Bone morphogenetic proteins (BMPs) have a powerful capacity
to elicit new-bone formation1,2.
There are several methods of treating segmental bone defects with
BMPs3-9, one of which is gene
therapy to deliver BMPs. Retrovirus, adenovirus, and adeno-associated
virus have been utilized experimentally to deliver genes including
BMPs in vitro and in vivo10-13. However, the direct use of these
vectors has several disadvantages. For example, retroviral vectors require
actively dividing cells for integration of the foreign gene, and
they incorporate randomly into the host DNA, possibly transforming
proto-oncogenes into oncogenes11.
The adenovirus may induce a host immune response that decreases
the efficiency of infection14,15.
In addition, many of these vectors are cytotoxic, and their systemic
effects on the host are largely unknown13,16.
One way to circumvent these disadvantages is to use ex
vivo gene therapy techniques7,9,17-19.
This method involves the isolation and cultivation of host cells,
transduction of the cells in vitro, and implantation
of these cells into the skeletal defect7. Ex
vivo gene therapy increases the efficiency of transduction
(thus increasing the expression of the desired protein), and it
allows the clinician to screen cells for tumorigenicity prior to
implantation. Furthermore, the use of autologous cells decreases
the possibility of immune rejection.
Another advantage of ex vivo gene therapy is
that the clinician can choose the type of cells to be transduced.
Any one of several types of cells that are easily isolated from
a patient can be used as a delivery vehicle for the BMP. The presence
of stem cells in bone marrow20-22 and
skeletal muscle20,23-27 suggests
that these cells might be used as a vehicle for BMP delivery. These
stem cells have the capacity to differentiate into osteogenic cells
upon stimulation with BMP1,3,26.
Thus, these cells can act as carriers of BMP as well as participate
in new-bone formation by differentiating into osteogenic cells.
Bone-marrow stromal cells have recently been used in ex
vivo gene therapy to elicit healing of segmental bone defects
in syngeneic rats7. Skeletal muscle
is also an abundant source of stem cells3,9,20,23-28.
A subpopulation of muscle-derived cells has been shown to differentiate
into osteogenic cells in vitro and in vivo upon
stimulation with BMP-23,29. Recent
studies have identified a population of cells in skeletal muscle
that have the same surface markers as bone-marrow-derived mesenchymal stem
cells20,29. Muscle-derived cells
are easy to obtain with muscle biopsy and are simple to culture
or cryopreserve3,26,30-32.
In this study, we sought to determine whether primary cultures
of muscle-derived cells from mice can deliver recombinant human
BMP-2. Using a critical-sized skull-defect model in mice, we showed
that primary cultures of muscle-derived cells, genetically engineered
to express BMP-2, enhanced bone-healing. In addition, we showed that
a subpopulation of these cells differentiated into osteogenic cells.
All methods involving animals were approved by the animal use
committee at the Children’s Hospital of Pittsburgh.
Isolation of Primary Cultures of Muscle-Derived Cells
Muscle from hindlimbs of six to eight-week-old normal male mice
(C57BL10J+/+; Jackson Laboratories, Bar
Harbor, Maine) was dissected free of bone and minced. The minced
muscle was digested by serial one-hour incubations at 37°C in 0.2% type-XI
collagenase, dispase (grade II, 240 U), and 0.1% trypsin.
The cell suspension was passed through 18, 20, and 22-gauge needles,
centrifuged at 3000 rpm for five minutes, resuspended in growth
medium (Dulbecco modified Eagle medium supplemented with 10% fetal
bovine serum, 10% horse serum, 0.5% chick embryo
extract, and 2% penicillin/streptomycin), and
plated in collagen-coated flasks. During the initial seven days,
floating cells in the medium were centrifuged at 3000 rpm for five
minutes and replated into the same culture flask with fresh medium.
Cells that did not adhere to the flask after seven days were discarded.
This method of cell isolation yielded a mixed population of 70% to
80% myofibroblasts and 20% to 30% myogenic
cells as analyzed by desmin-staining31,32.
