Acta Biomaterialia 8 (2012) 3777–3783

Contents lists available at SciVerse ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /actabiomat

In vivo osteointegration of three-dimensional crosslinked gelatin-coatedhydroxyapatite foams

J. Gil-Albarova a, M. Vila b,c, J. Badiola-Vargas a, S. Sánchez-Salcedo b,c, A. Herrera a, M. Vallet-Regi b,c,⇑a Orthopaedics Department, Miguel Servet University Hospital, Faculty of Medicine, Universidad de Zaragoza, Zaragoza, Spainb Inorganic and BioInorganic Chemistry Department, Universidad Complutense de Madrid, Plaza de Ramón y Cajal s/n, 28040 Madrid, Spainc Networking Research Center on Bioengineering, Biomaterials and Nanomedicine, 50018 Zaragoza, Spain

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

Article history:Received 6 April 2012Received in revised form 4 June 2012Accepted 13 June 2012Available online 20 June 2012

Keywords:Sol–gel processesOsteointegrationApatiteElastic adaptationMacroporous bioceramic scaffolds

1742-7061/$ - see front matter � 2012 Acta Materialhttp://dx.doi.org/10.1016/j.actbio.2012.06.019

⇑ Corresponding author at: Inorganic and BioInorgUniversidad Complutense de Madrid, Plaza de RamóSpain. Tel.: +34 913941861.

E-mail address: vallet@farm.ucm.es (M. Vallet-Reg

The main requirement of bone regenerative scaffolds is to enhance the chemical reactions leading to theformation of new bone while providing a proper surface for tissue in-growth as well as a suitable degra-dation rate. Calcium phosphate ceramics are conformed by different shaping methods. One requirementis to design implants and scaffolds with suitable shapes and sizes, but also with interconnected porosityto ensure bone oxygenation and angiogenesis. In this work we present the in vivo performance of hier-archically arranged glutaraldehyde crosslinked, gelatin-coated nanocrystalline hydroxyapatite (HABP)scaffolds (1–400 lm), with high potential as bone regenerators and excellent osteointegration perfor-mance, as well as an appropriate bioresorption rate. 6 � 10 mm bone defects were made in the lateralaspect of both distal femoral epiphysis of 15 mature (9 months old) male New Zealand rabbits. The bonedefect in the left femur was then filled by using HABP foam cylinders, allowing the surgeon to carve theappropriate shape for a particular bone defect with high stability intra-operatively. The foam becomesswollen with body fluid and fills the cavity, ensuring good fixation without the need for a cement. His-tological and radiographical studies after 4 months implantation showed healing of all treated bonedefects, with bone integration of the HABP foam cylinders and bone conduction over the surface. Thisin vivo behaviour offers promising results as a scaffold for clinical applications, mainly in orthopaedicsand dentistry.

� 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Calcium phosphates are widely used as biocompatible bone fill-ers and regenerators in tissue engineering, as they are very similarin hardness to natural bone tissue. These types of ceramics are con-formed by different synthesis and shaping methods. The mainrequirement of these scaffolds is to be precursors of newly formedbone as well as to ensure bone oxygenation and angiogenesis [1].For this purpose the designed materials must have interconnectedporosity which must also include a number of macropores [2]. Re-ported pore sizes supporting the osteointegration of implants havebeen in the range 1–500 lm for a variety of biomaterials, such ashydroxyapatite (HA), b-calcium phosphate and biphasic calciumphosphate [3,4]. It is also necessary to find a compromise betweena suitable degradation rate and a robust scaffold inducing tissueregeneration. These materials have attracted wide interest because

ia Inc. Published by Elsevier Ltd. A

anic Chemistry Department,n y Cajal s/n, 28040 Madrid,

i).

of their excellent biocompatibility and bioactivity, having a posi-tive influence on the biological responses of cells.

Some of the modern approaches to improving the performanceof metallic, polymeric and ceramic scaffolds have been to attachgrowth factors to the scaffold surface [5–7] or to seed them withosteoprogenitor cells to facilitate osteoblast colonization and in-duce cell penetration over the three-dimensional volume [8,9].Moreover, studies on the influence of the degree of porosity onscaffold resorption have shown it to be a key parameter in the bal-ance of all parameters involved in bone reconstruction [10]. Never-theless, it is expected that the simpler the scaffolds design, thesimpler would be the manipulation during the surgical procedure.Because of this, the synthesis of hierarchical porous scaffolds hasbeen under examination for several decades and remains so today,with the perfect combination of all requirements still underdevelopment.

