Human shaped thumb bone tissue engineered by hydrogel-beta-tricalciumphosphate/poly-epsilon-caprolactone scaffolds and magnetically sorted stem cells.

ORIGINAL ARTICLE
Human Shaped Thumb Bone Tissue Engineered by
Hydrogel-�-Tricalciumphosphate/Poly-�-Caprolactone
Scaffolds and Magnetically Sorted Stem Cells
Christian Weinand, MD, PhD,* Rajiv Gupta, PhD, MD,§ Eli Weinberg, BS,‡ Ijad Madisch, MD,§
Jesse B. Jupiter, MD,† and Joseph P. Vacanti, MD*
Abstract: Traumatic amputation of a thumb with bone loss leaves a
patient in severe disability. Reconstructive procedures are restricted
by limited shape and have the disadvantage of severe donor-site
morbidity. To overcome these limitations, we used a tissue engi-
neering approach to create a distal thumb bone phalanx, combining
magnetically sorted 133� human mesenchymal stem cells (hMSCs)
suspended in successful tested hydrogels for bone formation and
porous 3-dimensionally printed scaffolds (3DP) in the shape of a
distal thumb bone phalanx. Collagen I and fibrin glue hydrogels with
suspended hMSCs were first histologically evaluated in vitro for
bone formation after 6 weeks.
Then 3DP scaffolds, made from a mix of osteoinductive and
-conductive �-tricalciumphosphate (�-TCP) and poly-�-caprolac-
tone (PCL), with hydrogels and suspended hMSCs, were implanted
into nude mice subcutaneously for 15 weeks. Histologic evaluation,
high-resolution volumetric CT (VCT) scanning, and biomechanical
testing confirmed formation of bonelike tissue. Both hydrogels with
CD 133� hMSCs on 3DP scaffolds supported bone formation.
Collagen I resulted in radiologically better bone formation. Bone
tissue can be successfully tissue engineered with CD 133� hMSCs,
collagen I hydrogels, and porous 3DP �-TCP/PCL scaffolds.
Key Words: tissue engineering, thumb, bone, reconstruction,
hMSCs, CD 133�, magnetically cell sorting
(Ann Plast Surg 2007;59: 46–52)
The reconstruction of severely traumatized or congenitallydeficient digits in young patients who need multiple
complicated reconstructive procedures is a challenging prob-
lem. Most limbs are amputated by crushing or avulsing
injuries, which make surgical repair more difficult and lower
the percentage of viability.1–4 The degree of disfigurement
and function loss of the hand can be severe and permanent. In
general, single-digit replantation is recommended only in
selected circumstances, such as for patients with specialized
occupations like playing a musical instrument.
The exception is the thumb,5,6 provided the part and
circumstances are suitable. Thumb-reconstruction techniques
range from bone lengthening to bone grafting, microvascular
toe-to-thumb transfer, or even free whole-joint transfer, with
use of a microvascular anastomosis.7–10 However, these tech-
niques result in donor-site morbidity and limits in shape.
Bone as connective tissue does not regrow in length once
lost. Consequently, amputation of bone results in shortening and
leading to permanent loss of structure and function.11 The ideal
bone substitute should provide stability, motion, and durability.
Although a vascularized autogenous toe-to-hand trans-
fer has a potential to meet these demands, it is limited by
donor-site morbidity and the availability of tissue.12 The
major problem with regard to transplantation of allografts and
xenografts remains the necessity for lifelong immunosuppres-
sion. Another solution is needed.
More recently, techniques of engineering of new tissue
have been introduced to reimplant autologous, cultured os-
teocytes on a scaffold.13 The types of biomaterials currently
used include alginic acid, fibrin, polyglycolic acid, �-trical-
cium phosphate (�-TCP), poly (�) caprolactone (PCL), and
poly (lactic-glycolic acid).14–16 All these materials support
the initial growth and differentiation of cells but eventually
degrade, allowing the cells to produce extracellular matrix
components.
However, the amount of autologous osteocytes is lim-
ited. The use of mesenchymal stem cells could potentially
overcome these limitations. It was shown previously that
stem cells have a high proliferation potential and can be
differentiated into several cell lines.17 CD 133� stem cells
are described to be hematopoietic progenitor cells but also
potentially participate in the formation of connective tissue.18
Received February 20, 2007, and accepted for publication February 21, 2007.
