Characterization of the Properties of Cartilage in the
Hartley Guinea Pig Spontaneous Osteoarthritis Model
by
Bryant Yenfong Lin
Submitted to the Department of Electrical Engineering
and Computer Science in partial fulfillment of the requirements for the degree of
Master of Engineering in Electrical Engineering and
Computer Science
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
June 1997
© Bryant Yenfong Lin, 1997. All Rights Reserved
The author hereby grants to M.I.T. permission to reproduce and distribute paper and electronic copies of this thesis and3 grant ohers te right to do so ,
"............
..
................
.. ....
".
Author .........
Depart
gi ee ng and Computer Science
May 27, 1997
t of Ele tel
.. .......................
Certified by ................
Cetiie bAlan
Mechanical, and Bioengineering
)-)1h;esiis,ýrvisor
•
Professor of Electri
Accepted by ........
J.Grodzinsky
-..........
........ ,...
Arthur C.Smith
Chairman, Department Committee on Graduate Theses
.D
2 1997
"- .*. '
p.
. •t•?
,
..
Characterization of Properties of Cartilage in the Hartley Guinea Pig
Spontaneous Osteoarthritis Model
by
Bryant Yenfong Lin
Submitted to the Department of Electrical Engineering and Computer Science on May 27, 1997, in partial fulfillment of the requirements for the degree of Master of Engineering in Electrical
Engineering.
Abstract
A common cause of osteoarthritis (OA) is the degradation of the articular cartilage covering the ends of the bones in joints. This study examines the characteristics of the articular
cartilage tibial plateau of the Hartley guinea pig looking at the electromechanical (static
stiffness, dynamic stiffness, streaming potential) and biochemical (glycosaminoglycan,
collagen) properties at three sites. The cartilage thickness was also measured. The goal
was threefold. First, the ability to perform electrokinetic indentation measurements in a
nondestructive manner was quantified and any differences in the electromechanical properties of left and right joints in the guinea pig model examined. Second, the sensitivity of
these techniques to the age-related changes in the development of OA in the guinea pig
model was determined. Third, those changes in electrokinetic properties were to be compared to changes in biochemical ones. The indentation was found to be nondestructive.
The results of the age experiments show a large variation between the properties at different sites while differences among age groups is not clear. Preliminary work on the biochemistry is positive. Further statistical and biochemical analyses must be performed to
verify whether there are significant difference in physical properties with age.
Thesis Supervisor: Alan J. Grodzinsky
Title: Professor of Electrical, Mechanical, and Bioengineering
Acknowledgements
I would like to first thank Prof. Alan Grodzinsky for giving me the opportunity to learn about biomedical research this year. It has been an educational and enlightening experience.
Everyone in this lab has been extremely helpful and supportive each in their own way. Thank you
Linda, Eliot, Steve, Paula, Marc, Larry and Han-hwa. Special thanks to Beth who has been my
colleague in the project this year.
Most of all, thanks to my family for always being there for me.
This research was funded in part by a grant from Procter & Gamble.
Table of Contents
Abstract.....................................................
........................................................................ 2
Acknowledgem ents ... .................................................................................................. 3
Table of Contents................................................... ..................................................... 4
List of Figures...................................................... ....................................................... 6
1 Introduction and Background ....................................................... .. .................
1.1 O steoarthritis............................................................................................8..........
10
1.2 Cartilage Structure ...........................................................................................
11
1.3 Lesion Developm ent .............................................................................
1.4 Objective ..................................................................................
............
11
1.5 Overview ..........................................................................................................
13
2 Left versus Right D ifferences, Phase 1..................................
...........
14
2.1 Introduction and Sum mary .................................... . ...... ........
14
2.2 M aterials and M ethods..................................................
.......... .......... 15
2.3 Results ........................................................... ............................................. 20
2.4 D iscussion ............................................................................. ..................... 29
2.5 Conclusion ............................................................................ ..................... 31
3 Electromechanical Differences Between Age Groups, Phase2 .............................. 33
3.1 Introduction ............................................................................ .................... 33
3.2 M aterials and M ethods..................................................
.......... .......... 34
3.3 Results...........................................................................................................35
3.4 D iscussion .......................................................................
............................ 44
3.5 Conclusion ............................................................................ ..................... 46
4 Biochem ical A ssays .......................................................................... ................... 47
4.1 Introduction ...................................................................................................... 47
4.2 M ethods............................................................................................................ 48
4.3 Results........................................................... ............................................. 50
4.4 D iscussion ...............................................................................................
52
4.5 Conclusion ................................................................................................
53
5 Future Testing .....................................................................................................
54
5.1 Introduction .............................. .................................................................. 54
5.2 Phase 3 .................................
...................................................................... 54
5.3 Conclusion ..................................................................................................... 55
Appendix A Load Control...........................................................................................
56
A.1 Introduction ...............................
.................................................................
56
A.2 M ethods............................................................................................................56
A .3 Results..............................................................................................................57
A .4 D iscussion .................................................................................................
59
A.5 Conclusion ................................................................................................
59
Appendix B Resolution of Orientation Bias ............................................. 60
B.1 Introduction....................................
............................................................
60
B.2 M ethods............................................................................................................60
B.3 Results........................................................................................................61
B.4 D iscussion and Conclusion ........................................................................... 61
Appendix C Normalization .......................................................................................
63
C.1 Introduction ......................................................................................
.......... 63
C.2 M ethods..........................................................................
............................. 63
C.3 Results and Discussion ..................................... ........
.
........
C.4 Conclusion ..................................................
...............................................
Appendix D Sample Chart Recorder W aveforms .......................................
....
Appendix E Phase 2 Plots without Chart Recorder Data.................................
...
Bibliography ..................................................................................
.........................
63
64
65
67
82
List of Figures
Figure 1.1: Lesion SEM ....................................................................................................
9
Figure 2.1: Mounting Chamber.........................................................................
16
Figure 2.2: Indentor Probe. ........................................................................................ 16
Figure 2.3: Tibial Plateau Sketch..................................................
................... 17
Figure 2.4: Thickness Measurement ...................................................................
19
Figure 2.5: Static Stress, AM site ..............................................
........ 20
FIgure 2.6: Static Stress, MM and ML sites ........................................
..... 21
Figure 2.7: Dynamic Stiffness, AM site ................................................ 22
Figure 2.8: Dynamic Stiffness, MM and ML sites ......................................................... 23
Figure 2.9: Normalized Dynamic Stiffness, MM and AM sites ..................................... 24
Figure 2.10: Normalized Dynamic Stiffness, ML site ................................................. 25
Figure 2.11: Streaming Potential, AM and MM sites.....................................................
26
Figure 2.12a: Streaming Potential, ML site ................................................................. 27
Figure 2.12b: Normalized Streaming Potential, AM site .....................................
27
Figure 2.13: Normalized Streaming Potential, MM and ML sites ................................. 28
Figure 2.14: Phase 1 Thickness ......................................................................................
29
Figure 3.1: Static Stress, AM site ...................................................................................
35
Figure 3.2: Static Stress, MM and ML sites ...................................................................
36
Figure 3.3: Dynamic Stiffness, M L site..........................................................................37
Figure 3.4: Dynamic Stiffness, MM and AM sites..................................
.... 38
Figure 3.5: Normalized Dynamic Stiffness, ML and MM sites .................................. 39
Figure 3.6: Normalized Dynamic Stiffness, AM site ....................................
.... 40
Figure 3.7: Streaming Potential, ML and MM sites .....................................
... 41
Figure 3.8a: Streaming Potential, AM site.......................................
...... 42
Figure 3.8b: Normalized Streaming Potential, ML site................................
... 42
Figure 3.9: Normalized Streaming Potential, MM and AM sites ................................ 43
Figure 3.10: Phase 2 Thickness ......................................................
.. ........ 44
Figure 4.1: GAG Standard Curve ...................................................................................
51
Figure 4.2: Hydroxyproline Standard Curve ........................................
..... 52
Figure A. 1: Static Load and Displacement Control................................
..... 57
Figure A.2: Normalized Dynamic Stiffness and Streaming Potential ......................... 58
Figure B.1: Static Stress, Orientation Bias .....................................................................
61
Figure B.2: Dynamic Stiffness and Streaming Potential, Orientation Bias.................62
Figure C. 1: Dynamic Stiffness by Applied Displacement.....................64
Figure D. 1: Stress Relaxation, Chart Recorder.................................
...... 65
Figure D.2: Dynamic Sweep, Chart Recorder .......................................
..... 66
Figures E.1-E.15: Phase 2 Plots without Chart Recorder Data.............................. 67-81
List of Tables
Table 4. 1: GAG Content ........................................................................
