in vitro evaluation of the surface effects on magnesium-yttrium alloy degradation and mesenchymal...

In vitro evaluation of the surface effects on magnesium-yttrium alloydegradation and mesenchymal stem cell adhesion

Ian Johnson,1 Daniel Perchy,2 Huinan Liu1

1Department of Bioengineering, University of California, Riverside, California 925212Department of Biology, University of Pittsburgh, Pennsylvania 15219

Received 23 May 2011; revised 23 September 2011; accepted 4 October 2011

Published online 29 November 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/jbm.a.33290

Abstract: Magnesium (Mg) alloys present many advantages

over current materials used in medical implants and devices.

However, the rapid degradation of Mg alloys can raise the

local pH and create gas cavities. Fundamental understanding

of their biodegradation processes is necessary for their

success in clinical applications. This study investigated how

the oxidized and polished surfaces of a Mg-yttrium (Y) alloy

affected the degradation mode and rate in cell culture media

versus deionized water. The interactions of the alloy surfa-

ces with cells were examined in vitro using bone marrow

derived mesenchymal stem cells, since they are critical cells

for bone tissue regeneration. The polished surface was

more stable than the oxidized surface in cell culture media,

but less stable in water. When comparing polished and

oxidized surfaces, their degradation modes were similar in

water, but different in cell culture media. The microstruc-

ture, roughness, and oxygen content of the alloy surface

contributed to these differences. The presence or absence of

a stable degradation layer determined the rate of Y loss and

the inhibiting or promoting behavior of Y on degradation.

The initial alloy surfaces not only influenced the degrada-

tion, but also determined cell attachment, which is critical

for tissue integration. The polished surface showed more

cell adhesion than the oxidized surface, mainly because of

its slower degradation rate and lesser effect on the local

pH. In conclusion, this study demonstrated that both the

Mg alloy surfaces and the immersion fluids played impor-

tant roles in controlling the degradation and cellular interac-

tions. VC 2011 Wiley Periodicals, Inc. J Biomed Mater Res Part A:

100A: 477–485, 2012.

Key Words: magnesium-yttrium alloy, degradation, mesen-

chymal stem cells, surface characterization, microstructure,

cytocompatibility, medical devices, orthopedic implants

How to cite this article: Johnson I, Perchy D, Liu H. 2012. In vitro evaluation of the surface effects on magnesium-yttrium alloydegradation and mesenchymal stem cell adhesion. J Biomed Mater Res Part A 2012:100A:477–485.

INTRODUCTION

The promise of magnesium (Mg) forbiodegradable implantsMagnesium (Mg) has attracted great interest for use in bio-medical implants and devices because it can potentiallyaddress many of the problems associated with currentimplant materials.1 Ideal properties for a medical implantinclude satisfactory biodegradability, bioactivity, and bio-compatibility. Further, the implant should provide an appro-priate surface for cell adhesion and long-term functions, andhave mechanical properties similar to natural tissue. Mgalloys are biodegradable, and their degradation productscan be excreted or used in metabolic processes.2 Mg isosteoconductive1 and osteoinductive,3 and it promotes boneformation.4,5 Mg alloys have elastic moduli similar to bone,which addresses the stress-shielding problems associatedwith current metallic implants.6 The mechanical strength ofMg is similar to cortical bone, reducing the likelihood ofmechanical failure.6 For these reasons, Mg alloys have manyadvantages over current materials used for orthopedic andcraniofacial implants.

Titanium (Ti) alloys and polymers are still the mostwidely used orthopedic biomaterials, although they havetheir own strengths and weaknesses. Ti alloys have accepta-ble biocompatibility, high corrosion resistance, and highmechanical strength for load bearing applications.7 However,Ti alloys are nondegradable and cause stress shielding tobone due to their high elastic modulus, and their corrosionmay release harmful wear particles that often lead to revi-sion surgeries.7 Bioabsorbable polymers such as polylactide-co-glycolide (PLGA) have been developed in recent decadesto replace permanent metals. However, their low mechanicalstrength often leads to device/implant breakage and insome cases catastrophic failure.8,9 Table I summarizes thekey physical properties, mechanical properties, and biologi-cal interactions of these materials.10–16 It shows that themechanical properties of Mg alloys closely match corticalbone, and that Mg alloys have many desirable biologicalproperties.

Despite all the advantages of Mg alloys, they degrademuch too rapidly in physiological conditions to be medicallyeffective. The rapid degradation of Mg can raise the local

Correspondence to: H. Liu; e-mail: [email protected]

Contract grant sponsors: NSF ERC; NSF BRIGE; University of California

VC 2011 WILEY PERIODICALS, INC. 477

pH, create gas cavities, and cause premature mechanicalfailure or implant detachment.17 Therefore, fundamentalunderstanding of their degradation processes in physiologi-cal environment is necessary for their success in clinicalapplications.

