Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/5972835

Novelculturestrategyforhumanstemcellproliferationandneuronaldifferentiation

ArticleinJournalofNeuroscienceResearch·December2007

DOI:10.1002/jnr.21451·Source:PubMed

CITATIONS

16

READS

74

6authors,including:

Someoftheauthorsofthispublicationarealsoworkingontheserelatedprojects:

Developingcomplex3Dhepaticspheroidstostudyliverdiseases.Viewproject

CatarinaBrito

iBET-InstitutodeBiologiaExperimentaleTec…

68PUBLICATIONS691CITATIONS

SEEPROFILE

JúliaCosta

NewUniversityofLisbon

82PUBLICATIONS1,958CITATIONS

SEEPROFILE

ManuelJTCarrondo

iBET-InstitutodeBiologiaExperimentaleTec…

365PUBLICATIONS5,962CITATIONS

SEEPROFILE

PaulaMAlves

NewUniversityofLisbon

373PUBLICATIONS4,450CITATIONS

SEEPROFILE

AllcontentfollowingthispagewasuploadedbyCatarinaBritoon13March2014.

Theuserhasrequestedenhancementofthedownloadedfile.Allin-textreferencesunderlinedinbluearelinkedtopublicationsonResearchGate,lettingyouaccessandreadthemimmediately.

Novel Culture Strategy for HumanStem Cell Proliferation andNeuronal Differentiation

Margarida Serra,1 Sofia B. Leite,1 Catarina Brito,1 Julia Costa,1

Manuel J.T. Carrondo,1,2 and Paula M. Alves1*1Instituto de Biologia Experimental e Tecnologica/Instituto de Tecnologia Quımica e Biologica(IBET/ITQB), Oeiras, Portugal2Faculdade de Ciencias e Tecnologia, Universidade Nova de Lisboa, Monte da Caparica, Portugal

Embryonal carcinoma (EC) stem cells derived fromgerm cell tumors closely resemble embryonic stem (ES)cells and are valuable tools for the study of embryo-genesis. Human pluripotent NT2 cell line, derived froma teratocarcinoma, can be induced to differentiate intoneurons (NT2-N) after retinoic acid treatment. To realizethe full potential of stem cells, developing in vitro meth-ods for stem cell proliferation and differentiation is akey challenge. Herein, a novel culture strategy for NT2neuronal differentiation was developed to expand NT2-N neurons, reduce the time required for the differentia-tion process, and increase the final yields of NT2-Nneurons. NT2 cells were cultured as 3D cell aggregates(‘‘neurospheres’’) in the presence of retinoic acid, usingsmall-scale stirred bioreactors; it was possible to obtaina homogeneous neurosphere population, which can betransferred for further neuronal selection onto coatedsurfaces. This culturing strategy yields higher amountsof NT2-N neurons with increased purity compared withthe amounts routinely obtained with static cultures.Moreover, mechanical and enzymatic methods forneurosphere dissociation were evaluated for their abilityto recover neurons, trypsin digestion yielding the bestresults. Nevertheless, the highest recoveries were ob-tained when neurospheres were collected directly totreated surfaces without dissociation steps. This novelculture strategy allows drastic improvement in the neu-ronal differentiation efficiency of NT2 cells, insofar as afourfold increase was obtained, reducing simultane-ously the time needed for the differentiation process.The culture method described herein ensures efficient,reproducible, and scaleable ES cell proliferation anddifferentiation, contributing to the usefulness of stemcell bioengineering. VVC 2007 Wiley-Liss, Inc.

Key words: human carcinoma stem cells; expansion;neuronal differentiation; stirred tank bioreactors;neurospheres

Human embryonic stem cells (hESCs) have beenthe focus of intense research since the first successful iso-lation of human inner cell mass (ICM) in 1994 (Bongso

et al., 1994) and the establishment of the first hESC linenearly 4 years later (Thomson et al., 1998). These cellscan be propagated in culture for extended periods andcan differentiate into all somatic cell types (pluripotentcells), thus holding enormous potential for the develop-ment of novel approaches for improving processes in celltherapy, tissue engineering, drug discovery, and in vitrotoxicology (Jones and Thomson, 2000; Davila et al.,2004; McNeish, 2004). However, assessing and character-izing hESC lines can be difficult, time consuming, and ex-pensive because of demanding cell culture procedures andrequirements under existing patents. To overcome this,hESC model systems are very valuable for predicting cel-lular behavior and elucidating mechanisms involved inself-renewal and differentiation. Within this, several groupshave proposed using embryonal carcinoma (EC) cell linesbased on earlier observations that germ cells are pluripo-tent (Pal and Ravindran, 2006) and share characteristicswith embryonic stem (ES) cells (Andrews, 2002).

Pluripotent EC cells are derived from tumorsknown as teratocarcinomas that are understood to arisefrom transformed germ cells (Przyborski et al., 2004). Itis generally accepted that EC cells closely resemble EScells and are often considered to be the malignant coun-terparts of ES cells (Andrews et al., 2001). In fact, ECstem cell lines provide a useful alternative to embryos forthe study of mammalian cell differentiation (Stewartet al., 2003). More specifically, the well-establishedNTera-2/cl.D1 (NT2) lineage, which has been derivedfrom a human testicular cancer, offers a convenient and

Contract grant sponsor: European Commission; Contract grant number:

NMP4-CT-2004-500039 (Cell Programming by Nanoscaled Devices).

