Acta Biomaterialia 6 (2010) 429–435

Contents lists available at ScienceDirect

Acta Biomaterialia

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

Preparation and characterization of polysaccharidic microbeads by a microfluidictechnique: Application to the encapsulation of Sertoli cells

L. Capretto a,1, S. Mazzitelli a,1, G. Luca b, C. Nastruzzi a,*

a Department Chemistry and Technology of Drugs, Via del Liceo 1, 06100 Perugia, Italyb Department of Experimental Medicine and Biochemical Sciences, University of Perugia, Perugia, Italy

a r t i c l e i n f o

Article history:Received 20 February 2009Received in revised form 5 August 2009Accepted 19 August 2009Available online 22 August 2009

Keywords:MicrobeadsMicrofluidicsCell encapsulationSertoli cells

1742-7061/$ - see front matter � 2009 Acta Materialdoi:10.1016/j.actbio.2009.08.023

* Corresponding author. Address: Dipartimento dFarmaco, University of Perugia, Via del Liceo 1, 0610075 5852057.

E-mail address: nas@unipg.it (C. Nastruzzi).1 The first and second authors contributed equally t

a b s t r a c t

Polysaccharides (e.g. alginate or agarose) represent a class of polymers commonly employed for the prep-aration of microparticles for cell entrapment and tissue engineering applications. The present workdescribes the production and characterization, by a microfluidic approach, of microbeads constitutedof alginate and alginate/agarose blends, for the encapsulation of eukaryotic cells. The general productionstrategy is based on the formation of water-in-oil multiphase flow by a ‘‘Y” junction squeezing mecha-nism. The presented data demonstrate that the gelation step represents the crucial point for the produc-tion of morphologically excellent microbeads. In this respect, microfluidic methods appear to be aneffective procedure for the production of microbeads intended for cell encapsulation, as proved by thehigh viability and maintenance of functional capability demonstrated by the encapsulated Sertoli cells.

� 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Transplantation of encapsulated cells in spherical-shaped de-vices (microbeads or microcapsules) is an appealing strategy fortreatments of a wide variety of diseases such as cancer, diabetes,Parkinson’s and other endocrine disorders [1–3]. Living cellsembedded in microbeads, acting as scaffolds, have also foundapplication in other fields, including cell culture and tissue engi-neering [4]. The main advantage of these devices is representedby the presence of an immunoisolating membrane, which enablesthe transplantation of non-self cells and tissues without the needfor immunosuppressive regimens.

Successful clinical use of encapsulated cells strongly depends ona number of crucial characteristics such as: (i) the morphologicaland dimensional properties; (ii) mechanical stability; (iii) biocom-patibility; and (iv) molecular exchangeability of microbeads. Asscaffold materials, polysaccharides have largely been used. Poly-saccharides (such as alginate, agarose or chitosan) indeed possessadequate mechanical properties, permit the exchange of moleculesand can be transformed into spherical gelled solid particles by mildprocedures, preserving the cell viability.

In this respect, the gelling property of alginate is due to thestacking of guluronic acid (G) blocks with the formation of ‘‘egg-

ia Inc. Published by Elsevier Ltd. A

i Chimica a Tecnologia del0 Perugia, Italy. Tel./fax: +39

o the project.

box” calcium-linked junctions [5], while agarose provides the for-mation of thermoreversible gels by double helices, stabilized bythe presence of water molecules bound inside the double helicalcavity.

Alginate microspheres have been used for the encapsulation ofa wide variety of biologically active agents, including proteins,antibodies, DNA and cells [6,7], owing to the relatively mild cross-linking conditions of preparation.

Recently, the production of alginate microbeads was also accom-plished by microfluidic procedures. Choi and collaborators reportedthe preparation of alginate beads using a polydimethylsiloxane-based chip, including two injection lines for alginate solution andCaCl2, and an injection line for the cell suspension [8]. The capsulesproduced were intended for the encapsulation of GFP-yeast cells,which is a stronger cell model than primary mammalian cells. Inaddition, the capsules contained only a few cells per bead.

