ) that develop in the absence of T cell help. B quences were determined by routine cloning and sequencing (Fig. cells from these animals were then adoptively transferred into 3). Strikingly, mutations were abundant in sequences from wild-type C57BL/6-IgHa mice, which were subsequently immu- Foxj1Ϫ/Ϫ, but not Foxj1ϩ/ϩ FLCs, particularly the CDR1 and nized with Ti Ags and/or otherwise subjected to the assays in this CDR2 regions, reflecting an ϳ10-fold increase in somatic hyper- study. These Foxj1Ϫ/Ϫ BOCs recapitulated the Ab phenotypes of Ϫ Ϫ mutation frequency per base pair (Fig. 3, p Ͻ 0.0001). the Foxj1 / FLCs, including accentuated anti-hapten titers and Because somatic hypermutation, affinity maturation, as well as affinity, as well as germinal center formation, in response to NP- b isotype switching of B cells can generally be attributed to germinal Ficoll immunization (as judged by IgH -allotype-specific Ig as- Ϫ/Ϫ centers, these findings suggested that Foxj1 deficiency might ac- says; Figs. 2–4). In addition, all Foxj1 BOCs developed ϩ/ϩ centuate humoral responses by promoting germinal center forma- Crithidia-positive anti-dsDNA Abs, in contrast to Foxj1 coun- tion. Indeed, while Foxj1ϩ/ϩ FLCs failed to develop significant terparts, as judged 16 wk after reconstitution (10 of 10 vs 0 of 12 b Ͻ numbers of germinal centers, either spontaneously or in response sera, as judged by IgH -allotype-specific IgG assays; p 0.0001). Thus, Foxj1 is an intrinsic antagonist of B cell activation, and in its to Ti Ag immunization (Fig. 4, A and B, and data not shown; p Ͻ Ϫ Ϫ absence, B cells undergo uncontrolled activation. 0.001 comparing PBS- or NP-Ficoll-immunized Foxj1 / to their ϩ/ϩ counterparts), Foxj1Ϫ/Ϫ chimeras developed significant numbers of spontaneous germinal centers, which were dramatically Modulation of NF-␬B by Foxj1 Ͻ increased in number upon Ti Ag immunization ( p 0.0001, com- Foxj1 is known to repress NF-␬B activation in T cells (2), and Ϫ/Ϫ paring PBS- to NP-Ficoll-immunized Foxj1 mice to each other several members of the NF-␬B pathway play critical roles in ger- ϩ ϩ or to Foxj1 / mice). Altogether, these findings (Figs. 2–4) indi- minal center formation and humoral immunity (21). Consistent cate that Foxj1 antagonizes B cell responses, as characterized by with this, naive B cells from Foxj1Ϫ/Ϫ BOCs animals possess (auto)Ag-specific Abs, somatic hypermutation, and germinal enhanced spontaneous NF-␬B, but not NFAT, activity compared centers. with their Foxj1ϩ/ϩ counterparts, as demonstrated by luciferase The Journal of Immunology 957

FIGURE 6. Foxj1 regulates IL-6. A, Resting B cells from Foxj1 Ϫ/Ϫ vs ϩ/ϩ BOCs were stimulated in vitro with 25 ␮g/ml LPS for 4 h, and RNA was analyzed for IL-6 expression by real-time PCR. B, Production of IL-6 by B cells from Foxj1 Ϫ/Ϫ vs ϩ/ϩ B cell chimeras was assessed in culture supernatants by ELISA after 2 days in culture with LPS. Note that both parts of the figure are de- rived from a single experiment, with differences in y-axes used to emphasize the spontaneous (albeit low-level) production of IL-6 by Foxj1-deficient B cells in response to no or low-dose LPS stimulation. C, The activity of an IL-6 promoter- reporter construct was assessed in M12 B Downloaded from cell lymphoma cells in the presence (pcDNA-Foxj1) or absence (pcDNA) of Foxj1. D, The ability of NF-␬B decoy oligonucleotides to inhibit IL-6 secretion was assessed in B cells from Foxj1 Ϫ/Ϫ vs ϩ/ϩ BOCs stimulated with 2.5 ␮g/ml LPS for 2 days. SDs reflect triplicate http://www.jimmunol.org/ samples, representative of at least three separate experiments.

