Acta Biomaterialia 6 (2010) 1852–1860

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /ac tabiomat

Fe–Mn alloys for metallic biodegradable stents: Degradation and cellviability studies q

Hendra Hermawan a, Agung Purnama b, Dominique Dube a, Jacques Couet b, Diego Mantovani a,*

a Laboratory for Biomaterials and Bioengineering, Department of Mining, Metallurgy and Materials Engineering, Pav. Adrien-Pouliot, 1745-E, Laval University,1065 Ave de la Médecine, Québec City, QC, Canada G1V 0A6b Laval Hospital Research Center/Québec Heart Institute, Department of Medicine, Laval University, Québec, Canada

a r t i c l e i n f o a b s t r a c t

Article history:Received 27 March 2009Received in revised form 11 November 2009Accepted 17 November 2009Available online 23 November 2009

Keywords:Fe–Mn alloyBiodegradable stentDegradationCell viability

1742-7061/$ - see front matter � 2009 Acta Materialdoi:10.1016/j.actbio.2009.11.025

q Part of the Thermec’2009 Biodegradable MetaProfessor Diego Mantovani and Professor Frank Witte

* Corresponding author. Tel.: +1 418 656 2131x627E-mail address: Diego.Mantovani@gmn.ulaval.ca (

Biodegradable stents have shown their potential to be a valid alternative for the treatment of coronaryartery occlusion. This new class of stents requires materials having excellent mechanical propertiesand controllable degradation behaviour without inducing toxicological problems. The properties of thecurrently considered gold standard material for stents, stainless steel 316L, were approached by newFe–Mn alloys. The degradation characteristics of these Fe–Mn alloys were investigated includingin vitro cell viability. A specific test bench was used to investigate the degradation in flow conditions sim-ulating those of coronary artery. A water-soluble tetrazolium test method was used to study the effect ofthe alloy’s degradation product to the viability of fibroblast cells. These tests have revealed the corrosionmechanism of the alloys. The degradation products consist of metal hydroxides and calcium/phosphoruslayers. The alloys have shown low inhibition to fibroblast cells’ metabolic activities. It is concluded thatthey demonstrate their potential to be developed as degradable metallic biomaterials.

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

1. Introduction

Stents, tiny tubular mesh-like metallic structure implants, haveproved their effectiveness in treating narrowed arteries [1,2]. Nev-ertheless, they present two remaining complications: subacutestent thrombosis [3] and in-stent restenosis [4]. Two approacheshave been proposed to deal with these limitations: the drug elutingand the biodegradable stents. The current technology of drug elut-ing stents is still facing the problem of late stent thrombosis [5].Meanwhile, the implantation of biodegradable stents hopefullywill leave behind only a healed arterial vessel, preventing late stentthrombosis, in-stent restenosis and the prolonged antiplatelettherapy [6]. They are expected to degrade within a reasonable per-iod (12–24 months) [7] after the stented artery has been remod-elled (6–12 months) [8,9]. Biodegradable stents made of metalswere clinically implanted to treat congenital heart disease inbabies [10–12] and to treat critical limb ischemia in adults [13].Recently, a non-randomized multi-center clinical trial on biode-gradable magnesium stents was conducted with encouraginginitial results [14].

ia Inc. Published by Elsevier Ltd. A

ls Special Issue, edited by.0; fax: +1 418 656 5343.

D. Mantovani).

Two classes of materials have been used to prepare biodegrad-able stents: polymers, from the lactic acid, glycolic and caprolac-tone families [15–18], and metals, either magnesium alloys[13,19–25] or pure iron [26–28]. Metallic biodegradable stentswere more developed than their polymeric counterparts. In fact,metals have superior mechanical properties than polymers forreplicating the properties of stainless steel 316L (SS316L), the ref-erence material for coronary stent [29]. Nevertheless, improve-ments are needed mainly to decrease the degradation rate ofmagnesium alloys [14,19,23,25] or to accelerate that of iron-basedstents [26,27]. Some attempts have been made by developing newmagnesium alloys including Mg–Zn–Mn [30], Mg–Ca [31,32] andFe–Mn alloys [33,34]. Fe–Mn alloys, containing between 20 and35 wt.% manganese, exhibited mechanical properties comparableto those of SS316L alloy [33,34]. They possess a similar austenite(c) structure, even though the c forming elements are differentas nickel was used for the SS316L and manganese for the Fe–Mnalloys. The presence of this austenitic phase reduces the magneticsusceptibility compared to SS316L alloy which will give an en-hanced compatibility with the magnetic resonance imaging(MRI). From a biological point of view, the presence of an alloyingelement such as manganese for a biodegradable iron-based alloyappears more appropriate than nickel, the former being essentialto human [35,36] the latter classified as toxic and carcinogenic[37]. An overdosage of manganese could lead to intoxificationand neurotoxicity [38]; however, due to the extensive plasma

ll rights reserved.

H. Hermawan et al. / Acta Biomaterialia 6 (2010) 1852–1860 1853

protein binding that counteract the effect of manganese toxicity,excess of manganese is not reported to be toxic in cardiovascularsystem [39,40]. Moreover, considering the very light weight of astent which is 50–100 mg depending on design, the release ofelement from the alloy during a controlled degradation could beexpected to be lower than their toxic level in the blood.