At the moment, the typical methods of preparating cells con-taining agarose microbeads are based on the extrusion of awater-in-oil dispersion [9,10].

To the best of the authors’ knowledge, there is only one paperconcerning cell-enclosing alginate/agarose microcapsules. Algi-nate–agarose subsieve-size capsules were produced by a dropletgenerator-based procedure [11]. The capsules produced, enclosingfeline kidney cells, were relatively polydisperse, and the number ofenclosed cells/capsule was relatively scarce. This paper focuses ontwo novel features: (a) the new gelation process (ionic plus thermalgelation), which represents the key step to obtaining microbeads;and (b) the production of excellent microbeads in terms of perfectspherical shape, narrow size distribution and smooth surface.

ll rights reserved.

430 L. Capretto et al. / Acta Biomaterialia 6 (2010) 429–435

The authors chose to entrap Sertoli cells (SC), since this peculiarcell population has recently gained great importance, being able toimprove in vitro and in vivo functions and differentiation of manycell types, including pancreatic islets and neurons [12,13].

The production strategy presented here for microbead produc-tion is based on a microfluidic technique. Microfluidic devices wereconveniently applied as new tools for the formation of multiphasicregimes of flow [14], later converted in different ways, in highlymonodisperse spherical polymeric microparticles. As recently re-viewed, microfluidic procedures allow emulsion droplets to begenerated with production rates on the order of 1 kg day�1, a ratethat is sufficient for commercial production [15].

The chemical nature of the droplet phase (disperse phase)determines the next step, in which the droplets are transformedinto microparticles by a consolidation procedure. Droplets contain-ing monomers can be solidified by means of thermally initiated orUV-initiated polymerization [16]. Alternatively, droplets of poly-mer dispersions can be hardened by different procedures, includ-ing solvent evaporation [17], chemical reactions [18] or ioniccrosslinking [19].

The main aim of the present study was to produce microbeadsintended for cell encapsulation. Microbeads with a mean diameterbetween 300 and 600 lm are generally considered optimal for cellencapsulation. In fact, the size of cell clusters (i.e., pancreatic islets)can usually exceed hundreds of microns, and thus microbeads can-not be smaller than 200 lm.

2. Materials and methods

2.1. Materials

The polymers used were agarose Type VII (Low Gelling Temper-ature; appearance, white to off-white powder; solubility, clear tohazy, colourless to light yellow solution at 10 mg ml�1 in waterwith heat, gel strength, NTL 200 g cm�2 at 1.0%; gel point, 26–30 �C at 1.5%; electroendosmosis (-mr), NMT 0.1; loss on drying,NMT 10%) (Sigma Aldrich, Germany) and sodium alginate IE-1105 (viscosity, 20.0–40.0 cP; pH, 6.0–8.0, C = 1%, H2O) (InotechBiosystem International, Switzerland). For the gelation of polymerdroplets, barium chloride dehydrate (Sigma Aldrich, Germany) wasemployed. Sunflower-seed oil (Collina d’oro, Italy) was used as thecontinuous (external) phase. All other chemicals were from Flukaand were of the highest purity available.

For the preparation of microbeads a ‘‘Snake mixer slide” chip(Thinxxs, Germany) with squared channels (320 � 320 lm) wasused.

2.2. SC isolation

SC were isolated from testis of prepubertal neonatal Large-White piglets aged 7–15 days, as previously described [10]. Thepiglets were anaesthetized with 0.1 mg kg�1 azaperon (Stresnil40 mg ml�1, Janssen, Bruxelles, Belgium) and 15 mg kg�1 ketamine(Imalgene 100 mg ml�1, Gellini Farmaceutici, Aprilia, Italy) co-administered intramuscularly.

