Foxd3 suppresses interleukin‐10 expression in B cells

Summary

Interleukin‐10‐positive (IL‐10⁺) regulatory B (Breg) cells play an important role in restraining excessive inflammatory responses by secreting IL‐10. However, it is still unclear what key transcription factors determine Breg cell differentiation. Hence, we explore what transcription factor plays a key role in the expression of IL‐10, a pivotal cytokine in Breg cells. We used two types of web‐based prediction software to predict transcription factors binding the IL‐10 promoter and found that IL‐10 promoter had many binding sites for Foxd3. Chromatin immunoprecipitation PCR assay demonstrated that Foxd3 directly binds the predicted binding sites around the start codon upstream by −1400 bp. Further, we found that Foxd3 suppressed the activation of IL‐10 promoter by using an IL‐10 promoter report system. Finally, knocking out Foxd3 effectively promotes Breg cell production by up‐regulating IL‐10 expression. Conversely, up‐regulated Foxd3 expression was negatively associated with IL‐10⁺ Breg cells in lupus‐prone MRL/lpr mice. Hence, our data suggest that Foxd3 suppresses the production of IL‐10⁺ Breg cells by directly binding the IL‐10 promoter. This study demonstrates the mechanism for Breg cell production and its application to the treatment of autoimmune diseases by regulating Foxd3 expression.

Abbreviations

Breg cells

regulatory B cells

ChIP

chromatin immunoprecipitation

IL‐10

interleukin‐10

LPS

lipopolysaccharide

qPCR

quantitative PCR

shRNA

short hairpin RNA

TBS‐T

Tris‐buffered saline containing 0·1% Tween 20

Introduction

As a humoral immunity component, B cells play an important role in the adaptive immune system by secreting antibodies, presenting antigens and secreting cytokines.1 Recently, it has been shown that regulatory B (Breg) cells contribute to maintaining immune balance by restraining excessive inflammatory responses.2 Impairment of Breg cell development and function may cause inflammatory and autoimmune diseases.3 Defective development or function of Breg cells have been associated with chronic inflammation such as in systemic lupus erythematosus.4 On the other hand, Breg cell transfer can effectively reverse the autoimmune diseases in murine models, such as mice with experimental autoimmune uveitis mice5 and lupus‐prone mice.6 These studies suggest that Breg cells may be used to treat autoimmune diseases.

Regulatory B cells have been described in human and mouse models in a wide variety of inflammatory and autoimmune diseases.3, 7 Pivotal to Breg cell function is interleukin‐10 (IL‐10) which inhibits pro‐inflammatory cytokines.2 Immature B cells such as CD21^hi CD23^hi IgM^hi transitional 2‐marginal zone precursor8 and CD1d^hi CD5⁺ B109, 10 mature B cells such as splenic marginal zone,11, 12 and CD138⁺ CD44^hi plasmablasts13 and CD138⁺ MHC‐II^lo B220⁺ plasma cells14 all have the capacity to differentiate into IL‐10‐producing Breg cells. Remarkably, plasma blasts/plasma cells are the main types of activated B cells producing the cytokine IL‐10.13, 14 It is evident that transcriptional repressor Blimp‐1 can drive the plasma cell differentiation by suppressing transcription factors such as B‐cell‐associated Pax5 and germinal centre B cell‐associated Bcl6.15 The lack of blimp‐1 but not Bcl6 has a critical effect on regulatory plasmablast generation.13, 16 In addition, Blimp‐1 has been shown to mediate IL‐10 expression in CD4⁺ T cells.17, 18, 19 These studies suggest that plasma cell differentiation‐associated transcription factor may mediate Breg differentiation by regulating IL‐10 expression.

Until now, a critical transcriptional regulator that defines Breg cell function had not been discovered. Experimental evidence supports the concept that Breg cells are not lineage‐specific, but are expanded in response to stimuli.16 In vitro stimulation via lipopolysaccharide (LPS), together with PMA and ionomycin, induces the differentiation and an enrichment of IL‐10‐producing B cells.2, 6, 10

As one functional B‐cell subset, Breg cells suppress inflammatory response by secreting IL‐10. Hence, we explore which transcription factor plays a critical role in IL‐10 expression in B cells. We found here that Foxd3 suppressed the activation of IL‐10 promoter by predicting transcription factors binding IL‐10 promoter and using an IL‐10 promoter report system. Knock down of Foxd3 could effectively promote Breg cell production by up‐regulating IL‐10 expression. Our data suggest that Foxd3 suppresses the production of IL‐10⁺ Breg cells by limiting IL‐10 expression.

Methods and materials

Mice

Seven‐to‐nine‐week‐old C57BL/6 (Huafukang Corp., Beijing, China), female MRL/MpJ/lpr/lpr (MRL/lpr) mice (Nanjing Biomedical Research Institute of Nanjing University, Nanjing, China), and age‐matched MRL/MpJ/+/+ (MRL/++) mice (The Chinese Academy of Medical Sciences, Beijing, China) as previously reported7, 20, 21 were bred in our animal facilities under specific pathogen‐free conditions. Care, use and treatment of mice in this study were in strict agreement with international guidelines for the care and use of laboratory animals. This study was approved by the Animal Ethics Committee of the Beijing Institute of Basic Medical Sciences.

