Copper-containing stainless steel was designed and fabricated by the addition of the appropriate amount of copper to medical stainless steel. The experimental materials used in this study included a previously reported 316L−5Cu stainless steel with a nominal chemical composition (wt%) of Cr 19, Ni 13, Mo 3.5, Cu 4.5, and Fe in balance, which was obtained by vacuum induction melting. Purchased medical-grade 316L stainless steel was used for comparison (Ren et al., 2012, 2015). Samples were cut from forged bars into disks of 1 mm thickness and small disks of 5 mm diameter and 1 mm thickness for use in the in vitro experiments. Intramedullary nails, which were used to create tibia injury in C57 mice in vivo, were processed by cold-drawing 316L−5Cu stainless steel to 0.45 mm diameter and then cut into 9-mm lengths. All the experimental steels with single austenitic structures were solution treated at 1,050°C for 0.5 h, followed by water quenching, and then aged at 700°C for 6 h to achieve Cu-rich precipitation in the 316L−5Cu stainless steel; this was expected to promote the release of Cu ions from the surface of the steel. All the samples were ground with SiC sand papers up to grade 2000#, soaked in absolute ethanol, cleaned with ultrasonic and deionized water, and finally sterilized at 121°C prior to use in experiments.

Each steel sample was placed in one well of a 24-well culture plate. Sugar-rich Dulbecco's modified Eagle's medium (DMEM) was dropped onto the samples (1 ml medium per 3 cm² surface area of experimental material), followed by incubation at 4°C for 72 h. The media cocultured with the steel samples were extracted and collected.

RAW264.7 cells were purchased from Cyagen Biosciences (China). High-glucose DMEM (Life Technologies, USA) containing 1% (v/v) penicillin/streptomycin (Thermo, USA) and 10% fetal bovine serum (Thermo, USA) was used for culture of RAW cells. RAW cells were incubated at 37°C with 5% CO₂. Additional drugs, namely, Mito-TEMPO (a mitochondria-targeted ROS antagonist, 50 nM, Sigma), were added to the culture medium for 60 min before polarization.

Polarization was started by changing the culture medium for complete medium, 316L−5Cu extracted medium, or 316L extracted medium, supplemented with the following cytokines: 40 ng/ml recombinant murine interleukin (IL)-4 (PeproTech, 214-14) and 20 ng/ml recombinant murine IL-13 (PeproTech, 210-13) for M2a; and 40 ng/ml recombinant murine IL-10 (PeproTech, 210-10) for M2c. After 48 h of polarization, cells and culture supernatants were collected for further experiments.

Total protein was extracted from RAW 264.7 cells using the same procedures as described in detail elsewhere (Santa et al., 2016). Briefly, raw cells were lysed using radioimmunoprecipitation assay (RIPA) lysis buffer (Thermo, 89901) containing proteinase inhibitors and phosphatase inhibitors (KeyGEN BioTECH, KGP250). The protein concentrations were determined using a bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific, 23,225). Subsequently, the phenol–ethanol supernatant was mixed with isopropanol to isolate proteins. Equal amounts of protein and supernatants were separated by 12.5 or 15% sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred to 0.45 or 0.20 μm-pore polyvinylidene fluoride membranes (Merck Millipore, IPVH00010 or ISEQ00005). Membranes were incubated overnight with antibodies against CD163 (Abcam, ab182422), CD206 (Abcam, ab64693), PDGF-BB (Santa Cruz, sc365805), MMP9 (Abcam, ab38898), and β-actin (Cell Signaling echnology, 2118). The membranes were then incubated with peroxidase-conjugated goat antirabbit immunoglobulin G (IgG) (h + l) secondary antibody (1:5,000) for 1 h. Protein signals were detected using an enhanced chemiluminescence kit (Cat. WBKLS0500, Millipore), and Western blot bands were examined and analyzed with a chemiluminescence instrument (Guangzhou Ewell Bio Technology Co. Ltd., China).

Male C57BL/6 mice (10–12 weeks old) were housed at a controlled temperature (22 ± 1°C) with a light–dark cycle (7:00 am to 7:00 pm) and allowed food and water ad libitum. They were randomly divided into two evenly sized groups: a control (316L) group and a 316L−5Cu group. Mice were anesthetized, and hair was removed from the left hind limb. An incision was made on the skin over the medial aspect of the proximal tibia. Soft tissue was cleared from the distal end of the tibial crest, and a hole (0.8 mm in diameter) that penetrated through both the medial cortices and the intervening medulla was created in the bone using a 21-gauge needle. A stainless steel pin (316L or 316L−5Cu) with length of 8 mm and diameter of 0.45 mm was inserted into the tibial medullary cavity from the proximal tibia following skin closure.

