Cytokine Mixtures Mimicking Secretomes From Mesenchymal Stem Cells Improve Medication‐Related Osteonecrosis of the Jaw in a Rat Model

ABSTRACT Recently, several studies have demonstrated that intravenous administration of mesenchymal stem cells (MSCs) improve medication‐related osteonecrosis of the jaw (MRONJ), and paracrine effects of secretomes from MSCs have been hypothesized as the primary contributors. These secretomes in conditioned media from human MSCs (MSC‐CM) were previously demonstrated to promote bone and tissue regeneration. Because MSC‐CM contain cytokines monocyte chemoattractant protein (MCP)‐1, insulin growth factor (IGF)‐1, and vascular endothelial growth factor (VEGF) at relatively higher concentrations than other factors, these cytokines were considered as relevant active factors for tissue regeneration. By mixing the recombinant proteins of MCP‐1, IGF‐1, and VEGF, included at the same concentrations in MSC‐CM, we prepared cytokine mixtures mimicking MSC‐CM and then evaluated its therapeutic effects in a rat MRONJ model. In vitro, cytokine mixtures promoted osteogenic differentiation, migration, and proliferation of rat MSCs. In addition, these maintained osteoclastic function. In vivo, we used a rat MRONJ model to examine therapeutic effects of the cytokine mixtures through intravenous administration. In MSC‐CM or cytokine mixture group, open alveolar sockets in 66% or 67% of the rats with MRONJ, respectively, healed with complete soft tissue coverage and socket bones, whereas in the other groups, the exposed necrotic bone with inflamed soft tissue remained. Histological analysis revealed new bone formation and the appearance of osteoclasts in MSC‐CM or cytokine mixture group; however, osteoclasts were significantly reduced in the other groups. Thus, we concluded that intravenous administration of cytokine mixtures might be an effective therapeutic modality for treating patients with MRONJ. © 2017 The Authors JBMR Plus published by Wiley Periodicals, Inc. on behalf of American Society for Bone and Mineral Research


Introduction
B isphosphonates (BPs) were originally developed as drugs for the treatment of bone-resorbing diseases such as multiple myeloma and bone metastasis, whose typical origins are breast and prostate cancers as well as tumor-related hypercalcemia, (1)(2)(3)(4) and for preventing pathological fractures in patients with osteoporosis. Lately, several reports have suggested that a rare but potentially severe side effect of BPs is medicationrelated osteonecrosis of the jaw (MRONJ), (5) defined as an exposed bone in the maxillofacial region that persists for >8 weeks in patients with present or previous BP treatment without a history of radiation therapy in the jaw.
To date, various attempts to control this disorder have been unsuccessful, and standard osseous sequestrectomy usually results in further enlargement of the bone defects. (6,7) Therefore, conservative, nonsurgical approaches have been recommended for the management of MRONJ, which slow its deterioration but do not cure the disease. (8)(9)(10) Thus, development of an effective approach is an urgent concern for the prevention and treatment of patients with MRONJ.
Bone marrow-derived mesenchymal stem cells (MSCs) have been extensively investigated in clinical practice as tools for the regeneration of bones. (11,12) In fact, some studies have reported that an intravenous injection of MSCs improves MRONJ. (13,14) However, several issues with stem cells remain to be addressed, including tumorigenesis, (15) poor survival of implanted cells, (16) and transmission of infectious diseases. Because implanted MSCs may contribute to tissue regeneration not only via pluripotency but also via paracrine effects, (17,18) secretomes in conditioned media from human MSCs (MSC-CM), which are known to include various types of cytokines and other factors, may help address these issues.
We previously reported therapeutic effects of MSC-CM in a rat MRONJ model. (19) Furthermore, we reported that MSC-CM contains numerous cytokines, including monocyte chemoattractant protein (MCP)-1, insulin growth factor (IGF)-1, and vascular endothelial growth factor (VEGF). (19,20) These cytokines regulate migration, angiogenesis, antiapoptosis, and osteogenesis in host MSCs, osteoclast precursors, or immune cells in inflamed tissues and may thus accelerate the regeneration of bone and other tissues. (21,22) It is possible that only a small number of specific factors in MSC-CM are capable of regenerating bone and other tissues. In fact, we hypothesized that MCP-1, IGF-1, and VEGF are the relevant active factors because according to our previous reports, (20) these were present in MSC-CM at relatively higher concentrations than the other factors and are well known to affect cellular migration, angiogenesis, antiapoptosis, and osteogenesis. In the present study, these cytokines (MCP-1, IGF-1, and VEGF) were mixed at the concentrations detected in MSC-CM, and the mixtures were prepared to mimic MSC-CM. Therapeutic effects of these cytokine mixtures on MRONJ was evaluated and compared with those of MSC-CM by using a rat MRONJ model.

