Mechanical and Electrical Environments to …Mechanical and Electrical Environments to Stimulate Bone Cell Development Gwynne Hannay, BE (Medical)(Hons) Thesis submitted for the degree

Mechanical and Electrical

Environments to Stimulate

Bone Cell Development

Gwynne Hannay, BE (Medical)(Hons)

Thesis submitted for the degree of Doctor of Philosophy

Medical Engineering Program, School of Engineering Systems, Faculty

of Built Environment and Engineering, Queensland University of

Technology, Brisbane, Australia

August 2006

Keywords

PEMF, Pulsed electromagnetic fields, Electrical stimulation, Mechanical stimulation,

Mechanical strain, Cell substrate stretch, Biophysical stimuli, Dual stimuli,

Osteoblast, Bone healing, Electrical currents in bone, Cell proliferation, Cell

differentiation, Cell adhesion, Surface characteristics

II

Abstract

Healthy bone is bombarded with many different mechanical strain derived signals

during normal daily activities. One of these signals is present as a direct connective

tissue strain on the cells. However, there is also the presence of an electrically

charged streaming potential during this straining. The electrical potential is created

from the movement of charged fluid through the small bone porosities. To date, little

focus has been applied to elucidating the possible synergistic effects of these two

stimulants.

The aim of this project was to evaluate the effects of mechanical strain and indirect

electrical stimulation upon the development of bone forming osteoblast cells and any

possible synergistic effects of the two stimulants. This aim was achieved by using a

novel device, designed and developed with the capability of creating a cell substrate

surface strain along with an exogenous electrical stimulant individually or at the

same time. Proliferation and differentiation were determined as a measure of cellular

development.

The indirect electrical stimulation was achieved through the use of a pulsed

electromagnetic field (PEMF) while the mechanical strain was produced from

dynamic stretching of a deformable cell substrate. Strain and strain rate were

modelled from recent studies proposing that relatively high frequency, low strain

osteogenic mechanical stimulants are more indicative of what healthy bone would be

experiencing during normal activities. The PEMF signal mimicked a clinically

available bone growth stimulator signal.

Results showed a PEMF stimulus on monolayers of SaOS-2 and MG-63 osteoblast-

like cells leads to a depression in proliferation. A concomitant increase in alkaline

phosphatase production was also observed for the SaOS-2 cultures, but not for the

MG-63 cell line. It was hypothesised that this was due to the MG-63’s lack of

phenotypic maturity compared to the SaOS-2 cells. Mechanical strain of the cell

substrate alone, at a relatively high frequency (5Hz) but small strain, did not

significantly effect either cell proliferation or differentiation for the MG-63 cells.

III

However, when the electrical and mechanical stimulants were combined a significant

increase in cellular differentiation occurred with MG-63 cultures, revealing a

possible synergistic effect of these two stimulants on the development of bone cells.

IV

Contents KEYWORDS............................................................................................................. II

ABSTRACT............................................................................................................. III

CONTENTS............................................................................................................... V

LIST OF FIGURES ............................................................................................... XII

LIST OF TABLES ................................................................................................XXI

LIST OF ABBREVIATIONS ............................................................................ XXII

STATEMENT OF ORIGINALITY .................................................................XXIII

ACKNOWLEDGEMENTS...............................................................................XXIV

1 INTRODUCTION.............................................................................................. 1

2 BIOELECTRICAL STIMULI.......................................................................... 7

2.1 INTRODUCTION ............................................................................................. 7

2.2 INFLUENTIAL FACTORS IN ELECTRICAL STIMULATION STUDIES ................. 10

2.3 IN VITRO ELECTRICAL STIMULATION ......................................................... 11

2.4 IN VIVO ELECTRICAL STIMULATION ........................................................... 15

2.5 INFLUENCE OF PEMF CHARACTERISTICS ON BIOLOGICAL RESPONSE ........ 17

2.5.1 In Vitro ............................................................................................... 17

2.5.2 In Vivo ................................................................................................ 18

2.5.3 Summary and Conclusion .................................................................. 19

2.6 MECHANISMS OF ACTION............................................................................ 21

2.7 SAFETY ....................................................................................................... 23

2.8 CONCLUSIONS ............................................................................................. 24

3 BIOMECHANICAL STIMULI...................................................................... 26

3.1 IN VITRO MECHANICAL STRAIN.................................................................. 26

3.1.1 Organ culture / Explant studies ......................................................... 26

3.1.2 Fluid flow ........................................................................................... 27

3.1.2.1 Similarity to in vivo strain fields ......................................................28

3.1.3 Substrate Stretching ........................................................................... 28

3.1.3.1 Similarity to in vivo strain fields ......................................................29

V

3.1.4 Hydrostatic Pressure.......................................................................... 30

3.1.4.1 Similarity to in vivo strain field........................................................31

3.2 IN VIVO MECHANICAL STRAIN.................................................................... 31

3.2.1 Cortical Bone ..................................................................................... 34

3.2.2 Cancellous Bone................................................................................. 34

3.2.3 Overload............................................................................................. 35

3.2.4 Computational Mechanobiology ........................................................ 35

3.3 MECHANOTRANSDUCTION PROCESS............................................................ 38

3.4 CONCLUSIONS ............................................................................................. 42

4 CONVERGENCE OF STIMULI ................................................................... 44

4.1 IN VITRO STUDIES....................................................................................... 44

4.2 IN VIVO STUDIES......................................................................................... 45

4.3 CONCLUSIONS ............................................................................................. 47

5 INITIAL PEMF DEVICE ............................................................................... 49

5.1 INTRODUCTION............................................................................................ 49

5.2 PEMF SIGNAL GENERATOR........................................................................ 50

5.2.1 Design Specifications ......................................................................... 51

5.3 PEMF COIL ................................................................................................ 51

5.3.1 Measurement devices used in calibration of coil ............................... 55

5.3.2 Magnetic field from PEMF device ..................................................... 58

5.3.2.1 Magnetic Field map within PEMF coil apparatus............................58

5.3.2.2 Influence of Stainless Steel Shelf on Magnetic Field ......................60

5.3.2.3 Dynamic Magnetic Field Measurements..........................................61

5.3.3 Induced Electric Field from PEMF.................................................... 62

5.3.3.1 Comparison of induced EMF from each PEMF coil apparatus .......63

5.4 PEMF DEVICE CHARACTERIZATION IN CELL GROWTH INCUBATOR........... 64

5.5 COMPARISONS OF PEMF SIGNALS.............................................................. 66

5.6 DISCUSSION................................................................................................. 67

5.7 CONCLUSIONS ............................................................................................. 69

6 PEMF STIMULATION OF CULTURED BONE CELLS .......................... 70

6.1 MATERIALS AND METHODS......................................................................... 71

6.1.1 PEMF Device ..................................................................................... 71

VI

6.1.2 Cell Cultures ...................................................................................... 71

6.1.3 Experiments........................................................................................ 72

6.1.4 Proliferation....................................................................................... 73

6.1.5 Differentiation.................................................................................... 73

6.1.6 Statistical Analysis ............................................................................. 74

6.2 RESULTS ..................................................................................................... 75

6.2.1 Protocol 1........................................................................................... 75

6.2.2 Protocol 2........................................................................................... 76

6.2.3 Protocol 3........................................................................................... 76

6.2.4 Protocol 4........................................................................................... 76

6.3 DISCUSSION ................................................................................................ 78

6.4 CONCLUSIONS ............................................................................................. 80

7 DESIGN OF THE DUAL STIMULUS DEVICE (DSD).............................. 81

7.1 OVERVIEW OF DSD..................................................................................... 82

7.1.1 DSD Design Issues and Requirements............................................... 82

7.1.1.1 Stimuli Consistency .........................................................................82

7.1.1.2 Cell Line Flexibility .........................................................................83

7.1.1.3 Ease of Manufacture ........................................................................83

7.1.1.4 Ease of Maintenance ........................................................................83

7.1.1.5 Growth Media Fluid Flow Strain .....................................................83

7.1.1.6 Mechanical Strain Signal Flexibility................................................83

7.1.1.7 Reliability.........................................................................................84

7.1.1.8 Durability .........................................................................................84

7.1.1.9 Cost ..................................................................................................84

7.1.1.10 Ease of PEMF Integration ...............................................................84

7.1.2 Review of Current Technology........................................................... 85

7.1.2.1 Flow of growth media fluid across cell layer inducing shear strain 87

7.1.2.2 Pressure induced strain of a circular deformable membrane ...........88

7.1.2.3 Stretch of a circular deformable membrane via a piezoelectric

actuator ..........................................................................................................89

7.1.2.4 Uniaxial or biaxial stretch of a deformable membrane....................90

7.1.2.5 Intermittent pressurisation of the gaseous environment in the cell

growth incubator ..............................................................................................90

VII

7.1.3 Evaluation of Design Alternatives...................................................... 91

7.2 DSD DESIGN............................................................................................... 94

7.2.1 Original DSD Design ......................................................................... 94

7.2.2 Revised DSD Design .......................................................................... 95

7.2.3 Parts of DSD Design .......................................................................... 95

7.2.3.1 Base ................................................................................................100

7.2.3.2 Piezoelectric Actuator ....................................................................101

7.2.3.3 Indentor ..........................................................................................102

7.2.3.4 Central Pin Insert............................................................................103

7.2.3.5 Spacer Ring ....................................................................................104

7.2.3.6 Cell Substrate Annulus...................................................................105

7.2.3.7 O-Ring............................................................................................105

7.2.3.8 Deformable Cell Substrate Membrane...........................................105

7.2.3.9 Top .................................................................................................106

7.2.3.10 PEMF Coil and Former..................................................................107

7.2.3.11 Lid ..................................................................................................108

7.3 CONCLUSIONS ........................................................................................... 108

8 SPECIFICATION OF ACTIVE DSD COMPONENTS ............................ 109

8.1 DISPLACEMENT METER AND ITS CALIBRATION......................................... 109

8.2 PIEZOELECTRIC ACTUATOR....................................................................... 110

8.2.1 Operating Principle.......................................................................... 110

8.2.2 Static Calibration ............................................................................. 111

8.2.3 Dynamic Calibration........................................................................ 112

8.2.4 Blocking Force ................................................................................. 114

8.2.5 Conclusions ...................................................................................... 115

8.3 PEMF COIL .............................................................................................. 115

8.3.1 Measurement devices used in calibration of coil ............................. 115

8.3.2 Magnetic field strength from PEMF coil ......................................... 116

8.3.3 Induced EMF.................................................................................... 117

8.3.3.1 Vertical Spectrum...........................................................................117

8.3.3.2 Horizontal Spectrum ......................................................................117

8.4 CONCLUSIONS ........................................................................................... 118

9 CELL SUBSTRATE CHARACTERISATION AND TREATMENT ...... 120

VIII

9.1 CHOICE OF MATERIAL AND ITS PHYSICAL CHARACTERISTICS ................... 121

9.1.1 Tensile tests ...................................................................................... 121

9.1.2 Thickness tests.................................................................................. 123

9.2 SURFACE CHARACTERISTICS..................................................................... 124

9.2.1 X-ray Photoelectron Spectroscopy (XPS) ........................................ 124

9.2.2 Attenuated Total Reflectance (ATR) ............................................... 126

9.3 SURFACE TREATMENT............................................................................... 127

9.3.1 Description of Plasma Set-up and Procedures ................................ 128

9.4 SURFACE CHARACTERISTICS POST-TREATMENT....................................... 132

9.4.1 X-ray Photoelectron Spectroscopy (XPS) ........................................ 132

9.4.2 Attenuated Total Reflectance (ATR) ............................................... 137

9.4.3 Contact Angle................................................................................... 139

9.5 POST-PLASMA TREATMENT OPTIMISATION .............................................. 140

9.5.1 Treatment Methods .......................................................................... 142

9.5.2 Results and Discussion..................................................................... 144

9.5.3 Cell Counts....................................................................................... 150

9.5.4 Conclusions...................................................................................... 154

9.6 DISCUSSION .............................................................................................. 154

9.7 CONCLUSIONS ........................................................................................... 155

10 DSD SURFACE STRAIN CHARACTERISATION.............................. 157

10.1 EXPERIMENTAL SURFACE STRAIN............................................................. 157

10.1.1 Methods............................................................................................ 157

10.1.1.1 Static Experimental Strain .............................................................160

10.1.1.2 Dynamic Experimental Strain........................................................162

10.1.2 Results .............................................................................................. 162

10.1.2.1 Large Static Experimental Strain...................................................162

10.1.2.2 Small Static Experimental Strain...................................................164

10.1.2.3 Dynamic Strain Particle Tracking .................................................168

10.1.3 Discussion and Conclusions ............................................................ 168

10.2 THEORETICAL SURFACE STRAIN ............................................................... 171

10.2.1 Methods............................................................................................ 171

10.2.1.1 Sensitivity Analysis .......................................................................171

10.2.2 Results .............................................................................................. 172

IX

10.3 COMPARISON OF EXPERIMENTAL AND THEORETICAL SURFACE STRAIN ... 175

10.4 DISCUSSION............................................................................................... 178

10.5 CONCLUSIONS ........................................................................................... 179

11 DUAL STIMULATION OF CULTURED BONE CELLS .................... 181

11.1 INTRODUCTION.......................................................................................... 181

11.2 MATERIALS AND METHODS....................................................................... 182

11.2.1 Dual Stimulus Device (DSD)............................................................ 182

11.2.2 Experiments...................................................................................... 182

11.2.2.1 Mechanical Stimulation .................................................................183

11.2.2.2 Electrical Stimulation.....................................................................183

11.2.3 Cell Cultures .................................................................................... 184

11.2.4 Proliferation ..................................................................................... 186

11.2.5 Differentiation .................................................................................. 188

11.2.6 Statistical Analysis ........................................................................... 188

11.3 RESULTS.................................................................................................... 189

11.3.1 Proliferation ..................................................................................... 189

11.3.2 Differentiation .................................................................................. 191

11.4 DISCUSSION............................................................................................... 192

11.5 CONCLUSION............................................................................................. 196

12 DISCUSSION, FUTURE WORK AND CONCLUSIONS..................... 197

12.1 FUTURE WORK.......................................................................................... 199

12.2 CONCLUSIONS ........................................................................................... 202

APPENDIX A: RAW DATA FROM IN VITRO CELLULAR

EXPERIMENTATION.......................................................................................... 204

APPENDIX B: DUAL STIMULUS DEVICE (DSD) ENGINEERING

DRAWINGS ........................................................................................................... 206

APPENDIX C: GAS PLASMA CELL SUBSTRATE SURFACE

MODIFICATION PROCEDURE ........................................................................ 219

APPENDIX D: FINITE ELEMENT ANALYSIS RESULTS............................ 220

APPENDIX E: RESEARCH PRESENTATIONS AND PUBLISHED

MATERIAL............................................................................................................ 238

X

REFERENCES....................................................................................................... 240

XI

List of Figures

Figure 5-1 A diagram of the induced electric field trace from a clinically available

PEMF device that has been simulated for this project (Reproduced from Bassett,

C.A. (1989) Crit Rev Biomed Eng 17(5): 453). ................................................. 50

Figure 5-2 A diagram of the Initial PEMF Device. Shown are the two coil

apparatuses, pulse generator and voltage power supply. ................................... 52

Figure 5-3 The plastic tray shown is positioned underneath the PEMF coil during

calibration and cellular experimentation. ........................................................... 53

Figure 5-4 A picture of the gauss/tesla meter used for magnetic field measurements

during testing with incubator tray. ..................................................................... 53

Figure 5-5 Set up of initial PEMF device during biological testing. ......................... 54

Figure 5-6 The real time oscilloscope used to measure magnetic field and induced

electric fields and the DC power supply used for PEMF device. ...................... 56

Figure 5-7 The coil probe dosimeter, with shunt resistor shown on left, used to record

induced electric fields from PEMF coils............................................................ 56

Figure 5-8 A close up of the coil from the coil probe dosimeter. .............................. 57

Figure 5-9 Set up of initial PEMF device during calibration testing. ........................ 57

Figure 5-10 A top view of the PEMF coil showing marked positions used for

calibration measurements of the magnetic field................................................. 59

Figure 5-11 A side view of the PEMF coil, with the level of the magnetic field

measurements marked with a black dotted line. ............................................... 59

Figure 5-12 Maximum magnetic field with distance from centre of PEMF coil.

Shown are measurements along the centre line (25 and 50mm from centre of

coil; black line) of the PEMF coil apparatus and at the culture well positions

(~11 and ~32mm from centre of coil; grey line). See Figure 5-10 for diagram of

positions. ............................................................................................................ 60

Figure 5-13 Maximum magnetic field with each cell culture well position.

Measurements were taken when the PEMF apparatus was placed on a non-

metallic bench (black line) or on the metallic incubator tray (grey line). .......... 61

Figure 5-14 Comparison of dynamic magnetic field measurements from PEMF coil

apparatus 1 and 2. No discernable difference is seen between each trace. ........ 62

XII

Figure 5-15 The induced electric field trace in the coil probe dosimeter when placed

within the active PEMF coil............................................................................... 63

Figure 5-16 A comparison of the induced electric field from the PEMF coil apparatus

1 and 2. No discernable difference is seen between the traces. ......................... 64

Figure 5-17 A comparison of the induced electric field from the PEMF signal when

the device was located in the cell growth incubator and on the laboratory bench

during calibration. .............................................................................................. 65

Figure 5-18 A comparison of the magnetic field at PEMF exposed cell culture wells

(trace ‘a’) and those in the control cell culture plate (trace ‘b’)......................... 66

Figure 5-19 A comparison of the ideal and the induced voltage measured from the

PEMF device. Trace 'a' (dotted line) signifies the desired waveform while Trace

'b' (solid line) is the recorded signal................................................................... 67

Figure 6-1 PEMF exposure protocols on osteoblast-like SaOS-2 cells to quantify

effects on cellular proliferation and differentiation. Shaded sections denote

PEMF exposure while clear sections denote normal cell culture conditions..... 72

Figure 6-2 Proliferation, described as a percentage to controls, of PEMF exposed cell

cultures from each PEMF exposure protocol at 25,000 and 50,000 cells per well

seeding density. # Indicates statistical significance (P < 0.05). Error bars are +/-

standard error of the mean. ................................................................................ 77

Figure 6-3 Differentiation, described as a percentage to controls, of PEMF exposed

cell cultures from each PEMF exposure protocol at 25,000 and 50,000 cells per

well seeding density. Error bars are +/- standard error of the mean. ................. 77

Figure 7-1 Diagrams of previously reported in vitro cell straining devices. a) fluid

flow shear strain b) pressure induced substrate strain c) piezoelectric actuated

substrate strain d) direct mechanical straining of substrate e) hydrostatic gas

pressurization of culture environment................................................................ 86

Figure 7-2 The assembled dual stimulus device (DSD) without PEMF Coil and

Former or Lid..................................................................................................... 96

Figure 7-3 A diagram of the assembled dual stimulus device (DSD) without Lid. ... 96

Figure 7-4 A diagram of the fully assembled dual stimulus device (DSD). ............. 97

Figure 7-5 A diagram of the fully assembled dual stimulus device (DSD) with Lid

cut away to reveal cell culture well.................................................................... 97

Figure 7-6 A diagrammatical cross section of the assembled DSD. Boxed area on top

diagram is shown above with part numbers labelled. ........................................ 98

XIII

Figure 7-7 A diagrammatical exploded view of the cross-sectioned DSD, with part

numbers shown. Area 1 and 2 are used for adhesive containment and are

discussed in Section 7.2.3.1 and 7.2.3.3............................................................. 99

Figure 7-8 The Base from the dual stimulus device. See Section 7.2.3.1 for

description of numbers. .................................................................................... 101

Figure 7-9 The Piezoelectric Actuator used in the dual stimulus device................. 102

Figure 7-10 Diagram of Indentor from dual stimulus device. ................................. 103

Figure 7-11 Diagram of Central Pin Insert from dual stimulus device. .................. 104

Figure 7-12 Diagram of Spacer Ring from dual stimulus device. ........................... 104

Figure 7-13 Cell Substrate Annulus (clear) with attached O-Ring (black) from dual

stimulus device................................................................................................. 105

Figure 7-14 Silicone material used as cell substrate for dual stimulus device. Shown

is the Deformable Cell Substrate Membrane attached to a polymer backing

sheet.................................................................................................................. 106

Figure 7-15 Top from dual stimulus device. ............................................................ 107

Figure 7-16 PEMF Former from dual stimulus device. Shown without Copper Wire

Coils. ................................................................................................................ 107

Figure 7-17 Diagram of Lid from dual stimulus device........................................... 108

Figure 8-1 Piezoelectric Actuator displacement output with applied voltage when

unconstrained (but glued to the DSD Base; ‘Unassembled Device’) or

constrained (when the DSD was fully assembled; ‘Assembled Device’) with

attachment of DSD Top. Maximum rated voltage was +/- 180V..................... 111

Figure 8-2 Calibration displacement meter output voltage range with increasing

actuator driving voltage frequency................................................................... 113

Figure 8-3 The piezoelectric actuator output displacement with increasing blocking

force.................................................................................................................. 114

Figure 8-4 The magnetic field strength of the PEMF coil used in the DSD with an

increasing vertical distance from the cell substrate.......................................... 116

Figure 8-5 Maximum peak-to-peak voltage range of induced electric field in coil

probe dosimeter with increasing vertical distance from the cell substrate surface.

.......................................................................................................................... 117

Figure 8-6 Maximum peak-to-peak voltage range of induced electric field in coil

probe dosimeter with increasing horizontal distance from the centre of the

PEMF Coil in the DSD..................................................................................... 118

XIV

Figure 9-1 The typical stress vs strain normalised to 100% elongation for materials

cut in three different directions. Regression lines are shown with linear equation

and R2 values noted for each specimen angle. Slopes of the equations were used

to determine the material’s elastic modulus..................................................... 122

Figure 9-2 Typical tensile force vs elongation of PDMS silicone cell substrate

membrane material when taken to breaking point. Specimens were cut from

three different angles in the membrane sheet................................................... 123

Figure 9-3 An XPS survey scan of native PDMS cell substrate material with electron

volt binding energy peaks at 23(Oxygen 2s), 100(Silicon 2p), 150(Silicon 2s),

282(Carbon 1s) and 530(Oxygen 1s). .............................................................. 126

Figure 9-4 An Attenuated Total Reflectance (ATR) spectrum scan of native PDMS

cell substrate material with wave number (cm-1) peaks of interest at 3700-

3000(Hydroxyl groups), 1000(Silicon-Oxygen stretching), 1260(Si-CH3 stretch)

and 780(Si-CH3 stretch). .................................................................................. 127

Figure 9-5 Gas Plasma Machine showing glow from reactive gas.......................... 129

Figure 9-6 Gas Plasma Machine showing full length of vacuum chamber and

associated parts. ............................................................................................... 130

Figure 9-7 An XPS survey scan of native and 5 watt plasma treated PDMS cell

substrate material. Variation in concentration of material constituents between

the two is seen in the differing electron volt binding energy peak heights...... 135

Figure 9-8 An XPS detailed scan of carbon 1s for native and 5-watt plasma treated

PDMS cell substrate material. The decrease in peak height for the treated

substrate signifies a loss of carbon 1s. ............................................................. 135

Figure 9-9 An XPS detailed scan of oxygen for native and 5-watt plasma treated

PDMS cell substrate material. The increase in peak height for the treated

substrate signifies a gain in oxygen while the shift in binding energy is due to a

change in binding properties. ........................................................................... 136

Figure 9-10 An XPS detailed scan of silicone 2p for native and 5-watt plasma treated

PDMS cell substrate material. The decrease in peak height for the treated

substrate signifies a loss of Silicon 2p while the second binding energy peak is

due to an additional bind between silicon and oxygen (Si-O). ........................ 136

Figure 9-11 An ATR spectrum scan for native and 5-watt plasma treated PDMS cell

substrate material. Little to no difference can be seen between the two traces at

XV

wave number (cm-1) peak 3700-3000(Hydroxyl groups), 1000(Silicon-Oxygen

stretching), 1260(Si-CH3 stretch) and 780(Si-CH3 stretch). ............................. 137

Figure 9-12 An ATR detailed spectrum scan of the hydroxyl region (3700-3000cm-1)

for native and 5-watt plasma treated PDMS cell substrate material. ............... 138

Figure 9-13 An ATR detailed spectrum scan of the amine group region (1700-

1500cm-1) for native and 5-watt plasma treated PDMS cell substrate material.

.......................................................................................................................... 139

Figure 9-14 Water droplet contact angle for native and 5-watt plasma treated PDMS

cell substrate material. Error bars are ± standard error mean........................... 140

Figure 9-15 The post-plasma treatment protocols. Protocol 1 is negative control and

Protocol 5 is positive control. All protocols were conducted over nine days with

each day being a treatment of plasma, media, air or UV while over the last four

days cellular attachment and proliferation counts were conducted.................. 143

Figure 9-16 Photo of PDMS cell substrate after cells had attached. Also showing is

the numbered location of each cell count. Circled numbers indicate

predetermined areas used for cell counting...................................................... 144


for each post-plasma treatment protocol on PDMS cell substrate material, with

protocol numbers attached to each line. ........................................................... 145


for protocol 1 and plasma treated PDMS cell substrate material. Differences

signify the hydrophobic recovery of the surface after air contact.................... 146


1500cm-1) for each post-plasma treatment protocol on PDMS cell substrate

material, with protocol numbers attached to each line..................................... 147


1500cm-1) for protocol 1 and plasma treated PDMS cell substrate material. .. 148

Figure 9-21 The relative concentration of hydroxyl groups (area ratios) on the

surface of PDMS cell substrate material after each post-plasma treatment

protocol............................................................................................................. 149

Figure 9-22 The relative concentration of amine groups (area ratio) on the surface of

PDMS cell substrate material after each post-plasma treatment protocol. ...... 150

XVI

Figure 9-23 The number of cells attached to the surface of the PDMS cell substrate

for each post-plasma treatment protocol and native PDMS over 72 hours. #

Indicates significant difference from other protocols (p<0.05). ...................... 152

Figure 9-24 The percentage increase in number of cells attached to the surface of the

PDMS cell substrate over initial attachment counts for each post-plasma

treatment protocol and native PDMS over 72 hours. # Indicates significant

difference from other protocols (p<0.05)......................................................... 153

Figure 10-1 Holes drilled in polymer template used for marking ink dots on cell

substrate during strain calibration. ................................................................... 158

Figure 10-2 Specialised jig for tightly gripping the ink marking template during

CMC high speed drilling. This jig also includes a punch for accurate removal of

the template after drilling was completed. ....................................................... 158

Figure 10-3 A diagram of cell substrate and locations of ink dots used for

experimental strain calculations. Position 1 on subsequent figures for radial

strain is defined as the displacement/strain between the centre ink dot (c) and

ink dot 1 (1). This follows through to point 9. Circumferential strain at position

1 is defined as the displacement/strain between ink dot 1 (1) and ink dot 1’ on

the rotated axes (1’).......................................................................................... 159

Figure 10-4 A diagram of cell substrate during activation of the DSD. The diagram

shows bending membrane as dotted lines. Camera loaction resulted in correction

factors for measured strain to be implemented. ............................................... 161

Figure 10-5 Radial strain at different radial positions on cell substrate when substrate

membrane is deformed at different heights. See Figure 10-3 for description of

radial positions. ................................................................................................ 163

Figure 10-6 Radial strain with differing cell substrate deformations over all radial

positions on cell substrate. See Figure 10-3 for description of position number.

.......................................................................................................................... 163

Figure 10-7 Radial strain at different radial positions on cell substrate for two

membrane deformation heights of 0.1mm and 0.5mm. See Figure 10-3 for

description of position number. ....................................................................... 165

Figure 10-8 Circumferential strain at different radial positions on cell substrate for

two membrane deformation heights of 0.1mm and 0.5mm. See Figure 10-3 for

description of position number. ....................................................................... 165

XVII

Figure 10-9 Relationship between radial strain and membrane deformation for each

radial position on cell substrate. Regression equation lines are placed in order

from position 1 to 9. See Figure 10-3 for description of position number....... 166

Figure 10-10 Relationship between circumferential strain and cell substrate

membrane deformation for each radial position on cell substrate. Regression

equation lines are placed in order from position 1 to 9. See Figure 10-3 for

description of position number......................................................................... 166

Figure 10-11 Radial strain vs radial position for cell substrate. These values were

interpolated strains for 72µm central pin displacement, calculated from the

regression lines in Figure 10-9 and Equation 10-3. See Figure 10-3 for


Figure 10-12 Circumferential strain vs radial position for cell substrate. These values

were interpolated strains for 72µm central pin displacement, calculated from the

regression lines in Figure 10-10 and Equation 10-3. See Figure 10-3 for


Figure 10-13 Close up of cell substrate tethering with the DSD Top, Cell Substrate

Annulus and O-Ring. See Figure 7-6 for full cross sectioned DSD. ................ 170

Figure 10-14 Finite Element Analysis of maximum in-plane principal strain with

radial position from the centre (0) to the edge (10) of the cell culture well when

undergoing actuation in the DSD. .................................................................... 172

Figure 10-15 A Finite Element Analysis colour contour plot of cell substrate in plane

principal surface strains.................................................................................... 173

Figure 10-16 A Finite Element Analysis of circumferential strains on cell substrate

with radial position from the centre (0) to the edge (10) of the cell culture well

when undergoing actuation in the DSD. Variation of peak at point 0.5 is due to

inconsistencies in the FEA mesh...................................................................... 174

Figure 10-17 A Finite Element Analysis maximum strain vector plot of the dual

stimulus device surface during active deformation. The red arrows signify an in

plane principal strain as the predominant strain present. ................................. 174

Figure 10-18 A graph of radial strain with radial position comparison between

theoretical (green line) and experimental (red dots) studies. The central pin

displacement was set at 0.1mm. ....................................................................... 176

XVIII

Figure 10-19 A graph of radial strain with radial position comparison between

theoretical (red line) and experimental (green dots) studies. The central pin

displacement was set at 0.5mm........................................................................ 176

Figure 10-20 A graph of circumferential strain with radial position comparison

between theoretical (green line) and experimental (red dots) studies. The central

pin displacement was set at 0.1mm.................................................................. 177

Figure 10-21: A graph of circumferential strain with radial position comparison

between theoretical (green line) and experimental (red dots) studies. The central

pin displacement was set at 0.5mm.................................................................. 177

Figure 11-1 Timing of stimulant/s from the dual stimulus device (DSD) during 3-day

experimentation protocol. Shaded region signifies the activation of the DSD

stimulant/s. ....................................................................................................... 183

Figure 11-2 Timing of dual stimulants during activation of dual stimulus device.

Trace A represents the PEMF signal's repetitive pulse burst. Trace B represents

the mechanical deformation (and hence strain) of the cell substrate membrane.

.......................................................................................................................... 184

Figure 11-3 The specially designed cell culture well inserts for use in control

cultures. These effectively reduced the cell growth area to match that of the

DSD.................................................................................................................. 185

Figure 11-4 Control cell culture plates with well inserts push fit into position for

experimentation................................................................................................ 186

Figure 11-5 The raw absorbance results for LDH measured proliferation from each

method of DSD stimulation. # Represents statistical significance (p < 0.05).

Error bars are +/- standard error of the mean................................................... 190

Figure 11-6 The raw absorbance results for pNPP measured differentiation from each

method of DSD stimulation. # Represents statistical significance (p < 0.05).

Error bars are +/- standard error of the mean................................................... 191

Figure 11-7 Percentage change in proliferation and differentiation with respect to

controls for each method of DSD stimulation. Error bars were computed from

the addition of percentage errors in the original raw data presented in Figure 11-

5 and Figure 11-6. # Represents statistical significance (p < 0.05). ................ 192

Figure A - 1 Base; General View............................................................................. 207

Figure A - 2 Base ..................................................................................................... 208

XIX

Figure A - 3 Base; Detailed Drawing....................................................................... 209

Figure A - 4 Indentor; General View....................................................................... 210

Figure A - 5 Indentor ............................................................................................... 211

Figure A - 6 Central Pin Insert ................................................................................ 212

Figure A - 7 Spacer Ring.......................................................................................... 213

Figure A - 8 Cell Substrate Annulus ........................................................................ 214

Figure A - 9 Top; General View............................................................................... 215

Figure A - 10 Top ..................................................................................................... 216

Figure A - 11 PEMF Coil Former ........................................................................... 217

Figure A - 12 Lid ...................................................................................................... 218

Figure C - 1 Young’s Modulus = 1.75MPa.............................................................. 221

Figure C - 2 Young's Modulus = 2MPa ................................................................... 222

Figure C - 3 Young's Modulus = 2.25MPa .............................................................. 223

Figure C - 4 Young's Modulus = 2.5MPa ................................................................ 224

Figure C - 5 Poisson's Ratio = 0.35.......................................................................... 225

Figure C - 6 Poisson's Ratio = 0.5............................................................................ 226

Figure C - 7 Cell Substrate Thickness = 20µm ........................................................ 227



Figure C - 10 Cell Substrate Thickness = 60µm ...................................................... 230

Figure C - 11 Cell Substrate Thickness = 75µm ...................................................... 231

Figure C - 12 Initial Central Pin Displacement = 10µm .......................................... 232




Figure C - 16 Central Pin Insert Displacement = 100µm........................................ 236

Figure C - 17 Central Pin Insert Displacement = 500µm........................................ 237

XX

List of Tables

Table 7-1 Evaluation of In vitro mechanical straining techniques for dual stimulus

device design...................................................................................................... 93

Table 9-1 X-ray Photoelectron Spectroscopy (XPS) results from native PDMS

membrane, where columns 2 to 8 represent the following: X axis position

(binding energy); Full width at half - maximum (FWHM); Raw area underneath

peak (CPS); Relative sensitivity factor (RSF) – Used in calculating atomic

concentration; Atomic mass of element; Atomic Concentration (%); Mass

Concentration (%). ........................................................................................... 125

Table 9-2 XPS results from gas plasma treated PDMS membranes at differing

powers for determination of ideal plasma process. Columns are as defined in

Table 9-1. ......................................................................................................... 133

XXI

List of Abbreviations

AC – Alternating Current

αMEM – Minimum Essential Medium, Alpha Medium

B – Magnetic Field

BMP – Bone Morphogenetic Protein

BMU – Basic Multicellular Unit

Ca2+ – Calcium Ion

DC – Direct Current

DSD – Dual Stimulus Device

E – Electric Field

E(t) – Time Varying Electric Field

EBI – ElectroBiology Incorporated

ECM – Extracellular Matrix

EDTA – Ethylinediamine Tetra-acetic Acid, Disodium Salt

EMF – Electromotive Force

FEA – Finite Element Analysis

FCS – Foetal Calf Serum

G – Gauss (Tesla * 10-4)

HBSS – Hanks Balanced Salt Solution

LDH – Lactate Dehydrogenase

mRNA – Messenger Ribonucleic Acid

mT – Milli Tesla

NO – Nitric Oxide

NOS – Nitric Oxide Synthase

PDMS – Polydimethylsiloxane

PEMF – Pulsed Electromagnetic Field

pNPP – p-Nitrophenyl Phosphate

PTH – Parathyroid Hormone

V – Volts

µε – Microstrain (strain * 10-6)

XXII

Statement of Originality

The work contained in this thesis has not been previously submitted for a degree or

diploma at any other higher education institution. To the best of my knowledge and

belief, the thesis contains no material previously published or written by another

person except where due reference is made.

Signature: ___________________________

Date : ___________________________

XXIII

Acknowledgements

Throughout the last three and a half years of undertaking my doctorate I have been

blessed with the assistance of many different people. My sincere thanks and

appreciation must be extended to those people, as they have significantly helped me

along the path of this project and whom without I would have suffered. The support

they have provided goes well beyond the technical aspects of the project.

To my principal supervisor Professor Mark Pearcy, thank you for your continual

support and understanding in all facets of my project. You have provided many very

helpful and supportive comments that kept me on the correct path towards the final

hurdle. Your ability to ask the right question at the right time was very useful! To my

associate supervisor Dr David Leavesley, I appreciate the enthusiasm you have

shown in involving yourself with the engineering aspects of this project and how

they apply to cell biology. I also wish to thank you for the biological coaching you

have provided to me over the years.

I would like to thank Professor John Evans for the many encouraging and extremely

helpful comments you made to me. Many a time I thought I had come up against an

insurmountable problem at which point you provided me with a deceptively simple

solution! Thanks to Barry Wood in the Future Materials research centre at the

University of Queensland who answered a lot of questions and was willing to spend

time troubleshooting with me. Special thanks are extended to the technical and

workshop staff of Greg Tevlen, Terry Beach, Wayne Moore, Jon James and Mark

Hayne. They all contributed in solving the (numerous!) technical problems and went

a long way to helping me complete the successful device design. Work was

completed promptly and with cheerfulness, even in times of great workloads.

Acknowledgment must be made of the support and helpful comments received from

my fellow postgraduate students, with special mention made to Shobha and

Sanjleena (thank you for the lively lab conversations) and Cam Wilson (provided

answers to the endless questions I posed to him).

XXIV

Many thanks to my mother-in-law, Dana, who has given vast amounts of her time to

help support my wife and I in this very hectic time of thesis writing and looking after

a new baby. Much thanks to Dad and Sandi for reviewing the thesis at the last minute

and my Mum, brother Jeremy and Grandad for the many kind words of support and

love over the last 4 and a half years. You have always believed in my ability, even

when I didn’t.

Most important of all is the endless and loving support from my wife, Laila. She has

been the source of my strength throughout this project, providing me with enormous

amounts of encouragement to continue working through the many problems that

cropped up along the way and has repeatedly helped me get ‘back on track’! Thank

you from the bottom of my heart! Additionally, I would like to send out my love to

our beautiful daughter Eliri. When I look into your eyes I know that it was all worth

it!

Lastly, I would like to dedicate this thesis to the two most special people in my life,

Laila and Eliri.

XXV

Chapter 1: Introduction

1 Introduction

The form and function of bone is mediated by its loading history. Increased loading

of bone leads to hypertrophy while disuse leads to atrophy. Tennis players and other

sports people who repeatedly impart high mechanical loads on the same bone find

increased mass at and around the muscle attachment points due to the mechanical

stimulus. In contrast to this is the situation experienced by astronauts who undergo

long periods of weightlessness and hence a decreased loading of bone. Studies

conducted before and after space flight have shown there is a significant reduction in

bone mass over the short period of travel in orbit. However, more common is the loss

of bone from prolonged bed rest. Concomitant with the mechanical loading of bone

is the presence of an internal electric field.

It has been shown that each of these environments modify bone, however the

synergistic influence of these two factors has not been widely studied.

The rational for this research project stems from an effort to reproduce the electrical

and mechanical environment bone cells experience in vivo in an in vitro cell culture.

Literature has revealed the main influence on normal bone maintenance in vivo is

exogenous mechanical stimulation such as walking and running. The fundamental

mechanism by which this controls normal homeostasis of bone is unknown, however

the endogenous electrical signal and cellular level mechanical strain that are created

during loading of bone (walking, running, etc.) are the likely stimuli that initiate

cellular responses (Turner and Pavalko, 1998; Pilla, 2002b).

The levels of each of these stimuli during normal locomotion are:

• Extracellular matrix mechanical strain in the range of 0-3000µε (0 – 0.3%)

(Lanyon, 1984; Fritton et al., 2000).

• Electric signals in range of 0.1-10mV/cm pulsed at frequencies of <10Hz

with a distorted trapezoid pulse shape (Pilla, 2002b).

Evidence is emerging that both these stimuli modulate an identical sub-cellular

chemical pathway (Brighton et al., 2001). While literature has focused on one or the

1


other of these stimuli, no published work quantifies the response of a dual

mechanical and electrical stimulus on in vitro cultures of bone cells. This is the

motivation for this research project, which aims to induce electrical stimuli indirectly

and mechanical stimuli directly, through a novel device producing both stimulants in

isolation or in unison. The electrical stimulation is created by way of pulsed

electromagnetic fields (PEMFs) while the mechanical stimulation is induced via

stretching of the surface the cells are attached to.

Research Aim:

To clarify the effects of pulsed electromagnetic field stimulation and mechanical

strain stimulation in isolation and in unison on the development of osteoblast cell

cultures.

Bone is a vascularized connective tissue, consisting primarily of cells in an extensive

matrix of collagen fibres (organic material) and hydroxyapatites (inorganic material).

Its duty is to support, maintain form of the body and provide muscle attachment

points. A process of ossification forms these inorganic and organic components of

bone.

Bone is comprised of cancellous and cortical tissue. Cancellous bone is spongy in

nature and has irregularly arranged lamellae with very high porosity, while the

cortical, or compact, bone contains tightly packed osteons that are a circular, layered

type of structure with lacunae (porosities containing bone cells called osteocytes)

regularly dispersed in-between each of the layers. Each lacuna is connected with

numerous canaliculi; canals that contain osteocyte ‘fingers’ called cell processes

(Kamioka et al., 2001), and serves to transport nutrients to and waste from the

osteocytes to the Haversian canals (centre of the osteon) containing a vascular

supply.

Osteoblast and Bone Lining cells are bone-forming cells that secrete and deposit

bone extracellular matrix (collagen and calcium phosphate minerals). They are

located on all the bone’s internal surfaces: the lamellae in cancellous bone, Haversian

canals in cortical bone and also in the bone forming regions of the periosteum (outer

2


layer) and endosteum (inner layer). The osteocyte processes that sit in the canaliculi

are actin rich bundles that are 200 times stiffer than the cell body (Tanaka-Kamioka

et al., 1998) and have transverse proteoglycan elements connecting it to the

canalicular wall (Shapiro et al., 1995).

Long bones, such as the femur, are formed by endochondral ossification. During this

process, cartilage cells (chondrocytes) proliferate, hypertrophy and then undergo

apoptosis (cell death). This releases angiogenic factors such as vascular endothelial

growth factor (VEGF) that promote the ingrowth of the vascular supply into the

ossification centre, allowing osteoclasts and osteoblasts to begin the dissolution of

the calcified cartilage and the laying down of new bone respectively (Olsen et al.,

2000). At this stage there is bone modelling with little to no remodelling.

Modelling is defined as either the formation or resorption of bone at a particular site.

It occurs predominantly at the periosteal and endosteal surfaces. The ‘modelling

threshold’ is the mechanical strain threshold, as measured on the surface of bone, that

when exceeded will result in modelling of bone (net gain in bone) and below will

turn modelling off (no net gain in bone). This is centred on approximately 1000µε

(Frost, 2001). In contrast, remodelling continues throughout life and is a coupled

process of resorption and then formation of bone on the Haversian and trabecular

surfaces to maintain bone mass and healthy tissue. The ‘remodelling threshold’ is the

threshold that when exceeded will result in conservation mode remodelling of bone

(no net gain in bone) and below will turn disuse mode remodelling on (net loss in

bone mainly due to decreases in the formation of trabeculae and endocortical bone

after osteoclastic resorption). This is centred on approximately 100µε (Frost, 2001).

This remodelling process, which continues throughout life, is mainly controlled by

extracellular stimulation such as applied loading but is also affected by factors like

hormones, calcium, vitamin D and genes. These secondary factors may influence

around 10% of a bone’s postnatal strength while mechanical loading determines 40%

(Frost, 2001). Therefore the bone is able to structurally adapt to withstand the applied

strain as is shown within the trabeculae bone matrix of the femur neck, where linear

compressively loaded bands of bone matrix are positioned in line with the direction

of the compressive upper body weight force, thus reducing tissue strain.

3


Bone contains basic multicellular units (BMUs) that are a team of osteoclasts and

osteoblasts that dissolve old bone and then deposit new osteoid (uncalcified bone

matrix) respectively. The osteoblasts initiate the remodelling process when they

receive the signal from the osteocytes by secreting RANK-ligand which remains

bound to their cell membrane, activating cells of the monocyte-macrophage lineage

to fuse and differentiate into mature multinucleated osteoclasts that resorb the bone.

After this, more osteoblasts are recruited to start forming osteoid. When the layer of

osteoid reaches around 6µm in thickness then osteoblast-regulated calcification takes

place (Ott, 1998). Initially, nucleation of a calcium phosphate ion forms a crystal by

either super-saturation or attachment of the ion to a non-collagenous protein such as

osteonectin. This is followed by growth of the crystal (Sikavitsas et al., 2001).

Martin (Martin, 2000b) proposes the theory that BMUs transient refilling rate is due

to an inhibitory signal being sent from a newly formed osteocyte to a bone forming

osteoblast in close proximity, inhibiting its apposition of osteoid and differentiating it

into an osteocyte to be surrounded by bone. This theory correctly predicts differences

between osteonal and surface BMU apposition rates from previous experiments but

is still somewhat controversial (Martin, 2000a).

Mechanical stimulation of the in vivo bone causes three main mechanical stimulants

to occur: extracellular matrix strain, intra-porosity pressure and fluid flow in the

canaliculi and lacuna. These three are discussed in detail in Chapter 3, which reviews

past in vivo and in vitro research utilising a mechanical stimulant on bone tissue.

Relatively recent work on efforts made to mathematically model bone in response to

a mechanical strain is also covered. Mechanotransduction, or the pathway a

mechanical stimulant takes to transduce the applied signal into a biological result, is

reviewed in detail. Different theories of this process are also presented.

Fluid flow within the bone porosities is derived from the mechanical load. However

as discussed in Chapter 2, this flow also creates an electrical stimulus for the bone

cells (osteocytes). When early researchers first observed the creation of an electrical

potential difference across the shaft of longs bones that were undergoing mechanical

strain, it was assumed the bone structure itself was the cause (Pauwels, 1941).

However it was not until later that the true cause of the electrical current was

4


discovered to derive from the flow of charged species in the bone fluid (Chakkalakal

and Johnson, 1981).

Chapter 2 reviews the use and applications of electrical stimulation on in vivo and in

vitro studies of bone. Additionally, the effect of the electrical stimulant’s

characteristics on the cell’s biological response is reviewed. This chapter also

discusses the influential variables in vitro cell culture studies are required to either

control or properly define when reporting results. A discussion on the possible

transduction mechanisms electrical stimulants take is also covered.

Similarities in these transduction pathways with those from mechanical stimulation

are reviewed and discussed in Chapter 4. In vivo studies employing both stimulants

have been conducted and provide some interesting insight into a possible hierarchical

role present in the transduction pathway. Cellular level interactions and similarities

are also discussed and possible synergistic mechanisms of action are hypothesised.

To elucidate an electrical stimulus effect on the development of bone cells, a device

was designed to stimulate cultures with an externally applied pulsed electromagnetic

field (PEMF). Chapter 5 covers the development of this device and its associated

calibration. Measures of its magnetic and inductive electrical properties are made

with comparisons between signals when measured on a laboratory bench and when

in the cell culture environment during experimental testing.

Application of the PEMF to monolayer cell cultures induces an electrical current

within the cell layer. The resulting phenotypic state of the cells after this stimulation

was applied is quantified via measures of proliferation and differentiation compared

to unexposed control cultures. Timing of the PEMF stimulant within a 3-day cellular

experimentation protocol was also varied to elucidate any effects this factor may play

on the final results (Chapter 6).

Having established the effects of indirect electrical stimulation on the bone cells, a

novel dual stimulus device capable of imparting a mechanical cell substrate strain

and an electrical stimulus to the attached cells was designed, developed and

produced. Chapter 7 outlines the design criteria and the relevant technology currently

5


available with which this device could have been produced from. These options are

evaluated and the selected design is detailed. Chapter 8 outlines the specifications of

the active components of the dual stimulus device.

A particularly large segment of the device’s design involved characterisation,

treatment and optimisation of the cell substrate used, which are covered in Chapter 9.

Following on from this, experimental determination of the substrate surface strain is

outlined and discussed, with comparisons made to theoretical expectations (Chapter

10).

To test and establish a result for the main aim of the thesis, the dual stimulus device

was experimentally tested with cell cultures of osteoblasts and presented in Chapter

11. Electrical stimulation alone via PEMF exposure was studied again to facilitate

comparisons with the original data discussed in Chapter 6. A different cell line was

used which made for a more detailed and insightful discussion on the influence of the

cell’s phenotypic state in transduction of an external electrical stimulation.

Mechanical strain alone and the dual stimulus of mechanical and electrical

stimulation were tested and comparisons among all three sets of data are made.

