As shown in Fig. 1, the different composite raw materials were processed into particles by the banbury mixer. The composite raw materials PCL@Mg and PCL@Mg/Amp appeared gray. PCL and PCL@Amp appeared inherently white. These raw materials were printed by an FDM printer into scaffolds at 80 °C. The macro appearance size of the scaffold was basically consistent with the software design. Micro-CT images of scaffolds showed an interconnected pore structure. The above results showed that PCL mixed with Mg and Amp did not affect the whole printing process.

Composite raw material particles, scaffold printing, macro appearance and micro-CT images. (a) PCL, (b) PCL@Amp, (c) PCL@Mg and (d) PCL@Mg/Amp

As shown in Fig. 2A, Mg had a spherical structure, and the microspheres were 10—60 μm in diameter. The average diameter of the Mg microspheres was 29.38 μm. As shown in Fig. 2B, SEM images show the morphology of the surface and cross section of the scaffold. The filament diameter of the scaffold was 436 ± 7.2 μm, and the pore size of the scaffold was 420.66 ± 12.1 μm. The filaments and pores of the scaffolds were very uniform and regular. The scaffold surface of the PCL and PCL@Mg groups was relatively smooth. The surface of the PCL@Amp and PCL@Mg/Amp groups became slightly rough due to the addition of Amp powder. In the PCL@Mg and PCL@Mg/Amp groups, Mg microspheres were embedded on the surface of the scaffold filaments. SEM images of the cross section of the scaffold showed that Mg microspheres in the PCL@Mg and PCL@Mg/Amp groups were evenly dispersed inside the scaffold filaments.

A SEM images of Mg microparticles and diameter distribution. B SEM images of the surface and cross-section morphology of the scaffold. The green arrow indicates the Mg microsphere

XRD, TGA and FTIR were used to determine the composition of the scaffold. The phase composition was checked in accordance with the Joint Committee on Powder Diffraction Standards reference patterns of Mg (PDF No. 89–4894). With the addition of Mg microparticles to the PCL@Mg/Amp and PCL@Mg scaffolds, the diffraction pattern included several peaks caused by Mg microparticles (Fig. 3A). Amp showed an absorption peak in the region of 1730 – 1720 cm−1, which is caused by C = O β-lactam stretching. The peak at 1610 cm−1 belongs to C = O amide stretching [27] (Fig. 3B). TGA was used to analyse the contents (W/W) of Mg and Amp in different scaffolds. PCL in the scaffold began to decompose at 225 °C. The mass losses of PCL, PCL@Amp, PCL@Mg and PCL@Mg/Amp were 99.902%, 94.221%, and 83.482%, respectively (Fig. 3C). The PCL@Mg and PCL@Mg/Amp scaffolds showed a slow release of Mg2+ profile during the whole 21-day release period, while the scaffolds in the PCL and PCL@Amp groups did not release Mg2+. At 3 days, the cumulative release of Mg2+ in the PCL@Mg and PCL@Mg/Amp groups was 2.67 ± 0.58 mM and 2.89 ± 0.7 mM, respectively. At 21 days, the cumulative release of Mg2+ in the PCL@Mg and PCL@Mg/Amp groups was 5.75 ± 0.64 mM and 6.26 ± 0.94 mM, respectively (Fig. 3D). As shown in Fig. 4, the water contact angles of different groups of scaffolds were measured to assess the surface hydrophilicity and hydrophobicity. The contact angles of the PCL, PCL@Amp, PCL@Mg and PCL@Mg/Amp groups were 117.5 ± 4.4°, 102.2 ± 3.0°, 87.2 ± 5.2° and 83.5 ± 3.7°, respectively. PCL showed an obvious hydrophobic surface, which was consistent with previous literature [28]. With the addition of Amp and Mg, the hydrophilicity of the PCL composite scaffold surface was significantly enhanced.

