Comparison of hydrogels in the in vivo formation of tissue-engineered bone using mesenchymal stem cells and beta-tricalcium phosphate.

Comparison of Hydrogels in the In Vivo Formation
of Tissue-Engineered Bone Using Mesenchymal Stem Cells
and Beta-Tricalcium Phosphate
CHRISTIAN WEINAND, M.D., Ph.D.,1 RAJIV GUPTA, Ph.D., M.D.,2
ALBERT Y. HUANG, B.S.,1 ELI WEINBERG, B.S.,3 IJAD MADISCH, M.D.,2
RAMEEZ A. QUDSI, B.A.,1 CRAIG M. NEVILLE, Ph.D.,1
IRINA POMERANTSEVA, M.D.,1 and JOSEPH P. VACANTI, M.D.1
ABSTRACT
Availability of grafts and morbidity at the donor site limit autologous transplantation in patients requiring
bone reconstruction. A tissue-engineering approach can overcome these limitations by producing bone-
like tissue of custom shape and size from isolated cells. Several hydrogels facilitate osteogenesis on porous
scaffolds; however, the relative suitability of various hydrogels has not been rigorously assessed. Fibrin
glue, alginate, and collagen I hydrogels were mixed with swine bone marrow–derived differentiated
mesenchymal stem cells (MSCs), applied to 3-dimensionally printed porous beta-tricalcium phosphate
(b-TCP) scaffolds and implanted subcutaneously in nude mice. Although noninvasive assessment of osteo-
genesis in 3 dimensions is desirable for monitoring new bone formation in vivo, correlations with traditional
histological and mechanical testing need to be established. High-resolution volumetric computed to-
mography (VCT) scanning, histological examination, biomechanical compression testing, and osteonectin
(ON) expression were performed on excised scaffolds after 1, 2, 4, and 6 weeks of subcutaneous im-
plantation in mice. Statistical correlation analyses were performed between radiological density, stiffness,
and ON expression. Use of collagen I as a hydrogel carrier produced superior bone formation at 6 weeks,
as demonstrated using VCT scanning with densities similar to native bone and the highest compression
values. Continued contribution of the seeded MSCs was demonstrated using swine-specific messenger
ribonucleic acid probes. Radiological density values correlated closely with the results of histological
and biomechanical testing and ON expression. High-resolution VCT is a promising method for monitor-
ing osteogenesis.
INTRODUCTION
BONE REGENERATION is critical in the healing of fracturesand bone defects caused by tumor resection, birth de-
fects, or serious injury. Insufficient repair results in non-
unions, pseudarthrosis or both and are normally treated with
bone grafts. Bone is the second most commonly transplanted
tissue, with more than 650,000 procedures occurring annu-
ally in the United States alone, at a cost exceeding $3 bil-
lion.1 Approximately 45% of the grafts use autologous tissue
(most often derived from the top of the iliac crest), 45% use
allogeneic tissue, and the remaining 10% rely on synthetic
or nonhuman materials. For a variety of reasons, fully half
of all grafts eventually fail. Autologous grafts (bone tissue
1Department of Surgery, Massachusetts General Hospital, Harvard, Medical School, Boston, Massachusetts.
2Department of Radiology, Massachusetts General Hospital, Charlestown, Massachusetts.
3Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts.
TISSUE ENGINEERING
Volume 13, Number 4, 2007
# Mary Ann Liebert, Inc.
DOI: 10.1089/ten.2006.0083
1 (page numbers are temporary)

transplanted from one site in the patient to another) have the
best clinical outcome. The grafted tissue includes vital cells
and appropriate growth factors and is remodeled to form a
strong unified structure with the host site. Autologous grafts
are considered the criterion standard, but significant limita-
tions include availability of donor tissue (restricting size and
shape), and requirement of an additional surgical procedure
to harvest tissue that results in intense hip pain, greater risk
of deep tissue infection, hypersensitivity, greater blood loss,
longer operating time, greater cost, longer period of rehabil-
itation, and vascular injury. These complications occur in
10% to 35% of all patients and are particularly acute in the
elderly.2,3
Tissue engineering offers a promising new approach that
can overcome these limitations by producing living and func-
tional tissues of customizable size and shape using porous
3-dimensional (3D) scaffolds and tissue- or organ-specific
progenitor cells.4–7
Materials used as scaffolds for bone tissue engineering,
such as beta-tricalcium phosphate (b-TCP)8,9 poly(lactic-
co-glycolic acid),10,11 andmagnesium12,13 areosteoinductive
or osteoconductive and provide additional strength to the re-
pair site. In addition, b-TCP can be used for 3D printing.
Three-dimensional printing technology is similar to ink-jet
printing and can be used to create complex 3D structures by
integrating a computer-aided design model of an object with
a layer-by-layer printing process. Thus, control of texture and
internal microstructure can be achieved in 3D-printed scaf-
folds.14,15
Hydrogels are used in tissue engineering to facilitate new
tissue formation in a 3D environment.16,17 Specific hydrogels
(fibrin glue, alginate, and Pluronic F127) have been shown
to promote the formation of bone-like tissue from progenitor
cells and osteoblasts18–25 in vitro and in vivo. We have re-
cently evaluated the ability of several hydrogels, including
collagen I (Cellagen), to support formation of new bone-
like tissue using bone marrow–derived mesenchymal stem
cells (MSCs) on b-TCP scaffolds in vitro.26 Newly formed
bone from MSCs suspended in hydrogels has been demon-
strated.18,21,27,28 In our experiments, we used MSCs because
of their high proliferation rate and documented ability to
differentiate into osteocytes29 in hydrogels.30,31 Although
collagen I is a major component of bone, it is infrequently
used in bone tissue engineering; however, we demonstrated
superior bone formation using this hydrogel with b-TCP
scaffolds in vitro.26 These in vitro studies were extended to
evaluate the effectiveness of hydrogels in promoting bone
formation in vivo.
