Rapid prototyping of chitosan-coated alginate scaffolds through the use of a 3D fiber deposition technique

Journal ofMaterials Chemistry B

PAPER

Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27

.

View Article OnlineView Journal | View Issue

Rapid prototypin

aDepartment of Chemistry, Sapienza Univers

Italy. E-mail: [email protected] of Life, Health and Environmen

L'Aquila, ItalycDepartment of Biology and Biotechnology

00185 Rome, Italy

† Electronic supplementary informa10.1039/c4tb00732h

Cite this: J. Mater. Chem. B, 2014, 2,6779

Received 7th May 2014Accepted 18th July 2014

DOI: 10.1039/c4tb00732h

www.rsc.org/MaterialsB

This journal is © The Royal Society of C

g of chitosan-coated alginatescaffolds through the use of a 3D fiber depositiontechnique†

Cristina Colosi,a Marco Costantini,a Roberta Latini,a Serena Ciccarelli,a

Alessandra Stampella,c Andrea Barbetta,*a Mara Massimi,b Laura Conti Devirgiliisc

and Mariella Dentinia

Three dimensional, periodic scaffolds of chitosan-coated alginate are fabricated in a layer-by-layer fashion

by rapid prototyping. A novel dispensing system based on two coaxial needles delivers simultaneously

alginate and calcium chloride solutions permitting the direct deposition of alginate fibers according to

any designed pattern. Coating of the alginate fiber with chitosan and subsequent cross-linking with EDC

and genipin assured the endurance of the scaffold in the culture environment for a prolonged period of

time. The cross-linking protocol adopted imparted to the scaffold a hierarchical chemical structure as

evidenced by Confocal Laser Microscopy and FTIR spectroscopy. The core of the fibers making up the

scaffold is represented by alginate chains cross-linked by ester bonds only, the periphery of the fiber is

constituted by an inter-polyelectrolyte complex of alginate and chitosan cross-linked in all pair

combinations. Fibers belonging to adjacent layers are glued together by the chitosan coating.

Mechanical behavior of the scaffolds characterized by different layouts of deposition was determined

revealing anisotropic properties. The biocompatibility and capability of the scaffolds to sustain

hepatocyte (HepaRG) cultures were demonstrated. Typical hepatic functions such as albumin and urea

secretion and induction of CYP3A4 enzyme activity following drug administration were excellent, thus

proving the potential of these constructs in monitoring the liver specific function.

Introduction

Tissue engineering (TE)1 combines biological and engineeringexpertise to provide articially developed substitutes for tissuesand organs, hence promoting repair and replacement therapies.

A key element involved in TE processes is the matrix orscaffold, which serves as a substrate or framework for cellgrowth, aggregation, and tissue development. Scaffolds must beporous and interconnected to allow cell migration during thecolonization process and to allow the transport of nutrients andwaste to and from cells. The internal porous architecture ismostly determined by the technology used for scaffoldproduction and it can greatly inuence the overall success ofnew tissue formation. Therefore, the controlled fabrication of

ity of Rome, P.le A. Moro 5, 00185 Rome,

; Tel: +39-06-49913630

tal Sciences, University of L'Aquila, 67100

C. Darwin, Sapienza University of Rome,

tion (ESI) available. See DOI:

hemistry 2014

scaffold structures is becoming increasingly important inmodern approaches to regenerative medicine.

Among biomaterials, polymers, both natural and synthetic,are themostly used for scaffold fabrication. Hydrogels, thanks totheir high water content, have received a particular interest asscaffoldingmaterials since they resemble the native extracellularenvironment, enhance uid transport inside the matrix andpossess mechanical properties similar to those of so tissues.2

Rapid Prototyping (RP) is a manufacturing technology thathas emerged since 2000 in TE for the fabrication of scaffoldswith well-dened internal architecture.3 Generally, in RP threemotorized axis and a deposition system are used to lay down amaterial in a layer by layer manner, generating a three-dimen-sional physical model from Computer-Aided-Design (CAD) data.This technology enables the tissue engineer to have full controlover the design, fabrication and modeling of the scaffold beingconstructed, providing a systematic learning channel forinvestigating cell–matrix interactions. RP techniques are veryspecialized technologies in terms of material processability,and a great challenge is represented by the search for novel andeffective deposition strategies to prototype hydrogel scaffolds.4

Among three Dimensional Fiber Deposition techniques(3DFD), two major methods have been devised for printinghydrogel-based scaffolds. In the rst one, a concentrated

J. Mater. Chem. B, 2014, 2, 6779–6791 | 6779

http://crossmark.crossref.org/dialog/?doi=10.1039/c4tb00732h&domain=pdf&date_stamp=2014-09-13

http://dx.doi.org/10.1039/c4tb00732h

http://pubs.rsc.org/en/journals/journal/TB

http://pubs.rsc.org/en/journals/journal/TB?issueid=TB002039

Journal of Materials Chemistry B Paper

Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27

. View Article Online

solution of a biopolymer is deposited directly from a pneumaticsyringe dispenser on an either static or moving platform. Theshear-thinning experienced by the solution during the processof extrusion results in a fall in shear stress at higher extrusionpressures and is at the basis of the deposition of gel-like solu-tion. Such an approach was exploited in the manufacturing of asemi-interpenetrated network of hyaluronic acid and dextran-methacrylate based scaffolds5 and in the layered deposition ofcell-laden hydrogel strands.6 This method suffers from a lack inuniformity of the porous texture. Vertical pores are regularthroughout the samples, whereas transversal pores tend to fusetogether because of the relaxation downwards of the printeduncross-linked gel. Furthermore, the resolution of the printedscaffolds is rather poor. For instance, the strand distance mayvary from 0.5 to 2.5 mm.6a As a consequence while it is valid forprinting biopolymers and cells at the same time, this approachdoes not permit to exploit the full potential offered by RP inobtaining completely tailored scaffolds.

A second major 3DFD technique employing biopolymers isrepresented by bioplotting where a solution of a biopolymer isprinted in a coagulating bath. The key feature of this method isthe 3D dispensing of liquids and pastes into a liquid mediumwith matched density. The plotting material leaves the nozzleand solidies in the plotting medium aer bonding to theprevious layer. The liquid medium compensates for gravity andhence no support structure is needed.

This technique presents the disadvantage of a limited reso-lution, and fused horizontal pores have been mentioned insome cases. In addition, the 3D-Bioplotter technology is a timeconsuming technique due to the optimization of the plottingconditions for each different material7 and highly water solublematerials cannot be processed with this technique.8 We spec-ulated that some of the above illustrated limitations could beovercome if the ber-like, cross-linked (either physically orchemically) hydrogel making up the scaffold is formed simul-taneously with its extrusion. This approach can be realizedpractically by making recourse to a microuidic-like dispensingsystem. In the literature, examples of bers produced viamicrouidics are reported. For instance, a spider mimickingmicrouidic device has been employed for the production ofbroin bers.9 Aligned collagen brils were deposited on planarsubstrates from collagen solutions streaming through amicrouidic channel system.10

The aim of this article is to illustrate a bioplotting approachbased on a novel concept. Instead of printing in a coagulatingbath or depositing a pre-gelled viscous solution we envisagedthat by extruding at the same time the polymeric and thecoagulating solution through the inner and outer needles,respectively, of a coaxial dispensing system it would be possibleto deposit directly a solid lament. In this way the process ofscaffold printing is greatly simplied since all the abovereported limitations are bypassed. This approach relies on theability of several biopolymers to undergo a rapid gelation whenin contact with an appropriate chemical environment. Forinstance, acidied aqueous solutions of chitosan undergogelation when exposed to basic solution.7 Similarly, an aqueoussolution of broin solidies by re-establishing b-sheet junctions

6780 | J. Mater. Chem. B, 2014, 2, 6779–6791

when exposed to methanol or ethanol.11 Probably, the mostpopular and effective gelling biopolymer is alginate that whenexposed to divalent cations, usually Ca2+, instantaneously formsa physical gel by generating an egg-box network in which ionsare caged between two alginate chains.12 Furthermore, alginatehas been found to be particularly useful in pharmaceutical andbiomedical application areas. Therefore alginate seems to bethe optimal candidate to test the validity of the scaffoldproduction system we are here proposing.