Prediction of transcription factor binding sites of IL‐10 promoter

We chose the IL‐10 gene sequence from the start codon upstream −2000 to downstream +100 as candidate promoter. We used the promoter 2 prediction server (http://www.cbs.dtu.dk/services/Promoter/) to identify potential promoter sequences. As expected, the sequence had obvious characteristics of a promoter. Subsequently, we used the internet http://jaspar.genereg.net/ to predict transcription factors (PAX5, Bcl‐6, Blimp‐1 and Foxd3) binding sites of IL‐10 promoter. To further analyse Foxd3 binding sites, we used another website (http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3).

Chromatin immunoprecipitation

Chromatin was immunoprecipitated according to the manufacturer's instruction (#9002, Cell Signaling, Danvers, MA). Briefly, sorted cells were cross‐linked with 1% (vol/vol) formaldehyde at room temperature for 10 min, and incubated with glycine for 5 min at room temperature. Cells were then sequentially washed in ice‐cold buffer A and buffer B, followed by digesting with MNase. Nuclear pellet was suspended in chromatin immunoprecipitation (ChIP) buffer, sheared by sonication with an average size of sheared fragments of about 300 bp to 800 bp. After centrifugation at 9600 g for 10 min, sheared chromatin was diluted in ChIP buffer and pre‐cleared by addition of protein A/G plus agarose beads (sc‐2003) for 1 hr at 4°. Before antibody incubation, input samples were removed from the lysate and stored at −80° until extraction. The beads were discarded and the supernatant was then incubated with anti‐mouse Foxd3 antibody (sc‐133588, Santa Cruz Biotech, Santa Cruz, CA) or control anti‐IgG (Cell Signaling Tech), at 4° overnight. The next day, protein A/G plus agarose beads were added and incubated for 2 hr at 4°. Beads were harvested by centrifugation and went through three low‐salt washes and one high‐salt wash. Beads were then eluted with ChIP elution buffer. The elutes and input were then added with proteinase K and RNase A and heated at 65° for 2 hr to reverse the formaldehyde cross‐link. DNA fragments were purified with Chip DNA clean & concentrator™‐capped column (D5205, ZYMO Research Corp, Irvine, CA). The immunoprecipitated and input DNA, and A SYBR Green PCR kit (Bio‐Rad, Hercules, CA) were used for quantitative real‐time PCR analysis. PCR was conducted on an initial denaturing step of 3 min at 94° followed by 45 cycles of 94° for 10 seconds, 60° for 15 seconds and 72° for 10 seconds and then a final extension at 72° for 7 min. The results were quantified with an Icycler IQ (Bio‐Rad). The relative binding was defined by determining the immunoprecipitation level (ratio of the amount of immunoprecipitated DNA to that of the input sample).

IL‐10 promoter reporting gene analysis

The firefly luciferase reporter plasmid pGL2B with the 5′‐flanking region from start codon upstream −1538 to downstream +64 of mouse IL‐10 gene (Plasmid # 24942) were purchased from Addgene (Cambridge, MA). Foxd3 binding P3 and p4 sites (Fig. 1) were located in −620~−723 bp and −1413 bp, respectively, in the upstream of the IL‐10 start codon inside the luciferase plasmid. Truncated IL‐10 promoter (−1393~+64) was amplified by PCR and constructed into pGL2B vector. Empty vector Lv201 and Foxd3 expression vector Lv201/Foxd3 (EX‐Mm24691‐Lv201) were purchased from GeneCopoeia (Guangzhou Province, China). Truncated Foxd3 (1–133 amino acids) and Foxd3 (347–385 amino acids) were amplified by PCR and constructed into Lv201 vector. To ensure that the recombinant clone is correct, we isolated the vector and verified that no mutations were introduced during cloning by DNA sequencing and that the construct was in the correct orientation. Then, 0·5 μg Lv201/Foxd3 or Lv201/truncated Foxd3, 0·5 μg firefly luciferase reporter plasmid pGL2/IL‐10 promoter or pGL2/truncated IL‐10 promoter, and 0·05 μg Renilla luciferase reporter vector pRL‐SV‐40 vector (cat# E2231, Promega Corp., Madison, WI) were co‐transduced into 4 × 10⁵ RAW264.7 cells in 12‐well plate by using 6 μl Lipofectamine®2000 Reagent (Cat# 11668‐019, Invitrogen Corp., Carlsbad, CA). On day 3, sequential measurement of firefly luciferase (Reporter #1) followed by Renilla luciferase activity (Reporter #2) was assessed on 1420 Multilabel Counter (1420 Victor 3, PerkinElmer Corp., Waltham, MA), and analysed. The results were shown as the ratio of firefly to Renilla luciferase activity.

Figure 1.

Schematic diagram of mouse interleukin‐10 (IL‐10) promoter region illustrating the positions of the primer pairs used for chromatin immunoprecipitation assays. Sequences represent main predicted Foxd3 binding sites from Tables 2 and 3. Arrows represent the region of six primer pairs. TSS, Transcription start site.