From the *Laboratory for Tissue Engineering and Organ Fabrication and
†Hand and Upper Extremity Service, Massachusetts General Hospital,
Harvard Medical School, Boston, MA; ‡Massachusetts Institute of Tech-
nology, Cambridge, MA; and the §Department of Radiology, Massachu-
setts General Hospital, Harvard Medical School, Charlestown, MA.
Presented at the Annual Meeting of the Northeastern Society of Plastic
Surgeons, Boston, MA, November 30–December 3, 2006.
This study was funded by the Therics Corporation, NJ.
Reprints: Joseph P. Vacanti, MD, Chief, Department of Pediatric Surgery,
Massachusetts General Hospital, John Homans Professor of Surgery, Sur-
geon-in-Chief, Massachusetts General Hospital for Children, Director, Pe-
diatric Transplantation, Director, Laboratory for Tissue Engineering and
Organ Fabrication, Warren 11-1157, Massachusetts General Hospital, 55
Fruit Street, Boston, MA 02114. E-mail: JVacanti@Partners.org.
Copyright © 2007 by Lippincott Williams & Wilkins
ISSN: 0148-7043/07/5901-0046
DOI: 10.1097/01.sap.0000264887.30392.72
Annals of Plastic Surgery  Volume 59, Number 1, July 200746

These cells can be sorted by magnetic antibody labeling and
magnetic sorting.
To achieve bone formation in a desired shape, pre-
formed scaffolds are used. However, elaborate seeding con-
ditions are needed to achieve uniform cellular distribution,
sustain cell viability, and provide nutrients for tissue forma-
tion in scaffolds.19,20 To address limitations of current seed-
ing methods for cellular growth into scaffolds and bone
formation, we used hydrogels to facilitate delivery and dis-
tribution of cells in scaffolds made of �-TCP. In previous
experiments, collagen I and fibrin glue were superior in bone
formation than alginate or Pluronic F 127 in vitro and in
vivo.21,22 We chose collagen I and fibrin glue as hydrogels.
Hydrogels lack the initial mechanical strength needed for
weight bearing. Therefore, an additional support is needed.
Many synthetic and natural materials such as �-TCP,23 poly-
(lactic-coglycolic) acid (PLGA), and magnesium possess ade-
quate strength for bone reconstruction and have proven to be
osteoinductive and osteoconductive.24–27 Ultimately, for defects
with substantial curvature, the tissue-engineered constructs
should also have appropriate topography.
Here, 3-dimensional printing (3DP) can be used. 3DP
uses a technology similar to ink-jet printing. From a CAD
picture of the desired part, a slicing algorithm draws detailed
information for every layer. Each layer begins with a thin
distribution of powder spread over the surface of a powder
bed. Using a technology similar to ink-jet printing, a binder
material selectively joins particles where the object is to be
formed.28
In the current 2 experiments, 3DP scaffolds made from
�-TCP and PCL in the form of a distal phalanx of a human
thumb were used. We investigated the osteogenic potential of
magnetically sorted CD 133� human mesenchymal stem
cells (hMSCs) in 2 different hydrogels in vitro and vivo on
porous 3DP �-TCP/PCL scaffolds.
MATERIALS AND METHODS
Model for Human Phalanx-Shaped Scaffolds
An ultrahigh-resolution volumetric CT (VCT) image
was created from a right human cadaver hand. The picture of
the proximal and distal phalanx of the thumb was isolated
from this image series and digitally reconstructed by a slicing
algorithm to draw detailed information for every layer into a
3-dimensional CAD picture (Fig. 1). This image was trans-
ferred to a customized 3-dimensional printer at MIT.
3DP �-TCP/PCL Phalanx Scaffolds
The porous scaffolds were fabricated from a mix of
50% PCL (Sigma Aldrich Co, St. Louis, MO; �65,000 MW)
and 50% �-TCP (Cosmocel). Each layer was started with a thin
distribution of powder spread over the surface of a powder bed.
Using a technology similar to ink-jet printing, a binder material
selectively joins particles where the object was to be formed.
The proper placement of droplets can be used to create
surfaces of controlled texture and to control the internal
microstructure of the printed part. This layer-by-layer process
was repeated until the part was completed. Following a heat
treatment, unbound powder was removed, leaving the fabri-
cated part. The scaffolds were 2.5-cm long in the exact shape
of the high-resolution picture from the distal human phalanx
and had an overall porosity of �90% (Fig. 2). The scaffolds
were controlled by VCT scan for quality of the reconstruction
and density measurement before implantation.
