Optimized osteogenesis of porcine bone-derived xenograft through surface coating of magnesium-doped nanohydroxyapatite

As one of the key factors influencing the outcome of guided bone regeneration, the currently used xenografts possess insufficient capability in osteogenesis. With the aim of improving the osteogenic performance of xenografts, porcine bone-derived hydroxyapatite (PHA) was prepared and subsequently coated by magnesium-doped nano hydroxyapatite (nMgHA, 10%, 20%, and 30% of Mg/Ca + Mg) through a straightforward and cost-efficient approach. The physiochemical and biological properties of nMgHA/PHAs were examined in vitro and in vivo. The inherent three-dimensional (3D) porous framework with the average pore size of 300 μm was well preserved in nMgHA/PHAs. Meanwhile, excess magnesium released from the so-called ‘surface pool’ of PHA was verified. In contrast, slower release of magnesium at lower concentrations was detected for nMgHA/PHAs. Significantly more newly-formed bone and microvessels were observed in 20%nMgHA/PHA than the other specimens. With the limitations of the present study, it could be concluded that PHA coated by 20%nMgHA may have the optimized osteogenic performance due to the elimination of the excess magnesium from the ‘surface pool’, the preservation of the inherent 3D porous framework with the favorable pore size, and the release of magnesium at an appropriate concentration that possessed osteoimmunomodulatory effects on macrophages.


Introduction
Xenografts composed of biological apatite have been widely used in the treatment of bone defects caused by trauma, tumor, surgery, and congenital malformations due to their high accessibility and similarity to the inorganic component of human bone in structure and chemical composition [1]. However, the osteogenic efficacy of xenografts remains low since their traditional fabricating strategy makes them inert [2]. In order to improve the osteogenic efficacy of biomaterials and accelerate the process of bone reconstruction, strategies have been proposed to design them as osteoimmunomodulatory biomaterials that could promote osteogenesis through regulating immune cellular responses and modulating the local immune environment [3,4]. Two main approaches have been proposed, including the incorporation of bioactive molecules and the modification of chemical and/or topographical characteristics [4]. Although bioactive molecules have been incorporated into different materials [5][6][7], the selection of such molecules should be cautious, since the effects of cytokines and signaling factors involved in multiple signaling pathways in the osteoimmune microenvironment have not been fully elucidated [4]. In contrast, to modify the chemical and/or topographical characteristics of materials is safer, more promising, and more cost-effective [8]. For instance, xenografts may become osteoimmunomodulatory through the incorporation of bioactive inorganic ions which could modulate the immune response in the microenvironment [4,9,10].
As one of such bioactive inorganic ions, magnesium (Mg 2+ ) has drawn much attention since it plays an important role in bone health and has concentration-dependent effects on both the physicochemical properties and osteogenic capability of biomaterials [11][12][13][14][15][16][17]. Previously, both 5-8 mM [17] and ∼10 mM [11] extracellular Mg 2+ were reported to be beneficial to osteogenesis at the early stage of bone repair (∼1 week), whereas either higher concentration or longer exposure period of Mg 2+ may inhibit bone formation and remodeling [11]. It was also reported that about 30% of bone-contained magnesium was in a 'surface pool' , which could release rapidly when xenografts were implanted [18]. It remains unknown whether the rapidly-released magnesium causes inflammation and inhibits cellular functions. Therefore, it is challenging to improve the osteogenic efficacy of xenografts by controlling the extracellular concentration of Mg 2+ in the local microenvironment and modulating the immune response [19].
The present study was to explore the release of magnesium from a porcine bone-derived xenograft, and to provide a cost-effective and straightforward strategy for the controlled release of magnesium, so as to optimize the osteogenic efficacy of the currently used xenografts. Specifically, xenografts with an inherent three-dimensional (3D) framework were prepared from porcine bones which were highly accessible worldwide through a thermal process [10]. The surface of the xenografts was coated by magnesium-doped nanohydroxyapatite (nMgHA) through a hydrothermal approach. The physiochemical properties and osteogenic capability of the modified xenografts were evaluated.