In vitro B‐cell cultures

Lymphocytes were separated from spleen or lymph nodes by Ficoll separation solution. B220⁺ B cells were isolated using mouse B220 MicroBeads (AutoMACS; Miltenyi Biotec, Bergisch Gladbach, Germany) as described22 and cultured in RPMI‐1640 medium containing 10% fetal bovine serum, 2 mm glutamine, penicillin (100 IU/ml), streptomycin (100 mg/ml), and 50 mm 2‐mercaptoethanol. Cells were stimulated for 0, 1, 2 and 3 days with 1 μg/ml LPS (Sigma L2630 from Escherichia coli 0111:B4; Sigma, St Louis, MO).

Cytometric analysis and intracellular cytokine staining

To visualize IL‐10‐competent B cells, 1 μg/ml LPS (Sigma), 50 ng/ml PMA, 1 μg/ml ionomycin (Sigma‐Aldrich), 10 μg/ml brefeldin A and 2 μm monensin were added into cell culture before the cells were stained for cell surface B220. After 5 hr, cells were collected. Cells (1 × 10⁶ cells/sample) were washed with fluorescence‐activated cell‐sorting staining buffer (PBS, 2% fetal bovine serum or 1% bovine serum albumin, 0·1% sodium azide). All samples were incubated with anti‐Fc receptor antibody (BD Biosciences, Franklin Lakes, NJ), before incubation with other antibodies diluted in fluorescence‐activated cell sorting buffer supplemented with 2% anti‐Fc receptor antibody. After staining for cell surface B220, the cells were fixed for 50 min with 1 ml fixation buffer (eBioscience, San Diego, CA). After washing, the fixed cells were stained for IL‐10. The samples were filtered immediately before analysis or cell sorting to remove any clumps. The following antibodies were purchased from eBioscience: Peridinin chlorophyll protein‐ or allophycocyanin‐conjugated anti‐mouse B220 antibodies and phycoerythrin‐conjugated anti‐mouse IL‐10 antibodies. Data collection and analyses were performed on a FACS Calibur flow cytometer using cellquest software.

Quantitative PCR analysis

All RNA samples were DNA free. The cDNA synthesis and quantitative PCR (qPCR) analyses were performed as described previously.20, 21 Total RNA was extracted from B cells with Trizol (Invitrogen Life Technologies). The final RNA pellets were dissolved in 0·1 mm EDTA (2 μl/mg original wet weight). Reverse transcription reactions were carried out on 22 μl of sample using superscript II RNAse H‐Reverse Transcriptase (Invitrogen Life Technologies) in a reaction volume of 40 μl. All samples were diluted in 160 μl nuclease‐free water. Quantitative PCR was employed to quantify mouse Foxd3 gene expression from the cDNA samples. Mouse Foxd3 primers were designed using primer express version 1.0 software (Applied Biosystems, Foster City, CA) from the mouse Foxd3 gene sequences (GenBank/EBML databases). The primer pair used for qPCR analysis spanned at least an intron. A 171‐base‐length Foxd3 fragment was amplified using the following primers: forward 5′‐TCTTACATCGCGCTCATCAC‐3′ and reverse 5′‐TCTTGACGAAGCAGTCGTTG‐3′. PCR was conducted on an initial denaturing step of 3 min at 94° followed by 35 cycles of 94° for 10 seconds, 60° for 15 seconds and 72° for 20 seconds and then a final extension at 72° for 7 min. Mouse Foxd3 mRNA expression was normalized to the levels of the GAPDH gene.

Western blot analyses

Whole‐cell lysates were prepared for Western blotting. Twenty‐five micrograms of cell protein was electrophoretically separated on a 10% SDS–polyacrylamide gel and transferred to a PVDF membrane,which was then blocked by incubation for 1 hr at room temperature in 5% fat‐free dry milk in Tris‐buffered saline containing 0·1% Tween 20 (TBS‐T). The blots were then incubated overnight at 4° with rabbit antibodies against mouse GAPDH (sc‐20357, Santa Cruz Biotech) and Foxd3 (ab107248, abcam corp., Cambridge, MA) diluted 1 : 1000 in TBS‐T containing 5% bovine serum albumin, washed for 25 min with TBS‐T, and incubated for 1 hr at room temperature with horseradish peroxidase‐conjugated secondary F(ab')₂ (Zymed Laboratories, San Francisco, CA) (1 : 20 000 in TBS‐T containing 5% bovine serum albumin), then bound antibody was visualized using the ECL detection system (Amersham, Arlington Heights, IL).

Knock down of Foxd3 in B cells

B220⁺ B cells (1 × 10⁶ cells/ml) were infected with 1 ml 1 × 10⁷ IU (infected units) control and Foxd3‐specific short hairpin RNA (shRNA)‐contained lentivirus (sc‐145222‐V; Santa Cruz Biotech) in the presence of 1 μg/ml LPS in six‐well plates (total volume: 5 ml) in RPMI‐1640 medium containing 10% fetal bovine serum, 2 mm glutamine, penicillin (100 IU/ml), streptomycin (100 mg/ml) and 50 mm 2‐mercaptoethanol. On day 3, cells were collected and lysates were analysed by qPCR and Western blot.