Adenovirus Construct Encoding Recombinant Human
BMP-2 Gene
The BMP-2-125 plasmid that contains the recombinant human BMP-2
cDNA was provided by Genetics Institute, Cambridge, Massachusetts.
A replication-defective, E1 and E3-gene-deleted adenoviral vector
was engineered to encode BMP-2 under the human cytomegalovirus promoter.
The BMP-2-125 plasmid was digested with Sal I, resulting in a 1237-base-pair
fragment containing the BMP-2 cDNA. The BMP-2 cDNA was then inserted
into the Sal I site of the pAd.lox plasmid, which placed the gene
under a human cytomegalovirus promoter. Recombinant adenovirus was obtained
by co-transfection of pAd.lox with psi-5 viral DNA into CRE-8 cells.
The adenovirus-BMP-2 construct (adBMP-2) was stored at —80°C
until it was used.
Infection of Muscle-Derived Cells with adBMP-2
Freshly isolated muscle-derived cells were equally divided and
plated in two T-75 flasks (Fisher Scientific, Pittsburgh, Pennsylvania).
Cells in one flask were trypsinized, removed, and counted to calculate
the amount of viral particles needed. The second flask of cells
was washed with Hanks balanced salt solution. Adenovirus particles
(fifty infectious particles per cell; multiplicity of infection = 50) were
premixed into Hanks balanced salt solution and then layered onto
the cultured cells. (Multiplicity of infection equals the number
of particles used to infect one cell. For the adBMP-2, previous experiments
have demonstrated that a multiplicity of infection of 50 was optimal
for muscle-derived cells3,9,29.)
After four hours of incubation at 37°C, an equal volume of growth
medium was added, and cells were allowed to recover overnight. Cells
were then harvested, washed with Hanks balanced salt solution, and
counted. Approximately 0.5 to 1.0 ¥ 106 infected
cells were used for the skull-defect assay.
ELISA (Enzyme-Linked Immunosorbent Assay) for
Determination of Production of BMP-2 by Muscle-Derived Cells
Freshly isolated muscle-derived cells were transduced with adBMP-2
and plated in a T-25 culture flask with Optimem serum-free medium
(Gibco BRL, Grand Island, New York). The serum-free medium was necessary
to eliminate the BMP-2 present in the serum. After transduction,
cells were grown for an additional three days, and the medium was
collected for ELISA. Noninfected cells, treated in an identical
manner, were used as negative controls.
The monoclonal antibody (H3B2/17.8.1) and a biotinylated
antibody (bH4B2/5.10.24) to BMP-2 were donated by Genetics
Institute. H3B2/17.8.1 was coated onto a ninety-six-well
microtiter plate by adding 50 L/well of 8 g/mL
of antibody in the coating buffer (0.1M carbonate/bicarbonate
buffer, pH 9.6) and incubating at 2° to 8°C overnight. Wells were
then blocked by adding 200 L/well of the blocking solution
(4% w/v nonfat dry milk, 50mM Trizma base, 1mM
glycine, and 0.5M NaCl, pH 8.0) and incubating for one to two hours
at 37°C. After 1:4 dilution with the diluent (25% v/v Optimem
and 75% v/v THHST [1.5M NaCl, 50mM Trizma
base, 1mM glycine, pH 8.0]), 50 L/well of the
samples were added and were incubated at room temperature for two
to three hours. Plates were washed six times with the wash buffer
(50 mM Trizma base, 1mM glycine, 0.5M NaCl, pH 8.0, 0.05% v/v
Tween-20). The biotinylated antibody was diluted 1:300 in the wash
buffer and was added to each well (50 L) for incubation at room temperature
for 1.5 to 2.0 hours. Horseradish peroxidase conjugated to avidin
(Sigma Chemical, St. Louis, Missouri) was diluted 1:10,000 in buffer
and was added to each well for a one-hour incubation at room temperature.
Detection of bound BMP-2 was accomplished by adding 100 L/well
of the TMB solution (3,3¢,5,5¢-tetramethylbenzidine;
Kirkegaard and Perry Laboratories, Gaithersburg, Maryland) and incubating
for four to eight minutes at room temperature. Color reaction was
stopped with 100 L/well of 0.18M H2SO4 solution and read
at 490 nm on a plate reader. All samples were assayed in triplicate,
and known concentrations of BMP-2 were used as standards in each
plate.
