Schistosome infection promotes osteoclast-mediated bone loss

Infection with schistosome results in immunological changes that might influence the skeletal system by inducing immunological states affecting bone metabolism. We investigated the relationships between chronic schistosome infection and bone metabolism by using a mouse model of chronic schistosomiasis, affecting millions of humans worldwide. Results showed that schistosome infection resulted in aberrant osteoclast-mediated bone loss, which was accompanied with an increased level of receptor activator of nuclear factor-κB (NF-κB) Ligand (RANKL) and decreased level of osteoprotegerin (OPG). The blockade of RANKL by the anti-RANKL antibody could prevent bone loss in the context of schistosome infection. Meanwhile, both B cells and CD4+ T cells, particularly follicular helper T (Tfh) cell subset, were the important cellular sources of RANKL during schistosome infection. These results highlight the risk of bone loss in schistosome-infected patients and the potential benefit of coupling bone therapy with anti-schistosome treatment.

Introduction Schistosomiasis is a chronic helminthic disease that affects more than 200 million people worldwide, with another 700 million considered at risk of infection tab [1,2]. Although the primary cause of pathogenesis of schistosomiasis is the granulomatous inflammatory responses in the liver and intestines [3,4], there is evidence that schistosomiasis patients experience long-term hidden pathologies that have remained poorly characterized. For instance, physical growth impairment in children is linked to parasitic infectious diseases, regardless of nutritional status [5,6]. Indeed, schistosome infection has been associated with growth inhibition in children, even with low parasitic burdens [7], indicating that schistosome infection may influence the skeletal system. Despite the importance of bone tissue in health and development, little is known about the underlying mechanisms of how schistosome infection disrupts bone homeostasis.
Schistosome infection elicits the robust multi-type immune responses beyond the Th2-type response [8], including B-cell response, regulatory T (Treg) cell response, Th17-cell response, and T follicular helper (Tfh) cell response [9][10][11]. Strikingly, substantial evidence support that schistosome-induced immunity has important impacts on the progression and outcome of multiple other diseases, such as type 1 diabetes and experimental allergic encephalomyelitis [12][13][14]. Bone is a dynamic tissue that is constantly remodeled by bone-resorbing osteoclasts (OCs) and boneforming osteoblasts (OBs) in a tightly regulated process [15]. The differentiation of OC precursors into OCs is under the influence of the key osteoclastogenic cytokine receptor activator of nuclear factor-κB ligand (RANKL), which is moderated by osteoprotegerin (OPG), a RANKL's physiological decoy receptor [16]. In fact, an increase in the ratio of RANKL to OPG accelerates the rate of osteoclastic bone resorption in various infectious diseases such as HIV infection and bacterial infection-induced periodontitis [17,18]. Many cell types are capable of producing RANKL, but T cells and B cells are recognized as predominant sources of RANKL in some infectious diseases [18,19]. Given the pivotal roles of T cells and B cells in orchestrating bone metabolism [20,21], it raises the possibility that schistosome infection might influence the skeletal system by altering the host's immune responses, particularly T-cell and/or B-cell responses.
Here, we found that the increased RANKL was associated with schistosome infectioninduced bone loss, and showed that both B cells and CD4 + T cells, particularly Tfh cell subset, were the important contributors to an increased level of RANKL during schistosome infection. Thus, schistosome infection leads to bone loss that reflects the immunological consequences of infection in the host, suggesting that patients with schistosomiasis should pay more attention to bone problems caused by schistosome infection.

Schistosome infection results in bone loss in mice
First, we investigated whether schistosome infection had an impact on bone homeostasis. Xradiographic analysis of the femurs showed that mice in the chronic phase of the infection (since 11-13 weeks post-infection) had severe osteoporosis accompanied by a marked decrease in trabecular bone mineral density (Fig 1A). Micro-computed tomography (μCT) analysis of distal femurs showed that schistosome-infected mice had significantly reduced trabecular bone volume and number (Fig 1B). Bone morphometric analysis of distal femurs by μCT showed that trabecular and cortical bone mineral density (BMD, Fig 1C), bone volume (BV), bone volume fraction (BV/TV), trabecular thickness (Tb.Th), number (Tb.N.), and connectivity density (Conn.D., Fig 1D), as well as cortical bone thickness (Ct.Th), area (Ct.Ar), and cortical bone fraction (Ct.Ar/Tt.Ar, Fig 1E) were significantly decreased in schistosome-infected mice. However, no significant difference in tissue volume or total cross-sectional area (TV or Tt.Ar) or trabecular space (Tb.Sp) was detected, while bone architecture was dramatically altered reflected by an elevated structure model index (SMI) in schistosome-infected mice ( Fig  1D and 1E). Indeed, schistosome infection-induced bone loss was also observed in other mouse strains such as BALB/c and ICR (S1 Fig) and patients with schistosomiasis (S1 Table and S1 Text). These results demonstrated that schistosome infection leads to bone loss.

