ARTICLE

Shear Stress During Early Embryonic Stem CellDifferentiation Promotes Hematopoietic andEndothelial Phenotypes

Russell P. Wolfe, Tabassum Ahsan

Department of Biomedical Engineering, Tulane University, 500 Lindy Boggs Center, New

Orleans, LA 70118; telephone: 504-865-5899; fax: 504-862-8779; e-mail: tahsan@tulane.edu

ABSTRACT: Pluripotent embryonic stem cells (ESCs) are apotential source for cell-based tissue engineering and regen-erative medicine applications, but their translation intoclinical use will require efficient and robust methods forpromoting differentiation. Fluid shear stress, which can bereadily incorporated into scalable bioreactors, may be onesolution for promoting endothelial and hematopoietic phe-notypes from ESCs. Here we applied laminar shear stress todifferentiating ESCs using a 2D adherent parallel plateconfiguration to systematically investigate the effects ofseveral mechanical parameters. Treatment similarly promot-ed endothelial and hematopoietic differentiation for shearstress magnitudes ranging from 1.5 to 15 dyne/cm2 and forcells seeded on collagen-, fibronectin- or laminin-coatedsurfaces. Extension of the treatment duration consistentlyinduced an endothelial response, but application at laterstages of differentiation was less effective at promotinghematopoietic phenotypes. Furthermore, inhibition of theFLK1 protein (a VEGF receptor) neutralized the effects ofshear stress, implicating the membrane protein as a criticalmediator of both endothelial and hematopoietic differenti-ation by applied shear. Using a systematic approach, studiessuch as these help elucidate the mechanisms involved inforce-mediated stem cell differentiation and inform scalablebioprocesses for cellular therapies.

Biotechnol. Bioeng. 2013;110: 1231–1242.

� 2012 Wiley Periodicals, Inc.

KEYWORDS: pluripotent stem cells; endothelial; hemato-poietic; bioreactor; shear stress; bioprocessing

Introduction

Differentiated phenotypes derived from pluripotent stemcells are potential cell sources for tissue engineering andtherapeutic applications, but effective and robust methodsfor scale-up production have yet to be identified. Directeddifferentiation techniques can increase cell yield but remainfairly inefficient and often require exogenous growth factors,which are considered a costly resource for large-scale cellproduction. Increasing evidence suggests that appliedphysical forces, such as fluid shear stress, may be auseful tool for promoting differentiation of stem cells,including embryonic stem cells (ESCs) (Stolberg andMcCloskey, 2009). Thus, differentiation approaches togenerate the clinically relevant cell numbers (107–109)(Kehoe et al., 2010; Kirouac and Zandstra, 2008) may benefitfrom leveraging the shear stresses inherent to manyproduction-scale bioreactors.

Scale-up bioreactors are typically based on stir-basedsuspension systems, which not only remove gradients ofcytokines and metabolic byproducts caused by biologicalactivity, but also decrease the medium to cell ratio byminimizing areas of stagnant volume. ESCs differentiated ascell aggregates, known as embryoid bodies (EBs), andcultured in these bioreactor systems have been found to haveimproved EB homogeneity (Carpenedo et al., 2007;Gerecht-Nir et al., 2004), differentiation towards cardio-myocytes (Niebruegge et al., 2009; Sargent et al., 2009), andmaintenance of hematopoietic (Cameron et al., 2006)phenotypes. Even though these are macroscopically well-mixed environments, aggregation of 3D cell clusters (withsome cells in an interior core and others forming outerlayers) leads to an inherently varied microenvironment forthe individual cells with potential limitations in diffusion(Azarin and Palecek, 2010). This suggests an opportunity tofurther enhance for differentiation towards target pheno-types if inputs could be homogeneously presented to theentire cell population. Adherent culture of ESCs ontosurfaces of microcarriers is one potential alternative thatameliorates many of the heterogeneity problems while

The authors have no conflict of interest.

Correspondence to: T. Ahsan

Contract grant sponsor: NIH

Contract grant number: #P20 GM103629

Additional supporting information may be found in the online version of this article.

Received 9 July 2012; Revision received 30 September 2012; Accepted 30 October 2012

Accepted manuscript online 8 November 2012;

Article first published online 15 February 2013 in Wiley Online Library

(http://onlinelibrary.wiley.com/doi/10.1002/bit.24782/abstract)

DOI 10.1002/bit.24782

� 2012 Wiley Periodicals, Inc. Biotechnology and Bioengineering, Vol. 110, No. 4, April, 2013 1231

preserving the use of scalable suspension culture (Abrancheset al., 2007; Fernandes et al., 2007). Initial studies indicate,however, that cells cultured on microcarriers may be moresensitive to shear stress (Chisti, 2001; Gregoriades et al.,2000; Hua et al., 1993; Mollet et al., 2004) and thereforemotivate a more thorough understanding of the biologicalimpact of shear stress parameters.

Shear stress effects are often difficult to interpret inscalable systems due to complex patterns of fluid flow.Consequently, single cells or cell aggregates in these systemsare commonly exposed to shear stresses up to 20 dyne/cm2

that vary based on location within the bioreactor chamber,such as distance from the impeller (Vallejos et al., 2011;Venkat et al., 1996) or the chamber walls (Bilgen et al.,2006). Instead, use of simpler configurations, which allowapplication of a well-defined shear stress, can help decipherthe effect of shear stress parameters. For example, theparallel plate system, which applies a uniform shear stress tocells attached to a surface, is a suitable surrogate to study theeffect of shear stress magnitude and protein substrate foradherent cell culture on microcarrier surfaces.

Adherent model systems have shown that shear stresspromotes an endothelial phenotype when applied toendothelial progenitor cells (Obi et al., 2009; Yamamotoet al., 2003) or to ESC-derived populations sorted for FLK1(Yamamoto et al., 2005) or CD31 (Nikmanesh et al., 2012).Separate studies, which sorted for CD41þ cells, found thatshear stress can promote the maturation of cells in thehematopoietic compartment (Adamo et al., 2009). Withoutextensive differentiation or cell sorting, it has been foundthat the sustained application of shear stress during earlierstages of ESC differentiation can be used to promotethe mesodermal lineage and even inhibit endodermalspecification (Wolfe et al., 2012). Furthermore, shear stressapplied under those same conditions promoted anendothelial phenotype (Ahsan and Nerem, 2010; Zenget al., 2006). It has yet to be determined, however, if shearstress applied during early ESC differentiation, prior tohematopoietic commitment, can also be used to generatehematopoietic cells.

Studies in development have long observed thathematopoietic and endothelial cells arise with spatial andtemporal proximity in the embryo (Murray, 1932; Sabin,1920), suggesting that both phenotypes emerge in similarmicroenvironments in vivo. Recent in vitro studies usingESCs have shown that endothelial and hematopoietic cellsare derived from a common progenitor that expresses theVEGF receptor FLK1 (Eilken et al., 2009; Lancrin et al.,2009), which is now known to be expressed by progenitorsof several mesodermal phenotypes (Bautch, 2006; Ema et al.,2006). Taken together, this implies that identical cultureconditions may drive differentiation towards both endothe-lial and hematopoietic phenotypes through common stagesof early specification. Thus, a simultaneous analysis of bothendothelial and hematopoietic differentiation may helpelucidate the shear-mediated mechanism present duringearly ESC differentiation.

