TI

PC

a

ARRAA

KGCCM

1

tggDaa

vsaisc

vT

h0

International Journal of Biological Macromolecules 70 (2014) 10–19

Contents lists available at ScienceDirect

International Journal of Biological Macromolecules

j ourna l ho me pa g e: www.elsev ier .com/ locate / i jb iomac

ailoring the properties of gelatin films for drug delivery applications:nfluence of the chemical cross-linking method

. Coimbra ∗, M.H. Gil, M. FigueiredoIEPQPF, Chemical Engineering Department, University of Coimbra, Polo II, 3030-290 Coimbra, Portugal

r t i c l e i n f o

rticle history:eceived 5 February 2014eceived in revised form 5 June 2014ccepted 10 June 2014vailable online 24 June 2014

eywords:elatin filmshemical cross-linkingarbodiimideethacrylic anhydride

a b s t r a c t

Two types of chemically cross-linked gelatin films were prepared and characterized. The first typeof films was cross-linked with 1-ethyl-3-(3-dimethyl aminopropyl)carbodiimide hydrochloride (EDC)under heterogeneous conditions and are named Gel-E. In the second type of films, gelatin was previouslyfunctionalized with methacrylamide side groups by the reaction with methacrylic anhydride and forthat is named Gel-MA. The modified gelatin was subsequently cross-linked by a photoinitiated radicalpolymerization.

These films were characterized relatively to their degree of cross-linking, buffer uptake capacity, resis-tance to hydrolytic and proteolytic degradation, and mechanical and thermal properties. Results showthat the employed cross-linking method, together with the degree cross-linking, dictate the final prop-erties of the films. Gel-E films have significant lower buffer uptake capacities and higher resistance to

collagenase digestion when compared to Gel-MA films. Additionally, Gel-E films exhibit higher values ofstress at break and lower strains at break. Moreover, the films properties could be modified by varyingthe extent of the chemical cross-linking, which in turn could be controlled by varying the concentrationof EDC, for the first type of films (Gel-E), or by using gelatins with different degrees of functionalization,in the case of the second type of films (Gel-MA).

© 2014 Elsevier B.V. All rights reserved.

. Introduction

Gelatin is a water-soluble proteinaceous material obtained fromhe denaturation and partial hydrolysis of collagen [1,2]. Generally,elatin is obtained by an acidic or alkaline treatment of colla-en from animal byproducts such as cattle bones and pork skins.ue to the variety of collagen sources and production processes,

large diversity of gelatins, with variable chemical compositionsnd physical properties, can be produced [2].

As known, gelatins can be dissolved in water and form lowiscosity solutions at temperature above ca. 35 ◦C, but undergo aol–gel transition when cooled, forming physical gels. The mech-nism of this thermoreversible gelation has been extensivelynvestigated [3–6]. The gelation involves a conformational tran-

ition of part of the gelatin chains from a random coil state toollagen-like triple-helix structures.

∗ Corresponding author at: CIEPQPF, Department of Chemical Engineering, Uni-ersity of Coimbra, Rua Sílvio Lima, Polo II, 3030-790, Coimbra, Portugal.el.: +351 239798700; fax: +351 239798703.

E-mail address: patricia@eq.uc.pt (P. Coimbra).

ttp://dx.doi.org/10.1016/j.ijbiomac.2014.06.021141-8130/© 2014 Elsevier B.V. All rights reserved.

Gelatin has been used for a long time in the food and pharmaceu-tical industry in a large number of applications [2]. Besides beingbiodegradable and biocompatible, gelatin is nonimmunogeneticand has hemostatic properties [1]. Due to these favorable prop-erties, gelatin is also used in several biomedical applications, suchas a hemostatic sponges for surgical purposes [7], wound dressingmaterials [8,9], and sealants for vascular prostheses [10]. Further,gelatin has already been investigated as a material for the construc-tion of scaffolds for tissue engineering [11–16], vehicles for thecontrolled delivery of bioactive macromolecules, such as peptides,proteins, and oligo- and polynucleotides [17–19], and as carrier fortransplanted fragile tissues, like retinal sheets [20].

Since gelatin gels formed by simple thermal induced gelationreadily dissolve at physiological temperatures, the gelatin basedmaterials used in many of the described biomedical applicationsare usually submitted to an irreversible cross-linking treatment, inorder to enhance their thermal and mechanical stability, as well asto retard their rate of degradation in vivo [17]. Permanent cross-linkis achieved by the introduction of intermolecular covalent bonds

between gelatin chains. This can be accomplished either by physicalmethods like, for example, the treatment with heat or irradiation[21,22], or by chemical methods, by the reaction of gelatin withvarious types of chemical agents. Dialdehydes, like formaldehyde

f Biolo

oahim(Oiopib[t

wmdit[

lptcwto[mtcinaf

clwGgp

clsmtggtar

2

2

(t(m

of 0.15% relatively to the Gel-MA weight. The warm solution waspoured into polystyrene petri dishes (diameter = 8.5 cm), left to cool

P. Coimbra et al. / International Journal o

r glutaraldehyde, are recognized as very efficient cross-linkinggents of gelatin and collagen [21,23–25]. Unfortunately, their useas been frequently associated with problems such as cytoxic-

ty and calcification [26–28], which prompt the investigation ofore biocompatible cross-linkers, like the water soluble 1-ethyl-3-

3-dimethyl aminopropyl)carbodiimide (EDC) [7,10,18–20,29,30].ther synthetic chemical agents already investigated include var-

ous diisocyanates and diepoxides [24,31]. Alternatively, a varietyf non-toxic natural occurring molecules, like genipin [32–34] andhenolic compounds such as tannic acid [35] and nordihydrogua-

aretic acid [36], have been used. Gelatin can also be cross-linkedy the use of naturally occurring enzymes, like transglutaminase11,25,37,38]. In this case, the enzymes work by catalyzing reac-ions that create covalent bonds between gelatin chains.

Other approach involves the pre-functionalization of gelatinith pendant methacrylate side groups, or other photosensitiveoieties, and the subsequent chemical cross-linking by light irra-

iation in presence of an appropriate photoinitiator. This methods of particular interest for biomedical applications, since it permitshe 3D encapsulation of viable cells simultaneously to gel formation14,15], or even in situ gel formation [39].

Independently of the method chosen to prepare chemical cross-inked gelatins, the phenomenon of physical gelation is alwaysresent, and will happen, in a greater or lesser extent, as soon as theemperature is sufficiently low. Thus, the final network structure ofhemical cross-linked gelatins and, consequently, their properties,ill be determined essentially by the way both of these processes

ake place, i.e., if they occur at the same time or one after thether. This particular aspect has been investigated by some authors38,40–42], that used rheological and optical rotation measure-

ents to follow the formation of gelatin networks under differenthermal protocols, in which physical cross-linking and chemicalross-linking occurred simultaneously or sequentially. These stud-es demonstrated that the viscoelastic properties of the formedetworks are highly dependent upon the sequence of formationnd on the relative amount of triple helices and covalent bondsormed.

In the present work, we prepared and characterized two types ofhemical cross-linked gelatin films, with several degrees of cross-inking. The first type of films, designated hereinafter as Gel-E,

ere cross-linked with EDC whereas in the second type of films,el-MA, gelatin was first functionalized with methacrylamide sideroups and subsequently cross-linked by a photo-initiated radicalolymerization of the added moieties.

