Effect of Hydrostatic Pressure on Bone RegenerationUsing Human Mesenchymal Stem Cells

Chenyu Huang, M.D., Ph.D.,1,2 and Rei Ogawa, M.D., Ph.D., FACS1

Background: Mechanics is increasingly being recognized as the fourth essential factor in bone tissue engineeringnext to cell, scaffold, and growth factors. The development of bioprocessors has made it possible to simulate thein vivo mechanics that are needed to generate three-dimensional (3D) bone constructs. However, althoughhydrostatic pressure (HP) is a dominant and constant mechanical strain on bone cells in vivo, little is knownabout the effect of HP applied via perfusion bioprocessors on in vitro human bone marrow–derived mesen-chymal stem cell (hMSC) behavior.Methods: hMSCs underwent primary culture for three passages before being seeded into hydroxyapatite (HA)scaffolds. The scaffolds were incubated for 3 weeks in an automated bioprocessor under cyclic HP. Scaffoldsexposed to atmospheric pressure (AP) served as the comparator. Osteogenic differentiation medium was em-ployed for both the HP and AP groups. Immediately before and 1, 2, and 3 weeks after incubation, the scaffoldswere harvested for histological, immunohistochemical, and gene expression analyses.Results: Cells were only found in the AP scaffold surfaces, whereas in the HP group, they were distributedevenly throughout the scaffolds. Immunohistochemical analysis revealed that the HP group expressed higherlevels of osteocalcin (OC), osteopontin (OP), osteonectin (ON), and collagen type 1 (Col1) than the AP groupduring the 3-week process. Gene expression analysis revealed that the HP group expressed higher levels of ON,Col1, alkaline phosphatase, and integrin b5 than the AP group at the 1-, 2-, and 3-week timepoints. The HPgroup also expressed higher levels of core-binding factor a-1 (Cbfa1) at the 2- and 3-week timepoints and higherlevels of OP and OC at the 1-week timepoint. Their proliferating cell nuclear antigen levels were lower at the1- and 2-week timepoints.Conclusions: HP enhances cellular viability and improves osteogenic differentiation and maturation, althoughsomewhat at the expense of proliferation and self-renewal of MSCs. Possible negative effects of the bioprocessor-induced HP on bone regeneration were not observed. Further, the mechanotransductive molecule integrin b5was expressed at high levels after HP stimulation and may enhance migration, promote differentiation, andinhibit osteoclast maturation during HP-driven osteogenesis in vitro.

Introduction

Advances in the tissue engineering field have made itincreasingly possible to use natural bone regeneration

processes to fill bone defects that arise from trauma, tumorresection, infection, or congenital defects. The essential rolesplayed by osteogenic cells, osteoconductive scaffolds, andgrowth factors in such bone tissue engineering are now wellillustrated. However, it is increasingly being realized thatmechanics are a fourth important factor.1 This notion is at-tracting increasing interests from various fields, includingthe fields of functional and constitutive tissue engineering.

That mechanics participate in bone tissue engineering issupported by the well-known fact that is illustrated by

Wolff’s law and the mechanostat theory: the mass and ge-ometry of bone is physically remodeled in a dynamic fashionto function mechanically as needed by detecting and re-sponding to mechanical loads. This is known as bone ad-aptation.2 The recently developed mechanotransductiontheories have further promoted tissue engineering since theyexplain how physical forces are converted into biochemicalsignals and are then integrated into cellular responses.3 Theprogress of mechanotransduction in bone repair and regen-eration is described in detail by our review.4

These observations suggest that mechanics should beconsidered when engineering bone tissue. To facilitate this,mechanics-oriented functional tissue engineering in vitroshould be performed to elucidate further the effects of

1Department of Plastic, Reconstructive and Aesthetic Surgery, Nippon Medical School, Tokyo, Japan.2Department of Plastic Surgery, Meitan General Hospital, Beijing, China.

TISSUE ENGINEERING: Part AVolume 18, Numbers 19 and 20, 2012ª Mary Ann Liebert, Inc.DOI: 10.1089/ten.tea.2012.0064

2106

mechanical forces on bone regeneration. To this end,bioprocessors have been developed. These dynamic biopro-cessors, or bioreactors, particularly the perfusion biopro-cessors, have shown that mechanical stimuli promote theappropriate behaviors of cells in three-dimensional (3D)scaffolds as well as promoting oxygen and nutrient distri-bution.5 In fact, hydrostatic pressure (HP) is a dominant andconstant mechanical strain on bone cells in vivo, and HP inthe mammalian bone marrow is 10.7–120 mmHg, which isabout a quarter of the systemic blood pressure.6 However, todate, little is known about the effect of HP applied via per-fusion bioprocessors on in vitro human bone marrow–de-rived mesenchymal stem cell (hMSC) behavior.

