Journal of Controlled Release 135 (2009) 203–210

Contents lists available at ScienceDirect

Journal of Controlled Release

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

Bioactive cell-hydrogel microcapsules for cell-based drug delivery

Gorka Orive a, María De Castro a, Hyun-Joon Kong b, Rosa Ma Hernández a, Sara Ponce a,David J. Mooney b, José Luis Pedraz a,⁎a Laboratory of Pharmacy and Pharmaceutical Technology, Faculty of Pharmacy, University of the Basque Country, Vitoria-Gasteiz, Spainb School of Engineering and Applied Sciences, Harvard University, Cambridge, MA, USA

⁎ Corresponding author. Tel.: +34 945 013091; fax: +E-mail address: joseluis.pedraz@ehu.es (J.L. Pedraz).

0168-3659/$ – see front matter © 2009 Published by Edoi:10.1016/j.jconrel.2009.01.005

a b s t r a c t

a r t i c l e i n f o

Article history:

Improvement of long-term Received 6 October 2008Accepted 12 January 2009Available online 21 January 2009

Keywords:Biomimetic capsulesRGDAlginatesCell encapsulationDrug deliveryCell therapy

drug release and design of mechanically more stable encapsulation devices arestill major challenges in the field of cell encapsulation. This may be in part due to the weak in vivo stability ofcalcium-alginate beads and to the use of inactive biomaterials and inert scaffolds that do not mimic thephysiological situation of the normal cell milieu. We hypothesized that designing biomimetic cell-hydrogelcapsules might promote the in vivo long-term functionality of the enclosed drug-secreting cells and improvethe mechanical stability of the capsules. Biomimetic capsules were fabricated by coupling the adhesionpeptide arginine glycine aspartic acid (RGD) to alginate polymer chains and by using an alginate-mixtureproviding a bimodal molecular weight distribution. The biomimetic capsules provide cell adhesion for theenclosed cells, potentially also leading to mechanical stabilization of the cell-polymer system. Strikingly, thenovel cell-hydrogel system significantly prolonged the in vivo long-term functionality and drug release,providing a sustained erythropoietin delivery during 300 days without immunosuppressive protocols.Additionally, controlling the cell-dose within the biomimetic capsules enables a controlled in vitro and invivo drug delivery. Biomimetic cell-hydrogel capsules provide a unique microenvironment for the in vivolong-term de novo delivery of drugs from immobilized cells.

© 2009 Published by Elsevier B.V.

1. Introduction

The technology of cell microencapsulation is based on theimmobilization of therapeutically active cells within a polymermatrix, often alginate, surrounded by a semipermeable membrane.Such encapsulated cells are protected against immune cell- andantibody-mediated rejection and have the potential to produce anarray of therapeutically active substances [1–6]. These “living” drugdelivery systems are especially interesting for controlled andcontinuous expression of hormones [7] and growth factors, for thelocal and targeted delivery of drugs [8] and to improve thepharmacokinetics of easily degradable peptides and proteins, whichoften have short half-lives in vivo. One potential impact of this drugdelivery approach is that administration of immunosuppressants andimplementation of strict immunosuppressive protocols can bepartially or fully eliminated and therefore the serious risks associatedwith these drugs can also be avoided.

Although several attempts have been explored to improve thelong-term functionality of these bio-systems [9–11], designingmechanically more stable encapsulation devices and providing a

34 945 013040.

lsevier B.V.

more physiological microenvironment for the immobilized cells stillrepresent major challenges in the field. Typical calcium-alginate beadsare sensitive to non-gelling agents such as sodium and potassium,which in physiological conditions, results in osmotic swelling of thebeads, inevitably leading to increased pore size and destabilizationand rupture of the polymer scaffold. Addressing some of theselimitations may have a major impact on the long-term functionality ofthe enclosed drug-secreting cells. However, one major risk whendesigning strategies aimed to increase the stability of the devices isthat this enhancement may sacrifice other properties such as capsulepermeability, intra-cellular microenvironment and cell functionality.

Interestingly, until now, only inactive biomaterials and polymershave been used for microcapsule elaboration and all the reportedapproaches have considered the three dimensional capsules as inertscaffolds in which cells are physically entrapped and transported. Incontrast, we investigated the possibility of designing biomimetic cell-hydrogel capsules to promote the in vivo long-term functionality of theenclosed cells and improve the mechanical stability of the capsules. Toaddress these issues, first, we fabricated biomimetic alginates bycoupling the fibronectin-derived adhesion peptide arginine glycineaspartic acid (RGD) to alginate polymer chains. An alginate-mixtureproviding a bimodal molecular weight distribution was used forcapsule preparation with the aim of increasing the elastic modulus ofthe gel whileminimally raising the viscosity of the pre-gelled solution.Increasing the mechanical properties of the gels by simply raising the

204 G. Orive et al. / Journal of Controlled Release 135 (2009) 203–210

concentration of the highmolecularweight alginates typically used forcell encapsulation would also significantly raise the viscosity of thepre-gel solution, and increase cell damage duringmixing [12]. Becausealginate itself does not support cell attachment, the decoration of thepolymer with RGD will facilitate its interaction with the integrinreceptors of the enclosed cells and hypothetically prolong the survivaland functionality of the latter [13,14].