Mesenchymal Stem Cells, CD 133� Selection
Human bone marrow from human femoral heads was
minced by using a bone cutter, filtered from debris through a
100-�m and then a 40-�m mesh. For osteogenic differenti-
ation, growth medium was supplemented with 100-nM dexa-
methasone, 50 mg/mL ascorbic acid, and 10-mM �-glycer-
olphosphate (all from Sigma-Aldrich Co).
Half of the gained bone marrow was washed in a buffer
solution (PBS, pH 7.2, 0.5% FBS, and 2-mM EDTA), then
incubated with CD133 antibodies attached to magnetic beads
at 4°C for 30 minutes. Buffer solution was poured into a
10-mL syringe and placed into a stand with a magnet (all
from Miltenyi, CA). CD 133� cells were selected magneti-
cally by the surrounding magnet. Selected cells were then
washed into tissue culture wells with DMEM supplemented
with 100 U/mL penicillin/100 �g streptomycin and 10%
FBS. CD133� cell selection was confirmed by FITC-labeled
secondary antibody staining under fluorescent microscopy.
Immediately after selection, CD 133� hMSCs and CD
FIGURE 1. a, Three-dimensional VCT
picture of the human hand. b, Volu-
metric reconstructed CAD image of
human thumb.
Annals of Plastic Surgery  Volume 59, Number 1, July 2007 Human Shaped Thumb Bone Tissue
© 2007 Lippincott Williams & Wilkins 47

133�-depleted hMSCs were plated in cell culture flasks
(Corning Inc, Corning, NY) and adherent cells were ex-
panded for 10–12 days, using growth medium. hMSCs were
passaged at 2:3 ratio when 80% subconfluence was reached.
After a sufficient number of cells were reached, they were
cultured in osteogenic differentiation medium for 6 days. Cell
viability assessed by trypan blue staining was 90% and
higher.
In Vitro Constructs
Collagen I hydrogel (Cellagen; ICN Biomedicals Inc,
Aurora, OH) was mixed on ice according to the manufactur-
er’s instructions using 5� concentrated OS medium on ice.
hMSCs 200,000/mL were mixed with liquid hydrogel; 2-mL
cell hydrogel mix was placed into a culture dish in a 37°C
warm incubator for evaluation of bone formation. Gelation of
the hydrogel occurred after 1 minute of incubation.
Fibrin Glue Hydrogel
Differentiated hMSCs were dispersed in 0.113 mg/mL
porcine fibrinogen,29 (Sigma-Aldrich Co.) in OS medium,
and 1 mL of cell suspension was mixed with 1 mL (100 U)
of thrombin (Sigma-Aldrich Co.) so a final concentration of
200,000 cells/mL hydrogel resulted; 2-mL cell hydrogel mix
was directly poured into a cell culture dish and immediate
gelation was observed.
In Vivo Constructs
All scaffolds were ETO sterilized and kept under vac-
uum sterilely for 3 days until immediate use. Two constructs
of each type hydrogel with CD 133� and CD 133�-depleted
cells were prepared, resulting in 8 cellular constructs.
Collagen I Hydrogel
As described before, 2-mL cell hydrogel mix was
placed directly after preparation of the cell hydrogel mix onto
each scaffold in a 37°C warm incubator. Gelation occurred
after 1 minute incubation.
Fibrin Glue Hydrogel
Similarly as described above, 2-mL cell hydrogel mix
(200,000/mL) was directly applied onto each scaffold. Nearly
instantaneous gelation occurred.
Upon gelation, all constructs were implanted subcuta-
neously in the back of nude mice for 15 weeks. Acellular
controls were created with hydrogel only and without cells
and hydrogel and implanted subcutaneously.
Implantation
The Institutional Animal Care and Use Committee of
the Massachusetts General Hospital approved all animal pro-
cedures. Six- to 8-week-old male nude mice (Charles River
Laboratory, Boston, MA) were anesthetized with Avertin,
125–200 mg/kg IP. A dorsal subcutaneous pocket was
formed, and 1 construct with hMSCs suspended in hydrogel
(8 mice with �-TCP/PCL cellular constructs) or scaffolds
with or without hydrogel (6 mice) were inserted into the
pocket. The wound was closed with staples, each mouse
received buprenorphine 0.05–0.1 mg/kg intramuscularly, and
animals were allowed to recover from surgery.