To address this question, an HP/perfusion culture systemthat applies HP directly on the medium fluid phase wasdeveloped.7 This system is currently used for chondrogen-esis, where it promotes the 3D chondrogenic differentiationof adipose-derived stem cells, as described in our formerstudies.8 It also effectively enhances 3D osteogenesis whensarcoma cells are used.9

In the present study, our in vitro bioprocessor was used toapply HP to hMSCs. The cellular and molecular effects of HPon hMSC viability, proliferation, osteogenic differentiation,and maturation were analyzed. In addition, whether themechanotransductive molecule integrin b5 could participatein these HP-mediated effects was analyzed. The observationsof this study improve our understanding of the biologicaleffects of HP on hMSCs in vitro and may facilitate functionalbone engineering and promote the development of applica-tions that improve osteogenesis in vivo.

Materials and Methods

Cell culture

hMSCs were obtained from Lonza (Lonza Walkersville,Inc.; passage 2). The cells were thawed according to the

manufacturer’s instructions and plated at a density of 5000cells/cm2 in 58 cm2 tissue culture dishes (Cellstar) in 10 mLD-10 culture media composed of Dulbecco’s modified Ea-gle’s medium (DMEM)/F-12 Hams (Gibco) supplementedwith 10% fetal bovine serum (FBS; Gibco) and 1% antibiotic-antimyomic (Gibco). The cells were incubated at 37�C with5% CO2 and 90% humidity and expanded in the same me-dium until 80–90% confluence had been achieved (*5–7days in culture). The cultures were then passaged twice.Thus, P4 cells were used in this study.

Hydroxyapatite scaffolds

The porous hydroxyapatite (HA) scaffolds (Pentax) were5 mm in diameter, 2 mm in height, had 80% porosity, and apore size of 50–200 mm (Fig. 1). All 42 sterilized HA scaffoldsthat were used were prewetted in D-10 culture medium in ahumidified atmosphere of 5% CO2 and 90% humidity for 4 h.They were then placed in culture dishes in preparation forhMSC seeding.

FIG. 1. Photo of an HA scaffold. The porous HA scaffolds(Pentax) were 5 mm in diameter, 2 mm in height, had 80%porosity, and a pore size of 50–200mm. HA, hydroxyapatite.Color images available online at www.liebertpub.com/tea

FIG. 2. HP stimulation using a bioprocessor. By using aTEP-P02 bioprocessor (Takagi Industrial), cyclic HP wasplaced on human mesenchymal stem cells at 0–0.5 MPa(3750 mmHg, 4.93 atm), 0.5 Hz, with a medium replenishmentrate of 50 mL/min, while the cells were incubated at 37�C with5% CO2 and 20% O2 in air. HP, hydrostatic pressure. Colorimages available online at www.liebertpub.com/tea

EFFECT OF HYDROSTATIC PRESSURE ON BONE REGENERATION 2107

Cell seeding

The hMSCs were harvested by trypsinization and 15mL ofthe cell suspension (8.16 · 106 cells/mL) was dripped evenlyonto the top of the HA scaffold. After incubation at 37�C for20 min, the scaffold was turned over and another 15 mL of thecell suspension was dripped onto its top. Thus, 2.45 · 105

hMSCs were allowed to soak into and attach to each scaffoldover 6 h in the incubator. The scaffolds were then hydrated in11 mL of D-10 culture medium and incubated for 34 h at 37�Cwith 5% CO2 in air and 90% humidity before being used inthe experiments.

HP stimulation using a bioprocessor

The 42 HA scaffolds were divided into two groups. TheHP group scaffolds were placed in a TEP-P02 bioprocessor(Takagi Industrial) and incubated at 37�C, 5% CO2, and 20%O2 in air for 3 weeks with cyclic HP at 0.5 Hz, 0–0.5 MPa(3750 mmHg, 4.93 atm) in a sinusoidal waveform, and with amedium replenishment rate of 50 mL/min. The atmosphericpressure (AP) group scaffolds were incubated for 3 weeksin the same incubator without HP stimulation. The me-dium used for both groups was a differentiation mediumcomposed of DMEM/F-12 Hams, 10% FBS, 1% antibiotic-antimyomic, dexamethasone (2 · 10 - 7 mol/L), L-ascorbicacid 2-phosphate sesquimagnesium salt hydrate (1 · 10 - 4

mol/L), b-glycerol 2-phosphate disodium salt n-hydrate(2 · 10 - 2 mol/L), and 5% ITS� Premix (BD Biosciences)

(Figs. 2 and 3). The cells were fed by complete changes of thisosteogenic medium once per week under sterile conditions.Six scaffolds were harvested immediately before pressurewas applied (0 week) for histological, immunohistochemical,and gene expression analyses. After 1, 2, and 3 weeks, sixscaffolds were harvested per group and subjected to theseanalyses as well.