Scaffold characterizations: (A) XRD, (B) TGA, (C) FTIR patterns and (D) Mg2+ release from scaffolds

Measurement of the contact angle of different scaffolds, n = 3 p < 0.05

3D printing technology can be used to personalize and prepare various macro morphologies of bone repair scaffolds. In 1983, Chuck Hull et al. invented stereo lithography appearance (SLA) 3D printing technology [29], and in 1988, S. Scott Crump invented FDM 3D printing technology [30]. FDM 3D printing technology completes the whole printing process of the scaffold by melting the polymer material at high temperature and then extruding, printing and cooling. The printing process of FDM does not require the addition of photoinitiator (LAP) included in SLA printing technology. Previous literature has shown that LAP is obviously cytotoxic, which decreases the bone repair ability of scaffolds [31]. Based on the above reasons, the bone repair scaffold prepared in this study uses FDM printing technology. Currently, polylactic acid (PLA), poly(lactide-co-glycolide) (PLGA), PCL and other polymer materials are commonly used in the preparation of bone repair scaffolds. The melting point of PCL is 58 – 64 °C, which can be used to mix some drug components with poor thermal stability. In this study, PCL scaffolds containing Mg and AMP were successfully printed at 80 °C. SEM images showed that Amp and Mg were evenly mixed in the PCL material, which could be slowly released into surrounding tissues with the degradation of PCL in vivo (Fig. 2B). Compared with surface coating technology, Amp and Mg2+ were released slowly throughout scaffold degradation and new bone regeneration25. TGA, XRD and FTIR results showed that Mg and Amp were successfully mixed into the scaffold without any change in the phase pattern and molecular structure (Fig. 3). The contact angle of the scaffold surface has an important effect on cell adhesion and proliferation15. The surface with a 65° contact angle was the most favourable for osteoblast adhesion, while both high and low contact angles were unfavourable for osteoblast adhesion. The contact angle of the PCL@Mg/Amp group was 83.5 ± 3.7°, which was the closest scaffold to 65°, and it was speculated that it was the most conducive to cell adhesion and proliferation among all scaffolds. PCL is a biodegradable polyester approved by Food and Drug Administration (FDA) in tissue engineering applications [32]. PCL degrades more slowly than PLA and PLGA in vivo [33]. The molecular weight of PCL decreased gradually within 2 years, and degraded to fragments of low molecular weight after 2 years [34].The increased hydrophilicity of the scaffold can increase the rate of water diffusion into the PCL polymer, which may also increase the degradation rate of the scaffold [34].

As shown in Fig. 5A, the cell adhesion and proliferation abilities were measured at 1 and 3 days. The amount of cell adhesion was relatively low in the PCL and PCL@Amp groups. In contrast, there was a significant increase in cell adhesion in the PCL@Mg and PCL@Mg/Amp groups. In all groups, very few dead cells were present. As shown in Fig. 5B, the proliferation ability of the PCL@Mg and PCL@Mg/Amp groups was significantly higher than that of the PCL and PCL@Amp groups at 1 and 3 days. The proliferation ability of the PCL@Amp group was slightly higher than that of the PCL group, and the same trend was slightly higher in the PCL@Mg/Amp group than in the PCL@Mg group.

A Calcium AM staining of adherent cells on scaffolds and (B) CCK-8 assay of cell proliferation, n = 3 p < 0.05

As shown in Fig. 4, the addition of AMP and Mg microspheres significantly improved the hydrophilicity of the scaffold, and the cell adhesion and proliferation results were consistent with the contact angle results. Since the contact angle of the PCL@Mg/Amp group was closest to 65°, its cell adhesion and proliferation were best [15]. Previous literature has also demonstrated that Mg2+ is involved in platelet-derived growth factor (PDGF)-stimulated MC-3T3-E1 cell adhesion and proliferation [35]. The Mg2 + concentration is also important for cell adhesion and proliferation. Jie Shena et al. found that a Mg2+ concentration of approximately 4.11 mM was most conducive to cell adhesion and proliferation of MC-3T3-E1 cells. The cumulative release concentration of Mg2+ ranged from 2–7 mM in the PLC@Mg and PCL@Mg/Amp groups during the 3–21 days release period, which were all within the appropriate concentration range that could promote cell adhesion and proliferation (Fig. 3D). PDGF is generally believed to have the ability to promote the adhesion and proliferation of many cell types. Therefore, the PCL@Mg and PCL@Mg/Amp groups with bioactive molecular Mg significantly promoted cell adhesion and proliferation.

