Animal Surgery
Female New Zealand White rabbits were used under the guidelines of the
Institutional Animal Care and Use Committee and with the approval of our
institutional review board. The rabbits had an average age (and standard
deviation) of 7.5 ± 1.2 months and an average weight of 2.5 ±
0.5 kg. A group of seven rabbits that did not undergo an operation provided
the source of bone marrow-derived mesenchymal stem cells and Achilles tendons
in order to provide normal values. A second group of fifty-seven rabbits were
used for the experiment proper, which involved biomechanical testing at three
weeks (twelve rabbits), six weeks (sixteen rabbits), and twelve weeks (five
rabbits) as well as cell tracing, histological analysis, and morphometric
analysis at one week (four rabbits), three weeks (eight rabbits), six weeks
(seven rabbits), and twelve weeks (five rabbits). Both hindlimbs underwent
surgical repair, with one side serving as a control for the other.
Surgical Technique
The animals were anesthetized with a combination of an intramuscular and
subcutaneous ketamine-xylazine cocktail (at a ratio of 1:1 and a concentration
of 0.8 mg/kg) as well as inhalational 1% halothane. With use of an aseptic
technique, a longitudinal skin incision was made directly over the Achilles
tendon. The gastrocnemius tendon was dissected free from the plantaris and
soleus tendons. A complete transverse laceration was made with a surgical
blade through the midsubstance of the Achilles tendon. This laceration was
immediately repaired with a modified Kessler suture with use of Prolene 4/0
suture (Ethicon, Somerville, New Jersey) reinforced with a running
epitendinous Prolene 5/0 suture. The procedure was then repeated for the other
limb. On the basis of a randomization table, each limb was then allocated to
receive either bone marrow-derived mesenchymal stem cell treatment or control
treatment. For the experimental limb, the bone marrow-derived mesenchymal stem
cell preparation (4 million cells per repair) was pelleted down and mixed with
fibrin sealant for implantation. The fibrin sealant (Baxter AG, Vienna,
Austria) was prepared from a two-component sealant according to the
manufacturer's protocol. Briefly, thrombin was dissolved in calcium chloride
solution (40 mmol/L) to yield solution A. Subsequently, Tisseel was dissolved
in Aprotinin solution (3000 KIU/mL) with use of a heated stirring device
(Fibrinotherm; Baxter AG) to yield solution B. For the bone marrow-derived
mesenchymal stem cell group, 100 µL of solution A was mixed with the
pelleted bone marrow-derived mesenchymal stem cells prior to usage. A
Duploject System (Baxter AG) was used to simultaneously apply equal amounts
(100 µL) of the two solutions. The fibrin sealant was injected
intratendinously at the repair site as well as externally around the repair
site. The same procedure was done on the control side, except that the bone
marrow-derived mesenchymal stem cells were omitted from the fibrin sealant.
Skin closure was performed in a routine manner with use of Vicryl sutures
(Ethicon). The rabbits were not immobilized postoperatively and were fed ad
libitum.
Bone Marrow-Derived Mesenchymal Stem Cell Isolation and
Expansion
Bone marrow-derived mesenchymal stem cells were obtained from anesthetized
rabbits by means of bone marrow aspiration from the iliac crest, and
processing was performed as previously
described15. The
aspirate was mixed with bone marrow-derived mesenchymal stem cell growth
medium (ratio, 1:2) consisting of Dulbecco's modified Eagle medium (DMEM;
Sigma, St Louis, Missouri), 1% penicillin-streptomycin (Sigma) and 10% fetal
bovine serum (HyClone, Logan, Utah). Samples were washed twice with medium and
were centrifuged at 2000 rpm for five minutes. The supernatant was discarded,
and the resulting cell pellet was resuspended with medium to make up 15 mL and
was subsequently plated onto T-75 culture flasks. Cells were grown in an
incubator at 37°C, humidified with 5% CO2. After five days, the
contents of the flask were removed and washed with medium, leaving behind bone
marrow-derived mesenchymal stem cells that adhered to the bottom of the flask.
