lable at ScienceDirect

Biomaterials 34 (2013) 9863e9876

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

In vivo stimulation of bone formation by aluminum and oxygenplasma surface-modified magnesium implants

Hoi Man Wong a, d, 1, Ying Zhao a, b, d, 1, Vivian Tam a, Shuilin Wu b, Paul K. Chu b,Yufeng Zheng c, Michael Kai Tsun To a, d, Frankie K.L. Leung a, d, Keith D.K. Luk a,Kenneth M.C. Cheung a, Kelvin W.K. Yeung a, d, *

a Department of Orthopaedics and Traumatology, The University of Hong Kong, Pokfulam, Hong Kong, Chinab Department of Physics and Materials Science, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, Chinac Department of Materials Science and Engineering, Collage of Engineering, Peking University, Beijing 100871, Chinad Shenzhen Key Laboratory for Innovative Technology in Orthopaedic Trauma, The University of Hong Kong Shenzhen Hospital, 1 Haiyuan 1st Road, FutianDistrict, Shenzhen, China

a r t i c l e i n f o

Article history:Received 28 July 2013Accepted 19 August 2013Available online 20 September 2013

Keywords:Magnesium implantCyto-compatibilityOsteoblastPlasma surface treatment

* Corresponding author. Department of OrthopaeUniversity of Hong Kong, Pokfulam, Hong Kong, Chfax: þ852 2817 4392.

E-mail address: wkkyeung@hku.hk (K.W.K. Yeung1 The authors share the co-first authorship.

0142-9612/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.biomaterials.2013.08.052

a b s t r a c t

A newly developed magnesium implant is used to stimulate bone formation in vivo. The magnesiumimplant after undergoing dual aluminum and oxygen plasma implantation is able to suppress rapidcorrosion, leaching of magnesium ions, as well as hydrogen gas release from the biodegradable alloy insimulated body fluid (SBF). No released aluminum is detected from the SBF extract and enhancedcorrosion resistance properties are confirmed by electrochemical tests. In vitro studies reveal enhancedgrowth of GFP mouse osteoblasts on the aluminum oxide coated sample, but not on the untreatedsample. In addition to that a small amount (50 ppm) of magnesium ions can enhance osteogenicdifferentiation as reported previously, our present data show a low concentration of hydrogen can giverise to the same effect. To compare the bone volume change between the plasma-treated magnesiumimplant and untreated control, micro-computed tomography is performed and the plasma-treatedimplant is found to induce significant new bone formation adjacent to the implant from day 1 untilthe end of the animal study. On the contrary, bone loss is observed during the first week post-operation from the untreated magnesium sample. Owing to the protection offered by the Al2O3

layer, the plasma-treated implant degrades more slowly and the small amount of released magnesiumions stimulate new bone formation locally as revealed by histological analyses. Scanning electronmicroscopy discloses that the Al2O3 layer at the bone-implant interface is still present two monthsafter implantation. In addition, no inflammation or tissue necrosis is observed from both treated anduntreated implants. These promising results suggest that the plasma-treated magnesium implant canstimulate bone formation in vivo in a minimal invasive way and without causing post-operativecomplications.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Bone impairment arising from osteoporosis aswell as other bone-related diseases are a major public health problem [1,2]. Anti-catabolic drugs such as bisphosphonates (BPs) is a common

dics and Traumatology, Theina. Tel.: þ852 2255 4654;

).

All rights reserved.

medication to treat osteoporosis [3,4] and approximately 190millionpeople worldwide are administrated with this drug as a preventivemeasure against bone loss [5]. However, atypical bone fractures havebeen observed from patients after long-term therapies of BPs [6,7].Other complications including gastro-intestinal inflammation,osteonecrosis in jawbone have also been recorded [8,9]. Alterna-tively, other treatment options such as estrogen-replacementtherapy and calcitonin injection have been studied [10] but unfor-tunately, these therapeutic strategies share the same treatmentconcept as BPs targeting a smaller bone turnover rate during thebone remodeling process possibly resulting in the formation of morebrittle bone [11]. Other treatments by means of biological agents

Table 1Implantation and deposition conditions for implantingand depositing Al and O2 to form Al2O3.

Negative high voltage power supplyeimplantation

NH current 1.0 mANH voltage �15 kVPulse duration 300 msFrequency 10 Hz

Pulsed filtered cathodic arc sourceedeposition

Arc current 0.1 AArc voltage 92 VTriggering voltage 12.6 kVCoil current 2.3 APulse duration 250 msFrequency 10 Hz

0

10

20

30

40

50

60

70

80

90

100

Ato

mic C

on

c. (%

)

OAlMg

a

H.M. Wong et al. / Biomaterials 34 (2013) 9863e98769864

such as bone morphogenetic proteins (BMPs) and insulin-likegrowth factors can be theoretically applied to stimulate bone for-mation [12,13], but the effective dosage of these biological agents isunexpectedly high due to the fast degradation rate. Furthermore,they are easily deactivated by either enzymes or some physical andchemical reactions [14,15]. Hence, owing to the uncontrolled releasepattern and systemic side effects in musculoskeletal tissues [16,17],the development of new treatment for osteoporosis is slow [11].Recently, administration of parathyroid hormone (PTH) has beenproposed for osteoporosis and osteoporotic fracture healing [18,19].Despite the increased new bone formation in osteoporotic patients,an adverse effect related to the development of osteogenic sarcomasbased on animal studies has been reported [20].

In this work, a simple approach to stimulate bone formation byusing a newly developed magnesium implant is described. Mag-nesium is chosen as one of the major components in the implantbecause controlled release of magnesium ions (Mg2þ) upondegradation has been found to benefit bone growth [21,22]. Addi-tionally, magnesium ions are able to stimulate bone growth andbone healing by enhancing osteoblast and osteoclast activities[23,24]. However, the rapid degradation rate of magnesium andrelease of hydrogen gas hamper wider medical applications [25e27] and therefore, precise control of the degradation rate iscrucial to clinical adoption [28,29].