The technical procedure is based on the excision of testes: uponremoval of their fibrous cap, the testes were finely chopped to obtaina homogeneous dense tissue. The tissue obtained was enzymaticallydigested with a 2 mg ml�1 collagenase P (Roche Diagnostics, S.p.A.,Monza, Italy) solution in Hanks’ balanced salt solution (HBSS)(Sigma Chemical Co., St. Louis, USA) until the separation of seminif-erous tubules. The tubules collected were washed twice in HBSS andspun down at 500 rpm, then the tubules were incubated with a HBSSsolution containing trypsin (2 mg ml�1) and DNAse I (Sigma). Afterthe second digestion, the trypsin solution was diluted 1:1 with

HBSS + 20% FBS to halt trypsin digestion and washed twice by grav-ity for 15 min in order to sediment SC aggregates, while peritubularcells remained in the supernatant and were aspirated off. The tissuepellet was washed twice in HBSS and spun at 800 rpm for 3 min. Thepellet was then resuspended in HBSS, passed through a 500 lMstainless steel mesh, centrifuged at 800 rpm for 5 min and resus-pended by pipetting for 7 min in 1 M glycine, 2 mM EDTA HBSS pH7.2, in order to eliminate the residual Leydig and peritubular cells.The residual tubules were precipitated for 5 min at 800 rpm, washedtwice with HBSS, and finally collected and culture maintained inHAM F12 (Euroclone) supplemented with 0.166 nM retinoic acid(Sigma) and 5 ml/500 ml of insulin-transforming selenium (ITS)(Becton Dickinson #354352) in 95% air/CO2 at 37 �C. After 3 daysof in vitro culture maintenance, the SC were incubated with20 mM Tris(hydroxymethyl)aminomethane hydrochloride (Tris)buffer (Sigma) at pH 7.2, for 5 min, in order to eliminate the residualgerminal cells. After the Tris had been washed off, SC were culturedin the same conditions in 75 cm2 cell culture flasks, as reportedabove.

2.3. Microbeads preparation

For injection into the microfluidic chip of both dispersed andcontinuous phases, a syringe pump (KDS Model 100 Series, Kd Sci-entific) was employed. Polymer dispersions (alginate or alginate/agarose blends) were used as aqueous internal phase (water phase,WP) and slowly injected into a reagent inlet of the squeezinggeometry microchannel. The second immiscible liquid (oil phase,OP) was injected into the other inlet as a continuous phase. Poly-mer dispersion was forced into the OP at the junction of thesqueezing channel to form a multiphasic flow (droplets) repre-sented by a w/o emulsion. Finally, the microdroplets were gelledinto a BaCl2 solution (1.5%, w/v), in order to produce the final con-solidated microbeads (Fig. 1).

2.4. Morphological characterization of microbeads

Since a uniform shape and a smooth and regular microbeadsmembrane are essential to ensure the viability of the microencap-sulated cells and the biocompatibility of the immobilization device,the microparticle morphology, size and size distribution were as-sessed under inverted phase and stereomicroscopy, by countingat least 300 particles/batches.

2.5. Rheological studies

Rheological measurements were carried out by means of aStresstech HR controlled stress rheometer (Reologica Instruments,AB Milano, Italy) equipped with cone-plate geometry (diameter40 mm and angle 1�). The alginate/agarose gels were characterizedby oscillation measurements (oscillation stress sweep at 37 �C).The samples were applied to the lower plate to ensure that gelshearing did not occur. The frequency sweep measurements wereperformed using the frequency range 0.1–10 Hz, and the stressvalue previously determined in the linear viscoelastic region.

2.6. Determination of viability and function of encapsulated SC

The viability of SC encapsulated in alginate/agarose microbeadswas assessed after different lengths of in vitro culture (immediatelyafter microbead preparation and at days 3, 6, 9 and 12) by doublestaining with ethidium bromide (Sigma) and fluorescein diacetate(FDA, Sigma). Cells were visualized under a fluorescence micro-scope (Nikon, Optiphot-2, Nikon Corporation, Tokyo, Japan) usingthe filter block for fluorescein. Dead cells were stained in red, whileviable cells appeared green. The quantification of the viability of SC

Fig. 1. Schematic representation of the microfluidic system set-up for the preparation of polysaccharidic microbeads, including drawing of the multiphase flow generator bythe ‘‘Y” junction squeezing mechanism (upper panel). Microphotographs of the multiphase flow formation, bar = 300 lm (lower panel).

