gelation and biocompatibility of injectable alginate-calcium phosphate gels for bone regeneration

Gelation and biocompatibility of injectable alginate–calcium phosphategels for bone regeneration

D. Alves Cardoso,1,2 J. J. J. P. van den Beucken,2 L. L. H. Both,1 J. Bender,3 J. A. Jansen,2

S. C. G. Leeuwenburgh2

1EMCM B.V., Middenkampweg 17, 6545 CH Nijmegen, The Netherlands2Department of Biomaterials, Radboud University Nijmegen Medical Center, 6500 HB Nijmegen, The Netherlands3Bender Analytical Holding B.V., Beukstraat 73, 3581 XE Utrecht, The Netherlands

Received 4 February 2013; revised 26 March 2013; accepted 4 April 2013

Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/jbmm.34754

Abstract: An emerging approach toward development of

injectable, self-setting, and fully biodegradable bone substi-

tutes involves the combination of injectable hydrogel matri-

ces with a dispersed phase consisting of nanosized calcium

phosphate particles. Here, novel injectable composites for

bone regeneration have been developed based on the com-

bination of ultrapure alginate as the matrix phase, crystal-

line CaP [monetite and poorly crystalline hydroxyapatite

(HA)] powders as both a dispersed mineral phase and a

source of calcium for cross-linking alginate, glucono-delta-

lactone (GDL) as acidifier and glycerol as both plasticizer

and temporary sequestrant. The composites were maxi-

mized with respect to CaP content to obtain the highest

amount of osteoconductive filler. The viscoelastic and physi-

cochemical properties of the precursor compounds and

composites were analyzed using rheometry, elemental anal-

ysis (for calcium release and uptake), acidity [by measuring

pH in simulated body fluid (SBF)], general biocompatibility

(subcutaneous implantation in rabbits), and osteocompati-

bility (implantation in femoral condyle bone defect of rab-

bits). The gelation of the resulting composites could be

controlled from seconds to tens of minutes by varying the

solubility of the CaP phase (HA vs. monetite) or amount of

GDL. All composites mineralized extensively in SBF for up

to 11 days. In vivo, the composites also disintegrated upon

implantation in subcutaneous or bone tissue, leaving behind

less degradable but osteoconductive CaP particles. Although

the composites need to be optimized with respect to the

available amount of calcium for cross-linking of alginate, the

beneficial bone response as observed in the in vivo studies

render these gels promising for minimally invasive applica-

tions as bone-filling material. VC 2013 Wiley Periodicals, Inc. J

Biomed Mater Res Part A: 00B:000–000, 2013.

Key Words: biomaterials, bone, composites, natural polymer,

alginate, calcium phosphates, injectable

How to cite this article: Alves Cardoso D, van den Beucken JJJP, Both LLH, Bender J, Jansen JA, Leeuwenburgh SC. 2013. Ge-lation and biocompatibility of injectable alginate–calcium phosphate gels for bone regeneration. J Biomed Mater Res Part A2013:00: 000–000.

INTRODUCTION

As life expectancy increases and degenerative bone diseasesbecome more urgent, a rapidly expanding number of elderpatients will need effective bone regeneration and implantprosthesis procedures. These degenerative diseases imposehuge personal and societal burdens to future generationsbecause surgery is the only treatment option in most cases.Bone and joint diseases cause more functional limitations inthe adult population in most welfare states than any othergroup of disorders.1 As a consequence, bone is a common tis-sue for transplant procedures, second only to blood. Bone tis-sue distinguishes itself from any other soft tissue by itsremarkable regenerative capacity owing to the concertedaction of bone-forming osteoblasts and bone-resorbing osteo-clasts.2,3 Nevertheless, large critical size bone defects requirea bone graft that fills the defect and facilitates tissue regener-ation. Currently, the only treatment option that effectively

fulfills all requirements needed for the replacement of bonetissue still involves the use of autograft or allograft bone.3

Still, the use of auto- or allografts is associated with severedrawbacks,3,4 such as (i) extra surgery to harvest bone tissue,(ii) risk of septic complications, viral transmission or diseasetransfer, (iii) morbidity and severe pain complications at thesite of harvesting, (iv) limited bone volume at donor site, and(v) increasing organ donor shortages.

Consequently, a wide variety of synthetic ceramic andpolymeric bone substitutes has been developed and commer-cialized over the past three decades. Synthetic calcium phos-phate (CaP) ceramics represent the most widely usedbiomaterials for bone regenerative treatments due to theirbone-bonding and osteoconductive behavior.5–7 Still, pure CaPceramics are difficult to mold into the desired defect shapewhen applied as granules due to the intrinsic brittleness ofthis class of materials. From a clinical point of view,

Correspondence to: S. C. G. Leeuwenburgh; e-mail: [email protected]

VC 2013 WILEY PERIODICALS, INC. 1

injectability of biomaterials for reconstruction of osseousdefects offers several clinical and economical advantages ascompared to solid, prefabricated implants. Using flowablematerials, complete filling of the defect site can be establishedby means of minimally invasive techniques, thus avoidingvoids, which can lead to fibrous encapsulation. Althoughinjectable CaP cements have been developed to overcome thishandling problem,8 the degradation rates of these cements donot match with the rate of bone modeling. In most cases, CaPceramics consist of microstructured hydroxyapatite (HA) sincethis thermodynamically most stable polymorph is present inhuman bone tissue. Typically, HA bone substitutes reveal veryslow degradation rates, which impede complete regenerationof bone defects,8 prompting further research on developmentof injectable, biodegradable bone substitutes.

