Specimens from a total number of 100 male and female CD-1 mice with an age of 3–4 months (young adult, n = 50) and 16–18 months (aged, n = 50) were included in the present study. Notably, male and female mice were equally distributed among the study groups to avoid gender as a bias in bone regeneration. In a previous study we have already shown that CD-1 mice with an age of 16–18 months exhibit an altered bone regeneration after fracture [14]. Herein, we now analyzed angiogenesis and inflammation as well as the trabecular architecture and bone remodeling within the callus tissue at 1, 2, 3, 4 and 5 weeks after surgery (young adult and aged mice: n = 10 per group and per time point) (Fig. 1a). The animals were bred at the Institute for Clinical & Experimental Surgery, Saarland University, and kept in groups at a regular light and dark cycle with free access to tap water and standard pellet food (Altromin, Lage, Germany).

a: Experimental setup of the present study. Femoral fracture healing of young adult (3–4 months) and aged (16–18 months) CD-1 mice was analyzed by X-ray, µCT, histology and immunohistochemical analysis. The focus of the investigation included the process of angiogenesis and inflammation at the fracture site as well as the bone mineralization, the trabecular architecture and the bone remodeling within the callus tissue. b: X-ray of a fractured mouse femur with an inserted tungsten guidewire during surgery. Scale bar: 1 mm. c: Postoperative X-ray after insertion of the intramedullary screw. Scale bar: 1 mm

For the investigation of angiogenesis, inflammation and bone remodeling within the callus tissue of young adult and aged mice a closed femoral fracture model with an intramedullary screw (RISystem AG, Landquart, Switzerland) was used [14]. The model was described previously in detail by Holstein et al. [7].

The mice were anesthetized by an intraperitoneal (i.p.) injection of ketamine (75 mg/kg body weight; Ursotamin®, Serumwerke Bernburg, Bernburg, Germany) and xylazine (15 mg/kg body weight; Rompun®, Bayer, Leverkusen, Germany). To guarantee adequate precision, all procedures were performed using an operating microscope. A medial parapatellar incision of ~ 4 mm at the right knee was created and the patella was dislocated laterally. The intramedullary canal was opened by a drill bit (diameter: 0.5 mm) at the intercondylar notch. Then, an injection needle (0.4 mm) was inserted through the greater trochanter over the intramedullary cavity. Afterwards, a tungsten guide wire (diameter: 0.2 mm) was positioned through the needle (Fig. 1b). The femur was fractured by a 3-point bending device, as described previously [15]. Subsequently, the MouseScrew was implanted over the guide wire, providing stabilization of the femoral fracture by interfragmentary compression. The guide wire was removed, the patella was reduced and the wound closure was performed by 5–0 synthetic sutures. Postoperative X-rays confirmed the correct implant position and fracture reduction (MX-20, Faxitron X-ray Corporation, Wheelin, IL, USA) (Fig. 1c). For analgesia the mice received 5 mg/kg body weight carprofen (Rimadyl™, Zoetis GmbH, Berlin, Germany) subcutaneously at the day of surgery. Additionally, tramadol-hydrochloride (Grünenthal, Aachen, Germany) was added to the drinking water (1 mg/mL) one day prior to surgery until three days after surgery.

To guarantee adequate fracture reduction and detect possible fracture or implant dislocation, lateral radiographs were performed before harvesting the femora at the end of the experiments (MX-20, Faxitron X-ray Corporation, Wheelin, IL, USA).

The femora were scanned (Skyscan 1172, Bruker, Billerica, MA, USA) at a spatial resolution of 9 μm with a standardized setup (tube voltage: 50 kV; current: 200 μA; intervals: 0.4°; exposure time: 3500 ms; filter: 0.5 mm aluminum). Images were stored in 3-dimensional arrays. Gray values were expressed as mineral content (bone mineral density (BMD)). For this, calcium hydroxyapatite (CaHA) phantom rods with known BMD values (0.250 g and 0.750 g CaHA/cm3) were employed for calibration. The region of interest (ROI) defining the novel bone was contoured manually excluding any original cortical bone. The thresholding allowed the differentiation between poorly and highly mineralized bone. The thresholds to distinguish between poorly and highly mineralized bone were based upon visual inspection of the images, qualitative comparison with histological sections and studies investigating bone repair and callus tissue by μCT [16, 17]. A BMD with more than 0.642 g/cm3, resulting in gray values of 98–255, was defined as highly mineralized bone. Poorly mineralized bone was assumed to have a BMD value between 0.410 and 0.642 g/cm3, resulting in gray values of 68–97.

