The Wnt5a-Ror2 axis promotes the signaling circuit between interleukin-12 and interferon-γ in colitis

Wnt5a, which regulates various cellular functions in Wnt signaling, is involved in inflammatory responses, however the mechanism is not well understood. We examined the role of Wnt5a signaling in intestinal immunity using conditional knockout mice for Wnt5a and its receptor Ror2. Removing Wnt5a or Ror2 in adult mice suppressed dextran sodium sulfate (DSS)-induced colitis. It also attenuated the DSS-dependent increase in inflammatory cytokine production and decreased interferon-γ (IFN-γ)-producing CD4+ Th1 cell numbers in the colon. Wnt5a was highly expressed in stromal fibroblasts in ulcerative lesions in the DSS-treated mice and inflammatory bowel disease patients. Dendritic cells (DCs) isolated from the colon of Wnt5a and Ror2 deficient mice reduced the ability to differentiate naïve CD4+ T cells to IFN-γ-producing CD4+ Th1 cells. In vitro experiments demonstrated that the Wnt5a-Ror2 signaling axis augmented the DCs priming effect of IFN-γ, leading to enhanced lipopolysaccharide (LPS)-induced interleukin (IL)-12 expression. Taken together, these results suggest that Wnt5a promotes IFN-γ signaling, leading to IL-12 expression in DCs, and thereby inducing Th1 differentiation in colitis.

there are no gross abnormalities in the post-natal development of Wnt5a heterozygous (Wnt5a +/− ) mice, they have a bone-loss phenotype with decreased trabecular bone mass 8 . Aged Wnt5a +/− mice have an increased population of B cells and develop myeloid and B-cell leukemia 9 . In addition, it has been suggested that post-natal abnormalities in Wnt5a signaling are involved in inflammatory diseases, as well as cancers 1 .
For instance, expression of Wnt5a and Fz5 can be induced through Toll-like receptors (TLR) stimulated by Mycobacterium tuberculosis (M. tuberculosis) or lipopolysaccharide (LPS) in macrophages, and Wnt5a signaling is involved in producing pro-inflammatory cytokines, such as interleukin (IL)-12 and IL-6 [10][11][12] . Wnt5a has been detected in granulomatous lesions in the lungs of patients with M. tuberculosis, bone marrow macrophages in septic patients, and macrophages accumulated within the intima of atherosclerotic patients 10,11,13 . Thus, these results suggest that Wnt5a released from macrophages in response to inflammatory cues affects macrophages in an autocrine manner to release cytokines. However, the underlying mechanism by which Wnt5a regulates inflammation remains enigmatic. Although purified Wnts are useful tools to better understand the roles of Wnt signaling in the inflammatory responses, it has been reported that recombinant Wnt preparations contain TLR agonists that may lead to inflammatory cytokine production 14 . Therefore, further in vivo studies elucidating the roles of Wnt5a signaling in the immune responses using adult Wnt5a knockout mice are necessary.
Interferon-γ (IFN-γ ) is a key immunoregulatory protein that plays a major role in the host innate and adaptive immune responses 15 . IFN-γ is mainly produced in Th 1 cells, which are differentiated from naïve T cells by IL-12 released from antigen-presenting cells, including dendritic cells (DCs) and macrophages 15,16 . Engagement of IFN-γ with its receptor leads to the activation of Janus kinase (JAK) and the phosphorylation of signal transducer and activator of transcription (STAT)-1 17,18 . STAT1 then translocates into the nucleus where it binds to DNA and initiates the transcription of the STAT1 target genes. IFN-γ also has a priming function and increases inflammatory cytokine production, including tumor necrosis factor-α (TNF-α ), IL-6, and IL-12, in response to TLR ligands in DCs and macrophages 15,16 .
Thus, it appears that IFN-γ and IL-12 form a signaling circuit between Th 1 cells and antigen-presenting cells.
Here we use dextran sodium sulfate (DSS)-induced colitis in mice as a model for inflammatory diseases and show that disease symptoms were milder in Wnt5a and Ror2 conditional knockout mice than control mice. Blocking Wnt5a signaling also reduced the production of pro-inflammatory cytokines in the colon. Finally, we demonstrate that the Wnt5a-Ror2 axis enhances the priming action of IFN-γ to increase TLR-dependent production of IL-12 in DCs, thereby promoting immune responses.

