Nanoparticulate bioactive-glass-reinforced gellan-gum hydrogels for bone-tissue engineering

Materials Science and Engineering C 43 (2014) 27–36

Contents lists available at ScienceDirect

Materials Science and Engineering C

j ourna l homepage: www.e lsev ie r .com/ locate /msec

Nanoparticulate bioactive-glass-reinforced gellan-gum hydrogels forbone-tissue engineering

Ana Gantar a,b, Lucilia P. da Silva c,d, Joaquim M. Oliveira c,d, Alexandra P. Marques c,d, Vitor M. Correlo c,d,Saša Novak a,b,⁎, Rui L. Reis c,d

a Department for Nanostructured Materials, Jožef Stefan Institute, Sloveniab Jožef Stefan International Postgraduate School, Sloveniac 3B's Research Group, Department for Polymer Engineering, University of Minho, Portugald ICVS/3B's, PT Government Associate Laboratory, Braga, Guimarães, Portugal

⁎ Corresponding author at: Jožef Stefan Institute, DepJamova c. 39, SI-1000 Ljubljana, Slovenia. Tel.: +386 1 47

E-mail address: [email protected] (S. Novak).

http://dx.doi.org/10.1016/j.msec.2014.06.0450928-4931/© 2014 Published by Elsevier B.V.

a b s t r a c t

Article history:Received 20 March 2014Received in revised form 22 May 2014Accepted 30 June 2014Available online 9 July 2014

Keywords:ScaffoldGellan-gumBioactive-glassCompositeBone-tissue engineering

This work presents bioactive-glass-reinforced gellan-gum spongy-like hydrogels (GG-BAG) as novel hydrophilicmaterials for use as the scaffolding in bone-tissue engineering. The reinforcement with bioactive-glass particlesresulted in an improvement to the microstructure and to the mechanical properties of the material. Thesemechanical properties were found to be dependent on the composition and improved with the amount ofbioactive glass; however, values necessary to accommodate biomechanical loading were not achieved in thisstudy. Nevertheless, by incorporating the bioactive-glass particles, the composite material acquired the abilityto form an apatite layer when soaked in simulated body fluid. Furthermore, human-adipose-derived stem cellswere able to adhere and spread within the gellan-gum, spongy-like hydrogels reinforced with the bioactiveglass, and remain viable, which is an important result when considering their use in bone-tissue engineering.Thus, hydrogels based on gellan gum and bioactive glass are promising biomaterials for use either alone orwith cells, and with the potential for use in osteogenic differentiation.

t. for Nanostructured Materials7 3271.

© 2014 Published by Elsevier B.V.

1. Introduction

Three-dimensional scaffolds, in combination with bioactivesubstances and cells, have been widely accepted as the basis for moderntissue-regeneration strategies and extensively used in hard- as well assoft-tissue engineering applications [1]. These scaffolds are usuallyfabricated from bioresorbable and, preferably, bioactive materials, whichhave the potential to support and stimulate the regeneration of livingtissue. When designing a scaffold, several key characteristics have tobe taken into account, e.g., biocompatibility, biodegradability, mechanicalproperties and scaffold architecture [2]. Depending on the type ofmaterial and the fabrication techniques, these characteristics may varysignificantly [3].

Scaffold materials are typically divided into three main groups:inorganic, organic, and composites. Inorganic scaffolds, such as hydroxy-apatite (HAp), tri-calcium phosphate (TCP) and bioactive glass (BAG),have been primarily developed for hard-tissue regeneration [4,5].

Although HAp and TCP are quite popular, because they have achemical composition that is similar to bone mineral, their unfavourablemechanical properties and an inflammatory response after long-termexposure [6,7] indicate the need to develop better materials. Bioactiveglasses, as a group of fully resorbable amorphous materials, and withthe potential to increase solubilitywhile retaining the unique characteris-tics of inorganic materials, have lately been attracting a great deal ofattention. As an alternative to HAp, BAG induces HAp precipitation inthe presence of a biologicalfluid, resulting in the enhancedmineralizationof bone tissue [8,9]. Moreover, due to its specific chemical composition itis also osteoinductive. In addition, BAG possesses the ability to enhancethe proliferation and differentiation of osteoblasts, and to stimulatevascularization [10]. However, because of its intrinsic brittleness, BAGhas only been successfully applied as a filler or as a coating for metalimplants [9,11], while not being suitable for load-bearing applications.BAG is usually synthesized by melting–quenching–milling routes or thesol–gel technique [12,13], having different compositions, depending onthe concentrations of SiO2, CaO, Na2O, P2O5, K2O, MgO, B2O3 and sometrace metals [14]. Ceramic, glass and glass-ceramic scaffolds have beenfabricated using different methods, such as sponge replication, coatingor foaming techniques, organic-phase burn out, solid freeform fabrication,solvent casting/particulate leaching, and thermally induced phaseseparation [14]. The scaffolds obtained with these techniques are usually

Table 1Compositions of the gellan gum-bioactive glass composite hydrogels.

Samplename

Gellan gum Bioactive glassA⁎

Bioactive glassB⁎⁎

(% w/w) (% w/w) (% w/w)

GG 100GG-25A 75 25GG-50A 50 50GG-25B 75 25GG-50B 50 50

⁎ 70 SiO2–30 CaO (% n/n).⁎⁎ 66 SiO2–22 CaO–10 Na2O–2 P2O5 (% n/n).

28 A. Gantar et al. / Materials Science and Engineering C 43 (2014) 27–36

brittle and are defined by a highmechanical stiffness, negligible elasticityand a hard surface [14], which together lowers their applicability inload-bearing scaffolds for tissue engineering.

