Bone loss in Staphylococcus aureus (S. aureus) osteomyelitis poses a serious challenge to orthopedic treatment. The present study aimed to elucidate how S. aureus infection in bone might induce bone loss. The C57BL/6 mice were injected with S. aureus (106 CFU/ml, 100 μl) or with the same amount of vehicle (control) via the tail vein. Microcomputed tomography (microCT) analysis showed bone loss progressing from week 1 to week 5 after infection, accompanied by a decreased number of osteocalcin-positive stained osteoblasts and the suppressed mRNA expression of Runx2 and osteocalcin. Transcriptome profiles of GSE30119 were downloaded and analyzed to determine the differences in expression of inflammatory factors between patients with S. aureus infected osteomyelitis and healthy controls, the data showed significantly higher mRNA expression of granulocyte colony-stimulating factor (G-CSF) in the whole blood from patients with S. aureus infection. Enzyme-linked immunosorbent assay (ELISA) analysis confirmed an increased level of G-CSF in the bone marrow and serum from S. aureus infected mice, which might have been due to the increased amount of F4/80+ macrophages. Interestingly, G-CSF neutralizing antibody treatment significantly rescued the bone loss after S. aureus infection, as evidenced by its roles in improving BV/TV and preserving osteocalcin- and osterix-positive stained cells. Importantly, we found that G-CSF level was significantly up-regulated in the serum from osteomyelitis patients infected by S. aureus. Together, S. aureus infection might suppress the function of osteoblastic cells and induce progressive bone loss by up-regulating the level G-CSF, suggesting a therapeutic potential for G-CSF neutralization in combating bone loss in S. aureus osteomyelitis.
Osteomyelitis, a persistent infection of bone tissue, is characterized by severe inflammation and progressive bone destruction . Staphylococcus aureus (S. aureus) is the most common causative organism for osteomyelitis [2,3]. S. aureus infection in bone may cause excessive bone loss and destruction that contribute to the increased fracture risk in patients suffered by osteomyelitis . However, the underlying mechanisms how S. aureus infection induces bone loss remain to be elucidated.
The adult skeleton is renewed continuously by remodeling, which involves two distinct processes, bone resorption by osteoclasts and subsequent new bone formation by osteoblasts . Runt-related transcription factor 2 (Runx2) is a transcription factor required for osteoblastic differentiation from bone mesenchymal stromal cells (BMSCs) as well as for the appropriate functioning of mature osteoblasts [6,7]. It has been reported that Runx2 regulates expression of genes encoding osteocalcin, an osteoblastic marker, and receptor activator of NFκB ligand (RANKL), an absolutely essential factor that initiates the osteoclast differentiation pathway and promotes the survival and activity of differentiated osteoclasts [8,9]. Overall, the above work highlights the importance of these genes in bone remodeling, dysregulated expression of these genes may impede bone remodeling and induces bone loss.
S. aureus and its products are potent stimulators of resorptive bone loss. Some products of S. aureus, such as surface-associated proteins and protein A were reported to stimulate osteoclastogenesis, thereby promoting bone resorption [10,11]. In addition, S. aureus products can also inhibit osteogenic differentiation of BMSCs and suppress the proliferation of osteoblast [12,13]. Importantly, it is apparent that, in addition to increased osteolysis and decreased bone formation induced by S. aureus and its products, bone loss during osteomyelitis progression is associated with inflammation. It was reported that strong inflammatory response was induced by S. aureus infection, resulting in the migration of a large number of neutrophils and macrophages to the site of infection . Bone remodeling process can be greatly changed by acute and chronic inflammation, which is associated with bone non-union and delayed fracture healing [15,16]. In the process of inflammation, innate and adaptive immune-related cells release a large number of inflammatory factors, such as tumor necrosis factor-α, interleukin (IL)-1, IL-6 and IL-17, leading to enhanced bone degradation and decreased bone formation [17–19]. However, the key inflammatory factor that leads to bone loss in osteomyelitis remains unclear .
