Acta Biomaterialia 6 (2010) 1529–1535

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /actabiomat

Phase composition, mechanical performance and in vitro biocompatibilityof hydraulic setting calcium magnesium phosphate cement

Uwe Klammert a,*, Tobias Reuther a, Melanie Blank b, Isabelle Reske b, Jake E. Barralet c,Liam M. Grover d, Alexander C. Kübler a, Uwe Gbureck b

a Department of Cranio-Maxillo-Facial Surgery, University of Würzburg, Pleicherwall 2, 97070 Würzburg, Germanyb Department for Functional Materials in Medicine and Dentistry, University of Würzburg, Pleicherwall 2, D-97070 Würzburg, Germanyc Faculty of Dentistry, McGill University, Montreal, Quebec, Canadad Department of Chemical Engineering, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK

a r t i c l e i n f o

Article history:Received 29 May 2009Received in revised form 7 October 2009Accepted 13 October 2009Available online 1 November 2009

Keywords:Bone replacement materialCalcium magnesium phosphate cementNewberyite

1742-7061/$ - see front matter � 2009 Acta Materialdoi:10.1016/j.actbio.2009.10.021

* Corresponding author. Fax: +49 931 201 74845.E-mail address: klammert_u@klinik.uni-wuerzburg

a b s t r a c t

Brushite (CaHPO4�2H2O)-forming calcium phosphate cements are of great interest as bone replacementmaterials because they are resorbable in physiological conditions. However, their short setting timesand low mechanical strengths limit broad clinical application. In this study, we showed that a significantimprovement of these properties of brushite cement could be achieved by the use of magnesium-substi-tuted b-tricalcium phosphate with the general formula MgxCa(3–x)(PO4)2 with 0 < x < 3 as cement reac-tants. The incorporation of magnesium ions increased the setting times of cements from 2 min for amagnesium-free matrix to 8–11 min for Mg2.25Ca0.75(PO4)2 as reactant. At the same time, the compressivestrength of set cements was doubled from 19 MPa to more than 40 MPa after 24 h wet storage. Magne-sium ions were not only retarding the setting reaction to brushite but were also forming newberyite(MgHPO4�3H2O) as a second setting product. The biocompatibility of the material was investigatedin vitro using the osteoblast-like cell line MC3T3-E1. A considerable increase of cell proliferation andexpression of alkaline phosphatase, indicating an osteoblastic differentiation, could be noticed. Scanningelectron microscopy analysis revealed an obvious cell growth on the surface of the scaffolds. Analysis ofthe culture medium showed minor alterations of pH value within the physiological range. The concentra-tions of free calcium, magnesium and phosphate ions were altered markedly due to the chemical solubil-ity of the scaffolds. We conclude that the calcium magnesium phosphate (newberyite) cements have apromising potential for their use as bone replacement material since they provide a suitable biocompat-ibility, an extended workability and improved mechanical performance compared with brushite cements.

� 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Synthetic bone grafts offer the advantages of having a well-definedchemistry and architecture and are commonly used for both non- orlow-load-bearing defects. Additional advantages are the absence ofany donor-site morbidity compared with autologous bone grafts,and the fact that the risk of a transmission of infectious diseases whileusing homologous or xenogenous grafts can be excluded. Favouredsynthetic materials for bone replacement are based on calcium phos-phate chemistry due to the similar chemical composition to the min-eral phase of bone; materials are either applied as sintered monolithsor granules of b-tricalcium phosphate (b-TCP) [1] or hydroxyapatite(HA) [2] as well as self-setting cements [3].

Calcium phosphate cement (CPC) formation is based on thedifferent pH-dependent solubility of cement reactants and the final

ia Inc. Published by Elsevier Ltd. A

.de (U. Klammert).

setting product, which is either hydroxyapatite (HA) or dicalciumphosphate dihydrate (known as brushite) [4]. Above pH 4.2 HA isthe product, while at low pH < 4.2, orthophosphate is protonatedand the secondary calcium phosphates monetite (CaHPO4) and brush-ite (CaHPO4�2H2O) are the most insoluble calcium phosphate phases[5]. HA-forming cements have been successfully introduced into clin-ical applications for filling non-load-bearing defects in craniomaxillo-facial and orthopaedic surgery [6–8], but this cement type isthermodynamically non-resorbable due to the stability of HA underphysiological conditions. Degradation is only possible by osteoclasticbone remodelling and is limited to surface degradation since cells can-not penetrate the microporous cement structure. In contrast, brushitecements are resorbable under physiological conditions due to thehigher solubility of brushite at pH 7.4 compared to HA. The biocom-patibility and degradation of brushite cements have been demon-strated in several in vitro and in vivo studies [9–15].