We suggest that biomimetic cell-hydrogel capsules can increasethe stability of the cell-loaded capsules and provide an improvedmicroenvironment for the enclosed cells. In the present paper, wereport that appropriately designed cell-hydrogel scaffolds signifi-cantly increase the mechanical resistance of the capsules whileminimizing damage during cell encapsulation. The novel biomimeticcapsules provide cell adhesion for the enclosed cells and prolong theirlong-term functionality and drug release. In addition, controlling thecell-dose within the biomimetic capsules enables a controlled in vitroand in vivo drug delivery.

2. Materials and methods

2.1. Alginate preparation

Low viscosity and high guluronic acid containing alginate (LVG)and medium viscosity and high guluronic acid containing alginate(MVG) were purchased from Protanal LF 20/40 (FMC Technologies).The molecular weights (Mw) of alginates were adjusted by irradiatingthem with a cobalt-60 source. The irradiation dose was varied from 2to 5Mrads. The final Mw of alginate molecules were analyzed with gelpermeation chromatography (Viscotek).

Following irradiation, alginate molecules were functionalized withsynthetic oligopeptides containing a sequence of (glysine)4-arginine-glysine-aspartic acid-tyrosine (G4RGDY, Commonwealth BiotechnologyInc., Richmond, VA, USA) by using previously described carbodiimidechemistry [15]. A molar ration of 0.0016 RGD peptide per sugar residueof the alginate chains was maintain constant. Briefly, alginate solutionswere prepared with a 2-[N-morpholino]ethanesulfonic acid hydrate(MES, Sigma) buffer and sequentially mixed with N-hydroxysulfosucci-nimide (sulfo-NHS, Pierce Chemical), and 1-ethyl-3-[3-(dimethyla-mino)propyl] carbodiimide (EDC, Sigma). Oligopeptides were added tothe reaction, and the alginates were allowed to react for 24 h. Themodified alginates were purified by extensive dialysis against distilledwater which removed the molecules smaller than 3500 g/mol.Following filtration for sterilization and lyophilization, the modifiedalginate was re-constituted to the desired concentration.

2.2. Viscosity measurement

The viscosity of the alginates solutions was determined by arotational rheometer (AR 1000 TA Instruments) with a cone-platemeasuring head. The parameters of the cone were: diameter 40 mmand clearance between cone and plate 50 µm. The rheometer isconnected to a thermostated bath at 25 °C to maintain thetemperature constant during the measurement process. During themeasurements the top part of the sample was covered in order toprevent water evaporation. The frequency was varied from 0.05 to1000 s−1.

2.3. Cell culture

C2C12 myoblasts derived from the skeletal leg muscle of an adultC3H mouse and genetically engineered to secrete murine EPO werekindly provided by the Institute des Neurosciences (Ecole Polytechni-que Federale of Lausanne, Lausanne, Switzerland). Cells were grown inDulbecco'smodified Eaglemedium containing 10% fetal bovine serum,2 mM L-glutamine, 4.5 g/l glucose, and 1% antibiotic/antimycoticsolution. Cells were cultured in T-flasks, maintained at 37 °C in a

humidified 5% CO2/95% air atmosphere standard incubator, and werepassed every 2–3 days. All the components of the culture mediumwere purchased from Gibco BRL (Invitrogen S.A., Spain). Before cellencapsulation, EPO secretion from 106 cells/24 h was determinedusing a sandwich enzyme-linked immunoabsorbent assay (ELISA) kitfor human EPO (R&D Systems, Minneapolis, MN, USA). Cross-reactionof the kit allowed detection of murine EPO in culture supernatants.

2.4. Cell encapsulation

Genetically engineered C2C12 myoblasts were immobilized withinRGD and non-RGD microcapsules using an electrostatic dropletgenerator. Briefly, cells (5×106 cells/ml) were suspended in RGD-alginate mixture solution (2.75% w/v) (50:50 LVG-RGD and MVG-RGD alginate) and non-RGD alginate mixture solution (1.4% w/v)(50:50 LVG and MVG alginate). In the dose–response study, differentcell concentrations (5, 3.75, 2.5 or 1.25×106 cells/ml) weresuspended in the RGD-alginate mixture solution (2.75% w/v) (50:50LVG-RGD and MVG-RGD alginate).