Aluminum oxide has been widely adopted as the articulatingsurface in total hip prostheses and dental prostheses due to its

Fig. 1. Implantation of the untreated and plasma-treated implants in the lateral epi-condyle of SD rats after 2 months. Black arrow shows the implant position.

chemical inertness, high strength and hardness, as well as resis-tance to wear and corrosion [30e34]. However, the use of thiscoating on magnesium implant to suppress release of magnesiumions and integration between magnesium implant and bone havehitherto not been reported. Here, by means of dual aluminum andoxygen plasma immersion ion implantation and deposition(PIII&D), a robust coating is produced on a magnesium implant tocontrol the degradation rate [35]. The aluminum oxide (Al2O3eMg)coated implant is found to stimulate new bone formation as a resultof the controlled release of magnesium. The crux of the work is toimplant the plasma-treated magnesium implant into the site withbone loss by minimal invasive surgery to increase the bone massand density in the particular space. After new bone formation, therisk of bone fracture is reduced and our results indeed revealstimulation of bone growth on the plasma-treated magnesiumimplant in vivo.

2. Materials and methods

2.1. Sample preparation

An AZ91 magnesium ingot containing 9 wt% aluminum and 1 wt% zinc (JiaozuoCity Anxin Magnesium Alloys Scientific Technology Co., Ltd.) was cut into disks

0 40 80 120 160 200 240 280 320

Depth / nm

0 50 100 150 200 250 300 350 400

0

10

20

30

40

50

60

70

80

Ato

mic C

on

c. (%

)

Depth (nm)

O Al Mg

b

Fig. 2. a Depth profile of the untreated magnesium alloy. b Depth profile of thealuminum oxide plasma-treated magnesium alloy.

H.M. Wong et al. / Biomaterials 34 (2013) 9863e9876 9865

(5 mm in diameter and 4 mm thick) and rods (2 mm in diameter and 6 mm long). Allthe samples were ground and polished with sand paper to remove the surface oxide.The disks and rods were ultrasonically cleaned with ethanol prior to PIII&D and theimportant PIII&D parameters are listed in Table 1. All the samples were cleaned withethanol after PIII&D before surface characterization and other tests.

2.2. Surface composition and depth profile analyses

The surface chemical compositions and depth profiles were determined by X-rayphotoelectron spectroscopy (XPS, Physical electronics PHI 5802) using Al Ka irra-diation at an estimated sputtering rate of 5.3 nm per minute. High-resolution XPSspectra were acquired at different sputtered depths to investigate the chemicalstates and the binding energies were referenced to C 1s at 284.5 eV.

2.3. Corrosion resistance tests

2.3.1. Immersion testsImmersion tests were carried out at different time points to monitor the

degradation and release of magnesium ions from the plasma-treated and untreatedsamples. Four each of the plasma-treated and untreated disk samples were indi-vidually immersed in sealable capsules containing 10 ml of simulated body fluid(SBF) and then incubated at 37 �C for a total of 30 days. Leaching of magnesium ionsfrom the samples was monitored at 5 different time points of 1, 4, 7, 14, and 30 daysusing inductively-coupled plasma optical emission spectrometry (ICP-OES) (PerkinElmer, Optima 2100DV). The pH values of the samples weremeasured and the rate ofcorrosionwas determined by measuring the weight loss of the samples. The surfacemorphology was examined by scanning electron microscopy (SEM) before immer-sion as well as after 3 h and 28 days. The corrosion product and its composition afterimmersion for 28 days were analyzed by energy-dispersive X-ray spectroscopy (EDS,Leo 1530 FEG SEM equipped with Oxford Instruments, INCAx-sight EDS Detectors).Owing to the concern on ion toxicity, the Al ion concentrations were determined byusing ICP-OES after immersion for 7, 14, and 28 days.

2.3.2. Electrochemical testsThe electrochemical experiments were conducted an acellular simulated body

fluid (SBF) at a pH of 7.40 (ion concentration of Naþ 142.0, Kþ 5.0, Mg2þ 1.5, Ca2þ

2.5, Cl� 147.8, HCO�3 4.2, HPO2�

4 1.0, and SO2�4 0.5 mM) on a Zahner Zennium

electrochemical workstation using the three-electrode technique. The potentialwas referenced to a saturated calomel electrode (SCE) and a platinum sheet servedas the counter electrode. The samples were immersed in the SBF at 37 �C and EISwas carried out after stabilization in the solution for 5 min. The data wererecorded from 100 kHz to 100 mHz with a 5 mV sinusoidal perturbing signal at the

Fig. 3. High-resolution XPS spectra of O 1s and Al 2p obtained after the

open-circuit potential. The polarization curves were acquired by scanning thepotential at a rate of 1 mV/s from �300 mV to 600 mV following the EISmeasurement.

2.4. Cell culture

MC3T3-E1 pre-osteoblasts and enhanced green fluorescent protein osteoblasts(eGFPOB) fromGFPmicewere cultured in the DMEM culturemedium supplementedwith 10% (v/v) fetal bovine serum (FBS, Biowest, France), antibiotics (100 U/ml ofpenicillin and 100 mg/ml of streptomycin), and 2 mM L-glutamine and incubated at37 �C in an atmosphere of 5% CO2 and 95% air. The same culturing conditions wereused in all the experiments.

2.4.1. Cell viability in H2 enriched mediumHydrogen gas is formed as a by-product during magnesium degradation thus

there is a need to study its biological effect on osteoblasts. The MTT assay was usedto measure osteoblast viability in the presence of the hydrogen-enriched DMEMculture medium. The enriched medium was prepared by dissolving 40 ml ofhydrogen in 60 ml of the DMEM at a pressure of 0.1 MPa and then stabilized for 3days at �20 �C. The hydrogen level was measured by gas chromatography (GC-TCDHP 5890, USA) and the enriched media was used to culture MC3T3-E1 pre-osteoblasts for the MTT assay.

In brief, MC3T3-E1 pre-osteoblasts (7� 104 cells/cm2) were cultured for one dayon a 96-well tissue culture plate, after which the culture media in each well wasreplaced with the magnesium-supplemented DMEM and incubated for 3 days. 10 mlof the MTT solution (5 mg thiazolyl blue tetrazolium bromide powder in 1 ml ofphosphate buffered saline (PBS, OXOID Limited, England)) was added to each welland further incubated for one day after which 100 ml of 10% sodium dodecyl sulfate(SDS, Sigma, USA) in 0.01 M hydrochloric acid was added to each well and incubatedfor a further 18 h. The absorbance was recorded by the multimode detector (Beck-man Coulter DTX 880) at a wavelength of 570 nm, and the reference wavelength of640 nm was used to determine the cell viability in comparison to the control. Thepercentage cell viability was calculated by dividing the absorbance values of thesamples to that of the control.