In Vitro Assessment
After 6 weeks in vitro, hydrogels and cells were em-
bedded into 10% phosphate-buffered formalin for histologic
analysis. The samples were embedded into paraffin; 4-�m
sections were taken and evaluated histologically. Staining
was done by H&E for cell morphology and viability, tolu-
idine blue for pericellular proteoglycan, alkaline phosphatase
for bone tissue formation, and van Kossa staining for calci-
fication of the newly formed tissue.
In Vivo Assessment
Radiologic Evaluation
After 15 weeks, VCT scanning of each specimen was
performed at 120 kV tube voltage and 10-mA tube current on
a high-resolution VCT scanner (Siemens, Germany) in vivo.
The VCT scanner effective field of view was 25 � 25 � 18
cm, with isotropic resolution of �200 m, with 2 � 2 binning,
and volumetric coverage. The projection data were recon-
structed using a modified Feldkamp algorithm. Radiologic
density of the samples was assessed by measuring Hounsfield
density units. To calculate a mean value for bone formation
for each part of the scaffold, the scaffolds were separated into
3 parts of 0.6-cm length: tip, middle, and basis. Mean values
for 6 separate points from tip, middle, and basis, respectively,
FIGURE 2. a, b, Macroscopical appearance of the 3-dimensional printed scaffold of the distal phalanx. c, Scanning electron
microscope picture of the scaffold. The bar represents 50-�m length.
Weinand et al Annals of Plastic Surgery  Volume 59, Number 1, July 2007
© 2007 Lippincott Williams & Wilkins48

were calculated. Then mice were killed using an overdose of
pentobarbital (200 mg/kg IP), and constructs were explanted
and processed.
Gross and Histologic Evaluation
Then specimens were harvested and examined for appear-
ance. Then specimens were cut in 3 parts of 0.6 cm length to
assess bone formation through the length of the scaffold. From
each part of each specimen, 0.1-cm-thick cuts were taken from
the proximal part of the scaffold; they were embedded into 10%
phosphate-buffered formalin for histologic analysis. The sam-
ples were embedded into paraffin; 4-�m sections were taken and
evaluated histologically. Staining was done by H&E for cell
morphology and viability, toluidine blue for pericellular proteo-
glycan, alkaline phosphatase for bone tissue formation, and van
Kossa staining for calcification of the newly formed tissue.
Histologic evaluation of the samples was done by a histopathol-
ogist blinded to the study.
Biomechanical Compression Testing
Each construct was kept frozen until biomechanical
testing. Each part was tested separately for compression using
a Texture Analyzer TA-XT Plus (Texture Technologies,
Scarsdale, NY), and corresponding parts were compared,
referred to as tip, middle, and basis. Unconstrained uniaxial
compression was applied while compressive force and dis-
placement were recorded after the probe tip contacted the
sample. Experiments were run until the 35-kg limit of the
load cell was reached. The geometry of the samples made
measuring material moduli difficult, so we compared samples
by referring to one sample that required higher force to
compress as being “stiffer” than a sample that required less
force to compress the same distance. Native bone served as
control. Mean values for the entire phalanx of each construct
were calculated for each sample from tip, middle, and basis.
RESULTS
In Vitro Assessment
CD 133� cell selection was confirmed with FITC
labeling under fluorescent microscopy (Fig. 3). After 6
weeks’ culture, light microscopic analysis of CD 133� spec-
imens showed cellular growth, with confluence of cell layers
within collagen I and fibrin glue hydrogels; staining with
H&E confirmed viability of the new tissue. Toluidine blue,
alkaline phosphatase, and von Kossa staining of the speci-
mens confirmed bonelike tissue formation (results not
shown). Controls did not show bonelike tissue formation.
Radiologic Analysis
All VCT scanning showed densities above 250.0
Hounsfield units in CD 133� cellular �-TCP-PCL scaffolds,
comparable to native mouse spine and native human hand CT
in parts of the cellular scaffolds (Fig. 4). The highest densities
were found in the tips of the cellular scaffolds. CD 133�
cellular specimens of collagen I and fibrin glue showed
similar density values. The second-highest values were in the
basis (results not shown).