The cell suspension was extruded into a 55 mM calcium chloridesolution and the resulting alginate beads were successively ionicallylinked with a 0.05% (w/v) poly-L-lysine solution for 5 min (PLL;hydrobromide Mw: 15,000–30,000 Da; Sigma, St. Louis Mo, USA) andalginate solution 0.1% (w/v) for 5 min. Microcapsules were preparedat room temperature, under aseptic conditions, and were cultured incomplete medium. Morphology of microcapsules was assessed by aninverted optical microscopy (Nikon TMS) equipped with a camera(Sony CCD-Iris).

2.5. Mechanical stability studies: osmotic and compression resistancetests

The osmotic resistance of RGD and non-RGD microcapsules wasevaluated by a short-term stability study in which the swellingsuffered by the microcapsules after 1% citrate solution (w/v)treatment was determined. Briefly, 100 µl of microcapsule suspension(50–100microcapsules)wasmixedwith 900 µl of phosphate-bufferedsaline (PBS) and placed in a 24-well cell culture cluster. The cell clusterwas put in a shaker at 500 rpm and 37 °C for 1 h. Afterwards,supernatants were eliminated and 800 µl of 1% citrate solution (w/v)was added. The cluster containing the microcapsules was maintainedat static conditions at 37 °C for 24 h. The following day the diameter of20microcapsules of each typewasmeasured. Thewashing and shakingstep with PBS and the static conditions were repeated during thefollowing days until a 6-day period was completed. Results areexpressed as (Df/Di) where Di is the initial diameter of the capsuleand Df is the final diameter after the treatment.

The compression resistance of capsules was determined as themain force (g) needed to generate 70% uniaxial compression on aTexture Analyser (TA-XT2i, Stable Microsystems, Surrey, UK). Theforce exerted by the probe on the capsule was recorded as a functionof the compression distance leading to a force versus strain relation.Thirty capsules per batch were analyzed in order to obtain statisticallyrelevant data.

2.6. Animal protocols

All the procedures involving animals were in accordance with therelevant Spanish legislation and the European Community CouncilDirective on “Protection of Animals used in Experimental and OtherScientific Purposes” of 24 November 1986 (86/609/EEC).

2.7. Microcapsule implantation

Adult female Balb/c mice (Harlan Interfauna, Spain) were used asallogeneic recipients. Before implantation, microcapsules were

Fig.1. ComparisonofRGDandnon-RGDalginate solutions andmicrocapsules. (A)viscositymeasurement of RGD and non-RGD mixture solutions. (B) Schematic description of thebiomimetic cell-hydrogel capsules compared to non-RGD capsules. RGD modificationprovides a specific mechanism for cell adhesion. (C) and (D) RGD-modified scaffoldsprovide mechanical stabilization of the cell-polymer system. The swelling of biomimeticscaffoldswas significantly lower than non-RGDdevices (Pb0.05)whereas the resistance ofthe capsules against compression was significantly higher (Pb0.01).

205G. Orive et al. / Journal of Controlled Release 135 (2009) 203–210

washed several times in Hank's balanced salt solution (HBSS).Animals were anesthetized by isoflurane inhalation and a totalvolume of 0.2 ml of cell-loaded RGD and non-RGD microcapsuleswere implanted in the subcutaneous space of mice. In the second setof experiments 0.2 ml of RGD microcapsules containing 5, 3.75, 2.5and 1.25×106 cells/ml suspended in HBSS were also implantedsubcutaneously. An 18-gauge catheter was used for capsule implanta-tion (Nipro Europe N.V., Belgium). Control animals received only HBSSby the same route. Upon recovery, animals had access to food andwater ad libitum. No immunosuppressant protocols were applied tothe animals during the study.

At specific time points bloodwas collected by retroorbital punctureon anesthetized animals using heparinized capillaries (Deltalab,Spain). The hematocrit was measured by a standard microhematocritmethod. Results are expressed as mean±standard deviation.

EPO production from RGD and non-RGD cell loaded microcapsulesand from explanted capsules was measured using the sandwich ELISAfor human EPO. Results are expressed as mean±standard deviation.The cellular activity of immobilized cells was evaluated using thetetrazolium assay (MTT) (Sigma, St. Louis, MO, USA) as describedelsewhere [16].

2.8. Histological analysis

At day 330 after transplantation, animals were sacrificed andmicrocapsules were retrieved and fixed in a 4% paraformaldehydesolution in 0.1 M sodium phosphate, pH 7.2. Thereafter, serialhorizontal cryostat sections (14 µm) were processed for hematox-ylin-eosin staining.

2.9. Statistical analysis

Data are presented as mean±standard deviation. Differencesbetween control and experimental groups were analyzed forstatistical significance using a Student t test according to the resultsof the Levene test of homogeneity of variances. A P-valueb0.05 wasconsidered statistically significant.