In the article we will show that the scaffolds produced mayhave practical applications. In the last decade there has been anincreasing interest in the development of in vitro-systems,especially 3D-culture systems on a chip-basis, for drug testing.There is a wealth of information on how cells conned inconventional 2D monolayer cultures differ substantially in theirproperties from cells cultured in a 3D conguration. Lackingthe physical and chemical cues dening their natural in vivomicroenvironment, cells in 2D culture differ substantially intheir shape and organization, in contacts with neighboringcells, and in their physiology and metabolism from cellsobserved in more physiologically relevant 3D environments.13

Thus, the drug evaluation process will stand to gain greatbenets from the fairly accurate predictions of cellularresponses displayed by 3D-engineered tissue models whenexposed to the drugs of interest in vitro.

Liver toxicity is among the most common biological reasonsfor drug candidates to be pulled from clinical development, so itis important to be able to predict it. Even if a molecule does notharm the liver, that organ's detoxifying actions may harm themolecule, rendering potential drugs ineffective.

Since the liver, and the hepatocyte in particular, is the mainsite of xenobiotic biotransformation and detoxicationprocesses, and thus also of drug-induced toxicity in the body,cultures of liver cells provide a valid tool for systematic, repetitiveand quantitative investigations in drug analysis, allowing also asignicant reduction of the costs of preclinical drug develop-ment. Considerable attention has thus commonly paid to theestablishment of relevant hepatic in vitro models. The HepaRGcell line was recently found to be a valuable human-hepatocyte invitro model for investigating properties and potential effects ofvarious drug compounds.14 HepaRG cells are terminally differ-entiated hepatic cells that retain most characteristics of primarymature human hepatocytes. In contrast with other hepatomacells such as HepG2, HepaRG not only expresses a range of P450enzymes (CYP1A2, 2B6, 2C9, 2E1, 3A4) and responds to selectiveinducers of these enzymes, but also expresses other specicmarkers of adult hepatocytes, including nuclear receptors, drugtransporters and phase II xenobiotic metabolizing enzymes.15 Inaddition, results from recent studies show that HepaRG perfor-mance improves further when cultured in spheroids or in 3Dorganotypic cultures using scaffold free systems.16

A drug screening construct to be effective needs to bereproducible and rely on a fabrication technology which is ableto create a highly structured 3D culture microenvironment withdened shape and position on the micrometer scale.

As a consequence an additional aim of this work is todemonstrate the efficacy of the printing process in creating

This journal is © The Royal Society of Chemistry 2014


Fig. 1 (a) Schematic representation of the coaxial needle system for the extrusion of alginate fibers. (b) Photograph of the 3D fiber depositionsystem during scaffold production. (c–e) Optical micrographs of printed structures with (c) squared, (d) hexagonal and (e) radial patterns of fibers.Scale bars: 500 mm.

Paper Journal of Materials Chemistry B

Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27


scaffolds based on different geometrical webs and whether andhow these inuence the expression of liver specic functions incultured differentiated HepaRG cells.

Experimental sectionMaterials

Alginate (68% content in guluronic residues) was purchasedfrom Fluka and puried by precipitation in ethanol and dia-lysed against distilled water before use. Its average molecularweight was determined by capillary viscosimetry (Ubbelohdecapillary, F ¼ 0.53 mm, for dilution sequences) using theviscosity measuring unit AVS370 (Schott-Gerate, Hoeim,Germany). Measurements were carried out on a solution ofalginate in NaCl 0.01 M and in the elaboration of the experi-mental data the following Mark–Houwink constants K ¼ 4.8 �10�6 dl g�1, a ¼ 1.15 were used. The average viscosimetricmolecular weight (Mh) turned out to be 146 kg mol�1. Chitosan,low viscous, was purchased from Fluka and used withoutfurther purication. Hexa-Methyl-Di-Silazane (HMDS) waspurchased from Fluka. Calcium chloride, 1-ethyl-3,3-[3-(dime-thylamino)propyl]carbodiimide hydrochloride (EDC), 2-(N-morpholino)ethanesulfonic acid, 4-morpholineethanesulfonicacid monohydrate (MES) were purchased from Sigma. Genipinwas purchased from Wako.

Experimental set-up

The experimental set-up used to fabricate alginate scaffolds hasbeen assembled in our laboratory. It consists of three-motorizedprecision linear stages (Micos): X–Y movement is associatedwith the deposition plane, Z movement is associated with theextrusion system. Deposition paths are generated using a


custom Matlab algorithm in which ber position (spacing andorientation) and scaffolds macroscopic dimensions are set. Theextruder is constituted by two needles, assembled one insidethe other and xed with silicone glue (Fig. 1). The internalneedle (Hamilton) has a 26 G dimension (ID ¼ 260 mm, OD ¼470 mm) and supplies alginate solution; the external needle(Fishman) has a 19 G dimension (ID ¼ 690 mm, OD ¼ 1070 mm)and supplies calcium chloride solution. The internal needleprotrudes out from the external one of approximately 500 mm.Both ows are controlled volumetrically by means of twomicrouidic pumps (Harvard Apparatus).

Characterization of the bers deposited with the coaxial-needle system

To characterize the deposition step using the coaxial needlesystem, the dimension of the bers in scaffolds produced atdifferent ow rates of alginate (Qal) and calcium chloride (Qca)and at different velocities of the deposition plane (FR, feed rate)(Fig. 2) was measured using light microscopy images andImageJ soware. Before acquiring microscopy data, sampleswere kept in the cross-linking solution used during the printingprocess (CaCl2 0.22 M, MES 0.2 M, PH¼ 4.5) for 24 h to let themreach the swelling equilibrium. These measurements werecompared with a simple theoretical model in which the bersare supposed to be cylindrical and their dimension is predictedusing a volumetric equivalence:

Qal ¼ FR � Af

Af ¼ d2

4� p

J. Mater. Chem. B, 2014, 2, 6779–6791 | 6781


Fig. 2 Scheme of the fiber extrusion process and representation ofthe physical parameters responsible in defining the fiber diameter.


Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27


where Af and d are the area and diameter of the circular sectionof the cylindrical ber, respectively (Fig. 2).