Animal experiments
All animal experiments conducted in this study were in strict accordance with the protocols approved by the Guidelines for Animal Experimentation of the Nagoya University School of Medicine (approval nos. 25374 and 26063).

Cell preparation
Human MSCs (hMSCs) were purchased from Lonza Inc. (Walkersville, MD, USA) and cultured in MSC basal medium (Lonza Inc.) containing MSCGM SingleQuots (Lonza Inc.) at 37°C in 5% CO 2 /95% air. After primary culture, the cells were subcultured at a density of approximately 1 Â 10 4 cells/cm 2 . For the experiments, hMSCs at third to sixth passages were used.
Rat MSCs (rMSCs) were isolated from 7-week-old Wistar/ST male rats (Japan SLC, Shizuoka, Japan) as previously reported. (23) In brief, donor rats were euthanized and the femora were dissected out. Under sterile conditions, the edge of each bone was cut, Dulbecco's modified Eagle medium (DMEM; Gibco, Rockville, MD, USA) was injected into the bone marrow using an 18-gauge syringe, and bone marrow cells were flushed out to the opposite side; this procedure was repeated several times. The marrow was then seeded into each tissue culture flask in DMEM containing an antibiotic-antimycotic solution (100 units/ mL penicillin G, 100 mg/mL streptomycin, and 0.25 mg/mL amphotericin B; Gibco), and the medium was supplemented with 10% fetal bovine serum (FBS). Three days after seeding, the floating cells were removed and the medium was replaced with a fresh medium. The adherent, spindle-shaped cells were passaged when the cells approached confluence. The adherent cells were collected using trypsin/EDTA, resuspended in fresh medium, and transferred to new flasks at a density of 1 Â 10 4 cells/cm 2 .

Preparation of conditioned media
Eighty-percent confluent hMSCs were replenished with serumfree DMEM (Gibco) containing antibiotic-antimycotic solution. The cell-cultured conditioned media (CM) were collected after 48-hour incubation. Subsequently, the CM were centrifuged at 440g for 5 minutes at 4°C. The supernatant was collected and centrifuged at 17,400g for 3 minutes at 4°C, filtered at 0.22 mm (Millex-GP; Merck Millipore Ltd., Billerica, MA, USA), and used as MSC-CM. MSC-CM was stored at À80°C before being used for the experiments detailed below.
rMSCs and rOCs (Cosmo Bio Co. Ltd., Tokyo, Japan) were cultured at a density of 1 Â 10 4 and 3 Â 10 4 cells/well, respectively, in 96-well culture plates at 37°C in 5% CO 2 /95% air. When the cells became 80% confluent, these were treated with concentrations of 150 mM Zol (clinical concentrations) with or without MSC-CM or cytokine mixtures for 48 hours. After the indicated culture time, MTT assay was performed by incubation of cells with 0.2% MTT solution (Cell Counting Kit, DOJINDO, Tokyo, Japan) at 37°C for 4 hours. The resulting color was then analyzed by measuring the absorbance at 450 nm (A450), wherein A450 corresponded to the viability of the cells. (27) Pit formation assay To confirm the function of rOCs, we performed a pit formation assay. rOC precursors were cultured at a density of 6 Â 10 4 cells/ well in a 96-well Osseo Assay Stripwell Plate (Corning Inc., Corning, NY, USA) at 37°C in 5% CO 2 /95% air for 10 days. These were treated with BM, BM þ 150 mM Zol, BM þ 150 mM Zol þ MSC-CM, BM þ 150 mM Zol þ MIV, or BM þ 150 mM Zol þ MSC-CM-dep MIV for 10 days. Six wells were used per group. To measure the areas containing resorption, we performed the von Kossa staining. Cells in the well plates were fixed using 100% ethanol and then stained with 5% silver nitrate, placed under a UV lamp for 30 minutes, and rinsed with distilled water before treatment with 5% sodium thiosulfate for 2 to 3 minutes. Images of the resorbed area were measured using light microscopy and results were expressed as the total area resorbed in at least six random fields, as evaluated using Image J software (version 1.67; NIH, Bethesda, MD, USA).