3. Results and discussion

3.1. Biomimetic cell-hydrogel capsules are mechanically and chemicallymore resistant

We first fabricated biomimetic alginates by coupling the fibronec-tin-derived adhesion peptide arginine glycine aspartic acid (RGD) toalginate polymer chains. The molar ratio between RGD peptides andhigh MW alginate chains was kept constant at 0.016:1. Combiningboth polymers to form binary hydrogels realizes the advantages ofboth low and high MW alginates. The former contributes to thestiffness of the hydrogel while the latter maintains a high strain atfailure [14]. With this approach, we increased considerably the elasticmodulus of the gel while only minimally raising the viscosity of thepre-gelled solution, an important requirement for cell manipulationand immobilization as it dictates the magnitude of shear forcesexperienced by the immobilized cells.

We elaborated two different binary gels: the RGD alginate mixturesolution and the non-RGD alginate mixture solution. To confirm thesolutions were comparable, from a viscosity point of view, theviscosities were quantified (Fig. 1A). The viscosity of binary alginatesolutions was always lower than the viscosity that would result fromusing the same polymer concentration with only high molecularweight polymer, and low and similar overall viscosities were foundwith both binary gel types.

Alginate has no capacity to support cell interaction and attachmentdue to the lack of mammalian receptors recognizing its surfacechemistry [14,17]. RGDmodification provides a specificmechanism for

cell adhesion (Fig. 1B), as integrin receptors on the cell surface havethe ability to bind and crosslink multiple chains of the RGD-modifiedalginate scaffold, potentially also leading tomechanical stabilization ofthe cell-polymer system.

To assess the adhesive and biomimetic effects of RGD alginates,genetically engineered mouse skeletal C2C12 myoblasts secretingerythropoietin (EPO) were selected as a model cell system, and mixedwith the binary polymer solutions. It has been observed that RGDconcentration and alginate composition can control myoblast pheno-type [18]. Myoblasts are relatively easy to manipulate and onceencapsulated and implanted into the host, they irreversibly exit thecell cycle and fuse into multinucleated myofibrils. The latter is apotential advantage to control the therapeutic dose and to avoidpotential biosafety risks. EPO was selected as a model drug due to itsemerging and therapeutic effects and because it is easy to monitor its

206 G. Orive et al. / Journal of Controlled Release 135 (2009) 203–210

expression and bioactivity in vivo by following the hematocrit level(Hct). Moreover, because it exhibits poor pharmacokinetics, weexpect that cell encapsulation technology would avoid the repeatedEPO injections currently practiced.

Three dimensional (3D)microcapsuleswere fabricatedwith RGD andnon-RGD alginate solutions (Fig. 1B). To ensure an optimal biocompat-ibility, long-term functionality and a suitable zero-order kinetic release ofEPO, small diameter (450 μm)anduniformmicrocapsuleswereproduced[19–21]. We hypothesized that together with the bimodal molecularweight distribution of alginate mixture, the cell–ligand interactions maybe helpful to increase the mechanical properties of the matrices as cellsthemselves will take part in the matrix network formation. To evaluateand compare the mechanical integrity of the RGD and non-RGD alginatemicrocapsules, we studied the swelling behavior and the mechanicalresistance of the capsules against compression. Although developingreproducible and efficacious methods aimed to measure microcapsule'smechanical resistance is also another important challenge, the osmoticswelling study and the compression resistance study have been widelyacceptedby the scientific community involved in thefield. The swellingofRGD alginate microcapsules (1.12±0.01) was significantly lower than

Fig. 2. In vivo functionality of EPO-secreting cells immobilized in the biomimetic cell-hydrobiomimetic RGD capsules and non-RGD capsules compared to control animals. Cell-hydrogelclose to 85% during 10 months after only one administration (Pb0.01). (B) Photomicrograpimplantation. (C) Explanted biomimetic cell-hydrogel capsules showed significantly highedevices showed a minimal inflammatory reaction based on a thin fibroblastic layer surroun

non-RGD devices (1.16±0.01) (n=20, Pb0.05) (Fig. 1C), whereas theresistance of the capsules against compression was significantly higher(49.6±6.7 g/microcapsule versus 39.1±4.7 g/microcapsule; n=20,Pb0.01) (Fig. 1D). In addition, the binary gel system in either casesignificantly improved the mechanical resistance reported for classicalsingle alginate-poly-L-lysine capsules, which is close to 16 g/micro-capsule [22]. These results support the idea that a bimodal molecularweight distribution of alginate is an effective tool to increase the strengthof the gels. In addition, the interaction of myoblasts with the adhesionligands provides additionalmechanical integrity to the scaffolds byactingas cross-linkers.