IL‐10 concentration analysis by ELISA

The concentration of IL‐10 was measured by ELISA (Cat# BMS614/2*, eBioscience). Briefly, anti‐mouse IL‐10 capture antibody was coated in triplicate to the plate for overnight at 4° and diluted supernatants were added to the plate for 1 hr at 37°; Then after washing, 4 μg/ml biotin rat anti‐mouse IL‐10 detection antibody were added to the plate, and were incubated for another hour at 37°. Thereafter, unbinding antibodies were washed off, followed by addition of avidin‐horseradish peroxidase (1/1000 diluted). Plates were incubated for 1 hr at 37°. Finally, the colour was developed by incubation with o‐phenylenediamine. The optical density was read at 492 nm with an ELISA reader (Bio‐Rad). Standard curves were established to quantify the amounts of the respective cytokines.

Statistics

Data were analysed using graphpad prism (version 5.0, GraphPad Software Inc., San Diego, CA), and because D'Agostino–Person Omnibus and Shapiro–Wilk normality tests were failed, Mann–Whitney U test was used. Results were considered statistically significant at P < 0·05.

Results

Foxd3 was predicted to bind the IL‐10 promoter

To explore what transcription factor plays a critical role in IL‐10 expression in plasma cells, we chose the IL‐10 promoter sequence from start codon upstream −2000 bp to downstream +100 bp (see Supplementary material, Table S1) containing previously identified IL‐10 promoter activation sequences such as histone phosphorylation binding sites (−1535~+3 bp),23MARE motif (−500‐bp).24 To predict unknown transcription factor binding sites in the IL‐10 promoter, we chose a web‐based prediction software. We first used PAX5, Bcl6 and Blimp‐1 transcription factor to verify the software. We found that compared with PAX5 and Bcl6, Blimp‐1 (Prdm1) had more binding sites and higher predicted scores such as 19·8, 16·6, 16·5 and 11·1 (Table 1). The results are in line with those of previous studies.13, 15, 16, 17, 18, 19 Hence, the software could be used to identify the unknown transcription factors binding the IL‐10 promoter.

Table 1.

Predicted PAX5, Bcl6 and Prdm1 binding sites in interleukin‐10 promoter

Model name	Score	Relative score	Start	End	Strand	Predicted site sequence
PAX5	8·99	0·83	2070	2088	−1	GTGCTGAGCCAGGCATGAT
PAX5	8·27	0·82	2041	2059	1	CAAGGCAGCCTTGCAGAAA
PAX5	6·58	0·80	2063	2081	1	GAGCTCCATCATGCCTGGC
PAX5	6·26	0·80	394	412	1	GAAGGCAGCAAAGGGGATT
Bcl6	7·4	0·80	1445	1458	1	TATCCTCAAAGGAT
PRDM1	19·8	0·96	626	640	1	AGAAAGTGAAAGGGA
PRDM1	16·6	0·92	1856	1870	1	AAAAAGGGAAAGGAA
PRDM1	16·5	0·92	620	634	1	GGAAGGAGAAAGTGA
PRDM1	11·1	0·85	1873	1887	1	AAAAAAAGAAAGAAA
PRDM1	8·6	0·82	1085	1099	1	ACAGAATGAAAGGCA
PRDM1	8·3	0·82	1769	1783	1	AAAAATTAAAAGAGA
PRDM1	7·9	0·81	1877	1891	1	AAAGAAAGAAATTAA
PRDM1	7·5	0·81	614	628	1	CAAAAGGGAAGGAGA

We used a web‐based tool (http://jaspar.genereg.net/) to analyse the binding sites of PAX5, Bcl‐6 and Prdm1 in the interleukin‐10 promoter sequence from start codon upstream −2000 to downstream +100.

By using this software, we found that transcription factor Foxd3 had many more binding sites and higher predicted scores on binding the IL‐10 promoter (Table 2). In addition, we used another website to analyse Foxd3 binding sites in the IL‐10 promoter. Similarly, there are 12 binding sites of Foxd3 in the IL‐10 promoter (Table 3). Critically, two kinds of software predict the same binding sites about the start codon upstream −1400 bp (that is, 600 bp from −2000 bp of the IL‐10 promoter) (Tables 2 and 3).

Table 2.