A B

D

E F

0 100 200 300 400 5000

20

40

60

80

100

microbead diameter (um)

Cum

ulat

ive

frequ

ency

unde

rsiz

e (%

)

C

Fig. 2. Dark field stereophotomicrographs of polysaccharidic microbeads. (A) Alginate microbeads prepared by ‘‘external gelation” procedure; (B) microbeads consisting of(B) 9:1, (C) 4:1, (D) 1:1, w/w alginate/agarose blends. Alginate/agarose (1:1 blend, w/w) microbeads containing SC (E). In the insets are reported dark fieldstereophotomicrographs showing, at higher magnification, the microbeads surface and morphology. Bars correspond to 400 lm (A–D) and 150 lm (E).

L. Capretto et al. / Acta Biomaterialia 6 (2010) 429–435 431

Table 1Effect of the different experimental parameters on the morphological characteristics of alginate/agarose microbeads.

Batch Polymer Polymer ratio (w/w) Polymer concentration (%, w/w) Mean diameter, ±SD (lm) Notes

Alg1 Alginate – 2 – Presence of very pronounced tailsAlg9–1 Alginate/agarose 9:1 2 – Presence of pronounced to moderate tailsAlg6–1 Alginate/agarose 6:1 2 233.2 ± 28.5 Presence of moderate tailsAlg4–1 Alginate/agarose 4:1 2 239.7 ± 30.2 Presence of few moderate tailsAlg2–1 Alginate/agarose 2:1 2 247.9 ± 27.6 Presence of faint tailsAlg1–1 Alginate/agarose 1:1 2 249.3 ± 25.2 Spherical shape and narrow size distribution

Experimental parameters: water phase (flow rate 50 lL min�1); oil phase, sunflower-seed oil (flow 110 lL min�1); gelling bath, 1,5% (w/v) BaCl2, 5 �C, microchip channel size,320 � 320 lm.

0

1000

2000

3000

4000

5000

6000

7000

8000

0

500

1000

1500

2000

0 2 4 6 8 10 12

G' (

Pa)

G" (P

a)

frequency (Hz)

Fig. 3. Frequency sweep tests showing the frequency dependence of the elastic G0

(solid lines, open symbols) and loss G00 (dotted lines, closed symbols) moduli ofalginate/agarose gels. Gels were prepared using different weight ratios betweenalginate and agarose, namely: 9:1 (circles), 4:1 (squares) and 1:1 (diamonds). Datarepresent the average of three independent determinations.

432 L. Capretto et al. / Acta Biomaterialia 6 (2010) 429–435

encapsulated in alginate/agarose microbeads was determined afterdissolution of microbeads by treatment with an EDTA solution(50 mM, pH 8) for 10 min at 37 �C. Thereafter, cells were washedthree times with PBS, stained and examined under a fluorescencemicroscope, by counting at least 800 cell/batches. The final viabil-ity data represent the mean of three determinations made in trip-licate (on independent microbead batches) ± SD.

The evidence of the functions of the encapsulated SC was deter-mined by assessing the cell a-aromatase activity as previously de-scribed [12]. Briefly, 106 SC as monolayers or encapsulated inmicrobeads were cultured for 3 days in the absence or presenceof 1 lg ml�1 of follicular stimulating hormone (FSH) (Serono,Rome, Italy); thereafter, 0.2 lg ml�1 of testosterone enanthate(SIT, Pavia, Italy) were added to the cells’ culture medium. After8 h of treatment, the conditioned medium was withdrawn and as-sayed for the presence of 17-b-estradiol (E2). E2 concentrationswere determined by direct chemiluminescent technology (ADVIACentaur, Estradiol-6 III, Bayer Diagnostics, Germany) (intra-assayCV < 4.0%; inter-assay CV = 6.0%). The reported data represent themean of three determinations made in triplicate (on independentmicrobead batches) ± SD.

