This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Shape memory polymer foams for cerebral aneurysm reparation: Effectsof plasma sterilization on physical properties and cytocompatibility

Luigi De Nardo a,*, Rachele Alberti b, Alberto Cigada a, L’Hocine Yahia c,Maria Cristina Tanzi b, Silvia Fare b

a Dipartimento di Chimica, Materiali e Ingegneria Chimica, ‘‘G. Natta”, Politecnico di Milano, Via Mancinelli 7, 20133 Milan, Italyb Biomaterials Laboratory, Bioengineering Department, Politecnico di Milano, Piazza Leonardo da Vinci 32, 20133 Milan, Italy

c Institute of Biomedical Engineering, Ecole Polytechnique de Montreal, Montreal, QC, Canada

Received 3 July 2008; received in revised form 28 October 2008; accepted 25 November 2008Available online 13 December 2008

Abstract

Shape memory polyurethanes (SMPUs) represent promising candidate materials for aneurysm embolization, since they could enableclinical problems still associated with these clinical procedures to be overcome. In this work, we report on the characterization of phys-icochemical, thermomechanical and in vitro interface properties of two SMPU foams (Cold Hibernated Elastic Memory, CHEM), pro-posed as a material for embolization devices in minimally invasive procedures. Moreover, because device sterilization is mandatory forin vivo applications, effects on the properties of the foams after plasma sterilization were also evaluated. Both foams (CHEM 3520 andCHEM 5520) showed excellent shape recovery ability (recovery rate, Rr, up to 99%) in conventional shape recovery tests, performed atconstant heating rate. Transition temperatures (Ttrans), determined by tand peaks in dynamic mechanical analysis (DMA), were 32.2 and45.1 �C, for CHEM 3520 and 5520, respectively. The value of Ttrans affects shape memory ability in the recovery test at 37 �C, whichsimulates the behavior after implantation of the device: in fact, Rr was significantly higher for lower Ttrans foam (Rr � 82% andRr � 46%, respectively, for CHEM 3520 and CHEM 5520). After plasma sterilization performed by a Sterrad� sterilization system,an increase in open porosity was observed: this is probably due to the sterilization cycle; however, no effects on shape recovery behaviorwere observed. Furthermore, plasma treatment had no significant effect on L929 cells in in vitro cytotoxicity tests, performed on cell cul-ture medium extracts in contact with foams for up to 7 days. Moreover, direct cytocompatibility tests showed a good colonization andgrowth from L929 cells on CHEM foams, suggesting the effectiveness of an in vivo healing process. All these results seem to suggest thatCHEM foams could be advantageously used for manufacturing devices for mini-invasive embolization procedures of aneurysms.� 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Shape memory polymers; Minimally invasive structures; Plasma sterilization; Cytocompatibility; Polyurethane foams

1. Introduction

Endovascular surgery represents the more recent andinnovatory frontier in vascular surgery: when comparedto traditional vascular procedures, it is significantly lessinvasive, resulting in shorter hospitalization and fasterreturn to a normal life. For this main reason, together

with a reduction of associated medical costs, during thelast 10 years endovascular procedures have gained recog-nition as an alternative to open surgery for several vas-cular diseases treatments, including arterial aneurysms[1–3]. In 1991 the introduction of Guglielmi detachablecoils (GDCs) changed the therapeutic approach to cere-bral aneurysms [1,4]. GDCs are used in devices to beused in minimal invasive techniques for treating selectedtypes of cerebral aneurysms, with the insertion of anocclusive coil into the aneurysm, to promote its emboli-zation. The surgical procedure can be performed by

1742-7061/$ - see front matter � 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

doi:10.1016/j.actbio.2008.11.017

* Corresponding author. Tel.: +39 0 223993161; fax: +39 0 223993180.E-mail address: luigi.denardo@polimi.it (L. De Nardo).

Available online at www.sciencedirect.com

Acta Biomaterialia 5 (2009) 1508–1518

www.elsevier.com/locate/actabiomat

Author's personal copy

using a catheter inserted inside the blood vessel, eitherunder general anesthesia or light sedation, and usingimaging techniques to observe in real time the patient’svascular system [2].

The US Food and Drug Administration approved theuse of these devices in 1995; today, manifold variationsof the original GDC coil design are available [5]. Encourag-ing results obtained via GDC devices account for theirrapid and progressively broadened use: in many medicalcenters, endovascular treatments represent the primarychoice for at least 60–70% of all ruptured aneurysms andtheir effectiveness in aneurysm rupture prevention has alsobeen established by different multicenter studies [5–7].

