Time-dependent mechanical behavior of newly developing matrix of
bovine primary chondrocytes and bone marrow stromal cells
using Atomic Force Microscopy
by
BoBae Lee
ARCHIVES
Submitted to the Department of Materials Science and Engineering
in partial fulfillment of the requirements for the degree of
MASSAOHUSET
OF TECHNOLOGY
Doctor of Philosophy
at the
SEP 0 9 2009
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
LIBRARIES
September 2009
© Massachusetts Institute of Technology 2009. All rights reserved.
Author........................
Department of Materials Science and Engineering
/
C-rtified hb
y
...
. .
.
...
,-,
Aug 3rd, 2009
.....
J.Grodzinsky
........
...........
Professor of Electrical, Mechanical, and Biolog'ial Engineering
Thess Supervisor
Certified by.............................
Christine"iftz
Associate Professor of Materials Science and Eng 'ering
/
h
T esis
pe isor
Accepted by.......................................
Christine Ortiz, Associate Professor of Materials Science and Engineering
Chair, Departmental Committee on Graduate Students
Time-dependent mechanical behavior of newly developing matrix of
bovine primary chondrocytes and bone marrow stromal cells
using Atomic Force Microscopy
by
BoBae Lee
Submitted to the Department of Materials Science and Engineering
On August 3rd, 2009, in partial fulfillment of the requirements for the degree of
Doctor of Philosophy
Abstract
Thesis Supervisor: Alan J. Grodzinsky
Title: Professor of Electrical, Mechanical, and Biological Engineering
Thesis Supervisor: Christine Ortiz
Title: Associate Professor of Materials Science and Engineering
Acknowledgements
It has been almost five years since I came to MIT to pursue my PhD degree in
Materials Science and Engineering. It was my first time to study abroad and so I learned
a lot of valuable things through exciting, and sometimes challenging events. The most
valuable thing will be that I luckily met my two greatest advisors, Prof. Christine Ortiz
and Prof. Alan J Grodzinsky in my PhD program. At the first time, I wanted to study the
biological system based on my background of chemical engineering and polymer science.
At that time, I had an opportunity to talk to Christine in person about research projects in
her research group. One of them is to characterize mechanical properties of articular
chondrocytes using Atomic Force Microscopy (AFM). It was a great moment that I
realized this might be one that I wanted to do. That's because this would require the
knowledge of biology and engineering since biological system can be described as a
complex system composed of macromolecular structures. After another meeting with
Alan, I decided to join this research project with confidence. Although it was the very
first time for me to see AFM, Laurel, a senior graduate student, who was already one of
expert in the field of AFM, helped me a lot how to operate AFM step by step. With her
painstaking effort to train me, I was able to feel more comfortable and get reproducible
data using AFM. In addition, she had been always good to me as a good friend as well as
a counselor even if I asked silly questions sometimes. Lin was also always friendly and
resourceful to me. He was a very smart guy to know everything about AFM. I admitted
that my dynamic oscillatory compression experiment could be impossible without his
help. I hope his join will reveal another exciting discovery using AFM.
As a group member of both research groups, I enjoyed various activities. In
Christine's group, we attended MRS every winter and had a dinner together for
celebrating the successful poster presentations and talks. Last summer, we had also a
great time in Washington DC to visit Smithsonian museum and NIH. Meeting with the
researchers in NIH offered me a chance to obtain a valuable insight from other
backgrounds. In Alan's group, we had an Endicott meeting every summer for a whole day
full of talks, foods, and outdoor activities. It was a great time to catch up each other's
research interests and get along together. Every month, Han-Hwa baked a very delicious
cake herself to celebrate the birthday of the month for us. I am already going to miss her
yummy cake. And I should mention our margarita's chair. Once everyone finishes his/her
own ORS abstract on the last day of submission in summer, we had a great time in
making and sipping our own margaritas, thanks to our chairs' efforts: Cameron, Yi, and
Sangwon. I always love Linda's humors and gestures.
Now it's time to thank all of my friends in my both research groups. Juha has been
one of my best friends since she joined Christine's group. We chatted a lot while having
dinner together at night after group meeting. She always cheered me up. Hsu-Yi is also
good to me as always. She is now an expertise in AFM imaging. Her skill is so amazing
that she can show every aspect of aggrecan molecules. I also appreciate the supports of
Jae-Hyuck, Miao, Lifeng, Haimin, and Fevzi whenever I asked questions related to my
research. Matthew, Fei and Shuodan are very friendly to me to talk with. Hopefully,
Shuodan and I hopefully can graduate in the same year as we promised before. It's a
secret that she is indeed a good cook. Paul and Eric were very helpful to me for
conducting my research of bovine stem cells. Discussion with Eric finalizes the issue of
complexity in my statistics. Eliot is a great person to ask any difficult questions I once
have. Rachel generously donates her cells to me whenever our experiment schedule is
coincident. I enjoyed the memory of roommates with Yi and Yunna while I attended
ORS. Without Han-Hwa's help, my research in Alan's group might not be complete. Her
advice does not only give the very information I need for my experiment but also helps to
relieve my personal worries and concerns. And I am deeply thankful to my Korean big
sisters and brothers in MIT, Sangwon, Junsang, Jeongsoo, Eunjee, Inja, Miso, Jung Ah
for their supports.
Finally, I should mention how much I love, respect, and appreciate my two
greatest advisors. Despite of my clumsy English and skills, they always support me with
all the best and with patience. Without their dedication, my PhD in MIT could not be
realized. I also thank my families for always praying for me. I thank God that I have such
a wonderful time during my life in MIT.
Contents
1 Introduction
27
1.1 Chondrocytes with their pericellular matrix in articular cartilage .............
27
1.2 Single cell mechanics of articular chondrocytes ........
29
............................
1.3 Chondrogenic differentiation of bone marrow stromal cell ..................
31
1.4 O bjectives ............................................................................
33
2 Backgrounds
37
2.1 Matrix organization of articular cartilage ........................................ 37
2.2 Composition and organization of PCM in mature cartilage ..................... 38
2.3 Composition and organization of newly developing PCM .................... 40
2.4 Mechanical properties of PCM in mature chondron and newly developing
PC M ...............................................................................
.
41
2.5 Implication of mechanical properties of PCM in cartilage tissue engineering
43
3 Time-dependent mechanical properties of primary chondrocytes
and their newly developing PCM via AFM-based stress
relaxation
.. 47
3.1 Introduction ........................................
3.2 M ethods ........................ ........... ............................. ..
49
..
49
3.2.1 Cell source and culture ........................................
3.2.2 Microfabricated substrates for single chondrocyte mechanics
..
experim ents ......................................................
50
3.2.3 Atomic Force Microscopy probe tips ....................................... 50
3.2.4 Stress relaxation and dynamic nanoindentation ............................ 51
3.2.5 Data analyses and statistical tests ........................................
. 53
3.3 Results ................................................................ 54
3.3.1 Rate-Dependence of deflection sensitivity: control experiment ........... 54
3.3.2 Stress relaxation of chondrocytes with their newly developing PCM .... 56
3.3.3 Stress relaxation of newly developing PCM of cultured primary
chondrocytes (Day 7-28) ........................................
.......
3.4 Discussion .........................................
60
61
3.4.1 Viscoelastic properties of day 0 chondrocytes ............................. 61
3.4.2 Culture duration effect on Eo and E. of newly developing PCM .......... 62
3.4.3 Tip geometry dependence of Eo and E. of newly developing PCM ...... 63
3.4.4 Deformation rate dependence of viscoelastic time constants of newly
developing PCM ...................................................... 64
3.4.5 Comparison between newly developing PCM and mature PCM in
cartilage tissue engineering perspective ...................................... 67
3.5 Conclusion .........................................
69
4 Dynamic mechanical properties of the tissue-engineered matrix
associated with individual chondrocytes
71
4.1 Introduction ............................................................................... 71
4.2 Materials and Methods ........................................
4.2.1 Cell source and culture .........................................
............. 73
..........
73
4.2.2 Dynamic oscillatory compression .......................................
74
4.2.3 Data analysis and statistical tests ............................................ 75
4.3 Results .............................................................
78
4.3.1 Dynamic mechanical properties of individual chondrocytes and their
cell-associated matrix using micron-sized spherical tip .................. 78
4.3.2 Dynamic mechanical properties of individual chondrocytes and their
cell-associated matrix using pyramidal nanosized tip ................... 80
4.3.3 Indentation stiffness of single chondrocytes with growth factor
treatment during time in culture .......................................
86
4.4 Discussion ............................................................................
88
4.5 Conclusions ...............................................................
91
..........
5 Adhesion behavior of primary chondrocytes with their newly
developing PCM
95
5.1 Introduction ........................................................................
95
5.2 Methods ................................................................
97
......
5.2.1 Cell source and culture ..................................................
97
5.2.2 Adhesion measurement of single chondrocytes .............................. 97
5.2.3 Data analysis and statistical tests of adhesion measurement ............... 98
5.3 Results ........
100
........................................
5.3.1 Adhesion behavior of single chondrocytes and their newly developing
100
matrices ...........................................
5.3.2 Adhesion behavior of single chondrocytes (day 0) ....................... 107
5.3.3 Adhesion behavior of the cell-matrix composite (days 7-28) .........
108
5.3.3.1 The effect of surface dwell time on adhesion behavior of cellmatrix composite (days 7-28) ........................................ 108
5.3.3.2 The effect of tip geometry on the adhesion behavior of cellmatrix composite (days 7-28) ........................................ 109
5.3.3.3 The effect of culture duration on the adhesion behavior of
newly developing matrix (day 7-28) ............................... 110
5.3.3.4 The effect of growth factor treatment on the adhesion behavior
of newly developing matrix (day 7-28) ........................... 110
5.4 Discussion ........................................
112
5.5 Conclusions ........................................
115
6 Mechanical behavior of individual bovine bone marrow stromal
cells undergoing chondrogenesis and their newly developing cellassociated matrix
117
117
6.1 Introduction ........................................
6.2 Methods .......
...............................
6.2.1 Cell source and culture ............................
........
120
. ..............
120
6.2.2 Histology of newly developing PCM .........................................
121
6.2.3 Biochemical assays of newly developing PCM ..............................
121
6.2.4 Mechanical testing of BMSCs during chondrogenesis ...................
122
6.2.5 Data analysis and statistical tests ..............................................
123
6.3 Results ........................................
125
6.3.1 Biochemical composition of the newly developing cell-associated
matrix and its appearance of BMSCs-derived chondrocytes over 28
days .........................................
125
6.3.2 Elastic and transient mechanical properties of the cell-matrix
composites of BMSCs-derived chondrocytes ................................
127
6.3.3 Dynamic properties of newly developing PCM ...........................
131
6.3.4 Estimation of hydraulic permeability .....................................
132
6.4 Discussion ........................................
134
6.5 Conclusions ..............................................................................
140
7 Concluding Remarks
143
A Glossary
147
B Characteristic time constants for time-dependent behavior of
tissue-engineered cell-associated matrix
149
B.1 V iscoelastic m odel ..................................................................
149
151
B.2 Poroelastic model ........................................
C Time-dependent mechanical properties of single chondrocytes
C .1 Introduction ................
.. .. ... .............................................
...................... ................. 158
160
C.4 Discussion ........................................
C.5 Conclusion .....
155
156
C.2 M ethods .............................................................
C .3 Resu lts .........................................
155
......................
. ................. 163
165
D Cell fixation protocol
165
D. 1 Paraformaldehyde fixation ........................................
167
E Histology staining protocol
E.1 Fixing cells .........................................
167
E.2 Mounting cells onto a slide ......................................
167
E.3 Toluidine blue ...........................................
168
E.4 Aniline blue ....................................
.....
169
E.5 Mounting coverslip .......................................
170
E.6 Taking microscopy images ........................................
170
F Casting alginate beads with cells
171
F. Materials ..............................................
171
F.2 Formation of alginate beads ........................................
172
F.3 Dissolution of alginate beads .........................................................
G Bovine bone marrow stromal cells isolation
G. 1 Bone marrow harvest and processing ......................................
173
175
175
G.2 Nucleated cell count, red blood cell lysis, and cell plating .................... 176
References
179
14
List of Figures
1-1 Structural organization of native cartilage chondrocyte PCM: (a) the
assembly of aggregates and interactions with ECM components (1) and
(b) a schematic diagram of the mechanical environment of a single cell
within the cartilage ECM with each characteristic modulus (2) ............. 28
3-1 Scanning Electron Microscope (SEM, JEOL JSM 6060, Japan) image of
a Si3N4 nanotip used for stress relaxation and dynamic nanoindentation
measurements: (a) a cantilever probe tip (b) an enlarged image of a
cantilever tip .....................................................................
51
3-2 Schematic representation of stress relaxation for day 14 chondrocyte in
10% Fetal Bovine Serum (FBS) culture medium probed by a micronsized spherical tip: (a) constant indentation depth with three different
displacement rates and (b) change of force during stress relaxation
m easurem ent ..................................................................
52
3-3 Viscoelastic model prediction of stress relaxation for day 14
chondrocyte in 10% Fetal Bovine Serum (FBS) culture medium probed
by a micron-sized spherical tip:
(a) experimental data with fitting
parameters corresponding to time-dependent mechanical properties
using a 5-element Maxwell Weichert model, (b) Comparison between
experimental data and model prediction using 3- and 5- element
viscoelastic model in earlier transient relaxation behavior of newly
developing PCM, and (c) a 5-element Maxwell-Weichert model .......... 54
3-4 Deflection sensitivity of the cantilever for stress relaxation measurement
with different z-piezo displacement rates: (a) a nanotip and (b) a micronsized tip. Control experiment is carried out onto the silicon substrate
covered by 10% Fetal Bovine Serum (FBS) culture medium .............. 55
3-5 Model fitting of experimental data with a 5-element model for day 14
chondrocyte in 10% Fetal Bovine Serum (FBS) culture medium probed
by a nanotip: Logarithmic values of force and time indicate the change
of slope, which corresponds to two time constants .................
57
3-6 Instantaneous modulus, E0 , and equilibrium modulus, Eo of newly
developing PCM over 28 days: (a) Eo probed by a nanosized pyramidal
tip, (b) Eo probed by a micron-sized spherical tip, (c) Eo probed by a
nanotip, and (d) E. probed by a micron-sized spherical tip (Two-way
ANOVA, p<0.05; n=7 cells, mean+SE; A dotted line indicates the
separation between day 0 chondrocytes lack of PCM and cultured
chondrocytes with newly developing PCM) ................................. 58
3-7 Time constants obtained from stress relaxation measurement using a
nanotip (a, c) and a micron-sized spherical tip (b, d).
tl
corresponds to
the time constant of the short-time transient behavior and 12 to the
longer-time transient response of newly developing PCM (Two-way
ANOVA, p<0.05; n=7 cells, mean+SE; A dotted line indicates the
separation between day 0 chondrocytes lack of PCM and cultured
chondrocytes with newly developing PCM) ................................. 59
4-1 Dynamic oscillatory compression of individual chondrocytes and their
associated matrix in FBS culture using a spherical tip: (a) Experimental
setup: Chondrocyte-matrix composites in microfabricated silicon wells
(radius of cell = R e, radius of probe tip = Rtip, matrix thickness = 31am,
indentation depth =
tm, and small sinusoidal oscillation = 5nm; the
detailed structure of the matrix has been described previously (1, 3)), (b)
Force (red) and indentation depth (blue) curve showing the initial
indentation (depth, do = 1tm) and subsequent force relaxation
exemplified by a day 14 chondrocyte, and (c) inset showing the applied
dynamic sinusoidal deformation (d(f/)) and resulting dynamic force (F(f))
shifted from the displacement by the phase angle 68() ..................... 77
4-2 Dynamic mechanical properties of chondrocytes and their newly
developing cell-associated matrix up to 28 days in FBS culture (10%
FBS) using the spherical probe tip: (a) IE*(f) , (b) 6(f), (c) E'(f) and (d)
E"(f. n=number of cells, mean + SEM; maximum SEM is -4kPa for
E'(J) and -6kPa for E"(J), respectively .........................................
81
4-3 Dynamic mechanical properties of chondrocytes and their cellassociated matrix up to 28 days in GF culture (IGF-I+OP-1) using the
spherical probe tip: (a) E*(f) , (b) 6(f), (c) E'(f
and (d) E"(f).
n=number of cells, mean + SEM; maximum SEM is -2kPa for E'(f) and
-I kPa for E"(), respectively ...............................
82
4-4 Effect of culture duration and growth factor treatment on dynamic
oscillatory properties of chondrocytes and their newly developing matrix
using the spherical probe tip: (a) the effect of culture duration on
IE* 9)
for GF culture chondrocytes (one-way ANOVA: frequency)
and (b) the effect of growth factor treatment on 669 (one-way ANOVA:
GF treatment) .........................................
................. 83
4-5 Dynamic mechanical properties of chondrocytes and their newly
developed matrix up to 28 days in FBS culture using the pyramidal
probe tip: (a) E*0) I, (b) (69,(c) E'(f) and (d) E"(f). n=number of cells,
mean ± SEM; maximum SEM is -2kPa for E'(f) and -3kPa for E"(),
respectively ........................................................... 84
4-6 Dynamic mechanical properties of chondrocytes and their cellassociated matrix up to 28 days in GF culture (IGF-I+OP-1) using the
pyramidal probe tip: (a) I E*() , (b)
69, (c) E'()
and (d) E"(f).
n=number of cells, mean ± SEM; maximum SEM is -2kPa for both E'(f)
and E"(f) .............................................
85
4-7 Effect of culture duration and growth factor treatment on dynamic
oscillatory properties using the pyramidal probe tip: (a) the effect of
culture duration on IE*( I for GF culture chondrocytes (one-way
ANOVA: culture duration) and (b) the effect of frequency on 6() for GF
culture chondrocytes (one-way ANOVA: frequency) ...................
4-8 Force-displacement curves of individual chondrocytes and chondrocytematrix composites using each probe tip: (a,c) FBS culture and (b,d) GF
culture (IGF- I+OP-1). n = number of cells, mean ± SEM ...................
5-1 Adhesion behavior of day 28 chondrocyte with its PCM ................... 99
5-2 Adhesion behavior of single chondrocyte (day 0) ......................... 102
5-3 Adhesion behavior of day 14 chondrocyte with its PCM from FBS
... 103
culture ......................................
5-4 Adhesion behavior of day 14 chondrocyte with its PCM from GF culture
104
5-5 Adhesion behavior of day 28 chondrocyte with its PCM from FBS
culture .......................................
............ 105
5-6 Adhesion behavior of day 28 chondrocyte with its PCM from GF culture 106
5-7 The maximum adhesion force,
Fad,
and the adhesion energy,
Ead,
Of
single chondrocytes (day 0) and their newly developing matrices (days
7-28) using (a,c) FBS culture (10% FBS) and (b,d) GF treatment (IGF1+OP-1) probed via micron-sized spherical tip (mean±SEM) ............ 108
5-8 The maximum adhesion force, Fad, and the adhesion energy, Ead, Of
single chondrocytes (day 0) and their newly developing matrices (days
7-28) using (a,c) FBS culture (10% FBS) and (b,d) GF treatment (IGF1+OP-1) probed via nanosized pyramidal tip (mean±SEM) ................. 111
5-9 The temporal change in the ratio of Ead to Etot for the cell-matrix
composite over 28 days with surface dwell time (mean±SEM) ........
112
6-1 Dynamic oscillatory compression of individual BMSCs and their
associated matrix using a spherical tip: (a) Experimental setup: BMSCderived chondrocyte-matrix composite in microfabricated silicon wells
(radius of cell = Rcell, radius of probe tip = Rtip, matrix thickness =
3.51pm, indentation depth = Iym, and small sinusoidal oscillation = 5nm;
cell radius and matrix thickness are based on day 21-28) (b) Force (red)
and indentation depth (blue) curve showing the initial indentation (depth,
do = Ilm) and subsequent force relaxation exemplified by a day 21
BMSC, and (c) inset showing the applied dynamic sinusoidal
deformation (d(f)) and resulting dynamic force (F(f)) shifted from the
displacement by the phase angle 6(f) ......................................