Although most research has previously focused on apatitecements, interest in brushite cements increases as its potential

ll rights reserved.

1530 U. Klammert et al. / Acta Biomaterialia 6 (2010) 1529–1535

advantages become apparent. Indeed the first brushite product(chronOSTM Inject, Synthes) has been launched recently to themarket. Brushite cements are produced by b-tricalcium phosphate(b-TCP), monocalcium phosphate monohydrate (MCPM) [16,17],b-TCP-phosphoric acid [18] or b-TCP polyphosphoric acid [19] sys-tems. Because brushite cement has such a fast setting reaction(several times higher than HA), a low powder-to-liquid ratio com-bined with setting retarders has to be used for brushite cements tokeep the cement paste workable over an adequate time period.Retarders are either chondroitin-4-sulfate/silica gel [20], pyrophos-phate, citrate, sulfate [21] or magnesium ions [22], which are com-monly added as soluble salts in low concentrations to the cementliquid. Magnesium ions are in addition known to suppress in vivoHA formation in brushite cements [22] since magnesium acts as astrong inhibitor of hydroxyapatite crystal growth. Various studieshave synthesized calcium magnesium phosphate cements in thepast, e.g. by using mixtures of MgO and MgHPO4�3H2O with calciumphosphates like CaHPO4, Ca(H2PO4)2�H2O, b-TCP, a-TCP or tetracal-cium phosphate [23,24]. Although the setting times of some of thesecements were in a clinically appropriate range of 4–11 min, theirmechanical performance was relatively weak with a maximum com-pressive strength of approximately 11 MPa.

The current study followed a different approach by investigat-ing the effectiveness of using magnesium-substituted b-tricalciumphosphate ceramic as cement reactant to slow down the settingreaction of brushite cement. Magnesium-substituted compoundswith the general formula MgxCa(3–x)(PO4)2 (0 < x < 3) were pro-duced by sintering and reacted with monocalcium phosphatemonohydrate (MCPM, Ca(H2PO4)2�H2O) and a 0.5 mol l�1 citric acidsolution. The setting products were characterized with regard totheir phase composition and mechanical performance and thein vitro cytocompatibility was tested using the osteoblast-like cellline MC3T3-E1.

2. Materials and methods

2.1. Cement preparation

b-TCP was prepared by sintering CaHPO4 and CaCO3 (both Merck,Darmstadt, Germany) in a molar ratio of 2:1 at 1100 �C for 5 h. Mg-substituted b-TCPs with the general formula MgxCa(3–x)(PO4)2 with0 < x < 3 were prepared by replacing CaCO3 with Mg(OH)2 (Fluka,Seelze, Germany) and CaHPO4 by MgHPO4�3H2O (Sigma–Aldrich,Steinheim, Germany) in appropriate stoichiometry with similar sin-tering conditions. The sintered cake was crushed with mortar andpestle and sieved with 355 lm pore size mesh followed by ball mill-ing for 10–150 min at 200 rpm in a planetary ball mill (PM400, Rets-ch, Haan, Germany). Brushite cements were produced by mixing theTCP powder in an equimolar ratio with MCPM (Sigma–Aldrich, Ger-many) in a coffee grinder (PCML 2012, Quelle, Fürth, Germany) for30 s, followed by mixing these powders with 0.5 mol l�1 citric acidat a constant powder-to-liquid ratio (PLR) of 3.0 g ml�1 for all exper-iments unless otherwise stated. Cement pastes were prepared bymixing the cement powder with the appropriate amount of liquidon a glass slab for 20 s.

2.2. Phase composition and mechanical properties

Cement cuboids with an aspect ratio of 2:1 (12 � 6 � 6 mm) wereprepared using silicone rubber moulds and aged in double-distilledwater at 37 �C for 24 h prior to testing. Compressive strength testingwas performed at a cross-head speed of 1 mm min�1 using a staticmechanical testing device Zwick 1440 (Zwick, Ulm, Germany) witha 5 kN load cell. The initial setting times of the cements were mea-sured in a humidity chamber at 37 �C and >90% humidity using theGilmore needle test with a needle of 113.98 g and 2.117 mm diame-

ter according to ASTM standard [25]. The pH values of cement pasteswere measured at a P/L ratio of 3 g ml�1 for up to 6 h using a cut-in pHelectrode (Mettler-Toledo, Germany). X-ray diffraction patternswere recorded on a D 5005 diffractometer (Siemens, Karlsruhe, Ger-many). Data were collected from 2h = 20–40� with a step size of 0.02�and a normalized count time of 1 s/step using Cu–Ka radiation.