Along with the positive results, many endovascularteams of surgeons have progressively experienced differenttypes of negative events. Most endovascular interventionson aneurysms can have, in fact, important drawbacks,the more frequent being a significant incidence of residuallesions, deficient neck healing and recurrences [6,8,9].Therefore, basic research has been focused on identificationor design of new materials to allow the fabrication of moreeffective occlusive devices [10–12]. Among the possible can-didate materials, polyurethanes (PUs) represent an interest-ing option, due to the possibility of tailoringchemicophysical properties in a simple and effective way.Moreover, because of their low thrombogenicity, they arewidely used in cardiovascular applications [13]. Recently,shape memory polyurethanes (SMPUs) have been pro-posed by several authors [14] and their use is particularlyinteresting in the realization of minimally invasive devices[15]. SMPUs could be effectively used as occlusive emboli-zation materials because their shape and dimensions can beminimized to achieve a less invasive insertion; an originalapplication in this sense was proposed by L’H. Yahiaet al. [16,17]. A similar approach has been also adoptedby Maitland et al. for the fabrication of a device for aneu-rysm reparation based on SMP foams, in which shapememory recovery is activated by a laser source. Dataacquired in an in vitro basilar aneurysm model confirmedthe successfulness of this approach [18].

In this work, two shape memory polyurethane foams,Cold Hibernated Elastic Memory (CHEM), with potentialfor biomedical applications as an aneurysm filler, wereinvestigated. The materials were provided by W. Sokolow-ski from Jet Propulsion Laboratory (JPL, Pasadena, CA,USA) and have been synthesized by Mitsubishi HeavyIndustry (Nagoya, Japan) [19–21]. Materials were fullycharacterized in order to obtain a set of parameters forGDC device design: in particular, the ability of shaperecovery at constant heating rate [14] was compared torecovery at body temperature. Moreover, because a devicesterilization is mandatory for this application, effects onmorphological and thermomechanical properties of theCHEM foams after plasma sterilization were also evalu-ated. Finally, in vitro cytotoxicity and cytocompatibilityof the two CHEM foams were evaluated before and aftersterilization, using L929 fibroblasts cells line.

2. Materials and methods

2.1. Materials

Two poly(ether urethane) foams, kindly provided by JetPropulsion Laboratory (Pasadena, CA, USA) were stud-ied: CHEM 3520 (#0406352037) and CHEM 5520(#0406552012). For both materials, glass transition tem-perature activates the shape memory effect (Ttrans = Tg).Nominal glass transition temperature values wereTg = 35 �C and Tg = 55 �C, for CHEM 3520 and CHEM5520, respectively. Cylindrical specimens (diame-ter = 15 mm, h = 10 mm) were obtained from foams witha manual die and used (if not otherwise specified) in mor-phological, chemicophysical, thermomechanical, shapememory characterization and for in vitro cytotoxicity andcytocompatibility tests.

2.1.1. Plasma sterilization

Foam specimens were sterilized using a plasma Sterrad�

sterilization system (100S, Johnson & Johnson) at IstitutoNazionale per lo Studio e la Cura dei Tumori (Milan, Italy)according to producer indications. A schematic descriptionof temperature, pressure and gas of the plasma processinvolved in a Sterrad� sterilization system can be found inRef. [22]; briefly, 400 W radiofrequency power (RF,13.56 MHz) is applied at a pressure of 500 mTorr (67 Pa) tocreate a plasma, after injection of vaporized chemical phaseof H2O2. The overall plasma sterilization cycle is about 52 min.

2.2. Morphological and chemicophysical characterization

2.2.1. Scanning electron microscopy

Surface morphology was evaluated by scanning electronmicroscopy (SEM). Samples were gold sputter-coated(Edward S150B) and observed with a StereoScan 360SEM (Cambridge) at 10–20 kV. Surface microanalyseswere performed with an energy dispersion X-ray spectro-scope (EDS, Oxford INCA 200).

2.2.2. Average foam pore diameter

The average pore diameter was determined according tothe ASTM D3576, based on the number of pore wallsintersecting a reference line. Thin slices (0.2 � 0.3 mm,n = 3) were cut from the foams, and observed with a stereomicroscope (Leika, Wild Heerbrugg). Using a referenceline drawn (chord) on a glass slice, it was possible to eval-uate the pore size by using the following equation:

diameter ¼ Al0=n ð1Þwhere A = 1.623 (by assuming pores to be uniformly dis-tributed and spherical), l0 = 40 mm (chord length) and n

is the number of intersections.

2.2.3. Density

The density values of foams were calculated according tothe standard practice ISO 845, by using 12 cylindrical sam-ples (diameter = 15 mm, h = 10 mm) for each CHEM foam.

L. De Nardo et al. / Acta Biomaterialia 5 (2009) 1508–1518 1509

Author's personal copy

2.2.4. Open pore percentage

The percentage of open pores (open porosity), thataccounts only for the penetrable volume of the polymericfoam (i.e. interconnected pores), was evaluated by gasexpansion porosimeter with a purpose-built pycnometer[23]. The percentage of open porosity was calculated as:

Open porosity ð%Þ ¼ V g � V i

V i� 100 ð2Þ

where Vg is the geometric volume of the foams, and Vi isthe impenetrable volume of the foam. Open porosity wasevaluated performing measurements on cylindrical speci-mens, for each CHEM foam (diameter = 15 mm,h = 10 mm, n = 3).