Discussions on the possible meaning of these results in relation to proposed

mechanisms of action are outlined (previously discussed in Chapter 4).

Collating all the experimental results, Chapter 12 provides indications of the

influence a mechanical and electrical stimulus applied individually or in unison has

upon the development of osteoblastic cell cultures. Some technical refinements of

designs and experimental procedures are recommended for future work and

conclusions are drawn with reference to the collected data and the reviewed research.

Development of the dual stimulus device was a major component of this project

while the subsequent cell tests performed (discussed in Chapter 11) were limited to

small replicate numbers demonstrating functionality of the device and feasibility of

combining the two stimuli. Hence the results presented are to be considered

preliminary.

6

Chapter 2: Bioelectrical Stimuli

2 Bioelectrical Stimuli

2.1 Introduction

Low energy, time varying magnetic fields were first used to treat therapeutically

resistant problems of the musculoskeletal system 30 years ago, after the functional

significance of the electromechanical properties of bone were elucidated. When fully

hydrated or native bone and other structural tissues like cartilage and tendon are

dynamically deformed, they develop electric potentials in the range of 1 to 5 mV/cm

(Bassett, 1971).

This is the behavior of ‘piezoelectric’ materials, which exhibit an electric potential

while undergoing mechanical deformation. However, potentials in bone are

comprised of two factors, first is a piezoelectric effect and second is the

electrokinetic potential. The magnitude of each in the measured electric potential

waveform is strongly influenced by the frequency of the loading. During rapid

deformation, such as impact loading of bone, there is an initial high amplitude spike

on the electric potential waveform that appears to be governed by the piezoelectric

response (Reinish and Nowick, 1976; Chakkalakal and Johnson, 1981). The

subsequent waveform behavior is characterized by long relaxation times and is

created by fluid and ion flows past the fixed charges on the small canals between

osteocytes, called canaliculi (Cowin et al., 1995; Cowin, 1999). Since the early 70’s

researchers have recognised the therapeutic potential of using electric fields similar

to these naturally occurring in bone to augment the fracture healing process (Bassett

and Pawluk, 1975)

Undoubtedly the most common use of these electric fields has been through

implementation of what is termed ‘pulsed electromagnetic fields’ (PEMFs). These

are time varying magnetic fields set up from a pulsed current passed through a single

or a number of wire coils which are then located at or around a fracture site.

7


The magnetic field produces a current in the conducting medium (bone) proportional

to the rate of change of the magnetic flux (density of the magnetic field). Serway

describes this relationship as, “The EMF induced in a circuit (ε) is proportional to the

time rate of change of the magnetic flux through the circuit (-dΦ/dt)” (Serway,

2004). The majority of the clinically available PEMFs are made up of a ‘pulse burst’

system. This waveform contains a broad band of frequency components with most of

their energy at the lower end of the electromagnetic field spectrum (< 1 KHz) and is

displayed in Figure 5-1. The use of ‘pulsed’ fields was undertaken in the early 70’s

due to the observation that strain-generated potentials were also pulsed. Prior to this,

only DC electric fields had been used for bone healing which had been modeled on

the measured electric potential between the ends of a fracture site (Friedenberg et al.,

1971).

It is believed that it is this electric field that potentiates the bioeffects seen (Pilla,

1993; Pilla et al., 1993) and not the magnetic field, however, recent research

theorises an important interaction between magnetic fields and actin filaments in the

microvilli of the cell (Gartzke and Lange, 2002) plus it may also act to directly

inhibit cell growth through a mechanism independent of electric field mediated gap

junctional coupling (Vander Molen et al., 2000). Also, studies employing static

magnets (which would not include an induced electrical stimulus) placed upon cell

cultures have produced statistically significant results over control cultures

(McDonald, 1993; Fanellia et al., 1999; Markov and Pilla, 1999; Binhi et al., 2001;

Yamamoto et al., 2003; Qin et al., 2004).

However, there has been no definitive answer as to the magnitude of the magnetic

field’s influence over cellular function as compared to the induced electric field’s

when considering bone adaptation. A recent PhD thesis made a comparison of static

or pulsed magnetic fields and alternating current stimulation on production of

organic (collagen type 1) and inorganic (calcium) components of bone. The results

demonstrated that static magnets produced no significant changes in either the

calcium or collagen production while an alternating magnetic field exhibited down

regulation in both. The alternating current was the only stimulus that increased the

production in both constituents of bone (Supronowicz, 2002) supporting the

8


contention that enzyme activity is influenced by the electric field via gap junctional

coupling (Vander Molen et al., 2000).

Commercial devices employing PEMFs for treatment of fracture healing have been

available for over 20 years. These devices are not restricted to long bone fractures

(Frykman et al., 1986; Kahanovitz et al., 1994) and, among other pathologies, can be

used in osteoarthritic joints (Trock et al., 1994) and osteoporotic bone (Tabrah et al.,

1990; Chang and Chang, 2003). This method of stimulation has also been effective in

reversing femoral head necrosis and augmenting spinal fusions (Bassett et al., 1989;

Aaron, 1994; Guizzardi et al., 1994; Linovitz et al., 2002). Promotion of tibial bone

fracture union with these devices has been shown to be at least as effective as

surgical intervention, with an increased success rate for patients who have already

undergone failed surgical intervention (Gossling et al., 1992).

At a cellular level, osteocytes, osteoblasts (Zhang et al., 1998) and the bone

resorbing cells, the osteoclasts (Espinosa et al., 2002) are electrically active.

Osteoblasts and osteoclasts migrate in different directions when placed in an electric

field (Ferrier et al., 1986a). Osteoblasts migrate towards the negative potential,

which has been measured as the net charge of bone forming regions and fracture sites

(Friedenberg and Brighton, 1966; Borgens, 1984). The role of minute electric

currents around and within the cells is of critical importance for their normal

functioning (De Loof, 1986) and can accelerate normal cellular functions such as

endocytosis (Antov et al., 2004). PEMFs perturb these currents and charges (Smith et

al., 1991) and influence the process it initiates and thus external non-invasive

electrical stimulation is a very powerful tool in augmentation of cells, tissues and

organs.

As this project employed the use of PEMFs, the following in vivo and in vitro

reviews will focus on studies using this stimulant.

9


2.2 Influential Factors in Electrical Stimulation

Studies

Endogenous electrical events that occur naturally, and may be significant in the

transduction of the PEMF signal, could confound results if not controlled. These can

include (Bassett, 1995):

• Fixed charge on moving membranes

• Action potentials

• Transmembrane potentials

• Injury potentials

• Developmental potentials

• Piezoelectric potentials

• Electrokinetic (streaming potentials)

• Resultant biomagnetic fields

The exposed cell cultures are very sensitive to the many physical factors of the

PEMF field. These field factors include:

• Strength of field

• Homogeneity of the induced electric field (E vs. B)

• Static and time varying components of magnetic field (Bac and Bdc)

• Repetition rate and sequencing

• Pulse shape (symmetric or not)

• Rise and fall times of the induced electric field • Frequency content of signal (Fourier analysis)

While secondary (environmental) fields that could possibly confound results include

(Bassett, 1991):

• Geomagnetic (static and time varying)

• Switching transients

• Electron microscopes, NMR, ESR

10


• Powerlines

• R.F. and Microwave

• Magnetic door catches

• Electrostatic (fur, clothing)

It has been reported that cells respond to the rate of change in the magnetic field

(dB/dt) and not to the peak field magnitude or total flux exposure (O’Conner et al.,

1982; Dennis et al., 2003). This highlights the importance of properly masking any

spurious signals that may be contained in the PEMF signal such as switching

transients. Created from the on off switching of the signal generator these high

frequency, high magnitude electric fields introduce an influential confounder to

experimental results.

The importance of describing all the elements of the inductively coupled PEMF and

biological system when conducting experimental studies is highlighted by the

conflicting results seen between skin wound healing models from Ottani et al.

(Ottani et al., 1988) and McGrath et al. (McGrath et al., 1983) who have failed to

properly describe the orientation of the PEMF coils. Coil orientation has been

theoretically and empirically proven to induce different stimuli for cell cultures

(McLeod et al., 1983).

2.3 In Vitro Electrical Stimulation

The majority of the in vitro cell culture devices utilizing PEMFs as the source of the

electrical stimulus have employed an air coil system. Simply, it is coil of current

carrying wire connected to an output device capable of driving a pulsed current

through the system. The pulsed current in the coil produces the magnetic field

perpendicular to the flow of current. This time changing magnetic field produced is

then placed over the conductive biological tissue and the induced electric field is

produced. The majority of early work on PEMF induced cellular perturbations

(Bassett, 1982) began from Bassett’s work on bone augmentation.

11


Bone cell proliferation and differentiation are important factors during bone tissue

healing and exogenously applied stimuli, that specifically promote one or the other,

have great therapeutic potential. Clinical PEMF devices have been shown to affect

proliferation and differentiation of bone cells in vitro (McLeod et al., 1991;

Fitzsimmons et al., 1995; Yonemori et al., 1996; Sollazzo et al., 1997; Lohmann et

al., 2000; Chang et al., 2003; Lohmann et al., 2003). PEMFs stimulate many

subcellular responses in living systems and appear to demonstrate exquisite

specificity of action depending upon both the physical and biological factors

involved (Bassett, 1989). One proposed principal target for PEMFs is the plasma

membrane and transmembrane proteins, rather than the cytoplasm (Luben, 1993;

Adair, 1998). Gap junctions, specialized intercellular junctions, have been proposed

as mediators of the PEMF exposed cellular response (Vander Molen et al., 2000;

Lohmann et al., 2003). There is evidence that they act as electrical connections

during exposure to the PEMF, creating an amplification of the signal (Muehsam and

Pilla, 1999; Pilla, 2002b).

Three studies conducted by a team based at Ferrara University in Italy, tested the

effects of PEMFs on human osteoblast-like cells and osteosarcoma cells (MG-63 and

TE-85) from 1996 to 1999. They used a pulse burst PEMF containing a 1.3msec

pulse burst, pulsed at 75Hz. Each pulse contained a 200µsec, 15mV positive

amplitude and a 50µsec, 150mV negative amplitude with a maximum magnetic field

of 20 Gauss.

The first of these studies (Sollazzo et al., 1996) simply assayed the proliferation rates

of human osteoblast-like cells after 24 hours of PEMF exposure. They discovered a

two to five-fold increase in DNA synthesis, however this effect was only observed

when the cells were cultured in 10% Foetal Calf Serum (FCS). The next study

(Sollazzo et al., 1997) focused on culturing both human osteoblast cells and

osteosarcoma cells (MG-63) in differing amounts of FCS while undergoing PEMF

stimulation, to determine its effects upon the proliferation rate of cells. These results

confirmed the hypothesis that the proliferation rate of human osteoblasts and MG-63

cells were proportional to the amount of FCS present in the culture. When human

osteoblast cells were cultured in absence of FCS there was no increase in

12


proliferation rate over control cell cultures whereas MG-63 osteosarcoma cells did

show an increase. Aside from this, the proliferation rate of all cell lines increased

with the greater length of exposure to the PEMF.

The final of the three papers (De Mattei et al., 1999) took the approach of finding a

correlation between PEMF exposure time and a significant increase (p<0.05) of

proliferation over controls. For the normal human osteoblast cells (derived from bone

biopsies) used, the PEMF application required was 6-9 hours, while the two

osteosarcoma cell lines (TE-85 and MG-63) required only 30 minutes of exposure

before PEMF-stimulated proliferation was significantly greater than the control.

Once again, the osteosarcoma cells showed lower proliferation rates in a FCS-free

culture medium, while human osteoblast cells without FCS did not increase over

controls at all.

Luben (Luben, 1993) outlines the effects of the PEMF characteristics on the signal

transduction of G protein linked receptors. Osteoblasts, pineal cells, fibroblasts and

hepatoma cells were used to test the effects of PEMFs on the ß adrenergic receptors.

It was found that the receptors in the osteoblasts and the pineal cells were inhibited

but not the receptors in the fibroblasts or hepatoma cells. PEMFs inhibited

parathyroid hormone (PTH) signal transduction through the PTH receptor. PTH

signal transduction is integral in the recruitment of osteoclasts and an inhibition of

this pathway will lead to a decrease in bone resorption. It is proposed that bony tissue

deposited in regions of negative charge is caused by this inhibition of PTH (Marieb,

1995).

Bone loss due to disuse osteoporosis, where osteoclast outstrips osteoblast activity, is

blocked with the use of PEMFs (Cruess et al., 1983; Rubin et al., 1989). It is

believed the reduction in influence of the PTH pathway from the PEMF is a potent

mediator in these pathologic states.

As osteoclasts resorb bone, they play an important role in maintaining the bone’s

mass and mechanical integrity. PEMF stimulation of osteoclasts was reported to have

reduced the number of new osteoclasts being formed (Rubin et al., 1996). This

reduction is apparently dependant upon the intensity of the induced electric field on

13


the cells, as higher induced fields (12.2µV/cm) reversed the reduction from lower

fields (4.8µV/cm) (Chang et al., 2003). Others found that PEMFs can stimulate bone

resorption but only when rat osteoblasts were cultured with rat osteoclasts (Shankar

et al., 1998). Furthermore, PEMFs do not affect the action of calcitonin in inhibiting

bone resorption, however, PEMFs reverse calcitonin’s inhibitory affect on the Ca2+

receptor. These papers support the theory that PEMFs in part reduce osteopenia and

osteoporosis initiated from disuse, through inhibition of osteoclast cell recruitment.

Bodamyali (Bodamyali et al., 1998) produced interesting data on the PEMF effects

on osteogenesis and the rate of bone morphogenetic protein (BMP) transcription. The

PEMF was found to positively affect both. Bone-like nodules (osteogenesis)

increased in number (39% over controls) and size (70% larger over controls) after 6

hours of exposure (p < 0.05), while synthesis of BMP 2 and 4 only needed 30

minutes of exposure to significantly (p < 0.01) increase over controls. BMPs

individually induce bone formation and are thus a precursor to extracellular matrix

formation.

Lohmann (Lohmann et al., 2000) evaluated PEMF effects on differentiation and

local factor production in MG-63 osteosarcoma cells. The results from this study are

somewhat different from those in previous studies on PEMF application to

osteosarcoma cells. The proliferation of the treated cells was inhibited by the PEMF.

This result and the increase in alkaline phosphatase activity, osteocalcin and

collagen, indicates enhanced differentiation. Local factor production such as

prostaglandin E2 and transforming growth factor β1 also increased with the

application of PEMFs, corroborating this hypothesis. In contrast to this, Sollazzo et

al. (Sollazzo et al., 1997) and De Mattei et al. (De Mattei et al., 1999) have reported

increases in the cellular proliferation due to PEMF exposure.

The explanation of the discrepancy lies in the state of confluence of the cells before

exposure. Sollazzo et al. and De Mattei et al. were using cultures that had not yet

reached confluence. Lohmann et al. began testing after cells had reached confluence,

an even greater length of time than the entire 48-hour test of Sollazzo et al. and

DeMatti et al. Thus the phenotypic differentiation of the cells used in Sollazzo et al.

and DeMatti et al. were at an earlier stage. As stated by Lohmann et al., the

14


“proliferation is negatively correlated with [a differentiated] phenotypic expression

in osteoblasts”.

2.4 In Vivo Electrical Stimulation

PEMFs were first introduced to mimic locomotion-induced electric fields in bone

that were in the order of 1V/m at frequencies below 10 Hz. Inducing this electric

field was not practical as strong relaxation processes occur in surrounding tissue and

attenuate the signal before it reaches the bone. Therefore the higher frequency

component of impact-loaded bone was mimicked with a signal component (1-

10KHz) gated at a relatively low frequency (1-100Hz) (Otter et al., 1998).

Clinical studies into canine osteotomies exposed to 28 days of pulsed

electromagnetic fields, found that there was an increase in the repair response

(Bassett et al., 1974). The PEMF consisted of a 1Hz, 1ms duration pulse. These

animal studies expanded into numerous human clinical trials.

Such trials principally looked at surgically resistant bone fractures. Pseudoarthroses

and non-unions were focused upon in the late 70’s and early 80’s. Salvage of these

limbs (destined to be amputated) was achieved with a maximum PEMF strength of

2mT pulsed at 72Hz. The first of these human trials resulted in 70 % (Bassett et al.,

1977) and 87 % success (Bassett et al., 1981; Bassett et al., 1982a). However, similar

studies in the U.K found only 60 % success with the same pulse characteristics

(Sutcliffe and Goldberg, 1982).

Cane, Botti and Soana (Cane et al., 1993) have carried out work into PEMF

applications on osteoblast activity in repair of transcortical holes made in six adult

horses. The PEMF characteristics vary from those used in the work conducted by

Bassett. The specifications (28 Gauss peak and 75Hz-repetition rate) are modelled

off the IGEA Biostim™ device marketed within Europe. Results of this paper found

increased bone formation and mineral apposition rate, indicating PEMFs improve the

osteogenic phase of healing.

15


Another Italian group (Mammi et al., 1993) using the same IGEA Biostim™ PEMF

specifications as Cane et al.(1993), conducted a double blind study of PEMF

application on tibial osteotomies created on patients suffering degenerative arthrosis

of the knee. X-rays of the tibia were double blindly evaluated in a 1 to 4 grading

system (4 being the most advanced healing) 60 days after the patients were treated

with PEMFs. A significant 73.6% of the control group were included in the 1st and

2nd stage healing, while 72.2% of the PEMF stimulated group exhibited 3rd to 4th

stage healing.

Osseointegration (direct attachment of living bone to the surface of an implant) of

hydroxyapatite-coated titanium implants in cancellous bone has been increased with

the use of PEMFs as measured by micro hardness values at a distance of 200 and

500µm from the implant interface. The PEMF stimulated implants had micro

hardness values approaching those for normal bone (Fini et al., 2002).

A review of 44 English language publications employing a clinical PEMF

stimulation for ununited tibial fractures revealed, “PEMF treatment of ununited

fractures has proved to be … at least as effective as surgical therapies” (Gossling et

al., 1992). PEMF stimulated fractures that have had previously failed surgery were

reported to have a greater success rate than an additional surgical procedure, with

success rate increasing as number of prior surgeries increases. PEMF healing success

was also greater for infected fractures and closed fractures but lesser for open

fractures, as compared to surgical intervention.

Studies have shown the greatest long term cellular changes from PEMF treatment

occurs when the bone has already matured into the calcification stage of fracture

healing as opposed to during the initial inflammatory response of cellular

proliferation (Bassett, 1989). It should be noted however that this does not mean

applying the PEMF signal early in a fracture will not induce an effect, but that the

observed changes occur very quickly and controls ‘catch up’ by the time

measurements are taken. As discussed in Section 2.3, in vitro cultures of osteoblast

like cells showed that 30 minutes of PEMF exposure was enough to significantly

16


increase cellular proliferation over controls but tapered off by 24 hours (De Mattei et

al., 1999).

2.5 Influence of PEMF Characteristics on Biological

Response

The PEMFs biological transduction pathway in stimulating a specific cellular action

has not yet been fully understood. Evidence points towards a multitude of specific

events that ultimately result in a few coherent outcomes (Cleary, 1993). This creates

a complex problem to unravel when trying to explain specific cellular actions

stemming from PEMF stimulation.

Experimental work on the use of single pulse and pulse burst (PEMF) systems was

able to elucidate that the two produced different cellular responses. Single pulse

systems operate with a single repetitive pulse at one particular frequency. However,

each single pulse waveform contains a range of frequency content as measured via

Fourier analysis.

2.5.1 In Vitro

Goodman and co-workers (Goodman et al., 1983) assayed an increase in specific

activity of messenger RNA in dipteran (two winged insect) salivary gland

chromosomes when exposed to two different PEMFs for only 15 minutes. The more

proficient of the two signals tested, was a single pulse train of 72Hz, 380µsec pulse

duration, while the less proficient was the pulse burst of 5msec duration, with each

pulse consisting of 200µsec positive duration and 28µsec negative duration. The

same group then found that PEMFs alter cellular transcription and translation in

eukaryotic cells (Goodman and Henderson, 1986).

Subsequently, Goodman (Goodman et al., 1987) showed that each PEMF system

triggered characteristically different gene translation ‘signatures’ in X chromosomes

17


from sciara coprophila. Additionally, studies on calcium in avian chondrocytes

when exposed to these two signals showed single pulse EMFs reduced while the

pulse burst EMF increased the uptake (Bassett et al., 1979).

Endothelial cell response to PEMFs showed an increase in DNA synthesis and the

formation of new blood vessels (Yen-Patton et al., 1988), while simple sine waves

are very specific and only increase the DNA synthesis of foreskin fibroblasts (Liboff

et al., 1984).

Efforts to reduce the complexity of the PEMF signal discovered a pulsed 2.7MHz

sine wave, similar in frequency to the initial magnetic field ramping of PEMFs, is

capable of modifying calcium behaviour of mineralising tissues (Fitton-Jackson,

1985). Fitton-Jackson also found other sine wave components of the PEMF signal

were biologically active, however, these were not as potent as the PEMF stimulus.

Additionally, they observed there were different responses to pulsing or continuous

waveform patterns.

2.5.2 In Vivo

Modification of callus formation in rat tibial osteotomies (bone cuts), when using

different PEMF characteristics (Bassett et al., 1982b), found significant differences

upon the load bearing capability of the rat tibiae. Four PEMF waveforms were used;

two with a negative spike and two with a negative square shape. The major findings

concluded that:

1. PEMFs appear to facilitate osteogenesis by enhancing the preliminary step of

endochondral ossification, namely calcification of the fibrocartilage.

2. The pulse characteristic (15Hz negatively spiked 5msec pulse burst with each

pulse containing 200µsec, 17mV positive amplitude and 28µsec, 150mV

negative amplitude) with the highest total energy input into the system did

not produce any increase in load bearing above controls.

3. This same PEMF characteristic (as above) has inhibited the regenerative

thrust of salamander limb regeneration, while all other PEMF characteristics

tested accelerated regeneration (Smith and Pilla, 1981).

18


4. The most effective pulse (5Hz negative square wave 5msec pulse burst with

each pulse containing 250µsec, 17mV positive amplitude and 33µsec, 150mV

negative amplitude) increased mineralisation of bridging callus, which

increased load bearing 2.4 times over the controls.

However, it was shown that the pulse burst signal discussed in point two, which did

not affect the load bearing, has had positive results for previously failed fractures,

healing 80% of patients (Bassett et al., 1982c). This signal has been approved by the

food and drug administration for clinical use, and is the signal utilized in this project.

Revascularization of devascularized rabbit femoral heads was significantly greater

with the signal pulse system (75%) than the PEMF (45%) (Rinsky et al., 1980) and

seems to be more effective at treating osteonecrosis and disuse osteoporosis than the

pulse burst system (Bassett, 1983).

Rubin and co-workers undertook a correlation analysis of the spectral content of

PEMFs and their bone remodelling activity in animal models of disuse osteoporosis

(Rubin et al., 1989). They found that the maximum osteogenic effect occurred with a

time changing magnetic flux between 0.01 and 0.04 tesla per second and an optimum

induced electric field stimulated below approximately 75Hz (McLeod and Rubin,

1990).

In a follow up study to confirm the preferential sensitivity to stimuli below 75Hz, it

was found that an osteogenic influence from sinusoidal electric fields was dependent

on the frequency. A 150, 75, and 15-Hertz sinusoidal field generated a -3 %, + 5 %,

and + 20 % mean change in the bone area respectively. These results suggest a tissue

sensitivity that is specific to very low-frequency sinusoidal electric fields and that the

induced electric fields need not have complex waveforms to be osteogenic (McLeod

and Rubin, 1992).

2.5.3 Summary and Conclusion

The field effectiveness peaks between the 15Hz and 30Hz range, which corresponds

to that in normal function of bone (locomotion impact reaction forces and muscle

19


fibre dynamics). Mechanical modulation (Rubin et al., 2002) and PEMF modulation

(Fitzsimmons et al., 1989) in these frequencies are osteogenic and most effectively

influence Ca2+ release from the cell (Smith et al., 1987), which is a vital step in the

calcification process of bone.

Biological tissue predominantly displays ionic conduction, or the transfer of current

via charged species in the bone fluid, creating high voltage spiked energization and

deenergization waveforms as is experienced with all dispersive dielectric materials

(Bronzino, 1995). Therefore the position and orientation of the specimen in the

magnetic field is of critical importance when defining exposure conditions.

Of great importance when considering the induced electric waveform from PEMFs is

the nature of the tissue’s passive electrical properties. Clinical situations such as

disuse osteoporosis will modify the induced electric waveforms in the bone due to

the lack of ionic species available to conduct the induced electric field. These

biological environments would be more closely described by an electrically passive

bone tissue model which exhibit reduced high voltage spikes from the induced

electric field (Bassett, 1989).

In conclusion, these studies strongly suggest that there is a gene-specific activation

and deactivation by each PEMF waveform depending upon the target tissues

electrical properties. The signal characteristics have the ability to stimulate a

multiplicity of differing actions within the cell. This emphasises the advantages of

focussing research on tissue specific signals, suited for each pathological condition in

order to maximise the true potential of the PEMF stimulation.

20


2.6 Mechanisms of Action

General cellular mechanisms of action are believed to be through one or a

combination of:

1. Interaction of PEMFs with the cell membrane surface

2. Perturbation of cell membrane potential

3. Electric charge distributions on surface to which cells attach

Theoretical evidence suggests that the cellular changes seen from PEMF exposure

are not created through direct mediation of DNA transcription and are instead created

via cell membrane surface changes (Adair, 1998). Luben (Luben, 1991)

hypothesised that EM fields may alter signal transduction of hormone regulation and

in particular, desensitisation of the parathyroid hormone (PTH) receptor to create

increased bone formation.

Fitzsimmons et al. (Fitzsimmons et al., 1992) proposed that EMFs increase Insulin-

like Growth Factor II (IGF II) receptor translocation from the interior of bones cells

and/or increase release of IGF itself by increasing calcium influx into the cell and the

subsequent activation of a calcium dependant protein kinase. These studies show that

the signalling is an important step in the transduction process but do not clarify the

actual mechanism of the EMFs interacting with the flow of information across the

cell membrane. Coulombic forces arising at the surface of the cell from the EMFs

could possibly be a solution to this problem. These forces derive from the interaction

of the cell surface charge and the EMF (only possible if the cell is anchored to the

surrounding extracellular matrix), quantified to be approximately 10-12 N when in a

1mV/m field strength. Any local portion of the cell that is not directly attached to the

bone matrix is able to be mechanical perturbed, biasing the Brownian Ratchet

mechanism. Large amplifications due to this biasing can occur and thus large

deformations may distort transmembrane proteins (receptors and ion channels) (Otter

et al., 1996) and/or the intracellular actin cytoskeleton (Pavalko et al., 2003b).

Perturbation in the cell membrane potential as a mechanism of action has been

disputed in the past, as ~10nV perturbations in membrane potential caused by PEMF

21


signals are swamped by membrane noise which is in the order of 100µV (Tenforde,

1989; Adair, 1991). However, as bone cells exhibit gap junctions (Doty, 1981) then

amplification of this signal can take place. As reported by Muehsam and Pilla

(Muehsam and Pilla, 1999), the initial resting membrane potential of the cell and the

number of cells in an array joined by gap junctions is important. To be effective, the

cell array needs to be relatively large (~1mm) and the induced electric field in the

order of 1mV/m, which creates a signal to noise ratio of 1 with only 20 000 cells

(Otter et al., 1998; Muehsam and Pilla, 1999). This variation in EMF bioactive

ability is due to the voltage-dependent binding and transport mechanisms of bone

cells, with each process containing an inherent optimum frequency. For example, the

Hodgkin–Huxley (Hodgkin and Huxley, 1952) K+-conduction pathway across the

membrane has preferential sensitivity to applied field frequencies in the 1–100 Hz

range, centred at approximately 16 Hz (Muehsam and Pilla, 1999) while the binding

of Ca2+ to Calmodulin centres on 10Hz (Pilla, 2002b). The calcium causes

configurational changes in gap junction proteins, regulating intracellular potential

(Weinbaum et al., 1994) and thus the propagation of the electric signal through the

osteocytic network. Harrigan and Hamilton (Harrigan and Hamilton, 1993) also

proposed a model based on electric coupling between adjacent cells. It correctly

predicted the remodelling processes of bone, concluding that position, loading rate,

manner of loading (compression versus bending) and degree of cellular coupling

influences the results, as has been found by others (Weinbaum et al., 1994; Donahue,

2000; Hsieh and Turner, 2001).

Additionally, the extracellular electric currents will alter the electric charge

distributions on the surface to which the cells are attached. Surface charge directly

effects the attachment, proliferation and differentiation of cells (Qiu et al., 1998).

This alteration could be direct (surface charge) or indirect (adsorbed proteins such as

growth factors and adhesion molecules) on the substrate (McLeod et al., 1998).

Controlled charge induction experiments (1-10µC/m2) have established that cell

adhesion, growth (Vander Molen and McLeod, 1995), and phenotypic expression can

be perturbed, while substrates that have been exposed to ELF electric fields before

cell plating have also shown similar effects (McLeod and Rubin, 1994). PEMFs

affect calcium salt crystal formation (Madronero, 1990) and bony ingrowth into

22


inorganic substrates of tricalcium phosphate (low ingrowth) and hydroxy apatite

(high ingrowth) by way of surface charge (Shimizu et al., 1988).

It is widely agreed that the positive results from healing of clinical bone pathologies

with PEMFs is based on the calcification (initiation and acceleration) of the

fibrocartilage, (Bassett, 1989) and the vascularization of the new bony tissue (Yen-

Patton et al., 1988). Ca2+ release is a vital step in the calcification process, whose

movements have shown to be most effectively influenced by 15Hz PEMF signals

(Smith et al., 1987). This reduces the net negative charge of the fibrocartilage

allowing cells with fixed electronegative charges on their membrane to invade.

Osteoblasts, which naturally migrate towards negative charges (Ferrier et al., 1986a),

are already releasing the calcium in the gap. When there are high extracellular levels

of calcium, electric fields have a reduced ability to modulate intracellular calcium

concentration, thought to be due to calcium not allowing the electric fields to

‘destabilise’ the membrane (McLeod et al., 1991) by the biased Brownian Ratchet

mechanism mentioned earlier.

However, osteoblasts also undergo hyperpolarization of their cellular membranes

proportional to extracellular calcium concentration (Ferrier et al., 1985), which may

have an influence over the ability of the PEMF to create a great enough

transmembrane potential for the cellular array to be ‘seen’ above the normal thermal

noise fluctuations.

2.7 Safety

According to the published studies utilising pulsed therapeutic electromagnetic fields

there is no adverse health risk associated with its use. Former cancer patients

(successfully treated with chemotherapy) who underwent PEMF therapy for ununited

fractures showed no sign of ‘relighting’ of the malignancy (Bassett, 1989). Rats

exposed to PEMFs for one year, showed identical results with control cultures in

regard to tumour growth (Bassett, 1978).

23


Also, a ten-year review of treatment of delayed union and nonunion with a bone

growth stimulator revealed that all fractures had remained united, and normal bone

remodelling had occurred supporting the long-term safety and effectiveness of

PEMFs in treating non-uniting fractures (Cundy and Paterson, 1990).

2.8 Conclusions

PEMF stimulation has had therapeutic effects on surgically resistant ununited bone

fractures such as osteonecrosis, pseudoarthroses, fracture callus formation, mineral

apposition rates, osteotomies and non-unions. The in vivo results of clinical trials

have been positive in the majority of cases with both human and animal subjects. It

appears that a PEMF applied to various bone conditions has an advantageous result.

The exact mechanisms for PEMF action upon the bone tissue are not clearly

understood, however cell and tissue effects such as mineralisation, increased matrix

and DNA synthesis have undergone concentrated studies. These have concluded that

PEMF actions can upregulate all these variables through modification of 2nd

messengers such as Ca2+, inositol triphosphate (IP3), cAMP and protein kinases.

Growth factors such as TGFβ and soluble factors like prostaglandins have also been

shown to upregulate. Cell culture confluence levels mediate cellular proliferation

results for cells stimulated with PEMFs. Greater cell culture confluence levels

promote cellular differentiation when stimulated with PEMFs and not cellular

proliferation.

PEMFs have been shown in certain cases to be as effective if not greater than

surgical intervention for bone pathologies such as non-unions, pseudoarthroses and

osteonecrosis (Gossling et al., 1992).

Consideration needs to be made of the influence unaccounted external magnetic and

electrical fields may have on the final results of studies employing a PEMF stimulus.

Additionally, no single cellular transduction mechanism is present for PEMF stimuli.

Instead, specifications of the PEMF pulse will determine the final biological

outcomes seen with in vivo and in vitro experimentation. During in vitro studies,

biofactors such as the level of differentiation (Diniz et al., 2002), density of cells

24


(McLeod et al., 1993; Hart, 1996; De Mattei et al., 2001), presence of gap junctions

(Donahue, 2000) and the PEMF exposure pattern (McLeod et al., 1983) affect the

final results.

25

Chapter 3: Biomechanical Stimuli

3 Biomechanical Stimuli

3.1 In Vitro Mechanical Strain

Mechanical stimulation of in vitro cell cultures follows three principal methods: fluid

flow across the cell monolayer, substrate stretching and hydrostatic compression. All

three techniques mechanically ‘strain’ (stretch) the cells. Also used is the direct

loading of organ cultures to elicit responses. Techniques differ in their method of

strain application and therefore usually differ in their level of applied strain.

One unique way of mechanically stimulating cells was to restrict the natural

endocytosis mechanism, thereby creating an internal mechanical stress. This study

concluded that the applied mechanical stress promoted cell transdifferentiation from

myoblasts to osteoblasts (Rauch et al., 2002). However the involvement of factors

that could possibly initiate cellular differentiation, such as bone morphogenetic

protein 2 (BMP2), confirmed the interpretation of these results as inconsequential.

3.1.1 Organ culture / Explant studies

Loading of human adult cancellous bone resulted in a rise of intracellular glucose 6-

phosphate dehydrogenase (G6PD) in bone-lining cells immediately after the

application of the load. Also, an increase in RNA synthesis from osteocytes 6 h after

loading was observed (El Haj et al., 1990).

Indomethacin (inhibits prostaglandin) resisted both the loading-related G6PD and the

RNA increase response in El Haj’s study (El Haj et al., 1990) consistent with Pead

and Lanyon, who reported indomethacin reduced osteogenic capacity in vivo (Pead

and Lanyon, 1989). This production of prostaglandins in response to mechanical

loading has subsequently been repeated by others:

• 17-day old embryonic chick tibiotarsi (Dallas et al., 1993)

• Adult canine cancellous bone (Rawlinson et al., 1993)

26


• Mechanically stretched (3400µε, 600 cycles, 1Hz) cultures of rat long bone-

derived osteoblast-like cells (Zaman et al., 1997)

• Mechanical loading of 5-week-old rat bones over an 18-hour period (Cheng

et al., 1997)

This increased release of prostanoids (prostaglandins and prostacyclins) produced

increases in cell proliferation and matrix production and therefore it is proposed that

a prostanoid-dependent mechanism for bone cell development occurs in response to

mechanical stimuli.

3.1.2 Fluid flow

Two principal apparatus configurations have been used to impart fluid shear – The

cone-and-plate system and the parallel plate flow loop apparatus.

With the former, relative velocity and separation between the spinning cone and

stationary plate surfaces varies linearly with radial position. This configuration is

advantageous as it achieves homogeneous fluid shear stress on both surfaces.

The latter uses a pressure differential between two slit (manifold) openings at either

end of a rectangular chamber, causing uniform laminar flow to develop across the

culture surface. The vast majority of researchers have used this fluid shear method.

Gravity heads have been used for the pressure drop (Li et al., 1996) along with active

pumps (Jacobs et al., 1998), while a special version, incorporating a separate

"settling chamber" and a curvilinearly tapered inlet to optimise temporal/spatial flow

field development in pulsatile stimulus situations was developed by Ruel (Ruel et al.,

1995).

A parallel plate flow chamber configuration producing mean shear stresses of 0.4 to

1.2 Pa elicited production of nitric oxide (NO) and PGE2 by bone cells in a dose-

dependent manner with shear stress (Bakker et al., 2001). Sakai and co-workers

(Sakai et al., 1998) used the cone and plate system on SaOS-2 osteoblast-like cells.

After three hours of continuous exposure to physiologic shear (1.7 - 2 Pa) the cells

raised their TGF – β1 levels three-fold, while after six hours the osteoblasts increased

27


interleukin-11 (IL-11) four-fold with respect to controls (mediated by

prostaglandins).

Reich and Frangos (Reich and Frangos, 1991), also using osteoblasts subjected to

steady shear stress (6 dyn/cm2 and 24 dyn/cm2) in a cone and plate viscometer, found

that prostaglandin E2 increased 9- and 20- fold respectively while inositol

trisphosphate (IP3) increased dramatically after two hours of 24dyns/cm2.

3.1.2.1 Similarity to in vivo strain fields The current in vitro methods of fluid flow do not mimic the three-dimensional nature

of fluid flow in the porosities of bone, which are filled by osteocyte processes. It has

been proposed that any fluid flow responses from osteocytes elicited in vitro will be

invalid, as the fluid drag will stimulate the cellular processes and bodies of the

osteocytes. Pressures in the lacuna which house the osteocyte cell bodies do not

reach strain levels as high as those acting on the cell processes (You, 2002).

The shear stress produced in the canaliculi porosity due to normal locomotion is in

the order of 0.5 – 3 Pa (5 – 30 dyns/cm2) (Weinbaum et al., 1994; Zeng et al., 1994)

which has been shown to elicit intracellular Ca2+ release, matrix protein mRNA and

growth factor responses from osteoblastic and osteocytic cells in vitro (Reich and

Frangos, 1991; Williams et al., 1994; Hung et al., 1995; Jacobs et al., 1998; Sakai et

al., 1999; You et al., 2000).

As discussed in Chapter 2, fluid flow induces a flow of charged species across the

cell layer setting up an electrical potential, and may be associated with the

transduction of the mechanical strain.

3.1.3 Substrate Stretching

This technique incorporates stretching a cell substrate membrane that is seeded with

the cells. There are a number of different methods to achieve this stretch, each of

which is discussed in Chapter 7. Most studies employ out-of-physiologic strain levels

and are thus invalid for any hypothesis made regarding in vivo situations. However,

28


Bottlang and co-workers (Bottlang et al., 1997) designed a device that was able to

apply variable frequency, four-point bending to monolayer cell cultures producing

strain in the range encountered by bone in vivo (200 - 3000µε). Also, Pioletti and co-

workers (Pioletti et al., 2003) produced a micro-mechanical device simulating the

mechanical situation at the bone-implant interface, however both these studies did

not include biological results.

This focus on low amplitude, high frequency strain, has been made in an attempt to

replicate the exact environment the cells experience in vivo (Tanaka, 1999; Tanaka et

al., 2003b) which has been proven to have anabolic effects when applied to in vivo

situations (Rubin et al., 2002).

Culture surface strains depend upon a complex fluid/structure interaction with most

methods using movement of the substrate, which sets up cellular shear strains due to

the movement of growth medium fluid during deformation. Many studies have not

taken this into account, such as those by Stanford (Stanford et al., 1995) and Zaman

(Zaman et al., 1997). Calibration is complicated by the fact that the driving signal’s

magnitude, frequency and waveform plus the nutrient medium’s mass and viscosity

influence the culture surface stimulus (Brown, 2000).

3.1.3.1 Similarity to in vivo strain fields Direct cellular stretching of osteocytes and osteoblasts occurs during mechanical

loading in vivo. This is caused by the deformation created in the extracellular matrix

(ECM) to which the cells are adhered.

In vivo macro-level strain due to walking is predominantly in the range of 100-400µε

(Lanyon et al., 1975; Fritton et al., 2000), while most in vitro experiments require

strains in the magnitude of 1000-10000µε to elicit a response (Toma et al., 1997;

Brand et al., 2001) and thus an inherent amplification process must be present for the

mechanical strain to be transduced into cellular action.

There is no question that cellular stretching occurs within the processes of cartilage

formation, as the ECM, made predominantly of collagen, is flexible and elicits large

29


deformations on the cell membrane (Meyer et al., 2001a) evoking cellular

differentiation.

3.1.4 Hydrostatic Pressure

These devices are only useful on monolayer cultures as suspended cells are relatively

incompressible due to their high water content (Basso and Heersche, 2002). The

general method used employs a pressurisation of the gas phase the cells grow in. It

has high simplicity and spatial homogeneity of the stimulus. No streaming potential

effect is created that may confound results and the state of adhesion between cells

and their substrate, which is important for fluid flow techniques, is not required.

Some drawbacks include the high pO2 and pCO2 produced in the liquid nutrient

medium, which requires compensatory treatment steps before experimental

conclusions can be drawn (Ozawa et al., 1990).

Work conducted on growth plate cartilage by Klein-Nulend (Klein-Nulend et al.,

1986; Burger et al., 1991), concluded that uncalcified cartilage containing

hypertrophic chondrocytes responds directly to compressive force (air pressurization)

with an increased calcification of the cellular matrix, as experienced during

endochondral bone formation. Additionally, an increase in osteoclastic resorption

using a physiological level of dynamic strain loading was observed. Intermittent

compressive force (0.3Hz at 13 KPa) evoked a greater response than continuous

compressive force (13Kpa).

Salter and co-workers (Salter et al., 2000) have shown that the bone cellular

membrane hyperpolarizes in response to pressure induced strain. This process is

mediated by integrins (an intact actin cytoskeleton is essential), while interleukin-1

beta production, in response to mechanical stimuli, potentiates autocrine/paracrine

signaling. This response was only noted at a frequency of 0.33Hz and not 0.104Hz.

As discussed in Chapter 4, this phenomenon has implications for the dual stimulus

device, which includes the PEMF electrical stimulus.

30


3.1.4.1 Similarity to in vivo strain field Burger and co-workers (Burger et al., 1992) found intermittent pressurization

produced an anabolic effect on organ cultures of ossifying long bones as apposed to a

catabolic effect from continuous compressive force. This is after the same group

found that distortional stresses from the applied compressive force were created at

the interface between the mineralized/non-mineralized tissues in embryonic long

bone cultures, which have been proposed as a greater influence than the hydrostatic

stress over the mineralisation of the tissue (Tanck et al., 2000). However, Tanck and

colleagues (Tanck et al., 1999) had previously discovered, from their analysis of an

in vitro hydrostatic compressive force experiment on calcification of growth plate

cartilage in fetal mouse cartilaginous long bone rudiments (Klein-Nulend et al.,

1986), that distortional strain was highly influenced by the tissue matrix

compressibility and resulted in a strain too small to influence mineralisation.

Claes and colleagues (Claes et al., 1998) hypothesized that hydrostatic pressure less

than 0.15MPa, with small strains (<5%), produced intramembranous bone formation,

while strains less than 15% and hydrostatic pressure more than 0.15Mpa, stimulated

endochondral ossification in the fracture callus.

3.2 In Vivo Mechanical Strain

In vivo and ex vivo mechanical stimulus of bone has been applied through different

techniques such as exercise (Biewener and Bertram, 1993), osteotomies (Augat et al.,

1998) and loading devices that control the magnitude, rate and number of cycles of

mechanical stimulus on the bone (Turner et al., 1994b; Mosley and Lanyon, 1998).

To observe effects of decreased loading, approaches include neurectomy (removal of

nerve) (Frost, 2001), hind limb suspension (Uhthoff and Jaworski, 1978) and space

flight (van Loon et al., 1995). Disuse reduces osteocyte viability causing apoptosis,

followed by the release of factors stimulating osteoclast recruitment. The osteoclasts

will then resorb the dead cells and bone producing a net loss of bone mass (Mosley,

2000). Osteocyte apoptosis also appears to play a crucial role in the removal of

surrounding damaged or redundant bone (Noble and Reeve, 2000).

31


Studies of the influence that mechanical loading has upon skeletal tissue have

uncovered some general rules;

1. Cyclic compression forms stronger but more compliant healing fractures than

static compression (Panjabi et al., 1979).

2. Cyclically compressed fracture calluses are larger and stiffer than those

statically compressed (Goodship and Kenwright, 1985).

3. Vibration of bone in the range of 20-30 Hz with low strains (~250µε),

coinciding with muscular contraction frequency during normal posture

control, has strong anabolic effects on trabecular bone (Rubin et al., 2001a,b;

Rubin et al., 2002) while simulated muscular contractions via skin surface

electrodes promote callus development and mineralization (Park and Silva,

2004).

4. Loading can create different tissues (de Rooij et al., 2001) and cell

morphologies (Guldberg et al., 1997).

5. All effects are produced more strongly in growing rather than mature animals

(van der Meulen et al., 2002).

6. High shear strain and fluid flows in vivo deform bone precursor cells

stimulating formation of fibrous connective tissue while medium levels

stimulate formation of cartilage, and low levels cause ossification (Huiskes et

al., 1997; Lacroix and Prendergast, 2002).

7. Strains between 5% and 15% in the fracture callus will promote faster

healing, while strains outside this range will produce fibrocartilage (Claes and

Heigele, 1999).

8. Mechanical movement affects mesenchymal cell differentiation (Le et al.,

2001).

Previously, common belief was that peak strain from loading of the bone was

directly proportional to the production of new bone (Frost, 1983), however strain rate

has now been highlighted as the major determinate (O'Connor et al., 1982; Turner et

al., 1995; Mosley and Lanyon, 1998) with the shear stresses generated on bone cells

(due to fluid movement) proportional to this strain rate. Higher loading rates are

more effective for increasing bone formation than higher peak strains (Burr et al.,

2002), supporting this proportionality theory.

32


Earlier, O’Connor et al. (O'Connor et al., 1982) had applied bending and

compressive loads to the radius and ulna of experimental sheep. Loads were applied

at 0.5 Hz for one hour per day for six weeks with peak strains and strain rates never

exceeding the range attainable during normal locomotion. The definitive conclusion

to this study was that the most potent variable for load related remodelling of bone

was the ratio between maximum strain rate of the artificial regime and the maximum

strain rate during walking. This accounted for between 68 and 81% of the variation

in the measures of surface bone deposited and was the greatest influence over

intracortical secondary osteal remodelling. Observation of the influence strain rate

has upon bone remodelling is provided in similar work by Lanyon, who found that

remodelling in an avian ulna was promoted by dynamic strain (1 Hz, 525 N, ramped

square wave) but impeded by statically loaded (100 sec, 525 N, per day) and non-

loaded bones (Lanyon and Rubin, 1984). This work pioneered the concept of

“genetically programmed” regions of the skeleton (Rubin and Lanyon, 1987), where

differing strain thresholds, located in different regions of the skeleton, have to be met

before bone remodelling takes place.

It has been shown that a short loading stimulus can reinitiate osteogenesis after

disuse (Pead et al., 1988; Robling et al., 2000; Robling et al., 2002a,b; Hatton et al.,

2003). Bone only requires a “single short exposure to an osteogenic loading regime”

before “the full cascade of cellular events between quiescence and active bone

formation” occurs (Pead et al., 1988). Many have found that other variables such as

hysteresis energy (Kunnel, 2002) are also important physical parameters of bone

formation. A tendency of the proteoglycans to orientate closer to the collagen fibrils

after a short period of intermittent loading is thought to be a method of storing a

‘strain memory’ in bone tissue (Geiger, 1989; Skerry et al., 1990), however an

accurate study of the transient deposition rate of bone with respect to strain memory

effects has not been conducted.

Conflicting results from past work involving in vivo physiologic loading (running

chickens) (van der Meulen et al., 2002) has highlighted the need for a more

controlled mechanical stimulus on the bones such as those designed by Turner

(Turner et al., 1994b) and Lanyon (Mosley and Lanyon, 1998). Genetic background

33


is an important variable to take into consideration as different breeds of animal can

confound previously proven results of mechanical loading on osteogenesis (Pedersen

et al., 1999) as shown by Robling and co-workers who used differing genetic strains

of mice to show that mechanical strain transduction is under genetic control (Robling

and Turner, 2002). Inconsistencies in the response of human exercise studies (using

same loading regime) have highlighted the possibility of an individual mechano-

sensitivity level.