Results
Wnt5a knockout mice were less susceptible to DSS-induced colitis. Given that Wnt5a −/− mice suffer perinatal lethality 5 , we generated Wnt5a flox/flox (Wnt5a fl/fl ) mice (Figures S1A-C), in which exon 2 of the Wnt5a gene was flanked by loxP sites. These mice were crossed with different Cre-expressing mice, including CAG-Cre/ERT2 Tg mice, and the offspring were treated with DSS. CAG-Cre/ERT2 Tg mice show ubiquitous expression of Cre/ERT2 which is activated by administration of tamoxifen 19 . In the Wnt5a fl/fl ;CAG-Cre/ERT2 Tg mice, Wnt5a exon2 was deleted in the colon by administering tamoxifen (Wnt5a CAGΔ/Δ mice) ( Figure S1D). Wnt5a mRNA expression levels varied between Wnt5a fl/fl mouse tissues, but it was remarkably lower in the liver and bone marrow than other tissues (Fig. 1a). In Wnt5a CAGΔ/Δ mice, Wnt5a mRNA was dramatically decreased in the colon, small intestine, stomach, bone marrow, heart, muscle, and brain; and expression was reduced by half in the liver, spleen, thymus, and lung (Fig. 1b). DSS (2.5%) was delivered in drinking water to Wnt5a CAGΔ/Δ mice and their Wnt5a fl/fl littermates. Wnt5a CAGΔ/Δ mice showed less weight loss than the WT mice following colitis induction (Fig. 2a). DSS caused bleeding in the stools (Fig. 2b) and pasty stools (Fig. 2c) at 5-6 days after its administration. These phenotypes were observed later and to a lesser degree in Wnt5a CAGΔ/Δ mice (Fig. 2b,c). The overall severity of colitis measured as the disease activity index (DAI), which included scores for body weight loss, occult and gross stool bleeding, and stool consistency, was milder in Wnt5a CAGΔ/Δ mice than Wnt5a fl/fl mice (Fig. 2d). These results suggested that Wnt5a CAGΔ/Δ mice were less susceptible to DSS-induced colitis.
The histological patterns in the colon of DSS-fed mice were classified into 4 categories (Figures S2A-D); (1) intact crypts, (2) decreased crypt lesions, (3) monolayer lesions, and (4) ulcer lesions. In WT mice, the frequencies of decreased crypt, monolayer, and ulcer lesions increased gradually over 5 days of DSS administration ( Figure S2E). However, the histopathological scores of epithelium damage were reduced in the DSS-fed Wnt5a CAGΔ/Δ mice compared with Wnt5a fl/fl mice ( Fig. 2e and Figure. S2F). Although several Wnt mRNAs, including Wnt1, Wnt4, Wnt5a, Wnt5b, Wnt6, and Wnt11, were expressed in the colon, Wnt5a mRNA was expressed more highly than other Wnt mRNAs, and significantly increased by DSS administration (Fig. 2f). Thus, Wnt5a might be involved in DSS-induced colitis.
Wnt5a was required for inflammatory cytokine production in the colon. Many ulcer lesions were observed in the area within a 1 cm distance from the anus in DSS-induced colitis. In the lesions of Wnt5a fl/fl mice, DSS treatment induced increased mRNA expression levels of IL-6, TNF-α , IL-12a, IL-12b, and IFN-γ (Fig. 3a). The increases were suppressed in the colon of Wnt5a CAGΔ/Δ mice (Fig. 3a). The mRNA levels of IL-17a, IL-10, and transforming growth factor-β1 (TGF-β 1) were increased in the colons of both Wnt5a fl/fl and Wnt5a CAGΔ/Δ mice by DSS treatment, but their expression levels were not significantly Scientific RepoRts | 5:10536 | DOi: 10.1038/srep10536 reduced in Wnt5a CAGΔ/Δ mice (Fig. 3a). The IL-23a mRNA level was not changed in the colon of DSS-fed Wnt5a fl/fl and Wnt5a CAGΔ/Δ mice (Fig. 3a). Enzyme-linked immunosorbent assay (ELISA) confirmed that DSS-induced production of IL-6 and TNF-α was decreased in the Wnt5a CAGΔ/Δ mouse colon and that production of IL-10, IL-17A, and IL-23A was not changed in Wnt5a CAGΔ/Δ and Wnt5a fl/fl mice (Fig. 3b). These results suggested that DSS-induced pro-inflammatory cytokine synthesis in the colon was decreased in Wnt5a CAGΔ/Δ mice compared with control Wnt5a fl/fl mice.