2.2.5. Water uptake (WU)Water uptake kinetic was evaluated by the immersion of

specimens in vials containing 20 ml of distilled water: theywere incubated at 37 ± 1 �C up to 720 h and, at each timepoint, samples were blotted with laboratory paper andaccurately weighed. Water uptake, WU (%), was calculatedaccording to the following equation:

WU ð%Þ ¼ W t � W 0

W 0

� 100 ð3Þ

where W0 is the dry weight of the foams, and Wt is the wetweight at each time point. Tests were carried out intriplicate.

2.2.6. Extract analysis

CHEM foam samples (18–20 mg, n = 3) were incubatedat 37 �C for time periods ranging from 3 h up to 7 days in5 ml of phosphate buffered saline (PBS, pH 7.4) containing0.02% NaN3 as bacteriostatic agent. PBS extracts wereanalyzed by size-exclusion HPLC at 1 ml min�1 flow rate,with a Bio-Rad chromatography system (Bio-Gel TSK-30and TSK-40 columns, Bio-Rad and UV detector) equippedwith ValueChrom chromatography software. As control,PBS with NaN3 was also analyzed by HPLC. Tests wereperformed in triplicate.

2.3. Thermomechanical characterization

2.3.1. Uniaxial cyclic compression

Cyclic compression tests were performed according toUNI 6350–68, in wet conditions (deionized water) at differ-ent temperatures (Tg � 20 �C, Tg, Tg + 20 �C) [24], byusing a uniaxial electromechanical system (Instron� model4200, load cell 1 kN), equipped with a thermoregulatedchamber. Cylindrical specimens (diameter = 15 mm,h = 10 mm) were tested at 1 mm min�1 crosshead rate, per-forming three cycles up to 50% strain. For each tempera-ture and CHEM foam, tests were performed in triplicate.

2.3.2. Dynamic mechanical analysis

Dynamic mechanical analysis (DMA) was performedusing a DMA 2980 (TA Instrument) analyzer in compres-

sion mode in the temperature range 0 � 180 �C (or up tofoam melting temperature), with a heating rate of 1 �Cmin�1, 1 Hz frequency and 50 lm strain. Storage modulus(E0), loss modulus (E00) and tand (tand = E00/E0) wererecorded. Tests were performed in duplicate, according tothe ASTM 1640, using cylindrical specimens (diame-ter = 15 mm, h = 10 mm).

2.3.3. Shape memory recovery characterization

Shape memory recovery tests were performed in com-pression mode on cylindrical specimens (diame-ter = 15 mm, h = 10 mm): specimens were heated up toTH = Tg + 30 �C and then compressed to em = 50%. Atconstant strain (em), each sample was cooled down toTL = Tg � 30 �C to fix the temporary shape. At T = TL,stress was released and the sample was allowed to recoverto eu; eu represents the deformation after the release ofthe stress, and it is a consequence of the fact that resultingtemporary shape always differs from the shape achieved bydeformation [14]. Specimens were subsequently either: (i)heated at constant rate (Hr = 1 �C min�1) or (ii) main-tained at constant temperature (T = 37 �C), up to600 min. In both cases, values of recovered deformationas a function of temperature or time and the final deforma-tion (ep) were recorded. Shape memory properties of eachmaterial were quantified by calculating strain recovery rate(Rr), recovery behavior as a function of temperature(Rr(T)) at constant heating rate and recovery behavior asa function of time (Rr(t)) at constant temperature, accord-ing to the following equations:

RrðT Þ ¼eðT Þ � ep

emð4Þ

RrðtÞ ¼eðtÞ � ep

emð5Þ

Rr(T) and dRr(T)/dT were plotted vs. temperature, Rr(t)vs. time.

2.4. In vitro L929 cell interaction

In vitro cytotoxicity of the extracts and cytocompatibil-ity of CHEM foams were assessed using L929 murine fibro-blasts cells line (ECACC No. 85011425).

2.4.1. In vitro cytotoxicity tests

Extracts were obtained according to ASTM F619.Foam samples were disinfected by soaking in 70% etha-nol solution and exposure to UV light (10 min for eachside), and then rinsed in sterile PBS. Three samples(diameter = 6 mm, h = 3 mm) for each foam wereimmersed in Dulbecco’s Modified Eagle’s Medium(DMEM) with 10% fetal bovine serum (FBS) and 1%penicillin/streptomycin, maintaining a material/mediumratio of 0.04 cm3 ml�1. After 2 or 7 days of incubation,medium extracts were put in contact with L929 (den-sity = 2.5 � 104 cells ml�1) in 96-well tissue culture plates

1510 L. De Nardo et al. / Acta Biomaterialia 5 (2009) 1508–1518

Author's personal copy

(TCPs). Cells were cultured for 48 h, and cell viabilitywas investigated by MTT colorimetric assay (M5655,Sigma). MTT solution was added to each well andremoved after 4 h of incubation at 37 �C: formazan crys-tals were dissolved in dimethylsulfoxide (DMSO) and theabsorbance measured at 570 nm with a spectrophotome-ter; measured absorbance was expressed as relative ratioover control cells on TCPs in 2 and 7 day extracts. Testswere performed in triplicate.