reporter assays (Fig. 5A, p Ͻ 0.001). Because Foxj1 regulates Foxj1 inhibited the activity of an IL-6 promoter-luciferase re- by guest on September 27, 2021 NF-␬B activity in T cells via I␬B␤ (2), we speculated that a similar porter (Fig. 6C, p Ͻ 0.0001 comparing pcDNA to pcDNA- mechanism would account for NF-␬B hyperactivity in Foxj1Ϫ/Ϫ B Foxj1)—likely reflecting NF-␬B inhibition (2). In addition, NF-␬B cells. Indeed, Foxj1 Ϫ/Ϫ B cells were selectively deficient in I␬B␤, decoy oligonucleotides, but not control mismatch oligonucleo- as judged by both Western blot and real-time PCR, in contrast to tides, inhibited the IL-6 hypersecretion of Foxj1Ϫ/Ϫ B cells, re- other NF-␬B family members (Fig. 5, B and C, and data not ducing the amounts to the levels secreted by Foxj1ϩ/ϩ B cells (Fig. shown; p Ͻ 0.001 comparing I␬B␤ mRNA in Foxj1 Ϫ/Ϫ vs ϩ/ϩ 6D, p Ͻ 0.0001 comparing decoy- to mismatch oligonucleotide- samples). Because Foxj1 can transactivate the I␬B␤ promoter (2), treated Foxj1Ϫ/Ϫ samples). Finally, in vivo treatment of Foxj1Ϫ/Ϫ such findings together suggest that Foxj1 is required in B cells to BOCs with neutralizing Ab against IL-6 inhibited spontaneous and regulate I␬B␤ expression and antagonize NF-␬B activity. Indeed, immunization-induced germinal center formation (Fig. 4, D and E; in vivo treatment of Foxj1Ϫ/Ϫ BOCs with antisense oligonucleo- p Ͻ 0.0001 comparing isotype- to anti-IL-6-treated animals). tides against the RELA (p65) NF-␬B subunit inhibited spontane- Thus, hyperactivity of B cells in Foxj1 deficiency results from ous and immunization-induced germinal center formation (Fig. 4, overactivity of both NF-␬B and IL-6. D and E, p Ͻ 0.0001 comparing p65 antisense- to p65 missense- treated animals), suggesting that NF-␬B hyperactivity, in fact, ac- Discussion counted for the hyperactive B cell phenotype of Foxj1 deficiency. A previous study has indicated that Foxj1 is required to prevent autoimmunity, because Foxj1-deficient FLCs develop spontaneous Modulation of IL-6 by Foxj1 T cell-related end-organ inflammation, and defective expression of Several NF-␬B target genes could account for this phenotype, in- Foxj1 in T cells correlates with spontaneous lupus-like disease (2). cluding costimulatory molecules such as members of the CD28/B7 The present findings extend the relevance of Foxj1 to B cell (hu- system, or cytokines such as IL-6 (21). However, when assessed as moral) autoimmunity, because Foxj1-deficient BOCs exhibit ac- freshly isolated naive cells or 4 h after LPS stimulation in vitro, centuated spontaneous autoantibody formation, anti-hapten re- Foxj1Ϫ/Ϫ B cells demonstrated 1.5-fold or less differences in RNA sponses in immunization and germinal center formation, and and/or protein levels of CD80, CD86, ICOS, ICOSL, CD40, defective expression of Foxj1 in B cells correlates with spontane- CD154, TNF-␣, LT-␣, bcl-6, Blimp-1, IRF-4, Mitf, c-myc, Pax5, ous lupus-like disease (Figs. 1–4). Interestingly, the role of Foxj1 and Xbp1, at least as judged by real-time PCR and/or flow cytom- in B cells appears analogous to its role in T cells, antagonizing etry (our unpublished data). In contrast, they consistently ex- NF-␬B and IL-6 (Figs. 5 and 6). Still, the present results suggest pressed at least 3-fold more IL-6 RNA after4hofLPSstimulation that Foxj1 deficiency does not break tolerance in B cells per se, in (Fig. 6A, p Ͻ 0.0001), and secreted 5- to 10-fold or more IL-6 in contrast to T cells (2), but rather amplifies ongoing Ag- or autoan- vitro (Fig. 6B, p Ͻ 0.0001 comparing all LPS doses), compared tigen-specific responses (Fig. 2), likely via amplified NF-␬B with Foxj1ϩ/ϩ B cells. and/or IL-6 activity. 958 ROLE OF Foxj1 IN B CELLS