It is thought that intestinal inflammation is caused by an imbalance between the inflammatory response and tolerance 20,21 . Therefore, the frequencies of CD4 + Th 1 , Th 17 , and T reg cells in various tissues were compared in Wnt5a fl/fl and Wnt5a CAGΔ/Δ mice. In the colon, Wnt5a CAGΔ/Δ mice had reduced frequencies of IFN-γ -producing cells (Th 1 cells) compared with Wnt5a fl/fl mice in the absence or presence of DSS treatment (Fig. 3c,d). In contrast, the frequency of IL-10-producing T cells (T reg cells) was slightly, not significantly, decreased, and the frequency of IL-17-producing T cells (Th 17 cells) was not affected by Wnt5a deletion (Fig. 3c,d). In the small intestine, although the frequency of Th 17 cells was slightly, not significantly, decreased by Wnt5a deletion, the frequencies of others were not affected ( Figure S3A). Loss of Wnt5a did not affect the frequencies of Th 1 , Th 17 , and T reg cells in the mesenteric lymph nodes, spleen, and thymus ( Figure S3B). These results suggested that Wnt5a deficiency suppresses Th 1 polarization process, including T cell recruitment and differentiation 22,23 , in the colon, thereby attenuating DSS-induced colitis. However, it is unlikely that T cell recruitment from the mesenteric lymph nodes to the colon was decreased in Wnt5a CAGΔ/Δ mice, because the number of total CD4 + T cells in the colon and the mesenteric lymph nodes was not changed between Wnt5a fl/fl and Wnt5a CAGΔ/Δ mice ( Figure S3C). The number of Th 1 cells was reduced in the colon of Wnt5a CAGΔ/Δ mice rather than that of Wnt5a fl/fl mice, whereas that of T cells was not increased in the mesenteric lymph nodes ( Figure S3C). Therefore, Wnt5a might be involved in Th 1 differentiation. might be caused by the loss of Wnt5a in hematopoietic cells. To address this question, Wnt5a fl/fl ;Mx-Cre Tg mice were generated. Mx-Cre mainly deletes genes in hematopoietic cells 25 . In these mice Wnt5a mRNA expression levels were decreased in the liver, thymus, and bone marrow by the peritoneal administration of polyinosine-polycytosine (pIpC) (Wnt5a MxΔ/Δ mice) (Fig. 1b). Genomic DNA isolated from the bone marrow of Wnt5a MxΔ/Δ mice showed efficient digestion of Wnt5a exon2 compared with that of Wnt5a fl/fl mice ( Figure S4A). However, the Wnt5a MxΔ/Δ mice had colitis phenotypes that were similar in terms of severity to Wnt5a fl/fl mice ( Figures S4B and C). The results are shown as means ± SD (a-e) or SE (f). *P < 0.05, **P < 0.01 as calculated by one-way ANOVA (a-e) and by the Student's t-test (f). Villin-Cre depletes genes in the epithelial cells of the intestines 26 . In Wnt5a fl/fl ;Villin-Cre Tg mice (Wnt5a VilΔ/Δ mice), Wnt5a genomic DNA was digested in the colon ( Figure S4D). However, the reduction of Wnt5a mRNA in the colon was not clearly observed ( Figure S4D Figure 3c were expressed as means ± SD. The results are shown as means ± SE (a and b) or SD (d). *P < 0.05, **P < 0.01 as calculated by the Student's t-test (a and b) and by one-way ANOVA (d).

Loss of
Scientific RepoRts | 5:10536 | DOi: 10.1038/srep10536 Ror2 deficient mice were also less susceptible to DSS-induced colitis. Ror2 is one of the receptors for Wnt5a. Ror2 −/− mice are embryonic lethal and display phenotypes similar to Wnt5a −/− mice 27 . To examine whether Ror2 is involved in DSS-induced colitis, Ror2 fl/fl ;CAG-Cre/ERT2 Tg and Ror2 fl/fl ;Mx-Cre Tg mice were administrated tamoxifen (Ror2 CAGΔ/Δ ) and pIpC (Ror2 MxΔ/Δ ), respectively, and then DSS was administrated to these mice. Ror2 mRNA was variably expressed in Ror2 fl/fl mouse tissues, but the liver, thymus, and muscle had the lowest expression levels (Fig. 1c). Ror2 mRNA expression was reduced in most tissues of Ror2 CAGΔ/Δ mice, and especially in the liver, bone marrow, and heart of the Ror2 MxΔ/Δ mice (Fig. 1d). Compared with the Ror2 fl/fl mice, both types of Ror2 conditional knockout mice exhibited milder signs and pathological changes following DSS-induced colitis (Fig. 4a,b). The histopathological scores of epithelium damage were also reduced in the DSS-fed Ror2 MxΔ/Δ mice compared with Ror2 fl/ fl mice ( Fig. 4c and Figure. S2G). It is noteworthy that Ror2 MxΔ/Δ mice suppressed DSS-induced colitis in contrast to Wnt5a MxΔ/Δ mice. The production of pro-inflammatory cytokines, including IL-6 and TNF-α , was suppressed in the colon of Ror2 MxΔ/Δ mice, similar to the Wnt5a CAGΔ/Δ mice (Fig. 4d,e). Taken together, these results suggested that Ror2 in hematopoietic cells was involved in progression of DSS-induced colitis.