Alizarin red staining
For mineralization analysis, Zol rMSCs were cultured with continuous exposure to osteogenic induction medium (OM; Osteogenic Differentiation Medium BulletKit; Lonza Inc.) with MSC-CM, MIV, or MSC-CM-dep MIV for 14 days and were subsequently fixed using 90% ice-cold ethanol for 10 minutes. After fixation, ethanol was removed and the amount of mineralization was measured using 1% alizarin red S (pH 6.2) for 30 minutes at RT. The stained cultures were subsequently measured for A570 and photographed. (27) Wound healing assay The migratory properties of Zol rMSCs were examined using the CytoSelect Wound Healing Assay kit (Cell Biolabs, San Diego, CA, USA) according to the manufacturer's instructions. In brief, cell suspension was added to the well using a plastic insert. The insert was removed from the well after a monolayer of cells had formed, creating a wound gap of 0.9 mm. After washing, the cells were incubated at 37°C for 48 hours with MSC-CM, MIV, MSC-CM-dep MIV, 30% FBS, or serum-free DMEM (DMEM [À]). The extent of wound closure was determined using a light microscope (CK40; Olympus, Tokyo, Japan).

Real-time reverse transcription polymerase chain reaction (RT-PCR) analysis
Zol rMSCs were cultured with DMEM (À), MSC-CM, MIV, or MSC-CM-dep MIV for 48 hours. Total RNA was extracted using the RNeasy Mini Kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer's instructions. RT-PCR analysis was performed as previously described. (28,29) The sequence of specific primers and probes used for real-time RT-PCR analysis of alkaline phosphatase (alp), runt-related transcription factor 2 (runx2), vegf, rankl, osteoprotegerin (opg), and gapdh are listed in Table 1.
RT-PCR and the resulting relative increase in reporter fluorescent dye emission were monitored in real time using the 7000 Sequence Detector (Perkin-Elmer, Foster City, CA, USA). Signals were analyzed using Sequence Detector Software (version 1.1; Perkin-Elmer). The PCR conditions were as follows: one cycle at 50°C for 2 minutes, one cycle at 60°C for 30 minutes, one cycle at 95°C for 5 minutes, 50 cycles at 95°C for 20 seconds (denaturation), and then 60°C for 1 minute (annealing and extension). The relative amount of each mRNA in one sample was obtained by calculating the respective standard curves. The standard curves for each mRNA were drawn using different concentrations (625, 125, 25, 5, and 1 ng) of total RNA of rMSCs. The relative expression levels were normalized to gapdh expression.