In developing materials for biodegradable stents, studies ondegradation behaviour and degradation products’ cytotoxicityhave to be considered. The degradation study investigates themechanisms of decomposition and rate. A specific test bench sim-ulating conditions encountered in human coronary artery has beendeveloped by Levesque et al. [20]. The test bench was able to showthe important role of shear stress on degradation behaviour ofmetals which cannot be assessed with the common immersion orelectrochemical test methods [21]. This so-called dynamic degra-dation test method was suggested to be used for evaluating mate-rials proposed for biodegradable stents [21].

The cytotoxicity study can assess the early signs for biocompat-ibility such as acute cytotoxicity. At present time, it is difficult tofind literature on cytotoxicity study of degradable metals for car-diovascular applications. However, methods used in studies onmagnesium alloys dedicated for bone implants could be trans-posed to assess the early sign of cytotoxicity for other metals pro-posed for biodegradable stents. Li et al. have conducted anevaluation of cytotoxicity of alkali heat treated pure magnesiumusing marrow cells [41]. The same indirect contact cytotoxicity testmethod was carried out by Li et al. [31] on Mg–Ca alloys using L-929 cells. The ISO 10993-5:1999 standard was used as a guidefor preparing the extracts [42]. Therefore, in the present study,the new iron-based alloys containing manganese (Fe–Mn) were as-sessed for their degradation characteristics and their early sign ofcytotoxicity using a dynamic degradation test bench and using acell viability test, respectively.

2. Materials and methods

2.1. Materials

Materials used in this study were iron-based alloys containing20–35 wt.% manganese denoted as Fe20Mn, Fe25Mn, Fe30Mnand Fe35Mn. The alloys were prepared through powder sinteringprocess from high purity elemental powders of iron and manga-nese followed by a series of cold rolling and resintering cyclesresulting in a highly dense material. The details about these alloysincluding fabrication process, structure and properties are de-scribed elsewhere [33,34]. The Fe20Mn and Fe25Mn are consti-tuted of c + e phases, whereas Fe30Mn and Fe35Mn arecomposed of single c phase [34]; therefore for degradation tests,specimens of Fe25Mn and Fe35Mn alloys representing two differ-ent microstructure conditions were chosen. Specimens with an ex-posed surface area of �300 mm2 were mounted in acrylic resin.They were then polished using abrasive papers #1000, ultrasoni-cally cleaned in 75% ethanol, air-dried and stored for 24 h in a des-iccator prior to use.

For cell viability tests, powders were used in order to provide anextreme condition of high surface area. Powders of Fe–Mn alloyswith size of �53 lm were prepared by means of mechanical filingand sieving. This size was chosen based on preliminary experi-ments with different powder size where it was found that�53 lm powders induced the most severe effect to the cells. Pow-ders of commercial iron, manganese and SS316L were also used forcomparison purposes (Atlantic Equipment Engineers, Bergenfield,USA). The powders were sterilized with 75% ethanol and rinsedwith phosphate buffered saline prior the tests. The Fe35Mn alloy

was considered for further tests as it contains the highest manga-nese content among the other Fe–Mn alloys under study.

2.2. Degradation test

Test solution was prepared from modified Hank’s solution hav-ing ionic composition and concentration considerably similar tothose of human blood plasma. A pseudo-physiological-like shearstress of 4 Pa was generated by a predetermined laminar flowingsolution in a test bench designed to mimic blood flow conditionin human coronary artery. More details on the test bench and testparameters are reported elsewhere [21]. The specimens were takenout from the test bench after 1 week, 1 month and 3 months andwere then characterized. The temperature of the test solutionwas kept at 37 �C and its pH was also recorded every 24 h.

Solution samplings were carried out during the test and theirconcentration of iron and manganese was measured by a PerkinElmer 3110 atomic absorption spectrometer (AAS). A SiemensD5000 X-ray diffractometer (XRD) with Cu Ka radiation at a scan-ning rate of 0.02�/1.2 s�1 was employed to identify the structure ofdegradation products of the specimens. The microstructure wasstudied with an Olympus PME3 optical microscope (OM) and a JeolJSM-840A scanning electron microscope (SEM). Images obtainedfrom OM were analysed using a quantitative image analyser soft-ware (Clemex Vision) to measure corroded depth on at least 5fields to obtain averages and standard deviations. The chemicalcomposition of degradation products was analysed using X-rayphotoelectron spectroscopy (XPS) in a Phi VersaProbe and usingan energy dispersive spectrometer (EDS) which was coupled withthe SEM. A Cameca SX100 electron probe micro-analyser (EPMA)was used to map iron, manganese and oxygen. Specimens forOM, SEM and EPMA were mounted in acrylic resin to expose theircross-section area and were then mechanically polished usingabrasive papers up to #1200 and 0.1 lm diamond paste.

2.3. Cell viability test

Cell viability tests were carried out by indirect contact with the3T3 mouse fibroblast cell line (3T3). They were cultured inDulbecco’s modified Eagle’s medium (DMEM) with 10% fetal bo-vine serum (FBS), 100 U ml�1 penicillin and 100 mg ml�1 strepto-mycin at 37 �C in a humidified atmosphere of 5% CO2. The cellswere incubated in 24-well tissue culture plates at the density of50,000 cells/well for 24 h to allow attachment.