Gross Clinical and Histologic Analysis
Histologically new bonelike tissue surrounded the
�-TCP/PCL scaffolds integrating into the material, demon-
strated by toluidine blue and alkaline phosphatase staining.
Cells embedded into the matrix had large round morphology,
with basophile-staining plasma. More bonelike tissue was
present in the cellular collagen I than cellular fibrin glue
specimens (Fig. 5). Staining for calcium was positive in both
specimens; however, more positively stained tissue was
present in the collagen I specimen. CD 133�-depleted spec-
imens did not show bonelike tissue formation. Acellular
controls stained negative for bonelike tissue (not shown). The
observed matrix was amorphous, without formation of lamel-
lae like in the cellular specimens of collagen I or fibrin glue.
Biomechanical Compression Testing
The basis of each specimen had the highest stiffness
compared with all other samples. However, cellular CD 133�
collagen I specimens had the highest compression values for the
tips. For collagen I CD 133� specimens, a statistically higher
stiffness was observed than in CD 133� fibrin glue specimens.
CD 133� fibrin glue specimens showed slightly higher stiffness
than acellular controls; however, the difference was not statisti-
cally significant (P � 0.05). Control specimens had a stiffness
value similar to that of the cellular fibrin glue specimen (Fig. 6).
CD 133�-depleted specimens showed similar stiffness to con-
trol acellular scaffolds.
Statistical Correlation Analysis
Higher HU corresponded to increased biomechanical
strength of the scaffolds (Spearman correlation � 0.80, Pear-
son � 0.72, Kendall � 0.667). This association was robust
across the different correlation coefficients. Bivariate corre-
lations were tested nonparametrically using Spearman rank
correlation and Kendall � b statistics. Spearman rank corre-
lations showed biomechanical stiffness: HU � 0.87 (P �
FIGURE 3. FITC-labeled CD 133� cell directly after magneti-
cally sorting. Fluorescent microscopy, magnification 200�.
Annals of Plastic Surgery  Volume 59, Number 1, July 2007 Human Shaped Thumb Bone Tissue
© 2007 Lippincott Williams & Wilkins 49

0.06). Kendall � b statistics revealed biomechanical stiffness:
HU � 0.77 (P � 0.06).
DISCUSSION
Successful tissue engineering involves the implantation
of living cells with synthetic scaffolding and usually results in
the generation of new tissue. We have tissue engineered a
distal phalanx of a human thumb, using the novel 3DP
technique. This revolutionary technique of rapid processing
allows the construction of forms in any desired shape as long
as there is a CT picture available. We chose the distal phalanx
for the first engineering purposes since our laboratory already
has experience.13,30 In contrary to previous work, we were
able to print an exactly detailed anatomic scaffold by using
the high-resolution technique of the new VCT scanner to
create a detailed picture. Details of each scaffold were con-
firmed by VCT scanning before application of the hydrogel/
cell mix.
Tissue-engineered bone has many advantages over au-
tologous or cadaveric bone graft or synthetic materials. Bone
FIGURE 4. a, In vivo VCT scan of the implanted cellular scaffold after 15 weeks. The shadow of the implanted distal phalanx
scaffold is visible. b, Hounsfield density measurement. In the phalanx densities close to native mouse, vertebrae can be ob-
served (circles). c, Bone densities of a human hand are similar to the tissue-engineered phalanx.
FIGURE 5. Clinical appearance of
explanted cellular collagen I speci-
men after 15 weeks (a); histologic
evaluation of the same specimen
(b). A layer of bonelike tissue sur-
rounds the scaffold material. Stain-
ing H&E (b), toluidine blue (c), and
von Kossa (d); magnification 40�
and 200�.
Weinand et al Annals of Plastic Surgery  Volume 59, Number 1, July 2007
© 2007 Lippincott Williams & Wilkins50

autografts that are not anchored to adjacent bone are typically
resorbed31 and thus are not an effective long-term treatment.
We did not observe any resorption of the construct within 15
weeks, confirmed by VCT scanning. Furthermore, harvesting
of bone autografts often results in donor-site morbidity.12
Like bone autografts, cadaveric bone grafts are usually
resorbed, and they carry the risk of disease transmission32
from the donor tissue. This can be avoided by using autolo-
gous hMSCs.