Mg degradation in physiological fluidsDegradation of Mg alters its surface chemistry and micro-structure, mechanical properties, and interactions with sur-rounding cells, all of which are critical factors in determin-ing the success of implants. Many factors cause Mgdegradation, including chemical reactions, galvanic reactions,and mechanical stress.18 Reactions 1 and 2 below describesome common Mg degradation reactions.18–20 Mg reactswith water to produce Mg(OH)2 on the metal’s surface. Thisforms a protective but porous degradation layer that slowsfurther degradation, although not very effectively. Reaction1 shows how this protective degradation layer is formed.Reaction 2 shows how chloride ions (Cl�) convert this deg-radation layer to a more soluble form (i.e., MgCl2) that ismore easily dissolved.21

Mgþ 2H2O ! MgðOHÞ2 þ H2 (1)

MgðOHÞ2 þ 2Cl� ! MgCl2 þ 2ðOH�Þ (2)

The dissolution of the degradation layer exposes moreMg surface to the environment. The exposed Mg is morevulnerable to chemical and galvanic degradation reactions.The degradation reactions with Cl� also increase the localpH. Other constituents of the immersion media can havetremendous impact on the degradation rate and mode. Theincorporation of carbonates can produce a more stabledegradation layer and inhibit pitting degradation pro-cesses.22 Phosphate incorporation within the degradationlayer makes it denser and more resistant to Cl� mediateddegradation reactions. Phosphates also allow the accumula-tion of calcium within the degradation layer.23

Yttrium (Y) in Mg alloysY is often included in Mg alloys to increase strength,24,25

ductility,26 and degradation resistance.27 Y shows both

degradation inhibiting and promoting behaviors in Mg,depending on local environment, alloy composition, andmicrostructure. Y decreases Mg degradation by forming aY2O3 passivation layer,28–30 which inhibits cathodic reac-tions.30 Y also reduces Mg grain size. When the grain size issmall, the b phase can form a continuous network around aphase to protect it from degradation.25,31 Y promotes Mgdegradation through microgalvanic coupling. Although Y hasthe same standard electrochemical potential as Mg,32 thisdoes not eliminate microgalvanic degradation because the Y-rich intermetallic phase acts as a cathode to the a phase.28

The objective of this study was to investigate the effectsof the alloy surface and the environment upon the degrada-tion of a new Mg-Y alloy and cell adhesion on the surface ofthis alloy. Bone marrow derived mesenchymal stem cells(BMSCs) were used as the model cells in this study becausethey play important roles in bone tissue regeneration. Threemajor questions were addressed in this study:

1. How does the surface of this Mg-Y alloy affect its degra-dation mode and rate?

2. How does the surrounding fluid affect the alloy’s degra-dation mode and rate?

3. How do the alloy’s surface and degradation affect mesen-chymal stem cell adhesion?

MATERIALS AND METHODS

Mg-Y alloyMg-4 wt % Y (MgY) alloy was prepared by melting Mg with4 wt % Y in a protected environment and casting as aningot. The as-cast MgY alloy ingot was cut into 250 lm thickdiscs with a wire electric discharge machine (AgieChar-milles, Agiecut 200 VHP). The as-produced MgY alloy discshad thermal oxide layers on their surfaces and were calledMgY_O in this study. Some of the MgY_O samples werepolished using 600, 800, and 1200 grit silicon carbide abra-sive papers (PACE Technologies) to remove their oxidizedsurfaces, and were referred to as MgY_P in this study. BothMgY_O and MgY_P samples were cut to dimensions of 10 �10 mm2, cleaned in isopropanol (Sigma-Aldrich, CAS num-ber 67-63-0), weighed, and sterilized on all sides under

TABLE I. Summary of the Physical, Mechanical, and Biological Properties of Selected Implant Materials10–16

Cortical Bone Mg

Mg Alloys

PLGA Ti6Al4VWE43 AZ91E

Yield Strength (MPa) 104.9–114.3 126 250 260 41–55 760–880Elastic Modulus (GPa) 3–20 40 44 44 1.4–2.8 114Density (g/cm3) 1.8–2 1.74 1.84 1.81 1.3 4.43Biodegradable? – yes yes yes yes noDegradation/Corrosion Alters Local pH? – increases increases increases decreases noCan Degradation/Corrosion Products

Be Metabolized?– yes some some yes no

Osteoconductive – yes yes yes no noOsteoinductive – yes yes yes no no

The properties of these materials may vary depending upon the methods of processing and testing. Mg has mechanical properties similar to

cortical bone, and desirable interactions with surrounding tissue. An entry of ‘‘some’’ in the metabolism of degradation products row means

that the magnesium can be metabolized, but some constituents of the alloy may not, may have harmful effects, or their effects on tissue may

be still unclear.

478 JOHNSON, PERCHY, AND LIU Mg-Y ALLOY DEGRADATION AND STEM CELL ADHESION

ultraviolet (UV) radiation before degradation and cell cul-ture experiments.

In vitro cytocompatibility of Mg-Y alloyBMSCs were chosen as the model cells because they areimportant for osteointegration of orthopedic implants. Thesecells were harvested from a female goat with a matureskeleton using an established protocol.33,34 This protocolcomplied with the National Institute of Health guidelines forthe care and use of laboratory animals, and was approvedby the Institutional Animal Care and Use Committee. Thegoat was examined prior to the harvesting of cells to ensurethat it had no physiological abnormalities. The goat waseuthanized, and then bone marrow was harvested from theiliac crest under sterile conditions. Harvested tissue wasrinsed with phosphate buffered solution (PBS, Sigma) andthen cultured in Dulbecco’s Modified Eagle Medium (DMEM;Hyclone) supplemented with 10% fetal bovine serum (FBS;Hyclone) and 1% penicillin/streptomycin (P/S; Invitrogen)under standard cell culture conditions (37�C, 5% CO2/95%air, humidified sterile environment). The cell culturemedium was changed every other day. Once the cellsreached 80–90% confluence, they were detached usingtrypsin (Invitrogen) and passed to subculture. Cells fromthe second passage were used in this experiment.