Once confluent, the bone marrow-derived mesenchymal stem cells were detached
and serially subcultured. Second passage cells were used for implantation.
Cells isolated with use of this technique have a fibroblast-like appearance
and have been shown to be capable of multipotent
differentiation16-18.
Cytoplasmic Cell Labeling
In repairs that were assigned to histological examination, cells were
labeled with a fluorescent, amine-reactive carboxyfluorescein succinimidyl
ester (Vybrant CFDA SE Cell Tracer; Molecular Probes, Eugene, Oregon)
according to the manufacturer's instructions. The succinimidyl ester group
reacts with intracellular amines, forming fluorescent conjugates. The cells
were washed in PBS, resuspended in a 10-µM CFDA-PBS solution, and incubated
for ten minutes at 37°C. The cell suspension was then pelleted and
resuspended in complete medium for thirty minutes at 37°C to ensure
complete modification of the probe. The cells were then counted and washed
with PBS twice to make up a final yield of four million cells for
implantation. The same number of cells was used for each repair.
Histological Analysis
The rabbits were killed at one, three, six, and twelve weeks for analysis.
After the time of death, the animals that had been assigned to biomechanical
assessment had amputation of the lower limb at the level of the middle of the
femur and the specimens were stored at —80°C until the time of
analysis. For specimens undergoing histological examination, the repaired
tendon was dissected free, divided proximally at the musculotendinous junction
and distally at the calcaneal insertion, placed in liquid nitrogen
immediately, and then stored at —80°C until processing.
Tendon specimens were cryosectioned at 5 µm for cell tracing,
immunohistochemical analysis, and nuclear morphometry evaluation. Sections
that were used for fluorescent imaging of CFDA-labeled cells were not stained
with hematoxylin and eosin. These sections were examined with confocal
microscopy to assess the presence and distribution of the labeled bone
marrow-derived mesenchymal stem cells (with four specimens being examined at
each time-point of one, three, and six weeks, for a total of twelve
specimens).
Matrix organization was assessed by means of immunohistochemical staining
of collagen I and III fibers (with seven specimens being examined in both the
treatment and control groups at three weeks, five specimens being examined in
both groups at six weeks, and three specimens being examined in both groups at
twelve weeks, for a total of thirty specimens). Monoclonal primary antibodies
for collagen I and III (Sigma) were used in dilutions of 1:250. For detection,
the streptavidinbiotin (Lab Vision, Fremont, California) method was coupled
with a DAB chromogen system (Lab Vision). Negative control staining was
performed by omitting the primary antibody during the protocol. Positive
control staining was obtained with use of normal tendon (collagen I) and
healing tendon (collagen
III)19.
Cellular counts and organization were quantified on the basis of a
previously described method to assess tendon organization (with eight
specimens being examined in both groups at three weeks, seven specimens being
examined in both groups at six weeks, and five specimens being examined in
both groups at twelve weeks, for a total of forty
specimens)20,21.
Briefly, randomly chosen hematoxylin and eosin-stained sections were digitally
captured at the repair site with use of an Olympus IX70 microscope (Olympus
America, Melville, New York). Microscopic and nuclear measurements were made
with use of the MicroImage v4.5 imaging software (Olympus America). Images
were first enhanced with use of the display range to create distinct nuclear
shapes. The software then calculated the major and minor diameter of each
shape in the field as well as the angle between the major diameter and the
horizontal plane. From these readings, the nuclear aspect ratio (defined as
the ratio of the minor diameter to the major diameter) and the nuclear
orientation angle (defined as the angle between the major axis of the nucleus
and the longitudinal axis of the tissue) were obtained.
