Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Magnetic vector potential along z direction
B
Magnetic force
f
ext
External volume force
H
Magnetization vector
n1 Outward normal direction of material 1
n2 Outward normal direction of material 2
xiii
P
(3.1)
The first equation is also named as Amperes Circuital Law:
t
D
J H
c
c
+ = V
(3.2)
Equation (3.2) states that the curl of the magnetic field intensity H
due to the flow of charges and the time derivative of the electric flux
density D
magnetization vector P
magnetic force J
permeability
c
permittivity
o
conductivity
e
angular frequency
d thickness in z direction density
T stress tensor p air pressure
39
In this study case, the electromagnet is supposed to work under low frequency, less than
50Hz. The wavelength for a 50Hz signal is 6000km. On the other hand, the structure size of the
magnets is around 50x50mm, which means the length of the geometry is only a very small
fraction of the exciting wavelength. In this slow variation case, the electromagnetic field is
assumed to be a quasi-static field with
0 = c c t D
+ =
, equation (3.3) for quasi-static electromagnetic field can be extended to:
e
J B v E H
+ + = V ) ( o
(3.4)
Combining the constitutive relation,
H M H B
r
0 0
) ( = + =
, the definition of the
magnetic potential A B V =
, and electrical potential
t
A
V E
c
c
V =
, Maxwell-Amperes law
can be rewritten as:
e
J V A v M A
t
A
= V + V V V +
c
c
o o o ) ( ) (
1
0
(3.5)
Since the exciting current is a harmonic signal, this study case is a time-harmonic case. In
the time-harmonic case, Amperes law includes a displacement current
D je
, with the
constitutive relation
P E D + =
0
c
, equation (3.5) can be rewritten as following:
P j J V j A v M A A j
e
e ec o o c e eo + = V + + V V V +
) ( ) ( ) ( ) (
0
1
0 0
2
(3.6)
40
3.2.2 Electromagnetic Model Analysis
Since our objective is to study the field distribution and the flux density B of a quasi-
static magnetic field, we can ignore the coupling between the electric and magnetic field, and
only consider the induction currents that are relevant to the magnetic potential. Moreover, as we
are mainly interested in the B distribution on a horizontal surface (x-y direction), we can study
the magnetic field in two dimensions to simplify the modeling process. In the two-dimensional
case, with only the consideration of induction current, equation (3.6) can be simplified as:
e
z z z r
J L V A v A
+ A = V V V
/ ) ( ) (
1 1
0
o o (3.7)
Equation (3.7) is utilized in the electromagnet subdomain analysis [89]. To perform the
simulation, we make the following assumptions:
- The current source is a time-harmonic sinusoidal current with magnitude of 0.4A and
frequency of 1Hz.
- The entire coil is considered as a whole block with a constant external current density
corresponding to the current in each single wire. This assumption is more efficient than
modeling each turn of the coil as a separate wire.
- The eddy current between the turns in the coil is negligible.
- The material of the core is a solid soft iron, ignoring the laminations among the sheets
of the iron core
Based on the above assumptions, we choose the 2D, time-harmonic, quasi-static,
perpendicular induction current, and vector potential application mode in the COMSOL AC/DC
module. The complete set of geometrical and electrical parameters used for the simulation is
listed in Table 3.2 and 3.3. The meshed geometry of the electromagnet model is shown in Fig.
41
3.1. In the graph, there are three subdomains: air, coil (copper) and core (soft iron). The bobbins
between the coil and the core are ignored since they do not significantly influence the magnetic
field analysis. The air space surrounding the feet of the magnet is 30mm, just half the size of a
60mm standard culture dish. Therefore, the graph represents half of a stretch sample; the other
side of the stretch sample is symmetrical. The geometrical dimensions and material properties
of the magnet were provided by the product company. The mesh generator partitions the
subdomains into triangular elements. The finite element model has 2716 elements and the
number of degrees of freedom is 5491. The size and density of the elements are allocated based
on the geometrical dimensions. The areas with more boundaries have smaller elements with
higher density, while the outside areas have larger elements with lower density. This allocation
can reduce the computational expenses and facilitate the calculation procedure.
Table 3.2 Geometrical Parameters
Parameters Expression Description
r_coil 0.2mm Radius of the coil wire
N 550 Turns of the winding
Ww 4mm Width of the winding
Lw 28mm Length of the winding
Wc 12.5mm Width of the core
Lc 38mm Length of the core
w 1mm Width between the core and winding
42
Figure 3.1 Meshed Electromagnet Model
(a)
43
(b)
(c)
Figure 3.2 Distribution of Magnetic Flux Density
(a) Bx component, (b) By component, (c) Bn component
44
Table 3.3 Electrical Parameters
Parameters Expression Description
I_coil 0.4A Current of the coil
J_coil I_coil*N/(Ww*Lw) External current density
o
5.99e7S/m Conductivity of the copper wire
As the exciting source is a time-harmonic current, the magnetic field distribution of the
study space varies with time. The magnetic poles of the electromagnet alternate between the
two feet during a period. Fig. 3.2 (a), (b), and (c) exhibits the x component, y component and
normal distribution of the maximum flux density in a period at phase zero, respectively. The
units for x and y coordinates are meter. Since the magnetic force along the x direction serves as
the stretch force, the x component of flux density, denoted as Bx, is the main study object. In the
Fig. 3.2 (a), the red arrows act as magnetic field lines representing the magnetic field
distribution. The magnetic south pole (orange) of the horseshoe electromagnet is at the lower
foot and the magnetic north pole (blue) is at the upper foot. The distribution shows that the
magnetic flux density Bx has greater magnitude in the soft iron core than in the air space. The
maximum Bx in the soft iron is 0.204T at the inner corners. However the amplitude of Bx
dramatically reduces to 0.041T at the boundary of the iron core and air space. Except for the
two feet areas, the flux density B in most air space is near zero (green). For example, the
magnitude of the Bx is only 0.015T at 5mm distance of the edge, 0.012T at 10mm distance, and
0.009T at 15mm distance.
45
Figure 3.3 Magnetic Flux Density Bx along y Direction
Since the magnitude of the Bx determines the magnitude of the magnetic force along the x
direction, the effective area of the non-contact magnetic force is limited. In order to apply non-
contact magnetic force, the magnetic rods should be placed within 10mm from the edge of the
electromagnet. Fig. 3.3 shows the distribution of Bx along the y direction (16mm long) at 5mm
distance from the edge of the magnet feet. The plot demonstrates that the magnitude of Bx at
that location is very small, with a maximum value of 0.015T. However, It also shows that the
absolute value of the density Bx is almost symmetrical along the y direction symmetrical axial
at y=0.151, which means if a 16mm long magnetisable rod is put into the field along y
direction, the two sides of the rod will experience symmetric force. This plot will be compared
with Fig. 4.9 in Chapter 4 section 4.6.3. Besides being affected by the flux density, the
magnitude of the force is also determined by the distance between the magnet and the
magnetisable material and the materials property, which will be discussed in Chapter 4.
46
Although we focus our investigation on the distribution of magnetic flux in the x direction,
we still need to study the normal magnetic flux density B because it is the closest parameter
(2D) to the results of the physical test (3D). For the 2D model, the normal magnetic flux density
B was defined as the total magnitude of the X component and the Y component. The distribution
of the normal magnetic flux density is shown in Fig. 3.2 (c). The simulation results demonstrate
that the normal magnetic flux density B is 0.082T directly beside of the horseshoe magnet. This
result will be compared with the experimental results in the next section.
3.3 Experimental Verification of the COMSOL Model
Serial physical experiments were implemented to verify the COMSOL model of the
horseshoe-shaped electromagnet. In the experiments, the magnet was excited by a pulsed 0.4A
current train with 1Hz frequency, and 50% duty cycle. A Gauss meter with integral Hall Effect
Probe (Model 2010, Magnetic Instruments Inc.) was utilized to test the magnetic flux density B
[91]. The sample time for each test point is one minute. The verification and simulation results
are compared from the aspects of distance, exciting frequency and current.
3.3.1 Flux Density vs. Distance
The first experiment is conducted to verify the relationship between the flux density B and
distance. The corresponding density B was recorded along the x direction every 5mm from each
magnets foot. Fig. 3.4 (a) and (b) represent the density B along x direction, at y=0.0143 (south
pole) and y=0.0159 (north pole), respectively. The red curves in the plots are the simulation
results of B
norm
from the COMSOL model. The blue curves are the testing results of flux density
B, which were obtained from the average of twenty peak magnitudes during the sample time.
47
The y axis of Fig. 3.4 shows the absolute magnitude of B for at the south pole (a) and north pole
(b); while the x axis shows the distance from the test points to the edges of the electromagnet.
The graphs demonstrate that the south pole and the north pole possess the same trends.
The trend of B in these plots shows good correlation between the test and simulation results.
Moreover, the field distribution coincides with the Bio-Savart law. According to the Bio-Savart
law, the magnetic flux density produced at a point from an element of length l d
of a
filamentary wire carrying a steady current I is expressed as:
2
0
4 R
a l Id
B d
R
t
= (3.8)
where is an element of length in the direction of the current, is the unit vector
pointing from to the point, and R is the distance between the point and the current element
. The equation shows that flux density B is proportional to the current I and the reciprocal of
the distance square [90]. However, most of the testing results are, on average, higher than the
simulated results in the ambient area (distance<0.015m) of the magnet except for the zero
position, and lower that the simulated results in the remote area. This phenomenon will be
discussed in the next section.
l d
l d
R
a
l d
48
SimulationvsTestingResults
(SouthPolary=0.159)
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0 0.005 0.01 0.015 0.02 0.025 0.03
Distance[m]
M
a
g
n
e
t
i
c
F
l
u
x
D
e
n
s
i
t
y
[
T
]
si mul ati on
testi ng
(a)
SimulationvsTestingResults
(NorthPolary=0.143)
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0 0.005 0.01 0.015 0.02 0.025 0.03
Distance[m]
M
a
g
n
e
t
i
c
F
l
u
x
D
e
n
s
i
t
y
[
T
]
si mul ati on
testi ng
(b)
Figure 3.4 Simulation Results vs. Testing Results of Magnetic Flux Density
(a) South Pole at y=0.0159, (b) North Pole at y=0.0143
49
3.3.2 Flux Density vs. Exciting Frequency
In order to testify the relationship between flux density and the exciting frequency, the
probe of the Gauss meter was fixed at the lower corner of the north pole (around y=0.0159)
0.005m away from the magnet. The flux density was recorded as the average of twenty peak
magnitude during the sample time, with pulse train frequencies from 1Hz to 5Hz. The test
results show that the flux density ranges from 372.0G to 386.5G, with an average of 375.8G,
and a standard deviation of 5.63 (Table 3.4). The results demonstrate that frequency variations
in the lower range (1~5Hz) have little effect on the magnitude of flux density, which coincides
with the simulation assumption that the eddy current between the wires in the coil is negligible.
Due to the assumption of eddy current, the simulation results show that exciting frequency does
not affect the flux density.
Table3.4 Magnetic Flux Density vs. Exciting Frequency
Frequency
[Hz]
1 2 3 4 5
Flux Density
[G]
386.5 374.7 374.5 372.0 377.0
STD DEV 5.63
50
3.3.3 Flux Density vs. Exciting Current
The relationship between the magnitude of the exciting current and the flux density B was
also investigated in this section. In the interface of the apparatus (Fig. 2.10), only the stretch
related parameters are presented to the users, such as stretch frequency, duty cycle, and strength.
However, all of these mechanical parameters are fulfilled by defining related electrical
parameters. For example, the magnitude of stretch strength is directly controlled by exciting
current. When the stretch strength is set as 95%, the magnitude of the exciting current is 0.4A,
while when the stretch strength is set as 50%, the current is 0.2A.