Aberrant bone resorption causes bone loss in S.japonicum-infected mice
S.japonicum-infected mice displayed higher bone resorption compared to normal mice, as demonstrated by higher serum levels of the bone resorption marker CTx (C-terminal telopeptide of collagen; Fig 2A), comparable levels of the bone formation marker osteocalcin (Fig 2A), more osteoclasts, and larger osteoclast-covered surface (Fig 2B and 2C), as well as diminished trabecular structure (Fig 2D and 2E). These data suggested that S.japonicum-infected mice have reduced bone mass as a consequence of increased osteoclastic bone resorption.

The increased RANKL is associated with schistosome infection-induced bone loss in vivo
Our aforementioned results showed the increases in bone resorption and osteoclast-covered surface in S.japonicum-infected mice relative to normal mice (Fig 2A-2C). We next wondered whether the aberrant osteoclast generation was involved in the increased bone resorption during schistosome infection. Thus, we firstly analyzed the absolute number of OCPs (CD3 -B220 -NK1.1 -CD11b -/low CD115 + CD117 + cells) [22] in bone marrow (BM) of normal and infected mice. Results in Fig 3A showed a comparable number of OCPs in normal and S.japonicum-infected mice. Next, we evaluated the levels of osteoclastogenic cytokine RANKL in BM, and found a significant increase in RANKL level in S.japonicum-infected mice ( Fig 3B). Furthermore, the percentages of RANKL + cells were substantially increased in bone marrow (S2, S3A and S3B Figs), mesenteric lymph nodes (Fig 3C), and spleens ( Fig 3D) during infection. Notably, these effects were also associated with lower OPG expression, resulting in a higher ratio of RANKL to OPG expression in BM ( Fig 3B). X-radiographic analysis of the femurs showed an increased trabecular bone mineral density in S. japonicum-infected mice treated with anti-RANKL antibody, compared with that in isotype treated-infected mice ( Fig  3E). Furthermore, μCT analysis showed that BV, BV/TV (Fig 3F and 3G Overall, these results indicated that RANKL blockade ameliorated schistosome infectioninduced bone loss in mice. mesenteric lymph nodes and spleens during infection. We found that both CD4 + T cells and B cells were shown to be major in vivo sources of RANKL during schistosome infection, and accounted for more than 90% of total RANKL + cells (Fig 4A and 4B). We next analyzed the relative contribution of CD4 + T cells and B cells to the increased RANKL + cells in schistosomeinfected mice. Results showed that both CD4 + T cells and B cells were likely to be the major contributors to the increased RANKL + cells during schistosome infection (Fig 4B-4J). A similar increased frequency of RANKL + lymphocytes was observed in BM in schistosome-infected mice, most of which are constituted by T cells (S3C Fig). Our results, in conjunction with our observation that there was an evaluated level of osteoclastogenic cytokine RANKL in BM, suggest that CD4 + T cells are likely to be an important contributor to RANKL-mediated bone loss in schistosome-infected mice through releasing soluble RANKL.