Figure 1 illustrates the
nuclear aspect ratio and nuclear orientation angle. Nuclear aspect ratio
values that approach 0 suggest a flatter shape, whereas a value of 1
represents a perfect circle. Nuclear aspect ratio results were grouped in bins
of 0.3. Nuclear orientation angle values of 0° represent a nucleus that is
perfectly aligned toward the longitudinal axis. As the value increases, the
nucleus becomes more angled until it approaches 90°, where it lies
perpendicular to the long axis of the tendon. Nuclear orientation angle
results were grouped in bins of 30°. Each bin is represented as the
percentage of nuclei in the bin over the total number of nuclei in the field.
Normal values for the nuclear orientation angle, the nuclear aspect ratio,
cellularity, and cross-sectional area were established by measuring the values
in five uninjured Achilles tendons.
Biomechanical Testing
Tendons used for biomechanical testing were harvested at the calcaneus and
musculotendinous junction (with twelve specimens being harvested in both
groups at three weeks, sixteen specimens being harvested in both groups at six
weeks, and five specimens being harvested in both groups at twelve weeks, for
a total of sixty-six specimens). Measurements of the repair site were made
with use of a digital vernier caliper (Mitutoyo, Aurora, Illinois). The
calcaneal portion of the specimen was placed in a customized jig that
consisted of a metal cage with an adjustable pin to restrain the calcaneus.
The proximal end of the tendon was clamped directly with use of a coarse-grade
sandpaper interface to prevent slippage. The specimen was mounted onto a
mechanical testing machine (model 3343; Instron, Norwood, Massachusetts), and
a preload force of 2 N was applied to provide initial tension. Once in
tension, the tendons were marked with black silk sutures transversely, 5 mm
away from both the clamp and the jig. These markers were used to analyze the
surface strain with use of a noncontacting video extensometer (Instron). The
tendons were then loaded to failure at a constant speed of 10 mm/min. Data
collected from Merlin v5.31 software (Instron) were used to compute the
material and structural properties of each specimen.
Statistical Analysis
The number of animals used in the study was decided a priori after
consultation with our institution's biostatistician. The study has 90% power
with use of a two-sided test of 5% to detect a 5, 15 and 47-MPa difference in
modulus at three, six, and twelve weeks. These numbers were chosen after a
review of relevant
literature10 and
our expectation that cell therapy would likely improve the rate of healing
rather than increase the final healing strength. For the nuclear orientation
angle and nuclear aspect ratio, the results for corresponding groups of bins
were compared across treatments and repair times with use of analysis of
variance with the level of alpha set at 0.05. The biomechanical properties of
the tendon were analyzed across treatments and repair times with use of
analysis of variance with the level of alpha set at 0.05. Pairwise comparisons
were performed with use of two-tailed paired t tests. Statistical analysis was
performed with use of SPSS software (SPSS v12.0; SPSS, Chicago, Illinois).
Gross Findings
There were no macroscopic differences between the two groups of tendons and
no repair ruptures. At three weeks, the fibrin glue around the repair site had
almost completely degraded. The repair site was hypertrophic and appeared
semitranslucent. At six weeks, there was no trace of fibrin around the repair
site and the repair site was less thick. By twelve weeks, the repair site
appeared continuous with the proximal and distal ends of the tendon in terms
of thickness and opacity. Compared with the normal tendon, the cross-sectional
area of the repair site was increased at three weeks in both the treatment and
control groups (p < 0.05) (Fig.
2). The difference between the two groups was not significant. The
cross-sectional area in both groups decreased to almost normal values at six
weeks and remained similar at twelve weeks.
Cell Tracing of Implanted Bone Marrow-Derived Mesenchymal Stem
Cells
Fluorescent labeling showed that bone marrow-derived mesenchymal stem cells
remained viable at the repair site for as long as six weeks. Staining was most
intense but was concentrated at the site of introduction at one week. At three
and six weeks, the cells began to migrate around the repair site but not
proximally or distally into the normal tendon.