It is interesting to note that, at least in some circumstances, 5. Berland, R., and H. H. Wortis. 2002. Origins and functions of B-1 cells with notes pathogenic autoantibody production may (4, 22) or may not (23– on the role of CD5. Annu. Rev. Immunol. 20: 253–300. 6. Calame, K. L., K. I. Lin, and C. Tunyaplin. 2003. Regulatory mechanisms that 25) require germinal centers. Because different experimental sys- determine the development and function of plasma cells. Annu. Rev. Immunol. 21: tems were used in each of these studies, these apparently disparate 205–230. 7. Kopf, M., S. Herren, M. V. Wiles, M. B. Pepys, and M. H. Kosco-Vilbois. 1998. conclusions may reflect a differential relevance of germinal centers Interleukin 6 influences germinal center development and antibody production to different types of autoimmune disease: some lupus-like syn- via a contribution of C3 complement component. J. Exp. Med. 188: 1895–1906. dromes, in fact, may result in pathogenic autoantibody production 8. Deng, C., E. Goluszko, E. Tuzun, H. Yang, and P. Christadoss. 2002. Resistance to experimental autoimmune myasthenia gravis in IL-6-deficient mice is associ- without the development of germinal centers. Therefore, Foxj1 ated with reduced germinal center formation and C3 production. J. Immunol. 169: deficiency might be found in only a subset of affected individuals 1077–1083. or mice, and may only be relevant to specific disease subtypes. 9. Brody, S. L., X. H. Yan, M. K. Wuerffel, S. K. Song, and S. D. Shapiro. 2000. Ciliogenesis and left-right axis defects in forkhead factor HFH-4-null mice. Nonetheless, the present findings correlate well with the previously Am. J. Respir. Cell Mol. Biol. 23: 45–51. demonstrated importance of the NF-␬B pathway in the pathogen- 10. Starnes, H. F., Jr., M. K. Pearce, A. Tewari, J. H. Yim, J. C. Zou, and J. S. Abrams. 1990. Anti-IL-6 monoclonal antibodies protect against lethal Esch- eses of several autoimmune syndromes (26); the role of IL-6 in erichia coli infection and lethal tumor necrosis factor-␣ challenge in mice. J. Im- autoreactive germinal centers (7, 8); the spontaneous germinal cen- munol. 145: 4185–4191. ter formation that occurs in humoral autoimmune diseases, likely 11. Schlaak, J. F., A. P. Barreiros, S. Pettersson, P. Schirmacher, K. H. Meyer Zum Buschenfelde, and M. F. Neurath. 2001. Antisense phospho- in response to endogenous autoantigens (22, 27); as well as the rothioate oligonucleotides to the p65 subunit of NF-␬B abrogate fulminant septic proposed utility of IL-6 blockade in the treatment of humoral au- shock induced by S. typhimurium in mice. Scand. J. Immunol. 54: 396–403. toimmune diseases like lupus (28). As such, our present findings 12. Peng, S. L., S. J. Szabo, and L. H. Glimcher. 