2.4.2. Alkaline phosphatase (ALP) activity in H2 enriched mediumThe ability of MC3T3-E1 pre-osteoblasts to undergo osteoblastic differentia-

tion in the hydrogen-enriched medium was studied. The MC3T3-E1 pre-osteoblasts (1.4 � 104 cells/cm2) were cultured in DMEM for one day in a 24-well tissue culture plate. The culture medium in each well was subsequentlyreplaced with hydrogen-enriched DMEM and cultured for 3, 7, or 14 days. Afterincubation, the cells were washed with PBS three times and lysed with 0.1% Triton

aluminum oxide plasma treatment (sputtering rate 5.29 nm/min).

H.M. Wong et al. / Biomaterials 34 (2013) 9863e98769866

X-100 at 4 �C for 30 min. The cell lysates were centrifuged at 574 g at 4 �C for10 min (2-5 Sartorius, Sigma, USA), and 10 ml of the supernatant from each samplewas transferred to a 96-well tissue culture plate. The ALP activity was determinedby a colorimetric assay using an ALP reagent containing p-nitrophenyl phosphate(p-NPP) (Stanbio, USA) as the substrate. The absorbance was recorded by themultimode detector (Beckman Coulter DTX 880) at a wavelength of 405 nm. TheALP activity was normalized to the total protein level of the samples measured bythe Bio-Rad Protein Assay (Bio-Rad, USA) and the data were expressed as thespecific ALP activity per unit of protein.

2.4.3. Cytocompatibility test of Al2O3eMgThe eGFP osteoblasts fromGFP mice were cultured on the surface of the plasma-

treated samples to determine the growth and cytocompatibility. The samples wereplaced on a 96-well plate and 1.7 � 104 cells/cm2 GFPOB were seeded on eachsample and cultured for 1 or 3 days. After 1 or 3 days, the cell morphology wasobserved by fluorescence microscopy (Niko ECL IPSE 80i, Japan). The attached livingeGFP-expressive osteoblasts were visualized using a 450e490 nm incident filter, andthe fluorescence images emitted at 510 nmwere captured by a Sony DKS-ST5 digitalcamera.

2.5. In vivo animal study

2.5.1. Surgical proceduresThe anesthetic, surgical, and post-operative care protocols were examined by

and fulfilled the requirements of the University Ethics Committee of The

0 5 10 15 20 25 30 35

0

100

200

300

400

500

600

700

800

900

1000

1100

Untreated Mg

Al2

O3-treated Mg

Mg

io

n c

on

ce

ntra

tio

n in

S

BF

(p

pm

)

Day

*

0 5 10

0.000

0.002

0.004

0.006

0.008

0.010

0.012

0.014

0.016

Untreated

Al2O

3-tre

Weig

ht L

oss fro

m sam

ple (g

)

*

a

c

Fig. 4. a Magnesium ions concentration of the immersion extract from Al2O3-plasma treatbetween treated and untreated sample (p < 0.05). b pH values of the immersion extract frsignificant difference between treated and untreated sample (p < 0.05). c Total weight losignificant difference between treated and untreated sample (p < 0.05).

University of Hong Kong as well as Licensing Office of the Department of Healthof the Hong Kong Government. Ten two-month old female SpragueeDawley rats(SD rats) weighing between 200 and 250 g from the Laboratory Animal Unit ofThe University of Hong Kong were used. The rats were anaesthetized with ke-tamine (67 mg/kg) and xylazine (6 mg/kg) by intraperitoneal injection. Theoperation sites of the rats were shaved and sterilized, and followed by decorti-cation. A 2 mm diameter and 6 mm deep hole was made by a hand driller at thelateral epicondyle by a minimally invasive approach. Subsequently, the plasmatreated or untreated magnesium implants were implanted into the preparedholes on either the left or right femur of the rats (Fig. 1). The wound was suturedlayer by layer and a proper dressing was applied over the incision. After theoperation, all the rats received subcutaneous injection of 1 mg/kg terramycin(antibiotics) and 0.5 mg/kg ketoprofen. The rats were euthanized 2 months post-surgery.

2.5.2. Micro-CT evaluationIn order to monitor the in vivo degradation of the plasma-treated and untreated

implants and new bone formation around the implants, serial time points of 1, 2, 3,4, and 8 weeks were set. At each respective time point, the rats were scanned usingthe micro-CT device (SKYSCAN 1076, Skyscan Company) to view the extent ofcorrosion on the samples and new bone formation. After the 2D planes werereconstructed using the NRecon (Skyscan Company), the 3D models were generatedby CTVol (Skyscan Company). The residual implant was then analyzed using theCTAn program (Skyscan Company) which was used to examine the micro-CT datasets for morphometry and densitometry as well as new bone growth.

0 5 10 15 20 25 30 35

7.5

8.0

8.5

9.0

9.5

Untreated Mg

Al2O

3-treated Mg

pH

v

alu

es

Day

*

15 20 25 30 35

Mg

ated Mg

Day

*

b

ed and untreated AZ 91 magnesium alloy over time. *denotes a significant differenceom Al2O3-plasma treated and untreated AZ 91 magnesium alloy over time. *denotes ast from Al2O3-plasma treated and untreated magnesium alloy over time. *denotes a

H.M. Wong et al. / Biomaterials 34 (2013) 9863e9876 9867

2.5.3. Analysis of the magnesium ion concentration in bloodBlood was collected prior to surgery, and at 1, 2, 3, 4 and 8 weeks post-operation

to determine the magnesium ion concentration. The blood was centrifuged at1,339 g for 15 min at room temperature (2-5 Sartorius, Sigma, USA) and the serawere collected and stored at 4 �C. Prior to the analysis, the serawere diluted 10 timesin double distilled water. The serum magnesium ion concentration was determinedusing ICP-OES (Optical Emission Spectrometer, Perkin Elmer, Optima 2100DV) andthe concentrations in the animals with treated and untreated samples werecompared.

2.5.4. Histological analysisThe rats were euthanized two months after operation and the bone samples

underwent hard tissue processing as previously described [36]. Briefly, the implantswere harvested and fixed in 10% buffered formalin for 3 days. The tissues weredehydrated stepwise for 3 days in 70%, 95%, or 100% ethanol, followed by incubationin xylene for 3 days. Finally, all the samples were embedded in methyl-methacrylate(Technovit 9100 New�, Heraeus Kulzer, Hanau, Germany) according to the manu-facturer’s instructions. The embedded samples were cut into 200 mm sections andthen micro-ground to a thickness of 50e70 mm before Giemsa staining (MERCK,Germany). The morphological and histological analyses were performed using op-tical microscopy and scanning electron microscopy to observe bone on-growth andintegration with the host tissue. In addition, EDS (Leo 1530 FEG SEM equipped withOxford Instruments, INCAx-sight EDS Detectors) in conjunction with elementalmapping was performed to determine the surface composition of the treated anduntreated implants 2 months after implantation.

