lable at ScienceDirect

Biomaterials 31 (2010) 7485e7493

Contents lists avai

Biomaterials

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

The effect of a slow mode of BMP-2 delivery on the inflammatory responseprovoked by bone-defect-filling polymeric scaffolds

Gang Wu a,b, Yuelian Liu a,b, Tateyuki Iizuka c, Ernst Bruno Hunziker a,*aCenter of Regenerative Medicine for Skeletal Tissues, Department of Clinical Research, University of Bern, Murtenstrasse 35, CH-3010 Bern, Switzerlandb Section of Oral Implantology, Department of Oral Function, Academic Center of Dentistry Amsterdam (ACTA), Louwesweg 1, 10066 EA Amsterdam, The NetherlandscDepartment of Cranio-Maxillofacial Surgery, University of Bern, Inselspital, CH-3010 Bern, Switzerland

a r t i c l e i n f o

Article history:Received 8 April 2010Accepted 23 June 2010Available online 18 July 2010

Keywords:BiocompatibilityCalcium-phosphate coatingControlled drug releaseForeign-body giant cellsInflammationOsteogenesis

* Corresponding author. Tel.: þ41 31 632 86 85; faxE-mail address: ernst.hunziker@dkf.unibe.ch (E.B.

0142-9612/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.biomaterials.2010.06.037

a b s t r a c t

We investigated the inflammatory response to, and the osteoinductive efficacies of, four polymers(collagen, Ethisorb�, PLGA and Polyactive�) that bore either an adsorbed (fast-release kinetics) ora calcium-phosphate-coating-incorporated (slow-release kinetics) depot of BMP-2. Titanium-plate-supported discs of each polymer (n ¼ 6 per group) were implanted at an ectopic (subcutaneous) ossi-fication site in rats (n ¼ 48). Five weeks later, they were retrieved for a histomorphometric analysis of thevolumes of ectopic bone and foreign-body giant cells (a gauge of inflammatory reactivity), and the degreeof polymer degradation. For each polymer, the osteoinductive efficacy of BMP-2 was higher when it wasincorporated into a coating than when it was directly adsorbed onto the material. This mode of BMP-2carriage was consistently associated with an attenuation of the inflammatory response. For coatedmaterials, the volume density of foreign-body giant cells was inversely correlated with the volumedensity of bone (r2 ¼ 0.96), and the volume density of bone was directly proportional to the surface-areadensity of the polymer (r2 ¼ 0.97). Following coating degradation, other competitive factors, such as thebiocompatibility and the biodegradability of the polymer itself, came into play.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Fractures of the orbital, maxillary and zygomatic bones arecommon in both children [1] and adults [2], and these lesions oftenrequire surgical reconstruction. For this purpose, the grafting ofautologous bone is still considered to be the optimal therapy [3].However, such material is of limited availability, and its removal isassociated with donor-site morbidity and pain [4e6]. The graftingof allogeneic bone is an alternative [7,8], but the use of this materialalso has its drawbacks, which include the risk of transmittinginfectious (viral) micro-organisms from the donor to the host [9,10]and of triggering a host immune response [11]. Consequently, non-osseous scaffolding materials have been sought. Both natural andsynthetic polymers have been considered and tested in a clinicalsetting. The use of natural polymers, such as collagen [12], has notmet with unqualified success, since these materials are potentiallyimmunogenic and, moreover, are not readily amenable to fabrica-tion in a stable, solid form. Synthetic polymers, on the other hand,being non-immunogenic and fully resorbable, are a popular option,

: þ41 31 632 49 55.Hunziker).

All rights reserved.

and developments in this field have been rapid during the pastfew decades [13]. For the clinical repair of defects within laminarbone, several synthetic polymers are available. These includepolydioxanone [14e16] and mixtures of either polylactate andpolyglycolate [17e19] or polyethylene glycol and polybutyleneterephthalate (Polyactive�) [20,21].

Although these natural and synthetic polymers furnish a scaf-fold for the deposition of new tissue, they are not e indeed cannotbe e intrinsically osteoinductive. During their biodegradation, theyare replaced not by bone but by dense collagenous tissue [15,22,23].Consequently, the mechanical stability of the repair tissue issometimes insufficient. For example, when such polymers are usedto fill large bony defects within the orbital floor, the repair tissuethat is deposited cannot support the eyeball, thereby resulting inpersistent enophthalmos or diplopia [14,15].

The property of osteoinductivity can be conferred only by anosteogenic agent, such as BMP-2 (bone morphogenetic protein 2),which is a member of the transforming growth factor beta family[24]. BMP-2 can induce ossification at ectopic sites [25], and hasbeen used extensively to enhance bone formation at orthotopicones, both in experimental animals and in clinical trials [25e29].However, the mode of delivery of BMP-2 to its potential site ofaction is a crucial determinant of its osteoinductive efficacy. This

G. Wu et al. / Biomaterials 31 (2010) 7485e74937486

agent is water-soluble, and when it is administered topically byinjection, it diffuses away too rapidly to be effective [30].Attempts have been made to adsorb BMP-2 to polymeric bone-defect-filling materials. Although bone formation can be therebytriggered, the pharmacological dose that must be applied toinduce ossification is exceedingly high: the adsorbed depot isreleased so rapidly [31,32] that most of it diffuses away before itcan take effect. Moreover, apart from the unnecessary cost ofsuch wastage, the transiently very high concentrations of thisgrowth factor can trigger undesirable side-effects. These includean over-stimulation of local bone resorption, an over-stimulationof bone formation at both local and distant sites, and anaugmentation of local neuropathy [33,34].

In previous studies, we have refined a technique for thebiomimetic deposition of osteoconductive calcium-phosphatecoatings, which can be co-precipitated with an osteogenic agent,such as BMP-2 [35]. Using this technique, BMP-2 can be trulyincorporated into the inorganic crystalline latticework, and issubsequently liberated at a slow, steady rate [27,36] in a cell-mediated manner. Using pharmacologically low loading doses,BMP-2-functionalized calcium-phosphate coatings have beenshown to induce and sustain bone formation for several weeksaround titanium implants that were inserted at both ectopic [26]and orthotopic sites [27] in experimental animals.