2.7. In vivo biocompatibility of encapsulated SC

Under general anaesthesia, induced by intra-peritoneal co-administration of 100 mg kg�1 ketamine (Parke-Davis/Pfizer,Karlsruhe, Germany) and 15 mg kg�1 xylazine (Bayer, Leverkusen,Germany), microbeads were collected in a sterile transfer pipetteand delivered into the peritoneal cavity of the recipient femaleNOD mice, weighing �25 g. (Harlan, Italy) through a small abdom-inal incision. Each mouse received 106 SC encapsulated in alginate/agarose (1:1 blend, w/w) microbeads. The body weight of eachrecipient was monitored. At 4 months of transplant, the microcap-sules were retrieved to evaluate their morphology, mechanical sta-bility and biocompatibility. After anaesthesia of the animals,performed as above described, the microbeads were retrieved byperitoneal washing with warm saline. The general characteristicsof the retrieved microcapsules were determined by examiningthe extent of the fibrotic overgrowth under phase contrast micros-copy, and morphological features such as microcapsule sphericity,surface smoothness and, finally, viability by double staining withethidium bromide and fluorescein diacetate.

3. Results and discussion

This study describes the development and optimization of anew method of preparation for polysaccharidic microbeads basedon a microfluidic approach.

Special regard was given to the morphological and dimensionalcharacteristics of the microbeads, with the specific aim of entrap-ping living cells, possibly avoiding the frequently encounteredmorphological problems represented by the presence of coales-cences and irregular tail-shaped microbeads.

As a general approach, a chip with a ‘‘Y” shaped microchannelconfiguration was employed for the formation of the w/o multi-phase flow regime. The w/o flow consisted of a water dispersionof a polysaccharide into a continuous sunflower-seed oil phase.Afterwards, the microdroplets generated were conveniently con-verted into gelled microbeads by gelation procedures (see schemein Fig. 1).

First, the possibility of producing pure alginate microbeads byan ‘‘external gelation” procedure was tentatively tested. Injectingthe two immiscible phases into the microfluidic channels, a multi-phase flow was easily obtained by tuning their relative flow rates.Subsequently, the alginate microdroplets formed were collectedinto a gelling BaCl2 bath, allowing gelation of the alginate droplets.

As reported in Fig. 2A, the microbeads obtained were unfortu-nately characterized by a very pronounced tail-shape (see arrowsin the inset of Fig. 2A). This particular shape was attributed tothe slow pass of the Na–alginate liquid droplets through the OP/gelling bath interface.

In order possibly to solve this main drawback, subsequentexperiments were conducted using alginate/agarose blend disper-sions as water phases, with the purpose of starting the gelationprocess of the polymeric droplets when they were still residentinto the chip microchannel.

From analysis of the microphotographs reported in Fig. 2B–D, itis noticeable that the progressively increasing amounts of agarose(with the respect to alginate) results in an improvement of the finalmorphological characteristics of the microbeads obtained.

In fact, the microbeads consisting of small amounts of agarose(alginate/agarose from 9:1 to 4:1, w/w) still presented the un-wanted tail-shape (even if in a less pronounced way). Only whenthe agarose content was increased to a 1:1 (alginate/agarose,

L. Capretto et al. / Acta Biomaterialia 6 (2010) 429–435 433

w/w), were the microbeads produced characterized by a perfectlyspherical shape, without signs of coalescence (Fig. 2D). Interest-ingly, together with amelioration of the microbeads’ morphology,a slight and progressive increase in the mean particle diameterwas observed when the agarose content was raised (see the meandiameters reported in Table 1, and the size distribution analysis re-ported in Fig. 2F). This behaviour was tentatively attributed to thedifferent hardening process characteristic of agarose, in fact, aga-rose gelation does not involve shrinking, owing to the complexa-tion of calcium ions.