The objective of this study was to determine if theapplication of fluid shear stress is a robust method forgenerating endothelial and hematopoietic phenotypes frompluripotent cells. Using an in vitro 2D parallel plate modelsystem, we found that during early ESC differentiation shearstress promoted both endothelial and hematopoieticphenotypes for a range of shear stress magnitudes andunderlying protein substrates. The response was amplifiedwith increasing durations of shear stress, although delayedapplication of treatment was less effective at promotinghematopoietic phenotypes. Inhibition with SU1498 nullifiedthe effects of stress treatment, suggesting this process ismediated by activation of the FLK1 membrane protein.Together these results elucidate the mechanisms behindforce-mediated stem cell differentiation and provide insightfor improving scale-up bioreactor design which can helprealize the promising impact of stem cells on public health.

Materials and Methods

Expansion of Mouse Embryonic Stem Cells

Mouse D3 ESCs and embryonic fibroblasts (MEFs) werepurchased from ATCC (Manassas, VA) and cultured aspreviously described (Ahsan and Nerem, 2010). Briefly,ESCs were expanded on mitotically-inactivated MEFs andthen stored in liquid nitrogen. Prior to experiments, ESCswere thawed and expanded for one passage on gelatin-coated tissue culture plastic. Culture medium consisted ofDulbecco’s Modification of Eagles Medium with 15%ES-qualified fetal bovine serum (Invitrogen, Carlsbad, CA),2mM L-glutamine, 0.1mM non-essential amino acids,1,000U/mL Leukemia Inhibitory Factor (LIF; ESGRO1

from EMD Millipore, Billerica, MA) and antibiotics.

Application of Shear Stress

ESCs were differentiated without LIF on protein-coatedglass slides to promote cell attachment before shear stresstreatment, as described previously (Ahsan and Nerem,2010). Briefly, either collagen type IV (COL: BD Biosciences,Bedford, MA), fibronectin (FN: BD Biosciences) or laminin(LM: Sigma–Aldrich, St. Louis, MO) was adsorbed ontoglass slides for 1 h at a concentration of 3.5mg/cm2. ESCswere then seeded onto slides (10,000 cells/cm2) and culturedin an incubator (378C, 5% CO2) for up to 3 days indifferentiation medium, consisting of Minimum EssentialAlpha Media supplemented with 10% fetal bovine serum,0.1mM beta-mercaptoethanol, and antibiotics.

Fluid shear stress was applied using a parallel platebioreactor system (Ahsan and Nerem, 2010; Levesque andNerem, 1985) to cells attached to glass slides as describedabove. A media reservoir and peristaltic pump (both fromCole-Parmer, Vernon Hill, IL) were connected in series witha dampener and flow block. Slides were placed in the flow

1232 Biotechnology and Bioengineering, Vol. 110, No. 4, April, 2013

block such that t¼ 6Qm/(bh2), where the shear stress (t)depended on the flow rate (Q), viscosity of the media (m), aswell as the width (b) and height (h) of the flow channel.Using this system we applied steady laminar shear stress at15, 5.0 or 1.5 dyne/cm2 for up to 4 days (SHEAR samples).These magnitudes are comparable to the physiological levelsobserved in the adult (15 dyne/cm2) and embryonic(5.0 dyne/cm2) aortas (Adamo et al., 2009), in addition tolower magnitudes (1.5 dyne/cm2) which are known to effectadult phenotypes including hemo-vascular cells (Fukudaand Schmid-Schonbein, 2003; Gerszten et al., 1998). Timematched controls were cultured under static conditions byplacing slides in petri dishes (STATIC samples). Samplesfrom both groups were cultured with 125mL of differentia-tion medium at 378C and 5% CO2. For FLK1 inhibitionstudies, medium was supplemented with 4.5mg/mL SU1498(Calbiochem1 from EMD Millipore) during the 2-daySTATIC or SHEAR treatment.

Analysis for mRNA Expression

Gene expression of mesodermal, endothelial, and hemato-poietic markers were evaluated using real-time RT-PCR.Samples were lysed and RNA isolated using QIAshreddersand the RNeasy kit (Qiagen, Valencia, CA). RNAconcentrations were quantified using a Nanodrop1

spectrophotometer and 1mg of RNA from each samplewas converted into cDNA (Invitrogen Superscript1 IIIFirst-strand synthesis). SYBR1 Green (Applied Biosystems,Carlsbad, CA) and a StepOnePlusTM PCR system were usedto quantify mRNA expression using the standard curvemethod where flk1, tie2, and runx1 were evaluated as earlymarkers of mesodermal, endothelial, and hematopoieticdifferentiation, respectively. Downstream endothelial (pecam1and klf2) and hematopoietic (cd41, c-kit, cd34, cd133, scl,and vav1) genes were also assessed. All gene expressionlevels were normalized to the housekeeping gene gapdh.Descriptions and primer sequences for individual genes arelisted in Supplemental Table I.

Analysis for Protein Expression

Flow cytometry was used to quantitate changes in proteinexpression due to shear stress treatment, as done previously(Ahsan and Nerem, 2010). StemPro1 Accutase1

(Invitrogen) was used to generate single cell suspensionsthat were fixed in 4% formaldehyde for 15min and thenstored in a buffer solution consisting of PBS with 0.3% BSAand 0.001% TWEEN120 (Sigma–Aldrich). Cells werepermeabalized using 0.5% triton-X (Sigma–Aldrich),blocked with 10% serum, and then stained with primaryand secondary antibodies. Dilutions for primary andsecondary (as needed, AF488 Invitrogen) antibodies were1:100 and 1:200, respectively. Antibodies used were FLK1(Fitzgerald, Acton, MA), PE conjugated-TIE2 (Abcam,Cambridge, MA), as well as RUNX1, PECAM1, vWF, eNOS,and CD41 (all from Santa Cruz Biotechnology, Santa Cruz,

CA). Fluorescence was detected using a BD FACSCanto II(Supplemental Figs. 1–3). For each sample, the positive-expressing cells were defined as being above the 98% level ofthe flow cytometry controls. Characterization of certaindownstreammarkers used an upper quartile percentage (i.e.,the percent of cells above the 75% level of flow cytometrycontrols) to detect more subtle differences in expressionlevel. Values of samples from the same group were averagedfor statistical analysis.

Hematopoietic Colony Assessments

Quantitative analysis of colony formation was evaluatedto determine hematopoietic potential. At the end ofSTATIC and SHEAR culture, individual samples weretreated with Accutase1 and single cell solutions were putinto suspension culture on an orbital shaker system(40 rpm) to form aggregates. After 7 days, cell clusterswere dissociated and assessed for hematopoietic potentialusing MethoCult1 GF M3434 (STEMCELL Technologies,Vancouver, CA). Following the manufacturer’s instructions,cells were placed into low adherence 35mm petri dishescontaining the provided methylcellulose medium with thenumber of colonies in the entire dish counted after 10 days.

