Galectin-9 Ameliorates Clinical Severity of MRL/lpr Lupus-Prone Mice by Inducing Plasma Cell Apoptosis Independently of Tim-3

Galectin-9 ameliorates various murine autoimmune disease models by regulating T cells and macrophages, although it is not known what role it may have in B cells. The present experiment shows that galectin-9 ameliorates a variety of clinical symptoms, such as proteinuria, arthritis, and hematocrit in MRL/lpr lupus-prone mice. As previously reported, galectin-9 reduces the frequency of Th1, Th17, and activated CD8+ T cells. Although anti-dsDNA antibody was increased in MRL/lpr lupus-prone mice, galectin-9 suppressed anti-dsDNA antibody production, at least partly, by decreasing the number of plasma cells. Galectin-9 seemed to decrease the number of plasma cells by inducing plasma cell apoptosis, and not by suppressing BAFF production. Although about 20% of CD19−/low CD138+ plasma cells expressed Tim-3 in MRL/lpr lupus-prone mice, Tim-3 may not be directly involved in the galectin-9-induced apoptosis, because anti-Tim-3 blocking antibody did not block galectin-9-induced apoptosis. This is the first report of plasma cell apoptosis being induced by galectin-9. Collectively, it is likely that galectin-9 attenuates the clinical severity of MRL lupus-prone mice by regulating T cell function and inducing plasma cell apoptosis.


Introduction
Systemic lupus erythematosus (SLE) is a systemic autoimmune disease characterized by autoantibody production against selfantigens. Among SLE complications, lupus nephritis is the most serious and a major predictor of poor prognosis [1]. Until recently, glucocorticoids, aspirin and antimalarials were approved for treatment of SLE. B-cell stimulatory factors promote the loss of B-cell tolerance and drive autoantibody production. B cell activation mediated by B-cell activator factor belonging to the TNF family (BAFF) and a proliferation-inducing ligand (APRIL) have been implicated in SLE pathogenesis [2,3,4]. This suggests that B cell regulation, in addition to T cell regulation, is required for SLE treatment [2].
Gal-9 is a b-galactoside binding lectin that exhibits therapeutic effects in autoimmune disease models, such as autoimmune arthritis, experimental allergic encephalomyelitis, and Type 1 diabetes mellitus [5,6,7]. Such therapeutic effects of Gal-9 seem to be ascribed to the decrease of Th1 and Th17 effector cells expressing Tim-3 [8]. It has also been found that the decrease of Th1 and Th17 effector cells is likely induced by programmed cell death of effector cells through a Gal-9/Tim-3 interaction [8]. In contrast, Gal-9 expands Foxp3 + regulatory T cells (Tregs) in vivo and in vitro [5]. Furthermore, Gal-9 ameliorates immune complex (IC)-induced inflammation by suppressing IC-induced macrophage activation and C5a generation [9]. Collectively, Gal-9 seems to regulate a variety of immune cells to ameliorate autoimmune inflammation. Nevertheless, little is known about the effects of Gal-9 on B cell autoantibody production, although it is clear that B cells and B cell-derived autoantibody are associated with the pathogenesis of autoimmune disorders.
The purpose of the present study is to test whether Gal-9 ameliorates lupus signs and suppresses anti-dsDNA antibody production by inducing plasma cell apoptosis.

Materials and Methods
Mice MRL/lpr lupus-prone and MRL/lpr +/+ mice were purchased from Japan SLC (Shizuoka, Japan). All mice were housed in plastic boxes in groups of 3 to 4 under a 12:12 light cycle with food and water provided ad libitum.
The study protocol was approved by the Animal Care and Use Committee of Kagawa University, and mice used in this research received humane care to minimize suffering in accordance with international and national guidelines of humane laboratory animal care. Mice were sacrificed by CO 2 narcosis unless otherwise specified.