Bone regeneration and growth in vivo can be evaluated
using histology of biopsy specimens and analysis of pro-
collagen levels in blood;32–34 both are invasive procedures.
Non-invasive conventional X-ray techniques or computed to-
mography (CT) scanning are more commonly used in clinical
practice.Despite the significant improvements in the currently
employed imaging techniques, they have limited resolution
and cannot demonstrate the actual regeneration of bone in 3
dimensions.Thiswillbeespecially importantwhenengineered
tissue is transplanted into a defect and its in-growth into na-
tive bone must be noninvasively monitored.
A recently developed tomographic imaging detector, a
flat-panel volume CT (VCT) scanner, enables the acquisi-
tion of high-resolution 3D reconstruction images.35,36 At
this resolution, processes such as the formation of trabecula
within cortical bone can be directly visualized. Based upon
a spatial resolution equivalent to 10% modulation transfer
at 24 line-pairs per cm, VCT provides a resolution of ap-
proximately 250 mm and a contrast resolution of approxi-
mately 5 Hounsfield density units (HU). Therefore, a new
bone island larger than 250 mm with a density greater than 5
HU over the background should be visible.35,36
The present study was conducted to evaluate in vivo for-
mation of bone-like tissue from MSCs suspended in several
hydrogels and seeded onto porous b-TCP scaffolds and to
further validate the efficacy of VCT technique by correlating
parameters measured with this system with more-traditional
measures of bone formation such as histology and mechan-
ical testing.
MATERIALS AND METHODS
Mesenchymal stem cells
Bone marrow was aspirated from the iliac crests of seven
7-month-old Yucatan mini pigs into heparinized syringes
(100 U heparin/mL bone marrow) to prevent clotting. Bone
marrow from several animals was pooled after harvest to
eliminate donor-to-donor variability and plated in cell culture
flasks (Corning Inc., Corning, NY). Adherent cells were cul-
tured in Dulbecco’s modified Eagle medium (Gibco, Invitro-
gen Corp., Grand Island, NY) supplemented with 10% fetal
bovine serum (Sigma–Aldrich, St Louis, MO), 100U/mL
penicillin, and 100mg/mL streptomycin (Gibco). After a
sufficient number of cells had been reached, they were incu-
bated in osteogenic differentiation medium for 6 days. For
osteogenicdifferentiation,growthmediumwas supplemented
with 100 nM dexamethasone, 50mg/mL ascorbic acid, and
10mM b-glycerol phosphate (all from Sigma–Aldrich) as
previously described.29
Construct preparation
The 3D printed b-TCP microporous (100 –250mm) scaf-
folds (Therics, Princeton, NJ) were designed to allow hy-
drogel penetration into the interior of the scaffold and
improve delivery of nutrients and oxygen.26 The scaffolds
measured2.0�2.0�0.7 cm,with a grid of sixteen 2.0�2.0mm
channels (Fig. 1). Scaffolds were sterilized with ethylene ox-
ide. Differentiated MSCs were trypsinized and resuspended
in liquid hydrogels at a concentration of 2�106 cells/mL
2 WEINAND ET AL.

hydrogel. Liquid hydrogel–cell suspension (2.8mL/scaffold)
was applied and allowed to set. Hydrogels prepared without
cells were used as acellular controls.
Fibrin glue. Differentiated MSCs were dispersed in 0.113
mg/mL porcine fibrinogen (Sigma–Aldrich)24 in osteogenic
medium,and1.4mLofcell suspensionwasmixedwith1.4mL
(100 U) human thrombin (Sigma–Aldrich) and applied to
scaffolds. Nearly instantaneous gelation occurred.
Alginate hydrogel. Ultra-pure alginate powder (Novama-
trix, Brakerø´ya, Norway) with high-percentage b-D man-
nuronic acid was dissolved in osteogenic medium to produce
a 3% solution. Differentiated MSCs were dispersed in the
hydrogel, and 2.8mL of the suspension was applied to each
scaffold. Instantaneous gelation was achieved after adding
1mL of 100-mM calcium chloride (Sigma-Aldrich).
Collagen I hydrogel. Collagen I (Cellagen, ICN Bio-
medicals, Aurora, OH) was mixed on ice according to the
manufacturer’s instructions using 5� osteogenic medium.
MSCs were suspended in liquid hydrogel, and 2.8mL of the
suspension was placed onto each scaffold. Gelation occurred
within 1min of incubation at 378C. The scaffolds were pre-
pared on the day of surgery and were kept in medium on ice
until implantation.
Construct implantation
The Institutional Animal Care and Use Committee of
Massachusetts General Hospital approved all animal proce-
dures. Forty 6- to 8-week-old male nude mice (Charles River
Laboratories,Wilmington,MA)were anesthetizedwithAver-
tin 125 to 200mg/kg intraperitoneally. After a dorsal subcu-
taneous pocket was formed, a construct was inserted into the
pocket, and the wound was closed with staples. Each mouse
received an intramuscular injection of Buprenorphine (0.05–
0.10mg/kg) and allowed to recover from surgery. At each
time point (1, 2, 4, and 6 weeks), 10 mice were killed with an
overdose of pentobarbital (200mg/kg intraperitoneally). Six
cellular constructs (2 with each of the three hydrogels) and 4
acellular controls (1 with each of the three hydrogels and 1
scaffold with neither hydrogel nor cells) were explanted at
each time point (Table 1).
Scaffolds were divided into sections containing cubicles
from the inner and outer parts of the scaffold and fixed in
10% buffered formalin or snap-frozen in liquid nitrogen and
stored at –808C for further analysis.