Granulocyte colony-stimulating factor (G-CSF) is an inflammatory factor that affects the hematopoietic system by promoting differentiation, proliferation, and survival of granulocytes . In clinic, G-CSF treatment is often applied for patients with severe congenital neutropenia . However, long-term treatment with G-CSF is associated with development of clinically significant osteopenia, characterized by decreased bone mineral density (BMD) and vertebral compression fractures . Experimental study also revealed that G-CSF exposure led to decreased bone mass in cortical and trabecular bone in mice [24,25]. In addition, the mRNA expression and the serum level of G-CSF dramatically increased in the pneumonia mice model, orthopedic implant infection model induced by S. aureus infection, and in patients with S. aureus septicemia [26–28]. However, the role of G-CSF in bone loss induced by S. aureus osteomyelitis has not been investigated so far.
In the present study, we evaluated the time-dependent effects of S. aureus infection on bone metabolism and the levels of inflammatory factors in response to S. aureus infection. Herein, we characterized the levels of G-CSF in both the mice infected by S. aureus and the patients with S. aureus-induced osteomyelitis. Furthermore, we assessed the effect of G-CSF neutralizing antibody on bone mass in the mice infected with S. aureus. The present study is the first to show the role of G-CSF in mediating S. aureus infection-induced bone loss. We hope our findings may potentially help develop novel clinical therapies for osteomyelitis.
Bacterial strain and preparation of bacteria
S. aureus strain was isolated from the patient with chronic osteomyelitis was identified using PHOENIX 100 (Becton Dickinson Microbiology Systems, U.S.A.). The bacteria were cultured on tryptic soy agar containing 5% sheep blood and incubated overnight. Next, a single colony of S. aureus were selected and grown in tryptic soy broth (TSB) in tubes on a shaker at 37°C. When the optional density (OD) 600 reached 0.2, the bacteria were centrifuged at 3000 rpm for 10 min at 4°C. 15% glycerol/TSB solution was added to the pellet and the bacteria were aliquoted and stored at −80°C until use. On the day before inoculation, one frozen stock of S. aureus was thawed on ice and grown in TSB for 16 h, collected by centrifugation (10 min, 3000 rpm). The bacterial pellet was then washed with phosphate-buffered saline (PBS) for three times and resuspended in PBS. To determine the colony forming units (CFU)-to-OD relation of S. aureus, serial dilutions were prepared on tryptic soy agar plate containing 5% sheep blood in order to estimate bacterial colony numbers at each OD. The OD value at 1.0 was found to correlate with approximately 1 × 108 CFU/ml. Finally, the bacteria concentration was adjusted to the concentration required for injection.
Animals and treatments
Animal care and experimental procedures were approved by the Animal Care and Use Committee at the Southern Medical University. All the animal experiments were performed in the Key Laboratory of Bone and Cartilage Regenerative Medicine at Nanfang Hospital, Southern Medical University. After infection, mice were monitored daily in agreement with Animal Welfare Policies of Southern Medical University. Mice were housed in facility with 12 h light/dark cycle, 24°C room temperature and ad libitum access to water and food. To assess the time-dependent effects of S. aureus infection on bone mass, two groups of C57BL/6 mice were set with an uninfected control group undergoing intravenous (i.v.) injection of 100 μl PBS and an infected group receiving i.v. injection of S. aureus (1 × 106 CFU/ml, 100 μl). Animals were killed at 1 week and 5 weeks after injection, their femurs and tibias were dissected for further analysis.
To detect the effect of G-CSF neutralizing antibody on bone mass of S. aureus-infected mice, three groups of mice were set, with an infected group receiving i.v. injection of S. aureus (1 × 106 CFU/ml, 100 μl), a G-CSF neutralizing group receiving subcutaneous injection of G-CSF neutralizing antibody (clone 67604, R&D Systems, 10 mg/kg) 30 min prior to S. aureus infection, and a control group receiving i.v. injection of the same volume of PBS. Animals were killed 2 weeks after injection, and their femurs and tibias were dissected for further analysis.