Material preparation
Porcine bone-derived hydroxyapatite (PHA, S 0 ) was prepared with a simple thermal treatment modified from the previous study [10]. In brief, cancellous bone harvested from the femoral epiphysis of 5 month-old pigs (weight: ∼100 kg) was thoroughly cleansed and dissected into regular disks (Ø 9.0 mm × 2.0 mm) with a cutting machine (Accutom-50, Struers, Denmark) and a trephine bur. The disks were then calcinated at 800 • C (heating rate: 10 • C min −1 , holding time: 2 h) in a muffle furnace (Xigema Furnace Industry, Luoyang, China) to remove the organic component. A porcine bonederived xenograft (PHA, S 0 ) with a 3D porous framework was obtained.
For the preparation of material extract, PHA and nMgHA/PHAs (1000 mg) were disinfected by autoclaving and immersed in 10 ml Dulbecco's modified Eagle's medium (DMEM, ThermoFisher, America) in a humidified atmosphere with 5% CO 2 at 37 • C for 2 days. Then, the supernatant was collected by centrifugation at 1500 rpm for 5 min, which was further filtered (0.22 µm filter) to obtain the material extract.

Physicochemical characterization 2.2.1. Morphology and chemical composition
The surface morphology of all samples was observed with a scanning electron microscope (SEM, SU8220, HITACHI, Japan). The pore size of each sample was measured by randomly selecting five sites in each of the three SEM images by using Image Pro Plus (v. 6.0, Media Cybernetics, America). The elemental composition was examined by energy-dispersive xray spectroscopy (EDS, SU8220, HITACHI, Japan). The crystal characteristics were evaluated with an x-ray powder diffractometer (XRD, Empyrean, Netherlands). Powdered samples were mounted on glass stubs, using Cu Kα1 radiation (λ = 1.54056 Å) with a graphite-diffracted beam monochromator, over the range 4 • -60 • (2θ) using a step size of 0.02 • at 4 • min −1 . Diffraction patterns were analyzed in MDI Jade (v. 5.0, Materials Data, Livermore, America). Phase identification was done by comparison with the standard hydroxyapatite (HA) powder . In brief, the compact bones harvested from tibia and femora were excised into chips and digested with collagenase II (1 µg ml −1 , Gibco, America) for 1 h to obtain long bone marrow, which was then used to achieve the single-cell suspensions through filtering with a 70 µm cell mesh. Subsequently, the cell suspensions were seeded on dishes (100 mm, Jet, China) and cultured in a differentiation medium (Procell, China) for three days. Non-adherent cells were removed, while the adherent cells (mBMSCs) were then cultured in Minimum Essential Medium α (ThermoFisher, America) supplemented with 10% FBS and 1% penicillin-streptomycin. Specifically, P3-mBMSCs were used in the subsequent experiments of the present study.

Cytotoxicity
Cytotoxicity of all samples was assessed with a Cell Counting Kit-8 (CCK-8, Dojindo, Kumamoto, Japan). RAW 264.7 cells and mBMSCs were seeded at a density of 2000 cells/well in the 96-well plates, respectively. These cells in the experimental groups were treated with the conditioned medium as mentioned in 2.3.1 for 1, 3, and 5 d, whereas the NC was treated with the regular cell culture medium. Subsequently, the medium was replaced by CCK-8 solution and the plates were incubated in dark for 2 h. The absorbance was then measured at the wavelength of 450 nm using a microplate analyzer. A total of three parallel wells were set for each group and the process was repeated for three times.

RT-qPCR
For RT-qPCR test, RAW 264.7 cells in the NC group was treated with the regular cell culture medium, while those in the experimental groups were treated with the conditioned medium contained with material extract as mentioned in 2.1. Additionally, another group was set by using magnesium chloride solution (MgCl 2 aq, 1M, Leagene, China) as the excess magnesium source so as to avoid the confounding effect of other elements released from PHA. In this group, RAW 264.7 cells were stimulated with lipopolysaccharides (LPS, 1 µg ml −1 , Solarbio, China) for 6 h at first [20], and then treated with cell culture medium containing MgCl 2 at the concentration of 2, 5 and 10 mM for 24 h. A total of three parallel samples of RAW 264.7 cells were collected in each group after 3 days and the total RNA was extracted using an RNA-Quick Purification kit (ESscience, China). For the reverse transcript, complementary DNA was synthesized using a Hifair ® III 1st Strand cDNA Synthesis SuperMix (Yeasen, China) following the manufacturer's instructions. The primer pairs used in RT-qPCR (table 2)   . Two bilateral calvarial bone defects with the external diameter of 5 mm were created on each rat using a trephine drill. A total of 30 critical-sized calvarial defects were then randomly divided into six groups (n = 5) according to the grafts filled: NC (unfilled), positive control (PC, Bio-Oss ® , Geistlich, Switzerland), PHA (S 0 ), 10%nMgHA/PHA (S 1 ), 20%nMgHA/PHA (S 2 ), and 30%nMgHA/PHA (S 3 ). All grafting materials were ground into granules with a diameter of 0.5-1 mm. After the complete filling of grafts, all defects were covered with collagen membranes (Bio-Gide ® , Geistlich, Switzerland) and sutured in layers. The rats were euthanized after 4 weeks. The calvaria containing the entire defects were fixed in 4% paraformaldehyde solution for 24 h.