*Correspondence to: Paula Marques Alves, Instituto de Biologia Experi-

mental e Tecnologica, Apartado 12, 2781-901 Oeiras, Portugal.

E-mail: marques@itqb.unl.pt

Received 27 March 2007; Revised 24 May 2007; Accepted 31 May

2007

Published online 14 September 2007 in Wiley InterScience (www.

interscience.wiley.com). DOI: 10.1002/jnr.21451

Journal of Neuroscience Research 85:3557–3566 (2007)

' 2007 Wiley-Liss, Inc.

robust model for studying the commitment of humanEC stem cells to the neural lineage and their subsequentdifferentiation into neurons (Andrews, 1984). Upontreatment with retinoic acid (RA), the NT2 cells can beinduced to differentiate into postmitotic neurons (NT2-N) that express many neuronal markers (Pleasure et al.,1992). By using specific culture conditions, these NT2-N neurons form functional synapses (Hartley et al.,1999) and express a variety of neurotransmitter pheno-types (Pleasure and Lee, 1993; Yoshioka et al., 1997;Guillemain et al., 2000). This cell line has also beenused in several transplantation studies, including engraft-ments in experimental animals (Ferrari et al., 2000; Wat-son et al., 2003) and in human patients (Kondziolkaet al., 2001).

Although techniques have been developed to pro-duce and purify NT2-N neurons from other contami-nating cell types, these methods are laborious, yield rela-tively few neurons per culture, and are thus time con-suming. Therefore, the last 2 years have witnessed anincrease in experiments aimed at improving culture sys-tems for enhanced neuronal productivity and reduceddifferentiation time (Durand et al., 2003; Horroks et al.,2003). These studies are focused on the cultivation offree-floating cell spheres (neurospheres) under suspensionconditions, using nonadherent Petri dishes. However,these static culture systems have several limitations thataffect culture productivity, scalability, and especially re-producibility. To overcome these limitations, novel cul-ture systems for proliferation and differentiation of NT2cells are needed. So far, because of their characteristicsand wide use, stirred culture systems, including fullycontrolled bioreactors, appear as promising candidates,they are hydrodynamically well characterized, are easy toscale-up, enable better homogeneity of cultures, providereproducible experimental conditions, and allow easysampling. Within the field of stem cell expansion anddifferentiation, several reports have been published onstirred tank bioreactors (Collins et al., 1998a,b; Senet al., 2002; Youn et al., 2005; Cameron et al., 2006).However, these are previous approaches, and moreinnovative strategies should be explored. Up to now, thebiggest restriction is still the low efficiency of the stem

cell differentiation process, and this limitation is verycritical in aiming toward scaleable protocols.

The present work describes a novel culture systemthat allows the improvement of expansion and differen-tiation of NT2 cells. Herein, for the first time, a strategywas established to obtain neuronal differentiated fromnon-neuronal EC cells with stirred suspension condi-tions. Thus, culturing the NT2 cells as 3D aggregates(neurospheres) permitted increased yields of differenti-ated NT2-N neurons in a shorter process time. Theimprovement of the culture system described here opensnew perspectives for hESC technology, providing promis-ing strategies for stem cell proliferation and differentiation,which have been bottlenecks in the expansion of thisscience and technology.

MATERIALS AND METHODS

Cell Culture

Undifferentiated NT2 cells were routinely cultivated instandard tissue culture flasks (T75 culture flasks from Nunc)and maintained in OptiMEM medium (Gibco, Grand Island,NY) supplemented with 5% (v/v) fetal bovine serum (FBS;Hyclone, Logan, UT) and 100 U/ml penicillin-streptomycin(P/S; Gibco).

Systems for NT2 Neuronal Differentiation

To evaluate the effectiveness of the culture strategydeveloped herein (stirred conditions), protocols for culturingand differentiation NT2 cells in static conditions reported inthe literature were also performed. A brief description of bothmethodologies is summarized in Table I.

Experiments in Static Conditions

Control experiments of NT2 differentiation were per-formed according to the procedure described in Pleasure et al.(1992). Briefly, NT2 precursor cells were cultured in standardtissue culture flasks (T75 culture flasks from Nunc) and differ-entiated in Dulbecco’s modified Eagle’s medium (DMEM)-high glucose (HG; Gibco) with 10% (v/v) FBS, 100 U/ml P/S, and 10 lM retinoic acid (RA; Sigma) for 5 weeks intopostmitotic NT2-N neurons. This cell culture was then tryp-

TABLE I. Comparison of Neuronal Differentiation Protocols in Static and Stirred Culture Conditions for NT2 Cells

Static culture protocol (Pleasure et al., 1992) Stirred culture protocol

NT2 proliferation and differentiation Step 1: NT2 proliferation and differentiation

— Inoculation of precursor NT2 cells in T flasks — Inoculation of precursor NT2 cells in Spinner vessels

1 Day Culture of NT2 cells without RA 2 Days Culture of NT2 cells without RA (growth period)

34 Days RA treatments 14 Days RA treatments (differentiation period)

Medium replacement (100% twice per week) Medium replacement (50% three times per week)