Statistical Analysis

All quantitative data are presented as mean� SEM. Studentpaired t-tests were used to detect differences betweenSHEAR samples and time- and condition-matched STATICcontrols. For all statistical comparisons, P values<0.05 wereconsidered significant.

Results

Shear Stress Promotes Early Differentiation

Initial studies screened for the effects of shear stresstreatment on a variety of protein substrates and for a rangeof stress magnitudes. Allowing for 2 days of initialdifferentiation followed by 2 days of treatment (STATICvs. SHEAR), we evaluated gene changes for early differenti-ation markers flk1 (mesoderm), tie2 (endothelial), andrunx1 (hematopoietic) for cells seeded on COL-, FN- or LM-coated slides and exposed to 1.5, 5.0 or 15 dyne/cm2 shearstress. We have previously shown (Ahsan and Nerem, 2010)that protein expression of mesodermal and endothelialmarkers were upregulated in response to 15 dyne/cm2 onCOL-coated slides. Under these same conditions, here wefound significant (P< 0.05) gene upregulation for flk1(>2.6-fold), as well as tie2 (>1.5-fold) and runx1 (>1.5-fold) as compared to STATIC controls (Fig. 1). Here wefurther established that regulation of these genes was notfound to be stress magnitude dependent, as levels ofupregulation were similar at 1.5, 5.0, and 15 dyne/cm2. Toinstead look at the effects of protein coating, cells wereseeded onto slides coated with either FN or LM and thenexposed to shear stress (5.0 dyne/cm2). Cell populations,

Wolfe and Ahsan: Shear Stress and ESC Differentiation 1233

Biotechnology and Bioengineering

though heterogeneous within any given sample, were alllargely adherent and displayed similar morphologies acrossthe different substrate groups (Fig. 2A). Likewise, shearstress responses for the flk1, tie2, and runx1 genes were allsimilar for the different substrates (Fig. 2B). Thus, a range ofstress magnitudes and protein coatings seem equally suitableduring shear-stress mediated differentiation and this type ofmechanical modulation can simultaneously promote bothendothelial and hematopoietic phenotypes.

Extended Durations of Shear Stress PromoteEndothelial and Hematopoietic Phenotypes

To determine the effects of longer durations of treatment,ESCs were seeded onto protein-coated slides (COL was usedin all subsequent studies), cultured under static conditionsfor 2 days to allow for cell attachment, and then eithercontinued under STATIC conditions or exposed to SHEAR(5.0 dyne/cm2) for 1, 2 or 4 days. Gene expression(normalized to gapdh) was quantified and SHEAR samples

were compared to time-matched STATIC controls (Fig. 3).Early differentiation markers flk1, tie2, and runx1 weresignificantly upregulated for all tested durations (Fig. 3A, B,and E), with sustained treatment leading to the greatestlevels of increase (upregulation after 4 days by 6.8� for flk1and 2.9� for both tie2 and runx1). While the downstreamendothelial markers klf2 and pecam1 underwent little or nochange in expression after just 1 day of shear stress (Fig. 3C–D), extending treatment to 4 days led to upregulatedexpression by 3.7� and 2.7�, respectively. To determine ifthe effects on gene expression led to actual changes towardsan endothelial phenotype, protein expression was evaluatedin samples exposed to 4 days of treatment. Flow cytometryanalysis showed shear stress significantly (P< 0.01) in-creased the percentage of cells positively expressing FLK1 to57� 4% (compared to 34� 2% in the STATIC controls,Fig. 4A and B). Shear-treatment also increased thepercentage of cells positively expressing TIE2 from8� 0.3% to 10� 0.6% (P< 0.05) and RUNX1 from5� 1% to 9� 1% (P< 0.05) compared to STATIC controls(Fig. 4C and D). Protein expression of mature endothelialmarkers also increased (Fig. 5A and C): the percentage ofPECAM1þ cells increased from 34� 1% to 42� 3%(P< 0.05) with slight but significant increases in NOS3and VWF observed based on quartile analysis (P< 0.01 andP< 0.05, respectively). Thus, extended application of shearstress elevates gene and protein expression indicative of anendothelial phenotype.

The effect of extended durations of shear treatment onhematopoietic differentiation was also evaluated. Asdiscussed above, tie2 and runx1, early markers of endothelialand hematopoietic differentiation, respectively, were simi-larly upregulated with extended durations of sheartreatment (Fig. 3B and C). In contrast to gene changes indownstream endothelial markers, however, applied shearstress had a transient effect on gene expression of thedownstream hematopoietic marker cd41. After 1 or 2 days ofSHEAR treatment, cd41 was upregulated by >1.5-fold(P< 0.01). With 4 days of treatment, however, the effectswere lost and levels for SHEAR samples and STATICcontrols were almost identical (Fig. 3F). Similarly, theexpression of the hematopoietic marker c-kit was notdetectably different after 4 days of treatment (Fig. 3G). Tofurther explore the effects of shear stress on hematopoieticgene expression, additional markers (cd34, scl, cd133, andvav1) were evaluated and it was found that only modesteffects were observed overall (Supplemental Fig. 4).Although the transient upregulation of cd41 gene expressionwas over by 4 days of shear stress treatment, proteinexpression at that time was significantly (P< 0.05)increased (Fig. 5B). This indication of sustained phenotypicchanges prompted evaluation in hematopoietic potentialusing the colony forming assay. After maturation in rotary-suspension culture for 7 days, cells from both SHEARand STATIC samples generated hematopoietic colonies(Fig. 5B), with those from the SHEAR group generatingalmost twice as many in number (1.9�, P< 0.05). While the

1.5 5.0 150

2

4

6

flk1

[Nor

mal

ized

to S

tatic

Con

trol

s]

** ****

A

B

C1.5 5.0 15

0

1

2

3

tie2

[Nor

mal

ized

to S

tatic

Con

trol

s]

** *** *

1.5 5.0 150

1

2

3

Magnitude [dynes/cm2]

runx1

[Nor

mal

ized

to S

tatic

Con

trol

s]

******

Figure 1. A range of magnitudes induce similar shear stress responses. Cells

were seeded on slides coated with collagen type-IV, cultured under static conditions

for 2 days, and then for 2 days either exposed to 1.5, 5.0 or 15 dyne/cm2 of shear stress

(SHEAR samples) or maintained in culture under static conditions (STATIC controls).

flk1, tie2, and runx1 gene expression was assessed for SHEAR samples (&) and

normalized to trial-matched STATIC controls (&). Data presented are mean�SEM

(n¼ 7–12), where significant differences between STATIC and SHEAR groups are

indicated with asterisks (�P< 0.05, ��P< 0.01, and ���P< 0.001).