The objective of this work is to gain insight on how differenthemical cross-linking methods, thermal protocols, and cross-inking levels can be used to tailor the properties of gelatin films,uch as the buffer uptake capacity, enzymatic degradation, andechanical and thermal properties. We hope that the findings of

his work will serve as basis for the achievement of our ultimateoal, which is the development of an implantable drug releaseelatin film. More specifically, we intend to develop a gelatin filmhat can release, after being implanted in a body site subject to

surgical tumor resection, a chemotherapeutic agent, in order toeduce the probability of tumor recurrence.

. Materials and methods

.1. Materials

Gelatin type A (from porcine skin, 300 bloom), collagenase

from Clostridium histolyticum, EC 3.4.24.3, 245 CDU/mg), 2,4,6-rinitrobenzene sulfonic acid (TNBS) solution (5% w/v), 1-ethyl-3-3-dimethyl aminopropyl) carbodiimide hydrochloride (EDC), and

ethacrylic anhydride (MAA), were purchase from Sigma–Aldrich.

gical Macromolecules 70 (2014) 10–19 11

N-hydroxysuccinimide (NHS) was obtained from Acros Organ-ics and the photoinitiator 1-[4-(2-hydroxyethoxy)phenyl]-2-hydroxy-2-methyl-1-propan-1-one (Irgacure®2959) was a kindgift from Ciba Speciality Chemicals. All the other chemicals wereof reagent grade and were used as received without further purifi-cation.

2.2. Preparation of chemical cross-linked Gel-E films

These films were produced following the procedure reportedby Kuijpers et al. [18,29]. A gelatin solution of 8% (w/v) wasprepared by dissolving gelatin in deionized water at 50 ◦C. Afterbeing sonicated, in order to remove air bubbles, 20 ml solutionwere poured into polystyrene petri dishes (diameter = 8.5 cm), andleft to air-dry at room temperature (around 25 ◦C). The formedfilms, with a thickness of 156 ± 10 �m, were then cut into squaresof 1 cm width. After being dried in vacuum for 1 day, sampleswere cross-linked with variable amounts of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC). For each reaction,25 gelatin squares, with a total weight of approximately 625 mg(average weight of 25 mg each), were immersed in 25 ml of 2-morpholinoethanesulfonic acid (MES) buffer (pH 5.5, 0.1 M) andcross-linked with the addition of EDC and N-hydroxysuccinimide(NHS). The amount of EDC was calculated in order that the molarratio of EDC to the carboxylic acid groups of gelatin varied from0.25 to 4.0, assuming that there are 0.080 mol of carboxyl groupsper 100 g of gelatin type A [18,29]. The molar ratio of NHS to EDCwas kept constant at 0.2. The reaction was allowed to proceed for5 h at 4 ◦C and then it was quenched by submerging the samples ina solution of 0.1 M disodium hydrogenphosphate and 2 M sodiumchloride (pH ≈ 8.7) for 20 min. Finally, samples were repeatedlywashed with deionized water and dried in a vacuum oven at ambi-ent temperature.

2.3. Preparation of photo-cross-linked Gel-MA films

2.3.1. Synthesis of Gel-MAGelatin methacrylamide (Gel-MA) was prepared by the reaction

of methacrylic anhydride (MMA) with the primary amino groups ofgelatin, according to the procedure originally reported by Van denBulcke et al. [43]. In order to prepare Gel-MA with various degrees ofsubstitution, the amount of MAA used was varied between 0.13 mland 8.4 ml, in order that achieve molar ratios of MMA to free aminogroups of 0.25, 0.5, 1.0, 2.0, 4.0 and 16.0. In the calculations itwas assumed that there are 0.035 mol of free amino groups per100 g of gelatin [43]. For the reaction, 10 g of gelatin were dis-solved in 100 ml of phosphate buffer saline (PBS, pH 7.4) at 50 ◦C.MMA was then added under vigorously magnetic stirring condi-tions and left to react for 1 h at 50 ◦C. The reaction was stopped andthe mixture was poured in to dialysis bags and dialyzed for 4 daysagainst distilled water at room temperature. The reaction prod-ucts were then frozen with liquid nitrogen and freeze-dried during3 days.

2.3.2. Photo-cross-linking of Gel-MA filmsFor the preparation of the films, 1.6 g of freeze-dried Gel-MA was

dissolved in 20 ml distilled water at 50 ◦C. The water soluble pho-toinitiator Irgacure®2959 was added to the solution in the amount

during 5–10 min, and then irradiated with a long-wave ultraviolet(LWUV) light (10 mW/cm2) for 10 min. The formed films were thenleft to dry at room temperature. After complete drying, films witha thickness of 190 ± 15 �m were formed.

1 f Biolo

2

tolGoricb2

rtwpsatuttpat1cce

D

dg

bs

2

irobf

B

2

wmt

wCCw

2 P. Coimbra et al. / International Journal o

.4. Determination of the degree of cross-linking and substitution

The calculation of the degree of cross-linking in the gelatin filmsreated with EDC was based on the difference between the numberf free �-amino groups present in the films before and after cross-inking. In a similar way, the degree of substitution of the preparedel-MA was estimated based on the difference between the amountf �-amino groups in gelatin before and after the modificationeaction. The quantification of the free �-amino groups (presentn lysine and hydroxylysine residues) was carried out using theolorimetric assay developed by Bubnis and Ofner [44], which isased in the reaction of the primary amino groups with the reagent,4,6-trinitrobenzene sulfonic acid (TNBS).

Before the assays, samples were placed in a vacuum oven atoom temperature during 1 week in order to remove all the mois-ure. In the case of the Gel-E films, small pieces (2–4 mg, weightedith a precision of ±0.01 mg) were immersed in a mixture com-osed of 1 ml distilled water, 1 ml sodium hydrogencarbonateolution (4% w/v, pH ≈ 8.5), and 1 ml TNBS solution (0.5%, w/v),nd left to react at 40 ◦C during 3 h. The reaction solution washen acidified by the addition of 3 ml of HCl 6 M and left at 80 ◦Cntil completely dissolution of the samples (≈1 h). After coolingo room temperature, 4 ml of deionized water was added andhe absorbance, at 345 nm, was measured against a blank pre-ared exactly the same way, but without any gelatin sample. Thebsorbance was measured using a JASCO 550 UV/vis spectropho-ometer. Gelatin solutions, with concentrations between 0.5 and.5 mg/ml, treated in the same way as samples, were used toonstruct a standard curve. All experiences were done in quadrupli-ate. The degree of cross-linking was determined by the followingxpression:

egree of cross-linking (%) =(

1 − Ac

Ao

)× 100

In this formula, Ac is the absorbance of the solution with theissolved cross-linked gelatin sample and Ao the absorbance of aelatin solution with the same concentration.

The degree of substitution of Gel-MA derivatives was estimatedy a similar procedure, but without the addition of HCl and theubsequence heat treatment.