Mice tibias were fixed in 4% paraformaldehyde for 48 h and analyzed by Skyscan 1176 (Skyscan, Aartselaar, Belgium). The scanner was set at a voltage of 50 kV, a current of 400 μA, and a resolution of 8.88 μm/pixel. The region of interest (ROI) in trabecular bone was drawn from 0.15 to 1.15 mm blow the lowest point of growth plate. After 3D measurement of trabecular morphology, the trabecular bone volume fraction (the ratio of the segmented bone volume to the total volume, BV/TV), trabecular thickness (Tb. Th, the thickness of trabeculae), trabecular number (Tb. N, the average number of trabecular per unit length), and trabecular separation (Tb. Sp, the average thickness of space between trabeculae) were analyzed to evaluate the microarchitecture of trabecular bone.
Tibias were dissected free from muscle, fixed in 4% paraformaldehyde for 24 h, decalcified in 10% ethylenediaminetetraacetic acid disodium salt (EDTA) (pH 7.4) for 4 weeks, dehydrated through graded ethanol series and then embedded in paraffin. 5 μm sections were cut longitudinally and processed for hematoxylin and eosin (H&E) staining or tartrate-resistant acid phosphatase (TRAP) staining. For TRAP staining, the deparaffinized and rehydrated sections were stained with Acid Phosphatase Leukocyte kit (Sigma) according to the manufacturer’s protocol. The TRAP+ mononuclear cells and multinucleated cells containing at least three nuclei were identified as preosteoclasts and osteoclasts, respectively, and counted under an inverted microscope (Leica DMI6000B, Solms, Germany). The TRAP+ cells per millimeter of bone surface (No./mm) were quantified.
Immunohistochemistry and immunofluorescence
For immunohistochemical analyses, after deparaffinization and rehydration, antigen retrieval was performed using citrate buffer (0.01 M, pH 6.0) or 10 μg/ml proteinase K solution. 3% H2O2 was used to inactivate endogenous peroxidase before blocking with 10% goat serum. Next, the sections were incubated with anti-osteocalcin (ab93876, Abcam), or anti-F4/80 (18705-I-AP, Proteintech) overnight at 4°C, followed by incubation with secondary antibodies conjugated to biotin and avidin-conjugated to peroxidase using VECTASTAIN ABC HRP kit (Vector Laboratories). Immunoreactivity was visualized by the peroxidase substrate DAB kit (Vector laboratories) followed by hematoxylin counterstaining. Images were acquired with an optical microscope (CX31; Olympus, Hamburg, Germany). Positively stained cells or relative staining intensity were measured in three random visual fields per section, three sections per mouse, and three or four mice per group.
For immunofluorescence analysis, sections were blocked with 10% goat serum at room temperature and incubated with primary antibody against osterix (Osx) (ab22552, Abcam) overnight at 4°C, the secondary antibody conjugated with Alexa Fluor 594-conjugated goat anti-Rabbit IgG(H+L) (SA00006-4, Proteintech) was used to visualize signals and the sections were subsequently stained with 4′,6-diamidino-2-phenylindole (DAPI, Vector Laboratories, Inc). Images were acquired with a Leica DMI6000B fluorescence microscope (Solms, Germany). The numbers of senescent cells were determined by three random fields per section, three sections per mouse and four mice per group.