1234 Biotechnology and Bioengineering, Vol. 110, No. 4, April, 2013

various colony types formed by cells in the hematopoieticcompartment were not discerned, the negligible effects ofshear stress on downstream hematopoietic gene markersindicate that our treatment may primarily promote theinitial differentiation towards an early hematopoieticlineage. Thus, the application of shear stress for 4 dayspromotes a functional hematopoietic phenotype, which iscorrelated to gene expression of runx1 but not otherhematopoietic markers.

Shear Stress-Mediated Differentiation Depends onTime of Application

The initial attachment period prior to the application oftreatment induces early differentiation due to the removalof LIF (Wolfe et al., 2012; Supplemental Fig. 5). Two dayswas used for the studies described above, but for thesestudies the attachment period was varied from 1 to 3 daysfollowed by 1 day of treatment (STATIC vs. SHEAR at15 dyne/cm2). SHEAR treatment after 2 days of attachmentinduced a significant increase in gene expression of flk1,

tie2, and runx1, consistent with the 2-day treatment resultsdescribed above; however, SHEAR effects were found to bedependent on pre-treatment duration (Fig. 6). flk1 wasupregulated after 1 (4.4�, P< 0.05) or 2 (6.0�, P< 0.01)days of pre-treatment, but differences were no longerdetectable (P¼ 0.06) after 3 days (Fig. 6A). The earlyendothelial marker tie2 was significantly upregulated for 1,2 or 3 days of pre-treatment, but the effect droppedfrom a 1.9� increase after 1 day to <1.6� after 2 or 3 days(Fig. 6B). The early hematopoietic marker runx1, incontrast, was only significantly upregulated when treat-ment was applied after 2 days (3.1-fold, P< 0.05) but hadno effect outside this brief window (Fig. 6C). Thus, theeffect of shear stress is dependent on the stage ofdifferentiation.

Delayed Application of Shear Stress Is Less Effective atPromoting Hematopoietic Phenotypes

Endothelial and hematopoietic phenotypic changes wereevaluated by protein expression after either 2 or 3 days of

Figure 2. Various proteins induce similar shear stress responses. Cells were seeded on slides coated with collagen type-IV (COL), fibronectin (FN) or laminin (LM), cultured

under static conditions for 2 days, and then for 2 days either exposed to 5.0 dyne/cm2 of shear stress (SHEAR samples) or maintained in culture under static conditions (STATIC

controls). Phase images of cells for both STATIC controls and SHEAR samples on COL, FN, and LM are shown. flk1, tie2, and runx1 gene expression was assessed for SHEAR

samples (&) and normalized to trial-matched STATIC controls (&). Data presented are mean�SEM (n¼ 7–12), where significant differences between STATIC and SHEAR groups

are indicated with asterisks (��P< 0.01 and ���P< 0.001). Scale bar represents 100mm.

Wolfe and Ahsan: Shear Stress and ESC Differentiation 1235

Biotechnology and Bioengineering

0 1 2 3 4 5 60.0

0.2

0.4

0.6

Sta

ticC

ultu

re

Duration [days]flk1

[M/M

gapdh

]

*** **

*

0 1 2 3 4 5 60

2

4

Sta

ticC

ultu

re

Duration [days]

tie2

[mM

/M gapdh

] ********

0 1 2 3 4 5 60

2

4

Sta

ticC

ultu

reDuration [days]

runx1

[mM

/M gapdh

] ***

******

0 1 2 3 4 5 60

20

40

Sta

ticC

ultu

re

Duration [days]

pecam1

[mM

/M gapdh

]

*

***

0 1 2 3 4 5 60.0

0.1

0.2

0.3

Sta

ticC

ultu

re

Duration [days]

klf2

[M/M

gapdh

] ***

***

0 1 2 3 4 5 60.0

0.1

0.2

0.3

Sta

ticC

ultu

re

Duration [days]

cd41

[mM

/M gapdh

] ** **

0 1 2 3 4 5 60

10

20

Sta

ticC

ultu

re

Duration [days]

c-kit[

M/M

gapdh

]

****

A

B

C

D

E

F

G

Figure 3. Extending treatment duration promotes endothelial gene expression. All samples were initially allowed to adhere for 2 days under static conditions on COL

(indicated by gray shading) and then experimental samples were exposed to shear stress (5.0 dyne/cm2) for 1, 2 or 4 days (~, —) while time-matched controls were cultured under

static conditions (*, - - -). Gene expression normalized to gapdh is shown for flk1 (A), tie2 (B), pecam1 (C), klf2 (D), runx1 (E), cd41 (F), and c-kit (G). Data presented are mean�SEM

with n¼ 9–12 (3–4 independent trials), where significant differences between time-matched STATIC and SHEAR groups are indicated with asterisks (�P< 0.05, ��P< 0.01,���P< 0.001).

1236 Biotechnology and Bioengineering, Vol. 110, No. 4, April, 2013

attachment followed by 4 days of treatment (STATIC vs.SHEAR at 5.0 dyne/cm2). SHEAR treatment significantlyincreased the percentage of cells expressing the earlymesodermal marker FLK1 (Fig. 7A) after either 2 or 3 daysof pre-treatment (P< 0.01 and P< 0.001, respectively). Interms of the endothelial phenotype, the percentage ofcells expressing the early marker TIE2 was significantly(P< 0.05) increased for SHEAR samples at both pre-treatment durations (Fig. 7B). Thus, shear stress applied after3 days of pre-treatment still promoted an endothelialphenotype. Conversely, the significant upregulation due toSHEAR of the early RUNX1 (Fig. 7C) and the downstreamCD41 (Fig. 7D) hematopoietic markers after 2 days ofattachment was lost with the additional day of pre-treatment.Thus, the data after 2 days of pre-treatment confirms thatapplied shear stress promotes both endothelial andhematopoietic phenotypes, though it seems that applicationat later time points may be less effective at promotinghematopoietic differentiation.

FLK1 Is Required for Shear-Mediated Differentiation

FLK1, a receptor for the vascular endothelial growth factor,has long been identified as a mechanosensor of shear stress

in mature endothelial cells (Chen et al., 1999). To determinethe role of FLK1 in shear stress-mediated endothelial andhematopoietic differentiation of ESC’s, an inhibitor of FLK1activation (SU1498) was added to the medium. All sampleswere first cultured for 2 days under static conditions to allowfor cell attachment, then samples were designated for anadditional 2 days under static (STATIC) or shear stress(SHEAR, t¼ 5.0 dyne/cm2) conditions, in the presence orabsence of SU1498. For the trial-matched controls tie2and runx1 gene expression were significantly (P< 0.001)

B DC

FLK1

%Po

sitiv

eC

ells

0

40

80

**

TIE20

10

20

*

RUNX10

10

20

*

FLK1

Eve

nts

100 101 102 103 1040

50

100

STATIC SHEAR20 Control

A

Figure 4. Extended durations of shear stress promote hematopoietic and en-

dothelial phenotypes. Samples were exposed to 4 days of either STATIC (white striped

bars) or SHEAR treatment (t¼ 5.0 dyne/cm2: black striped bars). A representative

histogram of FLK1 protein expression (A) is shown for an immunostaining 28 antibody-only control (filled gray), a STATIC experimental control (gray line), and a SHEAR

experimental sample (black line). Fluorescence of both STATIC and SHEAR samples

were assessed and compared to flow cytometry control samples to calculate the

percentage of cells positive for markers of early differentiation. Assessed proteins

were FLK1 (B), TIE2 (C), and RUNX1 (D). Data presented are mean�SEM (n¼ 3–4),

where significant differences between STATIC and SHEAR groups are indicated by

asterisks (�P< 0.05, ���P< 0.001)