An emerging approach toward development of injectable,self-setting, and fully biodegradable bone substitutes involvesthe combination of injectable hydrogel matrices with a dis-persed phase consisting of nanosized CaP particles.9 Hydrogelsare ideal candidates for use in bone tissue regeneration anddrug delivery since they are generally biocompatible, biode-gradable, and injectable,10,11 whereas the combination of poorcrystallinity and high specific surface area renders nanosizedCaP particles more degradable than highly crystalline, sinteredmicrostructured CaP ceramics.9 Therefore, a plethora of bothsynthetic and natural hydrogels has been investigated over thepast decade for bone regenerative applications.12 In thatrespect, alginate hydrogels are particularly interesting sincetheir efficient cross-linking by means of divalent cations suchas calcium renders alginate hydrogels highly suitable forimplantation into human tissues. Alginate is a natural linear,unbranched copolymer derived from brown algae composedof (1–4)-linked b-D-mannuronate (M) and its C-5 epimera-L-guluronate (G) residues, which are linked together in differ-ent sequences or blocks. The carboxylic acid groups as presentin G residues are responsible for ionic cross-linking by electro-static interactions with divalent cations such as Ca21.

Since the 1970s, alginate hydrogels have been combinedwith CaPs but the vast majority of these formulations werecross-linked prior to implantation and did not gel in situ.13–18

Herein, an injectable, in situ setting alginate–CaP compositeshas been developed that can be extruded through a two-component dual syringe. These composites were composedof a matrix phase containing ionically cross-linkable algi-nate, a polyol (glycerol) as plasticizer and glucono-delta-lac-tone (GDL) as acidifier, whereas the dispersed phaseconsisted of nanosized crystalline CaP particles. The polyolwas added to the mixture in order to increase the homoge-neity, dryness and mechanical strength of the resulting com-posites,19 whereas GDL was introduced in order to releasecalcium ions from the nanosized CaP phase upon gradualhydrolysis of GDL to gluconic acid.19,20 The use of GDL toinduce cross-linking of alginate hydrogels has been reportedbefore with respect to sparingly soluble calcium salts, suchas CaCO3 or CaSO4,

21 but its potential for cross-linking ofalginate–CaP composites has been poorly investigated.

Therefore, the current study aimed to investigate thegelation mechanism and biocompatibility of alginate–

glycerol–CaP composites, which were internally cross-linkedby means of GDL-induced acidification of two types of crys-talline CaP (i.e., monetite and poorly crystalline HA). Precur-sor compounds and composites were characterized usingscanning electron microscopy (SEM), X-ray diffraction, Fou-rier-transform infrared spectroscopy, and rheology, whereasthe release of calcium from CaP precursor powders wasquantified using a colorimetric assay. Moreover, the calcifica-tion behavior of injectable composites was evaluated uponsoaking in simulated body fluid (SBF) by monitoring cal-cium levels and pH changes in the supernatant, whereas thebiocompatibility of HA-containing composites was evaluatedby means of ectopic (subcutaneous) and orthotopic (femoralcondyle) implantation in New Zealand white rabbits.

MATERIALS AND METHODS

ChemicalsUltrapure, high-molecular-weight HighG alginate was pro-vided by EMCM (Nijmegen, the Netherlands) obtainedaccording to the ASTM F 2064-00 (reapproved 2006) stand-ard. The purification procedure was performed by EMCMaccording to a previously disclosed method22 and resultedinto an endotoxin level of less than 30 EU (endotoxinunits)/g. The CaP powders used to prepare the compositeswere HA (VPP132; AAP Biomaterials) and dicalcium phos-phate anhydrous (monetite, C7263; Sigma-Aldrich). Glycerolwas purchased from Sigma-Aldrich (G5516) and GDL waspurchased from Merck (s6094494).

Preparation of alginate=CaP compositesAlginate=CaP composites were prepared using dual syringesby mixing an aqueous alginate phase with a mineral phaseconsisting of dispersed CaP powders in glycerol with orwithout the presence of an acidifier (GDL; see Table I forchemical composition of experimental groups). The compo-sites were maximized with respect to CaP content to obtainthe highest amount of osteoconductive filler. The CaP con-tents of 12.5% (w=v; for HA) and 25% (w=v; for monetite)corresponded to the maximum amount of CaP powder thatstill allowed the formation of injectable formulations.

The alginate powder was dissolved in water to obtainan alginate concentration of 3.5% (w=v). The mineral-con-taining phase was obtained by mixing CaP (either HA ormonetite), glycerol (and the acidifier GDL in case of HA-con-taining composites). Composites containing monetite formedgels spontaneously, whereas composites containing HA didnot form gels unless GDL was added to stimulate therelease of calcium ions from HA powder particles.

The alginate phase and mineral phase were loaded intothe separate chambers of a dual syringe system for mixingand injection (MEDMIXVR , L-system, 2.5 mm chamber, mixingtip 25 mm) leading to extrusion of a homogeneous compos-ite [Fig. 1(A,B)].