Samples to be tested were then added into 3 lm tissue cultureinserts in the same medium as for the cells. The content of samplesin the medium was varied and expressed as concentration(mg ml�1). At least 6 replicates were studied for each condition.After 48 h of incubation, the cell viability was assessed usingwater-soluble tetrazolium based assay (10% WST-1, 4-[3-(4-iodo-phenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio]-1,3-benzene-disulfon-ate) for 2 h. The dissolved degradation products were rinsed out asthe medium was aspirated after the treatment with the samples.The absorbances of the solutions were measured spectrophotomet-rically at 440 nm using a VersaMax microplate reader (MolecularDevices, Sunnyvale, USA) and were analysed using the Prism 5 soft-ware (GraphPad Software Inc., San Diego, USA).

3. Results

3.1. Degradation tests

Fig. 1 shows cross-sectional profiles of Fe–Mn specimens beforeand after dynamic degradation test. The degradation (corrosion)took place over the entire surface and then went deeper into the

Fig. 1. Cross-sectional profile of polished Fe–Mn specimens: (a) before and (b and c) after 1 week and 3 months of degradation test respectively, and (d and e) etched Fe25Mnand Fe35Mn specimens after 3 months of degradation test respectively (etchant: Nital 2%).

1854 H. Hermawan et al. / Acta Biomaterialia 6 (2010) 1852–1860

bulk (Fig. 1a–c). The microstructure shows a preferential intergran-ular attack of the Fe25Mn alloy (Fig. 1d), while a more invasive at-tack is visible in Fe35Mn specimen (Fig. 1e). In Fig. 1b and c,specimens were polished with abrasive paper #1200 in order toprevent the edge smoothing effect and preserve the corroded pro-file for measurement of the corroded depth (Fig. 2). Moreover,Fig. 2 shows a slightly deeper corrosion attack for Fe25Mn thanin Fe35Mn specimens. The pH of the test solution was recordedas 7.4 at the beginning and never exceeded more than 7.8 untilthe end of the tests (after 3 months).

Fig. 3 shows the evolution of the concentration of iron and man-ganese ions in solutions during 3 months of degradation test. Therewas no difference in ion concentration for both alloys up to14 days, but then specimens of Fe25Mn alloy released slightlymore ions than those of Fe35Mn alloy.

Fig. 2. Corroded depth as a function of test period for Fe25Mn and Fe35Mn alloys.

Fig. 3. Concentration of iron and manganese ions in test solution as a function ofimmersion time for specimens of Fe25Mn and Fe35Mn alloys measured by the AAS.

Fig. 4 shows microstructure and elemental mappings of thecross-section of Fe35Mn specimen after 3 months of degradationtest. Fig. 4a illustrates a backscattered electron (BSE) image show-ing three regions with different atomic density: bulk, intermediateand top degradation layers. The intermediate layer was probablycracked during curing of the mounting resin or due to a dehydra-tion event. Fig. 4b and c shows that iron and manganese were uni-formly distributed in the intermediate layer and in the bulk, tinydispersed manganese-rich inclusions being visible in the substra-tum. Somewhat less iron and manganese were found in intermedi-ate layers than in the bulk. However in the intermediate layer,manganese was sometimes more visible than iron although itsconcentration was much lower than that of iron. (Fig. 4c). Theconcentration in oxygen was superior in the intermediate than intop layer and also oxygen penetrated locally in the bulk (Fig. 4d).

Fig. 4. Images of cross-section area of Fe35Mn specimens after 3 months of degradation test: (a) BSE image, and (b, c and d) EPMA maps for iron, manganese and oxygen,respectively. The color represents the intensity of the mapped elements.

H. Hermawan et al. / Acta Biomaterialia 6 (2010) 1852–1860 1855

All the three elements demonstrated very low intensities in the toplayer.

The top layer of degradation product contained mainly carbon,oxygen, nitrogen, sodium and trace of phosphorus, chlorine and

Fig. 5. XRD pattern of degradation products and XPS pattern (insert) of top surface

sulphur as measured by the XPS (Fig. 5 insert). Meanwhile theXRD spectrum of the degradation layer (Fig. 5) showed a ratheramorphous pattern with a very low intensity, but it approachedthe pattern of magnetite, Fe3O4. The presence of hydrogen could

of degradation layer for Fe35Mn specimens after 3 months of degradation test.

Table 1Concentration of elements on the degradation surface of Fe35Mn specimen after3 months of degradation test measured by the EDS.

Element Concentration (wt.%)

Flat layer Agglomerates

Iron 47.5 (0.6) 45.9 (0.5)Manganese 7.1 (0.4) 6.9 (0.3)Oxygen 26.6 (0.6) 20.8 (0.6)Chlorine 3.5 (0.1) 1.5 (0.1)Calcium 2.0 (0.1) 4.4 (0.1)Phosphorus 2.1 (0.1) 4.3 (0.1)Sulphur 0.7 (0.1) 0.8 (0.1)

1856 H. Hermawan et al. / Acta Biomaterialia 6 (2010) 1852–1860

not be detected but it is expected as hydrated degradationproducts.

The degradation products adhered to the surface and were notcompletely washed out by the flowing solution in the test bench(Fig. 6a). The surface was covered by a flat degradation layer andspreading agglomerates formed on it (Fig. 6b). At higher magnifica-tion, the porous agglomerates appear like a coral (mineral) struc-ture (Fig. 6c and d). The EDS analysis shows a difference in themain detectable elements between the flat layer and the agglomer-ates shown in Fig. 6 as presented in Table 1. The agglomerates con-tain more calcium and phosphorus compared to the flat layerwhich contains more on oxygen and chlorine.

Standard deviation in parentheses.

Fig. 7. Relative metabolic activity of 3T3 fibroblast cells in presence of variousmetal powders as a function of concentration of the powders.