3.2. Biomimetic cell-hydrogel capsules prolong the in vivo long-termfunctionality of the enclosed cells

To characterize the viability and functionality of the myoblastswithin the matrices, we studied the metabolic cell activity and EPOrelease during 3 weeks in vitro. These short-term studies did not showmajor differences either in cell viability or EPO release between RGDand non-RGD microcapsules (data not shown). After 21 days in

gel capsules. (A) Long-term hematocrit levels of mice subcutaneously implanted withRGD capsules exhibit a more sustained functionality, maintaining the hematocrit levelshs of the explanted of the biomimetic RGD and the non-RGD capsules 330 days post-r EPO release than non-RGD devices (Pb0.05). (D) Photomicrographs of the explantedding the cell-loaded capsules.

207G. Orive et al. / Journal of Controlled Release 135 (2009) 203–210

culture, EPO secretion from 100 cell-loaded RGD-microcapsules wasmeasured as 71±9 mIU EPO/24 h whereas the production in non-RGD matrices was 72±5 mIU EPO/24 h.

To further evaluate the potential impact of RGD bioactivatedmatrices on the survival and long-term functionality of the encapsu-lated cells, we implanted 0.2 ml of RGD and non-RGD cell-loaded

Fig. 3. Biomimetic cell-hydrogel capsules provide control over cell dose and drug delivery inwith different cell densities. (B) and (C) In vitrometabolic activity of the immobilized cells anrecipients subcutaneously implanted with RGDmicrocapsules, independently of the dose, wecell-dose provided significantly higher hematocrit levels than the lower dose at different ti

microcapsules (5×106 cells/ml alginate) in the subcutaneous tissue ofimmunocompetent Balb/c mice (n=5) without implementation ofimmunosuppressive protocols. A subcutaneous site was selected asprevious reports testing this administration route resulted in poorimplant viability and inconsistency in the hematocrit response toEPO secretion, even in syngeneic recipients [23]. The subcutaneous

vitro and in vivo. (A) Photomicrographs of the biomimetic cell-hydrogel capsules loadedd EPO secretionwere correlated with immobilized cell dose. (D) Hematocrit levels of allre significantly higher than control mice for all the time-periods (*, Pb0.05). The higherme-points (†, Pb0.05).

208 G. Orive et al. / Journal of Controlled Release 135 (2009) 203–210

administration of EPO versus the continuous bolus administration ofthe molecule should be taken into account for future potential clinicalapplications [23]. In fact, patients receiving subcutaneous constantrelease of EPO require lower doses and less frequent injections thanpatients treated with intravenous injections, suggesting the former isa more effective treatment [16,24].

Results indicated that hematocrit levels of all implanted recipients(RGD and non-RGD microcapsules) were significantly increased(Pb0.01), as compared to the control mice which showed a constantbaseline during the study. Mice implanted with RGD-alginatedevices increased their hematocrit until day 28 (88%) and then theymaintained an asymptotic level between 80 and 90% until the end ofthe study. Recipients with non-RGD microcapsules showed a similarbehavior until day 120. Thereafter, a progressive decline in the Hct ofthe animals was detected. Moreover, in the last 100 days of the studythe hematocrit level of the recipients with RGD alginate matriceswas significantly higher than non-RGD devices (Pb0.01) (Fig. 2A).According to these results, biomimetic cell-hydrogel capsules exhibit amore sustained functionality, maintaining the hematocrit levels closeto 85% during 10 months after only one administration. To ourknowledge, this is the first report describing such a sustained EPOrelease from encapsulated cells implanted subcutaneously and with-out immunosuppressive protocols. In previous studies using micro-capsules elaborated with lowMWalginate, we reported sustained Hcl

Fig. 4. Morphological and histological characterization of the explanted devices. (A) The bioafter implantation. Capsules appeared as irregular aggregates surrounded by vascular netwresponse and the presence of capillaries within and surrounding the capsule aggregates.

during 100 and 130 days respectively [8,24]. Our new data representsa three-fold increase in the efficacy of the drug delivery system.

We explanted the cell-loaded microcapsules from the recipients330 after the implantation. At explantation, both RGD and non-RGDcapsules appeared together forming an irregular and neovascularizedstructure which was adhered to the subcutaneous tissue. Micro-capsules were easily separated from the aggregates and they stillmaintained a defined semipermeable membrane (Fig. 2B). However,from a morphological point of view RGD capsules presented a morespherical shape and homogeneous surface. To asses the functionalityof the explanted grafts, we measured the EPO released from the RGDand non-RGD devices (Fig. 2C). Explanted biomimetic cell-hydrogelcapsules (100 capsules) released 51.8±38 mIU EPO/24 h whereasnon-RGD devices secreted 4.9±3.9 mIU EPO/24 h (Pb0.05). Theseresults suggest that biomimetic design of the capsules not onlyprovides a physically and chemically favorable environment, but alsoan improved biological space by presenting signals that specificallybinds to cell receptors.