PECAM1

% P

ositi

ve C

ells

0

30

60

*

CD410

10

20

*

A B

NOS3

Even

ts

100

101

102

103

104

0

31

63

94

125Upper Quartile:

STATIC: 86±1.3%SHEAR: 95±0.7%

p<0.01

C

VWF

Even

ts

100

101

102

103

104

0

31

63

94

125Upper Quartile:

STATIC: 51±2.6%SHEAR: 63±2.5%

p<0.05

# C

olon

ies/

20k

cells

0

5

10

15

*

STATIC SHEAR

Figure 5. Extended durations of shear stress promote mature and definitive

phenotypes. Samples were exposed to 4 days of either STATIC (white bars, striped or

dotted) or SHEAR treatment (t¼ 5.0 dyne/cm2: black bars, striped or dotted). Fluores-

cence of both STATIC and SHEAR samples were assessed and compared to flow

cytometry control samples to calculate the percentage of cells positive for differenti-

ation markers. Assessed proteins were PECAM1 (A) and CD41 (B, Left). A represen-

tative histogram of NOS3 and VWF protein expression (C) is shown for an

immunostaining 28 antibody-only control (filled gray), a STATIC experimental control

(gray line), and a SHEAR experimental sample (black line). To detect subtle changes in

protein expression, quartile analysis was used to quantify the expression of NOS3 and

VWF. Following STATIC and SHEAR treatment, hematopoietic potential was assessed

using a colony counting assay (B, Right). Data presented are mean�SEM (n¼ 3–4),

where significant differences between STATIC and SHEAR groups are indicated by

asterisks (�P< 0.05) or displayed P-value.

Wolfe and Ahsan: Shear Stress and ESC Differentiation 1237

Biotechnology and Bioengineering

upregulated in response to shear stress, as was shown earlier,but here it was found that the addition of SU1498completely prevented (P¼ 0.62 and 0.93, respectively)this response (Fig. 8). Thus, these results indicate that FLK1may be a necessary mediator for early endothelial andhematopoietic differentiation of ESCs due to applied shearstress.

Discussion

Establishing ESC-derived cell sources for tissue engineeringand therapeutic applications will require methods forpromoting differentiation which are not only effective butrobust enough for scale up applications. Here, we found thatapplication of 2 days of shear stress during early ESCdifferentiation increased the genetic expression of earlymesoderm, hematopoietic, and endothelial markers for arange of shear stress magnitudes and underlying proteinsubstrates. Increasing the duration of shear stress furtheraugmented the effect on genetic markers for early

mesoderm, endothelial, and hematopoietic differentiation,as well as downstream endothelial differentiation, butthe overall effects on downstream hematopoietic markerswere less pronounced. Although application of shear stressafter an initial 2-day static pre-differentiation periodincreased the genetic and protein expression of endothelialand hematopoietic markers, the effects on hematopoieticdifferentiation diminished when shear treatment was appliedafter 3 days. Supplementing the media with SU1498 duringtreatment abated the shear-mediated upregulation ofendothelial and hematopoietic markers, suggesting thatthis response requires activation of the FLK1 membraneprotein. Overall, this process characterization indicates thatshear stress is a robust method for promoting ESCdifferentiation and informs several operating parametersfor scale-up bioreactors.

We found that applied shear stress promotes ESCdifferentiation towards both endothelial and hematopoieticphenotypes in an adherent model system which, in theabsence of externally applied forces or chemical cues, allowsfor very little endothelial and hematopoietic differentiation(Supplemental Fig. 5). In our adherent model system, 4 daysof applied shear stress increased the number of mesodermalFLK1þ, endothelial PECAM1þ, and hematopoietic CD41þcells by 1.7�, 1.2�, and 1.9�, respectively, though the actualpercentage of positive cells still remained fairly low.To better understand the effects of applied mechanicalforces, the medium used in the presented studies was notsupplemented with growth factors that would favor anyparticular phenotype. Hence, it may be possible to optimizethe medium composition to further increase cell yieldor perhaps allow for more specific targeting of a singlephenotype. In suspension-cultured EBs, other groups foundthat medium supplementation with VEGF approximatelydoubled the number of T-GFPþFLKþ mesodermal cells(Purpura et al., 2008b), increased the number of PECAM1þendothelial cells by 60–100% (Purpura et al., 2008a; Vittetet al., 1996), and approximately doubled the number ofCD41þ cells when additional growth factors were used(Pearson et al., 2008). Although these initial comparisonsare across culture paradigms and include medium supple-mented with serum, our results indicate that applied shearstress can promote mesoderm, endothelial, and hematopoi-etic differentiation on a scale similar to those achieved usinggrowth factor-based methods. This raises the meaningfulpossibility of utilizing mechanical cues that can readily beincorporated into scalable bioreactors to substitute forexogenously added expensive or xenogeneic reagents inclinical production of cells.

Large-scale bioprocessing needs to consistently andhomogenously ensure a high quality cell product. Chemicalgradients that can lead to unsought biological responses can beovercome by use of microcarriers and stir-or mixing-basedfeatures. While studies in these systems have exploredexpansion of ESCs and induced pluripotent cells (Abrancheset al., 2007; Fernandes et al., 2007; Kehoe et al., 2010), theeffects of the inherently present shear stresses on

1 2 30.0

0.5

1.0

1.5

flk1

[mM

/Mgapdh] **

*

1 2 30.0

0.1

0.2

0.3

Pre-treatment Duration [days]

runx1

[mM

/M gapdh

]

*

A

B

1 2 30

2

4

tie2

[mM

/Mgapdh] **

* *

C

Figure 6. Delayed exposure dampens the shear-mediated increase in gene

expression. Cells were cultured under static conditions for 1–3 days before 1 day of

STATIC (&) or SHEAR (t¼ 15 dyne/cm2, &) treatment. Gene expression (normalized

to gapdh) was calculated for flk1 (A), tie2 (B), and runx1 (C). Data presented are

mean� SEM (n¼ 6–8), where significant differences between trial-matched STATIC

and SHEAR groups are indicated by asterisks (�P< 0.05, ��P< 0.01).