.5. Films buffer uptake capacity

Small dried samples, accurately weighted (Wd), were immersedn PBS (pH 7.4) at 37 ◦C during 24 h. The swollen samples were thenetrieved from the buffer solution, the excess of water was wipedff with a filter paper, and the wet weight (W) was registered. Theuffer uptake capacity (gPBS/gfilm) was calculated according to theollowing expression:

uffer uptake capacity = (W − Wd)Wd

.6. Films proteolytic and hydrolytic degradation behavior

The films resistance to hydrolytic and enzymatic degradationas evaluated by incubating samples in two different aqueousediums and then monitoring their weight loss as a function of

ime.To evaluate the resistance against proteolytic enzymes, the films

ere incubated in TES buffer (0.05 M, pH 7.4), supplemented witha2+ (5 mM) and with 0.1 mg/ml (24.5 U/ml) of collagenase from. histolyticum. The dry samples, accurately weighted (≈25 mg),ere covered with 1 ml of the aforementioned medium, freshly

gical Macromolecules 70 (2014) 10–19

prepared, and left to rest in an oven at 37 ◦C. The enzyme mediumwas changed every 24 h. At predetermined intervals, samples wereremoved, washed with deionized water, and dried under vacuumat room temperature for 2 days. The samples were then weightedand the percentage of remaining weight determined according tothe expression:

Remaining weight (%) = Wt

Wo× 100

where Wo is the original weight and Wt is weight of the sample atthe degradation time t.

For assessing the films stability in an aqueous medium withoutproteolytic enzymes, the samples were immersed in a PBS solu-tion (pH 7.4) with 0.25% of sodium azide, at 37 ◦C. The weight lossalong time was determined following the same procedure reportedabove.

2.7. Films thermal properties

The thermal transitions of the prepared films were assessedthrough Modulate Differential Scanning Calorimetry (MDSC Q100,TA Instruments). Prior to measurements, the samples were con-ditioned, for at least 2 weeks at room temperature, in a sealedvessel containing a saturated solution of K2CO3, which assuresan atmosphere with a constant relative humidity of 44% [45].Conditioned samples, with a weight between 2 and 4 mg, weresealed in aluminum pans, cooled to 0 ◦C, and equilibrate during10 min. The samples were then scanned from 0 to 200 ◦C at arate of 5 ◦C/min, with a modulation with ±0.50 ◦C of amplitudeand a period of 40 s. An empty pan of the same type as the sam-ple pan was used as a reference. Thermograms were analyzedusing Universal Analysis software (version 4.2E, TA Instruments)for the identification of the glass transition temperatures and melt-ing temperatures. The glass transition temperature was detectedas a shift in the baseline of the reversible heat flow thermo-gram, while the melting temperature (helix-to-coil transition) wasdetect as an endothermic peak in the non-reversible heat flowthermogram.

2.8. Films mechanical properties

The tensile properties of the films were determined with atensile testing machine (Chatillon TCD 1000). Strip-shaped sam-ples, with a length of 40 mm, width of 7 mm, and thicknessesbetween 0.23 and 0.34 mm, were conditioned at 25 ◦C in a watervapor saturated atmosphere, using a desiccator containing a sat-urated solution of copper sulfate. The samples were conditionedfor at least 2 weeks prior to analysis. The initial gauge lengthwas set to 20 mm and the experiments were run at a crossheadspeed of 50 mm/min. The measurements were performed imme-diately after the sample was removed from the desiccator. Atleast five measurements were made for each type of film. Theobtained stress–strain curves were used to calculate the filmsstress at break (MPa), strain at break (%), and Young’s modulus(MPa).

2.9. Statistical analysis

Data are expressed as mean ± standard deviation. Statisticalanalysis was performed using Microsoft Excel software. Compar-isons between multiple groups were made with analysis of variance(ANOVA). Comparisons between two groups were made using Stu-dent’s t-test (two tail, unequal variance). A p-value of <0.05 wasconsidered to be statistically significant.

P. Coimbra et al. / International Journal of Biolo

Fgn

3

3u

Edwbtfroo1nwAttoawd

oGpatcomoa

ig. 1. Degree of cross-linking (triangles) and buffer uptake capacity (bars) of theelatin films cross-linked with EDC. All results are statistically different (p < 0.05,

= 4).

. Results and discussion

.1. Gel-E films preparation, degree of cross-linking, and bufferptake capacity

The cross-linking mechanism of proteinaceous materials withDC in the presence of NHS is well established and is amplyescribed in literature [18,46]. EDC is a zero-order cross-linker,hich means that this molecule enables the cross-link formation

ut do not remain as part of that linkage. In gelatin, EDC activateshe carboxylic groups of the aspartic and glutamic acid residues,orming an intermediate (O-acylisourea group) that permits theeaction of the activated carboxylic acids with the �-amino groupsf lysine and hydroxylysine residues, which results in the formationf an amide bond and a molecule of 3-(3-dimethylaminopropyl)--ethyl urea as a byproduct. However, and due to its instableature, the intermediate O-acylisourea can be easily hydrolyzed,hich results in the formation of the starting carboxylic acid group.lternatively, this intermediate can suffer a rearrangement and be

ransformed into a stable N-acylisourea group. In order to minimizehese undesirable side reactions, N-hydroxysuccinimide (NHS) isften included in EDC coupling protocols. NHS reacts with the EDCctivate carboxylic acid groups, forming a NHS ester intermediate,hich is considerable more stable than the O-acylisourea interme-iate.

By varying the amount of EDC (i.e., by increasing the molar ratiosf EDC to COOHgelatin groups), it was possible to obtain a series ofel-E films with a broad range of degrees of cross-linking (DC). Thisarameter, determined by the TNBS assay, represents the percent-ge of the initial �-amino groups present in gelatin that reacted withhe activated carboxylic groups to form inter- or intra-molecularovalent bonds between gelatin chains. As Fig. 1 shows, the degree

f cross-linking increases with the increase of EDC/COOHgelatinolar ratios (Gel-E films with DC between 23.4% and 74.0% were

btained by employing EDC/COOHgelatin molar ratios between 0.25nd 4.0). Inversely, the buffer uptake capacity at equilibrium of

Fig. 2. Scheme of gelatin’s m

gical Macromolecules 70 (2014) 10–19 13

the Gel-E films (also represented in Fig. 1) decreases monotoni-cally with the increase of EDC/COOHgelatin molar ratio, i.e. with theincrease of the degree of cross-linking. This expected behavior canbe explained by the equilibrium swelling theories of cross-linkedpolymer networks, namely the model propose by Flory and Rehner,that postulates that the equilibrium extent of swelling is deter-mined by two opposite driving forces [47]: the thermodynamiccompatibility of the polymer chains and the swelling liquid, andthe elastic retractive forces of the polymer chains. The reduction ofthe free energy of the system due to polymer and solvent mixingpromotes the swelling, while the increase in free energy due to thestretching of polymer chains opposes the swelling. When these twoopposite forces counterbalance each other, the swelling reachesequilibrium. As the cross-linking density increases, the averagelength of the polymer chains between network junctions decreases,causing the intensification of the retractive forces of the polymerchains that limit the swelling.

3.2. Gel-MA films preparation, degree of substitution, and bufferuptake capacity

As mentioned earlier, in order to prepare the photo-cross-linked gelatin films, gelatin must first be modified with pendentmethacrylamide groups, moieties that are susceptible to poly-merize when exposed to UV irradiation in the presence of aphotoinitiator.