Scaffold production and crosslinking

Puried alginate was dissolved in distilled water at a concen-tration of 4% w/v. Calcium chloride was dissolved in distilledwater at a concentration of 0.22 M. The two solutions were putin plastic syringes. Printed scaffolds were designed to have a 10mm � 10 mm � 1 mm or 5 mm � 5 mm � 5 mm totaldimension. Each layer is formed of parallel bers and theorientation of bers in adjacent layers was set to 0–90� for thesquared pattern, 0–60�–120� for the hexagonal pattern. Fiberspacing was set to 200 mm. Fiber diameter was adjusted to 130mm imposing equal alginate and calcium ow rates with the useof syringe microuidic pumps (Qal ¼ Qca ¼ 5 ml min�1) and thesubstrate speed was set to 235 mm min�1. Layer height was setto 90% of the ber diameter in order to favor the surface contactbetween consecutive deposited layers. At the end of the 3DFDprocess, prototyped alginate scaffolds were immersed in a 2 Mcalcium chloride aqueous solution for 24 h before proceedingwith the coating step with chitosan. Chitosan and calciumchloride were dissolved in 0.2 M MES buffer (pH ¼ 4.5) at aconcentration of respectively 1% w/v and 0.1 M. Scaffolds wereimmersed in the chitosan solution for 24 h, and then cross-linked in an ethanol–water mixture (8 : 2) containing genipin(1% w/v) and EDC (one third of total alginate weight, alginate:EDC ¼ 3 : 1) for 48 h. Finally, scaffolds were dialyzed against 1M sodium chloride solution to rapidly remove calcium ions andEDC and then against distilled water until reaching theirnominal conductivity.

6782 | J. Mater. Chem. B, 2014, 2, 6779–6791

Scaffold coding system

In order to identify the scaffolds obtained in the various steps ofcross-linking and to distinguish between those characterized bydifferent geometries of printing, a simple coding system wasadopted. Scaffolds made of calcium-alginate were indicated asCa-Alg. Alginate scaffolds cross-linked with EDC were indicatedas AlgEDC. The code for alginate scaffolds cross-linked with EDC,coated with chitosan and cross-linked with EDC and genipin isAlg–Chit. To distinguish scaffolds printed according to a squareor hexagonal geometry, the suffix s or h was added to Alg–Chit(Alg–Chits, Alg–Chith).

Scaffold characterization

Prototyped and cross-linked scaffolds were visualized by meansof scanning electron microscopy (LEO 1450VP, Oxford Instru-ments), stereo-microscopy (Leica Stereozoom S8 APO equippedwith a EC3 camera) and confocal microscopy (Zeiss LSM 510,Carl Zeiss). Before SEM characterization, samples were driedusing increasing graded ethanol–water mixtures (from 30% to100% ethanol), and then using increasing graded HMDS–ethanol mixtures (from 30% to 100% HMDS), and were nallygold-sputtered (SEM COATING UNIT 953, Agar Scientic). Forconfocal imaging, the uorescence of the genipin bound tochitosan was exploited. Samples immersed in water wereexcited with a 405 nm light source and transmitted light wasdetected. A scansion depth of 300 mm or 30 mm was used tovisualize respectively many layers of the scaffold or highlight theinternal structure of a single ber.

FTIR spectroscopy

The connectivity between alginate–alginate, chitosan–chitosanand alginate–chitosan networks in the printed hydrogels wasexamined using Fourier Transform IR spectroscopy.

The IR analysis of scaffold fragments was carried out, col-lecting a sequence of spectra by means of a Thermo ScienticNicolet 6700 FT-IR equipped with a Golden Gate to performmeasurements in Attenuated Total Reection (ATR) mode. Thesequence of spectra was acquired in the absorbance mode inthe 800–2000 cm�1 range, by executing 200 scans at 4 cm�1

resolution.Hydrogel specimens were oven dried at 80 �C for 4 hours

prior to analysis. The spectra were processed using Happ–Genzel apodization, ATR correction, smoothing and baselinecorrection algorithms. Deconvolution of spectra was carried outby means of PeakFit 4.12. The second and fourth derivatives ofthe input spectra were examined in order to reveal the presenceof hidden peaks.

Mechanical testing

The compressive stress–strain measurements were performedon the water swollen printed hydrogels (Alg–Chits and Alg–Chith) using a Instron 3365, equipped with a 10 N load cell. Thecubic gel samples of 0.5 cm of edge were set on the lower plateand compressed at a strain rate of 1.2 mm min�1 to 70%deformation by the upper plate, which was connected to the




Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27


load cell. Measurements were carried out at room temperature.Elastic moduli were derived from the regression of the linearportion of stress–strain curves. Each measurement was per-formed in quintuplicate. Results are reported as the mean �standard deviation.

Cell culture

The terminally differentiated human HepaRG cells (Life Tech-nologies™, Grand Island, NY, USA) were suspended in Wil-liams' Medium E supplemented with 10% fetal bovine serum,1% GlutaMAX™-I, 100 U per ml penicillin, 5 mg ml�1 of insulin,25 mg ml�1 of streptomycin. The cell seeding procedure wasmodied from that described previously.32b Briey, 5 � 105 cellsin 40 ml of medium were loaded into the UV sterilized bioma-terials. Scaffolds with the entrapped cells were incubated at37 �C in a humidied atmosphere of 5% CO2 for 5 h to allow cellsettlement. Aer this incubation period, a fresh medium wasadded to each well. In parallel experiments, cells, suspended inthe same medium, were seeded on Collagen I-coated plates (3�105 cells per cm2) and cultured in monolayer.

Scanning electron microscopy (SEM)

Cell-seeded materials were rinsed with culture medium andxed with 2% glutaraldehyde in 0.1 M Na-cacodylate buffer with0.1 M sucrose for 12 h. Aer treatment with 1% tannic acid in0.15 M Na-cacodylate buffer for 1 h, the samples were dehy-drated in ethanol solution (from 70% to 100%). The sampleswere then immersed in 100% hexamethyldisalazane (HMDS) for3 min, dried and mounted on aluminum stubs and gold-sput-tered with a thin layer of gold. Specimens were nally examinedusing an XL 30 CP (Philips) SEM.

Viability/cytotoxicity assays

Cell viability was determined throughout the culture by meansof a MTS assay. 20 ml of Cell Titer 96 AQueous One Reagent(Promega, Madison, WI, USA) were added to the cell-seededscaffolds cultured in 100 ml of complete medium and incubatedfor 1 h at 37 �C. MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxy-methoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] bio-reducedproduct was quantied by absorbance at 490 nm (GLOMAX,Promega). Values obtained in the absence of cells wereconsidered as background. A standard curve was assessed toconvert 490 nm O.D. values to the number of viable cells, usingcell densities ranging from 0 to 4 � 105 cells.

To determine the extent of cell damage and toxicity, therelease of lactate dehydrogenase (LDH) by impaired cells wasquantied in supernatants and compared with the LDH contentof the total cell lysates. A LDH-cytotoxicity assay kit (BioVisionInc., Mountain View, CA) was used according to the manu-facturer's instructions. Before the assay, wells were washed withPBS and a fresh medium was added; aer 24 h, this mediumwas withdrawn and 100 ml aliquots were assayed. Absorbancewas measured at 450 nm using a microplate reader. Highabsorbance equating to 100% LDH release was obtained bytreating the entrapped cells with the lysis solution providedwith the kit. Values obtained in the absence of cells were


considered as background. Each experiment was performed intriplicate.

Albumin secretion assay

An enzyme-linked immunosorbent assay (ELISA) Quantitationkit (Bethyl Laboratories Inc., Montgomery, TX, USA) was used tomeasure albumin levels during the culture period (at 2, 6, 9 and15 days). Before the assay, wells were washed with PBS and freshmediumwas added; aer 24 h, this mediumwas withdrawn andaliquots were used for the specic albumin secretion assay.Human albumin was used to establish the standard curve. Thelevels of albumin measured at each time point were normalizedto the cell number obtained by the MTS assay. Each experimentwas performed in triplicate.