BMSCs were seeded directly onto MgY_O and MgY_Psamples at a density of 40,000 cells/cm2. Bioactive glass(ThermanoxV

R

coverslips; Fisher Scientific) was used as apositive control since its surface was treated for optimal cellattachment. Plasma treated tissue culture polystyrene(TCPS; BD Biosciences) served as a reference. The cell cul-ture experiments were performed in triplicate. All the sam-ples were incubated in DMEM supplemented with 10% FBSand 1% P/S under standard cell culture conditions for 24 h.After that, nonadherent cells were removed by washingwith PBS. Adherent cells were fixed with 4% formaldehyde(10% neutral buffered formalin; VWR) and stained with40,6-diamidino-2-phenylindole dilactate (DAPI; Invitrogen).The adherent cells were visualized and counted using a flu-orescence microscope (Nikon Eclipse Ti). Cells in four fieldson each MgY sample were counted, and cell density wascalculated as cells per area. The average cell density wasdivided by the seeding density, and then multiplied by100 to calculate the percentage of cell adhesion. After 24 hof culture, the pH meter (Accumet AB15) was used to mea-sure the pH change of the media. The pH meter was cali-brated with known standards before use.

Surface characterization before and after cell cultureThe surface of the MgY alloy was characterized before andafter cell culture (as described above) using a field emissionscanning electron microscope (FESEM; Philips XL-30). Anaccelerating voltage of 20 kV was used to characterize theMgY surfaces before cell culture. The accelerating voltagewas reduced to 5 kV to image MgY after cell culture becausethe surfaces became less conductive due to the presence ofthe cells and proteins. Energy dispersive X-ray spectroscopy(EDS) was performed at 5000� magnification to quantify

the elemental composition of the MgY surfaces before andafter cell culture. Magnification of 5000� was chosen sothat a substantial portion of the sample’s surface would beanalyzed to show the average elemental distribution of thesamples.

The surface roughness of MgY_O and MgY_P was quanti-fied before the degradation and cell studies by using thetapping mode of an atomic force microscope (AFM; DigitalInstruments Dimension 5000). A probe with an 8 nm radiustip (Veeco, RTESP) was used. For each sample, two differentscan sizes (500 � 500 nm2 and 5 � 5 lm2) were used todetermine surface roughness at the nanoscale and themicronscale. All scans were repeated three times in threerandom sections on the surface. The average roughnessvalues were calculated on flattened AFM images using theNanoscope IV image analysis software (Digital Instruments).

Mg degradation in immersion mediaThe immersion method was used to investigate MgY degra-dation. All degradation experiments were performed in trip-licate. MgY samples were immersed in DMEM supplementedwith 10% FBS and 1% P/S under standard cell cultureconditions. Deionized (DI) water produced by a MilliporeMilli-QVR Biocel System was used as a control. All MgY sam-ples were incubated in DMEM and DI water according toprescribed sequential intervals. The incubation intervalswere set for an hour long at the beginning of degradation,but were increased to 48 h long after 72 h of degradation.A higher time resolution was necessary in the beginning ofdegradation to track the initial rapid changes of MgYsamples and their immersion media. The initial degradationperiod plays critical roles in cell attachment, the first stepof tissue-implant integration. When an incubation intervalended, the samples were removed from their media anddried in a 37�C isotemp oven for 12 h or until the samplereached a constant mass. Degradation products on thesurface of the MgY samples were left intact, but solubledegradation products remained in the media. The pH meter(Accumet AB15) was first calibrated with known standardsand then used to measure the pH of the immersion mediaat the end of every incubation interval. The samples weredried, weighed, photographed, sterilized under UV radiation,and then placed in fresh immersion media for the nextinterval. This procedure was repeated for each incubationinterval.

RESULTS

Mesenchymal stem cell adhesion on MgY surfaceFigure 1 shows the results of BMSC adhesion on the surfa-ces of MgY_O, MgY_P, and bioactive glass (control) after24 h of incubation in DMEM supplemented with 10% FBSand 1% P/S under standard cell culture conditions. Thecells cultured on MgY_O had a round morphology, implyingthat they were unhealthy and detached from the surface[Fig. 1(A)]. The cells cultured on MgY_P were mostly round,but some cells had a more elongated morphology [Fig.1(B)]. The BMSCs cultured on bioactive glass had an elon-gated, spindle like morphology [Fig. 1(C)]. The elongated,

ORIGINAL ARTICLE

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | FEB 2012 VOL 100A, ISSUE 2 479

spindle like morphology implied that the cells were health-ier and attached to the surface. The percentage of BMSC ad-hesion to the samples was quantified in Figure 1(D). Themost cell adhesion was observed on the surface of bioactiveglass and the least cell adhesion on MgY_O. The percent celladhesion on MgY_P was greater than on MgY_O, but lessthan on bioactive glass. The pH of the culture media wasmeasured after 24 h of incubation and reported in Figure1(E). The media for bioactive glass had the pH closest toneutral, while the media for MgY_O had the most alkalinepH. The pH of the media for MgY_P was in between the pHof the media for bioactive glass and MgY_O.