Three-dimensional (3-D) hydroxyapatite (HA) foams with ahigh degree of porosity were prepared by sol–gel routes, and in or-der to confer flexibility and handleability the foams were coatedwith gelatin–glutaraldehyde biopolymers. These foams have previ-ously been proposed by our group as new potential devices for thetreatment of heavy metal intoxication by ingestion and as water

ll rights reserved.

3778 J. Gil-Albarova et al. / Acta Biomaterialia 8 (2012) 3777–3783

purifiers [11,12]. However, both the composition and 3-D intercon-nected architectural design of these HA foam-like systems suggestthat they would allow cellular internalization and colonizationover the entire surface. Thus an exhaustive study of the in vitro re-sponse of osteoblast-like cells and the in vitro degradation processwas presented in a previous work, exhibiting excellent in vitro per-formance and a suitable degradation rate with non-cytotoxic deg-radation products [13].

The design of these scaffolds allows rapid bone in-growth to-gether with fast ceramic resorption as well as easy manipulationduring surgery.

Herein, for the first time, the potential of these HABP foams asbone regenerators in vivo has been studied in 15 mature New Zea-land rabbits. For this purpose radiological and histological studieshave been performed to asses the effects of the scaffolds onosteointegration.

2. Experimental

2.1. Scaffold fabrication and characterization

3-D HA foams have been synthesized and conformed in a one-step process following a sol–gel technique [11], including thenon-ionic surfactant Pluronic F127 (EO106PO70EO106) as a macro-pore inducer using an accelerated evaporation-induced self-assem-bly (EISA) method [12]. HA was obtained from the reaction of cal-cium nitrate tetrahydrate and triethylphosphite (TIP) (Aldrich,Steinheim, Germany) at a molar ratio of F127:TIP of 11. The result-ing foams were coated by immersion in a solution of a biocompat-ible polymer in use in US Food and Drug Administration (FDA)approved products (1.2% w/v type A gelatine (porcine skin) cross-linked with 0.05% w/v glutaraldehyde) as reported elsewhere[11]. Scanning electron microscopy (SEM) in a JEOL 6400 micro-scope (Tokyo, Japan) was used to characterize the macroporous3-D architecture of the biopolymer-coated HA foams (HABP). Hgporosimetry measurements were carried out in an AutoPore IIIporosimeter (Micromeritics Instrument Corp., Norcross, GA) andX-ray diffraction (XRD) in a Philips X’Pert diffractometer using CuKa radiation. Further details of the foam characterization have beenreported previously [11,14].

2.2. Surgical procedures

15 male mature New Zealand rabbits, 9 months old and weigh-ing on average 3.751 g (3.231–4.450 g) were used. The animalswere placed in individual cages, under standard conditions (roomtemperature 20 ± 0.5 �C, relative humidity 55 ± 5% and illumina-tion with a 12 h/12 h light/dark photoperiod), fed with full rabbitspecial fodder (Nantas�), given water ad libitum, and were withoutrestriction of movement. Surgery was performed under asepticconditions and general anaesthesia induced by injection of keta-mine (75 mg kg�1), supported by an inhalation mask of O2 and iso-flurane (2.5 and 0.8–1.5 l min�1, respectively). Analgesia wasmaintained by subcutaneous injection of buprenorfine (0.001–0.05 mg kg�1), and antibiotic prophylaxis was by means of twoinjections of cefazoline (50 mg kg�1). Using a motorized drill abone defect 6 mm in diameter and 10 mm in depth was made inthe lateral aspect of both distal femoral epiphyses in all animals,using continuous irrigation with physiological saline to preventbone necrosis. The bone defect in the left femur was then filledusing HABP foam cylinders previously carved into the requiredshape with a scalpel and sterilized with ethylene oxide. The bonedefects in the right femora served as controls. The designed bonedefects corresponded to critical bone defects according to the stan-dard requirements [15–19]. All the animals were killed pharmaco-

logically under general anaesthesia after 4 months. The 15 rabbitsformed one group and were numbered 1–15.