125
6-2 Biochemical characterization of the newly developing matrix of
BMSCs-derived chondrocytes over 28 days that were released from
alginate culture at designated time points: (a) Total GAG content
(measured
by
DMMB)
and
collagen
content
(measured
by
hydroxyproline), (b) optical microscopy (10x) images of individual
BMSCs stained for collagen and proteoglycan (PG) ....................... 127
6-3 Elastic and time-dependent properties of the cell-matrix composite of
BMSCs-derived chondrocytes over 28 days that were released from
alginate culture at designated time points: (a) Ey from the quasistatic
indentation,
(b) Eo and Eo, (c) ti and
t
2
from force relaxation
(n=number of cells, mean±SEM, *vs. other days, **vs. day 7-14, + vs.
day 0) ...................
..........
....
.........
129
6-4 Linear regressions of biomechanical properties vs. biochemical
composition of BMSCs-derived chondrocytes over 28 days: (a) Ey vs.
hydroxyproline, (b) E. vs. GAG, (c) Eo vs. hydroxyproline, and (d) E.
vs. the combinational
interaction of GAG and hydroxyproline)
(Parameters for each regression are shown in Table 6-2) .................. 130
6-5 The cell-matrix composite of BMSCs-derived chondrocytes over 28
days that were released from alginate culture at designated time points:
(a) E*(/)
and (b) .(f) (n=number of cells, mean±SEM, *vs. lower
frequencies on the same time points) ..........................
132
6-6 Plot I E'*()
vs. f/2 for the cell-matrix composite of BMSCs-derived
chondrocytes over 28 days using the micron-sized spherical probe tip
(mean+SEM ) ........................................
133
B-1 Comparison of initial transient behavior for day 14 chondrocytes with
matrix between experimental data and viscoelastic model prediction: (a)
experimental data with model prediction of 3- and 5- element model
(EM) (single time constant for 3-EM and 2 different time constants for
5-EM, respectively), and model prediction of the experimental data up to
0.03, 0.05, 0.5 seconds (labeled as 0.03s, 0.05s, and 0.3s, respectively)
with 3-element model (b) enlarged plot of Figure B-l(a) up to 0.1 sec.
(For experimental data, a limited number of data points are indicated
than the actual number of raw data points; all data sets are only shown
up to either 10 or 0.1 seconds to highlight the deviation of individual
model prediction from experimental data.) .................................. 150
B-2 Plot IE*() I vs.f/ 2 for newly developing PCM during 1-month culture:
(a) from FBS medium (10% FBS) and (b) from GF medium (IGF-I+OP1) using the micron-sized spherical probe tip, and (c) from FBS medium
(10% FBS) and (d) from GF medium (IGF-I+OP-1) using the nanosized
pyramidal probe tip (n=number of cells, mean±SEM) .....................
153
C-I Dynamic oscillatory compression of freshly isolated chondrocyte (day
0) probed by spherical tip: (a) Experimental setup: Chondrocyte in a
microfabricated silicon well, (b) Force and Indentation depth curve, and
(c) Enlarged plot of the square area in (b) ................................... 158
C-2 Dynamic properties of single chondrocytes with both probe tips:
(a) E*(f) , (b.)(f) (n=number of cells, mean+SEM) ...................... 159
C-3 Dynamic properties of single chondrocytes with both probe tips: (a,b)
micron-sized spherical tip, (c,d) nanosized pyramidal tip (n=number of
cells, mean±SEM, *,** vs. lower frequencies on the same time points)
160
List of Tables
6-1 Characterization of cell diameter (Dce,,)
and the newly developing matrix
thickness of BMSCs-derived chondrocytes estimated from optical
micrographs (n= number of cells, mean±SEM), compared to those of
primary chondrocytes reported by Ng, et al.(4) .............................. 127
6-2
Correlation
between
mechanical
properties
and
biochemical
composition of BMSCs-derived chondrocytes over 28 days using
multiple linear regressions .....................................
130
6-3 Linear regression correlation between I E*(f) I and fl2 (Figure 6-6) of
the cell-matrix composite of BMSCs-derived chondrocytes over 28 days
using micron-sized spherical probe tip .................................... . 134
6-4 Characterization of BMSCs-derived chondrocytes for GAG spacing b,
hydraulic permeability k, and poroelastic time constant Tp, based on the
total GAG amount synthesized during culture duration, along with the
elastic modulus H, determined by FEA simulation ......................... 134
B-i Characteristic time constant for transient relaxation behavior of day 14
cell-PCM composite from experimental data and model prediction: for X,
the exponential function of force (Force = A exp(-time/X)+B, where A
and B is fitting parameter (unit of nN) and k is the characteristic time
constant(sec.) is used; 3-EM is 3-element model with single time
constant and 5-EM is 5-element model with two time constants of t-(the
initial time constant) and T2(the longer relaxation constant) .............. 151
B-2 Linear regression correlation between I E*(f) and f12 (Figure B-2) of
the cell-matrix composite in both FBS and GF culture using both
micron-sized spherical probe tip and nanosized pyramidal probe tip ....... 154
B-3 Estimation of the hydraulic permeability, k, based on the total GAG
amount synthesized over 28 days and the characteristic poroelastic time
constant, c,[-1/fp], using chondrocytes in FBS culture (b: GAG spacing,
H: elastic modulus determined by FEA simulation) ........................ 154
Chapter 1
Introduction
1.1 Chondrocytes with their pericellular matrix in articular
cartilage
Articular cartilage chondrocytes are solely responsible for the synthesis, assembly,
and maintenance of the extracellular matrix (ECM) and yet occupy <10% of the cartilage
tissue volume. Chondrocytes (equilibrium modulus, E - 0.3-4 kPa) develop a
micrometer-thick pericellular matrix (PCM) in vivo and in vitro which is softer
(equilibrium modulus, E - 40 - 70 kPa) than the surrounding mature ECM (E - 0.5 MPa)
(4, 5) (Figure 1-1). PCM is a jim-thick hydrated, porous macromolecular region
surrounding chondrocytes, being rich in fibronectin, proteoglycans (e.g., aggrecan,
hyaluronan, and decorin) and collagen (types II, VI, and IX); PCM is primarily defined
by the presence of type VI collagen as compared to ECM (4-6). Since PCM transfers
loads from the ECM to the cell during physiological compression, it is important to cel
signaling and mechanotransduction. The structure of chondron, single chondrocyte and
surrounding region of PCM, is recently imaged with respect to the temporal evolution (4).
One promising approach to cartilage tissue engineering is embedding chondrocytes in
synthetic scaffolds and exposing to various growth factors and mechanical loads to
facilitate ECM synthesis and suppress catabolic degradation of ECM macromolecules.
(a)
(b)
XIV
Extracfflutar 3
Matrix
Ey~ 1 MPa
Chondeluar -Ey- 1 kPa
-
Figure 1-1 Structural organization of native cartilage chondrocyte PCM: (a) the assembly
of aggregates and interactions with ECM components (1) and (b) a schematic diagram of
the mechanical environment of a single cell within the cartilage ECM with each
characteristic modulus (2)
Chondrocytes are capable of developing a cell-associated matrix in vitro when
seeded in 3-D hydrogel scaffold with similar cell division rate, as compared to isolated
chondrons (7). However, the morphology of pericellular halo from cultured chondrocytes
are distinctively different from that of isolated chondrons (8) and mature cartilage (9),
reflecting the disparity in the gene expression of cartilage-related genes between the
tissue-engineered constructs and native cartilage (9). Despite of these deficiencies, the
use of chondrocytes can be beneficial to accomplish the cell expansion, cell manipulation,
a higher cell yield per volume of tissue with ease, as compared to the use of chondron (9).
With this regards, the understanding of mechanics at cellular level in articular cartilage
may provide a better picture of the composition-structure-function relations linked to the
growth/remodeling activity of articular cartilage in order to achieve the mechanically
functional cartilage tissue engineered constructs. While the biomechanical testing of the
macroscopic tissue would give an overall outcome of multiple interactions involved, the
biomechanical study at cellular level might better examine the consequences of such
interactions in a more focused manner.
1.2 Single cell mechanics of articular chondrocytes
Single cell mechanics has dealt with the cellular response to mechanical stimuli.
For example, it is reported that cytoskeletal structure and intracellular organelles are
elaborately regulated with mechanical stimuli at cellular level (10-12). Biochemical
reaction and gene expression are influenced by mechanical stimuli at cellular level as
well (13). Biomechanical properties of single cell are tested via micropipette aspiration
(14-18), AFM nanoindentation (4, 19-24), magnetic bead microrheology (25, 26), laser
tracking microrheology (27), osmotic loading (28, 29), and micromanipulation (30-33).
In micropipette aspiration, aspiration length of cell is measured with time at given
pressure difference (14-18). AFM nanoindentation obtains force curves in terms of
displacement, which gives elastic modulus (4, 19-24). In addition, hysteresis between
approach and retraction curves are used to evaluate the time-dependent modulus and time
constant (4, 34). Magnetic bead microrheology and laser tracking microrheology utilize
characteristic particles (i.e. magnetic microbeads and laser deflection particles) to
measure local viscoelastic properties of cells (25-27). Osmotic loading exposes cells to
different osmotic environment to determine the change of cellular volume with time (28,
29). Micromanipulation includes unconfined creep compression (30-33), uniaxial
compression or traction (35-37), and equibiaxial distension (38) to measure cell
viscoelasticity. Each measurement gives intrinsic material properties of single cell such
as characteristic modulus, viscosity, time constant, and Poisson's ratio. In addition,
mechanical properties of single cell would reflect the change in cellular structure,
composition, and function (10-16, 18, 39). Single cell mechanics would provide a
valuable means to understand mechanotransduction pathway. Furthermore, the effect of
drugs, mutations, and diseases on cellular metabolism can be elucidated by measuring the
change in mechanical properties of single cell that undergoes the development process
(10).
For articular chondrocytes, several mechanical testing methods were applied to
characterize cellular response to the mechanical deformation and loading. Transient
nanomechanical properties of chondrocytes have been studied via micropipette aspiration
(14-16, 18), atomic force microscopy (AFM)-based nanoindentation (4, 5), osmotic
loading (28, 29), and micromanipulation (31-33, 37). The biomechanical properties of
single chondrocytes (e.g., characteristic modulus, viscosity, time constant, and Poisson's
ratio) are measured with respect to the culture time (4), the effect of growth factors (4),
and the zonal variation (5). The biochemical assay and imaging of cultured chondrocytes
strongly suggest the development of PCM from in vitro culture with its characteristic
presence of type IV collagen (4). The finding that the hysteresis increases with
displacement rates above 2 lim/s in AFM-based nanoindentation of single chondrocytes
indicates the time-dependent deformation (4). The dynamic loading of single
chondrocytes has not been done so far, although the dynamic loading is applied to the
different type of cells (40-44). These studies open new possibilities to study the timedependent behavior of single chondrocytes with their newly developing PCM during the
culture period. Since articular cartilage experiences the various range of frequency in vivo
while in ambulation such as walking and running (23, 45), it is necessary to extend the
study of the time-dependent behavior with respect to the broad range of frequency as well
as the temporal evolution in culture. Characterization of time-dependent mechanical
behavior will directly reveal the change of cellular response to mechanical deformation
with temporal evolution and provide the valuable information to elaborate the current
theoretical model of articular cartilage.
1.3 Chondrogenic differentiation of bone marrow stromal cell
Stem cell-based tissue engineering holds great potential for the regeneration
and/or replacement of damaged cartilage. Indeed, chondrogenesis of mesenchymal stem
cells (MSCs) was observed to occur within I week of culture and demonstrated the
capability of MSCs-derived chondrocytes to produce cartilage-like extracellular matrix
(ECM), based on biochemical and gene expression analyses (46-48). ECM of articular
cartilage is of critical importance in the load-bearing function during joint motion, and
thereby loss and/or degradation of ECM is implicated with osteoarthritis and leads to the
failure of the functional integrity of articular cartilage. With this regard, MSCs are sought
as an alternative cell source for cartilage tissue repair instead of autologous chondrocytes
transplantation; MSCs can be easily obtained with low donor side morbidity of the
harvesting procedure, well expanded in vitro in an undifferentiated state, as well as
differentiated into both osteogenic and chondrogenic cell lines under specified conditions
(48). In particular, the latter is a beneficial effect of the use of MSCs on the repair of
osteochondral defect where the restoration of both the subchondral bone (i.e.,
endochondral ossification) and articular surface (i.e., chondrogenesis) is needed (48).
Furthermore, another study showed that the pretreatment of MSCs with transforming
growth factor pl (TGF-pl) and the maintenance of a three-dimensional culture
environment prevent transdifferentiation of MSCs from a chondrogenic to an osteogenic
lineage, which further supports the application of MSCs for cartilage implantation (46).
MSCs are obtained from various source (e.g., adipose (47, 49, 50), bone marrow (46-48,
50), and amniotic fluid (51)) and cultured with tissue-engineered constructs (e.g., agarose
(49, 50), alginate (46-49, 51), self-assembling peptide gels(50)), in which the change of
MSCs during differentiation process was examined in order to find the optimal condition
in the potential use of stem cells for cartilage tissue engineering. So far, stem-cell based
studies of cartilage tissue engineering have examined the characterization of cartilagespecific gene expression (46-48, 50), the synthesis of ECM components of cartilage(49,
50), and immunohistochemical evaluation for collagen type 11 (46-50), along with a
relatively few functional studies of MSCs-based tissue engineered constructs (49). In
addition, chondrogenic culture medium using TGF-3 superfamily resulted in a dramatic
increase of the gene expression and biosynthesis of ECM macromolecules, as compared
to the absence of growth factor (47, 49-51) and showed a dose-dependent increase (46).
These studies revealed the temporal evolution of MSCs during chondrogenesis
from the MSCs-based hydrogel constructs as a whole tissue with biochemical perspective.
Yet, the cell-to-cell heterogeneity and/or the functional aspects of MSCs-based tissue
constructs at single cell level are not yet fully addressed, which can be informative to
elucidate the cellular mechanism underlying chondrogenesis from the local measurement
rather than a macroscopic mechanical testing of the cell-scaffold constructs. Indeed,
mechanical behavior at single cell
level was distinctively different between
undifferentiated and differentiated cells, implying that the intracellular organization (e.g.,
cytoskeleton) is significantly altered during differentiation (52); while the parallel
orientation of cytoskeleton in undifferentiated cells is reorganized into the radial
distribution from the center of the cell during chondrogenesis, the parallel cytoskeleton
become robust through the realignment with thicker stress fibers. Such changes might
lead to distinctive Young's modulus for each cell type (i.e., the relative magnitude of
Young's
modulus:
undifferentiated
differentiated
osteoblasts
cells). Nonetheless,
>
differentiated
the continuous
change
chondrocytes
>
of MSCs during
chondrogenesis is largely unknown; hence, mechanical studies of individual MSC
cultured in vitro during chondrogenic differentiation can provide the valuable insights in
composition-structure-function relations in the temporal development of cartilage-like
neo-tissue.
1.4 Objectives
Previous studies of single chondrocytes measured characteristic modulus with
quasistatic/transient compression using AFM (4, 5, 53). Temporal evolution of modulus
was evaluated from single chondrocytes and their newly synthesized PCM cultured in
growth factor stimulated culture medium compared to those in 10% Fetal Bovine Serum
(FBS) culture medium (4). However, time-dependent mechanical properties such as
characteristic modulus (i.e., complex modulus of dynamic compression, instantaneous
and equilibrium modulus for stress relaxation) and time constant are not yet examined for
chondrocytes during culture in vitro. Since articular chondrocytes are exposed to broad
frequency ranges of mechanical loading during development in vivo (45), it is necessary
to extend the static loading experiment on single chondrocytes to the dynamic loading
nanoindentation with different frequencies. So far, several studies on single cell
mechanics of articular chondrocytes have focused on the characterization and
quantification of mechanical properties using chondrocytes and mature chondron (5, 15).
The early stage of PCM would be of great interest in the perspective of cartilage tissue
engineering and mechanobiology of cartilage in health and disease, as yet there are a little
information available at cellular (4) and tissue level (9). Therefore, the goal of this study
using primary chondrocytes or bone marrow stromal cells is to fully characterize
quantitatively the time-dependent mechanical properties of chondrocytes and to further
investigate the mechanical properties of their newly developing PCM that undergoes the
temporal evolution, which might provide the valuable insights in the underlying
mechanism of mechanical behavior. To describe the time-dependence of mechanical
response, stress relaxation and dynamic oscillatory compression were applied to newly
developing PCM and chondrocytes via AFM-based nanoindentation. To explore the
origin of dissipation mechanism, adhesion behavior of chondrocytes with newly
developing PCM was also investigated. For both dynamic oscillatory compression and
adhesion studies, the effect of growth factor stimulation on the time-dependent
mechanical properties was also assessed.
In addition, our study utilized MSCs obtained from bovine bone marrow and
cultured them in the condition where MSCs can undergo chondrogenesis (50). To explore
their potential use as a cell source in cartilage tissue engineering (48), MSC-derived
chondrocytes were extracted from 3D alginate scaffold at selected time points while
being cultured over 28 days with chondrogenesis culture medium. Mechanical properties
and biochemical properties of MSCs during chondrogenic differentiation were
characterized in order to compare with those of primary chondrocytes. Ultimately, these
studies would help understand the relationship between the time-dependence of
biomechanical properties of PCM and its temporal evolution of structure and composition,
which further improve cartilage tissue engineering.
Hence, our study of newly developing PCM can be summarized with a testing
method of mechanical characterization: for primary chondrocytes and their newly
developing PCM, stress relaxation (Chapter 3), dynamic oscillatory compression
(Chapter 4), and adhesion (Chapter 5) were performed using AFM-based nanoindentation.
For MSC-derived chondrocytes and their newly developing PCM, mechanical behavior
was investigated using quasistatic indentation, stress relaxation, and dynamic oscillatory
compression via AFM-based nanoindentation, along with biochemical characterization
(Chapter 6).
First, the time-dependence of primary chondrocytes with their newly developing
PCM in response to mechanical deformation was assessed using stress relaxation with the
various deformation rates (Chapter 3) and dynamic oscillatory compressive loading of
5nm small amplitude deformation (Chapter 4) via AFM-based nanoindentation. In
addition, the effect of growth factor stimulated culture on frequency dependent
mechanical properties of chondrocytes was compared to the effect of regular culture (i.e.,
10% Fetal Bovine Serum) (Chapter 4). Model prediction of time- and frequencydependent mechanical behavior of newly developing PCM was compared with
experiment data in order to elucidate the relative contribution of solid viscoelasticity and
poroelasticity to observed mechanical properties in each culture condition. To further
investigate the origin of dissipation mechanism at cellular level, adhesion behavior of the
cell-PCM composite was independently measured using primary chondrocytes cultured
either in regular medium or with growth factor treatment (Chapter 5). Finally, newly
developing PCM was obtained from bovine MSCs cultured in the chondrogenic medium
(Chapter 6). During chondrogenesis,
newly developing PCM of MSC-derived
chondrocytes and cell were characterized using biochemical assays and via AFM-based
nanoindentation to compare the results with those of primary chondrocytes and their
newborn PCM. This study was aimed to evaluate the potential use of MSCs as a cell
source in cartilage tissue engineering.
Overall, the information of this study will extend our current knowledge of
biomechanical characteristics of chondrocytes and their newly developing PCM, in
particular, at early stage of development. With this purpose, this study was aimed to
compare the time-dependent mechanical behavior at cellular level to that of macroscopic
level and elucidate the origin and possible consequences of our findings. Furthermore,
experimentally determined mechanical properties of chondrocytes and newly developing
PCM will be beneficial to improve the model prediction of articular cartilage. Ultimately,
the accumulation of such efforts will lead to the success of fully functional cartilage
tissue engineering.
38
Chapter 2
Background
2.1 Matrix organization of articular cartilage
Articular cartilage is a heterogeneous tissue that typically divided into four zones:
superficial, middle, deep, and the calcified cartilage layer. Each zone features distinctive
tissue structure and composition, cellular shape and arrangement, depending on the depth
from the tissue surface (54). Collagen and proteoglycan content, cell morphology, and
collagen fiber orientation vary across zones, leading to depth-dependent mechanical
properties of articular cartilage (e.g., compressive and tensile modulus (55), osmotic
pressure and electrical potential (56)). For example, superficial zone has a flattened shape
of chondrocytes, middle zone more spherical chondrocytes, and deep zone chondrocytes
in a columnar arrangement, reflecting the orientation of collagen fibers (57-59). In
addition, collagen fibers in the superficial layer are tangentially aligned, forming a
tension-resisting. On the other hand, the radial alignment of collagen fibers in the deeper
39
layers likely serves as a hydrostatic shock absorber to withstand a compression loading.