Gelatin was modified by the reaction with methacrylic anhy-dride in PBS [43]. In this reaction, schematically representedin Fig. 2, the �-amino groups of gelatin react with methacrylicanhydride (MMA), yielding gelatin with pendant methacrylamidefunctional groups and methacrylic acid as a side product. Byemploying different amounts of MMA, a range of methacrylamidegelatin derivatives, bearing different amounts of methacrylamideside groups, were obtained. As Fig. 3 illustrates, the degree ofsubstitution (DS) of these derivatives (i.e., the percentage of themodified �-amino groups, determined by the TNBS assay) increasesalmost linearly with the increase of MAA/NH2gelatin molar ratiosuntil a saturation plateau is reached (around 90% DS) at the high-est MAA/NH2gelatin molar ratios (4.0 and 16.0). These results agreewith the ones published by other authors [15,48,49] and demon-strate the efficiency of the modification reaction and the ability tocontrol the degree of substitution by employing different amountsof MMA.

The photo-cross-linked Gel-MA films were prepared from 8%(w/v) Gel-MA solutions in the presence of the photoinitiatorIrgacure®2959. Irgacure®2959 is the most commonly used pho-toinitiator in the preparation of photo-cross-linked hydrogels forbiomedical applications, due to its water solubility and goodcytocompatibility in comparison to other photoinitiators [50,51].When exposed to LW-UV, Irgacure®2959 forms radicals that willinitiate a free radical polymerization of the pendant methacry-lamide groups introduced along the gelatin backbone, originatingpolymethacryamide segments (PMA) connected to several gelatin

chains. The number and the length of the formed PMA chainswill dictate the final chemical network structure of the hydrogeland, in this way, determine several of its final properties, such asthe swelling capacity and mechanical properties [52]. Since the

odification reaction.

14 P. Coimbra et al. / International Journal of Biolo

Fig. 3. Degree of substitution (triangles) of the methacrylamide gelatins and bufferunt

nsctceefecst

ldtct

cmcof1fsrtoctcmttFbr

httw

ptake capacity (bars) of the respective photo-cross-linked films. Statistically sig-ificant differences were detected between groups (p < 0.05, n = 4), except betweenhe ones signalized by *.

umber and length of the formed PMA chains are influenced byeveral factors, such as the DS of Gel-MA derivatives, Gel-MAoncentration, initiator concentration, and photo-irradiation time,hese parameters can be manipulated in order to produce photo-ross-linked hydrogels with different properties. Van den Bulcket al. [43] investigated the influence of the aforementioned param-ters in the final viscoelastic properties of GM-MA hydrogels andound that all these variables affect the hydrogels rheological prop-rties. However, besides the DS and the cross-linking reactiononditions, the hydrogel properties are also greatly affected by theequence in which chemical cross-linking and physical gelationake place [43]. This aspect will be discussed later.

In this work, the different Gel-MA derivatives were photo-cross-inked immediately after the Gel-MA solutions were cast in petriishes and reached the room temperature, keeping the same reac-ion conditions. So the different behaviors of the produced filmsan only be attributed to the different degrees of substitution ofhe Gel-MA derivatives.

In Fig. 3, the buffer uptake capacity at equilibrium of the photo-ross-linked Gel-MA films is plotted against the MAA/NH2gelatinolar ratios used to prepare the different Gel-MA derivatives. As it

an be seen, the buffer uptake capacity decreases with the increasef MAA/NH2gelatin molar ratios (i.e. with the degree of substitution)or MAA/NH2gelatin molar ratios inferior or equal to 1.0 (DS between9.6% and 57.7%). For higher molar ratios, and regardless of theurther increase in DS, the buffer uptake capacity remains con-tant (∼5 gPBS/gfilm). Since the buffer uptake capacity is inverselyelated with the cross-linking density (as it could be observed forhe Gel-E films) it can be concluded that the cross-linking densityf the formed films only increases with the increase of DS until aertain value, being after that insensible to a further increase ofhe degree of substitution. This can be explained by the way thehemical cross-linkers are formed in this type of films: as the poly-erization reaction develops, and the PMA segments are formed,

he mobility of the gelatin chains diminishes rapidly, rendering theotal conversion of the methacrylamide groups almost impossible.or Gel-MA derivatives with high DS, the cross-linking reaction wille limited essentially by the loss of mobility of the gelatin chainsather than by the number of available methacrylamide groups.

Comparing Figs. 1 and 3, it is clear that Gel-MA films have

igher buffer uptake capacities than those of Gel-E. For example,he Gel-MA membrane prepared from the Gel-MA derivative withhe lowest DS (19.6%) has a buffer uptake capacity of 11.2 gPBS/gfilm,hich is almost three times higher than the buffer uptake

gical Macromolecules 70 (2014) 10–19

capacity of the Gel-E membrane with the lowest DC (23.4%), whichis 4.1 gPBS/gfilm. On the other extreme, the Gel-MA films preparedwith the Gel-MA derivatives with highest DS (77.7%, 88.4% and91.8%) exhibit buffer uptake capacities of around 5 gPBS/gfilm, whichis also almost three times higher than the buffer uptake capacityof the Gel-E membrane with the highest DC (74.0%, 1.7 gPBS/gfilm).The general higher buffer uptake capacities of the Gel-MA could beattributed, at least in part, to the markedly hydrophilic character ofmethacrylamide moieties introduced in the gelatin structure.

The buffer uptake capacity is a very important parameter thatmust be taken in consideration when developing a chemical cross-linked gelatin film for drug delivery applications. This parameterwill not only determine the quantity of drug that can be loadedinto the film (if drug loading is carried by diffusion, after film for-mation) but also, in diffusion controlled systems, the drug releaserate. Furthermore, and as the next results demonstrate, the higherthe buffer uptake capacity, the faster the enzymatic degradation ofthe films will be.

Figs. 1 and 3 also show that the films produced exhibit alarge range of degrees of crosslinking and degrees of substitu-tion. Three films of each type, representative of low, medium,and high chemical cross-linking degrees, were selected for fur-ther characterization. They are Gel-E0.25 (DC = 23.4%); Gel-E1(DC = 50.5%); Gel-E4 (DC = 74.0%) regarding the first type of films,and Gel-MA0.25 (DS = 19.6%); Gel-MA1 (DS = 57.7%) and Gel-MA4(DS = 88.4%) regarding Gel-MA films.

3.3. Films hydrolytic and enzymatic degradation

As described earlier, to assess the resistance to hydrolytic degra-dation, films where incubated in PBS at 37 ◦C and their weight losswas recorded as a function of time. The results, presented in Fig. 4a,show that both types of films are stable and remarkably resistant tohydrolytic degradation, especially the ones with high and mediumdegrees of cross-linking, that lost less than 10% of mass during a9 weeks period. Only the films with the lowest degrees of cross-linking (Gel-E0.25 and Gel-MA0.25) exhibited a gradual weight lossalong time, more prominent in the case of Gel-MA0.25 film.