2.3. Radiographical and histological evaluation

During follow-up standardized lateral digital radiographs weretaken monthly (high ionization Sedecal tube, 400 mA, 500 mAs,150 kV, indirect digital radiography, Fujifilm digital reader, FCRPrima).

After death the femora were extracted for histological study.Three 3 mm thick longitudinal sections from both femurs were ob-tained from each animal and placed in a buffered paraformalde-hyde (10%) solution for fixation by immersion for 15 days. Afterfixation the samples were washed and decalcified in 5% nitric acidsolution for 10 days.

For the histological study the samples were dehydrated througha graded ethanol series (70, 96 and 100%) and cleared with xylene,before being embedded in paraffin using a Leica TP 1050 tissueprocessor. The paraffin blocks were cut into 4 lm sections with aShandon Finesse 325 microtome, deparaffinized and rehydrated.The sections were stained with hematoxylin and eosin (H&E) andMasson–Goldner Trichrome for histological evaluation.

Histological sections were examined with a Zeiss AxiokoP 40microscope and microphotographs were obtained with a Zeiss Axi-oCam MRc5 camera.

All procedures were carried out under Project Licence PI45/10approved by the in-house Ethics Committee for Animal Experi-ments of the University of Zaragoza. The care and use of animalswere performed according to the Spanish Policy for Animal Protec-tion RD1201/05, which meets the European Union Directive 86/609 on the protection of animals used for experimental and otherscientific purposes.

3. Results

3.1. HABP foam characterization

HABP foams have a controlled distribution of interconnectedmacroporosity in the range 1–400 lm, determined by Hg intrusionporosimetry and observed by SEM in Fig. 1a and c. Low tempera-ture sol–gel synthesis allows crystal size control and results in ananocrystalline structure, as confirmed by the XRD pattern corre-sponding to pure nanocrystalline HA (ICDD PDF 9-432). The aver-age crystallite size was calculated based on the 100 and 001reflections by Rietveld refinement [20] giving a value of 20 nm.As shown in Fig. 1c, the Hg intrusion porosity technique performedbefore and after biopolymer coating (an SEM image of the coatedHABP foam is shown in the inset, with a pore volume value of70%) shows only a slight decrease in the total pore volume distri-bution in the range 100–300 lm [11]. The amount of crosslinkedgelatin on the coated HA foams was determined by thermogravi-metric analyses of three replicates, showing a biopolymer contentof 20 ± 2% (see Supplementary Fig. S1). As has been reported previ-ously, crosslinked gelatin-coated foams have a swelling ratio of400W% when immersed in aqueous solution due to the hydrophilicnature of gelatine crosslinked with glutaraldehyde [11] (see Fig. 1dand Supplementary Fig. S2). Swelling was calculated as W (swell-ing) (%) = 100 � (Wt �Wd)/Wd, where Wd is the weight of driedfoam and Wt is the weight of hydrated foams at time t (0–24 h).

3.2. Intra-operative findings

The carved HABP foam cylinders were easily managed duringsurgery and remained stable in the bone defect. After surgicalimplantation into the bone defect the carved HABP foam cylinders

Fig. 1. (a) SEM micrograph of uncoated HA foam (250�). (b) XRD experimental patterns and the calculated pattern and difference plot obtained by Rietveld refinement.Experimental (symbols) and calculated (solid line) powder X-ray diffraction patterns for HA foam. The lower trace is the difference between the observed and calculatedpatterns. The vertical lines mark the position of the calculated Bragg peaks for the apatite phase. (c) Pore size distribution and cumulative pore volume determined by Hgintrusion of the uncoated and coated HA. (Inset) SEM image of the coated HABP foam. (d) Digital micrograph of the hydrated foam.

Fig. 2. Digital images of the implantation procedure.

J. Gil-Albarova et al. / Acta Biomaterialia 8 (2012) 3777–3783 3779

were progressively infiltrated by blood from the surrounding bone,increasing the intra-operative mechanical stability by means ofelastic adaptation of its shape and contour to the bone defect wallsand limits (Fig. 2). Their swelling ability allows them to perfectlyfill the bone defect cavity when hydrated, leading to good fixationinside the bone wall.