— Replate 1: Replating of cells at lower density Step 2: Neurosphere harvesting

12 Days Mitotic inhibitor treatment 7 Days Mitotic inhibitor treatment

Medium replacement (100% – twice per week) Medium replacement (100% twice per week)

Replate 3: Selecting NT2-N neurons Step 3: Selecting NT2-N neurons

0–5 Days Cultivation in conditioned medium 0–5 Days Cultivation in conditioned medium

Total 47–53 days Harvesting of NT2-N neurons Total 23–28 days Harvesting of NT2-N neurons

3558 Serra et al.

Journal of Neuroscience Research DOI 10.1002/jnr

sinized and replated at a lower density (total of cells collectedfrom one T75 culture were equally divided by two T175) andcultured in DMEM-HG with 5% (v/v) FBS containing 100U/ml P/S and mitotic inhibitors [1 lM cytosine arabinoside(AraC), 10 lM fluorodeoxyuridine (Fudr), and 10 lM uridine(Urd); Sigma, St. Louis, MO] for 12 days to obtain an NT2-N culture with maximal purity. After this period of culture inT175 culture flasks, NT2-N neurons were selectively trypsi-nized using a 1:3 diluted trypsin solution (trypsin-EDTA 13,liquid 0.05% trypsin; Gibco) in DMEM-HG medium. Subse-quently, the cells were counted and transferred to plates/slidescoated with PDL and MG (PDL-MG) for further experi-ments, including culture characterization using immunocyto-chemistry tools.

Experiments in Stirred Suspension Conditions:Spinner Culture

This protocol was designed in three sequential steps;only the first one is performed under suspension and stirredconditions.

Step 1: NT2 proliferation and differentiation instirred suspension conditions. NT2 cells were inoculatedin a 125-ml spinner vessel (fromWheaton, Techne, NJ) equippedwith ball impeller, at a density of 53 105 cell/ml. Throughout thisstep, cells were cultured in stirred suspension conditions for16 days.

Growth period. During the first 2 days, undifferentiatedNT2– cells were cultured in 80 ml DMEM-HG medium(Gibco) supplemented with 10% (v/v) FBS (Hyclone) and100 U/ml P/S (Gibco).

Differentiation period. From the third day onward, every2–3 days (i.e., three times per week), RA (Sigma) was addedto the medium to yield a final concentration of 10 lM in100 ml of culture volume. On a first approach, the mediumreplacement was performed by letting neurospheres settlingdown, as previously reported by Moreira et al., 1994. How-ever, the sedimentation of NT2 aggregates was too slow,yielding necrotic centers and dramatic loss of viability. Toovercome that, a new strategy was attempted transferring halfof culture supernatant to centrifuge tubes. After centrifugationat 200g for 10 min, the collected cells/aggregates were ressus-pended in new culture medium containing RA. Then, the cellsuspension obtained was added to the remaining cell suspensionin the spinner vessel to yield a final volume of 100 ml. Theagitation rate was increased during cultivation to avoid aggregateclumping and to control neurosphere size (day 1 to day 9, 60 rpm;day 10 to day 14, 80 rpm; day 15 onward, 100 rpm).

Step 2: Neurosphere harvesting. After 16 days ofculture, corresponding to 2 weeks of differentiation time, theNT2 neurospheres were collected. After that, both intact and-dissociated neurospheres were plated (cell density 1–3 3 105

cells/cm2) on T75 culture flasks (Nunc) coated with poly-D-lysine (PDL; Sigma) and Matrigel (MG; Becton-Dickinson,San Jose, CA), and cultured under mitotic inhibitory condi-tions: DMEM-HG (Gibco) supplemented with 5% (v/v) FBS(Hyclone), 100 U/ml P/S (Gibco), 1 lM Ara C, 10 lMFudR, and 10 lM Urd (Sigma); the culture medium was par-tially (50%) replaced every 2 days.

Step 3: Selecting NT2-N neurons. After 7 days ofculture in T75 culture flasks, NT2-N neurons were selectivelytrypsinized with a 1:3 diluted trypsin solution (trypsin-EDTA13, liquid 0.05% trypsin; Gibco) in DMEM-HG medium. Sub-sequently, the cells were counted and transferred to plates/slidescoated with poly-D-lysine (PDL; Sigma) and Matrigel (MG;BD Biosciences) PDL-MG for further experiments, includingculture characterization with immunocytochemistry tools.

Aggregates Dissociation Test

Five protocols were compared for the dissociation ofNT2 neurospheres obtained after cultivating aggregates inspinner vessel for 16 days, including mechanical and enzymaticmethods. The same volume (1 ml) of aggregate suspensionculture was harvested and used for each dissociation protocol.In the mechanical assay, NT2 aggregates were dissociatedusing a fire-narrowed Pasteur pipette. In the enzymatic proto-cols, neurospheres were disrupted using four different solu-tions: TrypLe Select (TS; Gibco), trypsin-EDTA (13; Tryp,liquid 0.05% trypsin, Gibco), Accutase (At; Chemicon, Teme-cula, CA), and Accumax (Am; Chemicon). For all assays, asimilar procedure was used: 1 ml of cell suspension was col-lected for each protocol, and, after centrifugation (200g, 10min), the neurospheres were incubated with 200 ll dissocia-tion enzyme at 378C during 2 min. Then, 800 ll serum-sup-plemented medium was added to each sample. In all proce-dures, cell density and viability were evaluated after dissocia-tion. Cell density was assessed with a hemacytometer, andviability was determined by the standard trypan blue exclusiontest. The experiments were performed in triplicate.