Predicted Foxd3 binding sites in interleukin‐10 promoter

Model name	Score	Relative score	Start	End	Strand	Predicted site sequence
Foxd3	14·53	0·96	983	994	−1	GAATGTTTGCTT
Foxd3	13·21	0·93	136	147	−1	AATTATTTTTTT
Foxd3	13·13	0·93	593	604	−1	TTTTGTTTTTTT
Foxd3	12·74	0·92	380	391	−1	GCATGTTTTTTT
Foxd3	11·08	0·89	376	387	−1	GTTTTTTTTTTT
Foxd3	10·45	0·88	1765	1776	−1	TAATTTTTTTTT
Foxd3	9·60	0·86	372	383	−1	TTTTTTTTATTT
Foxd3	9·55	0·86	596	607	−1	GTTTTTTGTTTT
Foxd3	9·52	0·86	134	145	−1	TTATTTTTTTTT
Foxd3	9·46	0·86	1764	1775	−1	AATTTTTTTTTC
Foxd3	9·23	0·85	1604	1615	−1	AAATAATTGTTT
Foxd3	8·52	0·84	1869	1880	−1	CTTTTTTTTTTT
Foxd3	8·14	0·83	1600	1611	−1	AATTGTTTCCTC
Foxd3	7·99	0·83	1868	1879	−1	TTTTTTTTTTTC
Foxd3	7·86	0·83	1873	1884	−1	CTTTCTTTTTTT
Foxd3	7·73	0·82	1767	1778	−1	TTTAATTTTTTT
Foxd3	7·53	0·82	70	81	1	TAATGTTTTATG
Foxd3	7·41	0·82	64	75	1	ATATATTAATGT
Foxd3	7·37	0·82	1762	1773	−1	TTTTTTTTTCTT
Foxd3	7·24	0·81	1870	1881	−1	TCTTTTTTTTTT
Foxd3	7·20	0·81	373	384	−1	TTTTTTTTTATT
Foxd3	7·17	0·81	1937	1948	−1	AAAGGTTTTTGT
Foxd3	7·08	0·81	132	143	−1	ATTTTTTTTTGT
Foxd3	6·97	0·81	1733	1744	−1	AAAAGTTGTATT
Foxd3	6·92	0·81	1768	1779	−1	TTTTAATTTTTT
Foxd3	6·85	0·81	597	608	−1	TGTTTTTTGTTT
Foxd3	6·65	0·80	130	141	−1	TTTTTTTTGTCT
Foxd3	6·63	0·80	1116	1127	1	GAATGTTCTTCC
Foxd3	6·59	0·80	22	33	1	TTATATTGATAT
Foxd3	6·55	0·80	1882	1893	−1	GTTTAATTTCTT

We used a web‐based tool (http://jaspar.genereg.net/) to analyse Foxd3 binding sites in interleukin‐10 promoter sequence from start codon upstream −2000 to downstream +100.

Table 3.

Predicted Foxd3 binding sites in interleukin‐10 promoter

Factor name	Start position	End position	Dissimilarity	String
FOXD3 [T02290]	1385	1392	5·785782	AAACACGG
FOXD3 [T02290]	1282	1289	5·412889	AAACAGGG
FOXD3 [T02290]	1044	1051	5·320522	AAACAGGT
FOXD3 [T02290]	210	217	5·207923	TGGTGTTT
FOXD3 [T02290]	335	342	4·374352	AAACACAA
FOXD3 [T02290]	603	610	4·001457	AAACAGAA
FOXD3 [T02290]	383	390	1·875536	AAACATGC
FOXD3 [T02290]	539	546	1·851876	AAACATCA
FOXD3 [T02290]	596	603	1·018305	AAACAAAA
FOXD3 [T02290]	1603	1610	0·461834	AAACAATT
FOXD3 [T02290]	69	76	0·184733	TAATGTTT
FOXD3 [T02290]	986	993	0	AAACATTC

We used a web‐based tool (http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3) to analyse Foxd3 binding sites in interleukin‐10 promoter sequence from start codon upstream −2000 to downstream +100.

Foxd3 directly binds the IL‐10 promoter

To prove whether Foxd3 directly bound the IL‐10 promoter, we used ChIP‐PCR technology. We first chose the IL‐10 promoter sites which represent the main predicted Foxd3 binding sites from Tables 2 and 3 as PCR template and designed the primer pairs used for ChIP assays (Fig. 1). To perform ChIP, we first used B220 microbeads to sort B220⁺ B from 7‐ to 9‐week‐old C57BL/6 mice and LPS to stimulate B cells for 3 days. Anti‐mouse Foxd3 antibody or control IgG was used to probe the IL‐10 locus in LPS‐activated wild‐type B220⁺ B cells. The relative binding was defined by quantitative PCR (Fig. 2). The results suggest that Foxd3 can bind the sites amplified by p3 and p4 primer pairs (Fig. 2). Of note, ChIP‐PCR proves that Foxd3 directly binds the same binding sites about the start codon upstream −1400 bp (that is, 600 bp from −2000 bp of the IL‐10 promoter) predicted by two web sites. The same Foxd3 binding site p4 has higher scores 13·13 (Table 2) and 4·0 (Table 3) evaluated using two types of web‐based prediction software.

Figure 2.

Foxd3 directly binds with the interleukin‐10 (IL‐10) promoter. Chromatin immunoprecipitation assays of lipopolysaccharide (LPS)‐activated wild‐type (WT) B220⁺ B cells using a Foxd3 antibody (αFoxd3) or control IgG probing for the IL‐10 locus. Quantitative PCR (a) and PCR gel (b) were used to analyse the enrichment and the fold enrichments are represented from one of three independent experiments; p1, p2, p3, p4, p5 and p6 PCR products were amplified by primer pairs p1, p2, p3, p4, p5 and p6, respectively, in Fig. 1. White arrows show the potential binding sites of Foxd3 on the IL‐10 promoter. IgG versus Foxd3, *P < 0·05. (Mann–Whitney U‐test).