On the other hand, natural polymers, such as proteins (silk, collagen,gelatin, fibrinogen, etc.) or polysaccharides (cellulose, amylose, chitin,and gellan gum) have been receiving considerable interest from tissueengineers mainly due to their availability, intrinsic biocompatibility,elasticity and biodegradability. Amongst them, the polysaccharidegellan gum (GG) secreted from Sphingomonas pancimobilis hasrecently been introduced as a promising non-cytotoxic polymer fortissue-engineering applications [15,16]. It is composed of repeatingunits of glucose, rhamnose and glucuronic acid. Most importantly, thetransition from a coiled form at high temperature to a double-helixstructure at room temperature, results in the formation of GGhydrogels,which can then be cross-linked by the addition of counter-ions.Moreover, after a specific freeze-drying and re-hydration, the hydrogelscan give rise to spongy-like hydrogels that have a variety of characteris-tics. In addition to being easily processed in water, this natural,biodegradable and hydrophilic polymer exists in abundance and isrelatively cheap. When compared to other animal-extracted, naturalpolymers, for instance the hyaluronic acid extracted from roostercombs [17], it represents a lower risk of disease transmission. So far,the possibility of tuning GG hydrogels in terms of cross-linking degreeand related mechanical properties, as well as their ability for in situgelation, has allowedGGhydrogels to beproposed for different regener-ativemedical applications [18–21]. On the other hand, likemost naturalpolymers [22], its relatively poor mechanical properties narrow thescope of its applications in tissue engineering [23].

Thus, the combination of inorganic materials, with their limitedflexibility, and biopolymers, with their poor rigidity, becomes a logicaland attractive solution. Accordingly, it has been reported that themechanical properties of brittle, bioactive-glass scaffolds were improvedto some extent by impregnating themwith natural or synthetic polymers[24] and, conversely, that the polymers could be reinforced by incorporat-ing HAp [25–27], TCP [28] or BAG [29–32] particles.

To the best of our knowledge, GG-BAG composite materials for TEapplications have not been reported before, and in this work, compositeGG materials reinforced with fine BAG particles were developed.The preparation of GG-BAG composites was carried out to verify thepossibility of improving the mechanical properties and the bioactivityof the GG using nanoparticulate BAG, such that it would extend theapplication range of the GG to include bone repair. The effect of thisBAG reinforcement was evaluated by an analysis of the microstructure,the mechanical properties and the ability to form HAp in simulatedbody fluid. In addition, in vitro studies were performed by means ofculturing human-adipose-derived stem cells onto the compositematerials.

2. Material and methods

2.1. Preparation of gellan-gum-based composites

BAG powders were synthesized using the particulate sol–geltechnique described elsewhere [33] and thermally treated at 600 °C toremove any organic residues. BAGs with two compositions were used:(A) a 2-component BAG-A with 70% SiO2 and 30% CaO (% n/n), and(B) a 4-component BAG-B containing 66% SiO2, 22% CaO, 10% Na2Oand 2% P2O5 (% n/n, see also in Table 1). Gellan gum powder (Sigma,USA) was dissolved in distilled water (2.0% w/w) under constantstirring at 90 °C to promote the linearization of GG single chains.Meanwhile, the BAG powder was dispersed in a 0.18% w/w CaCl2(Merck, Germany) aqueous solution and admixed with the GG solutionunder the conditions described above and left stirring for an additional20 min. The suspensions with different BAG-to-GG ratios (0, 25 or 50%w/w BAG, see Table 1) were cast into a Petri dish, left to cross-link andspontaneously jellify for 30 min with a temperature decrease and

ionic cross-linking. Discs of Ø 10 ± 0.47 mm and height 5 ± 0.32 mmwere cut out and stored in phosphate-buffered saline (PBS, Sigma,USA) for a further stabilization period of 48 h. To produce driednetworks, hydrogel discs were frozen overnight at −80 °C and freeze-dried for 3 days in a Telstar Cryodos-80 freeze-dryer (Telstar, Spain).Rehydration of the dried networks to form spongy-like hydrogels wasperformed with 48 h of immersion in a PBS solution at room tempera-ture, which was regularly replenished. For the cell-culture studies, thesamples were sterilized with ethylene oxide.

2.2. Microstructural analysis

The microstructure of the GG-BAG composites in the form of driednetworks was observed in cross-section using a field-emission-gunscanning electron microscope (FEG-SEM) (JEOL JSM 7600F, USA)equippedwith an energy-dispersive X-ray spectrometer (EDXS) systemfrom INCA Oxford Instruments. The three-dimensional structure andporosity of the discs were further analysed (one representative sampleper composition) by micro-computed tomography (μ-CT) using ahigh-resolution μ-CT scanner (Skyscan, Konitch, Belgium) with apixel-size resolution of 11.3 μm and an integration time of 1.6 ms. TheX-ray source was set to 32 keV and 191 μA. The data sets werereconstructed using standard cone-beam reconstruction software(NRecon v1.4.3, Skyscan, Belgium). The spongy-like hydrogels wereobserved with a stereo microscope (Zeiss Discovery.V8, Zeiss,Germany).

2.3. Mechanical property assessment

The mechanical properties of the samples in all three forms(hydrogels, dried networks and spongy-like hydrogels) were analysedusing compression tests (Instron 5542, Instron, USA). Discs with thedimensions described above were loaded at a rate of 0.2 mm/minuntil failure. The apparent Young's modulus was estimated from theslope of the linear region of the stress–strain curves in the 0–1% strainrange. Five replicates per sample were tested.

2.4. Dissolution tests

The dissolution of the hydrogels was evaluated by measuring theweight loss during immersion in the PBS solution for 48 h. The sampleswere each immersed into 20 mL of PBS solution and placed on a shakerat an ambient temperature of 37 °C. After 48 h the samples wereremoved from the solution. Prior to weighing, the excess water wasremoved by placing the disc between sheets of filter paper. In addition,the water uptake of the dried networks was assessed by measuring theweight gain after immersion.