Urea secretion assay

Urea secretion was determined by a colorimetric test using theQuantiChrom Urea Assay Kit (BioAssay Systems, Hayward, CA)designed to measure urea directly in biological samples. Theintensity of the color measured is directly proportional to theurea concentration in the sample. Urea secretion was measuredat 450 nm at different intervals of culture (2, 6, 9, 15 days).Twenty-four hours before the assay, wells were washed with PBSand a fresh medium was added; at the time of performing thetest, this medium was withdrawn and aliquots were used for thespecic urea secretion assay. The levels of urea measured ateach time point were normalized to the cell number obtained bythe MTS assay and expressed as mg per 105 viable cells per 24 h.Each experiment was performed in triplicate.

CYP3A4 assay

CYP3A4 activity was measured using a luminescent method(P450-Glo™ Assay, Promega) following the manufacturer'sinstructions, as detailed elsewhere.17 Entrapped cells weretreated for 72 h with dexamethasone 10�4 M to induce theCYP3A4 enzyme activity. Dexamethasone was dissolved indimethyl sulphoxide (DMSO), whose nal concentration neverexceeded 0.1% (v/v). Dishes containing the correspondingamount of vehicle alone (DMSO) were considered as controls.Aer treatment, the medium was replaced with a fresh mediumcontaining a luminogenic CYP3A4 substrate 1 : 40 (luciferin–PFBE) and incubated for 31/2 h. Aer incubation, 25 ml ofmedium was combined with 25 ml of luciferin. Detectionreagent and luminescence were read aer 20 min at roomtemperature. Luminescence background was calculated byadding the luminogenic substrate to the medium without cells.Values obtained aer 6, 9, 15 days of culture were normalized tothe cell number obtained by the MTS assay and expressed asRLU per 105 cells. Each experiment was performed in triplicate.

Statistical analysis

The Student's t-test was used to analyze the statistical signi-cance of the data. Differences with a p value <0.05 wereconsidered signicant.

J. Mater. Chem. B, 2014, 2, 6779–6791 | 6783



Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27


Result and discussion

Three-Dimensional Fiber Deposition (3DFD) is probably themost simple and popular technique within rapid prototyping,in which a 3D object is created by the horizontal deposition ofbers, layer upon layer. Usually, this translates into a rougherapproximation of the 3D computer-generated models whencompared to other RP techniques, like stereolithography or 2-photon polymerization, but this limitation is balanced by amuch simpler and cost-effective hardware set-up.

In this work we present a Three Dimensional Fiber Deposi-tion (3DFD) strategy for the production of alginate hydrogelscaffolds. In contrast to so far developed systems for the 3Ddeposition of alginate bers8 that generally uses a suitableplotting medium to obtain the gelication of the polymer ordeposits pre-gelled alginate,6 we induce the physical gelation ofalginate simultaneously with its extrusion using two coaxialneedles (Fig. 1a): the internal one is fed with an alginateaqueous solution and the external one with a calcium chlorideaqueous solution. As soon as the two ows come into contact atthe tip of the dispensing system, the biopolymer solutionundergoes gelation. This gel adheres to the deposition planeand by moving the dispensing system, a gelled alginate ber islaid down, similarly to the wet-spinning process. The calciumsolution, continuously provided through the external needle,wets the scaffold preventing its drying during the time ofdeposition.

The dimension of the ber depends on the imposed volu-metric ows for the alginate and calcium chloride solutions andon the speed of movement of the extruder (Fig. 3). Predictingber dimension for different values of the control parameters iscrucial for the generation of the codes of movement for the

Fig. 3 Dependence of fiber diameter on the speed of movement of the econstant alginate flow rate and for a given value of FR (Qal¼ 5 ml min�1). Thof the external needle is 690 mm.

6784 | J. Mater. Chem. B, 2014, 2, 6779–6791

robotic dispensing system and for the subsequent success of theprinting process. In Fig. 3 the dependence of ber diameter (d)on FR and Qca is reported, while Qal was kept constant at 5 mlmin�1. These measurements show that the diameter of thedeposited bers decreases from that predicted by the volu-metric equivalence as the outer ow of CaCl2 is increased. Anexcessive outer ow can, on the other hand, disturb the depo-sition of the rst layers of the scaffold, and for Qca > 10 ml min�1

the delity of reproduction of the designed scaffold rapidlydecreases. As shown in Fig. 3, with the alginate concentrationused and dimension of the needles indicated, the range of berdiameter obtained spanned between 120 and 260 mm. Fromthese experiments also emerged that the diameter of theinternal needle through which alginate is supplied to theextruder represents an intrinsic superior limit for the dimen-sion of deposited bers. With our system, a decrease of FR orQca in order to obtain bers bigger than 260 mm producescreased or undulating rather than stretched out bers, oencausing the failure of the printing process.

As far as the inferior limit of ber dimension is concerned,we deposited bers with a maximum velocity of the depositionplane of 385 mm min�1 that led to bers of 100 mm. This doesnot represent a strict inferior limit to ber dimension, as anincrease of either FR or Qca, or a decrease of Qal, may induce tosome extent a further reduction of the ber dimension. If thedimension of the ber is too small with respect to the internalneedle, a loss in accuracy and reproducibility may occur as aconsequence of the possible unwanted contact between theneedle and the deposited structures during the printingprocess. It is also worth noticing that the dimension of the beris independently correlated with the deposition speed and thecalcium outer ow: bers produced at Qal ¼ 3 ml min�1, Qca ¼ 3

xtruder (FR) and (inset) on the outer flow rate of CaCl2 solution (Qca) fore internal diameter of the inner needle is 260 mm, the internal diameter




Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27


ml min�1, FR ¼ 140 mm min�1 have the same dimension asbers produced at Qal ¼ 5 ml min�1, Qca ¼ 3 ml min�1, FR ¼ 235mm min�1 (Qal/FR ¼ cost, d ¼ 160 mm).

The described co-owing extrusion system is mounted on amotorized Z axis while glass slides, used as depositionsubstrates, are moved in the X–Y direction. A computer gener-ated path designs the porosity and architecture of the scaffoldby dening the distance between adjacent bers and thedirection of deposition for each layer, until a three-dimensionalporous material of desired macroscopic dimension is obtained,as shown in Fig. 1b. In Fig. 1c–e a few examples of differentpatterns of deposition are provided. As it can be seen theprocess of writing is very precise and leads to a scaffold with aneat structure. The deposition of bers proceeds in a contin-uous fashion, and as it is clear in Fig. 1e and videos provided inthe ESI,† it maintains its delity also when it rapidly changesthe direction.

At the end of the prototyping step, a brous scaffold ofionically cross-linked alginate was obtained. Since the interac-tion between alginate and calcium is of physical nature, it isreversible and, in particular, destabilization of the egg-boxstructure occurs when the scaffold is exposed to solutionscontaining monovalent cations.18 Furthermore, since gelationof the alginate ber induced by exposure to calcium ions startsfrom the outer shell of the lament of solution extruded, theber deposited is wrapped by a sheet of calcium ions whichhinder the interaction and bonding among bers belonging tosuccessive planes. A relatively high concentration of Na+ ispresent in a common cell culture medium, thus the gelationprocess induced by calcium ions during the prototyping of thescaffold is insufficient to produce chemically stable scaffoldssuitable for cell seeding. In particular, it has experimentallyemerged that, aer being soaked in the RPMI-1640 cell culturemedium, the scaffolds rstly swelled (Fig. 4b), then lost theirordered structure because of the detachment of consecutivelayers (Fig. 4c) and nally the bers dissolved (Fig. 4d).