2.4.2. In vitro cytocompatibility tests

Foams (cylindrical specimens, diameter = 6 mm,h = 3 mm) were disinfected by soaking in 70% ethanolsolution, by UV and rinsed with sterile PBS before seedingL929 cells (density = 1 � 105 cells ml�1). After 24 h, 3 daysand 7 days of incubation, cell viability was assessed usingMTT colorimetric assay in triplicate. To evaluate cell mor-phology, 3 h and 1, 3 and 7 days after seeding, CHEMfoams samples (n = 2 for each time point) were preparedfollowing a standard procedure [25], dehydrated with dif-ferent ethanol concentrations and gold sputtered forSEM observation.

2.4.3. Statistical analysis

Statistical analysis (Origin 6.0 software) was performedusing a t-test (Student test), with significance levelp = 0.05. Before each statistical test, normal distributionwas verified by normal probability plots.

3. Results

3.1. Morphology and chemicophysical characterizations

3.1.1. Foam morphology

In Fig. 1A and C, SEM micrographs of CHEM 3520 and5520 can be observed; both materials showed high porosityvalues, although lack of pore interconnection is evidenced.No significant difference in density values was observed(p > 0.05) between the two foams (q � 0.12 g cm�3), despitethe difference in the average pore diameter (diame-ter = 419 lm and diameter = 375 lm, respectively, forCHEM 5520 and CHEM 3520 foams). A lower percentageof open pores was exhibited by CHEM 3520 compared toCHEM 5520 (Table 1). Plasma sterilization seems to havean effect on foam morphology, as shown in Fig. 1B andD. An increase in open pore percentage was observed forboth foams: after sterilization (PLS specimens, Table 1)the increase in open porosity resulted about +138% forCHEM 3520, and +27% for CHEM 5520.

3.1.2. Chemicophysical properties

In Fig. 2, water uptake (WU%) kinetics are reported.CHEM 5520 exhibited a faster water absorption thanCHEM 3520. After plasma sterilization (3520 PLS and5520 PLS), an increase in WU can be observed, in particu-lar for CHEM 5520, that, after 600 h, seems to have not yetreached the plateau (Fig. 2).

3.2. Thermomechanical properties

3.2.1. Uniaxial compression tests

Results of uniaxial compression tests, performed on wetsamples on both CHEM 3520 and 5520 at different temper-atures (T = Tg, T = Tg ± 20 �C), are reported in Fig. 3.The uniaxial cyclic compression tests carried out onCHEM 3520 foam at different temperature resulted in sim-ilar stress/strain curves: increase in stress/strain curve slopeis linear up to 25–35% strain (Fig. 3A, C, and E). Curvesshowed the typical behavior of a rubber-like material, withsmall hysteresis areas that decreased with higher number ofcycles and test temperature. Residual deformation was inthe range 5 � 15%.

CHEM 5520 foam showed a different compressivebehavior at the three test temperatures: a linear increasein the curves slope until 10% strain is followed by a plateauzone (Fig. 3B) for the first compression cycle, whereascurves (#2 and #3, Fig. 3B) showed a progressive decreasein the hysteresis areas. Although three cycle tests were per-formed for each testing temperature, in Fig. 3D and F onlythe first cycle is reported, as after the first cycle samplesfailed.

3.2.2. Dynamic mechanical analysis

Results of dynamic mechanical analysis (DMA) onboth CHEM foams, before (CHEM 3520 and CHEM5520) and after plasma sterilization (CHEM 3520PLSand CHEM 5520PLS), are reported in Fig. 4 andresumed in Table 2. As expected, Tg values for CHEM3520 foam are lower than those for CHEM 5520, andthey are also closer to the nominal Tg values (Table 2).On the contrary, CHEM 5520 foam exhibited a largerdifference between the Tg and nominal value. It is possi-ble to observe, for both foams, a two orders of magni-tude decrease for E0 from the glassy to the rubberystate (Fig. 4A and C). In addition, the melting tempera-ture (Tm) of CHEM 3520 was lower than the one forCHEM 5520 foam (Fig. 4B and D). Plasma sterilizationseems to have slight effects on dynamic mechanical prop-erties as well as on transition temperatures for bothfoams; in particular, an increase in the Ttrans is evidentfor 5520 PLS specimens (Fig. 4C).

3.2.3. Shape memory tests

The results of the shape memory recovery tests, per-formed at constant heating rate, on samples before(CHEM 3520 and CHEM 5520) and after plasma steriliza-tion (CHEM 3520 PLS and CHEM 5520 PLS), arereported in Fig. 5. High final shape recovery values arenoticeable (>95%), as well as a high shape retention(>98.5%) for both foams. Moreover, shape memory behav-ior does not seem affected by the plasma sterilization.

The shape memory recovery tests carried out at constanttemperature (T = 37 �C), showed a higher final shaperecovery for CHEM 3520 (�82%) than for CHEM 5520(�46%), as reported in Fig. 6.

L. De Nardo et al. / Acta Biomaterialia 5 (2009) 1508–1518 1511

Author's personal copy

3.3. In vitro L929 cell interaction

Cytotoxicity tests on extracts in contact with CHEMfoams before and after plasma sterilization were performedat two different extraction times (2 and 7 days). Absor-bance values of MTT tests (Fig. 9) resulted comparablebetween the two time points: no significant differences(p > 0.05) were noticed among untreated and plasma-trea-ted CHEM foams, for both extraction times.