Generation of a rat MRONJ model and therapeutic effects of MIV
We evaluated a rat MRONJ model in our previously report (19) and established that model. In brief, 5-week-old Wistar/ST male rats received Zol (35 mg/kg/wk) and Dex (1 mg/kg/d) subcutaneously for 2 weeks. (19,31) Thereafter, unilateral maxillary molars were extracted with minimal trauma to the surrounding tissues under deep anesthesia using intraperitoneal injection of ketamine (60-90 mg/kg) and xylazine (100-150 mg/kg). We defined only the rats with bone exposure as the MRONJ model. Two weeks after tooth extraction, we defined the following groups based on the material used for the intravenous injection: 1) nontreatment group, not administered; 2) DMEM (À) group (1 mL

Radiographic and histological analyses
Surgical sites were dissected, fixed in 4% PFA, and subjected to micro-computed tomography (micro-CT) analysis using a laboratory X-ray CT device (SkyScan1176; Bruker, Kontich, Belgium). Using the software supplied with the instrument, two-dimensional (2D) and three-dimensional (3D) images were reconstituted.
After the radiological assessment, explants were decalcified using K-CX solution (Falma Co., Tokyo, Japan), dehydrated with graded ethanol, cleared with xylene, and embedded in paraffin. The specimens were sagittally cut to create 5-mm-thick histological sections and stained using hematoxylin and eosin (H & E).
Empty osteocytic lacunae were counted in at least 3 random fields and expressed as a percentage of total bone lacunae. Osteoclasts were identified using tartrate-resistant acid phosphatase (TRAP) staining with the leukocyte acid phosphatase kit (Sigma-Aldrich, Tokyo, Japan) according to the manufacturer's instructions. Histological analysis was performed using light microscopy.

Immunohistochemistry
Fresh frozen samples were sectioned at 8 mm according to the Kawamoto method using the Multi-Purpose Cryosection Preparation Kit (Leica Microsystems, Tokyo, Japan). (32) Each section was fixed in 100% ethanol for 10 minutes at RT, washed, blocked with 5% bovine serum albumin/PBS for 30 minutes, and stained with primary antibodies in blocking buffer overnight at 4°C. The sections were then stained with secondary antibodies for 30 minutes, washed, stained with 4 0 ,6-diamidino-2phenylindole (DAPI; Invitrogen, Tokyo, Japan) for 15 minutes, mounted with SCMM R3 (Leica), and examined using fluorescence microscopy (BZ-9000, Keyence, Osaka, Japan). Purified mouse antibody against rat CD31 (Purified Mouse Anti-Rat CD31; BD Pharmingen, Tokyo, Japan) was used for immunostaining. Secondary antibodies were conjugated using Alexa Fluor 633 (Invitrogen).

Statistical analysis
All experiments were performed in triplicate and repeated at least twice. Group means and standard deviations (SDs) were calculated for each measured parameter. Statistical differences were evaluated using Tukey's honestly significant difference test. A p value <0.05 was considered statistically significant and that <0.01 was considered statistically very significant.

Effect of cytokine mixtures on rMSCs or rOCs viability
MTT assay was performed to examine its effect on cytotoxicity and cell viability of rMSCs and rOCs. The results in Fig. 1A, B indicate that cell viability was significantly increased with the addition of 150-mM Zol-added MSC-CM or MIV. However, the viability of rMSCs and rOCs was decreased with only Zol, single cytokine, two combination cytokines, or MSC-CM-dep MIV. Based on these results, three-cytokine mixtures, mainly MIV, may contribute to cell viability.

MIV maintains resorptive activity of osteoclasts
A pit formation assay was then performed to test the resorptive activity of osteoclasts ( Fig. 2A). A function maintenance effect on pit number and area was observed on the Osseo Assay Stripwell Plate (Corning Inc.) in Zol þ MSC-CM or Zol þ MIV group, whereas an inhibitory effect was observed in Zol or Zol þ MSC-CM-dep MIV group. Fig. 2B shows quantification of the resorptive areas. Thus, MIV showed function maintenance in osteoclasts.