3. Results

3.1. Surface composition and depth profile analyses

The surface chemical composition and elemental depth profilesof the untreated and plasma-treated magnesium alloys are shownin Fig. 2a and b, respectively. The untreated magnesium alloy has anatural thin oxide layer about 220 nm thick. Fig. 2b shows thepresence of both aluminum and oxygen after the surface treatment.The highest aluminum concentration is approximately 50 nm andthe penetration depth of Al is about 250 nm. The high-resolutionXPS O 1s and Al 2p spectra at different depths are shown in Fig. 3.

Fig. 5. Surface morphologies of the (a)(b) untreated and (c)(d) Al2O3-plasma treated magne(SEM). Cracks (red arrow) can be observed on the untreated sample after immersion. (For inweb version of this article.)

The O 1s peak at 531 eV and Al 2p peak at 74.5 eV correspond toAl2O3 formation suggesting formation of surface aluminum oxide.

3.2. Corrosion resistance tests

3.2.1. Immersion testThe immersion test is one of the direct ways to determine the

corrosion resistance properties. Fig. 4a shows the amount of mag-nesium ions release from the plasma-treated and untreated mag-nesium alloy. The amount of magnesium released from theuntreated sample is significantly higher (p < 0.05) at days 1 and 4(w201 ppm and 327 ppm, respectively) compared to the plasma-treated samples (177 ppm and 273 ppm, respectively). However,no significant difference is found between the untreated andplasma-treated samples between days 4 and 30. Moreover, nodetectable released aluminum ions are found from the SBF afterincubation for 7, 14, and 28 days. The corresponding pH values areshown in Fig. 4b. The pH increases rapidly fromdays 1e14, but slowsdown at day 30when the pH values range between 8.12 and 9.47 forthe untreated and 8.15 to 9.43 for the plasma-treated samples.

Fig. 4c shows the weight losses over time and correlation withthe degradation rate. The weight losses increase with incubationtime and are found to be significant at days 4 and 14. Larger lossesare observed from the untreated samples than treated samplesthroughout the immersion period. At day 30, the total weight lostfrom the untreated and plasma-treated samples are approximately13 mg and 11.5 mg, respectively.

Fig. 5 depicts the surface morphology of the untreated andplasma-treated samples after SBF immersion for 3 h. Localizedcorrosion with a large amount of cracks (red arrow) is observedfrom the untreated sample whereas no corrosion and cracks areobserved from the plasma-treated sample. The white areas on bothsamples represent the phases of the AZ 91 magnesium alloys.Fig. 6a shows the surface morphology of the untreated and treated

sium alloy after 3 h of SBF immersion by viewing under scanning electron microscopyterpretation of the references to color in this figure legend, the reader is referred to the

Fig. 6. a Surface morphologies of the untreated and Al2O3-plasma treated magnesium alloy before and after 28 days of SBF immersion by viewing under scanning electron mi-croscopy (SEM). b Composition of the corrosion products examined under energy-dispersive X-ray spectroscopy (EDS). The corrosion products were found to mainly containmagnesium (Mg) and oxygen (O).

H.M. Wong et al. / Biomaterials 34 (2013) 9863e98769868

90

100

110

120

a

H.M. Wong et al. / Biomaterials 34 (2013) 9863e9876 9869

samples before and after immersion in SBF for 28 days whereasFig. 6b shows the corrosion products examined by EDS. Severecorrosion is observed from the untreated sample compared to thetreated sample. The corrosion products which consist of mainlymagnesium and oxygen are likely magnesium oxide and magne-sium hydroxide. Calcium and phosphate are also detected fromboth the treated and untreated samples and sodium chloride (NaCl)is also present as an impurity from the SBF.

3.2.2. Electrochemical measurementsThe electrochemical polarization curves of the untreated and

plasma-treated samples are shown in Fig. 7a. The cathodic polari-zation curve is considered to represent the cathodic hydrogenevolution through water reduction, whereas the anodic polariza-tion curve represents dissolution of magnesium. The corrosionpotential (Ecorr) of the plasma-treated sample shifts the open-cir-cuit potential to a slightly more positive one. Although the differ-ence in the Ecorr between the treated and untreated sample is nothuge, the corrosion current (Icorr) of the plasma-treated samples isat least 2-fold lower than that of the untreated sample.

The representative EIS spectra (Nyquist plots) acquired from theuntreated andAl2O3-treated sample after immersion in SBF for 5minare shown in Fig. 7b. The capacitive arc at high frequencies results

Fig. 7. a Polarization curves of untreated and Al2O3-treated magnesium alloy in SBF. bEIS spectra of untreated and Al2O3-treated magnesium alloy after soaking in SBF for5 min.

from charge transfer and that arc at medium or low frequenciesresults from the effects of the surface film. It is obvious that after ionimplantation, the capacitive arcs are evidently enlarged. Moreover,an inductive arc is visible in the low frequency region. The inductivearc is usually related to the formation, adsorption, and desorption ofcorrosion products on the surface. It is known that a larger diameterarc represents better corrosion resistance and hence, both polari-zation and EIS results indicate that ion implantation appreciablyimproves the corrosion resistance of the AZ91 magnesium alloy.

3.3. In vitro studies

3.3.1. Effects of hydrogen on osteoblastic activityThe hydrogen concentration in the hydrogen-enriched medium

is approximately 1.85% but no H2 is found from the normal

Normal DMEM H rich DMEM

0

10

20

30

40

50

60

70

80

2

Cell V

iab

ility (%

)

0

50

100

150

200

250

300

AL

P a

ctiv

ity

(U

/mg

pro

te

in)

Day3

Day7

Day14

*

H2

rich DMEMNormal DMEM

b

Fig. 8. a Cell viabilities of MC3T3-El pre-osteoblasts cultured in medium with thehydrogen-enriched medium. The absorbance was detected at a wavelength of 570 nmwith a reference wavelength of 640 nm to determine the cell viability in comparison tothe cell cultured in normal medium. The percentage cell viability was calculated bydividing the absorbance values of the samples to the control. b The specific ALP ac-tivities of MC3T3-El pre-osteoblasts cultured in medium with the hydrogen-enrichedmedium on Day 3, Day 7 and Day 14. The readings were detected under the absor-bance reading at 570 nm wavelength and the ALP activity was normalized to the totalprotein level of the samples. *denotes a significant difference between hydrogen-enriched medium and the normal medium (p < 0.05).