The different cell suspensions were then plated on six-well plates (Nunc) precoated with PDL-MG, and cells werecultured for up to 7 days under mitotic inhibitory conditions.All cultures were monitored daily in an inverted microscope(Leica DM IRB).

Aggregate Size

The aggregate size in each culture sample was measuredby using a micrometer coupled to an inverted microscope(Leica DM IRB). The average diameter was calculated bymeasurement of two perpendicular diameters of a minimumof 15 aggregates. Aggregates less than 20 lm in diameter (gen-erally doublets or triplets) were not considered.

Immunocytochemistry: Cell Culture Slides andCell Aggregates

Aggregates and cell culture grown on glass coverslipswere rinsed in phosphate-buffered saline (PBS) with 0.5 mMMgCl2 and fixed 20 min in 4% (w/v) paraformaldehyde(PFA) solution in PBS with 4% (w/v) saccarose. After fixa-tion, cells were washed twice with PBS and then incubatedwith permeabilization solution, Triton X-100 0.3% (w/v)diluted in PBS, during 20 min. Then, cells were washed againwith PBS and kept in blocking solution composed of bovineserum albumine (BSA) 1% (w/v) and Triton X-100 0.1% (w/v) in PBS for 1 hr. Cells were then incubated with primaryantibodies diluted in PBS containing 0.1% (w/v) Triton X-100 for 2 hr at room temperature; the coverslips were washed

Novel Culture Strategy for Stem Cells 3559

Journal of Neuroscience Research DOI 10.1002/jnr

three times with PBS and overlaid with secondary antibodiesfor 1 hr at room temperature. After three washes with PBS,samples were mounted in ProLong medium supplemented withDAPI (Molecular Probes, Eugene, OR) for nucleus staining.

Primary antibodies and dilutions used were: anti-nestin(nestin, 1:200; Chemicon), anti-tubulin beta III isoform(b-TubIII, 1:10; Chemicon), anti-neurofilament light chain(NF-L, 1:200; Chemicon), anti-neurofilament heavy chain(NF200, 1:200; Chemicon), anti-microtubule-associated pro-tein-2 (MAP2a/b, 1:100; Chemicon), anti-glial fibrillary acidicprotein (GFAP, 1:200; Chemicon), and anti-oligodendrocytemarker O4 (O4, 1:200; Chemicon). The secondary antibodiesand dilutions used were: anti-mouse Alexa 488 (1:200) andanti-rabbit Alexa 594 (1:500) from Molecular Probes. Cellswere visualized in a fluorescence microscope (Leica DMRB).

RESULTS

NT2 Neurospheres Cultured in Spinner Vessel

The ability to grow, consistently and reproducibly,large numbers of human neural cells in vitro is oftenlimited by complex and demanding protocols. BecauseNT2 cells have been reported to proliferate and differen-tiate successfully on non-adhesive substrate as cell clusters(Durand et al., 2003; Horroks et al., 2003), here thechallenge was to cultivate the NT2 neurospheres inscaleable and robust culture systems.

Therefore, NT2 precursor cells were inoculated ina spinner vessel, at a density of 5 3 105 cells/ml. Duringthe following 2 days of culture (growth period), the agi-tation rate was set to low values to promote NT2 cellaggregation. Under these conditions, the NT2 precursorcells proliferated, with a twofold increase in cell concen-tration (results not shown), and successfully aggregatedinto small cell aggregates ranging from 50 to 100 lm(average diameter 60.1 6 7.6 lm; Fig. 1A).

Neuronal differentiation was induced by additionof RA 2 days after cell inoculation (differentiation pe-riod). Initially, such aggregates had a ragged appearance(Fig. 1B), but, as differentiation progressed, their diame-ter increased (ranging from 100 to 300 lm, average di-ameter 160.0 6 78.6 lm), and their structure stabilized,forming solid and spherical 3D aggregates, neurospheres(Fig. 1C,D). Here, an essential parameter that requiresmonitoring control is the aggregate/neurosphere size,which should be minimized to avoid detrimental diffu-sion gradients and/or decrease the differentiation poten-tial of the culture. As reported previously (Moreiraet al., 1994; Kallos et al., 2003), the agitation rate in sus-pension systems should be sufficient to control the diam-eter of neurospheres. Therefore, in this study, the rate ofagitation was increased during culture time to ensure ahomogenous neurosphere culture and avoid necrotic andapoptotic cells within the neurospheres.

After 2 weeks of RA treatments, neurosphereswere characterized in terms of cell composition by usingimmunofluorescence microscopy. Figure 1E shows that,at this time, neurospheres were composed of bothnestin-positive cells (undifferentiated stem cells and neu-

ral progenitor cells) and b-TubIII-positive cells (NT2-Nneurons).

Neurospheres Harvesting

At day 16, neurospheres were collected, and twodifferent strategies were tested to evaluate the bestmethod to recover the NT2-N neurons. In both strat-egies, the cells were seeded into PDL-MG-coated surfa-ces and cultured under mitotic inhibitory conditions for1 week.