................. 50
Chapter 1
Introduction and Background
1.1 Osteoarthritis
Joint and bone disorders affect over 80% of people over 55 years severely restricting their mobility and quality of life. Osteoarthritis (OA) is the most prevalent of these disorders and society
benefits greatly any time additional insight is obtained. Presently, there is no effective way to
identify changes in the articular cartilage in the early progression of OA. The use of electromechanics in developing endpoint markers to quantify the lesion growth in OA will aid in evaluating
the efficacy of pharmaceuticals created to treat it. In this project, the course of OA was studied in
guinea pigs by observing the changes in electrokinetic properties of articular cartilage in order to
create an evaluation tool and to gain insight on the development of the disease
The three parts of the joint, the cartilage, the synovial fluid, and the bone all change with the progression of OA. The articular cartilage and the synovial fluid combine to lubricate the joint and
absorb the compressive impact due to regular movement. A change in any of these three parts
will cause a deformation in the shape of the joint resulting in uneven, usually painful, stresses
(Collins et al. 1982).
A common cause of OA is the degradation of the articular cartilage covering the ends of the
bones in the joints, most commonly the knees. Visually, the advent of osteoarthritis appears in
the form of a fibrillated lesion on a central part of the cartilage. Because of the lack of available
human tissue samples, a useful model in studying the disease is the Hartley strain of guinea pigs
which develops osteoarthritis at an early age. These guinea pigs begin to show signs of a lesion
on the middle medial portion of the tibial articular cartilage by around 6 months (Figure 1.1).
In this thesis, we will discuss many aspects of electromechanical and biochemical testing of the
articular cartilage of guinea pigs. In order to get a better understanding of the results of those
tests, we must first cover the basic structure of cartilage and some of the theory behind
biomechanical testing.
Figure 1.1 An scanning electron micrograph of the tibial plateau from a guinea pig joint.
1.2 Cartilage Structure
The cancellous ends of bones are covered with articular cartilage, a thin, gel-like connective tissue. Articular cartilage consists primarily of chondrocytes, collagen, proteoglycans, and water.
1.2.1 Chondrocytes
Chondrocytes are cells which are responsible for the synthesis and maintenance of the the extracellular matrix. They migrate throughout the extracellular matrix (ECM) weaving collagen fibers
and the proteoglycan network. Various growth factors such as IGF-1 and TGF-B have been
shown to stimulate chondrocytes to increase ECM synthesis.
1.2.2 Collagen
Collagen is not single protein but a family of proteins that all share a similar structure of three
polypeptide chains wrapped in a a triple helical configuration. This family of supermolecules
makes up a large portion of the ECM of connective tissue. Type II is the most prevalent kind of
collagen in articular cartilage and the fibrillar nature of these molecules gives cartilage much of its
tensile strength. Collagen content typically stays constant throughout the onset of OA so it is a
useful measure with which to normalize GAG content. Since hydroxyproline is a constituent of
collagen, collagen content is usually found by assaying for hydroxyproline.
1.2.3 Proteoglycans
Proteoglycan aggreagates (2x108 Da) are macromolecules aggrecans (PG monomers) bound to a
hyaluronic acid backbone aided by link proteins. A variety of sulfated glycosaminoglycan (GAG)
chains (mainly chondroitin-4-sulfate, chondroitin-6-sulfate, and keratin sulfate) are bound to the
aggrecan core proteins with the relative amounts varying with the age and location of the cartilage.
The GAGs form a network of negative fixed charges which is balanced by a slight excess of positive ionic charge in the intratissue fluid (resulting in a swelling pressure). The repulsion of the
negative charges in the GAG network contributes to the equilibrium modulus of the tissue. When
the cartilage is compressed, the fluid and accompanying positive ions are convected past the fixed
negative charge groups. The slight separation between positive and negative charge that is
induced by the flow gives rise to a streaming potential gradient within and across the tissue.
1.3 Lesion Development
The primary method of assessing OA in humans has been by visual and histochemical grading of
lesion development. As mentioned previously, the spontaneously occurring OA in the Hartley
guinea pig strain we are studying is characterized by the appearance and growth of a lesion on the
middle medial portion of the tibial articular cartilage. With time, the lesion expands in area and
thickness. In our project, particular attention was paid to the electromechanical properties of the
cartilage properties at the lesion site.
1.4 Objective
The goal of this project was to first quantify the ability to perform electrokinetic indentation measurements in a nondestructive manner;, second to determine the sensitivity of these techniques to
the age-related changes in the development of OA in the guinea pig model, and third to compare
those changes in electrokinetic properties to changes in biochemical ones.
1.4.1 Phase 1
1.4.1.1 Nondestructive Indentation
The nondestructive nature of the indentation test was quantified by determining the threshold of
the static and dynamic strain that can be applied before damage to the cartilage results (as visualized by SEM micrographs).
1.4.1.2 Left/Right Differences
Using ten joints (5 animals) tested for the above mentioned electromechanical properties, we
determined if a difference exists between the left and right joints. The purpose of this step was to
identify if the disease intiates and develops in one limb prior to the other. Additionally in this
phase, the electromechanical testing protocol was further developed, and the number of joints for
obtaining statistical power in future phases studied.
1.4.2 Phase 2
Three groups (3, 8 and 12 months of age) of 15 guinea pigs each were tested for their electromechanical properties and will be tested for biochemical properties. Preliminary biochemical
assays on hydroxyproline and GAGs were performed. We determined whether any difference in
electromechanical properties existed between the age groups. The ultimate objective is to correlate electromechanical measures with unconfined compression, histological, physiochemical, and
biochemical markers and compare these measures with current methods of determining lesion
severity (e.g. Evan's Blue, SEM micrograph visual grading).
1.4.3 Phase 3
In order to identify the earliest age at which electromechanical properties significantly change,
more closely aged groups of guinea pig joints will be tested depending on the results of Phase 2.
1.4.4 Phase 4 (Pilot Compound Studies)
If the previous phases prove successful in determining clear markers characterizing OA, the electromechanical and biochemical properties of joints of animals treated with Proctor & Gamble
compounds will be measured.
1.5 Overview
This thesis covers work done on the first two phases of the project outlined above. Chapter 2 discusses the experimental protocol and the results of Phase 1. Chapter 3 covers the electromechanical tests done on the three age groups of guinea pigs for Phase 2. The biochemical measurements
for Phase 2 are now underway; chapter 4 covers preliminary work that has been done to adapt the
assays normally used for bovine cartilage to be used for guinea pig cartilage. The Appendices
cover some important points with regards to (A) the choice of displacement control versus load
control, (B) a possible mechanical bias in our testing system, (C) justification for a normalization,
and (D and E) some basic graphical data and waveform plots. In the last chapter, some new ideas
for testing and some ambiguities found in Phase 1 and 2 are suggested for exploration in the
future.
Chapter 2
Left versus Right Differences, Phase I
2.1 Introduction and Summary
The goal of this Phase 1 study was to lay the foundation for the next three phases by
determining whether the indentor probe damages the cartilage during the experiment and
whether there are differences between the progression of osteoarthritis on the left and right
joints. The damage experiments were conducted in August 1996. Procter & Gamble used SEM
to determine that no damage was visible in the areas the indentor probe tested. Also, on one of
our joints, we ran the experimental protocol through twice at the same site to determine whether
any damage occurred after the first series of compressions. We found that the second set of data
matched the first quite closely. Taken together, we feel confident that the indentor probe does
not damage the cartilage during testing.
The next step of Phase 1 was to examine whether any differences exist between the left and right
joints of the guinea pigs. We tested the electromechanical properties of the tibial plateau for
five animals (10 joints) at the age of eight months. Our experiments allowed us to calculate
static stiffness, dynamic stiffness, streaming potential, and thickness for each joint.
The results indicate that there were no statistically significant differences between left and right
joints for all the properties we tested, including static and dynamic stiffness, streaming potential
and thickness. Since the number of usable samples was rather small, our standard errors were
quite sensitive to outliers. This could be a result of animal to animal variations.
2.2 Materials and Methods
2.2.1 Dissection and Inhibitor Solution
Proctor & Gamble Laboratories, supplied 5 pairs of joints from 8 month old Hartley guinea
pigs. These joints were shipped frozen with dry ice and then stored wrapped in gauze and
aluminum foil at -80 OC. On the days of testing, we unwrapped the joint and defrosted it in PBE
(phosphate buffered EDTA) for 30 minutes to 1 hour. The beaker was kept in a bath of warm
tap water (approximately 30 OC) at room temperature. After defrosting, we then dissected the
tibia from the knee joint. The tibia was sawed about 1 inch down from the joint surface to fit it
into the mounting chamber.
To prevent enzymatic degradation of the cartilage during testing, the joint was immersed in 30
ml of a proteinase inhibitor "cocktail" (10 mmol EDTA, 1 mmol PMSF, 0.001 mmol pepstatin
A, and 0.001 mmol leupeptin).
2.2.2 Mounting Chamber and Dynastat
The tibiae were then mounted in the testing chamber (Figure 2.1). The chamber allows five
degrees of freedom to position the indentor probe normal to the desired testing sites. The joint
fits inside a V-groove mount which is then placed within the chamber.
