Targeting hepatic oxidative stress rescues bone loss in liver fibrosis

Objective Chronic liver diseases often involve metabolic damage to the skeletal system. The underlying mechanism of bone loss in chronic liver diseases remains unclear, and appropriate therapeutic options, except for orthotopic liver transplantation, have proved insufficient for these patients. This study aimed to investigate the efficacy and mechanism of transplantation of immature hepatocyte-like cells converted from stem cells from human exfoliated deciduous teeth (SHED-Heps) in bone loss of chronic liver fibrosis. Methods Mice that were chronically treated with CCl4 received SHED-Heps, and trabecular bone density, reactive oxygen species (ROS), and osteoclast activity were subsequently analyzed in vivo and in vitro. The effects of stanniocalcin 1 (STC1) knockdown in SHED-Heps were also evaluated in chronically CCl4 treated mice. Results SHED-Hep transplantation (SHED-HepTx) improved trabecular bone loss and liver fibrosis in chronic CCl4-treated mice. SHED-HepTx reduced hepatic ROS production and interleukin 17 (Il-17) expression under chronic CCl4 damage. SHED-HepTx reduced the expression of both Il-17 and tumor necrosis factor receptor superfamily 11A (Tnfrsf11a) and ameliorated the imbalance of osteoclast and osteoblast activities in the bone marrow of CCl4-treated mice. Functional knockdown of STC1 in SHED-Heps attenuated the benefit of SHED-HepTx including anti-bone loss effect by suppressing osteoclast differentiation through TNFSF11–TNFRSF11A signaling and enhancing osteoblast differentiation in the bone marrow, as well as anti-fibrotic and anti-ROS effects in the CCl4-injured livers. Conclusions These findings suggest that targeting hepatic ROS provides a novel approach to treat bone loss resulting from chronic liver diseases.