In the present study, we wished to apply these BMP-2-func-tionalized calcium-phosphate coatings to four polymericbone-defect-filling materials that are characterized by differentgeometries and surface densities, and to compare their inflamma-tion-provoking reactivities, their biodegradabilities and theirosteoinductive efficacies at an ectopic (subcutaneous) site in rats.The polymers tested were collagen (Helistat�), a combination ofpolyglactin 910/poly-p-dioxanone fleece and poly-p-dioxanone foil(Ethisorb�), a co-polymer of glycolic and lactic acids (PLGA), anda co-polymer of polyethylene oxide terephthalate and polybutyleneterephthalate (Polyactive�).

2. Materials and methods

2.1. Experimental design

The physical characteristics of the one natural polymer (collagen: Helistat�,Integra, Plainsboro, NJ, USA), and the three synthetic ones [Ethisorb� (Johnson andJohnson, USA), PLGA (Smith and Nephew, UK) and Polyactive� (Genci, USA)], as wellas those of the protein-functionalized layers of calcium phosphate with which theywere coated and the protein-release profiles in vitro, are described elsewhere [37].The purpose of the present study was to compare the inflammation-provokingreactivities, the biodegradabilities and the osteoinductive efficacies of the nativepolymers, with or without an adsorbed depot of BMP-2, and of the coated polymers,with or without an incorporated depot of BMP-2, at an ectopic (subcutaneous) site inrats, 5 weeks after implantation. The inflammatory response (gauged by deter-mining the volume density of foreign-body giant cells), the degradation of thepolymer, and the volume density of bone were evaluated histomorphometrically.

Table 1Initial loading doses of BMP-2 (expressed in micrograms per cubic millimetre ofpolymeric material). Mean values (n¼ 6) are represented together with the standarderror of the mean (SEM).

(mg/mm3) Collagen Ethisorb� PLGA Polyactive

Incorporated BMP-2Mean 0.66 0.10 0.70 0.21SEM 0.0577 0.0043 0.0787 0.0053

Adsorbed BMP-2Mean 0.13 0.21 0.26 0.06SEM 0.0010 0.0015 0.0016 0.0003

2.2. Biomimetic coating procedure

1-cm-diameter discs of each polymer type, which were of variable thickness(collagen: 1 mm; Polyactive�: 2 mm; Ethisorb�: 0.6 mm; and PLGA: 2 mm), werecoated with a layer of calcium phosphate according to a standard biomimeticprocedure [35,38]. Initially, a thin (1e3-mm-thick) layer of amorphous calciumphosphate was deposited, which served as a seeding substratum for the growth ofa more substantial crystalline one. The amorphous layer was produced byimmersing the polymer discs in five-times-concentrated simulated body fluid(683.8 mM NaCl, 12.5 mM CaCl2$2H2O, 5 mM Na2HPO4$2H2O, 21 mM Na2CO3), whichcontained 7.5 mM MgCl2$6H2O to inhibit crystal growth. The samples were incu-bated in this solution for 24 h at 37 �C. After freeze-drying in an evacuated chamberfor 24 h, they were transferred to a supersaturated solution of calcium phosphate(4 mM CaCl2$2H2O, 136 mM NaCl, 2 mM Na2HPO4), which was buffered with 50 mM

TRIS (pH 7.4). The samples were incubated in this medium (10 ml per sample) for48 h at 37 �C, and then freeze-dried in a vacuum. The entire coating procedure wasconducted under sterile conditions.

2.3. Incorporation of BMP-2 into calcium-phosphate coatings

Human recombinant BMP-2 [derived from a Chinese Hamster Ovary (CHO) cellline (InductOs�, Wyeth�, Cambridge, MA, USA)] was introduced into the supersat-urated solution of calcium phosphate during the coating procedure (see above) ata final concentration of 10 mg/ml. It is co-precipitated with the inorganic compo-nents, and is thereby incorporated into the crystalline latticework [35]. The amountof BMP-2 that was incorporated into each coating (Table 1) was determined using anenzyme-linked immunosorbent assay (ELISA) kit (PeproTech EC, London, UK), aspreviously described [26].

2.4. Adsorption of BMP-2 onto uncoated samples

A 20-ml drop of a stock solution (500 mg of BMP-2/ml) was deposited upon theupper and the lower surfaces of each polymer disc in turn, allowing for completeevaporation between the applications under sterile, ambient conditions. Theamount of adsorbed BMP-2 was pre-determined at 10 mg per sample for eachpolymer type (Table 1).

2.5. Animal model of ectopic bone formation: experimental set-up

One experimental and three control groups were established for each polymertype: (i) coated polymer bearing an incorporated depot of BMP-2 (experimentalgroup); (ii) uncoated polymer (negative control for the effects of a calcium-phos-phate coating and of BMP-2); (iii) coated polymer (negative control for the effects ofBMP-2); (iv) uncoated polymer bearing an adsorbed depot of BMP-2 (positivecontrol for the effects of BMP-2). To neutralize the effects of irregular stress-fieldsgenerated within the dorsal skin and body muscles of the rats, each polymer wasaffixed with Vicryl 5-0 thread (Ethicon) to a titanium disc (10 mm in diameter and1 mm in thickness; Ti Grade 4 medical, Fredec AG, Murten, Switzerland) via two 0.5-mm-diameter holes, one at each pole. Six samples per group were distributedamongst 48 rats. Under conditions of general anaesthesia, two discs were implantedper animal within the dorsal subcutaneous tissue, one on the left side and one on theright side. Each rat always received either BMP-2-containing samples or non-BMP-2-containing ones. This strategy was adopted to avoid the possibility of cross-reactivity. With this precondition, the samples were distributed amongst theanimals according to a systematic random protocol. The samples were retrieved 5weeks after surgery for the histomorphometric evaluation of the host inflammatoryresponse (gauged by estimating the volume density of foreign-body giant cells), thedegradation rate of the polymers, the volume density of bone and the osteoinductiveefficacy of BMP-2. On the basis of our previous findings, we know that a coating-incorporated depot of BMP-2 in the order of magnitude that was used in the presentstudy can sustain bone-formation activity for minimally 5 weeks [26].