3.3. Post-operative incidence

Two rabbits were excluded due to premature spontaneousdeath within the second post-operative month due to haemorrhag-ic gastroenteritis, as shown on autopsy. Another rabbit suffered aright supracondylar femoral fracture during first post-operative

3780 J. Gil-Albarova et al. / Acta Biomaterialia 8 (2012) 3777–3783

month, but this was consolidates and the rabbit may complete theplanned follow-up. All others survived the intervention and con-cluded the study uneventfully.

3.4. Radiographic analysis

No HABP foam cylinder migration or instability was observed inthe radiographic study. Images of progressive osteointegration,with bone growth and remodelling on the HABP foam cylinder inthe epiphyseal bone defect were observed during the whole post-operative period. The significant enhancement in the radiographicimage of the HABP foam cylinder during the first post-operativemonth, probably due to its high bioactivity, is striking. However,the central zone of the foam showed a lower radiographic density,perhaps an expression of a centripetal bone growth limit over thefoam. This central area of lower density declined in the secondmonth and almost disappeared in the third. The limits of the cre-ated bone defect and contour of the HABP foam cylinder were moreevident during the first two months of follow-up, with a higherradiographic density than the recipient bone, showing progressiveblurring in the third and fourth months. Fact suggests progressivecentripetal resorption of the HABP foam cylinder during this periodof follow-up. This resorption of the HABP foam would be accompa-nied by a simultaneous process of bone growth in the macroporousstructure of the HABP foam, in the same centripetal direction, as acreeping substitution, explaining these radiographic images. Fur-thermore, and supporting this observation, there was a blurringof the radiographic density of the foam in the last two months ofthe study, suggesting that bone healing in the foam was almostcomplete due to its porosity and interconnectivity. The radiologicalstudy suggests that bone healing did not follow a conventional pat-tern of callus formation, but a simultaneous process of HABP foamresorption and replacement by new bone (Fig. 3).

As shown in Fig. 4, the limits of the control bone defect, withoutHABP foam, could be clearly seen during the first two post-opera-tive months, showing progressive effacement in the last twomonths of follow-up. However, the cylindrical bone cavity had a

Fig. 3. Digital lateral radiographs of treated bone defect. Rabbit 9. I.P., immediately post-post-operative; 4m, 4 months post-operative.

Fig. 4. Digital lateral radiographs of a control bone defect. Rabbit 9. I.P., immediately3 months post-operative; 4m, 4 months post-operative.

lower bone density than the surrounding bone throughout thepost-operative period, suggesting partial healing of the lateral fem-oral bone surface, but insufficient bone repair within the bonedefect.

3.5. Histological analysis

No signs of inflammation or adverse tissue reaction were ob-served around implants during necropsy. Treated bone defectsshowed a satisfactory healing of the cavity, but complete repairof the bone defect was never observed in the control group, asjudged by the critical bone defect concept [15,17–19]. Control bonedefects showed only partial healing of the lateral femoral bonesurface.

The histological study showed suitable healing of all treatedbone defects, with bone integration of the HABP foams cylindersand osteoconduction over its surface (Fig. 5). Most of the samplesshowed varying degrees of new bone formation, located in the leftfemoral epiphysis of the rabbits. This new bone formation con-sisted of the formation of bone trabeculae of different morphology,usually arranged around well-differentiated, triangular shaped for-eign material, with a structure different to normal bone tissue. Thenewly formed bone trabeculae varied in size and shape and werepartially confluent, adopting a typical reticular morphology(Fig. 6a and b). The histological structure of these trabeculae corre-sponds to young bone tissue, with osteoblasts lining the peripheryand some osteoclasts. The repair and new bone formation observedin all treated bone defects were the result of angiogenesis on theporous surface of the scaffold, which served as a guide for bone for-mation in a centripetal direction.