Histological and immunohistochemical stains

The scaffolds (n = 3) were fixed in 4% paraformaldehydephosphate (WAKO) at 4�C overnight and then decalcifiedwith Kalkitox (WAKO) at 4�C for 10 h. The latter procedurewas stopped by suspending the samples in 5% sodium sul-fate solution (WAKO) overnight. The decalcified sampleswere dehydrated with 100% ethanol, embedded in paraffin,cut into 4-mm sections, and stained with hematoxylin andeosin. Immunohistological staining was performed by usingthe Vectastatin ABC kit (VECTASTAIN� Universal). Briefly,the deparaffinized and rehydrated sections were treated withcitrate acid (0.01 M, pH 6.0) for 5 min in a microwave tounmask the antigens, followed by incubation with 0.3%H2O2 for 10 min to quench the endogenous peroxidase ac-tivity. After rinsing with phosphate-buffered saline (PBS),the sections were blocked with normal horse serum at roomtemperature for 20 min in a humidified chamber. To stainosteocalcin (OC), osteopontin (OP), osteonectin (ON), orcollagen type 1 (Col1), the sections were incubated for 60 minat room temperature with rabbit antibodies against OC

FIG. 3. Hydrostaticpressure (HP)/perfusionculture system in abioprocessor. A medium bag,a perfusion pump, apressure-proof culturechamber, and a back pressurevalve are connected withunions and silicon tubing andinstalled in an incubator. Themedium is perfused with HP.The culture chamber isisolated with a thin film froma pressure chamber. Thepressure chamber is filledwith water, which iscompressed with a pistondriven with an actuator. HPis regulated at a setmagnitude with a backpressure valve controlledwith an actuator. Colorimages available online atwww.liebertpub.com/tea

2108 HUANG AND OGAWA

(Novus), OP (Acris), ON (Acris), and Col1 (Thermo) that hadbeen diluted 1:250, 1:1000, 1:1000, and 1:500, respectively,with antibody diluent (BD Pharmingen). After three rinseswith PBS, the sections were incubated with biotinylated goatanti-rabbit immunoglobulin G antibody (Vector Labora-tories) according to the ABC kit manufacturer’s instructions.After three rinses with PBS, the color was developed withNova Red (Vector NovaRED). Nuclei were counterstainedwith Mayer’s blue. Histological sections were examined witha light microscope (Olympus AX-80).

Real-time reverse transcriptase–polymerasechain reaction

Samples were homogenized using a handheld homogenizerand a QIA shredder with buffer RLT that contained b-mer-captoethanol. Total RNA was then extracted from the freshlycollected scaffolds (n = 3) by using RNeasy Mini Kits (Qiagen)according to the manufacturer’s instructions and quantified byusing the NanoDrop (NanoDrop Technologies) method.

Complementary deoxyribonucleic acid (cDNA) was syn-thesized by reverse transcriptase–polymerase chain reaction(RT-PCR) by using a SuperScript III First-Stranded SynthesisSystem (Invitrogen) and the Applied Biosystem Gene Amp�

PCR System (9700). Total RNA ( < 1mg) was mixed with ran-dom hexamers (50 ng/mL) and deoxyribonucleoside triphos-phate (10 mM) and incubated at 65�C for 5 min. The tubeswere cooled on ice, after which RT buffer (10 · ), magnesiumchloride (25 mM), dithiothreitol (0.1 M), RNaseOUT (40 U/mL), and Superscript III RT (200 U/mL) were added, yielding afinal volume of 21mL. The samples were then incubated at25�C for 10 min, 50�C for 50 min, and 85�C for 5 min, andcooled on ice. Thereafter, Escherichia coli RNase H (1.5mL) wasadded and the mixtures were incubated at 37�C for 20 min.

Quantitative real-time RT-PCR was performed in the ABIPrism 7500 System (Applied Biosystems) by using theRT2 SYBR Green/ROX PCR master mix (SA Biosciences)and primers for the genes for OC (also known as gamma-carboxyglutamate [gla] protein [BGLAP]), OP (also known assecreted phosphoprotein 1 [SPP1]), ON, alkaline phosphatase,liver/bone/kidney (ALPL), core-binding factor a-1 subunit(Cbfa-1, also known as runt-related transcription factor 2[RUNX2]), proliferating cell nuclear antigen (PCNA), collagentype Ia1 (COL1A1), and integrin b5 (ITGB5) (Table 1). ThecDNA samples were amplified in triplicate in 96-well platesin a final volume of 20mL for 40 PCR cycles that consistedof a denaturation step at 95�C for 15 s and an annealing/extension step at 60�C for 1 min. Fluorescence measurementswere taken by using the system software program v2.0.4(Applied Biosystems) and these were used to generate adissociation curve. Signal levels were normalized to the ex-pression of a constitutively expressed gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), as shown as a relativeratio. Relative expression was calculated using the 2 -DDCt

method with a correction for different amplification efficien-cies. The reference samples were 0-week samples.

Results

Histological and immunohistochemical analyses

As shown in Figure 4, histological analysis of the 0-weeksamples revealed that the cells were distributed evenly

Ta

bl

e1.