The zonal variation influences the mechanical properties of chondrocytes so that
chondrocytes of superficial zone have higher moduli and apparent viscosities than those
of middle and deep zones (5). While gene expression and protein synthesis are not
significantly different between middle and deep zones, superficial zone represents the
gene expression, protein synthesis, overall phenotype distinctive to middle and deep
zones (13).
In general, the deeper cartilage contains several distinct regions and each region
distinguished by pericellular, territorial, and interterritorial matrices, depending on the
proximity to the chondrocytes, whereas chondrocytes are only surrounded by PCM in the
superficial layer (58). Each matrix exhibits the unique composition and structural
organization; a pericellular matrix rich in hyaluronic acid, chondroitin sulfate, a territorial
matrix enriched with chondroitin sulfate, an interterritorial matrix primarily composed of
high molecular weight keratan sulfate.
2.2 Composition and organization of PCM in mature cartilage
Pericellular matrix (PCM) has received increasing research attention due to its
importance
in
mechanotransduction
between
extracellular
matrix
(ECM)
and
chondrocytes. PCM consists primarily of collagen (types II, VI, and IX), proteoglycans
(e.g., aggrecan, hyaluronan, and decorin), and fibronectin (54, 60, 61); PCM is
distinctively defined by the presence of type VI collagen as compared to ECM. It was
known that PCM interacts with chondrocytes via cell-surface receptors (e.g., integrins,
CD44, and annexin V) (8, 61, 62). In particular, CD44/HA receptors are of critical
importance to the retention and assembly of chondrocyte PCM, independent of
endogenous or exogenous origin of matrices (62). While one end of hyaluronan is bound
to cell surface via CD44/HA receptors, hyaluronan is complexed with aggrecan to give
rise to a brush-like configuration with tethered motion of the other end in PCM; A brush
configuration of the hyaluronan-aggrecan also resists the compressive force due to the
swelling pressure of highly negatively charged aggrecan in its hydrated form (3). Being
considered as a primary functional and metabolic unit of cartilage that maintain the
integrity of chondrocyte and its pericellular microenvironment under compressive
loading, chondron is coined to indicate the chondrocytes with its narrow region of
surrounding PCM enclosed by a 'felt-like' capsule as observed mostly in middle and deep
zones (54, 63). Indeed, newly synthesized aggrecan was initially sequestered in the
pericellular matrix and later migrated into more distant matrices (i.e., territorial and
interterritorial matrix) in a temporal- and spatial- dependent manner, supporting the role
of chondron as a metabolic unit of articular cartilage (54).
Composition and organization
of PCM were significantly altered with
osteoarthritis, as characterized by a loss of type IX collagen and the development of a
narrow glycocalyx around each chondrocytes with elevated levels of type VI, X collagen
and proteoglycan (8, 54); as a result, the superficial region reactive to the degradation has
the swelling pericellular microenvironment and displays several clusters of chondrocytes
within a pericellular microenvironment due to the enhanced cell division and migration.
While the mechanism of pericellular remodeling remains largely unknown, the
morphological and metabolic changes in PCM with osteoarthritis appear to reflect the
earliest process of cartilage catabolism as well (54).
2.3 Composition and organization of newly developing PCM
The characteristics of newly developing PCM formed by chondrocytes
significantly varies with the types of 3-D hydrogel cultures, being different from mature
PCM of isolated chondrons as well (8, 9, 64). In general, newly developing PCM is more
diffuse, compared to compact and uniform PCM of mature chondrons. One possible
explanation of the difference in the structural morphologies and material properties of
newly developing PCM may be that the biological environment specific to the hydrogel
scaffold induces the change in the interaction between cell and surrounding environments
(e.g., adhesive and mechanical properties of hydrogels) and affects the following
metabolic activity of cells, possibly different from that in native articular cartilage as well
(8, 9, 64, 65). Nonetheless, newly developing matrices obtained from tissue-engineered
constructs are referred to as newly developing PCM, cell-associated matrices, or PCM
interchangeably, despite of their immaturity compared to their mature PCM of in vivo
articular cartilage (4, 9, 59, 62, 66-68). In addition, the previous study of the same culture
condition clearly demonstrated that an individual cell with its newly developing matrices
exhibit the type VI collagen via immunohistochemistry, the important characteristic
component of PCM (4). Hence, newly developing PCM in our study is used throughout
the entire chapters to point the newly developing cell-associated matrices, as preceded by
'newly developing' to differentiate from PCM of mature chondron.
Newly developing PCM was well characterized by the presence of type VI
collagen, whereas the structural organization is remarkably different from that of mature
PCM. In addition to its more diffuse (i.e., porous) structure, the fibrils of newly
developing PCM are aligned in a radial direction from the chondrocyte, instead of a
tangential direction of fibrils in mature PCM. Furthermore, there was no polarity
observed in newly developing PCM (9). The composition of newly developing PCM was
also distinctive to that of native cartilage tissue as shown in the gene expression of
cartilage matrix constituents (9). Since the completion of skeletal tissue development
takes a long process after birth (59), the structural immaturity of newly developing PCM
in the early culture duration may reflect the time scale of normal tissue development.
2.4 Mechanical properties of PCM in mature chondron and
newly developing PCM
Mechanical properties and biochemical composition vary with species (69, 70), as
well as the joint surface and location (71, 72). In particular, zonal variation significantly
influences the matrix biosynthesis and gene expression (13), the presence and
organization of cytoskeletal filaments (73), the mechanical properties of chondrocytes (2,
5, 32), implying that these structural and compositional difference lead to the mechanical
properties of chondrocytes as well as ECM intrinsic to the origin of subpopulation.
Despite
of such
differences,
several
studies
on the
ultrastructure
and the
microcompression of articular cartilage generally agree that the resistance and tensile
properties of pericellular collagens sustain the swelling pressure generated by high
pericellular concentration of hyaluronan and aggrecan (54).
So far, mechanical properties of PCM have been characterized using mature
chondron which is obtained from mechanical isolation (15), or enzymatic digestion (7).
Elastic, viscoelastic, and poroelastic behavior of PCM from mature chondron was
quantified using micropipette aspiration technique (15, 16, 60) and theoretical modeling
(2, 74) from experimentally determined material parameters. These studies revealed
strikingly different mechanical behavior of PCM, compared to chondrocytes devoid of
PCM. In addition, shape and volume of chondrocytes with their nuclei are changed in a
coordinated manner with deformation of surrounding matrix, which has been seen as the
spatial- and time-varying stress-strain, fluid flow, osmotic pressure, and electric fields (28,
57, 75); further suggesting a possible mechanism for signal transduction (74). In this
context, an understanding of mechanical properties of PCM would provide valuable
insight on the mechanotransduction to link the biomechanical events of ECM, possibly
accompanying the biochemical reaction, to the following response of chondrocytes.
Newly developing PCM was also studied using atomic force microscopy (AFM)based nanoindentation at single cell level (4) as well as at macroscopic tissue level (9).
These studies of newly developing PCM indicated that an increased elastic modulus with
culture duration is concomitant to the increase in GAG and collagen, which is further
supported by the morphological change of PCM via tapping mode AFM and electron
microscopy (EM) imaging. Similar to PCM of mature chondron, newly developing PCM
also exhibits time-dependence of its mechanical behavior as shown in the increased
hysteresis behavior with deformation rate (4). However, the time-dependent mechanical
44
behavior of newly developing PCM is not yet fully understood, particularly at single cell
level, which can help elucidate its functional role in mechanical loading and approach to
the functional cartilage tissue in relation to the development of fully mature PCM.
2.5 Implication of mechanical properties of PCM in cartilage
tissue engineering
Experimental and theoretical studies of PCM strongly suggest that PCM is of
critical importance in mechanotransduction and implicated in degeneration followed by
injury (76). Indeed, the mechanical loading condition affects the biochemical
environment and vice versa; experimental studies show that abnormal loading condition
can lead to the initiation and progression of joint injury (e.g., degeneration of PCM/ECM).
Poor integrity of PCM (e.g., damaged PCM) in cartilage further influences loading
profile inside the cartilage matrix and incurs following degenerative or inflammatory
response which impedes the recovery of the injured cartilage. On the other hand, the
proper modulation of local tissue strain during PCM development can be beneficial to the
cartilage tissue engineering so that mechanical stimuli-mediated metabolic activity of
chondrocytes might lead to spatial localization of newly developing PCM components
(e.g., aggrecan, collagen) (77-79). The recent study applied the static deformation to
single chondrocytes and found that the gene expression of chondrocytes is significantly
altered by the static deformation even with a short duration (-1 min.) (79). However,
further studies are needed to determine how long such altered cellular response can be
effective or what the consequences of such altered cellular behavior is on the role of
chondrocytes in the matrix formation and maintenance, leading to influence the overall
biomechanical properties.
Similarly, theoretical studies reveal that the presence of PCM significantly
changes the strain amplication of chondrocytes under transient loading (80) and static
compression (74), further suggesting that the applied tissue loading magnitude might not
be a proper measure to interpret the chondrocytes' response to loading and injury where
the stains translated to the cell and its microenvironment are different from the applied
tissue loading magnitude (80). Since the model prediction of theoretical studies is largely
dependent on experimentally determined parameters (e.g., Poisson's ratio, characteristic
stiffness, and permeability), experimentally determined mechanical properties is crucial
to improve the reliability of the model prediction.
To achieve the tissue engineered constructs whose properties are comparable to
those of the fully-developed cartilage tissue, several distinctive features of the mature
native cartilage tissue should be considered. First, the complete development of cartilage
tissue in vivo is a long time process; for example, the time scale lasts up to several
months for rabbits even after birth (59), which can be a challenge to the in vitro cell
culture protocol, normally practiced for a few weeks or a month. In addition, the entire
process of native tissue formation is a synchronized resorption and neoformation, leading
to the mature cartilage tissue that exhibits structural, morphological, mechanical, and
metabolic anisotropies, which have not yet been achieved by tissue-engineered constructs
(9). With this regard, newly developing PCM from isolated chondrocytes in vitro culture
might be applicable as an intermediary to further obtain a more fully-developed
cartilaginous tissue with a reduced time in culture (67), an implant for the articular
cartilage repair (81), as well as a model system to probe its mechanical properties while it
undergoes developmental process in vitro.
48
Chapter 3
Time-dependent
mechanical properties of primary
chondrocytes and their newly developing PCM via
AFM-based stress relaxation
3.1 Introduction
Chondrocytes are solely responsible for the synthesis, assembly, and maintenance
of the extracellular matrix (ECM) and yet occupy <10% of the cartilage tissue volume.
Chondrocytes are embedded within solid matrix of ECM, mainly composed of collagen
(mainly type II) and proteoglycan aggrecans, in water dissolved with small electrolytes
(e.g., K+, Na , Cl-, Ca2+). Proteoglycan aggrecans such as keratan sulfate and chondroitin
sulfate glycosaminoglycan generate repulsive force and osmotic gradients when they are
hydrated within the interstitial fluid, due to numerous negatively charged carboxyl and
sulfate groups of proteoglycan aggrecans. This swelling pressure not only affects the
overall cartilage tissue macroscopically but also influences the local biomechanical
behavior of chondrocytes embedded within the solid matrix, indicating the time
dependence of its response to applied loading and deformation (57).
At cellular level, previous study reported that chondrocytes exhibit a significant
increase of hysteresis with displacement rates above 2gm/s (4), indicating time-dependent
deformation. So far, time-dependence of mechanical properties are explored at cellular
level using mature chondron via micropipette aspiration (60) and freshly isolated
chondrocytes devoid of pericellular matrix (PCM) via micropipette aspiration (14, 57),
via unconfined creep compression (32), or via AFM-based stress relaxation (5); a large
difference in time-dependent mechanical properties between mature chondron and
chondrocytes suggests that PCM is a critical component to sustain and transduce
mechanical loading between ECM and chondrocytes, also supported by the theoretical
model study (74, 76). However, temporal evolution of PCM under development to
mature PCM (i.e., newly developing PCM) has been not yet fully understood with respect
to time-dependence of its mechanical properties. This information could be useful to
simulate the loading or deformation profile of chondrocytes in native tissue cartilage that
undergo the formation and modification of their PCM, which can be promising to
cartilage tissue engineering as well. Therefore, we extended these studies to examine the
effect of the culture time and deformation rate on the time-dependent behavior of newly
developing PCM of individual chondrocytes. Since freshly isolated chondrocytes develop
PCM in vitro, thereby changing their composite mechanical properties (4), we
hypothesized that their time-dependent behavior would also change with the culture time
and deformation rate.
In this study, both nanosized pyramidal probe tip (end radius Rtip - 50 nm) and
micron-sized spherical probe tip (R - 2.5 pm) were used to perform stress relaxation
measurements on individual chondrocytes and their developing PCM, cultured up to 28
days in 3-D alginate gels. The results would shed light on the time dependent response of
chondrocytes exposed to various deformation rates during development of tissue
engineered constructs.
3.2 Methods
3.2.1 Cell source and culture
Chondrocytes are isolated from femoral condyle cartilage of 2-3 week old bovine
calves and sequentially digested using 0.2% pronase (Sigma) for 1 hr and 0.025%
collagenase (Boehrinher Mannheim) for 12 hrs as previously described (4); viability is
assessed by trypan blue (Sigma) exclusion. Cell suspension of 15x 106 cells/mL is mixed
with 2% (w/v) alginate (KelcoLVCR) in 0.9% (w/v) NaCl, and cell-seeded alginate beads
(- 3 mm diameter) are formed via polymerization in 102 mM CaCl 2 using a 22-gauge
needle. Culture medium is high-glucose Dulbecco's Modified Eagle Medium (DMEM)
with 10% Fetal Bovine Serum (FBS), 20 gg/mL L-ascorbic acid, and 1% (v/v) antibiotic-anti-mycotic. Each culture well contains seven alginate beads supplemented with
3ml culture medium. Medium is changed every other day up to 28 days. On days 7, 14,
21 and 28, cells are released from beads by depolymerization in 55mM NaCitrate (Fisher
Scientific) (82).
3.2.2 Microfabricated substrates for single chondrocyte mechanics experiments
Silicon wafers are microfabricated into an array of inverted square pyramidal
wells, each capable of holding a single chondrocyte for AFM nanoindentation (4, 83).
Anisotropic etchants such as alkali hydroxides (i.e., KOH, NAOH, etc.) make (111)
planes of silicon etched much slower than (100) and (110) planes, which give V-shaped
trenches with the angle of 55' from the vertical. Side dimension of each well is 15, 18, 20
or 22 pm to accommodate the chondrocytes with different size.
3.2.3 Atomic Force Microscopy (AFM) probe tips
Nanotip and micron-sized probe tip are used for stress relaxation and dynamic
nanoindentation. Nanotip is a standard Si3N4 AFM square pyramidal tip (Veeco, Rtip - 50
nm, k - 0.06 N/m). Micron-sized tip is prepared from tipless cantilever (Veeco, k - 0.06
N/m) by attaching the silica bead (Bang Labs, R - 2.5 gim) with low viscosity epoxy
adhesive (SPI, M-Bond 610). Two clean mica surfaces (lcm x lcm, SPI supplies, West
Chester, PA) are prepared and glued on the top of each metal disc; One drop of adhesive
is adsorbed onto the clean mica surface and small amount of silica beads is scattered in
the other mica surface. Tipless cantilever is loaded on the tip holder and the adhesive disc
is loaded onto the sample stage in Picoforce Nanoscope IV AFM (Veeco, Santa Barbara,
CA). The cantilever is moved above the adhesive surface by adjustment knobs and
slightly dipped in the adhesive. The adhesive disc is then replaced with silica beads disc
and the glued tip of cantilever is moved accurately to pick up single particle from the
silica beads disc. Bead-attached tip is dried in ambient conditions during 3 days.
(b)
(a)
Figure 3-1 Scanning Electron Microscope (SEM, JEOL JSM 6060, Japan) image of a
a
Si 3N4 nanotip used for stress relaxation and dynamic nanoindentation measurements: (a)
cantilever probe tip (b) an enlarged image of a cantilever tip
3.2.4 Stress relaxation and dynamic nanoindentation
Freshly isolated cells (day 0) or cells released from alginate beads (day 7-28)
were resuspended at 106 cells/mL in culture medium. Culture medium was allowed to
cover the substrate surface for 5 min prior to the addition of 100L of cell suspension
onto the surface. Each cell was moved into a well using the AFM probe tip.
The
Picoforce
Nanoscope
IV
AFM
(Veeco)
was
used
to
perform
compression/stress relaxation measurements of each cell within its silicon well. Both
nanosized (end radius, Rtip - 50 nm) and micron-sized colloidal probe tips (Rtip - 2.5 Rim)
were used to probe the same cells. For stress relaxation measurements, each tip was held
at a constant normal displacement (- I jpm) for 60 seconds while recording force
relaxation (Fig. la,b), where the penetration depth corresponds to less than -10% of the
cell diameter of about 10 gm (4). Z-piezo displacement ramp rates of 1, 6.25, and 14gm/s
were used; I gm/s was chosen since this is the maximum displacement rate without any
significant hysteresis incurred in force-displacement curve (4), 6.25 gm/s for the
comparison with the study of Darling, et al. using day 0 porcine chondrocytes (5), and 14
gm/s for the application of displacement rate analogous to a step-function (i.e.,
indentation of 1 gm takes only -0.07 seconds at a displacement rate of 14 Rm/s). Each
measurement was repeated 3 times at the same location for each displacement rate; the
mechanical properties of chondrocytes and their newly developing PCM are not
significantly changed during the repeated measurement to obtain the same results in F-d
curve and stress relaxation curve.
"1.5 -- ?-
(a)
S1.0
*0
14
-
I1
6.25
(b)
1 ijm/s
Z
S-0
0
5
6
30
50
Time (sec)
70
6
Single cell + PCM
Day14, 14 pm/s
-L
-3 (mean ± SD of 5 repeats)
0 10 20 30 40 50 60 70
Time (sec)
Figure 3-2 Schematic representation of stress relaxation for day 14 chondrocyte in 10%
Fetal Bovine Serum (FBS) culture medium probed by a micron-sized spherical tip: (a)
constant indentation depth with three different displacement rates and (b) change of force
during stress relaxation measurement
3.2.4 Data analyses and statistical tests
Data were compared to the predictions of a 5-element Maxwell-Weichert model
(Figure 3-3(c)) coupled to the Hertz model to account for geometric factors (5, 84) (Eq.
3-1, 3-2), which is a viscoelastic solution for small indentations of an homogeneous and
isotropic continuum with an infinitely hard indenter
4R 11 283/
2
Fmicro (t) = 3(1 )
3(1 - v)
Fao
2
tan k 3
2(1- v)
K
kl
k3 I
_
-t r
ke/
2
k3
ek3
-t / 2
(3-1)
k3
k2
-t/
2
(3-2)
k3
where 6 is penetration depth, Poisson's ratio v is assumed to be 0.4, and the nanotip
opening angle a is 350 (4, 5, 84, 85). The radius R in Eq. 3-1 is calculated from the radius
of micron-sized tip and cell (i.e., 1/R = 1/Rtip +1/Rcell). Spring constant kl, k2, k3 and
dashpot viscosity fli, Tl2 correspond to the element of Maxwell-Weichert model (Figure 33(c)). The ratio of the ith-dashpot viscosity to the ith-spring constant determines the ithtime constant.
Model fitting parameters were determined using a nonlinear least square method.
Effects of tip geometry, culture duration, and the displacement rate on the viscoelastic
stiffness and time constants were evaluated using three-way ANOVA with Tukey post
hoc test to compare groups when the significance was detected (p<0.05). Since of
primary interest is newly developing PCM of cultured chondrocytes, the pair-wise
comparison was performed onto groups between day 7 and other longer culture duration
(day 14-28). Being lack of 14 plm/s data for day 0 and 7, two types of pair-wise
comparison were executed; one pair-wise comparison was separately made onto the
groups that pooled the displacement rate of 1 and 6.25 ptm/s for each culture duration
(day 7-28). And then, the other set of pair-wise comparison was made onto the groups of
day 14-28 pooling the displacement rate of 1, 6.25, and 14 plm/s. All parameters
(instantaneous, equilibrium modulus and two time constants) are reported as mean+SE.