Most of the left femoral epiphysis examined showed scatteredand sparse areas of foreign body reaction, well delimited, withmultinucleated giant cells frequently containing the previously de-scribed foreign material within the cytoplasm (Fig. 7). A fibrousreaction and membranous ossification on residual fragments ofthe scaffold were commonly associated with these lesions(Fig. 6b). The presence of bone marrow was observed around the

operative; 1m, 1 month post-operative; 2m, 2 months post-operative; 3m, 3 months

post-operative; 1m, 1 month post-operative; 2m, 2 months post-operative; 3m,

Fig. 5. Left femur. Rabbit 9. Macroscopic appearance of the bone specimen afternecropsy and low magnification image of the histological staining, with blackarrows indicating bone integration of the HABP foam cylinders.

Fig. 6. Left femur. Rabbit 9. (a) New bone formation (white star) is observed,forming trabeculae around the foreign material (black arrow). Black asterisk, bonemarrow. Masson’s Trichrome. (b) Membranous ossification process (wavy arrow).Black arrow, foreign material; white star, new bone formation. Masson’s Trichrome.

J. Gil-Albarova et al. / Acta Biomaterialia 8 (2012) 3777–3783 3781

newly formed bone trabeculae, mostly formed by adipose tissue,blood vessels and haematopoietic tissue (Fig. 6).

The control samples of the right femoral epiphysis examineddid not show new bone formation such as observed in the left fem-oral epiphysis, and only the presence of adipose tissue and bloodvessels were observed, with very little evidence of bone healinginto the cavity, according to the expectations of critical bone de-fects (Fig. 8).

Fig. 7. Left femur. Rabbit 14. Foreign body granuloma (black snowflake) close tonew bone formation (white star). Black asterisk, bone marrow. Masson’s Trichrome.

4. Discussion

In this study the HABP foam tested allowed healing of epiphy-seal limited bone defects in rabbits. The results indicate that therewere considerable differences in healing potential between thecontrol group and the group that received carved cylinders of HABPfoam, showing adequate viability of the recipient bone. The HABPfoam tested showed evidence of osteointegration, bone growthover its surfaces and subsequent remodelling. Bone formation overthe HABP foam cylinders surfaces was predominantly of membra-nous type, with scarce fibrous tissue, and without evidence ofchondral cells. This suggests that it performed as an adequate scaf-fold according to Shors criteria [21] for bone growth over a bioac-tive implant under ideal conditions (stability, viability andproximity), by means of a process of mesenchymal cell recruitmentfrom the surrounding tissues and subsequent transformation intobone-forming cells. This observation has been reported by the

authors in previous experimental studies in rabbits with sol–gelglass and glass ceramic implants in the treatment of limited [22]and critical bone defects [23]. Moreover, the observed well-delim-ited foreign body reaction, with multinucleated giant cells contain-ing the material inside the cytoplasm, may be explained by

Fig. 8. Right femur. Rabbit 1. The bone defect has been filled with adipose tissue.H&E.

3782 J. Gil-Albarova et al. / Acta Biomaterialia 8 (2012) 3777–3783

incomplete resorption of the HABP ceramic scaffold during the fourmonths of the in vivo study, suggesting that it may disappear aftera longer follow-up than that used in our research.

As has been described, the ideal bone void filler should be oste-ogenic, biocompatible, bioresorbable, able to give structural sup-port, able to serve as carrier, provide a 3-D matrix to supportosteoblasts and precursor cells and, ultimately, support bone in-growth or on-growth during resorption and healing [21–24]. Therequirements range from an ability of the scaffold to providemechanical support during tissue growth and gradually degradeto biocompatible products to more demanding requirements, suchas an ability to incorporate cells or growth factors, providing osteo-conductive and osteoinductive environments [25–27]. In thissense, the mixture of HA and gelatine in the HABP foams allowsa combination of the biocompatibility and bioactivity of the bioce-ramic matrix with the flexibility, wettability, excellent resorbabil-ity and biological functional groups of the organic coating.Moreover, the nanocrystalline structure of HA, similar to the crys-tal size of natural bone HA, enhances bone cell differentiation [28].These foams are intended to be used in low mechanical require-ment applications, such as cavity filling or surface coverage, or incombination with firm synthetic bone to maintain bone functionuntil healing is complete if they are subjected to intense mechan-ical loads. The distribution of endochondral ossification and intra-membranous ossification depends upon the biological andbiomechanical environment, and load bearing is important in thehealing bone regaining its original mechanical strength, becausemechanical loading promotes an osteogenic response [29,30,25].