Micro-CT analysis
The specimens harvested above were scanned with a micro-computed tomography (micro-CT) scanner (SkyScan 1172 x-ray microtomograph, Kontich, Belgium) at 80 kV with a voxel size of 7.95 µm. Two phantom contained rods with a standard density of 0.25 and 0.75 g cm −3 were scanned with each sample for calibration. The newly-formed bone and xenografts were segmented from the marrow and soft tissue using a global thresholding procedure [21]. 3D reconstruction and calculation were completed using Mimics Research (v. 19.0, Materialise, Belgium). The greyscale thresholding from −900 to −820 represented bony tissue, while that from −820 to −769 represented the implanted bone substitutes. Bone volume (BV), bone surface (BS), material volume (MV), and total volume (TV) were calculated and presented as BV/TV (%), BS/TV (%), and MV/TV (%), respectively.

Histological analysis
The specimens were decalcified with the EDTA decalcification solution (Servicebio, China) for 5 weeks. Then, they were dehydrated in ethanol, embedded in paraffin, and cut into 5 µm-thick sections using a rotary microtome (RM215, Leica Microsystems, Germany). Haematoxylin and eosin staining (Servicebio, China) and Masson staining (Solarbio, China) were performed for each sample. Histological images were captured using a polarized light microscope (ECLIPSE LV100POL, Nikon, Japan). The cranial defect area was defined as the area of interest (AOI) in Image Pro Plus. Then, new bone regeneration rate and residual material rate were obtained by dividing the total area of new bone and residual material by AOI, respectively. The total number of microvessels was counted in three nonoverlapping views for each sample at 20 × magnification. The density of microvessels was defined as the total number of microvessels for each view at 20 × magnification (approximately 0.24 mm 2 ).

Physiochemical characterization
Representative morphological characteristics of specimens were shown in figure 1(A). At the magnification of 100× and 500×, a 3D porous framework was observed in PHA (S 0 ) and nMgHA/PHAs, with the average pore size of about 300 µm (ranging from 100 to 500 µm). At the magnification of 10 k×, micro pores could be observed for both Bio-Oss and PHA, which were not detected in nMgHA/PHAs whose surface was covered by a dense layer instead. At the magnification of 50k×, Bio-Oss was observed as polyhedral and rod like crystals with a length of 50-80 nm along the long axis. S 0 consisted of spheroid and ellipsoidal crystals with a diameter of 300-400 nm. The surface layer of nMgHA/PHAs was found to be spheroid crystals with a diameter of 30-50 nm. After soaking in DMEM for 7 d, little macro morphological changes were observed for S 0 and nMgHA/PHAs (500×, figures 1(A) and (B)). At the magnification of 10k×, micro pores could be observed in Bio-Oss, PHA and nMgHA/PHAs. The major chemical components of PHA and nMgHA/PHAs were calcium, phosphorus, and oxygen (table 3). Trace magnesium and carbon were also detected. After immersion in DMEM for 7 d, magnesium was almost undetectable in PHA (S 0 ) and 10%nMgHA/PHA (S 1 ), whereas 0.04 mol% and 0.08 mol% magnesium was detected in 20%nMgHA/PHA (S 2 ) and 30%nMgHA/PHA (S 3 ), respectively.
In figure 1(E), NC showed the baseline of Mg 2+ concentration in DMEM. The initial Mg 2+ released from PHA (S 0 ) was the highest (around 13.79 mM) in all samples, which dropped rapidly in the first two days. The release rate of Mg 2+ from Bio-Oss (PC) increased slowly while that from nMgHA/PHAs (S 1 , S 2 , and S 3 ) decreased from day 1 to day 7. The daily Mg 2+ released from all samples reached a platform at around 1.5 mM on day 7. As shown in figure 1(F