29A. Gantar et al. / Materials Science and Engineering C 43 (2014) 27–36

2.5. Evaluation of in vitro mineralisation

The ability of the GG and GG-BAG composites to induce apatiteformation was estimated by immersing the hydrogel discs in simulatedbody fluid (SBF) for up to 30 days at 36.5 °C, according to a proceduredescribed elsewhere [34]. After 14 and 30 days the samples weredried and their surfaces and cross-sections were observed with theFEG-SEM.

2.6. Analysis of human adipose stem cell response

The in vitro biological characterization was performed usinghuman adipose stem cells (hASC) obtained from the protocol of 3B'sResearch Group and Hospital da Prelada (Porto, Portugal). All thesubjects enrolled in this research responded to an Informed Consentform, which was approved by the ethical committees of both theseinstitutions (2007) and this protocol was found to be acceptableto all the subjects. The hASCs were isolated as described elsewhere[35] and cultured in an alpha-MEMmedium (Invitrogen, USA) supple-mented with 10% fetal bovine serum (FBS) (Invitrogen, USA) and 1%antibiotic/antimycotic (Invitrogen, USA). The cells in a concentrationof 5 × 105 cells/scaffold were seeded onto the dried networks, placedin 24-well plates and incubated for 1 h for full hydration of the structure

Fig. 1. FEG-SEM images of the BAG-A particles (A) and cross-sections of the GG (B, C), and GG-Bpresent the cross-sections of the pore walls in samples GG-50 A and GG-50B, respectively.

to form a spongy-like hydrogel. A total of 1 ml of the complete culturemedium was then added and the constructs were further incubatedfor 72 h in a 5% CO2 atmosphere at 37 °C. The culture constructs werethen stained with Calcein-AM (Ca-AM, Invitrogen, USA) and PropidiumIodide (PI, Invitrogen, USA) for the cell-viability evaluation. After thefixation with formalin cytoskeleton F-actin the fibres were stainedwith Phalloidin-TRITC (Sigma, USA) and the nuclei were counterstainedwith 4′,6-diamidino-2-phenylindole (DAPI, Sigma, USA). The sampleswere observed with an AxioImager Z1m microscope (Zeiss, Germany)or Olympus Fluoview FV1000 laser confocal microscope (Olympus,Japan).

3. Results

3.1. The addition of BAG alters the microstructure of the gellan-gumdried networks

Fig. 1A shows the globular BAG-A particles with an estimated size of~200 nm and Fig. 1B-I present cross-sections of the dried networksprepared from GG or GG-BAG. It is evident that the BAG additionstrongly affected the microstructure: in comparison to the GG(Fig. 1B), 25% w/w of BAG-A or BAG-B resulted in a drastic reductionof the pore size (Fig. 1D, G), which further decreased with 50% w/w

AG composites with 25% or 50% BAG-A (D, E) and 25 or 50% BAG-B (G, H). Figures F and I

30 A. Gantar et al. / Materials Science and Engineering C 43 (2014) 27–36

BAG (Fig. 1E, H). There is no clear effect of the BAG composition on themicrostructure. The BAG particles appeared to be homogeneouslydistributed in the GG matrix. Observations of the cross-sections athigher magnification (Fig. 1F, I) revealed that the original globularBAG particles with a size of around 200 nm (Fig. 1A), independent ofthe BAG composition, were embedded into the GG structures in theform of up-to-10-μm-large agglomerates. Moreover, in comparisonwith the ~10-μm-thick and apparently dense GG pore walls with asmooth surface (Fig. 1C), the pore walls of the GG-BAG compositesappeared thicker and their surface was rougher due to the embeddedBAG agglomerates.

The results of the SEM observation of the microstructures werefurther confirmedwith a μ-CT scan of the dried networks. In comparisonwith the GG structures presented in Fig. 2A, the compositeGG-50A (Fig. 2B) revealed a more homogeneous porosity and smallerpores. Moreover, the results of the analysis confirmed that the impactof the BAG composition on the porosity was minor, but an increasedaddition from 25 to 50% w/w resulted in a porosity decrease from 85%to 75% (Fig. 3A). The analysis also confirmed that the amount of BAGreinforcement greatly affected the pore size (Fig. 3B). While the poresin the GG were estimated from Fig. 2A to be larger than 1 mm, theaddition of BAG resulted in an average pore size of 500–600 μm, and180–230 μm for 25 and 50% w/w of BAG. It is also evident from Fig. 3Bthat the increased amount of BAG from 25 to 50% w/w results insignificantly narrower pore size distribution. Similar results wereobtained from the observation of the samples after rehydration(spongy-like hydrogels) under an optical microscope (not presented).

Fig. 2. Three-dimensional scan of dried n

3.2. Mechanical properties of GG-BAG composites depend on thecomposition and amount of BAG

Fig. 4A–D illustrates the effect of the incorporation of the BAGparticles on Young's modulus of the GG-based materials in the form ofas-prepared hydrogels, dried networks and spongy-like hydrogels.Young's modulus of the GG hydrogel increased after the addition of25% BAG-A, while no significant difference was observed for a furtheraddition of BAG-A (Fig. 4A). In contrast, the addition of 25% BAG-Bdecreased Young's modulus, while with 50% the modulus remainedunchanged.

After freeze-drying, independent of the composition and amount ofBAG, the Young's modulus of the composite dried networks was twiceas high as that for the GG (Fig. 4B); whereas for the GG the moduluswas 0.7 MPa, the value increased up to ~1.5 MPa for the GG-25A andup to ~1.8 MPa for the GG-50A. Upon rehydration (spongy-likehydrogels) this effect was reverted to the values similar to thoseobserved for the as-prepared hydrogels (Fig. 4C).