H.M. Wong et al. / Biomaterials 34 (2013) 9863e98769870

medium. The hydrogen-enriched medium is well tolerated by thepre-osteoblasts with more than 90% viability and it is not signifi-cantly different from the control (normal DMEM, Fig. 8a). Fig. 8bpresents the specific ALP activities of the cell cultured in thehydrogen-enriched and normal media. The peak specific ALP ac-tivity is found on day 7 and significantly different (p < 0.05) with273.3 U/mg protein for the cells cultured in the hydrogen-enrichedmedium and 185.5 U/mg protein for the cells cultured in the normalmedium. However, this is not significant at days 3 and 14.

3.3.2. Cyto-compatibilityFig. 9 shows the viable cells on the untreated and plasma-

treated samples after culturing for 1 and 3 days. On day 1, cellspreading is observed from the plasma-treated sample, but no cellattachment is observed from the untreated sample. After 3 days,the GFP osteoblasts exhibit good spreading and grow to 100%confluency on the plasma-treated sample, whereas no cell growthis observed from the untreated sample.

3.4. In vivo animal study

3.4.1. Micro-computed tomography analysisNew bone formation and implant volume reduction post-

operation are studied at selected time points using micro-computed tomography. Fig. 10a depicts the cross sections of thefemur with the implant 1, 2, 3, 4, and 8 weeks after surgery. All theimplants show direct contact with the newly formed bone after 2weeks. The percentage changes in the bone volume and implantvolume during the period are shown in Figs. 10b and 11, respec-tively. More than 54% bone resorption is found from the untreatedimplant after 1 week, whereas there is an approximate 46% in-crease in bone volume on the plasma-treated implant. An increaseof 138% in the bone volume is found from the plasma-treated

Fig. 9. Microscopic views of GFP mouse osteoblasts cultured on untreated (a)(c) and Al2O3-the Al2O3-plasma treated magnesium alloys. 5,000 GFPOB were cultured on the untreated

implant at week 8 compared to the untreated implant. After 4weeks, the implant volume reduction of the untreated implant issignificantly greater than that in the treated implant (97% and 99%,respectively). After 8 weeks, the volume of the untreated implantfurther drops to 93% compared to 98% on the plasma-treatedimplant.

Fig. 12 shows the 3D models of the newly formed bone on theuntreated and plasma-treated implants after 2 months. The un-treated implant shows significantly less new bone formation(1.06 mm3) than the plasma-treated implant (7.11 mm3).

3.4.2. Serum magnesium measurementsFig. 13 shows the percentage changes in the serum magnesium

levels of the rats during the 8 weeks post implantation. The serummagnesium concentrations in the rats implanted with either theuntreated or plasma-treated implants fluctuate between �7% and14%, which is not significant.

3.4.3. Histological evaluationFig. 14 shows the tissue response to the untreated and plasma-

treated implants two months after implantation at different mag-nifications using Giemsa staining. New bone tissues (black arrows)form around the implants and all the implants show direct contactwith the newly formed bone. More new bone is observed aroundthe plasma-treated implant fromwhere osteoblast-like cells can beobserved (red arrows).

Fig. 15 shows the histological analysis of the plasma-treatedimplant by SEM and EDS elemental maps. The elements foundaround the implants are shown semi-quantitatively. Bothaluminum (Al) and oxygen (O) are detected from the untreated andplasma-treated magnesium implants, with larger amounts ofaluminum (Al) and oxygen (O) detected from the edge of thetreated magnesium implant, indicating that the plasma treatment

treated (b)(d) magnesium alloy after 1 and 3 days to evaluate the cyto-compatibility ofand treated samples.

Fig. 10. a Micro-CT reconstruction images of the lateral epicondyle containing (a) untreated and (b) Al2O3-plasma treated implants immediately after surgery and 1, 2, 3, 4, and 8weeks of post-operation. New bone formation (yellow arrow) can be observed progressively through the time points. b The percentage changes in bone volume around theuntreated and Al2O3-treated implants immediately after surgery and 1, 2, 3, 4, and 8 weeks of post-operation. More than 50% bone resorption was observed on the untreatedimplant after 1 week of post-operation but not for the Al2O3-treated implant throughout the implantation period. (For interpretation of the references to color in this figure legend,the reader is referred to the web version of this article.)

H.M. Wong et al. / Biomaterials 34 (2013) 9863e9876 9871

Pre-o

p

Week 1

Week 2

Week 3

Week 4

Week 8

-40

-30

-20

-10

0

10

20

30

40

% c

ha

ng

e o

f s

eru

m m

ag

ne

siu

m

Implantation time

Untreated

Al2O

3-treated

Fig. 13. Percentage changes in serum magnesium levels before and after implantation.Whole blood was separated by centrifugation and the serum isolated and collected foranalysis of serum Mg levels. The magnesium ion concentration was determined byinductively-coupled plasma optical emission spectrometry (ICP-OES).

Day 0 Week 4 Week 8

0

90

92

94

96

98

100

% c

han

ge

in

im

pla

nt v

olu

me

Implantation time

Untreated

Al2O

3-treated

* *

Fig. 11. The implant volume of the untreated and Al2O3-treated implants immediatelyafter surgery and 4, 8 weeks of post-operation. The implant volume of the untreatedimplant dropped to 97% and 93% after 4 and 8 weeks post implantation, respectively,whereas the implant volume of Al2O3-treated implant was maintained above 98% after8 weeks. *denotes significantly different (p < 0.05).

H.M. Wong et al. / Biomaterials 34 (2013) 9863e98769872

effects still exist after 2 months (green arrow). Moreover, calcium(Ca) and phosphorus (P) are found from both the untreated andtreated implants suggesting new bone formation.