Foxd3 directly suppressed the activation of the IL‐10 promoter

For the role of Foxd3 in the activation of the IL‐10 promoter, we used a murine macrophage line RAW264.7 cells; these cells are suitable for inducing luciferase reporter vector pGL2/IL‐10 promoter (Fig. 3). Compared with the control, pGL3/IL‐10 promoter‐transfected RAW264.7 cells had a significantly higher ratio of firefly to renilla luciferase activity (Fig. 3). When full‐length Foxd3 was co‐transduced into RAW264.7 cells, the ratio of firefly to renilla luciferase activity was significantly reduced (Fig. 3a). These data suggest that Foxd3 directly suppresses the activation of IL‐10 promoter. To further prove whether the Foxd3 binding site (−1400 bp) of IL‐10 promoter played an important role in the suppressive function of Foxd3, we constructed the truncated IL‐10 promoter (−1393 ~ +64 bp). The results show that Foxd3 did not suppress IL‐10 promoter activation (Fig. 3b). The data suggest that Foxd3 suppresses IL‐10 promoter activation by binding −1400 bp site of the IL‐10 promoter. Previous study has demonstrated that the FoxD3 eh1/GEH motif is required for both repression of transcription and induction of the mesoderm.25 To identify whether the eh1/GEH motif was important in suppressive function of Foxd3 for IL‐10 promoter activation, we constructed two truncated Foxd3 (1–133 amino acids) without eh1/GEH motif and eh1/GEH motif‐contained Foxd3 (347–385 amino acids). Unexpectedly, we found that the N‐terminal domain of Foxd3 played an important role in suppressing IL‐10 expression (Fig. 3c).

Figure 3.

Foxd3 directly suppressed interleukin‐10 (IL‐10) promoter activation. Empty vector Lv 201 (Vector) or Lv201/Foxd3 (Foxd3) and luciferase reporter vector pGL2/IL‐10 promoter (−1538 ~ +64 bp) (a), Lv201/Foxd3 and pGL2/IL‐10 promoter (−1538 ~ +64 bp) or pGL2/IL‐10 promoter (−1393 ~ +64 bp) (b), or pGL2/IL‐10 promoter (−1538 ~ +64 bp) and full‐length Foxd3 (1–469 aa), truncated Foxd3 (1–133 aa) or Foxd3 (347–385 aa) (c) were co‐transduced into RAW264.7 cells. Dual luciferase reporter gene expression was analysed, and the results are shown as the ratio of firefly to Renilla luciferase activity. The data represent at least four independent experiments. Error bars, SEM. *P < 0·05, **P < 0·01, ***P < 0·001 (Mann–Whitney U‐test).

Foxd3 expression was up‐regulated in LPS‐stimulated B cells

To explore the role of Foxd3 in IL‐10 expression in B cells, we first developed the in vitro system of IL‐10⁺ Breg cell production. B220⁺ B from 7‐ to 9‐week‐old C57BL/6 mice were sorted using B220 microbeads and cultured for 1, 2 and 3 days in vitro with LPS. Flow cytometry (FACS) analysis showed that IL‐10 expression (IL‐10⁺ Breg cells) time‐dependently increased in LPS‐stimulated B cells (Fig. 4a).

Figure 4.

Both Foxd3 and interleukin‐10 (IL‐10) were up‐regulated in lipopolysaccharide (LPS) ‐activated B cells. B220⁺ B cells from 7‐ to 9‐week‐old C57BL/6 mice were sorted by B220 microbeads and cultured in vitro for 1, 2 and 3 days in the presence of 1 μg/ml LPS. (a) To visualize IL‐10‐competent B cells, LPS, PMA, ionomycin, brefeldin and monensin were added to the cultures 5 hr before the cells were stained for cell surface B220 and cytoplasmic IL‐10 expression and analysed by FACS. Representative histograms show IL‐10⁺ cell frequencies within the indicated gates. (b, c) On days 0, 1, 2 and 3 after LPS stimulation, the cells were collected and subjected to quantitative PCR (b) and Western blot (c). (a–c) Data represent four independent experiments. (b) Error bars, SEM. **P < 0·01 (Mann–Whitney U‐test).

Until now, no data demonstrated whether B cells expressed Foxd3. We first determined Foxd3 expression in B cells. We found that LPS effectively up‐regulated Foxd3 mRNA (Fig. 4b) and protein (Fig. 4c) expression. Critically, Foxd3 expression time‐dependently increased in LPS‐stimulated B cells (Fig. 4b,c). These results suggest that both Foxd3 and IL‐10 were up‐regulated in LPS‐activated B cells.

Knock down of Foxd3 expression up‐regulated IL‐10⁺ Breg cells by IL‐10 expression in LPS‐stimulated B cells

Lipopolysaccharide could effectively induce B cells to express Foxd3 and IL‐10. To explore the role of Foxd3 in IL‐10 expression in B cells, we used Foxd3‐specific shRNA to knock down Foxd3 expression in LPS‐activated B cells. Western blot and qPCR analysis demonstrated that Foxd3‐specific shRNA effectively reduced Foxd3 expression in LPS‐stimulated B cells (Fig. 5a,b). As expected, we found that IL‐10 levels were significantly up‐regulated in the supernatant from LPS‐stimulated Foxd3‐knocked down B cells (Fig. 5c). Accordingly, knock down of Foxd3 expression also up‐regulated IL‐10⁺ regulatory B cells in LPS‐stimulated B cells (Fig. 5d). Without stimulation conditions, knock down of Foxd3 expression also up‐regulated IL‐10 expression, although the IL‐10 level was far lower than that from LPS‐stimulated B cells (see Supplementary material, Fig. S1). Together, our data suggest that knock down of Foxd3 expression up‐regulated IL‐10⁺ regulatory B cells by IL‐10 expression in LPS‐stimulated B cells.