2002. T-bet regulates IgG class switching and autoantibody production. Proc. Natl. Acad. Sci. USA 99: raise the intriguing possibility that functional Foxj1 deficiency, by 5545–5550. Downloaded from whatever genetic mechanism, could perhaps underlie a large pro- 13. Lin, L., A. J. Gerth, and S. L. Peng. 2004. Active inhibition of plasma cell development in resting B cells by microphthalmia-associated transcription factor. portion of autoimmune syndromes. Further studies that address J. Exp. Med. 200: 115–122. how Foxj1, NF-␬B, and IL-6 relate to B and T cell tolerance, 14. Khaled, A. R., E. J. Butfiloski, E. S. Sobel, and J. Schiffenbauer. 1998. Use of therefore, are likely to be particularly enlightening. phosphorothioate-modified oligodeoxynucleotides to inhibit NF-␬B expression and lymphocyte function. Clin. Immunol. Immunopathol. 86: 170–179. Curiously, in addition to its demonstrated role in autoreactive 15. Overbergh, L., D. Valckx, M. Waer, and C. Mathieu. 1999. Quantification of germinal center formation (7, 8), IL-6 has been heavily implicated murine cytokine mRNAs using real time quantitative reverse transcriptase PCR. http://www.jimmunol.org/ in plasma cell survival (29). The ability of Foxj1-deficient B cells Cytokine 11: 305–312. 16. Laurencikiene, J., V. Deveikaite, and E. Severinson. 2001. HS1,2 enhancer reg- to generate significantly enhanced anti-hapten titers (Fig. 2B) are ulation of germline ⑀ and ␥2b promoters in murine B lymphocytes: evidence for potentially consistent with the enhanced development and/or ac- specific promoter-enhancer interactions. J. Immunol. 167: 3257–3265. 17. Peng, S. L., M. E. Robert, A. C. Hayday, and J. Craft. 1996. A tumor-suppressor cumulation of Ag-specific plasma cells. However, the increased function for Fas (CD95) revealed in T cell-deficient mice. J. Exp. Med. 184: affinity of these specificities in the absence of Foxj1, associated 1149–1154. with the development of significant numbers of germinal centers, 18. Muramatsu, M., K. Kinoshita, S. Fagarasan, S. Yamada, Y. Shinkai, and T. Honjo. 2000. Class switch recombination and hypermutation require activa- suggests that the role of IL-6 here, in fact, may be 2-fold, both tion-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell promoting germinal center formation as well as plasma cell sur- 102: 553–563. vival, as suggested by studies with IL-6 transgenic mice (30). 19. Aarden, L. A., E. R. de Groot, and T. E. Feltkamp. 1975. Immunology of by guest on September 27, 2021 DNA. III. Crithidia luciliae, a simple substrate for the determination of anti- Relatively little continues to be known regarding the role of the dsDNA with the immunofluorescence technique. Ann. NY Acad. Sci. 254: forkhead genes in B cells. Some studies have indicated that mem- 505–515. 20. Hahn, B. H. 1998. Antibodies to DNA. N. Engl. J. Med. 338: 1359–1368. bers of the Foxo forkhead subfamily may regulate apoptotic and/or 21. Li, Q., and I. M. Verma. 2002. NF-␬B regulation in the immune system. Nat. Rev. proliferative responses in B cells (31, 32), but such studies have Immunol. 2: 725–734. been primarily limited to transformed cultured cell lines, and/or 22. Grammer, A. C., R. Slota, R. Fischer, H. Gur, H. Girschick, C. Yarboro, G. G. Illei, and P. E. Lipsky. 2003. Abnormal germinal center reactions in sys- overexpression studies in primary cells in vitro. Therefore, the temic lupus erythematosus demonstrated by blockade of CD154-CD40 interac- present findings supplement this growing literature by demonstrat- tions. J. Clin. Invest. 