Once the joint is securely mounted, then the chamber is placed in a Dynastat mechanical
spectrometer. For this series of experiments, we chose displacement control mode. A desired
displacement was applied and the resulting load and streaming potential was measured. These
output signals were fed to a computer, which recorded and processed the signals using the
program DYNSSP. The data were also measured using a chart recorder as a backup for the
computer and to give us a real-time picture of the cartilage is behavior (Appendix D).
The testing probe is a Plexiglas indentor with a contact area of 0.79 m2 (Figure 2.2) which
enables "traditional" biomechanical measurements to be made on the intact joint surface. In
addition, a silver wire with a silver-chloride electrode tip is mounted centrally within the
Figure 2.1: Mounting chamber for the guinea pig tibia. (figure is to scale).
indentor probe. We load this probe in the Dynastat and displace it statically and then apply a
sinusoidal displacement. The Dynastat measures the resulting load. The contact point contains
a small silver chloride electrode. The indentor probe can thereby sense the potential difference
between the cartilage surface directly under the indentor and a reference electrode in the bath.
This potential difference is called the streaming potential.
Figure 2.2: Indentor probe used to measure load and streaming potential.
Figure 2.2: Indentor probe used to measure load and streaming potential.
2.3 Electromechanical Testing Protocol
The static and dynamic responses of the tibial articular cartilage were measured at three points:
the middle lateral (ML), the middle medial (MM), and the anterior medial (AM) (Figure 2.3).
At each site, we used the Dynastat to compress the cartilage by a given displacement. After
each static compression, we measured the static load. For applied dynamic displacements, both
the dynamic load and the streaming potential were measured.
AM
A1 Nil
MM
I
I..
Figure 2.3: Representation of the tibial plateau with the tested sites label on the surface.
2.3.1 Static Stiffness
To calculate static stiffness, the static load was measured at three distinct displacements: 25 ptm,
37 pm, and 49 pm. Initially, a preload of approximately 30 grams was used to be sure the
indentor probe was in contact with the cartilage surface before the experiment began. For each
static offset, the displacement was set to the desired value and the ensuing stress relaxation was
monitored for 15 minutes. When each step in displacement was applied, an initial peak in load
occurred followed by an exponential decay in stress (Appendix D).
To calculate the static stiffness, the static stress-strain curve was plotted and a line was fit to the
linear portion of the this curve. In some cases, strain-hardening of the material was observed as
the displacements increased so the initial slope between the 25 pm and 37 pm displacements
was used to represent the stiffness.
2.3.2 Dynamic Stiffness
After the 37 pm static compression, a dynamic strain was applied to the cartilage using a sweep
of sinusoids at frequencies of 1.0, 0.8, 0.05, 0.03, 0.02, 0.01, 0.08, 0.05, 0.03, 0.01, and 0.005
Hz (Appendix D). The amplitude of the sinusoid was 5 pm. At each frequency, the amplitude
of the load over four sinusoidal periods was averaged these amplitudes and used as the load for
the dynamic stiffness calculation. The dynamic stiffness (equilibrium modulus) was computed
as:
E-
P/A
u/8
where P is the dynamic load amplitude, 8 is the thickness, A is the area of the indentor probe tip
(which is in contact with the cartilage surface) and u is the dynamic displacement of 5 pm.
During each frequency sweep, the phase angle of the load and total harmonic distortion of each
signal was computed. The phase angle is the angle by which the load leads the displacement
(Appendix D). The total harmonic distortion (THD) is a measure of the nonlinearity of the load
signal (Appendix D). Low THD indicates a more linear response.
2.3.3 Streaming Potential
The streaming potential response was measured simultaneously with the dynamic load generated
by the applied dynamic displacement amplitude of 5 pm. The cartilage was stimulated at each
frequency for five periods and the steaming potential signal from the probe was fed to the
computer. The phase angle and THD were also measured for the streaming potential response
(Appendix D).
2.3.4 Thickness
To calculate the engineering strain, we needed to obtain the thickness of the cartilage at each of
the sites tested. The thickness was measured directly after each electromechanical test. The
indentor probe was replaced by a needle probe, a 1 cm long needle soldered in a 2 cm copper
cylinder.
For the thickness measurement, the level of the inhibitor solution was briefly drained below the
cartilage surface. The conductance between the needle and a reference electrode in the bath was
measured. The needle probe was then lowered at a constant velocity of 5 um/sec. As the needle
probe first touched the cartilage, a short circuit (a sharp increase in conductance) occurred
between the needle and the reference electrode. As the needle was further lowered, a relatively
sharp increase in load occurred, interpreted to be the position where the needle hit the
subchondral bone. Multiplying the time interval between the increases in conductance and load
by the velocity gave the thickness of the cartilage (Figure 2.4).
Sample Thickness Measurement (Joint 151 R, Site ML)
IIIACI
1U0V
Load
0
8
5 -1000
,Displaceine
o -2000
-d
o -3000
..J
if
1 -4000
Conductance
E
--- 5000
-250
(I)
-2501
-6000
.7nnn
C)
I
I
100
200
p
300
400
Thickness, um
p
500
600
Figure 2.4: Representative thickness measurement for right joint 151 at the middle lateral site.
2.3 Results
2.3.1 Number of Samples
For the static and dynamic tests, we completed tests for 5 right joints and 4 left joints at the AM
and MM sites. At the ML site, tests were completed for 4 right joints and 3 left joints. The data
at the other sites was unable to be obtained because of an initial problem with the computer.
2.3.2 Static Stiffness
The static stress-strain curves are pictured in Figures 2.5 and 2.6 below for the three sites. To
determine the static stiffness for the left and right joints, the slopes (i.e. the equilibrium
modulus) of these lines were calculated between the 25 um and 37 um displacements. The mean
and standard error of the slopes at each site were computed and compared (Figure 2.5).
2.3.3 Dynamic Stiffness
For the dynamic stiffness (Figures 2.7 amd 2.8), the trend of left and right joints versus
frequency was plotted at each site. Each line represents the mean of one subgroup, and the
magnitude of the error bar is one standard deviation.
Static Stress vs Static Eng Strain, site AM
Right stiffness = 7.9919 +/ - 3.6367
Left stiffness = 3.4461 +/ - 0.58307
Right = --Left = -
2 .5
1..5
0.5
I,
0.05
0.15
Strain
Figure 2.5 The static stress-strain curves for the anterior medial site.
0.25
I
I
It
I
I
Right stiffness = 3.1596 +/ - 0.64509
3.5
-
Left stiffness = 2.2314 +/ - 0.50547
Right=--Left = -
2. 5
..
0.5
. .. . .
.
.
.
I
I
I
n
.
0.05
0.15
!
0.25
Strain
(a)
Static Stress vs Static Eng Strain, site ML
a
Right stiffness = 2.04 +/- 0.91556
3.5
Left stiffness = 2.2817 +/ - 1.1625
E
Right = --Left = -
0.5
.....
...
-
0.05
0.15
Strain
0.25
(b)
Figure 2.6: The static stress-strain curves for (a) middle medial and (b) middle lateral sites.
Stiffness vs Log(Frequency), Site AM
100
Riaht = x (n=31
90
80
70
M 60
()
• 50L
40
30
20
10I
n
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
0
0.5
Figure 2.7: The dynamic stiffness vs frequency for the anterior medial sites of the left and right
joints.
The dynamic stiffness was then normalized to the static stress upon which the dynamic load was
superimposed (Figure 2.9 and 2.10). Since the thickness of the tissue varies widely from site to
site and animal to animal, the static load also varied. This difference could have an effect on the
dynamic load readings. Therefore, the dynamic stiffness was normalized by dividing by the
static stress for each data set.
Stiffness vs Log(Frequency), Site MM
100
Right = x (n=4)
E
90
70
Left = o (n=4)
-
-
50
E
20
-
±
-..
10
U
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
0
(a)
Stiffness vs Log(Frequency), Site ML
1 Aft3
Right = x (n=3)
120
1-
Left = o (n=4)
100
E
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
0
(b)
Figure 2.8: The dynamic stiffness versus frequency for the (a) MM and (b) ML sites.
Normalized Stiffness vs Log(Frequency), Site AM
-1.5
-2
-2.5
-1
-0.5
Log(Frequency), Hz
0.5
0
(a)
Normalized Stiffness vs Log(Frequency), Site MM
100
90
80
70
Right = x (n=4)
-
Left = o (n=4)
-
-
-
60
50
40
30
20
-
-
-
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
0
(b)
Figure 2.9: The normalized dynamic stiffness versus frequency for the (a) MM and (b) ML sites.
Normalized Stiffness vs Log(Frequency), Site ML
I
I
100
-
Right = x (n=3)
90
-
Left = o (n=4)
80
-2
70
2 60
Co
S50
40
30
20
10
n
-2.5
I
-2
I
-1.5
I
I
-1
-0.5
Log(Frequency), Hz
I
0
0.5
Figure 2.10: Normalized siffness versus frequency for the ML site.