1238 Biotechnology and Bioengineering, Vol. 110, No. 4, April, 2013

differentiation have not yet been thoroughly investigated.Using our surrogate 2D adherent model system we foundthat the underlying protein substrate has little effect on theshear-mediated differentiation, suggesting that the micro-carriers can be coated with a variety of proteins not limitedto specific lab or manufacturing capabilities. Furthermore,for early endothelial and hematopoietic differentiation, wefound that the presence or absence of shear stress was moreimportant than its magnitude. Thus, early differentiationto these phenotypes can be achieved in large volumebioreactors in which shear stress varies in magnitude as afunction of the impeller size, speed, and location (Vallejoset al., 2011; Venkat et al., 1996). Interestingly, other studiesin sorted populations of differentiating ESCs have foundthat applied shear stress at much later times points ofdifferentiation have found that FLK1 (Yamamoto et al.,2005) and runx1 (Adamo et al., 2009) expression was highlydependent on the magnitude of shear stress. The acquisitionof magnitude-sensitivity suggests that the shear activatedmechanisms may increase in complexity with furtherdifferentiation.

The mechanisms involved in shear-mediated earlyESC differentiation are largely unknown. Studies interminally differentiated phenotypes have identified severalmechanosensors of shear stress, including FLK1 (Chen et al.,

1999), PECAM1 (Osawa et al., 2002), heparin sulfateproteoglycans (Florian et al., 2003), G-protein coupledreceptors (Makino et al., 2006), and integrins (Wang et al.,2002). But it is unclear if all of these mechanisms are presentin ESCs and at which points in differentiation they areacquired or lost. Our findings with the various proteinsubstrates, known to bind to distinct integrin heterodimers,imply that shear-mediated differentiation does not dependon activation of unique integrin subunits. Initial studies byothers have indicated that shear-mediated differentiation ofESCs may involve epigenetic changes (Illi et al., 2005) andheparan sulphate proteoglycans (Toh and Voldman, 2011).Other studies have suggested that shear stress can promoteendothelial differentiation by activation of FLK1 (Zeng et al.,2006) in a ligand-independent manner (Yamamoto et al.,2005). This is consistent with published results that FLK1activation by shear stress, as opposed to VEGF, employsdifferent signaling pathways in mature endothelial cells(Wang et al., 2007). Here, we used inhibition experiments toshow that FLK1 activation is necessary in the early stages ofshear-mediated hematopoietic and endothelial differentia-tion, but the specific downstream signaling pathways stillneed to be determined.

Studies have suggested that the RUNX1 transcriptionfactor can inhibit flk1 gene expression (Sakai et al., 2009) by

Pre-treatment Duration [days]

% P

ositi

ve C

ells

2 30

40

80

*****

A

Pre-treatment Duration [days]

% P

ositi

ve C

ells

2 30

20

40

60

*

**B

Pre-treatment Duration [days]

% P

ositi

ve C

ells

2 30

5

10

15

*

*

C

Pre-treatment Duration [days]

% P

ositi

ve C

ells

2 30

10

20

30

*

D

Figure 7. Delayed exposure to shear stress is less effective at promoting a hematopoietic phenotype. Cells were cultured under static conditions for 2 or 3 days before 4 days

of either STATIC (striped white bars) or SHEAR (t¼ 5.0 dyne/cm2, striped black bars) treatment. Fluorescence of both STATIC and SHEAR samples were assessed and compared to

flow cytometry control samples to calculate the percentage of cells positive for markers of differentiation. Assessed proteins were FLK1 (A), TIE2 (B), RUNX1 (C), and CD41 (D). Data

presented are mean� SEM (n¼ 3–4), where significant differences between STATIC and SHEAR groups were determined by a t-test and indicated by asterisks (�P< 0.05 and���P< 0.001).

Wolfe and Ahsan: Shear Stress and ESC Differentiation 1239

Biotechnology and Bioengineering

a DNA binding mechanism (Hirai et al., 2005), but themeans by which FLK1 activation may upregulate RUNX1are unclear. In our studies, shear-induced FLK1 upregula-tion was associated with upregulation of both TIE2 andRUNX1, except at later stages of differentiation when theFLK1 and TIE2 responses were sustained but the RUNX1upregulation was lost. While FLK1 and endothelialdifferentiation have been linked in a variety of stem,progenitor, and differentiated cells, our studies imply thatthe relationship between FLK1 and shear-induced hemato-poietic differentiation may change as differentiationprogresses. This is consistent with studies by others thatfound that only during the first 5 days of EB differentiationdid FLK1 activation by VEGF simultaneously promote bothhematopoietic and endothelial differentiation (Purpuraet al., 2008a). Thus, the role of FLK1 in later stages ofhematopoietic specification is still unclear.

The recent identification of the FLK1 positive progenitorto both endothelial and hematopoietic cells has implicationsnot only for the field of development but also stem cellbiology. While traditionally identified as an endothelial-specific marker, FLK1 protein is expressed very early onduring ESC differentiation (Vittet et al., 1996) and is widelyexpressed by mesoderm-derived progenitors of several

phenotypes (Bautch, 2006; Ema et al., 2006), includinghematopoietic cells (Shalaby et al., 1997). These new insightssuggest that bioprocessing techniques may not differ intargeting early ESC differentiation towards endothelialversus hematopoietic phenotypes. Other studies found thatFLK1 activation by VEGF promotes ESC differentiationtowards an endothelial phenotype (Nourse et al., 2010;Vittet et al., 1996), but when present with other cues it canpromote a hematopoietic phenotype (Chadwick et al., 2003;Pearson et al., 2008). More recent studies have suggestedthat activation of FLK1 by hypoxia-induced endogenousVEGF can promote both endothelial and hematopoieticphenotypes (Purpura et al., 2008a). Our presented studiesare the first to examine the simultaneous effects of anidentical mechanical force (in the absence of additionalspecific growth factors) on differentiation towards bothendothelial and hematopoietic phenotypes. It has yet to bedetermined, however, whether these conditions also favorother clinically needed mesoderm phenotypes.

We found that applied shear stress in an adherent modelsystem promotes early ESC differentiation towards bothendothelial and hematopoietic phenotypes in a robustsubstrate-independent manner. In addition, we found thatduring early endothelial and hematopoietic specification,responses to shear stress are magnitude insensitive, whichmay be beneficial in large-scale bioreactors that inherentlyhave heterogeneous shear stress profiles. As differentiationprogresses, however, stress effects may independentlymodulate endothelial versus hematopoietic maturation.An extension of these findings is that ESCs cultured on thesurface of microcarriers within mixing bioreactors is areasonable approach to large-scale production of endothe-lial and hematopoietic cells. This can be used to build onprevious results that pluripotent cells can be expanded onmicrocarriers in scale up systems (Abranches et al., 2007;Fernandes et al., 2007; Kehoe et al., 2010) and that ESC-seeded microcarriers can be cryopreserved (Nie et al., 2009).Together this insinuates that a microcarrier-based approachmay utilize both early stress magnitude-insensitive andlater stress magnitude-sensitive stages, along with intermit-tent storage steps, to develop a flexible modular scale-upproduction process to yield mesodermal cells frompluripotent stem cells for clinical use. Future efforts todevelop this type of paradigm will need to investigate theeffects of different shear stress profiles (e.g., oscillatory,pulsatile), as well as optimization of medium compositionand exchange rates when in combination with shear stress.