Exciting Current vs Flux Density
0
50
100
150
200
250
300
350
400
10 20 30 40 50 60 70 80 90
Exciting Current (%)
M
a
g
n
e
t
i
c
F
l
u
x
D
e
n
s
i
t
y
[
G
]
Figure 3.5 Exciting Current vs. Flux Density
In order to study the relationship between the exciting current and the flux density, the probe
of the Gauss meter was fixed at the lower corner of the north pole at the distance of 0.005m
from the magnet. The flux density B was recorded when the stretch strength increases from10%
51
to 90% by every 10%. Fig. 3.5 shows the test results. It illustrates that the flux density decreases
linearly as the strength of the exciting current decreases. The standard deviation for each test
point is very small, ranging from 0.09 to 1.40. However, the simulation results demonstrate that
flux density is proportional to the exciting current. The test results agree with the simulation
results that the flux density B is proportional to the exciting current I in the low frequency
range. Therefore, the test results verified the design principle that the magnetic flux density and
force can be controlled by controlling the exciting current.
3.3.4 Comparisons of Simulation and Experiment Results
A series of physical experiments are implemented to test the relationships between the
flux density B and the distance, the exciting current and frequency. Although the experiment
results agree with the simulation results, there are still some differences. These differences will
be discussed in the following:
- Waveforms
For testing, the exciting current is a pulse train with 1Hz frequency, 50% duty cycle, and
0.4A amplitude. For simulation, we used a time-harmonic sinusoidal current with the same
frequency, duty cycle, and amplitude, but with different shapes. In experiments, the magnetic
field generates when the pulse train switches from OFF time to ON time, and disappears when
the pulse train switches from ON time to OFF time, but the magnetic polarities of the
electromagnet stay the same during the electrifying time. However, for modeling, as the
exciting current is a time-harmonic sinusoidal signal, the magnetic field exists in the testing
space during the whole period, while the magnetic polarities of the magnet alternate during
52
every half period. Although the waveforms of the simulation and the testing are different, the
maximum amplitudes of the two waveforms are the same. Therefore, the simulation results
shown in Fig. 3.2 (the maximum flux density in a period) can represent the real distribution.
- Dimensions
When we take a close look at Fig 3.3 (a) and (b), we can see although the testing results
and the simulation results share the same trend, there are three clear differences: (1) the testing
result is lower than the simulation result immediately adjacent to the magnet; (2) except for that
first point, the testing results are slightly higher than the simulation results close to the magnet
(distance<0.015m); (3) the testing results are slightly lower than the simulation results further
away from the magnet (0.015m<distance<0.03m).
The first difference may be caused by the thickness of the test probe. Since the probe of
the Gauss meter has a thickness, the first testing result is not the real density B at zero distance
(the boundary of between the core and the air). There is a small gap between the electromagnet
and the test probe. This may be the reason why the first testing result is lower that the
simulation result as density B decreases dramatically outside of the core. For the second
difference, the simulation results are obtained from a two-dimensional model, while the
experiment results are based on a real three-dimensional magnet. When the probe is positioned
at a specific place, it senses the flux density B from three dimensions (including Bz, density
along z direction); however, the simulation model did not consider this parameter. Therefore,
most of the testing results are slightly higher than the simulation results in the ambient area of
the magnet. With the increase of distance, density B radiates toward different directions, and the
difference between the simulation and testing gets smaller. For the third difference, when the
53
distance from the magnet is greater than 0.02m, the flux density is much lower than 0.01T. In
this situation, the sensitivity of the probe and the test errors greatly influence the experimental
results.
Despite the above differences, the testing results and the simulation results still agree
with each other well. Therefore, the modeling provides a useful tool to optimize the density
distribution and to calculate the non-contact magnetic force.
Figure 3.6 Magnetic Flux Density Distribution of the Stretch Device
(1-Electromagnet, 2-Shadow Area, 3-Scaffold, 4-Steel Bar)
The simulation results show that the spatial distribution of density B is non-uniform. The
ambient area of the magnet (distance<0.01m) has stronger density than other space, shown in
Fig. 3.4. We can apply this distribution feature to the designed apparatus. When a culture dish is
positioned between a pair of the modeled magnets (1), the shadowed areas in Fig. 3.6 (2) have
stronger density. If a pair of magnetic bars (4) is secured in these areas and mounted at the two
ends of an engineered tissue patch, the patch can experience a controllable stretch. Meanwhile,
54
the tissue patch (3) is exposed to a very weak magnetic field. The density B is less than 0.02T,
over a distance of 10mm (the edge of the shadow area). The uniaxial stretch apparatus described
in Chapter 2 is designed based on the above theory. The stretching force can be calculated
directly by the modeling, since the flux density has been verified by the physical experiment.
Hence, the time-varying magnetic field offers a promising approach to apply controlled
mechanical conditions to the tissue culture process via the non-contact magnetic force.
3.4 Effect on Cell Biology
The above analysis shows that magnetic field distribution along the x direction is non-
uniform. The maximum magnitude of the density B is about 0.08T, while the density reduces to
only 0.04T near where the steel bar is secured. The density B in most of the tissue culture area
(Fig. 3.6 (3)) is very low (B<0.01T). Two bio-experiments were performed to test the influence
of the magnetic field on cell biology. One experiment tests the cell proliferation and the other
investigates the cell morphology.
3.4.1 Cell Proliferation
- Cells and Cell Culture
To test the magnetic influence on the cell proliferation of engineered tissue patches, we
cultured rat smooth muscle cells on a gelatine sponge scaffold. The Gelfoam was fully soaked
in culture medium and incubated for three days before cell seeding. The mounting trays were
gas sterilized and placed in 60mm Petri dishes. The soaked scaffold (without clamps) was then
secured in the center groove of the mounting tray. Cell suspension (610
6
cells/250l culture
55
medium) was then seeded into each Gelfoam scaffold (40x20x10 mm). The cell seeding
procedure followed previously established methods [6], and the culture medium was gradually
added until the scaffold submerged.
The assembled dishes were placed at 37 in a humidified incubator with 5% CO
2
and
95% air for twenty-four hours for cell adhesion. Three samples were subsequently placed on the
newly designed bio-stretch apparatus with the modeled magnets. Since no stainless steel bars
were attached at the ends of the scaffold, the engineered tissue patches were only affected by
the time-varying magnetic field. The field was intermittently applied to the patches, alternating
between 15minutes of stretching time and 30minutes of rest time. The exposure time (15
minutes) referred previous studies [92, 93]. The frequency and strength of the exciting current
were set to be 1 Hz and 0.4A, respectively. In order to obtain a stronger magnetic field, the
strength of the current was set very high. However, the control groups were placed in the same
incubator without being exposed to the time-varying magnetic field. The culture medium was
changed every other day.
- DNA Assay
To determine the number of cells on each engineered tissue patch, the total DNA was
extracted using the DNeasy Blood&Tissue kit (Qiagen) following the manufacturers
instructions on days 3 and 14 after exposure to the magnetic field. The cell-seeded gelatine
sponge was lysed using the lysis buffer and the lysate was loaded onto the DNeasy mini spin
column. Total DNA was then eluted into TE buffer and read at 260 nm. Table 3.5 compares the
average and standard deviation of DNA at a magnetized condition with the control group from
the two time points (Day 3 and Day 14).
56
- Results
Table 3.5 shows that the control group had a slightly higher number of cells than the
magnetized samples. However, the T-test results for two groups show that there is no significant
difference among these groups because P>>0.05. The statistic power analysis was also
performed by SigmaStat software. The calculation results show that the powers of these two
groups are 1.0 and 0.99, respectively, which means there is no significant difference between
the magnetized group and the control group. Besides the analysis results, condensation was
observed on the lids of the magnetized group, which suggests that the temperature of the
magnetized group was different from that of the control group.
Table 3.5 Cell Numbers of Magnetized and Control Samples
3 Days 14 Days
Magnetized
(n=3)
Control
(n=3)
Magnetized
(n=3)
Control
(n=3)
DNA Ave
(g/ml)
56.2 68.9 69.7 72.7
DNA Std 1.36 0.93 0.33 0.57
T-Test P=0.25 P=0.47
57
3.4.2 Cell Morphology
- Cells and Cell Culture
To evaluate the magnetic influence on cell morphology, we used rat smooth muscle cells
(SMC) and mouse bone marrow stromal cells (BMSC) as cell sources. The cells were seeded
directly to a 35mm culture dish. Cell suspension with 0.210
6
cells was seeded into each dish.
After culturing 24 hours, three samples were placed into the same defined electromagnetic field
as section 3.4.1. The control dishes were placed into the same incubator without the magnetic
field. The lag period was intended to provide the cell enough time to recover from isolation
procedure and enable attachment.
- Results
We learn from the modeling that the magnetic field distribution of the electromagnet is
non-uniform. Therefore, we only observed the cells within 10mm to the magnet, where the
magnetic density is greater than 0.01T. Fig. 3.7 and Fig. 3.8 show the cell morphology of SMC
and BMSC, respectively. The images were taken by a NIKON Ti-S microscope for phase
contrast. The photographs are grouped into two columns. In Fig 3.7 columns (a) and (b),
comparison is made on the morphology of SMC after 48 hours and 72 hours; similarly, in Fig
3.8 columns (a) and (b), comparison is made on the morphology of BMSC after 48 hours and 72
hours, respectively. In each column, the control image is on the top panel while the image of the
magnetized sample is on the bottom panel. However, no significant difference in cell
morphology was observed for either cell type at the two time points by the observers who were
blinded to the study methods, which means that the time-varying magnetic field with the density
of 0.08T at zero distance has little effect on cell morphology.
58
(a) (b)
Figure 3.7 Cell Morphology of SMC
(a: 48hrs-SMC, b: 72hrs-SMC, magnification: 200x)
59
(a) (b)
Figure 3.8 Cell Morphology of BMSC
(a: 48hrs-BMSC, b: 72hrs-BMSC, magnification: 200x)
3.5 Heating Effect of the Electromagnet
When performing the bio-experiments, condensation was observed on the lids for the
magnetized group. This drew our attention to a possible side effect of heating when using an
electromagnet, a problem because temperature is an important parameter in tissue culture.
Normally, the temperature remains at 37 in the incubator. However, when we simulated the
electromagnet model, the heat dissipation problem was ignored, so that the increase of the
temperature cannot be predicted. Therefore, the variation of the temperature was tested by
experimentation.
3.5.1 Experiment Results
- Under Room Temperature
The Hall Effect Probe of the Gauss meter (Model 2010, Magnetic Instruments Inc.),
which we used to verify the flux density B, can also test temperature. When we studied the
60
relationship between flux density B and exciting frequency f, the probe was fixed at 5mm from
the edge of the magnet. The probe not only sensed the variation of the flux density, but also
recorded the variation of the temperature. Fig. 3.9 illustrates the temperature variation after the
electromagnet was excited. The electrify parameters are set as 1Hz frequency, 95% strength,
and 50% duty cycles. The plot shows that the temperature increased 1.3 after 9 minutes
exciting, and remained stable.
Temperature vs Time
25
25.5
26
26.5
27
0
:
0
0
:
0
0
0
:
0
0
:
2
7
0
:
0
1
:
3
8
0
:
0
2
:
0
6
0
:
0
3
:
1
7
0
:
0
3
:
4
4
0
:
0
5
:
0
8
0
:
0
5
:
3
5
0
:
0
6
:
0
2
0
:
0
7
:
0
8
0
:
0
7
:
3
5
0
:
0
8
:
4
6
0
:
0
9
:
1
3
0
:
0
9
:
4
0
Ti me [h:mm:ss]
T
e
m
p
e
r
a
t
u
r
e
[
]
Temperature
Figure 3.9 Temperature vs. Exciting Time under Room Temperature
- Under Incubator
The above test was performed under room temperature, while the real working
environment for the electromagnet is in the incubator. Therefore, a precision thermometer UNI-
T (UT328, UNI-TREND technology Inc., Guangdong) was utilized to test the temperature
variation under working conditions. The thermometer has two thermocouple input channels.
One thermocouple (t1) was fixed in the middle of a pair of electromagnets (the center of the
61
stretch board) to sense the temperature variation caused by the magnets; another thermocouple
(t2) was fixed at the metal shelf in the middle of the incubator to sense the temperature of the
incubator. After the test points fixed, the stretch board was electrified. The stretch plan was set
the same as in the bio-experiments in section 3.4 (electrifying 15 minutes and resting 30
minutes, 1Hz, 95% strength, 50% ON time). The thermometer was placed in the incubator for
11 hours. After three hours and forty minutes (five periods of the stretch plan) for satiability, it
started to record the two test points temperature every minute. Fig. 3.10 records seven and half
hours (10 successive periods) data.