Figure 5.

Knock down of Foxd3 expression up‐regulated interleukin‐10 (IL‐10) expression and IL‐10⁺ regulatory B cells in lipopolysaccharide (LPS) ‐stimulated B cells. B220⁺ B cells from 7‐ to 9‐week‐old C57BL/6 mice were sorted by B220 microbeads, infected with control or Foxd3‐specific shRNA, and cultured in vitro for 3 days in the presence of 1 μg/ml LPS. On day 3, the cells and supernatant were collected. The cells were subjected to quantitative PCR (a) and Western blot (b). In the supernatant, IL‐10 level was determined by ELISA (c). (d) On day 3, to visualize IL‐10‐competent B cells, LPS, PMA, ionomycin, brefeldin and monensin were added to the cultures 5 hr before the cells were stained for cell surface B220 and cytoplasmic IL‐10 expression and analysed by FACS. Representative histograms show IL‐10⁺ cell frequencies within the indicated gates. (a–d) Data represent four independent experiments. Error bars, SEM. **P < 0·01, ***P < 0·001 (Mann–Whitney U‐test).

Foxd3 expression was negatively associated with IL‐10 expression in B cells from lupus‐prone MRL/lpr mice

MRL/lpr mice bearing Fas/Fas ligand mutant genes develop autoimmunity and lymphoproliferation disease similar to human systemic lupus erythematosus and are considered a model of human systemic lupus erythematosus diseases.26, 27 Previous study has shown defective development or function of Breg cells in autoimmune disease such as systemic lupus erythematosus.4 As expected, compared with non‐lupus‐prone MRL/+ mice, lupus‐prone MRL/lpr mice reduced IL‐10⁺ Breg cells in the spleen and lymph nodes (Fig. 6a). Critically, we found that Foxd3 expression increased in B cells from lupus‐prone MRL/lpr mice (Fig. 6b). These results suggest that Foxd3 expression was negatively associated with IL‐10 expression in B cells from lupus‐prone MRL/lpr mice (Fig. 6a,b). Importantly, we found that the level of Foxd3 in B cells from lymph nodes was higher than that from splenic B cells (Fig. 6b) and the percentages of IL‐10⁺ Breg cells from lymph nodes were lower than those from spleen (Fig. 6a). Together, these results suggest that Foxd3 expression may be up‐regulated to suppress IL‐10⁺ Breg cells in lupus‐prone MRL/lpr mice.

Figure 6.

Foxd3 expression was negatively associated with interleukin‐10 (IL‐10) expression in B cells from lupus‐prone MRL/lpr mice. B220⁺ B cells from 12‐ to 16‐week‐old female non‐lupus‐prone MRL/+ and lupus‐prone MRL/lpr mice were sorted by B220 microbeads. (a) To visualize IL‐10‐competent B cells, B cells were cultured in vitro for 5 hr in the presence of lipopolysaccharide (LPS), PMA, ionomycin, brefeldin and monensin. Then, the cells were stained for cell surface B220 and cytoplasmic IL‐10 expression and analysed by FACS. Representative histograms show IL‐10⁺ cell frequencies within the indicated gates. Bar graphs indicate the mean ± SEM of IL‐10⁺ B‐cell frequencies from three independent experiments. (b) The cells were subjected to Western blot. (a, b) Data represent three independent experiments. Error bars, SEM. *P < 0·05, **P < 0·01 (Mann–Whitney U‐test).

Discussion

Foxd3 belongs to the forkhead protein family of transcription factors characterized by a DNA‐binding forkhead domain. Foxd3 expression has been reported in embryonic stem cells and trophoblast stem cells of the early mouse embryo.28 Foxd3 has shown a global role in all aspects of neural crest maintenance along the anterior–posterior axis, and establish an unprecedented molecular link between multiple divergent progenitor lineages of the mammalian embryo.29 Trophoblast progenitors in Foxd3^−/− embryos do not self‐renew and are not multipotent, but instead give rise to an excess of trophoblast giant cells.30 Accordantly, Foxd3 prevents the production of melanocyte progenitors from the developing neural crest and inhibits G(1)‐S progression in melanoma cells.31 The expression of Foxd3 was also found in the developing pancreas.28 However, its function is unclear in the developing pancreas. For the first time, we identified that Foxd3 expressed in B cells and suppressed IL‐10⁺ Breg cell production by limiting IL‐10 promoter activation.

FoxD3 functions as a transcriptional repressor.25 Foxd3 controls melanophore specification in the zebrafish neural crest by repressing the mitfa promoter in a subset of neural crest cells.32, 33 Neural crest stem cell multipotency requires Foxd3 to maintain neural potential and repress mesenchymal fates.34 Foxd3 is important for repressing melanogenesis in avian embryos35 and a mutant B‐RAF‐regulated inhibitor of G(1)‐S progression in melanoma cells.31 In accordance with these publications, our data demonstrated that as a transcriptional repressor, Foxd3 could effectively suppress IL‐10 production in B cells.

Previous study has demonstrated that the FoxD3 eh1/GEH motif is required for both repression of transcription and induction of the mesoderm.25 Unexpectedly, we found that the eh1/GEH motif is not important in suppressive function of Foxd3 for IL‐10 promoter activation. However, we found that the N‐terminal domain of Foxd3 played an important role in suppressing IL‐10 expression (Fig. 3c). It is very interesting to further explore the role of the N‐terminal domain of Foxd3 in suppressing IL‐10 promoter activation.