Surface characterization of Mg-Y alloyFigure 2 shows FESEM images of the surface of the MgYsamples before and after 24 h of BMSC culture in DMEM.MgY_O samples had rough, cracked surfaces before cell cul-ture [Fig. 2(A)]. These cracks were greatly enlarged after24 h of cell culture [Fig. 2(C)]. The surface of MgY_P seemedrather smooth before cell culture [Fig. 2(B)], and cracksappeared after cell culture [Fig. 2(D)]. The cracks on MgY_Pafter cell culture were smaller than those on MgY_O aftercell culture. The surface roughness of MgY_O and MgY_Pwas quantified before cell culture using AFM and summar-ized in Table II. The AFM data showed that the surface ofMgY_O was significantly rougher than the surface of MgY_P.

The elemental composition of the MgY surface wasmeasured by EDS before and after 24 h of BMSC culture, asshown in Figure 3. MgY_O began with a much larger amountof oxygen compared with MgY_P, and acquired larger

amounts of carbon, phosphorous, and calcium on its surfaceafter cell culture compared with MgY_P. MgY_P initially hadno detectable oxygen on its surface, and it gained smallamounts of oxygen, carbon, phosphorous, and calcium aftercell culture, which were less than that on MgY_O after cellculture. The percentage of surface Y in MgY_O decreasedafter cell culture, while the percentage of surface Y in MgY_Pincreased after cell culture. The standard deviation of the MgY_O surface elemental composition (wt %) before cell cultureranged from 0.08 to 0.28% (i.e., C 0.1%, O, 0.1%, Mg 0.08%,and Y 0.28%); and after cell culture ranged from 0.04 to3.26% (i.e., C 2.66%, O 2.69%, Na 2.29%, Mg 1.56%, Y0.04%, P 3.26%, and Ca 0.17%). EDS analysis of randomareas of the MgY_P sample was consistent up to the detectionlimit of EDS and the deviation was <0.01%.

Effects of surface and media on degradationmode and rate of Mg alloySequential snapshots of Mg-Y alloy degradation. Figure 4shows photographs of MgY degradation over time in DIwater and DMEM supplemented with 10% FBS and 1% P/S.The degradation layer on MgY_O in DI water appeared rela-tively uniform, with some pitting [Fig. 4(A)]. MgY_Odegraded slightly more rapidly in DI water than DMEM. Thedegradation layer on MgY_O in DMEM initiated fromthe edges of the sample and progressed inwards [Fig. 4(B)].The photographs of degradation over time in DMEM showedthat MgY_O shed many small fragments from its surface.The degradation layer on MgY_P in DI water appeared simi-lar to that of MgY_O, but in a shorter period due to more

FIGURE 1. The results of the BMSC adhesion on the surfaces of bioactive glass, MgY_O, and MgY_P samples after 24 h of culture. (A) Fluores-

cence image of BMSCs adhered on MgY_O surface. (B) Fluorescence image of BMSCs adhered on MgY_P surface. (C) Fluorescence image of

BMSCs adhered on a bioactive glass surface. Scale bars are 100 lm. BMSCs on MgY_O had an unhealthy round morphology and were fewest

in number. BMSCs adhered on MgY_P had a mixed morphology and a greater number of cells compared to MgY_O. BMSCs on bioactive glass

had a healthy spindle shape. (D) The percentage of BMSCs that adhered to the MgY and control surfaces. (E) The pH of the cell culture media af-

ter 24 h of incubation. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]


rapid degradation [Fig. 4(C)]. The degradation layer onMgY_P in DMEM initiated near the center of the sample andprogressed toward the edges [Fig. 4(D)]. The MgY_P samplebroke apart near its center into several large pieces duringthe final stages of degradation in DMEM.

In summary, the photographs demonstrated that thedegradation of MgY_O and MgY_P in DMEM initiated andprogressed with a different mode and rate. More pitting andlocalized degradation were observed on both MgY samplesin DMEM, while more gradual and uniform degradation wasobserved on both MgY samples in DI water.

Mass loss of Mg-Y alloy and pH change of immersionmedia. Figure 5 shows the mass loss of the MgY_O andMgY_P samples in DI water and DMEM supplemented with10% FBS and 1% P/S. Percent mass (%M) was equal to thesample mass at a given incubation interval (Mi) divided bythat sample’s initial mass (M0), and then multiplied by 100.The %M described the mass of the MgY samples at a giventime of degradation relative to their initial mass. At anygiven incubation interval, in the order of decreasing %Mvalue, the combination of surface and immersion mediafollowed this order: MgY_P in DMEM, MgY_O in DMEM,MgY_O in DI water, and finally MgY_P in DI water. The %Mfor MgY_O declined slightly more rapidly in DI water thanin DMEM. The %M for MgY_P declined much more rapidlyin DI water than in DMEM. Overall, MgY_P in DMEM showed

FIGURE 2. Scanning electron micrographs of MgY sample surfaces before and after 24 h of BMSC culture. (A) MgY_O before cell culture.