As shown in Fig. 6A, the COL-I and Runx2 genes were selected to evaluate the osteogenic differentiation ability of scaffolds at 3 days. qPCR results showed that the mRNA expression levels of COL-I and Runx2 in PCL@Mg and PCL@Mg/Amp scaffolds containing Mg were significantly higher than those in PCL and PCL@Amp scaffolds without Mg. The COL-I expression levels in the PCL@Mg and PCL@Mg/Amp groups were 1.47 and 1.67 times higher than that in the PCL group, respectively. Runx2 expression levels were 1.63 and 1.83 times higher in the PCL@Mg and PCL@Mg/Amp groups than in the PCL group, respectively. As shown in Fig. 3D, the Mg-containing scaffold prepared in this study can slowly release Mg2+ into the surrounding solution, which is mainly attributed to the reaction of Mg with water to generate Mg2+, hydrogen and OH−36. Many previous studies have confirmed that Mg2+ can promote osteogenic differentiation through upregulation of osteogenic gene expression, alkaline phosphatase (ALP) secretion and mineralization [36,37,38,39]. Qiangsheng Dong et al. used FDM 3D printing technology to fabricate Mg-containing PCL scaffolds. qPCR results showed that the osteogenic genes osteopontin (OPN), osteocalcin (OCN), COL-I and Runx2 were significantly upregulated during 7–21 days of culture [36]. Researchers have found that Mg2+ mainly regulates osteogenic differentiation through the PI3 K/Akt signaling pathway [40]. Mineralization capacity is also an important item to evaluate the osteogenesis of in-situ bone repair scaffolds. As shown in Fig. 6B, the surface of the scaffolds of PCL and PCL@Amp groups were relatively smooth, and the effect of mineralized deposition was not obvious. In contrast, the PCL@Mg and PCL@Mg/Amp groups had a rough and uneven surface, which was attributed to the surface mineralization of the scaffold. Degradation of Mg microspheres in scaffolds will release Mg2+ and form an alkaline environment, which will promote the formation of apatite [41]. The scaffolds containing Mg microspheres showed obviously mineralization ability, which was consistent with the results of the relative mRNA expression of COL-I and Runx2 genes.

A The relative mRNA expression of COL-I and Runx2, n = 3 p < 0.05. B The mineralization ability of different scaffolds evaluated by SEM

As shown in Fig. 7, the bacteriostatic ring test against E. coli and S. aureus was used to evaluate the antibacterial ability of the scaffolds. There was no obvious bacteriostatic ring E. coli or S. aureus around the PCL and PCL@Mg scaffolds. In contrast, the Amp-containing scaffolds PCL@Amp and PCL@Mg/Amp showed obvious bacteriostatic rings against E. coli and S. aureus. The diameters of bacteriostatic rings formed by PCL@Amp and PCL@Mg/Amp on the surface of agar plates coated with E. coli were 37.0 ± 3.6 mm and 38.2 ± 2.3 mm, respectively. The diameters of bacteriostatic rings formed by PCL@Amp and PCL@Mg/Amp on the surface of agar plates coated with S. aureus were 34.3 ± 2.0 mm and 34.8 ± 2.5 mm, respectively.

Antibacterial ability of different groups of scaffolds against E. coli and S. aureus, n = 3 p < 0.05

Antibacterial agents used to inhibit open bone trauma mainly include antimicrobial peptides, silver nanoparticles (AgNPs), antibiotics, etc. Lei Chen et al. used pDA molecules to attach antimicrobial peptides (ponericin G1) to the surface of PLGA scaffolds, and ponericin G1 showed good antibacterial effects on E. coli and S. aureus [25]. The advantage of antimicrobial peptides is that it is not easy to develop antimicrobial resistance, and they can have broad-spectrum antimicrobial activity [42]. However, antimicrobial peptides are composed of amino acids, and their thermal stability is relatively poor, so they cannot be directly used in FDM 3D printing. AgNPs with diameters between 1–100 nm have been widely used in the preparation of antibacterial scaffolds due to their strong antibacterial activity [43]. Jiayi Li et al. used FDM 3D printing technology to prepare a PCL bone repair scaffold with AgNP coating on the surface, which showed a significant bacteriostatic effect on S. aureus. In vivo animal experimental results showed that scaffolds containing AgNP coatings had the best effect on repairing infected bone defects at the external tibial epicondyle of rabbits [44]. Antibiotics such as vancomycin, tobramycin, tetracyclines, and Amp have been used to treat bone defects [24]. The pharmacological mechanisms of these antibiotics have been well studied and approved by regulatory authorities, so their safety in vivo was higher than that of other antibacterial agents [45]. Antibiotics also have higher thermal stability than antimicrobial peptides, so scaffolds containing antibiotics can be prepared using FDM 3D printing technology. Amp, the most commonly used antibiotic in clinical practice, can be used to treat osteomyelitis [46]. In this study, Amp was used as an antibacterial agent to treat bone defects. The PCL scaffold can also be mixed with other types of antibiotics to treat different bacterial infections.