In spite of these results indicating that only using a 1:1 alginateand agarose weight ratio resulted in optimal microbeads (seeabove), rheological characterization was performed by oscillatorymeasurements of gels consisting of alginate/agarose in different ra-tios (namely, 9:1, 4:1 and 1:1).

Non-destructive oscillatory measurements were indeed per-formed in order to obtain information about the effect of thechange in ratio between alginate and agarose.

B

C D

A

E

Fig. 4. Fluorescence micrographs showing the viability of SC encapsulated in alginate/aculture; (C) after 6 days of cell culture; (D) after 9 days of cell culture; (E) 12 days of cblend, w/w), as described in the experimental section. (F) Comparative analysis of the vmean of three determinations made in triplicate (on independent microbead batches) ±

After mechanical stress, viscoelastic materials (such as alginate/agarose gels) show contemporaneously viscous and elastic flow. Inthis respect, oscillation measurements furnish data regarding en-ergy stored (storage modulus G0) and dissipated energy (loss mod-ulus G00).

Gel measurements were performed first by increasing the stressfrom 0.1 to 100 Pa, working at a constant frequency of 1 Hz. Thistest permitted individuation of the linear viscoelastic region, inwhich it is possible to recognize the stress value useful to ensureinstantaneous recovery after the removal of the applied force. Suc-cessively, oscillatory measurements were performed over a fre-quency range 0.1–10 Pa and at constant stress, identified in thelinear viscoelastic region.

As shown in Fig. 3, storage modulus (G0) predominated in all gels,demonstrating that these formulations were able to keep the fur-nished energy (applied stress) maintaining the primitive structure.

In particular, increasing the agarose percentage content causeda progressive improvement in both G0 and G00.

F

0

20

40

60

80

100

0 4 8 12 16

Cel

l via

bilit

y (%

)

cell culture (days)

garose microbeads: (A) immediately after the preparation; (B) after 3 days of cellell culture. SC encapsulating microbeads were prepared with alginate/agarose (1/1iability of free (solid bars) and encapsulated (stripped bars) SC; data represent theSD. Bar = 100 lm.

0

50

100

150

200

250

300

350

0 1 2 3 4 5 6 7cell culture (days)

17-ß

-est

radi

ol/1

x106 c

ells

(pg/

ml) A

B

C

Fig. 5. (A) Functional analysis and (B,C) biocompatibility of alginate/agarose (1:1blend, w/w) microbeads containing SC. As a specific marker to evaluate thefunctional capacity of the encapsulated SC, the a-aromatase activity (expressed in

434 L. Capretto et al. / Acta Biomaterialia 6 (2010) 429–435

Once the optimal experimental set-up and polymer concentra-tion resulting in microbeads with an excellent morphology andnarrow size distribution (Fig. 2D) had been selected, a new seriesof experiments was started, aimed at validating the entire experi-mental set-up for the encapsulation of living eukaryotic cells,namely SC. As reported in Fig. 2E, the stereophotomicrographclearly indicated that SC were conveniently encapsulated into algi-nate/agarose microbeads prepared with alginate/agarose (1/1blend, w/w), with notable shape and surface, in terms of microbeadmorphology.

Moreover, the fluorescence photomicrographs reported inFig. 4A–D demonstrated that SC remain highly viable after encap-sulating up to 16 days of cell culture. Fig. 4E reports the entire via-bility profile comparing free and encapsulated SC. These resultsshow that the encapsulated cells display viability strictly compara-ble with the free ones. In this respect, it is to be underlined that,although immortalized cell lines have provided invaluable infor-mation about cell biology to researchers, primary cell culture offersa more relevant system for the study of cell function, disease statesand patient therapy. However, working with primary cells in cul-ture presents numerous challenges. Primary cell cultures are sensi-tive to apoptosis, owing to contact inhibition, serum concentrationand their three-dimensional (3D) environment, and many condi-tions for optimal growth and proliferation of primary cells remainunknown. Generally, primary cells can be maintained in culture forperiods ranging between 3 and 15 days; after that, cells usually die.To evaluate the functional capability of the encapsulated SC, thea-aromatase activity (in the presence or absence of FSH) was ana-lysed. As depicted in Fig. 5, the a-aromatase activity, expressed interm of 17-b-estradiol biosynthesis, remains statistically un-changed in free and encapsulated SC in both the presence (squares)and the absence (circles) of FSH, for the entire period analysed(6 days of cell culture).