Re

al

-T

im

eR

ev

er

se

Tr

an

sc

rip

ta

se

–P

ol

ym

er

ase

Ch

ain

Re

ac

tio

nP

rim

er

s

Gen

eG

enba

nk

cod

eF

orw

ard

Rev

erse

Gam

ma-

carb

ox

yg

luta

mat

e(g

la)

pro

tein

/O

steo

calc

in(B

GL

AP

)N

M_1

9917

3.4

5¢-G

GC

AG

CG

AG

GT

AG

TG

AA

GA

-3¢

5¢-C

CT

GA

AA

GC

CG

AT

GT

GG

T-3

¢

Sec

rete

dp

ho

sph

op

rote

in1/

Ost

eop

on

tin

(SP

P1)

NM

_000

582.

25¢

-AC

CT

GA

AC

GC

GC

CT

TC

TG

-3¢

5¢-C

AT

CC

AG

CT

GA

CT

CG

TT

TC

AT

AA

-3¢

Ost

eon

ecti

n(O

ST

EO

NE

CT

IN)

NM

_003

118.

25¢

-TG

CC

AC

TG

AG

GG

TT

CC

CA

-3¢

5¢-T

CG

GT

TT

CC

TC

TG

CA

CC

AT

C-3

¢A

lkal

ine

ph

osp

hat

ase

(AL

PL

)N

M_0

0047

8.4

5¢-A

GC

CC

TT

CA

CT

GC

CA

TC

CT

GT

-3¢

5¢-A

TT

CT

CT

CG

TT

CA

CC

GC

CC

AC

-3¢

Ru

nt-

rela

ted

tran

scri

pti

on

fact

or

2(R

UN

X2

)N

M_0

0434

8.3

5¢-G

CA

GC

AC

GC

TA

TT

AA

AT

CC

AA

AT

T-3

¢5¢

-AC

AG

AT

TC

AT

CC

AT

TC

TG

CC

AC

TA

G-3

¢P

roli

fera

tin

gce

lln

ucl

ear

anti

gen

(PC

NA

)N

M_0

0259

2.2

5¢-C

AG

GG

CT

CC

AT

CC

TC

AA

GA

A-3

¢5¢

-TC

TT

CA

TT

GC

CG

GC

GC

AT

T-3

¢C

oll

agen

,ty

pe

I,a1

(CO

L1

A1

)N

M_0

0008

8.3

5¢-C

CT

GC

GT

GT

AC

CC

CA

CT

CA

-3¢

5¢-A

CC

AG

AC

AT

GC

CT

CT

TG

TC

CT

T-3

¢In

teg

rin

,b5

(IT

GB

5)

NM

_002

213.

35¢

-CG

AG

CT

TG

GG

AT

AA

AG

CA

AG

-3¢

5¢-T

CA

AC

AG

GC

AT

CT

CA

AC

AG

C-3

¢G

lyce

rald

ehy

de-

3-p

ho

sph

ate

deh

yd

rog

enas

e(G

AP

DH

)N

M_0

0204

65¢

-GG

AT

TT

GG

TC

GT

AT

TG

GG

-3¢

5¢-G

GA

AG

AT

GG

TG

AT

GG

GA

TT

-3¢

EFFECT OF HYDROSTATIC PRESSURE ON BONE REGENERATION 2109

throughout the scaffold, which indicates satisfactory seeding.At 1-, 2-, and 3-week timepoints, the cells in the AP groupwere found on the surfaces of the scaffold. In contrast, thecells in the HP group were evenly distributed throughout thescaffolds. Neither group showed a tendency for cell numbersto increase gradually over time, indicating that little prolif-eration had occurred.

Immunohistochemical analysis revealed that the osteoidmatrix ingredients OC, OP, ON, and Col1 were expressed athigher levels in the HP group than in the AP group over the3-week process (Fig. 4).

Gene expression analysis

As shown in Figure 5, ON, COL1A1, ALPL, and ITGB5were significantly upregulated (95% confidence) at 1-, 2-, and3-week timepoints in the HP group as compared with theirexpression in the AP group. Compared with the AP group,the HP group expressed higher levels of Cbfa1 at 2- and3-week timepoints, higher levels of OP at 1-week timepoint,and higher levels of OC at 1- and 3-week timepoints. Com-pared with the AP group, the PCNA expression of the HPgroup was significantly (95% confidence) downregulated byHP stimulation at 1- and 2-week timepoints.