2.3.4 Streaming Potential
To plot the streaming potential, the voltage was normalized by the dynamic strain. Therefore,
the units of the y-axis are in milli-volts per percent strain. In Figure 2.11 and 2.12, these values
are plotted for the range of frequencies tested.
As before, the streaming potential data was normalized by the static stress over which the
sinusoidal displacement was applied (Figure 2.12b and 2.13).
,,...
Streaming Potential vs Log(Frequency), Site AM
1
Right = x (n=4)
Lnft= n In=41
0.08
0.06,
0.04
>
0.02-
cr -0.02
E
-0.06
-0.08
-0.1
-2.5
-2.5
·
-2
-1.5
·
·
·
-1
-0.5
Log(Frequency), Hz
0
(a)
Streaming Potential vs Log(Frequency), Site MM
Vi J
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
0
0.5
(b)
Figure 2.11: The streaming potential per percent strain vs frequency for the (a) AM, and (b) MM
sites.
Streaming Potential vs Log(Frequency), Site ML
0.1
I
III
I
Right = x (n=3)
0.08
Left = o (n=4)
0.06
C
0.04
> 0.02
E
E
0
0
a.
cn-0.02
.c
E
2 -0.04-0.06-0.08
-U..1
2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
0
(a)
Normalized Streaming Potential vs Log(Frequency), Site AM
0.35 7
Right = x (n=4)
0.3
Left = o (n=4)
-
0.25
0.2
0.15
E
0.1
0.05
-0.05
-
-n. i2
-2.•
5
-2
- -1.5
-I
I
-I -- - -0
-1
-0.5
Log(Frequency), Hz
0.5
(b)
Figure 2.12: The streaming potential per percent strain versus frequency (a) for the ML site and
(b) the normalized streaming potential per percent strain versus frequency for the AM site.
Normalized Streaming Potential vs Log(Frequency), Site MM
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
-0.05-U.1
-2'.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
0
(a)
Normalized Streaming Potential vs Log(Frequency), Site ML
fl4
0.35
Right = x (n=3)
0.3
Left = o (n=4)
0.25
0.2
0.15
0.1
0.05
0-0.05
-n1 I
-2.5
.
-2.5
I
-2
-2
I
-1.5
-1.5
-1I
-1
Log(Frequency), Hz
-0.5
-0.5
0I
0
0.5
(b)
Figure 2.13: The normalized streaming potential per percent strain for (a) the MM site and (b)
the ML site.
2.3.5 Thickness
The thickness was measured for the three sites and separated into left and right subgroups
(Figure 2.14). The bar represent the mean thickness for the subgroups. The magnitude of the
error bars is plus and minus one standard deviation.
2.4 Discussion
2.4.1 Static Stiffness
The Student's (two-way) T-test indicated that there was no significant difference between the
static stiffnesses of the left and right joints (P = 0.2061 for the AM site , P = 0.2872 for the MM
site, and P = 0.8737 for the ML site).
Thickness by Site
· rrr
0
1
2
3
site
4
5
6
7
Figure 2.14: The means and standard deviations of thickness for the three sites with left and
right joints separated.
2.4.2 Dynamic Stiffness
The Student's T-test revealed no significant difference at all sites between the left and right
joints for the dynamic stiffness measurements (Figures 2.7 and 2.8).
Since the experiments were performed under displacement and not load control, figures 2.7 and
2.8 represent dynamic load responses at different static stresses (possibly due to a difference in
the amount of preload applied to the joint prior to the experiment). To normalize the dynamic
stiffnesses, the dynamic stiffnesses were divided by the static stresses to which the dynamic
displacements were added. With the normalized data, no statistically significant left/right
differences were found at any site.
2.4.3 Streaming Potential
For the streaming potential measurement, the T-test showed no significant difference between
left and right joints. After static stress normalization (see 2.4.2), the T-test again showed no left/
right difference.
2.4.4 Thickness
T-tests on the thickness measurements revealed no left/right differences at any of the three sites
(P = 0.44 for the AM site, P = 0.29 for the MM site, and P = 0.81 for the ML site). The
differences between the three sites were also examined. The AM and ML sites did not show
statistically significant differences in thickness; however, the MM site was distinctly thicker
than the AM and ML sites. Since the lesion is located at the MM site, this thickness difference
warrants further examination in the future.
2.4.4 Sources of Error
For several of the tests, the standard deviation was quite high due to the low number of samples
and the learning curve of the experimental protocol. In many of the plots, there is a difference
between the means of the left and right joints; however, in most cases the means are within the
standard deviation of the two subgroups.
One source of error is the difference in preload applied to the joint prior to the experiment. In
most cases, this preload ranged from 10-30 grams. However, it is difficult to tighten up this
range because we are unsure of the point at which the indentor probe begins to deform the
cartilage. Also, since the cartilage is so thin, it is physically difficult and time-consuming to
consistently preload before the experiment. The re-normalization of our dynamic stiffness and
streaming potential values should decrease but not eliminate this error. Of course, joint-to-joint
and animal to animal variability in these cartilage electromechanical properties contribute to the
error.
During Phase 2, this preload value was more carefully monitored. A potential solution is to run
the experiment in load control instead of displacement control (Appendix A)
2.4.5 Possible Bias in Testing
One interesting trend is that the mean static and dynamic stiffnesses are consistently higher on
the left ML sites and the right AM and MM sites. This trend could indicate a bias in the testing
procedure, since the joint is normally positioned in the chamber with the posterior side toward
the experimenter. Since the joint chamber has five, rather than six, degrees of freedom, the left
and right joints might not be positioned exactly the same when a given site is tested.
Because of this possible bias in our experimental process, we tested one of the Phase 2 joints in
its normal position first, rotated the chamber 180 degrees so that the anterior site faces the
experimenter and re-ran the experiment. This procedure and the results which indicated no
obvious bias are discussed in Appendix B.
2.5 Conclusion
In analyzing the data from the left/right difference experiments, several weaknesses in the
testing protocol have become apparent. The potential bias of our testing apparatus mentioned
above in section was resolved (Appendix B). Additionally, the difficulty of maintaining a
consistent preload must be eliminated, possibly by switching to load control instead of
displacement control. Changing to load control turned out not to be feasible because of the
additional time constraints it imposes (Appendix A). In view of the impracticality of load
control and the apparent lack of any physical bias in the system, we proceeded to Phase 2 with
relatively small changes in the testing protocol.
Chapter 3
Electromechanical Differences Between Age Groups, Phase 2
3.1 Introduction
The purpose of the second phase was to determine, first, any changes in electromechanical properties between different age groups of guinea pigs; second, any changes in biochemical properties; and third, any correlation between the two.
After looking at the results of Phase 1, the experimental protocol was altered to make more economical use of testing time and to add more detail to the static measurements. Otherwise, the
dynamic measurements of Phases 1 and 2 were taken essentially the same way. Also, during
Phase 2, possibility of changing to Load Control from Displacement Control was explored and
rejected. The results of that experiment are presented in Appendix A.
This chapter covers only the electromechanical testing part of Phase 2 since, as was mentioned in
Chapter 1,the bulk of the biochemical assays have yet to be performed; however, preliminary biochemical assays are discussed in the next chapter. Overall, for the dynamic (normalized and
unnormalized) and static stiffness tests, differences in site but no differences in age were found at
the 95% significance level. The ANOVA revealed a statistically significant difference for both site
and age for the normalized streaming potential data (P=0.043) though post hoc analysis did not
show dignificance at a 95% leve. For the thickness measurement, a difference exists for both age
and site.
3.2 Methods and Materials
A total of 45 legs (15 per age group of 3, 8, and 12 months) were received from Procter & Gamble. All joints tested were from the right legs. Procter & Gamble kept the left legs for histological
studies. The guinea pig legs were dissected and the thickness was measured in exactly the same
manner described in the previous chapter. This part of the study was blind in that Procter & Gamble did not inform us of the groupings of the joints until the electromechanical experiments were
completed.
All the physical testing apparatus and the inhibitor solutions remained constant between Phase 1
and Phase 2. The major changes were the reduction of the stress relaxation period and the addition of eight more static loads to give more detail to the static compression characteristics of the
guinea pig cartilage.
3.2.1 Stress Relaxation
In Phase 1 testing, the cartilage was allowed to stress relax for about 15 minutes. Because a total
of 45 joints were to be tested, cutting down the stress relaxation time would allow us to test two
joints a day instead of only one. After examining the exponentially decaying static load response,
the time constant was determined to be approximately 30 seconds (Appendix D). This is consistent with the fact that the cartilage is very thin, and the relaxation time is proportional to the
square of the thickness. Instead of a relaxation time of 15 minutes, the time chosen was six times
the time constant or about 3 minutes.