To evaluate the resistance against proteolytic enzymes, the filmswere incubated in TES buffer with 24.5 U/ml of collagenase fromC. histolyticum. In these conditions, all films, except Gel-E4, weretotally disintegrated after one day (Fig. 4b). For each type of film, thedegradation rate decreased with the increase of the chemical cross-linking level. These results are in accordance to the ones reported inthe literature by several authors that show that, in vitro, an increaseof the chemical cross-linking density of gelatins materials leads toan increase in the resistance to collagenase degradation [21,34].The Gel-E films were found to be considerably more resistant tocollagenase digestion than Gel-MA films. In fact, all Gel-MA films,independently of the cross-linking level, were completely dissolvedafter 8 h, while the Gel-E1 membrane disintegrated only after 24 hand, notably, the membrane Gel-E4 resisted to disintegration andlost only about 40% of their original mass after an incubation periodof 4 days. These different behaviors can be explained by the differ-ent buffer uptake capacities exhibited by the two types of films.As already shown, Gel-MA films have significantly higher bufferuptake capacities than the Gel-E films. Since the higher the watercontent, the faster the diffusion of the enzymes throughout thehydrogel network and the lower the spatial restrictions imposedto the enzymes in their approach to the gelatin chains, it is thusexpected that the hydrogels with higher water uptake capacitiesbe more easily degraded by collagenases than those hydrogels with

lower buffer uptake capacities, as it is the case of the investigatedfilms.

The evaluation of the enzymatic degradation behavior is animportant step in the development of gelatin based drug delivery

P. Coimbra et al. / International Journal of Biolo

Fb(

sitpaditc

3

cmttcttatacir

ig. 4. Remaining weight of the cross-linked gelatin films as a function of the incu-ation time in: a) PBS at 37 ◦C; b) TES buffer with collagenases from C. histolyticum24.5 U/ml) at 37 ◦C.

ystem. This is especially true when the therapeutic agent can formnnumerous electrostatic interaction with the positively or nega-ively charged gelatin residues, such as the case of many peptides,roteins, and oligo- and polynucleotides. In this circumstance, ands several in vivo studies suggest, the mechanism that controls therug release rate is the enzymatic degradation of the gelatin carrier

tself [17]. Thus, and since the enzymatic degradation is affected byhe degree of cross-linking, the release rate of the therapeutic agentan be controlled by manipulating this last parameter.

.4. Thermal properties

The thermal behavior of dehydrated gelatin films formed byasting is strongly dependent upon their drying conditions andoisture content. Generally, in gelatin films cast and dried at

emperatures below the coil-to-helix transition temperature (gela-ion temperature), part of the gelatin macromolecules recovers aollagen-like triple-helix structure before complete evaporation ofhe water. Further, the film’s renaturation or structuring level (i.e.he percentage of gelatin chains in the film that recovered into

collagen-like helical structure) increases with the decrease ofhe drying temperature [53]. On the other hand, in films formed

nd dried at temperatures above the gelation temperature, gelatinhains remain almost entirely in a random coil conformation, form-ng, when dried, an amorphous structure. Gelatin films dried atoom temperature are usually in a semi-crystalline state, with

gical Macromolecules 70 (2014) 10–19 15

crystalline domains formed by the association of triple-helix struc-tures, that are characterized by a melting temperature associatedto the helix-to-coil transition, and amorphous domains, where thegelatin chains are in a random coil conformation, exhibiting a glasstransition temperature.

Table 1 lists the glass transition temperature (Tg), melting tem-perature (Tm) and enthalpy of melting (�Hm) obtained for the twotypes of chemically cross-linked films together with the resultsobtained for a gelatin film physically cross-linked (taken as con-trol) and a set of films obtained solely by the physical gelationof the Gel-MA derivatives (named p-Gel-MA). As complement ofTable 1, representative thermograms of all analyzed films are pro-vided in supplementary data. Since water acts as a plasticizer of thegelatin chains, the glass transition and melting temperature exhib-ited by gelatin samples are highly dependent upon their moisturecontent. For this reason, and in order to assure that all the sam-ples had similar moisture contents, the samples were conditionedin an atmosphere with a constant relative humidity (RH) of 44% forat least 2 weeks before being subject to thermal analysis. At theseconditions, the water content in (mammalian derived) gelatin castfilms is reported to be around 14% [45,54,55].

Table 1 shows that the gelatin membrane without any typeof chemical cross-links (control) exhibited a Tg of 62 ± 5 ◦C and aTm of 82 ± 6 ◦C. The obtained melting temperature is lower thanthose reported in literature for porcine or bovine gelatin filmsconditioned at the same RH and determined by conventional DSC(Tm between 100◦ and 110 ◦C [45,54,55]). More in line with ourresults are the transition temperatures reported by Dai et al. [53](Tg ∼ 60 ◦C, Tm ∼ 75 ◦C; samples conditioned at 50% RH), determinedalso by MDSC.

In Table 1 it is also displayed the melting enthalpy (�Hm) ofthe membranes, a parameter associated with the Tm temperaturethat reflects the extent in which renaturated gelatin is present. The�Hm values exhibited by physically cross-linked gelatin films dif-fer according to the chemical composition of gelatin (associatedwith the type of animal source and extraction method) [56,57],conditions in which the films are formed (temperature and relativehumidity) [53], and the moister content present during the analy-sis [45,55]. Generally, for mammalian derived gelatin films cast atroom temperature and conditioned at RH similar to the one usedin this work, the �Hm values reported in literature vary between20 and 30 J/g [25,32,53,55–57], even though, in a few works, muchhigher values are reported [58]. In this work, the �Hm obtainedfor the control film was of 64 ± 12 J/g, which is about three timeshigher than the general values referred above. This result is proba-bly due to the fact that, in these samples, melting takes place overa wide temperature range. This aspect, associated with instrumen-tal baseline curvature, make it difficult to determine the onset andoffset of crystal melting, which can lead to an overestimation of�Hm. In view of the above considerations, �Hm values obtainedfor the different films will be analyzed only in comparative terms,considering the �Hm value obtained for the physically cross-linkedgelatin as reference.

Generally, the introduction of covalent bonds between gelatinchains leads to an increase the thermal stability of gelatin, whichis usually reflected in the increase of the Tg and Tm values[23,25,32,33]. However, some reports have been published inwhich chemical cross-linking didn’t change the Tg and Tm values, oreven decreased them [24]. In this work, the melting temperaturesof the gelatin films cross-linked with EDC stayed approximately thesame as the Tm of the control sample. Due to the low energy asso-ciated to the Tg transitions, for some samples it was impossible to

identify unambiguously this transition in the reversible heat flow.This situation was observed for the films Gel-E0.25 and Gel-E1. Bycontrast, the Gel-E4 film, with the higher degree of cross-linking,exhibited two well defined glass transition temperatures – one at

16 P. Coimbra et al. / International Journal of Biological Macromolecules 70 (2014) 10–19

Table 1Glass transition temperature (Tg), melting temperature (Tm), and enthalpy of melting (�Hm) of the gelatin films.

Type of gelatin film Tg (◦C) Tm (◦C) �Hm (J/g)

Phsicallycross-linkedgelatin(control)

62 ± 5 (2) 82 ± 6 (3) 64 ± 12 (3)

Gelatincross-linkedwith EDC

Gel-E0.25 n.d. 83 ± 2 (2) 63 ± 9 (2)

Gel-E1 n.d. 80 ± 5 (2) 59 ± 10 (2)Gel-E4 54/117* 78 ± 5 (2) 56 ± 10 (2)

Physicallycross-linkedGel-MA

p-Gel-MA0.25 54 ± 2 (2) 75 ± 2 (2) 28 ± 5 (2)

p-Gel-MA1 54 ± 2 (2) 72 ± 3 (2) 35 ± 6 (2)p-Gel-MA4 54 n.d. n.d.