3.2.1 Cortical Bone

The main changes that occur due to mechanical loading are in cortical bone quantity,

not quality (van der Meulen et al., 2000). Studies focusing on the mechanical

regulation of cortical bone have revealed some common characteristics required for

osteogenesis. Strain rates and magnitudes of strain need to be high (Mosley and

Lanyon, 1998), while the resulting morphology of the tissue formed is dependant

upon the strain rate (Turner et al., 1994b). Frequency of strain is also very important

in determining the efficiency of the bone formation, where an optimum of 5 –10Hz is

required for a positive change in bone strength. Also, separate ‘bouts’ of loading, not

increased number of cycles, are required for greater endocortical bone formation

(Turner, 1998; Robling et al., 2000).

3.2.2 Cancellous Bone

Loading of cancellous bone causes increases in density and direction of the

trabeculae, depending on the direction of the principal stresses (Huiskes et al., 2000).

Quantitative measurements of cancellous bone formation in response to mechanical

stimulation have greater clinical relevance due to the trabecular bone’s significant

control over bone mass and mechanical integrity. Measurements are more difficult to

experimentally produce and control, however, hydraulic bone chambers that

encapsulate cancellous bone have been used with great success in vivo (Goldstein

and Guldberg, 1996; Guldberg et al., 1997; Lamerigts et al., 2000; Morgan et al.,

2001).

34


Vibration of bone in the range of 20-30 Hz, coinciding with muscle stimulation

frequency during normal posture control, has strong anabolic effects on trabecular

bone (Rubin et al., 2001a,b; Rubin et al., 2002). However the strains required for this

to occur are much smaller than those used for cortical bone.

3.2.3 Overload

Microdamage (microcracks, etc) is a result of overload. An overload threshold has to

be surpassed before microcracks will accumulate and cause fracture. This

microdamage threshold is centred on 3000µε (Frost, 2001). Normal mechanical

usage of bone creates microcracks that are remodelled (Frost, 2001). It has been

shown that a pooling of interstitial fluid in these microcracks impedes transport from

the blood supply, depleting the concentration of molecular entities in and

downstream from areas of damage. The osteocytes in these depleted areas lose

viability and are hence targeted for new remodelling activity (Tami et al., 2002).

Martin (Martin, 2000a) hypothesised that remodelling suppressing signals are

released from osteocytes during normal homeostasis and thus microcracks impede

these signals and the BMU will be activated into remodelling the area.

A strain overload, due to large gap sizes in a fracture callus, will result in excess

initial inflammatory tissue (Augat et al., 1998), more fibrous tissue, less

vascularization and less bone formation (Claes et al., 2002). The strains required for

fibrous tissue to form are in the range of >15% (Claes et al., 1998) equating to a gap

size of ≥6 mm in the fracture callus (Lacroix and Prendergast, 2002).

3.2.4 Computational Mechanobiology

Computational mechanobiology analyses and simulations created with the use of

finite element analysis (FEA) programs, have focused on functional adaptation of

particular types of bone to mechanical stress and the influence it has over the natural

progression of tissue differentiation. The latter has focused on such things as

arthroplasty interface movement and normal fracture healing, while the former is

based on cortical and trabecular bone adaptation due to external loading. The initial

basis for this avenue of research was the premise that mechanical deformation of the

35


tissue caused differentiation (Pauwels, 1941). Carter and associates (Carter et al.,

1998; Carter and Beaupre, 2001) defined endochondral bone formation with an

‘osteogenic index (I)’ that involves the addition of peak hydrostatic stress (D) with

the product of the peak octahedral shear stress term (S) and an empirical constant (k)

(Equation 3-1).

I = S + kD

Equation 3-1 Osteogenic Index of endochondral bone formation as proposed by Carter et al. (1998).

Successful simulations of skeletal development, fracture healing and healing around

orthopaedic implants have been completed, however each of these predicted different

values for (k), thus there is little strength in this ‘osteogenic index’ to anticipate

unique situations.

Huiskes and associates (Prendergast et al., 1997) developed an alternative to the

‘osteogenic index’ called the ‘mechano-regulation index (M)’ (Huiskes et al., 1997)

which incorporates an interstitial fluid flow velocity (ν) and maximum distortional

shear strain (γ) component (Equation 3-2). These are potent factors in the response of

skeletal tissue to mechanical loading (Turner et al., 1994a; Forwood and Turner,

1995; Turner and Pavalko, 1998; Sikavitsas et al., 2001; van der Meulen et al.,

2002). The equation is as follows:

M = γ/a + ν/b

Equation 3-2 Mechano-regulation Index of bone formation as proposed by Huiskes et al. (1997).

Where a = 0.0375 and b= 3µm/sec. The value of M determines the tissue present

with values of M >3 = Fibrous tissue, 1 ≤ M ≤ 3 = Cartilage tissue and M < 1 = Bone

tissue with each tissue containing a particular modulus value and permeability

(derived from (Søballe et al., 1992)). This simulation successfully replicated

experimental results from Søballe (Søballe et al., 1992) and was used to simulate

36


fracture healing (Lacroix and Prendergast, 2002) from a previous experimental study

(Claes et al., 1998). As noted by van der Meulen (van der Meulen et al., 2002), the

model of Huiskes and associates is a true simulation, as numerous iterations (500)

with feedback (fluid velocity and strain) into the index equation were tested.

Functional adaptation of bone to mechanical stress has been focused upon for longer

than tissue differentiation studies, as simulations only need to deal with bone tissue,

which requires less complexity in the models. Brown (Brown et al., 1990) concluded

from a FEA study of periosteal remodelling that “strain energy density, longitudinal

shear stress, and tensile principal stress/strain are the mechanical parameters most

likely related to the initiation of the remodelling response”. Hart (Hart, 1984)

simulated cortical bone adaptation in the early 80’s, while trabecular bone has been

simulated in an isotropic (Weinans et al., 1992) and anisotropic (Jacobs et al., 1997)

manner. However, these simulations only employ the activity of osteoblasts and

osteoclasts and need to take a more focused approach, such as that used by Adachi

(Adachi et al., 2001) and Mullender (Mullender and Huiskes, 1995) who have both

modelled the adaptation of bone on the surface of the trabeculae and not at arbitrary

points.

In summary, there are limitations to the level of complexity these computational

methods are able to compute and consequently their simulations are restricted to

provide only general outcomes. They lack many influential biochemical cellular

regulation pathways present in vivo. Some of the transitional constants used for tissue

differentiation thresholds are very influential on the final index value, leading to

large error amplification effects in the simulation (Carter et al., 1987). One area of

future computation research that has not appeared in literature is the effect of bone

disease, such as a reduced vascularization (avascular necrosis), on the computation

outcomes.

37


3.3 Mechanotransduction Process

Mechanotransduction is the term used to describe the process a cell follows to sense

and interpret a mechanical load before it elicits a cellular response. The term is

independent of cell type, as muscle cells (Vandenburgh and Karlisch, 1989; Miller et

al., 2000), neural cells (LaPlaca and Thibault, 1997), brain cells (Ellis et al., 1995;

Lazarowski et al., 1997) and even plant cells (Lynch and Lintilhac, 1997) have

exhibited sensitivity to mechanical loading. Normal mammalian development

involves mechanical stimulus on all cells in the body and this stimulus has been

proposed as a major determinate of a tissue’s natural form (Carter and Beaupre,

2001; Frost, 2001). In utero distortional strain (shear strain), due to muscular

twitching, influences normal development of embryonic mouse bones (Tanck et al.,

2000) and although not quantified it is more than likely the same would be true for

human skeletal development.

Specific loading information is transduced into a recognised biological signal that

mediates the cells’ normal processes. The process of mechanical loading, cell

reception and cell response is yet to be completely understood. However, the

mechanotransduction process can be described in four generalised steps:

1. Mechanocoupling – external mechanical force transduced into a local signal

(mechanical/electrical).

This is most likely conducted through fluid shear stress impinging on the osteocytes

within the canaliculi-lacunae porosity (Kufahl and Saha, 1990; Turner and Pavalko,

1998; Burger and Klein-Nulend, 1999; Hsieh and Turner, 2001; Sikavitsas et al.,

2001). The exact mechanics of how this is achieved is only just beginning to be

unravelled. An amplification theory described by You (You, 2002) is based on fluid

drag impinging transverse filaments, which tether the osteocyte cell processes to the

canalicular wall. Transverse filaments are linked to the intracellular actin

cytoskeleton (IAC), and when placed under fluid drag, a bone shaft strain

amplification of one order of magnitude at the cellular level results. However, this

38


stress is not produced in the lacunae, which houses the osteocytes cell body due to

the larger volumetric area with the same fluid pressure. You’s model explains the

current discrepancy between strains required to elicit a cellular response in vitro and

that occurring during mechanical loading in vivo.

It is documented that actin filaments reorganise into stress fibres in response to fluid

shear, and their connection with the transmembrane integrins is a crucial mediator of

the mechanotransduction process (Ajubi et al., 1996; Toma et al., 1997; Pavalko et

al., 1998). The mechanical information deforms the sensor cell membrane via the

transmembrane integrins, which in turn drives conformational changes in membrane

proteins. Some of these are linked to a solid-state signaling scaffold that releases

intracellular protein complexes capable of carrying mechanical information,

mechanosomes, into the nucleus. The mechanosomes translate this information into

changes in the geometry of target gene DNA altering gene activity (Pavalko et al.,

2003b).

Additionally, strain-generated potentials (SGP) are considered to be another

mechanocoupling mediator (Zeng et al., 1994; Pilla, 2002b). As discussed in Chapter

2, piezoelectric phenomenon is a characteristic of dry bone (dead) and is shielded in

wet bone (living) by ion relaxation processes (Otter et al., 1992). The flow of ions in

the intra-canalicular fluid moving past the fixed charges on the canaliculi, set up

streaming potentials commonly called strain generated potentials (Harrigan and

Hamilton, 1993; Zeng et al., 1994; Cowin et al., 1995; Cowin, 1999; Mow et al.,

1999; Beck et al., 2002) that may mediate the mechanotransduction process.

A unique study using a varying fluid viscosity (and hence shear strain) plus a

constant fluid velocity (and hence constant streaming potential) found that streaming

potentials and chemotransport played no role in the release of Nitric Oxide and

Prostaglandin E2 - two messengers in the bone signaling pathway (Bakker et al.,

2001). However, they failed to incorporate the transient nature of the streaming

potential signals experienced in vivo, which is similar to a pulsed distorted trapezoid.

Therefore these results cannot be interpreted as evidence that streaming potentials

and electric fields in general are not involved in the mechanotransduction process.

39


2. Biochemical coupling – or the transduction of a mechanical signal to a

biochemical response within a cell and its membrane.

Increased intracellular calcium levels, second messenger prostaglandins and nitric

oxide (NO) production (due to activation of endothelial and inducible nitric oxide

synthase (eNOS & iNOS)) are stimulated with fluid flow (Reich and Frangos, 1991;

Cheng et al., 1997; Bloomfield, 2001) but not physiological levels of mechanical

stretching (Owan et al., 1997; Smalt et al., 1997). However, these studies did not

include a high frequency component of strain, which has been shown to induce bone

formation. Additionally, mechanical stretching stimulates inositol triphosphate (IP3),

bone matrix mRNA levels and growth factors such as insulin-like growth factor

(IGF-I and IGF-II) (Brighton et al., 1992; Meyer et al., 2001b) among others.

Osteocytes are proposed as the initial sensors of loading and it has been found that

they are more sensitive to fluid flow and hydrostatic compression than osteoblasts

(Klein-Nulend et al., 1995b; Ajubi et al., 1996; Westbroek et al., 2000), with

prostaglandins and nitric oxide being immediately unregulated with fluid flow

(Klein-Nulend et al., 1995a). Marrow derived pre-osteoclast cells also exhibit release

of prostaglandins and nitric oxide after fluid shear stress stimulation. This may be a

sign of their ability to undergo autocrine signalling during normal locomotion

(McAllister et al., 2000).

The endocrine factors parathyroid hormone (PTH) and estrogen both modulate the

biochemical coupling by inhibiting nitric oxide synthase and enhancing

prostaglandin and nitric oxide (NO) production in osteoblasts respectively (Armour

et al., 2001; Joldersma et al., 2001; van't Hof and Ralston, 2001). Prostaglandin E2

and PTH both stimulate production of vascular endothelial growth factor (VEGF);

the protein that increases vascularization of the region (Harada et al., 1994; Cowin,

2001). NO involvement in cellular communication of strain measurement and

distribution between osteoblasts and osteocytes (as well as adaptive changes in bone

cell behaviour) is probable (Pitsillides et al., 1995) and seems to have a biphasic role,

where high levels will repress osteoclast resorption inhibiting growth and

differentiation of osteoblasts and low levels potentiate cytokine-induced bone

resorption (Ralston, 1997).

40


Also, G proteins and intracellular calcium play an important role in mediating

prostaglandins and nitric oxide (Reich et al., 1997). NO production seems to act

independent of the intracellular calcium pathway during steady flow but not during

flow bursts (rate of change in shear stress) (McAllister and Frangos, 1999). Mitogen-

activated protein kinases (MAPK) are unregulated due to mechanical loading (fluid

shear stress or stretch) (You et al., 2001a; Kletsas et al., 2002), one step away from

the regulation of Core binding factor A1 (Cbfa1) which is an important and potent

regulator of osteoblastic differentiation and function. It is proposed that this Cbfa1 is

the main target of mechanical signals (Ziros et al., 2002a; Ziros et al., 2002b).

Activation of this factor leads to autocrine/paracrine signalling and ultimately gene

expression.

3. Cell to cell communication – This is the transferral of the mechanical stimulus

between the cells in order to synchronise the effector response.

Cell to cell communication is very likely to be achieved through gap junctions (small

connections between cells that allow transfer of proteins) (Donahue, 2000; Vander

Molen et al., 2000) where the osteocytes (sensor cells) release the prostaglandins and

nitric oxide to communicate with the osteoblasts and osteoclasts (effector cells). In

support of this is the observation that inhibition of the prostaglandins and nitric oxide

(NO) can eliminate mechanically induced bone formation (Saunders et al., 2001).

However discrepancies arise (using the theory that stimulatory signals from the

osteocytes are released in proportion to increasing strain) when bone is in a state of

disuse. According to this, disuse would result in a decrease of bone remodelling,

however in reality remodelling increases with disuse. Therefore some believe an

inhibitory signal is the mediator of the remodelling process (Marotti, 1996; Martin,

2000a).

Osteoclasts and osteoblasts respond differently to prostaglandins (Ferrier et al.,

1986b). The osteoblasts undergo hyperpolarization in response to prostaglandins

(Salter et al., 2000), a response that may well effect transduction of applied electric

stimuli (discussed in Chapter 4).

41


4. Effector response – This is the final step of the mechanotransduction pathway

and involves the net production or removal of bone.

It is widely believed that the osteoblasts are the initiators of bone remodelling

(Rodan and Martin, 1981). Recruitment of non-dividing preosteoblasts and

osteoblasts from the bone lining occurs, along with the differentiation of

osteoprogenitor cells, to form sufficient levels of effector cells to undertake

osteogenesis.

3.4 Conclusions

Some of the most important revelations to be extracted from the in vivo mechanical

loading studies to date are that strain rate and strain frequency, not peak strain, are

the most potent stimuli on bone adaptation. Also interesting is the desensitisation

effect, whereby excessive loading cycles result in a loss of mechanically induced

bone formation.

The greatest osteogenic response is achieved through a dynamic strain regime,

meeting or surpassing a bone-location specific strain threshold before any net bone

deposition is achieved. Loading need only be a single short stimulus and not

continual dynamic strain, before the full cascade of cellular events between

quiescence and active bone formation is accomplished.

A low level of vibrational strain on bone, in the range of 20-30 Hz, which coincides

with muscular posture control, has strong anabolic effects. And may be used in

treating skeletal complications such as osteoporosis in a future non-pharmacological

intervention technique. For fracture healing situations, studies show that cyclic

compression forms larger, stiffer, stronger and more compliant healing fractures than

static compression. An optimum strain window in the fracture callus (5% and 15%)

promotes faster fracture healing, while loading affects tissue differentiation more

strongly in growing rather than mature animals through shear strain and fluid flow

dependant mechanisms. Mathematical simulations of bone adaptation and tissue

42


differentiation have revealed there are limitations in the complexity of the models

and they do not yet represent a true in vivo situation.

The in vitro mechanotransduction process from mechanical loading follows the steps

of mechanocoupling, biochemical coupling, cell-to-cell communication and then an

effector response. In all, the osteocyte, osteoblast and bone lining cell syncytium act

somewhat like a neuronal network whereby the sensing cells (osteocytes) transduce

the mechanical strain signal to biochemical pathways, which stimulate the effector

cells (osteoblasts and osteoclasts) into action. The use of nitric oxide and

prostaglandins as signalling molecules is the most likely mediator of this

mechanotransduction process; interestingly these factors also play a role as

neurotransmitters in the neural network (Marieb, 1995; Turner et al., 2002). The

cytoskeleton-integrin interaction has also shown to have mechano-sensing

capabilities, where the fluid flow induced mechanical strain directly induces

intracellular gene expression (Pavalko et al., 1998).

43

Chapter 4: Convergence of Stimuli

4 Convergence of Stimuli

4.1 In Vitro Studies

Studies have shown that hyperpolarization of the surface of the bone cell occurs

during pressure induced mechanical strain and application of physical shock waves.

The response is mediated via integrins and requires tyrosine kinase activity and an

intact actin cytoskeleton (Salter et al., 2000). Hyperpolarization initiates RAS

activation, used for signal transduction and protein synthesis, via the CBFA1

transcription factor (Wang et al., 2001).

Polarization of the cell membrane is capable of influencing cell division, speed and

direction of migration within electric fields and the speed of cellular differentiation

(Wang et al., 2001; Song et al., 2002; Zhao et al., 2002; Finkelstein et al., 2004) and

is a controlling factor in the ‘head-to-tail’ body axis of vertebrates during embryonic

development (Keller, 2002). This also effects cellular migration of biological cells

during wound healing. Each wound edge provides a positive electric polarity creating

an electric field, commonly called injury potentials, which attracts the negatively

charged cells towards the wound site and initially causes local vasoconstriction to

reduce blood loss (Kloth and McCulloch, 1996).

Interestingly, the level of hyperpolarization will influence the speed of the migration

as studied by Finkelstein et al. who increased cell membrane polarization via an over

expression of a CDC42-activated kinase, PAK4, resulting in migrational speed

increases compared to controls in the same electric field (Finkelstein et al., 2004).

Also, studies have determined that the orientation of the cells cleavage line is held

perpendicular to the electric field lines when close to the wound site and less so with

increasing distance, while the number of cells dividing increases with an increase in

electric charge at the wound site (Zhao et al., 1999; Song et al., 2002). These results

suggest the size of the electric field as a controlling factor in the level of cell

migration and division. Externally applied electrical stimuli will also change the

polarization of the cell membrane (Gross et al., 1986; Smith et al., 1991), but the

44


magnitude of this change is dependant upon the pre-existing surface charge (Gross,

1988).

It is also observed that the Bcl-2 gene family, which down-regulates apoptosis in

bone cells, causes hyperpolarization of membrane potential via the voltage gated K+

channel (Wang, 1999). Interestingly, fluid shear stress on osteocytes has been shown

to reduce cell apoptosis via the up-regulation of the Bcl-2 gene (Bakker et al., 2004),

suggesting that the mediator for such an effect may be by means of membrane

hyperpolarization. Others have also confirmed the ability of fluid shear to inhibit

osteoblast apoptosis (Pavalko et al., 2003a), highlighting that both the strain sensor

(osteocytes) and effector (osteoblast) cell numbers are maintained during mechanical

activation of bone formation.

Electrical stimulation with PEMFs appear to regulate the voltage gated channels to

direct cell migration in electric fields possibly via the electrical hyperpolarization of

the cellular membrane (Djamgoz et al., 2001).

Therefore, both electrical and mechanical stimulants have been observed to initiate

the same cellular control pathway by increasing cytosolic Ca2+ and activated

cytoskeletal calmodulin (El Haj et al., 1999; Brighton et al., 2001). Supporting the

theory that both stimulants are modulating cellular perturbations via the same sub

cellular actuator (Pilla 2002b).

As it is apparent that the polarization of the cell membrane is influential in many

normal cellular processes, then logically a mechanical strain leading to

hyperpolarization will ‘sensitise’ a cell to any naturally occurring internal electrical

field produced from a wound, and may play an integral role in the initiating and

possible acceleration of healing.

4.2 In Vivo Studies

Spadaro and co-workers, in two separate studies, used direct electrode implants in

the medullary canal of rabbit femurs and PEMFs (Spadaro et al., 1990, 1992), both

45


with and without mechanical micromovement, to find the efficacy of PEMFs on

osteogenesis. These studies concluded that bone formation takes place by the

mechanical medullary movement alone and that the electrical stimulation (PEMFs

and direct implanted current electrodes) did not unilaterally initiate osteogenesis.

However, the PEMF seemed to enhance medullary bone formation when it was

initiated by a concomitant mechanical stimulus. It was hypothesised that controlled

mechanical perturbations are fundamental while electrical perturbations require a

primary mechanical/chemical co-stimulus for effects to be enhanced (Spadaro,

1997).

It has been noted that cartilaginous tissue is more responsive to PEMFs than fibrous

tissue in the fracture gap (Gossling and Krompinger 1983), which has been proven

by in vivo studies noting successful healing once mechanical movement had been

controlled (Sharrard et al., 1982; Bassett, 1984). These studies highlight the level of

control a mechanical perturbation has over the efficacy of the applied PEMF

stimulus in a clinical situation.

Pilla (Pilla, 2002b) takes the approach that all cellular effects due to strain generated

potentials (SGP), ultrasound (US) and electromagnetic fields (EMF), are generated

from a time-varying electric field (E(t)). This field is derived from streaming

potentials (due to SGP), microcurrents (due to US) and directly from the EMF. He

states that E(t) controls “bone repair or remodelling” and “may be interchangeably

modulated using mechanical (including ultrasound) or electromagnetic signals. As

mentioned in Chapter 2, the time varying electric field modulates the voltage-

dependent binding, and in particular the binding of Ca2+ to Calmodulin. Intracellular

Ca2+ release is the most widely proven and accepted second messenger to be

mediated by both mechanical (Xia and Ferrier, 1992) and electrical (McLeod et al.,

1991; Lorich et al., 1998) stimuli and plays a major role in the ossification process

and signal transduction through the bone cell network. Brighton (Brighton et al.,

2001) also noted that both the electrical and mechanical stimulation led to an

increase in cytosolic Ca2+ and an increase in activated cytoskeletal calmodulin.

Direct mechanical modulation of the cell membrane (You et al., 2001; Pavalko et al.,

2003b) no doubt plays an important role in the mechanism of action for situations of

46


bone ossification and fracture repair. Applying mechanical stimulation will induce

fluid flow and hence streaming potentials. They produce time varying electrical

currents that are amplified by an array of bone cells in gap junction connection to

each other (Fear and Stuchly, 1998; Vander Molen et al., 2000). The amplification of

the transmembrane voltage then affects the kinetics of the calcium-calmodulin

binding process, speeding up growth factor release and hence bone formation (Pilla,

2002a,b).

Application of an exogenous electrical signal will directly induce the time varying

electrical field and hence the calcium/calmodulin binding process. It is conceivable

that both methods of stimulation work in unison with mechanical loading of bone.

It is also theorised that electrical stimulation may involve a direct cellular mechanical

perturbation from the biasing of the Brownian Ratchet mechanism (Otter et al.,

1997).

Comparisons of PEMF and mechanical stimulation effects have yielded many

interesting overlaps. The upregulation of intracellular 2nd messengers (Ca2+, IP3),

mRNA, prostaglandins and some growth factors (TGFβ) are consistent with both

PEMF and mechanical stimulation.

4.3 Conclusions

There has been limited work applied to the synergies of the different subcellular

theories into the mechanisms of action. The transmembrane potential perturbation

from time varying electrical fields (due to mechanical or electrical stimuli) and

chemical stimuli has not been rigorously correlated with the Brownian Ratchet

theory and the fluid flow amplification theory proposed by You (You et al., 2001)

has not taken into account the influence of strain generated potentials.

Mechanical actuation of bone sets up a cascade of events including the transduction

of force into a fluid shear stress within the canaliculi, impinging on the osteocytes,

which undergo a biochemical stimulation that is proposed to be prostanoid

47


dependant. These signals upregulate gene expression in the bone lining cells

(osteoblasts and osteoclasts) with autocrine/paracrine signalling using nitric oxide

(NO) and prostaglandins.

Additionally, the fluid flows create streaming potentials that are amplified by the

bone cells (via gap junctions). This has the power to affect Ca2+/ Calmodulin

binding, increasing the growth factor release for bone formation. Mechanical strain

on the bone cells also elicits hyperpolarization between the cell’s membrane and the

internal cytoskeleton. This will influence cell division, speed and direction of cell

migration within an electric field. Additionally it will also mediate cellular

differentiation. Thus a mechanical stimulus will ‘sensitise’ a cell to an external

electrical stimulus. This may explain the results reported by Spadaro who noted a

PEMF stimulus alone did not initiate intramedullary bone formation in rabbit tibia,

instead finding it enhanced the bone formation after initiation with mechanical

movement (Spadaro et al., 1990, 1992).

The specific combination of these two stimuli has only been mentioned a few times

in literature. Lee (Lee et al., 1982) stated, “general similarity in response to both

mechanical and electrical stress suggests common processes by which they modulate

cellular synthesis”. While Brighton (Brighton et al., 2001) noted the similarities in

the signal pathways: “all three forms of electrical stimulation (capacitive coupling,

inductive coupling and combined electromagnetic fields) as well as mechanical strain

led to…an increase in cytosolic Ca2+ and an increase in activated cytoskeletal

calmodulin”. Finally, Pilla (Pilla, 2002b) recently stated that these “combined

modalities” may lead to an “optimum therapeutic effect” for bone related disorders.

Hence this study sought to examine the effects of combined mechanical and

electrical stimuli on bone forming osteoblast-like cell cultures in vitro with the intent

to highlight any synergistic effects in cellular development that may be influential in

the in vivo environment.

48

Chapter 5: Initial PEMF Device

5 Initial PEMF Device

5.1 Introduction

In order to expose bone cell cultures to PEMFs, it was required to design and

develop a novel device. It was necessary for the device to produce a specific

magnetic and electric pulse specification while exhibiting a robust and consistent

signal over the cell culture stimulation period. Previous methods to stimulate bone

and bone cell cultures to varying electric currents (Bassett, 1989) have been

employed via solenoids, direct electrode implantation, permanent magnets and others

(Bassett et al., 1964; Zengo et al., 1976; Rubin et al., 1996; McCaig and Zhao, 1997;

Yamamoto et al., 2003). However, the most common in vitro method has employed

the use of an ‘air coil’ system.

A pilot study (Hodgkinson, 2001), conducted in 2001, evaluated different methods of

inducing a PEMF stimulus within bone cell cultures. It concluded that air coil

systems were superior to other methods of PEMF stimulation when considering

many factors such as size, usability, consistency, etc. Hence it was decided that this

method of stimulation would be used for this study.

The studied PEMF mimics a Food and Drug Administration (FDA) approved clinical

bone-healing device manufactured by EBI® (EBI Incorporated, Parsippany, NY,

USA) (Figure 5-1). This product has been used clinically since 1974 and is used to

combat pathologies such as fracture non-unions, congenital pseudoarthroses,

osteonecrosis and others (EBI, 2005).

49


Figure 5-1 A diagram of the induced electric field trace from a clinically available PEMF device that has been simulated for this project (Reproduced from Bassett, C.A. (1989) Crit Rev Biomed

Eng 17(5): 453).

The PEMF device was also designed with the capability of producing a second signal

that imitated the Biostim SPT® bone growth stimulator available throughout Europe

and manufactured by the Italian company, IGEA© (IGEA, Carpi, Italy).

Each of these devices affect in vitro cultures of bone cells (Sollazzo et al., 1997;

Lohmann et al., 2000). The previously mentioned pilot study conducted in 2001

(Hodgkinson, 2001) ascertained that the development of exposed bone cell cultures

did not show significant differences between the EBI® and IGEA® derived signals.

Thus it was decided to continue the research employing only the EBI® signal.

There are many gaps in the knowledge when trying to understand the principal

mechanisms of action from electrical stimulation of bone cells. One of these is the

effect PEMF timing has upon the cell cultures. This has been elucidated with

experimental testing and is discussed in Chapter 6.

5.2 PEMF Signal Generator

The PEMF Signal Generator creates the PEMF signal that is used to stimulate the

cell cultures. It controls the electrical current in the PEMF coil creating a magnetic

field (much like a solenoid) that induces an electric field in the bone cell culture

50

halla

This figure is not available online. Please consult the hardcopy thesis available from the QUT Library


layer. It is designed to amplify a DC voltage from the power supply, switching it at

the desired frequency. The pulse generator is powered by 20V direct current.

5.2.1 Design Specifications

This device was designed with PEMF signal specifications that consisted of a 5msec

pulse burst repeated at 15Hz (Figure 5-1). Each pulse burst creates an asymmetrical

‘quasi-square wave’ voltage trace within the cell layer at a frequency of ~4kHz.

Section 5.5 compares the design specifications with the actual signal output from the

PEMF device.

Peak PEMF coil current duration lasted for 204µs, producing a maximum magnetic

field of 1.3mT. Measurements of the magnetic field within the area used for

treatment of the bone cell cultures showed a ± 0.05mT fluctuation around this

maximum value.

The induced electric field trace contains 15mV positive amplitude and approximately

80mV negative amplitude on each individual pulse within the pulse burst and is

discussed further in Section 5.3.3.

5.3 PEMF Coil

All ‘air coil’ systems utilise a coil of conductive wire to produce the magnetic field,

which induces the subsequent electric field in the bone cell monolayer. Two PEMF

devices were developed for this study (Figure 5-2). Each consisted of two separate

coils connected together in parallel and placed 20mm apart, with dimensions 150 X

100 mm, designed for use with 24-well cell culture plates. Each coil was made up of

50 turns of 0.51mm diameter acrylic coated copper wire, together producing a

resistance of 2.3Ω.

PEMF coils were raised from the surface of the cell culture incubator tray by placing

them on specially designed plastic trays. The impetus for this design feature was to

minimise any possible distortion of the electromagnetic fields by currents induced in

51


the conductive shelves and to raise cell culture plates so the bone cell monolayer sits

in the centre of the two separate coil windings. Figure 5-3 shows the plastic tray built

for each PEMF coil, while Figure 5-4 shows the PEMF coil, tray and cell growth

incubator tray underneath. The coils were connected to the PEMF signal generator

producing the pulsed magnetic field perpendicular to the cell monolayer, inducing a

parallel-aligned electric field (see Section 5.3.2 and 5.3.3). Figure 5-5 uses a block

diagram to represent the set up of the PEMF device when used during biological

testing.

Figure 5-2 A diagram of the Initial PEMF Device. Shown are the two coil apparatuses, pulse generator and voltage power supply.

52


Figure 5-3 The plastic tray shown is positioned underneath the PEMF coil during calibration and cellular experimentation.

Figure 5-4 A picture of the gauss/tesla meter used for magnetic field measurements during testing with incubator tray.

53


2

PEMF

Pulse

Generator

DC Power

Supply

Bone Cell

Culture 2

Bone Cell

Culture 1

PEMF Coil Apparatus PEMF Coil Apparatus

1

Cell Growth Incubator

Figure 5-5 Set up of initial PEMF device during biological testing.

54


5.3.1 Measurement devices used in calibration of coil

A DC power supply (GW Laboratory DC Power Supply, Model GPD3030) was

connected to the PEMF signal generator for the creation of the PEMF signal. This

power supply was capable of a variable output voltage up to 30V. As mentioned

previously, it was set at 20V to supply to the PEMF signal generator (Figure 5-6).

A real time, two-channel oscilloscope (Tektronix TDS210 Two Channel Digital

Real time Oscilloscope) was used to measure and record all magnetic and induced

electric field data obtained from the PEMF coils. It contained a RS232 connection to

allow the downloading of captured waveforms to a personal computer for further

processing (Figure 5-6).

Magnetic field measurements from the PEMF coils were obtained via a Gauss/Tesla

meter (FW Bell Gauss/Tesla Meter, Model 5080). Measurements of maximum

magnetic field, minimum magnetic field and magnetic field RMS were taken using

the real time oscilloscope, which also facilitated visualisation of the real time,

dynamic, magnetic field (Figure 5-4).

All measurements of the induced electric field in the bone cell culture layer were

obtained by use of a coil dosimeter, also known as a coil probe. This small, tightly

wound coil of wire has a shunt resistor attached in parallel and when placed within

the energized PEMF coils, produces the first derivative waveform of the magnetic

field (i.e. induced electric field) as read from an oscilloscope. The coil probe consists

of 62 turns of 0.07mm nominal diameter wire coiled with 5mm internal diameter and

a 470Ω shunt resistor (Figure 5-7 and Figure 5-8). It was purpose built for

measurement of the induced electric fields from the PEMF.

The resistance and inductance of each of the PEMF coils and the coil dosimeter were

verified with an RCL meter (Fluke Programmable Automatic RCL meter, Model

PM6306). Figure 5-9 describes the set up of the PEMF device and the associated

instruments during calibration.

55


Figure 5-6 The real time oscilloscope used to measure magnetic field and induced electric fields and the DC power supply used for PEMF device.

Figure 5-7 The coil probe dosimeter, with shunt resistor shown on left, used to record induced electric fields from PEMF coils.

56


Figure 5-8 A close up of the coil from the coil probe dosimeter.

Gauss /

Tesla Meter

PEMF Coil Apparatus

PEMF Pulse Generator

DC Power

Supply

Coil

Dosimeter

Real Time

Oscilloscope

Figure 5-9 Set up of initial PEMF device during calibration testing.

57


5.3.2 Magnetic field from PEMF device

The magnetic field produced from the time changing current through the PEMF coils

was measured with a Gauss/Tesla meter (see section above for instrument

information). The meter was connected via an RS232 cable to a real time

oscilloscope for dynamic measurement. The magnetic field was then recorded and

downloaded from the oscilloscope to a personal computer for evaluation. During

static calibration of the magnetic field the PEMF signal generator was supplied with

30V DC. Therefore, maximum magnetic fields during characterisation were greater

than those experienced during cellular testing, which used a 20V DC power supply.

The higher voltage supply was provided to accentuate subtle differences in the

magnetic field measurements.

During the characterisation of the PEMF device all possible sources of secondary

(environmental) magnetic fields that could have possibly confounded results were

removed/masked or minimised. These included static (permanent magnets) and time

varying (power cords, etc) magnetic fields plus other peripherals such as R.F. and

microwave sources and magnetic door latches, etc. In order to reduce the possibility

of electrostatic (fur, clothing) interference non-synthetic clothing was worn and all

possible charge carrying equipment including the person conducting the

experimentation was earthed.

5.3.2.1 Magnetic Field map within PEMF coil apparatus Measurements of the maximum magnetic field within the energized PEMF coil were

quantified by making three horizontal sweeps along two axes that centred on the cell

culture wells and another along the centre line (when using a 24 well cell culture

plate) (Figure 5-10). As discussed in Chapter 6, only the centre 8 wells were seeded

with cells during experimental testing and therefore only these wells were tested

during horizontal sweeps. This was due to the magnetic field only showing

homogeneity within this cental region.

Symmetry determined that only one line of measurement was required for

representation of the magnetic field experienced by bone cell cultures (Line No. 1,

Figure 5-10). All measurements were taken half way between the two, parallel

58


wound, coils in each apparatus to maintain homogeneity of the magnetic field

measurements. Bone cell cultures, when undergoing PEMF stimulation, were

positioned at this same level (Figure 5-11).

Line

No.

3

2 1

Centre 8 wells seeded

with bone cells

Figure 5-10 A top view of the PEMF coil showing marked positions used for calibration measurements of the magnetic field.

Magnetic Field

measurement level

Figure 5-11 A side view of the PEMF coil, with the level of the magnetic field measurements marked with a black dotted line.

59


The maximum magnetic field along each line shown in Figure 5-10 is shown in

Figure 5-12. Line 1 and line 3 were tested for conformity due to symmetry and

therefore only one line is displayed. Line 2 was originally tested to achieve an overall

picture of the PEMF coil’s magnetic field along the centre line. This demonstrated

the normal magnetic field pattern associated with coiled conductors, whereby the

magnetic field increases to a maximum at the surface of the conductor proportionally

with the distance. Measurement for each cell culture demonstrated that the maximum

magnetic field of the PEMF device was 2.08mT, while the range between the centre

two and outer two wells was bounded within a 0.1mT variation.

0

0.5

1

1.5

2

2.5

-60 -40 -20 0 20 40 60

Distance from centre of coil (mm)

Max

imum

Mag

netic

Flu

x (m

T)

Centre Culture wells

Figure 5-12 Maximum magnetic field with distance from centre of PEMF coil. Shown are measurements along the centre line (25 and 50mm from centre of coil; black line) of the PEMF coil apparatus and at the culture well positions (~11 and ~32mm from centre of coil; grey line).

See Figure 5-10 for diagram of positions.

5.3.2.2 Influence of Stainless Steel Shelf on Magnetic Field During cellular testing, the PEMF device was housed within a cell growth incubator

(see Chapter 6 for more details). Cell cultures were placed on shelves manufactured

from copper enriched stainless steel. A study was conducted on the influence these

shelves had on the magnetic field experienced by the cell cultures. The centre four

wells of row 1 from Figure 5-10 were used, as the symmetrical relationship of the

60


coil negated the need for measurement of row 3. An increase of 0.02mT was

exhibited for the centre two wells while placed upon the tray (Figure 5-13) compared

with the non-metallic bench. This difference is five times smaller than the 0.1mT

variation between the centre and outer wells of the central portion of the 24-well cell

culture plate and was thus not of significance.

0

0.5

1

1.5

2

2.5

1 2 3 4 5 6

Cell Culture Well Location

Max

imum

Mag

netic

Flu

x (m

T)

W/O Tray Tray

Figure 5-13 Maximum magnetic field with each cell culture well position. Measurements were taken when the PEMF apparatus was placed on a non-metallic bench (black line) or on the

metallic incubator tray (grey line).

5.3.2.3 Dynamic Magnetic Field Measurements Dynamic magnetic field signals produced from the active PEMF device were

obtained from the gauss/tesla meter placed within the centre of the PEMF coil

apparatus. The power supply to the PEMF signal generator was 20V DC. When

signals from each apparatus are graphed together (Figure 5-14), no differences can be

seen. Measurements were also taken at each of the cell culture wells, however only

minor differences from those at the centre of the coil were noted.

61


-1.00

-0.50

0.00

0.50

1.00

1.50

0.00 1.00 2.00 3.00 4.00 5.00 6.00

Time (milliseconds)

Mag

netic

Fie

ld (m

T)

Coil 1 Coil 2

Figure 5-14 Comparison of dynamic magnetic field measurements from PEMF coil apparatus 1 and 2. No discernable difference is seen between each trace.

5.3.3 Induced Electric Field from PEMF

As mentioned previously, the induced electric field from the PEMF was measured by

means of an inductive search coil, also know as a coil dosimeter (Figure 5-7 and

Figure 5-8). When the coil was placed in a magnetic field, a current was induced in

the coil and hence a voltage across the parallel resistor (Figure 5-15).

This voltage provided an estimate of the induced electric field within the cell culture

and did not represent the exact electric field a bone cell would experience. However,

in vivo measurements of induced electric fields in deionised bone tissue undergoing

exogenous PEMF exposure have shown similarity to induced electric fields in coil

dosimeters (Hart, 1987).

Although the magnetic field is uniform across each of the seeded wells in the cell

culture plate, the electric field will not be because of the conductive growth medium

(assumed to have a conductivity of 1410mS/m). The electric field strength induced at

any position inside the well is dependant on radial distance from the centre. It has

been theoretically and experimentally validated that the current density follows a

62


sinh function decay, with the maximum of 2µA/cm2 at the outer well boundary

decaying to zero at the centre of the well (McLeod et al., 1983).

-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02

0.0000 0.0006 0.0012 0.0018 0.0024 0.0030 0.0036 0.0042 0.0048 0.0054

Time (sec)

Volta

ge (V

)

Figure 5-15 The induced electric field trace in the coil probe dosimeter when placed within the active PEMF coil.

5.3.3.1 Comparison of induced EMF from each PEMF coil

apparatus Both coil apparatuses were tested for conformity before cellular testing.

Measurements of the 15Hz EBI pulse trace were taken from both coil apparatuses

during the same continuous operation of the PEMF device. The pulse traces showed

no significant deviation from each other during the EBI 15Hz pulse burst, with only

minor differences in the negative voltage amplitudes (Figure 5-16).

63


-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02

0.04

0 0.0006 0.0012 0.0018 0.0024 0.003 0.0036 0.0042 0.0048 0.0054

Time (sec)

Volta

ge (V

)

Outlet 1 Outlet 2

Figure 5-16 A comparison of the induced electric field from the PEMF coil apparatus 1 and 2. No discernable difference is seen between the traces.

5.4 PEMF Device Characterization in Cell Growth

Incubator

To quantify any distortional effects the cell growth incubator may have had on the

PEMF, measurements of the induced EMF within the coil dosimeter were taken

while the coil apparatuses were located in the incubator. These were then compared

with results obtained while the device was calibrated on a laboratory bench (Figure

5-17). All care was taken to remove spurious sources of magnetic field error that may

have been present in the incubator such as highly magnetic cell sorting devices,

audio speakers and other ferromagnetic materials. Repeat measurements (n = 3) were

taken to show significance in the results.

There was an average reduction of 8.8mV in the incubator’s negative voltage spike

when comparing the induced EMF traces from the external calibration pulse trace

(laboratory bench) and incubator located pulse trace.

64


-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02

0.0000 0.0006 0.0012 0.0018 0.0024 0.0030 0.0036 0.0042 0.0048 0.0054 0.0060

Time (sec)

Volta

ge (V

)

Incubator Pulse Calibration Pulse

Figure 5-17 A comparison of the induced electric field from the PEMF signal when the device was located in the cell growth incubator and on the laboratory bench during calibration.

During experimental testing with cell cultures, control plates were placed in the same

incubator as the PEMF exposed cell cultures. This necessitated the need to measure

and record any peripheral PEMF signal within the incubator. These measurements

were taken at various locations in order to detect any change in the base level EMF

normally present in the incubator. These results highlighted that fringing fields from

the device drop off significantly with distance. Finally, an evaluation of the

magnetic field experienced by the control cultures during cellular testing was made

to detect any change induced in the control signal from the PEMF device. This

comparison is shown in Figure 5-18 and shows that there was no change in the

background magnetic field detected at control culture positions with the application

of the PEMF device in the same incubator.

65


-0.40

-0.20

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

0.0000 0.0008 0.0016 0.0024 0.0032 0.0040 0.0048 0.0056 0.0064 0.0072 0.0080 0.0088 0.0096

Time (sec)

Mag

netic

fiel

d (m

T)

PEMF Culture Control Culture

a

b

Figure 5-18 A comparison of the magnetic field at PEMF exposed cell culture wells (trace ‘a’) and those in the control cell culture plate (trace ‘b’).

5.5 Comparisons of PEMF Signals

As discussed in the introduction, the PEMF device was designed and constructed

with the aim of simulating a PEMF signal from a clinically available bone-healing

device. Efforts were made to imitate this signal precisely, however variations in the

size, number of turns, wire gauge and orientation of the PEMF coil apparatus plus

variations in the coil probe dosimeter used to quantify the induced EMF made exact

replication very difficult. Also, neither the clinical coil apparatus nor a description of

the coils was available for study. Figure 5-19 shows the comparison of the reported

clinical device signal (Figure 5-1) with that produced from the device whilst tested

on a laboratory bench. There was an average decrease of 70mV in the negative

voltage spike from the purpose built device compared with the specification of the

EBI signal.

66


Only a single pulse within the pulse burst is shown for clarity of observation. There

was no variation in the timing (horizontal axis) of individual pulses and pulse bursts

between the clinical signal and that from the designed PEMF device.

-0.1500

-0.1250

-0.1000

-0.0750

-0.0500

-0.0250

-

0.0250

0.0500

- 12 24

36 48

60 72

84 96

108 120

132 144

156 168

180 192

204 216

228 240

252 264

276 288

392 Time (microsec)

Volta

ge (V

)

b

a

Figure 5-19 A comparison of the ideal and the induced voltage measured from the PEMF device. Trace 'a' (dotted line) signifies the desired waveform while Trace 'b' (solid line) is the

recorded signal.

5.6 Discussion

Implementation of a PEMF stimulus imparted on in vitro bone cell cultures has been

conducted using many different techniques. The method employed in this project

used an air coil system, which consists of two coils of wire wound together in a

rectangular shape allowing 24 well cell culture plates to be exposed to the PEMF.

The coils of wire are connected to a PEMF signal generator powered with 20V DC.

The air coil system is simple and robust and is also straightforward for calibration

and characterisation. However, as discussed in the introduction, this project has not

quantified the exact induced electric signals in the bone cells.

67


Hydrated tissues show strong charge relaxation whereby the induced electric field

trace, from externally applied PEMFs, displays a large initial (positive) and final

(negative) spike in response to the initial and final time rate of change in the

magnetic field pulse. The induced electric field in between these initial and final

spikes is almost non-existent due to the movement of counter charge ions within the

tissue negating the potential difference.

However, induced electric waveforms (from PEMFs) in search coil dosimeters

exhibit a similar waveform pattern to that of non-living, deionised bone tissue.

Consequently, it has been used as a reference tool comparing the desired clinical

signal with that produced by the purpose built PEMF device for this project.

Another important factor with air coil systems when using PEMF stimulation is the

orientation of tissue/cells and PEMF coil. Cook and Bassett have shown that the

orientation of the tissue in the PEMF will vary the induced EMF (Cook and Bassett,

1983), especially for tissues with a distinct long axis such as bone or tendon. When

the magnetic field is parallel to the long axis, a greater initial and final voltage spike

occurs for each pulse of the PEMF. Furthermore, perturbation of the induced electric

fields within in vitro cell cultures is influenced by PEMF coil orientation. McLeod et

al. found that horizontal coil orientation (as used in this project) and vertical coil

orientation influences the magnitude of the electric stimulus at different locations

within the cell culture well (McLeod et al., 1983).

The PEMF device showed reasonable homogeneity over the centre rectangular area

of the PEMF coil apparatus (covering the 8 wells seeded in the cell culture plate

during cellular testing). There was a 0.1mT variation in maximum magnetic flux

from the centre to the edge of the rectangle (Figure 5-12). The stainless steel

incubator tray the coil apparatuses were placed on during cellular testing did not

effect these measurements and PEMF signals from both coil apparatuses conformed.

Differences between the desired EBI derived PEMF signal and that produced by the

purpose built device were apparent. There was a reduction of 70mV recorded in the

negative voltage spike when compared with the desired signal. This can be attributed

to differences in coil apparatus and coil dosimeter specifications used for recording

the signals from the EBI and purpose built PEMF devices.

68


Differences were also seen when the PEMF coil apparatuses were placed in the cell

growth incubator where they were located during cellular testing. There was an

average reduction of 8.8mV in the PEMF signal’s negative voltage spike whilst in

the incubator when compared to laboratory bench testing.

The influence a variation in the magnitude of the negative voltage spike might have

upon in vitro or in vivo situations is not discussed in the available literature.

However, as it has been observed that the time rate of change and not the magnitude

of the magnetic field is the more significant influence over bone cell cultures

(Pienkowski et al., 1992; Dennis et al., 2003). Therefore, it can be argued that this

variation is inconsequential and the timing of the pulse burst and its associated

individual pulses is more important.

No detectable PEMF signals were present at the control culture positions within the

incubator. Fringing PEMF fields dropped off rapidly from the stimulated region

within the centre of the PEMF coil apparatus reducing to zero at an approximate

distance of 30cm from the edge of each coil.

5.7 Conclusions

An air coil PEMF device has been designed, developed and produced that

successfully imparts a 15Hz pulse burst style electromagnetic field on in vitro

cultures of bone cells. The characterisation of the device has successfully described

the nature of the stimulus experienced by the bone cell cultures but not the exact

induced electric signal. This device can be used for the centre 8 wells of a 24 well

cell culture plate when testing within cell growth incubators (CO2 Incubators).