2.3. Cell culture

For cell culture assays the murine osteoblast-like cell line MC3T3-E1 was used. Cells were purchased from DSMZ (German Collection ofMicroorganisms and Cell Cultures, Braunschweig, Germany) andused after three passages. Cells were seeded as a solution with2 ml medium (MEM a, Invitrogen, Karlsruhe, Germany, supple-mented with 10% fetal calf serum (Invitrogen, Karlsruhe, Germany))per well in a concentration of 5 � 104 per well (24-well microtitreplates, Greiner, Frickenhausen, Germany) onto cylindrical scaffoldswith a diameter of 15 mm and height of 1.5 mm. The calcium mag-nesium phosphate scaffolds with the raw material phase composi-tion Mg2.25Ca0.75(PO4)2 will be termed below as newberyitescaffolds. The medium was changed every second day. The cultureperiod covered 21 days. All assays were performed on days 5, 9, 13,17 and 21 in three independent experiments, each time n = 4 permaterial per day. Time points were chosen to cover the osteoblasticproliferation and differentiation period from seeding until reachinga confluent cell culture stage. Standard positive (cp-titanium) andnegative (copper) control specimens with the same geometricaldimensions and smooth surfaces were established. The control sam-ples were used as received (cold rolled metal sheet), cleaned onlywith detergents without further surface treatment and autoclavedat 134 �C for 2 h. Ahead of cell culture, the newberyite scaffolds wererinsed with phosphate buffered saline (PBS), until reaching neutralpH values, soaked with 70% ethanol and air-dried. For visualizationof cell growth and cell morphology, the cell-bearing scaffolds under-went scanning electron microscopy (SEM) (FEI, Quanta 200, CzechRepublic). Prior to SEM, samples underwent a fixation procedurewith glutaraldehyde, graded dehydration with acetone, critical pointdrying procedure and coating with gold [26].

2.4. Cell viability

The WST-1 Kit (Roche Diagnostics, Mannheim, Germany) wasused for this assay according to the producer’s manual. The tetra-zolium salt is cleaved to formazan by mitochondrial dehydrogen-ases. The proliferation of viable cells leads to an increased overallactivity of these enzymes. The corresponding changes in absor-bance of the dye solution were measured photometrically at440 nm and normalized by a standard curve of various cell concen-trations. Hence, the cell numbers were calculated.

2.5. Activity of alkaline phosphatase

After removal of the culture medium, the cells were separatedenzymatically (Accutase, PAA Laboratories, Cölbe, Germany) fromthe scaffolds, washed with PBS and treated with 500 ll lysis bufferfor 1 h at 37 �C on a shaker. The lysis buffer was composed of0.75 M 2-amino-2-methyl-1-propanol, pH 10.3 (Merck, Darmtadt,Germany) and 2 mg ml�1 p-nitrophenylphosphate (Sigma–Aldrich,Taufkirchen, Germany). The reaction was stopped by addition of500 ll NaOH. The enzyme activity of the cell lysate was determinedby the conversion of p-nitrophenyl phosphate to p-nitrophenol,which was measured photometrically at 405 nm. Data were normal-ized by a standard curve using various standard concentrations of p-nitrophenol. The alkaline phosphatase (ALP) activity was expressedas nmol of p-nitrophenol produced per min per cell.

Fig. 1. X-ray diffraction patterns of magnesium-substituted tricalcium phosphatesafter sintering at 1100 �C for 5 h. Most relevant peaks are marked as (a) b-Ca3(PO4)2,(b) Ca3Mg3(PO4)4, and (c) Mg3(PO4)2 (farringtonite).