The histological analysis of the capsules revealed a minimalinflammatory reaction based on a thin fibroblastic layer surroundingboth RGD and non-RGD capsules (Fig. 2D). A similar response has beenreported previously using conventional alginate-poly-L-lysine cap-sules [9,25,26]. Although commercially purified alginates were chosenfor these experiments, and polymer decoration was done in sterile

mimetic capsules loaded with different cell-densities were carefully retrieved 200 daysorks. (B) The histological analysis of the RGD matrices showed minor inflammatory

209G. Orive et al. / Journal of Controlled Release 135 (2009) 203–210

conditions, a thin pericapsular layer was present around the capsuleaggregates. It has been reported that the implantation procedure byitself attracts and activates immune cells and stimulates cytokinerelease. Assuming this, even an ideal, absolutely non-immunogenicmicrocapsule might not prevent such events to occur [26]. In thisreport, the weak immune-reaction detected did not alter thefunctionality of the biomimetic cell-hydrogel capsules nor limitcapsule graft survival and function. However, further evaluationsmight be needed to determine if reducing this slight response mayfurther improve the long-term viability of the enclosed cells. Interest-ingly, EPO directly addresses cellular inflammation by inhibitingseveral pro-inflammatory cytokines including IL-6, TNF-α and MCP-1which exert deleterious effects and trigger graft failure [27,28].

The presence of a capillary network surrounding the capsuleaggregates may also contribute to reducing the distance between theenclosed cells and the vasculature. The latter may ensure a betternutrient and oxygen supply, and thus improve the long-term function-ality of the devices. In fact, effective diffusion of oxygen and nutrientsfrom capillaries to neighbouring immobilized cells can occur over amaximum distance of 200 μm [29,30].

3.3. Biomimetic cell-hydrogel capsules provide control over cell dose anddrug delivery in vitro

In a second set of experiments, we made a dose–response study toevaluate the potential of biomimetic cell-hydrogel capsules to providecontrolled drug delivery. We first, fabricated and loaded RGD-matriceswith different cell densities (5, 3.75, 2.5 and 1.25×106 cells/mlalginate solution). Microscopic evaluation showed an increasing celldensity within the biomimetic scaffolds as well as the uniformity andspherical shape of all types of devices (Fig. 3A). The in vitro metabolicactivity of the immobilized cells was correlated with the cell dose, aswas the EPO production (Fig. 3 B and C). In addition, both themitochondrial activity of the cells and drug release were maintainedconstant during 3 weeks in vitro. According to this, it is possible tofabricate tailored cell-hydrogel capsules which release continuousdrug doses.

3.4. Biomimetic cell-hydrogel capsules provide control over cell dose anddrug delivery in vivo

To assess the therapeutic potential of this approach, weimplanted subcutaneously 0.2 ml of the different cell-loadedbiomimetic cell-hydrogel capsules in Balb/c mice. Results indicatedthat hematocrit levels of all recipients implanted with RGDmicrocapsules, independently of the dose, were significantly higherthan control mice for all the time-periods (Pb0.05) (Fig. 3 D). Thehematocrit level of all mice implanted with RGD-microcapsules rosefrom a baseline value (46±7%) to a peak value at day 28. At thatpoint, the levels were 87±3%; 88±4%; 86.4±1.5%; and 81±8% forthe cell-densities of 5, 3.75, 2.5 and 1.25×106 respectively.

A dose response effect in the hematocrit levels was observed in theinitial 14 days. From day 28 till day 60, the Hct obtained with all theformulations was not significantly different, indicating that a max-imum hematopoietic effect was obtained with all the formulations. Inthe duration of the time period, mice receiving the higher doseshowed significantly higher Hct than recipients receiving the lowerdose (Pb0.05). Additionally, the hematocrit levels of mice implantedwith the lowest cell-dose scaffold decreased progressively whereasanimals receiving the other 3 formulations maintained the Hct levelsconstant until the end of the study. These results may be explainedconsidering the pharmacodynamic behaviour of EPO, which ischaracterized by high therapeutic effects obtained with relativelylow EPO doses. Although lower EPO levels will be recommended forthe treatment of anaemia, higher doses might be required forneuroprotection [31] and the treatment of schizophrenia [32].