Fig. 3.9 and Fig. 3.10 show clearly that the temperature increases when the
electromagnet is electrified. Both experiments verified that the local temperature around the
magnets is about 1.0 higher than the environment. Fig. 3.9 emphasizes the short time
increase trend under room temperature. The temperature increased 1.3.
Temperature vs Time
35.5
36
36.5
37
37.5
38
38.5
1
:
0
7
:
4
7
1
:
2
6
:
4
7
1
:
4
5
:
4
7
2
:
0
4
:
4
7
2
:
2
3
:
4
7
2
:
4
2
:
4
7
3
:
0
1
:
4
7
3
:
2
0
:
4
7
3
:
3
9
:
4
7
3
:
5
8
:
4
7
4
:
1
7
:
4
7
4
:
3
6
:
4
7
4
:
5
5
:
4
7
5
:
1
4
:
4
7
5
:
3
3
:
4
7
5
:
5
2
:
4
7
6
:
1
1
:
4
7
6
:
3
0
:
4
7
6
:
4
9
:
4
7
7
:
0
8
:
4
7
7
:
2
7
:
4
7
7
:
4
6
:
4
7
8
:
0
5
:
4
7
8
:
2
4
:
4
7
Time [hh: mm: ss]
T
e
m
p
e
r
a
t
u
r
e
[
]
t1
t2
Figure 3.10 Temperature vs. Exciting Time in incubator
62
Fig. 3.10 shows the temperature fluctuation in the incubator. Compared with the baseline
t2, t1 shows more severe fluctuation. During this test period, the temperature for t1 ranges from
37 to 37.9, with an average of 37.4; while the temperature for t2 ranges from 36.
to 36.7, with an average of 36.4. The temperature influence on cell biology was
investigated in the following section.
3.5.2 Bio-experiment Verification
As the experiment results show that the electromagnets causes the local temperature
fluctuated during electrifying, while temperature is an important culture parameter for cell
growth. The biological influence of temperature needs to be clarified.
- Cells and Cell Culture
To test the influence of temperature on the cell proliferation of engineered tissue
patches, we again used the gelatine sponge (Gelfoam: Pharmacia & Upjohn Co., Kalamazoo,
MI) as the scaffold and rat smooth muscle cells as the cell source. The experiment followed the
same cell-seeding protocol described in section 3.4.1. However, we used only half the number
of the cells. As the experiment was performed in 35 mm Petri dishes, the size of the scaffold
was only a half of the experiment in section 3.4.1. Hence, the cell suspension (310
6
cells/125l culture medium) was seeded into each Gelfoam scaffold (20x20x10 mm). The cell-
seeded patches were then assembled into 35 mm culture dishes. All the samples were first
placed into a humidified incubator with 5% CO
2
and 95% air at 37 for twenty-four hours.
63
The test and control groups were then placed into two different humidified incubators with 5%
CO
2
and 95% air for three days. The temperature for the test group was 38, while the control
groups temperature was 37.The number of samples in each group was three.
- DNA Assay
A DNA assay was again used to determine the number of cells on each engineered tissue
patch, following the same procedure as in section 3.4.1. Table 3.6 compares the average and
standard deviation of DNA at a high temperature condition (38) with the control group
(37).
Table 3.6 Cell Numbers at 38 and at 37
High Temperature 38
(n=3)
Control 37
(n=3)
DNA Ave
(g/ml)
48.2 52.0
DNA Std 2.30 5.26
T-Test P=0.32
- Results
Table 3.6 shows that the total cell number of the control group is slightly higher than the
group cultured under higher temperature. However, the T-test result for two groups shows that
there is no significant difference among these groups because P>>0.05. This result is in
64
accordance with the result obtained from section 3.4.1. The phenomenon may explain that one
degree of temperature difference does not significantly affect cell proliferation. In section 3.4.1,
the magnetized group was not only exposed to the time-varying magnetic field, but also affected
by the temperature variation. However, these two parameters also do not significantly affect the
experiment results.
3.5.3 Effect Factors
When an external electric current passes through an electromagnet, the heat conductivity
is known to be coupled with the electromagnetic field simulation. Therefore, the temperature
variation was examined.
There are several physical parameters affecting the heating problem of a pulsed
electromagnet, such as permeability , conductivity o, and permittivity . Heat is generated not
only in the coils but also from the core [65-68]. For example, when the iron core is exposed to a
pulsed magnetic field, the permeability of the material will change. The temperature
dependent parameter will introduce a phase shift between the magnetic flux density B and the
magnetic field strength H that causes the hysteresis pneumonia. Joule Heating can be called
resistance heating that is mainly heat loss of the coils. It is determined by the local current flux
and temperature-dependent conductivity. In a Joule Heating study, the heat transfer problem
couples with the electronic current problem. The calculation of the heat source is based on the
equation
2
V Q V = o
. However, the thermal energy in turn changes the electrical conductivity
of the coil. The temperature-dependent conductivity can be expressed as
)) ( 1 (
1
0 0
t t +
=
o
o , where 0
(
m O
) is a reference resistivity at a reference temperature
65
t
0
(K), and is proportionality constant (K
-1
) for the temperature dependence. The temperature-
dependent conductivity causes the variation of the exciting voltage, which results in the
coupling between the temperature and electromagnetic fields.
In future modelling, a multiple physical aspects need to be studied. The model should
consider not only magnetic field, but also heat transfer. The above effect factors should be
considered into the modelling, which is useful to accurately describe the real phenomena. To
reduce the effect of temperature fluctuation in later experiments, there are two methods can take
into account. First, reduce the temperature of the incubator 0.5 lower than the previous set
number, which can effectively reduce the average temperature of the stretch group, but the
control group will also be affected. Another way is to place the electromagnets onto a metal
base such as aluminum, which can disparate the heating and not affected by the magnetic field.
Therefore, the local temperature of the electromagnet can lower down.
3.6 Conclusions
The time-varying electromagnetic field generated by the stretch device is quantitatively
studied in this chapter, both through simulation and experimentation. The magnetic field is
modeled by COMSOL multiphysics 3.3 and verified by the Gauss meter Model 2010. The
investigation shows that the distribution of magnetic flux density along the stretch direction (X
direction) Bx is non-uniform; the magnitude of the density is proportional to the exciting current
I; and the magnitude of the density is not related to the exciting frequency at low frequency
range (1Hz to 5Hz). The relevant biological experiments demonstrate that the weak, low
frequency magnetic field generated by the stretch device does not significantly affect cell
proliferation and cell morphology. However, the difference between the exam group and the
66
control group is caused by the influence of both magnetic field and temperature field as the
electromagnet has heat dissipation problem.
67
Chapter 4
Effect of the Uniaxial Cyclic Stretch on
Cell Biology
After constructing the biostretch apparatus (Chapter 2) and studying its magnetic influence
(Chapter 3), we will investigate the mechanical stimulation provided by the device in this
chapter. We will quantitatively study the influence of the mechanical stimulation generated by
the developed apparatus from the aspects of strain distribution, strain rate, and stretch force.
Relevant bio-experiments are conducted to validate the performances of the apparatus.
This chapter is organized as follows. A brief introduction of the current state of stretch
applications is presented in section 4.1. Section 4.2 validates the apparatus performance by
examining cell proliferation and cell orientation. Section 4.3 investigates the strain distribution
on the surface of the scaffold with a nondestructive method. Section 4.4 studies the strain rate of
the apparatus with Photrons high-speed camera Ultima APX. Section 4.5 calculates the stretch
force provided by the apparatus. Section 4.6 draws conclusions from the above research.
68
4.1 Introduction
Tissue engineering shows great potential for treating diseases by providing functional cell-
based substitutes for human tissues [6, 94, 95]. In addition, engineered tissues can also serve as
in vitro tissue equivalents for drug testing and development, and even for studying the tissue
functions that can expand our understanding of tissue biology [69, 96, 97]. However, in
conventional three-dimensional tissue culture systems, engineered tissues could not develop the
similar structure as native tissues. Compared with engineered tissues, native tissues live in a
much more complicated environment in vivo. For example, connective tissues such as tendon
and ligaments experience tensile loading, while native cardiac tissue contracts under the control
of electrical pacing signals. The specific constructs of tissues are built to adapt their
environment. Currently, most engineered tissues are cultured under a relevantly static
environment. Their cell density and mechanical properties in tension and compression are
substantially lower than those of native tissues [59]. Therefore, external physical stimuli are
necessary for the culture systems to mimic various aspects of the in vivo environment.
Different types of cells are sensitive and responsive to different mechanical stimuli. As
described in Chapter 1 section 1.2.1, smooth muscle cells usually experience tensile forces [7].
Since the designed apparatus is developed for uniaxial cyclic stretch, smooth muscle cells are
chosen as the cell source for all the relevant bio-experiments. Moreover, the types of the
uniaxial biostretch apparatus are reviewed in Chapter 2 section 2.2. Based on these devices,
related bio-experiments have demonstrated that cyclic stretch is a potent stimulus, which can
improve cell proliferation, stimulate extracellular matrix production and the mechanical
properties of tissues [6, 12, 24, 24, 54, 55, 59, 62, 64, 97, 98].
69
The group of Vandeburgh [12, 54-56] studied the effects of uniaxial stretch for decades.
Their results show that repetitive stretch/relaxation for eight days increased the elasticity, mean
myofiber diameter, and myofiber area percent of human bioartifical muscles (HBAMs). In
addition, they noted that the muscle cells generated increasing internal forces during formation.
Studies also show that uniaxial cyclic stretch improved the structure and function of cultured
avian and rodent muscle [54, 99]. Eschenhagen et al. and Zimmermann et al. [16, 24]
demonstrated that engineered heart tissue (mixture of neonatal rat cardiac myocytes, collagen I
and matrix factors) displayed important hallmarks of differentiated myocardium when subjected
to unidirectional cyclic stretch. Shimko et al. [98] used murine embryonic stem cells (ESC) to
examine the direct effect of mechanical loading (uniaxial stretch) on the differentiation of ESC-
derived cardiomyocytes, and found that mechanical loading significantly affected gene
expression. Akhyari et al. [6] utilized human heart cells to demonstrate that uniaxial stretch
enhanced the formation of a three-dimensional tissue-engineered cardiac graft.
The above influences were investigated by different custom-built devices. Due to the
diversity of the experimental devices, the comparison of all the mechanical parameters provided
by the devices is difficult. The most commonly studied mechanical parameter for these devices
is strain. In our study, we employed noncontact magnetic force as the cyclic uniaxial stretch
force, and provided two different stretch methods. Compared with the previous device [66], we
can precisely controll the mechanical parameter, strain, by confining the distance between the
stoppers in the mounting tray. However, uniaxial stretch can vary with other parameters, such as
stretch pattern, strain distribution and rate, stretch force and frequency, which will be discussed
in the following sections. These parameters will all contribute to cell orientation. Nevertheless,
70
before studying these mechanical parameters, the validation of the apparatus will first be
examined by the bio-experiments.
4.2 Validation of the Uniaxial Cyclic Stretch
Although many results of strain-induced cell proliferation, matrix production, and
increased mechanical properties have been widely reported in previous literature review [13,
58], the performance of the designed apparatus still needs to be validated by the biological
experiments. Since cell proliferation has been utilized to study the influence of the magnetic
field and temperature, the same method will be used to test the influence of uniaxial stretch.
4.2.1 Cell Proliferation
- Cells and Cell culture
In the experiment, the gelatine sponge (Gelfoam: Pharmacia & Upjohn Co., Kalamazoo,
MI) was again used as the scaffold and rat smooth muscle cells served as the cell source. All the
stretch related accessories, such as clamps, steel bars, and boards, were sterilized. The
experiment followed the same cell-seeding protocol described previously. The cell suspension
for each Gelfoam scaffold (30x10x10 mm) was 310
6
cells/125l culture medium. The scaffold
and clamps were assembled in the hood with tweezers. The assembled scaffolds were then
mounted into 35mm Petri dishes, and cells were seeded on the top of the Gelfoam. The samples
were then placed at 37 in a humidified incubator with 5% CO
2
and 95% air for twenty-four
hours. Five sample dishes were subsequently placed on the newly designed bio-stretch
apparatus, and another five samples were used as controls in the same incubator. The stretching
71
group was exposed to intermittent stretch, alternating between 10 minutes of stretch time and 10
minutes of rest time for three days. The parameters of the mechanical stimulus are defined as
1Hz, 50% strength, 50% ON time and 20% strain. The culture medium was changed every day.