Chapter 6 discusses the use of the device.

69

Chapter 6: PEMF Stimulation of Cultured Bone Cells

6 PEMF Stimulation of Cultured Bone

Cells

As discussed in the Chapter 5, a pulsed electromagnetic field (PEMF) device was

designed and produced with the capability of imparting a 15Hz pulse burst style

electromagnetic field on in vitro cultures of bone cells. Although in vitro PEMF

exposed cell culture studies have been performed previously, there has been no

consistency with PEMF exposure dose and timing. In order to evaluate the

performance of the PEMF device built for this project, an investigation of the effect

PEMF exposure timing has on bone cell cultures was undertaken.

Studies focusing on exposure timing (Cane et al., 1997; De Mattei et al., 1999) used

differing doses of PEMF stimulation and varied exposure time. To the authors’

knowledge, no published studies have evaluated pure timing effects of PEMF

stimulation on osteoblast-like cell cultures. It is highly likely that cell culture

confluence and cell division rate have an effect on PEMF stimulation effects as

highlighted by a recent study (Diniz et al., 2002). Cultures in differing states of

cellular maturation, when exposed to PEMF stimulation, were found to have

different reactions. Initial seeding density is also a potent mediator of cell cultures

exposed to PEMF stimulation (Noda et al., 1987; McLeod et al., 1993; Pavlin et al.,

2002). Thus by evaluating the effects of PEMF stimulation at differing seeding

density and times of cell growth, the mechanism of cellular responses to PEMFs

might become clearer.

The aim of this particular study was to evaluate the effects of PEMF stimulation

timing and cell density on the development of the osteoblast-like cell cultures. As an

experimental model, an osteogenic human osteosarcoma cell line (SaOS-2) was used

and the effects of PEMFs on cell proliferation, measured by 3H-Leucine

incorporation and cell differentiation via alkaline phosphatase production was

evaluated.

70


6.1 Materials and Methods

6.1.1 PEMF Device

As discussed in Chapter 5, the PEMF signal mimicked the FDA approved clinical

bone-healing device manufactured by EBI® (EBI Incorporated, Parsippany, USA).

6.1.2 Cell Cultures

A human osteosarcoma cell line, SaOS-2 (ATCC No: HTB-85, Rockville, U.S.A.),

was used for these experiments. The cells were cultured in minimum essential

medium alpha (αMEM) supplemented with 10% fetal bovine serum, 1% penicillin -

streptomycin diluted from stock solution [both 5,000 U/ml] and 0.01% gentamicin

[10mg/ml] (All from Gibco, Grand Island, U.S.A.).

The cells were less than 10 passages from original cell stock. For experimental

procedures the cells were seeded (with 1ml of growth media) at densities of either

50,000 or 25,000 cells per well into the centre 8 wells of 24 well cell culture plates.

Each well has an effective growth area of 1.9cm2 (Nunc, Roskilde, Denmark). Only

the centre 8 wells of the plates were seeded due to inhomogeneity of the PEMF

stimulant at the edges of culture plates.

For each PEMF protocol, after seeding, two 24 well plates were assigned to the

PEMF exposed and two to the control group. The cells were allowed to attach for 2

hours before experimentation. All experimental procedures were conducted within a

CO2 incubator at a temperature of 37˚C in an atmosphere of 95% air/5% CO2 and

100% relative humidity. This was performed 3 times for each protocol except for

Protocol 1 that was repeated twice due to technical difficulties. Control cell culture

plates were prepared in an identical manner and placed within the same incubator.

Control plates experienced a background magnetic flux of ±1G.

71


Measurements of temperature within the PEMF exposed cell culture wells during

operation of the coils (with door of incubator closed) were conducted with a

thermocouple. These results showed no increase in temperature above the basal

37˚C, and demonstrated that there were no heating effects on the exposed cell

cultures.

6.1.3 Experiments

The cell cultures were subject to four protocols, each with a different period of

PEMF stimulation. All protocols had a total of 24 hours PEMF exposure, eliminating

dose response as a variable and allowing PEMF timing to be studied over the three-

day period (Figure 6-1). After 72 hours, cellular proliferation and differentiation

were assayed.

Da

y 1

Da

y 2

D

ay

3

8 Hr 24 Hr

1 2 3

Protocol Number

Experiment Stopped – Cellular proliferation and

alkaline phosphatase measured

4

Figure 6-1 PEMF exposure protocols on osteoblast-like SaOS-2 cells to quantify effects on cellular proliferation and differentiation. Shaded sections denote PEMF exposure while clear

sections denote normal cell culture conditions.

72


6.1.4 Proliferation

De novo protein synthesis was assayed from 3H-leucine incorporation into acid

precipitable proteins (3H-leucine, Amersham International, Little Chelford,

UK)(Freshney, 2000). This measure gives an indication of growth in cell number. At

the beginning of each experiment, the radiolabelled amino acid was added to the

growth media (1µCi per millilitre of growth media) and was thus available for

protein incorporation throughout the entire 72-hour experimentation protocol. Cells

were rinsed with Hanks Balanced Salt Solution (Gibco, Grand Island, NY, U.S.A.)

and treated with 5% trichloroacetic acid. Cell precipitates were then washed with

sterile distilled water and solubilized in 0.5M NaOH / 0.1% Triton X-100 for 12

hours on a shaker table. Samples were manually triturated prior to sampling to ensure

homogeneity.

Triplicate samples of 100µL from each of the 24 separate wells were then counted in

a liquid scintillation counter (Wallac MicroBeta TriLux, Boston, MA, U.S.A.) with

150µL of scintillation fluid (Optiphase SuperMix, Perkin Elmer, Boston, MA,

U.S.A.).

6.1.5 Differentiation

Levels of alkaline phosphatase in the culture medium were measured for an

indication of early stage osteoblastic differentiation. 20µL samples of conditioned

culture medium were admixed to 20mM of p-nitrophenyl phosphate (Sigma-Aldrich,

St. Louis, MO, U.S.A.). Phosphatase activity was determined by measuring light

absorption at a wavelength of 405nm using a spectrophotometer (Beckman Coulter,

Fullerton, CA, U.S.A.). Repeat measurements were obtained immediately, at one

minute and at two minutes after addition of the cultured medium (Sigma Diagnostics

Alkaline Phosphatase Procedure No. 245). These three measurements were then used

to calculate the rate of increase in light absorption. The rate was checked for

linearity, and used to calculate alkaline phosphatase volume by means of the

following equation:

Alkaline Phosphatase (Units/Litre) = (∆A per min * TV * 1000) / (18.45 * SV * LP)

73


Where TV represents total volume, SV is sample volume, LP is length of light path,

18.45 is millimolar absorptivity of p-nitrophenol at 405nm and ∆A per min is the

change in absorbance per minute. One unit of alkaline phosphatase activity is defined

as the amount of enzyme that will produce one micromole of p-nitrophenol per

minute.

6.1.6 Statistical Analysis

To facilitate comparisons, all PEMF exposed cultures in each protocol were

normalised against control values, which were considered as 100% percent.

All experimental data for each protocol and cell density were pooled and averaged to

produce each proliferation measurement. Error bars are expressed as ± Standard

Error of the Mean (SEM) on each graphed result.

Alkaline phosphatase error bars were calculated as:

(SEM of light absorbance rate/Average light absorbance rate) * Alkaline

Phosphatase volume.

A non-parametric analysis of variance test, Kruskal-Wallis, was performed between

results for all four protocols and seeding densities in both the proliferation and

alkaline phosphatase data sets. Student’s t tests were performed for all protocols and

seeding densities (that were normally distributed) to compare PEMF exposed and

control cultures. When data sets did not satisfy normality calculations, the non-

parametric Mann Whitney U Test was applied, which in the case of proliferation was

both seeding densities in Protocol 1, 25,000cells/well seeding density in Protocol 3

and 50,000cells/well seeding density in Protocol 4. P values of 0.05 or less were

considered significant. Alkaline phosphatase measurements were limited to low

replicate numbers (between n=4 and n=8 repeats) due to technical difficulties and

although statistical significance was not achieved, consistent trends were seen in the

data.

74


6.2 Results

The raw data from these experiments is presented in Appendix A. The results

presented here follow analysis of the data to assess the effects of the timing of the

PEMF stimulation and the differences between stimulated and non-stimulated

cultures.

Observation of the data revealed no statistical difference of either proliferation or

differentiation with the timing of the PEMF stimulation (this important finding has

been accepted for publication; Hannay et al., 2005). Whilst the timing did not

significantly affect the results, exposure to the PEMF did affect the cell cultures and

the significant findings are presented in the following sections. Results are reported

as a percentage of the control for each exposure protocol. Linear regression of light

absorbance rates, used to calculate alkaline phosphatase volume, ranged from an R2

value of 0.8508 to 0.9976.

6.2.1 Protocol 1

Eight hours of PEMF stimulation each day for three days was conducted. Compared

to the control cultures, PEMF exposed cultures showed a significant 9% and 14%

average reduction in proliferation as measured by de novo protein incorporation of 3H-leucine for the two seeding densities (Figure 6-2).

Assays for alkaline phosphatase activity exhibited a non-statistically significant

increase in PEMF exposed cultures with respect to control cell cultures at both

seeding densities (16% for 25,000 cells/well and 12% for 50,000 cells/well, Figure

6-3).

75


6.2.2 Protocol 2

Stimulating cultures with PEMFs for 24 hours on the first day followed by no

stimulation for the next two days produced a reduction in proliferation with respect

to control cultures. Variability was observed between cultures seeded at 25,000 and

50,000 cells/well. In cultures seeded at 50,000 cells/well, no significant reduction in

PEMF exposed proliferation relative to controls was observed. In contrast, a

significant 11% reduction was observed in cultures seeded at 25,000 cells/well

(Figure 6-2).

Alkaline phosphatase activity was non-significantly increased in PEMF exposed

cultures with respect to the controls (20% for 25,000 cells/well and 14% for 50,000

cells/well, Figure 6-3).

6.2.3 Protocol 3

Cells exposed to protocol 3 did not exhibit a significant down-regulation of

proliferation in PEMF exposed cultures (Figure 6-2).

Cultures exposed to protocol 3 showed increases of alkaline phosphatase activity.

There was a large increase of 38% for 50,000 cells/well, while 25,000 cells/well

cultures exhibited a 19% increase, consistent with other protocols at that seeding

density (Figure 6-3).

6.2.4 Protocol 4

Although there was a consistent trend toward decreased proliferation from both

seeding densities, this protocol was only significant for the 50,000cells/well result

(17% decrease, Figure 6-2).

The 50,000cells/well result for this protocol was the only one that did not show the

trend for increased alkaline phosphatase production. However, the 25,000 cells/well

seeding density showed an increase in enzyme production of 22% over control

cultures (Figure 6-3).

76


0

20

40

60

80

100

120

140

160

Protocol 1 Protocol 2 Protocol 3 Protocol 4

Per

cent

age

(%)

25,000 50,000 Control

# ## #

Figure 6-2 Proliferation, described as a percentage to controls, of PEMF exposed cell cultures from each PEMF exposure protocol at 25,000 and 50,000 cells per well seeding density. # Indicates statistical significance (P < 0.05). Error bars are +/- standard error of the mean.

0

20

40

60

80

100

120

140

160

Protocol 1 Protocol 2 Protocol 3 Protocol 4

Per

cent

age

(%)

25,000 50,000 Control

Figure 6-3 Differentiation, described as a percentage to controls, of PEMF exposed cell cultures from each PEMF exposure protocol at 25,000 and 50,000 cells per well seeding density. Error

bars are +/- standard error of the mean.

77


6.3 Discussion

This study has investigated the effects of pulsed electromagnetic field stimulation

timing and cell density upon the development of the osteoblast-like cell cultures.

The data indicate that cells exposed to PEMFs exhibit reduced proliferation and

suggest they also exhibit a more differentiated phenotype due to the increased

alkaline phosphatase production. These results are consistent with other studies using

osteoblast-like cells, which have also shown decreased proliferation and increased

differentiation of cultures exposed to PEMFs (McLeod et al., 1993; Lohmann et al.,

2000; Vander Molen et al., 2000). The study of (Lohmann et al.) exposed MG63

cells to an 8 hour period of PEMF stimulation over 1, 2 or 4 days. However, in

contrast to this study, the cells were grown to confluence before PEMF stimulation.

The phenotypic state of the cells has been shown to influence exogenous PEMF

stimulation effects on cell development, specifically, the greater a cell’s

differentiation as measured by means of alkaline phosphatase, the less proliferation is

achieved (Diniz et al., 2002). MG63 cells have very low basal levels of alkaline

phosphatase activity compared with the human derived SaOS-2 cells (Rodan et al.,

1987) and this may explain the discrepancy between our study and that of others who

have used immature osteoblast-like cell lines and found proliferation increases from

PEMF exposure (Sollazzo et al., 1997; De Mattei et al., 1999; Chang et al., 2004). It

also may explain why we have achieved similar results to that of Lohmann et al.

(2000) when our PEMF stimulation methods are different.

Important regulators of an osteoblast’s ability to communicate and respond to

exogenous stimuli, such as PEMFs, are gap junctions (Schiller et al., 2001). These

are specialized intercellular channels for movement of small molecules and ions

between adjacent cells and directly affect electrical conductance (induced from an

exogenous PEMF stimulant) within the cell monolayer (Sreedharan and Zhang,

2003). This electrical conductance is amplified via cell coupling and is a proposed

regulator of PEMF stimulation effects (Muehsam and Pilla, 1999; Pilla, 2002b).

78


Recent studies show that the PEMF stimulated decrease in proliferation is

independent of gap junctional coupling, while increased enzyme activity (alkaline

phosphatase) levels are still dependant on the electrical communication achieved

through gap junctions (Vander Molen et al., 2000). Studies on gap junctional

expression have concluded that PEMF exposure decreases the amount of gap

junctional communication via a decrease in the mRNA expression of the gap

junction protein connexin 43 in well-differentiated osteoblasts and osteocyte-like

cells (Lohmann et al., 2003). However, (Yamaguchi et al., 2002) reported that the

decreased intercellular communication observed in immature osteoblasts from PEMF

stimuli was nullified when using originally well differentiated cells. This suggests

that communication through gap junctions between adjacent SaOS-2 cells used in

this study may not have been affected by the PEMF stimulant. It has also been noted

that these cells naturally show very little gap junction communication (Donahue et

al., 1995). Therefore, it is possible that the SaOS-2 cell line is not as sensitive to the

PEMF stimulus as a cell line that expresses a more pre-osteoblast or greater gap

junctional signalling phenotype.

There was no obvious difference between the protocols, which suggests that the

timing of PEMF stimulation may not be a critical feature. It has been reported that as

little as 30 minutes of PEMF stimulus provides significant increases in proliferation

for in vitro cultures of osteoblast-like cells, while the effects of stimulation taper off

after 24 hours (De Mattei et al., 1999). Thus a shorter period of stimulation may have

a greater influence over cellular development and could explain why protocol 1 with

its repeated stimulation periods of 8 hours per day over the three days is the only

consistently significant protocol for both seeding densities, while the longer exposure

protocols (protocol 2, 3, 4) do not show as much consistency.

McLeod et al. [1993], using a protocol similar to our protocol 4, demonstrated that a

‘window’ effect occurs such that in vitro cultures of ROS 17/2.8 osteoblast-like cells

with high (50,000 cells/cm2) or low (6,000 cells/cm2) seeding densities exhibited an

apparent reversal in the general trend of increased PEMF induced alkaline

phosphatase. ROS 17/2.8 osteoblast-like cells have been matched as closely

resembling SaOS-2 cells in osteoblastic qualities (Rodan et al., 1987) and may

79


explain the differences seen in the differentiation result from protocol 4 compared

with the other protocols.

Some limitations of this study include the very small but consistent magnetic flux

experienced by the control cultures. This could potentially mute the results seen from

the PEMF exposed cultures when making comparisons with the controls. However,

controls from each protocol underwent the same small exposure, cancelling any

influence it may have had on the PEMF timing results. The number of repeat

measurements of alkaline phosphatase volume which stifled statistical significance

and the resolution of the temperature measurements with the thermocouple were also

not ideal, while another small confounding factor could be associated with vibration

of the cultures from movement in the incubator shelves initiated from the mechanical

expansion of the coil during operation. Again, all cultures underwent this exposure,

cancelling any influence it may have had on the PEMF timing results.

6.4 Conclusions

The results indicate that a 15Hz PEMF stimulus on monolayers of an osteoblast-like

cell line leads to a depression in proliferation with a concomitant increase in alkaline

phosphatase production. Since alkaline phosphatase is related to bone cell

differentiation and bone mineralisation, these results support the hypothesis that a

commercially available PEMF device will stimulate an osteoblast-like cell line into

an increasing state of maturity. Applying the stimulus at different times following

culture seeding did not appear to affect the response of the cells (although there is

evidence that this may be due in part to the density of the cell cultures during PEMF

exposure and/or sensitivity to PEMFs in the cell line studied). These results provide

more evidence to help explain the mechanism by which clinical PEMF stimuli alter

in vitro cultures of bone cells.

80

Chapter 7: Design of the Dual Stimulus Device (DSD)

7 Design of the Dual Stimulus Device

(DSD)

The first three chapters review the similarities in cellular signalling and other

biological transduction pathways of mechanical and electrical stimuli in bone. Both

are present in bone during normal development, therefore elucidation of stimuli

interaction effects will help clarify transduction pathways and equip researchers with

more tools to modify the process of bone regeneration for therapeutic benefit.

As outlined in Chapter 6, electrical stimulation of bone cells by way of externally

applied electromagnetic fields reduces the level of bone cell proliferation while

concomitantly increasing their phenotypic maturity. It can be argued that this is a

clear indication of an accelerated bone cell development. However, this previous

study fails to correctly mimic the situation present during normal bone development

or bone fracture healing, where both electrical and mechanical stimuli are present.

A study of simultaneous mechanical and electrical stimuli exposure on in vitro

cultures of bone cells has not been conducted previously, and was the motivation of

this section of the PhD project.

Chapters 2 and 3 each described how the individual stimuli have an influence over

the developing form and function of bone. Alterations in mechanical strain during

foetal development (Tanck et al., 1998) and bone regeneration after fracture (Le et

al., 2001) result in changes of activated cellular biochemical pathways and

subsequent bone formation. Chapter 4 outlines some of the known interactions

between the two stimuli and potential results.

For any study of the synergistic effects of the two stimuli to be conducted, a device

capable of imparting both stimuli at the same time is required. This was achieved by

the use of a novel Dual Stimulus Device (DSD), capable of creating a controlled

mechanical and electrical environment for in vitro cultures of bone cells.

81


7.1 Overview of DSD

An initial assessment of the key design issues for the DSD was made. This step

helped identify the most important factors to control during the design and

development of the device. A subsequent review of current technology and an

evaluation of the design alternatives were conducted.

In the interests of scientific consistency and design calibration, it was decided the

DSD would use a PEMF air-coil system similar to that discussed in Chapter 5.

Therefore the design of the DSD was required to include an externally attached

PEMF air-coil that was immune from any interference induced via the DSD’s

mechanical stimulus.

As discussed in Chapter 3, previous devices that mechanically stimulate biological

cells have focused on larger strains than those experienced by bone cells in vivo and

have failed to take into account the presence of higher frequency components in the

strain. Therefore, it was decided that the DSD would aim to more accurately replicate

the in vivo environment the bone cells experience and introduce both these factors.

7.1.1 DSD Design Issues and Requirements

Ten key design issues of the DSD were identified and each given a weighting out of

ten in accordance with its importance. This weighting was used to determine a score

for ranking DSD design alternatives as described in the following section.

7.1.1.1 Stimuli Consistency The importance of the DSD to provide consistent mechanical and electrical stimuli is

high. Small variations in either stimulus may affect the cellular metabolic processes

significantly. Weighting = 9/10.

82


7.1.1.2 Cell Line Flexibility The ability to use different cell lines in the DSD was desirable. An example would be

skin (keratinocytes) or endothelial cells, which are both placed under strain in vivo.

However, as the scope of this project did not encompass such studies this issue was

less of a factor in the design process and was given a lower weighting. Weighting =

3/10.

7.1.1.3 Ease of Manufacture A reduced complexity in the DSD design reduces the possibility for manufacturing

flaws to occur, which could undermine the success of the DSD. A simplified design

also decreases experimental downtime when making future design modifications as

manufacturing time is reduced. This also enhances the future commercial potential of

the device. However, this issue is still of moderate importance in relation to the other

ten design issues. Weighting = 5/10.

7.1.1.4 Ease of Maintenance Maintenance and ability to access/replace the cell substrate material with maximum

repeatability is crucial and thus the weighting will be high. Being able to access the

internal structure of the DSD is also very important for cleaning and calibration

before cellular testing. Weighting = 8/10.

7.1.1.5 Growth Media Fluid Flow Strain Most devices employing substrate stretching for their mechanical stimulus set up

fluid shear strains from the movement of growth media. This is due to changes in

fluid pressure and can possibly confound cellular effects seen from substrate

stretching. Increases in pressure will occur when displacements of the fluid are

increased such as used by devices employing large out-of-plane cell substrate strains

(1%ε +). This factor needs to be controlled and/or quantified to a sufficiently

accurate degree. Weighting = 6/10.

7.1.1.6 Mechanical Strain Signal Flexibility The ability of the device to accurately produce high frequency strain with variable

waveform patterns was desirable. For example, waveforms with fast rise times but

83


slow relaxations could be used to mimic situations such as impact loading of bone

and the associated cellular responses. Current studies have highlighted the

omnipresent low magnitude, high frequency strains of in vivo bone as potential

mediators of bone integrity maintenance. It is an aim of the project to study this

method of mechanical stimulation on bone cells and as such a maximum weighting

of ten has been given. Weighting = 10/10.

7.1.1.7 Reliability Reliability and repeatability of stimuli output needs to be of a very high order. Zero

strain drift or strain biasing would be expected for accurate results to be reported.

Each testing procedure runs for a number of days continuously, after which another

test is performed within a short period of time (maximum of 24 hours) until three

replicates of the three differing protocols are performed. This means the DSD is

placed under continual use for more than a month at a time. Therefore, reliability is

important when considering alternative DSD designs. Weighting = 8/10.

7.1.1.8 Durability The DSD is required to withstand the high humidity (95% relative humidity), high

temperature (37°C) environment of the cell growth incubator while maintaining

stability in the mechanical and electrical stimuli during the testing procedures.

Therefore, the design issue of durability is of a moderate to high importance.

Weighting = 7/10.

7.1.1.9 Cost Due to financial limitations of the project, the cost of particular parts was a restrictive

factor when designing the DSD. This issue is significant and as a result a weighting

of nine was given. Weighting = 9/10.

7.1.1.10 Ease of PEMF Integration A fundamental aspect of the DSD is a seamless integration of the correct PEMF

electrical stimulus with the mechanical strain. It is an essential design criterion for

the DSD that stimulants do not interfere. Similar to the mechanical strain signal

84


flexibility design issue, this is essential and has been given the top weighting.

Weighting = 10/10.

7.1.2 Review of Current Technology

After a thorough investigation of mechanical stimulation of in vitro cultures of bone

cells, five different methods were considered and are described in the following

section. These were:

1. Flow of growth media fluid across cell layer inducing shear strain.

2. Pressure induced strain of a circular deformable membrane (with cells

attached).

3. Stretch of a circular deformable membrane (with cells attached) via actuation

with a piezoelectric disc.

4. Uniaxial or biaxial stretch of a deformable membrane (with cells attached).

5. Intermittent pressurisation of the gaseous environment in the cell growth

incubator.

Figure 7-1 visually represents these strain methods for comparison.

85


a)

Cell culture well

Flow of growth media fluid across cell layer inducing shear strain

b)

c)

d)

e)

Growth Media

Cell culture well

Stretch of membrane via central pin connected to a piezoelectric actuator

Growth Media

Stretch of cell substrate membrane via relativemovement between clamps

Cell culture well

Clamp

Gas pressure

Cell culture well

Pressure differential drives membrane ‘bulge’ and hence strain

Clamp

Cell culture well

Hydrostatic stress on cells due to intermittent pressurization of gaseous environment

Figure 7-1 Diagrams of previously reported in vitro cell straining devices. a) fluid flow shear strain b) pressure induced substrate strain c) piezoelectric actuated substrate strain d) direct mechanical straining of substrate e) hydrostatic gas pressurization of culture environment.

86


7.1.2.1 Flow of growth media fluid across cell layer inducing shear

strain As discussed in Chapter 3, a wide range of cellular phenomena are influenced by

fluid shear, such as the activation of plasma membrane receptors, ion channels,

integrins/focal adhesions and protein kinase signalling, which are all involved in the

mechano-reception process. Other cellular responses to fluid shear include increases

in calcium, nitric oxide and prostacyclin release and the remodelling of the internal

cytoskeleton.

Flow induced shear strain is mainly employed through a device called the parallel

plate flow chamber. A pressure differential is created between two slit (manifold)

openings at either end of a rectangular chamber, causing uniform laminar flow to

develop across the culture surface. Gravity heads (Li et al., 1996) along with active

pumps (Jacobs et al., 1998) have been used to create this pressure drop.

Special versions of this type of flow chamber have incorporated separate "settling

chambers" with curvilinearly tapered inlets to optimise flow field development in

pulsatile stimulus situations (Ruel et al., 1995) and rectangular obstacles in order to

create shear stress gradients for observation of cellular migration responses (Tardy et

al., 1997).

In flow-stimulus systems, it needs to be recognized that estimates of shear stress

based on calculated velocity gradients are only nominal. Local irregularities in the

surface topography of the culture layer itself can cause substantial strain

heterogeneity at the cellular level, as demonstrated by computational fluid dynamics

studies using atomic force microscopy (ATM) mapped culture surfaces (Davies et

al., 1995).

As parallel plate flow chambers introduce strain irregularities, then a study of the

exact effects of a rated shear strain on bone cells is impossible to quantify. It is a

concern that this method of mechanical stimulation does not replicate the in vivo

environment of bone cells where extracellular matrix deformation causes direct

mechanical stretching of cells. There is also a significant possibility that movement

87


of electrokinetic particles in the growth media will introduce an electric current,

similar to streaming potentials in bone, affecting the induced electric field in the cell

layer from the externally applied PEMF.

7.1.2.2 Pressure induced strain of a circular deformable membrane This method of mechanical stimulation is achieved by use of cell cultures grown on

peripherally tethered deformable circular substrates, which undergo deflections via

air pressure differentials located underneath the membrane. These pressure

differentials used to drive the substrate deformations are isolated from the cellular

gaseous environment by the membrane itself. The pressure causes ‘spherical’

distension of the membrane substrate either up or down depending on the pressure

gradient used.

An example of such a device was used by Winston et al., who applied positive

pressure underneath circular membranes cut from 100µm thick polyurethane urea

sheets. The sheets were clamped peripherally by O-Rings (Winston et al., 1989).

However, a fundamental limitation of this method of mechanical stimulation is strain

heterogeneity and the limited frequency response. For the case of a very thin

deformable membrane, the radial component of strain is nearly homogeneous

(Williams et al., 1992) although prestrain-dependent (Brodland et al., 1992).

However, circumferential strain is heterogeneous, varying from zero at the periphery

to a maximum at the centre.

One method designed to work around the heterogeneity problem has been to restrict

the culture to specific regions of the substrate, either by spot plating techniques or by

masking rings. However, another route with much more success has been to modify

the substrate deformation itself, imposing a kinematic constraint by means of a post

or platen placed in the centre of the circular membrane (Hung and Williams, 1994;

Schaffer et al., 1994). As the negative (vacuum) pressure starts drawing down the

membrane, the platen restricts the movement of a large centre portion while the

periphery induces stretch of the membrane over the platen. This stretch is

88


equibiaxial, meaning the radial and circumferential strain is equal and independent of

radial position. However, small frictional artefacts are introduced from the stretching

of the membrane over the platen surface and lubrication with vegetable oils is

required (Sotoudeh et al., 1998).

It was decided that the pressure-induced stretch with platen was to be considered for

the DSD design.

7.1.2.3 Stretch of a circular deformable membrane via a

piezoelectric actuator Only recently devised, this method of substrate deformation is actuated by means of

a piezoelectric crystal that mechanically deforms in response to an applied voltage.

The original device devised by Tanaka was capable of inducing low strain (200 –

40000µε), uniaxial stretch of a deformable membrane with arbitrary strain

waveforms. Operational strain frequencies spanned a range of a few tenths of a Hertz

up to several hundred Hertz (Tanaka, 1999). The displacements achieved by this

device were dependant upon specimen stiffness and thus there was no direct strain

control of the substrate. A laser displacement meter was necessary in order to

empirically adjust the power amplifier gain and achieve the desired displacement.

Also, a problem with this uniaxial stretch device was the spurious fluid shear strains

introduced by the movement of the membrane (as discussed in Section 3.1.3).

It was decided that this method of piezoelectric actuation could be combined with

direct stretching of the cell substrate surface similar in nature to the pressure induced

or uniaxial/biaxial substrate stretch. This is a more direct method of strain control

and would overcome most of the problems described in the previous paragraph. As

discussed in Chapter 11, the frequency capabilities of the Piezoelectric Actuator are

also very advantageous. Also, these devices potentially have lower frictional drag

due to smaller contact area. Therefore, a device utilising piezoelectric actuation of

the cell substrate producing stretch was to be considered as the third possible DSD

design.

89


7.1.2.4 Uniaxial or biaxial stretch of a deformable membrane This is simply the intermittent stretching of the cell growth substrate by way of a

variety of different methods. Strain is quantified from displacement of the membrane

and can be utilized in either uniaxial or biaxial situations. This method of mechanical

stimulation has advantages in areas such as duty cycle parameter control, input

quantitation, economy and ease of use. Uniaxial stretch introduces a compressive

strain due to Poisson’s Ratio. This was overcome by Norton et al., who obtained

isotropic substrate strain by pulling a membrane segment in two perpendicular

directions (Norton et al., 1995).

The method of mechanical actuation varies considerably; the most common method

includes a motor/cam-driven arrangement for cyclic tension (De Witt et al., 1984;

Neidlinger-Wilke et al., 1994), cyclical distension with stepper motors

(Vandenburgh, 1988; Murray and Rushton, 1990; Heinrich and Lunderstaedt, 2001;

Smith et al., 2001) and electromagnetic related drivers such as solenoids (Xu et al.,

1996; Decker et al., 1997; Smalt et al., 1997).

This mechanical stimulation approach can be very demanding in terms of hardware

implementation, and may introduce pre-strains with the gripping of the substrate.

Depending on the system used, strain heterogeneity associated with frictional effects

between the specimen and the underlying platen is also of concern.

7.1.2.5 Intermittent pressurisation of the gaseous environment in the

cell growth incubator These systems induce a compressive load by pressurization of the gaseous

environment the cells are grown in. Cells are grown on solid surfaces and the

pressurized gaseous incubator environment forces the cells to deform. Advantages of

this system include the simplicity of equipment, spatial homogeneity of the strain

stimulus, ease of multiple loading replicates (manifolding) and no dependence on the

state of adhesion between cell cultures and substrate for correct strains to be imparted

to the cell. The ability to create static and cyclical pressurization is possible. The

90


frequency of pressure cycles is limited to approximately 2 or 3 Hz depending upon

factors such as the volume of gas and the desired pressure.

Although high frequency and/or low pressure minimally affect the physical make up

of the medium (Tanck et al., 1999), pressurization can create high partial O2 and CO2

pressures in the liquid nutrient medium, which require compensatory treatment steps

(Ozawa et al., 1990).

Cell culture strains depend upon a complex fluid/structure interaction between the

substrate and its overlying nutrient medium. Thus, calibration is complicated by the

fact that several routinely varied operating parameters influence the culture surface

stimulus. These include the magnitude, frequency, and waveform of the driving

pressure signal, the mass (i.e., depth) and viscosity of the nutrient medium, and the

assembly pre-tension existing in the substrate itself.

7.1.3 Evaluation of Design Alternatives

The five DSD design options were evaluated. A design’s ability to satisfy each of the

ten issues outlined in Section 7.1.1 was scored out of five. Weightings for each issue

were multiplied by the score out of five to produce a numerical rating.

Each design issue discussed in Section 7.1.1 was given a weighting of importance out

of 10 with respect to the requirements of the dual stimulus device. The weightings

were not linearly spaced over the 10 design issues, but were scored relative to each

other to provide a more realistic emphasis on the important issues (e.g. both strain

flexibility and PEMF integration were given weightings of 10 as they are both

critical to the design of the device).

Weightings were derived from ranking the design issues that were most important

from 10 to 1 with 10 being critical and 1 being an issue of minor importance.

Between those values a ‘weighting’ judgment was made for each design issue. This

same method of weighting was then apportioned to each of the five available

designs, based on the design’s ability to accommodate each particular design

requirement.

91


Overall scores for each DSD design were summated from these numerical ratings in

each of the ten design issues (Table 7-1).

Eg: Fluid Flow DSD Design Overall score = [9 (accuracy weighting) X 2 (fluid

flow score for accuracy)] + [3 (cell line flexibility weighting) X 5 (fluid flow score

for cell line flexibility)] + ……+ [10 (PEMF integration weighting) X 2 (fluid flow

score for PEMF integration)] = 188

In conclusion, mechanically stretching the cell layer with a piezoelectric crystal is the

most viable design for the DSD. Hydrostatic gas pressure scored well as it was a

much simpler design that required no moving parts. Although both were high

scorers, it was the piezoelectric substrate stretching method and its high level of

strain waveform flexibility and expected consistency that was considered to be most

appropriate for this study.

92

Tab

le 7

-1 E

valu

atio

n of

In v

itro

mec

hani

cal s

trai

ning

tech

niqu

es fo

r du

al st

imul

us d

evic

e de

sign

Des

ign

Issu

es

Wei

ghtin

g F

luid

Flo

w S

hear

Stra

in

Air

Pre

ssur

e C

ell

Subs

trat

e St

retc

h

Piez

oele

ctri

c C

ell

Subs

trat

e St

retc

h

Uni

/Bia

xial

Cel

l

Subs

trat

e St

retc

h

Hyd

rost

atic

Gas

Pres

suri

zatio

n

Con

sist

ency

X

9

2

44

31

Cel

l lin

e fle

xibi

lity

X

3

5

33

35

Eas

e of

man

ufac

ture

X

4

43

22

4

Eas

e of

mai

nten

ance

X

7

4

33

24

Flui

d st

rain

X

6

-

44

25

Stra

in fl

exib

ility

X

10

1

35

42

Rel

iabi

lity

X

8

3

43

23

Dur

abili

ty

X7

4

33

25

Cos

t X

9

3

31

32

PEM

F In

tegr

atio

n X

10

2

44

35

Ove

rall

Scor

e

188

252

258

197

245

93


7.2 DSD Design

7.2.1 Original DSD Design

Originally, the selected DSD design did not include the ‘Central Pin Insert’ part

(Italicised names refer to component parts, See Section 7.2.3). Instead it had been

designed to actuate the strain by means of a circular ring platen, which pinched the

deformable cell substrate into a circular groove located adjacent to the cell culture

well stretching it on a level plane in an equibiaxial manner across the lower edge of

the walls of the cell culture well. This method of substrate stretch (without

piezoelectric actuation) has been previously used with success when employing large

strains (Hung et al., 1994; Schaffer et al., 1994).

However, a large frictional resistance between the cell substrate and the DSD Top

during dynamic stretching was observed. The novel nature of this device meant that

there were no previous reports dealing with this problem.

Additionally, inadequate assembly tolerances also contributed to the lack of

equibiaxial strains. Silicone adhesive glue was used for attaching the Piezoelectric

Actuator to the Base of the DSD plus the attachment of the circular ring platen

(named Indentor) to the piezoelectric actuator. This method of adherence did not

produce a horizontally level ‘pinching’ ring for the Deformable Cell Substrate

Membrane, even though guide pins specially produced for attaching the Indentor

were used. The non-level Indentor caused the cell substrate to be dragged in one

direction. Multiple attempts to fit varying sized and shaped spacer rings of

polycarbonate between the Base and the Top were tried with the intension of

overcoming this problem, however these attempts were not successful.

Therefore the DSD was redesigned to include the Central Pin Insert (Section

7.2.3.4). A drawback of this design was the non-equibiaxial strain produced,

however this method was the only alternative apart from major re-design and re-

manufacturing which was unfeasible.

94


7.2.2 Revised DSD Design

The successful dual stimulus device works by imparting a mechanical displacement

within the middle of the Deformable Cell Substrate Membrane by means of the

Central Pin Insert. The insert is placed in the Indentor, which is attached to the

Piezoelectric Actuator. This actuator forces the Indentor (and hence Central Pin

Insert) upwards into the Deformable Cell Substrate Membrane creating a central

displacement and resulting in a ‘tenting’ arrangement in the membrane (Diagram (c);

Figure 7-1)

7.2.3 Parts of DSD Design

External dimensions of the assembled device are 100mm X 100mm X 43.5mm

(Figure 7-2, Figure 7-3, Figure 7-4, Figure 7-5). The DSD, when set up for cellular

testing is made up of the following 11 parts (engineering drawings are given in

Appendix B). Each part is labelled in Figure 7-6 and Figure 7-7:

1. Base (Section 7.2.3.1)

2. Piezoelectric Actuator Disk (Section 7.2.3.2)

3. Indentor (Section 7.2.3.3)

4. Central Pin Insert (Section 7.2.3.4)

5. Spacer Ring (Section 7.2.3.5)

6. Cell Substrate Annulus (Section 7.2.3.6)

7. O-Ring (Section 7.2.3.7)

8. Deformable Cell Substrate Membrane (Section 7.2.3.8)

9. Top (Section 7.2.3.9)

10. PEMF Coil and Former (Section 7.2.3.10)

11. Lid (Section 7.2.3.11)

For clarity of detail some of the components are represented by diagrams and others

by photographs.

95


Figure 7-2 The assembled dual stimulus device (DSD) without PEMF Coil and Former or Lid.

Figure 7-3 A diagram of the assembled dual stimulus device (DSD) without Lid.

96


Figure 7-4 A diagram of the fully assembled dual stimulus device (DSD).

Figure 7-5 A diagram of the fully assembled dual stimulus device (DSD) with Lid cut away to reveal cell culture well.

97


3 6 4

Central pin used to ‘tent’

cell substrate 7

8

Figure 7-6 A diagrammatical cross section of the assembled DSD. Boxed area on top diagram is shown above with part numbers labelled.

98


Area 1 Area 2

2

Displacement

meter fibre

optic cable Piezoelectric

actuator wire

access

1

5

9

10

Figure 7-7 A diagrammatical exploded view of the cross-sectioned DSD, with part numbers shown. Area 1 and 2 are used for adhesive containment and are discussed in Section 7.2.3.1 and

7.2.3.3.

99


7.2.3.1 Base The Base component, manufactured from Perspex®, incorporates the Piezoelectric

Actuator in a manner that leaves the inside surface of the Base flush with the surface

of the actuator. The actuator’s outer edge is adhered to the Base with silicone

adhesive in an outer ring groove, which is required during activation to prevent

vibrational movement between the actuator and the base (Area 1 on Figure 7-7 and

Figure 7-8). The Base includes:

• One non threaded hole, ported through the side of the Base, for Piezoelectric

Actuator wires (Figure 7-8, 1)

• Two threaded holes; each with counter bores accessed from the underside, for

attachment of the displacement meter’s fibre optic cable. The centre hole is

for measuring movement of the actuator, while the offset hole is for direct

measurement of the Indentor. This feature allowed calibration to be

conducted during experimental cellular testing. (Figure 7-8, 2)

• Three additional holes (only one noted in figure), bored from the topside with

threaded screw access from the underside, were used for housing the guide

pins during adhesive attachment of the Indentor to the Piezoelectric Actuator

(Figure 7-8, 3).

100


Outer ring groove for

silicone adhesive (Area 1)

1

2

3

Figure 7-8 The Base from the dual stimulus device. See Section 7.2.3.1 for description of numbers.

7.2.3.2 Piezoelectric Actuator The Piezoelectric Actuator used in the DSD is a flat disc with a diameter of 63.5mm

and a height of 0.41mm (Figure 7-9) and is attached to the Base. It is a bending

transducer, where displacement occurs in the centre of the disk, which bows in and

out, similar to a drumhead, when a potential difference is applied across its two

faces. Specifications of the actuator are described in Section 8.2.

101


7mm

Figure 7-9 The Piezoelectric Actuator used in the dual stimulus device.

7.2.3.3 Indentor Originally, this was used for ‘pinching’ the Deformable Cell Substrate Membrane

into a groove in the Top, creating equibiaxial strain (discussed in Section 7.2.1).

However, as this was not achieved, a Central Pin Insert was used to actuate the

deformable membrane. As can be seen from Figure 7-10, guide pin holes and the

circular ring platen are present. The centre-bored section is also noted at the top of

the part and was used to house the Central Pin Insert. This part was manufactured

from Lexan®, a polycarbonate material, to obtain a high polish finish and reduce

frictional forces for the circular ring platen, as this was originally intended to create

the cell substrate membrane strain. This part was adhered to the Piezoelectric

Actuator with a silicone adhesive located at Area 2 on Figure 7-7.

102


Circular

ring

platen

Centre

Bored

Section

Guide pin

holes

Figure 7-10 Diagram of Indentor from dual stimulus device.

7.2.3.4 Central Pin Insert This part includes a 1mm radial tip point for impingement on the Deformable Cell

Substrate Membrane (Figure 7-11). This part was weight relieved so as to not affect

the maximum displacement output of the Piezoelectric Actuator. It was

manufactured from Perspex®.

103


Central

pin

Figure 7-11 Diagram of Central Pin Insert from dual stimulus device.

7.2.3.5 Spacer Ring This was designed to control the position of the Deformable Cell Substrate

Membrane (when assembled to the Top and Cell Substrate Annulus) to sit exactly

above the tip of the Central Pin Insert when the DSD was inactive (Figure 7-12).

Therefore, the membrane was subjected to the full movement of the Central Pin

Insert. These spacers were also used during calibration of cell substrate strain. These

were manufactured from Perspex®.

Figure 7-12 Diagram of Spacer Ring from dual stimulus device.

104


7.2.3.6 Cell Substrate Annulus Screwed onto the Top, the annulus clamps the Deformable Cell Substrate Membrane

between the Top and the O-Ring. Six M2.5 countersink screws were used for

fastening (Figure 7-13). This was manufactured from Perspex®.

Figure 7-13 Cell Substrate Annulus (clear) with attached O-Ring (black) from dual stimulus device.

7.2.3.7 O-Ring A nitrile O-Ring (Part No: RR0223, Ludowici Seals, Brisbane, QLD, Australia) was

used to seal the attachment between the Deformable Cell Substrate Membrane and

the Cell Substrate Annulus (Figure 7-13). This material was rated for wet, high

temperature working conditions and was therefore suited for the DSD.

7.2.3.8 Deformable Cell Substrate Membrane The Deformable Cell Substrate Membrane consisted of a thin polydimethylsiloxane

(PDMS) film approximately 50µm thick (Product No 7-4107, Dow Corning,

Midland, MI, USA). Films were cross-linked resulting in superior mechanical

characteristics. Membranes were transported on polymer backing sheets (Figure

7-14). The decision to use these substrates is discussed in Section 9.1.

105


Polymer

backing

sheet

Figure 7-14 Silicone material used as cell substrate for dual stimulus device. Shown is the Deformable Cell Substrate Membrane attached to a polymer backing sheet.

7.2.3.9 Top The Top (Figure 7-15) was used to enclose the device and partially seal it from the

highly humid and high temperate incubator environment. The Deformable Cell

Substrate Membrane is clamped onto the Top by the Cell Substrate Annulus and the

O-Ring (Figure 7-6). This arrangement facilitated easy removal of the deformable

membrane before and after cellular testing. The Top was designed with a circular lip

on the underside, which was used for centring the part into the Base, and had three

holes for placement of guide pins during assembly of the Piezoelectric Actuator and

Indentor. This part was also manufactured from Lexan®, and was used to obtain a

high polish finish on the bottom lip of the cell culture well walls. This was aimed at

reducing the frictional forces on the Deformable Cell Substrate Membrane during

activation of the DSD.

106


Figure 7-15 Top from dual stimulus device.

7.2.3.10 PEMF Coil and Former The former was manufactured from PVC, while the wire was an acrylic coated

copper (Figure 7-16). Two coils, each 50 turns of wire, were wrapped around the

former which was then push fit around the Top to a position where the centre line

between the two coils was at the level of the cell growth substrate.

Figure 7-16 PEMF Former from dual stimulus device. Shown without Copper Wire Coils.

107


7.2.3.11 Lid Finally, the Lid for the DSD (manufactured from Perspex®) was designed to sit on

the Top with a controlled 0.5mm clearance between the Lid and the DSD (Figure

7-17). This clearance was copied from the control cell culture plates to maintain

scientific validity of the experimental data. The Lid protects cultures from media

evaporation and contamination.

Figure 7-17 Diagram of Lid from dual stimulus device.

7.3 Conclusions

In conclusion, it was found that mechanically stretching the cell substrate layer with

a piezoelectric crystal and imparting a PEMF to the cells through the use of an

externally applied conducting coil was the most viable design for the dual stimulus

device. Mechanical strain was achieved by forcing a central pin into the centre of a

Deformable Cell Substrate Membrane, stretching the surface of the substrate where

the cells attach. A DSD was designed and manufactured capable of simultaneously

exposing in vitro cell cultures to both electrical and mechanical stimuli.

108

Chapter 8: Specification of active DSD Components

8 Specification of active DSD

Components

This chapter describes the three active components used in the dual stimulus device

and its calibration. These are the measurement transducer used to measure

displacements of the moving elements, the Piezoelectric Actuator and the PEMF

coils.

8.1 Displacement Meter and its Calibration

A reflective fibre optic displacement meter was used for measuring displacements

(Sensor Model No: D10UPFP, Optic Fibre Cable Model No: PBT16U, Banner

Engineering, Minneapolis, MN, USA). This device consisted of a long, flexible,

plastic fibre optic cable connected to a sensor/amplifier on one end and the DSD on

the other. Displacement was measured by means of reflected light intensity. The

device was capable of measuring a maximum displacement of 6mm.

Initially a known displacement range, greater than that to be measured, was

calibrated into the meter. This step defines the maximum (10V) and minimum (0V)

output voltage range for display on an oscilloscope during experimental

displacement testing. However, output voltages responded in a non-linear fashion at

the maximum and minimum extremes of measurement with the magnitude of the

non-linearity varying with every recalibration of the meter. Therefore, maximum

range of the output voltage for maximum displacement of the actuator varied from

6V to 10V.

109


8 Specification of active DSD

Components

This chapter describes the three active components used in the dual stimulus device

and its calibration. These are the measurement transducer used to measure

displacements of the moving elements, the Piezoelectric Actuator and the PEMF

coils.

8.1 Displacement Meter and its Calibration

A reflective fibre optic displacement meter was used for measuring displacements

(Sensor Model No: D10UPFP, Optic Fibre Cable Model No: PBT16U, Banner

Engineering, Minneapolis, MN, USA). This device consisted of a long, flexible,

plastic fibre optic cable connected to a sensor/amplifier on one end and the DSD on

the other. Displacement was measured by means of reflected light intensity. The

device was capable of measuring a maximum displacement of 6mm.

Initially a known displacement range, greater than that to be measured, was

calibrated into the meter. This step defines the maximum (10V) and minimum (0V)

output voltage range for display on an oscilloscope during experimental

displacement testing. However, output voltages responded in a non-linear fashion at

the maximum and minimum extremes of measurement with the magnitude of the

non-linearity varying with every recalibration of the meter. Therefore, maximum

range of the output voltage for maximum displacement of the actuator varied from

6V to 10V.

109


8.2 Piezoelectric Actuator

These devices have been utilised in many applications since their discovery in 1880

by Pierre and Jacques Curie. They have found widespread use in sensor and control

technologies such as accelerometers, medical ultrasound equipment and vibrational

suppression.