As an illustration, the stress–strain curves for the average examinedmaterials of each group are presented in Fig. 5A–D. Fig. 5A demonstratesa significant difference in the behaviour of the GG and sintered BAG(BAG-A): under loading, the BAG exhibited no plastic deformation dueto the high Young's modulus (225 MPa), and much higher achievedvalues of the compressive stress than for the GG. More detailedbehaviours for both materials and their composites in various formsare presented in Fig. 5B for the hydrogels, 5C for the dried networksand 5D for the spongy-like hydrogels (note the lower y-axis scale than

etworks of GG (A) and GG-50A (B).

Fig. 3. Porosity (A) and pore size distribution (B) for the composite GG-BAG dried networks as determined by μ-CT scan of the representative samples.

31A. Gantar et al. / Materials Science and Engineering C 43 (2014) 27–36

in Fig. 5A). In contrast to BAG, which exhibits a brittle-fracture modeleading to a catastrophic failure, the GG and GG-BAG hydrogelsexhibited large elastic and plastic deformations, both of which aredependent on the amount and, in particular, on the composition of theBAG. For the as-prepared hydrogels (Fig. 5B), as well as for thespongy-like hydrogels (Fig. 5D), it is evident that the 25 or 50% BAG-Aresulted in a much higher mechanical strength than with the BAG-B,while the effect of the amount of BAG is less pronounced. As shown inFig. 5C, for the dried networks the failure did not occur up to at least20% of strain. In this case, the influence of the amount of BAG prevailedover its composition.

3.3. Weight and morphology changes in PBS and SBF

When the dried networkswere immersed in PBS (rehydration step),a rapid weight gain due to water uptake was observed. As is evidentfrom Fig. 6A, the highest weight gain of 900% was observed for GG,~750% for the composites with 25% BAG and ~600% for the compositewith 50% BAG. No significant effect of the BAG composition wasobserved.

In contrast, during 48 h of immersion in PBS at 37 °C, the weight ofthe GG and GG-BAG composite hydrogels was reduced due to partialdissolution, Fig. 6B. For the GG hydrogels the weight loss was 9.5%. Asimilar result was also observed for the composite GG-25A, while amuch larger weight loss (17%) was observed for GG-25B and GG-50B.The lowest weight loss (~5%) was observed for the sample GG-50A.

Fig. 4. Young's modulus for as-prepared hydrogels (A),

After the immersion of the GG and composite hydrogels in the SBF,for GG no change inmorphologywas observedwithin 30 days,whereascauliflower-like HAp crystals were observed on the GG-BAG hydrogelcomposites after 14 days. Moreover, HAp was not only found on thesurface, but also in the central parts of the samples. After 30 days, thesurface of the composite materials was fully covered with HAp crystals(Fig. 7), for which a Ca/P ratio of 1.67 was confirmed by EDS analysis.

3.4. hASCs are able to attach to and spread within all thespongy-like hydrogels

The effect of the composition and the amount of BAG on hASC'sadhesion and viability was assessed to provide information about thebiological performance of the GG and GG-BAG spongy-like hydrogelsformed at the time of seeding. The confocal (left column) and fluores-cent image (right column) of the hASC after 72 h of culture on thegellan-gum spongy-like hydrogels (A, B), GG-50A spongy-likehydrogels (C, D) and GG-50B (E, F) are presented in Fig. 8. After 72 hof culture the hASCswere able to attach to all the spongy-like hydrogelsand spread within the structures, as was observed by the organizedactin filaments and the typical fibroblastic morphology (Fig. 8 A, C, E).Equally important, the majority of cells within the spongy-likehydrogels were viable after 72 h of culture, as demonstrated by theCalcein/PI staining (Fig. 8 B, D, F). The hydrogels containing BAGparticles showed a strong, red, auto-fluorescence characteristic for theBAG particles.

dried networks (B) and spongy-like hydrogels (C).

Fig. 5. Compressive stress vs. strain curves for BAG and GG hydrogel (A), and their composites in the form of as-prepared hydrogel (B), dried network (C) and spongy-like hydrogel(D) with various compositions.

Fig. 6.Water uptake during 48 h rehydration of dried networks in PBS (A) and weight loss of hydrogels during 48 h of immersion in PBS (B).

32 A. Gantar et al. / Materials Science and Engineering C 43 (2014) 27–36

Fig. 7. FEG-SEM images of the GG-50A after immersion in SBF for 30 days at two magnifications.

Fig. 8. Representative confocal microscopy images of hASCs within GG (A, B), GG-50A (C, D) and GG-50B (E, F) spongy-like hydrogels after 72 h of culture after Phalloidin-TRITC (red)(A, C, E) and Calcein AM (green)/Propidium Iodide (red) (B, D, F) staining. Cell nuclei were stained with DAPI (blue) (A, C, E). (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)

33A. Gantar et al. / Materials Science and Engineering C 43 (2014) 27–36

34 A. Gantar et al. / Materials Science and Engineering C 43 (2014) 27–36

4. Discussion

Gellan gum, a biodegradable and bioinert natural biopolymer, wasrecently introduced as a novel material for tissue engineering, due toits numerous advantageous properties [17]. Unlike, for instance,polycaprolactone- and polylactic-acid-based synthetic polymers thathave also been used in combination with bioactive-glass particles[36–38], the GG is water soluble and hydrophilic by nature. Thus,no chlorinated solvents (e.g., dioxane) are needed in its preparation.In addition, the GG degrades into its four, naturally occurring, constitu-ent sugars, whereas most synthetic polymers form acidic products. Inspite of its valuable features [16], like many other biodegradablehydrogels, GG hydrogel lacks good mechanical properties and theability to form a bone-like mineral phase. Therefore, several studieshave reported on the incorporation of HAp or TCP particles intopolymeric matrixes, such as collagen [39], chitosan [28,40], andgelatine/gellan [41] to overcome these limitations. Bioactive glass hasalso been used as a reinforcement phase in some biopolymers, such aspolycaprolactone [36], polylactic acid [24,37,38], chitosan [29,30],collagen [42] and collagen-hyaluronic acid-phosphatidylserine [31].The reinforcement of GG was recently achieved by the introduction ofHAp particles [25], while to the best of our knowledge the incorporationof BAG particles into GG has not been previously reported.