- DNA Assay
After three days of stretching, DNA assay was utilized to determine the number of cells
on each engineered tissue patch following the same instruction described before. Table 4.1
compares the average and standard deviation of DNA concentration for each group.
Table 4.1 Cell Numbers of Stretched and Control Samples
Uniaxial Stretched
(n=5)
Control
(n=5)
DNA Average
(g/ml)
60.0 36.2
DNA Std 5.37 4.99
T-Test 0.0009
- Results
Table 4.1 shows that the average of the DNA concentration is much higher in the
stretched group than in the control group. The DNA concentration between the stretched group
and the control group is statistically significant (p<<0.05).The results demonstrate that this
device can provide uniaxial cyclic stretch to the tissue construct, and thus influence cell
proliferation. This result was in accordance with other research groups results [6, 55, 97].
72
However, the experiment results are the response to all the characteristics of a uniaxial cyclic
stretch regime (deformation magnitude, stretch strength, stretch frequency, and even stretch
methods). The independent contributions of each mechanical parameter are unclear.
4.2.2 Cell Orientation
After studying the stretching effect on cell proliferation over a period of in four days,
the influence of stretch on cell orientation was studied for a longer term (2-weeks).
- Cells and Cell culture
In this experiment, the gelatine sponge (Gelfoam: Pharmacia & Upjohn Co., Kalamazoo, MI)
was again used as the scaffold and rat smooth muscle cells served as the cell source. The
experiment followed the same procedure described in section 4.3.1., with the only difference
being that the culture period lasted 14 days. The sample number for each group is three.
- Hematoxylin and Eosin (HE) Staining
Fourteen days after being seeded with cells, the engineered tissue patches (stretched (n=3);
control (n=3)) were carefully removed from the clamps in the culture dishes. The samples were
then fixed in 10% neutral buffered formalin for 24 hours. Using standard methods, the samples
were embedded in paraffin blocks, with care being taken to maintain their final orientation from
the apparatus.
Two paraffin sections were cut out from each sample, parallel to the surface of the samples.
HE staining was used to examine the gross tissue structure. Hematoxylin stains cell nuclei and
polyribosomes, giving nucleic acid a blue color. Eosin stains protein and cytoplasm, giving it a
73
pink color. The stain images were observed under a Nikon Eclipse 80i fluorescence microscope.
Fig. 4.1 compares the stretched and control sample under the magnification of 100x and 200x,
respectively.
(a)
(b)
Figure 4.1 Gross Structure of the Engineered Tissue
(a) stretched (left: 200x, right: 100x ); (b) control (left: 200x, right: 100x )
74
- Results
The images for both the stretched group and control group show that fourteen days later
cells still alive on the scaffold. Further, the visual observation demonstrates that the stretched
constructs have more cells than the control group. Three different images were examined for
each section from vary parts of the scaffold. The number of cells in each image (0.2mm
2
) was
counted in Adobe Photoshop CS (Adobe Systems, San Jose, California). The stretched group
and control group were averaged and compared (18 images each group) in Fig. 4.2. The
average number of cells was 194 for the stretched group, and 81 for the control group. The T-
test result shows p<<0.05, and marked as * in the graph, which means the stretched group has
significant difference compared with control group.
Cell Proliferation
0
50
100
150
200
250
C
e
l
l
N
u
m
b
e
r
a
t
0
.
2
m
m
2
stretched
control
*
-
Figure 4.2 Comparison of Cell Numbers
(* represent the significant difference between two groups)
Moreover, for the control samples, the cells were scattered among the holes of the
scaffold (Fig. 4.1(b)). For the stretch samples, the cells demonstrated inhomogeneous cell
75
distribution. The cell clusters were parallel along the stretch direction, longitudinally oriented
(Fig. 4.1(a) arrow direction). Additionally, compared with the control group, the stretched
samples exhibited a good cell orientation at the free edges and two ends. Complexes of
multicellular aggregates and longitudinally oriented cell bundles were also found at the ends
and free edge areas (Fig. 4.3).
Figure 4.3 Edge Structure of the Engineered Tissue
(a) control; (b) stretched (magnification: 100x )
4.2.3 Discussion
The results of these two bio-experiments validated the performances of the apparatus. The
device is capable to provide uniaxial cyclic stretch during tissue culture with well-defined
mechanical parameters (strain amplitude, frequency, strength, and duty cycle), which can affect
cell proliferation and cell orientation. However, the results reflected the influence of all
characteristics of a stretch regime, not a single parameter.
76
- Cell Reorganization
The HE images show that without external stimulation, cells clustered among the holes of the
scaffold randomly. However, under cyclic stretch, these tissue-like bundles are not uniformly
distributed along the surface of the scaffold. Cells aggregated at the free edges and ends of the
scaffold, while the centre part had fewer bundles. This phenomenon can be explained as
multiple stimuli existed during stretch. The scaffold was not only subjected to uniaxial stretch,
but also strain-induced fluid flow around and within the porous scaffold. Therefore, this culture
system actually induced not only cyclic stretch but also shear stimulation. Additionally, due to
Poissons effects, when the substrate undergoes one dimensional stretch, the other two
dimensions will experience compression. During a stretch/relaxation period, the medium inside
the scaffold may experience a squeeze/inhale period, which in turn improved the mass
transportation condition [17][59]. In future studies, the effect of each mechanical parameter
needs to be isolated to help researchers determine the mechatransduction of the cells.
- Variation of Mechanical Property
For the long-term culture, it was observed that the scaffolds maintain a high degree of
elasticity during the first seven days. Later, a stronger force (60%) is needed to obtain the same
strain magnitude, which indicates that the stiffness of the structure changed during culture. This
kind of variation has been observed and recorded by other groups [55, 97].
4.3 Study of Strain Distribution
From the literature review listed in table 2.1, we know that the most frequently studied
parameter of cyclic stretch is strain. The average strain provided by the device is defined
77
as
0
/ L L A = c . L
0
is the distance between two symmetric inner stoppers, shown in Fig. 2.6. For
symmetric stretch, AL is the total deformation of the scaffold, which is twice the distance from
the inner stopper to the outer stopper. For asymmetric stretch, AL is the deformation of the
movable end of the construct. Unlike an elastic rigid-body, the scaffold is a vesoelastic material.
Although the average strain of the construct is clear, the strain distribution at the surface of the
scaffold is unknown. Therefore, a nondestructive method [70][100, 101] was utilized to
investigate the full-field surface strain distribution of the three-dimensional scaffold.
4.3.1 Measurement Methodology
To study the strain distribution, samples of both Gelfoam (41 x 20 x 7 mm, n=2) and GE
RTV 6166 silicone (38 x 16 x 3 mm, n=11) were used as scaffolds. Compared with Gelfoam,
which is porous and has a heterogeneous property, the GE RTV 6166 silicone is a homogeneous
material, allowing it to serve as a baseline to compare to. The Gelfoam samples were marked
with dots at 2mm intervals using a fine-tipped permanent marker; while the RTV samples were
marked by dark microbeads. These marks separated each RTV silicone sample and Gelfoam
sample into 16 sections along the stretching axes (Fig. 4.4). The rectangular areas in Fig. 4.4 are
the analysis area. After marking, the scaffolds were affixed to stainless steel clamps with
silicone glue. Scaffolds were then soaked in water for at least 24 hours. The final dimensions of
the scaffolds with clamps were slightly different due to the gap made by the glue and shrinkage
when the scaffolds were inserted in water, ranging from 40x18x7 mm to 40.5x18x7 mm for
Gelfoam, and 40x16x3 mm to 41.5x16x3 mm for RTV silicone. The scaffolds were then placed
in the aforementioned apparatus for cyclic stretch. Two different stretch methods were applied:
symmetric stretch and asymmetric stretch, which were discussed in Chapter 2 section 2.3.2.
78
In these trials the outer stopper pins were not used. For asymmetric stretch, the inner stopper
pins on the fixed end were placed to secure the scaffold directly against the side of the culture
dish, and on the free end were placed to prevent the scaffold from deforming to less than its
original length. For symmetric stretch, the inner stopper pins were placed only to prevent the
scaffold from deforming to less than its original length, and the scaffold was allowed to stretch
the full length of the culture dish. In this way, the stretch magnitudes of the two methods were
the same. For Gelfoam, the stretch parameters of the biostretch apparatus were set with
frequency of 1Hz, ON time of 50%, and strength of 35% of the maximum value. For RTV, the
parameters were set with frequency of 1Hz, ON time of 50%, and strength of 60% of the
maximum value. The average strain was about 20%.
Digital photographs of the scaffold in the stretched and unstretched states were taken with a
D50 digital camera (Nikon Corporation, Tokyo, Japan) and a Tamron SP 90mm macro lens
(Tamron Co., Ltd., Saitama, Japan). Photographs had a resolution of 60 pixels per millimetre.
The photographs were edited with Adobe Photoshop CS (Adobe Systems, San Jose, California)
using the High Pass and Threshold tools to identify dots. The centroid of each dot was
determined using Scion Image (Scion Corporation, Frederick, Maryland), and distances
between successive centroids along the stretch axis were calculated to determine the distribution
of strain, defined as deformation (L) divided by the original distance between dots (L
0
). For
each distance along the stretch axis, the strains at several equally spaced locations on the
perpendicular axis were calculated and averaged, and the standard distribution of those strains
was determined.
79
(a) (b)
Figure 4.4 Images of the Gelfoam and GE RTV 6166
(a) up: unstretched RTV, down: stretched RTV
(b) up: unstretched Gelfoam, down: stretched Gelfoam
4.3.2 Results
We investigated two types of strain distributions, global strain and local strain distributions.
Global strain distribution refers to the relationship between the strains of different sections;
local strain refers to only the strain of a specific section.
- Global Strain distribution of Gelfoam
Based on the above method, we investigated the global strain distribution in both the stretch
axis (x-axis) and the perpendicular horizontal axis (y-axis). In general, strain distribution was
found to be non-uniform, with a high standard deviation, but exhibiting no significant trends
80
along the stretch axis. Fig. 4.5 (a) shows the average deformation between each adjacent pair of
data points for 10 rows of one Gelfoam sample along the x-axis. No statistically different results
were found for global strain distribution of Gelfoam, which is due to high standard deviation in
measurements.
(a)
(b)
Figure 4.5 Strain Distribution of a Typical Sample of Gelfoam
(a) Up: the x- axes; (b) Down: the y- axes
81
A similar graph for strain distribution in the y-axis shows a clearer trend, with the edges of
the Gelfoam deforming least, since they were securely glued to solid metal clamps (Fig. 4.5
(b)). Fig. 4.5 (a) and (b) demonstrates that the scaffold is not only experienced longitudinal
tensile, but also compressed along the lateral direction owing to Poissons effects.
- Local Strain Distribution of Gelfoam
Compared with the global strain distribution, the deformation of small sections of Gelfoam
(local strain) was far more significant. Photographs of the Gelfoam indicated that shapes of the
dots were visibly distorted during stretching (Fig. 4.6), such that local strain in small regions
(100m) ranged from negative strains on the order of -5% to large positive strains on the order
of 100%. One explanation for the phenomena is the non-uniform construction of the Gelfoam
sponge, which caused areas with larger surface cavities to exhibit higher strain than areas with
smaller surface cavities. Additionally, parts of the Gelfoam may have torn under tensile stress
and hence exhibited higher strain. In comparison to the global strain (10-20%) and the size of a
typical cell (10-40m), these local strains are clearly significant. We note that the material
properties of the Gelfoam may be different when the Gelfoam is seeded with cells; this
difference will be discussed in section 4.6.
Figure 4.6 Distortion of a Marked Dot
Left: before stretching; Right: after stretching
82
The significance of the local strains is demonstrated by comparing Gelfoam to RTV silicone,
which has a visibly more uniform construction. Local strains between dots on Gelfoam spaced
2mm apart exhibit much higher standard deviation than those on RTV silicone, with a standard
deviation of about 30% of the average strain in comparison to less than 10% for RTV silicone.