3.2.2 Static Stiffness Tests
As can be seen in the static stiffness plots in the previous chapter, the tests did not reveal much
detail since the curves consisted of only three points. In order to add more resolution and thereby
get a better sense of the region where the stress-strain curve becomes non-linear, eight more static
displacement points were added. Instead of displacements of 25, 37, and 42 um, the Phase 2
joints were tested for displacements of 5, 12.5, 20, 25, 20, 37.5, 45, 50, 55, 60, 65, 70, and 75 um.
The cartilage was first indented to 5 um, allowed to stress relax for three minutes, indented to 12.5
um, etc., until the displacement reached 37.5 um at which time the cartilage was dynamically
stimulated as in Phase 1. After the dynamic measure was taken, the joint was further tested at
progressively increasing static displacements. After all static tests for a joint were completed, the
thickness was measured at the testing site.
3.3 Results
3.3.1 Number of Samples
Although all 45 joints were electromechanically tested, due to a computer error, the electronic
form of the data obtained from 14 joints was lost. Fortunately, the chart recordings of the data
were always printed during the experiments and some (4 to 10 joint/site sets) of the data was
retrieved from the 14 joints.
3.3.2 Age-Group Code
After the experiments were completed, the age-groupoing code was broken: G3 = 3 months, G1 =
8 months, and G2 = 12 months.
Static stress vs strain, site AM
G1 = ... n=13
G2= .. n= 11
n=11
G3=
/·
I
Strain, percent
Figure 3.1: The stress-strain curves at the AM site for the three age groups.
Static stress vs strain, site MM
G1 stiffness = 1.5197 +/ - 0.20811
G2 stiffness = 1.5651 +/ - 0.28663
G3 stiffness = 1.3697 +/ - 0.14967
G1 = ... n = 13
G2= ._. n=12
G3 =
n = 11
8
c
6
(0
4
2
I
.
(I
. . ..·
0
i
I
0.1
0.2
·
0.6
0.5
0.4
Strain, percent
0.3
..
0.8
0.7
0.9
1
(a)
Static stress vs strain, site ML
G1 stiffness = 1.1685 +/ - 0.27208
G2 stiffness = 1.467 +/ - 0.26549
G3 stiffness = 1.7159 +/ - 0.51111
G1 = ... n=13
G2 = ._. n = 13
G3=
n=12
Co
a_
9D
0
0.1
0.2
0.3
0.4
0.5
0.6
Strain, percent
0.7
0.8
0.9
(b)
Figure 3.2: The static stress-strain curves for the (a) MM and (b) ML sites.
1
3.3.2 Static Stiffness
To calculate the static stiffness in the curves in Figures 3.1 and 3.2, the first 6 points were taken, a
line was fit to them (least squares method) and the slope computed. The first six points represent
the stress vs strain curve between 5 um and 37.5 um displacement. The mean and standard errors
of the slopes were calculated and compared at each site and for each age group.
3.3.3 Dynamic Stiffness
For the dynamic stiffnesses (Figures 3.3 and 3.4), the trend of the different age groups versus the
logarithm of the frequency was plotted at each site. Each line represents the mean of each group
and the magnitude of vertical bars represent plus or minus one standard error at each frequency.
The dynamic stiffness was then normalized to the static stress upon which the dynamic load was
superimposed as was done in Phase 1 (Figures 3.5 and 3.6). The justification for the normalization is discussed in Appendix C.
Stiffness vs Log(Frequency), Site ML
·
ItI
·
·
G1 = X; n = 13
G2= O; n= 11
50
G3= +;n= 12
-
40
-
30
-
20
E
U
-2.5
VI
I
I
-1
-0.5
Log(Frequency), Hz
0
I
-2
-1.5
Figure 3.3: The dynamic stiffness versus frequency plot for site ML.
-60
Stiffness vs Log(Frequency), Site MM
G1 = X; n = 13
E
G2 = O; n = 12
G3= +; n= 11
Ca
n
d&
0
'p
v-
B 30
U)
20
E
A
L
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
0
0.5
(a)
Stiffness vs Log(Frequency), Site AM
60
50
_ 40
30
20
10
|1I
-2
-1.5
-1
Log(Frequency), Hz
-0.5
0
(b)
Figure 3.4: The dynamic stiffness versus frequency plots for the (a) MM and (b) AM sites.
Normalized Stiffness vs Log(Frequency), Site ML
11
I
100
i
i
I
G1 = X; n= 13
G2= O; n= 11
G3= +; n= 12
90
80
c 70
0
0
i
60
50
40
30
20
10A
I
-2.5
-2
I
-1.5
I
-1
Log(Frequency), Hz
I
-0.5
I
0
0.5
0
0.5
(a)
Normalized Stiffness vs Log(Frequency), Site MM
110
1
100-
G1 = X; n = 13
100
u -
1
G2 = O; n =12
GIXn13
-
80
. 70
0
60
•
50
40
30
20
02.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
(b)
Figure 3.5: The normalized dynamic stiffness versus frequency at the (a) MI and (b) MM sites.
Normalized Stiffness vs Log(Frequency), Site AM
100
90
80
E
70
60
E
30
E
t-I--
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
0
Figure 3.6: The normalized dynamic stiffness versus frequency at site AM.
3.3.4 Streaming Potential
Streaming potential was plotted normalized by the dynamic strain as was done in Phase 1 (Figures
2.8 and 2.9a). As with the dynamic stiffness graphs, the streaming potential was normalized by
the static stress over which the sinusoidal displacement was applied (Figures 3.8b and 3.9).
Streaming Potential vs Log(Frequency), Site ML
nI I ~r
0.04
0.03
G1 = X; n = 13
G2 = O; n = 11
G3= +; n= 12
0.02
0.01
E
_n rt
-2.5
L
-2
-2.5
-1.5
-1
-0.5
Log(Frequency), Hz
0
(a)
Streaming Potential vs Log(Frequency), Site MM
n n0
0.04
·
·
·
·
-1
-0.5
Log(Frequency), Hz
0
K
0.03
G1 =)
G2=(
G3 =
0.02
0.01
.-n -. n
I
2
-2.=
5
I
-2
I
-1.5
(b)
Figure 3.7: The streaming potential per percent strain at the (a) ML and (b) MM sites.
Streaming Potential vs Log(Frequency), Site AM
0.05
0.04
-
0.03
-
G1 = X; n =13
0.02
G2 =
G3 =
H
0.01
-
--00A12.5
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
0
(a)
Normalized Streaming Potential vs Log(Frequency), Site ML
0.08
G1 =X;n=13
G2= O; n= 11
0.06
G3= +; n= 12
E
0.04
0.02
I
-·
II· II'1
·
-2.5
I
-2
II
T
I
-1.5
-±zzir
I
I
I
-1
-0.5
Log(Frequency), Hz
Figure 3.8: The (a) streaming potential per strain versus frequency at site AM and (b) the normalized streaming potential per strain versus frequency at site ML.
Normalized Streaming Potential vs Log(Frequency), Site MM
I
I
0.08 i
Gi =
G2=(
0.06 7
G3=-
0.04
0.02 1-
-n no
-2.5
-1.
-2
-2
-1.5
--.
I
-1
-0.5
0
I
0.5
Log(Frequency), Hz
(a)
Normalized Streaming Potential vs Log(Frequency), Site AM
0.1
0.
n A08
G1 =X;n=13
G2= O; n = 11
E
- 0.06
G3= +; n= 10
0.04
E
7 0.02
Z
A002
-2. 5
-2
I
-1.5
- - -
- - -
-1
-0.5
Log(Frequency), Hz
0
(b)
Figure 3.9: The normalized streaming potential per strain versus frequency at the (a) MM and (b)
ML sites.
3.3.5 Thickness
The thickness was measured at each of the three sites on each joint (Figure 3.10). The bars represent the mean thicknesses for each group and site and the vertical lines represent plus and minus
one standard error.
Thickness by site and age group
0UU
450
E
400
350
-
Bars 1-3 = site AM; G1-3; n = 13, 12, and 10.
Bars 4-6 = site MM; G1-3; n = 13, 12, and 12.
Bars 7-9 = site ML; G1-3; n = 13, 13, and 14.
-
T
300
E
250
|
I
T
7V
E
200
T
-
150
-
100
-
T
I
T
T
1
T
J_
T
50
E
I
m
1
m
2
3
11
4
5
6
7
Groups of joints by site and age
I8
9
Figure 3.10: A bar graph of the cartilage thicknesses at three sites and age groups.
3.4 Discussion
3.4.1 Static Stiffness
An two-way Analysis of Variance (ANOVA) was performed by age and site for all electromechanical measures taken at a 95 % significance level. The results for the static stiffness data show
that static stiffness varies by site, but not age (site: P = 0.2313, age: P = 0.1610) with no interac-
tion effects (P = 0..2313). The Tukey and Student-Newman-Keuls Tests showed that the AM site
was significantly stiffer than the other two sites. This result is visually apparent in Figures 3.1 and
3.2 since the curves in Figure 3.1 are so much steeper than in the other two graphs. The higher
stiffness at the AM site could be due to effects from the thinner cartilage in that area.