Morphology of L929 cells adhered onto CHEM 3520and 5520 foams are respectively shown in Figs. 7 and 8,

and results of biochemical assays (MTT test) in Fig. 10.Both CHEM foams showed good cells proliferation up to7 days of culture: a homogeneous distribution of cells onthe surface of the two materials was observed, with a goodspreading onto the pore surfaces. SEM observations areconfirmed by the absorbance values related to the MTTassay (Fig. 10). Cell viability increased from the first tothe third day after seeding, and after 7 days the absorbancevalue had more than doubled than after day three. Onceagain, L929 cell proliferation was similar onto both foams.

4. Discussion

In interventional radiology, therapeutic approach forcerebral aneurysms can be based on either endovascularor surgical techniques, where the former is performed byplacing an occlusive device through a catheter, allowingits selective positioning as close as possible to the patholog-ical process [26]. Several attempts have been made to selectthe appropriate geometry and material to realize thesedevices: iron particles, detachable balloons and platinumcoils, for instance, have been successfully used in somepatients [4,7,26,27]. However, their applicability was lim-ited by high rates of particle migration, balloon deflationand, eventually, aneurysm rupture [7]. A large part of theseproblems arises from material properties and/or geometriesof such devices. For this main reason, different researchersfocused their studies on the identification or synthesis ofnew materials to design more effective occlusive devices[10–12,18]. Shape memory polymers are ideal candidatesfor applications in which a temporary shape has to be pre-served until the device is placed in the cavity to be filled,allowing minimally invasive surgical procedures [15,28].Cold Hibernated Elastic Memory foams (CHEM), a class

Fig. 1. SEM micrographs of CHEM polyurethane foams: (A) CHEM 3520, (B) CHEM 3520 plasma sterilized PLS (C) CHEM 5520 and (D) CHEM 5520plasma sterilized PLS (scale bar 500 lm).

Table 1Open porosity before and after plasma sterilization (PLS) for CHEM 3520and CHEM 5520 foams.

Non-treated (%) PLS (%)

CHEM 3520 21 ± 5 50 ± 9CHEM 5520 59 ± 5 75 ± 12

Fig. 2. Water uptake kinetics for CHEM 3520 and CHEM 5520 beforeand after (PLS) plasma sterilization.

1512 L. De Nardo et al. / Acta Biomaterialia 5 (2009) 1508–1518

Author's personal copy

of shape memory foams, were proposed by Yahia et al. aspotential cerebral aneurysm fillers for endovascular surgeryapplications in previous work [16]; from this study, weevaluated different material properties that are critical inthe design of these devices based on the principle of shapememory recovery.

4.1. Morphological parameters

Both CHEM 3520 and CHEM 5520 showed a highporosity value, with an overall low pore interconnection:dimension and percentage of pores are similar, whileCHEM 5520 foam seems to have a better pore interconnec-tion. These morphological properties depend on the foamsynthesis process [19–21]: in particular, porosity of these

foams is obtained via an appropriate selection of a blowingagent [21].

An increase in open porosity, due to collapse of porewalls, is evident after the sterilization process and wasobserved by SEM analysis, water uptake kinetics and openporosity values. Plasma sterilization seems in fact to have arole in terms of increasing the pore interconnection andthis effect is probably related to the pressure variation levelin the Sterrad� sterilization chamber [22]. In fact, items tobe sterilized are placed inside the vacuum chamber of Ster-rad� and the chamber is initially evacuated to 500 mTorr(67 Pa) for 20 min. Vaporized H2O2 is injected (at400 mTorr) and low temperature plasma, with 400 W RFpower at 500 mTorr pressure, is applied for 17 min. Thiscycle is repeated twice before returning the chamber to

Fig. 3. Uniaxial cyclic compressive tests for CHEM 3520 (A, C, and E) and CHEM 5520 (B, D, and F) foams performed at different temperatures: (A andB) T = Tg � 20 �C, (C and D) T = Tg, and (E and F) T = Tg + 20 �C.

L. De Nardo et al. / Acta Biomaterialia 5 (2009) 1508–1518 1513

Author's personal copy

atmospheric pressure: pores are hence exposed to a nega-tive pressure of approximately 100 kPa that could promotethe collapse of the thinner pore walls. Moreover, activeplasma species diffuse into the chamber through the perfo-rations of electrode plates and react with the polymer sur-face, but these interactions are probably limited by theplastic wrapping of the specimens [22]. However, effectson the water uptake could be also caused by modificationin surface chemical properties (i.e. foams became morehydrophilic), as reported in Ref. [29]: foams became morehydrophilic, affecting the water uptake kinetics. Most ofthe observed alterations, in terms of surface chemical prop-erties, are probably due to the chemical agent (H2O2): infact, several authors stated that, in the Sterrad� sterilizers,the efficacy of the system is correlated to the hydrogen per-oxide phase exposition instead of plasma itself [30].

