Effects of the local mechanical environment on vertebrate tissue differentiation during repair: does repair recapitulate development?
Dennis M. Cullinane1,3,*,
Kristy T. Salisbury1,3,
Yaser Alkhiary2,
Solomon Eisenberg3,
Louis Gerstenfeld1 and
Thomas A. Einhorn1
1 Orthopaedic Research Laboratory, Department of Orthopaedic Surgery, Boston
University Medical Center, 715 Albany Street, Housman-205, Boston, MA
02118-2526, USA
2 Department of Restorative Sciences and Biomaterials, Boston University
School of Dental Medicine, Boston, MA 02118, USA
3 Department of Biomedical Engineering, Boston University, Boston, MA 02115,
USA
View larger version (74K):
[in a new window]
|
Fig. 1. A finite element model of the defect. Cortical bone is represented as an
ideal tube of appropriate thickness, while the defect is represented as a
mid-segment of the tube. The mechanical properties of the defect tissues are
taken from the literature for callus mechanics, while the cortical bone is
modeled as an incompressible solid. The model incorporates geometry,
mechanical properties and load characteristics and generates stress and strain
distribution fields that are used to create tissue differentiation
predictions.
|
|
View larger version (65K):
[in a new window]
|
Fig. 2. An external fixator mounted on a rat femur with an AutoCad representation
of the device. (A) The pin clamping screws can be seen facing out from the
animal. The device is in the straight and locked position, maintaining rigid
fixation within the defect. The healed surgical incision site can be seen
below the fixator. (B) The bicortex pins are situated in the pin channels
(green arrows) and are fixed by clamp screws (yellow arrows). The black arrow
indicates the axis of the bending fixator. When the locking screws are in
place (white arrows), the device is capable of rigid fixation.
|
|
View larger version (98K):
[in a new window]
|
Fig. 3. The linkage system connecting the motor and torque sensor to the fixator,
which is inserted by pins into the rat femur. As the wheel (bottom right)
rotates, the horizontal actuator arm (bottom) drives the vertical accuator arm
(bottom left), which is attached to one side of the fixator (hidden by the
plate). As the fixator bends on its axis or displaces in shear, the defect is
stimulated. The rat is lying in a sling hammock with its tail protruding to
the left of the image.
|
|
View larger version (50K):
[in a new window]
|
Fig. 4. An example Fourier transform of collagen fibril architecture. The panels
represent the stages of Fourier analysis including a thresholded image,
ellipse and frequency distribution. The transform identifies the predominant
pattern angle within an image (in this case, fibrillar orientation). The
predominant angle, identified by highest frequency, is relative to the
original captured image's orientation, but in the final analysis the angle is
relative to a presumed defect midline, perpendicular to the bone long
axis.
|
|
View larger version (58K):
[in a new window]
|
Fig. 5. Finite element models (FEMs) illustrating strain distribution within the
defect for shear (A,B) and bending (C,D). The three-dimensional models are to
the left (A,C), while their respective cross-section representations are to
the right (B,D). The models presented are from an intermediate stage of
loading to illustrate the strain progression. The distribution of strain is
illustrated using the quantitative color bars to the right of each active
model. Altering the mechanical properties of the defect tissues alters the
results of the models. The number of brick elements and the composition of the
structure also play a role in determining the model's effectiveness in
estimating local mechanical loads.
|
|
View larger version (60K):
[in a new window]
|
Fig. 6. Graphic stress- and strain-based tissue prediction diagrams created from
finite element results for (A) 12° bending and (B) shear. The areas in
green represent putative cartilage, the areas in red represent fibrous
tissues, and the blue areas represent bone. The bending model predicts two
opposing bone elements but does not predict that the cartilage element will
completely segment between the proximal and distal halves. The shear model
predicts cartilage segmentation between the halves.
|
|
View larger version (22K):
[in a new window]
|
Fig. 7. Moment resistance data for bending, shear and combination loading. The
bending (dotted line), shear (solid line) and combination (dashed line) data
correlate highs and lows very well, except for an initial lag in the
combination group. Tissue appearance seems constrained by the timing of
physiological processes and mechanobiological principles, while the
architecture and maintenance of tissues seems to be controlled primarily by
the mechanical environment. Note how the timing of the peaks and troughs
correlates well among the three treatment regimens, while the magnitude of the
moment resistance varies by treatment.
|
|
View larger version (28K):
[in a new window]
|
Fig. 8. Radiographs of (A) control, (B) bending and (C) shear specimens with the
external fixators attached. In every case, the mechanically stimulated defects
resulted in non-union. The cartilage tissues in the experimental treatment
defects are represented by translucencies in the gaps between the segments.
All experimental treatments were for a 35-day (six-week) duration. The control
specimen example is from a four-week control specimen, demonstrating the rapid
bony bridging occurring in the controls.
|
|
View larger version (65K):
[in a new window]
|
Fig. 9. An illustrative example of the mechanobiological paradigm's predictive
value. Here, the magnitude of strain graphically dictates the differentiation
of cartilage (shown in red) versus fibrous tissue (shown in blue)
within the mechanically stimulated defects. The defect shown in A underwent
10% cortex diameter shear, whereas the defect shown in B underwent 25% shear.
Thus, a threshold exists between these shear magnitudes that determines
cartilage versus fibrous tissue outcomes.
|
|
View larger version (118K):
[in a new window]
|
Fig. 10. The histological results from (A) control, (B) bending, (C) bending and
shear, and (D) shear stimulations. Note that the control exhibits very little
cartilage, while the treatment groups all present cartilage bands (shown in
red) spanning the defect. Note also the arched nature of the cartilage band in
the bending specimen, a further mechanobiological response to the bending
action.
|
|
View larger version (45K):
[in a new window]
|
Fig. 11. Fourier transformation of control cartilage collagen fibril orientation.
The control cartilage collagen is randomly oriented due to a small-magnitude,
unpredictable loading environment. Very little cartilage is produced within
the control specimens due to rigid fixation.
|
|
View larger version (86K):
[in a new window]
|
Fig. 12. In situ hybridization analysis of type II collagen mRNA expression
in control murine femoral joint tissues and in mechanically induced tissues.
(A) 10x magnification analysis comparing light- and dark-field images of
in situ hybridization profiles of a normal murine joint and the
underlying epiphyseal plate. Note the intense primary expression of type II
collagen in the epiphyseal plate but not in the area of mature joint cartilage
collagen. The black arrowheads in A and B indicate expression of Type II
collagen. (B) 20x magnification analysis of the cartilage tissue
produced by controlled mechanical loading, forming an artificial joint-like
structure. Collagen type II expression is seen at low levels throughout the
tissue but shows higher levels of expression in a band of cells adjacent to
the area of fibrous tissue where cartilage cavitation is beginning to occur.
(C) The 40x magnification shows silver grain localization over cartilage
cells of native cartilage within the epiphysis and (D) over chondrocytes
within mechanically generated tissues.
|
|
View larger version (97K):
[in a new window]
|
Fig. 13. Immunohistologically stained sections of (A) bending-stimulated and (B)
control cartilages. The brown stain in the mechanically stimulated cartilage
indicates the expression of growth and differentiating factor 5 (GDF-5),
whereas the control section demonstrates no reaction. The positive reaction to
GDF-5 indicates that the mechanical stimulation therapy is causing the
expression of this joint-determining gene.
|
|
© The Company of Biologists Ltd 2003