Gelation and morphology of compositesThe viscoelastic properties of the composite gels were ana-lyzed using a rheometer (AR2000ex; TA Instrument)equipped with a flat steel-plate geometry (20 mm diameter)

2 ALVES CARDOSO ET AL. INJECTABLE ALGINATE–CALCIUM PHOSPHATE GELS FOR BONE REGENERATION

at 37�C in the presence of a water trap. Storage moduli (G0)and loss moduli (G0) were determined in oscillatory timesweep tests for 60 min at a gap distance of 500 mm by sub-jecting the samples (n 5 3) at a stress of 0.1 Pa and a fre-quency of 1 Hz. SEM (JEOL 6310) was performed to analyzethe morphology of the composites after freeze-drying for 24h and subsequent sputter coating with gold.

Calcium release from CaP precursorsCalcium release from CaP precursor powders was analyzedusing the orthocresolphtalein complexone (OCPC) assay.Briefly, 250 mg of HA=glycerol (weight ratio of 1:3) or mon-etite=glycerol (weight ratio of 1:3) were immersed in 0.5mL of milliQ for 10, 30, and 60 min. Pure, glycerol-freepowders of HA and monetite were used as control. To studythe effect of GDL on calcium release from the mineral phase,250 mg of HA=glycerol=GDL mixtures at a fixed HA contentof 25 wt % with GDL contents ranging from 0.5 to 2 wt %were also tested. The supernatants were incubated over-night in 1 mL 0.5 N acetic acid on a shaker table. For analy-sis, 300 lL work reagent was added to 10 lL sample orstandard in a 96-well plate. The plate was incubated for 10min at room temperature. The absorbance of each well wasmeasured on a microplate spectrophotometer at 570 nm.The standards (range: 0–100 mg=mL) were prepared using aCaCl2 stock solution. Data were obtained from triplicatesamples and measured in duplo.

Soaking of alginate=CaP composites in SBFSamples of the composites were injected in round molds (8mm diameter and 5 mm height) for 24 h at 37�C. Subse-quently, the prepared disks of the composites (n 5 3) wereimmersed in SBF for 21 days. The in vitro experiments werecarried out in conventional SBF with an ionic compositionalmost equal to human plasma.23 Ionic concentrations ofthis SBF are 142.0 mM Na1, 5.0 mM K1, 1.5 mM Mg21, 2.5mM Ca21, 103.0 mM Cl2, 4.2 mM HCO3

22, 1.0 mM HPO422,

and 0.5 mM SO422. Tris–HCl served as buffer to maintain a

constant pH value of 7.4. The solution was changed at days1, 3, 5, 7, 9, 11, 13, 15, 17, 19, and 21 and saved for analy-sis of the calcium content (using the OCPC assay asdescribed previously) and pH (Radiometer Copenhagen,PHM 210). The depletion of Ca was plotted cumulatively bymeasuring the difference between the Ca concentration in

the sample-free SBF control solutions and the SBF solutionin the presence of alginate=CaP composites.

Surgical procedureThe protocol for the animal experiment was approved bythe Animal Ethics Committee of the Radboud University Nij-megen Medical Centre (RUDEC 20010-147). A total of sixhealthy skeletally mature female New Zealand White rabbitswith an average weight of 5 kg were used as experimentalanimals. Surgery was performed under general inhalationanesthesia. Anesthesia was induced by an intravenous injec-tion of Hypnorm (0.315 mg=mL fentanyl citrate and

FIGURE 1. A, Dual-syringe dispenser and alginate=mineral phase com-

position; (B) extrusion of HA 1% GDL composite (12.5% hydroxyapa-

tite, 36.5% glycerol, 1% GDL, 1.75% alginate, and 48.25% water); (C)

scanning electron micrographs of freeze-dried sample of HA 1% GDL

composite (scale bar 10 lm). [Color figure can be viewed in the online

issue, which is available at wileyonlinelibrary.com.]

TABLE I. Chemical Composition of Composite Materials Tested

Experimental Group

Mineral Phase Alginate Phase

Calcium Phosphate(%, w=v)

Glycerol(%, w=v)

GDL(%, w=v)

Alginate(%, w=v)

H2O(%, w=v)

HA 0% GDL

Hydroxyapatite 12.5

37.5 0

1.75 48.25HA 0.25% GDL 37.25 0.25HA 0.5% GDL 37 0.5HA 1% GDL 36.5 125% monetite

Monetite25 25 0

12.5% monetite 12.5 37.5 0

ORIGINAL ARTICLE

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | MARCH 2013 VOL 00A, ISSUE 00 3

10 mg=mL fluanisone) and atropine, and maintained by amixture of nitrous oxide, isoflurane and oxygen through aconstant volume ventilator. To reduce the perioperativeinfection risk, the rabbits received antibiotic prophylaxis[BaytrilV

R

, 2.5% (Enrofloxacin), 10 mg=kg]. The animals wereimmobilized on their back and the hind limbs were shaved,washed and disinfected with povidone-iodine. After expo-sure of the distal femoral condyle from the medial side, a1.0-mm pilot hole was drilled. The hole was gradually wid-ened with drills of increasing size until a final defect size of4 mm in width and 6 mm in depth was reached. Low rota-tional drill speeds (max. 450 rpm) and constant physiologicsaline irrigation was used. After preparation, the defectswere thoroughly irrigated and packed with sterile cottongaze to stop bleeding. Surgery was performed in both legsof the rabbits and one defect was created in each condyle.After optimization of the formulation in the above-men-tioned in vitro studies, a formulation consisting of 12.5%HA, 36.5 glycerol, 1% GDL, 1.75% alginate, and 48.25%milliQ was selected for further in vivo testing and injectedinto the defect site (n 5 6). This composition was selectedin view of the crystallographic similarity of HA with themineral phase of bone, while its high stiffness was consid-ered beneficial for stimulation of cell differentiation into theosteogenic lineage.24 Pure alginate gels were not implantedsince these gels could not be extruded from the dual sy-ringe system as used in the current study without using dis-persed calcium sources, such as calcium sulfate or calcite,while the osteocompatibility of premade pure alginate gelswas proven previously.11,25 After injection of the injectablecomposite into the bone defect, soft tissues were closedlayer-by-layer using resorbable Vicryl sutures.