Skull-Defect Assay
For the skull-defect assay, we used female SCID (severe combined
immunodeficiency strain) mice (Jackson Laboratories), which are
immunodeficient mice unable to mount a sufficient immune response.
Because we were employing human BMP-2, use of immunodeficient mice
was necessary to avoid an immune response to the human protein as
a variable in our model. The mice were anesthetized with methoxyflurane
and were placed prone on the operating table. With use of a number-10
blade, the scalp was dissected to the skull and the periosteum was
stripped. A 5-mm-diameter full-thickness circular skull defect,
which is a nonhealing critical-sized defect6,
was created at the apex of the skull with use of a dental burr,
with minimal penetration of the dura.
The mice were divided into four groups (Table I), and a collagen
sponge (Helistat; Colla-Tec, Plainsboro, New Jersey)5 cut to form-fit the defect was used
as a matrix. Group 1 did not receive any implant in the skull defect,
Group 2 received collagen sponge only, Group 3 received collagen
sponge seeded with muscle-derived cells, and Group 4 received collagen
sponge seeded with muscle-derived cells transduced with adBMP-2.
The scalp was closed with use of a 4-0 nylon suture, and the animals were
allowed food and activity ad libitum. At two and
four weeks, the mice were killed and the skull specimens were dissected
free from the soft tissue for digital imaging. The percent closure of
the defect was calculated with use of the NIH Image program. The
area of the original defect was calculated with a 5-mm circular
standard on the digital image, and the area of the closed defect
was drawn and calculated with the NIH Image program. The area of
the defect filled with new bone divided by the area of the original
defect yielded the fraction of skull-defect closure.
Skull specimens were then flash-frozen in 2-methylbutane (buffered
in phosphate-buffered saline solution and precooled in liquid nitrogen).
Frozen samples were cut into 5 to 10-m sections with use of a cryostat
(Microm HM 505 E; Fisher Scientific) and stored at -20°C until they
were used.
For von Kossa staining, slides were fixed in 4% formaldehyde
and soaked in 0.1M AgNO3 solution for fifteen minutes. After exposure
to light for at least fifteen minutes, the slides were washed with phosphate-buffered
saline solution and stained with hematoxylin and eosin for viewing.
Fluorescent in Situ Hybridization (FISH) and Osteocalcin
Immunohistochemistry
Cryosections were fixed for ten minutes in 3:1 methanol/glacial
acetic acid (v:v), air-dried, and then denatured in 70% formamide
2 ¥ SSC (saline-sodium citrate; 0.3M NaCl, 0.03M sodium
citrate, pH 7.0), at 70°C for two minutes. Slides were immediately
dehydrated with a series of ethanol washes (70%, 80%,
and 95%) for two minutes at each successive concentration.
The Y-chromosome-specific probe33 was
biotinylated with use of a BioNick kit (Gibco BRL) according to
the manufacturer’s instructions. The biotinylated probe
was then purified with use of a G-50 Quick Spin Column (Boerhinger-Mannheim,
Indianapolis, Indiana), and the purified probe was lyophilized along
with 5 ng/mL of sonicated herring sperm DNA. The probe
was resuspended in 50% formamide, 1 ¥ SSC (0.15M
NaC1 and 0.015M sodium citrate), 10% dextran sulfate solution
prior to hybridization. After denaturation at 75°C for ten minutes,
the probe was placed on denatured sections and allowed to hybridize
overnight at 37°C. After hybridization, sections were rinsed with
2 ¥ SSC solution, pH 7.0, at 72°C for five minutes. Sections
were then rinsed in BMS solution (0.1M NaHCO3, 0.5M NaCl, and 0.5% NP-40,
pH 8.0). The hybridized probe was detected with fluorescein-labeled
avidin (Oncor, Gaithersburg, Maryland). Nuclei were counterstained
with 10 ng/mL of ethidium bromide or DAPI (4¢,6¢-diamidino-2-phenylindole)
in Vectashield mounting medium (Vector, Burlingame, California).