In terms of expression of these genes relative to the 0-weekbaseline, ON, ALPL, COL1, Cbfa1, OC, and integrin weresignificantly upregulated at 1-, 2-, and 3-week timepointsin the HP group; OP was upregulated at 1-week timepointand downregulated at 3 week timepoint; and PCNA wasdownregulated at 1-, 2-, and 3-week timepoints. Interest-ingly, the expression patterns along time axis in ON, ALPL,and OC were gradual climbing. Integrin and COL1 experi-enced climbing and sloping with the peak expression at2-week timepoint. OP switched on early and then graduallystopped being produced. And PCNA kept stable low ex-pression. Those trends showed to some extent the stagedexpression of different markers in osteogenesis. But no dose–

effect relationship between HP stimulation and osteogenicdifferentiation at mRNA level along time can be found. InAP group, in response to the osteogenic medium, ALPL ex-pression was higher at 1- and 3-week timepoints than 0-weekbaseline. OC and PCNA expression was higher at 2-weektimepoint and lower at 3-week timepoint. OP expression waslower at 1-week timepoint. Integrin expression was lower at1-, 2-, and 3-week timepoints. Expression of ON, COL1, andCbfa1 at 1-, 2-, and 3-week timepoints was similar to 0-weekbaseline.

Discussion

Effect of HP on bone regeneration

The present study showed that, at both the cellular andmolecular levels, HP markedly improves the viability ofhMSCs along with their osteogenic differentiation and mat-uration. First, HP improved the scaffold-wide distributionand viability of the cells in the HA scaffolds. Histologicalanalysis of the 0-week samples confirmed that seeding wassatisfactory. Over the following 3 weeks, only the HP groupcontinued to show even and scaffold-wide distribution of thecells; in contrast, the AP group cells were mainly located inthe superficial parts of the scaffolds. It has been shown thatpassive oxygen diffusion can only support cells in the out-ermost 150–200-mm layer of scaffolds.10 The difference be-tween the HP and AP groups with respect to viability showsclearly that HP promotes the diffusion of oxygen and nu-trients, as well as the removal of waste products in thepartitioned cell construct.

Second, HP promoted the differentiation of the hMSCstoward osteogenesis. This gradual cellular maturation is in-dicated by the increased expression of the osteogenic mark-ers OP, ON, and ALPL and the osteogenic transcriptionfactor Cbfa1 in the HP group relative to the AP group.Further supporting this is the increased expression of the

FIG. 4. Histological and immunohistochemical analyses of the scaffolds. Immunohistochemical analysis revealed that theosteoid matrix ingredients osteocalcin, osteopontin, osteonectin, and collagen type 1 accumulated at higher levels in the HP-treated group scaffolds than in the AP control group scaffolds during the 3 weeks of mechanical stimulation process. The cellswere distributed evenly throughout the HP scaffolds at the 1-, 2-, and 3-week timepoints but only occurred in the superficiallayer of the AP scaffolds at the same timepoints. AP, atmospheric pressure. Color images available online at www.liebertpub.com/tea

2110 HUANG AND OGAWA

main extracellular matrix (ECM) ingredient Col1. However,this osteogenic differentiation at the mRNA level in the HPgroup did not increase over time. The osteogenic effects ofHP on hMSCs were achieved in direct and indirect ways. Onthe one hand, the diffusion of oxygen and nutrients and theremoval of waste products are more efficient in the parti-tioned cell construct under cyclic HP stimulation, leading toenhanced cellular viability that indirectly improves the dif-ferentiation of the viable cells. On the other hand, the im-proved osteogenic differentiation can be the direct role ofHP. As seen in other studies, mechanical stimulation (e.g.,matrix elasticity) directed mesenchymal stem cell lineagespecification toward osteoblasts, myoblasts, or neurons, un-der identical serum conditions.11 The potential role playedby the mechanotransductive protein integrin will be dis-cussed later. The relative contributions of the direct and in-direct effects of HP on enhanced osteogenic differentiationwill be investigated in a future study.

HP did not have obvious mitogenic effects. The PCNAmRNA levels were lower in the HP group than in the APgroup during the first 2 weeks, and were similar to thePCNA mRNA levels in the AP group on the third week.Therefore, the larger cell densities in the HP group, as shownby histological staining, were due to improved cell viabilityrather than increased proliferation. There are two possibleexplanations for this: (1) HP directly regulates the transitionfrom proliferation to osteogenic differentiation. In otherwords, the differentiation promoted by HP was at the ex-pense of proliferation and self-renewal. Previous studies of

the response of bone cell proliferation to fluid flow stresshave yielded vastly conflicting results, with fluid flow stressbeing shown to enhance,12,13 reduce,14,15 or not alter16,17

cellular proliferation. The differences between these studiesin terms of mechanical stimuli and the microenvironmentsmake it impossible to draw any conclusions about the rela-tionship between fluid flow stress and proliferation. (2) Thegradual maturation of cells from immature hMSCs to matureosteoblasts was not only associated with the acquisition of amore differentiated phenotype but also with a reducedproliferative capacity. In other words, the proliferative effectwas gradually ‘‘diluted’’ by the gradual increase in maturecells. This possibility is supported by other studies compar-ing the mechanoresponsiveness, proliferation, and differen-tiation capabilities of osteoblast precursor cells, osteoblasts,and mature osteocytes.18–20

The possibility that HP might have unwanted effects wasalso considered. However, since the bioprocessor onlyplaced HP on the HA and the stem cells inside the scaffold,commonly seen side effects of shear stress due to the fluidflow (such as cell loss or cell lysis)21,22 were beyond con-sideration. This is supported by our observations of the HPand AP scaffolds.