INTRODUCTION
The liver is a central organ possessing complex metabolic and xenobiotic functions in the digestive system, and it also participates in the endocrine system. Liver metabolism is highly involved in bone metabolism under physiological conditions through the function of somatotropic axis hormones such as growth hormone, insulin-like growth factor-I, and insulin-like growth factor binding protein 3 and of calciotropic hormones, including parathyroid hormone and vitamin D. Chronic liver diseases can potentially cause abnormal metabolism in the skeletal system [1]. The abnormal bone metabolism is associated with reduced bone mineral density (BMD) and decreased trabecular bone structures and induces bone loss likely due to osteoporosis or osteopenia [2]. Severe osteoporosis is frequently suffered in patients with chronic liver disease, especially in end-stages and in chronic cholestasis, non-alcoholic fatty liver disease, haemochromatosis, and alcoholism [1,2]. However, the pathogenesis and mechanisms underlying bone reduction in chronic liver disease remain unclear. Thus, it is necessary to elucidate the critical factors to develop an alternative option for bone loss in chronic liver disease. Reactive oxygen species (ROS) are known to trigger the progression of chronic liver fibrosis [3]. ROS are released from injured hepatocytes and transform quiescent hepatic stellate cells (HSCs) into their active form. ROS-activated HSCs release several inflammatory cytokines that recruit immune cells into liver tissue. Among these inflammatory cytokines, interleukin 17 (IL-17) promotes HSCs to produce extracellular matrix in the context of liver fibrosis [4]. Thus, hepatic ROS function to trigger complex interactions between activated HSCs and recruited immune cells to exacerbate fibrosis and inflammation within the liver. However, the cellular mediators and molecular mechanisms of hepatic ROSmediated bone loss in chronic liver fibrosis have not been elucidated.
Abnormal mineral turnover occurs due to the accelerated osteoclast function that underlies bone loss in patients with chronic liver disease [4,5]. Osteoclasts are responsible for bone mineral resorption and differentiated from bone marrow cells (BMCs) via tumor necrosis factor (TNF) receptor superfamily member 11a (TNFRSF11A) during osteoimmune communication [6]. IL-17 is known to induce osteoclasts via osteoblast; IL-17-related inflammatory conditions stimulate osteoblasts to express TNF superfamily member 11 (TNFSF11) and enhance osteoclast differentiation via TNFSF11eTNFRSF11A signaling [7]. Several chronic inflammatory diseases cause secondary bone reduction through IL-17-mediated osteoclasts [8,9], indicating a preventive potency of anti-IL-17 therapy in bone loss. However, it is not fully understood if hepatic ROS can trigger bone loss by IL-17mediated osteoclasts in chronic liver fibrosis and whether targeting hepatic ROS is effective to bone loss, as well as hepatic dysfunction, in chronic liver fibrosis. Human deciduous pulp stem cells were first identified tissue specific mesenchymal stem cells (MSCs) with clonogenicity properties with self-renewal and multipotency within the dental pulp tissues of exfoliated deciduous teeth and are referred to as stem cells from human exfoliated deciduous teeth (SHED) [10]. Current studies evaluated identified patient-derived dental pulp-tissue stem cells from disease-specific and disease-non-specific tissues [11e13] and clarify that the pharmacologically rejuvenation can improve dysfunction of patient-derived dental pulp-tissue stem cells [11,14,15]. Manufacturing of clinical grade SHED is established under a quality control for use in regenerative medicine [16,17]. Thus, SHED-based therapy is realistically considered to be a novel option for regenerative medicine [18]. Recently, it was reported that SHED that were converted into immature hepatocyte-like cells with limited hepatic functions (referred to as SHED-Heps) were established under the stimulation of hepatogenic cytokines and SHED-Hep transplantation (SHED-HepTx) showed anti-hepatic dysfunction and antifibrotic effects in animal models of Wilson's disease, hemophilia A, and chronic liver fibrosis [19e21]. However, it remains unclear if SHED-Heps are a feasible option for bone loss in chronic liver fibrosis. The present study was designed to investigate the anti-bone loss therapy, as well as anti-hepatic dysfunction and anti-fibrotic therapies, of SHED-HepTx in chronic liver fibrosis. Furthermore, we examined whether targeting anti-hepatic ROS effect of SHED-HepTx is effective to bone loss, as well as hepatic dysfunction, in chronic liver fibrosis.

Ethics statement, human subjects, and animals
Human deciduous teeth were collected from discarded clinical samples from healthy pediatric donors (5e7 years old, n ¼ 3) with written informed consent from the guardian of each child donor at the Department of Pediatric Dentistry, Kyushu University Hospital. Procedures for handling human samples were approved by the Kyushu University Institutional Review Board for Human Genome/Gene Research (Protocol Numbers: 738-01, 02, 03, and 04). All animal experiments in this study were approved by the Institutional Animal Care and Use Committee of Kyushu University (protocol numbers: A20-041-0 and A21-044-1). All methods were performed in accordance with relevant guidelines and regulations.

Animals
C57BL/6J mice (female, 6 weeks old) and pregnant mice were obtained from the Jackson Laboratories Japan (Yokohama, Japan). The animals were housed individually and freely provided with sterile water and standard chow under controlled environmental conditions with a 12 h light/12 h dark cycle.
2.3. Culture of SHED and SHED-Heps SHED were isolated by a colony-forming unit fibroblast method, cultured, and characterized as previously described [22]. SHED-Heps were induced under hepatogenic conditions and characterized as previously described [16,19]. SHED-Heps were pretreated for 3 days with small interfering RNA (siRNA) specific for stanniocalcin 1 (STC1) or with scrambled control (referred to as siSTC1 and siCONT, respectively; Santa Cruz Biotechnology, Santa Cruz, CA, USA). The culture details are described in the Supplementary Methods. via the spleen, and additional CCl 4 was administered for four weeks (referred to as SHED-HepTx mice and CCl 4 mice, respectively. Agematched mice received olive oil alone (Nacalai Tesque), referred to as control mice. All animals did not receive any immunosuppression and conditioning throughout this study. Mouse livers, long bones, and serum were harvested eight weeks after CCl 4 treatment.