For the subcutaneous implants, the dorsum of the rab-bits was shaved, washed, and disinfected with iodine. Twolongitudinal incisions of about 15 mm were made paraverte-brally through the full thickness of the skin. Subsequently,lateral to the incisions a subcutaneous pocket was createdby blunt dissection with scissors. Composite disks (8 mmdiameter and 5 mm height) prepared before surgery (24 hcross-linking at 37�C, sterilization by autoclavation) wereinserted into the subcutaneous pockets. Implantation peri-ods were 6 and 12 weeks (n 5 6 for each time period), andthe rabbits were sacrificed by an overdose of NembutalV

R

atthe end of these implantation periods.

HistologyAfter harvesting the femoral condyles and removal of sur-rounding soft tissues, each condyle was divided into medialand lateral halves along their longitudinal axis. Subse-quently, only the medial half was analyzed and was fixed in4% formaldehyde for 2 days, dehydrated in a graded seriesof ethanol and embedded in methylmethacrylate. After poly-merization, at least three 10 lm sagittal cross-sections wereprepared using a sawing microtome technique. Sectionswere stained with methylene blue=basic fuchsin and exam-ined with a light microscope equipped with a digital camera(Axio Imager Microscope Z1; Carl Zeiss Micro imagingGmbH, G€ottingen, Germany).

The retrieved subcutaneous specimens were fixed in10% neutral-buffered formalin, dehydrated in a graded se-ries of ethanol, embedded in paraffin and cut on a micro-tome (RM 2165; Leica) in 6 lm-thick sections. The sectionswere subsequently mounted on glass slides and stainedwith hematoxylin and eosin (HE) to study the general tissueresponse as well as with Von Kossa reagent to visualizethe mineralized matrix after counterstaining the nuclei withNuclear Fast Red.

Statistical analysisAll data were analyzed using Student’s t-tests (GraphpadInstat) and expressed as mean 6 standard deviation. Avalue of p < 0.05 was accepted as statistically significant.

RESULTS

Morphology and viscoelastic propertiesof alginate=CaP compositesAll alginate=CaP composites were extruded as homogeneouspastes [Fig. 1(B)]. At microscale [Fig. 1(C)] individual CaPmicroparticles could not be discerned, indicating that themixing procedure in the dual syringe system was highly effi-cient. The viscoelastic properties of composites containingmonetite without addition of GDL are shown in Figure 2(A),which reveals that the storage modulus gradually increasedfor the entire duration of the time sweep (60 min), whereasthe loss modulus remained constant. The storage moduliexceeded the loss moduli after 3 versus 15 min for compo-sites containing 25% (w=v) versus 12.5% (w=v) monetite.These values indicated that gelation times (defined as the

FIGURE 2. A, Storage modulus and loss modulus of alginate=monetite

composites as a function of time; (B) storage modulus of alginate=HA

composites as a function of time. * for significant difference between

the values.


transition from liquid-like behavior [G0 < G0 or tan(d) > 1]to solid-like behavior [G0 > G0 or tan(d) < 1] decreasedwith increasing monetite content. The highest storage mod-uli (i.e., 500 Pa) were observed after 60 min for compositesof highest monetite content (25%), whereas compositescontaining less monetite (12.5%) exhibited a considerablylower final storage modulus of 38 Pa after 60 min.

Figure 2(B) shows the storage and loss moduli of HA-containing composites as a function of GDL content. Storagemoduli increased with time and increasing amount of GDLup to values of about 100 kPa after 60 min of time sweepand for the highest amount of GDL of 1% (w=v). GDL-freeformulations, on the contrary, did not gellify as evidencedby a constant storage modulus of about 80 Pa for the entireduration of the time sweep. All GDL-containing compositerevealed tan(d) values lower than 1 from the onset of thetime sweep (data not shown), indicating that these formula-tions were gels immediately after extrusion from the dualsyringe.

Calcium release from CaP precursorsRelease of calcium was higher from monetite powders com-pared to HA powders [Fig. 3(A)]. Glycerol increased therelease of calcium from monetite two- to threefold, but cal-cium release from HA was not stimulated by the presenceof glycerol. Figure 3(B) shows the effect of the addition ofGDL on the release of calcium from HA precursor powder.

Without any GDL present, the amount of calcium releasewas negligible. With increasing amount of GDL, however,the release of calcium increased strongly. The GDL-inducedrelease of calcium increased with time corresponding to thegradual acidification caused by the time-dependent hydroly-sis of GDL toward gluconic acid.

Soaking of alginate=CaP composites in SBFThe calcium concentrations of the SBF solution decreasedfor all tested formulations from day 0 until day 11, indicat-ing that the composites stimulated uptake of calcium frommetastable SBF solutions. The cumulative calcium depletionwas quantified using the OCPC assay as shown in Figure4(A). The amount of calcium uptake by alginate=CaP compo-sites increased with increasing amount of CaP phase (formonetite-containing formulations) and decreasing GDL con-tent. Formulations containing monetite stimulated calciumuptake to a lesser extent than HA-containing composites atall time points. After 11 days, however, all composites dis-played a decrease in calcium depletion corresponding torelease of calcium into SBF rather than consumption of cal-cium from SBF for cross-linking and=or precipitation of CaP.