(b) 8
Single cell+PCM (
Day 14
6
(a)
6
EE
Data
-- 5-element
- 3-element
(c)
I
Ek
k
-
o
cL-V
2
k3
Tzl
(mean + SD of 5 repeats)
-3
0 10 20 30 40 50 60 70
Time (sec)
0
5
Time (sec)
10
Figure 3-3 Viscoelastic model prediction of stress relaxation for day 14 chondrocyte in
10% Fetal Bovine Serum (FBS) culture medium probed by a micron-sized spherical tip:
(a) experimental data with fitting parameters corresponding to time-dependent
mechanical properties using a 5-element Maxwell Weichert model, (b) Comparison
between experimental data and model prediction using 3- and 5- element viscoelastic
model in earlier transient relaxation behavior of newly developing PCM, and (c) 5element Maxwell-Weichert model
3.3 Results
3.3.1 Rate-dependence of deflection sensitivity: control experiment
Deflection sensitivity measures the change of z-piezo displacement with unit
voltage change. Since force is obtained from multiplying the voltage change with the
deflection sensitivity [m/V] and spring constant [N/m] of cantilever, the rate dependence
of deflection sensitivity in AFM-based dynamic nanoindentation should be examined
(86). Spring constant is determined by the thermal noise oscillation, independent of the
displacement rate. In stress relaxation measurement, three different displacement rates (1,
6.25, and 14 m/s) were used. The microfabricated silica substrate covered by 10% Fetal
Bovine Serum (FBS) solution was used for control experiment. As shown in Figure 3-4,
deflection sensitivity does not show the rate dependence in the range of displacement rate
used in this experiment. On the other hand, another study showed that deflection
sensitivity changed slightly within the same range of displacement rate that was
performed in air (86).
(b)
(a)
60
C
o
>
58
57
s6
63
62
"
56
S
55
65
61
lIPJm/s 2pJm/s 6.25pJmls 14pm/s
z-piezo displacement rate
M
60
21mIs 6.251pm/s 14Jm/s
lpmls
z-piezo displacement rate
Figure 3-4 Deflection sensitivity of the cantilever for stress relaxation measurement with
different z-piezo displacement rates: (a) a nanotip and (b) a micron-sized tip. Control
experiment is carried out onto the silicon substrate covered by 10% Fetal Bovine Serum
(FBS) culture medium
3.3.2 Stress relaxation of chondrocytes with their newly developing PCM
As the deformation gradually proceeds and reaches I rLm, chondrocytes (day 0) or
their newly developing PCM (days 7-28; PCM thickness is -3-4 pm being constant
during culture duration after day 7 (4)) exhibit the force, increasing up to peak force
when the deformation reaches I plm and then monotonically decreases to the quasiequilibrium within 60 sec (Figure 3-3(a)). This relaxation behavior is quite consistent
with the observation in the previous study (5).
A standard 3-element viscoelastic model in other studies (4, 5) is preliminarily
used to analyze the whole set of data obtained from chondrocytes cultured in 10% Fetal
Bovine Serum (FBS) culture medium up to day 28 probed by a nanotip and a micronsized tip and found to be adequate for predicting the long-time stress relaxation behavior
(R2= 0.938 ± 0.05, mean + SD). However, a 5-element viscoelastic model with two time
constants is necessary to fit the short-time behavior, characterized by
T, (R2 =
0.989 +
0.03, mean + SD, Figure 3-5), which would correspond to higher frequency behavior in
dynamic loading. In addition, the viscoelastic model fitting to experimental data makes it
clear the difference between a 3-element viscoelastic model and a 5-element viscoelastic
model; a 3-element viscoelastic model fails to describe most of initial decaying of the
relaxation curve (Figure 3-3(b)). Therefore, a 5-element viscoelastic model is
subsequently used to fit all data for each probe tip.
3
2.5 .for
T2
Z2
(R2=0.992)
o 1l for ,
(R2 =0.976)
0..j
0.5
-5
-4
-3
-2
-1
0
1
Log(time(sec))
2
3
4
5
Figure 3-5 Model fitting of experimental data with a 5-element model for day 14
chondrocyte in 10% Fetal Bovine Serum (FBS) culture medium probed by a nanotip:
Logarithmic values of force and time indicate the change of slope, which corresponds to
two time constants.
The instantaneous modulus, E0, is calculated from the peak force and the
equilibrium modulus, E., from the relaxed force (Figure 3-3(a)), coupled to the Hertz
model to account for the geometric factor.
Viscoelastic properties of day 0 chondrocytes (no PCM) Day 0 chondrocytes
exhibited Eo of 0.9±0.2 kPa, E. of 0.7±0.2 kPa and
T2
of 18±2.7 sec, respectively
(mean±SE; n=7) measured at a displacement rate of 6.25 ptm/s using a micron-sized
spherical probe tip (Figure 3-6(b)). A nanosized pyramidal tip obtained higher Eo and E~,
lower
T2
for day 0 chondrocytes than a micron-sized spherical probe tip, whereas tz of
day 0 chondrocytes was similar for both probe tips (Figure 3-6,3-7).
SSpherical tip
r Pyramidal tip
(a)
4
c2
1
0
(Tukey post-hoc analyses; p<0.05 compared to +day7 and ++between day 14 and 28)
(c)
31
1
*
I (d) 31
Figure 3-6 Instantaneous modulus, E0, and equilibrium modulus, E", of newly
developing PCM over 28 days: (a) Eo probed by a nanosized pyramidal tip, (b) Eo probed
by a micron-sized spherical tip, (c) Eo probed by a nanotip, and (d) Eo probed by a
micron-sized spherical tip (Two-way ANOVA, p<0.05; n=7 cells, mean+SE; A dotted
line indicates the separation between day 0 chondrocytes lack of PCM and cultured
chondrocytes with newly developing PCM)
Spherical tip
Pyramidal tip
o lIm/s m 6.251Jm/s a 141Jm/s
(a)
•
T
S(b)
T
I
I
I
I
T
1.0
0.5
0
0l)'
yv
Cmh
Day 0
7
14
21
28
Figure 3-7 Time constants obtained from stress relaxation measurement using a nanotip
(a, c) and a micron-sized spherical tip (b, d). ti corresponds to the time constant of the
short-time transient behavior and _ 2 to the longer-time transient response of newly
developing PCM (Two-way ANOVA, p<0.05; n=7 cells, mean±SE; A dotted line
indicates the separation between day 0 chondrocytes lack of PCM and cultured
chondrocytes with newly developing PCM)
3.3.3 Stress relaxation of newly developing PCM of cultured primary chondrocytes
(Day 7-28)
Eo, t1 and c2 were dependent on the type of probe tips used (i.e., a nanosized
pyramidal probe tip and a micron-sized spherical probe tip), culture duration, and the
displacement rate (p<0.05, 3-way ANOVA), whereas E-, were not significantly affected
by those three factors.
Eo and Eoo For the measurement using a nanosized tip, a pair-wise comparison of
Eo shows that PCM of day 7 is significantly different from that of day 28, being both
measured at the displacement rate of 6.25
im/s (Figure 3-6(a)). However, the
deformation rate dependence of Eo was rather modest, and culture duration effect was
more dramatic on Eo of newly developing PCM when the subset of each displacement
rate was pooled together and compared with respect to different culture duration; for
example, the group of day 28 is significantly different from that of day 7 and day 14.
Similarly, the measurement using a micron-sized spherical tip also demonstrated that the
group of day 28 is significantly different from that of day 7 (Figure 3-6(b)).
Again, culture duration is more dominant factor to affect E.o of newly developing
PCM, compared to the effect of displacement rate, when probed by both tips. For
nanosized tips, the group of day 28 is significantly different from that of day 7 (Figure 36(c)). For micron-sized tips, the group of day 28 is significantly different from that of day
7 and 14 (Figure 3-6(d)).
zr and r2 Interestingly, the effect of displacement rate on tl was clearly
demonstrated for both probe tips. For day 14, 21 and 28, T-of higher deformation rate
(6.25 Jim/s (day 21) and 14 ptm/s (day 14-28)) was significantly different from that of 1
ptm/s on the same culture duration (Figure 3-7(a)) probed by a nanosized pyramidal tip.
Similar deformation rate dependence of ti was observed on day 21 using a micron-sized
spherical probe tip, although the effect was rather modest. For r2 , a nanosized probe tip
only revealed the deformation rate dependence on day 14. Meanwhile, Tz and z2 did not
vary much with culture duration probed by both tips (Figure 3-7).
Tr and
z2 was in the
range of 0.2-1.4 and 5-25 seconds, respectively, being similar for both probe tips.
3.4 Discussion
3.4.1 Viscoelastic properties of day 0 chondrocytes
For day 0 chondrocytes, the time-dependent moduli show a good agreement with
other studies despite of the difference in species of cell source and mechanical testing
method (5, 14, 32, 57). However, the long-term characteristic time constant varies with
the mechanical testing method (-20 sec. in our study vs. 2-50 sec.) (5, 31-33, 39, 57),
which is likely prone to the testing configuration of cell (cells put in pyramidal well vs.
free suspension for micropipette aspiration (39, 57), cells adhered to the substrate for
cytoindentation (31) and unconfined compression (32, 33), and AFM-based relaxation
(5)), the type of transient loading (stress relaxation for AFM-based indentation(5) vs.
63
creep for micropipette aspiration (39, 57), cytoindentation (31), and unconfined
compression (32, 33)), as well as the source of exerted force during deformation
(compressive force in cytoindentation (31), unconfined compression (32, 33) and AFMbased relaxation (5) vs. suction pressure in micropipette aspiration (39, 57)). Meanwhile,
the latter can be a minor effect on the time constant since chondrocytes are fairly
incompressible (v - 0.4) and shear force may be primarily dominant at micron-level
deformation (5).
Viscoelastic properties of chondrocytes are attributable to flow-dependent (fluidsolid interactions and fluid viscosity) and flow-independent mechanisms (viscoelasticity
of the cytoskeleton) (2). In particular, microfilaments and intermediate filaments play a
dominant role in the viscoelastic properties of chondrocytes, possibly due to their
mechanical properties and distribution within the chondrocytes (14). Furthermore,
cytoskeletal alteration is known to be associated with cartilage degeneration;
osteoarthritic chondrocytes represent higher stiffness and varying amounts of cytoskeletal
actin and vimentin, compared to non-osteoarthritic chondrocytes (87).
3.4.2 Culture duration effect on Eo and E. of newly developing PCM
Our study of stress relaxation characterized the temporal evolution of both E0 and
E. of newly developing PCM, as well as the effect of the probe tip geometry on the timedependent mechanical properties of newly developing PCM. In the previous study of Ng,
et al (4), an increase in cell-PCM moduli with culture duration was consistent with
increased PCM synthesis during time in culture (4). Similarly, our study demonstrated
that the time-dependent mechanical properties of day 28 PCM are significantly different
to those of day 7 and 14, though much lower equilibrium stiffness of newly developing
PCM over 28 days (- 2kPa of newly developing PCM vs. 40-70 kPa of mature PCM
(60)) remains to be improved. Previously, a theoretical model prediction showed that the
mechanical properties of PCM dramatically change the stress-strain profile and the local
fluid flow environment within the chondrocyte as well as at the cell-matrix boundary (2,
74), which can subsequently affect the gene expression of nucleus inside the
chondrocytes and the subsequent biosynthetic activity (60). Indeed, the stimulatory
loading of chondrocytes led to more biosynthesis of proteoglycan and higher elastic
moduli (77). With this regard, newly developing PCM of our study might require such a
mechanical regulation for the in vitro culture system to achieve the mechanical properties
comparable to those of mature PCM.
3.4.3 Tip geometry dependence of Eo and Eoo of newly developing PCM
Eo of newly developing PCM was more sensitive to tip geometry, culture duration,
and displacement rate (3-way ANOVA; p<0.05), compared to Eo (2-way ANOVA of
culture duration and displacement rate; p<0.05). Since Eo is measured from the very
initial response of PCM immediately after applying a I tm deformation and Eo from the
later response once the PCM is fully relaxed at -1 p.m deformation, our hypothesis is that
the tip geometry effect gets less significant as the structure of PCM becomes more
compact and homogeneous at fully relaxed state of PCM compared to the initial PCM
structure. With this regard, the previous study of Ng, et al (4) reported that a force-
displacement (F-d) curve at large deformation (i.e., displacement > -1 lm) represents
higher force for a nanosized pyramidal tip than for a spherical micron-sized probe tip at
day 28. In addition, the magnitude of prefactor at the indentation depth of -1
m in
equation 3-1 and 3-2 is -50% higher for a spherical micron-sized probe tip as compared
to a nanosized pyramidal tip (2.73 vs. 1.83). Taken together, higher force and smaller
prefactor for a nanosized pyramidal probe tip leads to higher E0 and E. as compared to a
spherical micron-sized probe tip, while the difference due to tip geometry is modest in
Eoo.
3.4.4 Deformation rate dependence of viscoelastic
time constants of newly
developing PCM
Interestingly, the change of initial relaxation time constant, tI, with the
displacement rate was detectable for both probe tips. T was significantly dependent on tip
geometry, culture duration, and displacement rate (3-way ANOVA; p<0.05). Our
findings of t suggest that the very initial time-dependent response is better captured with
a nanosized pyramidal tip to probe the local response of newly developing PCM,
compared to a micron-sized spherical probe tip. No significant change of TI with the
culture duration indicates that the initial time-dependence of PCM might be largely
dominated by flow-dependent
mechanism. Higher deformation rate results
in
significantly shorter z, compared to that of 1 plm/s on the same culture duration for days
14-28 (Figure 3-7(a)); the reduced time is given for the fluid to be effluent when the
probe tip rapidly deforms hydrated PCM. Furthermore, this observation of liis consistent
with more solid-like PCM behavior and higher Eo at higher deformation rate (Figure 3-6
(a,b); 3-way ANOVA of Eo for tip geometry, deformation rate, and culture duration,
p<0.05). Meanwhile, E. is estimated once the PCM reaches its equilibrium state with
ltm deformation, leading to E. less dependent on the deformation rate (Figure 3-6(c,d)).
2,
the longer-term relaxation time constant, shows large variation, weakly
sensitive to tip geometry and displacement rate. As the tip probes the deformed structure
with longer time, the overall mature state of PCM of each chondrocyte (e.g., crosslinking
density and the proportion of components composed of PCM) might play a more
dominant role in the time-course of loading profile. Therefore, z2 is likely more
influenced by the heterogeneity in the mature state of PCM, unique to each individual
cell, even among the same culture duration (Figure 3-7(c,d)). We hypothesize that the
relaxed structure of PCM might better reflect the overall structural feature as a whole
from all the contribution of each component of PCM in a specific cell. Taken together
with observation on E0 and E., the culture duration affects significantly the initial and
final mechanical state (i.e., E0 and E.) and yet its time-profile of the path between initial
and final state is more dependent on the properties of PCM specific to each chondrocyte.
One possible mechanism for the time dependent mechanical behavior of PCM
might be that aggrecans is involved in the dominant mechanism of energy dissipation in
shock-absorbing tissues. Its higher viscosity of aggregated form (i.e., aggrecan aggregate;
aggrecan monomers are physically bound onto a hyaluronic acid chain (88)), compared to
the monomer itself, provides an efficient mechanism of energy dissipation at low
frequency deformation (i.e., the reptation movement), and yet the internal dynamics of
aggrecan monomer dominates the rheological properties at high frequency deformation
(88). Interestingly, the increase in the relaxation time (i.e., the time for the polymer for
reptation) was observed as a function of the concentration, and the aggrecan aggregates
exhibited longer relaxation time than the aggrecan monomers in particle tracking
microrheology experiments (89). The modulus of aggrecan (-1 Pa at 5wt%) further
supports its role as a dissipative filler between relatively rigid collagen fibers (2- 40 MPa,
collagen in articular cartilage (90)) (89).
Our study of mechanical testing might fall into a low frequency regime of
viscoelasticity, while the weak dependence of the viscoelastic time constants on culture
duration is indicative of little change in the concentration of aggrecan aggregate, despite
of increasing biosynthesis of GAG over time (4). In this context, another study
demonstrated that the rate-dependence of lateral force is more pronounced with
increasing ionic strength given in the range of 0.001-1.0 M using AFM-based
nanoindentation, reflecting that the relative importance of non-electrostatic components is
increased with ionic strength (91); possible mechanism of the rate dependence might
include
viscoelastic
processes
(e.g.,
molecular
friction,
molecular
reorientation/reconfiguration, interpenetration and entanglement of adjacent aggrecan)
and poroelastic rate processes driven by local fluid flow within and through aggrecan
molecules. Yet, the contribution of such interpenetration/entanglement between aggrecan
molecules might be negligible in our system due to low packing density of aggrecans and
random orientation in newly developing PCM as compared to the study of Han, et. al.
(91) (See the next section (3.4.5) for the estimation of hydraulic permeability of newly
developing PCM indicative of more porous and permeable structure as compared to
mature PCM).
Even though we might improve the prediction of experimental data with more
complicated viscoelastic model, the biological phenomena behind its viscoelastic
component is largely uncorrelated. In particular, each characteristic time constant is hard
to interpret in relation to specific component of PCM. Therefore, our next goal is to apply
the dynamic oscillatory compression to cultured chondrocytes in order to capture the
time- and frequency-dependent response, which might provide an insight on the
molecular dissipation mechanism of newly developing PCM exposed to various
frequency of deformation.
3.4.5 Comparison between newly developing PCM and mature PCM in cartilage
tissue engineering perspective
So far, most studies investigated the time-dependence of mature PCM at cellular
level using micropipette aspiration or unconfined compression coupled with biphasic
model (60). Since biphasic theory models the articular cartilage as a mixture of a porous,
permeable elastic solid and a movable interstitial fluid, the theory yields the material
properties of the articular cartilage in terms of Young's modulus, hydraulic permeability,
and Poisson's ratio (92), making it difficult to directly compare between newly
developing PCM in our study and mature PCM in previous studies. So far, material
viscoelastic properties of mature PCM is either experimentally determined using biphasic
model (16) or theoretically calculated by parametric studies using finite element analysis
(2). Based on parameters experimentally determined in the previous study of
Alexopoulos et al (16), the slowest characteristic poroelastic time constant, t,, is about 15 sec., estimated by the poroelastic theory (z,=L 2 HK; L- 7.8 tm for PCM thickness, H
- 40kPa for Young's modulus of PCM, and
K
_10xl0 17 m4 /Ns for hydraulic
permeability of PCM) (92). Another study of Han, et al recently performed AFM-based
stress relaxation using bovine cartilage and obtained E0 - 166 kPa, E. - 146 kPa, and 25 sec., using a 3-element viscoelastic model (93). In this study, a significant decrease in
- was also observed in cartilage after proteoglycan was almost completely digested by an
enzymatic digestion (>95% of matrix proteoglycan), implying that depletion of GAGs
sharply affects the solution viscosity and the following movement of collagen fibrils.
In comparison to viscoelastic properties of ECM, mature PCM has 3-4 times
lower viscoelastic stiffness and newly developing PCM represents two orders lower in
magnitude of viscoelastic stiffness. Again, even at the macroscopic tissue level, a 3element viscoelastic model results in a poor fit of experimental relaxation curve, in
particular, for the time < I sec. from the peak force. One possible explanation of this
observation may be that the early response of the time-dependence in mechanical
behavior of cartilage at macroscopic tissue level, as well as in the local environment of an
individual cell with its PCM, is more influenced by fluid-flow and local pressure gradient,
rather than the intermolecular reptation and/or friction between component of ECM (93).
The latter factors may contribute more to the time-dependence of viscoelastic behavior at
longer time scope. Along with viscoelastic material properties, the hydraulic permeability
of newly developing PCM can be estimated to be 4.2x10- 3m4/N-s from the poroelastic
2 /p,,
theory (K =LL
where ,p=15 sec., L = 3.5 pm (4), H = 1.5 kPa for day 28 PCM),
implying that newly developing PCM has a very porous structure, compared to much
lower permeability of mature PCM (- 4 x 10- 17 M4/N's (16)). Possible implications of
higher permeability of newly developing PCM with its lower modulus in the mechanical
environment of chondrocytes might include the change in the flow-generated streaming
potentials, shear stress, or solute transport to the chondrocyte (74). Indeed, osteoarthritic
chondrocytes having decreased elastic modulus of PCM/ECM and increased hydraulic
permeability demonstrated a sharp increase of the compressive strain at the cell when
subjected to the same deformation as in normal chondrocytes (16, 57), supporting a
working hypothesis that different mechanical stimuli further perturb the following
metabolic activity and gene expression of chondrocytes in osteoarthritic cartilage.