Tfh cells are an important cellular source of RANKL during schistosome infection
CD4 + T cells can differentiate into several subsets including Th1, Th2, Th17, Tfh, and Treg cells in humans and mice during schistosome infection [9]. To investigate the exact cellular mechanisms responsible for schistosome infection-induced bone loss, we analyzed the distribution of RANKL + CD4 + T cells in mice 8 and 13 weeks after infection, and found that Tfh cells were a major cellular source of RANKL during schistosome infection (Fig 5A and 5B). More importantly, the percentage of RANKL + Tfh cells within CD4 + T cells had been increased by approximately 5-fold in mice since 8 weeks after infection (Fig 5C), suggesting that Tfh cells were an important contributor to the increased RANKL + cells in vivo, and involved in osteoporosis-associated bone loss during schistosome infection.
Next, we used ICOSL knockout (KO) mice as Tfh cell deficiency model ( Fig 5D) [23,24] to address the exact contribution of Tfh cells in the induction of bone loss in S. japonicum-infected mice. As expected, Tfh cell deficiency led to a significantly reduced frequency of RANKL-producing cells in CD4 + T cells (Fig 5E), which nicely supported our previous observation that Tfh cells were a major subset of CD4 + T cells producing RANKL during schistosome infection. While the reduction of Tfh cells was only slightly decreased the percentages of total RANKL-producing cells in total lymphocytes ( Fig 5F). Indeed, X-radiographic and μCT analysis of the femurs showed that Tfh cell deficiency alone was difficult to reverse schistosome infection-induced bone loss in vivo (Figs 5G-5I and S5), which may suggest a possible compensation from multiple sources of RANKL. Collectively, our results indicate that Tfh cells play an important role of contributor, which may be sufficient but not necessary, in schistosome infection-induced bone loss.

Discussion
The long-term hidden pathological consequences of schistosome infection are poorly defined. Schistosome infection could affect the outcome of many other diseases by altering their host's immunological states [12][13][14]. In particular, the possible relationship between growth stunting and schistosome infection in humans suggests that schistosome infection may has a negative effect on bone homeostasis [7]. However, there is limited knowledge on how schistosome infection affects bone homeostasis. Given that skeletal homeostasis is dynamically influenced   [20,21,25], it will be of interest to examine the role of schistosomeinduced immune responses in bone metabolism in the host. In this study, we showed that the increased RANKL, mainly produced by T cells and B cells, was associated with schistosomeinduced bone loss, suggesting that schistosome-infected patients are at higher risk for bone loss and potentially benefit from coupling bone therapy with anti-schistosome treatment.
Accumulating evidence has revealed that some infectious diseases caused by bacteria, viruses, or protozoan parasites, such as Staphylococcus aureus, human immunodeficiency virus, or plasmodium, can lead to aberrant bone remodeling that is commonly associated with bone diseases such as osteomyelitis, osteoporosis, or even pathologic fractures [18,[26][27][28][29]. Although some papers showed that S. japonicum or its products suppressed bone resorption in the context of collagen-induced arthritis (CIA), actual schistosome infection exacerbated the severity of arthritis when established CIA mice were infected with schistosome [30][31][32]. Interestingly, schistosome infection before CIA induction reduced the severity of CIA [31]. Here, we found that actual schistosome infection results in osteoclastmediated bone loss in mice. Bone homeostasis is maintained by the dynamic balance between osteoclastic bone resorption and osteoblastic bone formation [33]. Because no significant difference in the progress of osteoblast-mediated bone formation was observed between normal and infected mice, we reasoned that the osteoclast-mediated bone resorption was likely to be a major contributor to schistosome infection-induced bone loss. Although some studies indicated that IL-4, IL-10, IL-13, and IFN-γ produced by activated T cells in response to infection likely disrupts OC signaling, the RANK-RANKL signaling pathway is central in the differentiation of OCs [34,35]. A key finding of our study is that schistosome-induced bone loss is associated with the alterations in the RANKL/OPG ratio, and ameliorated by blocking the RANKL-RANK pathway. Indeed, the rates of osteoporosis in schistosomiasis patients were considerably higher than those in healthy controls in the non-endemic area. Thus, these results suggest that patients with schistosomiasis, affecting more than 200 million people worldwide [1,2], might potentially benefit from coupling bone therapy with anti-schistosome treatment.
In homeostatic conditions, osteocytes are major sources of RANKL production, contributing to bone remodeling as demonstrated by previous reports [36]. However, in pathological conditions such as rheumatoid arthritis and HIV infection, other cells including T cells and B cells are likely to be major contributors to increased pathogenic RANKL production [37,38]. Schistosomes are parasitic trematodes living in the inferior mesenteric veins that induce marked B cell and CD4 + T cell responses in the peripheral lymphoid tissues such as the spleen and MLN [4]. Thus, it is reasonable to hypothesize that schistosome infection results in bone loss through triggering B cell and/or T cell responses in the spleen and MLN, which induce the release of the soluble RANKL and/or home to bone marrow. Although we can not rule out the effects of other cells on RANKL-mediated bone loss in this study, our results indicated that both B cells and CD4 + T cells in the MLN and spleen are likely to be important contributors to RANKL-associated bone loss through releasing soluble RANKL or homing into bone marrow, in which antigen-experienced T and B cells return to specialized niches for the formation of long-lived memory cells [21]. Consistent with this notion, we observed both elevated levels of soluble RANKL and increased frequencies of RANKL + T cells in bone marrow in schistosomeinfected mice. In addition, several previous published literatures also show that RANKL-producing T cells or B cells in the peripheral lymphoid tissues were the dominant drivers of osteoclast-associated bone loss under certain pathological conditions [18,[39][40][41].