In the present study, two grades of BAG with a particle size ofaround 200 nm were employed as a reinforcement for GG and theresultingGG-BAG composites exhibited isotropic structureswith evenlydistributed and well-interconnected pores. Themost pronounced effectof the BAG powder's incorporation was a significant reduction in theaverage pore size, which was, however, still relatively large, i.e.,500–600 μm for 25% BAG and 180–220 μm for 50% BAG. The observednarrower pore size distribution for the samples with 50% BAG impliesa better distribution of particles in the matrix. Also, the porosity waskept high, 85 and75% for compositeswith 25 and 50% BAG, respectively.The effect was nearly independent of the composition of the bioactiveglass, which is probably connected to the similar size of the BAGagglomerates embedded in the porewalls. These ~10-μm-large agglom-erates were formed (or remained) during the addition to the GGfrom the originally globularly shaped BAG particles. Although such amicrostructure was not desired, the BAG agglomerates captured in thethin GG pore walls contributed to a significant improvement of the GGproperties. Amongst the other effects discussed below, their presenceincreased the roughness of the pore walls, which can be assumed tobe beneficial for the attachment of cells [43].

The observation of the samples at various stages of their preparationconfirmed that from the apparently non-porous, as-prepared hydrogel,a highly porous network is formed by drying, and after re-hydration theporosity is partly retained. Thus, the huge water uptake observed afterre-hydration proceeds from the swelling of the pore walls and fromfilling the open pores with the solution.

In addition to other characteristics, themechanical behaviour that isusually described by Young's modulus and the compressive strength, aswell as the deformation under normal load, are highly relevant for amaterial that is to be used as a scaffold in bone-tissue engineering. It iswell known that, on the one hand, biopolymer hydrogels are typicallytoo “soft”, meaning that even at a low normal load a critically largedeformation (or even failure) takes place, which corresponds to a lowYoung's modulus, typically around 100 kPa [44]. On the other hand,scaffolds produced from inorganic materials, such as HAp, TCP andBAG, are too brittle and break at a certain load after negligible elastic,and without any plastic, deformation. The difference between thesetwo groups of materials is clearly illustrated in Fig. 5A. For the sinteredBAG pellets, brittle fracture appeared. In contrast, the GG-basedmaterials exhibit a much lower stiffness, expressed as a low Young'smodulus in the range 0.2 to 2 MPa, a lower point of failure and a largerdeformation. As expected, the curves for the GG-BAG compositesappeared between the curves for the two individualmaterials, implying

a reasonable compromise in terms of properties. The values obtained forthe dried networks are generally higher than those for the as-preparedand spongy-like hydrogels; however, since they are mainly related tothe durability during the manipulation and storage, the behaviour ofthe spongy-like hydrogels was considered as the most relevant fordescribing the behaviour of the scaffold.

For the spongy-like GG hydrogel with a Young's modulus of~0.4 MPa, the failure occurred already at ~0.02 MPa and ~5% elasticdeformation. However, Young's modulus and the point of failureappeared at much higher loads when the GG was reinforced with BAGparticles. While the composition of the BAG particles did not affect themicrostructure of the GG-BAG composites, it was found to have asignificant effect on their mechanical properties. The incorporation of25 or 50% w/w of the two-component BAG-A significantly increasedthe stiffness (Young's modulus increased from 0.4 to ~1.2 MPa) andthe strength (the point of failure increased from 0.02 to N0.11 MPa).The improvement was less pronounced for the four-componentBAG-B, which might be attributed to the active role of the Ca2+, whichis known as an effective cross-linker for the GG [45], and is releasedfrom the BAG during the composite hydrogel's preparation, thuscontributing to the stiffness of the GG. In the four-component BAG-B,the CaO is partly replaced with Na2O, which causes a less effectivecross-linking due to the released monovalent ion Na+ [45]. This effectis less pronounced in the case of dried networks, which is probablythe consequence of the precipitation of oxides. It is interesting to notethat in dried networks the amount of BAG reinforcement plays a largerrole than in hydrogels, which can be ascribed to a decrease in theporosity with an increase in the amount of added BAG. For example,as a consequence of the porosity decrease from 85 to 75%, Young'smodulus increased from 1.2 MPa for 25% BAG to 1.9 MPa for 50% BAG.As expected, in addition to the intrinsic properties of the material,the porosity of the material also had a significant effect on Young'smodulus [46].

Considering the remarkably similar microstructures of all fourcomposites, the observed higher mechanical strength for the compos-ites with BAG-A in comparison with those containing BAG-B supportsthe above suggestion that in the case of the 2-component BAG-A, thereleased Ca2+ contributes to the cross-linking of the GG, while inthe 4-component BAG-B the released monovalent Na+ detrimentallyaffects the cross-linking. Thus, the above findings suggest that with anoptimal addition of BAG, preferably containing only two- and higher-valent ions, the behaviour of the scaffold material with a certain degreeof porosity can be tailored in termsof thedesiredmechanical properties.

Although the mechanical properties of the GG were significantlyimproved by BAG addition (in particular BAG-A), the achieved Young'smodulus of the GG-BAG hydrogel (1.9 MPa in the dried form) andcompression strength are still low, not only in comparison withreported values for sintered BAG scaffolds [47–49], but also whencompared with synthetic polymers and their composites [36–38,50].However, it should be pointed out that this relatively high strength isgained at the expense of a lower porosity of the scaffolds and aslow bioresorption. One further drawback of the BAG scaffolds inrelation to BAG-reinforced GG is its degradation-induced alkalinityof the micro-environment. In comparison with the hydrophobicBAG-containing PCL or PLA [36–38], the GG-BAG hydrogels have anadvantage in terms of hydrophilicity.