U. Klammert et al. / Acta Biomaterialia 6 (2010) 1529–1535 1531

2.6. Chemical analysis of the culture medium

To enable an assessment of the chemical behaviour of the scaf-folds under cell culture conditions, the supernatant culture med-ium was collected within the medium changes for investigationof pH values as well as the concentration of calcium, magnesiumand phosphate ions. The measurement of the ion concentrationswas performed by the use of commercial complex forming test sys-tems, which are also applied in the clinical routine (Cobas Integrain vitro diagnostics laboratory system, Roche Diagnostics, Mann-heim, Germany). The pH values were measured with a pH meter(inoLab pH Level 1 combined with a SenTix 61 electrode, WTH,Weilheim, Germany).

2.7. Statistics

Data were normal distributed. Variance analysis and consecu-tive statistical calculations were performed with the program SPSS17 (SPSS Inc., Chicago, USA). For comparison of newberyite againsttitanium, respectively, against the bare culture dishes, a t-test forindependent samples was chosen and significance levels were setat p < 0.05.

3. Results

3.1. Phase composition and mechanical properties

X-ray diffraction patterns of the synthesized calcium magnesiumphosphate compounds are shown in Fig. 1. The results were in accor-dance with the Ca3(PO4)2–Mg3(PO4)2 phase diagram from Ando [27].The magnesium-free compound is phase-pure b-tricalcium phos-phate; at low magnesium contents (Mg/(Ca + Mg) < 0.15) magne-sium-substituted b-TCP was the predominant phase as indicatedby a shift of diffraction peaks to lower 2h values. A higher magnesiumcontent altered the crystalline structure such that at an equimolarCa:Mg ratio only the diffraction pattern of the Ca3Mg3(PO4)4 phasewas obtained. Further increasing the magnesium content resultedin the subsequent formation of farringtonite (Mg3(PO4)2).

These compounds were subsequently ground for 30–60 minresulting in medium particle sizes of 1.5–30 lm (Table 1), whereasparticle size increased with higher magnesium content. This waslikely a result of the higher hardness of the sintering cake, whichwas more dense for the magnesium-substituted compounds thanit was for pure b-TCP.

When mixed with an equimolar amount of MCPM, the magne-sium-substituted TCPs formed self-setting cements after adding a0.5 mol l�1 citric acid solution as cement liquid. Initial settingtimes of these cements were recorded either depending on the de-gree of magnesium substitution (Table 2) or the grinding time ofMg2.25Ca0.75(PO4)2 as the most promising compound (Table 3). Ce-ments made of magnesium-free TCP had a short setting time ofonly 2 min. Low magnesium contents strongly retarded settingsuch that samples made from Mg0.75Ca2.25(PO4)2 almost showedno hardening even after more than 60 min. Increasing the degreeof magnesium substitution subsequently decreased setting timeto a minimum of 11 min for Mg2.25Ca0.75(PO4)2, which could fur-ther be reduced to 8 min after 150 min grinding.

Wet compressive strength of cements set for 24 h are shown inFig. 2 depending on the magnesium content and the grinding timeof cement reactants. The magnesium-free reference cement had acompressive strength of approximately 12–19 MPa; similar resultswere observed for the compounds MgxCa(3–x)(PO4)2 with x > 1.5and a grinding time of 60 min. Higher strengths of up to 40 MPa werefound for Mg2.25Ca0.75(PO4)2 after grinding for 120 min. Surprisingly,low magnesium content in the range of 8–33% strongly reduced thecompressive strength to less than 2 MPa after 24 h setting.

XRD analyses (Fig. 3) revealed that cements made from magne-sium-free TCP were composed predominately of brushite with onlysmall amounts of unreacted TCP. Increasing the magnesium con-tent led to the appearance of newberyite (MgHPO4�3H2O) as a sec-ond crystalline phase.

3.2. Cell viability

The number of viable cells within the newberyite cultures in-creased continually over the course of 21 days. Cells cultured on tita-nium scaffolds as well as cells cultured on the bare culture dishesserved as positive control cultures. A scanning electron microscopyof the calcium magnesium phosphate cell culture scaffolds at theend of the course of 21 days is shown in Fig. 4. Compared to the po-sitive controls, the cell viability of the test specimens was a littleinferior but generally in the same order of magnitude (Fig. 5). Signif-icant differences could be observed infrequently (Fig. 5, asterisks).On day 21, the cell viability of newberyite was 84.9% and 76.6% com-pared with titanium and the bare culture dish, respectively. As ex-pected, the negative control cultures showed no cell growth due tothe cytotoxicity of copper.