cytometry data plots and statistics show the distribution of B cells (red), CD4 + T (blue), CD4 -T (green), and the other cells (black) within total RANKL + cells in infected and normal mice. Areas represent the means of percentages of B cells, CD4 + T, CD4 -T, and the other cells within total RANKL + cells. Gated to RANKL + cells; (C-F) Flow cytometry data plots and statistics show the frequencies and absolute numbers of RANKL + cells in mesenteric lymph nodes (C-D) and spleens (E-F), Gated to CD3 + cells; The percentages of RANKL + cells of CD4 + T cells in the graphs were calculated by RANKL + CD4 + / (RANKL + CD4 + + RANKL -CD4 + ) in the plots. (G-J) Flow cytometry data plots and statistics show the frequencies and absolute numbers of RANKL + cells with CD19 + cells mesenteric
Based on the phenotype, cytokine profiles, and functionality, CD4 + T cells are divided into distinctive subsets including type 1 helper T (Th1) cells, Th2 cells, Th17 cells, follicular helper T (Tfh) cells, and regulatory T (Treg) cells [42]. Recent in vitro experiments show that Th17 cells can secrete RANKL, and are viewed as an osteoclastogenic helper T cell subset [40]. Consistent with the published literature, our results showed that most Th17 cells produced RANKL in S.japonicum-infected mice. Accumulating evidence showed that IL-17 produced by Th17 cells has a role in promoting osteocytes to secrete RANKL [40]. Given that the number of Th17 cells was relatively low in schistosome-infected mice, it is plausible that IL-17-mediated RANKL production may not be a dominant contributor to increased RANKL in schistosome-infected mice. However, the exact contribution of IL-17 to RANKL production during schistosome infection remains to be further investigated. Unexpectedly, here we revealed that Tfh cells were a major in vivo source of RANKL. Although Tfh cells deficiency alone reduced significantly the frequency of RANKL-producing cells in CD4 + T cells, it only decreased slightly the percentages of total RANKL-producing cells in total lymphocytes, which suggests that Tfh cells were identified as a predominant subset of RANKL-producing CD4 + T cells, however, other RANKL-producing cells could serve a complementary role.
In conclusion, we demonstrated that the increased RANKL was correlated with the host's osteoclast-mediated bone loss in the context of schistosome infection, and that both B cells and CD4 + T cells, particularly Tfh cells, a dominant subset of RANKL-producing CD4 + T cells, are involved in schistosome infection-induced bone loss in vivo. Thus, schistosome infection may be a substantial contributing factor to exacerbated bone loss by altering the host's immune homeostasis. These results highlight the risk of bone loss in schistosomiasis patients and the potential benefit of coupling bone therapy with anti-schistosome treatment.

Ethics statement
Animal experiments were performed in strict accordance with the Regulations for the Administration of Affairs Concerning Experimental Animals (1988.11.1), and all efforts were made to

Mice, infection and anti-RANKL blocking antibody treatments
Eight-week-old male C57BL/6J, BALB/c or ICR mice were purchased from the SLAC Laboratory (Shanghai, China). ICOSL -/-C57BL/6J mice were purchased from the Jackson Laboratory (Bar Harbor, ME). Mice were infected percutaneously with 12 S. japonicum cercariae (Chinese mainland strain) obtained from infected Oncomelania hupensis snails purchased from the Jiangsu Institute of Parasitic Diseases (Wuxi, China). Three, five, eight, and thirteen weeks post-infection (a time point of chronic infection), experiments were performed. In certain experiments, mice 5 weeks after S. japonicum infection were administered intraperitoneally with 300 μg/mouse of anti-mouse RANKL blocking antibody (Clone IK22-5; BioXcell, West Lebanon, NH) or isotype control (Clone 2A3; BioXcell) three times per week for 8 weeks.