It has been suggested by Misra et al. [51] that the size of the fillerparticles strongly affects the mechanical properties of compositematerials as a result of a different interaction between thefiller particlesand the polymer matrix. Accordingly, careful processing to achieve abetter dispersion of the BAG nanoparticles in the matrix is proposed tofurther improve the mechanical properties of the composite scaffoldsto approach more closely that of the bone.

While on the one hand, the biodegradability of the scaffolds in bodyfluid is desired, on the other, their mechanical properties are impairedduring biodegradation. Of course, ideally, the newly formed mineral

35A. Gantar et al. / Materials Science and Engineering C 43 (2014) 27–36

phase compensates for the deterioration in themechanical properties. Aswas also proven in this study, the GG alone is unable, by its very nature, toproduce a mineral phase, while the BAG is known for its relatively fastdissolution and strong ability to induce mineralisation. Therefore, inaddition to the above-discussed improvement of the microstructure andmechanical properties, the reinforcement of the GG with BAG particlesalso resulted inmineral phase formation in the presence of the simulatedbody fluid for all the BAG-containing samples. The appearance of thematerials after the test suggests that the HAp formation was faster thanthe dissolution of the polymer phase (see enlarged area in Fig. 7). Theappearance of HAp in the central part of the GG-BAG hydrogels proves ahigh rate of ion diffusion through the hydrogel, which is sometimesmistakenly considered as a low-permeability structure.

The scaffoldmaterial is expected to fully degradewithin a certain timeafter the implantation, leaving behind no solid residues, except for thenewly formed bone tissue. In this respect, BAG has advantages overHAp or TCP, which were often observed to degrade only partially or tooslowly. The BAG-B was confirmed in a previous study using in vivo teststo completely degrade within ten weeks, resulting in fully mineralisedbone within a porous titanium structure [9]. In the present study, onlythe short-term dissolution of the examined materials was examinedand it was found that the rate is dependent on the BAG's composition.The addition of the 4-component BAG-B resulted in an increasedrate of weight loss in comparison with the GG, while conversely, the2-component BAG-A slightly hindered the dissolution. This can beunderstood as an additional confirmation of the effect of the releasedmono- and divalent ions on the GG cross-linking and, in addition,suggests the possibility to tailor the degradation rate using the composi-tion of the BAG reinforcement.

It is known that the first cell-material contact is crucial for thescaffold's success and therefore the adhesion of the cells has to occurwithin a few hours so as to maintain good cell viability. Accordingly, theadhesion itself has great importance for various intercellular signallingpathways that direct the cell viability, proliferation and differentiation[52,53]. In the present study, the cellular response to the GG-BAGspongy-like hydrogels in comparison with the GG was assessed usinghuman-adipose stem cells (hASCs). After 72 h of hASCs culture no cleardifferences in terms of cell viability and adhesionwere observed betweenthe GG and the GG-BAG spongy-like hydrogels. Also, the hASCs exhibiteda favourable growth morphology, reflected in active cytoplasmextensions of the cell membrane and good cell viability, similar to theGG without the BAG. Moreover, no occlusion of the pores was observedduring this short culturing period. Due to the favourable pore size andconnectivity in the examined spongy-like hydrogels, hASCs could alsoeasily reach and adhere to the porewalls in the central part of the scaffold.Based on the known osteogenic properties of bioactive glasses resultingfrom the effect of the ionic dissolution products [54,55], it is believedthat the addition of BAG is beneficial for the GG material to promote celladhesion and proliferation; however, this could not be expressed withina short-term test (72 h) and it should therefore be verified in a further,longer-term culture of hASCs within the GG-BAG materials.

Taken together, the prepared GG-BAG spongy-like hydrogels, inparticular those prepared with BAG-A, exhibited characteristics thatclosely coincide with the recommendations for scaffold materials, i.e.,an open and well-interconnected porosity of ~80%, a pore size of~100–200 μm [56] and a rough surface [43]. However, the mechanicalproperties are still insufficient, but it is supposed that a furtherimprovement is possible by optimizing the processing. The structure incombination with the ions released from the BAG provides a favourabletemporary scaffold for the cells to attach to and proliferate.

5. Conclusions

This work presents bioactive-glass-reinforced gellan-gum hydrogels(GG-BAG) as novelmaterials proposed for the scaffolding in bone-tissueengineering. Although the nanoparticulate, globular, BAG particles were

agglomeratedwithin the GGmatrix, they contributed significantly to animproved microstructure and the mechanical properties of the originalGG hydrogels. It was also shown that the mechanical properties can betailored by the composition and the amount of BAG. The compositescaffold containing 50% w/w 2-component BAG exhibited a Young'smodulus of 1.9 MPa in dried form and ~1.2 MPa in hydrogel form.Both values are still well below the desired value that would allowload-bearing applications and lower than the highest achieved valuesreported for sintered BAG with moderate porosity and BAG-containinghydrophobic biopolymers. A further improvement of the mechanicalproperties is proposed by careful processing, i.e., mainly a betterdispersion of the nanoparticles in the matrix.

Moreover, the BAG particles incorporated into GG matrix conferredon the developed structures the possibility to mineralize in vitro,which can be further enhanced by combining them with adipose stemcells, as the developed materials supported its adhesion and spreadingwithin the structure without compromising its viability. The presentednanoparticulate bioactive-glass-reinforced gellan-gum hydrogelsshowed great promise for applications in bone regeneration, for eitheracellular or cellular strategies.

Conflict of interest

No benefit of any kindwill be received either directly or indirectly bythe authors.

Acknowledgements

The Slovenian Research Agency is acknowledged for its financialsupport of the PhD study of the first author, Ms. Ana Gantar. Thiswork was partly performed within the short-term scientific mission(STSM reference number: ECOST-STSM-MP1005-250912-020779) ofMs. Gantar in the 3B's group under the COST Action MP1005 “Fromnano to macro biomaterials (design, processing, characterization,modelling) and applications to stem cell regenerative orthopaedic anddental medicine – NAMABIO”. The authors wish to thank Dr PaulMcGuiness for proofreading the article.