3.2. Cell viability tests

The WST-1 metabolic assay showed that manganese has thehighest metabolic inhibition effect to the 3T3 fibroblast cells com-pared to iron, SS316L and Fe35Mn alloy (Fig. 7). Its relative meta-bolic activities (RMA) declined to less than 80% when the cellswere exposed to 0.01 mg ml�1 manganese powder, and was unde-tectable when concentration reached 4 mg ml�1. In contrast tomanganese, cells treated with iron and SS316L showed higherRMA which was similar to the control along the tested concentra-tions. For these two powders, the RMA remained high, more than95%, even when their concentration reached 16 mg ml�1. Differ-ently, when manganese is alloyed with iron to form Fe35Mn alloy,it showed less inhibition effect than that of pure manganese asdemonstrated by higher RMA values. The alloy gave a distinctpattern compared to all tested metals in describing its metabolicinhibition (Fig. 7). Its 0% RMA was noted at the concentration of16 mg ml�1.

Fig. 8 provides more evidence on the alloying effect in loweringmetabolic inhibition. The 0.5 mg ml�1 concentration of metalsshowed a good picture of inhibition whereas it was chosen sinceit gave 50% metabolic activity inhibition for manganese. The RMA

Fig. 6. Images of Fe35Mn surface after 3 months of degradation test: (a) macrograph showing the specimen mounted in resin, and (b, c and d) SEM images at differentmagnification.

Fig. 8. Relative metabolic activity of 3T3 fibroblast cells in presence of variousmetal powders at a fixed concentration of 0.5 mg ml�1. Note: Fe + Mn = mixture of65 wt.% of iron powder with 35 wt.% of manganese powder.

H. Hermawan et al. / Acta Biomaterialia 6 (2010) 1852–1860 1857

in the presence of SS316L, iron and Fe35Mn alloy resembled that ofcontrol. It was significantly higher than those of manganese andthe mixture of iron and manganese powders. Fig. 9 shows thatthere was nearly no difference in RMA demonstrated by the alloyscontaining different manganese content.

4. Discussion

4.1. Properties of Fe–Mn alloys

Four alloys with manganese content ranging between 20 and35 wt.% were developed via a powder metallurgy route followedby a series of cold rolling and resintering cycles [33,34]. The phasecomposition, magnetic and mechanical properties of these newly

Table 2Properties of Fe–Mn alloys compared to SS316L (ASTM F138).

Material Nominal main composition(wt.%)

Phase atTroom

Magnetic susceptibilitya

(lm3 kg�1)

Fe20Mn 20 Mn c + e 0.2 (1.1)Fe25Mn 25 Mn c + e 0.2 (0.2)Fe30Mn 30 Mn c 0.2 (0.2)Fe35Mn 35 Mn c 0.2 (0.2)SS316L 18 Cr, 14 Ni c 0.5 (1.7)

a The values in parenthesis are after the specimens were subjected to 20% of plastic d

Fig. 9. Relative metabolic activity of 3T3 fibroblast cells in presence of Fe–Mn alloyspowders with different manganese content at a fixed concentration of 1 mg ml�1.

developed Fe–Mn alloys are presented in Table 2. Their microstruc-ture is mainly composed of c phase with the appearance of e phasein alloys having a lower manganese content. As the manganesecontent increases, the maximum elongation increases while theyield strength decreases. The alloys possess an antiferromagneticbehaviour, giving the same non-magnetic nature as that ofSS316L with best behaviour upon plastic deformation producingminimal magnetic susceptibility [34].

4.2. Degradation rate

The progress of degradation is clearly shown in Fig. 1a–c wherethe corrosion attack progressively went deeper and wider into thebulk of the material as a function of the degradation time. The cor-rosion attack in Fe25Mn was found to be slightly deeper than inFe35Mn (Fig. 3). This could be related to the bi-phase compositionin Fe25Mn where e-epsilon and c-austenite phases coexist. There-fore, it presents more micro-galvanic sites susceptible to corrosioninitiation than in Fe35Mn, which has only c-austenite phase. It wasalso found that intergranular corrosion, which preferably attacksgrain boundaries, was more evident in Fe25Mn than in Fe35Mn(Fig. 1d and e). The intergranular corrosion will mostly lead to alocalized degradation [44] which should be avoided to prevent apremature failure of an implant.

In this study, degradation rate was approached quantitativelyby measuring the corroded depth as a function of degradation time(Fig. 2). During the 3 month test period, the total corroded depth ofFe25Mn specimens was �130 lm and that of Fe35Mn specimenswas �110 lm, corresponding to average degradation rates of�520 and �440 lm year�1, respectively. Compared to pure iron,which has a corrosion rate of 220–240 lm year�1 [26], bothFe–Mn alloys have shown higher corrosion rates. This could implya faster in vivo degradation rate than pure iron, which was notcompletely degraded in the aorta of New Zealand rabbits after18 months [26]. The in vivo degradation is known to be muchslower than that of in vitro [45]. The concentration of iron andmanganese ions in the solution (Fig. 3) presents complementarydata on the rate of ions release. The highest average concentrationreached after 3 months of degradation test was �2 ppm for ironand �1.4 ppm for manganese in Fe25Mn alloy. This measuredion concentration is very low compared to the experiment onAM60B magnesium alloys that reached 50–100 ppm in 14 days[21]. This could be related to the fact that the degradation productsof Fe–Mn alloys which contain iron and manganese was not solu-ble. They mostly adhered to the surface of specimens (Figs. 4 and 6)and thus slowed down the ion exchange between the substrate andsolution.

