The CTRP3-AdipoR2 Axis Regulates the Development of Experimental Autoimmune Encephalomyelitis by Suppressing Th17 Cell Differentiation

C1q/TNF-related proteins (CTRP) including CTRP3 are a group of secreted proteins which have a complement C1q-like domain in common, and play versatile roles in lipid metabolism, inflammation, tumor metastasis and bone metabolism. Previously, we showed that the expression of C1qtnf3, encoding CTRP3, is highly augmented in joints of autoimmune arthritis models and CTRP3-deficiency exacerbates collagen-induced arthritis in mice. However, the mechanisms how CTRP3-deficiency exacerbates arthritis still remain to be elucidated. In this study, we showed that CTRP3 was highly expressed in Th17 cell, a key player for the development of autoimmune diseases, and Th17 cell differentiation was augmented in C1qtnf3–/– mice. Th17 cell differentiation, but not Th1 cell differentiation, was suppressed by CTRP3 and this suppression was abolished by the treatment with a receptor antagonist against AdipoR2, but not AdipoR1, associated with suppression of Rorc and Stat3 expression. Furthermore, AdipoR1 and AdipoR2 agonist, AdipoRon suppressed Th17 cell differentiation via AdipoR2, but not AdipoR1. The development of myelin oligodendrocyte glycoprotein (MOG)-induced experimental autoimmune encephalomyelitis was enhanced in C1qtnf3 –/– mice associated with increase of Th17 cell population. CTRP3 inhibited MOG-induced IL-17 production from T cells by affecting both T cells and dendritic cells. These results show that CTRP3 is an endogenous regulator of Th17 differentiation, suggesting that the CTRP3-AdipoR2 axis is a good target for the treatment of Th17 cell-mediated diseases.


INTRODUCTION
C1q/TNF-related protein (CTRP) superfamily consists of more than 30 family members including TNF, adiponectin, CTRP3 (also known as CORS26, cartducin, or cartnectin) and CTRP6, which have a complement C1q domain-like structure in common (1,2). Many of CTRP family members are categorized as adipokines, because these CTRP family members are secreted from adipose tissues. Functionally, this superfamily was suggested to be involved in a wide range of physiological and pathological processes such as lipid metabolism, inflammation, tumor metastasis and bone metabolism (3).
Progestin and adipoQ receptor (PAQR) family members are the receptors for the CTRP family members, and PAQR1 (also known as AdipoR1), PAQR2 (AdipoR2) and PAQR3 (AdipoR3) are identified as the receptors for adiponectin (4). CTRP6 regulates adipocyte proliferation and differentiation as well as lipogenesis in myoblasts via AdipoR1 (5,6). CTRP9 protects against acute cardiac damage and high glucose-induced endothelial oxidative damage, and inhibits macrophage-mediated inflammatory response against oxidized LDL via AdipoR1 (7)(8)(9). Thus, some CTRP family members share AdipoR1 and AdipoR2 as the common receptor, although complete ligand-receptor relationship has not been elucidated yet.
Adiponectin is suggested to inhibit dendritic cell (DC) activation, IFN-g-producing helper T (Th1) cell differentiation and IL-17-producing helper T (Th17) cell differentiation via AdipoR1, AdipoR2 and unknown receptor(s) (10)(11)(12). AdipoRon, an agonist of AdipoR1 and AdipoR2 (13), treatment inhibited Th17 cell polarization in bleomycin-treated mice (14). However, other reports indicate that adiponectin induces DC activation and Th17 cell differentiation (15,16). Furthermore, adiponectin activates the complement classical pathway through binding to C1q, whereas it inhibits the complement alternative pathway through binding to factor H (17,18). CTRP6 inhibits the complement alternative pathway by competing C3(H 2 O) binding to factor B (19). CTRP6 also inhibits the lectin pathway (20). On the other hand, CTRP3 does not affect the complement system (21). Thus, the roles of each CTRP family member in the immune system seem very complex and further investigation is required to elucidate the mechanisms.
In this report, we investigated involvement of CTRP3 in the development of Th17 cells, because CTRP3 shares AdipoR2 with adiponectin, which is suggested to be involved in the differentiation of Th17 cells (10,11,34). Furthermore, we examined the effect of C1qtnf3-deficiency on the development of experimental autoimmune encephalomyelitis (EAE), in which Th17 cells play a crucial role (35). We found that CTRP3 suppresses Th17 cell differentiation via AdipoR2. We showed that Th17 cell differentiation is enhanced in C1qtnf3 -/mice and the development of EAE is exacerbated in these mutant mice. Interestingly, however, Th1 cell differentiation and DC function were intact in C1qtnf3 -/mice.