Fig. 4 Optical micrographs of ionically cross-linked alginate scaffolds (aand f) in water (a and e) and in the RPMI-1640 cell culture medium (b–d ahour; (d) 2 hours; (f) 5 weeks).


To overcome this problem, we decided to proceed with achemical post-cross-linking of alginate preceded by the depo-sition of a chitosan coating on the scaffold bers. Chitosan isanother naturally derived polysaccharide very used in TE whichis a linear copolymer of D-glucosamine and N-acetyl-D-glucos-amine.19 It has been proven to be biodegradable, biocompatible,non-antigenic, non-toxic, bioadhesive, anti-microbial, bioactiveand to have haemostatic capacity.20 It is a positively chargedbiopolymer that, through electrostatic interaction, can bedeposited on negatively charged alginate bers by simplysoaking the alginate scaffold in a chitosan aqueous solution.The formation of an inter-polyelectrolyte network limits berswelling, and the deposited chitosan coating reinforces theattachment between consecutive layers. Moreover, chitosan–alginate hybrids have been shown to induce amore pronouncedcell adhesion with respect to pure alginate21 or chitosan22 scaf-folds, therefore this coating step should also improve cellularaffinity for the substrate. In order to further improve thestability of the scaffolds produced, covalent cross-linking wasemployed to strengthen the chitosan membrane using a natu-rally derived cross-linker, genipin (Fig. 4e and f).23 In the cross-linking reaction, genipin, a small molecule, can freely diffusethrough the alginate–chitosan complex membrane and interactwith the chitosan bound to the alginate gel. The reaction ofgenipin occurs according to two different pathways. The rstand faster one involves a nucleophilic attack of the aminogroups of chitosan on the genipin olenic carbon atom at C3 toform an intermediate aldehyde group, followed by the openingof the dihydropyran ring. The resulting aldehyde group issubsequently attacked by the secondary amine group formed inthe rst step of the reaction. In this way a heterocycliccompound of genipin linked to the glucosamine residue ofchitosan is thereby formed.24 In the second one, the ester groupof genipin is attacked by the amino groups of chitosan leadingto the formation of secondary amide linkages.25 Eventually athree-dimensional interpenetrating network structure is

–d) and chitosan coated chemically cross-linked alginate scaffolds (end f) at different soaking times ((a and e): 24 hours; (b) 10 minutes; (c) 1

J. Mater. Chem. B, 2014, 2, 6779–6791 | 6785



Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27


created within the ber membrane. The uorescence behaviourof genipin bound to chitosan was exploited to inspect theeffective coating of the alginate scaffold bers. Imaging of thescaffold on the whole and of the single bers through CLSManalysis supported this hypothetic structure. From Fig. 5a andb, it is clearly visible that chitosan uniformly covers the scaf-folds. Genipin reacts selectively towards primary amines26 ofchitosan while is inert towards alginate, as a consequence thepresence of a bright external layer (Fig. 5c) is indicative of theformation of new uorescent chitosan–genipin conjugates andvisually demonstrates the presence of a shell-like cross-linkedmembrane of approximately 40 mm thickness (Fig. 5d)surrounding the ber core. The intensity of the uorescence inthe ber external layer is directly proportional to the degree ofcross-linking and to the distribution of the chitosan coating inthe membrane. The stronger the uorescence, the more geni-pin–chitosan conjugates were formed, and thus, the highercross-linking degree in the chitosan coating was attained. As itis evident from Fig. 5d the uorescence signal is stronger at theexternal border of the ber and tends to fade out proceedingtowards the center. This may be due to the restricted diffusionof chitosan into the alginate core blocked by initial chitosanbinding and/or a higher degree of cross-linking at the externalborder of the membrane where more chitosan was deposited. Inany case, this treatment confers to the scaffold the desiredstability in the culture medium as evidenced by Fig. 4e and f.

Since as evidenced by CLSM, chitosan coating and cross-linking by genipin involve only the outer shell of the ber andleave unaltered the core made of alginate, in order to obtaincross-linking throughout the whole of the bers and to enhancethe cross-linking density through the creation of additionalbridges among the different macromolecular pairs (alginate–alginate, alginate–chitosan beside chitosan–chitosan) we addedEDC to the reaction medium (ethanol : H2O ¼ 80/20), besidesgenipin.

The effective occurrence of the various types of cross-linkingamong the macromolecular pairs was investigated by means of

Fig. 5 Confocal microscopy images of squared (a), hexagonal (b) scaffoldcross-linked with EDC and genipin. Scale bars: 100 mm. (d) Plot of fibesoftware). Blue light emission from the scaffolds/fiber is due to the chito

6786 | J. Mater. Chem. B, 2014, 2, 6779–6791

IR spectroscopy. Since the spectrum of the nal scaffold is ofdifficult interpretation due to the complexity of the materialfrom a compositional point of view, we have recorded thespectra of the scaffold at each stage of cross-linking. Further-more, spectra analysis was facilitated by carrying out thedeconvolution of partially overlapped peaks.

In Fig. 6A the spectrum of the scaffold directly produced bythe RP machine i.e. alginate physically cross-linked withcalcium (Ca-Alg) is displayed. Typical bands relative to thesymmetric and asymmetric stretching of the carboxylate groupsat 1581, 1412 and 1026 cm�1 and skeletal vibrations of C–O–Hand C–O–C at 1300, 1244 and 1088 cm�1 are clearly discerniblein the deconvoluted spectrum. In Fig. 6B the spectrum ofcalcium-alginate cross-linked with EDC (AlgEDC) in ethanol–water ¼ 80 : 20 is reported. The additional bands appearing at1666 cm�1 provide clear evidence of the formation of esterlinkages among alginate chains. Furthermore, there is anevident increase of the intensity ratio of the peaks positioned at1088 and 1028 cm�1 in the AlgEDC spectrum with respect to theCa-Alg one. The O–C–C band of the ester group occurs in fact inthe 1150–1030 cm�1 range. The percentage of cross-linking canbe quantied from the ratio of the area of the peak at 1666 cm�1

and the sum of the areas of the peaks at 1666 and 1589 cm�1. Itturns out that the degree of cross-linking of AlgEDC is 12%.Cross-linking operated by EDC involves alginate chains within asingle ber but is not effective in bonding different bers toeach other. This causes the detachment of the bers in theculture medium which is rich of sodium ions, among theothers. This outcome induced us to coat the alginate scaffoldwith chitosan and to carry out a second and third cross-linkingprocess by using EDC and genipin at the same time. Again, tohelp in the interpretation of the spectrum of the nal scaffold,we proceeded rst in a comparative analysis of the IR spectra ofnative chitosan and a genipin-cross-linked chitosan scaffoldproduced by gas-in-liquid foam templating27 as shown inFig. 6C and D. The C]O stretching absorption band of thenewly formed amide at 1649 cm�1 overlaps with the amide I

s and (c) of a section of a fiber. Scaffolds were coatedwith chitosan andr color intensity profile for the segment A–B (obtained with ImageJsan–genipin complex.



Fig. 6 FTIR spectra of: (A) Ca2+-alginate printed scaffold; (B) Ca2+-alginate printed scaffold cross-linked with EDC; (C) chitosan; (D) chitosanscaffold obtained by gas-foaming and cross-linked with genipin; (E) printed scaffold of alginate coated with chitosan cross-linked with EDC andgenipin.


Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27


absorption band of chitosan (1655 cm�1) and makes it becomebroader.28 A new broad peak, that appeared around 1419 cm�1

aer cross-linking by genipin, indicated the presence of ring-stretching of heterocyclic amine. The C–N stretching of amide III

of chitosan at 1233 cm�1 shied to 1260 cm�1 aer cross-linking by genipin. New bands appearing at 1703, 1201 and1236 cm�1 are due to the conjugated acetate present in genipin.Bands at 1594 and 1379 cm�1 are due to the bending vibrationof –NH2 and the stretching absorption of C–N. The apparentabsence of C]C stretching in the cyclic structure of genipin at1628 cm�1 allows presuming that the reaction of the aminogroups of chitosan with the olenic group of genipin was almostquantitative.29

The spectrum of the Alg–Chit scaffold is very complex butbands characteristic of the alginate–chitosan complex arediscernible. A broad band at 1583 cm�1 with a prominentshoulder at 1523 cm�1 as well as a band at 1406 cm�1 can beobserved. The two main bands are assigned to the carbonylvibrations of the carboxylate groups of alginate. The broadnessof these bands arises from overlapping bands from the partly


protonated chitosan moieties and the prominent shouldercorresponding to the 1535 cm�1 band seen in the chitosanreference scaffold. The peak at 1716 cm�1 is due to the carbonylvibration of the carboxylate group in the protonated form.When chitosan was dissolved in hydrochloric acid and put intocontact with the alginate scaffold in water, one reaction thattakes place is the protonation of alginate.30 A related band isthat at 1300 cm�1 which refers to the stretching vibration ofC–O of the COOH group.

The groups of bands in the region between 900 and 1150cm�1 are clearly the overlapping of the skeletal vibrations of thetwo polysaccharides. The characteristic peak of genipin at 1201cm�1 is still clearly visible.

The picture that emerges from the combined analysis of theCLSM images and IR data is as follows: the whole of a ber isrepresented by a network of alginate chains auto-cross-linked byester bonds. The outer shell of the ber (thickness �40 mm), asevident from CLSM images (Fig. 5) and IR data, is composed ofan interpenetrating polymer network of alginate–chitosanchains cross-linked by amide bonds and chitosan–chitosan

J. Mater. Chem. B, 2014, 2, 6779–6791 | 6787



Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27


chains cross-linked by genipin bridges. On a higher lengthscale, bers making up the scaffold are glued together by thegenipin cross-linked chitosan network.

From SEMmicrographs of Fig. 7a, b, c and d it is evident thatthe printed scaffolds exhibit an anisotropic morphology. Fibersare oriented horizontally and have a 200 mm spacing in the x–yplane, and a 120 mm spacing in the z direction (layer height).This should translate into direction dependent mechanicalproperties which is an usual characteristic of many tissues inthe human body. Compressive testing was conducted on Alg–Chits and Alg–Chitq in a perpendicular (t) and parallel (k)direction (Fig. 8a) with respect to the deposition plane obtain-ing two distinct values of the Young modulus (denoted as Etand Ek, respectively). The compressive stress–strain curves ofAlg–Chits are displayed in Fig. 8b. As it can be observed, the t

curve exhibits a linear region for a very extended range of strain(up to 50%) and beyond this limit it tends to rise moderately.The corresponding value of Et is 23.3 kPa (Table 1). Thismechanical behavior matches well with the cross-sectionalmorphology of the scaffold. In the t direction, spacing amongbers (Fig. 7b) is relatively small and upon compression,

Fig. 7 SEM micrographs of squared (a and b) and hexagonal (c and d)alginate scaffolds coated with chitosan and cross-linked with EDC andgenipin. Scale bars: 100 mm.

Fig. 8 (a) Image of a scaffold sample used in compression tests. Arrowstress (s)–strain (3) curves of Alg–Chits determined in a direction perpeninset, an enlarged view of the portion of the curves in the range of strain

6788 | J. Mater. Chem. B, 2014, 2, 6779–6791

bending of bers is relatively less pronounced. The modest riseobserved above 50% strain is ascribed to scaffold densication.A totally different behavior can be observed for the curvedescribing compression in the k direction. In such a case,spacing among bers (Fig. 7a, b) is more relevant and thestress–strain curve consists of three different regions. The rstone is a linear region (up to approximately 5% deformation,inset of Fig. 8b) with a rst increase with stress, which corre-sponds to cell edge bending. The Young's modulus is calculatedfrom the linear slope of the stress–strain curve in the rst regionand, as reported in Table 1, is about three times higher than Et.This is because half of the bers are directed parallel to theapplied force and oppose with the as highest efficacy as possibleto the applied stress. The second one is a stress plateau region,which corresponds to progressive cell collapse by elastic buck-ling. When the opposite cell walls come into contact, cellcollapse ends. This is the third regime, called densication,which denotes a collapse of the cells throughout the material.Upon compression, the stiffness of the cellular materialincreases and converges towards the stiffness of the basematerial.

The mechanical behavior of Alg–Chith (data not shown) isthe same for what concerns compression in thet direction andit is less pronounced in the k one as evidenced by Young moduli(Table 1) since in this case only one third of the bers areparallel to the applied force and its nominal porosity is verysimilar to that of Alg–Chits.

Cell studies

Both squared and hexagonal scaffolds were seeded with Hep-aRG cells. Cells showed a favorable initial colonization of thescaffolds and appeared well organized in the three-dimensionalspace. Scanning electron micrographs of cells cultured onsquared (Fig. 9a–c) and hexagonal (Fig. 9d–f) scaffolds showed ahomogeneous distribution of cells and a high level of adhesionto the bers within the scaffolds. Cells showed a good ability tocolonize and survive on both these biomaterials; they adheredrmly to the materials and not only to each-others, spreadingon the bers rather than promoting the formation of spheroids

s indicate the directions of the applied compression. (b) Compressivedicular (black line) and parallel to the deposition plane (red line). In the0–20% is shown.



Table 1 Young moduli of the printed scaffolds characterized by asquare and hexagonal geometry determined according to a directionperpendicular (Et) and parallel (Ek) to the deposition plane

Sample Et (kPa) Ek (kPa) Nominal porositya

Alg–Chits 23.3 � 4.2 77.2 � 7.2 66.6Alg–Chith 25.7 � 3.8 51.3 � 6.4 66.3

a Porosity was calculated from the scaffold computer model by using analgorithm developed in Visual Basic.


Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27


that are oen observed with more proliferating cell types, suchas HepG2 cells.31 Due to their differentiate phenotype, HepaRGcells are strictly anchorage-dependent; aer adhesion, theyspread on bers and merge forming at multilayer aggregates,which remind the hepatic cords. In fact, due to oxygen andnutrient diffusion difficulties, spheroid culture could be limitedin its ability to be used for long-term cultures, thus making theproposed system more attractive. Cells also showed a high levelof vitality, as revealed by the MTS test (Fig. 10a), suggesting agood capacity of the scaffolds to allow hepatocyte entrapmentand maintenance.

To evaluate indirectly the toxicity of these scaffolds, lactatedehydrogenase (LDH) leakage was monitored in both cellsupernatant and total cell lysate and the ratio of released and

Fig. 9 Scanning electronmicrographs at differentmagnifications of Hepac) or hexagonal (d–f) chitosan-coated alginate scaffolds. A homogeneoobserved within both kinds of scaffolds. Scale bars: (a and d) 200 mm; (b

Fig. 10 Viability (a) and cytotoxicity (b) assays of HepaRG cultured on hscaffolds at 2, 6, 9 and 15 days. Results are the mean � SD of at least th


total LDH was calculated (Fig. 10b). Cells, aer the rst two daysof culture, maintained good viability throughout the entiretime, with very low LDH leakage ratio, considered satisfactoryfor this cell type.32 The somewhat higher levels of LDH leakageinto the medium observed at the rst two days of culture areprobably due to detachment of cells that were not rmlyadherent to the material. Therefore, the subsequent decreasedlevels of LDH leakage in combination with the high number ofviable cells, revealed by the MTS assay, exclude possible adverseeffects of biomaterials against entrapped hepatocytes.