Radiological analysis
Radiological analysis was performed using a VCT scanner
(Siemens, Germany) at 10mA, 80 kV, 100 mAs and 25-�
25-� 18-cm field of view. Radiological density of cubicles
in each 2.0-� 1.0-� 0.7-cm sample was measured in HU.
Mean HU values of 8 cubicles in each sample were calcu-
lated, allowing for comparison of the radiological density of
different samples.
Histological analysis
Samples were demineralized in 0.5M ethylenediamine-
tetraacetic acid for 48 h. They were then fixed in 10%
phosphate buffered formalin, embedded in paraffin, and
sectioned at 5 mm. The sections were stained with hema-
toxylin and eosin (H&E) to demonstrate cell morphology
FIG. 1. Three-dimensional printed Beta-Tricalcium Phosphate
scaffold. Macrophotograph of scaffold, inset: scanning electron mi-
croscope image of the microporous scaffold material forming a
channel.
TABLE 1. NUMBER OF SCAFFOLDS IMPLANTED PER GROUP
Scaffolds with mesenchymal
stem cells Acellular scaffolds
Week Fibrin Alginate Collagen I Fibrin Alginate Collagen I Scaffold only
1 2 2 2 1 1 1 1
2 2 2 2 1 1 1 1
4 2 2 2 1 1 1 1
6 2 2 2 1 1 1 1
IN VIVO FORMATION OF TISSUE-ENGINEERED BONE 3

and distribution and to evaluate newly formed tissue. To-
luidine blue staining for pericellular proteoglycan and
intracellular alkaline phosphatase staining were used to
assess bone formation. Calcification of the newly formed
tissue was assessed using von Kossa staining.
Biomechanical compression testing
Compression testing of the 2.0-�1.0-�0.7-cm specimens
with 8 cubicles was performed using a Texture Analyzer
TA-XT Plus (Texture Technologies, Scarsdale, NY). Uncon-
strained uni-axial compressionwas applied to each specimen;
compressive force and displacement were recorded after the
probe tip contacted the sample. The tip of the probe mea-
sured 2.0�2.0 cm. The geometry of the samples made mea-
suring material moduli difficult. To compare samples, we
first calculated an average stiffness Kave as follows. We
defined a uni-axial stiffness as K¼ dF/dx, where F is the
force and x is the displacement. The stiffness was calculated
at each data point, and Kave was calculated as the average
of the stiffness values in the range 0< x< 1.5mm. To re-
move the effects of complex geometry of the scaffolds and
compare material properties, we defined a relative stiffness
for each sample Krel¼Kave/Kcontrol, where Kcontrol is
the average stiffness of the acellular scaffold without gel.
Krel was used to compare the stiffness between the various
samples and native swine bone from the iliac crest.
Osteonectin gene transcription analysis
Total ribonucleic acid (RNA) was extracted from frozen
specimens using RNA STAT 60 according to the manu-
facturer’s instructions (Tel-Test, Friendswood, TX), quan-
tified using spectrophotometric analysis, and its integrity
verified using agarose gel electrophoresis.37 Threemg ofRNA
from each sample were converted to complementary deox-
yribonucleic acid (cDNA) in a 30-mL reaction containing 1x
polymerase chain reaction (PCR) buffer, 5mM deoxynu-
cleotide triphosphates, 0.5mg random hexamer primers, and
200 units Moloney murine leukemia virus reverse transcrip-
tase (Promega, Madison, WI). After incubation at 378C for
2 h, the reaction was heat-inactivated and diluted to 100mL
with water. The housekeeping glyseraldehyde-3-phosphate
dehydrogenase (GAPDH) was used to normalize for po-
tential variations in absorbance concentration measurements,
first-strand cDNA conversion efficiencies, and potential dif-
ferences in the relative numbers of swine MSC-derived cells
present. Osteonectin (ON) was measured in parallel reac-
tions. Swine-specific primers were designed based upon
GenBank sequences (GAPDH, locus AF017079; ON, locus
AY963262). The GAPDH primers (F-GTTCGAGGACTGG
TCCAAA, R-GCCAGAGTTAAAAGCAGCC) amplify a
332-base pair (bp) segment of the 50 untranslated region. The
ON primers (F-ACCCACCCGTCACTAAGACA, R-TTCC
CCTCCTCCTGTTCTC) amplify a 243-bp segment of the 30
untranslated region. Native swine bone and mouse tissue
were used as controls for transcription analysis. One-mL
aliquots were used in each 40-mL reverse transcriptase PCR
reaction using gene-specific primers. After 30 cycles, 12-mL
aliquots were fractionated on a 2% agarose gel (1% agarose,
1% NuSieve GTG agarose) with 20mg ethidium bromide per
100mL gel solution. Digital images of the gels were obtained
and relative band intensity determined using Scion Image
(Scion Corp, Frederick, MD).
Statistical analysis
Radiological density and the results of mechanical testing
were expressed as means with standard deviations. At each
time point, changes in the radiological density of each con-
struct were correlated with biomechanical strength and ON
expression values. Pearson, Kendall, and Spearman corre-
lation coefficients were calculated using MatLab 7.0 (The
MathWorks, Inc., Natick, MA). Bivariate correlations were
obtained non-parametrically using Spearman rank correla-
tion and Kendall tau b statistics.
RESULTS
Fibrin glue, alginate, and collagen I hydrogels were mixed
with swinebonemarrow–derivedMSCs, applied to3Dprinted
porous b-TCP scaffolds, and implanted subcutaneously in
nude mice. All animals were killed at the predetermined end-
points.
Gross and microscopic evaluation
At each time point, the scaffold material was still present
in all samples. The surface of all cellular samples was cov-
ered with newly formed tissue, with collagen I samples hav-
ing more coverage than alginate and fibrin glue.