(B) MgY_P before cell culture. (C) MgY_O after cell culture. (D) MgY_P after cell culture. MgY_O had a rough surface while MgY_P had a smooth

surface before cell culture. The surface cracks on MgY_O were greatly enlarged after 24 h of cell culture. The surface of MgY_P became rough

after cell culture, although it still appeared more smooth than MgY_O surface after cell culture.

TABLE II. Surface Roughness of MgY_O and MgY_P was

Calculated Based on AFM Scans

RoughnessMeasurements

Ra (nm)Z Range(nm)

RMS(nm)

ScanningArea Sample

500 � 500 nm2 MgY_O 14 6 2 131 6 41 19 6 3MgY_P 8 6 3 80 6 65 10 6 4

5 � 5 lm2 MgY_O 196 6 47 2104 6 262 274 6 53MgY_P 65 6 31 618 6 254 81 6 36

Values are average 6 standard deviation (n ¼ 3). Ra is the mean

roughness, Z range represents the maximum height range of the sur-

face, and RMS is the root-mean-square value. For each sample, two

different scan sizes (500 � 500 nm2 and 5 � 5 lm2) were used to

determine surface roughness at the nanoscale and the micronscale.

ORIGINAL ARTICLE


the slowest mass loss among any combination of surfaceand immersion media (that is, slower than MgY_O in DMEM,MgY_O in DI water, and MgY_P in DI water). MgY_P in DIwater had the most rapid mass loss among any combinationof surface and immersion media.

Figure 6 shows the effect of MgY degradation on the pHof the immersion media: DI water and DMEM supplementedwith 10% FBS and 1% P/S. The pH of the immersion mediawas measured at the end of each incubation interval. Allsamples showed a significant pH increase within 24 h ofimmersion, and all pH values became relatively stable afterthe initial period. The pH change of the immersion mediaover time followed a similar trend for both MgY_O andMgY_P in the same media, except at the beginning of degra-dation. The pH of the DI water containing MgY_P and the

DMEM containing MgY_O had an acute alkaline spike withinthe initial 24 h of degradation.

DISCUSSION

Cell adhesion on Mg-Y alloy surfacesThe in vitro BMSC culture results showed that MgY_P wassuperior to MgY_O as a substrate for cell adhesion. Not onlydid more cells adhere to MgY_P, but they had a healthiermorphology. This occurred for several reasons. For the first24 h of degradation, MgY_P had a less pronounced effect onthe acute pH increase of DMEM than MgY_O did. MgY_Pdegraded more slowly in DMEM and provided a more stablesurface for cells. The shedding of the surface layer fromMgY_O may have made it more difficult for cells to adhere.In short, more cells adhered to MgY_P because it induced

FIGURE 3. The surface elemental composition of the MgY samples before and after 24 h of BMSC culture. EDS analysis was conducted on

images with �5000 magnification. (A) MgY_O surface compostion before cell culture. (B) MgY_P surface composition before cell culture. (C)

MgY_O surface composition after cell culture. (D) MgY_P surface compostion after cell culture. (E) The quantitative surface composition of

MgY_O before and after cell culture. (F) The quantitative surface composition of MgY_P before and after cell culture. [Color figure can be viewed

in the online issue, which is available at wileyonlinelibrary.com.]


less pH increase of the media, had a more stable degrada-tion layer formed on its surface, and showed a slower rateof mass loss, which indicated slower degradation.

Degradation mode and rate of Mg-Y alloyMgY_O degraded more slowly than MgY_P in DI waterbecause MgY_O began with a protective thermal oxide layer.This thermal oxide layer was formed during alloy hot proc-essing. MgY_O degraded much more rapidly in DMEM thanMgY_P. This is similar to the findings that WE43 with anoxidized surface degraded more rapidly in simulated bodyfluid than WE43 with a polished surface.28 The photographsof degradation over time showed that degradation of MgY_Osamples in DMEM began around the edges and progressedinwards. MgY_O samples continually shed fragments fromtheir outer layer until completely degraded. FESEM imagesshowed that MgY_O initially had a rough surface with manycracks, and these cracks were greatly enlarged after 24 h ofcell culture in DMEM. It has been reported that AZ91 alloy

with rough surfaces had less effective passivation than thesame alloy with smooth surfaces.35 A similar phenomenonmight have occurred on MgY surfaces since MgY_O had amore rough surface than MgY_P. Deep cracks in MgY_Oallowed deep and rapid penetration by attacking ions inDMEM, especially Cl�. MgY_O was also more vulnerable toCl� mediated degradation due to its higher initial oxide con-tent. Cl� mediated degradation led to undermining of theoxidized surface on MgY_O, and as a result, portions of thesurface layer that were surrounded by cracks eventuallybroke off.29,31 This can be especially damaging to the oxi-dized surface layer if the Cl� mediated degradation propa-gates laterally from local sites.22 These phenomenadestroyed the thermal oxide layer and prevented the forma-tion of a stable degradation layer on MgY_O. In the absenceof a stable and protective degradation layer, Y had a netdegradation promoting effect in MgY_O in DMEM. RapidMgY_O degradation released large amounts of hydroxideions into the local environment. These mechanismsexplained why the initial alkaline pH peak was observed inDMEM containing MgY_O.