All the implanted biomimetic capsules were carefully retrieved200 days after implantation. Microcapsules again formed irregularaggregates surrounded by vascular networks (Fig. 4A). The vascular-ization of the aggregates seemed to be dose-dependent, withminor orno capillary formation in the low dose RGD-capsule aggregates(1.25×106). EPO has been shown to possess pro-angiogenic propertiesas it enhances matrix metalloproteinase 2 production, Janus Kinase-2phosphorylation and stimulates the proliferation and migration ofendothelial cells and bone-marrow derived endothelial progenitorcells [33,34]. Although the present results suggest that aminimumEPOdose might be necessary to stimulate capillary formation around thecapsule aggregates, no clear relations were found between the vesseloutgrowth around the capsules and the long-term functionality of thedevices loaded with the higher cell doses. The histological analysis ofthe RGD matrices showed minor inflammatory response and thepresence of capillaries within and surrounding the capsule aggregates(Fig. 4 B). After isolating the matrices from the aggregates, weobserved that all RGD capsules remained viable and releaseddetectable amounts of EPO (data not shown).

4. Conclusions

Biomimetic cell-hydrogel capsules are mechanically and chemi-cally more resistant and prolong drug release and long-term efficacyof immobilized cells. The overall results indicate that biomimetic cell-hydrogel capsules can alter the pharmacokinetics and dose regimen ofEPO, eliminating the need for repeated parenteral administrations.Unlike other approaches encapsulating the peptide alone, this cell-based strategy permits the continuous de novo secretion of EPO,avoiding drug instability. Biomimetic design of cell-loaded capsulescan improve cell performance as cells are immobilized in a functionalscaffold which mimics the natural microenvironment of the cells. Thisapproach may be broadly applicable to any therapeutic approach inwhich continuous and controlled drug or growth factor delivery isnecessary, including deficiencies of a gene product, central nervoussystem diseases and pathologies demanding chronic drug treatments.

Acknowledgement

This project was partially supported by the Ministry of Educationand Science (BIO2005-02659). EPO-secreting myoblasts were pro-vided by Dr. Patrick Aebischer and the Institute des Neurosciences ofLausanne (EPFL), Lausanne, Switzerland. We also thank the Depart-ment of Histology and Pathological Anatomy of the University ofNavarra (Pamplona, Spain) for technical assistance with histologicalanalyses.

References

[1] F. Lim, A.M. Sun, Microencapsulated islets as bioartificial endocrine pancreas,Science 210 (1980) 908–910.

[2] R.P. Lanza, J.L. Hayes, W.L. Chick, Encapsulated cell technology, Nat. Biotechnol. 14(1996) 1107–1111.

[3] G. Orive, R.M. Hernández, A.R. Gascón, et al., Cell encapsulation: highlights andpromise, Nat. Med. 9 (2003) 104–107.

[4] T.M.S. Chang, Therapeutic applications of polymeric artificial cells, Nat. Rev. DrugDiscov. 4 (2005) 221–235.

[5] G. Orive, R.M. Hernández, A.R. Gascón, M. Igartua, J.L. Pedraz, Cell microencapsula-tion technology for biomedical purposes: novel insights and challenges, TrendsPharm. Sci. 24 (2003) 207–210.

[6] A. Townsend-Nicholson, S.N. Jayasinghe, Cell electrospinning: a unique biotechni-que for encapsulating living organisms for generating active biological micro-threads/scaffolds, Biomacromolecules 7 (2006) 3364–3369.

[7] P. De Vos, M.M. Faas, B. Strand, R. Calafiore, Alginate-based microcapsules forimmunoisolation of pancreatic islets, Biomaterials 27 (2006) 5603–5617.

[8] M. Lörh, A. Hoffmeyer, J. Kroger, et al., Microencapsulated cell-mediated treatmentof inoperable pancreatic carcinoma, Lancet 357 (2001) 1591–1592.

[9] G. Orive, M. De Castro, S. Ponce, et al., Long-term expression of erythropoietin frommyoblasts immobilized in biocompatible and neovascularized microcapsules, Mol.Ther. 12 (2005) 283–289.

210 G. Orive et al. / Journal of Controlled Release 135 (2009) 203–210

[10] A.M. Rokstad, I. Donati, M. Borgogna, et al., Cell-compatible covalently reinforcedbeads obtained from a chemoenzymatically engineered alginate, Biomaterials 27(2006) 4726–4737.

[11] R. Calafiore, G. Basta, G. Luca, et al., Microencapsulated pancreatic islet allograftsinto nonimmunosuppressed patients with type 1 diabetes: first two cases,Diabetes Care 29 (2006) 137–138.

[12] H.J. Kong, M.K. Smith, D.J. Mooney, Designing alginate hydrogels to maintainviability of immobilized cells, Biomaterials 24 (2003) 4023–4029.

[13] E. Alsberg, K.W. Anderson, A. Albeiruti, J.A. Rowley, D.J. Mooney, Engineeringgrowing tissues, Proc. Natl. Acad. Sci. U.S.A. 99 (2002) 12025–12030.

[14] A.D. Augst, J.J. Kong, D.J. Mooney, Alginate hydrogels as biomaterials, Macromol.Biosci. 6 (2006) 623–633.