Figure 4(B) shows the effect of HA-containing compo-sites on the pH of the SBF solution. No pH changes wereobserved in SBF solutions containing composites withoutGDL or a low amount of GDL (0.25%) after 21 days of soak-ing. The pH values of SBF solutions containing compositesof higher GDL content (0.5 or 1%), however, were about 5

FIGURE 3. A, Release of calcium from monetite powder, monetite=gly-

cerol mixture (1:3 by weight), HA powder, and HA=glycerol mixture

(1:3 by weight) as a function of time; (B) release of calcium from HA=-

glycerol mixtures containing four different percentages of GDL as a

function of time. When not otherwise specified (by the symbol #) the

values for each compound tested are significantly different between

each other at the time points shown.

FIGURE 4. A, Cumulative calcium depletion from SBF solution as a

function of time;(B) pH change of SBF solution as a function of time.

ORIGINAL ARTICLE


after 1 day of soaking, after which the pH graduallyincreased to neutral pH values between 5 and 11 days ofsoaking. After 11 days of soaking, the pH value of compo-sites containing 0.5 and 1% GDL decreased again to about6.5 increasing again to pH neutral until day 21.

In vivo implantation studyClinical observations. All 6 rabbits recovered from the sur-gical intervention without wound complications and pre-sented good health throughout the entire duration of theexperiment. A total of twelve bone implants were harvested(six implants after 6 week implantation and six implants af-ter 12 weeks implantation), whereas five subcutaneousimplants could be harvested after 6 weeks and only 3 after12 weeks due to loss of integrity of irretrievable samples.At retrieval, no visual signs of inflammatory or adverse tis-sue reaction were observed.

Descriptive light microscopy. Subcutaneous implantation.Representative histological images of alginate=HA compo-sites after subcutaneous implantation are shown in Figure 5(after 6 weeks) and Figure 6 (after 12 weeks). After 6weeks of implantation, the composites had already disinte-grated to a large extent and extensive soft tissue infiltrationinside the composite disk was observed in addition to thepresence of large numbers of spherical structures of up to100 mm in diameter [Fig. 5(A,B)], which were identified asmineralized, phosphate-containing particles by Von Kossastaining [Fig. 5(C,D)]. After 12 weeks of implantation, theexplants displayed very low mechanical integrity and HEstaining confirmed that the composite disks were almostcompletely degraded and replaced by newly formed soft tis-sue [Fig. 6(A,B)]. Compared to 6 weeks of implantation, thenumber of mineralized spherical structures as observed af-ter 12 weeks of implantation had decreased considerably[Fig. 6(C,D)].

Osseous implantation. Representative histological imagesof the bone tissue response to injectable alginate=HA com-posites are shown in Figure 7(A,B; after 6 weeks and C,D;after 12 weeks), respectively. After both time periods, theoriginal defects were clearly visible in all the samples ana-lyzed although the composites were already degraded exten-sively [Fig. 7(A,C)]. The alginate=HA composites had losttheir structural integrity already at 6 weeks of implantation,and extensive bone ingrowth was observed throughout thedegraded material. The newly formed bone was always inclose contact with the remnants of the composite material(stained black) without the presence of intervening fibroustissue [Fig. 7(B,D)]. No differences were observed betweenboth implantation time periods in terms of composite degra-dation or new bone formation.

For defects created at or near the growth plate, a differ-ent tissue response was observed occasionally (1 of 6implants after 6 weeks and 1 of 6 implants after 12 weeks)as shown in Figure 8(A,B; after 6 weeks and C,D; after 12weeks). This tissue response was characterized by the pres-ence of dense fibrous tissue of high cell density inside thecenter of the original defect [Fig. 8(A,C)]. This soft tissuewas surrounded by a zone of acellular material remnants(stained black). A sharp interface between the fibrous cen-ter and surrounding osseous tissue was clearly recognizable[Fig. 8(B,D)]. Also for this type of tissue response, no differ-ences were observed between both implantation timeperiods.

DISCUSSION

The objective of the current study was to develop a compos-ite based on purified alginate and CaP with tunable inject-ability, CaP content and viscoelasticity to combine thebiocompatibility of alginate carrier matrices with the osteo-conductivity of CaP particles to stimulate bone regeneration.To this end, formulations have been developed based on the

FIGURE 5. A and B, Representative histological sections of the bone response to alginate=HA composites after 6 weeks of implantation at low

(A, scale bare 1000 lm) and high (B, scale bar 200 lm) magnification. C and D, Representative histological sections of the alginate=HA compos-

ite after 12 weeks of implantation at low (C, scale bare 1000 lm) and high (D, scale bar 200 lm) magnification. [Color figure can be viewed in

the online issue, which is available at wileyonlinelibrary.com.]


mixture of alginate as hydrogel matrix, two types of CaPphases as osteoconductive dispersed filler (i.e., monetite orHA), glycerol as plasticizer and temporary sequestrant andGDL as acidifier. HA was selected as CaP phase owing to itschemical similarity to the mineral phase of bone, whereasmonetite was selected because of its higher solubility thanHA, which could be beneficial for the cross-linking of algi-nate without using acidifiers.