8.2.1 Operating Principle

The principle of piezoelectricity is based on the property of a material to become

electrically charged when subjected to a mechanical stress. Conversely, an applied

electric field will result in mechanical deformation of the piezoelectric material. A

majority of the piezoelectric materials are a type of ceramic that act as

polycrystalline dielectric materials.

The magnitude of the strain can be varied depending upon the magnitude of the

voltage used to drive the deformations in the crystal. The piezoelectric crystal can be

driven at extremely high frequencies and they do not suffer any significant

hysteresis, making them extremely useful as direct mechanical actuators for substrate

stretch.

The Piezoelectric Actuator used in the DSD was in the shape of a flat disc, with a

diameter of 63.5mm and a height of 0.41mm (Figure 7-9). It is a bending transducer,

where displacement occurs in the centre of the disk, which ‘spherically’ bows in and

out when actuated by the applied voltage. It was rated at a maximum driving voltage

of +/- 180V creating an unblocked displacement output of +/- 476µm (when not

glued to the DSD’s Base) however, after assembly into the DSD, the output

displacement of the actuator dropped to 72µm.

110


8.2.2 Static Calibration

Static calibration determined the output displacement and the required input voltage

for the Piezoelectric Actuator to drive the displacement (Figure 8-1).

y = 0.404xR2 = 0.998

0.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

80.00

0 20 40 60 80 100 120 140 160 180 200

Voltage (V)

Dis

plac

emen

t (m

icro

ns)

Unassembled Device Assembled Device

Linear Regression (Assembled)

Figure 8-1 Piezoelectric Actuator displacement output with applied voltage when unconstrained (but glued to the DSD Base; ‘Unassembled Device’) or constrained (when the DSD was fully

assembled; ‘Assembled Device’) with attachment of DSD Top. Maximum rated voltage was +/- 180V.

Displacement measurements of the actuator were taken with the applied piezoelectric

driving voltage to determine a calibration constant. Equation 8-1 below is defined

from the linear regression of the displacement output with Piezoelectric Actuator

voltage of the assembled DSD in Figure 8-1. This regression showed good fit with an

R2 value of 0.998.

Do = 0.404 Vi

Equation 8-1 Piezoelectric Actuator output displacement in micrometers with Piezoelectric Actuator input voltage; when DSD is fully assembled.

Output displacement of the actuator (in microns) is represented by Do and the driving

voltage (in volts) for the Piezoelectric Actuator is Vi.

111


An interesting phenomenon was observed during this testing. The actuator, when

glued to the Base without attached Indentor and Central Pin Insert, did not travel as

far (approximately 50µm) as the fully assembled DSD with the resistive deformable

membrane clamped above the central pin (72µm, Figure 8-1). As discussed in

Section 8.2.4, the Piezoelectric Actuator exhibits a greater displacement output when

there is a preload present.

8.2.3 Dynamic Calibration

All measurements were made within the 0 – 30Hz range and thus dynamic

calibration relationships are only representative for this frequency range.

Dynamic calibration results highlighted the inverse fashion in which increasing

frequency decreases the output displacement of the Piezoelectric Actuator (Equation

8-2). This property was calibrated from the polymeric regression of the assembled

device results in Figure 8-2. This regression showed good fit with an R2 value of

0.995.

Vo = 0.0021 fi 2 – 0.1887 fi +7.94

Equation 8-2 Output voltage range of the displacement meter with varying driving input voltage frequency; when DSD is fully assembled.

Where Vo is displacement meter’s voltage output range and fi is the frequency of the

input voltage signal measured in Hertz.

112


y = 0.0021x2 - 0.1887x + 7.94R2 = 0.9947

0

1

2

3

4

5

6

7

8

0 5 10 15 20 25 30 35

Frequency (Hz)

Out

put V

olta

ge R

ange

(V)

Assembled Device

Polynomial Regression

Figure 8-2 Calibration displacement meter output voltage range with increasing actuator driving voltage frequency

The displacement meter’s maximum output voltage range during static calibration

was measured to be 7.72V, which was achieved when the Piezoelectric Actuator

driving input voltage was at a maximum of 180V. As described by Equation 8-1 this

created a maximum central pin output displacement of 72µm. Therefore the

relationship between the displacement meter’s maximum output voltage range and

the maximum central pin output displacement is described by the ratio of 1V

(displacement meter output voltage range) = 9.33µm (central pin output

displacement).

Therefore when substituting this ratio into Equation 8-2, the actuators output

displacement with input voltage frequency is described by Equation 8-3:

Do = 0.0196fi 2 – 1.7606 fi +72.08

Equation 8-3 Piezoelectric Actuator output displacement in micrometers (when a maximum of 180V is applied) with input voltage frequency; DSD is fully assembled.

113


8.2.4 Blocking Force

To investigate the interesting phenomenon observed during testing that the

Piezoelectric Actuator when glued to the Base but in an unassembled state, did not

travel as far as the fully assembled DSD, static weights were placed on the

Piezoelectric Actuator during dynamic actuation to measure output displacement.

This was plotted against force and the resulting graph is depicted in Figure 8-3. A

blocking force of approximately 1N helped the actuator to achieve the maximum

output of 72µm.

0

10

20

30

40

50

60

70

80

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Blocking Force (N)

Dis

plac

emen

t (m

icro

ns)

Figure 8-3 The piezoelectric actuator output displacement with increasing blocking force.

As displacement of the actuator in the assembled DSD and that from the preloaded

testing model matched, it is estimated that the preload of the assembled DSD is also

approximately 1N.

One of the main reasons a preload affects the actuator in this manner is believed to

be due to the arrangement in which the actuator was adhered to the DSD. A preload

may have compressed the piezoelectric disc onto the Base, stabilising its inferior

outer ring (Area 1 on Figure 7-7), which had been used to glue it to the Base. This

114


may have given the best possible leverage for actuation while also realigning its

central axis, maximising vertical displacement.

8.2.5 Conclusions

The bending Piezoelectric Actuator disc, which impinges on the Deformable Cell

Substrate Membrane via the Indentor and Central Pin Insert creating a surface strain,

has been calibrated for use in the DSD. The main working characteristics of the

actuator have been defined statically and over a dynamic frequency range of 0-30Hz

with good fit between experimental results and regression relationships and have

been defined as:

• Increasing the driving voltage to a maximum of 180V increases the actuator’s

displacement output in a linear fashion described by Equation 8-1

(R2=0.998).

• Increasing the driving voltage frequency, over the 0 - 30Hz frequency range

studied, decreases the actuator’s displacement output in a quadratic fashion

described by Equation 8-2 (R2=0.995).

• A preload of 1N causes a maximum actuator output displacement of 72µm,

which was the assumed load imparted on the actuator during assembly of the

DSD.

8.3 PEMF Coil

Measurements of the magnetic field strength and the induced EMF from the pulsed

electromagnetic field coil were taken. Magnetic field strength is stated as a gauss

measurement where 1 Gauss = 10-4 Tesla. All settings for the PEMF device were the

same as described in Chapter 5, where DC power was set at 20V.

8.3.1 Measurement devices used in calibration of coil

These are discussed in Section 5.3.1 and are summarised as:

• A DC power supply for powering the PEMF signal generator.

115


• A real time, two-channel oscilloscope used to measure and record all

magnetic and induced electric field data obtained from the PEMF coils.

• A Gauss/Tesla meter to measure the magnetic field from the PEMF coils

• A Coil Dosimeter, also known as a coil probe, used to estimate the induced

electric field in the bone cell culture layer

8.3.2 Magnetic field strength from PEMF coil

Magnetic field strength from the PEMF coil was mapped at 10mm intervals from the

level of the cell substrate membrane vertically upwards. Four measurements were

made, quantifying the characteristics of the magnetic field with distance. Figure 8-4

displays the measurements graphically.

0

5

10

15

20

25

30

0 0.5 1 1.5 2 2.5 3 3.5

Distance (cm)

Mag

Fie

ld S

tren

gth

(G)

4

Figure 8-4 The magnetic field strength of the PEMF coil used in the DSD with an increasing vertical distance from the cell substrate.

116


8.3.3 Induced EMF

8.3.3.1 Vertical Spectrum The induced EMF follows the same pattern as the magnetic field measurements,

where increasing distance from the coil reduced both the magnetic field and induced

EMF measurements (Figure 8-5). The maximum peak-to-peak induced EMF was

138mV at the level of the deformable cell substrate. Decreasing EMF measurements

due to increasing distance from the PEMF coil was indicated by a reduction in the

magnitude of the negative (not positive) voltage spike (see Figure 5-15 for

explanation of voltage spikes). This continued until approximately 5cm from the

level of the cell substrate membrane at which stage both the positive and the negative

voltage spikes decreased in size.

0

20

40

60

80

100

120

140

160

0 1 2 3 4 5 6 7 8 9 10

Distance (cm)

Volta

ge R

ange

(mill

iVol

ts)

Figure 8-5 Maximum peak-to-peak voltage range of induced electric field in coil probe dosimeter with increasing vertical distance from the cell substrate surface.

8.3.3.2 Horizontal Spectrum Although there was only one cell culture well within the DSD, mapping the

horizontal spectrum of the induced EMF within the PEMF coil provided an

indication as to any differences between the centre of the cell culture well and the

117


edge. Figure 8-6 shows there is effectively no difference between the centre (0cm)

and the edge of the well (1cm) validating the assumption that the entire cell culture in

the well is experiencing the same magnetic field.

0

50

100

150

200

250

300

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

Distance (cm)

Volta

ge R

ange

(mill

iVol

ts)

Figure 8-6 Maximum peak-to-peak voltage range of induced electric field in coil probe dosimeter with increasing horizontal distance from the centre of the PEMF Coil in the DSD.

8.4 Conclusions

Calibration of the Piezoelectric Actuator determined its maximum output

displacement and hence the maximum Central Pin Insert displacement impinging on

the deformable cell substrate to be 72µm. This output decreases with an increasing

Piezoelectric Actuator frequency in a non-linear, quadratic fashion (Equation 8-3).

Output frequency was calibrated over the 0 – 30 Hz range, although higher

frequencies are also capable of being produced.

The pulsed electromagnetic field used in the DSD was produced from a custom built

PEMF coil. This coil was designed and calibrated with the same characteristics as

118


discussed in Chapter 5, and is capable of producing a 15Hz pulse burst

electromagnetic field, that induces an electric field within the cell cultures.

119

Chapter 9: Cell Substrate Characterisation and Treatment

9 Cell Substrate Characterisation and

Treatment

Cell attachment to the deformable cell substrate is of critical importance to the

success of the dual stimulus device. Adherent cell lines such as the osteoblast-like

MG-63 bone cells used in experimental testing with the DSD (discussed in Chapter

11) attach themselves to the cell substrate with proteins that exhibit preferential

affinity to particular attachment points on the surface. Surface chemistry forces

largely control this process while topographical surface characteristics are less

influential (Britland et al., 1996) and only seem to show significance when

topographical cues are of a large scale (Brunette, 1986).

A rapid adsorption of proteins from serum in the culture medium is the first stage

during cellular attachment. Cellular attachment to the substrate is essential for

survival (Ruoslahti and Reed, 1994) and the resulting cell morphology (which is

determined by the type and location of the proteins adsorbed). This will influence the

cells’ growth and phenotypic behaviour (Ben-Ze'ev et al., 1980; Stein et al., 1990).

The make up of the adsorbed protein layer is dependant upon the concentrations and

properties of the proteins in the culture medium and the physicochemical properties

of the deformable cell substrate (Anselme, 2000). It is these physicochemical

properties that can be modified via a number of ‘surface only’ techniques in order to

increase the ability of the cells to adhere, but at the same time do not change the

underlying bulk properties of the material.

Hydrophilicity is the characteristic of a material exhibiting an affinity for water, and

is a vital quality required for the healthy attachment of cells to the surface. The

surface chemistry allows hydrophilic materials to be wetted forming a water film or

coating. These materials possess a high surface energy value and have the ability to

form hydrogen bonds with water. Through the use of a surface etching technique

discussed in Section 9.3, the cell substrate maybe rendered hydrophilic.

120


9.1 Choice of material and its Physical Characteristics

Literature suggests that a polymeric, cross-linked, silicone elastomer membrane

presents characteristics essential for mechanical stability and enhanced cellular

attachment when surface treated (Lateef et al., 2002; Moretti et al., 2004). This

material has been used previously in other mechanically strained cell culture

applications (Schaffer et al., 1994; Lee et al., 1996; Sotoudeh et al., 1998;

Jagodzinski et al., 2004). As this material was easily assessable and was available in

thin preformed sheets capable of assembly in the DSD it was chosen to be the cell

substrate material (Product No. 7-4107, Dow Corning, Midland, MI, USA).

Named polydimethylsiloxane (PDMS) the silicone elastomer has a unique flexibility

resulting in one of the lowest glass-transition temperatures of any polymer.

Furthermore, it shows a low elasticity change versus temperature, a high thermal

stability, chemical inertness, optical clarity, shear stability and high compressibility.

Because of its high flexibility and the very low drift of its properties with time and

temperature, PDMS is well suited for mechanically strained applications, such as the

DSD.

9.1.1 Tensile tests

Tensile tests were performed to determine the engineering properties of the material.

These results were used during theoretical determination of the surface strain once

undergoing active stretch. Tensile test specimens were prepared at three different

angles (0°, 45° and 90°) to the longitudinal axis of the preformed PDMS membrane

sheet. Three specimens at each angle were tested to 100% elongation and then to

failure. Also, three separate specimens at 0° were tested to examine losses due to

hysteresis during 1mm extension following a preload of 1N. These tests were

performed with four different strain rates (50, 100, 300 and 500mm/sec).

All tests were performed with tensile test specimens sized according to Australian

and New Zealand Standards (AS 1145-3).

121


Results concluded that the material was anisotropic (Figure 9-1). Regression lines

used to determine Young’s modulus (as derived from the slope of the stress-strain

graph) were robust with R2 values of 1. Results for the moduli varied between

2.3MPa and 3.1MPa depending on the material direction. Test specimens broke at

maximum strains of approximately 270%, 350% and 530% for the 0°, 90° and 45°

material directions respectively (Figure 9-2).

As strain rates increased, hysteresis decreased, signifying a reduction in energy

storage within the cell substrate membrane. This results in repeatable elastic

deformations and hence more repeatable surface strains. Therefore, the hysteresis

tests indicated virtually no energy loss would occur at the strain rate the membrane

will be subjected to in the DSD.

y = 3119000.96xR2 = 1.00

y = 2382691.98xR2 = 1.00

y = 2878561.06xR2 = 1.00

0.000

500000.000

1000000.000

1500000.000

2000000.000

2500000.000

3000000.000

3500000.000

0.000 0.200 0.400 0.600 0.800 1.000 1.200

Strain

Stre

ss (P

a)

0 Degrees 45 Degrees 90 Degrees

0 Degrees

45 Degrees

90 Degrees

Figure 9-1 The typical stress vs strain normalised to 100% elongation for materials cut in three different directions. Regression lines are shown with linear equation and R2 values noted for

each specimen angle. Slopes of the equations were used to determine the material’s elastic modulus.

122


0

1

2

3

4

5

6

0 50 100 150 200 250 300 350 400

Extension (mm)

Forc

e (N

)

0 Degrees 45 Degrees 90 Degrees

Figure 9-2 Typical tensile force vs elongation of PDMS silicone cell substrate membrane material when taken to breaking point. Specimens were cut from three different angles in the

membrane sheet.

9.1.2 Thickness tests

Tests of material thickness were performed to confirm the value of 75µm supplied

with the cell substrate membrane. Digital venia callipers (Mitutoyo Digimatic

Vernier Callipers, Accuracy of +/- 0.0025-mm) were used to measure the thickness

of the membrane. However, these results were questionable, as this method may have

imparted a deformable force onto the membrane reducing the measurement values.

To confirm the results from the vernier callipers, thickness was measured with a

surface profilometer. To measure the thickness with a profilometer a small cone

shaped stylus, usually used to record surface profiles along a straight line on the

surface, was drawn over the edge of the membrane with the resulting drop in profile

representing the membrane thickness. Measurements for both methods were repeated

a number of times in different locations of the sample membrane.

123


Both tests confirmed that the membrane was thinner than the 75µm supplied value

stated by the manufacturer, with an average thickness of 50µm. However, large

variability (~20%) in thickness measurements were seen, with values of between

55µm and 45µm recorded.

9.2 Surface Characteristics

As discussed above, the physicochemical nature of the cell substrate surface is vitally

important during cellular attachment and has a greater influence than surface

topography. To ascertain the surface structure, two analytical methods were

employed. These were X-ray Photoelectron Spectrometry (XPS) and Attenuated

Total Reflectance (ATR). XPS analyses were conducted using a standard procedure

in the Future Materials research centre of the Faculty of Biological and Chemical

Sciences at the University of Queensland. ATR data were obtained using standard

procedures in the School of Physical and Chemical Sciences at QUT.

9.2.1 X-ray Photoelectron Spectroscopy (XPS)

XPS is based on the photoelectric effect. A light photon is fired onto the surface,

which results in an ejection of electrons from the surface atoms. The XPS technique

is highly surface specific due to the short range of the photoelectrons that are excited

from the solid. The energy of the photoelectrons leaving the sample is determined by

producing a spectrum with a series of photoelectron peaks. The binding energy of the

peaks is characteristic of each chemical element making up the surface. The peak

areas can be used (with appropriate sensitivity factors) to determine the composition

of the materials surface. The shape of each peak and the binding energy can be

slightly altered by the chemical state of the emitting atom. Hence XPS can provide

chemical bonding information as well. XPS is not sensitive to hydrogen or helium,

but can detect all other elements. Due to the delicate sensitivity of the photoelectrons

XPS must be carried out in an ultra high vacuum of approximately 10-9 millibar.

124


One small square sample with an approximate side length of 10mm was prepared

with care to not touch the surface of the sample. An initial survey scan of intensity

(CPS) for a range of binding energies of zero to 1200eV was made of the surface. A

more detailed scan of each peak was then undertaken to confirm the presence of each

particular chemical constituent. Figure 9-3 shows the initial survey scan with

significant peaks at binding energies of approximately 23, 100, 150, 282, 530 and

970eV. These correspond to oxygen 2s (O 2s), silicone 2p (Si 2p), silicone 2s (Si 2s),

carbon 1s (C 1s), oxygen 1s (O 1s) and the oxygen Auger electron emission

respectively (Cardona and Ley, 1978). Suffixes correspond to the outer electron

valence shells of each atom. The peaks of noteworthiness in this case are the C 1s, O

1s and the Si 2p as these are at the level of structural binding with the other elements

in the PDMS surface structure. Results of atomic concentration of each element on

the surface of the material are shown in Table 9-1. Mass concentration was computed

from the atomic mass of each element and its level of concentration.

Table 9-1 X-ray Photoelectron Spectroscopy (XPS) results from native PDMS membrane, where columns 2 to 8 represent the following: X axis position (binding energy); Full width at half -

maximum (FWHM); Raw area underneath peak (CPS); Relative sensitivity factor (RSF) – Used in calculating atomic concentration; Atomic mass of element; Atomic Concentration (%); Mass

Concentration (%).

Peak Position

(eV)

FWHM

(eV)

Area

(CPS)

RSF Atom.

Mass

Atom.

Conc

Mass

Conc

C 1s 282 2.664 160395 0.278 12.011 51.16 37.03

O 1s 530 2.817 234120 0.780 15.999 27.01 26.04

Si 2p 100 2.828 80615 0.328 28.086 21.82 36.93

125


0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

1200

1165

1130

1095

1060

1025 99

0

955

920

885

850

815

780

745

710

675

640

605

570

535

500

465

430

395

360

325

290

255

220

185

150

115 80 45 10

Binding Energy (eV)

Inte

nsity

(CPS

)

Figure 9-3 An XPS survey scan of native PDMS cell substrate material with electron volt binding energy peaks at 23(Oxygen 2s), 100(Silicon 2p), 150(Silicon 2s), 282(Carbon 1s) and

530(Oxygen 1s).

This atomic concentration data confirms the element ratio of carbon to oxygen to

silicone is 2 : 1 : 1. This confirms the material contains the repeating PDMS

monomer structure of (C2H6OSi)n.

9.2.2 Attenuated Total Reflectance (ATR)

As a molecule sits on a surface, it will vibrate. Such vibrations can be studied by the

internal reflection of light from an infrared beam focused on the surface. If the

molecule has a dipole moment, that is one end of the molecule has a positive charge

and the other end a negative charge, then the molecule can absorb infrared light, but

only at certain fixed frequencies. An infrared spectrum of light reflected from the

surface will show absorption peaks, which are characteristic of the molecule and its

method of bonding to the surface. Hence, this method of surface analysis is very

powerful for determining the structure of the surface species as opposed to the

elemental constituents. This surface technique was used to confirm results of PDMS

surface chemistry from the XPS analysis.

126


0.00E+00

2.00E-01

4.00E-01

6.00E-01

8.00E-01

1.00E+00

1.20E+00

1.40E+00

5001000150020002500300035004000

Wavenumber (cm-1)

Abs

orba

nce

Figure 9-4 An Attenuated Total Reflectance (ATR) spectrum scan of native PDMS cell substrate material with wave number (cm-1) peaks of interest at 3700-3000(Hydroxyl groups),

1000(Silicon-Oxygen stretching), 1260(Si-CH3 stretch) and 780(Si-CH3 stretch).

Figure 9-4 shows a sharp absorption band about 1000cm-1, which is due to the Si-O

stretch. Si-CH3 stretching bands are near 1260cm-1 and 780cm-1. This spectrum

contains very little hydroxyl or amine functional groups as demonstrated by the flat

response over the approximate 3100 – 3500 and 1500 – 1700 wavenumber bands

respectively (Urban, 1996) confirming published descriptions of PDMS’s

hydrophobic aversion to water (Cifkova et al., 1990). These two groups are

important factors determining the material’s affinity for water.

9.3 Surface Treatment

As was determined from the ATR data above, and as is widely recognised, the native

PDMS cell substrate membrane does not exhibit ideal surface qualities for the

biological attachment of normally functioning cells due to its non-polar side chains,

which contain carboxyl (COOH) groups. Protein adsorption to a hydrophobic

127


surface, such as native PDMS, markedly inhibits the proteins’ functionality as a

mediator of cell attachment (Stephansson et al., 2002) by tightly tethering it to the

surface and not allowing its reorganisation by the cells (Norde and Giacomelli,

2000). An increase in the surface hydrophilicity permits the cell to more easily

organise its protein attachment to the cell substrate surface.

The use of a technique named ‘gas plasma treatment’ (high energy gas) has sufficient

energy to break the Si-CH3 bond in the monomers on the surface. This allows

reactive group addition and cross-linking to occur which results in hydrophilic

properties. In particular, PDMS exposure to plasma leads to oxidation and the

formation of a silica-like surface (Weikart and Yasuda, 2000). Modification by

plasma treatment is usually confined to the top several hundred angstroms and does

not affect the bulk properties.

‘Gas plasma’ is an ionised medium consisting of charged (electrons and ions) and

neutral groups. It is produced from the excitation of a reactive gas (such as argon,

oxygen, nitrogen, fluorine, carbon dioxide, and water) with a low power, high

frequency electrical field. There are equilibrium (thermal) and non-equilibrium (non-

thermal) plasmas, however the low-pressure plasmas used in this research were of

the non-equilibrium, non-thermal type. For a more rigorous description of these types

of plasmas the reader is referred to other texts (Strobel et al., 1994; Chan et al.,

1996).

9.3.1 Description of Plasma Set-up and Procedures

The gas plasma machine consists of the following parts (Figure 9-5 and Figure 9-6):

• Reactive gas supply

• Vacuum chamber for placement of sample during treatment

• High voltage power supply

• Radio frequency signal generator and coil

• Reflected power meter

• Rotary and diffusion pump for creating vacuum

• Vacuum meter

128


Power

supply

RF signal

generator

Vacuum

chamber

with

sample

Reactive

gas

supply

(water)

Gas

reacting

with

electric

field to

create

plasma

Figure 9-5 Gas Plasma Machine showing glow from reactive gas.

129

Re

fle

cte

d p

ow

er

me

ter

V

ac

uu

m m

ete

r

Ro

tar

y

pu

mp

Figure 9-6 Gas Plasm

a Machine show

ing full length of vacuum cham

ber and associated parts.

130


The procedure for sample preparation and treatment is described in Appendix C.

However, the procedure followed this general outline:

1. Articles to be treated were placed within the vacuum chamber.

2. A fixed degree of vacuum was reached; at which point exposure time and

power were set.

3. The reactive gas was passed through the high frequency electrical field

creating charged and neutral species, which reacted with the surface of the

PDMS modifying its surface properties.

A detailed study on the effect power, vacuum and time levels have on surface

wettability revealed that vacuum level is of a lower importance than the power

utilised during the procedure (Weikart and Yasuda, 2000). The authors concluded

that a “lower pressure (vacuum) and higher input power were the best conditions

under which to maximise wettability” and that H2O plasma is superior to O2 plasma

as it “consistently lessened hydrophobic recovery”, the process by which the surface

returns to its original native state. Others have also noted the advantages of H2O

plasma, which does not result in the flaking off of the silicate layer formed when

using O2 plasma (Lee et al., 1991; Lateef et al., 2002).

Therefore it was decided that treatment was to be undertaken with H2O plasma.

Initial values for vacuum, power and length of exposure time were 0.8 Torr, 20W

and 4 minutes respectively. These were simulated from a previous study looking at

cellular adhesion on plasma treated membranes of silicone elastomer undergoing

flexion (Lateef et al., 2002). As mentioned above, power makes a significant

contribution to the resultant surface hydrophilicity. Therefore, an analysis was

performed of the influence gas plasma power level has on the surface chemistry. It

was hypothesised that an increasing level of power to 60 Watts would result in an

increasing level of surface hydrophilicity. Therefore tests were undertaken with

different plasma powers up to a maximum of 60W. These results are discussed in

Section 9.4.1.

131


9.4 Surface Characteristics Post-Treatment

To determine the changes plasma modification had made to the surface of the

silicone elastomer, the previously mentioned surface analysis techniques of XPS and

ATR were used. The results were compared to the native elastomer and the

difference was quantified. Contact angle measurements were made to determine the

wettability of the PDMS surface.

9.4.1 X-ray Photoelectron Spectroscopy (XPS)

To determine the optimum power for plasma treatment, six samples, each treated

with different H2O plasma power, were evaluated with XPS. The results for each

setting of power and the native PDMS are provided in Table 9-2:

132


Table 9-2 XPS results from gas plasma treated PDMS membranes at differing powers for determination of ideal plasma process. Columns are as defined in Table 9-1.

Peak Position

(eV)

FWHM

(eV)

Area

(CPS)

RSF Atom.

Mass

Atom.

Conc

Mass

Conc

C 1s 282 2.664 160395 0.278 12.011 51.16 37.03

O 1s 530 2.817 234120 0.780 15.999 27.01 26.04

Nat

ive

Si 2p 100 2.828 80615 0.328 28.086 21.82 36.93

C 1s 282 2.718 95795 0.278 12.011 25.08 16.98

O 1s 530 2.929 551695 0.780 15.999 52.25 47.12

5 W

atts

Si 2p 101 3.474 102050 0.328 28.086 22.67 35.90

C 1s 282 2.751 80615 0.278 12.011 20.57 13.71

O 1s 530 2.926 606885 0.780 15.999 56.00 49.74

10 W

atts

Si 2p 101 3.244 108250 0.328 28.086 23.44 36.54

C 1s 282 2.782 84540 0.278 12.011 21.25 14.26

O 1s 530 2.921 615285 0.780 15.999 55.95 49.99

15 W

atts

Si 2p 101 3.357 106845 0.328 28.086 22.80 35.75

C 1s 282 2.787 76015 0.278 12.011 19.09 12.76

O 1s 530 2.929 641145 0.780 15.999 58.25 51.85

20 W

atts

Si 2p 101 3.251 106280 0.328 28.086 22.65 35.40

C 1s 282 2.875 81840 0.278 12.011 21.94 14.83

O 1s 530 2.991 579515 0.780 15.999 56.18 50.59

40 W

atts

Si 2p 101 3.365 96210 0.328 28.086 21.88 34.59

C 1s 282 2.798 75060 0.278 12.011 19.79 13.28

O 1s 530 2.911 607885 0.780 15.999 57.97 51.82

60 W

atts

Si 2p 101 3.328 99400 0.328 28.086 22.24 34.90

133


The measure of atomic concentration determines the elemental constituents of the

surface. Original concentrations of carbon, oxygen and silicone on the native PDMS

material confirmed the (C2H6OSi)n PDMS monomer repeating structure with the

correct ratio of 2 : 1 : 1 respectively. Five watt plasma treatment and each subsequent

plasma power changed this ratio to 1 : 2 : 1. Therefore the hypothesis that an

increasing level of power will result in an increasing level of surface hydrophilicity is

incorrect as there is saturation at 5 watts, after which an increasing level of gas

plasma power does not result in a greater surface affinity for water.

The results suggest that oxidisation had occurred where one of the outer carboxyl

groups was stripped away, allowing group addition of oxygen and hydrogen to form

a hydroxyl group and hence a greater level of hydrophilicity. As changes in power

did not significantly affect this ratio, it was safe to assume the five-watt power rating

was sufficient for all subsequent plasma treatment of PDMS.

A detailed scan of each constituent was taken for the five-watt plasma treatment for

comparison to the native PDMS surface. Figure 9-7 compares survey scans from the

five-watt plasma treated and native membranes while Figure 9-8, Figure 9-9 and

Figure 9-10 show the detailed carbon, oxygen and silicone region comparisons

respectively.

134


0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

1200

1166

1132

1098

1064

1030 996

962

928

894

860

826

792

758

724

690

656

622

588

554

520

486

452

418

384

350

316

282

248

214

180

146

112 78 44 10

Binding Energy (eV)

Inte

nsity

(CPS

)

Native Plasma Treated

Figure 9-7 An XPS survey scan of native and 5 watt plasma treated PDMS cell substrate material. Variation in concentration of material constituents between the two is seen in the

differing electron volt binding energy peak heights.

0

1000

2000

3000

4000

5000

6000

292

291

290

289

288

287

286

285

284

283

282

281

280

Binding Energy (eV)

Inte

nsity

(CPS

)

Plasma Treated Native

Figure 9-8 An XPS detailed scan of carbon 1s for native and 5-watt plasma treated PDMS cell substrate material. The decrease in peak height for the treated substrate signifies a loss of

carbon 1s.

135


0

2000

4000

6000

8000

10000

12000

14000

536

535

534

533

532

531

530

529

528

527

526

525

524

Binding Energy (eV)

Inte

nsity

(CPS

)


Figure 9-9 An XPS detailed scan of oxygen for native and 5-watt plasma treated PDMS cell substrate material. The increase in peak height for the treated substrate signifies a gain in

oxygen while the shift in binding energy is due to a change in binding properties.

0

200

400

600

800

1000

1200

104.

5

103.

5

102.

5

101.

5

100.

5

99.5

98.5

97.5

96.5

95.5

94.5

Binding Energy (eV)

Inte

nsity

(eV)


Figure 9-10 An XPS detailed scan of silicone 2p for native and 5-watt plasma treated PDMS cell substrate material. The decrease in peak height for the treated substrate signifies a loss of

Silicon 2p while the second binding energy peak is due to an additional bind between silicon and oxygen (Si-O).

136


Figure 9-8 and Figure 9-9 visually represent the loss of carbon and gain in oxygen.

The shift in the oxygen binding energy was due to the change in bonds between Si –

O – Si to the Si – OH bond of a hydroxyl group. When curve fitting was introduced

to the plasma treated silicone graph (Figure 9-10), it was shown that the plasma had

introduced two new binding curves, one at 99.906eV making up an atomic

concentration of 11.32% and the largest of the three at 100.974eV making up 65.32%

of the silicone binding. The more significant of the two corresponds to silicone oxide

(SiO). This signified that surface hydroxyl groups were present.

9.4.2 Attenuated Total Reflectance (ATR)

To confirm the presence of surface hydroxyl groups found with the XPS data, the

five-watt plasma treated PDMS was analysed with ATR (Figure 9-11). The area

underneath the ATR curve between wavenumber 3000 and 3700 signifies the

presence of oxygen-hydrogen bonding as would be associated with hydroxyl groups.

Figure 9-12 shows the treated surface contains a higher degree of the OH functional

groups than the native PDMS. This measure is discussed with more quantitative

detail in Section 9.5.

0.00E+00

2.00E-01

4.00E-01

6.00E-01

8.00E-01

1.00E+00

1.20E+00

550105015502050255030503550

Wavenumber (cm-1)

Abs

orba

nce

1.40E+00

Native Treated

Figure 9-11 An ATR spectrum scan for native and 5-watt plasma treated PDMS cell substrate material. Little to no difference can be seen between the two traces at wave number (cm-1) peak 3700-3000(Hydroxyl groups), 1000(Silicon-Oxygen stretching), 1260(Si-CH3 stretch) and 780(Si-

CH3 stretch).

137


0.00E+00

5.00E-03

1.00E-02

1.50E-02

2.00E-02

2.50E-02

30003100320033003400350036003700

Wavenumber (cm-1)

Abs

orba

nce

Native Treated

Figure 9-12 An ATR detailed spectrum scan of the hydroxyl region (3700-3000cm-1) for native and 5-watt plasma treated PDMS cell substrate material.

The area underneath the ATR curve between the 1500 – 1700 wavenumber limits

measures the presence of functional amine groups on the surface. The presence of

amines (nitrogen containing organic compound) would signify the PDMS surface

contains amino acids and hence proteins. It can be seen from Figure 9-13 that there

was no difference between the treated and the native PDMS highlighting that only

hydroxyl groups were imparted onto the surface during the plasma treatment.

138


0.00E+00

2.00E-03

4.00E-03

6.00E-03

8.00E-03

1.00E-02

1.20E-02

1.40E-02

1.60E-02

15001520154015601580160016201640166016801700

Wavenumber (cm-1)

Abs

orba

nce

Native Treated

Figure 9-13 An ATR detailed spectrum scan of the amine group region (1700-1500cm-1) for native and 5-watt plasma treated PDMS cell substrate material.

9.4.3 Contact Angle

Contact angle measurements were made to quantify the hydrophilicity of the surface

after plasma treatment. The experiments were conducted using a standard procedure

in the School of Physical and Chemical Sciences at QUT. The apparatus for the

contact angle measurements consisted of a lit sample stage, a mirror inclined at 45°,

a microscope and a digital camera. Samples were placed on the stage and using a

microlitre syringe, 5µL of deionised water was delivered onto the surface of the

PDMS sample. A photograph was recorded immediately. A further 5µL of water was

added and another photograph taken. This was repeated once more to provide three

images.

Contact angle (θ) between the sample surface and the outer edge of the water drop

was obtained from Equation 9-1.

θ = 2*[Tan-1(2h/d)]

Equation 9-1 Water contact angle between sample surface and water droplet. Based on height and length of drop in contact with surface.

139


Where h and d represents measurements of the height and the length of the drop in

contact with the substrate respectively.

Each measurement for height and length of drop was repeated 5 times for each

photograph with the graphed mean averaged from the 15 measurements obtained

from the three photographs. Image analysis software capable of angular measurement

was used to confirm results. Values displayed in Figure 9-14 are expressed as mean

+/- standard error of the mean (SEM).

40

42

44

46

48

50

52

54

56

Con

tact

Ang

le (t

heta

)

Untreated Treated

Figure 9-14 Water droplet contact angle for native and 5-watt plasma treated PDMS cell substrate material. Error bars are ± standard error mean.

Surface hydrophilicity, as characterised by a decreased water contact angle, is

elevated by plasma treatment of the PDMS.

9.5 Post-Plasma Treatment Optimisation

Hydrophobic recovery of the plasma treated PDMS occurs when the surface is in

contact with air (Weikart and Yasuda, 2000) and although samples were immediately

140


placed in an air-sealed sample bag, a post-plasma treatment protocol was required to

maintain the hydrophilic surface until the samples were ready to be used for

experimental testing. Thus the main objective of the optimisation process was to

maintain the hydrophilicity of the surface during the time between completion of the

plasma treatment and the initiation of the cellular tests.

As discussed in Section 9.4, adsorption of proteins onto a surface occurs when placed

into an aqueous environment containing growth media with serum, aiding in the

attachment of cells (Schneider and Burridge, 1994). While protein adsorption occurs

on hydrophobic materials too, cell adhesion functionality is inhibited (Stephansson et

al., 2002) and preference is for hydrophilic surfaces (Steele et al., 1993). Serum

proteins mediate cellular responses to the surface through initial cell attachment and

cellular morphology (Ben-Ze'ev et al., 1980).

Five different procedures, three including serum soaking, were tested with ATR

surface analysis technique for their ability to enhance the surface’s concentrations of

hydroxyl and amine functional groups. Concentrations were interpreted as ratios of

area underneath the spectra at the specific wavenumber of the functional group with

respect to the area underneath the peak located at 2964cm-1. Each protocol’s

spectrum was normalised to this peak before areas were computed. Ratios were

expressed as an area index.

It was hypothesised that the level of adsorbed hydroxyl and amine functional groups

would increase with the length of time the surface is in contact with the water and

serum proteins up to a saturation point. It was also hypothesised that cellular

attachment would mirror this pattern of increase and plateau after the saturation point

in response to the amine and hydroxyl group concentration.

Visual observation and cell counts of cellular attachment and morphology were

undertaken to determine cellular responses to each post-plasma protocol.

141


9.5.1 Treatment Methods

The five protocols are outlined in Figure 9-15. The protocol 1 control entailed

leaving the plasma treated sample in the air sealed sample bag, however this bag was

not ideal and would have allowed air to enter, accelerating hydrophobic recovery.

Protocols 2, 3 and 4 all involved soaking the samples in cellular growth media

containing 10% serum for 1, 2 and 3 days after plasma treatment respectively. As

mentioned above, adsorption of proteins onto a surface occurs when placed into an

aqueous environment containing growth media serum. However this process is

extremely fast, with the absorbed protein layer composition changing over time

(Vroman and Adams, 1986), and in a more pronounced manner for hydrophilic

surfaces (Arai and Norde, 1990). Therefore, protein attachment time requires

optimisation before being utilised for cell attachment and growth.

When the prescribed media soaking time had elapsed for protocol 2 and 3, samples

were left with moisture on their surface and placed back into air sealed sample bags.

This was decided upon as dehydration of the surface after protein adsorption leads to

conformational changes (Xia et al., 2002) that would affect the cells recognizing and

attaching to the surface.

Protocol 5 was used as a positive control for the media soaking protocol 4 and

involved soaking the material in ultra pure water. This control was used to eliminate

the possibility that differences seen in hydroxyl concentrations were due to moisture

still on the surface of the sample after drying.

142


Protocol 4 Protocol 5Protocol 1 Protocol 2 Protocol 3

UV UVUV UV UV

Media WaterAir Air Air

WaterAir Air Media

Cellular Attachment and Proliferation Counts

Plasma PlasmaPlasma Plasma Plasma

Media Water

DAY 4

DAY 5

DAY 6

DAY 7

DAY 1

DAY 2

DAY 3

Air Media Media

Media

DAY 8

DAY 9

Figure 9-15 The post-plasma treatment protocols. Protocol 1 is negative control and Protocol 5 is positive control. All protocols were conducted over nine days with each day being a treatment of plasma, media, air or UV while over the last four days cellular attachment and proliferation

counts were conducted.

UV sterilisation of all protocols was conducted overnight for approximately 16 hours

after post-plasma treatment finished. Digital photographs of each sample were taken

each 24 hours up until 72 hours after initial cellular seeding which took place

immediately after cessation of UV sterilisation.

Digital photographs were to be used for cell attachment and morphology

quantification via image analysis software counting and manual counting techniques,

however photo quality was very low due to the membrane’s surface texture,

impeding the definition of cells on the imaging software (Figure 9-16). However,

manual counts of cell number from 5 predetermined areas on the membrane (shown

on Figure 9-16) provided an estimate of cell attachment and proliferation for each

post-plasma treatment protocol. These were conducted 3, 24, 48 and 72 hours after

initial cell seeding. The first count was used to quantitate the level of cellular

attachment with subsequent counts used to determine cellular proliferation.

143


PDMS cell substrate membrane

1

5 2 4

3

Figure 9-16 Photo of PDMS cell substrate after cells had attached. Also showing is the numbered location of each cell count. Circled numbers indicate predetermined areas used for

cell counting.

9.5.2 Results and Discussion

ATR data for the presence of hydroxyl groups (OH) confirmed that each post-plasma

treatment protocol utilising an aqueous soaking environment resulted in a higher

hydroxyl concentration on the surface when compared to the protocol 1 control. An

approximately proportionate increase in OH concentration is observed with each

extra day the sample was soaked in media. Protocol 5, where samples were soaked in

water for 3 days after plasma treatment showed little difference from protocol 2

which involved one day of soaking in media followed by 2 days in an air sealed

sample bag.

144

-1.0

0E-0

2

-5.0

0E-0

3

0.00

E+

00

5.00

E-0

3

1.00

E-0

2

1.50

E-0

2

2.00

E-0

2

2.50

E-0

2

3.00

E-0

2

3.50

E-0

2

4.00

E-0

2

4.50

E-0

2

3000

3100

3200

3300

3400

3500

3600

3700

3800

Wav

enu

mb

er (

cm-1

)

Absorbance

Pro

toco

l 1P

roto

col 2

Pro

toco

l 3P

roto

col 4

Pro

toco

l 5

1

2

3

4

5

Figu

re 9

-17

An

AT

R d

etai

led

spec

trum

scan

of t

he h

ydro

xyl r

egio

n (3

700-

3000

cm-1

) for

eac

h po

st-p

lasm

a tr

eatm

ent p

roto

col o

n PD

MS

cell

subs

trat

e m

ater

ial,

with

pro

toco

l num

bers

att

ache

d to

eac

h lin

e.

14

5


Comparison of the plasma treated PDMS and that of the protocol 1 control sample

clearly demonstrates the effects of hydrophobic recovery (characterised by hydroxyl

group concentration) when the surface is in contact with air (Figure 9-18).

0.00E+00

5.00E-03

1.00E-02

1.50E-02

2.00E-02

2.50E-02

3.00E-02

3.50E-02

4.00E-02

4.50E-02

300031003200330034003500360037003800

Wavenumber (cm-1)

Abs

orba

nce

Plasma Treated Protocol 1 Post-Plasma Treated

Figure 9-18 An ATR detailed spectrum scan of the hydroxyl region (3700-3000cm-1) for protocol 1 and plasma treated PDMS cell substrate material. Differences signify the hydrophobic

recovery of the surface after air contact.

Amine group concentration on the PDMS surface increased above the control

(protocol 1) for all protocols. Contrasting with hydroxyl group results, amine

concentration from protocol 4 did not show a proportionate increase above protocol

3. However, protocol 2 and 3 did follow this pattern (Figure 9-19).

Comparisons of amine group concentrations on the plasma treated PDMS and that of

the protocol 1 control sample (Figure 9-20) demonstrate the effects of hydrophobic

recovery (noted in Figure 9-18) on the attachment of amine groups to the surface.

146

0.00

E+0

0

5.00

E-0

3

1.00

E-0

2

1.50

E-0

2

2.00

E-0

2

2.50

E-0

2

3.00

E-0

2

1491

1498

1506

1514

1522

1529

1537

1545

1552

1560

1568

1576

1583

1591

1599

1606

1614

1622

1630

1637

1645

1653

1660

1668

1676

1684

1691

1699

1707

1714

Wav

enu

mb

er (

cm-1

)

Absorbance

Pro

toco

l 1P

roto

col 2

Pro

toco

l 3P

roto

col 4

Pro

toco

l 5

1

23

4

5

Figu

re 9

-19

An

AT

R d

etai

led

spec

trum

scan

of t

he a

min

e gr

oup

regi

on (1

700-

1500

cm-1

) for

eac

h po

st-p

lasm

a tr

eatm

ent p

roto

col o

n PD

MS

cell

subs

trat

e m

ater

ial,

with

pro

toco

l num

bers

att

ache

d to

eac

h lin

e.

14

7


0.00E+00

2.00E-03

4.00E-03

6.00E-03

8.00E-03

1.00E-02

1.20E-02

1.40E-02

1.60E-02

15001550160016501700

Wavenumber (cm-1)

Abs

orba

nce

Plasma Treated Protocol 1 Post-Plasma Treated

Figure 9-20 An ATR detailed spectrum scan of the amine group region (1700-1500cm-1) for protocol 1 and plasma treated PDMS cell substrate material.

Increases in concentration of each functional group were interpreted as ratios of area

underneath the particular functional group curve to that of a normalised peak for all

post-plasma treatments located at the 2964cm-1 wavenumber. These results are

shown in Figure 9-21 and Figure 9-22. Protocol 4, which involves three days of

soaking in cell growth media supplemented with serum, showed the highest change

in the hydroxyl group concentration. An approximately proportional decrease in this

concentration with reduction in soaking time was observed. The protocol 5 positive

control utilizing water soaking was significantly different from protocol 4,

confirming that differences seen in hydroxyl concentrations were not due to moisture

still on the surface of the PDMS during the testing procedure.

Protocol 1 maintained a higher concentration for both hydroxyl and amine functional

groups than protocol 5 confirming previous reports that an aqueous storage

environment for plasma treated materials helps reduce hydrophobic recovery

(Weikart and Yasuda, 2000). Hence as hydroxyl concentration for protocol 2 was on

a par with that of protocol 5 it could be suggested that media soaking of the material

148


for one day only serves to maintain the hydroxyl functional groups already on the

surface and not make any contribution to the concentration itself.

0

0.5

1

1.5

2

2.5

0 1 2 3 4 5 6

Post-Plasma Treatment Protocol No.

Are

a R

atio

s

Figure 9-21 The relative concentration of hydroxyl groups (area ratios) on the surface of PDMS cell substrate material after each post-plasma treatment protocol.

Amine concentration peaked for protocol 3 which involved only 2 days of media

soaking. Reorganisation of proteins on the surface occurs over time (Vroman and

Adams, 1986; Arai and Norde, 1990) and may in part contribute to the shift of the

peak amine concentration towards the lower serum soaking time of protocol 3. An

interesting point to note is the tendency for different proteins to have a preferential

peak of adsorption at different levels of blood plasma concentration (analogous to

serum) and may have had an influence in these results (Horbett and Schway, 1988;

Green et al., 1999). Protocol 2 exhibited a slightly higher ratio than protocol 5,

implying that proteins had adsorbed onto the surface of the material over the first 24

hour period.

Of question is the result for protocol 5 (positive water soaking control) which

displayed a greater amine concentration than that of protocol 1 which consisted of no

intervention post-plasma. However, the water used for soaking the membrane

149


material might have contained amines that adsorbed to the surface of the membrane,

confounding the results.

0

0.5

1

1.5

2

2.5

0 1 2 3 4 5

Post-Plasma Treatment Protocol No.

Are

a R

atio

s

6

Figure 9-22 The relative concentration of amine groups (area ratio) on the surface of PDMS cell substrate material after each post-plasma treatment protocol.

9.5.3 Cell Counts

Cell counts were taken from 5 predetermined areas (Figure 9-16) on each post-

plasma treated cell substrate after an initial seeding of 50,000 cells per well. A native

PDMS cell substrate was also sterilised and tested in the same manner. Tests were

conducted via manual counts on digital photographs taken of each well. The five

areas were located along the symmetry lines at the top, centre, bottom, right and left

of the substrate. As previously mentioned, surface texture of the PDMS cell substrate

membrane precluded cell counting from imaging analysis software. Results of cell

number were obtained from averaging counts over the 5 separate areas and the 3 cell

substrates tested for each post-plasma treatment. Consequently, each value displayed

in Figure 9-23 is derived from the average of 15 individual counts. Values were

interpreted at cells per well +/- standard error of the mean.