As shown in Fig. 8, the tibial defect samples were scanned and analysed at 8 weeks after surgery by micro-CT. In the PCL group, only a small amount of new bone was formed along the edge of the defect, and no obvious new bone was formed in the center of the defect. In the PCL@Amp group, a small number of regular porous structures of new bone were observed forming in the center of the defect. In the PCL@Mg and PCL@Mg/Amp groups, the new bone almost completely filled the entire site of the bone defect, and the new bone showed a regular porous structure. In X-ray images, new bone in the PCL@Mg and PCL@Mg/Amp groups can be clearly seen, which is similar to the regular porous structure of the 3D-printed scaffolds. The 3D view images of the defect further proved the repairability effect of different groups of scaffolds on the tibial defect. The BV/TV of the PCL@Mg/Amp (46.20 ± 3.58%) group was significantly higher than that of the PCL@Mg (35.47 ± 3.64%), PCL@Amp (25.82 ± 3.45%) and PCL (22.48 ± 2.15%) groups. As shown in Fig. 9, tissue slices at the tibial defect were analysed by H&E and Masson staining. In the PCL and PCL@Amp groups, a small amount of new bone was formed, and a large amount of fibrous tissue was formed in the interconnected pores of the scaffold, which was not conducive to bone growth and bone healing. In the PCL@Mg and PCL@Mg/Amp groups, the scaffolds were tightly surrounded by a large amount of new bone. The thickness of new bone in the PCL@Mg/Amp group was significantly higher than that in the other groups.

Evaluation of tibial defect repair in rats by micro-CT. A Macro CT view and X-ray images, (B) 3D view of defect and (C) BV/TV assay, n = 3 p < 0.05)

H&E and Masson staining images of tissue slices in tibial defects 8 weeks after surgery. S indicates scaffold, F indicates fibrous tissue, and NB indicates new bone

Regeneration of bone defects can be delayed and inhibited by bacterial infection at the defect site. In the clinic, infected bone defects are usually treated by a systemic or local administration of antibiotics to control bacterial infections. This treatment requires additional intervention with antibiotics for the treatment of infected bone defects, which is a time-consuming and tedious procedure [47]. In this study, anti-infective and osteogenic effects were integrated to prepare scaffolds for repairing bone defects. This dual-function in situ bone repair scaffold can release antibiotics at the defect location to play an antibacterial role, and Mg in the scaffold can release Mg2+ to promote bone repair. A certain amount of hydrogen gas was also produced during the release of Mg2+. If a lot of hydrogen gas was released into the surrounding tissue it could delay bone healing. A large amount of hydrogen gas generated around the scaffold would form gas cavities, and in this study, tissue slices results showed that no gas cavities was formed around the scaffold, which indicated that hydrogen gas was generated in a small amount and fully absorbed by around tiusse [48] (Fig. 9). As shown in Fig. 7, the PCL@Mg/Amp scaffold contained antibiotics, which could reduce bacterial infection at the bone defect site and promote bone repair. Both the PCL@Mg and PCL@Mg/Amp groups contained Mg, which significantly promoted osteogenesis compared with the Mg—free PCL and PCL@Amp groups. These results indicated that the in situ scaffold of PCL@Mg/Amp had dual anti-infection and osteogenesis functions. Although there was still the problem of antibiotic resistance, many antibiotics had been approved by regulatory administration, and the safety in vivo was high, so the scaffold containing antibiotics was more easily approved for clinical use [49]. The antibiotic-containing scaffold can reduce the amount and frequency of systemic antibiotic administration in patients. Antimicrobial peptides and AgNPs demonstrate good antimicrobial activity and can treat antimicrobial resistance. However, the in vivo safety of antimicrobial peptides and AgNPs is still being evaluated, and intensive studies are needed to prove their in vivo safety [50, 51].