Gel-MAphoto-cross-linked

Gel-MA0.25 n.d. 65 ± 7 (2) 25 ± 4 (2)

Gel-MA1 67 ± 2 (2) 96 ± 12 (3) 6 ± 2 (3)Gel-MA4 69 ± 2 (2) 124 ± 15 (3) 4 ± 2 (3)

nU

nd ot

5Tpt

Fhos

.d.: not detected.nder brackets: number of replicates.* Two glass transition temperatures were detected for this sample: one at 54 ◦C a

4 ◦C and other at 117 ◦C (Fig. S1a in Supplementary Data). The first

g value is slightly lower than the one observed in the control sam-le, while the second is much higher. Usually, the Tg increases withhe increase of the cross-linking density due to the intensification

ig. 5. Proposed schemes for the Gel-E and Gel-MA networks formation. A) Gel-E filmeterogeneous conditions (2). B) Gel-MA films: the chemical cross-linking, formed by thccurs when physical gelation is starting. After being formed, the chemical cross-linkers (ptructures and the development of new ones.

her at 117 ◦C (see Fig. S1a in Supplementary Data).

of the restrictions to molecular mobility introduced by the cross-

linkers. The appearance of two Tgs suggests that there are regions ofthe amorphous phase more densely cross-linked than others. Thisnon-homogeneity of the cross-linking density can be attributed

s: the physical network is formed (1) followed by the chemical cross-linking ine UV initiated free radical polymerization of the pendant methacrylamide groups,olymethacrylamide segments) limit the growth of the already formed triple-helical

f Biological Macromolecules 70 (2014) 10–19 17

ttitmrrlu

edlvtptastGatthliboso

idepgTTsfc(bmtgtialctdtt

3

wpcD(s

Fig. 6. Tensile properties of the various gelatin films. a) Stress at break; b) strain atbreak; c) Young’s modulus. Light grey: physically cross-linked membrane (control);dark grey: films cross-linked with EDC. Black: photo-cross-linked Gel-MA films.

P. Coimbra et al. / International Journal o

o the heterogeneous conditions in which the cross-linking reac-ion occurs. As mentioned earlier, these films were cross-linked bymmersing the already physically structured membrane in a solu-ion containing EDC and NHS (Fig. 5A). In this way, the cross-linker

olecules must diffuse through the network structure in order toeact with the carboxylic groups of the aspartic and glutamic acidesidues. These will originate a concentration gradient of the cross-inker throughout the membranes cross-section, which can lead toneven cross-linking densities.

As mentioned, and for comparison purposes, the thermal prop-rties of the films formed only by the physical gelation of the gelatinerivatives Gel-MA, i.e., without being subject to chemical cross-

inking (p-Gel-MA) were also analyzed (Table 1). The Tg and the Tm

alues displayed by these films are slightly lower than the ones ofhe control sample of unmodified gelatin. The glass transition tem-erature has not affected by the degree of substitution, remaininghe same (54 ◦C) for the three films with increasing DS. Addition-lly, for the film formed by the Gel-MA with the higher degree ofubstitution (p-Gel-MA4.0), no Tm was detected, which means thathe film is essentially in an amorphous state. As for the films p-el-MA0.25 and p-Gel-MA1.0, their melting enthalpies (28 ± 5 J/gnd 35 ± 6 J/g, respectively) are about half of the one observed forhe control sample. These results are in general accordance withhe ones obtained by Hu et al. for physically cross-linked Gel-MAydrogels with high water content [59]. The authors attributed the

owering of Tm and �Hm in the Gel-MA physical gels to a reductionn the length and extension of the triple-helix structures, causedy the interference of methacrylamide side groups in the processf formation and stabilization of these structures. The same conclu-ions, based on dynamic rheological measurements, were pointedut by other authors [43].

As for the photo-cross-linked Gel-MA films, and as observedn Table 1, the gelatin derivative with the lowest DS, Gel-MA0.25,isplays a Tm and an enthalpy of melting comparable to the onesxhibited by the corresponding non-photo-cross-linked counterartner (p-Gel-MA0.25). In contrast, in the films formed by theelatin derivatives with higher DS (Gel-MA1 and Gel-MA4), them increases while the enthalpy of melting decreases significantly.his behavior can be explained by the order in which physicaltructuring and chemical cross-linking happened during the filmsormation (Fig. 5B). Contrary to the Gel-E films, where chemicalross-linking occurs only after the physical structuring is completeFig. 5A), in the Gel-MA films, the chemical cross-linking, formedy the UV initiated free radical polymerization of the pendantethacrylamide groups, occurs immediately after the Gel-MA solu-

ion reaches room-temperature, i.e. at the same time as physicalelation is starting. Since physical gelation is a much slower processhan the chemical cross-linking, when the latter process is fin-shed, the triple helices structures present in the gel are yet limitednd are still growing. However, the newly formed chemical cross-inkers will restrain the movements of the gelatin strands and,onsequentially, limit the further development of physical struc-uring [41–43]. The higher the degree of substitution of the Gel-MAerivatives, the more limited will be the physical structuring and,herefore, the melting enthalpy associated to the helical-to-coilransition of the films.

.5. Mechanical properties

The tensile properties of the Gel-E and Gel-MA films, togetherith those of a physically cross-linked gelatin film (control sam-le), are summarized in Fig. 6. Prior to analysis, the samples were

onditioned in a water vapor saturated atmosphere for 2 weeks.uring this period, the films absorbed around 30.6 ± 2.6% of water

measured by gravimetry). At this water content, the glass tran-ition temperature of gelatin is lower than the environmental

Statistically significant differences were detected between groups (p < 0.05, n ≥ 5),except between the ones signalized by * or **.

temperature [45], which means that the films were in a rubberystate when they were analyzed.

Regarding Gel-E films, Fig. 6a shows that the stress at break(�b) values of these samples increase with the increase of DC (from

6.1 ± 1.3 MPa of Gel-E0.25 to 11.3 ± 0.9 MPa of Gel-E4), and are sev-eral times larger than the �b obtained for the control membrane(1.8 ± 0.8 MPa). On the contrary, the values of strain at break (εb),

1 f Biolo

dboattctctMatatv(

GtptitMeAtfit1

etbscicwcioaGieMbgtlaagcoslabY

ime

[

[

[

8 P. Coimbra et al. / International Journal o

isplayed in Fig. 6b, are lower than the one of the control mem-rane (103 ± 9%) and decrease with the increase of DC (from 86 ± 5%f the Gel-E0.25 to 38 ± 4% of the Gel-E4). (Note that the gener-lly high values of the strain at break obtained for all films reflectheir rubbery state.) This mechanical behavior is in accordance withhe majority of the reports found in literature for gelatin filmsross-linked with a number of chemical agents [20,23,25,31,32]:he mechanical strength of the films increases due to the chemicalross-linking which is reflected in a decrease of the stretchability ofhe films and an increase of the stress at break. As for the Young’s

odulus (E) of the Gel-E films, Fig. 6c shows that, although theylso increase with the increase of DC, they are not always largerhan that of the control sample. In fact, only the Gel-E4 film has

superior Young’s Modulus (61.0 ± 6.1 MPa) than that of the con-rol (E = 42.5 ± 10.3 MPa), while the Gel-E1 film has a comparablealue of E (42.55 ± 4.5 MPa) and Gel-E0.25 has an inferior value24.45 ± 5.0 MPa).

