This study looked at the direct long-term effect of magnesium alloys on primary HRDs. Cell viability, differentiation and morphology as well as pH and calcium uptake were analysed to assess the overall biocompatibility of the tested materials. We evaluated the long-term effects of magnesium on human cells to simulate the in vivo situation as closely as possible.
Mg2Ag had high cell viability from day 1 (113.4 ± 29.8 %) to day 21 (98.5 ± 12.0 %). It was also shown in previous works  that Mg-Ag alloys have negligible cytotoxicity and sound cytocompatibility. Pure Mg had high viability at the very first day (93.4 ± 25.3 %), but then the viability decreased to 24.0 ± 19.5 % at day 21. Mg10Gd and WE43 impaired cell viability in this study. Previous studies have shown higher values for cell viability measured by MTT test compared with the present study [15–17]. The difference between this work and previous publications is that the present study applied the longest in vitro incubation times for magnesium alloys tested up to now. The HRDs were kept in direct contact with the magnesium samples and not in magnesium extract, as done in most studies [15–17].
An important drawback of tetrazolium-based tests is that the difference between cytotoxic (cell death) and cytostatic (reduced growth rate) effects cannot be distinguished . We thus looked at cell morphology under light microscopy, TEM and SEM.
After examination under SEM and light microscopy, it was revealed that the number of cells decreased in the presence of pure Mg, Mg10Gd and WE43. These materials seem to have long-term cytotoxic effects on HRD when placed in direct contact with the cells. This explains the low viability values.
The cell number was high and the cells had normal morphology in Mg2Ag groups. However, the cell viability was lowest at day 7 (63.3 ± 11.0 %) for Mg2Ag. TEM analysis revealed an elevated amount of lysosomes which contained degraded magnesium particles. Degradation particles were also found in the cytoplasm. The presence of high amounts of degradation products inside the HRDs could explain the lower cell viability values for Mg2Ag at day 7. It was shown in previous studies that uptake of material particles leads to induction of cell stress which triggers cytotoxicity .
ALP activity in HRDs is an important factor in bone mineral formation and shows a range of changes during differentiation. Inhibition of ALP activity in osteogenic differentiating HRD was caused by pure Mg at day 14 and 28. All other magnesium alloys did not affect the ALP activity. In this respect, our study shows similar results to previous research in this area , despite the fact that we observed osteogenic differentiation over much longer periods and using direct contact of cells with magnesium.
SEM analysis revealed that the cellular attachment was generally best to crystals generated by degradation products on the material surface. Crystals have been seen forming on magnesium alloys such as Mg-Ag in previous studies . Formation of calcium phosphates [Cax(PO4)x] was also observed in previous publications . Interestingly, the crystal distribution was not homogeneous throughout the corrosion layer. In this sense, our results are similar to earlier findings [11, 15].
The fact that the cells attached to the crystalline structures more readily than to the overall material surface and developed numerous pseudopodia can be explained by the rough structure of crystals, and by the chemical composition of these crystals. It was previously shown that cells attach better to certain surfaces with preferable average surface roughness of ~0.5 μm up to ~8.5 μm . Values below or above this range diminish the cells’ ability to bind to the surface.
The chemical composition of the crystals and the degradation layer formed on the magnesium’s surface can also explain the better attachment of the cells to these structures. Their chemical composition consists of calcium, phosphorus, magnesium and oxygen . Thus, the cells attach to already reacted material where they are not mechanically disturbed by hydrogen gas produced as a by-product of degradation. The formation of the degradation layer could also explain the increase in cell density around Mg10Gd and WE43 after 21 days of incubation.
Based on previous findings, the following model for the formation of a corrosion layer has been suggested : (1) initial metal corrosion based on contact with water molecules leads to release of Mg ions, and a thin Mg(OH)2 and MgCO3 layer is formed; (2) The corrosion slows down, and a second layer consisting of amino acids and organic matter is formed; (3) Both layers together shield the sensitive environment around the material and enable cells to grow on the material . Such a complex process of corrosion layer formation could explain the cell viability increase at day 21 for Mg2Ag, Pure Mg and Mg10Gd observed in our experiments.
All magnesium-based materials decreased the amount of Ca2+ in this study. As shown in previous studies, Mg2+ promotes formation of calcium phosphates and consequently decreases the amount of free Ca2+ ions in the medium [11, 14]. In this regard, our results are consistent with earlier works. Sufficient supply of calcium is vital to ensure that bone laid down by osteoblasts is normally mineralised . Calcification is thus advantageous for bone implants that are to be used in the orthopaedic and maxillofacial fields.
It was shown in previous studies that magnesium increases the pH and that high pH promotes Ca2+ binding [13, 14]. In this study, it was also revealed that pH shifts to alkaline values in the presence of magnesium, but to somewhat different degrees for the different alloys. However, no statistical correlation was observed between pH and consumed Ca2+.
In conclusion, our study reveals the long-term effects of magnesium materials on human HRDs seeded directly onto magnesium discs. In respect to cell morphology, cell density and the effect on the surrounding pH, Mg2Ag showed the most promising results. However, the mechanism of cell stress induction and cytotoxicity needs to be further studied to enable prediction of possible health risks.