Finally, in order to evaluate the mechanical stability during theimplantation procedure and the biocompatibility of alginate/aga-rose (1:1 blend, w/w) microbeads containing SC, the microbeadswere transplanted into the peritoneal cavity of NOD mice, as de-scribed in the experimental section. The results obtained, summa-rized in Fig. 5B and C, demonstrate that the microbeads weremechanically stable, as proved by the morphological intactness,with the majority of the microcapsules (over 85%) appearing freeof any fibrotic tissue overgrowth at 4 months from transplant(Fig. 5B). In addition, the viability of SC encapsulated in alginate/agarose (1:1 blend, w/w) microbeads at 4 months of transplanta-tion was extraordinarily high.

term of 17-b-estradiol biosynthesis) was analysed (A). Open symbols, plain lines:free SC, closed symbols; dotted lines, SC encapsulated in alginate/agarose micro-beads. a-aromatase activity was evaluated for the indicated length of time in theabsence (circles) or in the presence (squares) of 1 lg ml�1 of FSH. Data representthe mean of three determinations made in triplicate (on independent microbeadbatches) ± SD. Examination of a typical batch of microbeads containing SC uponretrieval after 4 months from the peritoneal cavity, bright field (B) and after doublestaining with ethidium bromide and fluorescein diacetate. Bar corresponds to70 lm.

4. Conclusions

In conclusion, this paper confirms that microfluidic methodsappear to be a promising procedure for the preparation of micro-particles intended for cell encapsulation and tissue engineeringapplications.

In particular, the flow characteristics within microfluidic chipchannels and the high dimension precision with which the chipsare produced result in hydrogel-based microparticles with highlycontrolled morphological and dimensional properties. In this re-spect, microfluidic techniques can be suitably used to obtainmicroparticles without the common morphological defects oftenassociated with other production methods [20]. Batch emulsifica-tion, electrostatic or vibrational technologies are indeed prone tothe production of microparticles with the presence of ‘‘tails”,coalescences, surface irregularity and large size distribution[21,22].

The data reported demonstrate that the conversion of liquiddroplets into solid particles represents the critical point for the

production of morphologically acceptable microbeads by themicrofluidic approach. In this respect, the use of alginate/agaroseblends appears to be particularly appropriate for producing spher-ical, smooth monodisperse microbeads.

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figures 2, 4, and 5, aredifficult to interpret in black and white. The full colour images canbe found in the on-line version, at doi:10.1016/j.actbio.2009.08.023.

L. Capretto et al. / Acta Biomaterialia 6 (2010) 429–435 435

References

[1] Joki T, Machluf M, Atala A, Zhu JH, Seyfried NT, Dunn IF, et al. Continuousrelease of endostatin from microencapsulated engineered cells for tumortherapy. Nat Biotechnol 2001;19:35–9.

[2] De Vos P, Fass MM, Strand B, Calafiore R. Alginate-based microcapsules forimmunoisolation of pancreatic islets. Biomaterials 2006;27:5063–617.

[3] Wikstro J, Elomaa M, Syva H, Kuokkanen J, Yliperttula M, Honkakoski P, et al.Alginate-based microencapsulation of retinal pigment epithelial cell line forcell therapy. Biomaterials 2008;29:869–76.

[4] Wang L, Shelton RM, Cooper PR, Lawson M, Triffitt JT, Barralet JE. Evaluation ofsodium alginate for bone marrow cell tissue engineering. Biomaterials2003;24:3475–81.