X-ray and μCT analysis
The femurs from normal and infected mice were analyzed by a MX-20 X-ray equipment (Faxitron, Tucson, AZ) or a Skyscan 1176 μCT scanner (Bruker, Bremen, Germany). For microcomputed tomography (μCT) analysis, the femurs were removed and kept in 10% neutralbuffered formalin. The parameters of the scan in the 1176 μCT scanner are as follows: Scan duration (6-7 minutes), X-ray voltage (50 kV), X-ray current (455 μA), number of slices (329), Filter (0.5 mm aluminium), Image pixel size (17.76 μm), Camera pixel size (12.59μm), Rotation step (0.6), and Frame averaging (1). Two-dimensional images were used to generate three-dimensional renderings using the CT-Analyser (CTAn; Bruker Micro-CT Software). Both trabecular (metaphyseal) and cortical (metaphyseal-diaphyseal) were selected with reference to the growth plate. The trabecular region commenced about 0.3552 mm (20 image slices) from the growth plate level in the direction of the metaphysis, and extended from this position for a further 1.4208 mm (100 image slices). The cortical region commenced about 2.4864 mm (140 image slices) from the growth plate level in the direction of the metaphysis, and extended from this position for a further 0.5328 mm (170 image slices). According to the published guidelines for the assessment of bone microstructure in rodents using μCT [43], the following indices were calculated with the software which includes trabecular mineral density (BMD), tissue volume (TV), bone volume (BV), bone volume fraction (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N), trabecular space (Tb.Sp), structure model index (SMI, an indicator of the trabecular structures), and connectivity density (Conn.D.), and cortical mineral density (BMD), cortical thickness (Ct.Th), total cross-sectional area (Tt.Ar), cortical bone area (Ct.Ar), and cortical bone fraction (Ct.Ar/Tt.Ar).

Immunohistochemical analysis
Bones were fixed in 4% formaldehyde and decalcified in 10% EDTA, and 5-μm-thick-paraffinembedded sections were obtained and stained for TRAP and hematoxylin and eosin (H&E). All images were captured using an Axiovert 200M microscope (Carl Zeiss GmbH, Jena, Germany). Static histomorphometric parameters were measured in sections stained for TRAP using Image-Pro Plus software. The following primary parameters were determined in five fields of view (×200) in one section per mouse: OC surface area and bone surface area. The values obtained from these 5 fields were then averaged and used as the value for that particular animal. From the primary data, the OC surface per bone surface area (Oc.S/BS) was calculated.

Preparation of BM plasma
BM plasma (BMp) was prepared as described previously with some modifications [44]. Briefly, BMp was prepared by flushing femurs and tibias from one or four mice with 250μL RPMI 1640 containing 0.5% BSA and 10 mM Hepes for OPG or RANKL detection, respectively. Suspended cells were transferred into Eppendorf-type centrifuge tubes, and then cells and debris were removed by repeated (2×) centrifugation at 5,000 g for 10 min.

Statistics
Data were analyzed by unpaired t-test for two-group comparison and ANOVA test for three or more-group comparison. Before analysis, the normality assumption was examined. P-values of less than 0.05 were considered as statistically significant. To account for the type I error inflation due to multiple comparisons, we applied the Bonferroni correction. The analyses were performed using SPSS 11.0 software (IBM), and data are presented as mean ± SD.  Table. The rates of bone loss and osteoporosis in schistosomiasis patients and healthy humans. Schistosomiasis patients (n = 238) from a village in Chizhou City, Anhui province, China. The bone mineral density was detected in schistosomiasis patients by the heel ultrasound. The rates of osteoporosis from the healthy controls in Zhejiang province, a previous schistosomiasis endemic area but announced the elimination of this disease in 1992, were published in Chinese academic journal [45]. The rates of osteoporosis or bone loss between schistosomiasis patients and healthy humans were analyzed by Pearson Chi-Square Test. (DOCX) S1 Text. Supporting text. This file contains detailed materials and methods for S1 Table. (DOC)