Such alterations in foam morphology, induced byplasma sterilization, have to be taken into account, when

both sterilization technology and foams are consideredfor a peculiar application: however, they should representan interesting instrument when designing endovasculardevices. In fact, the mechanism by which endovascularembolization devices are effective seems to be related tohemodynamic changes in the aneurysm cavity. The bloodstasis induced by the presence of an embolization deviceprovokes a thrombosis, which is then progressivelyreplaced by a fibrous reaction [26]. A better water uptaketogether with an increase in pore interconnection couldplay a positive role in this sense.

4.2. Thermomechanical characterization

The main advantage derived from using GDC systems isrelated to their controllable detaching from a guide, onlywhen its positioning in the aneurysm pouch is judged cor-rect [26], but they are time-consuming and often requireretreatment [18]. Several authors proposed the use ofSMP to realize fillers for pathological defects reparation[14,16,18] to overcome challenging problems: interest inthis class of materials derive from the fact that SMPs canbe formed into a specific primary shape and temporarilyreformed into a stable secondary shape, significantly smal-ler than the primary. By imposing an external temperature,a controllable deploying to the primary shape can be pro-moted. Evaluation of the shape recovery ability, in termsof final recovery, recovery rates and activating tempera-

Fig. 4. DMA outputs: (A) storage modulus and (B) tand for CHEM 3520 untreated and plasma sterilized (PLS); (C) storage modulus and (D) tand forCHEM 5520 untreated and plasma sterilized (PLS).

Table 2Transition temperatures values for CHEM 3520 and CHEM 5520 foamsuntreated and plasma treated (PLS), obtained by DMA analysis.

Tg [�C] Tm [�C]

E0 onset tand peak E0 onset

CHEM 3520 Non-treated 16.3 32.2 129.6PLS 17.3 32.3 117.2

CHEM 5520 Non-treated 29.5 45.1 158.0PLS 35.5 49 137.0

1514 L. De Nardo et al. / Acta Biomaterialia 5 (2009) 1508–1518

Author's personal copy

tures, is hence mandatory. In this work, both foams studiedshowed good thermomechanical properties, with differenttransition temperatures that affect the shape memorybehavior, in particular when tests are performed evaluatingthe shape recovery as a function of time at body tempera-ture. The different values of glass transition and meltingtemperatures can be related to the different stoichiometricratios between hard and soft segment domains in thefoam’s chemical composition [14,31,32]; in particular, the

appropriate selection of difunctional isocyanates, polyolsand chain extenders, their relative ratio and the curing pro-cess result in different glass transition and crystallinity [21].

With the aim of realizing fillers for embolization proce-dures, two different approaches, based on SMPs, have beenproposed in terms of driving forces for device deployment:(i) selection of a material with a Ttrans near the body tem-perature (Ttrans � 37 �C) and the use of body temperatureto activate the shape memory, or (ii) using of an externalheat source (as proposed mainly by Maitland et al. [18]).The choice of one of these approaches affects not onlythe device design, but also the way of performing shaperecovery tests. In fact, even if both foams showed goodshape recovery as function of temperature, differences inglass transition temperature, the temperature activatingthe shape recovery process, affect their selection for medicalapplications: this aspect became evident when tests wereperformed at constant temperature (T = 37 �C), as a func-tion of time. As discussed by Lendlein and Kelch [14],shape memory effect can be quantified by thermomechani-cal investigations. The measurements are generally per-formed by means of a mechanical tester equipped with athermochamber and different test protocols can be usedto evaluate shape memory ability: for example, in the pro-gramming procedure (imposing the temporary shape), in

Fig. 5. Shape memory recovery tests performed at constant heating rate (Hr = 1 �C min�1) for (A and B) CHEM 3520, and (C and D) CHEM 5520untreated and plasma sterilized (PLS) foams.

Fig. 6. Shape memory recovery tests carried out at T = 37 �C for CHEM3520 and CHEM 5520 untreated foams.

L. De Nardo et al. / Acta Biomaterialia 5 (2009) 1508–1518 1515

Author's personal copy

stress or strain controlling, and in terms of temperaturerate and final temperature applied. A correct selection ofthe characterization procedure is a crucial point whendesigning a medical device. When performing recoverytests at constant temperature (T = 37 �C), that is similarto actual device ambient application, it was evident that abetter recovery has been achieved by CHEM 3520 foam,which has a Tg lower than body temperature and generallyhigher than ambient temperature. These results, and the

test parameters used, have a strong influence on the choiceof the driving force allowing shape recovery (body temper-ature vs. external heat source).

4.3. In vitro L929 cells interaction

In this work, preliminary investigation of L929 cell cyto-toxicity and cytocompatibility was conducted. This cell linewas selected because the optimal response of an endovascu-

Fig. 7. SEM micrographs of adhesion and proliferation of L929 cells onto CHEM 5520 foam (A) 3 h, (B) 24 h, (C) 72 h and (D) 7 days after seeding. Scalebar 200 lm.