Wnt5a was released primarily from fibroblasts and Ror2 was expressed in DCs in the colon. To examine which cells are the major source of Wnt5a, the colons from mice fed with DSS were analyzed for Wnt5a expression. Wnt5a mRNA was expressed in fibroblasts more highly than epithelial cells and hematopoietic cells, including CD4 + T cells, B220 + B cells, CD11c + DCs, CD11b + CD11c − macrophages, and γ δ T cells in the colon, and its expression level was elevated only in the fibroblasts after DSS administration ( Figure S5A). In DSS-fed mice, Wnt5a protein and Wnt5a mRNA were not detected in the intact crypt regions of the colon, whereas the protein and mRNA were clearly visible in the mesenchyme of the ulcer lesions ( Fig. 5a and Figure. S6). Wnt5a-expressing cells were positive for vimentin (a mesenchymal cell marker), but not for F4/80 (a macrophage marker) and CD11c (a DC marker) (Fig. 5b), suggesting that they were fibroblasts. In addition, Wnt5a was still expressed in the mesenchyme of ulcer lesions in Wnt5a MxΔ/Δ mice but not in Wnt5a CAGΔ/Δ mice (Fig. 5c). Taken together, these results suggested that Wnt5a secreted from cells other than hematopoietic cells, probably fibroblasts, affects DSS-induced colitis. Wnt5a expression was then examined in the colon from Crohn's disease (9 cases) and ulcerative colitis (10 cases) patients. Wnt5a was not detected in the intact crypt regions (E-cadherin positive) (Fig. 5d), but it was clearly detected in the mesenchyme of ulcer lesions (E-cadherin negative) in about half of the inflammatory bowel disease cases (Fig. 5e). As in the mice, the Wnt5a was detected in vimentin-positive cells but not in the cells expressing CD68, a macrophage and monocyte marker (Fig. 5f,g).
In contrast to Wnt5a, Ror2 mRNA was expressed in hematopoietic cells although it was lower than in fibroblasts ( Figure S5B). Among hematopoietic cells, DCs as well as B cells, macrophages, and γ δ T cells, showed higher Ror2 mRNA expression compared with T cells. However, DSS treatment reduced Ror2 mRNA expression in B and γ δ T cells ( Figure S5B), suggesting that these cells are hard to respond to Wnt5a in DSS-induced colitis. In addition, cell numbers of γ δ T cells were not changed between Wnt5a CAGΔ/Δ , Ror2 MxΔ/Δ , and their control mice irrespective of the presence or the absence of DSS treatment ( Figure S7). Taken together with the observations that γ δ T cells play a role in protections against DSS-induced colitis 28 , DCs might respond to Wnt5a in colitis rather than B, T, and γ δ T cells.
Intestinal DC subsets have been identified by the combination and expression levels of specific cell surface antigens and characterized by their functions in immune responses 29,30 . We classified CD11c + DCs into three subsets, including CD11c + CD11b − CD103 + cells that induce Th 1 polarization, CD11c + CD11b + CD70 + CX 3 CR1 intermediate cells that induce Th 17 polarization, and CD11c + CD11b + CD70 + CX 3 CR1 high cells that inhibit T cell proliferation [29][30][31][32] . The Ror2 mRNA levels varied among subsets and CD11c + CD11b − CD103 + DC cells indeed expressed Ror2 mRNA ( Figure S5C), and loss of Wnt5a or Ror2 did not affect total cell numbers of DC subsets in the presence or the absence of DSS treatment ( Figure S7). These results suggest that Wnt5a-Ror2 signaling is involved in the activation of intestinal DCs expressing Ror2 rather than their differentiation.