Quantitative real-time polymerase chain reaction
To obtain RNA, femoral metaphyses bone was homogenized in TRIzol Reagent (Invitrogen, Carlsbad, CA, U.S.A.), and total RNA was isolated according to the manufacturer’s instructions. RNA was reverse-transcribed using PrimeScriptTM RT Master Mix (Takara Biotechnology, Japan), and cDNA was synthesized from 200ng of total RNA. Next, cDNA samples were subjected to quantitative real-time polymerase chain reaction (qPCR) analysis using FastStart Universal SYBR Premix ExTaqTM II (Takara Biotechnology, Japan) on an ABI PRISM® 7900HT System (Applied Biosystems, Foster City, U.S.A.). The cycle threshold (Ct) values were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and relative changes were calculated using the 2−ΔΔCT method. Primer sequences used for qPCR were as follows: osteocalcin (Ocn): forward, 5′-AAGCAGGAGGGCAATAAGGT-3′, and reverse, 5′-CAAGCAGGGTTAAGCTCACA-3′; runt-related transcription factor 2 (Runx2): forward, 5′-CGACAGTCCCAACTTCCTGT-3′, and reverse, 5′-CGGTAACCACAGTCCCATCT-3′; receptor activator for nuclear factor-κB ligand (RANKL): forward, 5′-AGGCTCATGGTTGGATGTG-3′, and reverse, 5′-GAGGACAGAGTGACTTTATGGG-3′, GAPDH: forward, 5′-TGTCGTGGAGTCTACTGGTG-3′, and reverse, 5′-GCATTGCTGAACAATCTTGAG-3′.
White blood cell count
For white blood cell counting in mice, 0.02 ml blood sample was collected from each mouse by retro-orbital puncture using a pipette, and diluted in 0.38 ml white cell lysis buffer (DaMao Chemical Reagent Factory, China). Then the white blood cells were suspended and dropped into the counting pool of hemocytometer, and counted 2–3 min later after the white blood cells sank. Since the white blood cells are round, the pulp is translucent and the nucleus is purple-black, slightly refracted under low magnification, these features were used for distinguishing white blood cells from impurities.
Expression profile data and processing
In order to identify the inflammatory factors involved in bone loss induced by S. aureus infection, transcriptome profiles of GSE30119 contributed by Banchereau et al.  were obtained from the National Centre of Biotechnology Information (NCBI) Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/), a genomics data repository publically available. The dataset was developed from whole blood samples of 44 healthy controls and 57 osteomyelitis patients with S. aureus infection. Total RNA was extracted from whole blood for gene expression microarrays. The TXT files were downloaded based on GPL6974 Illumina Human HT-12 V3.0 expression beadchip (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GPL6947) platform. Probe coordinates were converted from identification numbers (IDs) into gene symbols using the LiftOver tool from the UCSC Genome Browser (http://genome.ucsc.edu/). The expression levels of inflammatory factors were selected for further analysis.
Seventy-six patients with S. aureus-infected osteomyelitis at an average age of 37.4 ± 1.9 years old were recruited in the present study. The diagnosis of osteomyelitis was determined by the criteria created by Lew and Waldvogel . The infected bone from the patients involved one of the following: including humerus, radius, ulna, hand, femur, tibia, fibula, patella, and foot but not pelvis, clavicle or scapula (n=76). Blood samples were collected, and centrifuged for 10 min at 3000 rpm to separate the sera. Control serum samples were collected from patients suffering from fracture and internal fixation without infection (n=10). Serum samples were stored at −80°C until test. The study protocol was reviewed and approved by Ethical Review Board of Nanfang Hospital, Southern Medical University. Informed consent was obtained from all subjects.
Bone marrow extraction
To measure the level of G-CSF in bone marrow, the control mice and S. aureus-infected mice were killed at weeks 1 and 5 and the tibias were harvested for bone marrow extraction. Briefly, the left tibia of each mouse was exposed by cutting two ends of the tibia, and centrifuged for 20 s at 16100×g at room temperature in nested microcentrifuge tubes where the inner 0.2 ml tube had holes created with a 26-gauge needle to allow marrow to travel into the outer 1.5 ml tube. Each bone marrow sample was collected in 40 μl PBS containing 1000 U/ml heparin and centrifuged at 860×g for 10 min to obtain the supernatant, which was stored at –80°C until analysis.