2.6. Histological processing

The rats were sacrificed by administering an overdose of gaseous carbondioxide, which induced cardiac arrest. The implanted polymers were retrievedtogether with a minimum quantity of the surrounding tissue, which embraced thecapsule of fibrous connective tissue. The retrieved samples (as well as unimplantedcontrols) were chemically fixed by immersion in 10% formaldehyde solution forseveral days at ambient temperature. They were then rinsed in tap water, dehy-drated in ethanol and embedded in methylmethacrylate. Applying a systematicrandom-sampling protocol (with a random start at the left-hand margin of thespecimen) [39], each disc was cut perpendicular to its flat surface into 8e12 slices,600 mm in thickness and 1mm apart, using a diamond saw. The slices weremountedon plexiglas holders, polished, and surface-stained with McNeal’s Tetrachrome,basic Fuchsine and Toluidine Blue O [40]. Using this protocol, newly-formed bonestains deep red, cell nuclei blue, collagen fibres pink and the calcium-phosphatecoating pale red. The slices were then examined in a Nikon-Eclipse E-1000 lightmicroscope and photographed in colour at a final magnification of either �200(collagen and Polyactive�) or �300 (Ethisorb� and PLGA). Approximately 25photomicrographs were collected per sample using a systematic random-sampling

G. Wu et al. / Biomaterials 31 (2010) 7485e7493 7487

strategy. These prints were used for the histomorphometric analysis of the variousstereological estimators.

2.7. Histomorphometry

2.7.1. Total volume of subcapsular tissue: reference volumeThe total volume of tissue that was embraced by the capsule of fibrous

connective tissue (reference volume) was estimated using Cavalieri’s methodology[41]. This involves measuring the cross-sectional area of a defined number of tissuesections at a fixed distance apart through the reference volume. The cross-sectionalarea of each section was estimated using the point-counting technique [39]. Thereference volume is calculated bymultiplying the sum of the cross-sectional areas ofthe sections by the fixed distance between them.

2.7.2. Surface-area density at time-point zeroThe surface-area density (the area of a surface per unit volume) of each polymer

typewas estimated at time-point zero using a cycloid grid, which was superimposedon light micrographs of the sectioned material [42].

2.8. Total volume of bone, of foreign-body giant cells and of the remaining polymericmaterial, 5 weeks after implantation

The total volumes of bone, foreign-body giant cells and the remaining polymericmaterial were estimated 5 weeks after implantation from the volume densities (the

Fig. 1. Light micrographs of sections through discs of the indicated polymer types in theuncoated polymer bearing an adsorbed depot of BMP-2 (Ads. BMP-2); unfunctionalized coatweeks after implantation. Bone (B) was deposited only in association with polymeric matercells (FBGC) were encountered in each of the groups. P ¼ polymeric material; F ¼ fat tissuToluidine Blue O. Bars ¼ 200 mm.

volume proportion of the phase of interest within a reference volume) and thesubcapsular reference volume (see above). The volume density of each estimatorwas determined stereologically from its area density on tissue sections by the point-counting technique [39].

2.9. Osteoinductive efficacy of BMP-2

Five weeks after implantation, the coatings associated with each polymer typehad been completely degraded. Hence, the amount of BMP-2 liberated was equiv-alent to the initial loading dose. The depot of BMP-2 that was adsorbed ontouncoated polymers is known to be completely exhausted within the first few days ofimplantation [32]. The osteoinductive efficacy of BMP-2 was estimated by dividingthe total volume of bone that was deposited after 5 weeks by the total quantity ofBMP-2 that was either incorporated into the coated polymers or adsorbed ontouncoated ones.

2.10. Statistical analysis

All data are presented as mean values together with either the standard devi-ation or the standard error of the mean. Data pertaining to each group werecompared using a one-way analysis of variance (ANOVA). The level of significancewas set at p < 0.05. SPSS statistical software (version 11.0.4 for an Apple Macintoshcomputer) was used for this evaluation. Post-hoc comparisons were made usingBonferroni corrections.

four groups [coated polymer bearing an incorporated depot of BMP-2 (Inc. BMP-2);ed polymer (Coating only); and unfunctionalized uncoated polymer (Polymer only)], 5ials (coated or uncoated) that had been functionalized with BMP-2. Foreign-body giante. The sections were surface-stained with McNeal’s Tetrachrome, basic Fuchsine and

Fig. 2. Graph depicting the volume density of bone that was deposited 5 weeks after implantation within the subcapsular space (reference volume) of each polymer type. Eachpolymeric material was functionalized with BMP-2, which was either adsorbed directly onto the uncoated material or incorporated into a calcium-phosphate coating. Mean values(n ¼ 6 for each group) are represented together with the standard deviation.

G. Wu et al. / Biomaterials 31 (2010) 7485e74937488

3. Results

Five weeks after implantation, each polymer disc was sur-rounded by a capsule of dense, fibrous connective tissue. Withinthe confines of this capsule, viz., within the subcapsular space, nobone tissue had been deposited on or around any of the polymericmaterials that had not borne either a directly adsorbed or a coating-incorporated depot of BMP-2 (Fig. 1). The volume density of bonethat was associated with polymers bearing a calcium-phosphatecoating into which BMP-2 had been incorporated was highest forcollagen and Ethisorb�, lower for PLGA, and lowest for Polyactive�

(Fig. 2). For each polymer type, the volume density of bone asso-ciated with the uncoated material bearing an adsorbed depot ofBMP-2 was consistently lower than that associated with the coatedcounterpart bearing an incorporated depot of the drug (Fig. 2). Thevolume density of bone deposited after 5 weeks in associationwithpolymers bearing a coating-incorporated depot of BMP-2 wasproportional (r2 ¼ 0.97) to the surface-area density of the polymer(Fig. 3).