The Wnt5a-Ror2 axis was involved in pro-inflammatory cytokine synthesis in DCs. Intestinal
DCs and macrophages play important roles in the regulation of gut homeostasis through induction of helper T cell subsets 20,21 . Therefore, intracellular signaling cascades for cytokine productions involved in Th 1 differentiation were examined in DCs. LPS induced the expression of IL-12a, IL-12b, IL-23a, and IL-6 mRNA in CD11c + DCs from Wnt5a fl/fl and Ror2 fl/fl mice (Fig. 6a,b). However, LPS-induced increases in the expression of these cytokines were reduced in DCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice (Fig. 6a,b). Colon CD11c + DCs induces Th 1 differentiation from naïve T cells in vitro 33,34 . The ability of DCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice to induce Th 1 differentiation was decreased compared with those from Wnt5a fl/fl and Ror2 fl/ fl mice (Fig. 6c), suggesting that Wnt5a signaling through Ror2 in CD11c + DCs is required for cytokine expression which induces Th 1 differentiation. It is noteworthy that LPS-induced production of IL-12a, IL-12b, IL-23a, and IL-6 mRNA was not suppressed in colon CD11c + DCs from Wnt5a MxΔ/Δ mice (Fig. 6d), which are consistent with the observation that Wnt5a MxΔ/Δ mice showed the phenotypes similar to control mice in DSS-induced colitis (see Figures S4B and C).   The Wnt/β -catenin pathway in intestinal DCs is required for immunosuppression 35 . Although Wnt5a has been shown to inhibit the β -catenin pathway 1,4 , mRNA expression levels of Axin2, a target gene of β -catenin signaling, did not increase in CD11c + DCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice (Fig. 6a,b). This suggested that the phenotypes induced by Wnt5a and Ror2 deficiency were not due to the activation of the Wnt/β -catenin pathway.

The Wnt5a-Ror2 axis was also involved in functions of bone marrow-derived DCs.
To examine the roles of Wnt5a signaling in DC functions further, bone marrow cells were isolated from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice and incubated with granulocyte macrophage colony-stimulating factor (GM-CSF) for 7 days to generate bone marrow-derived DCs (BMDCs). Loss of Wnt5a or Ror2 suppressed IL-12a, IL-12b, IL-23a, and IL-6 mRNA expression in BMDCs stimulated with LPS (Fig. 7a,b). Ectopic expression of Wnt5a did not affect the basal levels of IL-12a, IL-12b, IL-23a, IL-6, and IL-10 mRNA expression, but did enhance LPS-induced pro-inflammatory cytokine mRNA expression (Fig. 7c). ELISA confirmed that ectopically expressed Wnt5a promotes LPS-induced secretion of IL-6 and IL-12p40 but not that of IL-10 (Fig. 7d). However, Wnt5a did not affect the expression of CD80 and CD86, which indicate DC maturation, in BMDCs ( Figure S8), suggesting that Wnt5a signaling did not affect the phenotype of DCs.
To test whether Wnt5a secreted from other cells affects DC functions, bone marrow cells were cocultured with mouse fibroblast cell lines, L cells stably expressing Wnt5a in the presence of GM-CSF for 7 days. However, this coculture experiment failed, because L cells were detached from dishes after 5 days. Instead, HeLaS3 cells stably expressing Wnt5a (HeLaS3/Wnt5a cells) were used in this experiment ( Figure S9A). Knockdown of Wnt5a efficiently reduced both endogenous and exogenous Wnt5a and the phosphorylation of Dishevelled 2 (Dvl2), which indicates the activation of Wnt5a signaling ( Figure S9A). HeLaS3 cells neither expressed IL-12 mRNA nor induced IL-6 mRNA in response to LPS regardless of Wnt5a expression levels ( Figure S9B). Under these conditions, coculture of bone marrow cells with HeLaS3/Control cells promoted LPS-induced secretion of IL-12p40 and IL-6 in the resultant BMDCs (Fig. 7e), because HeLaS3 cells express Wnt5a endogenously 36 ( Figure S9A). Coculture with HeLaS3/ Wnt5a cells further enhanced it (Fig. 7e). Enhanced secretions of IL-12p40 and IL-6 were suppressed by knockdown of Wnt5a in these HeLaS3 cells (Fig. 7e). Taken together, these results suggested that the Wnt5a-Ror2 axis between non-hematopoietic cells and DCs promoted pro-inflammatory cytokine production in DCs.