The study design and procedures were entirely in accordance with the US National Research Council (2011). This study was approved by the Animal Ethics Committee of Nanfang Hospital, Southern Medical University.

After being fixed in 10% formalin solution for 12 h at 4°C, microcomputed tomography (μCT) scanning was performed (Scanco Medical, AG, Switzerland) with X-ray energy of 55 kvp and a current of 145 mA, voxel size of 9 μm, and an integration time of 400 ms. The scanned images were reconstructed with NRecon (v1.6), and the data were analyzed using CTAn (v1.9) and three-dimensional model visualization software (μCTVol v2.0). A sequence of images within the bone defect were chosen for analysis. Three-dimensional histomorphometric analysis was performed using longitudinal images of the tibia. The region of interest was set as the cylindrical region bordered by the defect edge. The three-dimensional structural parameter of bone volume to tissue volume ratio was analyzed.

Mice (n = 12 per group) were sacrificed, and tibia specimens were harvested at days 7, 14, and 21 postinjury for observation of the histological and histochemical alterations after injury. After being fixed in 10% formalin solution for 12 h at 4°C, the specimens were decalcified in 10% ethylenediaminetetraacetic acid (EDTA) (Sigma) solution for 4.5 days. Two incisions, parallel with the major axis of the drilling hole, were made on the ventral and dorsal sides of the tibia, respectively. At the crevice of the incisions, the bone tissue surrounding the implant was carefully divided into two parts and forced apart from the implant surface, minimizing the damage to the interface tissues.

Samples were immersed in 30% (w/v) sucrose in 0.1 M phosphate buffer (pH 7.3) overnight at 4°C, embedded in paraffin (n = 6) and optimal cutting temperature compound (n = 6, Sakura Finetek), and then cut into 6 μm thick sections longitudinally for hematoxylin and eosin (H&E) and immunohistochemical staining, and 10 μm thick sections in a cryostat (Leica CM1800; Heidelberg, Germany) for immunofluorescence staining.

For H&E staining, some of the sections obtained previously were preheated in an air oven at 60°C, followed by deparaffinization and rehydration in xylene and ethanol solutions (reducing concentrations of 100 to 70%). The sections were then successively soaked in H&E dyes. After dehydration with graded alcohol, histological observation was performed using a microscope (BX63, Olympus, Tokyo, Japan), and the fracture healing process was evaluated.

Immunohistochemical staining was performed on the deparaffinized and rehydrated sections as described previously, with specific primary antibody (rabbit anti-CD206; ab64693, Abcam 1:200, UK). All the sections were counterstained using Mayer's hematoxylin (Sigma-Aldrich) and mounted using a permanent mounting medium (Thermo Fisher Scientific, Waltham, MA, USA). Quantification of positive cell numbers within the injury site was carried out on 4 μm serial sections stained for CD206 expression. Three representative images (×40 magnification) were taken within the intramedullar injury zone for each sample at two sectional depths at least 50 μm apart. The number of CD206+ cells was quantified by counting each positively stained cell in each field of view.

Frozen sections embedded in optimal cutting temperature (OCT) were processed for immunofluorescence staining of Emcn, CD31,F4/80, CD163, CD206, and PDGF-BB. Sections were incubated in blocking buffer (10% goat serum in phosphate-buffered saline; PBS) for 1 h at room temperature and incubated with primary antibodies overnight at 4°C. The following primary antibodies were used for immunostaining: rat antiendomucin (Emcn, SC-65495, Santa Cruz, 1:50, US), rabbit anti-CD31 (ab222783, Abcam, 1:50, UK), rat anti-F4/80 (71299,Cell Signaling Technology, 1:500, USA), rabbit anti-CD206 (ab64693, Abcam 1:200, UK), rabbit anti-CD163 (ab182422, Abcam, 1:200, UK), and mouse anti-PDGF-BB (sc-365805, Santa Cruz, 1:50, USA). Sections were washed three times in PBS and then incubated with secondary antibodies at room temperature for 1 h. The following secondary antibodies were used for immunostaining: Alexa Fluor 594-conjugated goat antirabbit IgG (8889, Cell Signaling Technology, USA), Alexa Fluor 488-conjugated goat antirat IgG (4416, Cell Signaling Technology, USA), and 594-conjugated goat antimouse (HA1112, HuaBio, China). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (S2110, Solarbio, China). All the sections were observed under a BX63 microscope (Olympus, Tokyo, Japan).