Ethics Statement
All experiments were approved by the Animal Care and Use Committee of Tokyo University of Science and Kansai Medical University, and were conducted according to the institutional ethical guidelines for animal experiments and safety guidelines for gene manipulation experiments.

DC Activation Assay
According to a previous report (38), BMDCs were differentiated from bone marrow cells (BMCs) in femurs and tibiae by GM-CSF stimulation. Briefly, BMCs (5 x 10 4 cells/well) were cultured in 96-well plates under 100 ml RPMI 1640 medium containing 10% FBS supplemented with 20 ng/ml mouse GM-CSF (PeproTech, London, UK), in the absence or presence of recombinant human CTRP3, and a half volume medium was changed by adding the same volume of the same medium at day 3 and day 6. At day 8, cell proliferation was measured by using the Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan) and absorbance at 450 nm was measured using a microplate reader (MTP-300, Corona, Ibaragi, Japan).
For activation assay, BMCs (3 x 10 6 cells/well) were cultured in 100-mm dish (IWAKI, Shizuoka, Japan) under 10 ml of RPMI 1640 medium containing 10% FBS supplemented with 20 ng/ml mouse GM-CSF, and a half volume medium was changed by adding the same volume of the same medium at day 3 and day 6.
To measure the population of infiltrated IL-17 + T cells and IFN-g + T cells into the spinal cords at day 21 after immunization, mice under anesthesia were perfused with PBS to rinse blood off from the spinal cords through the intracardiac route using a peristaltic pump set (ATTO, Tokyo, Japan). The entire spinal columns were removed by gross dissection and spinal cords were ejected by a flush with a syringe attached in an 18-gauge needle, and then, spinal cords were incubated with 200 U/ml collagenase from Clostridium histolyticum (Sigma-Aldrich, St. Louis, MO, USA) in HBSS for 30 min at 37°C. The cell suspension of the spinal cords in 30% Percoll (Sigma-Aldrich, St. Louis, MO, USA) was overlaid on 70% Percoll in a 15 ml polypropylene tube and centrifuged at 2,200 rpm for 20 min at room temperature. Cells were collected from the interface between 30% and 70% Percoll and the population of IL-17 + T cells and IFN-g + T cells were analyzed by flow cytometry according to the standard techniques.

Histological Analysis
According to the previous report (40), sections of the spinal cord were prepared at day 28 after immunization with MOG/CFA. Serial sections (5 mm) were stained with hematoxylin and eosin (H&E) or luxol fast blue (LFB, Nacalai tesque, Kyoto, Japan), and images were taken using a slide scanner (NanoZoomer, Hamamatsu Photonics, Shizuoka, Japan).

Co-Culture of T Cells and BMDCs
T cells were purified from the spleen of 2D2 mice using antimouse Thy1.2 Ab-conjugated magnetic beads (#130-121-278, Miltenyi Biotec, Gladbach, Germany) and an AutoMACS (Miltenyi Biotec, Gladbach, Germany). The T cells were cocultured with BMDCs in the presence of 5 mg/ml MOG 35-55 peptide for 5 days. IL-17 and IFN-g concentrations in the culture supernatants were measured by ELISA.

Statistics
Unless otherwise specified, all results are shown as averages and SEM. Student's t-test was used to evaluate the statistical significance of the results, except that c 2 -test was used to evaluate incidence of EAE and Mann-Whitney U-test was used to evaluate severity score and mean maximum score of EAE. p values of < 0.05 were considered to be statistically significant.