The Wnt5a-Ror2 axis promoted the priming action of IFN-γ in DCs. Finally, the mechanism by which Wnt5a promotes inflammation was examined. LPS activates nuclear factor kappa B (NF-κ B), JNK, and p38 through TLR4 to induce the expression of IL-12, IL-6, and TNF-α in DCs 37 . When control BMDCs from Wnt5a fl/fl and Ror2 fl/fl mice were stimulated with LPS, inhibitor of kappa B-α (Iκ B-α ) was degraded to induce the nuclear translocation of the NF-κ B p65-subunit (NF-κ Bp65), and JNK and p38 were also phosphorylated ( Figures S10A-D). There were no significant differences in the activation of these signaling pathways in BMDCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice compared with Wnt5a fl/ fl and Ror2 fl/fl mice, respectively ( Figures S10A-D). Therefore, the Wnt5a-Ror2 axis might promote LPS-dependent cytokine production without affecting the TLR4 signaling pathway directly.
IL-12 from DCs and macrophages plays a critical role in Th 1 differentiation 15,38 , and FACS analyses revealed that IFN-γ -producing Th 1 cells in the colon were decreased in Wnt5a CAGΔ/Δ mice (see Fig. 3c,d). Therefore, the mechanism underlying how the Wnt5a-Ror2 axis affects the transcription of the IL-12b gene in DCs was examined using the chromatin immunoprecipitation (ChIP) assay. LPS induces the recruitment of RNA polymerase II (Pol II) and NF-κ Bp65 to the IL-12b promoter in macrophages 39 . Pol II and NF-κ Bp65 were recruited to the IL-12b promoter region by LPS stimulation in BMDCs from Wnt5a fl/fl and Ror2 fl/fl mice, but their recruitment was diminished in BMDCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice (Fig. 8a,b). LPS also enhanced histone H4K8-acetylation (H4K8-Ac) at the IL-12b gene in BMDCs from Wnt5a fl/fl and Ror2 fl/fl mice, which was decreased in BMDCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice (Fig. 8a,b). Thus, the Wnt5a-Ror2 axis may enhance LPS-induced transcription of the IL-12b gene in DCs through the formation of Pol II-containing transcription initiation complexes.
IFN-γ primes macrophages for the expression of IL-12b gene 16,40,41 . Given that the actions of Wnt5a are similar to the effects of IFN-γ priming, we examined whether the Wnt5a-Ror2 axis affects IFN-γ signaling in DCs. IFN-γ induced the phosphorylation of JAK1 and STAT1 in BMDCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice, but at a lower level than was observed in control BMDCs from Wnt5a fl/fl and Ror2 fl/fl mice (Fig. 8c,d). In addition, the IFN-γ -induced recruitment of STAT1 to the IL-12b promoter was decreased in BMDCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice compared with their control BMDCs (Fig. 8e,f). However, Src activity, which is involved in the Wnt5a and STAT signaling 42,43 , was not changed in BMDCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice (Figures S10E and F). Taken together, these results suggested that the Wnt5a-Ror2 axis enhances the priming action of IFN-γ , although the signaling pathway by which Wnt5a activates to regulate this action is not known at present.

Discussion
Our results support that the Wnt5a-Ror2 axis promotes DSS-induced colitis by enhancing pro-inflammatory cytokine production in the colon. There are two major pathways for Wnt signaling.
Scientific RepoRts | 5:10536 | DOi: 10.1038/srep10536 The first is the β -catenin dependent pathway, and the second is the β -catenin independent pathway, which is activated by Wnt5a. The β -catenin dependent pathway is activated in intestinal DCs and is required for the secretion of immunosuppressive cytokines 35 . Although Wnt5a signaling is able to suppress the β -catenin dependent pathway 1,4 , our results showed that the β -catenin dependent pathway was not activated in CD11c + DCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice. Therefore, the β -catenin independent pathway may regulate intestinal inflammatory responses independently of the β -catenin dependent pathway at least in DCs. It is intriguing to speculate that Wnt signaling may trigger anti-or pro-immune responses through the β -catenin dependent or independent pathway, respectively, in intestinal DCs.
The clinical signs, histological damage, and pro-inflammatory cytokine levels in DSS-induced colitis were suppressed in Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice compared with control mice. In turn, the Wnt5a MxΔ/Δ and Wnt5a VilΔ/Δ mice showed colitis phenotypes similar to control mice. The latter finding suggests that Wnt5a released from hematopoietic cells (Wnt5a MxΔ/Δ ) or epithelial cells (Wnt5a VilΔ/Δ ) in the colon was not involved in DSS-induced colitis. Our results demonstrated that in mice, the basal expression level of Wnt5a in intestinal fibroblasts is much higher than hematopoietic cells or epithelial cells, and that DSS administration induces Wnt5a expression in fibroblasts located in ulcer lesions. The observation that LPS-induced pro-inflammatory cytokines produced from CD11c + DCs in Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice were decreased but not in Wnt5a MxΔ/Δ mice also support that hematopoietic cells are not the primary source of Wnt5a in the colon. In addition, our human studies show that Wnt5a was clearly observed in vimentin-positive cells, but not macrophages, in the ulcer lesions of patients with ulcerative colitis and Crohn's diseases. Taken together, these results suggest that fibroblasts might be the primary source of Wnt5a at least in the colon during inflammatory bowel disease.