Quantitative analysis was carried out using three representative images (×40 magnification) taken within the intramedullar injury zone for each specimen at three sectional depths at least 50 μm apart. The number of positive cells was quantified by counting each positively stained cell in each field of view.

By combining both Mitotracker Green (MTG) and Mitotracker Orange (MTO) (Invitrogen, CA, USA), the relative mitochondrial potential with its mitochondrial mass as baseline can be obtained (Agnello et al., 2008). Polarized and pretreated RAW264.7 cells were seeded in six-well chamber slides and grown to 50–70% confluence. After washing twice with warm 1× PBS, cells were stained with 100 μM MTG, 500 μM MTO, and Hoechst 33324 (HuaBio, China) in DPBS + Ca/Mg/glucose to maintain their normal metabolic state. Cells were then incubated for 45 min at 37°C under standard culture conditions. The u-slide was then transferred to a confocal microscopic station with a 37°C heated chamber supplied with 5% CO₂ for live cell imaging. Tile scan images were captured randomly using the same TCS-SP2 confocal microscopy system (Leica Microsystem, Nussloch GmbH, Germany) with sequential detection for both stains. The fluorescence intensities of MTO (excitation/emission, 554/576 nm) and MTG (excitation/emission, 490/516 nm) were measured. For five samples from each group, a total of 20 different microscopic fields per sample, containing comparable numbers of cells, were integrated to obtain fluorescence mean pixel intensity values. The ratio of MTO/MTG intensities was then calculated and denoted the mean index of MTO/MTG, reflecting the relative mitochondrial potential.

To determine mitochondrial superoxide production, polarized and pretreated RAW 264.7 cells were incubated for 30 min with 100 μM MitoSOX dye (Thermo Fisher Scientific, M36008) and Hoechst 33324 diluted in cell imaging solution (10 mM HEPES, 1 mg/ml bovine serum albumin, 1 mg/ml glucose, 1 mM MgCl₂, and 1.8 mM CaCl₂ in PBS) at 37°C in a humidified atmosphere containing 5% CO₂. Then, cells were washed with PBS. The fluorescence was read using a SpectraMax i3x microplate reader at 510/580 nm, and the cell number in each well was determined using a SpectraMax MiniMax 300 imaging cytometer. Fluorescence values were first normalized to the cell number in each well; then, all the conditions for each genotype were normalized to the non-treated control group.

Statistically significant differences were determined using unpaired t-tests with one- or two-tailed distributions using PRISM 7 (GraphPad software, La Jolla, CA, USA). A value of p < 0.05 was deemed statistically significant. In all cases, data are represented as mean ± standard error of the mean (SEM).

M2a and M2c, two subtypes of macrophages, are known to be closely associated with angiogenesis, linked by IL4, IL13, and IL10 (Wu et al., 2013; Spiller et al., 2014). To identify the subtype with the major role under CCM stimulation, Western blotting was performed on M2 macrophages in vitro. Upon copper stimulation, M2a cells with high expression of CD206 and M2c cells with higher CD163 expression were detected. The expression levels of PDGF-BB and CD206 were higher in M2a macrophages stimulated by 316L−5Cu extract (Figure 1A), and CD206 and CD163 were identified in this distinct type of macrophage (Figures 1A,B). Notably, the expression levels of MMP9 and CD163 showed no difference between the 316L and 316L−5Cu groups (Figure 1B), indicating that M2a macrophages predominantly promoted the expression of PDGF-BB under the stimulation of 316L–Cu extract.

To determine the role of M2a macrophages in bone regeneration, a tibia drilling hole injury model in C57BL/6 mice was established, and the closure degree of the drilling hole was measured and assessed by μCT and histomorphological procedures. The model was established by inserting 316L−5Cu/316L pins into the bone marrow cavity of the injured tibia as intramedullary nails. Formation of new cortical bone in the drilling hole was faster in the 316L−5Cu group than in the 316L group at days 14 and 21 (Figure 2A). The bone mineral content within the drill hole was almost 2-fold higher in the 316L−5Cu group compared with the 316L group at day 21 (Figure 2B). Using H&E staining, no difference was observed in the callus between the two groups at day 7, whereas the formation of cortical bone callus in the 316L−5Cu group was accelerated at the later stage. New and integral cortical bones were observed in the 316L−5Cu group, covering the top of the drilling hole on day 21 (Figure 2A).