The resulting composites were maximized with respect toCaP content to obtain the highest amount of osteoconductivefiller. The viscoelastic and physicochemical properties of theprecursor compounds and composites were analyzed usingrheometry, elemental analysis (for calcium release anduptake), acidity (by measuring pH in SBF), general biocom-patibility (subcutaneous implantation in rabbits) and osteo-compatibility (implantation in femoral condyle bone defectsof rabbits).

The composites were maximized with respect to CaPcontent to obtain the highest possible amount of osteocon-ductive filler without compromising the injectability fromconventional dual syringe systems. All composites exhibiteda homogeneous structure without the presence of large CaPaggregates as evidenced by SEM, indicating that organic andinorganic components were mixed efficiently in the dual sy-ringe system at this high CaP content without clustering oraggregation of CaP particles. Gelation times of monetite-con-taining composites could be controlled between 3 and 15minutes by varying the amount of monetite from 12.5 to25%. Apparently, the amount of calcium released from mon-etite powder was sufficient to allow for cross-linking of algi-nate macromolecules, whereas insufficient calcium ionswere released from HA powders to facilitate cross-linking of

alginate. These observations were supported by the quantifi-cation of calcium release from pure CaP powders as well asCaP-glycerol mixtures. Calcium release from monetite pow-ders was considerably higher than release from HA, whichcan be understood from the fact that monetite has a muchhigher solubility product at 25�C than HA (1.8 3 1027 vs.6.6 3 102137 g=mL).26 In contrast to release of calciumfrom HA, calcium release from monetite powders increasedwith time as well as addition of glycerol. In addition to itsplasticizing function, glycerol also acted as a sequestrant ofcalcium after release from monetite, thereby delaying gela-tion until the point of gel extrusion from the dual syringeand allowing for homogeneous gel formulations.19 Glyceroldid not increase the amount of calcium release from HA dueto the very poor solubility of HA, but acidification usingGDL resulted into tunable release of calcium from HA up tovalues twice as high (for formulations containing 1% ofGDL) than release of calcium from monetite=glycerol mix-tures. It can be concluded that the combination of the tem-porary sequestration of calcium by glycerol and gradual,time-dependent acidification induced by GDL resulted intogradual release of calcium after extrusion from the dualsyringe, thereby preventing premature cross-linking thatwould compromise the injectability of the composites.

Storage moduli of HA-containing composites increasedfrom less than 100 Pa without GDL to about 100 kPa forhighly elastic and cohesive composites containing 1% GDL.Composites made of either monetite or HA and GDL dis-played a continuously increasing storage modulus as a func-tion of time, indicating that the release of calcium continuedafter gel extrusion. Due to the presence of the acidifier GDLin composite gels containing HA, the amount of calcium

FIGURE 6. A and B, Representative histological sections of occasional soft tissue formation inside alginate=HA composites after 6 weeks of im-

plantation at low (A, scale bare 1000 lm) and high (B, scale bar 200 lm) magnification. C and D, Representative histological sections of occa-

sional soft tissue formation inside alginate=HA composites after 12 weeks of implantation at low (C, scale bare 1000 lm) and high (D, scale bar

200 lm) magnification. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

ORIGINAL ARTICLE


released from HA was higher than the calcium releasedfrom monetite (see Fig. 3), resulting into higher cross-link-ing density and storage moduli than monetite-containingformulations. Furthermore, it was observed that HA=glycerolsuspensions were more viscous than mixtures of glyceroland monetite, most likely caused by differences in particlecharacteristics, such as size and shape, thereby contributingto the higher storage moduli as observed for the HA-con-taining composites.

From a clinical point of view, the immediate gelation ofHA-containing gels and the short gelation times of mone-tite-containing gels of less than 15 minutes can be consid-ered as a significant improvement over CaCO3–GDL–alginate gel systems as reported by Kuo et al.,21 sincethese formulations typically gellified at considerably slowerrates of up to several hours. Moreover, the use of CaPs assource for calcium cross-linking of alginate is advantageouscompared to systems that use calcium carbonate or cal-cium sulfate as source since synthetic CaPs closely resem-ble the apatitic mineral phase in bone and teeth. Similar tothe majority of mechanically weak hydrogels or brittle CaPceramics, it should be stressed that the hydrogel

composites developed herein are not suitable for load-bearing conditions.

To test the efficacy of HA- and monetite-containing al-ginate gels, their capacity to induce mineralization wastested in vitro. Since the composites disintegrated within 3days after incubation in calcium-free phosphate-bufferedsaline (PBS, data not shown), calcium-containing SBF wasselected as incubation medium, thereby allowing for moni-toring of calcium concentrations in the supernatant.Although calcium can also consumed upon ionic cross-link-ing, we observed using rheometry that viscoelastic proper-ties (G0 and G0) of the various gel composites did notchange after 24 h (data not shown), which seems to sug-gest that that observed calcium depletion in SBF wasmainly caused by precipitation of CaP onto the dispersedmineral phase in the alginate matrices rather than by ioniccross-linking.