It has been reported that Wnt5a can act in a paracrine manner on neighboring cells. For example, the Wnt5a-Ror2 axis between osteoblast-lineage cells and osteoclast precursors enhances osteoclastogenesis, while Ror2 deficiency in osteoclast precursors leads to impaired osteoclastogenesis 8 . In addition, the Wnt5a-Fz8/Flamingo axis functionally maintains HSCs in their niche through interactions between N-cadherin-positive osteoblasts, which express Wnt ligands including Wnt5a, and long-term HSCs that express Fz8 and Flamingo 44 . Similarly, our results suggest that Wnt5a from non-hematopoietic cells (probably fibroblasts) act on DCs, which express Ror2, to promote inflammation in response to infectious cues. However, knockdown of Wnt5a in murine macrophages suppresses the secretion of IFNs in vitro 12 . Wnt5a secreted from human DCs is involved in IL-12 secretion in autocrine manner and enhances IFN-γ secretion from CD4 + T cells in paracrine manner 45 . Thus, it cannot be excluded the possibility that Wnt5a promotes cytokine production from hematopoietic cells in an autocrine manner, at least in vitro, in different conditions. Intestinal DCs are the first cells recruited to and activated at sites of infection or injury. DCs play an important role in the gut where they continually sample intestinal antigen and participate in inducing distinct immune responses 21 . IL-12 production from DCs in particular is essential for Th 1 differentiation 15,38 , and IFN-γ -producing Th 1 cells in the colon were indeed decreased in the Wnt5a CAGΔ/Δ mice in this study. Although it might be possible that the Wnt5a-Ror2 axis is involved in the recruitment of naïve T cells into the colon 22,23 , this possibility is unlikely, because the T cell number in the colon was not changed in Wnt5a fl/fl and Wnt5a CAGΔ/Δ mice (see Figure S3C). Therefore, our results provided one possible mechanistic link between Wnt5a signaling and DC functions.
Wnt5a reportedly promoted NF-κ B signaling through the activation of PKC or Rac in endothelial cells and macrophages 12,46 , but this is not the case in DCs. In our study, LPS induced degradation of Iκ B-α and NF-κ Bp65 nuclear localization in BMDCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice, as well as BMDCs from control mice. Instead, the Wnt5a-Ror2 axis was required for LPS-induced IL-12 expression by enhancing Pol II recruitment, NF-κ Bp65 occupancy, and histone acetylation. Therefore, it is likely that the Wnt5a-Ror2 axis is involved in promoting chromatin remodeling to increase recruitment of TLR-activated transcription factors at the IL-12b gene. This mechanism of action resembles to that of IFN-γ . As another receptor of Wnt5a, Ryk (related to receptor tyrosine kinase) atypical tyrosine kinase receptor, is involved in Wnt5a signaling-mediated axon guidance 47,48 and Xenopus gastrulation movement 49 , and it is expressed in immature DCs and CD4 + T cells 45 . To clarify whether Ryk mediates Wnt5a-dependent immune responses would be necessary for the understanding of roles of Wnt5a signaling in inflammation.
An important function of IFN-γ is to prime DCs and macrophages for synergistic transcription of pro-inflammatory cytokine genes by inflammatory factors, through STAT1 activation 15,16,38 . For instance, IFN-γ induces TLR4 dependent transcription factors and RNA polymerase II to occupy cytokine gene promoters and enhancers, thereby increasing expression of TNF-α , IL-6, and IL-12, in macrophages 16 . Our results demonstrated that IFN-γ dependent JAK activation and STAT1 phosphorylation were decreased in BMDCs from Wnt5a CAGΔ/Δ and Ror2 MxΔ/Δ mice. It is possible that the Wnt5a-Ror2 axis promotes the IFN-γ priming action in DCs. Loss of Wnt5a or Ror2 did not affect the expression of JAK1 and STAT1 (see Fig. 8c,d) and the basal levels of IFN-γ receptor1/2 mRNA (data not shown). Thus, Wnt5a-induced expression of other genes might decide the threshold for IFN-γ signaling although the detailed mechanism is not known at present. The reduction in the Th 1 differentiation in Wnt5a CAGΔ/Δ mice could be due primarily to the decrease in IL-12 production from DCs. IFN-γ is mainly produced by Th 1 cells, natural killer cells, and CD8 + T cells. IL-12 is important for Th 1 differentiation and IFN-γ synthesis, and IFN-γ promotes IL-12 production in DCs. Therefore, the Wnt5a-Ror2 axis would regulate the priming action of IFN-γ and contributes to the signaling circuit between IL-12 and IFN-γ .