In the first strategy, neurospheres were dissociatedby five different protocols (including enzymatic and me-chanical methods). To investigate and select the bestprocedure, these results were compared in terms of finalsingle cell concentration and viability after each specificdissociation procedure. As shown in Figure 2A, enzy-matic methods led to higher cell viabilities (>90% in allprotocols) than the mechanical procedure. In addition,higher amounts of cell debris were present in mechani-cally dissociated cultures. These findings suggest that themechanical procedure is very aggressive for NT2 neuro-sphere dissociation. When the enzymatic dissociationprotocols were compared for final single cell concentra-tion, the highest value was observed for samples dissoci-ated with trypsin, which was also the method thatenhanced better neurosphere dissociation when moni-tored by microscopic inspection (Fig. 2B, day 0). Theremaining enzymatic dissociation assays (TrypLE Select,Accutase, and Accumax) were less efficient, insofar asseveral cell aggregates were detected in each cell suspen-sion following the dissociation procedure (Fig. 2B, day0). Moreover, the values of final cell concentration werelower compared with the samples dissociated with tryp-sin (Fig. 2A).

After 7 days of culture, it was observed that tryp-sin-dissociated cultures ensured better NT2-N neuronpreservation (Fig. 2B), yielding higher fractions of NT2-N neurons (quantified by cell counting after selectivetrypsinization; see Materials and Methods) from the totalnumber of NT2 cells at neurosphere harvesting. Thispercentage of neuronal differentiation was higher intrypsin-dissociated cultures (approximately 6.8%) than inthe other dissociated cultures (less than 1%), wheresmaller amounts of NT2-N neurons were detected onPDL-MG-coated surfaces (Fig. 2B). Trypsin digestion wasselected as the best method for neurosphere dissociationbecause it promoted greater recovery of NT2-N neurons.

In the second strategy, nondissociated neurosphereswere directly plated onto PDL-MG-coated surfaces. Af-ter 7 days of culture, no cell aggregates were detected inthe cultures plates. In addition, greater amounts of NT2-N neurons with extensive neuritic networks were cover-ing the entire surface. Here, the fraction of NT2-N cellsachieved from the initial neurosphere harvesting was12.3%, approximately two times higher than the fractionreached in the trypsin-dissociated cultures. These resultssuggested that harvesting non-dissociated neurospheres isthe best strategy to recover NT2-N neurons.

3560 Serra et al.

Journal of Neuroscience Research DOI 10.1002/jnr

Neurospheres and trypsin-dissociated NT2 cellswere also plated directly on uncoated tissue culturesflasks at harvesting. However, neither neurospheres nordissociated cells attached to those untreated surfaces(results not shown).

NT2-N Selection and Characterization

After neurosphere harvesting to coated surfaces,non-dissociated neurospheres cultures were further char-acterized in order to describe the changes and evolutionof the cultures. After 1 day of neurosphere plating, non-

Fig. 1. Phase-contrast photographs of NT2 cells cultured in stirred conditions (spinner). Small cellaggregates (�50 lm) were observed at day 2 (A). Afterward, these neurospheres grew in number(B; day 5) and in size (C; day 10), yielding an average diameter of approximately 300 lm at day15 (D). E: Immunofluorescence image of NT2 neurospheres collected at day 16: double labellingof b-TubIII (green) and nestin (red). Scale bars 5 100 lm in A–D; 50 lm in E.

Novel Culture Strategy for Stem Cells 3561

Journal of Neuroscience Research DOI 10.1002/jnr

dissociated NT2 neurospheres adhered well on PDL-MG-coated surfaces. During the first days, neurospheressprout large numbers of non-neuronal cells and alsosome NT2-N neurons; at this time, cultures are com-posed of stem cells, neural progenitor cells, and differen-tiated cells as detected by phase-contrast microscopy(Fig. 3A). This observation was confirmed by doublestaining the cell cultures with antibodies specific for ei-ther the undifferentiated and the neural progenitor NT2cells (rabbit anti-nestin) or NT2-N neurons (mouse anti-

b-TubIII; Fig. 3B). Incubation with mitotic inhibitoryconditions led to an elimination of proliferating NT2precursor cells, and the cultures gradually becameenriched into differentiated NT2-N neurons and exten-sive neuritic networks. After 7 days of treatment, wedetected that 95% of the cells present in culture plateswere NT2-N neurons; the rest of the cell populationappeared as flat cells with extensive cytoplasm (Fig.3C,D). At this point, no cell aggregates were detected inthe culture plates.

Fig. 2. A: Effect of neurosphere dissociation protocol in concentration of viable (white bars) andnonviable (gray bars) cells. B: Phase-contrast micrographs showing the non-dissociated and dissoci-ated neurosphere cultures, obtained immediately after each dissociation procedure (Day 0), or after7 days of culture under mitotic inhibitory conditions (Day 7). The column in the right corre-sponds to a fourfold magnification of the corresponding culture shown in the center column. Scalebars 5 50 lm.