In the context of autoimmune disorders, B cells can be pathogenic effectors through their production of autoantibodies. On the other hand, B cells appear to regulate autoimmune conditions by secreting IL‐10 independently of their ability to produce autoantibodies.36 Remarkably, plasma‐blasts and plasma cells are the main types of IL‐10‐producing Breg cells,13, 14 suggesting that B cells can differentiate into both antibody‐producing plasma cells and IL‐10‐producing Breg cells. Previous studies have suggested that plasma cell differentiation‐associated transcription factors such as Blimp‐1, but not B cells or germinal centre B‐cell‐associated transcription factors such as PAX5 and Bcl6 may mediate Breg cell differentiation by regulating IL‐10 expression.13, 15, 16, 17, 18, 19 In line with these studies, we found that compared with PAX5 and Bcl6, Blimp‐1 (Prdm1) had more binding sites on the IL‐10 promoter and higher predicted scores such as 19·8, 16·6, 16·5 and 11·1 (Table 1) by web‐based prediction software.

However, it is still unclear what decides the B‐cell differentiation fate. Previous studies have shown that Blimp‐1, a critical transcription factor for the generation of plasma cells,37, 38, 39 directly activated transcription factor Foxd3.40 Our data demonstrated here that Foxd3 could effectively suppress the production of IL‐10⁺ Breg cells by directly binding the IL‐10 promoter. Hence, Blimp‐1 directs the differentiation of B cells into plasma cells by activating Foxd3 to suppress IL‐10⁺ B‐cell production.

It is unclear whether there are other factors aside from Blimp‐1 involved in up‐regulated Foxd3 expression in LPS‐stimulated B cells or in vivo‐activated B cells from lupus‐prone mice. We found that naive B cells expressed very low levels of Foxd3 (Figs 4c and 6b). Unexpectedly, naive B cells, cultured in unstimulated conditions, also expressed very high levels of Foxd3 (see Supplementary material, Fig. S1). The data suggest that apoptotic B cells may express Foxd3. It is worthwhile to explore further the role of Foxd3 in apoptotic B cells.

Systemic lupus erythematosus is a female‐predominant heterogeneous systemic autoimmune disease characterized by autoantibodies‐producing plasma cells.41 In addition, we found that Foxd3 expression was up‐regulated in B cells from lupus‐prone MRL/lpr mice. Importantly, we demonstrated that up‐regulated Foxd3 expression was negatively associated with IL‐10⁺ Breg cells in lupus‐prone MRL/lpr mice. These results suggest that Foxd3 may participate in directing the differentiation of B cells into plasma cells by suppressing IL‐10⁺ B‐cell production.

Foxd3 has shown a global role in all aspects of neural crest maintenance.29 We found here that Foxd3 suppressed IL‐10⁺ Breg cell production by limiting the activation of IL‐10 promoter. In addition, Foxd3 may participate in directing the differentiation of B cells into plasma cells by suppressing IL‐10⁺ B‐cell production. Hence, Foxd3 may also have a broad role in the fate of B‐cell differentiation. Apart from IL‐10 promoter, Foxd3 represses the mitfa promoter in a subset of neural crest cells.32, 33 Hence, it is worthwhile to further explore whether Foxd3 regulates other transcription factors and/or cytokines in B‐cell or other cell differentiation.

Regulatory B cells have been shown to modulate the severity experimental autoimmune uveitis and other autoimmune diseases by producing IL‐10.42 For Breg cells to be useful therapeutically, greater clarity regarding the phenotype, induction and stability of these cells in vivo is needed.16 Critically, we showed here that the Foxd3 transcription factor suppressed the production of IL‐10⁺ Breg cells by directly binding IL‐10 promoter, whereas knock‐down of Foxd3 up‐regulated IL‐10⁺ Breg cells. The results suggest that targeting Foxd3 expression to induce or suppress Breg cells may provide a novel way to treat autoimmune diseases, infectious diseases and tumours.

In conclusion, we found that when activated in vitro and in vivo, B cells up‐regulated Foxd3 expression. Foxd3 suppressed IL‐10 promoter activation by directly binding the IL‐10 promoter. Critically, Foxd3 knock down effectively up‐regulated IL‐10⁺ Breg cells. The relationship between Foxd3 expression and IL‐10⁺ Breg cells was further proved in lupus‐prone MRL/lpr mice. Together, our data suggest that Foxd3 suppresses the production of IL‐10⁺ Breg cells by directly binding the IL‐10 promoter. Hence, the Foxd3 signalling pathway may be explored to regulate Breg cell production and its application for autoimmune diseases.

Authorship

YZ, ZW, HX, XL, GZ and CH performed the experiments, GH, GC, NM, BS and YL contributed essential reagents and materials for the experiments. RW conceived and designed the studies. All authors contributed to data analysis and manuscript preparation.

Disclosures

The authors declare no commercial or financial conflict of interest.

Supporting information

Table S1. Mouse interleukin‐10 promote r sequence is used to predict transcription factor binding site.

Figure S1. Knock down of Foxd3 expression up‐regulated interleukin‐10 expression.