A visual comparison of typical strain distributions for the two kinds of scaffolds, as in Fig. 4.7,
supports this claim. Since these local strains were calculated by measuring distances between
centroids of dots, they are in effect average values, and the standard deviation for local strains is
actually even higher than calculated.
Figure 4.7 Strain Distribution for the Middle Rows of the Gelfoam and RTV Silicone
Samples
- Symmetric and Asymmetric Stretch
Two different stretch methods, asymmetric and symmetric stretch, are compared for two
kinds of samples. However, there are no significant trends along the stretch axis under different
stretch methods. But the symmetric stretch method provides a good chance for real-time strain
monitoring since the sample could stay centered below a camera. Moreover, other simple
83
physical variables, such as displacement and speed of particles, may be easier to correlate with
differences in cell growth. For example, under one-way stretch, the free end is displaced further
and has a higher average speed than the fixed end, which may correlate with cell growth.
- Edge Effects
We note that the strain distribution in the edge sections, where the clamps are attached, is
heavily dependent on the method of application of force, and is subject to high measurement
error. Before utilizing the clamps described in Chapter section 2.3.2, other clamps were tested.
When force was applied to metal bars inserted through the Gelfoam, strain measured at the
edges was less than half of that measured at the centre. Conversely, when clamps were affixed
to the Gelfoam, strain at the edges appeared more than double the strain measured at the centre.
These are genuine differences in strain; however, they are not necessarily the cause of
differences in cell expression, as there are many other differences between the centre and edges
of the Gelfoam. We also note that the substrate undergoes deformation in the vertical direction
near the edges, which causes strain to appear different from its actual value. For these reasons,
we omit the outer 2 millimetres in all of our measurements.
Since the analyzed data was only taken when the Gelfoam was stretched either fully or not at
all, no information was obtained about the partially stretched state of Gelfoam, and hence strain
rate distribution is unknown. Previous biomechanics studies show that biological tissues are not
elastic materials, the living tissues are normally recognized as viscoelastic materials with
nonlinear stress-strain characteristics [102]. Instead of the strain magnitude itself, the history of
84
strain affects the stress more. Therefore, the study interest was shifted from strain distribution to
strain rate.
4.4 Study of Strain Rate
From the design principle of the apparatus, we know that when the magnet is electrified by a
pulsed current, the magnet will generate a time-varying magnetic field. When the assembled
dish (Fig. 2.6) is positioned under such a field, a magnetic force will act on the magnetizable
bar, which will then stretch the connected tissue construct longitudinally. When the magnetic
field disappears, the tissue construct will restore to its original shape due to the elasticity of the
scaffold. Under a confined strain value, if the stretch force varies, the strain rate will change.
Therefore, the relationship between stretch force and strain rate at the defined strain magnitude
was investigated.
4.4.1 Measurement Methodology
The engineered tissue was prepared according to the protocol provided in section 4.3. After
the tissue patch was cultured for 16 days, the assembled dish was positioned under the scope of
Photrons high-speed camera ultima APX (Photron, San Diego, California), and continuously
subjected to symmetric stretch. The stretching parameters were set as: 1Hz, 50% ON time, and
the total deformation of 5mm. As the original length of the Gelfoam sample is 20mm, the
maximum strain is 25%. The stretching process was recorded by the digital imaging system.
The high-speed digital imaging system was set to 500 fps at full mega pixel image resolution
(1024 x 1024). The recording time was 4.1 seconds, with 2048 frames, such that each record
85
includes four successive stretch periods. The recordings were saved in AVI format. Four video
outputs were recorded, with stretch force ranging from 40% to 90%.
4.4.2 Results
The time period of each frame of the video represents 2ms. The images were edited by GIMP
(GNU General Public License). With the help of the calibration picture, the stretch strain rate,
the restore strain rate, and the frequency response of the device for a specific material and strain
were determined from the video.
- Stretch Strain Rate
The data show that the stretch strain rate varies under different stretch strengths. Fig. 4.8
(a) and (b) show the stretch rates with different stretch strengths of 90% and 40%, respectively,
and the stretch processes demonstrate very good repetition for the four successive periods.
When the strength of the stretch force reaches 90%, 16ms is needed to reach the total
deformation. When the strength is only 40%, 22ms is needed to reach the same deformation.
Despite the difference in stretch time, the two plots share the trend that the stretch rates vary
and increase nonlinearly during the stretch process. These plots reflect the nonlinear viscoelastic
property of the material or indicate that the stretch force is not constant during the stretch
process. Thus, this stretch force will be calculated in section 4.6.
86
stretch rate at 90% strength
0
0.5
1
1.5
2
2.5
1 2 3 4 5 6 7
frames
d
e
f
o
r
m
a
t
i
o
n
[
m
m
]
round1
round2
round3
round4
(a)
stretch rate at 40% strength
0
0.5
1
1.5
2
2.5
1 2 3 4 5 6 7 8 9 10
frames
d
e
f
o
r
m
a
t
i
o
n
[
m
m
]
round1
round2
round3
round4
(b)
Figure 4.8 Stretch Strain Rate at Different Stretch Strength
(a) 90% strength; (b) 40% strength
87
The strain rate during the stretch process shares the exactly same plots as stretch rate (Fig.
4.8), since the strain is obtained by deformation (AL) divided original length (L
0
). In the test
range, the maximum strain rate for 90% stretch strength is about 107s
-1
, while the maximum
strain rate for 40% stretch strength is about 102s
-1
.
- Retreat Rate
When the external force (magnetic force) is removed, the construct will retreat back to its
original place due to the elasticity of the material. The analysis data show that the retreat
process takes 48ms, which is much longer than the stretch process. However, these two plots
demonstrate the same trend, which shows that the retreat process is not affected by whats
happened in the past but the materials own material properties. The plots demonstrate the creep
property (time dependent strain rate) of the material during the unloading process. These results
are in accordance with the facts that the scaffold is a viscoelestic material. But the plots (Fig.
4.9) do not match the creep functions of the ordinary linear viscoelastic mechanical models (the
Maxwell model, the Voigt model, and the Kelvin model) [102]. Therefore, the material property
of the engineered construct needs further study.
The scaffold of the sample is Gelfoam, an absorbable gelatin sponge. The material
properties of the Gelfoam may change when it is used as a scaffold, due to biodegradation and
cell reorganization. This non-invasive test method provides a new avenue to monitor the change
of the mechanical properties of the engineered tissue during culture process.
88
retreat rate at 90% strenght
0
0.5
1
1.5
2
2.5
3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
frames
d
e
f
o
r
m
a
t
i
o
n
[
m
m
]
round1
round2
round3
round4
(a)
stretch rate at 40% strength
0
0.5
1
1.5
2
2.5
3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
frames
d
e
f
o
r
m
a
t
i
o
n
[
m
m
]
round1
round2
round3
round4
(b)
Figure 4.9 Retreat Strain Rate at Different Stretch Strength
(a) 90% strength; (b) 40% strength
89
- Frequency Response
The above data provides us a method to study the frequency response of the device for a
specific material and strain. The stretch force is generated from the pulsed exciting current,
shown in Fig. 4.10 (a). Theoretically, the frequency of the current pulse can reach very high
values, as the maximum source frequency of the timer is 20MHz. However, according to the
recorded data, the stretch time and the retreat time are 16ms and 48ms, respectively. For a 10Hz
pulse train, the response stretch pulse is shown in Fig. 4.10 (b). The shortest response time that
the construct needs for a period is 64ms (Fig. 4.10 (c)). Therefore, the frequency response of the
scaffold is about 15Hz for 25% strain. The data demonstrate that frequency response is related
to not only the stretch force, but also the strain (deformation) and the material properties.
However, this calculation ignores the delay between the exciting current and the generated
magnetic field.
Figure 4.10 Stretch Frequency
90
4.4.3 Discussion
All the collected data were recorded under the real culture circumstance. During the process,
the Gelfoam were soaked in medium. The video shows that during stretching, the fluctuation of
the medium distorted the image, especially at the moment that the metal bar hit the wall of the
culture dish or the stoppers of the mounting tray, which is the major measurement error during
the test. Therefore, it was hard to capture the accurate position of the steel bar at the hitting
momentary. Due to this reason, the deformation of the last frame is not plotted in Fig. 4.8.
This test demonstrated that for a specific material the stretch rate is related to the strength of
the stretch force. However, the stretch rate is not proportional to the stretch force. When the
force doubled (from 40% to 90%), the full stretch time did not reduce to half (from 22ms to
16ms). Additionally, the material property affects both the stretch rate and the retreat rate. In
order to obtain the same strain magnitude, different scaffolds need different initial force to
stretch. This claim has been supported in section 4.4.1. To test the strain distributions of
Gelfoam and RTV silicone, different stretch force were applied. Therefore, this nondestructive
method also provides researchers a tool to real time monitor the variation of the material
property during cell orientation.
4.5 Study of Stretch Force
From section 4.5, we noticed that the stretch force may not be a constant force. In this
section, we will estimate the distribution and magnitude of the magnetic force generated
between a magnetisable bar and a horseshoe-shaped electromagnet. The geometrical model of
the magnet is based on MCI-3404 (Mci Limited Co., ON, CA) utilized in the biostretch
91
apparatus, the magnetizable bar (16x 3.5x 1mm). We focus our model study on two dimensions,
particularly along horizontal (x) direction because the horizontal magnetic force is used as the
uniaxial stretch force. In this model, the two dimensions of the bar are 16 x 1 mm, and s the
distance between the bar and the magnet is 5mm. The simulations were performed using
COMSOL Multiphysics 3.3.
4.5.1 Electromagnetic Force Calculation
In COMSOL Multiphysics, Maxwells stress tensor T is used to calculate the magnetic
force act on the surface of the object. Cauchys equation of continuum mechanics states
ext
f T
dt
r d
+ V =
2
2
(4.1)
where is the density, r denotes the coordinates of a material point, T is the stress
tensor, and f
ext
is an external volume force. In the stationary case, the equation representing the
force balance is
ext
f T + V = 0 (4.2)
If the stress tensor is continuous across a stationary boundary between two materials, the
equation can be expressed as:
0 ) (
1 2 1
= T T n (4.3)
where T
1
and T
2
represent the stress tensors of material 1 and 2, and n
1
is the outward
normal direction of material 1. The equation shows when material 1 (solid) is surrounded by
material 2 (air/vacuum), the surface force is on the boundary between the materials with the
solid. The user guide of COMSOL [89] deducted the electromagnetic stress tensor expression in
air:
92
T T
HB ED I B H D E pI T + + + = )
2
1
2
1
(
2
(4.4)
where p is the air pressure, I is the identity 3*3 tensor, E is the electric field intensity, D
is the electric flux density, H is the magnetic field intensity, B is the magnetic flux density. For
the quasi-static magnetic field, the stress tensor can be simplified as:
T
B H n B H n T n ) ( ) (
2
1
1 1 2 1
+ = (4.5)
4.5.2 Electromagnetic Model Analysis
In order to study the magnetic force magnitude and distribution in a specific area
according to the equation (4-5), the magnitude and distribution of the magnetic flux density B in
the study area must first be calculated. To perform the simulation, beside the assumptions we
made for the model discussed in Chapter 3, we added two more assumptions:
- The material of the steel bar is soft iron instead of Ph-174
- The surrounding area is air instead of medium
The meshed geometry of the electromagnet model has 3096 elements and the number of
degrees of freedom is 6251 (Fig. 4.11). The mesh generator partitions the subdomains into
triangular elements, which are allocated based on the geometrical dimensions.
93
Figure 4.11 Meshed Model of an Electromagnet and a Steel Bar
Figure 4.12 Magnetic Flux Density, x component
94
Fig. 4.12 shows the x component distribution with the maximum flux density in a period
at phase zero. In the graph, the magnetic south pole (orange) of the horseshoe electromagnet is
at the lower foot and the magnetic north pole (blue) is at the upper foot. The distribution shows
that the magnetic flux density Bx has greater magnitude in the soft iron core than in the air
space. The maximum B value in the soft iron is 0.221T located at the inner corners of the core.