Fig. 3. Graph depicting the linear relationship existing between the volume density ofbone that was deposited after 5 weeks in association with polymers bearing a coating-incorporated depot of BMP-2 and the surface-area density of the polymer beforeimplantation. Mean values (n ¼ 6) are represented together with the standard devi-ation. The correlation coefficient (r2) ¼ 0.97.

Inflammatory activity was gauged by estimating the volume offoreign-body giant cells (Fig. 4) within the fibrous capsule. The totalvolume of foreign-body giant cells was lowest in association withcollagen, Ethisorb� and PLGA, and highest in association withPolyactive� (Fig. 5). For collagen, the total volume of foreign-bodygiant cells was negligible in all groups except that in which thepolymer bore an unfunctionalized coating. For Ethisorb� andPLGA, the total volume of foreign-body giant cells was low andsimilar in each of the four groups. For Polyactive�, the total volumeof foreign-body giant cells was lower in association with thepolymer that bore a coating-incorporated depot of BMP-2 than inany of the other three groups, for which the values were similar.However, if the total volume of foreign-body giant cells is expressedrelative to the subcapsular volume (Fig. 6), then a different pictureemerges: the highest volume density of foreign-body giant cellswas associated with coated, but unfunctionalized, collagen. More-over, the negligible differences between the various groups forEthisorb� and PLGAwere accentuated, revealing the lowest valuesfor the polymer that bore a coating-incorporated depot of BMP-2.The volume density of foreign-body giant cells was inverselyproportional (r2 ¼ 0.96) to the volume density of bone that wasdeposited after 5 weeks in association with polymers bearinga coating-incorporated depot of BMP-2 (Fig. 7).

The influence of functionalizing the coatings with BMP-2 on thedegradation of the polymeric materials themselves was evaluated.Coated collagen, Ethisorb� and PLGA underwent significantdegradation during the 5-week implantation period, and func-tionalization of the coating with BMP-2 tended to accelerate thisprocess, although the effect attained significance only for collagenand PLGA (Fig. 8). Coated Polyactive� underwent no significantdegradation during the 5-week implantation period, and func-tionalization of the coating with BMP-2 had no influence on theprocess.

On the other hand, the coatings of each polymer type under-went complete degradation during the 5-week implantationperiod. Hence, within this timespan, all coatings into which BMP-2had been incorporated had liberated their initial depot of the drug(see Table 1). Uncoated polymers bearing an adsorbed depot ofBMP-2 were assumed, on the basis of existing evidence [32], tohave been exhausted of their entire initial loading dose during thefirst few hours of implantation. The osteoinductive efficacy of BMP-2 in association with each polymer type was estimated by dividing

Fig. 4. Light micrographs of foreign-body giant cells that were associated withuncoated, non-functionalized discs of Polyactive� (A), Ethisorb� (B) and PLGA (C) 5weeks after implantation. These cells, which contain numerous nuclei (arrows), inti-mately follow the surface contours of the longitudinally (A), transversely (B, C) andobliquely sectioned (C) polymeric material (asterisks) (N.B.: By the 5th postoperativeweek, the uncoated, non-functionalized collagenous polymer had been completelydegraded). The sections were surface-stained with McNeal’s Tetrachrome, basicFuchsine and Toluidine Blue O. Bars ¼ 30 mm.

G. Wu et al. / Biomaterials 31 (2010) 7485e7493 7489

the total volume of bone that was deposited after 5 weeks by thetotal quantity of BMP-2 that was initially either incorporated intothe coated materials or adsorbed onto the uncoated ones (Fig. 9).For each polymer type, the osteoinductive efficacy of BMP-2 was

higher when the drug was incorporated into a coating thanwhen itwas directly adsorbed onto the uncoated material (Fig. 9). Thedifference was least (non-significant) for Polyactive� and greatestfor Ethisorb� and collagen; PLGA held an intermediate position.Since the initial loading doses of BMP-2 were of the same order ofmagnitude for each polymer type (see Table 1), and since thepatterns of coating deposition and protein incorporation have beenpreviously shown to be the same [37], the different osteoinductiveefficacies of BMP-2 in association with the four polymer types mayreflect, either directly or indirectly, differences in their surface-areadensity, since, as aforementioned, this parameter was proportionalto the volume density of bone deposited within the 5-weekimplantation period (Fig. 3).

4. Discussion

In the fields of dental, maxillofacial and orthopaedic surgery,non-immunogenic, biodegradable and osteoconductive materialsthat can be functionalized with an osteogenic agent are needed asviable options to autologous (or even allogeneic) bone for thereconstruction of voluminous osseous defects. Several syntheticpolymers are commercially available and have been clinicallyapplied for this purpose. However, like all foreign materials, thesepolymers trigger an inflammatory response in which macrophagesand foreign-body giant cells participate [43,44]. The pro-inflam-matory cytokines that are released by migrant T-lymphocytessuppress bone formation [45]. This osteoinhibitory cytokine signalfar outweighs the potentially beneficial osteoinductive one thatemanates from macrophages in the form of BMP-2 release. Asa result of this inflammatory reactivity, the polymeric scaffoldbecomes ensheathed by a capsule of dense, fibrous connectivetissue. This walling-off of the implanted material impedes itsosseointegration with the surrounding tissue.