3.5 Conclusions
Our study of newly developing PCM using AFM-based stress relaxation
techniques demonstrated that time-dependent moduli of newly developing PCM
significantly increase with culture duration and characteristic time constants are
distinctive to tip geometry and deformation rate. Here, we quantified the timedependence of mechanical parameters under a constant deformation that was previously
difficult to measure accurately or directly using other techniques. Such information
would serve to improve the model prediction of the local mechanical environment of cellPCM. In particular, our study of newly developing PCM can be valuable to the cartilage
tissue engineering since it explores the local mechanical environment of cell-PCM that
undergoes the very initial stage of development. However, due to the limitation of a
viscoelastic model used, the very initial decaying response of newly developing PCM in
stress relaxation curve was not fully addressed, which can further illuminate the detailed
mechanism of poro-viscoelastic behavior inside. For this purpose, dynamic oscillatory
compression will be utilized to probe the newly developing PCM.
Chapter 4
Dynamic mechanical properties of the tissue-engineered
matrix associated with individual chondrocytes
4.1 Introduction
Cell-based tissue engineering holds great potential for therapies involving
regeneration and/or replacement of damaged cartilage (94, 95). Such approaches typically
involve seeding cells (e.g., chondrocytes, stem cells) within macromolecular scaffolds or
in scaffold-free environments and subjecting them to stimulatory biochemical and/or
mechanical factors in vitro to promote synthesis and growth of neo-tissue for subsequent
implantation in vivo. The ultimate goal and challenge is to produce a material with
structural, biochemical and biomechanical properties similar enough to healthy cartilage
so that upon maturation in vivo it can restore physiological function. To achieve this
objective, it is important to understand the temporal evolution of the properties of the
newly synthesized matrix. At early stages, matrix is formed around individual cells
within the scaffold and these cell-matrix composites are often isolated from each other.
As culture time increases, growth of neighboring cell-associated matrices leads to the
formation of a more continuous neo-tissue which undergoes further evolution in structure
and properties in vitro or in vivo. Throughout this process, the cell-associated matrix is
important in facilitating cell signaling and mechanotransduction (96, 97). Knowledge of
the properties of the in vitro-generatedmatrix provides an assessment of the quality and
ultimate success of a given tissue engineering approach and has great potential to be
utilized for optimization.
Recently, we reported atomic force microscope (AFM)-based indentation of the
newly synthesized cell-associated matrices of individual chondrocytes to quantify their
quasistatic mechanical properties (4). This approach provides an advantage over
macroscopic testing of the entire tissue-engineered construct, which is complicated by the
response of the cells and scaffold, a significant effect at early times (77). In addition, such
local measurements enable the quantification of cell-to-cell heterogeneity and the tissue
engineering process in general, while macroscopic experiments yield a homogenized
response. For chondrocytes cultured in fetal bovine serum (FBS)-supplemented medium
in alginate scaffolds, the stiffness of the newly-formed cell-associated matrix increased
from -0.1 kPa on day 7 to 0.17 kPa by day 28 (4), a value much less than that of the
pericellular matrix of mature intact cartilage chondrons (-60 kPa) (15), as well as
cartilage extracellular matrix (-1 MPa) (98). The use of insulin-like growth factor-i
(IGF-1) and osteogenic protein-i (OP-1) supplementation significantly increased matrix
stiffness -25-fold to 4.15 kPa on day 28 compared to FBS culture on day 28, consistent
with observed increases in total GAG accumulation, and indicative of a more mature
structure (4).
The present study investigates the dynamic oscillatory mechanical behavior (4042, 44, 99) of the cell-associated matrices of individual chondrocytes cultured in alginate
scaffolds up to 28 days. Such time-dependent behaviors are important because tissueengineered constructs implanted in vivo are expected to undergo cyclic and impact
loading that includes frequency components as high as I kHz, just as native cartilage
does (45). Given the known hydrated and porous microstructure of the tissue-engineered
cell-associated matrix (9), rate-dependent poro-viscoelastic processes are expected (76).
The dynamic nanomechanical experiments involved al pm indentation displacement
followed by a transient force relaxation. After 180 seconds, a dynamic oscillatory
displacement was then applied (5nm amplitude, frequencies, f-I -316 Hz) corresponding
to loads of 0.01-0.1 nN in amplitude. The objectives of this study were (1) to establish the
experimental and theoretical methodologies for quantifying the dynamic mechanical
properties of newly developed tissue-engineered matrix associated with single cells, (2)
to investigate the temporal evolution of these properties with increasing time in 3D
alginate gel culture, and (3) to quantify the effect of specific growth factors (IGF-I and
OP-1) on the dynamic mechanical functionality of the newly developed matrix.
4.2 Materials and Methods
4.2.1 Cell source and culture
Chondrocytes were obtained from femoral condyles of 1-2 week old bovine
calves and seeded into 2% alginate scaffolds, as described previously (4). Cell-seeded
alginate beads were maintained either in high-glucose Dulbecco's Modified Eagle
Medium (DMEM) with 10% FBS or in mini-ITS (Insulin-Transferrin-Selenium) with the
combination of IGF-1 (100 ng/mL) and OP-1 (100 ng/mL) (growth factor (GF) culture)
(100). On days 7, 14, 21, and 28, cells were released from beads and suspended with
high-glucose DMEM. Previous studies demonstrated that, under these same culture
conditions, the cell-associated matrix was -3 lim thick from day 7-28 (4).
4.2.2 Dynamic oscillatory compression
Using an AFM probe tip, each chondrocyte and its cell-associated matrix was
positioned into an inverted pyramidal well of a silicon substrate in DMEM (Figure 4-la).
A multimode Nanoscope IV AFM with PicoForce piezo (Veeco, Santa Barbara, CA) was
then used to perform compression measurements on each cell-matrix composite within its
well. Both pyramidal (Si 3N4, cantilever spring constant, K-0.06 N/m, end radius, Rtip -50
nm, NP-20, Veeco) and spherical colloidal probe tips (Silica beads from Bang Labs,
#SS06N, tipless cantilever from Veeco, NP-020, K-0.06 N/m, Rtip -2.5 jam) were used.
Script mode (Nanoscope IV software, V613rl) was programmed to apply a I lam ramp
indentation at a displacement rate of I jim/s. Force relaxation at constant displacement
was then measured during the subsequent 180-second interval (Figure 4-1b,c). After the
force reached quasi-equilibrium within -100 seconds, a sinusoidal displacement (5 nm
amplitude) was superimposed on the I ptm static offset displacement using an external
wave generator (Rockland 5100) connected to the AFM piezoelectric scanner. The
magnitude and phase angle of the resulting force were measured at discrete frequencies
of 1, 3.16, 10, 31.6, 100, and 316 Hz (Figure 4-1c). The force magnitude at lower
frequencies (<1 Hz) was below the AFM detection limit and software limitations
prevented testing above 316 Hz. Measurements were repeated 3 times at the same
location on each cell-matrix composite.
4.2.3 Data analysis and statistical tests
Dynamic force and displacement data for both probe tips were analyzed using
Fast Fourier Transforms to obtain the magnitude of the complex modulus, IE*(f)l, and
phase angle,
(f), between measured force and applied displacement. Application of
sinusoidal displacements even as low as 5 nm in amplitude, with either micron-sized or
nano-sized probe tips, resulted in measured sinusoidal force waveforms having total
harmonic distortion less than 5% over the entire frequency range of interest. 6 ) was
corrected for the effect of hydrodynamic drag (101). For the micron-sized spherical probe
tip, IE*(f) was obtained from a Taylor series expansion of the Hertzian contact model
(42, 102):
E h e c la( f
E* (
)
_1-v V22 F(f)
F(f)
2sphecald(f)
2Rd
(4-1)
where do is indentation depth at zero frequency, Poisson's ratio, v, was assumed to be 0.4
(103), R is the effective radius of the probe tip and cell-matrix composite (i.e., 1/R =
[1/Rspherical tip]+[ 1/Rcell-matrix composite]),
F(f) is the magnitude of the force at frequency f and
d(f) is the magnitude of the deformation at frequency f during dynamic oscillatory
compression (4) (Figure 4-1c). The Hertzian model approximation used in this study is
sufficient for assessing relative trends in the temporal evolution of matrix stiffness with
culture time, which is the primary focus of this study. The real [E'()] and imaginary
[E"(/)] parts of the complex modulus (i.e., storage and loss moduli) are also reported, as
well as
E*pyramtidal(f)
obtained using the pyramidal probe tip.
For the pyramidal probe tip, the magnitude of the complex modulus, E*(f) , and
the phase angle, (f), between the measured force and the applied sinusoidal displacement
(Figure 4-id) can be similarly expressed in terms of (the real) storage modulus, E'(f, and
the (imaginary) complex modulus, E"(f), as follows:
E (f) = E'(f) + jE"(f)
(4-2)
E (f)l = E'(f) 2 + E"2 (f)
(4-3)
tan (f) =E"(f)
(4-4)
E'(f)
For the pyramidal probe tip, JE*(J), was calculated using a Taylor series
expansion of the Hertzian contact model (42, 102):
1-v
Epyramidal
2
F(f)
= 27rndo tan a d(f)
(4-5)
where do is indentation depth at zero frequency, Poisson's ratio v is assumed to be 0.4
(103), the pyramidal tip opening angle a is 350(4), F(f) is the magnitude of the force and
d(f) is the magnitude of the applied displacement at frequency f during dynamic
oscillatory compression. For both AFM probe tips, it is assumed that IE*(f)
is not
significantly affected by the compression of the cell into the silicon well, and that the cell
or cell-matrix composite is modeled as a uniform, isotropic medium during compression
of the sphere (4).
To test our working hypothesis that culture duration and GF treatment
significantly influence the dynamic properties of cell-matrix composites, each effect
(culture duration, GF treatment, frequency and tip geometry) on |E*(f) and
$()
was
tested using one-way ANOVA followed by Tukey's post hoc test for comparisons. All
data are expressed as mean±SEM, with p <0.05 as statistically significant. Statistical
analyses were performed using Systat-12.0 (Richmond CA).
(a)
AFM cantilever
oscillatory displacement
-5 nm amplitude
colloid, R,, - 2.5 gm
stkindentation
1 pm
microfabri cated Si substrate
in cul ture medium
cell, Rce. ~i 3.8 pm
matrix, 3 tpm thick
square pyramidal well
0.05 -n
0
0
z
-0.05
Time(sec)
Time(sec)
Figure 4-1 Dynamic oscillatory compression of individual chondrocytes and their
associated matrix in FBS culture using a spherical tip: (a) Experimental setup:
Chondrocyte-matrix composites in microfabricated silicon wells (radius of cell = Rcell,
radius of probe tip = Rtip, matrix thickness = 3jtm, indentation depth = 1l m, and small
sinusoidal oscillation = 5nm; the detailed structure of the matrix has been described
previously (1, 3)), (b) Force (red) and indentation depth (blue) curve showing the initial
indentation (depth, do = 1l m) and subsequent force relaxation exemplified by a day 14
chondrocyte, and (c) inset showing the applied dynamic sinusoidal deformation (d(f)) and
resulting dynamic force (F(f)) shifted from the displacement by the phase angle 6(l)
4.3 Results
4.3.1 Dynamic mechanical properties of individual chondrocytes and their cellassociated matrix using micron-sized spherical tip
Effect offrequency. E*I and 5 increased nonlinearly with applied displacement
frequency in both FBS (Figure 4-2) and GF (Figure 4-3) cultures at each time point. The
effect of frequency on I E*I and 5was significant for both FBS and GF culture medium
with both the spherical micron-sized tip (Figures. 4-2a,b, 4-3a,b) and the pyramidal
nanosized tip (Figures. 4-5a,b, 4-6a,b, 4-7b) by one-way ANOVA (p<0.05). In FBS,
IE*I and
6 at 316 Hz were significantly different from that at all lower frequencies
during the entire culture period (Figure 4-2a,b, p<0.05). Similarly, for the nanosized tip,
IE*I and 6 at 316 Hz were significantly different from all lower frequencies during the
entire culture period for both FBS and GF culture (Figure. 4-5a,b, 4-6a,b, p<0.05).
Effect of culture duration. IE'() I,E'(f), E"(f) increased and f)decreased with
culture duration in both FBS (Figure 4-2) and GF (Figure 4-3). The effect of culture
duration on IE*(f) and 6() was significant for both FBS and GF medium with the
micron-sized (Figure 4-2a,b, 4-3a,b) and nanosized tip (Figures. 4-5a,b, 4-6a,b) (one-way
ANOVA, p<0.05). For GF culture, pair-wise comparisons showed that IE*f) I obtained
with the micron-size tip on day 14-28 at higher frequencies were significantly higher
from that on day 7 (p<0.05 for 316 Hz, Figure 4-4a; p<0.005 for 100 Hz (not shown)).
Effect of GF treatment. I E*(f) I, E'(f)increased significantly with the addition of
GF compared to FBS-supplementation (ANOVA, p<0.05). In GF culture, 6(f) was
significantly lower than that in FBS at both low and high frequencies (1-10, 316 Hz)
(p<0.05; Figure 4-4b). There was also a significant effect of GF treatment using the
nanosized tip (p<0.05, Figure 4-6). Interestingly, the quasi-static force-displacement
(indentation) curves of individual chondrocyte-matrix composites (Figure 4-8a,b)
demonstrated that GF-cultured samples exhibited a much higher indentation stiffness than
FBS-cultured samples when utilizing alI pm indentation displacement. In contrast, while
the dynamic stiffness
IE*f)l
measured at 5 nm displacement amplitude increased
significantly with GF-treatment, the increase was modest (-1.5-fold).
Individual contributionsfrom cell and matrix. Indentation data analyzed from the
initial force-displacement (ramp) portion (Figure 4-lb) of the experiments (Figure 4-8a,c)
showed that day 7-28 cell-matrix composites cultured in FBS were much more compliant
than day 0 cells with no surrounding matrix, consistent with our previous study (4). Finite
element analysis of this system by Ng, et al. indicated that with FBS treatment, cell
deformation was negligible compared to matrix deformation and, therefore, the measured
indentation stiffness was dominated by the matrix (4). Hence, the dynamic mechanical
properties reported here for the FBS culture conditions (Figure 4-2) can be assumed to be
dominated by the properties of the cell-associated matrix. However, for the GF cultures,
nanoindentation data (Figure 4-8b,d) show that the indentation stiffness of the cell-matrix
composite increased substantially over 28 days, eventually approaching the stiffness of
the day-0 cell with no matrix. Therefore, under dynamic loading conditions, the dynamic
mechanical properties of Fig. 4-3 likely include contributions from both the cell and
matrix.
4.3.2 Dynamic mechanical properties of individual chondrocytes and their cellassociated matrix using pyramidal nanosized tip
IE() I and 6(f) for the cell-matrix composites cultured in FBS (Figure 4-5) and
GF (Figure 4-6) observed using the pyramidal nano-sized probe tip (ANOVA, p<0.05)
exhibited similar trends compared to that obtained with the spherical micron-sized probe
tip (Figures 4-2, 4-3). In both FBS and GF culture, IE* Iincreased significantly with f
though most dramatically at higher frequencies. The effect off on 5was also significant;
for example, Figure 4-7b shows the frequency dependence for day-28 cells in GF culture.
The effect of culture duration on IE*9 I was also significant at each frequency
(ANOVA, p<0.05); Figure 4-7 shows the trend at 316 Hz between day 7 and 21 in GF
culture (p<0.05). GF treatment led to higher IE*(f) I at longer culture times, particularly
at higher frequencies (100-316 Hz) compared to FBS culture (Figure 4-6a vs. Figure 45a). Also, GF treatment led to a significant reduction in 6(f) over time in culture
compared to FBS culture (p<0.05).
The effect of tip geometry on IE*09
1 was
not significant in this 28 day culture
study. A possible interpretation is that after initial application of the -1 lm deformation,
the small amplitude oscillation (5 nm) involves approximately similar contact areas for
both probe tips sizes once embedded within the matrix. In contrast, the matrix stiffness
detected using larger 1 pim indentation deformation was much larger for the micron-sized
colloidal tip (Figure 4-8).
1
(a) 30' 10% FBS
-a- Day
-r- Day
CL 20 -o- Day
-X-Day
7 (n=5)
14 (n=5)
21 (n=4)
28 (n=5)
(b)
0
10
0
1
10
100
1
Frequency (Hz)
I,'A
Iul
.4IA
10
100
Frequency (Hz)
In'
(d) II
u0 110% FBS
JU
10% FBS
10
0.1 so*
0.1
Fq10
100
Frequency (Hz)
0.
I'a •
1i
100
Frequency (Hz)
Figure 4-2 Dynamic mechanical properties of chondrocytes and their newly developing
cell-associated matrix up to 28 days in FBS culture (10% FBS) using the spherical probe
tip: (a) I E*(f),
(b) 60), (c) E'(f) and (d) E"(f). n=number of cells, mean ± SEM;
maximum SEM is -4kPa for E'(f) and -6kPa for E"(f), respectively.
(a) 30
(b) 90 IGF-1+OP-1
60
6.20
w
(mean±SEM)
0
1
10
(Hz00
Frequency (Hz
(c) 100
0
10
10
Frequency (Hz)
(d) 1 0 0 IGF-1+OP-1
0.1
0.1
1
10ency (
Frequency (Iz) 0
0.011
1
FFrequency
10
(Hz
Figure 4-3 Dynamic mechanical properties of chondrocytes and their cell-associated
matrix up to 28 days in GF culture (IGF-I+OP-1) using the spherical probe tip:
(a) E*(fl I, (b) 3(j), (c) E'(f) and (d) E"Y). n=number of cells, mean ± SEM; maximum
SEM is -2kPa for E'(fP and -IkPa for E"()f, respectively.
O
0%&%.
(a) U IGF-I+OP-1 (316Hz) § 8p<0.05
%
§t
-20'
(b) 90
'
10
10% FBS
* IGF-1+OP-1
*p<0.005
60 (mean+SEM)
30
1;*
m
LU
7
21
14
Day
28
0
Day 21
(1 Hz)
=
Day 28
(316 Hz)
Figure 4-4 Effect of culture duration and growth factor treatment on dynamic oscillatory
properties of chondrocytes and their newly developing matrix using the spherical probe
tip: (a) the effect of culture duration on IE*9 I for GF culture chondrocytes (one-way
ANOVA: frequency) and (b) the effect of growth factor treatment on 6(f (one-way
ANOVA: GF treatment).
AA
(a) 30- 10% FBS
(b) ~U 110% FBS
-a- Day 7 (n=5)
Day 14 (n=5)
20 o -e- Day 21 (n=4)
-x-Day 28 (n=5)
30
~10w
0'
(c) 100
1
10
100
Frequency (Hz)
10% FBS
O
o
:1
10
iio
100
Frequency (Hz)
(d) 100
10% FBS
10
10
Eu
60
meCS~
0.1
0.1
110
100
Frequency (Hz)
(mean±SEM)
1
10
100
Frequency (Hz)
Figure 4-5 Dynamic mechanical properties of chondrocytes and their newly developed
matrix up to 28 days in FBS culture using the pyramidal probe tip: (a) I E*() I , (b) 69,
(c) E'(f and (d) E"(f). n=number of cells, mean + SEM; maximum SEM is -2kPa for
E'(f and -3kPa for E"(f, respectively.
(a) 30
(b)
S60
M120
30
I10
U
IGF-1+OP-1
0.
1
10
10
100
Frequency (Hz)
(c) 100 IGF-1+OP-1
100
Frequency (Hz)
(d)100
:-1 +OP-1
10I
10
0.1
0.1
0.01
1
1Frequency0
1(H00
Frequency (Hz)
1
(mean±SEM)
1requency 1(Hz00
Frequency (Hz)
Figure 4-6 Dynamic mechanical properties of chondrocytes and their cell-associated
matrix up to 28 days in GF culture (IGF-I+OP-1) using the pyramidal probe tip:
(a) IE*(f 1, (b) 6~(), (c) E'(f) and (d) E"(f). n=number of cells, mean ± SEM; maximum
SEM is -2kPa for both E'(f) and E"(f).
(a) 30 IGF-I+OP-1 (316Hz)
p<0.05
(b) 90 IGF-1+OP-1(Day 28) §p<0.05
§c
,-
Z
O60
7
14
21
Day
28
§
-------
....