Importance of integrin b5 in bone regeneration

Integrins are important mechanotransduction moleculesthat connect the extracellular matrix to the intracellular cy-toskeleton. To our knowledge, we are the first to show that

FIG. 5. Real-time RT-PCR analysis. HP-treated and AP control group scaffolds were harvested before and 1, 2, and 3 weeksafter initiating HP or AP (n = 3 per timepoint for each group) and analyzed by real-time RT-PCR. The expression of the osteogenicmarker genes osteonectin, osteopontin, osteocalcin, and ALPL was higher in the HP group than in the AP group. The HP groupalso expressed higher levels of the osteogenic transcription factor Cbfa1 (also known as runt-related transcription factor 2[RUNX2]) and the mechanosignaling molecule ITGB5. However, the HP and AP groups did not differ significantly during the 3weeks of culture in terms of PCNA and COL1A1 expression. *Significantly up- or downregulated with 95% confidence in the HPgroup compared with the AP group in terms of relative quantification. j Significantly up- or downregulated with 95% confi-dence in the 1-, 2-, or 3-week timepoint compared with the 0-week timepoint in terms of relative quantification. RT-PCR, reversetranscriptase–polymerase chain reaction; ALPL, alkaline phosphatase, liver/bone/kidney; Cbfa1, core-binding factor a-1 subunit;ITGB5, integrin b5; PCNA, proliferating cell nuclear antigen; COL1A1, collagen 1a1; BGLAP, gamma-carboxyglutamate (gla)protein; SPP1, secreted phosphoprotein 1; RQ, relative quantification. Gray bar, AP group; black bar, HP group.

EFFECT OF HYDROSTATIC PRESSURE ON BONE REGENERATION 2111

the b5 subunit of integrin is expressed in response to me-chanical stimulation. Since it is known that integrins cantransduce mechanical signals, our observations suggest thatintegrin b5 plays a mechanotransductive role in bone re-generation. While several studies on mechanotransduction inbone cells have focused specifically on the avb3, and b1subunits of integrin,23,24 none have examined the role thatthe b5 subunit might play. Integrin b5 is known to promotecellular migration,25 stimulate osteoblast differentiation,26

involve in bone formation,27 and inhibit osteoclast matura-tion28 in bone pathophysiology, but none of these studieshave related such abilities to the mechanoresponsiveness ofcells that promotes bone regeneration, let alone examined theeffect of mechanical stimuli on hMSCs. Only one study onteeth has suggested that integrin b5 may serve as a me-chanotransducer and matricellular regulator and that it in-teracts with HA to adapt cells to changes in mechanicalstresses. It is also noteworthy that integrin b5 is specificallylocalized to the periodontal ligament.29

Supporting the notion that integrin b5 participates in HP-driven osteogenesis is that, in the present study, integrin b5mRNA expression in hMSCs that were cyclically stimulatedby HP correlated well with the expression of ON, Col1, andCbfa1. In fact, b5 integrin is a potential ON receptor and acandidate target for direct ON activity, because RNA inter-ference knockdown of integrin b5 expression can reduce cellmigration in vitro.25 Notably, integrin b5 is expressed inimmature osteoclast precursors,30 and osteoclast formationand maturation are substantially enhanced and acceleratedin mice that lack the integrin b5 subunit.28 This indicates thatthis subunit could inhibit osteoclast formation and subse-quent bone resorption. Thus, the improved bone formationafter HP stimulation that was observed in the HP group mayhave been due to both enhanced osteogenic differentiationand decreased osteoclast maturation that was caused by theincreased expression of integrin b5.

Future prospects of the use of mechanical forcesto promote bone regeneration

The present study showed that HP with a sinusoidalprofile between 0 and 0.5 MPa, 0.5 Hz, promoted the osteo-genic differentiation of hMSCs. These mechanical parameterswere based on those used for 3D tissue culture, which wasstable at a constant HP of up to 5 MPa and at a cyclic HP ofup to 0–5 MPa at 0.5–0.03 Hz.31 More importantly, formerwork in our laboratory has shown that HP can inducechondrogenesis of human-adipose-derived stem cells.8 Otherin vitro HP profiles and algorithms, such as a constant HPand different cyclic settings, should be explored to optimizethe quality of the cell construct. Other influencing factors,such as the geometrical and structural features of the scaf-fold,32 should also be considered and optimized.

The present study also showed for the first time that in-tegrin b5 may play a significant mechanotransductive role inour model. Further studies to confirm this and elucidate themechanosignaling pathway(s) at RNA and protein levels arewarranted. This could be achieved by using specific siRNAsor antibodies that attenuate or block integrin b5 expression.The ability of this pathway to crosstalk with other pathways,such as the Rho-ROCK and Wnt 10b pathways, should alsobe examined.