2.5.
In vivo fibroinflammatory and ROS production assays in CCl 4 treated mice Serum levels of aspartate aminotransferase (AST), alanine aminotransferase (ALT), total bilirubin, and hepatic hydroxyproline (HYP) were measured by colorimetric analyses. The hepatic distribution of collagen was analyzed by Sirius Red staining. The hepatic localization of actin alpha 2 and smooth muscle (ACTA2) was analyzed by immunohistochemical analysis. The hepatic expression of Acta2, collagen type 1 alpha 1 (Coll1a1), Il-17, peroxisome proliferator-activated receptor gamma (Pparg), and nicotinamide adenine dinucleotide phosphate oxidase 4 (Nox4) was analyzed by quantitative reverse transcription polymerase chain reaction (RT-qPCR). Malondialdehyde (MDA) levels and glutathione peroxidase (GSH-Px) activity were measured in mouse livers by colorimetric analysis. The expression of serum amyloid A1 (Saa1) in mouse livers were analyzed by RT-qPCR. Serum levels of granulocyte stimulating factor (G-CSF), IL-17, SAA1, and transforming growth factor beta (TGFB) were analyzed using ELISA.

Immunological localization of donor cells in CCl 4 -treated mice
The hepatic distribution of human leukocyte antigens A, B, and C (HLA-ABC) and of human hepatocyte paraffin 1 antigen (HepPar1) was analyzed by immunohistochemical analysis. The co-distribution of HepPar1 and ACTA2 in mouse livers was analyzed using double immunofluorescence analysis.

Bone mineral assays in CCl 4 -treated mice
Trabecular bones of mouse tibiae were analyzed by micro-computed tomography (microCT) assays performed on a SkyScan 1076 scanner (Bruker, Billerica, MA, USA) using CT-Analyzer and CT-Volume software (Bruker) [23]. Serum levels of C-terminal telopeptide of type I collagen (CTX-I) and tartrate-resistant acid phosphatase 5 b (TRAP-5b) were analyzed by an enzyme-linked immunosorbent assay (ELISA).

2.9.
In vitro osteoclast inductive assay Mouse BMCs were isolated from femurs and tibiae of mice and cocultured with newborn calvarial osteoblasts and the osteoclast formation was determined as reported previously [24,25] , 5% non-essential amino acids (Nacalai Tesque), and premixed antibiotics containing 100 U/mL penicillin and 100 mg/mL streptomycin (Nacalai Tesque) in Dulbecco's modified Eagle's medium (Nacalai Tesque). The ROS content in the conditioned medium and the cell viability of mHeps were both analyzed by colorimetry.

Statistical analysis
Each test was performed in triplicate, and the results were expressed as the mean AE standard error of the mean. Comparisons between two groups were performed using independent two-tailed Student's t-test. Multiple group comparisons were performed using one-way repeated measures analysis of variance followed by the Tukey's post hoc test. Statistical significance was set at P ˂ 0.05. All statistical analyses were performed using PRISM 6 software (GraphPad Software, La Jolla, CA, USA).

Donor SHED-Heps engraft into the periportal area of CCl 4injured mouse livers
Isolated SHED exhibited characteristics of MSCs, including attached colony formation, immunophenotypes, and mesenchymal multipotency (Supplementary Figure 1). SHED-Heps exhibited immature hepatocytelike characteristics and limited-hepatic function compared to primary human hepatocytes by hepatocyte-specific gene expression and hepatic function analyses (Supplementary Figure 2). In vivo imaging demonstrated that DiR-fluorescence activity was detected the donor cell in situ in the liver region of DiR-labeled SHED-HepTx mice but not in that of the non-infused mice at 5 days after infusion ( Figure 1A). Using ex vivo imaging, DiR-fluorescence activity was detected in the livers and spleens but not in the lungs and kidneys of the DiR-labeled SHED-HepTx mice ( Figure 1B). No fluorescent activity was detected in the lung, livers, spleens, and kidneys of noninfused mice ( Figure 1B). Immunohistochemical analysis revealed that HLA-ABC-positive and HepPar1-positive cells were engrafted into the periportal region of the livers of SHED-HepTx mice (SHED-HepTx livers) but not in the livers of control and CCl 4 mice (control and CCl 4 livers, respectively) ( Figure 1C, D). ELISA detected serum human albumin in the SHED-HepTx mice but not in the control and CCl 4 mice ( Figure 1E). Double immunofluorescence analysis demonstrated that donor HepPar1-positive cells were localized close to recipient ACTA2-positive cells in the periportal region of SHED-HepTx livers ( Figure 1F). Immunohistochemical control tests using mouse IgG1 and IgG2a revealed no immune reactions (Supplementary Figure 3).