CTRP3 Suppresses Th17 Cell Differentiation via AdipoR2 Receptor
Because C1qtnf3 was highly expressed in Th17 cells ( Figure 1A), we assessed the role of CTRP3 in Th17 cell differentiation. We found that C1qtnf3 -/naïve T cells differentiated to Th17 cells more efficiently compared with naïve T cells from WT mice under Th17 cell-polarizing culture conditions ( Figure 1B). Furthermore, exogenous CTRP3 suppressed Th17 cell differentiation in a dosedependent manner ( Figure 1C). However, CTRP3 did not affect Th1 cell differentiation at all under Th1 cell-polarizing culture conditions ( Figure 1D). These results suggest that CTRP3 specifically inhibits Th17 cell differentiation, but not Th1 cell differentiation, in an autocrine manner.
Consistent with a previous report (15), Adipor1 and Adipor2 mRNA were highly expressed in Th17 cells compared to naïve T cells. Adipor1 mRNA levels in Th1 cells were higher than those in naïve T cells, while Adipor2 mRNA levels were comparable to those in naïve T cells ( Figure 1E). Previously, we identified that AdipoR2, but not AdipoR1, is a functional CTRP3 receptor on chondrogenic cells using siRNAs and blockers for AdipoR1 and AdipoR2 (32). Then, we examined whether or not CTRP3 inhibits Th17 cell differentiation via AdipoR2 using a peptide blocker. We found that AdipoR2 blocker, but not AdipoR1 blocker, abolished Th17-differentiation-inhibitory effect of CTRP3 ( Figure 1F). Next, we analyzed Ampka and Ppara mRNA expression, because AdipoR1 activates AMPK pathway and AdipoR2 activates PPARa pathway, respectively (41). We found that CTRP3 increased Ppara, but not Ampka, mRNA expression in T cells under Th17 cell-polarizing culture conditions ( Figure 1G). AdipoR1 blocker did not affect Ampka mRNA expression, while AdipoR2 blocker abolished the increment of Ppara mRNA expression ( Figure 1G). Furthermore, CTRP3 decreased Rorc and Stat3 mRNA expression. The decreased expression was abolished by AdipoR2 blocker, but not AdipoR1. CTRP3 did not significantly affect Mtor mRNA expression ( Figure 1G). Thus, AdipoR2 is the functional receptor for CTRP3 on naïve T cells to inhibit Th17 cell differentiation.
Furthermore, we assessed the role of AdipoR1 and AdipoR2 in the Th17 cell differentiation inhibition by AdipoRon. We found that AdipoRon suppressed Th17 cell differentiation of naive T cells from C1qtnf3 -/mice, in which no endogenous CTRP3 was present (Figure 2A). AdipoR2 blocker, but not AdipoR1 blocker, abolished the Th17-differentiation-inhibitory effect of AdipoRon ( Figure 2B). These results indicate that AdipoR2 mediates the suppression of Th17 cell differentiation.

Development of EAE Is Exacerbated in
C1qtnf3 -/-Mice Associated With an Increase of Th17 Cells Th17 cells play a pivotal role in the pathogenesis of multiple sclerosis (MS) (42,43) and the mouse model EAE (35,44). Because CTRP3 regulates the differentiation of Th17 cells, we examined the role of CTRP3 in the pathogenesis of EAE. The incidence of EAE was increased in C1qtnf3 -/mice than WT mice ( Figure 3A). The onset of paralysis in C1qtnf3 -/mice was earlier than WT mice ( Figure 3B). Chronological disease scores and the maximum score in C1qtnf3 -/mice were markedly increased compared to WT mice ( Figures 3C, D). The infiltration of immune cells and demyelination in spinal cords at the lumber level were more severe in C1qtnf3 -/mice ( Figures 3E, F). Since infiltration of Th17 cells into central nervous system is required for the initiation of EAE (45,46), we examined infiltration of Th17 cells in the spinal cord. We found that CD4 + T cell ( Figure 3G) and CD4 + IL-17 + T cell (Th17 cell) infiltration increased in C1qtnf3 -/mice compared to WT mice, but infiltrated CD4 + IFN-g + T cell (Th1 cell) population was comparable between WT and C1qtnf3 -/mice ( Figure 3H).
Consistent with our previous report (21), T cell recall proliferative response against MOG peptide was comparable between WT mice and C1qtnf3 -/mice ( Figure 3I). The culture supernatant of MOG-stimulated T cells from C1qtnf3 -/mice contained higher concentrations of IL-17, but not IFN-g, compared to that of WT mice ( Figure 3J). Furthermore, exogenous CTRP3 suppressed IL-17 production from MOGspecific 2D2 cells upon incubation with MOG peptide ( Figure 3K). In contrast, IFN-g production was not affected by CTRP3 at all.
Next, we assessed contribution of T cell-derived CTRP3 and DC-derived CTRP3 on the suppression of Th17 cell differentiation. MOG-specific TCR-expressing T cells were prepared from 2D2 and C1qtnf3 -/-2D2 mice and BMDCs were prepared from WT and C1qtnf3 -/mice, respectively, and they were cocultured in the presence of MOG 35-55 for 5 days and IL-17 and IFN-g production were measured. Higher concentration of IL-17 was detected in the culture supernatant in the co-culture of C1qtnf3 -/-2D2 T cells and WT BMDCs compared with that in WT T cell and WT DC coculture, indicating that T cell-derived CTRP3 suppressed Th17 cell differentiation in an autocrine manner. Furthermore, coculture of WT 2D2 T cells with C1qtnf3 -/-BMDCs also produced higher concentrations of IL-17 than the co-culture of WT 2D2 T cells and WT BMDCs, indicating that Th17 cell differentiation was also suppressed by the CTRP3 produced by DCs in a paracrine manner ( Figure 3L). As CTRP3 was highly expressed in DCs than T cells ( Figure 4A), IL-17 concentration in the co-culture of WT T cells and C1qtnf3 -/-BMDCs was higher than that of C1qtnf3 -/-T cells and WT BMDCs. These observations suggest that CTRP3 regulates the development of EAE by inhibiting Th17 cell differentiation in an autocrine and paracrine manner.

