The following materials were used to produce alloys for this study: magnesium (99.99 %, Xinxiang Jiuli Magnesium Co. Ltd., China), yttrium (99.95 % Y, Grirem Advanced Materials Co. Ltd., China), gadolinium (99.95 % Gd, Grirem Advanced Materials Co. Ltd., China), rare-earth mixture (Grirem Advanced Materials Co. Ltd., China) and silver (99.99 % Ag, ESG Edelmetall-Handel GmbH & Co. KG, Germany).

Three magnesium-based materials were produced: Mg2Ag (1.89 % Ag, the rest was Mg), Mg10Gd (8.4 % Gd, the rest was Mg) and WE43 (3.45 % Y, 2.03 % Nd, 0.84 % Ce, the rest was Mg). Pure magnesium (99.99 % Mg) was used as control. The concentrations of magnesium (Mg), Y, neodymium (Nd) and cerium (Ce) were determined by spark emission spectrometer (Spectrolab M, Spektro, Germany), and the concentrations of Ag and Gd were determined by X-ray fluorescence spectrometer (Bruker AXS S4 Explorer, Bruker AXS GmbH, Germany). The materials were cast at Helmholtz-Zentrum Geesthacht Magnesium Innovation Center (HZG-MagIC).

The three magnesium alloys (Mg2Ag, Mg10Gd and WE43) were produced by permanent mould gravity casting. After melting pure Mg, the melt was held at 720 °C and the preheated alloying elements were added with continuous stirring for 15 min. The melt was poured into a preheated (550 °C) permanent steel mould treated with boron nitride. During the casting process, cover gas was used (SF₆ and Ar mixture). The alloys were homogenised with a T4 heat treatment prior to extrusion in Ar atmosphere at 550 °C (Mg10Gd and WE43) or 420 °C (Mg2Ag) for 6 h. Afterwards, the alloys were extruded indirectly with extrusion ratio of 4:25. The chamber of the extrusion machine was set to 370 °C, and the 30-mm-diameter billets were preheated for 1 h at 370 °C (Mg2Ag), 390 °C (WE43) or 430 °C (Mg10Gd). The extrusion speed was between 3 and 4.5 mm/s. Pure Mg was cast by permanent mould direct chill casting.²⁷ The cast billet (d = 110 mm) was extruded indirectly with an extrusion ratio of 1:84. The billet temperature was 340 °C, and the speed of the extrusion was 0.7 mm/s. Discs (10 mm diameter, 1.5 mm thickness) were machined from the extruded bars.

The samples were sonicated for 20 min in dry isopropanol, dried and gamma-sterilised at the BBF Sterilisation Service GmbH facility (Kernen, Germany) with total dose of 29 kGy. Before seeding the cell culture, the samples were pre-incubated to form a protective degradation layer. Per 0.2 g of sample, 3 mL of F12 K (Gibco^®, Life Technologies, USA) was used. The samples were kept at 37 °C in 5 % CO₂ atmosphere for 12 h.

HRD cells were cultured from various patients, with the approval of the local Ethics Commission, as described by Wenisch et al. [18]. The adult patients were of different genders and ages and did not display any disease related to bone metabolism. In total, cells from six different patients were taken for this study.

The reaming debris was cultured in Petri dishes with F12 K medium including 20 % foetal calf serum (FCS), 100 U/ml penicillin and 100 μg/g streptomycin. After 4–7 days, HRD cells started to grow out of the debris. When the cells reached confluence after 2–3 weeks, they were trypsinised and transferred to cell culture flasks. All cells were kept at 37 °C in 5 % CO₂atmosphere.

To determine cell viability, an MTT assay was conducted according to Mosmann et al. [19]. Briefly, 10,000 cells/cm² were seeded into 12-well plates containing pre-incubated magnesium discs and F12 K medium with 20 % FCS and 100 μg/g streptomycin. Controls with no magnesium discs but only HRDs were used as well. The cell medium was changed every second day during the experiment. Duplicates were used for each material and patient, resulting in testing of 12 wells per specimen. After 24 h, 7 and 21 days, MTT solution was added to cell medium. The cells were then incubated in the dark for 4 h at 37 °C. Subsequently, the cell medium was discarded, and the cells were lysed with 0.004 N HCl in isopropanol. The cell lysates were centrifuged, and supernatants were transferred as triplets to a 96-well plate. Adsorption was measured at 570 and 630 nm using a Synergy HT microplate reader (BioTek, Bad Friedrichshall, Germany). The MTT assay was also performed for magnesium discs not seeded with cell culture to exclude the material’s effect on the test and see only how the cells reacted during the assay.