4.3. In vitro degradation mechanism

Fig. 10 illustrates the degradation mechanism of the Fe–Mnalloy during the dynamic degradation test in modified Hank’ssolution. The mechanism can be divided into four steps:

Yield (0.2%) strength(MPa)

Ultimate strength(MPa)

Maximum elongation(%)

420 700 8360 720 5240 520 20230 430 30190 490 40

eformation. Data compiled from [33,34] and ASTM F138-03 [43].

Fig. 10. Illustration of the corrosion mechanisms for Fe–Mn alloys: (a) initial corrosion reaction, (b) formation of hydroxide layer, (c) formation of pits, and (d) formation ofcalcium/phosphorus layer.

1858 H. Hermawan et al. / Acta Biomaterialia 6 (2010) 1852–1860

1. Initial corrosion reaction (Fig. 10a)

Immediately after immersion in the test solution, the alloy wasoxidized to metal ions, in anodic spots following Eqs. (1) and (2).The electrons from the anodic reaction were consumed by a corre-sponding cathodic reaction and the reduction of oxygen dissolvedin water, following Eq. (3). These reactions occurred randomly overthe entire surface where a difference in potential existed, at grainboundaries and at the interface between different phases.

Fe! Fe2þ þ 2e� ð1Þ

Mn!Mn2þ þ 2e� ð2Þ

2H2Oþ O2 þ 4e� ! 4OH� ð3Þ

2. Formation of hydroxide layers (Fig. 10b)

The released metal ions then reacted with the hydroxyl ion(OH�) released from the cathodic reaction to form insolublehydroxides (hydrous metal oxides) according to Eqs. (4) and (5).Since iron was the main composition of the alloy, the equationsare hereafter written for iron only.

2Fe2þ þ 4OH� ! 2FeðOHÞ2or2FeO:2H2O ð4Þ

4FeðOHÞ2 þ O2 þ 2H2O! 4FeðOHÞ3or2Fe2O3:6H2O ð5Þ

The oxidized (corroded) surface layer of iron alloys normallyconsists of FeO.nH2O at the bottom, Fe3O4.nH2O in the middleand Fe2O3.nH2O on the top [46]. Under visual observation on theFe–Mn specimens, those hydroxides appeared as red–brown(Fe2O3) layer on the top and black (Fe3O4 and FeO) layer on the bot-

tom. The XRD analysis rather detected an amorphous pattern withweak peaks corresponding more to those of magnetite, Fe3O4

(Fig. 5).

3. Formation of pits (Fig. 10c)

Since the hydroxide layers did not homogenously cover the sur-face, Cl� ions from the solution penetrated to compensate the in-crease of metal ions beneath the hydroxide layer. The formedmetal chloride was then hydrolysed by water to the hydroxideand free acid, Eq. (6), lowering the pH value in the pits while thebulk solution remains neutral. This autocatalytic reaction leads tothe formation of pits [47] that grew wider and deeper as shownin Fig. 1.

Fe2þ þ 2Cl� ! FeCl2 þH2O! FeðOHÞ2 þHCl ð6Þ

4. Formation of calcium/phosphorus layer (Fig. 10d)

As the degradation process continued, there was a formation ofa new layer (agglomerates) over the previously formed flat degra-dation layer (Fig. 6). These agglomerates, which appeared like acoral (mineral) structure, contained significant amount of calciumand phosphorus (Table 1). This finding is interesting since the pres-ence of those two elements could lead to the formation of hydroxy-apatite, which is considered biocompatible especially for boneimplants.

4.4. The effect of Fe–Mn alloys on cells

In most cases, metal toxicity arises only when reaction occursbetween metals and body fluids acting as electrolytes. The reaction

H. Hermawan et al. / Acta Biomaterialia 6 (2010) 1852–1860 1859

is electrochemical in nature and it is the metal ion, formed by ano-dic reaction, e.g. Eqs. (1) and (2), that can provide toxic effects [48].This depends on the nature of the metal ions itself against cell met-abolic activities. Fig. 7 shows that the powders of iron and SS316Lat a concentration up to 16 mg ml�1 demonstrated low inhibitionto the 3T3 fibroblast cells. In contrast, the powder of manganeseshowed high inhibition to the cells starting from very low concen-tration. Meanwhile, the powder of Fe35Mn alloy showed a moder-ate inhibition to the cells.

Iron is an essential element with a high toxic level, i.e. 350–500 lg dl�1 in serum [49]. Extracellular iron exclusively bound totransferrin, which maintains iron-soluble and non-toxic [50].Iron-loaded transferrin binds to its specific receptor on the cell sur-face, and undergoes endocytosis. The internalized excess of iron isdetoxified by sequestration into ferritin, an iron storage protein[50]. Those two characteristics of iron explained the low inhibitionof metabolic activity when cells exposed to iron at the range of thetested concentration. Meanwhile, the low inhibition of SS316L wasmainly due to the relatively inert behaviour conferred to this metalby its passive layer composed of chromium oxide (Cr2O3).

On the other hand, despite its essential functions in humanbody, manganese shows a potential to be toxic, i.e. a level of3–5.6 lg dl�1 can cause neurologic symptoms [51]. In tissues,manganese may exist primarily in the form of Mn2+ [52] andmay be oxidized to Mn3+, which is rather reactive and more toxicthan Mn2+ [53]. Manganese exerts its toxic effect by targetingmitochondria, causing a high level of lactic acid [54]. This will de-crease the cells’ ability to cleave WST-1 into soluble formazan,which is the parameter to measure the cell viability. This mecha-nism explained the result which showed that manganese has thehighest inhibition effect to the 3T3 fibroblast cells.