By using hMSC, we used the high proliferation ability
and multilineage of the cells17 to create new human bonelike
tissue. We sorted from hMSC population a subpopulation,
CD 133� cells, because these hematopoietic progenitor cells
are known to participate in the formation of bone and even
vascularized bone.18,33 New bonelike tissue formation was
observed in vitro and in vivo in the CD 133� specimens
when suspended in collagen I and fibrin glue hydrogel in vitro
and in vivo, indicating that this subpopulation of hMSCs is
involved in the formation of bone. However, we did not
observe the formation of vascularized bone but the formation
of islands of cortical bone only. This might be due to the fact
that the number of CD 133� cells in a hMSC population is
small (0.3%).18,33
Hydrogels provide a 3-dimensional structure for embed-
ded cells and are used in tissue-engineering bone.21,22,34,35 They
also provide a substrate for suspended cells and support uniform
cellular distribution into scaffold material. We demonstrated this
in all cellular specimens. After 15 weeks, newly formed bone-
like tissue most resembling native cortical bone was in the
cellular collagen I specimen. Lesser new bonelike tissue was
found in the cellular fibrin glue specimen. This might be due to
the fact that collagen I is an essential component of native bone
and its abundance in hydrogels surrounding differentiated
hMSCs proved to be beneficial for bone formation.
A 15-week end point was chosen for radiologic and
biomechanical stiffness evaluation since bone fracture heal-
ing is normally completed at this time point and the newly
formed matrix bone is radiologically visible. Higher radio-
logic densities were found in cellular CD 133� specimens
throughout the scaffold, in concordance with bonelike tissue
formation. Higher density values were found in collagen I
specimens, indicating that collagen I, naturally abundant in
bone, as hydrogel might be a better choice than fibrin glue for
tissue-engineering bone.
PCL is known to be only osteoconductive and can be
used in 3DP, thus allowing customized scaffold formation for
bone defects that need to be bridged. PCL has a degradation
rate of 2 years.36,37 Therefore, newly formed bonelike tissue
cannot easily replace the scaffold material within 15 weeks.
However, biomechanical stiffness was about three fourths of
that of human bone after 15 weeks’ time. Although cellular
specimen stiffness was higher than in acellular specimens, the
observed difference was minimal in fibrin glue samples. This
was in concordance with VCT Hounsfield values in histologic
findings. Radiologic densities of CD 133� specimens were
higher than acellular controls, conforming with histologic
formation of new bonelike tissue.
We demonstrated that higher HU corresponds to in-
creased biomechanical strength of the scaffolds (correlation
between 0.8). This association was robust across the different
correlation coefficients. This finding, in combination with our
histologic evaluation, demonstrates that an increase in HU is
a true marker of osteoblastic activity. In all specimens, one
could visually appreciate the change in HU in the VCT
images when the bone growth was successful. The imaging
results were concordant with biomechanical testing and his-
topathological findings.
This study demonstrates that hMSC CD 133� cells in
hydrogels in combination with 3DP �-TCP/PCL scaffolds
support bone formation in vitro. The best bonelike tissue was
formed using collagen I hydrogel, confirmed by histologic
and radiologic examination, and by the biomechanical stiff-
ness. However, the chosen material �-TCP/PCL is not ideal,
because of its long degradation rate of 2 years. Although only
2 samples of each hydrogel were tested biomechanically, the
quality of new bone was confirmed by histologic results and
radiologic findings. We conclude that the combination of
differentiated 133� hMSCs in collagen hydrogel with 3DP
�-TCP/PCL scaffolds can be used to engineer tissue that
meets individual needs for bone reconstruction. Future work
will concentrate on evaluation of different scaffold materials and
hydrogels for bone formation. Patients who have suffered bone
loss could potentially profit from tissue-engineered bone.
ACKNOWLEDGMENTS
The authors thank Dr. Andrew A. Freiberg, Chief of
Arthroplastic Surgery in the Orthopaedic Department, Mas-
sachusetts General Hospital, for generously providing human
mesenchymal stem cells. We would like to thank Dr. Frederik
S. Shapiro, Children’s Hospital, Boston, for the histopatho-
logical evaluation of the samples.
REFERENCES
1. Arakaki A, Tsai TM. Thumb replantation: survival factors and re-
exploration in 122 cases. J Hand Surg. 1993;18b:152–156.
2. Chiu HJ, Shieh SH, Hsu HY. Multivariate analysis of factors influencing
the functional recovery after finger replantation or revascularization.