4. Discussion

Biodegradable magnesium-based materials were first intro-duced for orthopedic applications in the beginning of the 20thcentury [37,38]. As Mg has been reported to be beneficial for boneformation [21,22], there is the idea of using a magnesium implantas an alternative treatment for bone loss in order to preventfracture. However, rapid degradation of the magnesium implantand subsequent release of hydrogen are major concerns [25,26].Hence, the clinical use of magnesium-based materials requiresmodification to decrease the degradation and hydrogen releaserates. Plasma immersion ion implantation and deposition (PIII&D)is one such approach to achieve this objective due to its uniquecharacteristics [39]. Al2O3 is extensively used on dental implantsand prostheses due to its excellent wear and corrosion resistance[40e42]. In this study, the corrosion resistance properties and

Fig. 12. Micro-CT 3D reconstruction models of newly formed bone (white in color) aroundmonths post-operation.

cyto-compatibility of the aluminum and oxygen PIII&DMg implantare assessed systematically.

The SBF immersion test indicates that there is significantly moremagnesium leached from the untreated magnesium alloys duringthe early time points. As reported by Song et al. [43], the corrosionresistance of the AZ91 magnesium alloy largely depends on thevolume fractions of the a and b phases. Since the corrosion currentdensity of the b phase Mg17Al12 is much smaller than that of the aphase, a more substantial and continuous b phase formed on themagnesium alloys should inhibit corrosion. Therefore, on accountof the larger amount of the continuous b phase on the Al2O3-treatedsample than the untreated sample, the corrosion resistance of thetreated sample is superior especially in the early time points.

Without the protection of a plasma coating, the lack of cellattachment on the untreated magnesium alloys may be due tocontinuous oxidation during rapid corrosion. This correlates withlow cell viability when cells are cultured on the untreated mag-nesium samples. Moreover, it is consistent with the cell viabilityobserved for different Mg ion concentrations as reported in ourprevious study [44], in which larger amounts of Mg ion result inlower cell viability. On the other hand, the enhanced corrosion

the implant (gray) on both untreated (a) and Al2O3-plasma treated (b) implants after 2

Fig. 14. Giemsa-stained hard tissue sections of the implants after 8 weeks’ implantation in the lateral epicondyle. Black arrows represent the newly formed bone and red arrowsrepresent the presence of osteoblast-like cells. (a) Untreated and (b) Al2O3-treated implants. (For interpretation of the references to color in this figure legend, the reader is referredto the web version of this article.)

H.M. Wong et al. / Biomaterials 34 (2013) 9863e9876 9873

resistance of the plasma-treated samples results in better cellattachment and growth as smaller amounts of Mg ion are releasedduring corrosion. However, the amounts of Mg ion leached fromthe untreated and plasma-treated samples are similar at the latertime points in the immersion test and can be explained by theformation of surface magnesium oxide and magnesium hydroxideon the untreated sample [28].

The plasma-treated Mg implant is able to stimulate new boneformation. Micro-computed tomography is used to visualize thecorrosion morphology of the implants and newly formed bonearound the implants, in addition to quantifying the in vivo corrosionrate. A high percentage of bone loss is found from the animalsimplanted with the untreated implants 1 week post-operation. Onthe contrary, it is not observed from the animals implantedwith theplasma-treated implants due to the slower corrosion rate. Rapidcorrosion of the untreated magnesium alloy in close contact withbone marrow after implantation results in a large amount ofleached magnesium which inhibits the growth of osteoblasts andconsequently bone formation [45]. However, owing to the forma-tion of the protective magnesium oxide and magnesium hydroxidelayer after 1 week, the corrosion rate is retarded and bone volumeincreases again afterwards [28]. On account of the protectionrendered by the aluminum oxide layer, no bone loss is observed 1week after implantation and the rate of new bone formation con-tinues to increase to 148% after 8 weeks. This also correlates withthe amount of magnesium released from the implant since lowconcentrations of magnesium have been reported to enhance theosteoblastic activity and generate a stimulatory effect on thegrowth of new bone tissues [28,46]. Our previous study [44] hasalso identified that a certain amount of magnesium ions (i.e.50 ppm) can stimulate pre-osteobalst differentiation whereasdown-regulation of osteogenic differentiation genes such as ALP,

Runx2, OPN and Type I collagen is observed at higher magnesiumion concentrations. This shows the importance of controlling therelease of magnesium ions and also explains the reason why boneloss is found from the untreated sample in the early stage afterimplantation but on the other hand, bone formation is alreadyobserved in this early stage from the Al2O3eMg implant. Further-more, the volume reduction of the plasma-treated implant is 5%less than the untreated implant, further suggesting that a smallamount of magnesium leached from the plasma-treated implant.These findings provide evidence that Al and O PIII&D can preventpost-operative bone loss and enhance new bone formation aftersurgery.

A cause of concern for the use of magnesium-based implants isthe release of hydrogen gas during corrosion and subsequentaccumulation especially in the early stages thereby inhibiting bonehealing [26]. The formation of gas bubbles has been previouslyreported to occur within one month after implantation [28,47,48].However, based on our previous findings [36], the hydrogen releaserate is related to the animal model used, implant size, as well assurgical site. Reduction in the corrosion rate may also decreasehydrogen release to a level that can be naturally absorbed by thebody without accumulation at the implant site. The presence ofhydrogen in the medium gives rise to a significantly higher level ofspecific ALP activity and hence, if the degradation rate is reduced inconjunction with low level of hydrogen release, bone formation isstimulated.

In addition to the micro-CT analysis, the histological analysisreveals more new bone formation around the plasma-treatedimplant compared to the untreated implant. Although less newbone is formed around the untreated implant, no inflammationor necrosis is observed from the animals implanted with eitherthe treated or untreated implants. This is in line with our

Fig. 15. Histology of untreated (A) and Al2O3-treated (B) implants viewed under scanning electron spectroscopy. Element mapping was conducted in the red rectangular box in (a)as magnified in (b) and the examined elements were as shown in (c). Aluminum and oxygen can be detected on the Al2O3-treated implant surface as indicated by green arrow. (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

H.M. Wong et al. / Biomaterials 34 (2013) 9863e98769874

previous findings [36]. Aluminum oxide is still detected 2 monthsafter implantation, showing that the degradation rate ofaluminum oxide is not large. It is noted that aluminum is sus-pected to have adverse effects on the nervous systems and maycause Alzheimer’s disease, but according to our present knowl-edge [49,50], it is still controversial as there is no direct andunequivocal proof that aluminum is related to the disease.Additionally, according to our SBF immersion tests, no releasedaluminum is detected even after 28 days of immersion, thereby

suggesting that the degradation rate is so low that there are nocytotoxic effects.