Enzyme-linked immunosorbent assay
The levels of G-CSF in the supernatant of mice bone marrow and in the serum were measured using a Quantikine® ELISA Mouse G-CSF Immunoassay (R&D Systems, U.S.A.). The G-CSF levels in the serum from the osteomyelitis patients were detected using a G-CSF Human ELISA Kit (Abcam, U.S.A.) according to the manufacturer’s protocol. The optical density values were measured in a microplate reader (SpectraMax M5, Molecular Devices, U.S.A.). Wavelength correction was set to 450 and 570 nm for human and mouse G-CSF immunoassay, respectively. The level of each sample was calculated according to the standard curve.
Normality distribution of data was assessed with the Kolmogorov–Smirnov test. Continuous variables were expressed as the mean, median, standard deviations (SD), and inter-quartile range (IQR). For normally distributed data, Student’s t-test, one-way analysis of variance (ANOVA), Kruskal–Wallis test or Mann–Whitney U-test was used. All statistical analyses were performed in SPSS 17.0 software (SPSS Inc, Chicago, Illinois, U.S.A.). A P value less than 0.05 was defined as statistical significance.
S. aureus infection induces bone loss in mouse
To evaluate the effects of S. aureus infection on bone metabolism, microcomputed tomography (microCT) was performed at weeks 1 and 5 after S. aureus infection. Results showed that S. aureus infection induced bone loss in mice, and trabecular bone mass was considerably decreased after 1 week of infection in the proximal metaphysis of tibia in the infected mice and progressively lost after 5 weeks of infection (Figure 1A). Quantitative analyses of structural parameters confirmed the above observations. As shown in Figure 1B–E, compared with the tibias of control mice, the trabecular bone volume fraction (BV/TV) in the tibias of S. aureus-infected mice decreased by 20.1% and 58.7% at weeks 1 and 5, respectively. The significant decrement in BV/TV was mainly associated with the reduced trabecular number (Tb. N, 2.4 ± 0.1 mm−1 in S. aureus-infected mice vs 3.1 ± 0.07 mm−1 in control group) at week 1 after infection, and with significant loss of Tb. N (1.6 ± 0.08 mm−1 in S. aureus-infected mice vs 2.8 ± 0.13 mm−1 in control group) and reduced trabecular thickness (Tb. Th, 0.054 ± 0.001 mm in S. aureus-infected mice vs 0.065 ± 0.002 mm in control group) at week 5 after infection. Meanwhile, trabecular separation (Tb. Sp) was increased by 0.03 and 0.04 mm in the tibias of infected mice at weeks 1 and 5 after infection, respectively.
S. aureus infection induces bone loss in mouse
S. aureus infection reduces bone formation
To understand the cellular mechanism by which S. aureus induced bone loss at weeks 2 and 5 after infection, immunohistochemical staining was performed to detect the mature osteoblastic marker osteocalcin (Ocn). As shown in Figure 2A,B, compared with control mice, the numbers of Ocn-positive stained cells were significantly decreased at week 2 in the tibia of S. aureus infected mice and progressively decreased after 5 weeks of infection. Interestingly, TRAP staining revealed significant increase in the number of osteoclasts at week 2 after infection compared with control mice, whereas there was no change in the number of osteoclasts at week 5 (Figure 2C,D). Moreover, qPCR results (Figure 2E) indicated that S. aureus infection caused significant reduction in the mRNA expression of Runx2 and Ocn compared with control group. Additionally, the mRNA expression of RANKL was significantly increased in the tibia of S. aureus infected mice compared with the control ones at week 2, whereas no significant difference in the mRNA expression of RANKL was observed at week 5 after infection, consistent with the result of TRAP staining. These data suggest that bone formation rather than bone resorption is critical in mediating the bone loss induced by S. aureus infection in chronic stage of infection.