Quite apart from the triggering of this osteoinhibitory inflam-matory reactivity, the available polymeric scaffolds must be appliedin conjunction with an osteogenic agent, such as BMP-2, to beosteoinductive. Hitherto, BMP-2 has been directly injected into theimplantation site. But since the drug is water-soluble, and sincemarked inflammatory reactivity is associated with the scaffold, itmust be applied at extremely high concentrations to be effective,with the undesirable consequences that bone is also formed withinnon-osseous tissues, such as muscle. More recently, attempts havebeen made to adsorb BMP-2 to the surfaces of polymeric scaffolds.But an adsorbed depot of the drug is released too rapidly to belocally effective. And if the loading dose is raised, then the afore-mentioned side-effects ensue.

We have refined a technique for the biomimetic production ofBMP-2-functionalized calcium-phosphate implant coatings. Sincethe BMP-2 is thereby incorporated into the inorganic crystallinelatticework, it is released gradually, at a low pharmacological dose,and in a cell-mediated, physiological-like manner. This mode ofBMP-2 delivery is conducive to sustained bone-formation activity.In the present study, we wished to investigate the inflammatoryresponses to, and the osteoinductive efficacies of, four types ofpolymeric material (with different chemical compositions andsurface characteristics) that bore either an adsorbed depot ofBMP-2 (fast-release kinetics) or a calcium-phosphate coating intowhich the drug had been biomimetically incorporated (slow-release kinetics). For this purpose, we drew on a well-establishedectopic ossification model in rats [46].

With respect to osteogenic activity, the findings of the presentstudy confirm those of a previous investigation in which titaniumimplants rather than polymeric scaffolds served as a carrier for theBMP-2-functionalized calcium-phosphate coating [46]: This modeof BMP-2 delivery was more efficacious in inducing and sustaining

Fig. 5. Graph depicting the total volume of foreign-body giant cells that was found 5 weeks after implantation within the subcapsular (reference) space of each polymer type; thefour different groups are indicated. Mean values (n ¼ 6) are represented together with the standard deviation.

G. Wu et al. / Biomaterials 31 (2010) 7485e74937490

bone formation over a 5-week period thanwas a directly-adsorbeddepot of the drug (Figs. 2 and 9).

One of the most striking and novel findings of the present studyrelates to the influence of the mode of BMP-2 carriage on theinflammatory response that is triggered by the polymeric materials.As gauged by the volume density of foreign-body giant cells withinthe subcapsular space, the inflammatory response was less severewhen the polymeric material bore a BMP-2-functionalized calcium-phosphate coating than when it bore a directly-adsorbed depot ofthe drug. In a clinical context, this finding is of paramount impor-tance, since it points to a means not only of curbing inflammatoryreactivity but also of dramatically lowering the pharmacologicaldose of the applied osteogenic agent to a safe level. The relevant

Fig. 6. Graph depicting the volume density of foreign-body giant cells within the subcapsutogether with the standard deviation. *: p < 0.05; **: p < 0.01; ***: p < 0.001.

literature affords several clues that might help to account for ourfinding. In the first place, an augmentation of osteogenic activityhas been shown to dampen inflammation by suppressing theparticipation of foreign-body giant cells. The effect is mediated byosteopontin, which, by occupying the CD44 surface receptors onmacrophages, inhibits the multi-nucleation process [47] that isindispensable for the formation of foreign-body giant cells [48]. Thepresence of osteoprogenitor cells is also known to dampeninflammatory reactivity [49].

Secondly, the inflammatory response may be modulated bydifferences in the controlled, macrophage-activated degradation ofthe polymeric material itself, since the degradation products canfurther promote the recruitment and activation of macrophages. In

lar (reference) space, 5 weeks after implantation. Mean values (n ¼ 6) are represented

Fig. 7. Graph depicting the inverse linear relationship existing between the volumedensity of bone that was deposited after 5 weeks in association with polymers bearinga coating-incorporated depot of BMP-2 and the volume density of foreign-body giantcells that had accumulated after 5 weeks, likewise in association with polymersbearing a coating-incorporated depot of BMP-2. Mean values (n ¼ 6) are representedtogether with the standard deviation. The correlation coefficient (r2) ¼ 0.96.

G. Wu et al. / Biomaterials 31 (2010) 7485e7493 7491

this context, the nature of the polymeric material, as well as theabsence or presence of a BMP-2-functionalized calcium-phosphatecoating, will play a role. During the early post-implantation phase(up to 7 days), the presence of a calcium-phosphate coating wouldbe expected to afford the host some protection against the pro-inflammatory effects of polymer degradation. Indeed, during thisphase, which was not investigated in the present study, theinflammatory response that is triggered by the calcium-phosphatecoatings should not be influenced by the nature of the underlyingpolymer. However, after about 2e3 weeks, the coating will haveundergone complete degradation, and the underlying material willbe exposed. The degradation of the polymeric material will thenprovoke a secondary inflammatory response, which will varyaccording to the pro-inflammatory characteristics of the break-down products. The influence of differences in the mechanicalproperties of the four materials could be excluded, since eachpolymer was stabilized by a titanium disc (to neutralize the effectsof irregular stress-fields generated within the dorsal skin and body

Fig. 8. Graph depicting the influence of functionalizing coated forms of each polymer type wthe remaining polymeric material is represented for each polymer type (n ¼ 6) together w