Fig. 8. SEM micrographs of adhesion and proliferation of L929 cells onto CHEM 3520 foam (A) 3 h, (B) 24 h, (C) 72 h and (D) 7 days after seeding. Scalebar 200 lm.

1516 L. De Nardo et al. / Acta Biomaterialia 5 (2009) 1508–1518

Author's personal copy

lar embolization device is represented by the formation of aneo-endothelium on aneurysm access, which can beachieved only when fibrous reaction around the device iscompleted [26].

Tests confirmed that CHEM foams hold a potential forbiological tissue interaction: both materials showed a goodL929 cytotoxicity response for tests performed on extractup to 7 days. These results were confirmed by HPLC anal-ysis (data not shown) of the extract: no release of lowmolecular weight products had been noticed, on bothuntreated and PLS foams. Cell proliferation tests, referredto a similar time period, showed a good response of cells tofoam substrate, suggesting the effectiveness of the healingprocesses. Moreover, polyurethane substrates are interest-ing because a modulation of the host response can be easilyrealized by incorporating or functionalizing the surfacewith appropriate drugs or biological factors [33].

5. Conclusions

This study was aimed at evaluating the application ofCHEM foams as material for aneurysm devices: physico-

chemical, thermomechanical and in vitro cell interactionhave been evaluated in order to understand the feasibilityof the use of porous structure with shape memory proper-ties in design of embolization devices. Thermomechanicaltests for both CHEM 3520 and 5520 showed a high shaperecovery: this property is necessary for a potential use inendovascular procedures because devices should be com-pressed and inserted in the human body by using a catheterand then recover their original shape at body temperature.

Preliminary studies on plasma sterilization effects haveshown interesting results: increase in volume percentageof open cells noticed as consequence of sterilizationcould be a positive factor for the cells colonization, over-coming the lack of pore interconnection observed onuntreated foams. Finally, in vitro cytotoxicity testsshowed no differences between Sterrad� sterilized anduntreated foams.

All these results confirm the possibility of using CHEMfoams as material for the realization of aneurysm fillers,although further studies should be performed. In particu-lar, shape recovery behavior in wet conditions at 37 �Cand the pressures exerted during the shape recovery phasewhen the expansion’s volume is limited appears necessary.Moreover, hemocompatibility and platelets adhesion onfoams should be also evaluated since, in vivo, they will bein contact with blood.

Acknowledgements

The authors thank W. Sokolowski (Jet Propulsion Lab-oratory, Pasadena, CA, USA) for providing the CHEMSMPu foams, and Prof. M. Nava (Istituto Nazionale perlo Studio e la Cura dei Tumori, Milan, Italy) for plasmasterilization treatments. The authors thank also Dr. M.Moscatelli for her helpful suggestions and discussionsabout the manuscript and her support in mechanical test-ing, and Dr. L. Draghi for text revisions.

References

[1] Murayama Y, Vinuela F, Duckwiler GR, Gobin YP, Guglielmi G.Embolization of incidental cerebral aneurysms by using the Guglielmidetachable coil system. J Neurosurg 1999;90:207–14.

[2] Ng P, Khangure MS, Phatouros CC, Bynevelt M, ApSimon H,McAuliffe W. Endovascular treatment of intracranial aneurysms withGuglielmi detachable coils: analysis of midterm angiographic andclinical outcomes. Stroke 2002;33:210–7.

[3] Vignali C, Cioni R, Neri E, Petruzzi P, Bargellini I, Sardella S, et al.Endoluminal treatment of abdominal aortic aneurysms. AbdomImaging 2001;26:461–8.

[4] Guglielmi G, Vinuela F, Dion J, Duckwiler G. Endovasculartreatment of aneurysms with coils–response. J Neurosurg1992;76:338–9.

[5] Yuki I, Murayama Y, Vinuela F. Development of medical devices forneuro-interventional procedures: special focus on aneurysm treat-ment. Expert Rev Med Devices 2005;2:539–46.

[6] Patterson MA, Jean-Claude JM, Crain MR, Seabrook GR, CambriaRA, Rilling WS, et al. Lessons learned in adopting endovasculartechniques for treating abdominal aortic aneurysm. Arch Surg2001;136:627–34.

Fig. 9. MTT absorbance values for L929 cells on culture medium extractsafter 2 and 7 days for CHEM untreated and plasma sterilized (PLS)foams.

Fig. 10. MTT absorbance values for L929 cells seeded onto CHEM 3520and CHEM 5520.

L. De Nardo et al. / Acta Biomaterialia 5 (2009) 1508–1518 1517

Author's personal copy

[7] Qureshi AI, Janardhan V, Hanel RA, Lanzino G. Comparison ofendovascular and surgical treatments for intracranial aneurysms: anevidence-based review. Lancet Neurol 2007;6:816–25.

[8] Raymond J, Desfaits AC, Roy D. Fibrinogen and vascular smoothmuscle cell grafts promote healing of experimental aneurysms treatedby embolization. Stroke 1999;30:1657–64.

[9] Raymond J, Roy D. Safety and efficacy of endovascular treatment ofacutely ruptured aneurysms. Neurosurgery 1997;41:1235–45, discus-sion 1245–1236.