MIV promotes osteogenic differentiation in Zol rMSCs
The osteogenic differentiation potential of Zol rMSCs was assessed using alizarin red staining. When treated with OM, Zol rMSCs exhibited weak alizarin red staining. However, when treated with MSC-CM or MIV together with OM, Zol rMSCs exhibited strong alizarin red staining (Fig. 3A, B) and when treated with MSC-CM-dep MIV together with OM, these exhibited weak alizarin red staining (Fig. 3A, B). Based on these

Effects of MIV on the migration of Zol rMSCs
The proportion of Zol rMSCs in the wound area of DMEM (À) was 6.91% AE 2.11%. There were 38.6% AE 7.46% Zol rMSCs in the wound area of positive controls (30% FBS). MSC-CM and MIV induced significant effects (p < 0.05) and closed the wound to 21.9% AE 1.52% and 22.0% AE 2.88% Zol rMSCs, respectively (Fig. 4A,  B). However, compared with MSC-CM and MIV, MSC-CM-dep MIV did not induce any significant effects and did not close the wound. Thus, compared with DMEM (À), MIV increased the migration of Zol rMSC more than threefold. These differences were statistically significant (p < 0.05), which indicates that MIV enhanced the migration of Zol rMSC (Fig. 4A, B).

MIV enhances osteogenic and angiogenesis gene markers or proteins and also enhances the central regulator of migration
To investigate the effects of MIV on Zol rMSCs, we performed RT-PCR and Western blot analyses. The expression levels of alp, runx2, and vegf genes and the runkl/opg ratio were significantly upregulated in Zol rMSCs cultured in MIV compared with those cultured in DMEM(À) or MSC-CM-dep MIV (Fig. 5A). Of note, mRNA expressions of these four genes were similar in MIV and MSC-CM (Fig. 5A).
Next, we examined the effects of MIV on ALP, RUNX2, VEGF, and m-TOR (a central regulator of proliferation, migration, and survival (33) ) in Zol rMSCs using Western blot analysis. In addition to mRNA levels, protein levels in Zol rMSCs also confirmed that ALP, RUNX2, VEGF, and m-TOR levels were higher in the MIV  group than those in the MSC-CM-dep MIV group. However, the protein levels in the MIV group were the same as those in the MSC-CM group (Fig. 5B, C). These results suggest that MIV upregulates osteogenesis, angiogenesis, and migration.

MIV injection cured MRONJ in rats as same as MSC-CM injection
An equivalent treatment protocol of high-dose intravenous nitrogen-containing BPs, Zol, and an immunosuppressive drug, Dex, is required for the development of a MRONJ disease model. Fig. 6A describes the protocol that we previously reported. (19) We used a rat MRONJ model revealing incomplete mucosal healing and the presence of exposed bone at 2 weeks after unilateral surgical extraction.
In the MSC-CM and MIV groups, micro-CT examinations revealed healing of bone defects compared with that in the nontreatment, DMEM(À), and MSC-CM-dep MIV groups at 2 weeks after intravenous injection (Fig. 6B). Two weeks after intravenous injection, open alveolar sockets of 66% or 67% of rats (n ¼ 8) with MRONJ in the MSC-CM or MIV groups, respectively, healed with complete soft tissue coverage, whereas in the nontreatment, DMEM(À), and MSC-CM-dep MIV groups, the exposed necrotic bone with inflamed soft tissue remained (Fig. 7B). Histological analysis revealed new bone formation and the appearance of osteoclasts in the MSC-CM and MIV groups; however, these were significantly reduced in the nontreatment, DMEM(À), and MSC-CM-dep MIV groups. In the MSC-CM and MIV groups, osteoclasts were observed on socket surfaces (Fig. 7A).
To further characterize impaired mucosal and bone healing, we assessed blood vessels. Immunohistochemical staining against CD31 for vascular endothelial cells showed increased endothelial cells in the MSC-CM and MIV groups compared with those in the other groups (Fig. 8). Particularly, CD31-positive endothelial cells decreased in the MSC-CM-dep MIV group. Based on these results, at 2 weeks after extraction, MRONJ rats receiving MIV injections showed complete mucosal healing, bone regeneration, and angiogenesis in the extracted alveolar sockets.