150


To confirm cell proliferation results, cell numbers at each time period for all

substrates tested were normalised to their respective initial attachment numbers and

presented as a percentage increase (Figure 9-24). Results confirm protocol 3 as

maintaining the highest number of cells per well and the highest rate of cellular

proliferation increase over the 3-day attachment and proliferation period. At all time

points, the native cell substrate showed a significant reduction in the number of

attached cells with respect to each of the post-plasma treated cell substrates

(Student’s T- test, p < 0.05). A significant reduction in proliferation rate compared

with the other cells substrates for the 48 and 72-hour time points was also observed

(Student’s T- test, p < 0.05)

151

0

50

00

0

10

00

00

15

00

00

20

00

00

25

00

00

30

00

00

Atta

chm

en

t 2

4 H

ou

rs 4

8 H

ou

rs 7

2 H

ou

rs

Cells per cell culture well

Pro

toco

l 1P

roto

col 2

Pro

toco

l 3P

roto

col 4

Pro

toco

l 5N

ative

#

#

#

Figure 9-23 The num

ber of cells attached to the surface of the PDM

S cell substrate for each post-plasma treatm

ent protocol and native PDM

S over 72 hours. # Indicates significant difference from

other protocols (p<0.05).

152

0

10

0

20

0

30

0

40

0

50

0

60

0

70

0

24

Ho

urs

48

Ho

urs

72

Ho

urs

Percentage Increase

Pro

toco

l 1P

roto

col 2

Pro

toco

l 3P

roto

col 4

Pro

toco

l 5N

ativ

e

#

#

Fi

gure

9-2

4 T

he p

erce

ntag

e in

crea

se in

num

ber

of c

ells

att

ache

d to

the

surf

ace

of th

e PD

MS

cell

subs

trat

e ov

er in

itial

att

achm

ent c

ount

s for

eac

h po

st-p

lasm

a tr

eatm

ent p

roto

col a

nd n

ativ

e PD

MS

over

72

hour

s. #

Indi

cate

s sig

nific

ant d

iffer

ence

from

oth

er p

roto

cols

(p<0

.05)

.

15

3


9.5.4 Conclusions

The results suggest that the hypothesis there is a saturation point for hydrophilicity

when soaked in media is partially correct. Protocol 3 which used two days of media

soaking exhibiting greater amine concentration on the surface than protocol 4 with

three days of media soaking. However, hydroxyl group concentration on the surface

peaked at 3 days of media soaking (Protocol 4), which was above all the other

protocols. Cell counts countered this finding and supported the original hypothesis,

with protocol 3 displaying the maximum cell attachment after three days.

Therefore, these results suggest that protocol 3 would provide an appropriate

hydrophilic surface for cellular attachment and spreading. Cell counts for this

protocol showed a greater proliferation and rate of increase of cell numbers over the

other cell substrate post-plasma treatments. Hence this protocol was used to prepare

membranes for cell studies using the DSD.

Of noteworthiness is that the 3-day timeframe was also used in in vitro tests of the

dual stimulus device, discussed in Chapter 11.

9.6 Discussion

Mechanical testing of the silicone based cell substrate demonstrated that it is a highly

elastic, anisotropic material during large deformations.

Surface characterisation of the silicone with x-ray photoelectron spectroscopy

confirmed the surface contains the monomeric repeating unit representative of a

polydimethylsiloxane (PDMS) material. These results were verified with an

Attenuated Total Reflectance (ATR) surface characterisation study, which also

confirmed the lack of hydroxyl and amine functional groups on the surface.

The ‘Gas Plasma Treatment’ creates a hydrophilic surface while physically etching

the surface (Weikart and Yasuda, 2000). The power used to drive the high frequency

154


electric field was found to produce the required properties at a value of five watts.

XPS analysis of the surface constituents verified the modification of the surface.

During plasma treatment this power value varied occasionally, independent of the

operator or initial settings. Although minutely small, these variations may have

created variations in surface etching depth. The distribution of plasma within the

vacuum chamber and over the PDMS sample during operation of the machine was

observed as inhomogeneous and worth noting is the possibility that surface

modifications may have occurred in localised regions at a higher degree than across

the rest of the PDMS surface. Depending upon a number of factors such as position

of the sample in the plasma chamber, turbulence in water vapour flow and variations

in radio frequency tuning (leading to power creep) the plasma treatment could

possibly leave regions of greater hydrophobicity than others resulting in different

protein adsorption mechanisms. However, even if this did occur, a combination of

water affinity and repulsion has been shown to facilitate fibronectin (a cell

attachment promoting protein) adsorption and an associated increase in cellular

spreading (Horbett and Schway, 1988).

Optimisation of post-plasma treatment was undertaken by examining 5 different

protocols. Hydroxyl and amine concentrations on the surface after post-plasma

treatment had elapsed were quantitated with ATR and used to compare each

protocol.

It is acknowledged that no studies of the attached proteins or their confirmations,

which have the ability to affect cellular function (Horbett, 2003), were undertaken.

However, this was due to the limited time available for these studies.

9.7 Conclusions

A polydimethylsiloxane (PDMS) deformable membrane of approximately 50µm

thickness, exhibiting hydrophobic surface properties, was physicochemically

characterised. Gas plasma treatment of the surface was undertaken to increase the

material’s affinity for water (hydrophilicity). The resulting treated surface was fully

155


characterised as containing a greater concentration of hydroxyl functional groups,

essential for cellular attachment and growth.

A post-plasma treatment protocol, used after surface treatment but before cellular

experimentation, was optimised to maintain and maximise the concentration of

hydroxyl groups and proteins on the surface. The result of this was a procedure

involving soaking the PDMS material for two days with serum supplemented cell

growth media followed by a day without, before UV sterilisation and cellular

experimentation was initiated.

The next chapter describes the measurement of the surface strain that is induced in

the PDMS cell substrate from activation of the DSD.

156

Chapter 10: DSD Surface Strain Characterisation

10 DSD Surface Strain Characterisation

10.1 Experimental Surface Strain

10.1.1 Methods

A number of different techniques to record the surface strain on the deformable

membrane in the DSD were evaluated. These included:

• Use of miniature strain gauges attached to the surface,

• Quantifying the change in surface area and thus surface strain,

• Using Poisson’s ratio,

• Visualisation and tracking of surface marks during dynamic strain and

• Visualisation and tracking of surface marks during static strain.

The first three methods were dismissed, as the available measurement tools were not

capable of quantitating the minute strains employed by the DSD. Other fundamental

problems, such as the measuring tool’s effect on the surface strain itself, ruled out

these techniques (i.e. attached strain gauges).

Visualisation and tracking of marks on the surface was undertaken with the use of

ink dots placed at 1mm intervals along the diameter of the cell substrate. Ink dots

were positioned with the help of a specifically designed template that was placed

between the central pin actuator and the underside of the cell substrate membrane

while dots were manually placed on the cell substrate from the topside. This template

(Figure 10-1) was produced from a 20mm diameter, 0.1mm thick plastic shim

material. Two perpendicular diametrical lines, each with nineteen 0.35mm diameter

holes were drilled at 1mm intervals. A specialised jig (Figure 10-2) was designed and

produced to tightly grip the template during CMC high speed drilling and removal

via a manual punch after drilling was completed.

157


2mm

Figure 10-1 Holes drilled in polymer template used for marking ink dots on cell substrate during strain calibration.

Figure 10-2 Specialised jig for tightly gripping the ink marking template during CMC high speed drilling. This jig also includes a punch for accurate removal of the template after drilling

was completed.

158


Ink dots were placed along the two diametrical lines dictated by the template and

again at an arbitrary but small-rotated position (<5°) to facilitate both radial and

circumferential strain measurement.

Figure 10-3 diagrammatically represents the result of this process. The DSD with ink

dot marked cell substrate was placed underneath a stereomicroscope (Model No.

MZ8, Leica Microsystems) with a digital camera (Model No. CoolPix 4500, Nikon)

attached. When maximum zoom on the camera and microscope were used together,

the system was capable of providing a viewfinder screen dimension of 1.32mm X

0.993mm. This equated to a ~0.5µm pixel dimension resolution. This equates to an

approximate strain variation of 0.5µε, well below the measured values of strain.

Figure 10-3 A diagram of cell substrate and locations of ink dots used for experimental strain calculations. Position 1 on subsequent figures for radial strain is defined as the

displacement/strain between the centre ink dot (c) and ink dot 1 (1). This follows through to point 9. Circumferential strain at position 1 is defined as the displacement/strain between ink

dot 1 (1) and ink dot 1’ on the rotated axes (1’).

159


10.1.1.1 Static Experimental Strain For this test, the Piezoelectric Actuator was powered to give its maximum

displacement of 72µm. Strain (ε) was computed from the change in the length of the

original distance between successive points via the engineering strain relationship in

Equation 10-1. An example of this would be between point 4 to 5 on the membrane

for radial strain and point 4 to 4’ on the membrane for circumferential strain.

ε = (Lf – Lo) / Lo

Equation 10-1 Engineering strain equation used to compute cell substrate strain between adjacent dots.

Where Lo is the original length between the dots and Lf is the length after strain has

been applied to the cell substrate. As strain and measurement magnitudes were small,

distances between adjacent points for circumferential strain calculations were

assumed linear.

Four megapixel digital photographs (Nikon Coolpixs 4500, Japan), before and after

the strain was applied, were taken to measure displacements between ink dots with

use of the ImageTool program (University of Texas Health Science Centre in San

Antonio, San Antonio, Texas, U.S.A.). This program allows images to be calibrated

with a known distance, facilitating direct measurement of displacements.

The ‘tenting’ of the cell substrate membrane from the vertically moving central pin

causes strain out of plane from the microscope focus thus effecting strain

measurements taken from the microscope’s digital photographs. Therefore, all

measurements of the strain were corrected for this by using an idealised

trigonometric ratio. From Figure 10-4 we can see that photographs would record

strain along the original plane of the cell substrate (εm), therefore to correct for this

discrepancy Equation 10-2 is used to obtain real strain (εr). i.e:

160


εr = (εm)/cos θ

Equation 10-2 Strain correction equation for measuremed strains.

Theta (θ) was obtained from the trigonometric relationship of:

tan θ = (H / Cell culture well radius)

= (0.072/10)

= 0.0072

θ = 0.00719989 rad

Camera

θ

Deformed

Membrane

Central

Pin

Cell culture well

H

εm

εr

Figure 10-4 A diagram of cell substrate during activation of the DSD. The diagram shows bending membrane as dotted lines. Camera loaction resulted in correction factors for measured

strain to be implemented.

Only maximum strain (from maximum Piezoelectric Actuator displacement) was

recorded due to the minute nature of the strains. Maximum resolution of both the

camera and microscope resulted in photographs containing only two adjacent dots

(e.g. point 1 and 2 only or 1 and 1’ only, etc) during the straining process.

161


The radial displacement and hence radial strain measurements for the entire radius of

the cell substrate was compiled from 18 successive photographs (9 in the unstrained

position, 9 in the strained position). Each radial line of ink dots was appraised in this

manner to confirm homogeneity.

10.1.1.2 Dynamic Experimental Strain Dynamic experimental strain was recorded by tracking ink dots in video files of the

active DSD. Short five second ‘.mov’ video files were taken during dynamic

actuation of the cell substrate at frequencies up to 8Hz (limit of camera’s recorded

frames per second rate). These files were then converted to ‘.avi’ format for use in an

image software package called ImageJ (National Institutes of Health, Bethesda,

Maryland, U.S.A.). Each frame in the movie file was extracted with the use of a

frame grabber plug in accessed through the NIH ImageJ website. Ink dots were

digitally marked in the program (by using a threshold image of each frame) and

tracked over the entire movie file sequence.

10.1.2 Results

10.1.2.1 Large Static Experimental Strain It was decided that the maximum cell substrate displacement imparted from the

Piezoelectric Actuator (~72µm) would not be large enough to make accurate surface

strain measurements. Therefore a process of interpolation from larger strains was

undertaken. Increasing central pin displacements of 0.1mm were created by placing

small spacers below the Central Pin Insert with the resulting cell substrate surface

strains recorded via the methods described above. Initially only radial strain was

measured to ascertain the viability of this strain quantification method.

Results show a relatively large variability in the measured strains for all radial

positions at low central pin heights (Figure 10-5). It is also clear there is no

consistent trend in levels of radial strain with radial position as required for an

accurate interpolation of strain levels for a 72µm Central Pin Insert displacement

162


(Figure 10-6). It was decided that a more detailed analysis of the smaller Central Pin

Insert heights between 0 – 0.5mm was required.

-2

0

2

4

6

8

10

12

14

16

18

0 0.5 1 1.5 2 2

Central Pin Insert Height (mm)

Stra

in (p

erce

ntag

e)

.5

Position 1 Position 2 Position 3 Position 4 Position 6 Position 7 Position 8 Position 9

Figure 10-5 Radial strain at different radial positions on cell substrate when substrate membrane is deformed at different heights. See Figure 10-3 for description of radial positions.

0

2

4

6

8

10

12

14

16

18

0 1 2 3 4 5 6 7 8 9 10

Position Number

Stra

in (p

erce

ntag

e)

0.1mm 0.5mm 1mm 1.5mm 2mm

Figure 10-6 Radial strain with differing cell substrate deformations over all radial positions on cell substrate. See Figure 10-3 for description of position number.

163


10.1.2.2 Small Static Experimental Strain As mentioned above, focus was applied to the 0 – 0.5mm range of Central Pin Insert

heights in an attempt to establish a clearer trend between the Central Pin Insert

height and level of cell substrate surface strain. Two Central Pin Insert heights

(0.1mm and 0.5mm) were tested during the process. Methods followed the same

pattern as previously described. In the interests of statistical validity, 5 photographs

of the same radial position (each with 5 repeated measurements of displacement)

were taken in each of two seperate strain calibration experiments. Therefore,

determination of strain at each radial and circumferential location was made from 50

measurements.

Results presented a situation similar to large experimental strain; no clear trend

between strain and radial position was found (Figure 10-7 and Figure 10-8).

However when reinterpreting the data as a variation in strains with the 0.1mm and

0.5mm Central Pin Insert heights, a slight trend appears for the radial strain

measurements. Gradient values of the increasing strain when moving from 0.1mm to

the 0.5mm Central Pin Insert displacement appear to decrease with an increasing

radial position (Figure 10-9 and Figure 10-10).

These linear gradients were used to interpolate the cell substrate surface strain that

would be experienced during DSD actuation with biological cells where Central Pin

Insert displacement reached 72 microns (Figure 10-11 and Figure 10-12). Surface

strain values were calculated using linear interpolation with Equation 10-3.

Strain (percentage) = -[Linear gradient of 0.1mm to 0.5mm line * (0.0001-

0.000072)] –[Strain at 0.1mm Central Pin Insert height]*100

Equation 10-3 Interpolated surface strain values during DSD activation.

Maximum radial strain, as calculated by the interpolation, starts at approximately

2000µε (Position 1) decreasing towards the edge of the cell culture well until the

final point, which displays an very large strain value of ~5600µε. Circumferential

164


strain varies around an approximate mean of –10000µε, an extremely large value

when compared to radial strain.

0

1000

2000

3000

4000

5000

6000

7000

8000

0 1 2 3 4 5 6 7 8 9

Radial Position

Stra

in (m

icro

stra

in)

10

0.1mm 0.5mm

Figure 10-7 Radial strain at different radial positions on cell substrate for two membrane deformation heights of 0.1mm and 0.5mm. See Figure 10-3 for description of position number.

-16000

-14000

-12000

-10000

-8000

-6000

-4000

-2000

0

0 1 2 3 4 5 6 7 8 9

Radial Position

Stra

in (m

icro

stra

in)

10

0.1mm 0.5mm

Figure 10-8 Circumferential strain at different radial positions on cell substrate for two membrane deformation heights of 0.1mm and 0.5mm. See Figure 10-3 for description of

position number.

165


y = -1572.6x + 5789.4

y = 5766.9x + 1642.5

y = 3381.6x + 1486.7

y = 8880.5x + 673.95

y = 1456.7x + 922.49

y = 6728.6x - 220.31

y = 12621x + 313.08

y = 12980x + 790.5

y = 4500.6x + 1402.2

0

1000

2000

3000

4000

5000

6000

7000

8000

0 0.1 0.2 0.3 0.4 0.5 0.6


Stra

in (m

icro

stra

in) Position 1

Position 2Position 3Position 4Position 5Position 6Position 7Position 8Position 9

Figure 10-9 Relationship between radial strain and membrane deformation for each radial position on cell substrate. Regression equation lines are placed in order from position 1 to 9. See

Figure 10-3 for description of position number.

y = 12253x - 14197

y = -2110.5x - 5505.1

y = 12682x - 12115

y = -163.76x - 12080

y = 3905.3x - 11084

y = 3718.3x - 8075.2

y = 12902x - 14664

y = 4139.9x - 8560.2

y = 8193.4x - 12181

-16000

-14000

-12000

-10000

-8000

-6000

-4000

-2000

0

0 0.1 0.2 0.3 0.4 0.5 0.6


Stra

in (m

icro

stra

in)

Position 1

Position 2

Position 3

Position 4

Position 5

Position 6

Position 7

Position 8

Position 9

Figure 10-10 Relationship between circumferential strain and cell substrate membrane deformation for each radial position on cell substrate. Regression equation lines are placed in

order from position 1 to 9. See Figure 10-3 for description of position number.

166


0

1000

2000

3000

4000

5000

6000

0 1 2 3 4 5 6 7 8 9 10

Radial Position Number

Stra

in (m

icro

stra

in)

Figure 10-11 Radial strain vs radial position for cell substrate. These values were interpolated strains for 72µm central pin displacement, calculated from the regression lines in Figure 10-9

and Equation 10-3. See Figure 10-3 for description of position number.

-16000

-14000

-12000

-10000

-8000

-6000

-4000

-2000

0

0 1 2 3 4 5 6 7 8 9 10

Radial Positon Number

Stra

in (m

icro

stra

in)

Figure 10-12 Circumferential strain vs radial position for cell substrate. These values were interpolated strains for 72µm central pin displacement, calculated from the regression lines in

Figure 10-10 and Equation 10-3. See Figure 10-3 for description of position number.

167


10.1.2.3 Dynamic Strain Particle Tracking Unfortunately, the minute cell substrate strains created by the DSD and the reduced

resolution of the camera (due to a lack of optical zoom during movie capture)

resulted in unreliable data. However, the process did confirm that strains were within

the range determined by static strain calibration.

10.1.3 Discussion and Conclusions

At larger heights of approximately > 1mm, a quasi-exponential pattern between

radial position and strain appeared (Figure 10-6). This can be explained by the

elimination of bending stresses, which keep the deformable membrane acting much

like an elastic plate with linear strain relationships. Therefore eliminating these

bending stresses results in non-linear strain behaviour with radial position. From the

observed data (Figure 10-5 and Figure 10-6), it can be assumed that the displacement

of the Central Pin Insert from the Piezoelectric Actuator (72µm) would not force the

membrane into that non-linear strain state.

Large variation in radial strains occurred at the final point (point 9) in both the small

experimental strain pin heights and the interpolated values for 72µm, as displayed in

Figure 10-7. Clamping the membrane slightly in from the edge of the cell culture

well creates frictional resistance of the membrane rubbing across the Top, impeding

displacement and hence strain (Figure 10-13). Therefore ink dot 9 does not move

significantly resulting in relatively large movements between it and ink dot 8 and the

resulting discrepancies in recorded strain.

Results showed significant variability in circumferential strain, much greater than

radial strain. However, similarities in strain variability between the two Central Pin

Insert heights (0.1mm and 0.5mm) with respect to radial position for both the radial

and circumferential strain graphs (Figure 10-7 and Figure 10-8) lend credence to the

theory that cell substrate surface inhomogeneities have played a significant part in

causing the variability seen with strain measurements.

168


Determination of the true surface radial strain from experimental results can only be

estimated by interpolation to be within 2000µε for the majority of the cell culture

well, increasing significantly at the edge due to frictional drag of the Deformable

Cell Substrate Membrane over the Top. Circumferential strain appears to be in

compression varying around –10000µε for all radial positions.

169

M

em

br

an

e c

lam

pe

d s

ligh

tly in f

ro

m c

ell c

ultu

re

wa

ll ca

us

ing

fr

ictio

na

l dr

ag

O

-Rin

g

Ce

ll

Su

bs

tra

te

To

p

Ce

ll

su

bs

tra

te

an

nu

lus

Figure 10-13 Close up of cell substrate tethering w

ith the DSD

Top, Cell Substrate A

nnulus and O-R

ing. See Figure 7-6 for full cross sectioned DSD

.

170


10.2 Theoretical Surface Strain

The difficulty in characterising the cell substrate surface strain prompted a theoretical

analysis to analyse the problems encountered during experimental testing.

10.2.1 Methods

The finite element analysis (FEA) software package, Abaqus (Abaqus Inc.,

Providence, RI, USA), was used to model the thin Deformable Cell Substrate

Membrane, initially with the following details:

• 20mm diameter

• 0.05mm thick

• Initial central pin displacement of 0µm

• 72µm Central Pin Insert displacement at the centre after actuation

• Clamped edge boundary conditions (Encastre – no translation or rotation)

• Cell substrate membrane properties of

o 2.6MPa Young’s Modulus (Determined experimentally – Section

9.1.1)

o 0.45 Poissons Ratio (Callister, 2005)

• Hexagonal elements at a seeding of 0.5 (global element size)

• Linear ramping of strain over 10 increments

• Cell substrate modelled as a deformable membrane material

10.2.1.1 Sensitivity Analysis To determine the influence of these particular factors on the final analysis outcome,

variations of their values were made and FE analyses were recomputed. Variations

were made for factors when either experimental observation or theoretical deviations

based on reference texts were noted. Each factor was iterated within this range of

possible values. When changes were made, other factors were held at their original

values.

The following is a list of values used for each factor:

171


• Young’s Modulus 1.75Mpa, 2Mpa, 2.25MPs, 2.5Mpa

• Poisson’s Ratio 0.35, 0.5

• Membrane Thickness 20µm, 30µm, 40µm, 60µm, 75µm

• Initial central pin displacement 10µm, 20µm, 30µm, 40µm

• Central Pin Insert displacement of 100µm and 500µm (as tested

experimentally - discussed in next section)

10.2.2 Results

The initial FE analysis of radial strain along with radial position presented a

significantly non-linear pattern. Figure 10-14 shows maximum in-plane principal

strain is ~120µε at the central pin position, decaying at a quasi-exponential rate

towards the edge of the cell culture well. Approximately 80% of the surface was

experiencing a maximum in-plane principal strain in the radial direction of 30µε or

less. A visual illustration was made with a colour contour plot in Figure 10-15.

Figure 10-14 Finite Element Analysis of maximum in-plane principal strain with radial position from the centre (0) to the edge (10) of the cell culture well when undergoing actuation in the

DSD.

172


Figure 10-15 A Finite Element Analysis colour contour plot of cell substrate in plane principal surface strains.

Circumferential strain (Figure 10-16) on the other hand seems to lie very close to

zero for the entire length of the radius apart from the central pin region, however this

large variation is due to the size of the elements in the FE mesh at the central region.

A strain vector plot on the surface in Figure 10-17 demonstrates that the magnitude

of the radial (red vectors) strongly outweighs the circumferential (yellow vectors)

strain.

173


Figure 10-16 A Finite Element Analysis of circumferential strains on cell substrate with radial position from the centre (0) to the edge (10) of the cell culture well when undergoing actuation in

the DSD. Variation of peak at point 0.5 is due to inconsistencies in the FEA mesh.

Figure 10-17 A Finite Element Analysis maximum strain vector plot of the dual stimulus device surface during active deformation. The red arrows signify an in plane principal strain as the

predominant strain present.

174


In the interests of space, results from all iterations of varied factors will not be shown

(see Appendix D). All variations in factors, apart from those noted below, did not

significantly affect the results from the initial analysis. A change in strain with a

change in Poisson’s ratio was noted. When Poisson’s ratio is set at 0.35, radial strains

showed higher strain at the centre and lower strain at the edge of the cell culture well

than when Poisson’s ratio was increased to 0.5. The higher ratio results in strain

maintaining a higher level across the radial path even though a lower peak strain at

the central pin was experienced.

10.3 Comparison of Experimental and Theoretical

Surface Strain

Analyses with the 0.1mm and 0.5mm central pin displacements were undertaken to

facilitate comparison of experimental and theoretical results. Experimental radial

strains were plotted with the maximum in-plane principal strains, as these were

predominately radial in nature (Figure 10-18 and Figure 10-19). The results do not

match in either of the central pin displacements of 0.1mm and 0.5mm, however the

discrepancy is more pronounced in the 0.1mm displacement.

As discussed in the previous section, circumferential strains showed a strongly

negative nature as would be experienced by compression of the cell substrate

membrane. These results were significantly different from those in the FE analysis

(Figure 10-20 and Figure 10-21).

175


Figure 10-18 A graph of radial strain with radial position comparison between theoretical (green line) and experimental (red dots) studies. The central pin displacement was set at 0.1mm.

Figure 10-19 A graph of radial strain with radial position comparison between theoretical (red line) and experimental (green dots) studies. The central pin displacement was set at 0.5mm.

176


Figure 10-20 A graph of circumferential strain with radial position comparison between theoretical (green line) and experimental (red dots) studies. The central pin displacement was

set at 0.1mm.

Figure 10-21: A graph of circumferential strain with radial position comparison between theoretical (green line) and experimental (red dots) studies. The central pin displacement was

set at 0.5mm.

177


10.4 Discussion

Disagreement between the measured and the theoretical surface strains of the

deformable membrane was observed during the calibration process. Although visual

observation through different microscopes confirmed the presence of membrane

strain during operation of the DSD, experimental and theoretical results differed

significantly. Experimental measurements were larger than theoretically derived

results for all testing procedures. These errors may be attributed to the minute scale

of the strains produced and limitations in the method of strain quantification (manual

measurements from digital photographs).

Large variations in the cell substrate thickness were highlighted in Section 9.1.2

while its anisotropic behaviour was highlighted in Section 9.1.1. These results

support the theory that the silicone elastomer used as the cell substrate displays

inherent inhomogeneity and is the cause of the large variations in experimental

surface strain calculations.

When larger calibration strains from the central pin displacement were created there

was still no union of experimental and theoretical strain, although results did tend

towards a closer match (Figure 10-19).

If we follow the experimental maximums, radial strains produced by the DSD are of

a physiological magnitude with 1000µε at the edge of the cell culture well increasing

in a quasi-linear fashion to a maximum of approximately 2000µε occurring at the

centre of the membrane where the Central Pin Insert is located (Figure 10-11).

However, if we are to believe the theoretical FEA results, there is a non-linear

change in radial strain with radial position and as a result approximately two thirds of

the radius experiences a strain of 30µε or less while the maximum strain occurring at

the central pin region is approximately 120µε (Figure 10-14).

Circumferential strains showed more variability between experimental and

theoretical results than those from the radial strain. Interpolation of the experimental

strain experienced during operation of the DSD concluded that circumferential strain

178


varied around –10000µε for all radial positions (Figure 10-12), however as this strain

is 5 times the size of radial strain near the central pin then this would appear to be an

unfeasible estimate. Supporting this theory is the FEA’s analysis confirming that the

circumferential strain is close to zero for all the radial positions (Figure 10-16).

Bone cell strain in a mechanical stimulus device depends upon many factors, such as

the cells’ level of attachment to the substrate, orientation and their state of cellular

division (Brown, 2000). Also, others have observed that cells only experience

approximately 60% of the applied cell substrate strain (Winston et al., 1989).

A purely flat substrate containing no topographic irregularities, such as that modelled

by the finite element analysis, describes an idealistic in vitro strain situation such as a

glass slide cell substrate. Surface strain results from the FE analysis assume attached

cells experience the same strain as that of the substrate surface. In reality cell

attachment is a complex event, which will determine the magnitude and direction of

the strain vector transmitted to the cell. Attachment can be made through focal

adhesions, close contact or extracellular matrix contact. Strain between adjacent

surface asperities at a different magnitude than the homogenised bulk material strain

may occur if the material is of a compliant nature such as the silicone elastomer used

in this study. If these surface peaks are used as focal attachment points for a cell,

then the applied strain level to the cell may differ from that applied to the cell

substrate, resulting in incorrect data.

10.5 Conclusions

Experimental tests of cell substrate surface strains during DSD activation were

inconclusive. The minute scale of the strains produced by the DSD, limitations in the

method of manual strain quantification and the observed inhomogeneities of the

deformable cell substrate are attributed as the causes of the experimental strain

variation with the theoretical FEA surface strain results.

However, the range of experimental radial strains across the cell substrate were of a

magnitude bone cells would experience in vivo, thus warranting the DSD’s use on in

179


vitro cultures of bone cells. A more rigorous strain characterisation was not justified

due to equipment cost and time constraints.

Dual mechanical and PEMF stimulation on cultures of bone cells using the DSD is

described in the next chapter (Chapter 11).

180

Chapter 11: Dual Stimulation of Cultured Bone Cells

11 Dual Stimulation of Cultured Bone

Cells

11.1 Introduction

Bone cells are mechano- and electro- responsive, meaning they adapt their phenotype

in response to a mechanical or electrical stimulus. As discussed in Chapter 2, 3 and

4, bony tissue exhibits a finely tuned adaptive mechanism to both these stimuli,

exhibiting thresholds and non-linear responses to the many different factors involved

in each of these two exogenously applied stimulants. Combinations of both

stimulants occur during normal loading of bone, each individually generating a

cellular response. However there is no data available on the synergistic or hierarchal

role between the two in creating bone adaptation and is yet to be studied in vitro.

As discussed previously, there is a growing amount of evidence that bone is exposed

to a constant level of background strain during daily activities. This strain is smaller

in magnitude (~5µε) than locomotion impact strains, which can reach levels of 100-

3000µε (Burr et al., 1996; Adams et al., 1997) and occurs throughout the life of the

bone originating from the dynamic movement of muscle attachment points during

postural control. Concomitant with these mechanical strains, bone cells in vivo

experience an electric field due to the movement of charged fluid through the bone

porosities in the order of 0.1-10mV/cm, pulsed at frequencies of <10Hz with a

distorted trapezoid pulse shape (Otter et al., 1998; Pilla, 2002b).

As bone is bombarded with both these stimulants in vivo, a study of their hierarchical

and synergistic effects upon in vitro cultures of bone cells is central to the elucidation

of the electro- and mechanotransduction pathways.

181


11.2 Materials and Methods

11.2.1 Dual Stimulus Device (DSD)

As discussed in Chapter 8 and 10 respectively, the dual stimulus device is capable of

imparting a pulsed electromagnetic field and dynamic strain to cell cultures.

Stimulants were used in unison or individually during 3 separate stimulation

protocols (see Section 11.2.2).

11.2.2 Experiments

Three stimulation protocols were experimentally tested with the dual stimulus

device. The first was the electrical stimulation of cell cultures with PEMFs.

Secondly, mechanical straining of the cell substrate was used during the

experimental procedure. The last protocol stimulated the cell cultures with both the

electrical and mechanical stimulants at the same time.

All three protocols were repeated three times with each experiment conducted over a

3-day timeframe. From observations in preliminary cell culture studies (Section

9.5.3) this length of time was required for the cultures to reach a logarithmic rate of

growth. This was desirable for robust LDH proliferation measurements to be

conducted. DSD stimulation/s were applied over an 8-hour period each day, followed

by no stimulation. This stimulation timing shows greater consistency in results for

PEMF exposed cultures as was described in Section 6.2. Figure 11-1 represents the

timing of stimulation in a graphical format. Before tests were begun, each part of the

DSD in contact with cells and/or growth media were autoclaved for sterilisation

(apart from the cell substrate which was UV sterilised).

182


Da

y 1

D

ay

2

Da

y 3

8 Hr

Figure 11-1 Timing of stimulant/s from the dual stimulus device (DSD) during 3-day experimentation protocol. Shaded region signifies the activation of the DSD stimulant/s.

11.2.2.1 Mechanical Stimulation Displacements of the cell substrate, located at the centre of the cell culture well, were

actuated at a frequency of 5Hz and approached a maximum of 65µm (graphically

represented by trace B in Figure 11-2). This was less than the 72µm maximum

output of the actuator during static actuation, as increasing actuator frequency

decreases displacement output (discussed in Section 8.2.3).

As discussed in Chapter 10, the surface strain in both the radial and circumferential

directions is a function of radial position and central pin displacement.

11.2.2.2 Electrical Stimulation Induced electrical fields in the cell culture from the PEMF stimulus are represented

by trace A in Figure 11-2. As displayed in Figure 5-15, each dark band in Figure

11-2 is made up of 20 individual quasi-square wave pulses. The maximum change in

induced voltage is 138mV.

183


A 138mV

0

10

20

30

40

50

60

70

80

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5

Time (sec)

Dis

plac

emen

t (m

icro

ns)

B

Figure 11-2 Timing of dual stimulants during activation of dual stimulus device. Trace A represents the PEMF signal's repetitive pulse burst. Trace B represents the mechanical

deformation (and hence strain) of the cell substrate membrane.

11.2.3 Cell Cultures

A human osteosarcoma cell line, MG-63 (ATCC No: CRL-1427, Rockville,

U.S.A.), was used for these experiments. These are a human bone cancer derived cell

line with a fibroblastic morphology. The cells were routinely cultured in phenol red

free minimum essential medium alpha (αMEM) supplemented with 10% fetal bovine

serum, 1% penicillin - streptomycin diluted from stock solution [both 5,000 U/ml]

and 0.01% gentamicin [10mg/ml] (All from Gibco, Grand Island, NY, USA). All

cells used were less than 8 passages from original cell stock. For experimental

procedures, a density of 50,000 cells per well were seeded (with 1ml of growth

media) into the dual stimulus device and control culture plates. This density was

determined to be ideal from initial seeding density-confluence tests within the in-

active DSD, confirming MG-63 cells were undergoing log phase growth after 3 days;

the length of the testing procedure.

184


Cell culture wells had an effective growth area of 3.14cm2. The cells were allowed

to attach for 2 hours before stimulation with the DSD began. All experimental

procedures were conducted within a CO2 incubator at a temperature of 37˚C in an

atmosphere of 95% air/5% CO2 and 100% relative humidity. Each procedure was

repeated 3 times for statistical validity.

A six-well control cell culture plate was prepared with specially designed well inserts

(Figure 11-3) manufactured from the same material (Lexan Polycarbonate) as the

DSD Top. Control culture wells contained identical PDMS cell substrates as used in

the dual stimulus device, mimicking the environment of an inactive DSD. Inserts had

an outer diameter of 35mm to facilitate push fit into the cell culture wells (Figure

11-4). Inner diameters were the same as the DSD at 20mm. Visual observation from

pre-experimental testing confirmed the absence of growth medium from around or

underneath the insert, proving that it effectively confined the medium to the central

region of the insert where the cells were located.

Control cultures were placed as far from the DSD in the cell growth incubator as

possible. From calibration studies, they experienced a background magnetic flux of

+/- 0.1 G.

Figure 11-3 The specially designed cell culture well inserts for use in control cultures. These effectively reduced the cell growth area to match that of the DSD.

185


Figure 11-4 Control cell culture plates with well inserts push fit into position for experimentation.

Measurements of temperature within the PEMF exposed cell culture wells during

operation of the coils (with door of incubator closed) were conducted with a

thermocouple. These results demonstrated no increase in temperature above the basal

37˚C, and as a result heating effects on the exposed cell cultures were not present.

11.2.4 Proliferation

Cellular proliferation was assessed via a lactate dehydrogenase (LDH) based

toxicology assay kit (Product No: TOX-7, Sigma Aldrich, St. Louis, MO, U.S.A).

Over the course of this PhD the potential for error from the radiolabelled leucine

proliferation test protocol was identified (Parker et al., 2005). This prompted the

adoption of the Lactate Dehydrogenase (LDH) proliferation assay that has been used

186


previously for the analyses of osteoblast cell cultures undergoing mechanical

straining (El Haj et al., 1990; Neidlinger-Wilke et al., 1994). This assay was tested

rigorously by colleagues in the laboratory, and shown to produce reliable results. It

was subsequently adopted as the standard laboratory technique for proliferation and,

hence, was used for the dual stimulus device.

LDH based proliferation measurement involves measuring the total number of cells

after lysis via total cytoplasmic lactate dehydrogenase. The assay is based on the

reduction of a substrate solution (NAD) to NADH, by LDH, which is utilised in the

stoichiometric conversion of a tetrazolium dye. The resulting coloured compound

was measured spectrophotometrically.

At the completion of each experiment, 100µL aliquots of the conditioned culture

medium from stimulated and control cultures were taken before lysis to determine

relative cell viability. Lysis of cells was achieved with the addition of 100µL of cell

lysis solution (Product No. L-2152, Sigma Aldrich, St. Louis, MO, U.S.A - provided

in LDH assay kit) to the culture well of the stimulated well and each culture well of

the control plate, followed by incubation for 45minutes within the incubator. Upon

completion, wells were triturated and three separate 100µL aliquots from the

stimulated and each of the control culture wells was transferred to a 96 well cell

culture plate (Nuncleon, Roskilde, Denmark) and supplemented with 50µL of LDH

assay mixture. This mixture had a constitution of equal parts LDH substrate, enzyme

and dye solutions.

Immediately following addition of the LDH mixture, the 96 well plate was covered

from light with aluminium foil and incubated at room temperature for 30 minutes

after which the reduction reaction was terminated with 15µL of 1N hydrochloric acid

(HCL).

Bubbles in each well of the 96 well plate were burst with the use of pipette tip before

the plate was placed in the spectrophotometric plate reader (Dynex MRX, Dynex

Technologies, Chantilly, VA, USA). Readings of light absorbance at a wavelength of

490nm were taken, with background absorbance at a wavelength of 690nm

187


automatically subtracted from results. Three separate readings, each one-minute

apart, were taken to confirm the reduction reaction had been stopped.


Levels of the enzyme alkaline phosphatase in the cell lysate were measured for an

indication of early stage osteoblastic differentiation (Dworetzky et al., 1990). To

measure alkaline phosphatase production, a reagent containing p-nitrophenyl

phosphate (pNPP), a chromogenic alkaline phosphatase substrate, is added to the cell

lysate. Cleavage of the phosphate by the enzyme results in a yellow coloured p-

nitrophenyl, when in basic solutions. This allows detection of the enzyme by

measuring light absorption at a specific wavelength using a spectrophotometer.

Three 50µL samples of cell lysate from each of the stimulated and control cell

culture wells were admixed to 150µL of p-nitrophenyl phosphate substrate (1mg/ml

pNPP, 0.2M Tris buffer - Product No: N2770, Sigma-Aldrich, St. Louis, MO,

U.S.A.). Samples were placed in a 96 well plate (Nucleon, Roskilde, Denmark) and

allowed to incubate at room temperature for 30 minutes, fully covered from light by

aluminium foil. Following incubation, reactions were stopped with 50µL of 3M

Sodium Hydroxide (NaOH).

Alkaline phosphatase activity was determined by measuring light absorption at a

wavelength of 405nm using the same spectrophotometer as used in the proliferation

assays (Dynex MRX, Dynex Technologies, Chantilly, VA, USA). Three repeat

measurements were obtained at one-minute intervals to confirm the reaction had

been stopped with the addition of the Sodium Hydroxide.

11.2.6 Statistical Analysis

Triplicate samples from each culture well (stimulated and controls) were taken for

the proliferation and differentiation light absorbance tests. All experimentation

procedures were repeated 3 times; therefore when the triplicate samples from each

culture well were averaged to one value, each value was the mean and standard error

188


of the mean of 18 results (6 cultures x 1 average of triplicate samples x 3 repeated

experiments).

These results of absorbance for proliferation and differentiation were tested for

normality using a Kolmogorov-Smirnov (K-S) goodness-of-fit test to determine use

of a parametric Student’s t-test or non-parametric Mann Whitney U Test. Stimulated

cultures were tested for significance with respect to controls for each type of test

stimulation and for both proliferation and differentiation. An analysis of variance

(non-parametric Kruskal-Wallis test) between control cultures from each stimulation

type was conducted. A null hypothesis probability less than 0.05 was considered

significant.

Error bars for the percentage change graph (Figure 11-7) were computed from the

addition of percentage errors in the original raw data presented in Figure 11-5 and

Figure 11-6.

11.3 Results

The raw data from these experiments is presented in Appendix A. The results

showed no significant difference in cell viability between the stimulated and control

cultures during the testing procedure signifying there were no elevated levels of

apoptosis in the stimulated cultures. Raw absorbance (Figure 11-5 and Figure 11-6)

variance tests between control groups in each type of stimulation showed a

significant difference between controls for each type of stimulation. Thus to facilitate

comparisons, results were presented as a percentage change with respect to controls

(Figure 11-7).

11.3.1 Proliferation

Pooled control data was non-normal for PEMF stimulation only and PEMF plus

mechanical strain stimulation data sets. Furthermore, pooled stimulated data for

mechanical strain stimulation only was also non-normal. Therefore the non-

189


parametric Mann Whitney U significance test was used to compare the means of the

control and stimulated data.

Exposure of the cultures to the indirect electrical stimulation (PEMF) only, resulted

in a significant reduction in proliferation as evidenced by a p value of 0.048. The

percentage reduction in absorbance with the stimulated culture was 13.8% (Figure

11-5). Using only mechanical strain as the stimulation for the culture resulted in a

non-significant decrease in proliferation from the stimulated culture of 3.9%.

The application of both stimulants resulted in a 6.8% increase, although this was also

non-significant with a p value of 0.10.

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

PEMF Stimulation Only MECH Stimulation Only PEMF + MECH Stimulation

Abs

orba

nce

Control Stimulated

#

Figure 11-5 The raw absorbance results for LDH measured proliferation from each method of DSD stimulation. # Represents statistical significance (p < 0.05). Error bars are +/- standard

error of the mean.

190



Pooled control data was non-normal for all data sets. All pooled stimulated data sets,

apart from PEMF stimulation only, were also non-normal. Therefore the non-

parametric Mann Whitney U significance test was used again.

Exposure of the cultures to PEMF stimulation only and the mechanical strain

stimulation only resulted in non-significant decreases in differentiation of 10.8% and

10% respectively (Figure 11-6). This trend was reversed with the application of both

stimulants resulting in a significant 11.1% increase (p < 0.04).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8


Abs

orba

nce

Control Stimulated

#

Figure 11-6 The raw absorbance results for pNPP measured differentiation from each method of DSD stimulation. # Represents statistical significance (p < 0.05). Error bars are +/- standard

error of the mean.

191


-40

-30

-20

-10

0

10

20


Perc

enta

ge C

hang

e

Proliferation Differentiation

##

Figure 11-7 Percentage change in proliferation and differentiation with respect to controls for each method of DSD stimulation. Error bars were computed from the addition of percentage

errors in the original raw data presented in Figure 11-5 and Figure 11-6. # Represents statistical significance (p < 0.05).

11.4 Discussion

A novel device capable of imparting a pulsed electromagnetic field (PEMF) and a

direct mechanical strain onto the cell culture was used to quantify the effects of these

stimulants on the development of osteoblast-like MG-63 cell cultures.

The data indicate that cells exposed to the PEMF stimulus exhibited a significantly

reduced proliferation over control cultures (Figure 11-7). However, the results did

not show a concomitant increase in the cellular differentiation as was experienced

with previous results (discussed in Chapter 6).

Although it is commonly accepted that there is a reciprocal and functionally coupled

relationship between the down-regulation of proliferation and the initiation of

expression of osteoblast phenotype markers such as alkaline phosphatase, MG-63

cells seem to exhibit a fairly heterogenous transition between these defined stages of

192


cellular maturation (Pautke et al., 2004) with a fairly immature osteoblastic

phenotype. SaOS-2 cells, on the other hand, exhibit a more mature gene expression,

which has been associated with greater cellular differentiation when undergoing

PEMF exposure (Diniz et al., 2002). This could explain the discrepancies seen

between results from this study (Figure 11-7) and that discussed in Chapter 6 (Figure

6-3).

The degree of cellular differentiation also effects the transduction of a mechanical

strain. Although osteoporotic donor cells do not show any increases in proliferation

when undergoing cyclic stretching (Rubenacker et al., 1995), healthy cells that are

initially further along the differentiation pathway before application of the strain will

respond with proliferation, a phenomenon that is reversed when cells are initially less

differentiated (Weyts et al., 2003). Very young and very old osteoblasts (osteocytes)

from rat calvaria showed no response to mechanical stretching of magnitude 4000µε

and frequency 1Hz. However, cells at differentiation levels between these two

extremes showed increases in proliferation and levels of cAMP, insulin-like growth

factor I, bone Gla protein and mineral accumulation (Mikuni-Takagaki et al., 1996).

Higher density seeding causes cultures to reach confluence too quickly, moving

beyond the log phase growth required for optimised LDH absorption measurements.

Higher confluence levels can also reduce the transduction of strain from the cell

substrate to the cell population (Winston et al., 1989).

The non-significant reduction in proliferation and differentiation of the cells when

exposed to direct mechanical stretch of the cell substrate was inconsistent with other

reports that have resulted in significant perturbations of these factors (Neidlinger-

Wilke et al., 1994; Kaspar et al., 1998; Kaspar et al., 2000; Simmons et al., 2003;

Jagodzinski et al., 2004). However, these studies were utilising much larger strains

(1%+) and lower frequency strain rates (1Hz) to elicit their responses.

It is theorised with experimental evidence that there is a resulting increase in

proliferation for high strain magnitudes and differentiation for low strain magnitudes

corresponding to the two phases of bone formation in vivo, i.e. growth of trabecular

bone and mineral apposition respectively (Burger and Veldhuijzen, 1993; Lacroix

193


and Prendergast, 2002). Therefore as strains employed in this study were of a low

magnitude, results would be expected to show an increased differentiation above

controls. Also of note is the phenomenon that uniaxial strain, which is the dominant

strain source produced from the dual stimulus device, promotes cellular

differentiation over and above equiaxial strains (Park et al., 2004).

Fritton and co-workers studying average strain histories of a dog tibia over a 24-hour

period of 5Hz stimulation (after data sampling and processing with a Fast Fourier

Transform) observed that strains on the bone shaft surface were in the range of 3µε.

This occurred on average 130,000 times within the 24 hours (Fritton et al., 2000), a

value very similar to the number of individual strains over the same period of time

used in this study, which applied 144,000 of them over the 8-hour stimulation period

per day.

However, the strains employed by the dual stimulus device and those occurring on

the shaft of the dog tibia are very different (approximately 1 order of magnitude). A

possible explanation for this discrepancy in strain could be provided by an

amplification theory initially proposed by Weinbaum (Weinbaum et al., 1994) and

refined by You (You, 2002). It predicts fluid flow within the small canaliculi of bone

will induce a membrane strain on the osteocytic processes, which reside within these

porosities. Strain is induced from the fluid drag of the bone sera on the pericellular

matrix of actin filaments, which tether the cellular process to the wall of the

canaliculi. According to this theory, applying a bone strain of 3µε at 5Hz, as

experienced by the dog tibia in the aforementioned study by Fritton (Fritton et al.,

2000), the strain on the osteocyte process membrane will be in the order of 30µε; a

value similar to the surface strain provided by the DSD (according to the theoretical

finite element analysis).

Prevention of bone loss in ovariectomized rats has been achieved with a high

frequency vibrational stimulus (Flieger et al., 1998), while increased expression of

mRNA in osteoblasts has also been observed (Tjandrawinata et al., 1997).

194


Tanaka (Tanaka et al., 2003b) using a 3-day cell loading protocol, found cultured

osteoblasts responded to broad frequency (average of 50Hz) white noise vibration

with a significant increase in production of MMP-9, a matrix metalloproteinase

expressed during osteogenesis, over and above a pure 3Hz sinusoidal strain

waveform. Furthermore, combination of the low frequency, high magnitude (3Hz,

3000µε) strain with the underlying vibration (a more physiologically realistic strain

waveform) elicited enhanced osteocalcin production after the 7-day protocol had

elapsed. This protein is a terminal differentiation marker associated with bone

formation and signifies the bone cells are being forced to differentiate.

It was hypothesised that this synergistic effect was a result of stochastic resonance,

or the ‘sensitisation’ of bone cells to larger strains (3Hz sine wave) with the

application of a vibrational stimulus (Tanaka et al., 2003a). The stimulus used in the

original study by Tanaka (Tanaka et al., 2003b) was consistent with the addition of

physiological levels of rarely occurring, high magnitude strains in bone with the very

commonly occurring, low magnitude strains originating from muscle attachment

(Fritton et al., 2000). It could then be argued that this strain signal is more

‘physiological’ in nature and therefore more potent, yielding positive results.