As the magnesium ion concentration is regulated by thekidneys, severe side-effects associated with a high level ofmagnesium are extremely rare [51]. The serum magnesiumconcentration measured in this study is within the normalphysiological ranges [52e54]. Hence, even as the surface coatinggradually degrades, there is no toxic effect to the surroundingtissues. All in all, reduction of Mg release in vitro and in vivo is

H.M. Wong et al. / Biomaterials 34 (2013) 9863e9876 9875

crucial to stimulate bone formation. With these promising re-sults, the use of the plasma-treated magnesium implant to treatbone loss provides better treatment results without majorshortcomings.

5. Conclusion

Dual aluminum and oxygen plasma treatment is conducted onmagnesium alloys to enhance the surface properties. After thesurface treatment, the corrosion resistance and biological perfor-mance are enhanced both in vitro and in vivo. The results providethe foundation for more long-term clinical studies and demon-strate the potential of using this plasma-treated magnesiumimplant in the treatment of bone loss.

Acknowledgments

This study was jointly financially supported by the AO TraumaResearch Grant 2012, Hong Kong Research Grant Council Compet-itive Earmarked Research Grant (#718913, #718507), HKU Univer-sity Research Council Seeding Fund, City University of Hong KongApplied Research Grant (ARG) No. 9667066, National Natural Sci-ence Foundation of China (NSFC) 2013, National Science Fund forDistinguished Young Scholars (Grant No. 51225101) and ShenzhenKey Laboratory for Innovative Technology in Orthopaedic Trauma,The University of Hong Kong Shenzhen Hospital. The authorswould like to thank Dr. Sun Xuejun of Department of DivingMedicine, Second Military Medical University and Dr. Yun-WahLam of Department of Biology and Chemistry, City University ofHong Kong for their kind support.

References

[1] Barrette-Connor E. The economic and human costs of osteoporotic fracture.Am J Med 1995;98:3Se8S.

[2] Iqbal MM. Osteoporosis: epidemiology, diagnosis, and treatment. SouthamptMed J 2000;93:2e18.

[3] Russell RG, Watts NB, Ebetino FH, Rogers MJ. Mechanisms of action ofbisphosphonates: similarities and differences and their potential influence onclinical efficacy. Osteoporos Int 2008;19(6):733e59.

[4] Banffy MB, Vrahas MS, Ready JE, Abraham JA. Nonoperative versus prophy-lactic treatment of bisphosphonate-associated femoral stress fractures. ClinOrthop Relat Res 2011;469(7):2028e34.

[5] Advisory Task Force on Bisphosphonate-Related Osteonecrosis of the Jaws.American association of oral and maxillofacial surgeons position paper onbisphosphonate-related osteonecrosis of the jaws. J Oral Maxillofac Surg2007;65(3):369e76.

[6] Abrahamsen B. Adverse effects of bisphosphonates. Calcif Tissue Int2010;86(6):421e35.

[7] Koh JS, Goh SK, Png MA, Ng AC, Howe TS. Distribution of atypical fractures andcortical stress lesions in the femur: implications on pathophysiology.Singapore Med J 2011;52(2):77e80.

[8] Bannon M. Bisphosphonates and osteoporosis: which therapeutic agent isbest? Qjm-Int J Med 2011;104(4):279e80.

[9] Giusti A, Hamdy NAT, Dekkers OM, Ramautar SR, Dijkstra S, Papapoulos SE.Atypical fractures and bisphosphonate therapy: a cohort study of patientswith femoral fracture with radiographic adjudication of fracture site andfeatures. Bone 2011;48(5):966e71.

[10] Eastell R. Treatment of postmenopausal osteoporosis. N Engl J Med1998;338(11):736e46.

[11] Baron R, Hesse E. Update on bone anabolics in osteoporosis treatment:rationale, current status, and perspectives. J Clin Endocrinol Metab2012;97(2):311e25.

[12] Kong SW, Kim JS, Park KS, Cha BH, Shim JH, Kim JY, et al. Surface modificationwith fibrin/hyaluronic acid hydrogel on solid-free form-based scaffolds fol-lowed by BMP-2 loading to enhance bone regeneration. Bone 2011;48(2):298e306.

[13] Brown KV, Li B, Guda T, Perrien DS, Guelcher SA, Wenke JC. Improving boneformation in a rat femur segmental defect by controlling bone morphogeneticprotein-2 release. Tissue Eng Part A 2011;17(13e14):1735e46.

[14] Ben-David D, Srouji S, Shapira-Schweitzer K, Kossover O, Ivanir E, Kuhn G,et al. Low dose BMP-2 treatment for bone repair using a PEGylated fibrinogenhydrogel matrix. Biomaterials 2013;34(12):2902e10.

[15] Jun SH, Lee EJ, Jang TS, Kim HE, Jang JH, Koh YH. Bone morphogenic protein-2(BMP-2) loaded hybrid coating on porous hydroxyapatite scaffolds for bonetissue engineering. J Mater Sci Mater Med 2013;24(3):773e82.

[16] Jeon O, Song SJ, Yang HS, Bhang SH, Kang SW, Sung MA, et al. Long-termdelivery enhances in vivo osteogenic efficacy of bone morphogenetic protein-2 compared to short-term delivery. Biochem Biophys Res Commun2008;369(2):774e80.

[17] Vo TN, Kasper FK, Mikos AG. Strategies for controlled delivery ofgrowth factors and cells for bone regeneration. Adv Drug Deliv Rev 2012:1292e309.

[18] Neer RM, Arnaud CD, Zanchetta JR, Prince R, Gaich GA, Reginster JY, et al.Effect of parathyroid hormone (1-34) on fractures and bone mineral density inpostmenopausal women with osteoporosis. N Engl J Med 2001;344(19):1434e41.

[19] Lin W, Lin J. Parathyroid hormone for osteoporosis treatment. Formos JMusculoskelet Disord 2011;2:113e7.

[20] McGreevy C, Williams D. Safety of drugs used in the treatment of osteopo-rosis. Ther Adv Drug Saf 2011;2(4):159e72.

[21] Song G. Control of biodegradation of biocompatable magnesium alloys. CorrosSci 2007;49(4):1696e701.

[22] Boccaccini AR, Hoppe A, Guldal NS. A review of the biological response to ionicdissolution products from bioactive glasses and glass-ceramics. Biomaterials2011;32(11):2757e74.