However the magnitude of Bx dramatically reduces to 0.086T at the boundary of the iron core
and air space. Compared Fig. 4.12 and Fig. 3.2 (a), we can find that the steel bar was
magnetized in the field. The Bx in the area between the magnet and the steel bare increased.
4.5.3 Results
- Flux Density Bx VS Distance
X Direction Fig. 4.7 (a) and (b) shows the cross-section plot of magnetic flux density Bn
normal along the X direction (from x=0.08 to x=0.085) at y=0.144 (south pole), and at y=0.159
(north pole), respectively. Compared to Fig. 3.4, the magnitude of the normal magnetic flux
density in Fig. 4.13 dramatically increased when the distance increased to 4mm. The
phenomenon is caused by the magnetizing of the steel bar.
95
(a)
(b)
Figure 4.13 Magnetic Flux Density Bn along x Direction
(a) at Y=0.143; (b) at Y=0.159
96
Figure 4.14 Magnetic Flux Density Bx along y Direction
Bx distributions @ x=0.085
0
0.02
0.04
0.06
0.08
0.1
0.12
0
.
1
5
9
0
.
1
5
8
0
.
1
5
7
0
.
1
5
6
0
.
1
5
5
0
.
1
5
4
0
.
1
5
3
0
.
1
5
2
0
.
1
5
1
0
.
1
5
0
0
.
1
4
9
0
.
1
4
8
0
.
1
4
7
0
.
1
4
6
0
.
1
4
5
0
.
1
4
5
0
.
1
4
4
y [m]
M
a
g
n
e
t
i
c
F
l
u
x
D
e
n
s
i
t
y
,
x
c
o
m
p
o
n
e
n
t
[
T
]
without rod
with rod
Figure 4.15 Comparison of Magnetic Flux Density Bx
97
Y Direction Fig. 4.14 shows the cross-section plot of the x component of the magnetic flux
density along y direction (from y=0.143 to y=0.159) at x=0.085. Fig. 4.14 demonstrates the
same trend as Fig. 3.3, but the magnitude is much greater. Fig. 4.15 compares the difference
between these two plots in absolute terms, which shows the flux density at the edge of the bar is
increased six times. This plot also implies that the force is mostly applied on the edge of the bar,
such that the center part will not be subject to magnetic force. Both Fig. 4.13 and Fig 4.14
illustrate the distribution of the flux density along x and y direction, which directly affects the
calculation of the magnetic force.
- Magnetic Force Fx vs. Exciting Current I
The magnetic force acting on the steel bar is calculated by the stress tensor method
described in section 4.6.1. The simulation results show that the magnetic force that the bar is
subjected to along the x direction is 5.1N when the electromagnet is electrified by a 0.4A time-
harmonic sinusoidal current. This is the maximum force the device can provide at 5mm distance
far because the maximum current designed for the device is 0.4A. The force varies as the
exciting current (sinusoidal signal) changes during a period. The relationship between the
magnetic force along x direction and the exciting current is illustrated in Fig. 4.16. In the
previous bio-experiments, the Gelfoam worked as the scaffold, and the stretch force was
normally set as 50%, which means the exciting current was 0.2A. From Fig. 4.16, we can see
the stretch force generated would be about 1.25N.
98
Figure 4.16 Exciting Current vs. Magnetic Force
- Magnetic Force Fx VS Distance
Fig. 4.17 shows that the distribution of the flux density is non-linear along the x direction,
which means the magnetic force will also vary along the x direction. The plot demonstrates that
when the bar is closer to the electromagnet, it will be subject to a much stronger force. For a
35mm culture dish, the maximum deformation of the substrate is set to be 2.5mm, so the force
ranges form 1.25N to 4N.
99
Rod_force_x versus Distance
0
10
20
30
40
50
60
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Di stance[mm]
R
o
d
_
f
o
r
c
e
_
x
[
N
]
Figure 4.17 Distance vs. Magnetic Force
In this thesis, the area around the electromagnet was modeled as air instead of medium, and the
steel bar was modeled by soft iron; more realistic modeling wound give more accurate results.
4.6 Conclusions
In order to validate the performance of the designed apparatus, the effects of uniaxial
cyclic stretch on cell biology have been studied from the aspects of cell proliferation and cell
morphology. The DNA assay results show that the apparatus improved cell proliferation in a
similar way as existing uniaxial stretch devices. The HE staining demonstrated the gross tissue
structure, and the results showed that stretching guided the cell reorganization along the stretch
direction. Moreover, the mechanical parameters (strain distribution, strain rate, and stretch
force) provided by the apparatus have also been investigated. A nondestructive method was
employed to investigate the strain distribution on the surface of the porous scaffold. The results
demonstrate that the strain distribution along the stretch direction is non-uniform, with a high
100
standard deviation. The strain rate was investigated by Photrons high-speed camera ultima
APX. The results demonstrated the viscoelastic property (time dependent strain) of the scaffold,
and also indicated that the strain rate varies with the strength of the stretch force. The stretch
force was calculated to be 1.25N at 5mm distance from the edge of the electromagnet by
COMSOL Multiphysics, the force was shown to be related to distance and current.
101
Chapter 5
Summary and Future Research
5.1 Summary
Mechanical stimulation is an important regulator in tissue culture for tissue engineering.
Among different forms of mechanical stimuli, uniaxial cyclic stretch is the most commonly
used stimulus for muscle tissues due to its precisely controlled end to end average strain. The
most typical method of applying stretch is motor driven. However, the motor driven method
involves a physical connection between the culture construct and the motor, which requires a
diversity of laboratory devices and increases the risk of contamination.
This thesis proposed a unique idea to design a uniaxial cyclic stretch apparatus based on
non-contact magnetic force. The apparatus developed in this thesis allows researchers to utilize
standard culture dishes, and applies a well-defined strain magnitude to an engineered tissue
patch during culture process, which greatly simplifies the assembly procedure and reduces the
risk of contamination. A prototype of the apparatus has been constructed and tested in a
research lab at the University Health Network in Toronto.
Since the applied mechanical stimulus is generated from magnetic force, the engineered
tissue construct is not only affected by mechanical force, but also exposed to a magnetic field.
102
Thus, the effects of the time-varying magnetic field during the culture process were
investigated. One side effect of using electromagnets, that of a temperature increase, is also
investigated. However, the biomedical experiment results show that neither a weak low
frequency magnetic field (0.1T, 1Hz) nor an increase of 1 in temperature has a significant
effect on the cell culture.
In order to verify the performance of the designed apparatus, the effects of uniaxial
cyclic stretch on cell biology have been studied from the aspects of cell proliferation and cell
orientation. The results show that the apparatus improved cell proliferation and cell
reorganization similarly to other known uniaxial stretch devices. Moreover, the mechanical
parameters (strain distribution, strain rate, and stretch force) provided by the apparatus have
also been quantitatively investigated. The strain distribution on the surface of the porous
scaffold along the stretch direction was shown to be non-uniform, with a high standard
deviation. The results of strain rate test indicated that the strain rate varies during stretching due
to the variation of the stretch force. It also proved the viscoelastic property (time dependent
strain) of the scaffold, and provided a new way to monitor the mechanical property of the
engineered tissue during tissue formation. The stretch force, which was calculated using
COMSOL Multiphysics, was shown to be related to distance and current.
5.2 Future Research
This section discusses the future research work that will be pursued after the completion
of the thesis, from the aspects of modeling, bio-experiments, and improvements.
103
5.2.1 Optimizing the Electromagnet Model
The model of the electromagnet discussed in this thesis only considered the quasi-static
electromagnetic field itself, while ignoring the influence of the energy dissipation problems
caused by resistance heating and hysteresis. In the future, the study areas need to cover not only
electromagnetics, but also heat transfer coupled with thermal-electric analysis. The multi-
physical model can help us optimize the relationships among the current, force, and
temperature. Additionally, some of the assumptions need to be revised in order to obtain more
accurate results, such as more specifically characterizing the features of the coil. Moreover, the
model should be upgraded from two-dimensional to three-dimensional.
5.2.2 Implementing Bio-experiments
Although the bio-experiments validated the performance of the apparatus, and
demonstrated that uniaxial stretch can influence cell culture from the aspects of cell
proliferation and cell orientation, the relationship between cell characteristics and the
mechanical parameters are still unclear. The results were affected by multiple parameters such
as stretch frequency, strength, strain, and even stretch pattern and shear flow. The independent
contribution of each parameter is not characterized. Therefore, future experiments should isolate
these mechanical parameters, and systematically examine their independent influence. Such an
experiment would provide researchers an effective platform to explore the mechatransduction of
cells. The upcoming experiment is studying tissue formation by leading the human embryonic
stem cells differentiated to cardiac tissue cells with the defined mechanical stimulation.
104
5.2.3 Calibrating the stretch force
At current stage, the primary controllable parameter of mechanical stimulation in tissue
engineering is strain magnitude, and cultured tissue patches are subjected to predefined
mechanical strain. However, the mechanical force that the patches are subjected to is not fully
studied. According to the simulation results of the COMSOL model, the stretch force is 1.25N
at 5mm away from the magnet, when the magnet is excited by a 0.2A pulsed current. Although
the simulated results of magnetic flux density B are verified by physical experimentation, and
the magnetic force is calculated based on the verified magnetic field, the magnetic force itself is
unverified. Further experiment needs to conduct to determine the stretch force. With the
information of the tested stretch force and filmed deformation, the mechanical properties
(stiffness or elasticity) of the scaffold can be monitored during tissue culture.
5.2.4 Integrating the mechanical and electrical stimuli together
Current bioreactor systems only apply uniaxial cyclic stretch while engineering a tissue
construct. However, the next step of the research will be to combine cyclic stretch and electrical
field stimulation together. The proposed bioreactor is designed to deliver a pulsed electrical
signal to the cell-seeded scaffold with the same frequency as the cyclic stretch. For cardiac
tissue engineering, these external stimuli will mimic the conductive and contractile properties
existing in native heart tissue. Hopefully, these stimuli can lead to better ultrastructural
organization of the engineered tissues.
105
The proposed research aims to develop a novel, compact electrical and mechanical
stimulation apparatus for 3D tissue culture. As shown in Fig. 5.1, the apparatus will consist of a
computerized controller, an electrified mounting tray, and a pair of magnetic metal bars. The
two ends of the cell-seeded scaffold (1) will be fixed by a pair of silver-coated stainless steel
bars (2). The integrated construct (3) will be secured onto the electrified mounting tray (4). The
tray will contain a battery (5) and relevant circuits (6), which will be coated by a biocompatible
material; it will also contain two pairs of metal stoppers (7) serving as electrodes. The mounting
tray will be designed to fit into a standard culture dish (8). Finally, the assembled culture dish
will be mounted between a pair of electromagnets (9). When the electromagnets are electrified
by the controller, the construct integrated with metal bars will experience uniaxial cyclic stretch
between the stoppers and the wall of the dish. As the metal bars touch the metal stoppers, the
construct will also undergo electrical stimulation.
Figure 5.1 Schematic Diagram of the Upgraded Device
106
The device will provide simultaneously and synchronously both uniaxial mechanical cyclic
stretch and electrical stimulus to a 3D engineered-tissue construct, which will subtly mimic the
spontaneous excitation-contraction coupling of native cardiac tissue. With the non-contact
magnetic force and encapsulated battery, the tissue patch will still be cultured in a standard
culture dish. Such an isolated environment will greatly limit the risk of contamination. Using
the new device, researchers will be able to study the effects of both mechanical and electrical
stimuli on cell proliferation, the increased expression of cardiogenesis, and the organization of
the extracellular matrix. Biological experiments will be conducted in the Division of
Cardiovascular Surgery at the University of Toronto. The new apparatus will provide an
effective tool to investigate the mechanism of spontaneous beating of cardiac tissues.
107
References
[1] Fuchs, J. R., Nasseri, B. A., and Vacanti, J. P., 2001, "Tissue Engineering: A 21st Century
Solution to Surgical Reconstruction," Annals of Thoracic Surgery, 72(2) pp. 577-591.
[2] Eschenhagen, T., and Zimmermann, W. H., 2005, "Engineering Myocardial Tissue,"
Circulation Research, 97(12) pp. 1220-1231.