Once iron is alloyed with manganese, i.e. at 35 wt.% ofmanganese, they formed a solid solution where the two atomsare homogenously arranged in a face-centered cubic (fcc) crystalstructure known as the c phase. A structure that is different fromthose of the forming elements, i.e. body-centered cubic (bcc) foriron and simple cubic for manganese. As a consequence, the alloyhas very different characteristics than its forming elements. There-fore, it is expected that the cytotoxic behaviour of the alloys is dif-ferent from that of the forming elements, i.e. less toxic thanmanganese to cells. Fig. 8 clearly shows that alloying significantlyreduced the inhibition effect of manganese to the cells. This is re-lated to slow (small) release of manganese ion during dynamicdegradation test (Fig. 3). If the two elements – iron and manganesepowders – were only mixed together without forming an alloy, themixture, as expected, still has a high inhibition effect as manganeseto the 3T3 fibroblast cells (Fig. 8).

Variations in manganese content in Fe–Mn alloys did notdisplay significant difference in RMA (Fig. 9). However, the inhibi-tion effect of the alloys should be addressed to their degradationbehaviour during the cell viability test period, which took only48 h. It is shown in Fig. 3 that during the first 48 h both alloysshowed no difference in the quantity of the released manganeseions. There could be a slightly different concentration of manga-nese, but this was under the detection limit of the AAS. For longerdegradation period, it is shown that Fe25Mn corroded slightlymore severe and released more manganese ions than Fe35Mn(Figs. 2 and 3). Fig. 9 also shows that the optimum RMA was givenby Fe30Mn, which could be the optimum composition of Fe–Mn al-loys under study in giving compromise between inhibition effectand physical and mechanical properties.

Finally, we would like to highlight some limitations of thiswork. The cell-based assays are simplistic and do not adequatelytest toxic effects. Cell viability is a single test of cellular metabolicactivity, and it can be anticipated that a range of complementaryassays could have been very appropriate. In particular, comple-

mentary assays devoted to compare potential for apoptosis,necrosis, DNA damage, mitochondrial membrane potential, andcytoskeletal rearrangement would greatly increase the power tocompare effects on different toxic and functional endpoints inthe cells. Second, the degradation studies were conducted in a saltsolution that did not contain some of the normal sequestering pro-teins of serum. This more realistic solution would probably changeboth the flow dynamics in the degradation assays and the profile ofmetals on the surface of the degrading pieces. However, it is alsoknown that tests carried out in solution tests containing serumproteins provide synergistic effects that complicated the analyticaltechniques for the characterization and the analysis of the results.In this context, mainly considering that these are pioneering worksin the new field of degradable metals, we decided to start from asimple model. Third, we are conscious that the use of indirect con-tact of the metal powders in the transwells might appear disput-able. Certainly the degradation process of metal particles in thestatic transwell is probably greatly different from the processesthat occur under flow conditions. Vascular stents are in direct con-tact with the cells. However, again, we decided to start from a sim-ple model. These directions will be the focus of future works.

5. Conclusion

The in vitro degradation of Fe–Mn alloys is governed by themechanism of corrosion involving the formation of pits over theentire surface which then went deeper and wider. The alloys werecorroded at an average rate up to �520 lm year�1 which is abouttwo times faster than that of pure iron. The degradation productsconstituted of iron hydroxides and calcium/phosphorus containinglayers. They adhered to the substrate and were not completely sol-uble in the test solution. The release of iron and manganese ionsinto the solution was limited due to the barrier effect of the insol-uble degradation layer.

The Fe–Mn alloys possess a low inhibition effect to 3T3 fibro-blast cells metabolic activities compared to pure manganese. Theinhibition effect increases as the concentration of the alloys in cel-lular medium increases. Its 50% inhibition effect reached at con-centration of 6 mg ml�1 while its 100% inhibition effect wasreached when the concentration exceeded 16 mg ml�1. Finally, itcan be concluded that the study demonstrates evidences of thepotentiality of Fe–Mn alloys to be a biocompatible degradablebiomaterial.

Acknowledgements

The authors would like to acknowledge the kind help and guid-ance of Elise Roussel from Laval Hospital during cell viability tests.Iron powder for the cell viability tests as well as for the Fe–Mn al-loys development was kindly provided by the Quebec Metal Pow-ders Inc. This work was partially supported by the Natural Scienceand Engineering Research Council (NSERC) of Canada, the Collabo-rative Health Research Projects of NSERC and the Canadian Insti-tutes of Health Research (CIHR).

Appendix A. Figures with essential colour discrimination.

Certain figures in this article, particularly Figs. 3, 4, 7 and 10 aredifficult to interpret in black and white. The full colour images canbe found in the on-line version, at doi:10.1016/j.actbio.2009.11.025.

References

[1] Fischman DL, Leon MB, Baim DS, Schatz RA, Savage MP, Penn I, et al. Arandomized comparison of coronary-stent placement and balloon angioplasty

1860 H. Hermawan et al. / Acta Biomaterialia 6 (2010) 1852–1860

in the treatment of coronary artery disease. Stent restenosis studyinvestigators. N Engl J Med 1994;331:496–501.