During the first 2 weeks, residual hydrogels could be iden-
tified in histological sections. Elongated spindle-shaped cells
with eosinophilic cytoplasm were evenly distributed on the
surface and interior of the scaffolds in all cellular samples.
Toluidine blue stain did not detect any proteoglycan.
At 2 weeks, multiple layers of cells were observed on the
surface of and within the channels in collagen I and fibrin
hydrogel samples and mostly on the surface of alginate sam-
ples. Large round cells with basophilic cytoplasm resembling
osteoblasts were found in collagen I and fibrin hydrogels, and
staining with toluidine blue and alkaline phosphatase was
positive (not shown). Calcium deposition could not be deter-
mined within the scaffold cubicles using von Kossa stain at
this time point.
Formation of bone-like tissue within the cubicles and on
the surface of the scaffold was first demonstrated in col-
lagen I and fibrin glue samples at 4 weeks, with positive
toluidine blue and von Kossa staining (not shown). Bone-
like tissue formation could not be determined in alginate gel
at this time.
At 6 weeks (Fig. 2), bone-like tissue was detected on the
surface of and within all cellular scaffolds, as indicated by
4 WEINAND ET AL.

positive toluidine blue and vonKossa staining; collagen I hy-
drogel showed the most homogeneous distribution of cellu-
lar and bone-specific extracellular matrix components within
and on the surface of the scaffold.
Radiological evaluation
Radiological densities were visible in the cubicles of the
cellular scaffolds with alginate and collagen gels starting at
2 weeks, with fibrin glue at 4 weeks, and in the cubicles of
all cellular scaffolds by 6 weeks. VCT images of collagen
I scaffolds with MSCs at all time points are presented in
Figure 3.
Mean radiological densities of the cubicles are presented
in Figure 4. The densities of the fibrin glue and collagen I
samples started to increase at 4 weeks, whereas in alginate
gel, they remained low. Mean HU of the cubicles in colla-
gen I samples were comparable with those of native porcine
bone at 6 weeks. The radiological densities of the cubicles
in the acellular scaffolds with hydrogels were negligible at
all time points (data not shown). They were comparable with
the densities of scaffolds without hydrogels, which did not
change during the experiment (6-week time point shown in
Fig. 4).
Biomechanical compression testing
All samples demonstrated gradually increasing stiffness
during the experiment, but stiffness remained lower than
that of native swine bone (Fig. 5). Relative stiffness of sam-
ples with cells suspended in collagen I hydrogel increased
FIG. 2. Histochemical and biochemical evaluation of cellular scaffolds at 6 weeks. Hematoxylin and eosin (bar¼ 75mm), and von
Kossa staining (bar¼ 50mm) of b-TCP scaffolds with mesenchymal stem cells in fibrin glue (A, D), alginate (B, E), and collagen I
(C, F) hydrogels. Bone formation with osteoblasts (O) embedded in bone matrix was observed in the pores of the scaffold (S) and in the
open channels. Color images available online at www.liebertpub.com/ten.
FIG. 3. Volumetric computed tomography images of the cel-
lular scaffolds with collagen I hydrogel. Note newly formed high-
density tissue within the channels (arrows) increasing with time.
FIG. 4. Mean radiological densities of the scaffold cubicles.
Collagen I containing scaffolds generated densities comparable to
those of native bone (250 Hounsfield density units (HU)). Radio-
logical densities of the cubicles in the acellular scaffolds without
hydrogels did not change during the experiment (6-week time point
shown).
IN VIVO FORMATION OF TISSUE-ENGINEERED BONE 5

significantly during the first week after implantation ( p<
0.05). Collagen I samples had higher stiffness values than
samples with fibrin glue and alginate hydrogels at all time
points. All scaffolds with MSCs had higher stiffness than
acellular controls with hydrogels (data not shown). The stiff-
ness of acellular scaffolds with hydrogels was similar to that
of scaffolds without hydrogels at all time points and did not
change during the experiment (Fig. 5).
Gene transcription analysis
RNA was isolated from segments of 6-week time-point
explanted scaffolds and examined for expression of ON.
Swine-specific ON expression was detected in tissue derived
from all cellular scaffolds at the 6-week time point and na-
tive swine bone (Fig. 6A) but not mouse. Likewise, all sam-
ples except mouse tissue expressed swine GAPDH (Fig. 6B).
After normalization to GAPDH levels, relative ON tran-
script levels were collagen (1.0)> fibronectin (0.53)>
alginate (0.41).
Correlation analysis
Higher radiological density corresponded to greater bio-
mechanical strength of the samples (Spearman correla-
tion¼ 1.0, Pearson¼ 1.0, Kendall¼ 0.85). This association
was robust across the different correlation coefficients of the
specimensover time.Therewas also high correlation between
radiological density and ON expression values (Spearman
correlation¼ 0.98, Pearson¼ 1.0, Kendall¼ 0.95). In addi-
tion, robust associations were found between biomechanical
stiffness values and ON expression (Spearman correlation¼
0.96, Pearson¼ 0.99, Kendall¼ 0.91). Bivariate correlations
tested using Spearman rank correlations produced the fol-
lowing correlation coefficients: biomechanical stiffness com-
pared with radiological density¼ 0.92 ( p< 0.008), biome-
0.008), biomechanical stiffness compared with ON expres-
sion values¼ 0.98 ( p< 0.001), and radiological density
compared with ON expression values¼ 0.96 ( p<0.002).
Kendall tau b statistics revealed high correlations between
biomechanical stiffness and radiological density (0.90, p<
0.001), biomechanical stiffness and ON expression (0.94,
p< 0.003), and radiological density andONexpression (0.90,
p< 0.001).