Experimental Protocol
All Gal-9 preparations used in the present experiment were .95% pure by SDS-PAGE with less than 0.001 endotoxin units/ mg, as assessed by a limulus turbimetric kinetic assay using a Toxinometer ET-2000 (Wako, Osaka, Japan). Nine-week-old MRL/lpr lupus-prone mice were injected intraperitoneally with human stable Gal-9 with no linker peptide (30 mg/mouse, 3times/week) or PBS as a control, to assess the therapeutic effects of Gal-9. Proteinuria, paw volume, and hematocrit were monitored until mice were 20 weeks of age.
Eight-week-old mice were treated with Gal-9 for 4 weeks to assess the effects of Gal-9 on the level of anti-dsDNA antibody and the frequency of splenic T and B cell subpopulations.

Laboratory Methods
Proteinuria was measured using the BCA Protein Assay Reagent Kit (Takara Bio Inc., Otsu, Japan). Clinical signs of arthritis (i.e., paw swelling) were monitored during the course of disease by water displacement plethysmometry. Paw swelling was expressed as increased paw volume. Hematocrit values were collected from the tail vein (70 ml) in 1 mm heparinized tubes. The tubes were spun and hematocrit was determined using a Hawksley Micro-haematocrit reader (Lancing, Sussex, UK).

Apoptosis
Plasma cells were purified from spleen in MRL/lpr lupus-prone mice using MACS CD138 + Plasma Cell Isolation Kit (Miltenyi Biotec) as recommended by the manufacturer. The cell population contained .98% CD138 + cells. The isolated plasma cells were used for apoptosis assay. The single cell suspensions were incubated for 5 h with 30 nM Gal-9 in 96 well flat-bottom plates in humidified incubators in the presence of 5% CO2. After the incubation period was over, the cells were stained for Annexin V (BioLegend) with or without 7AAD and analyzed immediately by flow cytometry.

Statistical Analysis
Student's paired or unpaired t-tests and one-or two-way ANOVA were used for statistical comparisons. All statistical analyses were performed with Prism 5 software (Graphpad Software, La Jolla, CA).

Galectin-9 Ameliorates Clinical Severity of MRL/lpr Lupusprone Mice
We first evaluated the therapeutic effects of Gal-9 by treating 9week-old MRL/lpr lupus-prone mice with Gal-9. Gal-9 significantly suppressed the level of proteinuria ( Figure 1A). Since it is well known that MRL/lpr lupus-prone mice develop autoimmune arthritis and hemolytic anemia [11,12], we also examined the effects of Gal-9 on these diseases. Gal-9 clearly retarded the onset of arthritis and suppressed the increase of paw volume ( Figure 1B). Hematocrits of 20-week-old MRL/lpr lupus-prone mice were significantly lower than that of MRL/lpr +/+ mice. When MRL/lpr lupus-prone mice were treated with Gal-9, hematocrit increased significantly ( Figure 1C). These results suggest that Gal-9 helps to ameliorate disease onset in MRL/lpr lupus-prone mice.
We assessed ELISA assay to compare the level of IL-17 in the culture supernatants of spleen CD4 T cells from PBS-and Gal-9treated mice with or without PMA+ionomycin stimulation. Surprisingly, the levels of IL-17 in the culture supernatants were negligible even after the stimulation, and there was no significant difference in IL-17 levels between PBS-and Gal-9-treated MRL/ lpr lupus-prone mice, suggesting that release of IL-17 was impaired in most CD4 T cells of MRL/lpr lupus-prone mice ( Figure 3B). We also assessed ELISA assay for IL-17 levels in plasma of PBS-treated and Gal-9-treated MRL/lpr lupus prone mice. Expectedly, ELISA assay revealed that plasma IL-17 level was negligible in both PBS-and Gal-9-treated MRL/lpr lupusprone mice ( Figure 3C).