3562 Serra et al.

Journal of Neuroscience Research DOI 10.1002/jnr

To enrich for neurons further, these cells wereselectively dislodged (see Materials and Methods) andplated in new culture dishes coated with PDL and MG.As shown in Figure 3E, pure cultures of NT2-N neu-rons with extensive neuritic networks covering theentire surface were obtained upon selection. The iden-tity of those final pure cultures was confirmed by posi-tive staining for the neuronal markers NF-L, NF200, b-TubIII, and MAP2a/b (Fig. 3E–I). By the end of theculturing strategy, the cells were also stained for neuralprecursor cells (nestin) and glial markers (GFAP forastrocytes and O4 for oligodendrocytes). Neither nestin-nor GFAP- and O4-positive cells was detected in thefinal cultures of NT2-N cells (not shown).

Expansion and Differentiation: Apparent Rates ofNT2 Cells in Stirred and Static Cultures

To assess the level of functionality of this novelNT2 differentiation process, the apparent rates ofexpansion, i.e., the rate of cell expansion obtained afterRA treatment period (35 days in static and 16 days instirred culture) and the apparent rate of differentiation,i.e., the rate of cell differentiation achieved during theRA treatment period (35 days in static and 14 instirred) were compared for the static and stirred suspen-sion culture. For this, it was assumed that no NT2-Ncells were present at inoculation day (day 0) and thatdifferentiation occurs during incubation with RA. Sub-sequent steps for neurons selection and recovery were

Fig. 3. Characterization of NT2-N cultures. A: Nondissociated NT2neurospheres adhered well on PDL-MG-coated surfaces after 1 dayof neurosphere platting. At this time, neurospheres sprout large num-bers of non neuronal cells and also some neurons. B: Immunofluo-rescence image of NT2 cultures after 1 day of neurosphere harvest-ing: Double labelling of b-TubIII (green) and nestin (red). C: After7 days, the culture became enriched into well-differentiated NT2-N

neurons. D: Immunofluorescence images of NT2 cultures after 7days of neurosphere harvesting: Double labelling of b-TubIII (green)and nestin (red). Nuclei are labelled with DAPI (blue). E: Pure cul-tures of NT2-N neurons obtained upon selection. F–I: Immunofluo-rescence images in pure NT2-N cultures: Labelling of b-TubIII (F),NF-L (G), MAP2a/b (H), and NF200 (I) in NT2-N neurons. Scalebars 5 50 lm in A, C-I; 100 lm in B.

Novel Culture Strategy for Stem Cells 3563

Journal of Neuroscience Research DOI 10.1002/jnr

not considered at the apparent differentiation rates cal-culation.

Figure 4 shows that the lengthy static adherentculture protocol promotes a higher apparent rate ofexpansion (5.6 3 103 cells � ml–1 � hr–1) compared withcells cultured in stirred suspension vessels (2.5 3 103

cells � ml–1 � hr–1). Concerning the apparent rate of differ-entiation, the highest value was achieved in stirred suspen-sion culture (5.43 102 NT2-N cells �ml–1 � hr–1 comparedwith 1.93 102 NT2-N cells � ml–1 � hr–1 obtained in staticculture). These findings suggest that higher amounts ofcontaminant cells are produced in static cultures.

Because of the specific limitations of each culturesystem (surface area in static system and minimal cellconcentration in stirred culture), different cell inoculationdensities were used, approximately 2 3 105 cells/ml and5 3 105 cells/ml for static and stirred culture, respectively.Therefore, the apparent differentiation rate results suggestthat the neuronal differentiation potential of both culturestrategies was practically similar.

Neuronal Differentiation of NT2 Cells in Stirredand Static Cultures

At the end of each neuronal differentiation strategy,the final concentrations of NT2-N neurons were calcu-lated. Table II shows that higher concentrations ofNT2-N neurons were achieved with the stirred culturestrategy in 23–28 days (1.8 3 105 cells/ml). In the lengthystatic culture protocol, the final concentration of NT2-Nneurons was about 1.6 3 105 cell/ml after 47–53 days ofculture.

Neuronal differentiation efficiency was also com-pared by analyzing the ratio between the number ofNT2-N neurons reached and the total amount of NT2cells harvested after the RA treatment period (Table II).For this, three experimental trials were evaluated foreach differentiation system. Reproducible results showed

that the lowest value of differentiation efficiency wasachieved by the static culture protocol (3.2% 6 0.8%),which was shown to be about four times less efficientthan the stirred culture protocol (13.6% 6 1.8%).

DISCUSSION

Several static culture protocols used for NT2 differ-entiation have been proposed to date. However, theyhave been found to be time consuming, to be extremelylaborious, to produce low yields of neurons per culture,and not to be scalable (Pleasure et al., 1992; Durandet al., 2003; Horroks et al., 2003). Here, a novel strategyto expand, accelerate, and enhance the neuronal differ-entiation process of NT2 cells was developed, overcom-ing these process limitations. Therefore, the challengewas to improve the proliferation and differentiationsteps, by culturing NT2 cells as 3D aggregates in thepresence of RA, using suspension stirred bioreactors,without compromising the differentiation potential.