1 3.16 10 31.6 100 316
Frequency (Hz)
Figure 4-7 Effect of culture duration and growth factor treatment on dynamic oscillatory
properties using the pyramidal probe tip: (a) the effect of culture duration on IE*(f) for
GF culture chondrocytes (one-way ANOVA: culture duration) and (b) the effect of
frequency on 6(f) for GF culture chondrocytes (one-way ANOVA: frequency).
4.3.3 Indentation stiffness of single chondrocytes with GF treatment during time in
culture
For all of the experiments of Figures 4-2, 4-3, 4-4, a force-displacement (F-d)
curve was obtained during the initial loading phase of each cell (Figure 4-1c). This F-d
curve was also used to characterize the quasistatic behavior of each cell prior to force
relaxation and subsequent dynamic loading. The F-d curves for cells in FBS and GF
culture as a function of time in culture are shown in Figure 4-8 for both micron-sized and
nano-sized probe tips. F-d curves varied with culture duration and tip geometry in both
medium conditions (two-way ANOVA, p<0.05) (Figure 4-8). For each probe tip, the
stiffness (slope of the F-d curve) depended significantly on the GF treatment and culture
duration as well (two-way ANOVA, p<0.05). Chondrocytes cultured in FBS medium
(Figure 4-8a,c) up to day 28 exhibited much lower stiffness than day 0 chondrocytes
which had no cell-associated matrix, while chondrocytes treated with GF exhibited
stiffness similar to or greater than day 0 cells after day 21 Figure 4-8b,d). These trends
are consistent with those reported in a previous study (4).
(b)
(a) 2.0 10% FBS(n=4-5)
Day 28
1.5.
i
i
Day 0
SDay 0
,
44 Day 21
1I
Day 14
1.0
0.5
Day 28
Day 21
II
,a~d
,"_ilil
al,,=tlll
i
Day 14
Day 7
I Day 7
500
101
500
Indentation depth (nm)
(d) 2.0 IIGF-1+OP-1 (n=5)
(c) 2.0 10% FBS(n=3-5)
Z
1.5
z
,Day 0
• 1.0
u.5
0.5
1000
Indentation depth (nm)
ih
ii4
,s',,
f~ir
1.5.
i Day 28
SI4 Day 21
1. 0
i Of'Day 0
Day 28 0
0. 5
,Day 21
x SDay 14
'Day 7
1000
500
Indentation depth (nm)
7
Day 14
aI1111I Day
500
1000
Indentation depth (nm)
Figure 4-8 Force-displacement curves of individual chondrocytes and chondrocytematrix composites using each probe tip: (a,c) FBS culture and (b,d) GF culture (IGFl+OP- 1). n = number of cells, mean ± SEM
4.4 Discussion
An AFM-based single cell mechanics approach was utilized to confirm that the
dynamic mechanical properties of the newly synthesized tissue-engineered matrix
associated with individual chondrocytes are frequency-dependent and change with
culture duration and GF treatment. Dynamic oscillatory nanomechanics methods (99)
were combined with a unique experimental protocol incorporating an inverted pyramidal
Si well (4) (Figure 4-la) to hold each cell-matrix composite in place, facilitating a more
physiologic-like morphology. Here, we focus on data obtained using the micron-sized
probe tip and cells cultured in FBS, for which the matrix was much softer than that of the
cell itself, and hence, the dynamic mechanical data for this system can be assumed to
represent solely the properties of the matrix. Following, we use scaling approximations to
identify the rate-dependent processes and energy dissipation mechanisms that underlie
the observed dynamic behavior, and then we discuss the effect of culture duration on
indentation versus dynamic stiffness and, finally, the effect of GF stimulation.
Mechanisms underlying dynamic frequency behavior. In general, the frequency
dependence of IE* I , E', E", and 5 may be associated with fluid flow-dependent
poroelasticity
and/or flow-independent viscoelastic processes.
To explore these
possibilities, we first estimated the characteristic poroelastic relaxation time,, IP
[L2/(Hk)] (also called the 'gel diffusion time') (92, 104-107), where L is the characteristic
length scale over which fluid flows, H is the elastic modulus, and k is the hydraulic
permeability of the cell-associated matrix. L can be approximated as the radius of the
probe tip, 2.5 pim (see Appendix B.2 for detailed discussion); H is approximated by
quasi-static indentation modulus, -0. 12 kPa for day 14 and 0.17 kPa for day 28 (4); and k
can be estimated from the measured GAG content (4) to be -6.6x10 -2' m4/N for day 14
and 1.7x10-12 m4/N-s for day 28 in FBS culture by using the unit cell model (108, 109)
assuming that GAGs are the primary determinant of the permeability (see Appendix B.2).
-pof the newly developing matrix was thereby estimated to be -0.01-0.02s from days 1428, corresponding to characteristic frequencies fp-[1/p] of 50-90 Hz, which is within the
1-316Hz range of the testing frequencies of the present experiments and, in part,
motivated the range of testing frequencies used.
Previous measurements and poroelastic models of the dynamic stiffness of native
cartilage (98, 104, 110) have shown that I E* I increases withf'/"in the neighborhood off,.
IE* I in the entire
1-316Hz range of Figs. 4-2a,3a does, indeed, show an f
'
dependence
(R2>0.94, p<0.001; Figure B-2, Table B-2). This result, coupled with the value of fp
estimated above, suggests the importance of poroelastic behavior in the experimentally
tested frequency range (104, 106, 110). Interestingly, for native cartilage explants, H is
much larger and k much smaller than the cell-associated matrix, but the [Hk] product is
on the same order (10-9 m2/s). However, the much larger L-I mm for cartilage explants
results in frequencies fp that are several orders of magnitude lower for the dynamic
behavior of cartilage (0.001-0.01Hz) in confined or unconfined compression (78, 98,
111). For cartilage explants, it was possible to perform experiments at frequenciesf>>f,
where it was observed that
(f)would begin to decrease with frequency. It would
therefore be very useful to obtain dynamic stiffness data for the cell-associated matrix at
higher frequencies up to the 10 kHz range to more fully compare with theoretical
predictions (98). At present, however, AFM instrumentation limits data acquisition
beyond a maximum frequency of -350 Hz. The values of H and k estimated for our
matrix may, indeed, reflect the structure and organization observed in previous studies (8,
9) in which newly developing matrix revealed more diffusive and porous morphologies
than that of freshly isolated, mature chondrons (8).
Effect of culture duration. The increase in IE*(/) with time in culture is consistent
with our previously measured increase in biosynthesis and deposition of GAG and
collagen within the cell-associated matrix (4). The decrease in ( ) with culture duration
corresponds to a decrease in energy dissipation during dynamic compression (i.e., a more
elastic-like response), consistent with increased biomacromolecular deposition and
altered matrix structure/organization
(e.g., physical crosslinking, macromolecular
entanglements). k is also predicted to decrease with culture time (Appendix B.2),
indicative of a denser matrix structure. It is hypothesized that the modest increase in
IE*(f)l and 6(f) with culture time compared to the quasi-static indentation stiffness may be
due to the fact that the newly developing matrix is more sensitive to the higher
compaction associated with the I[tm indentation displacement compared to the much
smaller -5nm dynamic displacement amplitude.
Effect of GF treatment. The increase in IE*(/) and decrease in 6() with GF
treatment compared to FBS at higher frequencies (31.6-316 Hz) again suggests a more
elastic-like response likely due to compositional and structural changes, including
proteoglycan accumulation (9, 65, 112). The effect of GF treatment on IE*(f)l (Figure 4-
2a, 4-3a) was modest compared to the more dramatic increase in the indentation stiffness
(Figure 4-8), suggesting that the effect of GF on functional mechanical properties is
sensitive to the magnitude of the applied displacement.
In summary, our AFM-based approach has enabled the measurement of the
poroelastic dynamic mechanical behavior of the newly developing matrix associated with
individual chondrocytes, and can be applied generally to study the dynamic behavior of
extremely compliant (-kPa) biological, porous, and hydrated systems at nm-length scales
over a broad range of frequencies. The high resolution dynamic mechanical approach
described here is able to discern fine differences in the development and maturation of
the cell-associated matrix temporally and with the addition of growth factors. The
methodologies reported here can thus be employed to assess maturation of matrix
synthesized by primary chondrocytes and, additionally, by other cells such as stem cells
undergoing chondrogenesis in applications for cartilage tissue engineering. It may also be
possible to use this methodology to study the mechanobiological response of single
chondrocytes (113) and stem cells to applied dynamic loads over a wide frequency range
4.5 Conclusions
The success of cell-based tissue engineering approaches in restoring biological
function will be facilitated by a comprehensive fundamental knowledge base of the
temporal evolution of the structure and properties of the newly synthesized matrix. Here,
we quantify the dynamic oscillatory mechanical behavior of the engineered matrix
associated with individual chondrocytes cultured in vitro for up to 28 days in alginate
scaffolds. The magnitude of the complex modulus (IE*|) and phase shift (6) were
measured in culture medium using AFM-based nanoindentation in response to an
imposed oscillatory deformation (amplitude - 5nm) as a function of frequency (f = 1-316
Hz), probe tip geometry (2.5 jLm radius sphere and 50 nm radius square pyramid), and in
the absence and presence growth factors (GF, insulin growth factor-I, IGF-1, and
osteogenic protein-1, OP-1). |E*| for all conditions increased nonlinearly with frequency
dependence approximately f/2 and ranged between -1-25 kPa. This result, along with
theoretical calculations of the characteristic poroelastic relaxation frequency, fp, (-50-90
Hz) suggested that this time-dependent behavior was governed primarily by fluid flowdependent poroelasticity, rather than flow-independent viscoelastic processes associated
with the solid matrix. Fluid flow-dependent poroelasticity reflects the pressure gradients
and associated fluid outflow through the porous matrix, and subsequent recovery, during
dynamic oscillatory compression, which occurs on a short time scale at the single cell
level (-0.01 sec.). ]E*(f)| increased, 6(f) decreased, and the hydraulic permeability, k,
decreased with time in culture and with growth factor treatment. This trend of a more
elastic-like response was thought to be associated with increased macromolecular
biosynthesis, density, and a more mature matrix structure/organization.
Acknowledgements
Supported by NSF-NIRT 0403903, NSF-CMMI 0758651, NIH Grant AR33236,
and the MIT Institute for Soldier Nanotechnologies, US Army Research Office (contract
number DAAD-19-02-D0002). The content does not necessarily reflect the position of
the government and no official endorsement should be inferred.
Conflicts of interest
None of the authors have a conflict of interest associated with this work.
96
Chapter 5
Adhesion behavior of primary chondrocytes with their
newly developing PCM
5.1 Introduction
Chondrocytes and their surrounding pericellular matrix (PCM) are known to
exhibit a viscoelastic solid-like behavior under transient loading(16) or deformation (4, 5).
Viscoelastic material properties of chondrocytes with their PCM are likely to influence
the metabolic activity and gene expression of chondrocytes, also depending on the
physiological status of chondrocytes (e.g., normal vs. osteoarthritic chondrocytes). Hence,
the local microenvironment of chondrocytes is of critical importance to account for the
mechanics and physiology of chondrocytes which incorporate the mechanical, chemical,
and electrical fields in the local microenvironment. Previously, aggrecan was studied at
molecular level and showed a self-adhesion and shear deformation using AFM-based
nanoindentation (91, 114). Such information would provide a valuable insight on
understanding of the molecular mechanism of cartilage loading/deformation. So far,
mechanical environment of chondrocytes is largely examined using mature intact or
osteoarthritic cartilage (74). Newly developing PCM is a precursor to understand PCM of
mature chondron and PCM of osteoarthritic chondrocytes; mechanically, it has lower
modulus and higher permeability than mature intact PCM. Biochemically, newly
developing PCM of non-osteoarthritic chondrocytes has the biomechanical properties
different from PCM of osteoarthritic chondrocytes (7), suggesting alteration/degeneration
of PCM in osteoarthritic chondrocytes are not the same as newly developing PCM of
non-osteoarthritic chondrocytes. With this regard, newly developing PCM gives a
valuable opportunity to study the temporal evolution to mature intact PCM having fully
established microenvironment or to provide new insights on the remodeling mechanism
of the alteration of PCM with osteoarthritis (54).
Here, our study is aimed to investigate the possible source of energy dissipation
during the quasistatic indentation of the chondrocyte-matrix composites. For this purpose,
adhesion
behavior
of newly developing
PCM
was tested using AFM-based
nanoindentation to obtain F-d (Force-displacement) curve with varying the surface dwell
time. Bovine articular chondrocytes are cultured in two culture conditions: one in 10%
FBS (fetal bovine serum) medium and the other in growth factor (the combination of
insulin growth factor-I, IGF-1, and osteogenic protein-I, OP-1) supplemented medium.
OP-1 is known to stimulate gene expression and production of matrix component (e.g.,
HAS-2, CD44, aggrecan mRNA, proteoglycan and type II collagen) (115, 116). In fact,
the combination of IGF-I and OP-1 is more effective in cell proliferation and enhanced
matrix production than each growth factor used alone (100). Thus, working hypothesis in
98
this study is that mechanical properties (i.e., total dissipation energy, adhesion energy,
and the maximum adhesion force) are affected by the stimulant effect of growth factor
IGF- and OP- over time in culture.
5.2 Methods
5.2.1 Cell source and culture
Chondrocytes were obtained from femoral condyles of 1-2 week old bovine
calves and either used directly for day 0 biomechanical measurements or seeded into 2%
alginate scaffolds as described previously (4). Cell-seeded alginate beads were
maintained either in high-glucose Dulbecco's Modified Eagle Medium (DMEM) with
10% FBS (FBS culture) or in mini-ITS with the combination of IGF-1 (100 ng/mL) and
OP-1 (100 ng/mL) (GF culture) (100). On days 7, 14, 21, and 28, cells were released
from beads and suspended with high-glucose DMEM.
5.2.2 Adhesion measurement of single chondrocytes
Each chondrocyte was placed into an inverted pyramidal well of Si substrate
covered by liquid medium. Multimode Nanoscope IV AFM with PicoForce piezo (Veeco,
Santa Barbara, CA) was used to perform adhesion measurements of each cell within a
silicon well. Both pyramidal (end radius, Rtip-50 nm) and spherical colloidal probe tips
(Rtip -2.5 gm) were used to probe the same cells to investigate whether different length
scales might reveal different cellular response. During F-d (force-displacement) curve
measurement, the surface dwell time at a constant normal force (F - 2.0 nN) and z-piezo
displacement rate (1lm/s) was increased up to 180 seconds in order to capture the
adhesion properties of individual cell-PCM composites. Each loading-unloading curve
contains the single surface dwell time data using each probe tip. Measurement was
repeated 3 times with both probe tips and employed the discrete surface dwell time range
of 0, 15, 30, 60, 90, 120, and 180 sec.
5.2.3 Data analysis and statistical tests of adhesion measurement
Adhesion data was analyzed to obtain the adhesive interaction energy, Ead, the
total dissipation energy, Etot, and the maximum adhesion force, Fad (Figure 5-1). Since
individual chondrocytes exhibit the significant amount of hysteresis behavior during the
entire culture period, independent of the presence of PCM, as shown in Figure 5-1 as
well as in previous study of Ng. et al. (4), Ead is estimated from the area between loading
and unloading curve corresponding to negative force (i.e., attractive force exerted in
adhesion between the cell-PCM composite and the probe tip). Similarly, Etot is the sum of
the area between loading and unloading curve including Ead and hysteresis. By contrast,
little hysteresis behavior was observed in the adhesion of aggrecan at a physiologic-like
packing density (114).
100
For statistical test, effects of surface dwell time and culture duration on Fad, Ead,
and Ead/Etot in each culture medium for each probe tip were evaluated using two-way
ANOVA. At the same culture duration, effects of growth factor treatment and tip
geometry on Fad, Ead, and Ead/Etot were compared using two-way ANOVA. Additionally,
effects of growth factor treatment and tip geometry on Fad, Ead, and Ead/Etot were
compared using two-way ANOVA. Tukey post hoc test was performed when the
significant effects were observed (p<0.05).
Day 28 chondrocytes
3
t=60 sec
--
"2
Hysteresis
S1
Etotal
O
LL.-O
0 bow--\E
F
-1
-4
-3
-2
-1
0
1
2
3
Displacement (pm)
Figure 5-1 Adhesion behavior of day 28 chondrocyte with its PCM
101
5.3 Results
5.3.1 Adhesion behavior of single chondrocytes and their newly developing matrices
The overall adhesion behavior of single chondrocytes or their newly developing
matrices tends to increase the maximum adhesive force, Fad, the adhesion energy, Ead, and
the total dissipation energy, Etot, over time and then exhibit similar Fad, Ead, and Etot
between 90 and 180 sec. The adhesion behavior is approximately similar once the surface
dwell time is longer than the threshold time (- 90 sec). This trend was observed both on
single chondrocytes (Figure 5-2) and on newly developing PCM (Figure 5-3, 5-4, 5-5, 56). Single chondrocytes (day 0) generally exhibit higher Fad and Ead, particularly at longer
surface dwell time (t = 90 - 180 sec) in nanosized probe tip measurements, as compared
to micron-sized spherical tip measurements (Figure 5-7 (a, c) vs. 5-8 (a, c); the same
results of day 0 were plotted in the data for GF culture). Similarly, the difference in Fad
and Ead of newly developing matrices (day 7-28) using both probe tips continuously
maintained over time in culture. Interestingly, Fad and Ead of newly developing matrices
were more dramatically changed by GF treatment; the growth factor treatment resulted in
lower Fad and Ead, as compared to 10% FBS culture via micron-sized spherical probe tips
(Figure 5-3 vs. 5-4 for day 14; Figure 5-5 vs. 5-5 for day 28); the findings of a single F-d
curve were further supported by the entire data (Figure 5-7). Similar results of Fad for
newly developing matrices were obtained via nanosized tips, whereas the effect of
growth factor on Ead was modest via nanosized tips over 28 days (Figure 5-8). Longer
culture duration with growth factor treatment resulted in lower Ead/Etot in micron-sized
102
spherical tip measurements (Figure 5-8 (b, d) vs. Figure 5-8 (a, c)). However, nanosized
tip measurements displayed the ratio (i.e., Ead/Etot) increasing with the surface dwell time,
suggesting that the adhesion becomes more predominant than the hysteresis for the
contribution to the total dissipation energy (Figure 5-9 (c, d)).
103
Day 0 Single Chondrocyte (no PCM)
(b),
(a)
z
C:
Z
z
C
t=15 sec
jL
a)
0
U.
L
0
U.
-4
-2
0
-4
2
Displacement (pm)
(c)z
Z
)
2
t=60 sec
I
0
U.
0
U-
L.
-4
-2
0
2
Displacement (pjm)
-4
-2
0
2
Displacement (jim)
(e),
Z
0
Displacement (pm)
(d)-
t=30 sec
-2
(f)
t=90 sec
Z
t=120 sec
C)
j
0
U.
U-
0L
UL
-4
-2
0
2
Displacement (pm)
-4
-2
0
2
Displacement (pm)
C
)
t=180 sec
S-I
0
U.
-4
-2
0
2
Displacement (pm)
Figure 5-2 Adhesion behavior of single chondrocyte (day 0)
104
Day 14 Single Chondrocyte with newly developing PCM
10% FBS
2
(b)
(a)
z
t=15 sec
0
C
c,
LO
0
U.
U.
-2
-4
(c)
0
0
2
-4Displacement2
Displacement (gm)
2
Displacement (pm)
2
(d)
2
t=60 sec
t=30 sec
1
1.
z
0
----- -- - - - - - 0
U-1
U
1. -1
0_
-2
2
0
-4 -2
2
0
-4 -2
Displacement (m)
Displacement (jpm)
.Oft
(e)
3
z
r. 2
1
0
L 0
U-1
(g)
ZC4
(f)
t=120 sec
0
U
u
2
0
-4 -2
Displacement (pm)
2
0
-2
-4
Displacement (ipm)
t=180 sec
-- -
- - - - - - - - - - -
0
U
L
2
0
-2
-4
Displacement (tpm)
Figure 5-3 Adhesion behavior of day 14 chondrocyte with its PCM from FBS culture
105
Day 14 Single Chondrocyte with newly developing PCM
IGF-I+OP-1
(a).
(b),
t=0 sec
t=15 sec
z
L.
LM
(c)
£LM
0
U
-4
-2
0
2
Displacement (
3
t=30 sec
2
z
U
U-
Z4z
0
U-
-4
-2
0
2
Displacement (jim)
-4 -2
0
2
Displacement (lm)
3
1
0
(g)
t=120 sec
U.