MgY_P degraded rapidly in DI water because it initiallyhad no protective oxide surface layer. A degradation layerformed on MgY_P’s surface during immersion in DI wateraccording to reaction 1 in the introduction. In DI water, thisporous degradation layer provided inferior degradation pro-tection when compared to the thermal passivated surface ofMgY_O. More rapid initial degradation of MgY_P releasedmore hydroxide ions and thus produced a higher initialalkaline peak in DI water than MgY_O.

MgY_P degraded much more slowly in DMEM because ofits surface structure and chemistry. In DMEM, the degrada-tion layer initially formed near the center of the MgY_P sam-ple and spread outwards. MgY_P samples tended to eventu-ally break apart in the center into several large fragments.MgY_P initially had little roughness on its surface layer andno cracks. After cell culture, the surface still had less rough-ness and fewer cracks than MgY_O. A smoother surface

FIGURE 4. Photographs of MgY degradation over time in DI water and

DMEM supplemented with 10% FBS and 1% P/S under standard cell

culture conditions. (A) MgY_O in DI Water. (B) MgY_O in DMEM. (C)

MgY_P in DI water. (D) MgY_P in DMEM. Both MgY samples showed a

similar degradation mode in DI water, although MgY_P degraded more

rapidly. MgY_P lasted much longer in DMEM than MgY_O. MgY_O deg-

radation in DMEM began around the edges of the sample and pro-

gressed inwards. MgY_P degradation in DMEM initiated near the center

and progressed outwards. [Color figure can be viewed in the online

issue, which is available at wileyonlinelibrary.com.]

FIGURE 5. The mass loss of MgY samples versus the incubation time

in DI water and DMEM. Percent Mass (%M) is the mass at a given

time interval (Mi) divided by the initial mass (M0) of that sample, mul-

tiplied by 100. MgY_O degradation was slightly slower in DMEM than

in DI water. MgY_P degraded much more slowly in DMEM than in DI

water. [Color figure can be viewed in the online issue, which is avail-

able at wileyonlinelibrary.com.]

ORIGINAL ARTICLE


could increase the effectiveness of passivation and decreasepitting formation.35 The fewer and smaller cracks on MgY_Pprovided fewer opportunities for invading ions. The surfacelayer of MgY_P had less oxygen content than that of MgY_Oduring the initial degradation, making MgY_P less suscepti-ble to Cl� attack. MgY_P was able to incorporate carbon andphosphorous from DMEM into its degradation layer sincethe very beginning of degradation. These protective ele-ments significantly improved the stability of the degradationlayer and its resistance to Cl� attack. Degradation productsthat accumulated on the surface blocked the degradation ofthe Y-rich intermetallic phase. The degradation mode ofMgY_P in DMEM allowed the formation of a stable protec-tive degradation layer, in contrast to the unstable surfacelayer formed on MgY_O. Y provided a net degradation inhibi-ting effect on MgY_P in DMEM due to the stable protectivelayer.

The results showed that the surface of MgY had tremen-dous impact on the mode and rate of degradation inresponse to their environment. Both MgY samples encoun-tered a similar mode of degradation in DI water, althoughthe MgY_O samples degraded more slowly. MgY samplesdemonstrated strikingly different modes and rates of degra-dation in DMEM. MgY_P formed a stable degradation layerin DMEM that greatly decreased degradation. MgY_O wasunable to form a stable protective degradation layer inDMEM due to three possible reasons. (1) The high oxygencontent in the thermal oxide layer made it particularly vul-nerable to Cl� attack. (2) The surface shedding behavioralso contributed to the loss of the protective layer. (3) Therough surface reduced the effectiveness of passivation andincreased pitting formation.

Furthermore, the degradation results confirmed that thepresence of ions and proteins in the physiological fluids(e.g., DMEM) versus the absence of ions and proteins (e.g.,DI water) played an important role in the degradation ofMg alloys. Specifically, both MgY_O and MgY_P showed aslower percent mass change in DMEM compared to that inDI water, indicating slower degradation in DMEM than in DIwater. This may partially explain why different degradation

rates of Mg alloys in vivo (i.e., abundant ions and proteins)versus in vitro (i.e., limited ions and proteins) werereported in literature.36

Y in the degradation layer before and after cell cultureThe combination of immersion media and sample surfacedetermined whether Y had a net degradation inhibiting orpromoting effect on the alloy samples. An intriguing discov-ery is that after cell culture MgY_O experienced a reductionin Y surface percentage while MgY_P experienced anincrease in Y surface percentage. Y has relatively high mobil-ity within Mg28 and migrated to the sample surface to beconsumed in oxidation reactions. As the degradation layerbecame thicker, it became more difficult for oxygen to reachY to react with.28 As discussed in the previous sections,MgY_O did not have a stable protective layer in DMEM. Thelack of a stable barrier on the MgY_O surface allowed forcontinued migration and loss of Y from the surface. Further-more, Y might have a net degradation promoting effect onthe MgY_O samples due to the lack of a stable degradationlayer. Another possibility is that the increased deposition ofother elements from DMEM on MgY_O may have diluted Yon the surface. The opposite trend was observed for MgY_Pbecause it had a much more stable degradation layer, whichallowed Y to express a net degradation inhibiting behavior.Y migrated to the surface of MgY_P and accumulated in thedegradation layer until that layer became thick enough toprevent further reactions. A higher concentration of Y in thedegradation layer would lead to better protection for thebulk Mg because the passivation layer with Y2O3 could in-hibit cathodic reactions.