DC Differentiation Is Normal in C1qtnf3 -/-Mice
We found that C1qtnf3 mRNA is highly expressed in DCs and neutrophils among other immune cells and stimulation of DCs with zymosan or LPS did not affect expression of C1qtnf3 mRNA expression ( Figure 4A). Then, we examined the effects of CTRP3 on in vitro differentiation of BMDCs from BMCs. Regarding this,  we and other researchers showed that CTRP3 promotes proliferation of specific types of cells such as chondrogenic cells, endothelial cells and myoblasts (32,(47)(48)(49)(50)(51)(52)(53). We found that cell number of BMDCs from C1qtnf3 -/-BMCs was similar to WT BMC-derived BMDCs ( Figure 4B). Furthermore, treatment with exogenous CTRP3 did not affect the cell number of BMDCs in vitro ( Figure 4C). The expression levels of maturation and activation markers on C1qtnf3 -/-BMDCs were comparable with BMDCs from WT mice, indicating that CTRP3 does not affect BMDC differentiation ( Figure 4D). Next, we stimulated BMDCs with zymosan and LPS, and examined cytokine production. IL-6 and TNF-a production from BMDCs by zymosan stimulation were comparable between WT and C1qtnf3 -/mouse-derived cells ( Figure 4E). Furthermore, exogenous CTRP3 did not affect zymosan-induced IL-6 and TNF-a production in C1qtnf3 -/-BMDCs ( Figure 4F). On the other hand, cytokine productions by LPS stimulation were increased in C1qtnf3 -/-BMDCs ( Figure 4G). Exogenous CTRP3 suppressed LPS-induced IL-6 and TNF-a expression in C1qtnf3 -/-BMDCs ( Figure 4H). This is probably because CTRP3 antagonizes LPS binding to TLR4 by binding to LPS (26). These results suggest that CTRP3 is dispensable for DC maturation and activation.