112: 1506–1520. ing that at least the Foxj subfamily member Foxj1 indeed modu- 23. Toellner, K. M., W. E. Jenkinson, D. R. Taylor, M. Khan, D. M. Sze, D. M. Sansom, C. G. Vinuesa, and I. C. MacLennan. 2002. Low-level hypermu- lates B cell effector function in vivo. Continued investigation into tation in T cell-independent germinal centers compared with high mutation rates the forkhead family, including members of other Fox subfamilies, associated with T cell-dependent germinal centers. J. Exp. Med. 195: 383–389. 24. Weller, S., A. Faili, C. Garcia, M. C. Braun, F. F. Le Deist, G. G. de Saint Basile, therefore, will hopefully reveal additional insights into the mech- O. Hermine, A. Fischer, C. A. Reynaud, and J. C. Weill. 2001. CD40-CD40L anisms of B as well as T cell immunoregulation, as well as the independent Ig gene hypermutation suggests a second B cell diversification path- relationship between their target genes and autoimmunity. way in humans. Proc. Natl. Acad. Sci. USA 98: 1166–1170. 25. William, J., C. Euler, S. Christensen, and M. J. Shlomchik. 2002. Evolution of autoantibody responses via somatic hypermutation outside of germinal centers. Acknowledgments Science 297: 2066–2070. ␬ 26. Peng, S. L. 2004. Transcription factors in the pathogenesis of autoimmunity. Clin. We thank Markus Neurath for advice regarding NF- B antisense oligonu- Immunol. 110: 112–123. cleotide strategies. 27. Luzina, I. G., S. P. Atamas, C. E. Storrer, L. C. daSilva, G. Kelsoe, J. C. Papadimitriou, and B. S. Handwerger. 2001. Spontaneous formation of Disclosures germinal centers in autoimmune mice. J. Leukocyte Biol. 70: 578–584. The authors have no financial conflict of interest. 28. Cross, J. T., and H. P. Benton. 1999. The roles of interleukin-6 and interleukin-10 in B cell hyperactivity in systemic lupus erythematosus. Inflamm. Res. 48: References 255–261. 1. Shlomchik, M. J., J. E. Craft, and M. J. Mamula. 2001. From T to B and back 29. Iwakoshi, N. N., A. H. Lee, and L. H. Glimcher. 2003. The X-box binding again: positive feedback in systemic autoimmune disease. Nat. Rev. Immunol. 1: protein-1 transcription factor is required for plasma cell differentiation and the 147–153. unfolded protein response. Immunol. Rev. 194: 29–38. 2. Lin, L., M. Spoor, A. J. Gerth, S. L. Brody, and S. L. Peng. 2004. Modulation of 30. Kovalchuk, A. L., J. S. Kim, S. S. Park, A. E. Coleman, J. M. Ward, H. C. Morse III, Th1 activation and inflammation by the NF-␬B repressor Foxj1. Science 303: T. Kishimoto, M. Potter, and S. Janz. 2002. IL-6 transgenic mouse model for ex- 1017–1020. traosseous plasmacytoma. Proc. Natl. Acad. Sci. USA 99: 1509–1514. 3. Lin, L., J. D. Hron, and S. L. Peng. 2004. Regulation of NF-␬B, Th activation, 31. Dijkers, P. F., R. H. Medema, J. W. Lammers, L. Koenderman, and P. J. Coffer. and autoinflammation by the forkhead transcription factor Foxo3a. Immunity 21: 2000. Expression of the pro-apoptotic Bcl-2 family member Bim is regulated by 203–213. the forkhead transcription factor FKHR-L1. Curr. Biol. 10: 1201–1204. 4. Hron, J. D., L. Caplan, A. J. Gerth, P. L. Schwartzberg, and S. L. Peng. 2004. 32. Yusuf, I., X. Zhu, M. G. Kharas, J. Chen, and D. A. Fruman. 2004. Optimal SH2D1A regulates T-dependent humoral autoimmunity. J. Exp. Med. 200: B-cell proliferation requires phosphoinositide 3-kinase-dependent inactivation of 261–266. FOXO transcription factors. Blood 104: 784–787.