SHED-HepTx improves hepatic ROS-induced fibroinflammation in CCl 4 -injured mice
Eight weeks after CCl 4 treatment, the CCl 4 mice exhibited increased liver fibrosis compared to the control mice ( Figure 2). The SHED-HepTx mice exhibited the reduced serum levels of AST, ALT, and total bilirubin compared to the CCl 4 mice by colorimetric analyses (Figure 2A). By Sirius Red staining and immunohistochemical analysis, the SHED-HepTx livers reduced the deposition of fibrous tissues and expression of ACTA2 compared to the CCl 4 livers ( Figure 2B,C). By colorimetric analysis and RT-qPCR, the SHED-HepTx livers decreased the HYP content and expression of Acta2 and Col1a1 compared to the CCl 4 livers ( Figure 2D, E). Hepatic ROS production is consistently manifested by measuring MDA and GSH-Px in liver [26]. Pparg and Nox4 are unambiguous markers of activated HSCs under ROS stimulation [27,28]. IL-17 is a critical mediator to produce extracellular matrix by HSCs [4]. RT-qPCR revealed that the CCl 4 livers showed higher Il-17 expression than the control livers, while the SHED-HepTx livers expressed lower Il-17 level than the CCl 4 livers ( Figure 2F). The enhanced ROS production and increased HSC activation was determined in the CCl 4 livers compared to the control livers, as indicated by the increased levels of MDA and GSH-Px by colorimetric analysis and the reduced Pparg and increased Nox4 levels by RT-qPCR, meanwhile, the SHED-HepTx livers suppressed the ROS production and HSC activation ( Figure 2G, H).

SHED-HepTx improves trabecular bone density and suppresses osteoclast differentiation via IL-17 in the bone marrow of CCl 4 -injured mice
MicroCT analysis revealed that the femurs of CCl 4 mice exhibited decreased trabecular bone structure, reduced trabecular BMD and abnormal trabecular bone parameters, including bone volume/ trabecular volume, trabecular thickness, trabecular number, and trabecular separation, compared to that of the control mice ( Figure 3AeD). ELISA demonstrated that the CCl 4 mice expressed the increased serum levels of CTX-I and TRAP-5b compared to the control mice ( Figure 3E). BMCs were isolated from control, CCl 4 -treated, and SHED-Hep-Tx mice (Cont-BMCs, CCl 4 -BMCs, and SHED-Hep-TX-BMCs, respectively) and cocultured with calvarial osteoblasts.
In vitro osteoclastogenic assay revealed that the CCl 4 -BMCs increased the number of TRAP-positive multinuclear cells (MNCs) and levels of Tnfrsf11a, nuclear factor of activated T-cell (Nfatc1), and cathepsin K (Ctsk), compared to the Cont-BMCs by TRAP staining and RT-qPCR ( Figure 3F, G and Supplementary Figure 4). Conversely, the SHED-HepTx mice exhibited recovery of bone structure, BMD, and parameters of trabecular bone and improved serum levels of CTX-I and TRAP-5b compared to the CCl 4 mice by microCT analysis and ELISA (Figure 3AeE). In vitro osteoclastogenic assay revealed that the SHED-HepTx-BMCs exhibited lower osteoclast differentiation than the CCl 4 - Original Article   Figure 3H). Mouse BMSCs were isolated from control, CCl 4 , and SHED-Hep-Tx mice (Cont-BMSCs, CCl 4 -BMSCs, and SHED-HepTx-BMSCs, respectively). In vitro osteogenic assay revealed that the CCl 4 -BMSCs reduced the osteogenic capacity compared to the Cont-BMSCs, as indicated by the decreased formation of mineralized nodules and suppressed expression of runt-related transcription factor 2, alkaline phosphatase, and bone gamma-carboxyglutamate protein, four and two weeks after osteogenic induction by Alizarin Red staining and RT-qPCR, respectively (Supplementary Figure 5). Meanwhile, the SHED-HepTx-BMSCs showed the improved osteogenic capacity compared to the CCl 4 -BMSCs (Supplementary Figure 5). Immunofluorescence analysis showed that HLA-ABC-positive cells were not detected in the bone marrow of SHED-HepTx mice four weeks after transplantation ( Figure 3I), indicating that the bone marrow is not the direct target of donor SHED-Heps.