3.3. Activity of alkaline phosphatase

The activity of ALP was correlated with the cell number (nmolconverted p-NPP min�1 cell�1). As there was no ALP detectableon day 5, a continual increase could be noticed on the subsequent

Table 2Initial setting times according to the Gilmore needle test of calcium magnesium phosphate cement matrices with a 0.5 mol l�1 citric acid solution as liquid and a powder-to-liquidratio of 3.0 g ml�1. The calcium magnesium phosphate raw materials were dry ground for 60 min and then mixed with an equimolar amount of MCPM.

Compound Mg3(PO4)2 Mg2.25Ca0.75(PO4)2 Mg1.5Ca1.5(PO4)2 Mg0.75Ca2.25(PO4)2 Ca3(PO4)2

Setting time (min) 21 11 25 >60 2

Table 3Setting times according to the Gilmore needle test of Mg2.25Ca0.75(PO4)2 after grindingfor up to 150 min.

Grinding time (min) – 30 60 90 120 150

Setting time (min) 60 32 11 10 9 8

Fig. 2. Wet compressive strength of calcium magnesium phosphate cements. Thecalcium magnesium phosphate raw materials were (A) either dry ground for 60(120) min or (B) ground for up to 120 min and then mixed with a equimolar amountof MCPM and a 0.5 mol l�1 citric acid solution as liquid at a powder-to-liquid ratioof 3.0 g ml�1. Data are expressed as mean of n = 6 specimens ± standard deviation.

Fig. 3. X-ray diffraction patterns of set cements after 24 h setting at 37 �C with0.5 mol l�1 citric acid as liquid, grinding time 120 min: the strongest reflectionpeaks of brushite (b) and newberyite (n) are marked.

Table 1Medium particle sizes d50 in lm before and after grinding.

Grinding time (min) d50 (lm)

Mg3(PO4)2 Mg2.25Ca0.75(PO4)2 Mg1.5Ca1.5(PO4)2 Mg0.75Ca2.25(PO4)2 Ca3(PO4)2

– 50.8 56.6 36.2 29.4 10.930 min 32.0 34.6 27.1 25.6 6.460 min 30.0 14.9 22.4 15.4 1.5

1532 U. Klammert et al. / Acta Biomaterialia 6 (2010) 1529–1535

days, apart from titanium on day 21 (Fig. 6). The ALP activity of thenewberyite cultures was most often significantly lower than ALP ofthe positive controls (Fig. 6, asterisks), but in a comparablemagnitude.

Fig. 4. Scanning electron microscopy of the calcium magnesium phosphate cellculture scaffolds at the end of the course of 21 days. To be seen are the cell-bearingscaffolds (arrows) in the background and the abundant osteoblastic cells in front(asterisks).

Fig. 5. Cell growth of MC3T3-E1 cultured on calcium magnesium phosphatescaffolds compared with positive (titanium and bare culture dishes) and negative(copper) controls. The differences are partly significant (p < 0.05) as indicated(asterisks). Results are displayed as mean of n = 12 specimens ± standard deviation.

Fig. 6. ALP activity of MC3T3-E1 cells cultured on calcium magnesium phosphatescaffolds compared with positive (titanium and bare culture dishes) and negative(copper) controls. The differences are partly significant (p < 0.05) as indicated(asterisks). Results are displayed as mean of n = 12 specimens ± standard deviation.

U. Klammert et al. / Acta Biomaterialia 6 (2010) 1529–1535 1533

3.4. Chemical analysis of the culture medium

The pH values of the newberyite cultures have never been out-side a physiological range and rated in the same magnitude com-pared with the control cultures without any significantdifferences (Fig. 7A). Compared with the control cultures as wellas with the pure cell culture medium (data not shown), the med-ium of the newberyite cultures presented significantly higher con-centrations of free PO3�

4 ions. The values were nearly constant andabout sixfold of the controls (Fig. 7B). The concentration of freeMg2+ ions of the newberyite cultures showed a slight increase untilday 21 (Fig. 7C); the differences to the controls were significant(about eightfold). The concentration of free Ca2+ ions presented aconverse behaviour: the values were degraded significantly, alto-gether approximately 65–80% of the controls (Fig. 7D). Further-more, a slight increase of free Ca2+ ions can be noticed over thecourse of 21 days.