S. aureus infection reduces bone formation
Screening of inflammatory factors related to
S. aureus infection
To examine the severity of infection after S. aureus injection, peripheral blood was collected weekly for white blood cell count. Results showed a significantly higher number of white blood cells in the infection group after 1 week of infection which continued to 3 weeks after infection, indicating the continuous inflammation in mice (Figure 3A).
Screening of inflammatory factors related to
S. aureus infection
During infection, inflammatory cytokines affect osteoblastic differentiation and activity, thereby disturbing the bone formation. Next, transcriptional dataset GSE30119 was downloaded and analyzed to screen the changes of inflammatory factors in patients with S. aureus infection. As shown in Figure 3B, the mRNA expression of several inflammatory factors in human peripheral blood cells was altered by S. aureus infection. The expression of G-CSF was significantly up-regulated in peripheral blood cells from the osteomyelitis patients infected with S. aureus (P=0.021) while the mRNA expression of IL-4 and IL-7 was dramatically down-regulated in the osteomyelitis patients. Since IL-4 and IL-7 have been reported as suppressors of osteoclast formation and bone formation, respectively [30,31], the down-regulated expression of IL-4 and IL-7 might not have been associated with the bone loss induced by S. aureus infections.
S. aureus infection increases the level of local and systematic G-CSF
To confirm whether the level of G-CSF might be regulated by S. aureus infection, enzyme-linked immunosorbent assay (ELISA) was conducted to detect the level of G-CSF in the serum from patient with osteomyelitis induced by S. aureus infection. As shown in Figure 4A, the control patients had a low level of G-CSF (2.65 ± 0.084 pg/ml) but the patient with osteomyelitis a 4.5-fold higher increase in the level of G-CSF (11.93 ± 1.075, P=0.001). To further explore the relationship between S. aureus infection and change in G-CSF level, we obtained serum and bone marrow extractions from control mice and mice with S. aureus infection at weeks 1, 2, and 5 after injection. Consistent with the aforementioned results, we found that S. aureus infection resulted in remarkable 205.95, 334.7, and 155.78 pg/ml increments in the G-CSF level of serum at three time points (Figure 4B), respectively. Interestingly, we observed significant 409.5 and 200 pg/ml increments of G-CSF level also in bone marrow at weeks 1 and 5 after infection, respectively (Figure 4C).
S. aureus infection increases the level of local and systematic G-CSF
As macrophages are one of the main cell types which express G-CSF in bone marrow . To further explore the cellular mechanisms by which S. aureus infection might up-regulate the level of G-CSF, we performed immunohistochemistry to detect the changes in macrophages at week 5 after infection. As shown in Figure 4D,E, S. aureus infection greatly increased the number of F4/80-positive stained macrophages, and quantification of F4/80-positive cells confirmed the above result. These data strongly suggest that S. aureus infection might stimulate the formation of macrophages and activate the production of G-CSF in bone marrow.
Blocking G-CSF rescues bone loss induced by
S. aureus infection
It was reported that excessive G-CSF had detrimental effects on bone mass [23–25]. As we observed that S. aureus infection induced time-dependent bone loss and a continuous high level of G-CSF in bone marrow, we reason that bone loss induced by S. aureus infection may be due to the elevated level of G-CSF. In contrast with significant bone loss in S. aureus infected mice, the mice treated with G-CSF neutralizing antibody showed a significant improvement in the bone mass of trabecular area (Figure 5A). As shown in Figure 5B, compared with control mice, S. aureus infection led to a significant reduction in BV/TV (P=0.026). Compared with simple S. aureus infection, additional G-CSF neutralizing antibody treatment considerably rescued the bone loss induced by S. aureus infection, and this improvement in bone mass was closely correlated to the increased Tb. N (2.8 ± 0.06 mm−1 for S. aureus infection plus G-CSF neutralization treatment vs 2.5 ± 0.03 mm−1 for merely S. aureus infection) and increased Tb. Th (0.063 ± 0.001 mm for S. aureus infection plus G-CSF neutralization treatment vs 0.055 ± 0.001 mm for merely S. aureus infection).