CONCLUSION

This study revealed that the surfaces of Mg alloys must betaken into account when controlling their mode and rate ofdegradation in physiological environments. The degradationrate is obviously important, but the degradation mode alsohas vital implications. An implant that degrades graduallyfrom the surface without releasing large fragments is gener-ally more desirable than an implant that constantly sheds

FIGURE 6. The pH change of immersion media over time (hours) after MgY samples were incubated in that media under standard cell culture

conditions. (A) DI water was used as the immersion media. The pH of the DI water containing MgY_P had an acute alkaline spike within the ini-

tial 24 h of degradation. (B) DMEM with 10% FBS and 1% P/S was used as the immersion media. The pH of the DMEM containing MgY_O had

an acute alkaline spike within the initial 24 h of degradation. [Color figure can be viewed in the online issue, which is available at

wileyonlinelibrary.com.]


fragments from its outer layer. The initial alloy surface alsodetermined cell attachment, which is critical for tissue inte-gration. Specifically, more cells with a healthier morphologyadhered to the polished MgY surface. The initial alloysurface and immersion media together determined the netdegradation inhibiting or promoting behavior of Y and theelemental composition of the degradation layer. MgY_Pdegraded the most slowly in DMEM, followed by MgY_O inDMEM, MgY_O in DI water, and MgY_P in DI water. Thepresence or absence of a stable degradation layer deter-mined the rate of Y loss, which may further affect the degra-dation and mechanical properties of Mg alloys. Furthermore,this study confirmed that the presence of ions and proteinsin the physiological fluids played an important role in thedegradation of Mg alloys. Therefore, the fluid conditions atthe intended implantation sites should also be considered.

ACKNOWLEDGMENTS

The authors would like to thank Dr. Prashant Kumta and Dr.Abhijit Roy for their kind help with alloy cutting, Dr. YadongWang for his permission to use the Nikon Fluorescence Micro-scope in his laboratory, Dr. Dong Yan for AFM training at theCenter of NanoSciences and NanoEngineering at the Universityof California at Riverside, and Mitch Boretz for proof-reading.

REFERENCES1. Pietak AM, Staiger MP, Huadmai J, Dias G.Magnesium and its

alloys as orthopedic biomaterials: A review. Biomaterials 2006;27:

1728–1734.

2. Witte F.The history of biodegradable magnesium implants: A

review. Acta Biomater 2010;6:1680–1692.

3. Tamimi F, Ibasco S, Meszaros R, Le Nihouannen D, Vengallatore

S, Harvey E, Barralet JE.Magnesium-sputtered titanium for the

formation of bioactive coatings. Acta Biomater 2009;5:2338–2347.

4. Yamasaki Y, Yoshida Y, Okazaki M, Shimazu A, Kubo T, Akagawa

Y, Uchida T.Action of FGMgCO(3)Ap-collagen composite in pro-

moting bone formation. Biomaterials 2003;24:4913–4920.

5. Witte F, Janning C, Willbold E, Vogt C, Nellesen J, Meyer-Linden-

berg A, Windhagen H, Thorey F.Magnesium hydroxide temporar-

ily enhancing osteoblast activity and decreasing the osteoclast

number in peri-implant bone remodelling. Acta Biomater 2010;6:

1861–1868.

6. Brar HS, Platt MO, Sarntinoranont M, Martin PI, Manuel MV.Mag-

nesium as a biodegradable and bioabsorbable material for medi-

cal implants. Jom 2009;61:31–34.

7. Breen DJ, Stoker DJ.Titanium lines: A manifestation of metallosis

and tissue response to titanium alloy megaprostheses at the

knee. Clin Radiol 1993;47:274–277.

8. Tran N, Webster TJ.Nanotechnology for bone materials. Wiley

Interdiscip Rev Nanomed Nanobiotechnol 2009;1:336–351.

9. Lembeck B, Wulker N.Severe cartilage damage by broken poly-L-

lactic acid (PLLA) interference screw after ACL reconstruction.

Knee Surg Sports Traumatol Arthrosc 2005;13:283–286.

10. Witte F, Hort N, Vogt C, Cohen S, Kainer KU, Willumeit R, Feyera-

bend F.Degradable biomaterials based on magnesium corrosion.

Curr Opin Solid State Mater Sci 2008;12:63–72.

11. Zheng YF, Gu XN, Cheng Y, Zhong SP, Xi TF.In vitro corrosion

and biocompatibility of binary magnesium alloys. Biomaterials

2009;30:484–498.