Type H vessels have been reported to participate in bone fracture healing (Stefanowski et al., 2019). In the drilling hole injury model, expression levels of Emcn and CD31 were examined; these are markers of generation of type H vessels in immunofluorescence-stained tissue slices. Emcn+CD31+ immunofluorescence staining was observed on the bone callus; this was more intense from days 7 to 21 in the 316L−5Cu group, indicating the generation of type H vessels (Figures 3A,B). These signals were located particularly beneath the new cortical bone at day 21. These results show that 316L−5Cu significantly accelerated the formation of callus, accompanied by promotion of the generation of type H vessels.

Macrophages have been shown to infiltrate into the callus and periosteum during the healing of bone fractures; in particular, CD206+ M2 macrophages were primarily assembled in the periosteum, with very few present in the fracture gap at day 14 in a bone fracture model (Stefanowski et al., 2019). According to our results, CD206+ cells in both groups were located in the callus around the newly formed bone between the hole gaps. Furthermore, a significant increase in CD206+ cells was observed in the callus of the 316L−5Cu group (Figure 4A). We further detected CD163+ cells (also known as M2c macrophages) with an association with angiogenesis by immunohistochemistry and immunofluorescence. Few CD163+ cells were found during the process of bone regeneration by immunohistochemistry (data not published), but it was notable that a few 163+ (M2c) macrophages were located in hole gaps and in the bone marrow cavity, especially near the endosteum. There was no difference between the two groups (Figure 4C). We further detected the relationship between the CD206 and F4/80 immunofluorescence signals; coexpression of CD206 and F4/80 indicates M2a macrophages. From days 7 to 21, coexpression of CD206 and F4/80 exceeded 70% of the whole CD206 signal in the 316L−5Cu group and was significantly higher than 316L at days 7 and 14 (Figure 4E). The coexpression of the above two signals was also higher in the 316L−5Cu group compared with the 316L group during the bone regeneration process (Figure 4D). CD206+ M2a macrophages were found mainly in the callus of the drilling hole in the early stage (day 7) and around the surface of the newly formed bone (day 14). In the later stage, cells were observed beneath the new cortical bone (day 21), indicating a close relationship between M2a macrophages and newly formed cortical bone (Figure 4B).

PDGF-BB, which is secreted by M2a macrophages and TRAP– preosteoclasts, is critical for enhancing the formation of type H vessels and bone during bone modeling and remodeling (Xu et al., 2018). We examined PDGF-BB in callus (Figure 5A) and found that it showed significantly higher expression in the 316–5Cu group than in the 316L group at days 14 and 21 and that it was located in the F4/80+ macrophages beneath the new cortical bone (Figures 5A,B). At day 7, no difference in PDGF-BB expression in callus was found between the 316L−5Cu and 316L groups. These results indicate that the M2a macrophages secreted PDGF-BB in the revolution process of bone regeneration, which was promoted by 316–5Cu.

Low-concentration copper preparations have been reported to have promising immunomodulatory potential and to contribute to oxidative stress in macrophages (Videla et al., 2003; Steinborn et al., 2017; Zhao and Zhao, 2019). Mitochondrial-derived reactive oxygen species (mtROS) are believed to derive mainly from NADH dehydrogenase when there is a high NADH/NAD+ ratio in the mitochondrial matrix. The inhibitory effect of copper on NADH dehydrogenase has been demonstrated in adenine–copper complexes and chelating 2-valent copper complexes (Hammud et al., 2008; Roy et al., 2015). To determine the role of CCM in inducing PDGF-BB expression by M2a macrophages, the oxidative stress caused by copper in macrophages was detected. According to our results, inhibition of NADH dehydrogenase activity occurred in M2a macrophages in the 316L−5Cu group, compared with 316L and control group (Figure 6D). In addition, levels of mitochondrial potential (Figures 6A,E) and mtROS (Figures 6B,F) were significantly elevated in IL4- and IL13-induced M2a macrophages by 316L−5Cu. Furthermore, the elevation of PDGF-BB expression in M2a macrophages induced by 316L−5Cu was found to be inhibited by a selective mtROS inhibitor (Mito-TEMPO) (Figures 6C,G).