The good efficiency of HepaRG cells cultured in these scaf-folds was evaluated by measuring the levels of albumin secre-tion and urea production, typical functions of welldifferentiated hepatocytes (Fig. 11a and b). The levels of theseproducts were considerably high, with a maximum reachedbetween 6 and 9th day for both the hexagonal (Alg–Chith) andthe squared (Alg–Chits) morphology. Cells grown on both kindsof scaffolds showed better hepato-specic activities than cells inthe monolayer. Similar results were obtained with cells culturedon Alg–Chith and on Alg–Chits, suggesting that both morphol-ogies could represent adequate support and optimal conditionsfor the maintenance of hepatocyte functions.

HepaRG cells, as said above, express a wide range of P450enzymes and respond to selective inducers of these enzymes.Within these new scaffolds, cells are expected to be in an ideal

RG cells cultured for 6 (a, b, d and e) and 9 days (c and f) on squared (a–us distribution of cells and a high level of adhesion to the fibers areand e) 50 mm; (c and f) 20 mm.

exagonal (Alg–Chith) or squared (Alg–Chits) chitosan-coated alginateree different experiments done in triplicate.

J. Mater. Chem. B, 2014, 2, 6779–6791 | 6789


Fig. 11 Albumin secretion (a) and urea production (b) from HepaRG cultured on hexagonal (Alg–Chith) or squared (Alg–Chits) chitosan-coatedalginate scaffolds or in the monolayer (Mono) at 2, 6, 9 and 15 days. At each time point, rates were normalized to the number of cells obtained bythe MTS assay and expressed as mg per 105 cells per 24 h. Results are the mean � SD of at least three different experiments done in triplicate.


Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27


condition to perform detoxication processes. This promptedus to assess the responsiveness to CYP3A4 inductors of theHepaRG cells entrapped into these scaffolds. The activationdegree of CYP3A4 enzyme by 10�4 M dexamethasone (DXM) wasvery high with respect to monolayer cell cultures (Fig. 12) ateach time point examined. The induction was not signicantlydifferent in the two kinds of scaffolds.

These data are in agreement with the morphological obser-vations at the optic and scanning electron microscopy anddemonstrate that both kinds of scaffolds provide an agreeableenvironment to support hepatocyte attachment, growth andperformance, thus promising to become a valuable in vitrosystem on which to study liver specic functions and metabo-lism or to be used as a tool for the screening of new drugs orpotentially toxic substances.

Conclusion

The use of a coaxial dispensing system coupled with a RPmachine permitted the fabrication of an alginate scaffold withan ordered structure and a fully interconnected morphology.The post-treatment of the scaffolds, consisting of coating the

Fig. 12 CYP3A4 activity, in the absence or presence of the inductorDXM 10�4 M, in HepaRG cells cultured on hexagonal (Alg–Chith) orsquared (Alg–Chits) chitosan-coated alginate scaffolds or in themonolayer (Mono) at 6, 9 and 15 days. Values are reported as thepercentage of monolayer control, considered as 100% (black contin-uous line). Results are the mean � SD of at least three differentexperiments done in triplicate.

6790 | J. Mater. Chem. B, 2014, 2, 6779–6791

assembly of bers making up the scaffold with a sheet of aninterpolyelectrolyte complex with chitosan and its subsequentcross-linking with EDC and genipin, assured their structuralstability in the culture medium for a prolonged period of time.The scaffolds based on different deposition layouts are char-acterised by anisotropic mechanical properties.

Preliminary culture tests of a well differentiated humanhepatocyte line showed that the scaffolds were biocompatibleand were densely populated from seeded cells. The high level ofalbumin and urea secretion coupled with the high responsive-ness to the induction of CYP3A4 enzyme make this system anattractive tool for testing the toxicity of new drugs.

The present dispensing technique has the potential ofprinting materials with sophisticated and unmet hierarchicalstructures. For instance, one can imagine to connect the innerneedle of the dispensing system to the outlet of a microuidicdevice which can be used to generate an oil-in-water emulsion.If the dispersed phase is represented by either a non-polymer-izing or polymerizing phase and the external phase by anaqueous solution of a gelling biopolymer it is possible to obtaineither printed materials with porous bers or a compositematerial in which individual bers are made by an inner threadof a hydrophobic polymer and an outer sheet of a hydrophilicpolymer obtaining in this way the double target of reinforcedmaterials and a hydrophilic and biocompatible surface.Furthermore, the method of scaffold manufacturing throughber layering presented here is compatible with differentchemistries. A promising approach is represented by clickchemistry which, under appropriate conditions, is an almostinstantaneous reaction that can be utilized with many properlyfunctionalized biopolymers.

Acknowledgements

The authors thank the “Sapienza” University of Rome (Ateneofunds) for nancial support.

References

1 E. Lavik and R. Langer, Appl. Microbiol. Biotechnol., 2004, 65,1–8.




Publ

ishe

d on

21

July

201

4. D

ownl

oade

d by

Uni

vers

itat P

olitè

cnic

a de

Val

ènci

a on

27/

10/2

014

13:2

1:27


2 (a) J. L. Drury and D. J. Mooney, Biomaterials, 2003, 24, 4337–4351; (b) S. Van Vlierberghe, P. Dubruel and E. Schacht,Biomacromolecules, 2011, 12, 1387–1408.

3 (a) W.-Y. Yeonga, C.-K. Chuaa, K.-F. Leonga andM. Chandrasekaran, Trends Biotechnol., 2004, 22, 643–652;(b) D. W. Hutmacher, M. Sittinger and M. V. Risbud,Trends Biotechnol., 2004, 22, 354–362; (c) S. J. Hollister, Nat.Mater., 2005, 4, 518–524; (d) S. M. Peltola,F. P. W. Melchels, D. W. Grijpma and M. Kellomaki, Ann.Med., 2008, 40, 268–280; (e) F. P. W. Melchelsa,M. A. N. T. Domingosc, J. Kleina, J. Maldaa, P. J. Bartolocand D. W. Hutmacher, Prog. Polym. Sci., 2012, 37, 1079–1104.

4 (a) S. A. Park, S. H. Lee and W. D. Kim,Macromol. Res., 2011,19, 694–698; (b) A. Ovsianikov, A. Deiwick, S. Van Vlierberghe,P. Dubruel, L. Moller, G. Dragerand and B. Chichkov,Biomacromolecules, 2011, 12, 851–858; (c) T. H. Ang,F. S. A. Sultana, D. W. Hutmacher, Y. S. Wong, J. Y. H. Fuh,X. M. Mo, H. T. Loh, E. Burdet and S. H. Teoh, Mater. Sci.Eng., C, 2002, 20, 35–42; (d) T. Billiet, M. Vandenhaute,J. Schelout, S. Van Vlierberghe and P. Dubruel,Biomaterials, 2012, 33, 6020–6041.

5 L. Pescosolido, W. Schuurman, J. Malda, P. Matricardi,F. Alhaique, T. Coviello, P. Rene van Weeren,W. J. A. Dhert, W. E. Hennink and T. Vermonden,Biomacromolecules, 2011, 12, 1831–1838.