Conclusions

In summary, the present study indicates that HP loadingwithin an HP/medium perfusion culture system enhancescellular viability and improves osteogenic differentiationand maturation, although somewhat at the expense of theproliferation and self-renewal of the hMSCs. Possible neg-ative effects of this system on bone regeneration were notobserved. Further, the mechanotransductive molecule in-tegrin b5 was found to be highly expressed after HP stim-ulation, which indicates that it may enhance migration,promote differentiation, and inhibit osteoclast maturationduring HP-driven osteogenesis in vitro. A better under-standing of the biological effects of HP on hMSCs in vitrowill improve functional bone engineering and could lead tothe development of applications that promote osteogenesisin vivo.

Author Contributions

Chenyu Huang: conception and design, collection and/orassembly of data, data analysis and interpretation, andmanuscript writing. Rei Ogawa: conception and design,provision of study materials, collection and/or assembly ofdata, data analysis and interpretation, and article writing.

Disclosure Statement

No competing financial interests exist.

References

1. Butler, D.L., Goldstein, S.A., and Guilak, F. Fuctional tissueengineering: the role of biomechanics. J Biomech Eng 122,

570, 2000.2. Frost, H.M. Bone’s mechanostat: a 2003 update. Anat Rec A

Discov Mol Cell Evol Biol 275, 1081, 2003.3. Huang, H., Kamm, R.D., and Lee, R.T. Cell mechanics and

mechanotransduction: pathways, probes, and physiology.Am J Physiol Cell Physiol 287, C1, 2004.

4. Huang, C., and Ogawa, R. Mechanotransduction in bonerepair and regeneration. FASEB J 24, 3625, 2010.

5. Brindley, D., Moorthy, K., Lee, J.H., Mason, C., Kim, H.W.,and Wall, I. Bioprocess forces and their impact on cell be-havior: implications for bone regeneration therapy. J TissueEng 2011, 620247, 2011.

6. Gurkan, U.A., and Akkus, O. The mechanical environ-ment of bone marrow: a review. Ann Biomed Eng 36, 1978,2008.

7. Mizuno, S., Ushida, T., Tateishi, J., and Glowacki, J. Effectsof physical stimulation on chondrogenesis in vitro. Mater SciEng C6, 301, 1998.

8. Ogawa, R., Mizuno, S., Murphy, G.F., and Orgill, D.P. Theeffect of hydrostatic pressure on three-dimensional chon-droinduction of human adipose-derived stem cells. TissueEng Part A 15, 2937, 2009.

9. Mueller, S.M., Mizuno, S., Gerstenfeld, L.C., and Glowacki,J. Medium perfusion enhances osteogenesis by murineosteosarcoma cells in three-dimensional collagen sponges.J Bone Miner Res 14, 2118, 1999.

10. Folkman, J., and Hochberg, M. Self-regulation of growth inthree dimensions. J Exp Med 138, 745, 1973.

11. Engler, A.J., Sen, S., Sweeney, H.L., and Discher, D.E. Matrixelasticity directs stem cell lineage specification. Cell 126,

677, 2006.

2112 HUANG AND OGAWA

12. Li, Y.J., Batra, N.N., You, L., Meier, S.C., Coe, I.A., Yellow-ley, C.E., and Jacobs, C.R. Oscillatory fluid flow affectshuman marrow stromal cell proliferation and differentia-tion. J Orthop Res 22, 1283, 2004.

13. Riddle, R.C., Taylor, A.F., Genetos, D.C., and Donahue, H.J.MAP kinase and calcium signaling mediate fluid flow-inducedhuman mesenchymal stem cell proliferation. Am J Physiol CellPhysiol 290, C776, 2006.

14. Scaglione, S., Wendt, D., Miggino S., Papadimitropoulos, A.,Fato, M., Quarto, R., and Martin, I. Effects of fluid flow andcalcium phosphate coating on human bone marrow stromalcells cultured in a defined 2D model system. J Biomed MaterRes A 86, 411, 2008.

15. Kreke, M.R., Huckle, W.R., and Goldstein, A.S. Fluid flowstimulates expression of osteopontin and bone sialoproteinby bone marrow stromal cells in a temporally dependentmanner. Bone 36, 1047, 2005.

16. Kreke, M.R., and Goldstein, A.S. Hydrodynamic shearstimulates osteocalcin expression but not proliferation ofbone marrow stromal cells. Tissue Eng 10, 780, 2004.

17. Kang, K.S., Lee, S.J., Lee, H.S., Moon, W., and Cho, D.W.Effects of combined mechanical stimulation on the prolifer-ation and differentiation of pre-osteoblasts. Exp Mol Med 43,

367, 2011.18. van der Plas, A., and Nijweide, P.J. Isolation and purification

of osteocytes. J Bone Miner Res 7, 389, 1992.19. van der Plas, A., Aarden, E.M., Feijen, J.H., de Boer, A.H.,

Wiltink, A., Alblas, M.J., de Leij, L., and Nijweide, P.J.Characteristics and properties of osteocytes in culture.J Bone Miner Res 9, 1697, 1994.