Additionally, Fig. 6 clearly shows that the photo-cross-linkedel-MA films possess rather different mechanical properties than

hose cross-linked with EDC. Indeed, and contrary to the Gel-E sam-les, the Gel-MA film with the highest stress at break is the one withhe lowest DS (Fig. 6a). This film has a �b of 7.1 ± 0.5 MPa, whichs higher than the �b of the control sample (and comparable tohe �b of Gel-E0.25). However, the other two Gel-MA films – Gel-

A1 and Gel-MA4 – formed by gelatin derivatives with higher DS,xhibit �b values comparable to the one of the control membrane.s for the Young’s Modulus of the Gel-MA films, Fig. 6c shows a con-

inuous decreases with the increase of DS: while the Gel-MA0.25lm has a E of 45.5 ± 6.2 MPa, comparable with the one of the con-rol membrane, the E values of Gel-MA1 and Gel-MA4 decrease to4.1 ± 0.4 MPa and 5.5 ± 1.4 MPa, respectively.

With respect to the values of strain at break, the Gel-MA filmsxhibit significantly higher values (around 144%) compared to bothhe control and the Gel-E films. Furthermore, they are not affectedy the DS (Fig. 6b). This result suggests that the methacrylamideide groups have a plasticizer effect in the gelatin films. This isorroborated by the MDSC results, where a decrease of approx-mately 10 ◦C in the glass temperature, in comparison with theontrol membrane, was observed for the Gel-MA films, with andithout chemical cross-linking (Table 1). As mentioned above, the

hemical cross-linking of the Gel-MA films happens when the phys-cal structuring is beginning (Fig. 5B). In this situation, the formationf helical structures will be hindered by the chemical cross-links,s already been demonstrated by several authors [38,40–42]. Forel-MA with a low DS, such as Gel-MA0.25, the formed chem-

cal network will be sparse which will permit, at least to somextent, the further development of the physical gelation. For Gel-A derivatives with higher DS, the formed chemical network will

e stronger and denser and will critically limit the mobility of theelatin chains, restraining in this way the further development ofhe physical gelation. Therefore, the Gel-MA0.25 sample has higherevels of renaturated gelatin than the films Gel-MA1 and Gel-MA4,s the melting enthalpies of these three films suggest (Table 1). In

series of studies, Bigi et al. [57,60] investigated the influence ofelatin’s renaturation level in the mechanical properties of physi-ally cross-linked gelatin films and has concluded that the increasef the renaturation level led to an increase in the stress at break,train at break, and Young’s modulus. Therefore, the decrease of theevel of renaturated gelatin with the increase of DS, alongside thepparent plasticizer effect of the methacrylamide side groups, cane responsible for the general decrease of the stress at break andoung’s Modulus of the Gel-MA films.

In the Gel-MA films, the positive effect of increasing the chem-cal cross-linking density (due to the increase of DS) on the

echanical properties of the films is exceeded by the negativeffect caused by the decrease of the renaturation level, which means

[

[

gical Macromolecules 70 (2014) 10–19

that the contribution of the chemical cross-links (PMA segmentsconnecting several gelatin chains) for the overall strength of thenetwork is probably weaker than the contribution of the physicalcross-links (formed by the association of triple–helices structures).In the Gel-E films, the gelatin’s renaturation level is not affectedby the extension of the chemical cross-linking, since the physi-cal network is already formed. In this situation, and contrary towhat happens in the Gel-MA films, the increase of the chemicalcross-linking density reinforces the strength and stability of phys-ical cross-links, which results in the increase of stress at break andof the Young’s modulus and in the decrease of the strain at break.

4. Conclusions

The present study has shown that both the degree of cross-linking and the cross linking method greatly affect the finalproperties of gelatin films. In fact, the gelatin films cross-linkedwith EDC exhibited much lower buffer uptake capacities and agreater resistance to collagenase digestion than the Gel-MA films.Moreover, the tensile tests also showed that Gel-E samples areconsiderable more tough and rigid. These differences are mostlikely a consequence of the rather distinct nature of the chemi-cal cross-links introduced in each type of film and also of the orderin which the chemical and physical networks were formed. Theresults obtained demonstrate clearly that it is possible to tailor theproperties of the chemically cross-linked gelatin films for a spe-cific application by selecting the adequate chemical cross-linkingmethod and also by controlling the thermal conditions in whichit occurs. Furthermore, and for a particular method, a further con-trol of the films properties can be achieved by the variation of thechemical cross-linking level.

Acknowledgment

Patrícia Coimbra gratefully acknowledges Portuguese ScienceFoundation (FCT) for financial support (SRFH-BPD-73367-2010).

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.ijbiomac.2014.06.021.

References

[1] S. Gorgieva, V. Kokol, in: R. Pignatello (Ed.), Biomaterials Applications forNanomedicine, InTech, Rijeka, 2011, pp. 17–52.

[2] K.B. Djagny, Z. Wang, S. Xu, Crit. Rev. Food. Sci. 41 (2001) 481–492.[3] M. Djabourov, J. Leblond, P. Papon, J. Phys. 49 (1988) 319–332.[4] V. Normand, S. Muller, J.-C. Ravey, A. Parker, Macromolecules 33 (2000)

1063–1071.[5] L. Guo, R.H. Colby, C.P. Lusignan, A.M. Howe, Macromolecules 36 (2003)

10009–10020.[6] V.I. Uricanu, M.H.G. Duits, R.M.F. Nelissen, M.L. Bennink, J. Mellema, Langmuir

19 (2003) 8182–8194.[7] M. Kabiri, S.H. Emami, M. Rafinia, M. Tahriri, Curr. Appl. Phys. 11 (2011)

457–461.[8] S.R. Hong, S.J. Lee, J.W. Shim, Y.S. Choi, Y.M. Lee, K.W. Song, M.H. Park, Y.S. Nam,

S.I. Lee, Biomaterials 22 (2001) 2777–2783.[9] V. Rattanaruengsrikul, N. Pimpha, P. Supaphol, J. Appl. Polym. Sci. 124 (2012)

1668–1682.10] M. Madaghiele, A. Piccinno, M. Saponaro, A. Maffezzoli, A. Sannino, J. Mater.

Sci.-Mater. Med. 20 (2009) 1979–1989.11] C.W. Yung, L.Q. Wu, J.A. Tullman, G.F. Payne, W.E. Bentley, T.A. Barbari, J. Biomed.

Mater. Res. A 83A (2007) 1039–1046.12] Y.-C. Chen, W.-Y. Su, S.-H. Yang, A. Gefen, F.-H. Lin, Acta Biomater. 9 (2013)

5181–5193.13] M.-k. Yeh, Y.-m. Liang, K.-m. Cheng, N.-T. Dai, C.-c. Liu, J.-j. Young, Polymer 52

(2011) 996–1003.14] Y.-C. Chen, R.-Z. Lin, H. Qi, Y. Yang, H. Bae, J.M. Melero-Martin, A. Khademhos-

seini, Adv. Funct. Mater. 22 (2012) 2027–2039.

f Biolo

[

[[

[

[

[[[

[

[

[[[

[

[

[

[

[

[

[

[

[[[

[

[

[[[

[[

[[

[

[

[

[

[[[[

[

P. Coimbra et al. / International Journal o

15] J.W. Nichol, S.T. Koshy, H. Bae, C.M. Hwang, S. Yamanlar, A. Khademhosseini,Biomaterials 31 (2010) 5536–5544.