4. Discussion and conclusion

Brushite-forming calcium phosphate cements are of clinicalinterest because of their dissolution under physiological condi-tions. Limiting factors for their extensive clinical application werein the past mainly short setting times, low mechanical strengthand a lack of sufficient fluidity to enable application through injec-tion cannulae. Successful attempts have been made in the past toprolong the setting time by the use of retardants, e.g. pyrophos-phates, sulfates or citrates [21]. Another setting retarder for brush-ite cements are magnesium ions, which were also shown tosuppress phase transformation of the setting product brushite tohydroxyapatite in vivo [22]. A previous study has used magne-sium-substituted hydroxyapatite powders as brushite cementreactants, which were prepared by precipitation from solutionwith a magnesium content (Mg/(Mg + Ca)) of approximately 5%[28]. Our current study followed a different approach by usingmagnesium-containing compounds with the general formulaMgxCa(3–x)(PO4)2 to introduce magnesium ions in the cement at abroad range of 0 < Mg/(Mg + Ca) < 1. These compounds were pre-pared by sintering of magnesium and calcium phosphate powdersources according to Eq. (1):

CaHPO4 þ CaCO3 þMgHPO4 � 3H2OþMgðOHÞ2!MgxCað3�xÞðPO4Þ2 ð1Þ

Depending on their stoichiometry, the sintered products wereeither (magnesium-substituted) tricalcium phosphate, farrington-ite (Mg3(PO4)2) or Ca3Mg3(PO4)4. All compounds formed hydraulicsetting cements when combined with acidic monocalcium phos-phate and an aqueous solution of citric acid and formed either purebrushite or a mixture of brushite and newberyite (MgHPO4�3H2O)as setting product, depending on the magnesium content of the ce-ment raw material. Setting times were found to vary over a broadrange of 10 to >60 min, whereas cements from long groundMg2.25Ca0.75(PO4)2 provided clinical acceptable setting times of8–10 min, which were in the same range as those obtained for acommercial brushite cement (chronOSTM inject), where a hardeningtime period of 6–12 min is given by the manufacturer [29] afterwhich the cement is set (6 min) and provides enough primary sta-bility to permit wound closure (12 min). The strengths of thesenewberyite cements in this study were in the range 30–40 MPawithout pre-compaction. All strengths were measured after 24 h

Fig. 7. Assessment of pH values of the cell culture medium (A). Merely slight differences compared to the control cultures without significance. The concentrations of freePO3�

4 (B) and Mg2+ (C) ions of the newberyite cultures were approximately, respectively, sixfold and eightfold higher than that of the controls. The concentration of free Ca2+

ions of the newberyite cultures was obviously lower than that of the controls (D). Results are displayed as mean of n = 12 specimens ± standard deviation.

1534 U. Klammert et al. / Acta Biomaterialia 6 (2010) 1529–1535

setting and incubation of the samples in PBS buffer without dryingthem before testing, thus strength values represented the clinicalattainable mechanical properties. Strength data for brushite ce-ments reported in literature are in the range 1–51 MPa for com-pressive strength [30]. However, caution must be exercised whencomparing literature strength values because of differences in testmethods and conditions. Many previous studies have used samplesthat were only stored in humid atmospheres or dried before test-ing. It is well known that the degree of hydration strongly affectsthe measured strength, e.g. Pittet and Lemaitre [31] showed thatstrengths were nearly double for dried samples compared withsamples stored in a high humidity atmosphere.

For evaluation of the biocompatibility of the calcium magnesiumphosphate cement, osteoblast-like cells were cultured on prefabri-cated scaffolds, since in vitro cell testing of biomaterials is a well-established method to determine a material’s cytocompatibility[32–34]. The cell proliferation and the expression of the typicalosteoblastic marker ALP served as parameters in our study. The re-sults show the proliferation of osteoblast cells and the expressionof a typical osteogenic marker in the context of positive and negativecontrols. The data of cells cultured on newberyite scaffolds werealtogether a little below the positive controls (bare cell culture dish,respectively, pure titanium, which is known to have a very suitable