Up to 11 days of soaking, all formulations consumed cal-cium from SBF. After 11 days, however, all compositesstarted to release calcium, which was most likely caused bythe replacement of divalent calcium by monovalent sodiumcations. Calcium uptake by gels containing monetite

FIGURE 7. A and B, Representative histological sections after hematoxylin=eosin staining of subcutaneous samples implanted for 6 weeks at

low (A, scale bare 1000 lm) and high (B, scale bar 500 lm) magnification. C and D, Representative histological sections after Von Kossa staining

of subcutaneous samples implanted for 6 weeks at low (C, scale bare 1000 lm) and high (D, scale bar 500 lm) magnification. [Color figure can

be viewed in the online issue, which is available at wileyonlinelibrary.com.]


increased with increasing monetite content due to the avail-ability of a higher number of nucleation points for homolo-gous CaP crystal growth, whereas the addition of GDLdecreased the tendency of HA-containing gels to mineralize.This latter phenomenon can be explained by the acidifica-tion caused by hydrolysis of GDL, which reduces localsupersaturation and precipitation of CaP. Formulations con-taining 0.5% or 1.0% of GDL displayed a drop of about 2pH units upon incubation in (buffered) SBF, stressing thatthe acid production by GDL exceeded the buffering capacityof SBF.

To test the effect of the above-mentioned acid produc-tion on the biological behavior of the composites, an in vivoimplantation study was performed in both rabbit femoralcondyles (orthotopic) and subcutaneous tissue (ectopic). Aformulation containing 25% HA and 1% GDL was selectedfor further in vivo studies since this gel displayed the opti-mal combination of injectability and viscoelastic properties(i.e., highest storage modulus). Since signs characteristic forinflammation were not observed in any of the explants, it

can be concluded that the acidification as observed in vitrodid not cause any inflammatory response in vivo, most likelydue to the higher perfusion and related buffer capacityunder in vivo conditions. Generally, the bone response tothe HA-containing hydrogels was characterized by extensivedegradation of the organic component of the hydrogel andabundant formation of new bone tissue. This newly formedbone was always observed to be in direct contact with rem-nants of remaining HA particles, confirming the osteocom-patible nature of the dispersed CaP phase. Since nodifferences were observed between the bone tissueresponse after 6 versus 12 weeks, it can be concluded thatthe hydrogel matrix was already degraded after relativelyshort time periods of less than 6 weeks, leaving behindpoorly degradable osteoconductive HA particles. Theseobservations were in line with histological observations ofsubcutaneous explants, which confirmed that prematuredegradation of the organic alginate–glycerol matrix phaseresulted into fast disintegration of the composites while lessdegradable CaP particles were retained. These results

FIGURE 8. A and B, Representative histological sections after hematoxylin=eosin staining of subcutaneous samples implanted for 12 weeks at

low (A, scale bare 1000 lm) and high (B, scale bar 500 lm) magnification. C and D, Representative histological sections after Von Kossa staining

of subcutaneous samples implanted for 12 weeks at low (C, scale bare 1000 lm) and high (D, scale bar 500 lm) magnification. [Color figure can

be viewed in the online issue, which is available at wileyonlinelibrary.com.]

ORIGINAL ARTICLE


suggest that the cross-linking density of the compositeswere sufficient for gelation upon extrusion from dualsyringes but insufficient for prolonged stability in vivo.Occasionally, the premature disintegration of the hydrogelcomposites might have caused soft tissue infiltration insidebone defects. Apparently, the dynamic and highly perfusedin vivo conditions induced considerable exchange of divalentcalcium ions with monovalent sodium ions, resulting intodisintegration of the gels. Based on the calcium concentra-tions as measured in the current study (ranging from 50 to600 mg=mL), we speculate that the amount of calcium ionsavailable for cross-linking of alginate was considerablylower than calcium concentrations as used for preparationof premade alginate scaffolds (ranging between 1 and20 mg=mL).27–30 Therefore, the current study warrantsfuture research on increasing the amount of calcium releasefrom CaP precursor powders using the proposed CaP-GDLsystem without provoking inflammatory responses due toexcessive acidification or compromising the extrudability ofthe hydrogel composites.

CONCLUSIONS

Novel injectable composites for bone regeneration havebeen developed based on the combination of purified algi-nate of high molecular weight as matrix phase, crystallineCaP powders as dispersed mineral phase as well as sourceof calcium for cross-linking, GDL as acidifier and glycerol asboth plasticizer and temporary sequestrant. The gelationkinetics of the resulting composites could be controlledfrom seconds to tens of minutes by varying the solubility ofthe CaP phase (HA vs. monetite) or amount of GDL. All com-posites mineralized extensively in SBF for up to 11 dayswhereafter the composites started to disintegrate. In vivo,the composites also disintegrated upon implantation in sub-cutaneous or bone tissue, leaving behind less degradablebut osteoconductive CaP particles. Although the compositesneed to be optimized with respect to the available amountof calcium for cross-linking of alginate, the beneficial boneresponse as observed in the current in vivo render thesegels promising for minimally invasive application as bone-filling material.

ACKNOWLEDGMENTS

The research leading to these results has received fundingfrom the European Community’s Seventh Framework Pro-gramme (MultiTERM, grant agreement No. 238551).

References1. United States Bone and Joint Initiative: The Burden of Musculo-

skeletal Diseases in the United States, 2nd ed. Rosemont, IL:

American Academy of Orthopaedic Surgeons; 2011. p 219–222.

2. Kartsogiannis V, Ng KW. Cell lines and primary cell cultures in

the study of bone cell biology. Mol Cell Endocrinol 2004;228:79–

102.

3. Barrere F, Mahmood TA, de Groot K, van Blitterswijka CA.

Advanced biomaterials for skeletal tissue regeneration: Instructive

and smart functions. Mater Sci Eng R 2008;59(16):38–71.

4. Lalan S, Pomerantseva I, Vacanti JP. Tissue engineering and its

potential impact on surgery. World J Surg 2001;25(11):1458–1466.