In vitro evaluation of nMgHA/PHA
Compared with the regular cell culture medium (NC), improved cell viability of RAW 264.7 was observed for all experimental groups on day 3 and 5 ( figure 2(A)). The viability of mBMSCs was comparable between the experimental groups and the control on day 1 and day 3. However, significantly lower viability was found in nMgHA/PHAs (S 1 and S 3 ) on day 5 in comparison with the control ( figure 2(B)).
Although there was no statistical difference, a descending trend from 10%nMgHA/PHA (S 1 ) to PHA (S 0 ) was observed in the expression level of M1 related genes (CD11c and IL-6). The expression level of M2 related genes (TGF-β and VEGF) of S 1 and 10%nMgHA/PHA (S 2 ) seemed to be higher than that of NC and PHA (S 0 ) without significant difference (figure 2(C)).
In the additional test with MgCl 2 , the expression level of both CD11c and IL-6 in the group of 10 mM was significantly lower than that of the other groups. Besides, the expression level of both TGF-β and VEGF in the groups of 2 mM and 5 mM was comparable, and significantly higher than that of NC and 10 mM groups (figure 2(D)).
The relative ALP staining area of mBMSCs was comparable among S 0 , S 1 and S 3 , while the area of S 2 was significantly larger than that of NC and S 1 (figure 2(E)).

In vivo evaluation of nMgHA/PHA
As shown in figure 3(A), few newly-formed tissues were observed in the defects of NC (critical bone defects without bone grafting). In contrast, the defects in the other groups were almost completely filled by the complex of newly-formed tissues and grafting materials. The BV/TV of 20%nMgHA/PHA (S 2 ) was slightly higher than that of 10%nMgHA/PHA (S 1 ), but significantly higher than that of the other groups ( figure 3(B)). The BS/TV of S 2 was comparable with that of Bio-Oss (PC), and statistically higher than that of S 1 and 30%nMgHA/PHA (S 3 ) ( figure 3(C)). The MV/TV of S 2 was close to that of S 1 , and significantly higher than that of S 3 and PC ( figure 3(D)).
In the histological images, bone grafts surrounded by newly-formed bone, microvessels, blood cells and collagen fibers were observed in the defects of the PC and experimental groups, whereas much less newly-formed tissues were detected in the defects of the NC (figures 4 and 5). Quantitatively, the newlyformed bone in the group of S 2 (20%nMgHA/PHA) was close to that of S 1 (10%nMgHA/PHA), and significantly more than that of the other groups (figures 4 and 5(B)). The residual material granules of S 3 (30%nMgHA/PHA) were substantially less than those of S 1 and S 2 (figures 4 and 5(C)). The microvessel density of S 2 was comparable with that of S 1 , and statistically higher than that of S 0 , S 3 , NC, and PC (figures 4 and 5(D)).

Magnesium released from PHA and nMgHA/PHAs
In the present study, magnesium ions were detected to release rapidly from PHA (S 0 , figure 1(E)), corresponding to the 'surface pool' contained magnesium [18]. The cumulative concentration of magnesium on day 2 and day 7 of PHA was 22.54 and 29.47 mM, respectively, which was double and even triple of the recommended concentration of 10 mM for bone regeneration [11]. Such excess magnesium was prone to aggravate inflammation and inhibit bone regeneration [18,22]. Therefore, the excess magnesium contained in the 'surface pool' of the original porcine-bone derived xenograft should be treated with caution.
In contrast, nMgHA/PHAs showed slower release of magnesium, indicating that the rapid release of magnesium from the 'surface pool' of the PHA part in nMgHA/PHAs had probably completed in the hydrothermal process which was not taken for the original PHA [18]. The slow release of magnesium was due to the dissolving of the surface coating (nMgHA), resulting in the exposure of the micro pores of PHA after immersion (figures 1(A) and (B), 10k×). The cumulative magnesium released from 10%nMgHA/PHA (S 1 , 4.03 mM), 20%nMgHA/PHA (S 2 , 5.92 mM), and 30%nMgHA/PHA (S 3 , 9.35 mM) was lower than 10 mM on day 2, which was likely to be safe and beneficial to cellular responses (figure 1(E)) [11]. On day 7, the cumulative content of magnesium of S 1 (8.13 mM) and S 2 (10.72 mM) was close to the recommended concentration, while that of S 3 (17.51 mM) was much higher than 10 mM, indicating an inhibition effect of S 3 on cellular performance [11]. This was supported by the significantly less newly-formed bone and microvessels of S 3 compared with S 1 and S 2 (figures 5(B) and (D)).