Discussion
BPs have been reported to promote osteoprotegerin, also known as osteoclastogenesis inhibitory factor, by MSCs and osteoblasts. (34) BPs inhibit the expressions of interleukin (IL)-6, M-CSF, and MCP-1 via inhibition of the Ras/MEK/ERK pathway and activation of p38MAPK in MSCs and osteoblasts, (35) thereby inhibiting the migration and differentiation of osteoclast precursors. Another study reported that BPs suppressed mineralization through direct cytotoxicity and inhibited osteoblast differentiation (27) or endothelial cell migration and differentiation.
Therapeutic effects of MSCs in tissue engineering and regenerative medicine can be partly attributed to paracrine pathways (18,24) triggered by several factors secreted into the CM. In fact, we have reported that MSC-CM contains numerous cytokines (20) such as MCP-1, IGF-1, and VEGF, which can synergistically affect migration, osteogenesis, and angiogenesis. The effects of these three cytokines are thought to be very complex, although IGF-1 is believed to regulate the migration of osteoblasts and MSCs, (36) and sustained systemic or local infusion of IGF-1 was shown to enhance bone formation. (37) On the other hand, VEGF is a master regulator of angiogenesis and enhances survival and differentiation in endothelial cells that promote osteogenesis, (38) and MCP-1 is a chemokine that recruits stem cells and immune cells to inflamed tissues. (39) MSC-CM contains numerous other soluble factors that may positively and negatively regulate bone and tissue regenerations; hence, it is necessary to identify the specific factors involved in bone and soft tissue regenerations. In this study, we hypothesized MCP-1, IGF-1, and VEGF as the relevant factors, and accordingly evaluated the activity and pharmacological effects of a defined mixture of these cytokines as well as those of MSC-CM from which these cytokines had been depleted. Concentrations of these MCP-1, IGF-1, and VEGF in MSC-CM, measured using ELISA, were 1103.96 AE 395.50, 1515.60 AE 211.83, and 772.55 AE 261.82 pg/mL, respectively; the cytokine mixture was prepared using 1100 pg/mL MCP-1, 1500 pg/mL IGF-1, and 770 pg/mL VEGF to match these concentrations.
Surprisingly, our data indicated that similar to MSC-CM, a defined mixture of cytokines enhanced the migration (Fig. 4) and expression of osteogenic and angiogenic marker genes in Zol rMSCs in vitro (Fig. 5) and had antiapoptotic effects on osteoclast precursors (Fig. 2). The clinically obvious MRONJ, specifically, open sockets with exposed necrotic bone and no mucosal lining, persisted beyond 2 to 3 weeks, and normal course of healing was observed in nontreated rats. (19) In our rat MRONJ model, MIV injection improved MRONJ in a similar manner as MSC-CM injection, but necrotic bone was not removed and sockets remained open in the other groups (Figs. 6 and 7).
Studies have also reported cooperative effects of these cytokines, which they exert on each other: IGF-1 upregulates VEGF expression through hypoxia-inducible factor-2a. (40) In addition, Chen and colleagues reported that MSCs could secrete microvesicles that contain various factors contributing to their angiogenesis-promoting function; among these, VEGF and MCP-1 might be of greater importance than the other cytokines. (41) Moreover, Lu and colleagues reported that MCP-1 is a major ligand of CC chemokine receptor 2 (CCR2) to recruit monocytes and stem cells into injured tissues in order to perform phagocytosis and produce IGF-1 for injury repair. (42) In our previous study that used a rat calvarial bone defect model, we demonstrated that three-cytokine mixture promotes the migration of stem cells and osteogenesis compared with that by a single cytokine or two-cytokine mixture. (43) Based on these findings, we assume that the cytokines in the three-cytokine mixtures may effectively activate multiple stages of the regeneration of bone and soft tissue, individually or in cooperation with each other.
In summary, this is the first study to report defined cytokine mixtures mimicking MSC-CM as a potential treatment for MRONJ. Such defined cytokine mixtures can be used as standardized formulations; however, the standardization of cells or cell-derived products from hMSCs can be challenging. Moreover, those can be directly used during surgery without any need to preculture patient MSCs. Intravenous administration of MIV provides an effective therapeutic modality for treating patients with MRONJ, which also provides insights regarding potential clinical applications.

Disclosures
All authors state that they have no conflicts of interest.