We would like to thank Dr. W.T. Godbey for the generous access to

his flow cytometer. The authors would also like to thank Tulane

University, the Louisiana Board of Regents (R.W.) and the NIH (#P20

GM103629) for supporting this work.

References

Abranches E, Bekman E, Henrique D, Cabral JM. 2007. Expansion of mouse

embryonic stem cells on microcarriers. Biotechnol Bioeng 96(6):1211–

1221.

0

1

2

3

tie2

[Nor

mal

ized

to S

tatic

]

0

1

2

3

runx1

[Nor

mal

ized

to S

tatic

]

-- -

+ +++

-Shear stress SU1498

***

A

B

***

Figure 8. FLK1 inhibition negates the effects of shear stress on early differenti-

ation. Cells were exposed to 2 days of STATIC (&) or SHEAR (t¼ 5.0 dyne/cm2, &)

treatment, where media from select STATIC and SHEAR samples was supplemented

with a FLK1 inhibitor (SU1498: 4.5mg/mL) during treatment. Gene expression was

assessed and each value was normalized to the average value of the trial-matched

STATIC controls. Genes assessed include tie2 (A) and runx1 (B). Data presented are

mean� SEM (n¼ 3–4), where significant differences between trial-matched STATIC

and SHEAR are indicated by asterisks (���P< 0.001).

1240 Biotechnology and Bioengineering, Vol. 110, No. 4, April, 2013

Adamo L, Naveiras O, Wenzel PL, McKinney-Freeman S, Mack PJ,

Gracia-Sancho J, Suchy-Dicey A, Yoshimoto M, Lensch MW,

Yoder MC, Garcıa-Cardena G, Daley GQ. 2009. Biomechanical forces

promote embryonic haematopoiesis. Nature 459(7250):1131–1135.

Ahsan T, Nerem RM. 2010. Fluid shear stress promotes an endothelial-like

phenotype during the early differentiation of embryonic stem cells.

Tissue Eng Part A 16(11):3547–3553.

Azarin SM, Palecek SP. 2010. Development of scalable culture systems for

human embryonic stem cells. Biochem Eng J 48(3):378.

Bautch VL. 2006. Flk1 expression: Promiscuity revealed. Blood 107(1):3–4.

Bilgen B, Sucosky P, Neitzel GP, Barabino GA. 2006. Flow characterization

of a wavy-walled bioreactor for cartilage tissue engineering. Biotechnol

Bioeng 95(6):1009–1022.

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

human embryonic stem cell-derived embryoid bodies by stirred vessel

cultivation. Biotechnol Bioeng 94(5):938–948.

Carpenedo RL, Sargent CY, McDevitt TC. 2007. Rotary suspension culture

enhances the efficiency, yield, and homogeneity of embryoid body

differentiation. Stem cells 25(9):2224–2234.

Chadwick K, Wang L, Li L, Menendez P, Murdoch B, Rouleau A, Bhatia M.

2003. Cytokines and BMP-4 promote hematopoietic differentiation of

human embryonic stem cells. Blood 102(3):906–915.

Chen KD, Li YS, Kim M, Li S, Yuan S, Chien S, Shyy JYJ. 1999. Mechan-

otransduction in response to shear stress—Roles of receptor tyrosine

kinases, integrins, and Shc. J Biol Chem 274(26):18393–18400.

Chisti Y. 2001. Hydrodynamic damage to animal cells. Crit Rev Biotechnol

21(2):67–110.

Eilken HM, Nishikawa SI, Schroeder T. 2009. Continuous single-cell

imaging of blood generation from haemogenic endothelium. Nature

457(7231):896–900.

Ema M, Takahashi S, Rossant J. 2006. Deletion of the selection cassette, but

not cis-acting elements, in targeted Flk1-lacZ allele reveals Flk1 expres-

sion in multipotent mesodermal progenitors. Blood 107(1):111–117.

Fernandes AM, Fernandes TG, DiogoMM, da Silva CL, Henrique D, Cabral

JM. 2007. Mouse embryonic stem cell expansion in a microcarrier-

based stirred culture system. J Biotechnol 132(2):227–236.

Florian JA, Kosky JR, Ainslie K, Pang ZY, Dull RO, Tarbell JM. 2003.

Heparan sulfate proteoglycan is a mechanosensor on endothelial cells.

Circ Res 93(10):E136–E142.

Fukuda S, Schmid-Schonbein GW. 2003. Regulation of CD18 expression on

neutrophils in response to fluid shear stress. Proc Natl Acad Sci U S A

100(23):13152–13157.

Gerecht-Nir S, Cohen S, Itskovitz-Eldor J. 2004. Bioreactor cultivation

enhances the efficiency of human embryoid body (hEB) formation and

differentiation. Biotechnol Bioeng 86(5):493–502.

Gerszten RE, Lim YC, DingHT, Snapp K, Kansas G, Dichek DA, Cabanas C,

Sanchez-Madrid F, Gimbrone MA, Jr., Rosenzweig A, et al. 1998.

Adhesion of monocytes to vascular cell adhesion molecule-1-

transduced human endothelial cells: Implications for atherogenesis.

Circ Res 82(8):871–878.

Gregoriades N, Clay J, Ma N, Koelling K, Chalmers JJ. 2000. Cell damage of

microcarrier cultures as a function of local energy dissipation created by

a rapid extensional flow. Biotechnol Bioeng 69(2):171–182.

Hirai H, Samokhvalov IM, Fujimoto T, Nishikawa S, Imanishi J. 2005.

Involvement of Runx1 in the down-regulation of fetal liver kinase-1

expression during transition of endothelial cells to hematopoietic cells.

Blood 106(6):1948–1955.

Hua J, Erickson LE, Yiin TY, Glasgow LA. 1993. A review of the effects of

shear and interfacial phenomena on cell viability. Crit Rev Biotechnol

13(4):305–328.

Illi B, Scopece A, Nanni S, Farsetti A, Morgante L, Biglioli P, Capogrossi

MC, Gaetano C. 2005. Epigenetic histone modification and cardiovas-

cular lineage programming in mouse embryonic stem cells exposed to

laminar shear stress. Circ Res 96(5):501–508.

Kehoe DE, Jing D, Lock LT, Tzanakakis ES. 2010. Scalable stirred-

suspension bioreactor culture of human pluripotent stem cells. Tissue

Eng Part A 16(2):405–421.

Kirouac DC, Zandstra PW. 2008. The systematic production of cells for cell

therapies. Cell Stem Cell 3(4):369–381.

Lancrin C, Sroczynska P, Stephenson C, Allen T, Kouskoff V, Lacaud G.

2009. The haemangioblast generates haematopoietic cells through a

haemogenic endothelium stage. Nature 457(7231):892–895.

Levesque MJ, Nerem RM. 1985. The elongation and orientation of cultured

endothelial cells in response to shear stress. J Biomech Eng 107(4):

341–347.

Makino A, Prossnitz ER, Bunemann M, Wang JM, Yao WJ, Schmid-

Schoenbein GW. 2006. G protein-coupled receptors serve as mechan-

osensors for fluid shear stress in neutrophils. Am J Physiol Cell Physiol

290(6):C1633–C1639.