Surface coating of PHA
The surface chemistry of biomaterials plays a paramount role in their interaction with the biological microenvironment, including protein adsorption and cellular responses, which can be optimized by the controlled release of bioactive elements [4] such as fluoride [9,23] and magnesium [24]. In the present study, the controlled release of magnesium was achieved by coating PHA with nMgHA through a cost-effective and straightforward hydrothermal method. The surface coating of nMgHA not only preserved the inherent 3D porous structure of PHA that is beneficial for angio-and osteo-genesis, but also functioned as the basis of extracellular magnesium regulation in the micro-environment.
The inherent 3D framework preserved in nMgHA/PHAs (figures 1(A) and (B)) was one of the key factors influencing the plasticity of immune cells and their interaction with bone-forming cells [4,25]. A framework with an appropriate pore size of biomaterials may determine the fate and polarization of macrophages by allowing the sufficient infiltration of oxygen and nutrients [22], and inducing a moderate hypoxia environment that hinders the inflammatory response and promotes the angiogenic effects [4]. To construct bone substitutes with favorable porosity  and appropriate pore size remains challenging [26]. Pores with a diameter of 350 µm could promote both the osteogenic and angiogenic processes [27]. In the present study, the average pore size of nMgHA/PHAs was 300 µm, which was likely to favor the angio-and osteo-genic processes, as demonstrated by the newlyformed bone and microvessels (figures 4 and 5).
In addition, the incorporation of magnesium could cause proportion-dependent changes to the physiochemical and biological properties of HA, including the lowered crystallinity, increased solubility and bioactivity [17]. This enabled the modulation of the biological performance of nMgHA/PHAs through tuning the incorporation proportion of magnesium into HA. As discussed above, S 1 (10%nMgHA/PHA) and S 2 (20%nMgHA/PHA) in the present study showed slow release of magnesium within the beneficial concentration range, indicating their potential as bone grafting materials after further evaluation.

Evaluation of nMgHA/PHAs in vitro
It was previously reported that magnesium had a concentration-dependent osteoimmunomodulatory effect on bone regeneration in vitro [11,14,15]. With the increasing concentration of magnesium, the production of M1 related pro-inflammatory cytokines was depressed by inhibiting the toll-like receptor signaling pathway [28]. However, there was no significant difference in the RT-qPCR results between PHA and nMgHA/PHAs (figure 2(C)), which may be ascribed to the confounding effects of other trace elements released from biological apatite [1,29]. Biological apatite prepared from natural resources contains trace amounts of elements such as magnesium, zinc, and fluoride [30,31]. It was found that the existence of these trace elements played regulatory roles in osteoimmunomodulation [8,30]. For instance, trace fluoride ion at 0.24-24 µM was likely to regulate the expression of genes including IL-6 and VEGF in macrophages [32].
Given this, as an additional test, the previously reported effect of magnesium on macrophages was verified. As shown in figure 2(D), the decreased expression level of CD11c and IL-6 at 10 mM Mg 2+ was observed, corresponding to the inhibition of M1 polarization by magnesium as reported before [11]. However, at the same concentration, the expression level of TGF-β and VEGF was downregulated as well, indicating that 10 mM Mg 2+ may inhibit the polarization of both M1 and M2. By contrast, with 2 mM and 5 mM Mg 2+ , although there was no substantially inhibitory effect on the expression level of CD11c and IL-6, the significantly increased expression level of TGF-β and VEGF demonstrated that the polarization of M2 was promoted.
As mentioned above, the day-2 material extract of S 1 , S 2 , S 3 , and S 0 contained 4.03, 5.92, 9.35, and 22.54 mM of Mg 2+ , respectively, which was then diluted to 89% and used for cell culture. In other words, the Mg 2+ concentration in the conditioned cell culture medium was around 3.58, 5.27, 8.32, and 20.06 mM for S 1 , S 2 , S 3 , and S 0 , respectively. Given the group of S 2 possessed the closest Mg 2+ content to 5 mM, it was likely to promote the expression . Histological analysis of calvarial defects area Masson's trichrome staining, circles ⃝ for newly-formed bone, black triangles ▲ for material granules, arrows ↑ for microvessels. (NC, negative control, critical bone defects without bone grafting; PC, positive control, bone defects filled with Bio-Oss; S0, bone defects filled with PHA, original porcine bone derived hydroxyapatite; S1, bone defects filled with 10%nMgHA/PHA, PHA coated by 10% magnesium incorporated hydroxyapatite; S2, bone defects filled with 20%nMgHA/PHA, PHA coated by 20% magnesium incorporated hydroxyapatite; S3, bone defects filled with 30%nMgHA/PHA, PHA coated by 30% magnesium incorporated hydroxyapatite). of TGF-β and VEGF. In addition, the group of S 2 showed the highest ALP activity in the coculturing system (figure 2(E)), indicating that S 2 may have the best potential of bone regeneration among the experimental groups.