[3] Langer, R., and Vacanti, J. P., 1993, "Tissue Engineering," Science, 260(5110) pp. 920-926.
[4] Leor, J., Amsalem, Y., and Cohen, S., 2005, "Cells, Scaffolds, and Molecules for
Myocardial Tissue Engineering," Pharmacology and Therapeutics, 105(2) pp. 151-163.
[5] Radii, M., and Vunjak-Novakovi, G., 2005, "Cardiac Tissue Engineering," Journal of the
Serbian Chemical Society, 70(3) pp. 541-556.
[6] Akhyari, P., Fedak, P. W. M., Weisel, R. D., 2002, "Mechanical Stretch Regimen Enhances
the Formation of Bioengineered Autologous Cardiac Muscle Grafts," Circulation, 106(13
SUPPL.) pp. I137-I142.
[7] Griffith, L. G., and Naughton, G., 2002, "Tissue Engineering - Current Challenges and
Expanding Opportunities," Science, 295(5557) pp. 1009-1010+1012-1014.
[8] Discher, D., Dong, C., Fredberg, J. J., 2009, "Biomechanics: Cell Research and Applications
for the Next Decade," Annals of Biomedical Engineering, 37(5) pp. 847-859.
108
[9] Jawad, H., Ali, N. N., Lyon, A. R., 2007, "Myocardial Tissue Engineering: A Review."
Journal of Tissue Engineering and Regenerative Medicine, 1(5) pp. 327-342.
[10] Radisic, M., Park, H., Gerecht, S., 2007, "Biomimetic Approach to Cardiac Tissue
Engineering," Philosophical Transactions of the Royal Society B: Biological Sciences,
362(1484) pp. 1357-1368.
[11] Butler, D. L., Goldstein, S. A., and Guilak, F., 2000, "Functional Tissue Engineering: The
Role of Biomechanics," Journal of Biomechanical Engineering, 122(6) pp. 570-575.
[12] Vandenburgh, H. H., Karlisch, P., and Farr, L., 1988, "Maintenance of Highly Contractile
Tissue-Cultured Avian Skeletal Myotubes in Collagen Gel," In Vitro Cellular and
Developmental Biology - Animal, 24(3 PART I) pp. 166-174.
[13] Terracio, L., Miller, B., and Borg, T. K., 1988, "Effects of Cyclic Mechanical Stimulation
of the Cellular Components of the Heart: In Vitro," In Vitro Cellular and Developmental
Biology - Animal, 24(1) pp. 53-58.
[14] Radisic, M., Park, H., Shing, H., 2004, "Functional Assembly of Engineered Myocardium
by Electrical Stimulation of Cardiac Myocytes Cultured on Scaffolds," Proceedings of the
National Academy of Sciences of the United States of America, 101(52) pp. 18129-18134.
[15] Brown, T. D., 2000, "Techniques for Mechanical Stimulation of Cells in Vitro: A Review,"
Journal of Biomechanics, 33(1) pp. 3-14.
[16] Fink, C., Ergn, S., Kralisch, D., 2000, "Chronic Stretch of Engineered Heart Tissue
Induces Hypertrophy and Functional Improvement," FASEB Journal, 14(5) pp. 669-679.
109
[17] Ethier, C.R., and Simmons, C.A., 2007, "Introductory Biomechanics From Cells to
Organisms," CAMPRIDGE UNIVERSITY PRESS, New York, pp. 511.
[18] El Haj, A. J., Minter, S. L., Rawlinson, S. C. F., 1990, "Cellular Responses to Mechanical
Loading in Vitro," Journal of Bone and Mineral Research, 5(9) pp. 923-932.
[19] Williams, G. M., Lin, J. W., and Sah, R. L., 2007, "Cartilage Reshaping Via in Vitro
Mechanical Loading," Tissue Engineering, 13(12) pp. 2903-2911.
[20] Hung, C. T., Mauck, R. L., Wang, C. C. -., 2004, "A Paradigm for Functional Tissue
Engineering of Articular Cartilage Via Applied Physiologic Deformational Loading,"
Annals of Biomedical Engineering, 32(1) pp. 35-49.
[21] Knippenberg, M., Helder, M. N., Doulabi, B. Z., 2005, "Adipose Tissue-Derived
Mesenchymal Stem Cells Acquire Bone Cell-Like Responsiveness to Fluid Shear Stress on
Osteogenic Stimulation," Tissue Engineering, 11(11-12) pp. 1780-1788.
[22] Triplett, J. W., O'Riley, R., Tekulve, K., 2007, "Mechanical Loading by Fluid Shear Stress
Enhances IGF-1 Receptor Signaling in Osteoblasts in A PKC-Dependent Manner," MCB
Molecular and Cellular Biomechanics, 4(1) pp. 13-25.
[23] DePaola, N., Davies, P. F., Pritchard Jr., W. F., 1999, "Spatial and Temporal Regulation of
Gap Junction connexin43 in Vascular Endothelial Cells Exposed to Controlled Disturbed
Flows in Vitro," Proceedings of the National Academy of Sciences of the United States of
America, 96(6) pp. 3154-3159.
110
[24] Zimmermann, W. -., Schneiderbanger, K., Schubert, P., 2002, "Tissue Engineering of a
Differentiated Cardiac Muscle Construct," Circulation Research, 90(2) pp. 223-230.
[25] Tandon, N., Cannizzaro, C., Chao, P. -. G., 2009, "Electrical Stimulation Systems for
Cardiac Tissue Engineering," Nature Protocols, 4(2) pp. 155-173.
[26] Erickson, C. A., and Nuccitelli, R., 1984, "Embryonic Fibroblast Motility and Orientation
can be Influenced by Physiological Electric Fields," Journal of Cell Biology, 98(1) pp. 296-
307.
[27] Sun, S., Titushkin, I., and Cho, M., 2006, "Regulation of Mesenchymal Stem Cell
Adhesion and Orientation in 3D Collagen Scaffold by Electrical Stimulus,"
Bioelectrochemistry, 69(2) pp. 133-141.
[28] Au, H. T. H., Cheng, I., Chowdhury, M. F., 2007, "Interactive Effects of Surface
Topography and Pulsatile Electrical Field Stimulation on Orientation and Elongation of
Fibroblasts and Cardiomyocytes," Biomaterials, 28(29) pp. 4277-4293.
[29] Guido, S., and Tranquillo, R. T., 1993, "A Methodology for the Systematic and
Quantitative Study of Cell Contact Guidance in Oriented Collagen Gels. Correlation of
Fibroblast Orientation and Gel Birefringence," Journal of Cell Science, 105(2) pp. 317-331.
[30] Tranquillo, R. T., Girton, T. S., Bromberek, B. A., 1996, "Magnetically Orientated Tissue-
Equivalent Tubes: Application to a Circumferentially Orientated Media-Equivalent,"
Biomaterials, 17(3) pp. 349-357.
111
[31] Hashimoto, Y., Kawasumi, M., and Saito, M., 2007, "Effect of Static Magnetic Field on
Cell Migration," Electrical Engineering in Japan (English Translation of Denki Gakkai
Ronbunshi), 160(2) pp. 46-52.
[32] Goodman, R., and Blank, M., 2002, "Insights into Electromagnetic Interaction
Mechanisms," Journal of Cellular Physiology, 192(1) pp. 16-22.
[33] Freed, L. E., Guilak, F., Guo, X. E., 2006, "Advanced Tools for Tissue Engineering:
Scaffolds, Bioreactors, and Signaling," Tissue Engineering, 12(12) pp. 3285-3305.
[34] Bilodeau, K., and Mantovani, D., 2006, "Bioreactors for Tissue Engineering: Focus on
Mechanical Constraints. A Comparative Review," Tissue Engineering, 12(8) pp. 2367-2383.
[35] Chen, H. -., and Hu, Y. -., 2006, "Bioreactors for Tissue Engineering," Biotechnology
Letters, 28(18) pp. 1415-1423.
[36] Carrier, R. L., Rupnick, M., Langer, R., 2002, "Perfusion Improves Tissue Architecture of
Engineered Cardiac Muscle," Tissue Engineering, 8(2) pp. 175-188.
[37] Bancroft, G. N., Sikavitsas, V. I., and Mikos, A. G., 2003, "Design of a Flow Perfusion
Bioreactor System for Bone Tissue-Engineering Applications," Tissue Engineering, 9(3) pp.
549-554.
[38] Schwarz, R. P., Goodwin, T. J., and Wolf, D. A., 1992, "Cell Culture for Three-
Dimensional Modeling in Rotating-Wall Vessels: An Application of Simulated
Microgravity," Journal of Tissue Culture Methods, 14(2) pp. 51-57.
112
[39] Leor, J., and Cohen, S., 2004, "Myocardial Tissue Engineering: Creating a Muscle Patch
for a Wounded Heart," Annals of the New York Academy of Sciences, 1015pp. 312-319.
[40] Carrier, R. L., Papadaki, M., Rupnick, M., 1999, "Cardiac Tissue Engineering: Cell
Seeding, Cultivation Parameters, and Tissue Construct Characterization," Biotechnology
and Bioengineering, 64(5) pp. 580-589.
[41] Sodian, R., Lemke, T., Fritsche, C., 2002, "Tissue-Engineering Bioreactors: A New
Combined Cell-Seeding and Perfusion System for Vascular Tissue Engineering," Tissue
Engineering, 8(5) pp. 863-870.
[42] Dumont, K., Yperman, J., Verbeken, E., 2002, "Design of a New Pulsatile Bioreactor for
Tissue Engineered Aortic Heart Valve Formation," Artificial Organs, 26(8) pp. 710-714.
[43] Seidel, J. O., Pei, M., Gray, M. L., 2004, "Long-Term Culture of Tissue Engineered
Cartilage in a Perfused Chamber with Mechanical Stimulation," Biorheology, 41(3-4) pp.
445-458.
[44] Vance, J., Galley, S., Liu, D. F., 2005, "Mechanical Stimulation of MC3T3 Osteoblastic
Cells in a Bone Tissue-Engineering Bioreactor Enhances Prostaglandin E2 Release," Tissue
Engineering, 11(11-12) pp. 1832-1839.
[45] Feng, Z., Matsumoto, T., Nomura, Y., 2005, "An Electro-Tensile Bioreactor for 3-D
Culturing of Cardiomyocytes," IEEE Engineering in Medicine and Biology Magazine, 24(4)
pp. 73-79.
113
[46] Butler, D. L., Hunter, S. A., Chokalingam, K., 2009, "Using Functional Tissue Engineering
and Bioreactors to Mechanically Stimulate Tissue-Engineered Constructs," Tissue
Engineering - Part A, 15(4) pp. 741-749.
[47] Osman, N. F., McVeigh, E. R., and Prince, J. L., 2000, "Imaging Heart Motion using
Harmonic Phase MRI," IEEE Transactions on Medical Imaging, 19(3) pp. 186-202.
[48] Garvin, J., Qi, J., Maloney, M., 2003, "Novel System for Engineering Bioartificial Tendons
and Application of Mechanical Load," Tissue Engineering, 9(5) pp. 967-979.
[49] Soma, S., Matsumoto, S., and Takano-Yamamoto, T., 1997, "Enhancement by Conditioned
Medium of Stretched Calvarial Bone Cells of the Osteoclast-Like Cell Formation Induced
by Parathyroid Hormone in Mouse Bone Marrow Cultures," Archives of Oral Biology, 42(3)
pp. 205-211.
[50] Lee, C. H., Shin, H. J., Cho, I. H., 2005, "Nanofiber Alignment and Direction of
Mechanical Strain Affect the ECM Production of Human ACL Fibroblast," Biomaterials,
26(11) pp. 1261-1270.
[51] Gilbert, J. A., 1994, "Strain Profiles for Circular Cell Culture Plates Containing Flexible
Surfaces Employed to Mechanically Deform Cells in Vitro," Journal of Biomechanics, 27(9)
pp. 1169-1177.
[52] Sotoudeh, M., Jalali, S., Usami, S., 1998, "A Strain Device Imposing Dynamic and
Uniform Equi-Biaxial Strain to Cultured Cells," Annals of Biomedical Engineering, 26(2)
pp. 181-189.