muscles of the rats). A direct correlation existed between thesurface-area density of the polymeric material and the volumedensity of bone [r2 ¼ 0.97 (Fig. 3)], and an inverse one between thevolume density of bone and the volume density of foreign-bodygiant cells [r2 ¼ 0.96 (Fig. 7)]. Hence, for polymers that bore a BMP-2-functionalized calcium-phosphate coating, the surface-areadensity of the polymer itself overrode the contribution of othercharacteristics, such as chemical composition or macroscopic form(fibrillar or sponge-like), on bone formation. Nevertheless, theinfluence of competitive factors, such as biocompatibility andbiodegradability, on the inflammatory response, and thus on boneformation, cannot be neglected at a later stage. During the 5-weekmonitoring period, the four types of polymer were degraded tovarying degrees (Fig. 8), the two extremes being represented byPolyactive� (less than 10%) and collagen (more than 90%). Thesedifferences in degradation are reflected in the volume density offoreign-body giant cells that occurred within the subcapsular space(Fig. 6). Since the discs of collagen had been almost completelydegraded by the 5-week juncture, the peak in foreign-body-giant-cell activity had passed. Discs of Polyactive�, on the other hand, hadonly just begun to undergo degradation at this time. Hence,foreign-body-giant-cell activity was still rising at the 5-weekjuncture, and the peak had not yet been attained. The biodegrad-ability of a material will depend not only upon intrinsic factors,such as chemical composition, but also upon extrinsic ones, such asthe local enzymatic apparatus, as well as upon co-operative ones,such as the biocompatibility of the material with the bodilycompartment in which it is implanted. As aforementioned, thepresence of a BMP-2-functionalized calcium-phosphate coatingdelays the biodegradation of the underlying polymer. In the case ofcollagen particularly, which is characterized by a high pro-inflam-matory response [50] and a high degradation rate, osteogenicactivity is probably favoured by the delay in, and perhaps also theattenuation of, inflammatory reactivity (as being triggered at a laterstage, viz., after the coating has been degraded and when bone-formation activity is well underway).

The amounts of BMP-2 that were incorporated into the calcium-phosphate coatings of the four different polymer types ranged from0.1e0.7 mg/mm3 of polymeric material (Table 1). The amounts ofBMP-2 that were adsorbed directly onto the surfaces of eachpolymeric disc ranged from 0.06e0.26 mg/mm3 of material(Table 1). Hence, the incorporated and the adsorbed depots of the

ith BMP-2 on the degradation of the polymeric material itself. The mean total volume ofith the standard deviation. *: p < 0.05; **: p < 0.01; ***: p < 0.001.

Fig. 9. Graph depicting the osteoinductive efficacies of BMP-2 in association with each polymer type. Each polymeric material was functionalized with BMP-2, which was eitheradsorbed directly onto the uncoated material or incorporated into a calcium-phosphate coating. Mean values (n ¼ 6) are represented together with the standard deviation.

G. Wu et al. / Biomaterials 31 (2010) 7485e74937492

drug that were associated with the four types of polymer lay withinan order of magnitude of each other. Albeit so, the osteoinductiveefficacy of BMP-2 varied greatly in association with the differentmaterials, although, as aforementioned, this parameter wasconsistently higher for a coating-incorporated than for a directly-adsorbed depot of the drug (Fig. 9). The difference in osteoinductiveefficacy between the two modes of BMP-2 delivery was least forPolyactive�. This finding is not surprising, since a higher volume offoreign-body giant cells was associated with Polyactive� than withany of the other three materials (Fig. 5). Such a high level ofinflammatory activity appears to suppress osteogenesis. Thenegative influence of inflammation on osteogenic activity ismediated predominantly by cells of the monocyte/macrophagelineage, which align the polymer surface and are induced by thetoxic chemical properties of the material itself or of its degradationproducts to undergo apoptosis. However, these cells escape deathby fusing to form foreign-body giant cells [51,52], which triggera chronic inflammatory response.

Although the volumes of foreign-body giant cells that wereassociated with Ethisorb� and PLGA were similar, the osteoinduc-tive efficacyof BMP-2washigher in associationwith the former thanwith the latter material (Fig. 9). This finding is probably accountedfor by the higher surface-area density of Ethisorb� (Fig. 3).

The data gleaned from the present study indicate that severalfactors contribute to the osteoinductive efficacy of BMP-2-func-tionalized polymeric scaffolds. However, the surface-area density ofthe material and its biocompatibility are important attributes.Furthermore, if the BMP-2 is incorporated into a calcium-phos-phate coating rather than adsorbed directly onto the surface of thepolymeric material, which is the mode of delivery that is currentlyadopted in clinical settings, then the characteristically slower, cell-mediated liberation of the drug is associated with a significantattenuation of the inflammatory response and a consequentaugmentation of osteogenic activity.

5. Conclusion

Bone-defect-filling polymeric materials can be renderedosteoinductive by functionalizing them with a BMP-2-bearingcalcium-phosphate coating. This mode of BMP-2 carriage is moreefficacious in inducing and sustaining bone formation at an ectopic

site in rats than is a directly-adsorbed depot of the agent. It is alsoassociated with a greater attenuation of the inflammatory responsethat is triggered by the native polymer, with the consequence thatbone-formation activity is correspondingly augmented. Further-more, during the initial 2e3 weeks of implantation that precede itsdegradation, the coating protects the host from the pro-inflam-matory effects of the products of polymer degradation, therebypermitting bone-formation activity to gain a firm foothold. Ina clinical setting, BMP-2-functionalized calcium-phosphate coat-ings would thus afford a means of curbing inflammatory reactivityand of dramatically lowering the pharmacological dose of theosteogenic agent to a safe level.

Acknowledgements

This study was supported by a grant from the Swiss NationalScience Foundation [to TI and EBH (no. 320000-107639/1)].

Appendix

Figureswith essential colour discrimination. Figs.1, 4e6 and 8 inthis article are difficult to interpret in black and white. Thefull-colour images can be found in the on-line version, at doi:10.1016/j.biomaterials.2010.06.037.

References

[1] Iizuka T, Thoren H, Annino Jr DJ, Hallikainen D, Lindqvist C. Midfacial fracturesin pediatric patients. Frequency, characteristics, and causes. Arch OtolaryngolHead Neck Surg 1995;121(12):1366e71.

[2] Iizuka T, Randell T, Guven O, Lindquist C. Maxillofacial fractures related towork accidents. J Craniomaxillofac Surg 1990;18(6):255e9.

[3] Carson JS, Bostrom MP. Synthetic bone scaffolds and fracture repair. Injury2007;38(Suppl. 1):S33e7.