Bone cells is impaired by ROS, but not by CCl 4 , in vitro
Since CCl 4 causes necrosis of hepatocytes [29], the toxicity of CCl 4 to mouse BMCs and BMSCs were analyzed under the stimulation of  Figure 10E).
ELISA demonstrated that the CCl 4 mice expressed the increased serum levels of TGFB compared to the control mice (Supplementary Figure 14A), which is correlated with previous studies that reflect to the osteoblast dysfunction in bone loss of chronic liver disease [31,32]. siSTC1-SHED-HepTx attenuated the recovered serum levels of TGFB of siCONT-SHED-HepTx in the CCl 4 mice (Supplementary Figure 14A).
In vitro osteogenic capacity demonstrated that the CCl 4 -BMSCs exhibited the decreased osteogenic capacity compared to the Cont-BMSCs by Alizarin Red staining and RT-qPCR, but the BMSCs of siSTC1-SHED-HepTx mice, siSTC1-SHED-HepTx-BMSCs, suppressed the improved osteogenic capacity compared to the BMSCs of siCONT-SHED-HepTx mice, siCONT-SHED-HepTx-BMSCs (Supplementary Figure 14BeD). SEMA3A is known as an osteoprotective factor produced by osteoblasts and regulates bone volume by a balance between osteoclastogenic suppression and osteogenic promotion [33]. The CCl 4 -BMSCs exhibited the decreased expression of Sema3a compared to the Cont-BMSCs two weeks after osteogenic induction by RT-qPCR but the siSTC1-SHED-HepTx-BMSCs improved the Sema 3a level in the siCONT-SHED-HepTx-BMSCs (Supplementary Figure 14D).