DISCUSSION
One of our goals was to compare the ability of fibrin glue,
alginate, and collagen I hydrogels to support formation of
bone-like tissue by differentiated MSCs in porous b-TCP
scaffolds in vivo. In our previous experiments, we demon-
strated that collagen I enhanced development of engineered
bone in vitro better than other hydrogels.26
In the present study, formation of bone-like tissue was
first observed in the constructs with collagen I at 4 weeks;
by 6 weeks, collagen I hydrogel showed the most homo-
geneous distribution of cellular and bone-specific extracel-
lular matrix components within and on the surface of the
scaffold, as demonstrated using H&E, toluidine blue, and
von Kossa staining. In fibrin glue and alginate hydrogels,
formation of bone-like tissue at 6 weeks was slower than in
collagen I. Collagen hydrogel has been successfully used
for cartilage38 and ligament tissue engineering;39 for bone
engineering, collagen I is used mostly in the form of min-
eralized collagen40 or 3D sponges.41
The lack of adequate initial strength of hydrogel–cell
constructs requires additional mechanical support for the
implantation of engineered bone in a weight-bearing posi-
tion. We chose b-TCP as a scaffold material for its docu-
FIG. 5. Biomechanical compression testing. Relative stiffness
of cellular scaffolds is presented as a ratio relative to that of acel-
lular controls. Scaffolds with collagen I hydrogel demonstrated
the highest stiffness at 6 weeks. The stiffness of acellular scaffolds
without hydrogels did not change during the experiment.
FIG. 6. Reverse transcriptase polymerase chain reaction gene
expression analysis. Expression of osteonectin (Panel A) and
glyseraldehyde-3-phosphate dehydrogenase (Panel B) were exam-
ined in scaffolds explanted after 6 weeks that contained MSCs
seeded with the hydrogels collagen I (C), alginate (A), and fibrin
glue (F). Native mouse (M) and swine bone (S) were used to
demonstrate species specificity of primers. A deoxyribonucleic
acid size marker was run in the first lane of each gel.
6 WEINAND ET AL.

mented ability to be osteoconductive and to provide ade-
quate initial strength.8,9 Additionally, this material can be
used for 3D printing, which allows fabrication of scaffolds
with custom size and shape corresponding to the bone de-
fect that needs to be repaired.15,18 The slower formation of
tissue in fibrin glue and alginate hydrogels in porous b-TCP
scaffolds may be due to a misbalance between the rates of
formation of new bone-like tissue and degradation of the
hydrogels. This study did not address the degradation rates
of the hydrogels; however, others have demonstrated that
formation of new bone matrix depends on the type of hy-
drogel and its degradation rate.42–46
Formation of appropriate bone-specific matrix and calcium
deposition in the engineeredbone can alsobe assessedby eval-
uating construct stiffness. We performed an unconstrained
uni-axial compression test and calculated the average and
relative (compared with scaffold without cells or hydrogel)
stiffness of each sample. This allowed for comparison of the
material properties, disregarding the effects of the scaffold’s
complex geometry. Collagen I samples had higher stiffness
values than samples with fibrin glue and alginate hydrogels at
all time points, consistent with histological results and ra-
diological density measurements. All cell-containing samples
had higher compression values than acellular samples at all
time points, suggesting that the newly formed tissue in cel-
lular specimens contributed to mechanical stability.
Detected expression of swine-specific messenger RNA
transcripts indicates that the implanted MSCs had survived
the experimental period and continued to contribute to the
developing engineered bone. There was greater expression
of bone-specific ON in collagen I scaffolds at 6 weeks than
with other hydrogels; this is consistent with histological
findings and the results of biomechanical testing. ON is ex-
pressed in areas of active remodeling and is one of the most
abundant noncollagenous proteins in bone. Although its func-
tion is not well understood, the expression of ON indicates the
presence of mature bone-forming osteoblasts.47,48 In our pre-
vious in vitro studies, expression of ONwas a better indicator
of bone formation than that of bone morphogenetic protein 4,
alkaline phosphatase, or osteopontin.26 Large round cells with
basophilic cytoplasm resembling osteoblasts were first found
in collagen and fibrin glue hydrogels and increased over time,
which may explain the higher ON expression values in these
2 hydrogels.
The second goal of our study was to validate the efficacy
of a recently introduced high-resolution VCT scanner35,36
for monitoring the formation of engineered bone. Standard
CT or quantitative CT scanners have been used to assess
trabecular bone structure and mineral density. Good corre-
lation has been found between measured density values
and biomechanical properties.49–51 We demonstrated in our
previous experiments that a high-resolution VCT scanner
provides resolution superior to conventional radiological
techniques and is valuable in the detection of engineered
bone developing in vitro.26 In the present study, the high-
resolution VCT scanner allowed us to monitor the formation
of bone-like structures in the channels of the scaffolds. Al-
though we performed ex vivo imaging of explanted scaffolds
to validate the efficacy of VCT to provide high-resolution
images, the required calibrations for subcutaneous or deeply
embedded structures result in VCT images with minimal
effect on HU resolution.35,36 This technique is potentially
useful for noninvasive monitoring of engineered bone after
implantation. The highest mean HU density values were
found in the cubicles of the scaffolds with MSCs suspended
in collagen I hydrogel after 4 weeks; the radiological density
approximated that of native swine bone at 6 weeks. Low
HU density values of alginate during the first 4 weeks may
be due to the slower degradation rate of this hydrogel42–46,52
delaying the formation of bone-like tissue more than colla-
gen I.
The results of radiological evaluation were consistent with
histological findings. We also found high correlation coeffi-
cients between mean HU density values, results of compres-
sion testing (0.85–1.0) over time, and expression of bone-
specific ON at 6 weeks (0.95–1.0). All acellular controls had
lower radiological densities than cellular specimens, sug-
gesting that newly formed bone-like tissue contributed to the
increasing HU density values.