Culturing NT2 cells in spinner flasks yielded neu-rospheres with homogenous morphology (Fig. 1) andavoided neurosphere clumping, resulting in a more effi-cient NT2 population than that found in static cultures(Pleasure et al., 1992; Durand et al., 2003). Moreover,the protocol developed drastically improves neuronal dif-ferentiation efficiency compared with static cultures,with a a fourfold increase of NT2 cells expressing severalneuronal markers together with a reduced time requiredfor the differentiation process (Table II, Fig. 4). Thefinding that NT2 differentiation is faster and more effi-cient in suspension than in adherent systems agrees withthe results previously reported by Durand et al. (2003).Although the studies reported by Durand et al. wereperformed in static conditions, the authors clearly showthat both cell–cell contacts (within the neurospheresstructures) and RA contribute to a more rapid neuronaldifferentiation process than in the conventional protocolwith cell monolayer culture (Pleasure et al., 1992).

The decrease in the NT2 apparent expansion rateobtained for the suspension stirred system does not com-promise the differentiation potential of NT2 precursors,suggesting that the proposed culturing protocol directsmore efficiently the neuronal differentiation process,reducing the percentage of contaminating cells (Fig. 4).These are important advances towards the developmentof scaleable protocols for stem cell differentiation, i.e.,significant improvement of differentiation process effi-ciency and, ultimately reaching higher concentrations of

Fig. 4. Apparent rates of NT2 cell expansion (white bars) and neuro-nal differentiation (gray bars) obtained by static and stirred (spinnerculture) differentiation protocols.

TABLE II. Comparison of Static and Stirred Culture Conditions

for NT2 Neuronal Differentiation

Static culture Stirred culture

Differentiation time 47 Days 23 Days

NT2-N final concentration 1.6 3 105 Cells/ml 1.8 3 105 Cells/ml

Differentiation efficiency 3.2% 6 0.8% 13.6% 6 1.8%

Laborious Yes No

Scaleable No Yes

3564 Serra et al.

Journal of Neuroscience Research DOI 10.1002/jnr

human neurons, one of the current drawbacks of stemcell technology (for review see Kallos et al., 2003;Ulloa-Montoya et al., 2005).

After NT2 expansion and differentiation, an effi-cient neurosphere harvesting process is required toensure a good recovery of differentiated neurons. Severalprotocols for NT2 cells and neurosphere dissociationhave been proposed to date (Pleasure et al., 1992; Senet al., 2004). However, because of the sensitivity of neu-ronal cells and neurosphere complexity, the preservationof NT2 neurons after each dissociation procedure couldbe compromised. In the present study, both mechanicaland enzymatic methods were evaluated, trypsin digestionbeing selected as the best strategy for neurosphere disso-ciation and NT2-N preservation. Nevertheless, the high-est recovery of NT2-N neurons was achieved whennondissociated neurospheres were collected directly totreated surfaces (with PDL-MG). This result indicatesthat both neurosphere dynamics and the matrix surfaces(PDL-MG) are sufficient to provide an optimal systemfor NT2-N neuron selection in a second step to opti-mize the neuronal differentiation process.

Overall, this study demonstrates that culturinghuman NT2 cells in stirred tank bioreactors increase byfourfold the culture differentiation efficiency. Further-more, the possibility of including an additional periodfor precursor cells proliferation is also of great advantage,insofar as it brings flexibility in what concerns the possi-bility to expand precursor NT2 cells before induction ofneuronal differentiation, yielding higher numbers of finalNT2-N cells per volume of culture. Static systems donot allow this design, because culture surface limits theNT2 neuronal differentiation potential.

In conclusion, this culture method is promising forNT2 proliferation and differentiation, ensuring a robust,easy to operate, scaleable, and reproducible system toculture NT2 pluripotent cells under a fully controlledenvironment. Studies on bioprocess optimization (in-cluding the effect of several process parameters, such aspO2, medium composition, growth factor concentra-tions, feeding strategies, and agitation rate) are ongoing,seeking further improvement of NT2 differentiationyields. The knowledge gained using this cell modelshould allow for a more straightforward application forthe culture of hESCs in bioreactors, a challenge in stemcell bioengineering.

ACKNOWLEDGMENTS

The authors acknowledge Prof. Virginia Lee andProf. John Trojanowski (Center for NeurodegenerativeDisease Research, University of Pennsylvania Schoolof Medicine) for the kind gift of NT2 cells. AntonioRoldao is also acknowledged for his support and helpfuldiscussions on the manuscript. The financial supportreceived from Merck Sharp and Dohme (Portugal) andfrom the European Project CellPROM (No. 500039-2)is greatly acknowledged.

REFERENCES

Andrews PW. 1984. Retinoic acid induces neuronal differentiation of a

cloned human embryonal carcinoma cell line in vitro. Dev Biol

103:285–293.

Andrews PW. 2002. From teratocarcinomas to embryonic stem cells. R

Soc 357:405–417.

Andrews PW, Przyborski SA, Thomnson JA. 2001. Embryonal carci-

noma cells as embryonic stem cells. Stem Cell Biology Monograph 40.

Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.

Bongso A, Fong CY, Ratnam S. 1994. Isolation and culture of inner cell

mass cells from human blastocysts. Hum Reprod 9:2110–2117.

Cameron CM, Hu WS, Kaufman DS. 2006. Improved development of

human embryonic stem cell-derived embryoid bodies by stirred vessel

cultivation. Biotechnol Bioeng 94:938–948.

Collins PC, Nielsen LK, Patel SD, Papoutsakis ET, Miller WN. 1998a.