DISCUSSION
We demonstrate that hepatic fibro-inflammation is caused by hepatic ROS released from damaged hepatocytes in CCl 4 -induced chronic liver disease model mice. MSC-releasing STC1 play an important role in treating several ROS-induced diseases, including retinal degeneration, obesity-induced hepatitis, and lung fibrosis [34e36]. Recent transcriptomic and proteomic analyses reveal that the molecular mechanism of CCl 4 -damaged liver fibrosis is related to oxidative stress and PPAR signaling pathway [37]. A previous study demonstrates that SHED-Hep-secreting STC1 suppressed ROS-mediated hepatocyte necrosis in Wilson's disease model rats [19]. The present siSTC1-SHED-HepTx showed the attenuation of therapeutic benefits of SHED-HepTx on abundant ROS production and fibro-inflammation in chronically CCl 4 -induced liver fibrosis. These findings suggest that hepatic ROS-targeting may offer a novel modality for treating chronic liver fibrosis in SHED-Hep-based therapy. The present study demonstrates that CCl 4 exhibits liver toxicity, but does not cause bone toxicity, indicating that the liver-releasing factors affects bone metabolism in CCl 4 -induced chronic liver disease, as correlated with the previous studies [32,38]. However, the regulation in the liverebone axis was unknown. Recently, liver releasing factors including insulin-like growth factor binding protein 1, vitamin D, and TGFB participate in a liverebone axis to cause the bone reduction in chronic liver disease [32,38]. It is known that IL-17 increases the expression of TNFSF11 on osteoblasts to induce osteoclast differentiation via TNFRSF11A [39]. Our in vivo and in vitro studies indicate that liver releasing ROS induced expression of Il-17 and Tnfrsf11a in BMCs enhances the osteoclast formation via TNFSF11eTNFRSF11A signaling. Moreover, we show that hepatic ROS-induced osteoblast dysfunction is associated with the bone reduction in CCl 4 -induced mice, as reported previously in cholestasis of patients and bile duct ligated or CCl 4 -treated mice [31,32,40,41], suggesting that liver releasing ROS exacerbate the bone loss through the imbalance between osteoclast and osteoblast activities in CCl 4 -induced mice. Further functional knock-down of STC1 in donor SHED-Heps attenuate the suppressed osteoclast and inducible osteoblast functions of SHED-HepTx in CCl 4induced mice. Thus, these findings suggest that liver releasing ROS target the bone cells including BMCs and BMSCs to cause bone loss through the imbalance between osteoclast and osteoblast differentiation in chronic liver fibrosis and indicated that SHED-Hep-based therapy targets liver releasing ROS to regulate the bone metabolism, as well as fibro-inflammation, in chronic liver fibrosis. We speculate another pathological sequence of gained expression of IL-17 in BMCs of CCl 4 induced mice. Liver-releasing ROS recruits IL-17-producing immune cells into the injured liver of chronic liver disease [42,43]. Recent study shows that Saa-overexpression recruits IL-17-secreting neutrophils in bone marrow, leading to exacerbating bone loss [44]. An increased level of circulating Th17 cells is implicated in bone loss in primary sclerosing cholangitis [45]. Given the present results that CCl 4 -induced hepatic ROS enhances the expression of bone marrow Il-17 and secretion of hepatic SAA1 in chronically CCl 4treated mice, we speculate that IL-17-secreting immune cells may contribute the liverebone axis to induce bone loss in chronic liver diseases. Further study will be necessary to elucidate the mechanism of requiting IL-17-producing cells into bone marrow under hepatic ROS condition.
Given the present findings that liver-releasing ROS targets BMSCs in chronically CCl 4 -treated mice, we suppose another possibility of antibone loss efficacy in SHED-Hep-based therapy that the recipient BMSC-targeting STC1 released from SHED-Heps might contribute the bone recovery in chronic liver disease. Recent STC1 knock-in and knock-down study shows the STC1-enriched EVs released from adipose MSCs participate in angiogenesis in carotid endarterium mechanical injury [46]. SHED-releasing EVs play a crucial role in targeting the recipient bone marrow MSCs in osteoporotic and autoimmune model mice [22,23]. STC1 enhances the differentiation and bone formation of osteoblasts in an autocrine/paracrine manner [47]. Locally secreted STC1 is also known to act as a hormone to regulate distant tissues/organs [48,49]. We demonstrate that siSTC1-SHED-HepTx attenuate the improvement of the osteogenic and osteoclastogenic functions of the recipient BMSCs in chronically CCl 4 -treated mice with SHED-Hep-Tx. Further study will be necessary to elucidate the mechanism of SHED-Hep-releasing STC1 target the recipient BMSCs in chronic liver disease. Taken together, the present findings suggest hepatic ROS-induced chronic liver fibrosis causes bone loss by the imbalance of osteoclast and osteoblast activities in a liverebone axis. The present study also indicate that targeting of hepatic ROS may provide a valuable means for anti-bone loss treatment, as well as anti-fibro-inflammatory treatment, in chronic liver fibrosis. This hepatic ROS-targeting SHED-Hep-based approach may provide a feasible tool for the development of effective therapies for various liver disorders and their associated secondary disorders.

DATA AVAILABILITY
No data was used for the research described in the article.