Microsurgery. 1995;16:713–717.
FIGURE 6. Biomechanical compression testing of �-TCP/PCL
scaffolds. Stiffness of the collagen I cellular specimens is
higher than the stiffness of the controls. Results for CD
133�-depleted scaffolds not included, as they were similar
to the control scaffold (without hydrogel and cells).
Annals of Plastic Surgery  Volume 59, Number 1, July 2007 Human Shaped Thumb Bone Tissue
© 2007 Lippincott Williams & Wilkins 51

3. Weiland AJ, Raskin KB. Philosophy of replantation 1976–1990. Micro-
surgery. 1990;11:223–228.
4. Benisse L. Amputation. In: Benisse L, ed. The Acute Hand. Stamford,
CT: Appleton & Lange; 1999:287–326.
5. Schlenker JD, Kleinert HE, Tsai TM. Methods and results of replanta-
tion following traumatic amputation of the thumb in sixty-four patients.
J Hand Surg. 1980;5:63–69.
6. Ward WA, Tsai TM, Breidenbach W. Per primam thumb replantation for
all patients with traumatic amputations. Clin Orthop. 1991;266:90–95.
7. Foucher G, Merle M, Merneaud M, et al. Microsurgical free partial toe
transfer in hand reconstruction: a report of 12 cases. Plast Reconstr Surg.
1980;65:616–627.
8. Buncke HJ Jr, Daniller AI, Schultz WP, et al. The fate of autogenous
whole joint transplanted by microvascular anastomosis. Plast Reconstr
Surg. 1967;39:333–341.
9. Tsai TM, Jupiter JB, Kutz JE, et al. Vascularized autogenous whole joint
transfer in the hand: a clinical study. J Hand Surg. 1982;7:335–342.
10. Wray RC, Mathes SM, Young VL, et al. Free vascularized whole-joint
transplant with ununited epiphyses. Plast Reconstr Surg. 1981;67:519–525.
11. Mankin HJ, Mow VC, Buckwalter JA, et al. Form and function of articular
cartilage. In: Simons SR, ed. Orthopedic Basic Science. 1994:18.
12. Eufinger H, Leppanen H. Iliac crest donor site morbidity following open
and closed methods for bone harvest for alveolar cleft osteoplasty.
J Craniomaxillofac Surg. 2000;28:31–38.
13. Vacanti CA, Bonassar LJ, Vacanti MP, et al. Replacement of an avulsed
phalanx with tissue-engineered bone. N Engl J Med. 2001;344:1511–1514.
14. Lorenz HP, Hedrick MH, Chang J, et al. The impact of biomolecular
medicine and tissue engineering on plastic surgery in the 21st century.
Plast Reconstr Surg. 2000;105:2467–2481.
15. Cima LG, Vacanti JP, Vacanti C, et al. Tissue engineering by cell
transplantation using degradable polymer substrates. J Biomech Eng.
1991;113:143–151.
16. Kim WS, Vacanti JP, Cima LG, et al. Cartilage engineered in predeter-
mined shape employing cell transplantation on synthetic biodegradable
polymers. Plast Reconstr Surg. 1994;94:233–237.
17. Pittenger MF, Mackay AM, Beck SC, et al. Multilineage potential of
adult human mesenchymal stem cells. Science. 1999;1:143–147.
18. National Institutes of Health. Stem cell information �database online�.
Available at: http://stemcells.nih.gov/info/scireport/appendixE.asp. Ac-
cessed June 23, 2005.
19. Meinel L, Karageorgiou V, Fajardo R, et al. Bone tissue engineering
using human mesenchymal stem cells: effects of scaffold material and
medium flow. Ann Biomed Eng. 2004;32:112–122.
20. Kreke MR, Goldstein AS. Hydrodynamic shear stimulates osteocalcin
expression but not proliferation of bone marrow stromal cells. Tissue
Eng. 2004;10:780–788.
21. Weinand C, Pomerantseva I, Neville CM, et al. Hydrogel beta-TCP
scaffolds and stem cells for tissue engineering bone. Bone. 2006;38:555–
563.
22. Weinand C, Gupta R, Huang AY, et al. Comparison of hydrogels in the
in vivo formation of tissue engineered bone using mesenchymal stem
cells and beta-tricalcium phosphate. Tissue Eng 2007;1 �Epub ahead of
print�.