The poorer bone healing observed in the control bone defectscompared with the treated bone defects supports the hypothesisthat the HABP foam provides the optimal conditions of intercon-nected porosity, surface chemistry and suitable resorption ratefor bone repair in our experimental model. Moreover, the 1–400 lm porosity of the scaffold, as suggested by the in vitro models[14], allowed 3-D colonization of the defect, avoiding calcificationof the material. This experimental critical bone defect design hasbeen used successfully by others authors [16,22].

The observed bone growth and bone remodelling over bone de-fects filled with the HABP foam may be difficult to explain by aninert implant. The epiphyseal and metaphyseal location of theHABP foam cylinders imply a continuous process of bone remodel-ling which showed an intimate union to bone. The designed bonedefect in our study corresponded to a critical bone defect [15–19]. The experimental comparison showed significant differencesin bone healing and new bone formation between the treated

and control bone defects. Resorption of the HABP foam and bonegrowth in a centripetal direction, as a true creeping substitution,in the treated bone defects may explain the pattern of bone healingand remodelling, which was almost complete after a 4 month fol-low-up period. This fact has been observed by others authors insimilar studies in rabbits with a calcium phosphate bone cement[16] and with a b-tricalcium phosphate [24]. The limited intercon-nectivity between adjacent pores has been reported as a cause ofincomplete bone tissue penetration in the central part of the im-plant [10,16,24]. Similar findings have been reported in goats witha porous bioglass [31]. Moreover, in previous experimental studiesin rabbits with sol–gel glass and glass ceramic implants in thetreatment of limited [22] and critical bone defects [23] we ob-served peripheral surface dissolution and fragmentation of the im-plants, with bone growth over the surfaces as evidence of excellentosteointegration, but without bone in-growth due to a lack ofporosity in the implants tested in those studies.

The poorest bone healing was observed in the control bone de-fects, with partial bone formation on the lateral bone surface. Onlyadipose tissue and blood vessels were present in the control bonecavities at the end of the follow-up period. This experimental bonedefect design has been used successfully by other authors [16,22].

Moreover, according to their biofunction and biological perfor-mance, the in vivo behaviour of the HABP foam tested matches thedefinition of a third generation material for orthopaedic clinicalusage [32,33]. This means the biopolymer-coated HABP foam, inthe form of an implant with three-dimensional structures, is biode-gradable and biocompatible with non-cytotoxic degradation prod-ucts. This material was designed to stimulate specific cellularresponses at the molecular level while degrading at the same rateas the tissue is repaired.

An ideal scaffold should allow the integration of osteoblasts andosteoclasts by means of a highly interconnected porous network,with 100–400 lm pores being optimal for bone progenitor cells[29,34] and 1–20 lm pores responsible for cellular developmentand fluid exchange [3]. The fabrication and characterization ofthe 3-D HABP foam tested in our study met these requirementsclosely, not only performing satisfactorily in terms of 3-D cell col-onization in the in vitro studies [14], but also supporting osteoin-tegration in vivo.

Osteoprogenitor cells are necessary to replace the inserted scaf-fold and to create new bone tissue. These cells can arise from theperiosteum at the fracture site or by chemotaxis and blood vesselsentering the haematoma [32]. In this experimental model theHABP foam cylinders were progressively infiltrated by blood fromthe surrounding healthy bone, allowing cell colonization. In addi-tion, this involved an intra-operative increase in its stability byelastic adaptation of its shape and contours to the recipient bonedefect. Afterwards specific mechanical and biological stimulantscause the cells to differentiate into osteoblasts [32]. This environ-ment promotes the optimal conditions for bone growth over a bio-active scaffold, by a process of cell recruitment from thesurrounding bone and subsequent progressive osteointegration,after blood vessels and fibroblasts proliferated in the fibrin clot.Although bone repair was successful, remodelling in the treatedbone defects was not complete after 4 months follow-up. A longerpost-operative follow-up period could lead to complete boneremodelling.

5. Conclusions

The in vivo behaviour of this macroporous HABP scaffold wasshown to promote the optimal conditions for bone growth overits surface by a process of cell recruitment from the surroundingbone and subsequent progressive osteointegration. The HABP foam

J. Gil-Albarova et al. / Acta Biomaterialia 8 (2012) 3777–3783 3783

was highly permeable to body fluids, accommodating its shape tothe limits of the bone defect by means of elastic adaptation of itscontours to the recipient bone defect walls, while increasing its ini-tial mechanical stability.