Histology and Immunohistochemistry
Hematoxylin and eosin stains revealed fibrosis and the presence of fibrin
in the midsubstance of the repair site in both groups in the early stage of
healing. Inflammatory responses in both groups were similar and were
characterized by the aggregation of lymphocytes and macrophages mainly around
the suture material. The distribution of inflammatory cells was similar in
both groups, and we did not observe an apparent immune response such as
lymphocyte infiltration elicited by the allogeneic cells. There were no traces
of fibrin, and inflammatory responses were markedly diminished at six weeks
and were minimal at twelve weeks. Cellularity was increased at the repair
site. At three weeks, the cell count at the repair site showed a 205% cellular
increase in the treatment group as compared with a 233% increase in the
control group. By twelve weeks, cell counts dropped to a 128% increase in the
treatment group and to a 137% increase in the control group (p < 0.05)
(Fig. 3). In both groups,
collagen I fiber staining at three weeks was visually assessed to be less
intense and organized around the repair site as compared with the proximal and
distal ends. However, the treatment group exhibited denser collagen I
structures that appeared more organized in comparison with those in the
control group (Fig. 4). At six
and twelve weeks, specimens from both the treatment group and the control
group were similar histo-logically and appeared to have more intense staining
and more organized collagen fibers at subsequent time-points. Collagen III,
which is normally present at the paratenon, was found at the repair site and
was diminished proximal and distal to the repair site in both groups. Collagen
III staining was most obvious at three weeks and diminished by twelve
weeks.
Morphometric Analysis
There was significant improvement in nuclear morphometric findings (the
nuclear aspect ratio and nuclear orientation angle) across the time-periods in
both groups (p < 0.05, analysis of variance); these improvements correlated
with tendon-healing and maturation (Figs.
5-A and 5-B). In addition, at three weeks, the percentage of cells
with a nuclear aspect ratio value of <0.3 and the percentage of cells with
a nuclear orientation angle value of <30° were both significantly
higher in the treatment group than in the control group (36.3% compared with
29.1%, respectively, for the mean nuclear aspect ratio and 85.2% compared with
73.9%, respectively, for the mean nuclear orientation angle, p < 0.05). At
six weeks, there was no significant difference between the two groups. By
twelve weeks, the values of both the nuclear aspect ratio and the nuclear
orientation angle were approaching normal.
Biomechanical Testing
The structural properties of both groups of tendons improved at each
successive time-point (Figs. 6-A and
6-B). At three weeks, the treatment group had a 32% higher modulus
compared with the control group (88.9 MPa compared with 67.2 MPa, p <
0.05). However, at six and twelve weeks, there was no significant difference
in material properties between the treatment and control groups.
Our results indicate that bone marrow-derived mesenchymal stem cells
implanted within the tendon-repair site remain viable and contribute to early
tendon-healing following primary repair, but, with the number of samples
available in the present study, no significant changes were identified at
twelve weeks.
The fluorescent dye labeling studies showed clear evidence of cell
viability for at least six weeks. Comparison of the studies that were
performed at one, three, and six weeks revealed that the cells become more
diffusely spread out. The dye is passed on to daughter cells with cell
division, and the fluorescence decreases with time. Labeled cells have been
detected for as long as eight weeks in
vivo22. It is for
this reason that our time-points for tracer studies were chosen. In a study
involving cells that were isolated with use of the same technique, viable
allogeneic bone marrow-derived mesenchymal stem cells were demonstrated at
eight weeks after implantation in a rabbit patellar tendon
defect23.
The better morphometric parameters and modulus at three weeks in the
treatment group suggest that bone marrow-derived mesenchymal stem cells
improve the rate of tendon-healing and maturation. These findings parallel
data from other studies that have shown that bone marrow-derived mesenchymal
stem cell-seeded tissue-engineered substitutes can improve the biomechanical
properties compared with non-seeded scaffolds in tendon
defects8,9,20.