Gal-9 Suppresses Anti-dsDNA Antibody Production
We hypothesized that Gal-9 suppresses autoantibody production in MRL/lpr lupus-prone mice, because Gal-9 improved hematocrit of MRL/lpr lupus-prone mice. MRL/lpr lupus-prone mice (8-week-old) were treated with Gal-9, since their anti-dsDNA antibody levels began to increase. Gal-9 treatment significantly suppressed anti-dsDNA antibody production in MRL/lpr lupusprone mice ( Figure 5A). Moreover, ANOVA analysis confirmed that the level of anti-dsDNA antibody in Gal-9-treated MRL/lpr lupus-prone mice did not increase, whereas the level in PBStreated mice significantly increased ( Figure 5B). Next, we performed experiments to ask whether Gal-9 also suppresses the levels of total IgG. In contrast, there was no significant difference in the levels of total IgG between PBS-and Gal-9-treated MRL/ lpr lupus-prone mice ( Figure 5C and D). Gal-9 Decreases the Frequency of CD19 2/low CD138 + Plasma Cells The above results raised the possibility that Gal-9 suppresses anti-dsDNA antibody production through B cell regulation. FACS analysis revealed that the frequency of CD19 + B cells was increased in 12-week-old Gal-9-treated MRL/lpr lupus-prone mice ( Figure 6A). In contrast, Gal-9 significantly reduced the frequency of splenic plasma cells (CD19 2/low CD138 + ) but not plasmablasts (CD19 + CD138 + ) ( Figure 6A). It was thus suggested Gal-9 preferentially decreases plasma cells in MRL/lpr lupusprone mice.
Since BAFF levels are up regulated in SLE patients and MRL/ lpr lupus-prone mice to induce plasma cell differentiation [18,19,20,21], we assessed the effects of Gal-9 on serum level of BAFF. ELISA analysis unexpectedly revealed that Gal-9 does not reduce BAFF level ( Figure 6B). These results suggest that Gal-9 does not decrease plasma cells by downregulating BAFF production, at least in this model.

Gal-9 Induces Plasma Cell Apoptosis
Because Gal-9 induces apoptosis of Th1, Th17, and CD8 T cells through Gal-9/Tim-3 interaction, we first asked whether plasma cells express Tim-3, and intriguingly found that about 20% of plasma cells (CD19 2 CD138 + ) in MRL/lpr lupus-prone mice express Tim-3 ( Figure. 7A). In order to ask whether Gal-9 induces plasma cell apoptosis, CD19 2 CD138 + cells were prepared by MACS. Gal-9 significantly increased the frequency of Annexin V + apoptotic plasma cells ( Figure 7B). The apoptosis was significantly suppressed by lactose, whereas anti-Tim-3 antibody unexpectedly did not ( Figure 7B). Further experiments revealed that Gal-9 induced both early apoptosis (Annexin V + 7AAD 2 ) and late apoptosis (Annexin V + 7AAD + ) of plasma cells ( Figure 7C). Further experiments, however, revealed that Gal-9 induced Tim-3 + plasma cell apoptosis than Tim-3 2 plasma cells ( Figure 7D). Taken together, Tim-3 may not be directly involved in the apoptosis, and Tim-3 may be associated with vulnerability to Gal-9-mediated plasma cell apoptosis, at least, in MRL/lpr lupus-prone mice.