-1
-4
-2
0
2
Displacement (tm)
3
t=180 sec
(9 %%o
2
O
t=60 sec
...
3
t=90 sec
2
-1
-4 -2
0
2
Displacement (m)
dm)
(d)
1
0
-1
(e)
.
U.
-4 -2
0
2
Displacement (m)
1
LM
O 0
U.
-1
-4
-2
0
2
Displacement (gm)
Figure 5-4 Adhesion behavior of day 14 chondrocyte with its PCM from GF culture
106
Day 28 Single Chondrocyte with newly developing PCM
10% FBS
(a).
(b) , 3
t=15 sec
sec
t=0
z
c.
oO
0
U.
U.
2
0
-4 -2
Displacement (dm)
(c)
1
t=30 sec
2
0
-4 -2
Displacement (pm)
(d)
£
t=60 sec
0t
C
z
C
U.
0
U)
O
L..
0
L.L
LM
0
U-
2
0
-4 -2
Displacement (pm)
(e)
Z
2
0
-4 -2
Displacement (pm)
(f) .
z
t=90 sec
3
t=120 sec
2
01
1.
LL
.)
2
0
-4 -2
Displacement (pm)
2
0
-4 -2
Displacement (pm)
(g)
Z
£
t=180 sec.
0
L.
0
U.
-4
-2
0
2
Displacement (pm)
Figure 5-5 Adhesion behavior of day 28 chondrocyte with its PCM from FBS culture
107
Day 28 Single Chondrocyte with newly developing PCM
IGF-1 +OP-1
(a)
Z
t=0 sec
(b).
Z
1.
%No
C.
0
O
0
U.
UL
-4
-2
0
2
Displacement (jim)
(c)
t=30 sec
-4 -2
0
2
Displacement (lm)
(d)
z
.
3
2 t=60 sec
uO0
U.
U.
-4
-2
0
2
Displacement (ljm)
(e)
z
t=15 sec
t=90 sec
-4 -2
0
2
Displacement (ljm)
(f)
0
C)
(U,
%%o
0
U-
0
L.
-4 -2
0
2
Displacement (gm)
-4
-2
0
2
Displacement (lm)
(g)
Z
t=180 sec
U..
C)
Ln
0
U. -1
-4
-2
0
2
Displacement (pm)
Figure 5-6 Adhesion behavior of day 28 chondrocyte with its PCM from GF culture
108
5.3.2 Adhesion behavior of single chondrocytes (day 0)
Fad,
the maximum adhesive force, of single chondrocytes (day 0) was significantly
changed with tip geometry and surface dwell time (2-way ANOVA, p<0.001). For a
micron-sized spherical probe tip, longer surface dwell time (t = 15 - 180 sec.) obtained
similar Fad significantly different to Fad at t = 0 sec. (Figure 5-7(a,b)). By comparison, a
nanosized probe tip demonstrated a gradual increase of Fad from t = 0 to 120 sec. of the
surface dwell time and then similar Fad (- 0.48 nN), independent of the surface dwell time
(Figure 5-8(a,b)). In addition, a nanosized tip displayed significantly higher
Fad
(- 0.91
nN) for t= 120 - 180 sec. as compared to the micron-sized tip measurement.
For Ead, the adhesion energy, of single chondrocytes, the effects of tip geometry
and the surface dwell time were both significant (two-way ANOVA, p<0.05). Similar to
the observation on Fad, Ead of micron-sized tip measurements is similar for t > 0 sec.,
except for t = 120 sec. (Figure 5-7(c,d)), whereas Ead of nanosized tip measurements is
gradually increased up to t = 120 sec. and approximately constant afterwards (Figure 58(c,d)). A nanosized tip obtained higher Ead than a micron-sized tip during the entire
surface dwell time (t = 0-180 sec.; p<0.05).
The ratio of Ead to Etot of single chondrocytes (Ead/Etot; the total dissipation energy
includes both contribution from hysteresis and adhesion) was also significantly changed
with tip geometry and surface dwell time (2-way ANOVA, p<0.005). Micron-sized tips
revealed similar contribution of
Ead
to
Etotal
(- 60%; Figure 5-9(a,b)). On the other hand,
nanosized tips exhibited the ratio increasing with the surface dwell time (Figure 5-9 (c,d)).
109
- -
Day 0.--Day 7 - Day 14.-.Day21 -w Day 28
(a) 1.5
z
4no/
(b)
Il
eDoI o
IGF-1+OP-1
1.1
1.01
, 1.0
IZ
0.5
u. 0.5
0
60
120
Time (sec)
60
120
Time (sec)
180
180
(d)
(c) 2.5
2.0
W
1.5
A
1.0
0.5.
60
120
Time (sec)
0
180
60
120
Time (sec)
180
Figure 5-7 The maximum adhesion force, Fad, and the adhesion energy, Ead, of single
chondrocytes (day 0) and their newly developing matrices (days 7-28) using (a,c) FBS
culture (10% FBS) and (b,d) GF culture (IGF-I+OP-1) probed by a micron-sized
spherical tip (mean±SEM, n=5, the number of cell measured)
5.3.3 Adhesion behavior of the cell-matrix composite (days 7-28)
5.3.3.1 The effect of surface dwell time on adhesion behavior of cell-matrix
composite (days 7-28)
Fad, Ead, Etot, and the ratio of Ead to Etot (Ead/Etot) were generally minimum for the
quasistatic instantaneous loading (i.e., t = 0 sec.) and stayed higher at t > 0 sec. in FBS-
110
and GF- supplemented culture probed by both micron-sized spherical probe tip and
nanosized pyramidal probe tip at each time point. Indeed, the effect of surface dwell time
on Fad, Ead, Etot, and Ead/Etot was significant for both FBS and GF culture medium using
both probe tips (Figures 5-7, 5-8, 5-9; one-way ANOVA, p<0.05). However, longer
surface dwell time did not affect significantly Fad, Ead, Etot, and Ead/Etot at t = 15-180 sec.
5.3.3.2 The effect of tip geometry on the adhesion behavior of cell-matrix composite
(days 7-28)
The effect of tip geometry on Fad, Ead, and Ead/Etot was significant in both FBS
and GF culture (Figures 5-7 vs. 5-8 for Fad and Ead; Figure 5-9 (a,b) vs. (c,d) for Ead/Etot;
one-way ANOVA (p<0.05)). However, the response was more distinctively different,
depending on the culture medium. In both FBS and GF culture, micron-sized tips
revealed lower Fad and Ead as compared to nanosized tip measurements (Figures 5-7 vs. 58; p<0.05). This might reflect the different biosynthetic activity between FBS and GF
culture over 28 days (4); GF culture resulted in higher GAG and lower collagen than FBS
culture. Generally, the overall architecture of newly developing cell-associated matrices
is largely based on collagen matrices and the rest space is filled with GAG molecules (1).
Therefore, nanosized tips might probe more sensitively the collagen matrices of GF
culture that are likely less compact as compared to those of FBS culture based on the total
accumulation of collagen over 28 days (4), than micron-sized tips. In FBS culture,
micron-sized tips might better detect the overall matrices than nanosized tips due to
111
higher collagen proportion leading to less porous structure. Similarly, Ead/Etot in both FBS
and GF culture demonstrated the tip geometry dependence (Figure 5-9 (a,b) vs. (c,d)).
5.3.3.3 The effect of culture duration on the adhesion behavior of newly developing
matrix (day 7-28)
Fad, Ead,
and Ead/Etot were significantly affected by culture duration in FBS and
GF culture medium in measurements using both probe tips (Figures 5-7, 5-8, 5-9; oneway ANOVA (p<0.05)). In FBS culture, Fad,
Ead,
and Ead/Etot exhibited decreased from
day 7 to day 14 and then maintained the level comparable to day 7 (Figures 5-7(a,c), 58(a,c), 5-9(a,c)). On the other hand, GF culture showed a decrease of Fad and Ead/Etot with
culture duration (Figures 5-7(b), 5-8(b), 5-9(b,d)). Interestingly, the effect of culture
duration on Ead was distinctive to each tip geometry in GF culture; A micron-sized tip
probed
Ead
decreasing with culture duration (Figure 5-7(d)), while a nanosized tip
revealed Ead with modest change over time in culture (Figure 5-8(d)).
5.3.3.4 The effect of growth factor treatment on the adhesion behavior of newly
developing matrix (day 7-28)
Fad, Ead/Etot decreased significantly over time in GF culture, as compared to FBS
culture using both probe tips (one-way ANOVA (p<0.05); Figures 5-7(b vs. a) and 5-9(b
vs. a) for a micron-sized tip, Figures 5-8(b vs. a) and 5-9 (d vs. c) for a nanosized tip).
The effect of GF treatment on
Fad,
and Ead/Etot was enhanced over time in culture as well
112
for both probe tips. However, the effect of GF treatment was more evident by a steady
decrease of Ead in measurements using a micron-sized tip (Figure 5-7 (d vs. c));
nonetheless, Ead of GF culture was significantly different from that of FBS culture using
both probe tips (ANOVA, p<0.05). Although Ead of GF culture was comparable to that of
FBS culture in nanosized tip measurement, Ead/Etot was significantly lower for GF culture
than for FBS culture over 28 days since hysteresis was also increased with culture
duration in GF culture (Figure 5-9 (b,d) vs. (a,c)).
-4. Day 0 -a-Day 7
Day 14 -4- Day 21 -
Day 28
(b) 1.5
(a) 1
IGI
-
-1.0
LL 0.5
0
(c) 2
(d) 2.5
2,
2.0
-
120
60
Time (sec
IGI
w
Figure 5-8 The maximum adhesion force, Fad, and the adhesion energy, Ead, of single
chondrocytes (day 0) and their newly developing matrices (days 7-28) using (a,c) FBS
culture (10% FBS) and (b,d) GF culture (IGF-I+OP-1) probed via nanosized pyramidal
tip (mean+SEM, n=5, the number of cell measured)
113
-4- Day 0 -a-Day 7
(a)
1
Day 14 -- Day 21 -*- Day 28
(b)
()
10% FBS
1
IGF-1+OP-1
0 0.8
0 0.8
0.6-
0.6
-
o 0.2
0.2
0
0
(C)
60
120
Time (sec)
i
0
(d)
10% FBS
o0.8
10.4
180
120
Time (sec)
1
180
IGF-1+OP-1
0.8
T
a
60
00.4 -
0.2
o
o .
0
60
120
Time (sec)
180
0
60
120
Time (sec)
180
Figure 5-9 The temporal change in the ratio of Ead to Etotal (% of Etotal dissipation) for the
cell-matrix composite over 28 days with surface dwell time (mean±SEM, n=5, the
number of cell measured) using (a,c) FBS culture (10% FBS) and (b,d) GF culture (IGFI+OP-1) probed by micron-sized tip and nanosized tip.
5.4 Discussion
Our study successfully demonstrated the temporal evolution of Fad, Ead, and
Ead/Etot with the surface dwell time over 28 days in different culture medium
supplementation. Our current study investigated the quasistatic equilibrium adhesion
behavior over 180 seconds.
114
Adhesion behavior ofDay 0 chondrocytes
Day 0 chondrocytes showed that Fad
and Ead are minimal at t=0 sec, and afterwards increase with the surface dwell time > 0
sec. (2-way ANOVA, p<0.05), while being constant after 120 sec. for both probe tips.
Interestingly, a nanosized tip is likely to penetrate the local area rather than to probe, as
evidenced by higher Fad and Ead than the micron-sized tip measurement. The ratio of Ead
to Etot (Ead/Etotal) probed by a nanosized tip is also higher than that probed by a micron-
sized tip, consistent with these findings.
Adhesion behavior of the cell-matrix composite For the cell-matrix composites
(day 7-28), similar trends of Fad, Ead, and Ead/Etot were observed with both probe tips
during culture duration, while the surface dwell time (> 0 sec.) did not significantly affect
Fad, Ead, and Ead/Etot. The interaction distance of nonlinear attractive force with jagged
(i.e., saw-tooth like) contours on the unloading curve of cultured chondrocytes with their
newly developing matrix increased over time in FBS culture, generally longer than that of
day 0 chondrocytes. However, the interaction distance was largely dependent on the
medium supplementation; GF supplementation led to a shorter range of the interaction
distance and lower Fad and Ead with increasing culture duration. As a result, GF
supplementation resulted in lower Ead/Etot than FBS culture, in particular using micronsized tips. Our findings can be attributable to the fact that the biosynthetic activity varies
with the medium supplementation. While the amount of collagen was more synthesized
and/or retained in FBS culture after day 14, the level of GAG accumulation maintained
much higher for GF supplementation after day 14 (4). Indeed, F-d curve of the cellmatrix composite indicated more enhanced adhesion in FBS culture (Figure 5-3, 5-5). By
115
contrast, the adhesion was sharply reduced in GF supplementation by day 28 (Figure 5-4,
5-6). Taken together, GF supplementation might lead to more compact matrix structure
via the enhanced synthesis and/or retention of GAG having lower
Fad, Ead,
and Ead/Etot.
Meanwhile, the effect of GF supplementation on the temporal evolution of
Ead
depends on the probe tip used. This is partly due to our estimation of Ead from F-d curve;
it should be noted that our separation of
Ead
from Etot is discretionary, since the area
between loading and unloading curve includes the mixed contribution of hysteresis and
adhesion. Since Fad is higher for nanosized tip measurements, Ead might be overestimated
and this prevents from the exact estimation of GF treatment effect in nanosized tip
measurements.
Contribution of hysteresis and adhesion to energy dissipation In comparison to
our study at single cell level, the previous study of Han, et al is noteworthy, where it
showed that aggrecan self-adhesion can be a possible molecular mechanism to support
both the tensile loading and the energy dissipation simultaneously (114); in this study,
AFM probe tips were coated with aggrecan and probed the single aggrecan layer
absorbed onto the gold-coated silicon substrate. Opposing layers of aggrecan at a
physiologic-like packing density showed higher Fad (-3nN) and
Ead
(-0.8fJ) than single
aggrecan layer with neutral hydroxyl-terminated self-assembled monolayer (OH-SAM)
colloidal probe tip, while both displayed similar trend of Fad and Ead with the surface
dwell time. The interaction between single aggrecan layer and the OH-SAM tip led to
similar Fad and lower Ead. In addition, the maximum adhesive interaction distance using
the OH-SAM tip was -410 nm, comparable to the contour length of fetal epiphyseal
116
aggrecan (Lc -398 nm) (117), both of which are much shorter than our measurement (up
to 3 Jpm; Figure 5-2, 5-3, 5-4, 5-5, 5-6). Since the total amount of GAG and collagen for
chondrocytes in early culture is much less than that of native cartilage (118), along with
more diffuse PCM morphology (4), the packing density of aggrecan in our study might
be suboptimal to the physiological level. Even in the lower packing density, higher Ead Of
the cell-matrix composite might be attributable to the contribution of collagen to add
more configurational/conformational complexities to the molecular interaction (114).
This hypothesis is further supported by the fact that the equilibrium time to reach plateau
of Fad and Ead was - 30-60 sec., much longer than that for aggrecan (-5 sec.) (114).
Overall, our study examined the adhesion behavior of single chondrocytes with
their newly developing matrix over time in culture using either FBS or GF
supplementation. At single cell level, Fad and Ead gradually changed with culture duration
and dramatically reduced in GF culture. Simultaneously, hysteresis increased with culture
duration in GF supplementation. Along with the enhanced biosynthetic activity of GAG
in GF supplementation, lower Fad and Ead in GF supplementation further supports the
efficacy of GF treatment on the matrix formation.
5.5 Conclusions
The adhesion behavior of chondrocytes with their newly developing matrix
revealed that maximal adhesive force and the adhesive energy are significantly reduced
over time and with GF treatment for both micron-sized and nanosized tips. During
117
measurement, a nanosized tip is more likely to penetrate the cell-matrix composite,
leading to a distinct profile of the adhesion behavior with surface dwell time. Nonetheless,
the total dissipation mechanism was largely attributable to hysteresis (-80%), rather than
adhesion behavior at later time in GF supplemented culture probed by both tips.
118
Chapter 6
Mechanical behavior of individual bovine marrowderived stromal cells undergoing chondrogenesis and
their newly developing cell-associated matrix
6.1 Introduction
Stem cell-based tissue engineering holds great potential for the regeneration
and/or replacement of damaged cartilage (119). Indeed, chondrogenesis of bone marrow
stromal cells (BMSCs), specifically mesenchymal stem cells (MSCs), was observed to
occur within 1 week of culture and demonstrated the capability of MSCs-derived
chondrocytes to produce cartilage-like extracellular matrix (ECM), based on biochemical
and gene expression analyses (46-48). ECM of articular cartilage is of critical importance
in the load-bearing function during joint motion, and thereby loss and/or degradation of
ECM is implicated with osteoarthritis and leads to the failure of the functional integrity of
articular cartilage. With this regard, BMSCs are sought as an alternative cell source for
119
cartilage tissue repair (119); MSCs can be easily obtained with low donor side morbidity
of the harvesting procedure, well expanded in vitro in an undifferentiated state, as well as
differentiated into both osteogenic and chondrogenic cell lines under specified conditions
(48). In particular, the latter is a beneficial effect of the use of BMSCs on the repair of
osteochondral defect where the restoration of both the subchondral bone (i.e.,
endochondral ossification) and articular surface (i.e., chondrogenesis) is needed (48, 119).
Furthermore, another study showed that the pretreatment of MSCs with transforming
growth factor pl (TGF-13) and the maintenance of a three-dimensional culture
environment prevent transdifferentiation of MSCs from a chondrogenic to an osteogenic
lineage, which further supports the use of BMSCs in cartilage tissue engineering (46).
Stem cells are obtained from various source (e.g., adipose (47, 49, 50), bone marrow (4648, 50, 120), and amniotic fluid (51)) and cultured with tissue-engineered constructs (e.g.,
agarose (49, 50, 120), alginate (46-49, 51), self-assembling peptide gels(50)) in order to
understand the chondrogenic differentiation as well as to find the optimal condition in the
potential use of stem cells for cartilage tissue engineering. So far, stem-cell based studies
of cartilage tissue engineering have examined the characterization of cartilage-specific
gene expression (46-48, 50), the synthesis of ECM components of cartilage(49, 50, 120),
and immunohistochemical evaluation for type II collagen (46-50, 120), along with a
relatively few functional studies of MSCs-based tissue engineered constructs at
macroscopic level (49, 120). Chondrogenic culture medium using TGF-3 superfamily
resulted in a dramatic increase of the gene expression and biosynthesis of ECM
macromolecules, as compared to the absence of growth factor (47, 49-51) and showed a
dose-dependent increase (46).
120
These studies revealed the temporal evolution of MSCs during chondrogenesis in
a hydrogel construct with macroscopic biochemical properties; for example, the cell-tocell heterogeneity and the regional variations in matrix accumulation are evidenced by
histological features of MSCs-based tissue engineered constructs (120). However, the
functional aspects at single cell level are not yet fully addressed. Indeed, mechanical
behavior at single cell level was distinctively different between undifferentiated and
differentiated cells, implying that the intracellular organization (e.g., cytoskeleton) is
significantly altered during differentiation (52). Elastic and viscoelastic properties of
primary chondrocytes were tested at single cell level to obtain lower elastic and
relaxation modulus, as compared to those of BMSCs (53). These features might reflect
the distinctive cytoskeletal structure for each cell type that is qualitatively described in
other studies of the cell topography using AFM-based imaging techniques (121, 122). To
connect the biomechanical link between undifferentiated and differentiated cells, the
temporal changes in the mechanical behavior of MSCs during chondrogenesis need to be
investigated, which can be informative to understand the cellular mechanism underlying
chondrogenesis from the local measurement of microenvironment at the nano/microscale
level rather than the macroscopic mechanical testing of cell-scaffold constructs.
Furthermore, the biomechanical knowledge of individual MSCs cultured in vitro during
chondrogenic differentiation can provide valuable insights on the composition-structurefunction relations during the development of cartilage-like neo-tissue.