A limitation of the present study is associated with our
animal model. Implantation of scaffolds into the subcutane-
ous dorsal pockets of nude mice is a standard technique used
for the evaluation of in vivo osteogenic potential of different
cell sources, their carriers, and scaffolds.53 This is a relatively
inexpensive model that allows for testing of nonautologous
cells; however, it represents an ectopic non-weight bearing
position for the evaluation of bone formation and may pro-
duce marked angiogenic and inflammatory response54 that
can affect the development of engineered bone.
In conclusion, we demonstrated that bone-like tissue
could be successfully engineered in an ectopic position
in vivo, using porous b-TCP scaffolds seeded with MSCs
suspended in hydrogels. Samples with collagen I had greater
mineralization, secretion of bone-specific matrix, and gene
expression of bone-specific markers than fibrin glue and
alginate hydrogels. Our results suggest that high-resolution
VCT is a valuable method for the evaluation of bone-like
tissue development and can potentially be used as a reli-
able noninvasive technique for monitoring engineered bone
in vivo. Future work will concentrate on further evaluation
of osteogenesis in hydrogel–b-TCP scaffolds with autolo-
gous differentiated MSCs implanted into a bone defect in a
weight-bearing position in an animal model.
ACKNOWLEDGMENTS
This study was generously funded by Therics Inc. We
thank Dr. Frederic Shapiro (Children’s Hospital, Boston) for
histopathological evaluation and Dr. Harutsugi Abukawa
(Massachusetts General Hospital, Boston) for reviewing the
manuscript.
IN VIVO FORMATION OF TISSUE-ENGINEERED BONE 7

Dr. Gupta has served as a consultant to Siemens Cor-
poration, Forchheim, Germany.
REFERENCES
1. FDA Meeting. Human Bone Allograft: Manipulation and Ho-
mologous Use in Spine and Other Orthopedic Reconstruction
and Repair. Bethesda, MD: National Institutes of Health Clin-
ical Center, 2000.
2. Gazdag, A.R., Lane, J.M., Glaser, D., Forster, R.A. Alterna-
tives to autogenous bone graft: efficacy and indications. J Am
Acad Orthop Surg 3, 1, 1995.
3. Muscheler, G.F., Nakamoto, C., Griffith, L.G. Engineering
principles of clinical cell-based tissue engineering. J Bone
Joint Surg Am 86, 1541, 2004.
4. Langer, R., Vacanti, J.P. Tissue engineering. Science 260,
920, 1993.
5. Wald, H.L., Sarakinos, G., Lyman, M.D., Mikos, A.G., Vacanti,
J.P., Langer, R. Cell seeding in porous transplantation devices.
Biomaterials 14, 270, 1993.
6. Montjovent, M.O., Burri, N., Mark, S., Federici, E., Scaletta, C.,
Zambelli, P.Y., Hohlfeld, P., Leyvraz, P.F., Applegate, L.L.,
Pioletti, D.P. Fetal bone cells for tissue engineering. Bone 35,
1323, 2004.
7. Shin, M., Yoshimota, H., Vacanti, J.P. In vivo bone tissue
engineering using mesenchymal stem cells on a novel elec-
trospun nanofibrous scaffold. Tissue Eng 10, 33, 2004.
8. Hinz, P.,Wolf, E., Schwesinger, G., Hartelt, E., Ekkernkamp, A.
A new resorbable bone void filler in trauma: early clinical ex-
perience and histologic evaluation. Orthopedics 2002;25:s597.
9. Porter, B.D., Oldham, J.B., He S.L., Zobitz M.E., Payne,
R.G., An K.N., Currier, B.L., Mikos, A.G., Yaszemski, M.J.
Mechanical properties of a biodegradable bone regenera-
tion scaffold. J Biomech Eng 122, 286, 2000.
10. Dean, D., Topham, N.S., Meneghetti, S.C., Wolfe, M.S., Jep-
sen, K., He, S., Chen, J.E., Fisher, J.P., Cooke, M., Rimnac, C.,
Mikos, A.G. Poly (propylene fumarate) and poly (DL-lactic-
co-glycolic acid) as scaffold materials for solid and foam-
coated composite tissue-engineered constructs for cranial
reconstruction. Tissue Eng 9, 495, 2003.
11. Lieb, E., Tessmar, J., Hacker, M., Fischbach, C., Rose, D.,
Blunk, T., Mikos, A.G., Gopferich, A., Schulz, M.B. Poly
(D,1-lactic acid)-poly(ethylene glycol)-monomethyl ether di-
block copolymers control adhesion and osteoblastic differen-
tiation of marrow stromal cells. Tissue Eng 9, 71, 2003.
12. Dhert, W.J., Klein, C.P., Jansen, J.A., van der Velde, E.A.,
Vriesde, R.C., Rozing, P.M., de Groot K. A histological and
histomorphometrical investigation of fluorapatite, magnesium
whitlockite, and hydroxylapatite plasma-sprayed coatings in
goats. J Biomed Mater Res 27, 127, 1993.
13. Witte F., Kaese, V., Meyer-Lindenberg, A., Switzer, E.,
Haferkamp, H., Wirth, C.J., Windhagen H. Degrading Mag-
nesium Implants Increase Periosteal and Endosteal Bone For-
mation (Paper no. 0256). ORS Meeting, San Francisco, CA,
2004.
14. Sastry, S.V., Nyshadham, J.R., Fix, J.A. Recent technological
advances in oral drug delivery—a review. Pharm Sci Technol
Today 3, 138, 2000.
15. Curodeau, A., Sachs, E., Caldarise, S. Design and fabrication
of cast orthopedic implants with freeform surface textures
from 3-D printed ceramic shell. J Biomed Mater Res 53, 525,
2000.
16. Fuchs, J.R., Pomerantseva, I., Ochoa, E.R., Vacanti, J.P.,
Fauza, D.O. Fetal tissue engineering: in vitro analysis of
muscle constructs. J Pediatr Surg 38, 1348, 2003.