3.4.2 Dynamic Stiffness
An ANOVA on the dynamic stiffness data showed no significant difference due to age, but a variablity due to site with little interaction effects (site: P= 0.0001, age: P=0.3303, interaction:
P=0.6165). The Tukey and Student-Newman-Keuls test showed site AM with a higher stiffness.
After normalization, the ANOVA results were the same (site: P = 0.087, age: P = 0.0717, interaction: P = 0.8753). Post hoc analysis revealed that the ML site was stiffer.
3.4.3 Streaming Potential
The ANOVA of the streaming potential data showed a difference by site, and again, no difference
by age (site: P = 0.0066, age: P = 0.0830, interaction: P = 0.0755). After normalization by static
stress, there is both an age and a site difference (site: P = 0.008, age: P = 0.0427, interaction: P =
0.0643). Although the ANOVA revealed a difference, neither the Student-Newman-Keuls nor the
Tukey Test showed a significant difference with age at each site at the 95% level. Clearly, from
Figures 2.8 and 2.9, the ML site exhibits the lowest streaming potential as was confirmed by post
hoc analysis.
The lower streaming potential at the ML site is most probably related to the relative thinness of
that site to the AM and MM sites. Also, it is important to note that when retrieving the lost computer data from the paper plots, some of the streaming potential data was unretrievable due to an
streaming potential that was too small to be detected visually. Since most of the irretrievable
streaming potential data was at the ML site, the results shown in Figures 3.7a and 3.8bcould actually be skewed too high.
3.4.4 Thickness
A two-way ANOVA revealed a significant difference with both age and site with little interaction
effects (age: P = 0.0079, site: P = 0.001, interaction: P = 0.5421). The Student-Newman-Keuls
and Tukey Tests confirmed these results showing that age group 3 (3 months) is the thinnest and
the MM site is the thickest. This comes as no surprise since from Watson's MRI experiments, the
younger guinea pigs appear to have the thinnest cartilage (Watson 1996).
3.5 Conclusion
The statistical analyses performed to date did not show significant age related changes with the
exception of the normalized streaming potential (by ANOVA only). The relatively large standard
error may be associated with the indentor-joint geometry, the thickness of the cartilage, and the
natural variations between animals. We note, however, that the data on stiffness and streaming
potential should be further correlated (normalized) to the biochemical measures on a joint-byjoint basis. These biochemical assays are now in progress. Based on those results, we will be in a
better position to know whether more animals in each age or a wider range of ages may be necessary to best characterize the changes in physical and biochemical properties of this guinea pig cartilage OA model.
Chapter 4
Biochemical Assays
4.1 Introduction
As mentioned in Chapter 1, Glycosaminoglycans (GAGs) play a major part in giving articular cartilage its tensile strength. The GAG content of cartilage is also strongly related to the strength of
the streaming potential. In studies of the bovine cartilage of OA, small cylindrical disks of articular cartilage are tested electromechanically in a geometrically simple compression chamber. The
sample is then assayed for GAG content and the GAG to wet or dry weight ratio is reported. The
central difficulty in replicating this technique in guinea pigs is one of joint and tissue size. The
task of harvesting a uniform cylindrical plug of cartilage is virtually impossible; therefore we try
to shave off with a scalpel as much cartilage from the joint surface as possible. Since the resulting
sample is still very small and the tissue water content is therefore difficult to measure, it is difficult
to standardize the GAG content by normalization to wet weight.
A potentially more definitive approach is to normalize GAG content in cartilage is by total collagen content, since the collagen content remains fairly constant. It is relatively simple to measure the amount of hydroxyproline in cartilage. Since hydroxyproline is a an amino acid
constituent group unique to collagen, the resulting measure of hydroxyproline is proportional to
the total amount of collagen in the assayed sample. Therefore, our solution is to assay the same
samples for both GAG and hydroxyproline and normalize the former to the latter.
These biochemical assays will be performed on the 45 joints from Phase 2, but to date we have
tested our methods on several joints in order to adapt the bovine protocols for use with guinea pig
tissue. The ratio of GAG to hydroxyproline was from 5.2 to 5.8 which is well within the expected
range for bovine samples.
4.2 Methods
4.2.1 Harvesting Samples
After the joint has been dissected and electromechanically tested, it is placed in a -20 OC freezer in
a conical vial for storage. Then the joint is removed from the conical vial and defrosted in a beaker of PBE at room temperature for 10 minutes.
After defrosting, the joint is placed under a steroscopic dissecting microscope in a large petri dish.
The joint is held in one hand so that the tibial plateau is at about a twenty degree tilt to the focal
plane of the microscope. From each joint, two samples are taken, the medial side and the lateral
side. Starting with the medial side, a #10 blade scalpel is held at a slight angle to the surface of
the cartilage. Then the cartilage is shaved off cutting from the outer edge of the plateau to the center line. All the cartilage from each side is removed and placed in two cryovials and frozen at -200
C.
4.2.2 Papain Digestion
The cartilage samples now must be enzymatically digested before they can be assayed. The standard method is to use papain to digest the samples. After digestion, both the GAG assay and the
hydroxyproline assay can be performed on aliquots from the same digest solution; hence, a useful
normalization can result.
Papain is a sulfhydryl protease with a wide specificity for proteins. It remains in zymogen form
until activated by cysteine. For 10 ml papain solution, 17.6 mg of cysteine is weighed and placed
in a 15 ml conical vial. Using sterile technique, 10 ml of PBE is added to the conical and the new
solution was sterile filtered (0.22 um) into another 15 ml conical. Finally, 20 ul papain is aliquoted and the final digest solution mixed by inversion.
The cryovials containing the cartilage samples are removed from the -200 C freezer. After
defrosting for 5 minutes, 0.5 ml of the papain digest solution is added. The cryovials are then
heated in a 60 OC water bath for three hours or until the cartilage is completely digested. After
incubation, the digested cartilage is again frozen at -200 C.
4.2.4 GAG Assay
This is a general assay for sulfated GAGs. Shark chrondroitin 6-Sulfate from Sigma (C4S is also
present in amounts that vary with age) is used as the standard. To make a stock solution of 2 mg/
ml, 20 mg of C6S is added using sterile technique to 10 ml of PBE. We use 7 test tubes for the
serial dilutions to make the samples for our standard curve. To six of those tubes, 100 ul PBE is
added while the last tube will contain 200 ul of stock solution. Starting with tube containing the
stock solution, the standards are serially diluted by aliquoting 100 ul from one tube to the next.
To run the assay, 20 ul of papain digested sample or 20 ul of a standard is added to a 1.5 ml
cuvette. Then 2 ml of 1, 9 dimethylmethylene blue chloride (DMMB) dye is added to the
cuvette. The optical density at 525 nm is then read using a Perkin Elmer, Lambda 3B spectrophotometer. Typically, the standardards are read first, then the samples, and lastly duplicates of the
standards.
4.2.5 Hydroxyproline Assay
The day before the assay, the papain digested samples are hydrolyzed in 1 ml of 6N HCI (100 ul
digest into 9 ul 6N HC1) at 115 C for 16 to 24 hours in a small oven.
Six tubes are set up for the standard curve with final concentrations of hydroxyproline of 0, 0.5,
1.0, 2.0, 3.0, and 4.0 ug/ml from a stock solution of 100 ug/ml. The sample tubes are removed
from the oven and cooled to room temperature under running water. The samples are vortexed
and then 2 drops of methyl red are added under a fume hood. NaOH is added until neutralised
(around 2.0 ml of 2.5 M NaOH). Then 0.5 M HCI is added dropwise until the pink reappears and
a few drops of 0.5 M NaOH are added until the straw color returns.
To bring the salt concentration down, distilled water is added to bring the volume to 15 ml. To a
new tube, 1 ml of the solution is transferred. Under a fume hood, 0.5 ml Chloramine-T is added
to each tube (both standards and samples), vortexed, allowed to sit at room temperature for 20
minutes. Then 0.5 ml pDAB solution is added while vortexing. The tubes are incubated at 60 C
for 30 minutes.
After incubation, the tubes are cooled in water for 5 minutes. The optical density samples and
duplicates of the standards are then read at 550 nm.
4.3 Results
A total of three joints (all from 8 month old animals) were tested. Each of those joints yielded a
medial and lateral cartilage specimen for a total of six samples. We have no confidence in the
weight of these samples due to varying hydration levels. The GAG assay was performed on all
three joints. Only one of those joints was assayed for hydroxyproline.
4.3.1 GAG
The GAG standard curve is shown in Figure 4.1 where the points represent the average of the ODs
from two duplicates. The vertical bars represent plus or minus one standard error. A line was fit
using the least square method to the first five points of the curve (the linear region). The OD
readings of the six samples and the total ug GAG in each sample calculated from the standard
curve line are shown below in Table 4.1.