Discussion
We found that in MRL/lpr lupus-prone mice, Gal-9 attenuates the severity of various symptoms, such as lupus nephritis, arthritis, and hemolytic anemia. It has been shown that Gal-9 induces apoptosis of Th1 and Th17 cells through Gal-9/Tim-3 interactions [8]. Our recent studies revealed that Gal-9 downregulates Th17 cell differentiation whereas it upregulates differentiation of Foxp3 + Tregs, independently of Tim-3 [5]. The beneficial effects of Gal-9 on lupus symptoms in MRL/lpr lupus-prone mice seem partially ascribed to Gal-9-induced decrease of Tim-3 + Th1 and Th17 cells because imbalance of Th17 and Th1 cells in SLE and Th17 and Tregs are critical for SLE pathogenesis. Interestingly, Gal-9 failed to expand Foxp3 + Tregs in MRL/lpr lupus-prone mice, suggesting that this is attributed to T cell abnormality in MRL/lpr lupus-prone mice [24]. In the present experiments, we also found that CD4 + T cells from MRL/lpr lupus-prone mice did not release IL-17 even by PMA stimulation though most of them surprisingly expressed IL-17 in the cytoplasm, suggesting T cell abnormality in the mice. Oppositely, Hou et al recently reported that only less than 1% of cells in CD4 T cells of MRL/lpr mice expressed IL-17 in the cytoplasm [25]. Although heterogeneity in MRL/lpr mice according to the source may be one explanation for the above discrepancy, further studies are, of course, required to ascertain it.
In the present experiments, we show that Gal-9 suppresses anti-dsDNA antibody levels in MRL/lpr lupus-prone mice though the level of total IgG was not changed by Gal-9. Although Gal-9 treatment increased CD19 + cells in MRL/lpr lupus-prone mice, it reduced CD19 2 CD138 + plasma cells but not CD19 + CD138 + plasmablasts. This plasma cell reduction may result in the suppression of anti-dsDNA antibody production.
Recently it was shown that Belimumab, a specific inhibitor of BAFF, and atacicept (TACI-immunoglobulin), a receptor molecule for APRIL, effectively ameliorates clinical symptoms in SLE patients [30,31,32]. Although family receptors for BAFF and APRIL vary in their expression patterns and levels across different B-cell subsets, biologic action of BAFF and APRIL may be primarily on memory and/or plasma cells [33,34]. Moreover, the innate immune system initiates and perpetuates autoimmunity [35]. A small number of patients, undergoing type I IFN therapy for cancer or viral infections, developed SLE. Similarly, IFNa accelerates SLE in some murine models and is associated with increased BAFF serum levels. Blockade of TNF-a induces increased levels of BAFF via upregulation of type I IFNs and has been associated with development of anti-nuclear antibodies in up to 50% of patients with clinical SLE [37]. B cell depletion in patients treated with anti-CD20 also results in high levels of BAFF, likely an attempt to maintain B cell homeostasis [38].
However, Gal-9 treatment fails to reduce BAFF levels, although it attenuates disease severity. Instead, Gal-9 induces plasma cell apoptosis, suggesting Gal-9 induced plasma cell apoptosis is, at least, partly involved in the suppression of anti-dsDNA antibody production. Of course, It can be raised an alternative explanation that Gal-9 block maturation of CD19 + B cells to CD19 2 plasma cells, because of increased CD19 + B cells and decreased plasma cells in MRL/lpr mice. Furthermore, we cannot, however, exclude the possibility of involvement of Gal-9 in the biological function of BAFF. It remains to be clarified whether Gal-9 inhibits binding between BAFF and BAFF receptors, Gal-9 downregulates BAFF receptors on B cells, and suppresses BAFF/BAFF receptorinduced signal transduction. Furthermore, plasma cells in the spleen of MRL/lpr lupusprone mice express Tim-3, and the Tim-3 + plasma cells are more susceptible to Gal-9-induced apoptosis than the Tim-3 2 plasma cells. Curiously, a blocking Tim-3 antibody does not abrogate the Gal-9-induced apoptosis. These observations raise several questions. Firstly, as far as we know, this is the first example of Tim-3 expression in the cells of B cell linage. It is, thus, urgently required to ascertain whether Tim-3 expression on plasma cells is limited in MRL/lpr lupus-prone mice or plasma cells in general, including in WT mice and in humans. The second question is whether there are any functional differences between Tim-3 + and Tim-3 2 plasma cells. Tim-3 expression is associated with exhausted phenotypes in virus-infected T-cells and those T-cells are also known to be susceptible to Gal-9-induced apoptosis [40]. The third question is how Tim-3 expression renders plasma cells susceptible to Gal-9, even though Tim-3 may not be the direct target molecule for Gal-9-induced apoptosis. Last, it will be intriguing to clarify whether Gal-9 is also involved in the regulation of B cell development other than plasma cells, including germinal center B cells as they regulate Th17 and Foxp3+ regulatory T cells [7]. Further in-depth studies will be, thus, required to answer the above questions to understand the exact molecular mechanisms of Gal-9-induced regulation of antibody production.
In conclusion, Gal-9 ameliorates clinical severity of MRL/lpr lupus-prone mice by decreasing the level of anti-dsDNA antibody in addition to T cell regulation. The suppressive mechanism was supposed to be the result of induction of apoptosis of plasma cells but not of decrease of BAFF production. Although some CD19 2/ low CD138 + plasma cells in MRL/lpr lupus-prone mice express Tim-3, Gal-9 induced plasma cell apoptosis independently of Tim-3. This suggests that Gal-9 may be a novel candidate for SLE therapy, which would involve an orchestrated mode of action on T cells [41,42], macrophages [43,44] and plasma cells.