References

[1] J.D. Kretlow, A.G. Mikos, From material to tissue: biomaterial development, scaffoldfabrication, and tissue engineering, AIChE J. 54 (12) (2008) 3048–3067.

[2] F.J. O'Brien, Biomaterials & scaffolds for tissue engineering, Materials Today 14 (3)(2011) 88–95.

[3] B. Dhandayuthapani, et al., Polymeric scaffolds in tissue engineeringapplication: a review, Int. J. Polym. Sci. 2011 (2011) 1–19.

[4] L.L. Hench, Bioceramics, J. Am. Ceram. Soc. 81 (7) (1998) 1705–1728.[5] F. Baino, C. Vitale-Brovarone, Three-dimensional glass-derived scaffolds for bone

tissue engineering: current trends and forecasts for the future, J. Biomed. Mater.Res. A 97A (4) (2011) 514–535.

[6] D. Zaffe, C. Bertoldi, U. Consolo, Accumulation of aluminium in lamellar bone afterimplantation of titanium plates, Ti-6Al-4 V screws, hydroxyapatite granules,Biomaterials 25 (17) (2004) 3837–3844.

[7] W.R. Nunery, et al., Exposure rate of hydroxyapatite spheres in the anophthalmicsocket — histopathological correlation and comparicon with silicone sphere,Ophthal. Plast. Reconstr. Surg. 9 (2) (1993) 96–104.

[8] I.D. Xynos, et al., Bioglass ®45S5 stimulates osteoblast turnover and enhances boneformation in vitro: implications and applications for bone tissue engineering, Calcif.Tissue Int. 67 (4) (2000) 321–329.

[9] N. Drnovsek, et al., Bioactive glass enhances bone ingrowth into the porous titaniumcoating on orthopaedic implants, Int. Orthop. 36 (8) (2012) 1739–1745.

[10] R.M. Day, Bioactive glass stimulates the secretion of angiogenic growth factors andangiogenesis in vitro, Tissue Eng. 11 (5–6) (2005) 768–777.

[11] E. Schepers, et al., Bioactive glass particulate material as a filler for bone lesions, J.Oral Rehabil. 18 (5) (1991) 439–452.

[12] L.L. Hench, J.K. West, The sol–gel process, Chem. Rev. 90 (1990) 33–72.[13] L.L. Hench, J. Wilson, An Introduction to Bioceramics, Vol. 1World Scientific, 1993.[14] J.R. Jones, Review of bioactive glass: from Hench to hybrids, Acta Biomater. 9 (1)

(2013) 4457–4486.[15] A.M. Smith, et al., An initial evaluation of gellan gum as a material for tissue

engineering applications, J. Biomater. Appl. 22 (3) (2007) 241–254.[16] J.T. Oliveira, et al., Gellan gum: a new biomaterial for cartilage tissue engineering

applications, J. Biomed. Mater. Res. A 93A (3) (2010) 852–863.

36 A. Gantar et al. / Materials Science and Engineering C 43 (2014) 27–36

[17] T. Yamada, T. Kawasaki, Microbial synthesis of hyaluronan and chitin: newapproaches, J. Biosci. Bioeng. 99 (6) (2005) 521–528.

[18] D.R. Pereira, et al., Hydrogels in acellular and cellular strategies for intervertebraldisc regeneration, J. Tissue Eng. Regen. Med. 7 (2) (2013) 85–98.

[19] J. Silva-Correia, et al., Angiogenic potential of gellan-gum-based hydrogels forapplication in nucleus pulposus regeneration: in vivo study, Tissue Eng. A 18(11–12) (2012) 1203–1212.

[20] D.R. Pereira, et al., Gellan gum-based bilayered scaffolds for application inosteochondral tissue engineering, J. Tissue Eng. Regen. Med. 6 (2012) 185.

[21] J.T. Oliveira, et al., Gellan gum injectable hydrogels for cartilage tissue engineeringapplications: in vitro studies and preliminary in vivo evaluation, Tissue Eng. A 16(1) (2010) 343–353.

[22] N. Goonoo, et al., An assessment of biopolymer- and synthetic polymer-basedscaffolds for bone and vascular tissue engineering, Polym. Int. 62 (4) (2013)523–533.

[23] D.F. Coutinho, et al., Modified gellan gum hydrogels with tunable physical andmechanical properties, Biomaterials 31 (29) (2010) 7494–7502.

[24] Q.Z. Chen, A.R. Boccaccini, Poly(D, L-lactic acid) coated 45S5 Bioglass®-basedscaffolds: processing and characterization, J. Biomed. Mater. Res. A 77A (3) (2006)445–457.

[25] M.G. Manda-Guiba, et al., Gellan gum–hydroxyapatite composite hydrogels for bonetissue engineering, J. Tissue Eng. Regen. Med. 6 (2012) 15.

[26] C. Du, et al., Three-dimensional nano-HAp/collagen matrix loading with osteogeniccells in organ culture, J. Biomed. Mater. Res. 44 (4) (1999) 407–415.

[27] A. Bigi, et al., Bonelike apatite growth on hydroxyapatite-gelatin sponges fromsimulated body fluid, J. Biomed. Mater. Res. 59 (4) (2002) 709–715.

[28] Y. Zhang, M. Zhang, Synthesis and characterization of macroporous chitosan/calciumphosphate composite scaffolds for tissue engineering, J. Biomed. Mater. Res. 55 (3)(2001) 304–312.

[29] M. Peter, et al., Development of novel α-chitin/nanobioactive glass ceramiccomposite scaffolds for tissue engineering applications, Carbohydr. Polym. 78 (4)(2009) 926–931.