23. Hinz P, Wolf E, Schwesinger G, et al. A new resorbable bone void filler
in trauma: early clinical experience and histologic evaluation. Orthope-
dics. 2002;25:s597–s600.
24. Dean D, Topham NS, Meneghetti SC, et al. Poly(propylene fumarate)
and poly(DL-lactic-co-glycolic acid) as scaffold materials for solid and
foam-coated composite tissue-engineered constructs for cranial recon-
struction. Tissue Eng. 2003;9:495–504.
25. Lieb E, Tessmar J, Hacker M, et al. Poly(D,L-lactic acid)-poly(ethylene
glycol)-monomethyl ether diblock copolymers control adhesion and
osteoblastic differentiation of marrow stromal cells. Tissue Eng. 2003;
9:71–84.
26. Dhert WJ, Klein CP, Jansen JA, et al. A histological and histomorpho-
metrical investigation of fluorapatite, magnesiumwhitlockite, and hy-
droxylapatite plasma-sprayed coating in goats. J Biomed Mater Res.
1993;27:127–138.
27. Witte F, Kaese V, Meyer-Lindenberg A, et al. Degrading magnesium
implants increase periosteal and endosteal bone formation. ORS Paper
Nr. 0256, ORS meeting in San Francisco, 2004.
28. Sherwood JK, Riley SL, Palazzolo R, et al. A three-dimensional osteo-
chondral composite scaffold for articular cartilage repair. Biomaterials.
2002;23:4739–4751.
29. Silverman RP, Passaretti D, Huang W, et al. Injectable tissue-engineered
cartilage using fibrin glue polymer. Plast Reconstr Surg. 1999;103:
1809–1818.
30. Isogai N, Landis W, Kim TH, et al. Formation of phalanges and small
joints by tissue engineering. J Bone J Surg. 1999;81A:306–316.
31. Buck-Gramcko D. The role of non-vascularized toe phalanx transplan-
tation. Hand Clin. 1990;6:643–659.
32. Tomford WW. Transmission of disease through transplantation of mus-
culoskeletal allografts. J Bone J Surg. 1995;77:1742–1754.
33. Jay KE, Rouleau A, Underhill TM, et al. Identification of a novel
population of human cord blood cells with hematopoietic and chondro-
cytic potential. Cell Research. 2004;14:268–282.
34. Rutherford RB, Gu K, Racenis P, et al. Early events: the in vitro
conversion of BMP transduced fibroblasts to chondroblasts. Connect
Tissue Res. 2003;44(suppl 1):117–123.
35. Rutherford RB, Moalli Franceschi RT, Wang D, et al. Bone morphoge-
netic protein-transduced human fibroblasts convert to osteoblasts and
form bone in vivo. Tissue Eng. 2002;8:441–452.
36. Schantz JT, Teoh SH, Lim TC, et al. Repair of calvarial defects with
customized tissue-engineered bone grafts, I: evaluation of osteogenesis
in a three-dimensional culture system. Tissue Eng. 2003;9(suppl 1):
S113–S126.
37. Schantz JT, Hutmacher DW, Lam CX, et al. Repair of calvarial defects
with customized tissue-engineered bone grafts, II: evaluation of
cellular efficiency and efficacy in vivo. Tissue Eng. 2003;9(suppl
1):S127–S139.
OPEN DISCUSSION
Joseph E. Losee, MD (Pittsburgh, PA): Where did you get your
stem cells? And how did you know they were stem cells?
Dr. Weinand: We sorted them from the femoral heads.
We crushed these, and then we directly attached magnetic
beads to these stem cells. We drove them into the lineages of
bone, cartilage, and fat.
David Low, MD (Philadelphia, PA): A number of years
ago there was a great deal of excitement that made its way
into the lay press when chondrocytes were seeded into an ear
framework, which was then implanted into the back of nude
mice, and it looked like a perfect ear. Now you don’t hear
much about that. My presumption was that it didn’t stand the
test of time, and it actually degraded. Does your bone main-
tain its shape so that you really can implant it?
Dr. Weinand: So far it maintains its shape until 15
weeks. However, I don’t think we are there yet in implanting
it. There is always a critical size defect, and for a critical size
defect, you need to have vascularized bone, and that is what
the next step would be.
Weinand et al Annals of Plastic Surgery  Volume 59, Number 1, July 2007
© 2007 Lippincott Williams & Wilkins52