[4] Banwart JC, Asher MA, Hassanein RS. Iliac crest bone graft harvest donor sitemorbidity. A statistical evaluation. Spine 1995;20(9):1055e60.

[5] Heary RF, Schlenk RP, Sacchieri TA, Barone D, Brotea C. Persistent iliac crestdonor site pain: independent outcome assessment. Neurosurgery 2002;50(3):510e6, discussion 6e7.

[6] Silber JS, Anderson DG, Daffner SD, Brislin BT, Leland JM, Hilibrand AS, et al.Donor site morbidity after anterior iliac crest bone harvest for single-levelanterior cervical discectomy and fusion. Spine 2003;28(2):134e9.

[7] Ayerza MA, Aponte-Tinao LA, Abalo E, Muscolo DL. Continuity and function ofpatellar tendon host-donor suture in tibial allograft. Clin Orthop Relat Res2006;450:33e8.

G. Wu et al. / Biomaterials 31 (2010) 7485e7493 7493

[8] Muscolo DL, Ayerza MA, Aponte-Tinao LA, Ranalletta M. Use of distal femoralosteoarticular allografts in limb salvage surgery. Surgical technique. J BoneJoint Surg Am 2006;88(Suppl. 1 Pt 2):305e21.

[9] Buck BE, Malinin TI, Brown MD. Bone transplantation and human immuno-deficiency virus. An estimate of risk of acquired immunodeficiency syndrome(AIDS). Clin Orthop Relat Res 1989;240:129e36.

[10] Moreau MF, Gallois Y, Basle MF, Chappard D. Gamma irradiation of humanbone allografts alters medullary lipids and releases toxic compounds forosteoblast-like cells. Biomaterials 2000;21(4):369e76.

[11] Lewandrowski KU, Rebmann V, Passler M, Schollmeier G, Ekkernkamp A,Grosse-Wilde H, et al. Immune response to perforated and partially demin-eralized bone allografts. J Orthop Sci 2001;6(6):545e55.

[12] Bakker NP, van Erck MG, Zurcher C, Faaber P, Lemmens A, Hazenberg M, et al.Experimental immune mediated arthritis in rhesus monkeys. A model forhuman rheumatoid arthritis? Rheumatol Int 1990;10(1):21e9.

[13] Guan L, Davies JE. Preparation and characterization of a highly macroporousbiodegradable composite tissue engineering scaffold. J Biomed Mater Res A2004;71(3):480e7.

[14] Dietz A, Ziegler CM, Dacho A, Althof F, Conradt C, Kolling G, et al. Effectiveness ofa new perforated 0.15mmpoly-p-dioxanon-foil versus titanium-dynamicmeshin reconstruction of the orbital floor. J Craniomaxillofac Surg 2001;29(2):82e8.

[15] Kontio R, Suuronen R, Salonen O, Paukku P, Konttinen YT, Lindqvist C. Effec-tiveness of operative treatment of internal orbital wall fracture with poly-dioxanone implant. Int J Oral Maxillofac Surg 2001;30(4):278e85.

[16] Merten HA, Luhr HG. Resorbable synthetics (PDS foils) for bridging extensiveorbital wall defects in an animal experiment comparison. Fortschr KieferGesichtschir 1994;39:186e90.

[17] Bergsma JE, de Bruijn WC, Rozema FR, Bos RR, Boering G. Late degradationtissue response to poly(L-lactide) bone plates and screws. Biomaterials1995;16(1):25e31.

[18] Rozema FR, Bos RR, Pennings AJ, Jansen HW. Poly(L-lactide) implants in repairof defects of the orbital floor: an animal study. J Oral Maxillofac Surg 1990;48(12):1305e9. discussion 10.

[19] Sasserath C, Van Reck J, Gitani J. The use of a polyglycolic acid membrane inthe reconstruction of the orbital floor and in loss of bone substance in themaxillofacial region]. Acta Stomatol Belg 1991;88(1):5e11.

[20] Puumanen K, Kellomaki M, Ritsila V, Bohling T, Tormala P, Waris T, et al.A novel bioabsorbable composite membrane of Polyactive 70/30 and bioactiveglass number 13e93 in repair of experimental maxillary alveolar cleft defects.J Biomed Mater Res B Appl Biomater 2005;75(1):25e33.

[21] van Dorp AG, Verhoeven MC, Koerten HK, van Blitterswijk CA, Ponec M.Bilayered biodegradable poly(ethylene glycol)/poly(butylene terephthalate)copolymer (Polyactive) as substrate for human fibroblasts and keratinocytes. JBiomed Mater Res 1999;47(3):292e300.

[22] Barbolt TA, Odin M, Leger M, Kangas L, Hoiste J, Liu SH. Biocompatibilityevaluation of dura mater substitutes in an animal model. Neurol Res 2001;23(8):813e20.

[23] Seiler RW, Mariani L. Sellar reconstruction with resorbable vicryl patches,gelatin foam, and fibrin glue in transsphenoidal surgery: a 10-year experiencewith 376 patients. J Neurosurg 2000;93(5):762e5.

[24] Blobe GC, Schiemann WP, Lodish HF. Role of transforming growth factor betain human disease. N Engl J Med 2000;342(18):1350e8.

[25] Lou J, Xu F, Merkel K, Manske P. Gene therapy: adenovirus-mediated humanbone morphogenetic protein-2 gene transfer induces mesenchymal progen-itor cell proliferation and differentiation in vitro and bone formation in vivo. JOrthop Res 1999;17(1):43e50.

[26] Liu Y, de Groot K, Hunziker EB. BMP-2 liberated from biomimetic implantcoatings induces and sustains direct ossification in an ectopic rat model. Bone2005;36(5):745e57.

[27] Liu Y, Enggist L, Kuffer AF, Buser D, Hunziker EB. The influence of BMP-2 andits mode of delivery on the osteoconductivity of implant surfaces during theearly phase of osseointegration. Biomaterials 2007;28(16):2677e86.