20. Aarden, E.M., Wassenaar, A.M., Alblas, M.J., and Nijweide,P.J. Immunocytochemical demonstration of extracellularmatrix proteins in isolated osteocytes. Histochem Cell Biol106, 495, 1996.

21. Mardikar, S.H., and Niranjan, K. Observations on the sheardamage to different animal cells in a concentric cylinderviscometer. Biotech Bioeng 68, 697, 2000.

22. Agashi, K., Chau, D.Y., and Shakesheff, K.M. The effect ofdelivery via narrow-bore needles on mesenchymal cells.Regen Med 4, 49, 2009.

23. Lee, D.Y., Li, Y.S., Chang, S.F., Zhou, J., Ho, H.M., Chiu, J.J.,and Chien, S. Oscillatory flow-induced proliferation of os-teoblast-like cells is mediated by alphavbeta3 and beta1-integrins through synergistic interactions of focal adhesionkinase and Shc with phosphatidylinositol3-kinase and theAkt/mTOR/p70S6K pathway. J Biol Chem 285, 30, 2010.

24. Weyts, F.A., Li, Y.S., van Leeuwen, J., Weinans, H., andChien, S. ERK activation and alpha v beta 3 integrin sig-naling through Shc recruitment in response to mechanicalstimulation in human osteoblasts. J Cell Biochem 87, 85,2002.

25. Sangaletti, S., Di Carlo, E., Gariboldi, S., Miotti, S., Cappetti,B., Parenza, M., Rumio, C., Brekken, R.A., Chiodoni, C., and

Colombo, M.P. Macrophage-derived SPARC bridges tumorcell-extracellular matrix interactions toward metastasis.Cancer Res 68, 9050, 2008.

26. Lai, C.F., Bai, S., Uthgenannt, B.A., Halstead, L.R.,McLoughlin, P., Schafer, B.W., Chu, P.H., Chen, J., Otey,C.A., Cao, X., and Cheng, S.L. Four and half lim protein 2(FHL2) stimulates osteoblast differentiation. J Bone MinerRes 21, 17, 2006.

27. Le Gat, L., Bonnel, S., Gogat, K., Brizard, M., Van DenBerghe, L., Kobetz, A., Gadin, S., Dureau, P., Dufier, J.L.,Abitbol, M., and Menasche, M. Prominent beta-5 gene ex-pression in the cardiovascular system and in the cartilagi-nous primordiae of the skeleton during mouse development.Cell Commun Adhes 8, 99, 2001.

28. Lane, N.E., Yao, W., Nakamura, M.C., Humphrey, M.B.,Kimmel, D., Huang, X., Sheppard, D., Ross, F.P., and Tei-telbaum, S.L. Mice lacking the integrin beta5 subunit haveaccelerated osteoclast maturation and increased activity inthe estrogen-deficient state. J Bone Miner Res 20, 58, 2005.

29. Luan, X., Ito, Y., Holliday, S., Walker, C., Daniel, J., Galang,T.M., Fukui, T., Yamane, A., Begole, E., Evans, C., andDiekwisch, T.G. Extracellular matrix-mediated tissue re-modeling following axial movement of teeth. J HistochemCytochem 55, 127, 2007.

30. McHugh, K.P., Hodivala-Dilke, K., Zheng, M.H., Namba,N., Lam, J., Novack, D., Feng, X., Ross, F.P., Hynes, R.O.,and Teitelbaum, S.L. Mice lacking beta3 integrins are os-teosclerotic because of dysfunctional osteoclasts. J Clin In-vest 105, 433, 2000.

31. Watanabe, S., Inagaki, S., Kinouchi, I., Takai, H., Masuda, Y.,and Mizuno, S. Hydrostatic pressure/perfusion culturesystem designed and validated for engineering tissue.J Biosci Bioeng 100, 105, 2005.

32. Eggli, P.S., Muller, W., Schenk, R.K. Porous hydroxyapatiteand tricalcium phosphate cylinders with two different poresize ranges implanted in the cancellous bone of rabbits. Acomparative histomorphometric and histologic study ofbony ingrowth and implant substitution. Clin Orthop RelatRes 232, 127, 1988.

Address correspondence to:Rei Ogawa, M.D., Ph.D., FACS

Department of Plastic, Reconstructive and Aesthetic SurgeryNippon Medical School

1-1-5 Sendagi Bunkyo-kuTokyo 113-8603

Japan

E-mail: r.ogawa@nms.ac.jp

Received: February 2, 2012Accepted: May 15, 2012

Online Publication Date: August 3, 2012

EFFECT OF HYDROSTATIC PRESSURE ON BONE REGENERATION 2113