Together with the promising in vivo results presented above,the synthesis and surgical management of this HABP foam is quitesimple, allowing the surgeon to pre-operatively approximatelycarve the appropriate scaffold shape, according to the clinicalrequirements and pre-operative measurements of the bone defect.

Our experimental model allowed evaluation of the HABP foamin a critical stable bone defect, with adequate viability of the reci-pient bone and without protective synthetic bone. Our results indi-cate biocompatibility of the HABP foam. The observed bone in-growth and bone remodelling over the bone defect filled withcrosslinked gelatin-coated HABP foam offer promise as an scaffoldfor low load bearing clinical applications, mainly in orthopaedicsand dentistry.

Acknowledgements

This work was financially supported by the Spanish CICYTthrough project MAT-2008-00736, Spanish National CAM projectS2009/MAT-172 and the Network of Excellence CSO2010-11384-E. M.V. would like to thank the Spanish Ministry for an RyC grantand the FSE.

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figs. 1, 2, and 5–8, aredifficult to interpret in black and white. The full colour images canbe found in the on-line version, at http://dx.doi.org/10.1016/j.actbio.2012.06.019.

Appendix B. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.actbio.2012.06.019.

References

[1] Chen QZ, Rezwan K, Françon V, Armitage D, Nazhat SN, Jones FH, et al. Surfacefunctionalization of bioglass-derived porous scaffolds. Acta Biomater2007;3:551–62.

[2] Hollister SJ. Porous scaffold design for tissue engineering. Nat Mater2005;4:518–24.

[3] Sánchez-Salcedo S, Nieto A, Vallet-Regí M. Hydroxyapatite/b-tricalciumphosphate/agarose macroporous scaffolds for bone tissue engineering. ChemEng J 2008;137:62–71.

[4] Levengood SKL, Polak SJ, Wheeler MB, Maki AJ, Clark SG, Jamison RD, et al.Multiscale osteointegration as a new paradigm for the design of calciumphosphate scaffolds for bone regeneration. Biomaterials 2010;31:3552–63.

[5] Lozano D, Manzano M, Doadrio JC, Salinas AJ, Vallet-Regí M, Gómez-Barrena E,et al. Osteostatin-loaded bioceramics stimulate osteoblastic growth anddifferentiation. Acta Biomater 2010;6:797–803.

[6] Oh SH, Kim TH, Lee JH. Creating growth factor gradients in three dimensionalporous matrix by centrifugation and surface immobilization. Biomaterials2011;32:8254–60.

[7] Odedra D, Chiu LLY, Shoichet M, Radisic M. Endothelial cells guided byimmobilized gradients of vascular endothelial growth factor on porouscollagen scaffolds. Acta Biomater 2011;7:3027–35.

[8] Torigoe I, Sotome S, Tsuchiya A, Yoshii T, Takahashi M, Kawabata S, et al. Novelcell seeding system into a porous scaffold using a modified low-pressuremethod to enhance cell seeding efficiency and bone formation. Cell Transplant2007;16(7):729–39.

[9] Godbey WT, Stacey Hindy BS, Sherman ME, Atala A. A novel use of centrifugalforce for cell seeding into porous scaffolds. Biomaterials 2004;25:2799–805.

[10] Doernber MC, Rechenberg B, Bohner M, Grunenfelder S, Lenthe GH, Müller R,et al. In vivo behaviour of calcium phosphate scaffolds with four different poresizes. Biomaterials 2006;27:5186–98.

[11] Sanchez-Salcedo S, Vila M, Izquierdo-Barba I, Cicuéndez M, Vallet-Regí M.Biopolymer-coated hydroxyapatite foams: a new antidote for heavy metalintoxication. J Mater Chem 2010;20:6956–61.

[12] Brinker CJ, Lu YF, Sellinger A, Fan HY. Evaporation-induced self-assembly:nanostructures made easy. Adv Mater 1999;11:579–85.