[2] Serruys PW, de Jaegere P, Kiemeneij F, Macaya C, Rutsch W, Heyndrickx G, et al.A comparison of balloon-expandable-stent implantation with balloonangioplasty in patients with coronary artery disease. Benestent study group.N Engl J Med 1994;331:489–95.

[3] Movahed MR, Vu J, Ahsan C. Simultaneous subacute stent thrombosis of twodrug-eluting stents in the left anterior descending and the circumflex coronaryarteries. Case report and review of the literature. J Invasive Cardiol 2006;18:E198–202.

[4] Mitra AK, Agrawal DK. In stent restenosis: bane of the stent era. J Clin Pathol2006;59:232–9.

[5] Ong AT, McFadden EP, Regar E, de Jaegere PP, van Domburg RT, Serruys PW.Late angiographic stent thrombosis (LAST) events with drug-eluting stents. JAm Coll Cardiol 2005;45:2088–92.

[6] Waksman R. Update on bioabsorbable stents: from bench to clinical. J IntervCardiol 2006;19:414–21.

[7] Serruys PW, Kutryk MJ, Ong AT. Coronary-artery stents. N Engl J Med2006;354:483–95.

[8] El-Omar MM, Dangas G, Iakovou I, Mehran R. Update on in-stent restenosis.Curr Interv Cardiol Rep 2001;3:296–305.

[9] Schomig A, Kastrati A, Mudra H, Blasini R, Schuhlen H, Klauss V, et al. Four-yearexperience with Palmaz-Schatz stenting in coronary angioplasty complicatedby dissection with threatened or present vessel closure. Circulation 1994;90:2716–24.

[10] Zartner P, Cesnjevar R, Singer H, Weyand M. First successful implantation of abiodegradable metal stent into the left pulmonary artery of a preterm baby.Catheter Cardiovasc Interv 2005;66:590–4.

[11] Schranz D, Zartner P, Michel-Behnke I, Akinturk H. Bioabsorbable metal stentsfor percutaneous treatment of critical recoarctation of the aorta in a newborn.Catheter Cardiovasc Interv 2006;67:671–3.

[12] Zartner P, Buettner M, Singer H, Sigler M. First biodegradable metal stent in achild with congenital heart disease: evaluation of macro and histopathology.Catheter Cardiovasc Interv 2007;69:443–6.

[13] Peeters P, Bosiers M, Verbist J, Deloose K, Heublein B. Preliminary results afterapplication of absorbable metal stents in patients with critical limb ischemia. JEndovasc Ther 2005;12:1–5.

[14] Erbel R, Di Mario C, Bartunek J, Bonnier J, de Bruyne B, Eberli FR, et al.Temporary scaffolding of coronary arteries with bioabsorbable magnesiumstents: a prospective, non-randomised multicentre trial. Lancet 2007;369:1869–75.

[15] Stack RS, Califf RM, Phillips HR, Pryor DB, Quigley PJ, Bauman RP, et al.Interventional cardiac catheterization at Duke Medical Center. Am J Cardiol1988;62:3F–24F.

[16] Tamai H, Igaki K, Kyo E, Kosuga K, Kawashima A, Matsui S, et al. Initial and 6-month results of biodegradable poly-l-lactic acid coronary stents in humans.Circulation 2000;102:399–404.

[17] Tsuji T, Tamai H, Igaki K, Kyo E, Kosuga K, Hata T, et al. Biodegradablepolymeric stents. Curr Interv Cardiol Rep 2001;3:10–7.

[18] van der Giessen WJ, Lincoff AM, Schwartz RS, van Beusekom HM, Serruys PW,Holmes Jr DR, et al. Marked inflammatory sequelae to implantation ofbiodegradable and nonbiodegradable polymers in porcine coronary arteries.Circulation 1996;94:1690–7.

[19] Heublein B, Rohde R, Kaese V, Niemeyer M, Hartung W, Haverich A.Biocorrosion of magnesium alloys: a new principle in cardiovascular implanttechnology? Heart 2003;89:651–6.

[20] Levesque J, Dube D, Fiset M, Mantovani D. Investigation of corrosion behaviourof magnesium alloy AM60B-F under pseudo-physiological conditions. MaterSci Forum 2003;426–432:521–6.

[21] Levesque J, Hermawan H, Dube D, Mantovani D. Design of a pseudo-physiological test bench specific to the development of biodegradablemetallic biomaterials. Acta Biomater 2008;4:284–95.

[22] Di Mario C, Griffiths H, Goktekin O, Peeters N, Verbist J, Bosiers M, et al. Drug-eluting bioabsorbable magnesium stent. J Interv Cardiol 2004;17:391–5.

[23] Waksman R, Pakala R, Kuchulakanti PK, Baffour R, Hellinga D, Seabron R, et al.Safety and efficacy of bioabsorbable magnesium alloy stents in porcinecoronary arteries. Catheter Cardiovasc Interv 2006;68:606–17.

[24] Waksman R, Pakala R, Okabe T, Hellinga D, Chan R, Tio MO, et al. Efficacy andsafety of absorbable metallic stents with adjunct intracoronary beta radiationin porcine coronary arteries. J Interv Cardiol 2007;20:367–72.

[25] McMahon CJ, Oslizlok P, Walsh KP. Early restenosis following biodegradablestent implantation in an aortopulmonary collateral of a patient withpulmonary atresia and hypoplastic pulmonary arteries. Catheter CardiovascInterv 2007;69:735–8.