Characterization of hematopoietic cell expansion, oxygen uptake and

glycolysis in a controlled, stirred-tank bioreactor system. Biotechnol

Prog 14:466–472.

Collins PC, Miller WN, Papoutsakis ET. 1998b. Stirred culture of pe-

ripheral and cord blood hematopoietic cells offers advantages over tradi-

tional static systems for clinical relevant applications. Biotechnol Bioeng

59:534–543.

Davila JC, Cezar GG, Thiede M, Strom S, Miki T, Trosko J. 2004. Use

and application of stem cells in toxicology. Toxicol Sci 79:214–223.

Durand FP, Tan S, Bicker G. 2003. Turning teratocarcinoma cells into

neurons: rapid differentiation of NT-2 cells in floating spheres. Brain

Res Dev Brain Res 142:161–167.

Ferrari A, Ehler E, Nitsch RM, Gstz J. 2000. Immature human NT2

cells grafted into mouse brain differentiate into neuronal and glial cell

types. FEBS Lett 486:121–125.

Guillemain I, Alonso G, Patey G, Privat A, Chaudieu I. 2000. NT2 neu-

rons express a large variety of neurotransmission phenotypes in vitro. J

Comp Neurol 422:380–395.

Hartley RS, Margulis M, Fishman PS, Lee VM-Y, Tang C-M. 1999.

Functional synapses are formed between human NTera2 (NT2N,

hNT) neurons grown on astrocytes. J Comp Neurol 407:1–10.

Horroks GM, Lauder L, Stewart R, Przyborski S. 2003. Formation of

neurospheres from human embryonal carcinoma stem cells. Biochem

Biophys Res Commun 304:411–416.

Jones JM, Thomson JA. 2000. Human embryonic stem cell technology.

Semin Reprod Med 18:219–223.

Kallos MS, Sen A, Behie LA. 2003. Large-scale expansion of mammalian

neural stem cells: a review. Med Biol Eng Comp 41:271–282.

Kondziolka D, Wechsler L, Bynum L. 2001. Tranplantation of cultured

human neuronal cell for patients with stroke. Neurology 55:565–569.

McNeish J. 2004. Embryonic stem cells in drug discovery. Nat Rev

Drug Discov 3:70–80.

Moreira JL, Alves PM, Rodrigues JM, Cruz PE, Aunins JG, Carrondo

MJ. 1994. Studies of baby hamster kidney natural cell aggregation in

suspended batch cultures. Ann N Y Acad Sci 745:122–33.

Pal R, Ravindran G. 2006. Assessment of pluripotency and multilineage

differentiation potential of NTERA-2 cells as a model for studying

human embryonic stem cells. Cell Prolif 39:585–98.

Pleasure SJ, Lee VM-Y. 1993. NTera 2 cells: a human cell line which

displays characteristics expected of a human commited neuronal pro-

genitor cell. J Neurosci Res 35:585–602.

Pleasure SJ, Page C, Lee VM-Y. 1992. Pure, postmitotic, polarized

human neurons derived from NTera2 cells provide a system for

expressing exogenous proteins in terminally differentiated neurons. J

Neurosci 12:1802–1815.

Przyborski SA, Chritie VB, Hayman MW, Stewart R, Horrocks GM.

2004. Human embryonal carcinoma stem cells: models of embryonic

development in humans. Stem Cells Dev 13:400–408.

Novel Culture Strategy for Stem Cells 3565

Journal of Neuroscience Research DOI 10.1002/jnr

Sen A, Kallos MS, Behie LA. 2002. Passaging protocols for mammalian

neural stem cells in suspension bioreactors. Biotechnol Prog 18:337–

345.

Sen A, Kallos MS, Behie LA. 2004. New tissue dissociation protocol for

scaled-up production of neural stem cells in suspension bioreactors.

Tissue Eng 10:904–913.

Stewart R, Christie VB, Przyborski SA. 2003. Manipulation of human

pluripotent carcinoma stem cells and the development of neural sub-

types. Stem Cells 21:248–256.

Thomson JA, Itzkovitz-Eldor J, Sharpiro SS, et al. 1998. Embryonic

stem cell lines derived from human blastocysts. Science 282:1145–

1147.

Ulloa-Montoya F, Verfaillie CM, Hu W. 2005. Culture systems for plu-

ripotent stem cells. J Biosci Bioeng 1:12–27.

Watson DJ, Longhi L, Wolf JH. 2003. Geneticaly modified NT2N

human neuronal cells mediate long-term gene expression as CNS grafts

in vivo and improve functional cognitive outcome following experi-

mental traumatic brain injury. J Neuropathol Exp Neurol 62:368–380.

Yoshioka A, Yudkoff M, Pleasure D. 1997. Expression of glutamic acid

decarboxylase during human neuronal differentiation: studies using the

Ntera-2 culture system, Brain Res 767:333–339.

Youn BS, Sen A, Kallos MS, Behie LA, Girgis AG, Kurpios N, Barcelon M,

Hassel JA. 2005. Large-scale expansion of mammariay epithelial stem cell

aggregates in suspension bioreactors. Biotechnol Prog 21:984–993.

3566 Serra et al.

Journal of Neuroscience Research DOI 10.1002/jnr