16] X. Hu, L. Ma, C. Wang, C. Gao, Macromol. Biosci. 9 (2009) 1194–1201.17] S. Young, M. Wong, Y. Tabata, A.G. Mikos, J. Control. Release 109 (2005)

256–274.18] A.J. Kuijpers, G.H.M. Engbers, P.B. van Wachem, J. Krijgsveld, S.A.J. Zaat, J.

Dankert, J. Feijen, J. Control. Release 53 (1998) 235–247.19] A.J. Kuijpers, P.B. van Wachem, M.J.A. van Luyn, G.H.M. Engbers, J. Krijgsveld,

S.A.J. Zaat, J. Dankert, J. Feijen, J. Control. Release 67 (2000) 323–336.20] J.-Y. Lai, Y.-T. Li, Mater. Sci. Eng. C 30 (2010) 677–685.21] M. Ozeki, Y. Tabata, J. Biomater. Sci. Polym. Ed. 16 (2005) 549–561.22] K. Terao, N. Nagasawa, H. Nishida, K. Furusawa, Y. Mori, F. Yoshii, T. Dobashi, J.

Biomater. Sci. Polym. Ed. 14 (2003) 1197–1208.23] A. Bigi, G. Cojazzi, S. Panzavolta, K. Rubini, N. Roveri, Biomaterials 22 (2001)

763–768.24] M. Catalina, G.E. Attenburrow, J. Cot, A.D. Covington, A.P.M. Antunes, J. Appl.

Polym. Sci. 119 (2011) 2105–2111.25] R.A. de Carvalho, C.R.F. Grosso, Food Hydrocolloid. 18 (2004) 717–726.26] J.Y. Lai, J. Mater. Sci. Mater. Med. 21 (2010) 1899–1911.27] Y.Z. Zhang, J. Venugopal, Z.M. Huang, C.T. Lim, S. Ramakrishna, Polymer 47

(2006) 2911–2917.28] M.J.A. van Luyn, P.B. van Wachem, L. Olde Damink, P.J. Dijkstra, J. Feijen, P.

Nieuwenhuis, J. Biomed. Mater. Res. 26 (1992) 1091–1110.29] A.J. Kuijpers, G.H.M. Engbers, J. Feijen, S.C. De Smedt, T.K.L. Meyvis, J. Demeester,

J. Krijgsveld, S.A.J. Zaat, J. Dankert, Macromolecules 32 (1999) 3325–3333.30] H.-C. Liang, W.-H. Chang, H.-F. Liang, M.-H. Lee, H.-W. Sung, J. Appl. Polym. Sci.

91 (2004) 4017–4026.31] G. Tronci, A.T. Neffe, B.F. Pierce, A. Lendlein, J. Mater. Chem. 20 (2010)

8875–8884.32] A. Bigi, G. Cojazzi, S. Panzavolta, N. Roveri, K. Rubini, Biomaterials 23 (2002)

4827–4832.33] C.-H. Yao, B.-S. Liu, C.-J. Chang, S.-H. Hsu, Y.-S. Chen, Mater. Chem. Phys. 83

(2004) 204–208.34] D.M. Kirchmajer, C.A. Watson, M. Ranson, M.i.h. Panhuis, RSC Adv. 3 (2013)

1073–1081.35] X. Zhang, M.D. Do, P. Casey, A. Sulistio, G.G. Qiao, L. Lundin, P. Lillford, S.

Kosaraju, J. Agric. Food Chem. 58 (2010) 6809–6815.

[[[[

gical Macromolecules 70 (2014) 10–19 19

36] T.J. Koob, D.J. Hernandez, Biomaterials 24 (2003) 1285–1292.37] H. Babin, E. Dickinson, Food Hydrocolloid. 15 (2001) 271–276.38] F. Bode, M.A. da Silva, A.F. Drake, S.B. Ross-Murphy, C.A. Dreiss, Biomacro-

molecules 12 (2011) 3741–3752.39] H. Okino, Y. Nakayama, M. Tanaka, T. Matsuda, J. Biomed. Mater. Res. 59 (2002)

233–245.40] S. Giraudier, D. Hellio, M. Djabourov, V. Larreta-Garde, Biomacromolecules 5

(2004) 1662–1666.41] D. Hellio-Serughetti, M. Djabourov, Langmuir 22 (2006) 8509–8515.42] D. Hellio-Serughetti, M. Djabourov, Langmuir 22 (2006) 8516–8522.43] A.I. Van den Bulcke, B. Bogdanov, N. De Rooze, E.H. Schacht, M. Cornelissen, H.

Berghmans, Biomacromolecules 1 (2000) 31–38.44] W.A. Bubnis, C.M. Ofner, Anal. Biochem. 207 (1992) 129–133.45] I. Yakimets, N. Wellner, A.C. Smith, R.H. Wilson, I. Farhat, J. Mitchell, Polymer

46 (2005) 12577–12585.46] J.V. Staros, R.W. Wright, D.M. Swingle, Anal. Biochem. 156 (1986) 220–222.47] F. Ganji, S. Vasheghani-Farahani, E. Vasheghani-Farahani, Iran. Polym. J. 19

(2010) 375–398.48] T. Billiet, B.V. Gasse, E. Gevaert, M. Cornelissen, J.C. Martins, P. Dubruel, Macro-

mol. Biosci. 13 (2013) 1531–1545.49] E. Hoch, C. Schuh, T. Hirth, G.E.M. Tovar, K. Borchers, J. Mater. Sci. Mater. Med.

23 (2012) 2607–2617.50] C.G. Williams, A.N. Malik, T.K. Kim, P.N. Manson, J.H. Elisseeff, Biomaterials 26

(2005) 1211–1218.51] I. Mironi-Harpaz, D.Y. Wang, S. Venkatraman, D. Seliktar, Acta Biomater. 8

(2012) 1838–1848.52] J.L. Ifkovits, J.A. Burdick, Tissue Eng. 13 (2007) 2369–2385.53] C.-A. Dai, Y.-F. Chen, M.-W. Liu, J. Appl. Polym. Sci. 99 (2006) 1795–1801.54] P.J.A. Sobral, A.M.Q.B. Habitante, Food Hydrocolloid. 15 (2001) 377–382.55] I. Yakimets, S.S. Paes, N. Wellner, A.C. Smith, R.H. Wilson, J.R. Mitchell,

Biomacromolecules 8 (2007) 1710–1722.56] D. Achet, X.W. He, Polymer 36 (1995) 787–791.

57] A. Bigi, S. Panzavolta, K. Rubini, Biomaterials 25 (2004) 5675–5680.58] M.S. Rahman, G.S. Al-Saidi, N. Guizani, Food Chem. 108 (2008) 472–481.59] X.H. Hu, D. Li, F. Zhou, C.Y. Gao, e-Polymers, 2012, pp. 14.60] A. Bigi, B. Bracci, G. Cojazzi, S. Panzavolta, N. Roveri, Biomaterials 19 (1998)

2335–2340.