In terms of the mechano-transduction of this signal, it is conceivable that the

proposed viscoelastic hardening of the bone cells in response to high frequency

strains (Warden and Turner, 2004) may increase the magnitude of the low frequency

strains directly transferred to the cells via their focal adhesions (Pavalko et al.,

2003b). However, it has also been noted that an increasing strain rate increases the

streaming potential current in bone, which could also be a transduction factor for the

applied stimuli (Gross and Williams, 1982).

The surface treated silicone used as the cell substrate for all cultures transfers the

applied mechanical stretch directly to the adherent cells via integrins (Banes et al.,

2001; Cavalcanti-Adam et al., 2002). These receptors are vitally important in

attachment and the transduction of mechanical strain signals (Carvalho et al., 1998;

Pavalko et al., 2003b). The form of attachment to the surface effects cellular

morphology, which is also directly mediated by the applied mechanical strain

(Akhouayri et al., 1999). Morphology can effect the strain sensing ability of the cells

195


as has been observed by Winston and co-workers who concluded that a rounded

morphology exhibits no membrane strain when attached to a periodically straining

substrate, however this is reversed if the cells are fully flattened (Winston et al.,

1989). Cells in this study were observed to appear flattened and closely attached to

the surface as evidenced by the inability to differentiate between cells and surface

topography during observation and would not have been affected by this.

As mechanical strain alone did not elucidate significant results, it could be argued the

effects seen from the two stimuli together was purely a result of the PEMF stimulus.

However, a significant increase in differentiation (alkaline phosphatase expression)

above the control that was observed with both stimuli was not seen with either of the

individual stimuli, providing a unique result. Therefore, it is probable a subtle

hierarchical cellular mechanism of action is taking place during this stimulus, such

that the cell is ‘sensitised’ by either the PEMF or mechanical strain followed by the

‘active’ stimulus potentiating the cellular results seen. In vivo studies of the

relationship between mechanical movement and electrical stimulation have

concluded the mechanical stimulant as dominate over the PEMF (Spadaro, 1997)

with the combination of both enhancing average bone formation by 44% in

intramedullary wire implants placed in the femur of rabbits (Spadaro et al., 1990).

11.5 Conclusion

In summary, a newly developed dual stimulus device capable of imparting a dynamic

surface strain and pulsed electromagnetic field has been experimentally validated

with osteoblast-like MG-63 cell cultures. The results using this device found that

with either the PEMF or the mechanical strain alone both proliferation and

differentiation were inhibited. However, when both stimuli were applied this was

reversed with proliferation and differentiation being enhanced, implying some level

of biological synergy.

196

Chapter 12: Discussion, Future Work and Conclusions

12 Discussion, Future Work and

Conclusions

This project aimed to determine if cultures of osteoblasts exposed to mechanical and

electrical stimuli, individually and in unison, exhibited enhanced cellular

development. This aim covered the design and development of a novel dual stimulus

device (DSD) capable of applying a non-invasive electrical stimulant and mechanical

strain via cell substrate stretch to an osteoblast cell culture.

During the calibration of the DSD it was observed that experimental and theoretical

determinations of the cell substrate surface strain when undergoing deformation were

not equivalent. The deformable silicone membrane used as the cell substrate was

observed to exhibit inhomogeneity in thickness and anisotropic strains when

undergoing large-scale deformations. In addition, frictional effects and the

misalignment of the actuator due to design tolerances could have been some of the

confounding factors. The inhomogeneity would also affect the small strains induced

by the DSD and hence it was not possible to fully characterise the strains

experienced by the cells. However, it was clear that strains in the order of those

designed for were produced, therefore tests were conducted to assess the effect of

dual stimulation and to evaluate if further refinement of the device would be worth

pursuing.

Cell substrate membranes were surface treated to maximise cell attachment and

growth before being placed in the device for biological testing. This was undertaken

with gas plasma, which etches the surface replacing hydrophobic entities with

hydrophilic hydroxyl groups. A post-plasma treatment protocol involving soaking

the cell substrate in growth media supplemented with serum for two days prior to

biological testing was used to maximise cell attachment and growth.

The influence of PEMF stimulation timing on in vitro cell cultures has been poorly

defined in previous literature. Therefore, the experiments discussed in Chapter 6

were designed to focus on a particular ‘window’ of time whereby significant changes

197


in the development of osteoblast cultures occurred. Four protocols, each with 24

hours of PEMF stimulation, were used. Three of the protocols involved one 24 hour

block of stimulation on the first, second and third day of the protocol while the last

employed 8 hours of PEMF stimulation per day on each of the 3 days.

No significant differences were seen between the protocols, however the 8 hours a

day protocol showed the most consistent response of cell proliferation and

differentiation over the repeated tests. Thus it was subsequently used in further

experiments with the dual stimulus device (DSD).

The DSD was evaluated using MG-63 osteoblast-like cell cultures with 3 different

types of stimulus; firstly a PEMF stimulus, secondly a mechanical stimulus and

finally the two stimuli in unison together. Tests were repeated three times to verify

results. In agreement with the data presented in Chapter 6, the PEMF stimulus alone

significantly reduced proliferation in comparison to the control cultures. However the

results did not show a concomitant increase in alkaline phosphatase as was originally

noted in Chapter 6.

The MG-63 cell line used with the DSD exhibits a more immature phenotype than

the SaOS-2 cells used in the initial studies (Pautke et al., 2002) thus expressing

alkaline phosphatase to a lesser degree than the SaOS-2 cells, possibly explaining the

differences in results.

The combination of the two stimuli on the cell cultures significantly increased

cellular differentiation above controls. This data contrasts with that for the PEMF

only and mechanical stimulation only, where both tests decreased cellular

differentiation with exposure to their respective stimulants. It is probable a subtle

hierarchical cellular mechanism of action is taking place during this stimulus. Some

in vivo studies of the relationship between mechanical movement and electrical

stimulation have concluded the mechanical stimulant is dominant over the PEMF

(Spadaro, 1997), noting that a combination of both can enhance average bone

formation by 44% in intramedullary wire implants placed in the femur of rabbits

(Spadaro et al., 1990).

198


Also, it has been observed that both stimulants create an increase in cytosolic Ca2+

and a subsequent increase in activated cytoskeletal calmodulin (Brighton et al.,

2001). Others have claimed that the mechanism of action for both stimulants is based

on the time varying electric field created from either the mechanically induced

streaming potentials past bone cells or the applied PEMF (Pilla, 2002b).

Due to the novel nature of this project, results must be considered as groundwork for

further research and are only indicative of the trends in bone cell development when

exposed to the mechanical and electrical stimulants. Key aspects of this research will

need to be covered more comprehensively (outlined in Section 12.1) before any firm

conclusions can be drawn.

12.1 Future Work

This project has provided novel data on the biophysical stimulation of bone cells.

PEMF stimulation parameters used for the dual stimulus device were based on

clinical devices used for augmentation of bone healing and were not derived from

recorded streaming potentials in bone. It is acknowledged that experimentation with

cells exposed to this style of electrical stimulus may offer a different response to

those in this study. Greater control over the magnitude, location and waveform of the

electrical stimulus would also benefit.

The level of influence streaming potentials play in the in vivo mechanotransduction

process is yet to be defined. Therefore induction of the exact specifications of these

in vivo fields either via in vitro fluid flow or an exogenous electrical current could

provide pertinent data on this process.

It would also prove very insightful if both the mechanical strain and electrical

stimulus parameters were varied with a particular focus on analysing fracture-healing

environments as would be experienced by the osteoblast-like cells in vivo. This

would entail increasing the level of strain, maintaining high and low frequency

components and varying environmental pressure. Culturing the cells in a three

dimensional scaffold to conduct the tests would also provide a more realistic

199


environment. Also, as the principal strain vectors within the fracture callus vary

according to the position of the cells, these could also be varied in vitro.

It has been widely accepted that electrical currents are set up in wounds, such as

fractures, between the wound edge and the surrounding tissue. It is proposed that

these currents augment or even initiate invasion of blood vessels and galvanotaxis

(electrically induced migration) of the bone forming cells. Therefore this stimulus

could well play a contributing factor to the fracture healing process and would be a

worthy study to undertake.

The dual stimulus device provided many design challenges during its development.

While it was designed and produced with the best available materials and design

criteria at the time, in hindsight there is scope for a review of some of its features. A

change in the method of attachment (silicone adhesive) between the Piezoelectric

Actuator and the DSD and that between the Indentor and the actuator would reduce

the potential for misalignment. It had been anticipated that a greater output

displacement from the actuator when assembled in the DSD would have resulted,

providing a greater available strain range for experimental testing. Thus an improved

procedure involving a spatially defined and reliable mechanical attachment designed

to maintain the maximum actuator displacement could be produced. An example

could include the use of an appropriately dimensioned cylindrical annulus ring, used

to clamp the peripheral edge of the actuator while also aligning the dynamically

moving Indentor with the central axis of the actuator.

The original dual stimulus device design uses an outer ring platen that effectively

dragged the cell substrate membrane over the lower edges of the cell culture wall

creating an equi-biaxial strain. Although all theoretical design issues that had been

highlighted as potential sources of error were taken into account before manufacture

and assembly, this method of mechanical strain did not perform correctly and

showed inhomogeneous surface strain. However, this method of cell substrate

surface strain is still a unique and viable technique for further in vitro cell culture

experimentation. It is only the restrictive nature of the thesis timeline and funding

that has impeded success with this design.

200


Assumed sources of the inhomogeneous strain are any/all of the following:

1. Anisotropic silicone elastomer mechanical properties,

2. Increase in frictional forces between the cellular growth medium and the

lower edges of the cell culture wall due to increased loading frequency, and

3. The aforementioned misalignment of the Indentor due to the method of

attachment with piezoelectric actuator.

Solutions to the third point are made above. A possible method of rectifying point

two could involve the individual forming of each sheet of silicone elastomer as

opposed to the purchase of pre-formed sheets, which would have introduced the

anisotropic characteristics seen. This process would have the potential to afford

greater control over factors such as membrane thickness, surface topography and

surface chemistry. This leaves point one which is fundamentally a frictional problem.

Increasing the frequency of movement between the membrane and its adjacent

surface on the dual stimulus device may result in the protein-rich cellular growth

medium ‘hardening’ and subsequently resisting movement. One way of overcoming

this problem would be dynamically moving a flat platen, slightly smaller than the

outer diameter of the culture well, underneath the cell substrate thus applying a

biaxial stretch across the membrane surface. The platen could be lubricated with an

air barrier, provided through a port drilled in the centre that is connected to a

constant air pressure source. As membrane deformations are small and air does not

show viscous hardening, this design could potentially overcome the inherent friction

problems experienced.

Linear strain testing of the membrane (whereby an equal strain level is induced

across the entire surface) with equally spaced ink dots would clarify if surface texture

affects strains between each pair of dots and confirm whether or not this is an

influential factor in measurement of surface strains. Surface treatment of the cell

substrate could include coating the membrane with specific proteins, which provide

specific binding sites for the attachment and spreading of cells. This could be

conducted in addition to the plasma treatment, which renders the surface hydrophilic.

Experimental support for this idea is provided by Yang and co-workers who have

shown that the hydrophilicity of a surface is less influential on cell attachment and

spreading than the presence of specific binding sites (Yang et al., 2004). Subsequent

201


to this application of a protein layer on the cell substrate surface would be a proper

study on the structure, location, constitution and density of the protein layer

adsorbed. Ideally this project should have also included this step after the application

of the post-plasma soaking treatment in media.

Finally, the experimental design could have included control studies using static

magnetic fields of similar magnitude to the peak magnetic field created via the

PEMF. Data from these studies could have possibly confirmed that induced electric

fields from the PEMF and not the magnetic field was the mediating factor for cellular

development (Pilla et al., 1993). Other variations in the applied mechanical stimulus

such as strain frequency, strain magnitude and different cell types could have been

performed to clarify current understanding of the influence electrical and mechanical

stimulants possess over bone cell development.

It is acknowledged that the novel results presented in this thesis are preliminary and

require further clarification before solid recommendations can be made, yet

elucidation of the interactions between mechanical and electrical stimulants on bone

cell development in vitro has relevance to an in vivo environment. Clearly it is not

practical for in vitro experimentation to attempt to recreate the exact in vivo

environment experienced by the bone cells, and therefore a comprehensive

determination of the interactions between mechanical and electrical stimulants will

take time to be revealed. However, isolating synergistic mechanisms of action in

vitro will set the groundwork to help guide future researchers focus upon the most

pertinent areas of study.

12.2 Conclusions

A novel device was designed and built to impart both an electrical and mechanical

stimulus on bone cells in culture.

A PEMF stimulus on monolayers of SaOS-2 and MG-63 osteoblast-like cells leads to

a depression in proliferation. A concomitant increase in alkaline phosphatase

production was also observed for the SaOS-2 cultures, but not for the MG-63 cell

202


line. It was hypothesised that this was due to the MG-63’s lack of phenotypic

maturity compared to the SaOS-2 cells. Using the MG-63 cells, mechanical strain of

the cell substrate at a relatively high frequency (5Hz) but small strain, did not

significantly effect either cell proliferation or differentiation. When the electrical and

mechanical stimulants were combined, cultures of MG-63 cells exhibited a

significant increase in cellular differentiation, revealing a possible synergistic effect

of these two stimulants on the development of bone cells.

These results warrant further development of the dual stimulus device to enable

better characterisation of the synergy between the two stimulants.

203

Appendix A: Raw Data from In Vitro Cellular Experimentation

Appendix A: Raw Data from In Vitro

Cellular Experimentation

1. PEMF Stimulation of Cultured Bone Cells

(Chapter 6)

PROLIFERATION: 3H-Leucine Incorporation (Counts Per Minute) Seeding

Density PROTOCOL 1 PROTOCOL 2 PROTOCOL 3 PROTOCOL 4

25,000 1124 1527 855 820 CONTROL

(n=24) 50,000 1614 2052 1250 1239

25,000 1027 1355 796 740 STIMULATED

(n=24) 50,000 1395 2027 1224 1020

DIFFERENTIATION: Alkaline Phosphatase (Units/Litre)

Seeding

Density PROTOCOL 1 PROTOCOL 2 PROTOCOL 3 PROTOCOL 4

25,000 15.24 19.38 16.55 15.71 CONTROL

(n=24) 50,000 19.49 37.62 22.13 25.59

25,000 17.62 23.42 19.62 19.23 STIMULATED

(n=24) 50,000 21.84 43.05 30.68 25.04

204

Appendix A: Raw Data from In Vitro Cellular Experimentation

2. Dual Stimulation of Cultured Bone Cells

(Chapter 11)

2.1 PEMF Stimulation only PROLIFERATION: Lactate

Dehydrogenase (Light Absorbance)

DIFFERENTIATION: Alkaline

Phosphatase (Light Absorbance)

CONTROL

(n = 18) 1.37 0.5

STIMULATED

(n = 18) 1.18 0.45

2.2 Mechanical Stimulation Only PROLIFERATION: Lactate




CONTROL

(n = 18) 1.45 0.70

STIMULATED

(n = 18) 1.40 0.63

2.3 PEMF and Mechanical Stimulation PROLIFERATION: Lactate




CONTROL

(n = 18) 1.15 0.49

STIMULATED

(n = 18) 1.23 0.55

205

Appendix B: Dual Stimulus Device (DSD) Engineering Drawings

Appendix B: Dual Stimulus Device

(DSD) Engineering Drawings

1. Base; General View (Figure A - 1)

2. Base (Figure A - 2)

3. Base; Detailed Drawing (Figure A - 3)

4. Indentor; General View (Figure A - 4)

5. Indentor (Figure A - 5)

6. Central Pin Insert (Figure A - 6)

7. Spacer Ring (Figure A - 7)

8. Cell Substrate Annulus (Figure A - 8)

9. Top; General View (Figure A - 9)

10. Top (Figure A - 10)

11. PEMF Coil Former (Figure A - 11)

12. Lid (Figure A - 12)

206


Figure A - 1 Base; General View

207


Figure A - 2 Base

208


Figure A - 3 Base; Detailed Drawing

209


Figure A - 4 Indentor; General View

210


Figure A - 5 Indentor

211


Figure A - 6 Central Pin Insert

212


Figure A - 7 Spacer Ring

213


Figure A - 8 Cell Substrate Annulus

214


Figure A - 9 Top; General View

215


Figure A - 10 Top

216


Figure A - 11 PEMF Coil Former

217


Figure A - 12 Lid

218

Appendix C: Gas Plasma Cell Substrate Surface Modification Procedure

Appendix C: Gas Plasma Cell Substrate

Surface Modification Procedure 1. Check vacuum meter current level and turn on

2. Check reflected power meter and turn on

3. Set power supply to desired level and leave turned off.

4. Pour a small volume of plain tap water into H2O chamber and correctly seal to

vacuum tube with tap closed

5. Turn rotary pump on after checking to confirm all taps and valves are shut

6. Allow vacuum level to reach approximately 0.5 torr

7. Open needle valve of H2O chamber very slightly until degassing (bubbling) is

observed -- continue until all bubbling has ceased

8. Close H2O chamber needle valve to maintain vacuum

Preparation and Placement of PDMS Sample

9. Bleed air into vacuum chamber with valve and open chamber

10. Place PDMS sample (on glass block), with surface to be etched upright, in centre

of radio frequency signal generator coil

11. Seal up vacuum chamber and allow vacuum to reach desired level

12. Once reached, very carefully open needle valve of H2O chamber until vacuum

level on meter holds steady at desired level

13. Set countdown timer to desired length for surface modification

14. Turn power supply and radio frequency signal generator on and start countdown

timer

15. After time has elapsed immediately turn off the power supply and close H2O

chamber needle valve.

Removing Treated PDMS Sample

16. Repeat step number 9

17. Remove the treated sample and immediately place into a sealed sample bag

making sure to remove as much air from the inside of the bag as possible before

closing

18. Repeat steps 10 onwards for further sample treatments

219

Appendix D: Finite Element Analysis Results

Appendix D: Finite Element Analysis

Results

As discussed in Section 10.2, iterations of the experimentally derived and

theoretically questionable computational values used in cell substrate strain finite

element analyses were computed in an attempt to explain deviations between

experimental and theoretical surface strain results. Results for each iteration are

displayed below. For all iterations, other values were held constant

1. Young’s Modulus = 1.75MPa (Figure C - 1)

2. Young’s Modulus = 2MPa (Figure C - 2)



5. Poisson’s Ratio = 0.35 (Figure C - 5)

6. Poisson’s Ratio = 0.5 (Figure C - 6)

7. Cell Substrate Thickness = 20µm (Figure C - 7)





12. Initial Central Pin Displacement = 10µm (Figure C - 12)




16. Central Pin Insert Displacement = 100µm (Figure C - 16)

17. Central Pin Insert Displacement = 500µm (Figure C - 17)

220


Radial Strain vs Radial Position Circumferential Strain vs Radial Position

Figure C - 1 Young’s Modulus = 1.75MPa

221



Figure C - 2 Young's Modulus = 2MPa

222



Figure C - 3 Young's Modulus = 2.25MPa

223



Figure C - 4 Young's Modulus = 2.5MPa

224



Figure C - 5 Poisson's Ratio = 0.35

225



Figure C - 6 Poisson's Ratio = 0.5

226



Figure C - 7 Cell Substrate Thickness = 20µm

227




228




229




230




231



Figure C - 12 Initial Central Pin Displacement = 10µm

232




233




234




235



Figure C - 16 Central Pin Insert Displacement = 100µm

236



Figure C - 17 Central Pin Insert Displacement = 500µm

237

Appendix E: Research Presentations and Published Material

Appendix E: Research Presentations and

Published Material

Journal Papers:

Hannay, G. G., D. I. Leavesley and M. J. Pearcy. (2005). "Timing of pulsed

electromagnetic fields does not affect bone cell development." Bioelectromagnetics

26(8): 670-676.

Research Presentations:

Hannay, G. G (2005). Mechanical and Electrical Environments to Stimulate Bone

Cell Development. Australian and New Zealand Orthopaedic Research Society 11th

Annual Scientific Meeting, 6-8th October, Perth, Australia (Oral Presentation)

Hannay, G. G., D. I. Leavesley and M. J. Pearcy. (2004). Electrical Environments to

Stimulate Bone Cell Development. The Annual Conference of Engineering and the

Physical Sciences in Medicine, 14-18th November, Geelong, Australia (poster)

Hannay, G. G., D. I. Leavesley and M. J. Pearcy. (2004). Electrical Environments to

Stimulate Bone Cell Development. The 11th Meeting of the Combined Orthopaedic

Associations, 24-29th October, Sydney, Australia (poster)

Hannay, G. G., D. I. Leavesley and M. J. Pearcy. (2004). Exogenous PEMF

Stimulation Promotes Osteoblast Development. Australian Society for Medical

Research Postgraduate Student Conference, 11th June, Brisbane, Australia (poster)

Hannay, G. G., D. I. Leavesley and M. J. Pearcy. (2003). Clinical Pulsed

Electromagnetic Fields Promote Osteoblast Cell Development Independently of

Exposure Timing. 6th International Conference on Cellular Engineering, 20-22nd

August, Sydney, Australia (poster)

238

Appendix E: Research Presentations and Published Material

239

halla

This article is not available online. Please consult the hardcopy thesis available from the QUT Library

References

References

Aaron, R. K. (1994). "Treatment of osteonecrosis of the femoral head with electrical

stimulation." Instr Course Lect 43: 495-498.

Adachi, T., K. Tsubota, et al. (2001). "Trabecular Surface Remodelling Simulation

for Cancellous Bone Using Microstructural Voxel Finite Element Models." J

Biomech Eng 123(5): 403-409.

Adair, R. K. (1991). "Constraints on biological effects of weak extremely low

frequency electromagnetic fields." Phys Rev A 43: 1039-1048.

Adair, R. K. (1998). "Extremely low frequency electromagnetic fields do not interact

directly with DNA." Bioelectromagnetics 19(2): 136-137.

Adams, D. J., A. A. Spirt, et al. (1997). "Testing the daily stress stimulus theory of

bone adaptation with natural and experimentally controlled strain histories." J

Biomech 30(7): 671-678.

Ajubi, N. E., J. Klein-Nulend, et al. (1996). "Pulsating fluid flow increases

prostaglandin production by cultured chicken osteocytes--a cytoskeleton-

dependent process." Biochem Biophys Res Commun 225(1): 62-8.

Akhouayri, O., M. H. Lafage-Proust, et al. (1999). "Effects of static or dynamic

mechanical stresses on osteoblast phenotype expression in three-dimensional

contractile collagen gels." J Cell Biochem 76(2): 217-30.

Anselme, K. (2000). "Osteoblast adhesion on biomaterials." Biomaterials 21(7): 667-

81.

240

References

Antov, Y., A. Barbul, et al. (2004). "Electroendocytosis: stimulation of adsorptive

and fluid-phase uptake by pulsed low electric fields." Exp Cell Res 297(2):

348-62.

Arai, T. and W. Norde (1990). "The behaviour of some model proteins at solid-liquid

interfaces 2. Sequential and competitive adsorption." Colloids and Surfaces

51: 17-28.

Armour, K. E., K. J. Armour, et al. (2001). "Defective bone formation and anabolic

response to exogenous estrogen in mice with targeted disruption of

endothelial nitric oxide synthase." Endocrinology 142(2): 760-766.

Augat, P., K. Margevicius, et al. (1998). "Local Tissue Properties in Bone Healing:

Influence of Size and Stability of the Osteotomy Gap." J Orthop Res 16: 475-

81.

Bakker, A. D., K. Soejima, et al. (2001). "The production of nitric oxide and

prostaglandin E(2) by primary bone cells is shear stress dependent." J

Biomech 34(5): 671-7.

Bakker, A. D., J. Klein-Nulend, et al. (2004). "Shear stress inhibits while disuse

promotes osteocyte apoptosis." Biochem Biophys Res Commun 320(4):

1163-1168.

Banes, A. J., G. Lee, et al. (2001). "Mechanical forces and signalling in connective

tissue cells: cellular mechanisms of detection, transduction, and responses to

mechanical deformation." Curr Opin Orthop 12: 389-396.

Bassett, C. A., R. J. Pawluk, et al. (1964). "Effects of Electric Currents on Bone in

Vivo." Nature 204: 652-4.

241

References

Bassett, C. A. (1971). Biophysical principals affecting bone structure. Biochemistry

and Physiology of Bone. G. Bourne. New York, Academic Press.

Bassett, C. A., R. J. Pawluk, et al. (1974). "Acceleration of fracture repair by

electromagnetic fields. A surgically noninvasive method." Ann N Y Acad Sci

238: 242-62.

Bassett, C. A. and R. J. Pawluk (1975). "Non-invasive methods for stimulating

osteogenesis." J Biomed Mater Res 9(3): 371-4.

Bassett, C. A., A. A. Pilla, et al. (1977). "A non-operative salvage of surgically-

resistant pseudarthroses and non-unions by pulsing electromagnetic fields. A

preliminary report." Clin Orthop Relat Res (124): 128-43.

Bassett, C. A. (1978). Pulsing electromagnetic fields: a new approach to surgical

problems. Metabolic Surgery. H. Buchwald and R. L. Varcho. New York,

Grune and Stratton: 255.

Bassett, C. A., H. R. Chokshi, et al. (1979). The effect of pulsing electromagnetic

fields on cellular calcium and calcification of non unions. Electrical

Properties of Bone and Cartilage: Experimental Effects and Clinical

Applications. C. T. Brighton, J. Black and S. R. Pollack. New York, Grune

and Stratton: 427.

Bassett, C., S. Mitchell, et al. (1981). "Treatment of ununited tibial diaphyseal

fractures with pulsing electromagnetic fields." J Bone Joint Surg Am 63(4):

511-1523.

Bassett, C., S. Mitchell, et al. (1982a). "Treatment of therapeutically resistant non-

unions with bone grafts and pulsing electromagnetic fields." J Bone Joint

Surg Am 64(8): 1214-1523.

242

References

Bassett, C., M. Valdes, et al. (1982b). "Modification of fracture repair with selected

pulsing electromagnetic fields." J Bone Joint Surg Am 64(6): 888-895.

Bassett, C. A., S. N. Mitchell, et al. (1982c). "Pulsing electromagnetic field treatment

in ununited fractures and failed arthrodeses." JAMA 247(5): 623-8.

Bassett, C. A. (1982). "Pulsing electromagnetic fields: a new method to modify cell

behaviour in calcified and noncalcified tissues." Calcif Tissue Int 34(1): 1-8.

Bassett, C. A. (1983). "Biomedical implications of pulsing electromagnetic fields."

Surgical Rounds: 22-31.

Bassett, C. A. (1984). "The development and application of pulsed electromagnetic

fields for ununited fractures and arthrodeses." Orthop Clin North Am 15: 61-

87.

Bassett, C. A. (1989). "Fundamental and practical aspects of therapeutic uses of

pulsed electromagnetic fields (PEMFs)." Crit Rev Biomed Eng 17(5): 451-

529.

Bassett, C. A., M. Schink-Ascani, et al. (1989). "Effects of pulsed electromagnetic

fields on Steinberg ratings of femoral head osteonecrosis." Clin Orthop

1(246): 172-185.

Bassett, C. A. (1991). Physical and biological principals affecting weak, extremely

low frequency, electromagnetic bioresponses. Electromagnetics in medicine

and biology. C. T. Brighton and S. R. Pollack. San Francisco, San Francisco

Press Inc.

Bassett, C. A. (1995). "Why are the principles of physics and anatomy important in

treating osteoporosis?" Calcif Tissue Int 56(6): 515-6.

243

References

Basso, N. and J. N. M. Heersche (2002). "Characteristics of in vitro osteoblastic cell

loading models." Bone 30(2): 347-351.

Beck, B. R., X. Qin, et al. (2002). "On the Relationship Between Streaming Potential

and Strain in an in vivo Bone preparation." Calcif Tissue Int 71(1): 334-43

Ben-Ze'ev, A., S. R. Farmer, et al. (1980). "Protein synthesis requires cell-surface

contact while nuclear events respond to cell shape in anchorage-dependent

fibroblasts." Cell 21(2): 365-72.

Biewener, A. A. and J. E. A. Bertram (1993). Mechanical loading and bone growth

in vivo. Bone. B. K. Hall. Boca Raton, CRC Press. 7: 1-36.

Binder, A., G. Parr, et al. (1984). "Pulsed electromagnetic field therapy of persistent

rotator cuff tendinitis. A double-blind controlled assessment." Lancet

1(8379): 695-8.

Binhi, V. N., Y. D. Alipov, et al. (2001). "Effect of static magnetic field on E. coli

cells and individual rotations of ion-protein complexes." Bioelectromagnetics

22(2): 79-86.

Bloomfield, S. A. (2001). "Cellular and molecular mechanisms for the bone response

to mechanical loading." Int J Sport Nutr Exerc Metab 11(Supplement): S128-

S136.

Bodamyali, T., B. Bhatt, et al. (1998). "Pulsed electromagnetic fields simultaneously

induce osteogenesis and upregulate transcription of bone morphogenetic

proteins 2 and 4 in rat osteoblasts in vitro." Biochem Biophys Res Commun

250(2): 458-461.

Borgens, R. B. (1984). "Endogenous ionic currents traverse intact and damaged

bone." Science 225(4661): 478-82.

244

References

Bottlang, M., M. Simnacher, et al. (1997). "A cell strain system for small

homogeneous strain applications." Biomedizinische Technik. (Biomedical

Engineering) 42(11): 305-309.

Brand, R. A., C. M. Stanford, et al. (2001). "Primary adult human bone cells do not

respond to tissue (continuum) level strains." J Orthop Sci 6(3): 295-301.

Brighton, C. T., B. J. Sennett, et al. (1992). "The inositol phosphate pathway as a

mediator in the proliferative response of rat calvarial bone cells to cyclical

biaxial mechanical strain." J Orthop Res. 10(3): 385-93.

Brighton, C. T., W. Wang, et al. (2001). "Signal Transduction in Electrically

Stimulated Bone Cells." J Bone Joint Surg Am 83(10): 1514-1523.

Britland, S., H. Morgan, et al. (1996). "Synergistic and hierarchical adhesive and

topographic guidance of BHK cells." Exp Cell Res 228(2): 313-25.

Brodland, G. W., A. T. Dolovich, et al. (1992). "Pretension critically affects the

incremental strain field on pressure-loaded cell substrate membranes." J

Biomech Eng 114(3): 418-20.

Bronzino, J. D., Ed. (1995). The Biomedical Engineering Handbook. The electrical

engineering handbook series. Boca Raton, Florida, USA, CRC Press.

Brown, T. D., D. R. Pedersen, et al. (1990). "Toward an identification of mechanical

parameters initiating periosteal remodelling: a combined experimental and

analytic approach." J Biomech 23(9): 893-905.

Brown, T. D. (2000). "Techniques for mechanical stimulation of cells in vitro: a

review." J Biomech 33(1): 3-14.

245

References

Brunette, D. M. (1986). "Fibroblasts on micromachined substrata orient

hierarchically to grooves of different dimensions." Exp Cell Res 164(1): 11-

26.

Burger, E. H., J. Klein-Nulend, et al. (1991). "Modulation of osteogenesis in foetal

bone rudiments by mechanical stress in vitro." J Biomech 24(Suppl 1): 101-9.

Burger, E. H., J. Klein-Nulend, et al. (1992). "Mechanical stress and osteogenesis in

vitro." J Bone Miner Res 7(Suppl 2): S397-401.

Burger, E. H. and J. P. Veldhuijzen (1993). Influence of mechanical factors on bone

formation, resorption and growth in vitro. Bone. B. K. Hall. Boca Raton, Fl,

CRC Press. 7: 37-56.

Burger, E. H. and J. Klein-Nulend (1999). "Mechanotransduction in bone--role of the

lacuno-canalicular network." FASEB J 13(Suppl): S101-12.

Burr, D. B., C. Milgrom, et al. (1996). "In vivo measurement of human tibial strains

during vigorous activity." Bone 18(5): 405-410.

Burr, D. B., A. G. Robling, et al. (2002). "Effects of biomechanical stress on bones

in animals." Bone 30(5): 781-786.

Callister, W. D. (2005). Fundamentals of materials science and engineering : an

integrated approach. Hoboken, NJ, John Wiley & Sons.

Cane, V., P. Botti, et al. (1993). "Pulsed magnetic fields improve osteoblast activity

during the repair of an experimental osseous defect." J Orthop Res 11(5):

664-70.

Cane, V., D. Zaffe, et al. (1997). "Correlation between PEMF-exposure time and

new bone formation." Ital J Anat Embryol 102S: 22.

246

References

Cardona, M. and L. Ley (1978). Photoemission in Solids. Berlin, Springer-Verlag.

Carter, D. R., D. P. Fyhrie, et al. (1987). "Trabecular bone density and loading

history: regulation of connective tissue biology by mechanical energy." J

Biomech 20(8): 785-794.

Carter, D. R., G. S. Beaupre, et al. (1998). "Mechanobiology of skeletal

regeneration." Clin Orthop Relat Res (355 Suppl): S41-55.

Carter, D. R. and G. S. Beaupre (2001). Skeletal Function and Form:

Mechanobiology of Skeletal Development, Aging and Regeneration.

Cambridge, Cambridge University Press.

Carvalho, R. S., J. L. Schaffer, et al. (1998). "Osteoblasts induce osteopontin

expression in response to attachment on fibronectin: demonstration of a

common role for integrin receptors in the signal transduction processes of cell

attachment and mechanical stimulation." J Cell Biochem 70(3): 376-90.

Cavalcanti-Adam, E. A., I. M. Shapiro, et al. (2002). "RGD Peptides Immobilized on

a Mechanically Deformable Surface Promote Osteoblast Differentiation." J

Bone Miner Res 17(12): 2130.

Chakkalakal, D. A. and M. W. Johnson (1981). "Electrical properties of compact

bone." Clin Orthop 161: 133-45.

Chakkalakal, D. A., T. J. Mollner, et al. (1999). "Magnetic field induced inhibition of

human osteosarcoma cells treated with adriamycin." Cancer Biochem

Biophys 17(1-2): 89-98.

Chan, C.-M., T.-M. Ko, et al. (1996). "Polymer surface modification by plasmas and

photons." Surface Science Reports 24(1-2): 1-54.

247

References

Chang, K. and W. H.-S. Chang (2003). "Pulsed electromagnetic fields prevent

osteoporosis in an ovariectomized female rat model: A prostaglandin E2-

associated process." Bioelectromagnetics 24(3): 189-198.

Chang, K., W. H.-S. Chang, et al. (2003). "Effects of different intensities of

extremely low frequency pulsed electromagnetic fields on formation of

osteoclast-like cells." Bioelectromagnetics 24(6): 431-439.

Chang, W. H.-S., L.-T. Chen, et al. (2004). "Effect of pulse-burst electromagnetic

field stimulation on osteoblast cell activities." Bioelectromagnetics 25(6):

457-465.

Cheng, M. Z., G. Zaman, et al. (1997). "Enhancement by sex hormones of the

osteoregulatory effects of mechanical loading and prostaglandins in explants

of rat ulnae." J Bone Miner Res 12(9): 1424-30.

Cifkova, I., P. Lopour, et al. (1990). "Silicone rubber-hydrogel composites as

polymeric biomaterials: I. Biological properties of the silicone rubber-

p(HEMA) composite." Biomaterials 11(6): 393-396.

Claes, L. E., C. A. Heigele, et al. (1998). "Effects of mechanical factors on the

fracture healing process." Clin Orthop Relat Res 355(Supp): S132-47.

Claes, L. E. and C. A. Heigele (1999). "Magnitudes of local stress and strain along

bony surfaces predict the course and type of fracture healing." J Biomech

32(3): 255-266.

Claes, L. E., K. Eckert-Hubner, et al. (2002). "The effect of mechanical stability on

local vascularization and tissue differentiation in callus healing." J Orthop

Res 20(5): 1099-1105.

248

References

Cleary, S. F. (1993). "A review of in vitro studies: low-frequency electromagnetic

fields." Am Ind Hyg Assoc J 54(4): 178-85.

Cook, I. and C. A. Bassett (1983). "Effects of tissue type and orientation of

electromagnetically induced voltage." Orthop. Trans. 7: 361.

Cowin, S. C., S. Weinbaum, et al. (1995). "A case for bone canaliculi as the

anatomical site of strain generated potentials." J Biomech 28(11): 1281-1297.

Cowin, S. C. (1999). "Bone poroelasticity." J Biomech 32(3): 217-38.

Cowin, S. C., Ed. (2001). Bone Mechanics Handbook. Boca Raton, CRC Press.

Cruess, R. L., K. Kan, et al. (1983). "The effect of pulsing electromagnetic fields on

bone metabolism in experimental disuse osteoporosis." Clin Orthop Relat Res

(173): 245-50.

Cundy, P. J. and D. C. Paterson (1990). "A ten-year review of treatment of delayed

union and non-union with an implanted bone growth stimulator." Clin Orthop

(259): 216-22.

Dallas, S. L., G. Zaman, et al. (1993). "Early strain-related changes in cultured

embryonic chick tibiotarsi parallel those associated with adaptive modelling

in vivo." J Bone Miner Res 8(3): 251-9.

Davies, P. F., T. Mundel, et al. (1995). "A mechanism for heterogeneous endothelial

responses to flow in vivo and in vitro." J Biomech 28(12): 1553-1560.

De Loof, A. (1986). "The electrical dimension of cells: the cell as a miniature

electrophoresis chamber." Int Rev Cytol 104: 251-352.

249

References

De Mattei, M., A. Caruso, et al. (1999). "Correlation between pulsed electromagnetic

fields exposure time and cell proliferation increase in human osteosarcoma

cell lines and human normal osteoblast cells in vitro." Bioelectromagnetics

20(3): 177-182.

De Mattei, M., A. Caruso, et al. (2001). "Effects of pulsed electromagnetic fields on

human articular chondrocyte proliferation." Connect Tissue Res 42(4): 269-

279.

de Rooij, P. P., M. A. N. Siebrecht, et al. (2001). "The fate of mechanically induced

cartilage in an unloaded environment." J Biomech 34(7): 961-966.

De Witt, M. T., C. J. Handley, et al. (1984). "In vitro response of chondrocytes to

mechanical loading. The effect of short term mechanical tension." Connect

Tissue Res 12(2): 97-109.

Decker, M. L., D. M. Janes, et al. (1997). "Regulation of adult cardiocyte growth:

effects of active and passive mechanical loading." Am J Physiol 272(6 Pt 2):

H2902-18.

Dennis, R. G., T. Goodwin, et al. (2003). "Effect of low-level time-varying magnetic

fields on cell proliferation, metabolism, and gene expression in vitro. (in

preparation)." (See http://www-

personal.umich.edu/~bobden/nasa_collaborations.html).

Diniz, P., K. Shomura, et al. (2002). "Effects of pulsed electromagnetic field (PEMF)

stimulation on bone tissue like formation are dependent on the maturation

stages of the osteoblasts." Bioelectromagnetics 23(5): 398-405.

Djamgoz, M. B. A., M. Mycielska, et al. (2001). "Directional movement of rat

prostate cancer cells in direct-current electric field: involvement of

voltagegated Na+ channel activity." J Cell Sci 114(14): 2697-2705.

250

http://www-personal.umich.edu/~bobden/nasa_collaborations.html

http://www-personal.umich.edu/~bobden/nasa_collaborations.html

References

Donahue, H. J., K. J. McLeod, et al. (1995). "Cell-to-cell communication in

osteoblastic networks: cell line-dependent hormonal regulation of gap

junction function." J Bone Miner Res 10(6): 881-889.

Donahue, H. J. (2000). "Gap junctions and biophysical regulation of bone cell

differentiation." Bone 26(5): 417-22.

Doty, S. B. (1981). "Morphological evidence of gap junctions between bone cells."

Calcif Tissue Int 33: 509-512.

Drissi, H., M. Zuscik, et al. (2005). "Transcriptional regulation of chondrocyte

maturation: potential involvement of transcription factors in OA

pathogenesis." Mol Aspects Med 26(3): 169-79.

Dworetzky, S. I., E. G. Fey, et al. (1990). "Progressive changes in the protein

composition of the nuclear matrix during rat osteoblast differentiation." Proc

Natl Acad Sci U S A 87(12): 4605-9.

EBI (2005). EBI - Products. http://www.ebimedical.com/products/index.cfm?s=0A

Accessed: 5th July 2005.

El Haj, A. J., S. L. Minter, et al. (1990). "Cellular responses to mechanical loading in

vitro." J Bone Miner Res 5(9): 923-32.

El Haj, A. J., L. M. Walker, et al. (1999). "Mechanotransduction pathways in bone:

calcium fluxes and the role of voltage-operated calcium channels." Med Biol

Eng Comput 37(3): 403-9.

Ellis, E. F., J. S. McKinney, et al. (1995). "A new model for rapid stretch-induced

injury of cells in culture: characterization of the model using astrocytes." J

Neurotrauma 12(3): 325-339.

251

References

Espinosa, L., L. Paret, et al. (2002). "Osteoclast spreading kinetics are correlated

with an oscillatory activation of a calcium-dependent potassium current." J

Cell Sci 115(19): 3837-3848.

Fanellia, C., S. Coppolaa, et al. (1999). "Magnetic fields increase cell survival by

inhibiting apoptosis via modulation of Ca2+ influx." FASEB J 13: 95-102.

Fear, E. C. and M. A. Stuchly (1998). "Biological cells with gap junctions in low-

frequency electric fields." IEEE Trans Biomed Eng 45(7): 856-866.

Ferrier, J., A. Illeman, et al. (1985). "Transient and sustained effects of hormones

and calcium on membrane potential in a bone cell clone." J Cell Physiol.

122(1): 53-8.

Ferrier, J., S. M. Ross, et al. (1986a). "Osteoclasts and osteoblasts migrate in

opposite directions in response to a constant electrical field." J Cell Physiol

129(3): 283-288.

Ferrier, J., A. Ward, et al. (1986b). "Electrophysiological responses of osteoclasts to

hormones." J Cell Physiol 128(1): 23-6.

Fini, M., R. Cadossi, et al. (2002). "The effect of pulsed electromagnetic fields on

the osteointegration of hydroxyapatite implants in cancellous bone: a

morphologic and microstructural in vivo study." J Orthop Res 20(4): 756-

763.

Finkelstein, E., W. Chang, et al. (2004). "Roles of microtubules, cell polarity and

adhesion in electric-field-mediated motility of 3T3 fibroblasts." J Cell Sci

117(Pt 8): 1533-45.

Fitton-Jackson, S. (1985). Biophysical studies of pulsing magnetic field interaction

with biological systems. Interactions Between Electromagnetic Fields and

252

References

Cells. A. Chiabrera, C. Nicolini and H. P. Schwan. New York, NATO ASI.

Series A: 537.

Fitzsimmons, R. J., J. R. Farley, et al. (1989). "Frequency dependence of increased

cell proliferation, in vitro, in exposures to a low-amplitude, low-frequency

electric field: evidence for dependence on increased mitogen activity released

into culture medium." J Cell Physiol. 139(3): 586-91.

Fitzsimmons, R. J., D. D. Strong, et al. (1992). "Low-amplitude, low-frequency

electric field-stimulated bone cell proliferation may in part be mediated by

increased IGF-II release." J Cell Physiol. 150(1): 84-9.

Fitzsimmons, R. J., J. T. Ryaby, et al. (1995). "IGF-II receptor number is increased

in TE-85 osteosarcoma cells by combined magnetic fields." J Bone Miner

Res 10(5): 812-819.

Flieger, J., T. Karachalios, et al. (1998). "Mechanical Stimulation in the Form of

Vibration Prevents Postmenopausal Bone Loss in Ovariectomized Rats."

Calcif Tissue Int 63(6): 510-514.

Forwood, M. R. and C. H. Turner (1995). "Skeletal adaptations to mechanical usage:

results from tibial loading studies in rats." Bone 17(S): 197-205.

Freshney, R. I. (2000). Culture of animal cells: a manual of basic technique. New

York, Wiley Press.

Friedenberg, Z. B. and C. T. Brighton (1966). "Bioelectric potentials in bone." J

Bone Joint Surg Am 48(5): 915-23.

Friedenberg, Z. B., P. G. Roberts, Jr., et al. (1971). "Stimulation of fracture healing

by direct current in the rabbit fibula." J Bone Joint Surg Am 53(7): 1400-8.

253

References

Fritton, S. P., K. J. McLeod, et al. (2000). "Quantifying the strain history of bone:

spatial uniformity and self-similarity of low-magnitude strains." J Biomech

33(3): 317-325.

Frost, H. M. (1983). "A determinant of bone architecture - the minimum effective

strain." Clin Orthop 200: 198-225.

Frost, H. M. (2001). "From Wolff’s Law to the Utah Paradigm: Insights About Bone

Physiology and Its Clinical Applications." Anat Rec 262: 398–419.

Frykman, G. K., J. Taleisnik, et al. (1986). "Treatment of non-united scaphoid

fractures by pulsed electromagnetic field and cast." J Hand Surg [Am] 11(3):

344-349.

Gartzke, J. and K. Lange (2002). "Cellular target of weak magnetic fields: ionic

conduction along actin filaments of microvilli." Am J Physiol Cell Physiol

283(5): C1333-1346.

Geiger, J. M. (1989). Residual electric polarizations from residual mechanical strains

in bone. Mechanical Engineering, Case Western Reserve University: 259.

Goldstein, S. A. and R. E. Guldberg (1996). "Mechanical influences on trabecular

bone architecture and extracellular matrix organization during formation."

Bone 19(1): 142S.

Goodman, E. M. and A. S. Henderson (1986). "Some biological effects of

electromagnetic fields." Bioelectrochem Bioenerg 15(1): 39-55.

Goodman, R., C. A. Bassett, et al. (1983). "Pulsing electromagnetic fields induce

cellular transcription." Science 220(4603): 1283-5.

254

References

Goodman, R., J. Abbott, et al. (1987). "Transcriptional patterns in the X

chromosome of Sciara coprophila following exposure to magnetic fields."

Bioelectromagnetics 8(1): 1-7.

Goodship, A. E. and J. Kenwright (1985). "The influence of induced micromovement

upon the healing of tibial fractures." J Bone Joint Surg- British Volume 67:

650-655.

Gossling, H. R. and W. J. Krompinger (1983). "The use of fracture gap biopsy in

predicting response of non unions to electrical bone stimulation." Trans.

Bioelectr. Growth Repair Soc. 3: 33

.Gossling, H. R., R. A. Bernstein, et al. (1992). "Treatment of ununited tibial

fractures: a comparison of surgery and pulsed electromagnetic fields

(PEMF)." Orthopedics 15(6): 711-719.

Green, R. J., M. C. Davies, et al. (1999). "Competitive protein adsorption as

observed by surface plasmon resonance." Biomaterials 20(4): 385-391.

Gross, D. and W. S. Williams (1982). "Streaming potential and the

electromechanical response of physiologically-moist bone." J Biomech 15(4):

277-95.

Gross, D., L. M. Loew, et al. (1986). "Optical imaging of cell membrane potential

changes induced by applied electric fields." Biophys J 50(2): 339-48.

Gross, D. (1988). "Electromobile surface charge alters membrane potential changes

induced by applied electric fields." Biophys J 54(5): 879-84.

Guizzardi, S., M. Di Silvestre, et al. (1994). "Pulsed electromagnetic field

stimulation on posterior spinal fusions: a histological study in rats." J Spinal

Disord 7(1): 36-40.

255

References

Guldberg, R. E., N. J. Caldwell, et al. (1997). "Mechanical Stimulation of Tissue

Repair in the Hydraulic Bone Chamber." J Bone Miner Res 12(8): 1295-

1302.

Hannay, G., D. Leavesley, et al. (2005). "Timing of pulsed electromagnetic fields

does not affect bone cell development." Bioelectromagnetics 26(8): 670-676.

Harada, S., J. A. Nagy, et al. (1994). "Induction of vascular endothelial growth factor

expression by prostaglandin E2 and E1 in osteoblasts." J Clin Invest 93(6):

2490-6.

Harrigan, T. P. and J. J. Hamilton (1993). "Bone strain sensation via transmembrane

potential changes in surface osteoblasts: Loading rate and microstructural

implications." J Biomech 26(2): 183-200.