Blocking G-CSF rescues bone loss induced by
S. aureus infection
As loss of functional osteoblasts is one of the characteristics of bone loss induced by S. aureus infection, we next performed immunohistochemistry to test the effect of G-CSF neutralizing antibody on osteoblast damage induced by S. aureus infection. As shown in Figure 5C,D, S. aureus significantly reduced the number of Ocn-positive stained cells in the tibia while neutralizing G-CSF treatment greatly increased the number of Ocn-positive cells by 8 cells/mm2. These data indicate that G-CSF neutralization treatment rescued the detrimental effect of S. aureus infection on osteoblasts. As blocking G-CSF, a potent stimulus for producing granulocytes, might suppress the production of granulocytes , we tested the effect of neutralizing G-CSF on the number of white blood cells in mice. Results showed that neutralizing G-CSF greatly blocked the increment of white blood cells 2 weeks after S. aureus infection (Figure 5E).
It is known that bone formation requires a constant supply of osteoblasts from osteoprogenitors. Therefore, we performed immunofluorescence to detect the changes in Osx, an osteoprogenitors marker. Results showed that S. aureus infection significantly reduced Osx-positive cells and neutralizing G-CSF treatment rescued the above effects caused by S. aureus infection (Figure 5F,G).
Osteomyelitis usually manifests an inflammatory reaction accompanied by bone erosion . It usually occurs as a complication of open fracture, internal fixation surgery, diabetic foot, and hematogenous bone infection [1,4]. S. aureus infection in bone may cause damages to bone formation which is closely associated with fracture nonunion. However, little is known about the mechanisms by which S. aureus infection induces delay in bone formation. Here we found time-dependent elevation of G-CSF in the serum and bone marrow from the mice infected by S. aureus, and elevated level of G-CSF in the serum from the patients with chronic osteomyelitis. Furthermore, we demonstrated that blocking G-CSF was effective in suppressing the bone loss induced by S. aureus infection in mice, highlighting the potential of G-CSF neutralizing antibody in the treatment of damaged bone formation induced by infection.
Long-term state of inflammation is closely linked to excessive bone loss [34,35]. It is known that S. aureus stimulates the expression of a variety of cytokines, including IL-1β, IL-6, TNF-α, IL-11, leukemia inhibitory factor and oncostatin M, which are potent stimulators of osteoclastogenesis and bone loss [17–19]. Although neutralizing antibodies against those cytokines could not abolish the effect of S. aureus on increased bone resorption , we found that G-CSF neutralizing antibody partly rescued the bone loss induced by S. aureus infection in mice, suggesting a critical role of G-CSF in mediating bone loss induced by S. aureus infection. In the present study, G-CSF neutralizing antibodies were applied before S. aureus infection in mice and resulted in the alleviated detrimental effects induced by S. aureus infection on bone formation. This suggests a potential therapeutic strategy of elective surgery for patient with risk of infection. For more clinical approaches, it would be interesting to evaluate whether G-CSF neutralization treatment after S. aureus infection still prevent the bone loss effectively.
Infection may induce a substantial elevation in serum G-CSF in patients and mice infection models [36,37]. We also found dramatically elevated G-CSF levels in the serum and bone marrow from mice infected by S. aureus at an acute stage of infection, which decreased over time at weeks 1 and 2 after infection. Notably, the extent of increased level of serum G-CSF was reported to be coupled with the number of bacteria burden in infected organs . Furthermore, with the infection progressing from an acute stage to a chronic stage, bacterial burden reduced and stabilized infection from week 1 to week 5 [17,39,40]. It is likely that decrease in serum G-CSF level observed in our study might be due to reduced bacteria burden in a chronic stage of infection. Therefore, this may suggest that the dose of G-CSF neutralizing antibody should be optimized depending on the stage and severity of infection.