[10] Mottu F, Gailloud P, Massuelle D, Rufenacht DA, Doelker E. Invitro assessment of new embolic liquids prepared from preformedpolymers and water-miscible solvents for aneurysm treatment.Biomaterials 2000;21:803–11.

[11] Murayama Y, Vinuela F, Tateshima S, Song JK, Gonzalez NR,Wallace MP. Bioabsorbable polymeric material coils for embolizationof intracranial aneurysms: a preliminary experimental study. JNeurosurg 2001;94:454–63.

[12] Oowaki H, Matsuda S, Sakai N, Ohta T, Iwata H, Sadato A, et al.Non-adhesive cyanoacrylate as an embolic material for endovascularneurosurgery. Biomaterials 2000;21:1039–46.

[13] Tanzi MC. Bioactive technologies for hemocompatibility. Expert RevMed Devices 2005;2:473–92.

[14] Lendlein A, Kelch S. Shape-memory polymers. Angew Chem Int Edit2002;41:2034–57.

[15] Fare S, De Nardo L, De Cicco S, Jovenitti M, Tanzi MC. Differentprocessing methods to obtain porous structure in shape memorypolymers. Mater Sci Forum 2007;539–543:663–8.

[16] Metcalfe A, Desfaits AC, Salazkin I, Yahia L, Sokolowski WM,Raymond J. Cold hibernated elastic memory foams for endovascularinterventions. Biomaterials 2003;24:491–7.

[17] Sokolowski W, Metcalfe A, Hayashi S, Yahia LH, Raymond J.Medical applications of shape memory polymers. Biomed Mater2007;2:S23–7.

[18] Maitland DJ, Ward SI, Ortega JM, Buckley PR, Rodriguez J,Hartman J, et al. Prototype laser-activated shape memory polymerfoam device for embolic treatment of aneurysms. J Biomed Opt2007;12:030504.

[19] Sokolowski WM. Cold hibernated elastic memory self-deployableand rigidizable structure and method therefor. US6, 702, 9762004:3.

[20] Sokolowski WM, Chmielewski AB, Hayashi S, Yamada T. Coldhibernated elastic memory (CHEM) self-deployable structures. Pro-

ceedings of SPIE—The International Society for Optical Engineering1999;3669:179–185.

[21] Hayashi S, Fujimura H. Shape memory polymer foam. US Patent 5,049, 591, 1991, p. 5.

[22] Lerouge S, Wertheimer MR, Yahia LH. Plasma sterilization: a reviewof parameters, mechanisms, and limitations. Plasmas Polym2001;6:175–88.

[23] Draghi L, Resta S, Pirozzolo MG, Tanzi MC. Microspheres leachingfor scaffold porosity control. J Mater Sci Mater Med 2005;16:1093–7.

[24] Li F, Zhang X, Hou J, Xu M, Luo X, Ma D, et al. Studies onthermally stimulated shape memory effect of segmented polyure-thanes. J Appl Polym Sci 1997;64:1511–6.

[25] Visai L, Rindi S, Speziale P, Petrini P, Fare S, Tanzi MC. In vitrointeractions of biomedical polyurethanes with macrophages andbacterial cells. J Biomater Appl 2002;16:191–214.

[26] Laurent A. Materials and biomaterials for interventional radiology.Biomed Pharmacother 1998;52:76–88.

[27] Bradac GB, Bergui M, Stura G, Fontanella M, Daniele D, Gozzoli L,et al. Periprocedural morbidity and mortality by endovasculartreatment of cerebral aneurysms with GDC: a retrospective 12-yearexperience of a single center. Neurosurg Rev 2007;30:117–25,discussion 125–116.

[28] Wong YS, Xiong Y, Venkatraman SS, Boey FYC. Shape memory inun-cross-linked biodegradable polymers. J Biomat Sci-Polym E2008;19:175–91.

[29] Lerouge S, Tabrizian M, Wertheimer MR, Marchand R, Yahia L.Safety of plasma-based sterilization: surface modifications of poly-meric medical devices induced by Sterrad and Plazlyte processes. Bio-Med Mater Eng 2002;12:3–13.

[30] Krebs MC, Becasse P, Verjat D, Darbord JC. Gas-plasma steriliza-tion: relative efficacy of the hydrogen peroxide phase compared withthat of the plasma phase. Int J Pharm 1998;160:75–81.

[31] Lin JR, Chen LW. Study on shape-memory behavior of polyether-based polyurethanes. I. Influence of the hard-segment content. J ApplPolym Sci 1998;69:1563–74.

[32] Lin JR, Chen LW. Study on shape-memory behavior of polyether-based polyurethanes. II. Influence of soft-segment molecular weight. JAppl Polym Sci 1998;69:1575–86.

[33] De Nardo L, Fare S, Di Matteo V, Cipolla E, Saino E, Visai L, et al.New heparinizable modified poly(carbonate urethane) surfacesdiminishing bacterial colonization. J Mater Sci Mater Med2007;18:2109–15.

1518 L. De Nardo et al. / Acta Biomaterialia 5 (2009) 1508–1518