Hart, F. X. (1987). "Pulse shape distortion by tissue." J. Bioelect. 6: 93.

Hart, F. X. (1996). "Cell culture dosimetry for low-frequency magnetic fields."


Hart, R. T. (1984). "A computational methods for stress analysis of adaptive elastic

materials with a view toward applications in strain-induced bone

remodelling." J Biomech Eng 106: 342-350.

Hatton, J. P., M. Pooran, et al. (2003). "A Short Pulse of Mechanical Force Induces

Gene Expression and Growth in MC3T3-E1 Osteoblasts via an ERK 1/2

Pathway." J Bone Miner Res 18(1): 58.

Heinrich, T. and R. A. Lunderstaedt (2001). "Quantification of mechanical properties

of human skin in vivo." Proceedings of SPIE - The International Society for

Optical Engineering 4472: 11-20.

256

References

Hirata, M., K. Kusuzaki, et al. (2001). "Drug resistance modification using pulsing

electromagnetic field stimulation for multidrug resistant mouse osteosarcoma

cell line." Anticancer Research 21(1A): 317-20.

Hodgkin, A. L. and A. F. Huxley (1952). "A quantitative description of membrane

current and its application to conduction and excitation in nerve." Journal of

Physiology 117: 500.

Hodgkinson, G. G. (2001). Pulsed Electromagnetic Field effects on Osteoblast-like

Cell Cultures. School of Mechanical, Manufacturing and Medical

Engineering. Brisbane, Queensland University of Technology: 85.

Horbett, T. A. and M. B. Schway (1988). "Correlations between mouse 3T3 cell

spreading and serum fibronectin adsorption on glass and

hydroxyethylmethacrylate-ethylmethacrylate copolymers." J Biomed Mater

Res 22(9): 763-93.

Horbett, T. A. (2003). Biological Activity of Adsorbed Proteins. Biopolymers at

interfaces. M. Malmsten. New York, Marcel Dekker: 393-413.

Hsieh, Y. F. and C. H. Turner (2001). "Effects of loading frequency on mechanically

induced bone formation." J Bone Miner Res 16(5): 918-24.

Huiskes, R., W. D. van Driel, et al. (1997). "A biomechanical regulatory model for

periprosthetic fibrous-tissue differentiation." J Mater Sci Mater Med 8(12):

785-788.

Huiskes, R., R. Ruimerman, et al. (2000). "Effects of mechanical forces on

maintenance and adaptation of form in trabecular bone." Nature 405(6787):

704-6.

257

References

Hung, C. T. and J. L. Williams (1994). "A method for inducing equi-biaxial and

uniform strains in elastomeric membranes used as cell substrates." J Biomech

27(2): 227-232.

Hung, C. T., S. R. Pollack, et al. (1995). "Real time calcium response of cultured

bone cells to fluid flow." Clinical Orthopaedics 313: 256-269.

Ieran, M., S. Zaffuto, et al. (1990). "Effect of low frequency pulsing electromagnetic

fields on skin ulcers of venous origin in humans: a double-blind study." J

Orthop Res 8(2): 276-82.

Jacobs, C. R., J. C. Simo, et al. (1997). "Adaptive bone remodelling incorporating

simultaneous density and anisotropy considerations." J Biomech 30(6): 603-

13.

Jacobs, C. R., C. E. Yellowley, et al. (1998). "Differential effect of steady versus

oscillating flow on bone cells." J Biomech 31(11): 969-76.

Jagodzinski, M., M. Drescher, et al. (2004). "Effects of cyclic longitudinal

mechanical strain and dexamethasone on osteogenic differentiation of human

bone marrow stromal cells." Eur Cell Mater 7: 35-41.

Joldersma, M., J. Klein-Nulend, et al. (2001). "Estrogen enhances mechanical stress-

induced prostaglandin production by bone cells from elderly women." Am J

Physiol - Endocrinology & Metabolism 280(3): E436-42.

Kahanovitz, N., S. P. Arnoczky, et al. (1994). "The effect of electromagnetic pulsing

on posterior lumbar spinal fusions in dogs." Spine 19(6): 705-709.

Kamioka, H., T. Honjo, et al. (2001). "A three-dimensional distribution of osteocyte

processes revealed by the combination of confocal laser scanning microscopy

and differential interference contrast microscopy." Bone 28(2): 145-149.

258

References

Kaspar, D., W. Seidl, et al. (1998). "Physiological dynamic strain amplitudes

increase human osteoblast proliferation but decrease osteocalcin synthesis in

vitro." J Biomech 31(1): 171.

Kaspar, D., W. Seidl, et al. (2000). "Dynamic cell stretching increases human

osteoblast proliferation and CICP synthesis but decreases osteocalcin

synthesis and alkaline phosphatase activity." J Biomech 33(1): 45-51.

Kaspar, D., W. Seidl, et al. (2002). "Proliferation of human-derived osteoblast-like

cells depends on the cycle number and frequency of uniaxial strain." J

Biomech 35(7): 873-880.

Keller, R. (2002). "Shaping the Vertebrate Body Plan by Polarized Embryonic Cell

Movements." Science 298(5600): 1950-1954.

Klein-Nulend, J., J. P. Veldhuijzen, et al. (1986). "Increased calcification of growth

plate cartilage as a result of compressive force in vitro." Arthritis &

Rheumatism 29(8): 1002-9.

Klein-Nulend, J., C. M. Semeins, et al. (1995a). "Pulsating fluid flow increases nitric

oxide (NO) synthesis by osteocytes but not periosteal fibroblasts--correlation

with prostaglandin upregulation." Biochem Biophys Res Commun 217(2):

640-8.

Klein-Nulend, J., A. van der Plas, et al. (1995b). "Sensitivity of osteocytes to

biomechanical stress in vitro." FASEB Journal 9(5): 441-5.

Kletsas, D., E. K. Basdra, et al. (2002). "Effect of protein kinase inhibitors on the

stretch-elicited c-Fos and c-Jun up-regulation in human PDL osteoblast-like

cells." J Cell Physiol 190(3): 313-321.

259

References

Kloth, L. C. and J. M. McCulloch (1996). "Promotion of wound healing with

electrical stimulation." Adv Wound Care 9(5): 42-5.

Kufahl, R. H. and S. A. Saha (1990). "A theoretical model for stress generated flow

in the canaliculi-lacunae network in bone tissue." J Biomech 23: 171-180.

Kunnel, J. G. (2002). Micromechanical testing of viable bone. Department of

Biomedical Engineering. Chicago, University of Illinois: 72.

Lacroix, D. and P. J. Prendergast (2002). "A mechano-regulation model for tissue

differentiation during fracture healing: analysis of gap size and loading." J

Biomech 35(9): 1163-1171.

Lamerigts, N. M. P., P. Buma, et al. (2000). "Incorporation of morsellized bone graft

under controlled loading conditions. A new animal model in the goat."

Biomaterials 21(7): 741-747.

Lanyon, L. E., W. G. Hampson, et al. (1975). "Bone deformation recorded in vivo

from strain gauges attached to the human tibial shaft." Acta Orthopaedica

Scandinavica 46(2): 256-68.

Lanyon, L. E. (1984). "Functional strain as a determinant for bone remodelling."

Calcif Tissue Int 36(Suppl 1): S56-61.

Lanyon, L. E. and C. T. Rubin (1984). "Static vs dynamic loads as an influence on

bone remodelling." J Biomech 17(12): 897-905.

LaPlaca, M. C. and L. E. Thibault (1997). "An in vitro traumatic injury model to

examine the response of neurons to a hydrodynamically-induced

deformation." Ann Biomed Eng 25(4): 665-677.

260

References

Lateef, S. S., S. Boateng, et al. (2002). "GRGDSP peptide-bound silicone

membranes withstand mechanical flexing in vitro and display enhanced

fibroblast adhesion." Biomaterials 23(15): 3159-3168.

Lazarowski, E. R., L. Homolya, et al. (1997). "Direct demonstration of mechanically

induced release of cellular UTP and its implication for uridine nucleotide

receptor activation." J Biol Chem 272(39): 24348-24354.

Le, A. X., T. Miclau, et al. (2001). "Molecular aspects of healing in stabilized and

non-stabilized fractures." J Orthop Res 19(1): 78-84.

Lee, J. H., J. W. Park, et al. (1991). "Cell adhesion and growth on polymer surfaces

with hydroxyl groups prepared by water vapour plasma treatment."

Biomaterials 12(5): 443-448.

Lee, R. C., J. B. Rich, et al. (1982). "A comparison of in vitro cellular responses to

mechanical and electrical stimulation." American Surgeon 48(11): 567-74.

Li, S., R. S. Piotrowicz, et al. (1996). "Fluid shear stress induces the phosphorylation

of small heat shock proteins in vascular endothelial cells." The Am J Physiol

271(1): C994-1000.

Liboff, A. R., T. Williams, Jr., et al. (1984). "Time-varying magnetic fields: effect on

DNA synthesis." Science 223(4638): 818-20.

Linovitz, R. J., M. Pathria, et al. (2002). "Combined magnetic fields accelerate and

increase spine fusion: a double-blind, randomized, placebo controlled study."

Spine 27(13): 1383-1389.

Lohmann, C. H., Z. Schwartz, et al. (2000). "Pulsed electromagnetic field

stimulation of MG63 osteoblast-like cells affects differentiation and local

factor production." J Orthop Res 18(4): 637-646.

261

References

Lohmann, C. H., Z. Schwartz, et al. (2003). "Pulsed electromagnetic fields affect

phenotype and connexin 43 protein expression in MLO-Y4 osteocyte-like

cells and ROS 17/2.8 osteoblast-like cells." J Orthop Res 21(2): 326-334.

Lorich, D. G., C. T. Brighton, et al. (1998). "Biochemical pathway mediating the

response of bone cells to capacitive coupling." Clin Orthop Relat Res (350):

246-56.

Luben, R. A. (1991). "Effects of low-energy electromagnetic fields (pulsed and DC)

on membrane signal transduction processes in biological systems." Health

Physics 61(1): 15-28.

Luben, R. A. (1993). Effects of low-energy electromagnetic fields (EMF) on signal

transduction by G-Protein linked receptors. Electricity and magnetism in

Biology and Medicine. M. Blank. San Francisco, San Francisco Press Inc.:

57-62.

Lynch, T. M. and P. M. Lintilhac (1997). "Mechanical Signals in Plant Development:

A New Method for Single Cell Studies." Developmental Biology 181(2):

246-256.

Madronero, A. (1990). "Influence of magnetic fields on calcium salts crystal

formation: an explanation of the 'pulsed electromagnetic field' technique for

bone healing." J Biomed Eng 12(5): 410-4.

Mammi, G. I., R. Rocchi, et al. (1993). "The electrical stimulation of tibial

osteotomies. Double-blind study." Clin Orthop Relat Res (288): 246-53.

Marieb, E. N. (1995). Human anatomy and physiology. Redwood City,

Benjamin/Cummings.

262

References

Markov, M. S. and A. A. Pilla (1999). Static microT level magnetic fields modulate

myosin phosphorylation via kinetic effects on calcium binding to calmodulin.

Electricity and Magnetism in Biology and Medicine. F. Bersani. New York,

Plenum Pub. Corp.: 605-608.

Marotti, G. (1996). "The structure of bone tissues and the cellular control of their

deposition." Ital J Anat Embryol 101: 25-79.

Martin, R. B. (2000a). "Toward a unifying theory of bone remodelling." Bone 26(1):

1-6.

Martin, R. B. (2000b). "Does osteocyte formation cause the non-linear refilling of

osteons?" Bone 26(1): 71-78.

McAllister, T. N. and J. A. Frangos (1999). "Steady and Transient Fluid Shear Stress

Stimulate NO Release in Osteoblasts Through Distinct Biochemical

Pathways." J Bone Miner Res 14(6): 930-936.

McAllister, T. N., T. Du, et al. (2000). "Fluid shear stress stimulates prostaglandin

and nitric oxide release in bone marrow-derived preosteoclast-like cells."

Biochem Biophys Res Commun 270(2): 643-648.

McCaig, C. D. and M. Zhao (1997). "Physiological electrical fields modify cell

behaviour." BioEssays 19(9): 819-826.

McDonald, F. (1993). "Effect of static magnetic fields on osteoblasts and fibroblasts

in vitro." Bioelectromagnetics 14(3): 187-96.

McGrath, M. H., L. S. Glassman, et al. (1983). "Effect of external pulsing

electromagnetic fields on the healing of soft tissue." Surg. Forum 34: 615.

263

References

McLeod, B. R., A. A. Pilla, et al. (1983). "Electromagnetic fields induced by

Helmholtz aiding coils inside saline-filled boundaries." Bioelectromagnetics.

4(4): 357-370.

McLeod, K. J. and C. T. Rubin (1990). "Frequency specific modulation of bone

adaptation by induced electric fields." J Theor Biol 145(3): 385-396.

McLeod, K. J., H. J. Donahue, et al. (1991). Low-frequency sinusoidal electric fields

alter calcium fluctuations in osteoblastic-like cells. Electromagnetics in

Biology and Medicine. C. T. Brighton and S. R. Pollack. San Francisco, San

Francisco Press: 111-115.

McLeod, K. J. and C. T. Rubin (1992). "The effect of low-frequency electrical fields

on osteogenesis." J Bone Joint Surg Am 74(6): 920-9.

McLeod, K. J., H. J. Donahue, et al. (1993). "Electric fields modulate bone cell

function in a density-dependent manner." J Bone Miner Res 8(8): 977-984.

McLeod, K. J. and C. T. Rubin (1994). Regulation of cell growth rates in vitro by

alteration of induced charge density. The Annual review of research on

biological effects of electric and magnetic fields from the generation, delivery

and use of electricity. Frederick, MD, W/L Associates, Ltd. 80: 65.

McLeod, K. J., L. Porres, et al. (1998). "Induced surface charge density effects on

protein adsorption and cell adhesion in vitro." Trans. Bioelectromagnet. Soc.

20: 243-244.

Meyer, U., T. Meyer, et al. (2001a). "Tissue differentiation and cytokine synthesis

during strain-related bone formation in distraction osteogenesis." Br J Oral

Maxillofac Surg 39(1): 22-29.

264

References

Meyer, U., T. Meyer, et al. (2001b). "Mechanical tension in distraction osteogenesis

regulates chondrocytic differentiation." Int J Oral Maxillofac Surg 30(6):

522-530.

Mikuni-Takagaki, Y., Y. Suzuki, et al. (1996). "Distinct responces of different

populations of bone cells to mechanical stress." Endocrinology 137: 2028-

2035.

Miller, C. E., K. J. Donlon, et al. (2000). "Cyclic strain induces proliferation of

cultured embryonic heart cells." In Vitro Cell Dev Biol Anim 36(10): 633-9.

Moretti, M., A. Prina-Mello et al. (2004). "Endothelial cell alignment on cyclically-

stretched silicone surfaces." J Mater Sci Mater Med 15(10): 1159-1164.

Morgan, T. G., X. Yang, et al. (2001). "Trabecular bone adaptation to controlled in

vivo loading." Trans Orthop Res Soc 26: 238.

Mosley, J. R. and L. E. Lanyon (1998). "Strain rate as a controlling influence on

adaptive modelling in response to dynamic loading of the ulna in growing

male rats." Bone 23(4): 313-8.

Mosley, J. R. (2000). "Osteoporosis and bone functional adaptation:

mechanobiological regulation of bone architecture in growing and adult bone,

a review." J Rehabil Res Dev 37(2): 189-99.

Mow, V. C., C. C. Wang, et al. (1999). "The extracellular matrix, interstitial fluid

and ions as a mechanical signal transducer in articular cartilage."

Osteoarthritis Cartilage 7(1): 41-58.

Muehsam, D. J. and A. A. Pilla (1999). "The sensitivity of cells and tissues to

exogenous fields: effects of target system initial state." Bioelectrochem

Bioenerg 48(1): 35-42.

265

References

Mullender, M. G. and R. Huiskes (1995). "Proposal for the regulatory mechanism of

Wolff's law." J Orthop Res 13(4): 503-12.

Murray, D. W. and N. Rushton (1990). "The effect of strain on bone cell

prostaglandin E2 release: a new experimental method." Calcif Tissue Int

47(1): 35-39.

Neidlinger-Wilke, C., H. J. Wilke, et al. (1994). "Cyclic stretching of human

osteoblasts affects proliferation and metabolism: a new experimental method

and its application." J Orthop Res 12(1): 70-8.

Noble, B. S. and J. Reeve (2000). "Osteocyte function, osteocyte death and bone

fracture resistance." Mol Cell Endocrinol 159(1-2): 7-13.

Noda, M., D. E. Johnson, et al. (1987). "Effect of electric currents on DNA synthesis

in rat osteosarcoma cells: dependence on conditions that influence cell

growth." J Orthop Res 5(2): 253-260.

Norde, W. and C. E. Giacomelli (2000). "BSA structural changes during homo-

molecular exchange between the adsorbed and the dissolved states." J

Biotechnol 79(3): 259-268.

Norton, L. A., K. L. Andersen, et al. (1995). "A methodical study of shape changes

in human oral cells perturbed by a simulated orthodontic strain in vitro." Arch

Oral Biol 40(9): 863-872.

O'Connor, J. A., L. E. Lanyon, et al. (1982). "The influence of strain rate on adaptive

bone remodelling." J Biomech 15(10): 767-81.

Olsen, B. R., A. M. Reginato, et al. (2000). "Bone development." Annu Rev Cell

Dev Biol 16: 191-220.

266

References

Ott, S. (1998). Osteoporosis and Bone Physiology, Department of Medicine,

University of Washington. 2002.

Ottani, V., V. De Pasquale, et al. (1988). "Effects of pulsed extremely-low-frequency

magnetic fields on skin wounds in the rat." Bioelectromagnetics 9(1): 53-62.

Otter, M. W., K. J. McLeod, et al. (1998). "Effects of electromagnetic fields in

experimental fracture repair." Clin Orthop Relat Res (355 Suppl): S90-104.

Otter, M. W., V. R. Palmieri, et al. (1992). "A comparative analysis of streaming

potentials in vivo and in vitro." J Orthop Res 10(5): 710-9.

Otter, M. W., L. Porres, et al. (1996). "An investigation of the Brownian ratchet in

MC-3T3-E1 osteoblast-like cells using atomic force microscopy."

Transactions from Society for Physical Regulation in Biology and Medicine

16: 10-11.

Otter, M. W., C. T. Rubin, et al. (1997). "Can the response of bone to extremely

weak stimuli be explained by the Brownian ratchet?" Ann Biomed Eng

25(Suppl 1): S76.

Owan, I., D. B. Burr, et al. (1997). "Mechanotransduction in bone: osteoblasts are

more responsive to fluid forces than mechanical strain." Am J Physiol 273(3

Pt 1): C810-5.

Ozawa, H., K. Imamura, et al. (1990). "Effect of a continuously applied compressive

pressure on mouse osteoblast-like cells (MC3T3-E1) in vitro." J Cell Physiol

142(1): 177-185.

Panjabi, M. M., A. A. White III, et al. (1979). "A biomechanical comparison of the

effects of constant and cyclic compression on fracture healing in rabbit long

bones." Acta Orthopaedica Scandinavica 50: 653-661.

267

References

Park, J. S., J. S. F. Chu, et al. (2004). "Differential effects of equiaxial and uniaxial

strain on mesenchymal stem cells." Biotechnology And Bioengineering

88(3): 359-368.

Park, S.-H. and M. Silva (2004). "Neuromuscular electrical stimulation enhances

fracture healing: results of an animal model." J Orthop Res 22(2): 382-387.

Parker, T., Z. Upton, et al. (2005). "Potential pitfalls of radiolabel adsorption to

ceramic biomaterials." J Biomed Mater Res A 72A(4): 363-72.

Pautke, C., M. Schieker, et al. (2002). "Comparison of human osteoblasts to different

osteosarcoma cell lines." Bone 30(3 (S1)): 9.

Pautke, C., M. Schieker, et al. (2004). "Characterization of osteosarcoma cell lines

MG-63, Saos-2 and U-2 OS in comparison to human osteoblasts." Anticancer

Res 24(6): 3743-8.

Pauwels, F. (1941). Grundrib einer Biomechanik der Fraktrheilung. 34th Kongress

der Deutschen Orthopadischen Gesellschaft Ferdinand Enke Verlag, Stuttgart

(1980) (Biomechanics of the locomotor apparatus). P. Manquet and R.

Furlong. Berlin, Spinger: 375-407.

Pavalko, F. M., N. X. Chen, et al. (1998). "Fluid shear-induced mechanical

signalling in MC3T3-E1 osteoblasts requires cytoskeleton-integrin

interactions." Am J Physiol 275(6 Pt 1): C1591-601.

Pavalko, F. M., R. L. Gerard, et al. (2003a). "Fluid shear stress inhibits TNF--

induced apoptosis in osteoblasts: A role for fluid shear stress-induced

activation of PI3-kinase and inhibition of caspase-3." J Cell Physiol 194(2):

194-205.

268

References

Pavalko, F. M., S. M. Norvell, et al. (2003b). "A model for mechanotransduction in

bone cells: The load-bearing mechanosomes." J Cell Biochem 88(1): 104-

112.

Pavlin, M., N. Pavselj, et al. (2002). "Dependence of induced transmembrane

potential on cell density, arrangement, and cell position inside a cell system."

IEE Trans Biomed Eng 49(6): 605-612.

Pead, M. J., T. M. Skerry, et al. (1988). "Direct transformation from quiescence to

bone formation in the adult periosteum following a single brief period of

bone loading." J Bone Miner Res 3(6): 647-56.

Pead, M. J. and L. E. Lanyon (1989). "Indomethacin modulation of load-related

stimulation of new bone formation in vivo." Calcif Tissue Int 45(1): 34-40.

Pedersen, E. A., M. P. Akhter, et al. (1999). "Bone Response to In Vivo Mechanical

Loading in C3H/HeJ Mice." Calcif Tissue Int 65(1): 41-46.

Pienkowski, D., S. R. Pollack, et al. (1992). "Comparison of asymmetrical and

symmetrical pulse waveforms in electromagnetic stimulation." J Orthop Res

10(2): 247-55.

Pilla, A. A. (1993). State of the Art in Electromagnetic Therapeutics. Electricity and

Magnetism in Biology and Medicine. M. Blank. San Francisco, San


Pilla, A. A., M. Figueiredo, et al. (1993). Broadband EMF acceleration of bone

repair in a rabbit model is independent of magnetic component. Electricity

and Magnetism in Biology and Medicine. M. Blank. San Francisco, San


269

References

Pilla, A. A. (2002a). "Electromagnetic and mechanical modalities in therapeutic

applications: from mechanisms to the clinic." Bioelectromagnetics 24th

Annual Meeting.

Pilla, A. A. (2002b). "Low-intensity electromagnetic and mechanical modulation of

bone growth and repair: are they equivalent?" J Orthop Sci 7(3): 420-428.

Pioletti, D. P., J. Muller, et al. (2003). "Effect of micromechanical stimulations on

osteoblasts: development of a device simulating the mechanical situation at

the bone-implant interface." J Biomech 36(1): 131-135.

Pitsillides, A. A., S. C. Rawlinson, et al. (1995). "Mechanical strain-induced NO

production by bone cells: a possible role in adaptive bone (re)modelling?"

FASEB Journal 9(15): 1614-22.

Prendergast, P. J., R. Huiskes, et al. (1997). "Biophysical stimuli on cells during

tissue differentiation at implant interfaces." J Biomech 30(6): 539-548.

Qin, K., L.-H. Qiu, et al. (2004). "The effect of static magnetic field on bone

morphogenetic protein-2 in periodontal membrane of the rat." Shanghai Kou

Qiang Yi Xue 13(4): 275-277.

Qin, Y.-X., T. Kaplan, et al. (2003). "Fluid pressure gradients, arising from

oscillations in intramedullary pressure, is correlated with the formation of

bone and inhibition of intracortical porosity." J Biomech 36(10): 1427-1437.

Qiu, Q., M. Sayer, et al. (1998). "Attachment, morphology, and protein expression of

rat marrow stromal cells cultured on charged substrate surfaces." J Biomed

Mater Res 42: 117-127.

Ralston, S. H. (1997). "The Michael Mason Prize Essay 1997. Nitric oxide and bone:

what a gas!" Br. J. Rheumatol. 36(8): 831-838.

270

References

Rauch, C., A.-C. Brunet, et al. (2002). "C(2)C(12) myoblast/osteoblast

transdifferentiation steps enhanced by epigenetic inhibition of BMP2

endocytosis." Am J Physiol 283(1): C235-C243.

Rawlinson, S. C., S. Mohan, et al. (1993). "Exogenous prostacyclin, but not

prostaglandin E2, produces similar responses in both G6PD activity and RNA

production as mechanical loading, and increases IGF-II release, in adult

cancellous bone in culture." Calcif Tissue Int 53(5): 324-9.

Reich, K. M. and J. A. Frangos (1991). "Effect of flow on prostaglandin E2 and

inositol trisphosphate levels in osteoblasts." Am J Physiol 261(1): C428-

C432.

Reich, K. M., T. N. McAllister, et al. (1997). "Activation of G proteins mediates

flow-induced prostaglandin E2 production in osteoblasts." Endocrinology

138(3): 1014-8.

Reinish, G. B. and A. S. Nowick (1976). "Effects of moisture on the electrical

properties of bone." J Electrochem Soc 201: 145.

Rinsky, L. A., A. Halpern, et al. (1980). "Electrical stimulation of experimentally

produced avascular necrosis of the femoral head." Orthop Trans 4: 238.

Robling, A. G., D. B. Burr, et al. (2000). "Partitioning a daily mechanical stimulus

into discrete loading bouts improves the osteogenic response to loading." J

Bone Miner Res 15(8): 1596-602.

Robling, A. G., K. M. Duijvelaar, et al. (2001). "Modulation of appositional and

longitudinal bone growth in the rat ulna by applied static and dynamic force."

Bone 29(2): 105-13.

271

References

Robling, A. G., F. M. Hinant, et al. (2002a). "Improved bone structure and strength

after long-term mechanical loading is greatest if loading is separated into

short bouts." J Bone Miner Res 17(8): 1545-1554.

Robling, A. G., F. M. Hinant, et al. (2002b). "Shorter, more frequent mechanical

loading sessions enhance bone mass." Med Sci Sports Exerc 34(2): 196-202.

Robling, A. G. and C. H. Turner (2002). "Mechanotransduction in bone: genetic

effects on mechanosensitivity in mice." Bone 31(5): 562-569.

Rodan, G. A. and T. J. Martin (1981). "Role of osteoblasts in hormonal control of

bone resorption - a hypothesis." Calcif Tissue Int 33: 349-351.

Rodan, S. B., Y. Imai, et al. (1987). "Characterization of a human osteosarcoma cell

line (Saos-2) with osteoblastic properties." Cancer Res 47(18): 4961-4966.

Rubenacker, S., C. Neidlinger-Wilke, et al. (1995). "Human osteoblasts from

younger normal and osteoporotic donors show differences in proliferation and

TGF[beta]-release in response to cyclic strain." J Biomech 28(12): 1411-

1418.

Rubin, C. T. and L. E. Lanyon (1987). "Osteoregulatory nature of mechanical

stimuli: function as a determinant for adaptive remodelling in bone." J Orthop

Res 5(2): 300-10.

Rubin, C. T., K. J. McLeod, et al. (1989). "Prevention of osteoporosis by pulsed

electromagnetic fields." J Bone Joint Surg Am 71(3): 411-7.

Rubin, C. T., H. J. Donahue, et al. (1993). "Optimization of electric field parameters

for the control of bone remodelling: exploitation of an indigenous mechanism

for the prevention of osteopenia." J Bone Miner Res 8(Suppl 2): S573-81.

272

References

Rubin, C. T. and K. J. McLeod (1994). "Promotion of bony ingrowth by frequency-

specific, low-amplitude mechanical strain." Clin Orthop Relat Res(298): 165-

174.

Rubin, C. T., K. J. McLeod, et al. (1996). "Formation of osteoclast-like cells is

suppressed by low frequency, low intensity electric fields." J Orthop Res

14(1): 7-15.

Rubin, C. T., G. Xu, et al. (2001a). "The anabolic activity of bone tissue, suppressed

by disuse, is normalized by brief exposure to extremely low-magnitude

mechanical stimuli." FASEB J 15(12): 2225-2229.

Rubin, C. T., A. S. Turner, et al. (2001b). "Anabolism: Low mechanical signals

strengthen long bones." Nature 412: 603 - 604.

Rubin, C. T., A. S. Turner, et al. (2002). "Mechanical strain, induced noninvasively

in the high-frequency domain, is anabolic to cancellous bone, but not cortical

bone." Bone 30(3): 445-452.

Rubin, C. T., R. R. Recker, et al. (2004). "Prevention of Postmenopausal Bone Loss

by a Low-Magnitude, High-Frequency Mechanical Stimuli: A Clinical Trial

Assessing Compliance, Efficacy, and Safety." J Bone Miner Res 19(3): 343-

351.

Ruel, J., J. Lemay, et al. (1995). "Development of a parallel plate flow chamber for

studying cell behavior under pulsatile flow." ASAIO Journal 41(4): 876-883.

Ruoslahti, E. and J. C. Reed (1994). "Anchorage dependence, integrins, and

apoptosis." Cell 77(4): 477-8.

273

References

Sakai, K., M. Mohtai, et al. (1998). "Fluid shear stress increases transforming growth

factor beta 1 expression in human osteoblast-like cells: modulation by cation

channel blockades." Calcif Tissue Int 63(6): 515-20.

Sakai, K., M. Mohtai, et al. (1999). "Fluid Shear Stress Increases Interleukin-11

Expression in Human Osteoblast-like Cells: Its Role in Osteoclast Induction."

J Bone Miner Res 14(12): 2089-2098.

Salter, D. M., W. H. Wallace, et al. (2000). "Human bone cell hyperpolarization

response to cyclical mechanical strain is mediated by an interleukin-1beta

autocrine/paracrine loop." J Bone Miner Res 15(9): 1746-1755.

Saunders, M. M., J. You, et al. (2001). "Gap junctions and fluid flow response in

MC3T3-E1 cells." Am J Physiol 281(6): C1917-25.

Schaffer, J. L., M. Rizen, et al. (1994). "Device for the application of a dynamic

biaxially uniform and isotropic strain to a flexible cell culture membrane." J

Orthop Res 12(5): 709-719.

Schiller, P. C., G. D'Ippolito, et al. (2001). "Inhibition of gap-junctional

communication induces the trans-differentiation of osteoblasts to an

adipocytic phenotype in vitro." J Biol Chem 276(17): 14133-8.

Schneider, G. and K. Burridge (1994). "Formation of focal adhesions by osteoblasts

adhering to different substrata." Exp Cell Res 214(1): 264-9.

Serway, R. A. (2004). Physics for scientists and engineers. Belmont, CA, Thomson-

Brooks/Cole.

Shankar, V. S., B. J. Simon, et al. (1998). "Effects of electromagnetic stimulation on

the functional responsiveness of isolated rat osteoclasts." J Cell Physiol

176(3): 537-544.

274

References

Shapiro, F., C. Cahill, et al. (1995). "Transmission electron microscopic

demonstration of vimentin in rat osteoblast and osteocyte cell bodies and

processes using the immunogold technique." Anat Rec 241(1): 39-48.

Sharrard, W. J. W., M. L. Sutcliffe, et al. (1982). "The treatment of fibrous non-

union of fractures by pulsing electromagnetic stimulation." J Bone Joint Surg

British 64-B(2): 189-193.

Shimizu, T., J. E. Zerwekh, et al. (1988). "Bone ingrowth into porous calcium

phosphate ceramics: influence of pulsing electromagnetic field." J Orthop Res

6(2): 248-58.

Sikavitsas, V. I., J. S. Temenoff, et al. (2001). "Biomaterials and bone

mechanotransduction." Biomaterials 22(19): 2581-2593.

Simmons, C. A., S. Matlis, et al. (2003). "Cyclic strain enhances matrix

mineralization by adult human mesenchymal stem cells via the extracellular

signal-regulated kinase (ERK1/2) signaling pathway." J Biomech 36(8):

1087-1096.

Skerry, T. M., R. Suswillo, et al. (1990). "Load-induced proteoglycan orientation in

bone tissue in vivo and in vitro." Calcif Tissue Int 46(5): 318-26.

Smalt, R., F. T. Mitchell, et al. (1997). "Induction of NO and prostaglandin E2 in

osteoblasts by wall-shear stress but not mechanical strain." Am J Physiol

273(1): E751-E758.

Smith, D. H., J. A. Wolf, et al. (2001). "A New Strategy to Produce Sustained

Growth of Central Nervous System Axons: Continuous Mechanical Tension."

Tissue Eng 7(2): 131-139.

275

References

Smith, O. M., E. M. Goodman, et al. (1991). "An increase in the negative surface

charge of U937 cells exposed to a pulsed magnetic field."


Smith, S. D., B. R. McLeod, et al. (1987). "Calcium cyclotron resonance and diatom

mobility." Bioelectromagnetics. 8: 215.

Smith, S. D. and A. Pilla (1981). Modulation of Newt limb regeneration by

electromagnetically induced low level pulsating current. Mechanisms of

Growth Control. R. O. Becker. Springfield, Charles C. Thomas: 137-152.

Søballe, K., E. S. Hansen, et al. (1992). "Tissue ingrowth into titanium and

hydroxyapatite-coated implants during stable and unstable mechanical

conditions." J Orthop Res 10(2): 285-99.

Sollazzo, V., L. Massari, et al. (1996). "Effects of Low Frequency Pulsed

Electromagnetic fields on Human Osteoblast-like cells in vitro." Electro- and

Magnetobiology 15(1): 75-83.

Sollazzo, V., G. C. Traina, et al. (1997). "Responses of human MG-63 osteosarcoma

cell line and human osteoblast-like cells to pulsed electromagnetic fields."


Song, B., M. Zhao, et al. (2002). "Electrical cues regulate the orientation and

frequency of cell division and the rate of wound healing in vivo." Proc Natl

Acad Sci U S A 99(21): 13577-82.

Sotoudeh, M., S. Jalali, et al. (1998). "A strain device imposing dynamic and

uniform equi-biaxial strain to cultured cells." Ann Biomed Eng 26(2): 181-

189.

276

References

Spadaro, J. A., S. A. Albanese, et al. (1990). "Electromagnetic effects on bone

formation at implants in the medullary canal in rabbits." J Orthop Res 8(5):

685-93.

Spadaro, J. A., S. A. Albanese, et al. (1992). "Bone formation near direct current

electrodes with and without motion." J Orthop Res 10(5): 729-38.

Spadaro, J. A. (1997). "Mechanical and electrical interactions in bone remodelling."


Sreedharan, V. and D. Zhang (2003). "Finite element modelling of cellular responses

of gap junction connected osteocytes under extremely low-frequency

electromagnetic fields." Bioengineering, Proceedings of the Northeast

Conference: 160-161.

Stanford, C. M., J. W. Stevens, et al. (1995). "Cellular deformation reversibly

depresses RT-PCR detectable levels of bone-related mRNA." J Biomech.

28(12): 1419-27.

Steele, J. G., B. A. Dalton, et al. (1993). "Polystyrene chemistry affects vitronectin

activity: an explanation for cell attachment to tissue culture polystyrene but

not to unmodified polystyrene." J Biomed Mater Res 27(7): 927-40.

Stein, G. S., J. B. Lian, et al. (1990). "Relationship of cell growth to the regulation of

tissue-specific gene expression during osteoblast differentiation." FASEB J

4(13): 3111-23.

Stephansson, S. N., B. A. Byers, et al. (2002). "Enhanced expression of the

osteoblastic phenotype on substrates that modulate fibronectin conformation

and integrin receptor binding." Biomaterials 23(12): 2527-2534.

277

References

Strobel, M., C. S. Lyons, et al. (1994). Plasma surface modification of polymers :

relevance to adhesion. Utrecht, Netherlands, VSP.

Supronowicz, P. R. (2002). The effects of biophysical stimuli on select bone cell

functions pertinent to osteogenesis. Biomedical Engineering. New York,

Rensselaer Polytechnic Institute: 157.

Sutcliffe, M. L. and A. A. Goldberg (1982). "The treatment of congenital

pseudarthrosis of the tibia with pulsing electromagnetic fields. A survey of 52

cases." Clin Orthop Relat Res (166): 45-57.

Tabrah, F., M. Hoffmeier, et al. (1990). "Bone density changes in osteoporosis-prone

women exposed to pulsed electromagnetic fields (PEMFs)." J Bone Miner

Res 5(5): 437-442.

Tami, A. E., P. Nasser, et al. (2002). "The Role of Interstitial Fluid Flow in the

Remodelling Response to Fatigue Loading." J Bone Miner Res 17(11): 2030 -

2037.

Tanaka, S. M. (1999). "A new mechanical stimulator for cultured bone cells using

piezoelectric actuator." J Biomech 32(4): 427-430.

Tanaka, S. M., I. Alam, et al. (2003a). "Stochastic resonance in osteogenic response

to mechanical loading." FASEB J 17: 313-314.

Tanaka, S. M., J. Li, et al. (2003b). "Effects of broad frequency vibration on cultured

osteoblasts." J Biomech 36(1): 73-80.

Tanaka-Kamioka, K., H. Kamioka, et al. (1998). "Osteocyte shape is dependant on

actin filaments and osteocyte processes are unique actin-rich projections." J

Bone Miner Res 13(10): 1555-1568.

278

References

Tanck, E., L. Blankevoort, et al. (1998). "The influence of muscular activity on local

mineralization patterns in metatarsals of the fetal mouse." J Biomech 31(1):

23.

Tanck, E., W. D. van Driel, et al. (1999). "Why does intermittent hydrostatic

pressure enhance the mineralization process in fetal cartilage?" J Biomech

32(2): 153-161.

Tanck, E., L. Blankevoort, et al. (2000). "Influence of muscular activity on local

mineralization patterns in metatarsals of the embryonic mouse." J Orthop Res

18(4): 613-619.

Tardy, Y., N. Resnick, et al. (1997). "Shear stress gradients remodel endothelial

monolayers in vitro via a cell proliferation-migration-loss cycle." Arterioscler

Thromb Vasc Biol 17(11): 3102-3106.

Tenforde, T. S. (1989). "Electro-reception and magneto-reception in simple and

complex organisms." Bioelectromagnetics 10(3): 215-21.

Tjandrawinata, R. R., V. L. Vincent, et al. (1997). "Vibrational force alters mRNA

expression in osteoblasts." FASEB J 11(6): 493-7.

Toma, C. D., S. Ashkar, et al. (1997). "Signal Transduction of Mechanical Stimuli Is

Dependent on Microfilament Integrity: Identification of Osteopontin as a

Mechanically Induced Gene in Osteoblasts." J Bone Miner Res 12(10): 1626-

1636.

Trock, D. H., A. J. Bollet, et al. (1994). "The effect of pulsed electromagnetic fields

in the treatment of osteoarthritis of the knee and cervical spine. Report of

randomized, double blind, placebo controlled trials." J Rheumatol 21(10):

1903-1911.

279

References

Turner, C. H., M. R. Forwood, et al. (1994a). "Mechanotransduction in bone: do

bone cells act as sensors of fluid flow?" FASEB J 8(11): 875-8.

Turner, C. H., M. R. Forwood, et al. (1994b). "Mechanical loading thresholds for

lamellar and woven bone formation." J Bone Miner Res 9(1): 87-97.

Turner, C. H., I. Owan, et al. (1995). "Mechanotransduction in bone: role of strain

rate." Am J Physiol 269(3 Pt 1): E438-42.

Turner, C. H. (1998). "Three rules for bone adaptation to mechanical stimuli." Bone

23(5): 399-407.

Turner, C. H. and F. M. Pavalko (1998). "Mechanotransduction and functional

response of the skeleton to physical stress: the mechanisms and mechanics of

bone adaptation." J Orthop Sci 3(6): 346-55.

Turner, C. H., A. G. Robling, et al. (2002). "Do bone cells behave like a neuronal

network?" Calcif Tissue Int 70(6): 435-442.

Uhthoff, H. K. and Z. F. G. Jaworski (1978). "Bone loss in response to long-term

immobilisation." J Bone Joint Surg- British Volume 60: 420-429.

Urban, M. W. (1996). Attenuated total reflectance spectroscopy of polymers : theory

and practice. Washington, DC, American Chemical Society.

van der Meulen, M. C., M. Moro, et al. (2000). "Mechanobiology of femoral neck

structure during adolescence." J Rehabil Res Dev 37(2): 201-8.

van der Meulen, M. C., C. H. Marjolein, et al. (2002). "Why mechanobiology?; A

survey article." J Biomech 35(4): 401-414.

280

References

van Loon, J. J., D. J. Bervoets, et al. (1995). "Decreased mineralization and increased

calcium release in isolated fetal mouse long bones under near

weightlessness." J Bone Miner Res 10(4): 550-7.

Vandenburgh, H. H. (1988). "A computerized mechanical cell stimulator for tissue

culture: effects on skeletal muscle organogenesis." In Vitro Cell Dev Biol

24(7): 609-619.

Vandenburgh, H. H. and P. Karlisch (1989). "Longitudinal growth of skeletal

myotubes in vitro in a new horizontal mechanical cell stimulator." In Vitro

Cell Dev Biol 25(7): 607-616.

Vander Molen, M. A. and K. J. McLeod (1995). "Reduced surface charge density

extends the G2/M phase of the cell cycle in proliferating osteoblastic cell

lines." Trans Bioelectromagnetics Soc 18.

Vander Molen, M. A., H. J. Donahue, et al. (2000). "Osteoblastic networks with

deficient coupling: differential effects of magnetic and electric field

exposure." Bone 27(2): 227-231.

van't Hof, R. J. and S. H. Ralston (2001). "Nitric oxide and bone." Immunology

103(3): 255-261.

Vroman, L. and A. L. Adams (1986). "Adsorption of proteins out of plasma and

solutions in narrow spaces." J Colloid Interface Sci 111(2): 391-402.

Wang, F. S., C. J. Wang, et al. (2001). "Physical shock wave mediates membrane

hyperpolarization and Ras activation for osteogenesis in human bone marrow

stromal cells." Biochem Biophys Res Commun 287(3): 648-55.

281

References

Wang, L. (1999). "Protection from Cell Death by mcl-1 Is Mediated by Membrane

Hyperpolarization Induced By K+ Channel Activation." J Membr Biol

172(2): 113-120.

Warden, S. J. and C. H. Turner (2004). "Mechanotransduction in the cortical bone is

most efficient at loading frequencies of 5-10 Hz." Bone 34(2): 261-270.

Weikart, C. M. and H. K. Yasuda (2000). "Modification, Degradation and Stability

of Polymeric Surfaces Treated with Reactive Plasmas." J Polym Sci A:

Polym Chem 38: 3028-3042.

Weinans, H., R. Huiskes, et al. (1992). "The behaviour of adaptive bone remodelling

simulation models." J Biomech 25: 1425-1441.

Weinbaum, S., S. C. Cowin, et al. (1994). "A model for the excitation of osteocytes

by mechanical loading-induced bone fluid shear stresses." J Biomech 27(3):

339-360.

Westbroek, I., N. E. Ajubi, et al. (2000). "Differential stimulation of prostaglandin

G/H synthase-2 in osteocytes and other osteogenic cells by pulsating fluid

flow." Biochem Biophys Res Commun 268(2): 414-9.

Weyts, F., B. Bosmans, et al. (2003). "Mechanical Control of Human Osteoblast

Apoptosis and Proliferation in Relation to Differentiation." Calcif Tissue Int

72(4): 505-512.

Williams, J. L., J. H. Chen, et al. (1992). "Strain fields on cell stressing devices

employing clamped circular elastic diaphragms as substrates." J Biomech Eng

114(3): 377-384.

Williams, J. L., J. P. Iannotti, et al. (1994). "Effects of fluid shear stress on bone

cells." Biorheology 31(2): 163-170.

282

References

Winston, F. K., E. J. Macarak, et al. (1989). "A system to reproduce and quantify the

biomechanical environment of the cell." J Appl Physiol 67(1): 397-405.

Xia, N., C. J. May, et al. (2002). "Time-of-flight secondary ion mass spectrometry

analysis of conformational changes in adsorbed protein films." Langmuir

18(10): 4090-4097.

Xia, S. L. and J. Ferrier (1992). "Propagation of a calcium pulse between osteoblastic

cells." Biochem Biophys Res Commun 186(3): 1212-9.

Xu, J., M. Liu, et al. (1996). "Mechanical strain induces constitutive and regulated

secretion of glycosaminoglycans and proteoglycans in fetal lung cells." J Cell

Sci 109 (Pt 6): 1605-1613.

Yamaguchi, D. T., J. Huang, et al. (2002). "Inhibition of gap junction intercellular

communication by extremely low-frequency electromagnetic fields in

osteoblast-like models is dependent on cell differentiation." J Cell Physiol

190(2): 180-188.

Yamamoto, Y., Y. Ohsaki, et al. (2003). "Effects of static magnetic fields on bone

formation in rat osteoblast cultures." J Dent Res 82(12): 962-966.

Yang, Y., J. Magnay, et al. (2004). "Effects of substrate characteristics on bone cell

response to the mechanical environment." Med Biol Eng Comput 42(1): 22-

29.

Yen-Patton, G. P., W. F. Patton, et al. (1988). "Endothelial cell response to pulsed

electromagnetic fields: stimulation of growth rate and angiogenesis in vitro."

J Cell Physiol 134(1): 37-46.

Yonemori, K., S. Matsunaga, et al. (1996). "Early effects of electrical stimulation on

osteogenesis." Bone 19(2): 173-180.

283

References

You, J., C. E. Yellowley, et al. (2000). "Substrate deformation levels associated with

routine physical activity are less stimulatory to bone cells relative to loading-

induced oscillatory fluid flow." J Biomech Eng 122(4): 387-93.

You, J., G. C. Reilly, et al. (2001a). "Osteopontin gene regulation by oscillatory fluid

flow via intracellular calcium mobilization and activation of mitogen-

activated protein kinase in MC3T3-E1 osteoblasts." J Biol Chem 276(16):

13365-71.

You, L., S. C. Cowin, et al. (2001). "A model for strain amplification in the actin

cytoskeleton of osteocytes due to fluid drag on pericellular matrix." J

Biomech 34(11): 1375-1386.

You, L. (2002). A new view of mechanotransduction in bone cells. Graduate Faculty

in Engineering. New York, City University of New York: 97.

Zaman, G., R. F. Suswillo, et al. (1997). "Early responses to dynamic strain change

and prostaglandins in bone-derived cells in culture." J Bone Miner Res 12(5):

769-77.

Zeng, Y., S. C. Cowin, et al. (1994). "A fiber matrix model for fluid flow and

streaming potentials in the canaliculi of an osteon." Ann Biomed Eng. 22(3):

280-92.

Zengo, A. N., C. A. Bassett, et al. (1976). "In vivo effects of direct current in the

mandible." J Dent Res 55(3): 383-90.

Zhang, D., S. Weinbaum, et al. (1998). "Electrical signal transmission in a bone cell

network: the influence of a discrete gap junction." Ann Biomed Eng. 26(4):

644-59.

284

References

285

Zhao, M., J. V. Forrester, et al. (1999). "A small, physiological electric field orients

cell division." Proc Natl Acad Sci U S A 96(9): 4942-4946.

Zhao, M., J. Pu, et al. (2002). "Membrane lipids, EGF receptors, and intracellular

signals colocalize and are polarized in epithelial cells moving directionally in

a physiological electric field." FASEB J 16(8): 857-9.

Ziros, P. G., A.-P. R. Gil, et al. (2002a). "The bone-specific transcriptional regulator

Cbfa1 is a target of mechanical signals in osteoblastic cells." J Biol Chem

277(26): 23934-23941.

Ziros, P. G., A.-P R. Gil, et al. (2002b). "Targeting of mechanical signals to the

osteoblast-specific transcriptional regulator CBFa1." Bone 30(3-S1): 15.

Documents

Mechanical and Electrical Environments to …Mechanical and Electrical Environments to Stimulate Bone Cell Development Gwynne Hannay, BE (Medical)(Hons) Thesis submitted for the degree