Mollet M, Ma N, Zhao Y, Brodkey R, Taticek R, Chalmers JJ. 2004.

Bioprocess equipment: Characterization of energy dissipation rate

and its potential to damage cells. Biotechnol Prog 20(5):1437–1448.

Murray PDF. 1932. The development in vitro of the blood of the early chick

embryo. Proc R Soc B: Biol Sci 111(773):497–521.

Nie Y, Bergendahl V, Hei DJ, Jones JM, Palecek SP. 2009. Scalable culture

and cryopreservation of human embryonic stem cells on microcarriers.

Biotechnol Prog 25(1):20–31.

Niebruegge S, Bauwens CL, Peerani R, Thavandiran N,Masse S, Sevaptisidis

E, Nanthakumar K, Woodhouse K, Husain M, Kumacheva E. et al.

2009. Generation of human embryonic stem cell-derived mesoderm

and cardiac cells using size-specified aggregates in an oxygen-controlled

bioreactor. Biotechnol Bioeng 102(2):493–507.

Nikmanesh M, Shi ZD, Tarbell JM. 2012. Heparan sulfate proteoglycan

mediates shear stress-induced endothelial gene expression in mouse

embryonic stem cell-derived endothelial cells. Biotechnol Bioeng

109(2):583–594.

Nourse MB, Halpin DE, Scatena M, Mortisen DJ, Tulloch NL, Hauch KD,

Torok-Storb B, Ratner BD, Pabon L, Murry CE. 2010. VEGF induces

differentiation of functional endothelium from human embryonic stem

cells: Implications for tissue engineering. Arterioscler Thromb Vasc

Biol 30(1):80–89.

Obi S, Yamamoto K, Shimizu N, Kumagaya S, Masumura T, Sokabe T,

Asahara T, Ando J. 2009. Fluid shear stress induces arterial differentia-

tion of endothelial progenitor cells. J Appl Physiol 106(1):203–211.

Osawa M, Masuda M, Kusano K, Fujiwara K. 2002. Evidence for a role of

platelet endothelial cell adhesion molecule-1 in endothelial cell

mechanosignal transduction: Is it a mechanoresponsive molecule?

J Cell Biol 158(4):773–785.

Pearson S, Sroczynska P, Lacaud G, Kouskoff V. 2008. The stepwise

specification of embryonic stem cells to hematopoietic fate is driven

by sequential exposure to Bmp4, activin A, bFGF and VEGF Develop-

ment 135(8):1525–1535.

Purpura KA, George SH, Dang SM, Choi K, Nagy A, Zandstra PW. 2008a.

Soluble Flt-1 regulates Flk-1 activation to control hematopoietic and

endothelial development in an oxygen-responsive manner. Stem Cells

26(11):2832–2842.

Purpura KA, Morin J, Zandstra PW. 2008b. Analysis of the temporal and

concentration-dependent effects of BMP-4, VEGF, and TPO on devel-

opment of embryonic stem cell-derived mesoderm and blood progeni-

tors in a defined, serum-free media. Exp Hematol 36(9):1186–1198.

Sabin FR. 1920. Studies on the origin of blood-vessels and of red

blood-corpuscles as seen in the living blastoderm of chicks during

the second day of incubation. Washington, D.C.: Carnegie Institution

of Washington.

Sakai E, Kitajima K, Sato A, Nakano T. 2009. Increase of hematopoietic

progenitor and suppression of endothelial gene expression by Runx1

expression during in vitro ES differentiation. Exp Hematol 37(3):

334–345.

Sargent CY, Berguig GY, McDevitt TC. 2009. Cardiomyogenic differentia-

tion of embryoid bodies is promoted by rotary orbital suspension

culture. Tissue Eng Part A 15(2):331–342.

Shalaby F, Ho J, StanfordWL, Fischer KD, Schuh AC, Schwartz L, Bernstein

A, Rossant J. 1997. A requirement for Flk1 in primitive and definitive

hematopoiesis and vasculogenesis. Cell 89(6):981–990.

Wolfe and Ahsan: Shear Stress and ESC Differentiation 1241

Biotechnology and Bioengineering

Stolberg S, McCloskey KE. 2009. Can shear stress direct stem cell fate?

Biotechnol Prog 25(1):10–19.

Toh YC, Voldman J. 2011. Fluid shear stress primes mouse embryonic

stem cells for differentiation in a self-renewing environment via

heparan sulfate proteoglycans transduction. FASEB J 25(4):1208–

1217.

Vallejos JR, Brorson KA, Moreira AR, Rao G. 2011. Integrating a 250mL-

spinner flask with other stirred bench-scale cell culture devices: A mass

transfer perspective. Biotechnol Prog 27(3):803–810.

Venkat RV, Stock LR, Chalmers JJ. 1996. Study of hydrodynamics in

microcarrier culture spinner vessels: A particle tracking velocimetry

approach. Biotechnol Bioeng 49(4):456–466.

Vittet D, Prandini MH, Berthier R, Schweitzer A, MartinSisteron H, Uzan

G, Dejana E. 1996. Embryonic stem cells differentiate in vitro to

endothelial cells through successive maturation steps. Blood 88(9):

3424–3431.

Wang Y, Chang J, Chen KD, Li S, Li JY, Wu C, Chien S. 2007. Selective

adapter recruitment and differential signaling networks by VEGF vs.

shear stress. Proc Natl Acad Sci U S A 104(21):8875–8879.

Wang Y, Miao H, Li S, Chen KD, Li YS, Yuan S, Shyy JY, Chien S. 2002.

Interplay between integrins and FLK-1 in shear stress-induced signal-

ing. Am J Physiol Cell Physiol 283(5):C1540–C1547.

Wolfe RP, Leleux J, Nerem RM, Ahsan T. 2012. Effects of shear stress on

germ lineage specification of embryonic stem cells. Integr Biol 4(10):

1263–1273.

Yamamoto K, Sokabe T, Watabe T, Miyazono K, Yamashita JK, Obi S,

Ohura N, Matsushita A, Kamiya A, Ando J. 2005. Fluid shear stress

induces differentiation of Flk-1-positive embryonic stem cells into

vascular endothelial cells in vitro. Am J Physiol Heart Circ Physiol

288(4):H1915–1924.

Yamamoto K, Takahashi T, Asahara T, Ohura N, Sokabe T, Kamiya A, Ando

J. 2003. Proliferation, differentiation, and tube formation by endothe-

lial progenitor cells in response to shear stress. J Appl Physiol 95(5):

2081–2088.

Zeng L, Xiao Q, Margariti A, Zhang Z, Zampetaki A, Patel S, Capogrossi

MC, Hu Y, Xu Q. 2006. HDAC3 is crucial in shear- and VEGF-induced

stem cell differentiation toward endothelial cells. J Cell Biol 174(7):

1059–1069.

1242 Biotechnology and Bioengineering, Vol. 110, No. 4, April, 2013