Methods
Animals. The protocols used for all animal experiments in this study were approved by the Animal Research Committee of Osaka University, Japan (No. 21-048-1). All animal experiments were carried out according to the guidelines for the care and use of experimental animals of Osaka University. Details of Wnt5a fl/ fl mouse line generation are described in Supplementary Methods. Ror2 fl/fl mice were generated as described 8 .
CAG-Cre/ERT2 Tg and Villin-Cre Tg mice were purchased from Jackson Laboratory. To treat the mice, a stock solution of tamoxifen (4-hydroxitamoxifen; Sigma-Aldrich, St. Louis, MO) in ethanol (150 mg/ml) was diluted in corn oil to 15 mg/ml. The tamoxifen suspension (0.1 ml) was administered to adult Wnt5a fl/fl or Ror2 fl/fl ;CAG-Cre/ERT2 Tg mice 4 weeks of age and relevant control Wnt5a fl/fl or Ror2 fl/fl mice three times at two day intervals to induce the activation of the Cre-ERT2 recombinase and to remove the floxed Wnt5a or Ror2 coding region 19 . Mx-Cre expression was induced by the intraperitoneal injection of 500 μ g pIpC (Sigma-Aldrich) six times at two day intervals to remove the floxed Wnt5a or Ror2 coding region 25 . The primers used for genotyping in this study were listed in Supplementary Table S1.
Induction of colitis. DSS was administered to eight to nine weeks old male mice in their drinking water at a final concentration of 2.5% (w/v) to induce colitis. Animals were observed daily for weight, stool consistency, and the presence of gross blood in feces and at the anus. Detailed evaluation of colitis is described in Supplementary Methods. Preparation of colon extraction. After 7 days administration of DSS, 2 cm length of colon from anus was isolated, and homogenized in homogenization buffer (25 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM phenylmethylsulfonylfluoride, 1 μ g/ml leupeptin, 1 μ g/ml aprotinin) by downs homogenizer (15 strokes). After the homogenization, 1% Nonidet P-40, 1% deoxycholic acid and 0.1% sodium dodecyl sulfate (SDS) were added to the homogenates. The homogenates were sonicated 6 times for 15 sec by Ultra S homogenizer (TAITEC, Nagoya, Japan), incubated at 4 ˚C with rotation for 1 h and then centrifuged at 20,000 x g for 10 min. The supernatants were used in ELISA.

Patients and tissues of inflammatory bowel diseases.
Tissue samples from 20 patients who underwent surgery for Crohn's disease and ulcerative colitis at Osaka University Hospital from 2011 to 2013 were examined. Histological specimens were fixed in 10% formalin and routinely processed for paraffin embedding. Paraffin-embedded specimens were stored in the dark room in the Department of Pathology at Osaka University Hospital at room temperature and cut into 4 μm thick sections at the time of staining. The study was approved by the ethics review board of Graduate School of Medicine, Osaka University (No. 13455).

Isolation of lamina propria DCs, bone marrow-derived DCs (BMDCs), and lymphocytes.
Isolations of lamina propria DCs, BMDCs, and lymphocytes were performed as described by Ueda Y. et al. 50 , Kayama H. et al. 51 , and Kusu T. et al. 52 , respectively. Details of experimental procedures are described in Supplementary Methods.

Coculture of naïve CD4 + T cells with lamina propria DCs.
Coculture of naïve CD4 + T cells with lamina propria DCs were performed as described by Atarashi K. et al. 53 and Kayama H. et al. 51 . Details of experimental procedures are described in Supplementary Methods.

General experimental procedures. General experimental procedures are described in Supplementary
Methods.

Statistical analysis.
The experiments in each figure were performed three to four times, and differences between control and experimental groups were evaluated using the Student's t-test and a one-way ANOVA with a Bonferroni test for multiple group comparison. A P value < 0.05 was considered a significant difference.
Others. Quantification data was calculated based on at least three blots or gels from different experiments. All the gels were run under the same experimental condition as detailed in the Supplementary Methods.