The following parameters were calculated from the callus ROI for each of the femora: poorly mineralized (pm) bone volume (BV) (mm3), highly mineralized (hm) BV (mm3), pm BV fraction (pm BV/total tissue volume (TV) (%)), hm BV fraction (hm BV/(TV) (%)), trabecular thickness (mm), trabecular separation (mm), trabecular number (1/mm) and BMD (g hydroxyapatite (HA)/cm3).

After µCT analysis the femora were prepared for histological analyses. Therefore, the bones were fixed in paraformaldehyde for 24 h. Subsequently, the specimens were embedded in a 30% sucrose solution for another 24 h and then frozen at −80 °C. Longitudinal sections through the femoral axis with a thickness of 4 µm were cut by means of the Kawamotos film method [18]. To analyze the process of bone remodeling, tartrate-resistant acid phosphatase (TRAP) activity was analyzed to detect osteoclasts within the callus tissue. For this purpose, longitudinal sections of 4 μm were incubated in a mixture of 5 mg naphthol AS-MX phosphate and 11 mg fast red TR salt in 10 mL 0.2 M sodium acetate buffer (pH 5.0) for 1 h at 37 °C. Sections were counterstained with methyl green and covered with glycerin gelatin. At a 400 × magnification, TRAP-positive cells were counted within the callus tissue in a standardized manner. In 1-, 2- and 3-week specimens, one high-power field (HPF, 400 × magnification) was placed in the central region of the callus (former fracture gap), while five additional HPFs were placed at each site within the periosteal region of the callus. Due to the reduced size of the callus at 4 and 5 weeks, only three additional HPFs were placed at each site within the periosteal region of the callus (Figs. 2a and b). The number of TRAP-positive osteoclasts within each HPF was counted and the mean for each specimen was determined.

a and b: Schematic illustration of the cellular analysis of the callus tissue at 1, 2 and 3 (a) as well as 4 and 5 (b) weeks after fracture. At 1, 2 and 3 weeks after fracture a single HPF was placed in the central region of the callus (former fracture gap, orange box), while five additional HPFs (green boxes) were placed at each site within the periosteal region of the callus (a). At 4 and 5 weeks only 3 additional HPFs were placed at each site within the periosteal region of the callus due to the reduced callus size (b). Scale bars: 0.5 mm

For the investigation of angiogenesis and inflammation within the callus tissue, additional longitudinal tissue sections were cut for immunohistochemical analyses. For the immunohistochemical detection of microvessels, sections were stained with a monoclonal rat anti-mouse antibody against the endothelial cell marker CD31 (1:100; Abcam, Cambridge, UK). For detecting mature microvessels with a α-smooth muscle actin (SMA) cell layer, the sections were additionally stained with a polyclonal rabbit anti-mouse α-SMA antibody (1:100, Abcam). Cell nuclei were stained with Hoechst 33342 (2 µg/mL; Sigma-Aldrich, Taufkirchen, Germany). In 1-, 2- and 3-week specimens, one HPF (400 × magnification) was placed in a standardized manner in the central region of the callus (former fracture gap), while five additional HPFs were placed at each site within the periosteal region of the callus. Due to the reduced size of the callus at 4 and 5 weeks, only three additional HPFs were placed at each site within the periosteal region of the callus (Figs. 2a and b). The number of CD31-positive endothelial cells and the fraction of mature α-SMA-positive microvessels (%) within each HPF was counted and the mean for each specimen was determined.

For the analysis of the inflammatory response within the callus tissue, the neutrophilic granulocyte marker myeloperoxidase (MPO) and the macrophage marker CD68 were detected. For this purpose, sections were stained with a polyclonal rabbit anti-mouse antibody against MPO (1:100; Abcam) and a polyclonal rabbit anti-mouse antibody against CD68 (1:200; Abcam). The number of MPO-positive granulocytes and CD68-positive macrophages was determined in a standardized manner according to the analysis of microvessels.

All data are given as means ± standard error of the mean (SEM). After proving the assumption for normal distribution (Kolmogorov–Smirnov test) and equal variance (F-test), comparisons between the two experimental groups were performed by an unpaired Student´s t-test. For non-parametrical data, a Mann–Whitney U-test was used. Statistics were performed using the SigmaPlot 13.0 software (Systat Software GmbH, Erkrath, Germany). A p-value < 0.05 was defined to indicate significant differences.