Joint Medial, OD
GAG, ug
Weight, g
Lateral, OD
GAG, ug
Weight, g
1
0.535
111.0
0.0042
0.442
42.8
0.0038
2
0.514
95.8
0.0038
0.408
17.5
0.0037
3
0.569
136.0
0.0049
0.432
35.1
0.0034
Table 4.1: The OD readings and calculated GAG content from three joints.
Elmer, GAG standard curve
1
0.9
0.8
E0.7
0 0.6
0.5
0.4
n03
-500
0
500
1000
1500
GAG concentration, ug/ml
2000
2500
Figure 4.1: The standard curve from the GAG assay using the spectrophotometer.
4.3.2 Hydroxyproline
The hydroxyproline standard curve is shown in Figure 4.2. Each point represents the average of
the ODs from two duplicates. The vertical bars represent plus or minus one standard error. As
with the GAG standard curve, a line was fit using the least squares method to the linear region of
the standard curve (the first 6 points). Then the concentration of the samples was calculated from
that equation and multiplied by 75 since the 0.5 ml sample is diluted 150 times when the OD is
taken.
Only joint 1 was assayed (Medial: 19.0301 ug OH-Pro, Lateral: 8.2067 ug OH-Pro). The resulting GAG to hydroxyproline ratio was 5.8344 for the medial side and 5.2125 for the lateral side.
These values are within the expected range for bovine articular cartilage.
Hypro standard curve
I
I
I
I
I
I
I
2
1.5
C
0
LO
LO
1
0.5
0
-5
I
0
5
10
15
Hypro concentration, ug/ml
I
20
25
Figure 4.2.: The standard curve from the hypro (hydroxyproline) assay using the spectrophotometer.
4.4 Discussion
Although the GAG and hydroxyproline assays described above are fairly standard for use on
bovine and other cartilage explants, we had little previous experience using them with smaller
guinea pig samples. Fortunately, for the sample size harvested and the papain digest volume chosen, we were able to get preliminary results that were within the readable range of the spectrophotometer. Also, a reasonable ratio of GAG to hydroxyproline was found (Medial: 5.8, Lateral: 5.2)
for the one joint on which both assays were tested.
One possible source of error is that in harvesting the cartilage, some bone might be included in the
sample as well. Bone also contains collagen, and so the collagen normalization might be skewed.
4.5 Conclusion
The preliminary results from the biochemical assays are encouraging enough to begin testing all
45 joints from Phase 2. The protocol described in the Methods section appears to be easily repeatable with little chance for inconsistency except at the harvesting stage. The results from the biochemical analysis should shed more light on the electromechanical data presented in Chapter 3.
Chapter 5
Future Testing
5.1 Introduction
Phase 1 of this OA animal study showed that no damage to the joint cartilage was produced by the
indentor for a wide range of loads from the indentor apparatus. Although the tested sample size
was low, the left and right joints did not appear statistically significantly different. However, the
results from the electromechanical tests of the three age groups of guinea pig joints from Phase 2
are not yet conclusive. Differences in site were detected but, for the most part, differences in age
were not. The biochemical assays must be completed in order to further normalize the streaming
potential and stiffness data.
5.2 Phase 3
The rationale behind this step in the project is to determine the age at which the electromechanical
properties change most. Conditional on the results from Phase 2 biochemistry, testing a wider
range of ages instead of a smaller range might also be helpful. Since the standard errors in the
electromechanical experiments were high, perhaps adding more joints at the age groups already
tested would better characterize the statistical distributions of the measured parameters and so
boost the confidence of the results.
Increasing the sample size should not pose a time-contraint problem since the current protocol
described in Chapter 3 allows two joints a day to be tested. Between Phase 1 and 2, more static
displacements were added to give more resolution to the static characteristic of the cartilage. For
Phase 3, more dynamic frequency sweeps could be added at different static displacements to give
more detail to the dynamic character of the joints. The procedure would be much like what was
done for the joint described in Appendix C in which the normalization method was justified. By
increasing the number of dynamic tests, the dynamic stiffnesses between joints could be more
easily compared without the normalization factor. Because the consistency of the electromechanical testing seems to be at a reasonable level, any additional change in the methods from Phase 2
to Phase 3 is not recommended.
5.3 Conclusion
Although further statistical and biochemical analyses must be performed to verify whether there
are significant differences in the physical properties with age, the Phase 1 and 2 studies have
developed a clear, well-defined electromechanical testing procedure for guinea pig joints. Also,
the preliminary results from the GAG and hydroxyproline assays show that there is a good prospect for obtaining crucial biochemical data quickly from a small amount of cartilage. With additional experiments (adding more joints to the three Phase 2 age groups, increasing the number of
dynamic stiffness tests, completing the biochemistry), the data will hopefully add to the utility of
the Hartley guinea pig OA model for evaluating pharmaceuticals for OA in the near future.
Appendix A
Load Control
A.1 Introduction
After looking at the results of Phase 1, the use of load control instead of displacement control was
examined as an alternative way to compare dynamic stiffnesses between joints. The difficulty
with displacement control is that the dynamic frequency sweeps take place at the same static displacement at all sites on all joints but, because of the varying thicknesses of the cartilage, not at
the same static load. Therefore, a possible solution is to operate the Dynastat under load control
so that the dynamic measurements might be more easily compared. The disadvantage of load
control is that the creep relaxation time would be expected to be about four times longer than the
streass relaxation time observed in displacement control. To determine the feasibility of the load
control, a joint was tested on both under load control and displacement control at the same sites.
A.2 Methods
The joint was mounted with inhibitor solution in the chamber with inhibitor solution as described
in Chapter 2. At each site, the indentor probe was preloaded onto the cartilage surface and tested
under displacement control as in Chapter 3. Then, the Dynastat was switched to load control and
then tested at 12 static loads of 10 to 120 grams in 10 gram increments with the frequency sweep
(amplitude 30 grams) at the 60 grams static load. At each static load, the sample was allowed to
"creep" (analogous to the stress relaxation under displacement control) for 15 minutes.
A.3 Results
The static stress versus strain cuves hardly changed between the displacement control and load
control (Figure A. 1). This result was to be expected since, after the sample cquilibrates in both
cases, the indentor remains lixed at a constant displacement.
The dynamic stiffness and streaming potential plots were plotted normalized to static stress as
described in Chapter 2 (Figures A.2). Since the dynamic tests under displacement control took
place under different static loads, normalization was necessary for comparison.
Static Stress vs Static Eng Strain by Location
-0.5
rC
-.
cc
H
09
-1
H
I-"
-1.5
-,
-100
-80
-40
-60
STATIC ENG STRAIN, %
-20
Figure A.1: "Load" is load control. Data, otherwise, are displacement control.
Normalized Stiffness vs Log(Frequency), Joint 280
-2
-1.8
-1.6
-1.4
-1.2
-1
-0.8
Log(Frequency), Hz
-0.6
-0.4
-0.2
(a)
Normalized Streaming Potential vs Log(Frequency), Joint 280R
Log(Frequency), Hz
(b)
Figure A.2
A.4 Discussion
Although the static stress versus strain characteristic turned out to be easily reproduceable under
load control, the dynamic measurements did not appear as comparable. For joint 280R, the ML
and AM dynamic stiffnesses were closely clustered for load and displacement control. However,
the difference between the two controls at the MM site was large (Figure A.2). The separation of
the two curves could be due to the normalization factor. It is reassuring to note that the two curves
have similar slopes.
Likewise, for the streaming potentials, the AM and ML sites were closely clustered but the MM
site measurements were different for the two controls. The results at the AM and ML sites do not
add much information since, due to the relative low thicknesses, the potential amplitude is small.
The MM site streaming potential under displacement control displays an unusual convex shape
which does not appear under load control. The only explanation, besides noise, is that the odd
shape is due to drift.
A.5 Conclusion
Because only one joint was tested under load control, no strong statement about the benefits of
load control over displacement control can be made. Comparing dynamic measurements between
joints under load control is theoretically easier and more reasonable than under displacement control. Comparing different joints tested under different controls does not appear to be possible.
The main benefit to displacement control is the relative shorter testing period ( approximately four
hours per joint as opposed to eight hours). For this reason, a switch to load control would not be
beneficial.
Appendix B
Resolution of Orientation Bias
B.1 Introduction
Because the testing chamber used for the electromechanical tests only has five and not six degrees
of freedom and the tibial plateau surface is non-uniform, a potential testing bias could appear. If
the joints are tested in exactly the same manner every time, any physical bias could skew the data.
For example, if only testing right joints, the indentor could be consistently hitting the cartilage
surface at an incline for the medial sites but not for the lateral site. During Phase 1, a trend of
higher dynamic stiffnesses on the left ML sites and the right AM and MM sites was noticed. This
trend could be naturally occurring or a result of a physical bias in the system. The hypothesis can
be tested by comparing the data from the same site with the chamber in two different orientations.