Additionally, cell morphology was studied by inverted light microscopy using a Leica microscope type 090-135.002 (Leica Microsystems GmbH, Wetzlar, Germany) equipped with a Nikon Ds-Fi1 digital camera (Nikon, Düsseldorf, Germany).

As an indicator of changes in the differentiation behaviour of the bone-forming cells caused by the test substances, a SensoLyte^® pNPP alkaline phosphatase assay (AnaSpec, Fremont, CA) was applied after 24 h and 7, 14, 21 and 28 days of culturing in Dulbecco’s modified Eagle’s medium (DMEM), low glucose with l-glutamine, 10 % FCS, 100 U/ml penicillin, 100 μg/g streptomycin, 0.1 μM dexamethasone, 0.005 μM ascorbic acid and 10 mM β-glycerol phosphate to induce osteogenic differentiation. The cell medium was changed every second day during the experiment. Duplicates were used for each material and patient, for a total of 12 wells per specimen. Controls with no magnesium discs but only HRDs were used as well.

The cells were washed and frozen at −80 °C. After thawing, the cell number was measured using a PicoGreen^® dsDNA quantitation assay (Invitrogen, Eugene, OR) according to the manufacturer’s protocol. Cells were lysed with 1 % Triton X-100 in phosphate-buffered saline. The cell lysates were centrifuged, and the supernatants were mixed with PicoGreen^® working solution in a 96-well plate. The samples were excited at 485 nm, and the fluorescence emission intensity was measured at 528 nm. The cells that were lysed for the PicoGreen^® assay were centrifuged, and the supernatants were diluted in a specific assay buffer. ALP substrate was applied to the diluted samples, and the absorbance was measured at 405 nm. The absolute amounts of ALP were correlated with the cell numbers obtained from the PicoGreen^® assay. Both ALP and PicoGreen^® assays were additionally performed for magnesium discs not seeded with cell culture to exclude the material’s effect on the test and see only how the cells reacted during the assays.

HRD cells seeded in chamber slides (Nalge Nunc International, Rochester, NY) were incubated with and without magnesium discs for 21 days. The cell layer was fixed for 30 min with 2 % paraformaldehyde in 0.1 M sodium phosphate buffer (pH 7.2–7.4) with 2 % glutaraldehyde and 0.02 % picric acid, followed by 20 min fixation with 1 % osmium tetroxide in 0.1 M sodium cacodylate buffer (pH 7.2–7.4). The samples were dehydrated and embedded in Epon before ultrathin sections (80–100 nm) were applied to collodion-coated copper grids. Analysis was done with a LEO 912 transmission electron microscope (Carl Zeiss AG, Oberkochen, Germany) at 80 kV accelerating voltage, equipped with a TRS SharpEye slow-scan dual-speed charge-coupled device (CCD) camera (Albert Troendle Prototypentwicklung, Moorenweis, Germany).

HRD cells were cultivated on magnesium discs for 7 and 21 days. Controls with no magnesium discs but only HRDs were used as well. Subsequently, the cells were fixed in 2 % glutaraldehyde in 0.1 M Na-phosphate buffer (pH 7.4) for 1 h at room temperature, followed by dehydration in a graded series of ethanol and critical-point drying. The specimens were mounted together on aluminium pin stubs with the help of adhesive carbon pads. The specimens were then sputter-coated with gold/palladium (SC7640 sputter coater, VG Microtech, Uckfield, East Sussex, UK) and assessed using a LEO 1530 (LEO Elektronenmikroskopie GmbH, Oberkochen, Germany) field-emission scanning electron microscope operated at 7.5 or 15 kV.

At established time points, medium was collected and analysed for Ca²⁺ in solution and pH. The concentration of Ca²⁺ was measured using a calcium analyser (9180 electrolyte analyzer, Roche, Mannheim, Germany), and pH measurements were carried out using a pH-meter (Titan X, Fisher Scientific GmbH, Schwerte, Germany) for each time point. The control group for this investigation consisted of well plates that contained only HRDs and medium but no magnesium.

Data were analysed using the Statistical Package for the Social Sciences (SPSS, v18, SPSS Inc., Chicago, USA). The significance level was set at 5 %. The treatment groups were compared with the control group without magnesium by Mann–Whitney U-test. The results are expressed as means. The graphs were plotted with SPSS.

It was seen that all magnesium materials increased the pH of the medium compared with the control group. The following general pattern was observed for all groups: the pH values were stable up to day 7, a sudden pH drop occurred on day 14, then the pH tended to increase slightly up to day 28. The pH values for Mg2Ag were most similar to the control. Pure Mg caused the greatest increase in pH among all groups, and this increase was statistically significant compared with Mg2Ag (p ≤ 0.003) but not compared with Mg10Gd (p ≥ 0.02) or WE43 (p ≥ 0.02). No correlation between pH and Ca²⁺ consumption was found in this study.