17. Vinatier, C., Guicheux, J., Daculsi, G., Layrolle, P., Weiss, P.
Cartilage and bone tissue engineering using hydrogels. Biomed
Mater Eng 16:S107, 2006.
18. Schantz, J.T., Teoh, S.H., Lim, T.C., Endres, M., Lam, C.X.,
Hutmacher, D.W. Repair of calvarial defects with customized
tissue-engineered bone grafts I. Evaluation of osteogenesis in
a three-dimensional culture system. Tissue Eng 9:S113, 2003.
19. Lawson, M.A., Barralet, J.E., Wang, L., Shelton, R.M.,
Triffitt, J.T. Adhesion and growth of bone marrow stromal
cells on modified alginate hydrogels. Tissue Eng 10, 1480,
2004.
20. Simmons, C.A., Alsberg, E., Hsiong, S., Kim, W.J., Mooney,
D.J. Dual growth factor delivery and controlled scaffold
degradation enhance in vivo bone formation by transplanted
bone marrow stromal cells. Bone 35, 562, 2004.
21. Schantz, J.T., Hutmacher, D.W., Lam, C.X., Brinkmann, M.,
Wong, K.M., Lim, T.C., Chou, N., Guldberg, R.E., Teoh, S.H.
Repair of calvarial defects with customized tissue engineered
bone grafts II. Evaluation of cellular efficiency and efficacy
in vivo. Tissue Eng 9:S127, 2003.
22. Fowler, E.B., Cuenin, M.F., Hokett, S.D., Peacock, M.E.,
McPherson, J.C. 3rd, Dirksen, T.R., Sharawy, M., Billman,
M.A. Evaluation of Pluronic polyols as carriers for grafting
materials: study in rat calvaria defects. J Periodontol 73, 191,
2002.
23. Chang, S.C., Wei, F.C., Chuang, H., Chen, Y.R., Chen, J.K.,
Lee, K.C., Chen, P.K., Tai, C.L., Lou, J. Ex vivo gene therapy
in autologous critical-size craniofacial bone regeneration.
Plast Reconstr Surg 112, 1841, 2003.
24. Silverman, R.P., Passaretti, D., Huang, W., Randolph, M.A.,
Yaremchuk, M.J. Injectable tissue-engineered cartilage using
a fibrin glue polymer. Plast Reconstr Surg 103, 1809, 1999.
25. Johnson, T.S., Xu, J.W., Zaporojan, V.V., Mesa, J.M.,
Weinand, C., Randolph, M.A., Bonassar, L.J., Winograd,
J.M., Yaremchuk, M.J. Integrative repair of cartilage with
articular and nonarticular chondrocytes. Tissue Eng 10, 1308,
2004.
26. Weinand, C., Pomerantseva, I., Neville, C.M., Gupta, R.,
Weinberg, E., Madisch, I., Abukawa, H., Troulis, M.J., and
Vacanti, J.P. Hydrogel–b-TCP scaffolds and stem cells for
tissue engineering bone. Bone 38, 555, 2006.
27. Bensaid, W., Triffitt, J.T., Blanchat, C., Oudina, K., Sedel, L.,
Petite, H. A biodegradable fibrin scaffold for mesenchymal
stem cell transplantation. Biomaterials 24, 2497, 2003.
28. Chung, T.W., Yang, J., Akaike, T., Cho, K.Y., Nah, J.W.,
Kim, S.I., Cho, C.S. Preparation of alginate/galactosylated
chitosan scaffold for hepatocyte attachment. Biomaterials 23,
2827, 2002.
29. Abukawa, H., Shin, M., Williams, W.B., Vacanti, J.P., Kaban,
L.B., Troulis, M.J. Reconstruction of mandibular defects with
autologous tissue-engineered bone. J Oral Maxillofac Surg
62, 601, 2004.
8 WEINAND ET AL.

30. Rutherford. R.B., Gu, K., Racenis, P., Krebsbach, P.H. Early
events. The in vitro conversion of BMP transduced fibroblasts
to chondroblasts. Connect Tissue Res 44, 117, 2003.
31. Rutherford, R.B., Moalli, M., Franceschi, R.T., Wang, D., Gu,
K., Krebsbach, P.H. Bone morphogenetic protein-transduced
human fibroblasts convert to osteoblasts and form bone
in vivo. Tissue Eng 8, 441, 2002.
32. Nenonen A., Cheng, S., Ivaska, K.K., Alatalo, S.L., Lehtimaki,
T., Schmidt-Gayk H., Uusi-Rasi, K., Heinonen, A., Kannus, P.,
Sievanen, H., Vuori, I., Vaananen, H.K., Halleen, J.M. Serum
TRACP 5b is a useful marker for monitoring alendronate
treatment: comparison with other markers of bone turnover.
J Bone Miner Res 20, 1804, 2005.
33. Chen, P., Satterwhite, J.H., Licata, A.A., Lewiecki, E.M.,
Sipos, A.A., Misurski, D.M., Wagman, R.B. Early changes in
biochemical markers of bone formation predict BMD re-
sponse to teriparatide in postmenopausal women with osteo-
porosis. J Bone Miner Res 20, 962, 2005.
34. Dobnig, H., Sipos, A., Jiang, Y., Fahrleitner-Pammer A.,
Ste-Marie, L.G., Gallagher, J.C., Pavo, I., Wang, J., Eriksen,
E.F. Early changes in biochemical markers of bone formation
correlate with improvements in bone structure during ter-
iparatide therapy. J Clin Endocrinol Metab 90, 3970, 2005.