The present study did not demonstrate any difference in outcome at twelve
weeks. This finding contrasts with the findings of previous studies of tendon
defects that have demonstrated differences in biomechanical properties at
twelve weeks9.
However, we cannot conclusively state that the bone marrow-derived mesenchymal
stem cells have no effect on final tendon strength, as our hypothesis was
based on the rate of tendon-healing rather than final strength. For this
reason, our sample sizes, and thus the power of the study, were chosen to
reflect this hypothesis. The numbers that were chosen for analysis at twelve
weeks could be insufficient to detect a real difference at that
time-point.
The exact role of implanted bone marrow-derived mesenchymal stem cells on
tendon-healing remains uncertain. One possibility is that they become
differentiated into tenocytes within the healing tendon environment and
participate in healing through collagen production and remodeling.
Alternatively, it has been suggested that bone marrow-derived mesenchymal stem
cells may contribute to healing by acting as "growth factor pumps"
rather than through terminal
differentiation24.
Our finding of viability of these cells intratendinously (as demonstrated by
CFDA labeling) supports either of these possibilities.
The New Zealand White rabbit is outbred. We chose to use allogeneic bone
marrow-derived mesenchymal stem cells for a number of reasons. Allogeneic
cells have previously been used successfully in tissue-engineering
applications8,9.
We did not observe any apparent lymphocytic immune response from the
allogeneic bone marrow-derived mesenchymal stem cells in histological
analysis, and there is convincing evidence that allogeneic bone marrow-derived
mesenchymal stem cells can modulate host immune
responses25-27.
In vitro and in vivo studies have shown that mesenchymal stem cells avoid
normal alloresponses through mechanisms such as hypoimmunogenicity and the
prevention of normal T-cell
responses28,29.
A recent review of mesenchymal stem cell transplant immunology is
available30. Our
findings are consistent with the findings of those studies. The
immunohistochemistry and morphometric results at three weeks provide evidence
that the improvement in tendon material properties is due to accelerated
healing rather than inflammation or scar formation.
However, one study showed that allogeneic BMP-12-engineered bone
marrow-derived mesenchymal stem cells could enhance bone-healing in a
critical-sized bone defect model only if short-course immunosuppression was
given31. This
difference in findings between the various studies cannot be explained at the
present time and may be related to differences in the choice of animal model,
the type of tissue studied, or the experimental protocol.
The clinical importance of the improvement in biomechanical properties at
three weeks also needs to be addressed. This improvement may need to be
optimized by increasing the number of bone marrow-derived mesenchymal stem
cells implanted or the addition of growth factors or may even be enhanced with
gene therapy.
Fibrin was chosen as a cell carrier because it has been used clinically for
nerve coaptation and hemostasis. The bone marrow-derived mesenchymal stem
cells remain viable in the fibrin, which then resorbs during the healing
process. We have thus used it as a delivery vehicle for the bone
marrow-derived mesenchymal stem cells. Published research indicates that the
addition of fibrin does not affect the tensile strength and histological
characteristics in repaired rabbit Achilles
tendons32.
Our experimental protocol allowed the rabbits to move freely following
surgery. Numerous studies on tendon-healing have demonstrated that early
mobilization enhances
tendon-healing1,33,34.
In our pilot studies, the rabbits were immobilized in a plaster cast after
tendon repair. This removed the positive effect of mobilization on the healing
process, with the aim of allowing the effect of bone marrow-derived
mesenchymal stem cells to predominate. However, because of problems with skin
wounds and mortality following immobilization, we changed the protocol to
allow free movement following repair. Surrounding tendons provided some
stress-shielding of the repair, but, nevertheless, the tendon was subjected to
motion.
Our findings suggest that using allogeneic bone marrow-derived mesenchymal
stem cells would be a means of providing an "off-the-shelf" cell
therapy solution for tendon
surgery35.?