For osteocalcin staining, the primary antibody was goat anti-mouse
osteocalcin (1:100 in phosphate-buffered saline solution; Chemicon,
Temecula, California), and labeled cells were visualized with anti-goat
antibody conjugated to Cy3.
Statistical Methods
Data are presented as the mean and standard deviation of the
mean. Statistical differences between groups were calculated with
use of a Student t test and a two-way analysis of variance with post
hoc tests.
We demonstrated that genetically engineered muscle-derived cells
can produce BMP-2 and can substantially enhance the healing of a
critical-sized bone defect. The Y-probe and osteocalcin analysis suggest
that a fraction of muscle-derived cells is incorporated into the
newly formed bone, differentiating in vivo into
osteogenic cells. The observation that the muscle-derived cells
alone were not capable of improving bone healing suggests that the
osteogenic stimuli (BMP-2) are a major determinant to differentiate
the muscle cells toward osteogenic lineage.
The existence of stem cells in the bone marrow has been well
documented20-22,28. Recent reports
have indicated that these stem cells can be used in experimental
models to treat clinical diseases. For example, Gussoni et al.20 used bone-marrow-transplantation
techniques to restore dystrophin expression in mdx mice, an animal
model of Duchenne muscular dystrophy. Also, Lieberman et al.7 recently reported repair of segmental
femoral defects in rats with use of bone-marrow stromal cells genetically
engineered to produce BMP-2.
However, the use of skeletal-muscle stem cells that can differentiate
into osteogenic cells remains relatively unexplored. Katagiri et
al.26 first reported that a subpopulation
of mouse muscle-derived cells can become osteogenic when exposed
to BMP-2. In their study, an immortalized cell line from muscle
showed increased expression of alkaline phosphatase, osteocalcin,
and parathyroid-dependent 3¢,5¢-cAMP in response
to BMP-2 in vitro26.
Recently, we identified a subset of muscle-derived cells that expresses
higher levels of alkaline phosphatase in response to BMP-2 than
do other fractions. This suggested that some muscle-derived cells
differentiate into osteogenic cells when exposed to BMP-2 in
vivo3. We also reported
isolation of a clone from the BMP-2-responsive population that expresses
several putative stem-cell markers29.
This clone expressed markers Flk-1, a mouse homologue of KDR, and
Sca-1, which were shown to be expressed by hematopoietic stem cells29,34. In addition, in vitro experiments
showed that our clone was able to differentiate into osteogenic
cells expressing alkaline phosphatase and osteocalcin in response
to BMP-2. In vivo, the clonal population was able
to enhance skull-defect closure as well as to differentiate into
osteocytes29. Results of other
experiments demonstrate that, when BMP-2 is delivered with fibroblasts
or chondrocytes, the cells do not seem to express alkaline phosphatase
in response to BMP-29, are uniformly
found outside of the bone matrix, and do not differentiate into
osteogenic cells3,9. Therefore,
muscle-derived cells, like the bone-marrow stromal cells, have the
ability to act as a vehicle for BMP-2 delivery as well as to differentiate
into osteocytes and participate in new-bone formation.
Skeletal muscle represents an abundant source of easily accessible
tissue. A simple surgical biopsy specimen can provide enough cells
to use in ex vivo gene therapy. Thus, we envision
a system in a clinical setting where a muscle biopsy specimen is obtained
from the patient, muscle cells are cultured and transduced in a
laboratory setting, and BMP-2-producing cells are reimplanted into
the bone defect within a week’s time. Our results suggest
that the number of stem cells with potential to differentiate into
osteogenic cells in muscle is relatively small (about 5%).
However, the remaining 95% of cells can act as a delivery
vehicle for BMP-2, recruiting host osteogenic cells to promote bone-healing.
The existence of muscle-derived stem cells and the abundance of
tissue make skeletal muscle an attractive tool for tissue-engineering
and gene therapy.
Another table, showing the total number of Y-probe-positive cells
per high-power field and the number and percentage in new bone,
is available with the electronic versions of this article, on our web
site (www.jbjs.org) and on our CD-ROM (call 781-449-9780, ext. 140,
to order).