35. Gupta, R., Grasruck, M., Suess, C., Bartling, S.H., Schmidt, B.,
Stierstorfer, K., Popescu, S., Brady, T., Flohr, T. Ultra-high
resolution flat-panel CT: fundamental principles, design ar-
chitecture, and system characterization. Eur Radiol 16, 1191,
2006.
36. Gupta, R., Bartling, S.H., Basu, S.K., Ross, W.R., Becker, H.,
Pfoh, A., Brady, T., Curtin, H.D. Experimental flat-panel
high-spatial-resolution volume CT of the temporal bone. Am.
J. Neuroradiol 25, 1417, 2004.
37. Rosen, K.M., Villa-Komaroff, L. An alternative method for
the visualization of RNA in formaldehyde agarose gels. Focus
845, 23, 1990.
38. Yamaoka, H., Asato, H., Ogasawara, T., Nishizawa, S.,
Takahashi, T., Nakatsuka, T., Koshima, I., Nakamura, K.,
Kawaguchi, H., Chung, U.I., Takato, T., Hoshi, K. Cartilage
tissue engineering using human auricular chondrocytes em-
bedded in different hydrogel materials. J Biomed Mater Res A
78, 1, 2006.
39. Noth, U., Schupp, K., Heymer, A., Kall, S., Jakob, F., Schutze,
N., Baumann, B., Barthel, T., Eulert, J., Hendrich, C. Anterior
cruciate ligament constructs fabricated from human mesen-
chymal stem cells in a collagen type I hydrogel. Cytotherapy 7,
447, 2005.
40. Domaschke, H., Gelinsky, M., Burmeister, B., Fleig, R.,
Hanke, T., Reinstorf, A., Pompe, W., Rosen-Wolff, A. In vitro
ossification and remodeling of mineralized collagen I scaf-
folds. Tissue Eng 12, 949, 2006.
41. Matsuno, T., Nakamura, T., Kuremoto, K., Notazawa, S., Na-
kahara, T., Hashimoto, Y., Satoh, T., Shimizu, Y. Development
of beta-tricalcium phosphate/collagen sponge composite for
bone regeneration. Dent Mater J 25, 138, 2006.
42. Weng, Y., Cao, Y., Silva, C.A., Vacanti, M.P., Vacanti, C.A.
Tissue-engineered composites of bone and cartilage for man-
dible condylar reconstruction. J Oral Maxillofac Surg 59, 185,
2001.
43. Alsberg, E., Anderson, K.W., Albeiruti, A., Franceschi, R.T.,
Mooney., D.J. Cell-interactive alginate hydrogels for bone
tissue engineering. J Dent Res 80, 2025, 2001.
44. Burdick, J.A., Mason, M.N., Hinman, A.D., Thorne, K., An-
seth, K.S. Delivery of osteoinductive growth factors from
degradable PEG hydrogels influences osteoblast differentia-
tion and mineralization. J Control Release 18, 53, 2002.
45. Shin, H., Zygourakis, K., Farach-Carson, M.C., Yaszemski,
M.J., Mikos, A.G. Modulation of differentiation and miner-
alization of marrow stromal cells cultured on biomimetic
hydrogels modified with Arg-Gly-Asp containing peptides.
J Biomed Mater Res 69A, 535, 2004.
46. Temenoff, J.S., Park, H., Jabbari, E., Conway, D.E., Sheffield,
T.L., Ambrose, C.G., Mikos, A.G. Thermally cross-linked
oligo (poly (ethylene glycol) fumarate) hydrogels support
osteogenic differentiation of encapsulated marrow stromal
cells in vitro. Biomacromolecules 5, 5, 2004.
47. Bruder, S.P., Horowitz, M.C., Mosca, J.D., Haynesworth, S.E.
Monoclonal antibodies reactive with human osteogenic cell
surface antigens. Bone 21, 225, 1997.
48. Buttery, L.D., Bourne, S., Xynos, J.D., Wood, H., Hughes,
F.J., Hughes, S.P., Epis, V., Polak, J.M. Differentiation of
osteoblasts and in vitro bone formation from murine embry-
onic stem cells. Tissue Eng 7, 89, 2001.
49. Jiang, Y., Zhao, J., Augat, P., Ouyang, X., Lu, Y., Majumdar,
S., Genant, H.K. Trabecular bone mineral and calculated
structure of human bone specimens scanned by peripheral
quantitative computer tomography: relation to biomechanical
properties. J Bone Min Res 13, 1783, 1998.
50. Rho, J.Y., Hobatho, M.C., Ashman, R.B. Relation of me-
chanical properties to density and CT numbers in human
bone. Med Eng Phys 17, 347, 1995.
51. Korkusuz, F., Akin, S., Akkus, O., Korkusuz, P. Assessment of
mineral density and atomic content of fracture callus by quan-
titative computerized tomography. J Orthop Sci 5, 248, 2000.
52. Wang, L., Shelton, R.M., Cooper, P.R., Lawson, M., Triffitt,
J.T., Barralet, J.E. Evaluation of sodium alginate for bone
marrow cell tissue engineering. Biomaterials 24, 3475, 2003.
53. Muraglia, A., Martin, I., Cancedda, R., Quarto, R. A nude
mouse model for human bone formation in unloaded condi-
tions. Bone 22, 131, 1998.S.
54. Rucker, M., Laschke, M.W., Junker, D., Carvalho, C.,
Schramm, A., Mulhaupt, R., Gellrich, N.C., Menger, M.D.
Angiogenic and inflammatory response to biodegradable scaf-
folds in dorsal skinfold chambers of mice. Biomaterials 27,
5027, 2006.
Address reprint requests to:
Joseph P. Vacanti, M.D.
Department of Surgery
Massachusetts General Hospital
Warren 1151
55 Fruit Street
Boston MA 02114
E-mail: jvacanti@partners.org
IN VIVO FORMATION OF TISSUE-ENGINEERED BONE 9