[23] Precival M. Bone health and osteoporosis. Appl Sci Nutr 1999;5:1e5.[24] Bigi A, Boanini E, Gazzano M. Ionic substitutions in calcium phosphates syn-

thesized at low temperature. Acta Biomater 2010;6(6):1882e94.[25] Song G, Song SA. Possible biodegradable magnesium implant material. Adv

Eng Mater 2007:298e302.[26] Zberg B, Uggowitzer PJ, Löffler JF. MgZnCa glasses without clinically observable

hydrogen evolution for biodegradable implants. NatureMater 2009;8:887e91.[27] Xue D, Yun Y, Tan Z, Dong Z, Schulz M. In vivo and in vitro degradation

behavior of magnesium alloys as biomaterials. J Mater Sci Technol 2012;28(3):261e7.

[28] Witte F, Kaese V, Haferkamp H, Switzer E, Meyer-Lindenberg A, Wirth CJ, et al.In vivo corrosion of four magnesium alloys and the associated bone response.Biomaterials 2005;26(17):3557e63.

[29] Walker J, Shadanbaz S, Kirkland NT, Stace E, Woodfield T, Staiger MP, et al.Magnesium alloys: predicting in vivo corrosion with in vitro immersiontesting. J Biomed Mater Res B Appl Biomater 2012;100(4):1134e41.

[30] Xin SG, Song LX, Zhao RG, Hu XF. Properties of aluminium oxide coating onaluminium alloy produced by micro-arc oxidation. Surf Coat Technol2005;199(2e3):184e8.

[31] Fischer H, Niedhart C, Kaltenborn N, Prange A, Marx R, Niethard FU, et al.Bioactivation of inert alumina ceramics by hydroxylation. Biomaterials2005;26(31):6151e7.

[32] Sharma V, Ranawat AS, Rasquinha VJ, Weiskopf J, Howard H, Ranawat CS.Revision total hip arthroplasty for ceramic head fracture a long-term follow-up. J Arthroplasty 2010;25(3):342e7.

[33] Stappert CF, Baldassarri M, Zhang Y, Hanssler F, Rekow ED, Van PT. Reliabilityand fatigue failure modes of implant-supported aluminum-oxide fixed dentalprostheses. Clin Oral Implants Res 2011:1173e80.

[34] Kim JM, Mudgal CS, Konopka JF, Jupiter JB. Complications of total elbowarthroplasty. J Am Acad Orthop Surg 2011;19(6):328e39.

[35] Anders A. Handbook of plasma immersion ion implantation and deposition.New York: Wiley; 2000.

[36] Cheung KMC, Wong HM, Yeung KWK, Lam KO, Tam V, Chu PK, et al.A biodegradable polymer-based coating to control the performance of mag-nesium alloy orthopaedic implants. Biomaterials 2010;31(8):2084e96.

[37] Brar HS, Platt MO, Sarntinoranont M, Martin PI, Manuel MV. Magnesium as abiodegradable and bioabsorbable material for medical implants. Jom2009;61(9):31e4.

[38] BarfieldW, ColbathG, DesJardins J, An Y, Hartsock L. The potential ofmagnesiumalloy use in orthopaedic surgery. Curr Orthop Pract 2012;23(2):146e50.

[39] Chu PK, Tang BY, Cheng YC, Ko PK. Principles and characteristics of a newgeneration plasma immersion ion implanter. Rev Sci Instrum 1997;68(4):1866e74.

[40] Gao Y-l, Wang C-s, Yao M, Liu H-b. The resistance to wear and corrosion oflaser-cladding Al2O3 ceramic coating on Mg alloy. Appl Surf Sci 2007;253(12):5306e11.

[41] Shan DY, Zhang RF, Han EH. Effect of Al2O3 thin film on corrosion resistance ofpure Mg and its anodic coating. Mater Sci Forum 2005;488-489:865e8.

[42] Nie X, Meletis EI, Jiang JC, Leyland A, Yerokhin AL, Matthews A. Abrasive wear/corrosion properties and TEM analysis of Al2O3 coatings fabricated usingplasma electrolysis. Surf Coat Technol 2002;149(2e3):245e51.

[43] Song G, Atrens A, Wu X, Zhang B. Corrosion behaviour of AZ21, AZ501 andAZ91 in sodium chloride. Corros Sci 1998;40(10):1769e91.

[44] Wong HM, Wu S, Chu PK, Cheng SH, Luk KD, Cheung KMC, et al. Low-modulusMg/PCL hybrid bone substitute for osteoporotic fracture fixation. Biomaterials2013;34(29):7016e32.

[45] Serre CM, Papillard M, Chavassieux P, Voegel JC, Boivin G. Influence of mag-nesium substitution on a collagen-apatite biomaterial on the production of acalcifying matrix by human osteoblasts. J Biomed Mater Res 1998:626e33.

[46] Zreiqat H, Howlett CR, Zannettino A, Evans P, Tanzil GS, Knabe C, et al. Mech-anisms of magnesium-stimulated adhesion of osteoblastic cells to commonlyused orthopaedic implants. J Biomed Mater Res 2002;62(2):175e84.

H.M. Wong et al. / Biomaterials 34 (2013) 9863e98769876

[47] Staiger MP, Pietak AM, Huadmai J, Dias G. Magnesium and its alloys as or-thopedic biomaterials: a review. Biomaterials 2006;27(9):1728e34.

[48] Li Z, Gu X, Lou S, Zheng Y. The development of binary Mg-Ca alloys for use asbiodegradable materials within bone. Biomaterials 2008;29(10):1329e44.

[49] Perl DP, Moalem S. Aluminum and Alzheimer’s disease, a personal perspectiveafter 25 years. J Alzheimers Dis 2006;9(3 Suppl.):291e300.

[50] Percy ME, Kruck TPA, Pogue AI, LukiwWJ. Towards the prevention of potentialaluminum toxic effects and an effective treatment for Alzheimer’s disease.J Inorg Biochem 2011 Nov;105(11):1505e12.

[51] Vormann J. Magnesium: nutrition and metabolism. Mol Aspects Med 2003;24:27e37.

[52] Saris N-EL, Mervaala E, Karppanen H, Khawaja JA, Lewenstam A. Magnesium:an update on physiological, clinical and analytical aspects. Clin Chim Acta2000;294(1e2):1e26.

[53] Pybus J. Determination of calcium and magnesium in serum and urine byatomic absorption spectrophotometry. Clin Chim Acta 1968;23(2):309e17.

[54] Rettig R, Virtanen S. Composition of corrosion layers on a magnesium rare-earth alloy in simulated body fluids. J Biomed Mater Res 2009;88A(2):359e69.