114
[53] Meikle, M. C., Reynolds, J. J., Sellers, A., 1979, "Rabbit Cranial Sutures in Vitro: A New
Experimental Model for Studying the Response of Fibrous Joints to Mechanical Stress,"
Calcified Tissue International, 28(2) pp. 137-144.
[54] Vandenburgh, H. H., Swasdison, S., and Karlisch, P., 1991, "Computer-Aided
Mechanogenesis of Skeletal Muscle Organs from Single Cells in Vitro," FASEB Journal,
5(13) pp. 2860-2867.
[55] Powell, C. A., Smiley, B. L., Mills, J., 2002, "Mechanical Stimulation Improves Tissue-
Engineered Human Skeletal Muscle," American Journal of Physiology - Cell Physiology,
283(5 52-5) pp. C1557-C1565.
[56] Yung, Y. C., Vandenburgh, H., and Mooney, D. J., 2009, "Cellular Strain Assessment Tool
(CSAT): Precision-Controlled Cyclic Uniaxial Tensile Loading," Journal of Biomechanics,
42(2) pp. 178-182.
[57] Zimmermann, W. H., Fink, C., Kralisch, D., 2000, "Three-Dimensional Engineered Heart
Tissue from Neonatal Rat Cardiac Myocytes," Biotechnology and Bioengineering, 68(1) pp.
106-114.
[58] Yost, M. J., Simpson, D., Wrona, K., 2000, "Design and Construction of a Uniaxial Cell
Stretcher," American Journal of Physiology - Heart and Circulatory Physiology, 279(6 48-6)
pp. H3124-H3130.
115
[59] Joshi, S. D., and Webb, K., 2008, "Variation of Cyclic Strain Parameters Regulates
Development of Elastic Modulus in fibroblast/substrate Constructs," Journal of Orthopaedic
Research, 26(8) pp. 1105-1113.
[60] Webb, K., Hitchcock, R. W., Smeal, R. M., 2006, "Cyclic Strain Increases Fibroblast
Proliferation, Matrix Accumulation, and Elastic Modulus of Fibroblast-Seeded
Polyurethane Constructs," Journal of Biomechanics, 39(6) pp. 1136-1144.
[61] Wakatsuki, T., Kolodney, M. S., Zahalak, G. I., 2000, "Cell Mechanics Studied by a
Reconstituted Model Tissue," Biophysical Journal, 79(5) pp. 2353-2368.
[62] Wille, J. J., Elson, E. L., and Okamoto, R. J., 2006, "Cellular and Matrix Mechanics of
Bioartificial Tissues during Continuous Cyclic Stretch," Annals of Biomedical Engineering,
34(11) pp. 1678-1690.
[63] Langelier, E., Rancourt, D., Bouchard, S., 1999, "Cyclic Traction Machine for Long-Term
Culture of Fibroblast-Populated Collagen Gels," Annals of Biomedical Engineering, 27(1)
pp. 67-72.
[64] Mol, A., Bouten, C. V. C., Znd, G., 2003, "The Relevance of Large Strains in Functional
Tissue Engineering of Heart Valves," Thoracic and Cardiovascular Surgeon, 51(2) pp. 78-
83.
[65] Pfister, B. J., Weihs, T. P., Betenbaugh, M., 2003, "An in Vitro Uniaxial Stretch Model for
Axonal Injury," Annals of Biomedical Engineering, 31(5) pp. 589-598.
116
[66] Liu, M., Montazeri, S., Jedlovsky, T., 1999, "Bio-Stretch, a Computerized Cell Strain
Apparatus for Three-Dimensional Organotypic Cultures," In Vitro Cellular and
Developmental Biology - Animal, 35(2) pp. 87-93.
[67] Smalt, R., Mitchell, F. T., Howard, R. L., 1997, "Induction of NO and Prostaglandin E2 in
Osteoblasts by Wall-Shear Stress but Not Mechanical Strain," American Journal of
Physiology - Endocrinology and Metabolism, 273(4 36-4) pp. E751-E758.
[68] Liu, M., and Post, M., 2000, "Invited Review: Mechanochemical Signal Transduction in
the Fetal Lung," Journal of Applied Physiology, 89(5) pp. 2078-2084.
[69] Liu, M., Tanswell, A. K., and Post, M., 1999, "Mechanical Force-Induced Signal
Transduction in Lung Cells," American Journal of Physiology - Lung Cellular and
Molecular Physiology, 277(4 21-4) pp. L667-L683.
[70] Yang, X., Vezeridis, P. S., Nicholas, B., 2006, "Differential Expression of Type X
Collagen in a Mechanically Active 3-D Chondrocyte Culture System: A Quantitative
Study," Journal of Orthopaedic Surgery and Research, 1(1) .
[71] Auluck, A., Mudera, V., Hunt, N. P., 2005, "A Three-Dimensional in Vitro Model System
to Study the Adaptation of Craniofacial Skeletal Muscle Following Mechanostimulation,"
European Journal of Oral Sciences, 113(3) pp. 218-224.
[72] Boublik, J., Park, H., Radisic, M., 2005, "Mechanical Properties and Remodeling of
Hybrid Cardiac Constructs made from Heart Cells, Fibrin, and Biodegradable, Elastomeric
Knitted Fabric," Tissue Engineering, 11(7-8) pp. 1122-1132.
117
[73] Shimizu, K., Ito, A., and Honda, H., 2007, "Mag-Seeding of Rat Bone Marrow Stromal
Cells into Porous Hydroxyapatite Scaffolds for Bone Tissue Engineering," Journal of
Bioscience and Bioengineering, 104(3) pp. 171-177.
[74] Yang, S. -., Sun, J. -., Liu, C. -., 2008, "Ex Vivo Magnetofection with Magnetic
Nanoparticles: A Novel Platform for Nonviral Tissue Engineering," Artificial Organs, 32(3)
pp. 195-204.
[75] Ito, A., Jitsunobu, H., Kawabe, Y., 2007, "Construction of Heterotypic Cell Sheets by
Magnetic Force-Based 3-D Coculture of HepG2 and NIH3T3 Cells," Journal of Bioscience
and Bioengineering, 104(5) pp. 371-378.
[76] Dobson, J., Cartmell, S. H., Keramane, A., 2006, "Principles and Design of a Novel
Magnetic Force Mechanical Conditioning Bioreactor for Tissue Engineering, Stem Cell
Conditioning, and Dynamic in Vitro Screening," IEEE Transactions on Nanobioscience,
5(3) pp. 173-177.
[77] Dobson, J., 2008, "Remote Control of Cellular Behaviour with Magnetic Nanoparticles,"
Nature Nanotechnology, 3(3) pp. 139-143.
[78] Moulder, J. E., 2000, "The Electric and Magnetic Fields Research and Public Information
Dissemination (EMF-RAPID) Program," Radiation Research, 153(5 II SUPPL.) pp. 613-
616.
[79] Edmonds, D. T., 1993, "Larmor Precession as a Mechanism for the Detection of Static and
Alternating Magnetic Fields," Bioelectrochemistry and Bioenergetics, 30(C) pp. 3-12.
118
[80] Cane, V., Botti, P., and Soana, S., 1993, "Pulsed Magnetic Fields Improve Osteoblast
Activity during the Repair of an Experimental Osseous Defect," Journal of Orthopaedic
Research, 11(5) pp. 664-670.
[81] Pilla, A. A., 2002, "Low-Intensity Electromagnetic and Mechanical Modulation of Bone
Growth and Repair: Are they Equivalent?" Journal of Orthopaedic Science, 7(3) pp. 420-
428.
[82] Denegre, J. M., Valles Jr., J. M., Lin, K., 1998, "Cleavage Planes in Frog Eggs are Altered
by Strong Magnetic Fields," Proceedings of the National Academy of Sciences of the
United States of America, 95(25) pp. 14729-14732.
[83] Fassina, L., Visai, L., Benazzo, F., 2006, "Effects of Electromagnetic Stimulation on
Calcified Matrix Production by SAOS-2 Cells Over a Polyurethane Porous Scaffold,"
Tissue Engineering, 12(7) pp. 1985-1999.
[84] De Mattei, M., Pasello, M., Pellati, A., 2003, "Effects of Electromagnetic Fields on
Proteoglycan Metabolism of Bovine Articular Cartilage Explants," Connective Tissue
Research, 44(3-4) pp. 154-159.
[85] Funk, R. H. W., Monsees, T., and zkucur, N., 2009, "Electromagnetic Effects - from Cell
Biology to Medicine," Progress in Histochemistry and Cytochemistry, 43(4) pp. 177-264.
[86] Miyakoshi, J., 2005, "Effects of Static Magnetic Fields at the Cellular Level," Progress in
Biophysics and Molecular Biology, 87(2-3 SPEC. ISS.) pp. 213-223.
[87] Stavroulakis, P., 2003, "Biological Effects of Electromagnetic Radiation,"Springer, Berlin, .
119
[88] Komazaki, S., and Takano, K., 2007, "Induction of Increase in Intracellular Calcium
Concentration of Embryonic Cells and Acceleration of Morphogenetic Cell Movements
during Amphibian Gastrulation by a 50-Hz Magnetic Field," Journal of Experimental
Zoology Part A: Ecological Genetics and Physiology, 307(3) pp. 156-162.
[89] Anonymous 2008, "COMSOL Multiphysics Users Guide," .
[90] Bansal, R., 2006, "Fundamentals of Engineering Electromagnetics," Taylor&Francis, .
[91] Magnetic Instruments Inc., "Gaussmeter model 2010 manual," Indianapolis.
[92] Skinner, S. J. M., Sommervell, C. E., and Olson, D. M., 1992, "The Effects of Mechanical
Stretching on Fetal Rat Lung Cell Prostacyclin Production," Prostaglandins, 43(5) pp. 413-
433.
[93] Liu, M., Xu, J., Souza, P., 1995, "The Effect of Mechanical Strain on Fetal Rat Lung Cell
Proliferation: Comparison of Two- and Three-Dimensional Culture Systems," In Vitro
Cellular and Developmental Biology - Animal, 31(11) pp. 858-866.
[94] Shimizu, T., Yamato, M., Isoi, Y., 2002, "Fabrication of Pulsatile Cardiac Tissue Grafts
using a Novel 3-Dimensional Cell Sheet Manipulation Technique and Temperature-
Responsive Cell Culture Surfaces." Circulation Research, 90(3) .
[95] Zimmermann, W. -., Melnychenko, I., Wasmeier, G., 2006, "Engineered Heart Tissue
Grafts Improve Systolic and Diastolic Function in Infarcted Rat Hearts," Nature Medicine,
12(4) pp. 452-458.
120
121
[96] Radisic, M., Yang, L., Boublik, J., 2004, "Medium Perfusion Enables Engineering of
Compact and Contractile Cardiac Tissue," American Journal of Physiology - Heart and
Circulatory Physiology, 286(2 55-2) pp. H507-H516.
[97] Kim, B. -., Nikolovski, J., Bonadio, J., 1999, "Cyclic Mechanical Strain Regulates the
Development of Engineered Smooth Muscle Tissue," Nature Biotechnology, 17(10) pp.
979-983.
[98] Shimko, V. F., and Claycomb, W. C., 2008, "Effect of Mechanical Loading on Three-
Dimensional Cultures of Embryonic Stem Cell-Derived Cardiomyocytes," Tissue
Engineering - Part A., 14(1) pp. 49-58.
[99] Okabo, T., and Matsuda, T., 1997, "Hybrid Muscular Tissues: Preparation of Skeletal
Muscle Cell- Incorporated Collagen Gels," Cell Transplantation, 6(2) pp. 109-118.
[100] Clark, C. B., Burkholder, T. J., and Frangos, J. A., 2001, "Uniaxial Strain System to
Investigate Strain Rate Regulation in Vitro," Review of Scientific Instruments, 72(5) pp.
2415-2422.
[101] Vande Geest, J. P., Di Martino, E. S., and Vorp, D. A., 2004, "An Analysis of the
Complete Strain Field within FlexercellTM Membranes," Journal of Biomechanics, 37(12)
pp. 1923-1928.
[102] Fung, Y.C., 1993, "Biomechanics Mechanical Properties of Living Tissues," Springer-
Veriag, New York, .