[26] Peuster M, Wohlsein P, Brugmann M, Ehlerding M, Seidler K, Fink C, et al. Anovel approach to temporary stenting: degradable cardiovascular stentsproduced from corrodible metal-results 6–18 months after implantation intoNew Zealand white rabbits. Heart 2001;86:563–9.

[27] Peuster M, Hesse C, Schloo T, Fink C, Beerbaum P, von Schnakenburg C. Long-term biocompatibility of a corrodible peripheral iron stent in the porcinedescending aorta. Biomaterials 2006;27:4955–62.

[28] Waksman R, Pakala R, Baffour R, Seabron R, Hellinga D, Tio FO. Short-termeffects of biocorrodible iron stents in porcine coronary arteries. J Interv Cardiol2008;21:15–20.

[29] Balcon R, Beyar R, Chierchia S, De Scheerder I, Hugenholtz PG, Kiemeneij F,et al. Recommendations on stent manufacture, implantation and utilization.Eur Heart J 1997;18:1536–47.

[30] Song G. Control of biodegradation of biocompatible magnesium alloys. Corr Sci2007;49:1696–701.

[31] Li Z, Gu X, Lou S, Zheng Y. The development of binary Mg–Ca alloys for use asbiodegradable materials within bone. Biomaterials 2008;29:1329–44.

[32] Kannan MB, Raman RKS. In vitro degradation and mechanical integrity ofcalcium-containing magnesium alloys in modified-simulated body fluid.Biomaterials 2008;29:2306–14.

[33] Hermawan H, Alamdari H, Mantovani D, Dubé D. Iron-manganese: New classof degradable metallic biomaterials prepared by powder metallurgy. Powdermetall 2008;51:38–45.

[34] Hermawan H, Dube D, Mantovani D. Degradable metallic biomaterials: designand development of Fe–Mn alloys for stents. J Biomed Mater Res A; in press:doi:10.1002/jbm.a.32224.

[35] Emsley J. The elements. Oxford: Clarendon Press; 1998.[36] Keen CL, Ensunsa JL, Clegg MS. Manganese metabolism in animals and humans

including the toxicity of manganese. In: Siegel A, Siegel H, editors. Manganeseand its role in biological processes. New York: Marcel Dekker; 2000. p. 89–121.

[37] McGregor DB, Baan RA, Partensky C, Rice JM, Wilbourn JD. Evaluation of thecarcinogenic risks to humans associated with surgical implants and otherforeign bodies – a report of an IARC Monographs Programme Meeting. Eur JCancer 2000;36:307–13.

[38] Aschner M, Guilarte TR, Schneider JS, Zheng W. Manganese: recent advances inunderstanding its transport and neurotoxicity. Toxicol Appl Pharmacol2007;221:131–47.

[39] Jynge P, Brurok H, Asplund A, Towart R, Refsum H, Karlsson JO. Cardiovascularsafety of MnDPDP and MnCl2. Acta Radiol 1997;38:740–9.

[40] Crossgrove J, Zheng W. Manganese toxicity upon overexposure. NMR Biomed2004;17:544–53.

[41] Li L, Gao J, Wang Y. Evaluation of cyto-toxicity and corrosion behavior of alkali-heat-treated magnesium in simulated body fluid. Surf Coat Technol2004;185:92–8.

[42] ISO 10993-5. Biological evaluation of medical devices – part 5: tests for in vitrocytotoxicity. Arlington, VA: ANSI/AAMI; 1999.

[43] ASTM F. 138: Standard specification for wrought 18chromium-14nickel-2.5molybdenum stainless steel bar and wire for surgical implants (UNSS31673). West Conshohocken: ASTM International; 2003.

[44] Steigerwald R. Metallurgically influenced corrosion. In: Korb LJ, Olson DL,editors. ASM handbook volume 13: corrosion. Materials Park: ASM; 1987. p.123–35.

[45] Witte F, Fischer J, Nellesen J, Crostack H-A, Kaese V, Pisch A, et al. In vitro andin vivo corrosion measurements of magnesium alloys. Biomaterials2006;27:1013–8.

[46] Roberge PR. Handbook of corrosion engineering. Toronto: McGraw-Hill; 2000.[47] Shreir LL, Jarman RA, Burstein GT. Corrosion, metal/environment reactions. 3rd

ed. Oxford: Butterworth-Heinemann; 2000.[48] Flint GN. A metallurgical approach to metal contact dermatitis. Cont Dermat

1998;39:213–21.[49] Frey RJ. Iron tests. Gale encyclopaedia of medicine. Gale Research; 1999.[50] Papanikolaou G, Pantopoulos K. Iron metabolism and toxicity. Toxicol Appl

Pharmacol 2005;202:199–211.[51] Ellenhorn MJ, Schonwald S, Ordog G, Wasserberger J. Ellenhorn’s medical

toxicology: diagnosis and treatment of human poisoning. Baltimore: Williams& Wilkins; 1997.

[52] Harris RW, Chen Y. Electron paramagnetic resonance and differenceultraviolet studies of Mn2+ binding to serum transferrin. J Inorg Biochem1994;54:1–19.

[53] Chen JY, Tsao GC, Zhao Q, Zheng W. Differential cytotoxicity of Mn(II) andMn(III): Special reference to mitochondrial [Fe–S] containing enzymes. ToxicolAppl Pharmacol 2001;175:160–8.

[54] Hirata Y. Manganese-induced apoptosis in PC12 cells. Neurotoxicol Teratol2002;24:639–53.