[28] Starr AJ. Recombinant human bone morphogenetic protein-2 for treatment ofopen tibial fractures. J Bone Joint Surg Am 2003;85-A(10):2049. authorreplies-50.

[29] ZhaoM, Zhao Z, Koh JT, Jin T, Franceschi RT. Combinatorial gene therapy for boneregeneration: cooperative interactions between adenovirus vectors expressingbone morphogenetic proteins 2, 4, and 7. J Cell Biochem 2005;95(1):1e16.

[30] Laurencin CT, Lane J. Poly (lactic acid) and poly (glycolic acid): orthopaedicsurgery applications. In: Brighton CT, Friedlaender G, Lane JM, editors. Boneformation and repair. Rosemont, Illinois: The American Academy ofOrthopaedic Surgeons; 1994. p. 325.

[31] Uludag H, D’Augusta D, Palmer R, Timony G, Wozney J. Characterization ofrhBMP-2 pharmacokinetics implanted with biomaterial carriers in the ratectopic model. J Biomed Mater Res 1999;46(2):193e202.

[32] Uludag H, Gao T, Porter TJ, Friess W, Wozney JM. Delivery systems for BMPs:factors contributing to protein retention at an application site. J Bone JointSurg Am 2001;83-A(Suppl. 1(Pt 2)):S128e35.

[33] Smith DM, Cooper GM, Mooney MP, Marra KG, Losee JE. Bone morphogeneticprotein 2 therapy for craniofacial surgery. J Craniofac Surg 2008;19(5):1244e59.

[34] Shields LB, Raque GH, Glassman SD, Campbell M, Vitaz T, Harpring J, et al.Adverse effects associated with high-dose recombinant human bonemorphogenetic protein-2 use in anterior cervical spine fusion. Spine 2006;31(5):542e7.

[35] Liu Y, Hunziker EB, Layrolle P, De Bruijn JD, De Groot K. Bone morphogeneticprotein 2 incorporated into biomimetic coatings retains its biological activity.Tissue Eng 2004;10(1e2):101e8.

[36] Liu Y, de Groot K, Hunziker EB. Cell-mediated protein release from calcium-phosphate-coated titanium implants. J Control Release 2005;101(1e3):346e7.

[37] Wu G, Liu Y, Iizuka T, Hunziker EB. Biomimetic coating of organic polymerswith a protein-functionalized layer of calcium phosphate: the surface prop-erties of the carrier influence neither the coating characteristics nor theincorporation mechanism or release kinetics of the protein. Tissue Eng Part CMethods; 2010 [Epub ahead of print].

[38] Liu Y, Layrolle P, de Bruijn J, van Blitterswijk C, de Groot K. Biomimeticcoprecipitation of calcium phosphate and bovine serum albumin on titaniumalloy. J Biomed Mater Res 2001;57(3):327e35.

[39] Gundersen HJ, Jensen EB. The efficiency of systematic sampling in stereologyand its prediction. J Microsc 1987;147(Pt 3):229e63.

[40] Schenk RK, Olah AJ, Herrmann W. Preparation of calcified tissues for lightmicroscopy. In: Dickson GR, editor. Methods of calcified tissue preparation.Amsterdam: Elsevier Science Publishers; 1984. p. 1e56. B.V.

[41] Cavalieri B. Geometria Indivisibilibus Continuorum. Bononi: Typis ClemetisFeronij1635. Reprinted as Geometria degli Indivisibili. Torino: UnioneTipografico-Editorice Torinese; 1966.

[42] Baddeley AJ, Gundersen HJ, Cruz-Orive LM. Estimation of surface area fromvertical sections. J Microsc 1986;142(Pt 3):259e76.

[43] Rodriguez A, Meyerson H, Anderson JM. Quantitative in vivo cytokine analysisat synthetic biomaterial implant sites. J Biomed Mater Res A; 2008.

[44] Ratner BD, Bryant SJ. Biomaterials: where we have been and where we aregoing. Annu Rev Biomed Eng 2004;6:41e75.

[45] Lacey DC, Simmons PJ, Graves SE, Hamilton JA. Proinflammatory cytokinesinhibit osteogenic differentiation from stem cells: implications for bone repairduring inflammation. Osteoarthr Cartil; 2008.

[46] Liu Y, Huse RO, de Groot K, Buser D, Hunziker EB. Delivery mode and efficacyof BMP-2 in association with implants. J Dent Res 2007;86(1):84e9.

[47] Sterling H, Saginario C, Vignery A. CD44 occupancy prevents macrophagemultinucleation. J Cell Biol 1998;143(3):837e47.

[48] Tsai AT, Rice J, Scatena M, Liaw L, Ratner BD, Giachelli CM. The role ofosteopontin in foreign body giant cell formation. Biomaterials 2005;26(29):5835e43.

[49] Ohtaki H, Ylostalo JH, Foraker JE, Robinson AP, Reger RL, Shioda S, et al. Stem/progenitor cells from bone marrow decrease neuronal death in globalischemia by modulation of inflammatory/immune responses. Proc Natl AcadSci U S A 2008;105(38):14638e43.

[50] Liu Y, Hunziker EB, Van de Valk C, de Groot K. Biomimetic coatings vs.collagens as a carrier material for BMP-2: an in-vivo comparision of osteo-genic responses using an ectopic rat model. Bioceramics; 2003:619e22.

[51] Brodbeck WG, Patel J, Voskerician G, Christenson E, Shive MS, Nakayama Y,et al. Biomaterial adherent macrophage apoptosis is increased by hydrophilicand anionic substrates in vivo. Proc Natl Acad Sci U S A 2002;99(16):10287e92.

[52] Brodbeck WG, Shive MS, Colton E, Nakayama Y, Matsuda T, Anderson JM.Influence of biomaterial surface chemistry on the apoptosis of adherent cells.J Biomed Mater Res 2001;55(4):661e8.