Intriguingly, it was reported that G-CSF receptor (G-CSFR) was not expressed in osteoblastic cells  but in neutrophils, monocytes, and hematopoietic stem cells [41,42]. Condition medium from the G-CSF treated bone marrow nucleated cells decreased the number and activity of osteoblasts and induced osteoblast apoptosis [24,43]. Thus, it is likely that G-CSF might suppress osteoblasts indirectly. Studies have shown that G-CSFR signaling in the monocytic lineage cells and neutrophils are all associated with reduction of osteoblasts [25,41], which can be attenuated by neutrophil-specific anti-Ly6G antibody or by blockade of neutrophil reactive oxygen species (ROSs) [25,45]. Therefore, G-CSFR signaling in neutrophils might mediate the detrimental effect of G-CSF on osteoblasts. After binding with G-CSFR, G-CSF can activate STAT family (STAT 1, 3, and 5), the major cytokine-response proteins that directly regulate gene expression . A recent study showed that over-activation of G-CSFR/STAT3 pathway enhanced neutrophil activity with excessive production of ROSs . Thus, the decreased number and function of osteoblasts during infection might be caused, in part, by ROSs released from the neutrophils evoked by G-CSF. However, the pathway leading from bone marrow neutrophils activated by G-CSF to osteoblast damage remains to be confirmed.
It was reported that osteoclastogenesis and subsequent bone loss increased significantly at an early stage of infection [46,47] induced by robust early inflammatory responses to S. aureus infection in bone . Our findings are consistent with these studies in that formation of osteoclasts increased significantly at week 2 after infection. In addition, we found the number of osteoclasts unaltered in a chronic stage of S. aureus infection. This might be likely due to the falling levels of inflammatory cytokines [39,47] because of the reduced and stabilized bacterial burden with an acute infection progressing to a chronic infection [17,39,40].
In conclusions, our research suggests that G-CSF may be one of the key inflammatory factors that mediate the bone loss in S. aureus-induced osteomyelitis though the underlying mechanism is no known. The present study provides substantial evidence to conclude that therapeutic agents targeting at G-CSF might be considered as a novel treatment of bone damage in osteomyelitis.
Bone loss in S. aureus osteomyelitis poses a serious challenge to orthopedic treatment. The present study aimed to elucidate how S. aureus infection in bone might induce bone loss.
We found that S. aureus infection induced progressive bone loss in mice, with significantly increased level of G-CSF in bone marrow and serum from infection mice. Interestingly, G-CSF neutralizing antibody treatment significantly rescued the bone loss after S. aureus infection. Importantly, we found that G-CSF level was significantly up-regulated in the serum from osteomyelitis patients infected by S. aureus.
S. aureus infection might suppress the function of osteoblastic cells and induce progressive bone loss by up-regulating the level G-CSF, suggesting a therapeutic potential for G-CSF neutralization in combating bone loss in S. aureus osteomyelitis.
The authors thank Allen P Liang for English proofreading of this manuscript.
YH and XZ designed the experiments and analyzed the data; YH and HQ performed the experiments and analyzed the data; NJ, GL, HW, and LB performed the experiments; YH and XZ wrote the manuscript; BY and XZ supervised the experiments, revised and approved the manuscript.
The authors declare that there are no competing interests associated with the manuscript.
This work was supported by the National Natural Science Foundation of China [grant numbers 81772366, 81572165].
bone mesenchymal stromal cell
colony forming units
enzyme-linked immunosorbent assay
granulocyte colony-stimulating factor
receptor activator of NFκB ligand
Runt-related transcription factor 2
tartrate-resistant acid phosphatase
tryptic soy broth
These authors contributed equally to this work.