biocompatibility on bone tissue). However, the calcium magnesiumphosphate cement enabled the osteoblast cells to proliferate and toexpress the differentiation marker ALP. Cytotoxic attributes, as dem-onstrated by the negative control specimens (copper), are unlikely.The slightly compromised proliferation and activity of ALP (partlysignificant) of the newberyite cell cultures could be caused by chem-ical features of the scaffold surface. Furthermore, the altered concen-trations of free ions, especially phosphate and magnesium, may playa role. However, these effects observed in vitro should be negligiblein vivo at a sufficiently vascularized implantation site with a con-stant fluid exchange. The altered concentrations of free phosphate,magnesium and calcium ions of the culture medium are due to thesolubility of the cement material. Consequently, free phosphateand magnesium are increased. The decreased concentration of freecalcium ions could be explained by an oversaturated solution ofphosphate, magnesium and calcium, leading to precipitation of low-er soluble calcium phosphates. However, the concentrations of freeelectrolytes remained nearly in the same magnitude over the course.This can additionally be explained by the circumstance that electro-lyte analyses were performed periodically every fourth day using theconsumed culture medium. Thus a further accumulation decrease offree electrolytes did not occur. Precipitation of lower soluble calciumphosphates results in a decrease of free calcium ions, because cal-

U. Klammert et al. / Acta Biomaterialia 6 (2010) 1529–1535 1535

cium ions originally of the culture medium are precipitated as wellby the abundant phosphate ions. Free magnesium ions are not de-creased because of the higher solubility of magnesium phosphates.Since magnesium acts as a crystallization inhibitor, the precipitationof amorphous calcium phosphates but not apatite is likely. Hence,the solubility of the cement is not reduced in contrast to pure brush-ite-based cements.

Due to the current results in vitro we conclude that the calciummagnesium phosphate cement introduced here provides promis-ing features for its use as scaffold for hard tissue regeneration. Fur-ther in vitro investigations are required concerning the mechanicalstrength over time as well as in vivo models to prove other biolog-ical features like resorption kinetics at the implantation site andosteoconductive properties.

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figs. 2, 4–7, are diffi-cult to interpret in black and white. The full colour images can befound in the on-line version, at doi:10.1016/j.actbio.2009.10.021.

References

[1] Le Geros RZ. Properties of osteoconductive biomaterials: calcium phosphates.Clin Orthop 2002;39:81–98.

[2] Rosen HM, Ackermann JL. Porous block hydroxyapatite in orthognathicsurgery. Angle Orthod 1991;61:185–91.

[3] Bohner M, Gbureck U, Barralet JE. Technological issues for the development ofmore efficient calcium phosphate bone cements: a critical assessment.Biomaterials 2005;26:6423–9.

[4] Dorozhkin SV. Calcium orthophosphate cements for biomedical applications. JMater Sci 2008;43:3028–57.

[5] Mirtchi AA, Lemaitre J, Terao N. Calcium phosphate cements: study of the beta-tricalcium phosphate–monocalcium phosphate system. Biomaterials 1989;10:475–80.

[6] Bajammal SS, Zlowodzki M, Lelwica A, Tornetta P, Einhorn TA, Buckley R, et al.The use of calcium phosphate bone cement in fracture treatment. J Bone JointSurg Am 2008;90A:1186–96.

[7] Oda H, Nakamura K, Matsushita T, Yamamoto S, Ishibashi H, Yamazaki T, et al.Clinical use of a newly developed calcium phosphate cement (XSB-671D). JOrthop Sci 2006;11:167–74.

[8] Mattsson P, Alberts A, Dahlberg G, Sohlman M, Hyldahl HC, Larsson S.Resorbable cement for the augmentation of internally-fixed unstabletrochanteric fractures – a prospective, randomised multicentre study. J BoneJoint Surg Brit 2005;87B:1203–9.

[9] Penel G, Leroy N, Van Landuyt P, Flautre B, Hardouin P, Lemaitre J, et al. Ramanmicrospectrometry studies of brushite cement: in vivo evaluation in a sheepmodel. Bone 1999;25:81S–4S.

[10] Fraysinnet P, Gineste L, Conte P, Fages J, Rouquet N. Short-term implantationeffects of a DCPD based calcium phosphate cement. Biomaterials 1998;19:971–7.

[11] Grover LM, Knowles JC, Fleming GJP, Barralet JE. In vitro dissolution of Brushitecement. Biomaterials 2003;24:4133–41.

[12] Marinno FT, Torres J, Tresguerres I, Jerez LB, Cabarcos EL. Vertical boneaugmentation with granulated brushite cement set in glycolic acid. J BiomedMater Res A 2007;81:93–102.

[13] Xia Z, Grover LM, Huang Y, Adamopoulos I, Gbureck U, Triffit JT, et al. In vitrobiodegradation of three brushite calcium phosphate cements by a macrophagecell-line. Biomaterials 2006;27:4557–65.