[5] Martinsen A, Skjak Braek G, Smidsrod O. Alginate as immobilization material: I.Correlation between chemical and physical properties of alginate beads.Biotechnol Bioeng 1989;33:79–89.

[6] Coppi G, Iannuccelli V, Leo E, Bernabei MT, Camerini R. Protein immobilizationin crosslinked alginate microparticles. J Microencapsul 2002;19:37–44.

[7] Albarghouthi M, Fara DA, Saleem M, El-Thaher T, Matalka K, Badwan A.Immobilization of antibodies on alginate-chitosan beads. Int J Pharm2000;206:23–34.

[8] Chang-Hyung C, Jae-Hoon J, Young Woo R, Dong-Pyo K, Sang-Eun S. Chang-SooLee Generation of monodisperse alginate microbeads and in situ encapsulationof cell in microfluidic device. Biomed Microdevices 2007;9:855–62.

[9] Sakai S, Hashimoto I, Kawakami K. Production of cell-enclosing hollow-coreagarose microcapsules via jetting in water–immiscible liquid paraffin andformation of embryoid body-like spherical tissues from mouse ES cellsenclosed within these microcapsules. Biotechnol Bioeng 2008;99:235–43.

[10] Gu YJ, Miyamoto M, Cui WX, Xu BY, Kawakami Y, Yamasaki T, et al. Effect ofneovascularization-inducing bioartificial pancreas on survival of syngeneicislet grafts. Transplant Proc 2000;32:2494–5.

[11] Sakai S, Hashimoto I, Kawakami K. Development of alginate–agarose subsieve-size capsules for subsequent modification with a polyelectrolyte complexmembrane. Biochem Eng J 2006;30:76–81.

[12] Luca G, Calvitti M, Nastruzzi C, Bilancetti L, Becchetti E, Angeletti G, et al.Encapsulation, in vitro characterization, and in vivo biocompatibility of Sertolicells in alginate-based microcapsules. Tissue Eng 2007;1:641–8.

[13] Bilancetti L, Tosi A, Di Croce S, Mazzitelli S, Bozzuto N, Luca G, et al. Design andproduction of alginate-based microdevices for Sertoli’s cells encapsulation.Minerva Biotech 2005;18:57–63.

[14] Baroud CN, Willaime H. Multiphase flows in microfluidics. Comptes RendusPhysique 2004;5:547–55.

[15] Shah RK, Shum HC, Rowat AC, Lee D, Agresti JJ, Utada AS, et al. Designeremulsions using microfluidics. Mater Today 2008;11:18–27.

[16] Nisisako T, Torii T, Higuchi T. Characterization of spontaneous transformation-based droplet formation during microchannel emulsification. Chem Eng J2004;101:23–9.

[17] Seo M, Nie Z, Xu S, Mok M, Lewis PC, Graham R, et al. Novel microreactors forfunctional polymer beads. Langmuir 2005;21:11614–22.

[18] Cohen I, Li H, Hougland JL, Mrksich M, Nagel SR. Using selective withdrawal tocoat microparticles. Science 2001;292:265–7.

[19] Huang KS, Lai TH, Lin YC. Manipulating the generation of Ca–alginatemicrospheres using microfluidic channels as a carrier of gold nanoparticles.Lab Chip 2006;6:954–7.

[20] Khademhosseini A, Langer R. Microengireered hydrogels for tissueengineering. Biomaterials 2007;28:5087–92.

[21] Sakai S, Kawabata K, Ono T, Ijima H, Kawakami K. Development of mammaliancell-enclosing subsieve-size agarose capsules for cell therapy. Biomaterials2005;26:4786–92.

[22] Grellier M, Granja PL, Fricain J, Bidarra SJ, Renard M, Bareille R, et al. The effectof the co-immobilization of human osteoprogenitors and endothelial cellswithin alginate microspheres on mineralization in a bone defect. Biomaterials2009;30:3271–8.