DISCUSSION
Here, we showed that CTRP3 inhibits Th17 cell differentiation via AdipoR2 in autocrine and paracrine manner. On the other hand, CTRP3 did not affect Th1 cell differentiation and BMDC activation, although these cells also express AdipoR2. Development of EAE was exacerbated in C1qtnf3 -/mice associated with over-expansion of Th17 cells.
Adiponectin receptors AdipoR1 and AdipoR2 share 66.7% amino acid identity (4). These receptors have common and different biological functions. We and other researchers showed that AdipoR1 and AdipoR2 express on Th17 cells (15). Although the effect of adiponectin on Th17 cell differentiation remains controversial (10,11,15), we showed that CTRP3mediated inhibition of Th17 cell differentiation was abolished by AdipoR2 blocker, but not AdipoR1 blocker. Furthermore, AdipoRon regulated Th17 cell differentiation via AdipoR2, but not AdipoR1. These results demonstrate that AdipoR2 has a crucial role in the regulation of Th17 cell differentiation.
On the other hand, Ruiz et al. showed that AdipoR1 and AdipoR2 maintain membrane fluidity in several types of cells, although adiponectin is not required for this biological function of AdipoRs (65). Adiponectin, CTRP6 and CTRP9 share AdipoR1 as a common receptor, and adiponectin and CTRP3 share AdipoR2 (5,6,8,32,66). Thus, distinct CTRPs may exert a common or similar function (3), depending on cell types. Furthermore, Zha et al. reported that   in the presence or absence of CTRP3 (0, 1, 10, 100 ng/ml) for 5 days, and Th17 cell content (left), and IL-17 (center) and IFN-g (right) concentrations in the culture supernatant were determined (n = 4 wells each). Average and SEM are shown. *p < 0.05 and **p < 0.01. Student's t-test. (L) WT or C1qtnf3 -/-2D2 T cells were cultured with WT or C1qtnf3 -/-BMDCs in the presence of 5 mg/ml MOG 35-55 for 5 days (n = 3 each), and IL-17 (left) and IFN-g (right) concentrations in the culture supernatant were determined. Average and SEM are shown. *p < 0.05, **p < 0.01 and ***p < 0.001. Student's t-test. All data were reproduced in another independent experiment with similar results. angiotensin II receptors (AT1 and AT2) form heterodimer with AdipoR1 and AdipoR2 and inhibit the biological action of AdipoRs in renal tubular epithelial cells (67). Thus, AdipoR signalings may also be regulated by antagonizing receptors, making the regulation of the CTRP-AdipoR system very complex. Many apparently inconsistent results are reported related to the CTRP family ligand-receptor relationships and their functions. We showed in this report that CTRP3 inhibits specifically Th17, but not Th1, cell differentiation through AdipoR2-mediated mechanism using AdipoR-specific blocking peptides. We did not observe any effects of AdipoR1 inhibition on CTRP3-regulated Th17 cell differentiation. On the other hand, adiponectin inhibits both Th1 and Th17 cell differentiation via AdipoR1, AdipoR2 and unknown receptor(s)   (69), suggesting an enhancing role of AdipoR1 for the differentiation of Th17 cells. Although this apparently seems contradict with previous reports (11)(12)(13), this is probably because signaling through AdipoR2 is enhanced in the absence of AdipoR1. Actually, we found that AdipoRon, an agonist of AdipoR1 and AdipoR2, inhibits Th17 cell differentiation. The effect of AdipoRon on Th17 cell differentiation was abolished by AdipoR2 blocker, but not AdipoR1 blocker. Similar to Th17 cells and chondrocytes, DCs also express AdipoR1 and AdipoR2. Previously, we reported that T cell recall response was normal in C1qtnf3 -/mice (21), and in this report, we showed that DC functions including proliferation and maturation are normal in CTRP3-deficient DCs. Consistent with our results, DC maturation and phagocytic capacity remain unaffected by the treatment with adiponectin (70). However, in this report, we found that IL-17 production was enhanced in a MOG-specific T cell-C1qtnf3 -/-DC-coculture, suggesting DC function is also suppressed by CTRP3. Consistent with our observations, Tan et al. showed that adiponectin arrests DC function by blocking NF-kB activation via AdipoR1 and AdipoR2 (68). Kupchak et al. also showed that DC differentiation and maturation are enhanced through inhibition of AdipoR1 and AdipoR2 by TNF-a derived from GM-CSF-activated BMDCs (71). However, Jung et al. reported that adiponectin induces functional maturation of DCs independently from AdipoR1 and AdipoR2 and Th1 and Th17 cell differentiation are enhanced in coculture with adiponectin-primed DCs (16). Thus, AdipoR functions are still controversial depending on the cellular and experimental settings. Clearly, further investigation is necessary to elucidate the detailed mechanisms.
In this report, we have demonstrated that CTRP3 ameliorates development of EAE by inhibiting Th17 cell differentiation. We have already shown that CTRP3 ameliorates development of autoimmune arthritis (21). Furthermore, CTRP3 is implicated in the regulation of myocardiac dysfunction, inflammatory bowel disease, severe acute pancreatitis and chronic kidney diseases (21)(22)(23)(24)(25). Because involvement of Th17 cells is suggested in these diseases (72)(73)(74)(75), it is possible that CTRP3 regulates excess inflammation in these diseases. Taken together, these results suggest that the CTRP3-AdipoR2 axis is a good target for the treatment of Th17 cell-mediated autoimmune diseases such as MS and RA.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The animal study was reviewed and approved by the Animal Care and Use Committee of Tokyo University of Science and Kansai Medical University.

AUTHOR CONTRIBUTIONS
MAM and YI designed research study and wrote the manuscript. MAM performed and analyzed most experiments with technical support from MM and TO, and H-HC. H-HC significantly contributed to the revision. All authors contributed to the article and approved the submitted version.