Evaluation of nMgHA/PHAs in vivo
As shown in figures 3-5, significantly more newlyformed bone and micro-vessels were observed in S 2 (20%nMgHA/PHA), indicating that S 2 had significantly higher capacity of osteogenesis and vascularization in vivo than the other materials. In comparison with S 2 , S 1 (10%nMgHA/PHA) showed the comparable amounts of newly-formed bone and microvessels in vivo. Besides, PHA showed less newly-formed bone and microvessels than PC (Bio-Oss) (figure 5), whereas S 3 (30%nMgHA/PHA) had the least residual material granules, newly-formed bone and microvessels among all groups.
The rapid release of magnesium from the 'surface pool' of PHA may cause inhibition of cellular responses, resulting in the unsatisfactory tissue regeneration in vivo [33]. The difference in tissue regeneration and material resorption between nMgHA/PHAs  (H), (E) staining, circles ⃝ for newly-formed bone, black triangles ▲ for material granules, arrows ↑ for microvessels; (B). Comparison of the newly-formed bone between groups; (C). Comparison of the residual materials between groups; (D). Comparison of the microvessels between groups (NC, negative control, critical bone defects without bone grafting; PC, positive control, bone defects filled with Bio-Oss; S0, bone defects filled with PHA, original porcine bone derived hydroxyapatite; S1, bone defects filled with 10%nMgHA/PHA, PHA coated by 10% magnesium incorporated hydroxyapatite; S2, bone defects filled with 20%nMgHA/PHA, PHA coated by 20% magnesium incorporated hydroxyapatite; S3, bone defects filled with 30%nMgHA/PHA, PHA coated by 30% magnesium incorporated hydroxyapatite).

Figure 6.
Optimizing the osteogenic capacity of PHA through magnesium-based osteoimmunomodulation (20% nMgHA/PHA released an appropriate concentration of Mg 2+ , which could not only regulate gene expression in macrophages, but also show the best potential in bone and vascular regeneration). may be mainly resulted from the differing incorporation proportion of magnesium in the surface coating of nMgHA. Although S 1 and S 2 had different incorporation proportion of magnesium, the released magnesium in the micro-environment was close between them, leading the quantitatively comparable new bone and microvessels. S 3 was coated by nMgHA with 30 mol% magnesium that was suggested as the highest ratio for the doping of magnesium into HA [34]. Such a high content of magnesium could substantially increase the solubility and instability of nMgHA, leading to the rapid biodegradation of nMgHA/PHA and less new bone than S 1 and S 2 (figure 5(D)). Therefore, the ideal doping proportion of magnesium may be 20% in the preparation of nMgHA/PHA (figure 6).

Limitations
This study was limited to a rat model with calvarial defects. Further studies are in need with larger animal models whose macro-/micro-structure and remodeling mode of bone are closer to those of human bone [25]. Besides, the material extract of xenografts was a compromised option for the observation of cellular responses since the topographical features of such biomaterials were lost and the contents of ions varied between extracts [1,29].

Conclusions
In the present study, the rapid release of excess magnesium from the 'surface pool' of xenografts was verified, which was likely to have negative influence on cellular responses. Meanwhile, a porcine bone-derived xenograft (PHA) was modified by coating nano magnesium-doped HA (nMgHA) through a straightforward and cost-effective hydrothermal approach, with its inherent 3D porous framework and favorable pore size well preserved. With the limitations of the present study, the osteogenic capacity of PHA was optimized by the surface coating of nMgHA. More researches on the stronger evidence of the osteoimmunomodulatory effect of magnesium released from nMgHA are needed in the future. With the ideal incorporation ratio of magnesium (20%), nMgHA/PHA could be a promising xenograft for bone reconstruction.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).