[13] Vila M, Sanchez-Salcedo S, Cicuéndez M, Izquierdo-Barba I, Vallet-Regí M.Novel biopolymers-coated hydroxyapatite foams form removing heavy metalsfrom polluted water. J Hazard Mater 2011;192:71–7.

[14] Cicuéndez M, Izquierdo-Barba I, Sánchez-Salcedo S, Vila M, Vallet-Regí M.Biological performance of hydroxyapatite–biopolymer foams: in vitro cellresponse. Acta Biomater 2012;8:802–10.

[15] Lasanios NG, Kanakaris NK, Giannoudis PV. Current management of long bonelarge segmental defects. J Orthop Trauma 2009;24:149–63.

[16] Miño-Fariña N, Muñoz-Guzón F, López-Peña M, Ginebra MP, del Valle-FresnoS, Ayala D, et al. Quantitative analysis of the resorption and osteoconduction ofa macroporous calcium phosphate bone cement for the repair of a critical sizedefect in the femoral condyle. Vet J 2009;179:264–72.

[17] Reichert JC, Saifzadeh S, Wullschleger ME, Epari DR, Schütz MA, Duda GN, et al.The challenge of establishing preclinical models for segmental bone defectresearch. Biomaterials 2009;30:2149–63.

[18] Caccioli A, Spaggiari B, Ravanetti F, Martini FM, Borghetti P, Gabbi C. Thecritical size bone defect: morphological study of bone healing. Ann Fac MedicVet Parma 2006;26:97–100.

[19] Cugala Z, Gogolewski S. Regeneration of segmental diaphyseal defects in sheeptibiae using resorbable polymeric membranes: a preliminary study. J OrthopTrauma 1999;13:187–95.

[20] Rodríguez-Carvajal J. Recent advances in magnetic structure determination byneutron powder diffraction. Physica B 1993;192:55–69.

[21] Shors EC. Coraline bone graft substitutes. Orthop Clin N Am 1999;30:599–613.[22] Gil-Albarova J, Garrido-Lahiguera R, Salinas AJ, Román J, Bueno-Lozano AL, Gil-

Albarova R, et al. The in vivo performance of a sol–gel glass and a glass-ceramicin the treatment of limited bone defects. Biomaterials 2004;24:4639–45.

[23] Gil-Albarova J, Salinas AJ, Bueno-Lozano AL, Román J, Aldini-Nicolo N, García-Barea A, et al. The in vivo behaviour of a sol–gel glass and a glass-ceramicduring critical diaphyseal bone defects healing. Biomaterials2005;26:4374–82.

[24] Walsh WR, Vizesia F, Michael D, Auld J, Langdown A, Oliver R, et al. b-TCP bonegraft substitutes in a bilateral rabbit tibial defect model. Biomaterials2008;29:266–71.

[25] Lane JM, Tomin E, Bostrom MPG. Biosynthetic bone grafting. Clin Orthop1999;367S:107–17.

[26] Rikli DA, Regazzoni P, Perren SM. Is there a need for resorbable implants orbone substitutes? Injury 2002;33:2–3.

[27] Gunatillake PA, Adhikari R. Biodegradable synthetic polymers for tissueengineering. Eur Cells Mater 2003;5:1–16.

[28] Nejati E, Firouzdor V, Eslaminejad MB, Bagheri F. Needle-like nanohydroxyapatite/poly(L-lactide acid) composite scaffold for bone tissueengineering application. Mater Sci Eng C 2009;293:942–9.

[29] Chao EYS, Inoue N. Biophysical stimulation of bone fracture repair,regeneration and remodelling. Eur Cells Mater 2003;6:72–85.

[30] Wolff J. The Law of Bone Remodelling. Berlin: Springer; 1986.[31] Nandi SK, Kundu B, Datta S, De DK, Basu D. The repair of segmental bone

defects with porous bioglass: an experimental study in goats. Res Vet Sci2009;86:162–73.

[32] Vallet-Regí M. Evolution of bioceramics within the field of biomaterials. CRChim 2010;13:174–85.

[33] Navarro M, Michiardi A, Castano O, Planell JA. Biomaterials in orthopaedics. J RSoc Interf 2008;5:1137–58.

[34] Schoreder JE, Mosheiff R. Tissue engineering approaches for bone repair:concepts and evidence. Injury 2011;42:609–13.