12. Zheng YF, Gu XN, Zhou WR, Cheng Y, Wei SC, Zhong SP, Xi TF,

Chen LJ.Corrosion fatigue behaviors of two biomedical Mg

alloys-AZ91D and WE43-In simulated body fluid. Acta Biomater

2010;6:4605–4613.

13. Gu X-N, Zheng Y-F.A review on magnesium alloys as biodegrad-

able materials. Front Mater Sci China 2010;4:111–115.

14. Jones DS. Pharmaceutical applications of polymers for drug

delivery. Shawbury, U.K.: Rapra Technology Ltd.; 2004. p 124.

15. Wise DL. Handbook of pharmaceutical controlled release technol-

ogy. New York: Marcel Dekker; 2000. p 890.

16. Gupta M, Sharon NML. Magnesium, magnesium alloys, and mag-

nesium composites. New York: Wiley; 2011. p 257.

17. Seal CK, Vince K, Hodgson MA.Biodegradable surgical implants

based on magnesium alloys – A review of current research. IOP

Conf Ser: Mater Sci Eng 2009;4:012011.

18. Song GL.Recent progress in corrosion and protection of magne-

sium alloys. Adv Eng Mater 2005;7:563–586.

19. de Mele MFL, Mueller WD, Nascimento ML, Zeddies M.Degrada-

tion of magnesium and its alloys: Dependence on the composi-

tion of the synthetic biological media. J Biomed Mater Res 2009;

90A:487–495.

20. Lindstrom R, Johansson LG, Thompson GE, Skeldon P, Svensson

JE.Corrosion of magnesium in humid air. Corros Sci 2004;46:

1141–1158.

21. Li LC, Gao JC, Wang Y.Evaluation of cyto-toxicity and corrosion

behavior of alkali-heat-treated magnesium in simulated body

fluid. Surf Coat Technol 2004;185:92–98.

22. Chu PK, Xin YC, Huo KF, Tao H, Tang GY.Influence of aggressive

ions on the degradation behavior of biomedical magnesium alloy

in physiological environment. Acta Biomater 2008;4:2008–2015.

23. Virtanen S, Rettig R.Composition of corrosion layers on a magne-

sium rare-earth alloy in simulated body fluids. J Biomed Mater

Res 2009;88A:359–369.

24. Socjusz-Podosek M, Litynska L.Effect of yttrium on structure and

mechanical properties of Mg alloys. Mater Chem Phys 2003;80:

472–475.

25. Aghion E, Gueta Y, Moscovitch N, Bronfin B.Effect of yttrium

additions on the properties of grain-refined Mg-3%Nd alloy.

J Mater Sci 2008;43:4870–4875.

26. Wu BL, Zhao YH, Du XH, Zhang YD, Wagner F, Esling C.Ductility

enhancement of extruded magnesium via yttrium addition. Mater

Sci Eng 2010;527:4334–4340.

27. Yao HB, Li Y, Wee ATS.Passivity behavior of melt-spun Mg-Y

alloys. Electrochim Acta 2003;48:4197–4204.

28. Uggowitzer PJ, Hanzi AC, Gunde P, Schinhammer M.On the bio-

degradation performance of an Mg-Y-RE alloy with various sur-

face conditions in simulated body fluid. Acta Biomater 2009;5:

162–171.

29. Uggowitzer PJ, Gunde P, Furrer A, Hanzi AC, Schmutz P.The influ-

ence of heat treatment and plastic deformation on the bio-degra-

dation of a Mg-Y-RE alloy. J Biomed Mater Res 2010;92A:409–418.

30. Zhang EL, He WW, Yang K.Effect of Y on the bio-corrosion behav-

ior of extruded Mg-Zn-Mn alloy in Hank’s solution. Mater Sci Eng

C 2010;30:167–174.

31. Song GL, Atrens A.Understanding magnesium corrosion-A frame-

work for improved alloy performance. Adv Eng Mater 2003;5:

837–858.

32. Atrens A, Liu M, Schmutz P, Uggowitzer PJ, Song GL.The influ-

ence of yttrium (Y) on the corrosion of Mg-Y binary alloys. Corros

Sci 2010;52:3687–3701.

33. Butler DL, Gooch C, Kinneberg KRC, Boivin GP, Galloway MT, Nir-

malanandhan VS, Shearn JT, Dyment NA, Juncosa-Melvin N.The

use of mesenchymal stem cells in collagen-based scaffolds for

tissue-engineered repair of tendons. Nat Protoc 2010;5:849–863.

34. Gulotta LV, Kovacevic D, Ehteshami JR, Dagher E, Packer JD,

Rodeo SA.Application of Bone Marrow-Derived Mesenchymal

Stem Cells in a Rotator Cuff Repair Model. Am J Sports Med

2009;37:2126–2133.

35. Kannan MB, Walter R.Influence of surface roughness on the

corrosion behaviour of magnesium alloy. Mater Des 2011;32:

2350–2354.

36. Witte F, Fischer J, Nellesen J, Crostack HA, Kaese V, Pisch A,

Beckmann F, Windhagen H.In vitro and in vivo corrosion mea-

surements of magnesium alloys. Biomaterials 2006;27:1013–1018.

ORIGINAL ARTICLE


in vitro evaluation of the surface effects on magnesium-yttrium alloy degradation and mesenchymal...

Documents