5. Dorozhkin S. Calcium orthophosphate-based biocomposites and

hybrid biomaterials. J Mater Sci 2009;44(9):2343–2387.

6. Navarro M, Michiardi A, Castano O, Planell JA. Biomaterials in

orthopaedics. J R Soc Interface 2008;5(27):1137–1158.

7. Tran N, Webster TJ. Nanotechnology for bone materials. Wiley

Interdiscip Rev Nanomed Nanobiotechnol 2009;1(3):336–351.

8. Brown WE, Chow LC. Dental restorative cement pastes. US patent

number 4,518,430; 1985.

9. Dorozhkin SV. Nanosized and nanocrystalline calcium orthophos-

phates. Acta Biomater 2010;6(3):715–734.

10. Mano JF, Silva GA, Azevedo HS, Malafaya PB, Sousa RA, Silva

SS, Boesel LF, Oliveira JM, Santos TC, Marques AP, Neves NM,

Reis RL. Natural origin biodegradable systems in tissue engineer-

ing and regenerative medicine: Present status and some moving

trends. J R Soc Interface 2007; 4(17):999–1030.

11. Augst AD, Kong HJ, Mooney DJ. Alginate hydrogels as biomateri-

als. Macromol Biosci 2006;6(8):623–633.

12. Gkioni K, Leeuwenburgh SC, Douglas TE, Mikos AG, Jansen JA.

Mineralization of hydrogels for bone regeneration. Tissue Eng

Part B Rev 2010;16(6):577–585.

13. Klein C, van der Lubbe H, de Groot K. A plastic composite of algi-

nate with calcium phosphate granulate as implant material: An in

vivo study. Biomaterials 1987;8(4): 308–310.

14. Rowley JA, Madlambayan G, Mooney DJ. Alginate hydrogels as

synthetic extracellular matrix materials. Biomaterials 1999;20(1):

45–53.

15. Zhang SM, Cui FZ, Liao SS, Zhu Y, Han L. Synthesis and biocom-

patibility of porous nanohydroxyapatite=collagen=alginate com-

posite. J Mater Sci Mater Med 2003;14(7):641–645.

16. Lin HR, Yeh YJ. Porous alginate=hydroxyapatite composite scaf-

folds for bone tissue engineering: Preparation, characterization,

and in vitro studies. J Biomed Mater Res B Appl Biomater

2004;71(1):52–65.

17. Shiraishi N, Anada T, Honda Y, Masuda T, Sasaki K, Suzuki O.

Preparation and characterization of porous alginate scaffolds con-

taining various amounts of octacalcium phosphate (OCP) crystals.

J Mater Sci Mater Med 2010;21(3):907–914.

18. Tan R, Niu X, Gan S, Feng Q. Preparation and characterization of

an injectable composite. J Mater Sci Mater Med 2009;20(6):1245–

1253.

19. Bender JCME, Mann MGA, Kamps M, Pellikaan HC. In situ gelling

alginate systems. 2010. WO 2010=117266.

20. Flink JM, Johansen A. A novel method for immobilization of

yeast cells in alginate gels of various shapes by internal liberation

of Ca-ions. Biotechnol Lett 1985;7(10):765–768.

21. Kuo CK, Ma PX. Ionically crosslinked alginate hydrogels as scaf-folds for tissue engineering: Part 1. Structure, gelation rate andmechanical properties. Biomaterials 2001;22(6):511–521.

22. Bender JCME, Vermeulen PS. Method for removing impurities

from biopolymer material. 2009. WO 2009=154440.

23. Cho SB, Nakanishi K, Kokubo T, Soga N, Ohtsuki C, Nakamura T.

Dependence of apatite formation on silica gel on its structure:

Effect of heat treatment. J Am Ceram Soc 1995;78:1769–1974.

24. Discher DE, Janmey P, Wang YL. Tissue cells feel and respond to

the stiffness of their substrate. Science 2005;310(5751):1139–1143.

25. Stevens MM, Marini RP, Schaefer D, Aronson J, Langer R, ShastriVP. In vivo engineering of organs: The bone bioreactor. Proc NatlAcad Sci USA 2005;102(32):11450–11455.

26. Elliott JC. Structure and Chemistry of the Apatites and Other Cal-cium Orthophosphates. Amsterdam: Elsevier; 1994. p 1–62.

27. Stevens MM, Qanadilo HF, Langer R, Prasad Shastri V. A rapid-

curing alginate gel system: Utility in periosteum-derived cartilage

tissue engineering. Biomaterials 2004;25(5):887–894.

28. Lin HR, Yeh YJ. Porous alginate=hydroxyapatite composite scaf-

folds for bone tissue engineering: Preparation, characterization,

and in vitro studies. J Biomed Mater Res B Appl Biomater

2004;71(1):52–65.

29. Turco G, Marsich E, Bellomo F, Semeraro S, Donati I, Brun F, Grand-

olfo M, Accardo A, Paoletti S. Alginate=hydroxyapatite biocompo-

site for bone ingrowth: A trabecular structure with high and

isotropic connectivity. Biomacromolecules 2009;10(6):1575–1583.

30. Lakshmi SN, Cato TL. Polymers as biomaterials for tissue engi-

neering and controlled drug delivery. Adv Biochem Eng Biotech-

nol 2006(102):47–90.


gelation and biocompatibility of injectable alginate-calcium phosphate gels for bone regeneration

Documents