[15] J.A. Rowley, D.J. Mooney, Alginate hydrogels as synthetic extracellular matrixmaterials, Biomaterials 20 (1999) 45–53.

[16] M. Sohmiya, T. Kakiba, Y. Kato, Therapeutic use of continuous subcutaneousinfusion of recombinant human erythropoietin in malnourished predialysisanemic patientswith diabetic nephropathy, Eur. J. Endocrinol.139 (1998) 367–370.

[17] A.B. Lansdown, M.J. Payne, An evaluation of the local reaction and biodegradationof calcium sodium alginate (Kaltostat) following subcutaneous implantation in therat, J. R. Coll. Surg. Edinb. 39 (1994) 284–288.

[18] J.A. Rowley, D.J. Mooney, Alginate type and RGD density control myoblastphenotype, J. Biomed. Mat. Res. 60 (2006) 217–223.

[19] R. Robitaille, J.F. Pariseau, F.A. Leblond, M. Lamoureux, Y. Lepage, J.P. Hallé, Studieson small (b350microm) alginate-poly-L-lysinemicrocapsules. III. Biocompatibilityof smaller versus standard microcapsules, J. Biomed. Mater. Res. 44 (1999)116–120.

[20] Z.P. Lum,M. Krestow, I.T. Tai, I. Vacek, A.M. Sun, Xenografts of rat islets into diabeticmice. An evaluation of new smaller capsules, Transplantation 53 (1992)1180–1183.

[21] D. Chicheportiche, G. Reach, In vitro kinetics of insulin release by microencapsulatedrat islets: effect of the size of the microcapsules, Diabetologia 31 (1998) 54–57.

[22] L. Grandoso, S. Ponce, I. Manuel, et al., Long-term survival of encapsulated GDNFsecreting cells implanted within the striatum of parkinsonized rats, Int. J. Pharm.343 (2007) 69–78.

[23] B.L. Schneider, F. Schwenter, W.F. Pralong, P. Aebischer, Prevention of the initialhost immuno-inflammatory response determines the long-term survival ofencapsulated myoblasts genetically engineered for erythropoietin delivery, Mol.Ther. 7 (2003) 506–514.

[24] I.C. Macdougall, S.J. Gray, O. Elston, et al., Pharmacokinetics of novel erythropoiesisstimulating protein compared with epoetin alfa in dialysis patients, J. Am. Soc.Nephrol. 10 (1999) 2392–2395.

[25] S. Ponce, G. Orive, R.M. Hernández, et al., In vivo evaluation of EPO-secreting cellsimmobilized in different alginate-PLL microcapsules, J. Control. Rel. 116 (2006)28–34.

[26] A. Murua, M. De Castro, G. Orive, R.M. Hernandez, J.L. Pedraz, In vitrocharacterization and in vivo functionality of erythropoietin-secreting cellsimmobilized in alginate-poly-L-lysine-alginatemicrocapsules, Biomacromolecules8 (2007) 3302–3307.

[27] R. Robitaille, J. Dusseault, N. Henley, K. Desbiens, N. Labrecque, J.P. Hallé,Inflammatory response to peritoneal implantation of alginate-poly-L-lysinemicrocapsules, Biomaterials 26 (2005) 4119–4127.

[28] N.V. Olsen, Central nervous system frontiers for the use of erythropoietin, Clin.Infect. Dis. 37 (2003) S323–S330.

[29] M.S. Martijin de Groot, T.A. Schuurs, H.G.D. Leuvenink, M.R.D. Van Schilfgaarde,Macrophage overgrowth affects neighboring nonovergrown encapsulated islets,J. Surg. Res. 115 (2003) 235–241.

[30] G. Orive, S.K. Tam, J.L. Pedraz, J.P. Hallé, Biocompatibility of alginate-poly-L-lysinemicrocapsules for cell therapy, Biomaterials 27 (2006) 3691–3700.

[31] M. Brines, A. Cerami, Emerging biological roles for erythropoietin in the nervoussystem, Nat. Rev. Neurosci. 6 (2005) 484–494.

[32] H. Ehrenreich, D. Degner, J. Meller, et al., Erythropoietin: a candidate compound forneuroprotection in schizophrenia, Mol. Phychiatry 9 (2004) 42–54.

[33] D. Ribatti, M. Presta, A. Vacca, et al., Human erythropoietin induces a pro-angiogenic phenotype in cultured endothelial cells and stimulates neovascular-ization in vivo, Blood 93 (1999) 2627–2636.

[34] K.J. Zwezdaryk, S.B. Coffelt, Y.G. Figueroa, et al., Erythropoietin, a hypoxia-regulated factor, elicits a pro-angiogenic program in human mesenchymal stemcells, Exp. Hematol. 35 (2007) 640–652.