6 (a) N. E. Fedorovich, J. R. De Wijn, A. J. Verbout, J. Alblas andW. J. Dhert, Tissue Eng., Part A, 2008, 14, 127–133; (b)G. Vozzi, A. Previti, D. De Rossi and A. Ahluwalia, TissueEng., 2002, 8, 1089–1098; (c) A. Tirella, A. Orsini, G. Vozziand A. Ahluwalia, Biofabrication, 2009, 1, 045002; (d)M. Mariani, F. Rosatini, G. Vozzi, A. Previti andA. Ahluwalia, Tissue Eng., 2006, 12, 547–557.

7 T. H. Ang, F. S. A. Sultana, D. W. Hutmacher, Y. S. Wong,J. Y. H. Fuh, X. M. Mo, H. T. Loth, E. Burdet andS. H. Teoh, Mater. Sci. Eng., C, 2002, 20, 35–42.

8 (a) R. Landers, A. Pster, U. Huebner, H. John,R. Schmelzeisen and R. Muelhaupt, J. Mater. Sci., 2002, 37,3107–3116; (b) R. Landers, U. Hubner, R. Schmelzeisen andR. Mulhaupt, Biomaterials, 2002, 23, 4437–4447.

9 E. Kang, G. S. Jeong, Y. Choi, K. H. Lee, A. Khademosseiniand S.-H. Lee, Nat. Mater., 2011, 10, 877–883.

10 B. Lanfer, U. Freudenberg, R. Zimmermann, D. Stamov,V. Korber and C. Werner, Biomaterials, 2008, 29, 3888–3895.

11 S. Ghosh, S. T. Parker, X. Wang, D. L. Kaplan and J. A. Lewis,Adv. Funct. Mater., 2008, 18, 1883–1889.

12 (a) T. Andersen, J. E. Melvik, O. Gaserød, E. Alsberg andB. E. Christensen, Biomacromolecules, 2012, 13, 3703–3710;(b) M. Dentini, G. Rinaldi, D. Risica, A. Barbetta andG. Skjak-Bræk, Carbohydr. Polym., 2005, 59, 489–499; (c)M. Dentini, G. Rinaldi, A. Barbetta, D. Risica, C. Anselmiand G. Skjak-Bræk, Carbohydr. Polym., 2007, 67, 465–473;(d) A. Barbetta, E. Barigelli and M. Dentini,Biomacromolecules, 2009, 10, 2328–2337.

13 (a) K. M. Yamada and E. Cukierman, Cell, 2007, 130, 601–610; (b) K. L. Schmeichel and M. J. Bissell, J. Cell Sci., 2003,116, 2377–2388; (c) F. Pampaloni, E. G. Reynaud andE. H. K. Stelzer, Nat. Rev. Mol. Cell Biol., 2007, 8, 839–845.


14 K. P. Kanebratt and T. B. Andersson, Drug Metab. Dispos.,2008, 36, 1444–1452.

15 (a) A. Guillouzo, A. Corlu, C. Aninat, D. Glaise, F. Morel andC. Guguen-Guillouzo, Chem.-Biol. Interact., 2007, 168(1), 66–73; (b) S. Antherie, C. Chesne, R. Li, C. Guguen-Guillouzoand G. Guillouzo, Toxicol. in vitro, 2012, 26, 1278–1285.

16 (a) S. B. Leite, I. Wilk-Zasadna, J. M. Zaldivar, E. Airola,M. A. Reis-Fernandes, M. Mennecozzi, C. G. Guillouzo,C. Chesne, C. Guillou, P. M. Alves and S. Coecke, Toxicol.Sci., 2012, 130, 106–116; (b) P. Gunness, D. Mueller,V. Shevchenko, E. Heinzle, M. Ingelman-Sundberg andF. Noor, Toxicol. Sci., 2013, 133, 67–78; (c) M. Yamada,R. Utoh, K. Ohashi, K. Tatsumi, M. Yamato, T. Okano andM. Seki, Biomaterials, 2012, 33, 8304–8315.

17 A. Stampella, A. Papi, G. Rizzitelli, M. Costantini, C. Colosi,A. Barbetta, M. Massimi, L. Conti Devirgiliis and M. Dentini,J. Mater. Chem. B, 2013, 1, 3083–3098.

18 M. A. LeRoux, F. Guilak and L. A. Setton, J. Biomed. Mater.Res., 1999, 47, 46–53.

19 M. Rinaudo, Prog. Polym. Sci., 2006, 31, 603–670.20 T. Dai, M. Tanaka, Y. Y. Huang and M. R. Hamblin, Expert

Rev. Anti-Infect. Ther., 2011, 9, 857–879.21 N. Iwasaki, S.-T. Yamane, T. Majima, Y. Kasahara,

A. Minami, K. Harada, S. Nonaka, N. Maekawa,H. Tamura, S. Tokura, M. Shiono, K. Monde andS.-I. Nishimura, Biomacromolecules, 2004, 5, 828–833.

22 H.Chen,W.Ouyang,M. Jones,C.Martoni, T.Haque,R.Cohenand S. Prakash, Cell Biochem. Biophys., 2006, 47, 159–167.

23 F. L. Mi, H. W. Sung and S. S. Shyu, J. Polym. Sci., Part A:Polym. Chem., 2000, 38, 2804–2814.

24 F. L. Mi, Biomacromolecules, 2005, 6, 975–987.25 S. Fujikawa, S. Nakamura and K. Koga, Agric. Biol. Chem.,

1988, 52, 869–870.26 (a) A. Bartkowiak and D. Hunkeler, Chem. Mater., 1999, 11,

2486–2492; (b) O. Gaserod, O. Smidsrod and G. Skjak-Braek, Biomaterials, 1998, 19, 1815–1825.

27 (a) A. Barbetta, A. Carrino, M. Costantini andM. Dentini, SoMatter, 2010, 6, 5213–5224; (b) A. Barbetta, G. Rizzitelli,R. Bedini, R. Pecci and M. Dentini, So Matter, 2010, 6,1785–1792; (c) I. Chimenti, G. Rizzitelli, R. Gaetani,F. Angelici, V. Ionta, E. Forte, G. Frati, O. Schussler,A. Barbetta, E. Messina, M. Dentini and A. Giacomello,Biomaterials, 2011, 32, 9271–9281.

28 F.-L. Mi, H.-W. Sung and S.-S. Shyu, J. Polym. Sci., Part A:Polym. Chem., 2000, 38, 2804–2814.

29 K. Zhang, Y. Qian, H. Wang, L. Fan, C. Huang, A. Yin andX. Mo, J. Biomed. Mater. Res., Part A, 2010, 95, 870–881.

30 G. Lawrie, I. Keen, B. Drew, A. Chandler-Temple, L. Rintoul,P. Fredericks and L. Grøndahl, Biomacromolecules, 2007, 8,2533–2541.

31 J. Friedrich, C. Seidel, R. Ebner and L. A. Kunz-Schughart,Nat. Protoc., 2009, 4, 309–324.

32 (a) M. R. McGill, H. M. Yan, A. Ramachandran, G. J. Murray,D. E. Rollins and H. Jaeschke, Hepatology, 2011, 53, 974–982;(b) M. De Colli, M. Massimi, A. Barbetta, B. Di Rosario,S. Nardecchia, L. Conti Devirgiliis and M. Dentini, Biomed.Mater., 2012, 7, 1–13.

J. Mater. Chem. B, 2014, 2, 6779–6791 | 6791


Documents

Rapid prototyping of chitosan-coated alginate scaffolds through the use of a 3D fiber deposition technique