B.2 Methods
The electromechanical testing protocol as in Chapter 3 was used with one modification. After
each site was tested electromechanically and before the thickness was measured, the chamber was
rotated 180 degrees and the same site tested again. Every attempt was made to reposition the
indentor in the same location after realignment of the chamber.
B.3 Results
In Figure B.1, the static stress curves reveal almost no difference in the two orientations of the
chamber. Each curve is labelled am, am2, mm, mm2, ml, or m12 where the "2" indicates the second testing at that site.
Some separation is seen for the dynamic stiffness curves (Figure B.2a). The nun curve is above
the mrm2 curve and the alm2 curve is above the am curve. The streaming potential (Figure B.2b)
shows a strong difference between orientations except at the nml site.
B.4 Discussion and Conclusion
The static and dynamic stiflifness plots indicate no physical bias in the system whereas the streaming potential data appears to argue the opposite. However, the streaming potential data is
extremely sensitive so, in the face of the strength of the other two measures, the orientation bias
does not seem to exist.
Static Stress vs Static Eng Strain by Location
'II
2
-0.5
F-
(I)
0
-1
F-
-1.0 I-100
I
...
I°D
-80
-40
-60
STATIC ENG STRAIN, %
Figure B.1
I
-20
0
Stiffness vs Frequency by Location
30
20
10
0
0.01
0.1
1
(a)
Streaming Potential vs Frequency by Location
0.25
E
0.2
Fz
w
O
F0
0~
(9
z
0.15
0.1
w 0.05
cc
FU)
n
0.01
0.1
1
(b)
Figure B.2
Appendix C
Normalization
C.1 Introduction
The rationale behind the normalization of dynamic stiffness and streaming potential discussed in
Chapter 2 is that the dynamic measures change linearly with the static loads over which they are
superposed. Since the dynamic frequency sweeps under displacement control occur at varying
static loads, it was reasonable to normalize (literally dividing) the dynamic stiffnesses and streaming potentials by the static stress. To be rigorous, we tested our linearity hypothesis with one joint
by dynamically stimulating the cartilage at different static displacements.
C.2 Methods
The protocol defined in Chapter 3 was followed except dynamic frequency sweeps were performed after the 13, 25, 38, 50, 60 and 70 um static displacements. The dynamic stimulation was
executed at the standard frequencies.
C.3 Results and Discussion
The dynamic stiffnesses and streaming potentials were plotted at each static displacement. We
have included the plot of the MM site data for reference (Figure C.1). As can be seen in the plots,
the spacing of the curves is very linear over all static loads and frequencies. This result shows
that great confidence can be held for the normalization procedure.
C.4 Conclusion
Although the normalization appears justified, such an inelegant solution can be avoided in the
future if all the dynamic tests were performed at different static displacements as in this expcri-
ment. After testing a group of joints, we could just chose the dynamic measurements that wcre
taken at similar static loads instead of using a mathematical fix.
Displacement Eng Stiffness vs Frequency by Applied Displacement
28
21
mmIn
14
1I
7
n
0.01
0.1
FREQUENCY, Hz
Figure C.i: This is a plot of the dynamic stiffness data at the MM site taken at different static displacements.
Appendix D
Sample Chart Recorder Waveforms
. LO/S
~0
-=-5-g
~
0,g
'o0
I-
-mI
I
7=a
II
--
irc
_1
!O
II
I
I
-0-
~I
rrrr~
II,
=
~
OE=
3sa!
~C3~1P
I
I
now..
i
i
Lli
D.I
-~
=1Z
-3-
Figure
Figure D.1
Figure D.2
66
Appendix E
Phase 2 Plots without Chart Recorder Data
Static stress vs strain, site AM
G1 stiffness 2.4648 +/ - 0.65491
G2 stiffness 5.9591 +/ - 2.9207
G3 stiffness 2.0455 +/ - 0.44617
G1 = ... n=13
G2= .. n=8
n=6
G3=
I
/
//
/
I
S/
SI
S/
S/
/
/
2
I
0.1
0.2
0.3
0.4
0.5
0.6
Strain, percent
Figure E.1
67
0.7
0.8
0.9
1
Static stress vs strain, site MM
.4V -
I2
G1 stiffness = 1.5197 +/ - 0.28179
G2 stiffness = 1.1896 +/ - 0.32024
G3 stiffness = 1.3107 +/ - 0.070432
G1 = ... n = 13
G2=
.
n = E3
n= E
G3=
.°
..
.. : .-.
. .; ..
•
___
0
,"
•.....,..•~~~~~~~~~~
'
' '·'?
,
·
0.1
·
0.2
·
--
0.3
...
,• °
.. ... .. .. . •..
...............
·
·
0.6
0.5
0.4
Strain, percent
Figure E.2
0.7
0.8
0.9
Static stress vs strain, site ML
G1 stiffness = 1.1685 +/ 0.3684
G2 stiffness = 1.5514 +/ 0.23695
G3 stiffness = 1.0841 +/ 0.32922
G1 = ... n=13
G2=
G3=
.
. n=8
n=6
Co
n
a-
C,
CD
a)
L.
0
0
0.1
0.2
0.3
0.4
0.5
0.6
Strain, percent
Figure E.3
0.7
0.8
0.9
1
Stiffness vs Log(Frequencv), Site ML
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
Figure E.4
0
0.5
Stiffness vs Log(Frequency), Site MM
,-.''-
I/
60
50
S40
C)
t4-
: 30
20
10
0
-2.5
-2
-1.5
-1
Log(Frequency), Hz
Figure E.5
-0.5
0
0.5
Stiffness vs Log(Frequency), Site AM
60
-
G1 =X: n =13
...
=
X:
.
n71
.
G2 = 0; n = 8
50
-
-
Uo = +, " =
r-I
)
C_1
1
cu
a 40
3/C
cn
c
l/r
c~7
30
L
r
K
L
Flj
20
L-
E
L
10
H
I'
U
-2.5
I
I
-1.5
I
I
-1
-0.5
Log(Frequency), Hz
Figure E.6
0.5
Normalized Stiffness vs Log(Frequency), Site ML
110
100
90
80
70
60
50
G1 =X;n= 13
G2= O; n= 8
G3=+; n =6
-
-
-
-
-
-
40
1"
I-2.5
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
Figure E.7
0
0.5
Normalized Stiffness vs Log(Frequency), Site MM
I IU
100
90
80
ca 70
w 60
CD 50
40
30
20
1i
-2.5
-2
-1.5
-0.5
-1
Log(Frequency), Hz
Figure E.8
0
0.5
Normalized Stiffness vs Log(Frequency), Site AM
4 4 d
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
Figure E.9
0
0.5
rr
II
hr
U.UO
Streaming Potential vs Log(Frequency), Site ML
·
0.04
0.03
0.02
0.01
-
G1 = X; n = 13
G2 = O; n = 8
G3 = +; n = 6
-
SW
_-0 n1
-.2
-2.[
5
•
TT
T
I
T
III
-1.5
-1
-0.5
Log(Frequency), Hz
Figure E. 10
0.5
I
"I-
U.UO
0.04
0.03
Streaming Potential vs Log(Frequency), Site MM
II
· I
H
-
G1 = X; n = 13
0.02
0.01
A
-V.V
-
G2 =O; n = 8
G3 = +; n = 6
-
.1
I
I
5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
Figure E. 11
0
0.5
Streaming Potential vs Log(Frequency), Site AM
0.0U
"
'fI
0.04
0.03
0.02
0.01
0
-0.01
-2.5
-2
-1.5
-1
-0.5
Log(Frequency), Hz
Figure E.12
0
0.5
Normalized Streaming Potential vs Log(Frequency), Site ML
0.08
E
G1 =X;n=13
G2 = O; n = 8
0.06
G3 = +; n = 6
E
0.04
H
0.02
-
S
-
A)2.
--2. 5
I
I
-1.5
5•
5
I
.
I
-1
-0.5
Log(Frequency), Hz
Figure E. 13
I
0.5
Normalized Streaming Potential vs Log(Frequency), Site MM
0.1
0.08
G1 = X; n = 13
G2 = 0; n = 8
G3 = +; n = 6
0.06
E
0.04
E
0.02
n •0)
-V.V.
-
-2.5
I
-2
I
-1.5
I
I
-1
-0.5
Log(Frequency), Hz
Figure E. 14
I
0
0.5
Normalized Streaming Potential vs Log(Frequency), Site AM
0.1
0.08
0.06
1-
G1 = X; n = 13
G2=O; n=8
G3 = +; n = 6
0.04
F
0.02
E
00 A'
-V1.-
-2.[
5
I
-2
I
-1.5
I
I1
-1
-0.5
Log(Frequency), Hz
Figure E.15
0
0.5