[14] Oberle A, Theiss F, Bohner M, Muller J, Kastner SB, Frei C, et al. Investigationabout the clinical use of brushite- and hydroxylapatite-cement in sheep.Schweizer Archiv Tierheilkunde 2005;147:482–90.

[15] Grover LM, Gbureck U, Tremayne MJ, Barralet JE. Biologically mediatedresorption of brushite cement in vitro. Biomaterials 2006;27:2178–85.

[16] Lemaitre J, Mirtchi A, Mortier A. Calcium phosphate cements for medical use:state of the art and perspectives of development. Silicates Ind 1987;10:141–6.

[17] Gbureck U, Dembski S, Thull R, Barralet JE. Factors influencing calciumphosphate cement shelf-life. Biomaterials 2005;26:3691–7.

[18] Grover LM, Gbureck U, Wright AJ, Barralet JE. Cements formed from the calciumphosphate–H2O–H3PO4–H4P2O7 system. J Am Ceram Soc 2005;88:3096–103.

[19] Lilley KJ, Gbureck U, Wright AJ, Knowles JC, Farrar DF, Barralet JE. Brushitecements from polyphosphoric acid, calcium phosphate systems. J Am CeramSoc 2007;90:1892–8.

[20] Tamimi-Marino F, Mastio J, Rueda C, Blanco L, Lopez-Cabarcos E. Increase ofthe final setting time of brushite cements by using chondroitin 4-sulfate andsilica gel. J Mater Sci: Mater Med 2007;18:1195–201.

[21] Bohner M, Lemaitre J, Ring TA. Effects of sulfate, pyrophosphate and citrateions on the physiochemical properties of cements made of b-tricalciumphosphate–phosphoric acid–water mixtures. J Am Ceram Soc 1996;79:1427–34.

[22] Rousseau S, Lemaitre J, Bohner M, Frei C. GRIBOI: Shanghai, China; 14–17 March,2002.

[23] Ginebra MP, Boltong MG, Driessens FCM, Bermudez O, Fernandez E, Planell JA.Preparation and properties of some magnesium-containing calcium phosphatecements. J Mater Sci: Mater Med 1994;5:103–7.

[24] Driessens FCM, Boltong MG, Zapatero MI, Verbeeck RMH, Bonfield W,Bermudez O, et al. In vivo behaviour of three calcium phosphate cementsand a magnesium phosphate cement. J Mater Sci: Mater Med 1995;6:272–8.

[25] ASTM-Standard C266-99: Standard test method for time of setting of hydrauliccement paste by Gilmore needles. ASTM International 2002.

[26] Ewald A, Ihde S. Salt impregnation of implant materials. Oral Surg Oral MedOral Pathol 2009;107:790–5.

[27] Ando J. Phase diagrams of Ca3(PO4)2–Mg3(PO4)2 and Ca3(PO4)2–CaNaPO4

systems. Bull Chem Soc Jpn 1958;31:201–4.[28] Lilley KJ, Gbureck U, Knowles JC, Farrar DF, Barralet JE. Cement from magnesium

substituted hydroxyapatite. J Mater Sci: Mater Med 2005;16:455–60.[29] chronOS inject user manual, see http://www.synthes.com/html/chronOS-

Inject.7253.0.html.[30] Hofmann M, Mohammed AR, Perrie Y, Gbureck U, Barralet JE. High strength

resorbable brushite bone cement with controlled drug releasing capabilities.Acta Biomater 2009;5:43–9.

[31] Pittet C, Lemaitre J. Mechanical characterisation of brushite cements: a Mohrcircles approach. J Biomed Mater Res 2000;53:769–80.

[32] Klammert U, Reuther T, Jahn C, Kübler AC, Gbureck U. Biocompatibility ofbrushite and monetite cell culture scaffolds made by 3D powder printing. ActaBiomater 2009;5:727–34.

[33] Jalota S, Bhaduri SB, Tas AC. In vitro testing of calcium phosphate (HA, TCP, andbiphasic HA–TCP) whiskers. J Biomed Mater Res 2006;78A:481–90.

[34] Clarke SA, Hoskins NL, Jordan GR, Henderson SA, Marsh DR. In vitro testing ofAdvanced JAXTM Bone Void Filler System: species differences in the response ofbone marrow stromal cells to b-tricalcium phosphate and carboxymethylcellulosegel. J Mater Sci: Mater Med 2007;18:2283–90.