[30] M. Peter, et al., Nanocomposite scaffolds of bioactive glass ceramic nanoparticlesdisseminated chitosan matrix for tissue engineering applications, Carbohydr.Polym. 79 (2) (2010) 284–289.

[31] Y. Wang, et al., Development and characterization of novel biomimetic compositescaffolds based on bioglass-collagen-hyaluronic acid-phosphatidylserine for tissueengineering applications, Macromol. Mater. Eng. 291 (3) (2006) 254–262.

[32] J.J. Blaker, et al., In vitro evaluation of novel bioactive composites based on Bioglass(R)-filled polylactide foams for bone tissue engineering scaffolds, J. Biomed. Mater.Res. A 67A (4) (2003) 1401–1411.

[33] N. Drnovšek, Development of Coatings on Ti6Al4V Alloy for New GenerationImplants Bone with Improved Osseointegration, Jožef Stefan InternationalPostgraduate School, Ljubljana, 2012. 132.

[34] T. Kokubo, H. Takadama, How useful is SBF in predicting in vivo bone bioactivity?Biomaterials 27 (15) (2006) 2907–2915.

[35] M.T. Cerqueira, et al., Human adipose stem cells cell sheet constructs impactepidermal morphogenesis in full-thickness excisional wounds, Biomacromolecules14 (11) (2013) 3997–4008.

[36] P. Fabbri, et al., Highly porous polycaprolactone-45S5 Bioglass (R) scaffolds for bonetissue engineering, Compos. Sci. Technol. 70 (13) (2010) 1869–1878.

[37] Z. Hong, R.L. Reis, J.F. Mano, Preparation and in vitro characterization of novelbioactive glass ceramic nanoparticles, J. Biomed. Mater. Res. A 88A (2) (2009)304–313.

[38] N. Barroca, et al., Tailoring the morphology of high molecular weight PLLA scaffoldsthrough bioglass addition, Acta Biomater. 6 (9) (2010) 3611–3620.

[39] J.P. Gleeson, N.A. Plunkett, F.J. O'Brien, Addition of hydroxyapatite improvesstiffness, interconnectivity and osteogenic potential of a highly porouscollagen-based scaffold for bone tissue regeneration, Eur. Cell. Mater. 20 (2010)218–230.

[40] P.B. Malafaya, R.L. Reis, Bilayered chitosan-based scaffolds for osteochondral tissueengineering: influence of hydroxyapatite on in vitro cytotoxicity and dynamicbioactivity studies in a specific double-chamber bioreactor, Acta Biomater. 5 (2)(2009) 644–660.

[41] N. Barbani, et al., Hydroxyapatite/gelatin/gellan sponges as nanocomposite scaffoldsfor bone reconstruction, J. Mater. Sci. Mater. Med. 23 (1) (2012) 51–61.

[42] H.-W. Kim, J.-H. Song, H.-E. Kim, Bioactive glass nanofiber–collagen nanocompositeas a novel bone regeneration matrix, J. Biomed. Mater. Res. A 79A (3) (2006)698–705.

[43] H.-I. Chang, Y. Wang, E. D, Cell responses to surface and architecture of tissueengineering scaffolds, Regenerative Medicine and Tissue Engineering — Cells andBiomaterials, 2011.

[44] H. Yamaoka, et al., Cartilage tissue engineering using human auricular chondrocytesembedded in different hydrogel materials, J. Biomed. Mater. Res. A 78A (1) (2006)1–11.

[45] E. Miyoshi, T. Takaya, K. Nishinari, Rheological and thermal studies of gel–soltransition in gellan gum aqueous solutions, Carbohydr. Polym. 30 (2–3) (1996)109–119.

[46] L.M. Mathieu, et al., Architecture and properties of anisotropic polymer compositescaffolds for bone tissue engineering, Biomaterials 27 (6) (2006) 905–916.

[47] N.D. Doiphode, et al., Freeze extrusion fabrication of 13–93 bioactive glass scaffoldsfor bone repair, J. Mater. Sci. Mater. Med. 22 (3) (2011) 515–523.

[48] X. Liu, M.N. Rahaman, Q.A. Fu, Oriented bioactive glass (13–93) scaffolds withcontrollable pore size by unidirectional freezing of camphene-based suspensions:microstructure and mechanical response, Acta Biomater. 7 (1) (2011) 406–416.

[49] X. Liu, et al., Bioactive borosilicate glass scaffolds: improvement on the strength ofglass-based scaffolds for tissue engineering, J. Mater. Sci. Mater. Med. 20 (1)(2009) 365–372.

[50] B. Cao, et al., Study on surface modification of porous apatite-wollastonite bioactiveglass ceramic scaffold, Appl. Surf. Sci. 255 (2) (2008) 505–508.

[51] S.K. Misra, et al., Comparison of nanoscale and microscale bioactive glass on theproperties of P(3HB)/Bioglass (R) composites, Biomaterials 29 (12) (2008)1750–1761.

[52] R.O. Hynes, Integrins: bidirectional, allosteric signaling machines, Cell 110 (6)(2002) 673–687.

[53] B.G. Keselowsky, D.M. Collard, A.J. García, Surface chemistry modulates focaladhesion composition and signaling through changes in integrin binding,Biomaterials 25 (28) (2004) 5947–5954.

[54] P. Valerio, et al., The effect of ionic products from bioactive glass dissolution onosteoblast proliferation and collagen production, Biomaterials 25 (15) (2004)2941–2948.

[55] V.G. Varanasi, et al., The ionic products of bioactive glass particle dissolutionenhance periodontal ligament fibroblast osteocalcin expression and enhanceearly mineralized tissue development, J. Biomed. Mater. Res. A 98A (2) (2011)177–184.

[56] P. Kasten, et al., Porosity and pore size of β-tricalcium phosphate scaffold caninfluence protein production and osteogenic differentiation of humanmesenchymal stem cells: an in vitro and in vivo study, Acta Biomater. 4 (6)(2008) 1904–1915.