Our objective was to test the hypothesis that BMSCs undergoing chondrogenesis
could synthesize a cell-associated matrix (i.e., pericellular matrix (PCM)) de novo having
nanomechanical properties and biochemical composition comparable to those of primary
121
chondrocytes in culture. For nanomechanical testing, we measured the temporal evolution
of the quasi-static and time-dependent viscoelastic mechanical properties of single bovine
BMSCs and their associated newly developing cell-associated matrix released from 3-D
alginate scaffolds during chondrogenesis over 28 days using chondrogenic culture
medium (50). In addition, a dynamic mechanical single cell protocol was applied to probe
the frequency-dependent mechanical behavior of the newly developing matrix (123). To
relate these changes to cellular synthesis of ECM macromolecules, dimethyl methylene
blue (DMMB) dye binding and hydroxyproline assays were performed to quantify
glycosaminoglycan and total collagen accumulation within the alginate scaffolds over
time, respectively.
6.2 Methods
6.2.1 Cell source and culture
Bovine bone marrow stromal cells (BMSCs) were obtained and expanded in two
passages (P2) as previously prepared (50). BMSCs (P2) were either used directly for day
0 biomechanical measurements or seeded into 2% alginate scaffolds as described
previously (4). Cell-seeded alginate beads were maintained in chondrogenic culture
medium of high-glucose Dulbecco's Modified Eagle Medium (DMEM) supplemented
with 1% ITS+ (Insulin-Transferrin-Selenium; Sigma-Aldrich, St. Louis, MO), 100nM
dexamethasone (Sigma-Aldrich, St. Louis, MO) and 10ng/mL recombinant human
transforming growth factor 31 (TGF-31; R&D Systems, Minneapolis, MN) (50, 120).
122
Culture medium was changed every other day. On days 3, 7, 14, 21, and 28, cells with
their associated matrices were released from beads and suspended with high-glucose
DMEM for biomechanical measurement.
6.2.2 Histology of newly developing PCM
At each time point, cells with their associated matrices released from alginate
beads were resuspended in culture medium (-106 cells/mL) and fixed using SafefixllI
(Fisher Scientific). The fixed cells were mounted onto a glass slide using Cytospin (CytoTek Miles, Diagnostic Division, Mishawaka, IN) by spinning cells at 1400 rpm for 10
min (4). After air dried, each slide of mounted cells was stained either with Toluidine
Blue (Toluidine BlueO, Sigma-Aldrich) for proteoglycans or with phosphomolybdic acid
followed by aniline blue (Rowley Biochemical Inc.) for collagen (124). Once the slides
were dried, one drop of Toluene (Permount, Fischer Scientific) was dripped onto the cell
area of each slide and a coverslip was placed over the drop. Each slide stained for sGAG
and collagen was used to estimate the newly developing cell-associated matrix thickness
of cells by day 21 and 29 using optical micrographs of cells.
6.2.3 Biochemical assays of newly developing PCM
Dimethyl methylene blue dye binding (DMMB) (125) and hydroxyproline (OHProline) contents (126) were used as measures of sGAG and total collagen of the PCM of
123
released cells at each selected time point. These data were normalized to DNA content
measured with Hoechst 33258 dye solution at the same time point (127).
6.2.4 Mechanical testing of MSCs during chondrogenesis
Freshly isolated cell or each cell with varying culture period (day 3-28) was
placed into an inverted pyramidal well of Si substrate with liquid medium by spherical
colloidal probe tip (Silica beads from Bang Labs, #SS06N, tipless cantilever from Veeco,
NP-020, spring constant, K- 0.06N/m, Rtip -2.5
tm) (Figure 6-1(a)). A multimode
Nanoscope IV AFM with PicoForce piezo (Veeco, Santa Barbara, CA) was used to
perform dynamic oscillatory compression measurements on each cell (day 0-3) or each
cell-matrix composite (day 7-28) within its silicon well. Script mode (Nanoscope IV
software, V613rl) was programmed to apply a 1 plm ramp indentation at a displacement
rate of I tm/s. Force relaxation at constant displacement of - 1tm was then measured
during the subsequent 180-second interval (Figure 6-1(b)) . After the force reached quasiequilibrium within -100
seconds, a sinusoidal displacement (5 nm amplitude) was
superimposed on the 1l m static offset displacement using an external wave generator
(Rockland 5100) connected to the AFM piezoelectric scanner. The magnitude and phase
angle of the resulting force were measured at discrete frequencies of 1, 10, 100, and 316
Hz (Figure 6-1(c)). Measurements were repeated 3 times at the same location on each
cell-matrix composite.
124
6.2.5 Data analysis and statistical tests
First, elastic modulus, Ey, was evaluated using the Hertzian contact mechanical
model from initial loading data with the indentation depth of-1 ptm (102). An indentation
depth up to -300nm was used to satisfy the model assumption of a small strain region
(E<0.05) (4).
Second, time-dependent mechanical properties were estimated from the relaxation
data using a 5-element Maxwell-Weichert model (84) to yield the instantaneous modulus
Eo (at t=0 sec.), the quasi-equilibrium modulus Eo (at t = l00sec.), the initial relaxation
time constant Ti, and the long-term relaxation time constant r2.
Third, dynamic loading data was analyzed via Fast Fourier Transform of
measured force and deformation by fitting the measured force and displacement to a
sinusoidal function to obtain the magnitude of complex modulus, IE*(f) 1, and phase
angle, 6 (), between force and deformation at each frequency. 6(f) was corrected for
hydrodynamic drag (101). For the micron-sized colloidal probe tip, I E*(j)
was obtained
from serial Taylor expansion of Hertzian contact model for spherical colloid probe tip
(Eq. 6-1) (42, 102):
E
1 - v2 F(f)
spherical (f ) =d(f)
(6-
(6-1)
where do is indentation depth at zero-frequency, Poisson's ratio v is assumed to be 0.4
(37), and R is the effective size of cantilever (i.e., 1/R
= [1/Rspherical tip] +[l/Rchondrocytes])
(4). F(f) is the magnitude of the force at frequency f and d(f) is the magnitude of the
deformation at frequency f during dynamic oscillatory compression. Since our study
125
mainly focuses on the assessment of the relative trends in the temporal evolution of
matrix stiffness with culture time, Hertzian contact model is sufficient to estimate cellular
mechanical behavior.
For statistical test, effects of frequency and culture duration on either magnitude
of complex modulus or phase angle were evaluated using one-way ANOVA. For elastic
and transient mechanical properties, the effect of culture duration was tested using oneway ANOVA. The Tukey post hoc test was performed when the significant effects were
observed (p<0.05). To determine the correlation between mechanical properties and
biochemical composition, multiple linear regressions were performed. All data were
reported as mean+SEM, with p < 0.05 as statistically significant. Statistical analyses
were performed using Systat-12.0 (Richmond, CA).
126
osclllatory
displacement
AFM cantilever
-5 nm amplltud
- 258 pmun
Sloid,
indentation - 1 pm
cell,RIc, - 4L8 pm
1
matrix, -3.5 pm thick
are pyramidal well
(b)
amic
-
(c)
0.04 -ri
o0
z
,-0.04%00
120.1
Tlme(sec)
50
100
Time(sec)
Figure 6-1 Dynamic oscillatory compression of individual BMSCs and their associated
matrix using a spherical tip: (a) Experimental setup: BMSC-derived chondrocyte-matrix
composite in a microfabricated silicon well (radius of cell = Rcell, radius of probe tip =
Rtip, matrix thickness = 3.5 pm, indentation depth = l tm, and small sinusoidal oscillation
= 5nm; cell radius and matrix thickness are based on day 21-28) (b) Force (red) and
indentation depth (blue) curve showing the initial indentation (depth, do = 1~tm) and
subsequent force relaxation exemplified by a day 21 BMSC, and (c) the applied dynamic
sinusoidal deformation (d(f)) and resulting dynamic force (F(f)) shifted from the
displacement by the phase angle 6() enlarged from the dotted square areas in (b)
6.3 Results
6.3.1 Biochemical composition of the newly developing cell-associated matrix and its
appearance of BMSCs-derived chondrocytes over 28 days
The total GAG content was initially - 0 at day 0 and -0.3 jig/jpg DNA at day 3
(Figure 6-2(a)). A negligible amount of GAG was largely due to a few GAG synthesis
127
normalized by high DNA content from the cell population seeded in alginate hydrogel (10 million cells/ ImL alginate solution). For day 7-14, the rate of increase in the total
GAG content was comparable to that of cell population estimated from DNA content,
leading to 2.7-2.9 !ag/!tg DNA. The total GAG content was then increased gradually up
to 4 tg/ tg DNA at day 21 and later reached 8 pg/ Lg DNA at day 28.
Similarly, the total hydroxylproline content was detected minimally at day 0 and
slightly higher than the total GAG amount by day 3 (Figure 6-2(a)). The total amount of
hydroxyproline changed with culture duration in a linear trend from 1.3
g/Ltg DNA at
day 3 to 5.0 jlg/ltg DNA at day 28. For day 3-21, the total hydroxyproline content
maintained the amount comparable to the total GAG amount on the same day. However,
the total GAG content showed - 2-fold increase at day 28, and the total hydroxyproline
content changed - 1.3-fold increase at day 28.
Optical microscopy image of BMSCs stained for collagen and PG revealed a
diffusive morphology at day 21 and an improved staining density for both at day 28
(Figure 6-2(b)). Increased intensity of staining for GAG and collagen at later time point
(day 29) was consistent with the increased synthesis and retention of GAG and collagen
at those later times (Figure 6-2(a)).
Cell diameter and the thickness of the cell-associated matrix were -9.2 plm and
-3.6 pm at day 21, respectively (Table 6-1). On day 28, cell diameter was slightly
increased to -10.2 jlm, whereas the thickness of the cell-associated matrix was somewhat
decreased to -3.3 jm. This change might be attributable to the increased cell count
(-30% increase) at day 28, predominated by the heterogeneity of cell population varying
its size and thickness even on the same culture condition.
128
(a) 12 4-GAG
6
-9-hydroxyprolin
l8
(b)
4
M
S3
Collagen PG
Day 21
4
2 5
23um
0
7
14
21
Day
28
3
Z
Collagen PG
Day 29
Figure 6-2 Biochemical characterization of the newly developing matrix of BMSCsderived chondrocytes over 28 days that were released from alginate culture at designated
time points: (a) Total GAG content (measured by DMMB) and collagen content
(measured by hydroxyproline), (b) optical microscopy (10x) images of individual
BMSCs stained for collagen and proteoglycan (PG)
Table 6-1 Characterization of cell diameter (Dce,,) and the newly developing matrix
thickness of BMSCs-derived chondrocytes estimated from optical micrographs (n=
number of cells, mean±SEM), compared to those of primary chondrocytes reported by
Ng, et al.(4)
Dcel (m)
PCM (im)
Day 21 (n=155)
Day 29 (n=202)
Primary Chondrocytes
9.22±0.19
10.19±0.14
7.65+0.85 (day 0)
3.64±0.12
3.29±0.14
3-4 (days 7-28)
6.3.2 Elastic and transient mechanical properties of the cell-matrix composites of
BMSCs-derived chondrocytes
Ey significantly increased with culture duration (one-way ANOVA, p<0.05; Fig.
6-2(a)). A significant change in Ey (Fig. 6-3(a)), coincident with GAG accumulation by
129
MSCs during chondrogenesis (Fig. 6-2(a)), was first detectable on day 7 as compared to
day 0 (p<0.05; day 0 and day 3 are not significant). A gradual increase of Ey, followed by
a sharp decrease on day 7 was similar to what observed in the previous study using
primary chondrocytes (4). Interestingly, day 28 showed Ey. significantly different from
that of day 7-14 (p<0.05), and comparable to that of day 0-3 (p>0.5).
Similarly, the effect of culture duration was significant on E0, E., and t2 (Fig. 63(b,c); one-way ANOVA, p<0.05). Day 0 exhibited significantly lower Eo and E., as
compared to longer culture duration (p<0.05). Eo of day 21 shows a decrease,
significantly different from that of day 7 and day 28 (p<0.05). E 0 of day 28 displays an
significant increase from early culture duration (day 0-21; Fig. 6-3(b)). Day 7 and 28
showed a significant increase in 12 from day 0 (p<0.05; Fig. 6-3(c)). However, the effect
of the culture duration was not significant on ti; they are almost constant (- 1.2 sec.) over
28 days (Fig. 6-3(c)).
Our findings revealed that Ey and Eo are significantly correlated with the
biosynthesis (Figure 6-4, Table 6-2); Ey of days 7-28 was significantly correlated with
the amount of hydroxyproline (p<0.001), and Em of days 7-28 was correlated with the
increase in GAG, hydroxyproline, and the combination of GAG and hydroxyproline
(p<0.05).
130
(a) i.
0#
A
I. I
-.
(n~.1
I r1
-pCo.o5)
..
37t
D(b).
(b) 1.5u
14
21
is
W.
00.51
ur
A
(n.1O-14"*pC&O.
•
Day0
MYa
3 ;7
14
21
2
Figure 6-3 Elastic and time-dependent properties of the cell-matrix composite of
BMSCs-derived chondrocytes over 28 days that were released from alginate culture at
designated time points: (a) Ey from the quasistatic indentation, (b) Eo and Eoo, (c) T1 and
T2 from force relaxation (n=number of cells, mean±SEM, *vs. other days, **vs. day 7-14,
+ vs. day
0)
131
(a) 0.6
(b) 1.0
CL
~v0.3
w >-
8
u0
(c)
.............
'0.5
2
4 D
9g OH-Prolinel jig DNA
2
4
6 8
1g GAG/ lpg DNA
(d) 1 .0
A
1.0
a. 0.5'
0 .5
8
8
u0
I
-
0
2
4
6
JLgOH-Prolinel p.g DNA
0
10
-
20
30
50
40
jig GAG x jig OH-Proline/ (gg DNA) 2
Figure 6-4 Linear regressions of biomechanical properties vs. biochemical composition
of BMSCs-derived chondrocytes over 28 days: (a) Ey vs. hydroxyproline, (b) E, vs.
GAG, (c) E. vs. hydroxyproline, and (d) E. vs. the combinational interaction of GAG
and hydroxyproline) (Parameters for each regression are shown in Table 6-2)
Table 6-2 Correlation between mechanical properties and biochemical composition of
BMSCs-derived chondrocytes over 28 days using multiple linear regressions
Slope
Intercept
R2
p
EY
Pro(Day7-28)
0.055
0.234
0.998
0.001
Eo,
GAG
0.045
0.508
0.791
0.018
GAG(Day7-28)
0.035
0.565
0.941
0.030
Pro
0.074
0.450
0.794
0.017
GAG*Pro
0.008
0.549
0.665
0.048
GAG*Pro(Day7-28)
0.005
0.623
0.912
0.045
132
6.3.3 Dynamic properties of newly developing PCM
Effect of frequency Both IE*()I and 6(1) showed a nonlinear increase with
frequency (one-way ANOVA, p<0.05). IE*()
of 100 and 316 Hz are significantly
different from that of I and 10 Hz over the entire culture duration (p<0.01; Figure 6-5(a)).
6(f) indicated a significant difference among all the frequency range measured at each
culture duration (p<0.05; Figure 6-5(b)).
Effect of culture duration Both E*(/)| and 6() were significantly affected by
culture duration as well (one-way ANOVA, p<0.05). IE*(f)| of day 28 fell below that of
other culture duration, being significantly lower than day 7 over all the applied frequency
range (p<0.05; Figure 6-5(a)). On the other hand, 6() did not change much over time in
culture except for day 21 (Figure 6-5(b)); 6(f) of day 21 at 100 Hz was comparable to that
of day 0, both of which are significantly higher than that of other culture duration
(p<0.05). 6(f) of day 7-28 was significantly lower than that of day 0 at the frequency of 1100 Hz (p<0.05).
133
-4- Day 0
4Day 3 --
Day 7 -- Day 14 -*- Day 21 -x- Day 28
(a) 15 (n=10-13; *p<0.005)
(b)
0 (n=10-13; *p<0.005)
C1
0
0
S
1
10
100
1
o
1io
Frequency (Hz)
Frequency (Hz)
Figure 6-5 The cell-matrix composite of BMSCs-derived chondrocytes over 28 days that
were released from alginate culture at designated time points: (a) E*() and (b) .(f)
(n=number of cells, mean±SEM; *vs. lower frequencies on the same time points)
6.3.4 Estimation of hydraulic permeability
Previous study showed that the time-dependence of cellular mechanical behavior
is largely dominated by heterogeneous pressure gradient within the cell, reflecting the
fluid flow-dependent poroelastic processes (128). Given more diffuse morphology of
cell-associated matrix of MSCs, as compared to primary chondrocytes (9), rate-dependent
poro-viscoelastic processes are expected (76). Indeed,
I E*(f) I showed a //2
-dependence
at the applied frequency range (1-316 Hz) (p<0.05; Figure 6-6, Table 6-3) as observed in
native cartilage (98, 110). To further explore the poroelastic frequency-dependence of
cell-matrix composite, the characteristic poroelastic time constant, Tp
-
[L2 /(Hk)], was
calculated (92) from the estimation of hydraulic permeability, k, using the unit cell model
(108, 109) assuming that GAGs are the primary determinant of the permeability (129).
The hydraulic permeability was calculated based on GAG spacing, b, (130) from the
134
experimentally measured GAG concentration, CPG (Figure 6-2(a); the detailed procedure
was shown in the previous publication (123)). For each concentration of GAG during
days 7-28 (Table 6-4), k of newly developing matrix of MSC-derived chondrocytes
4
decreased with culture duration from 13.74X 10-12 m /N s at day 7 to from 4.03
10-12
m4/N-s at day 28, being up to 3 times higher than that of PCM from primary
chondrocytes (123) due to lower GAG concentration of BMSCs. Therefore, ,Tp of BMSCderived chondrocyte-matrix composite yielded - 0.008-0.011 sec. for day 21-28,
corresponding to 90-126 Hz (Table 6-4), and this in part motivated the range of testing
frequencies in the present experiments (1-316 Hz).
14- Day
a-Day7
-- Day 14 -- Day 21 -M-Day 28
5
0
10
20
VFrequmWHz)
Figure 6-6 Plot IE) I vs. /'2 for the cell-matrix composite of BMSCs-derived
chondrocytes over 28 days using the micron-sized spherical probe tip (mean+SEM; the
number of cells, n= 1 0-13)
135
Table 6-3 Linear regression correlation between I E*) I and f/2 (Figure 6-6) of the cellmatrix composite of BMSCs-derived chondrocytes over 28 days using micron-sized
spherical probe tip
Tip geometry
Micron-sized
spherical tip
Day
0
3
Slope
0.530
0.370
R2
0.956
0.990
p
0.022
0.005
7
0.365
0.997
0.001
14
0.398
0.983
0.008
21
0.486
0.923
0.039
28
0.331
0.981
0.009
Table 6-4 Characterization of BMSCs-derived chondrocytes for GAG spacing b,
hydraulic permeability k, and poroelastic time constant tp, based on the total GAG
amount synthesized during culture duration, along with the elastic modulus H,
determined by FEA simulation
BMSCs-derived chondrocytes over 28 days
Day
GAG (jg/ 0P Cells)
b (nm)
k (x10-12m'/N-s)
7
20.8
95.37
13.74
14
22.8
91.09
12.40
21
31.2
77.87
28
62.3
55.11
H (kPa)
, (sec)
f, (Hz)
8.75
0.09
0.0079
126
4.03
0.14
0.0111
90.3
6.4 Discussion
Our study demonstrated that the functional properties of MSCs undergoing
chondrogenesis at single cell level, along with the quantification of ECM synthesis, vary
with culture duration and the time/frequency domain. The elastic Young's modulus of our
study, Ey, was -0.6kPa at day 0 and dramatically reduced to -0.3kPa at day 7. the
136
temporal change in Ey with culture duration showed a similar trend to that of primary
chondrocytes where the quasistatic stiffness of day 7 chondrocytes is much lower than
day 0 and gradually increased by day 28 (4). Therefore, it is hypothesized that day 7 in
the current study appears to detect the initial appearance of newly developing matrix.
Meanwhile, the recent study of Darling, et al. reported the elastic modulus of human
BMSCs to be - 2.5 kPa (53), much higher than that of bovine BMSC (- 0.6 kPa) in our
study. The large discrepancy of Ey might be attributable to the difference in species,
substrate, and testing configuration. Nonetheless, viscoelastic properties of human
BMSCs in the same study of Darling, et al. (53) have similar Eo (- 0.5 kPa), much
higher Eo (- 5 kPa), and similar c (- 10 sec., the longest viscoelastic relaxation time
constant) compared with those of BMSCs at day 0 (Figure 6-3), implying different
biomechanical components/loading mechanism might take part in the viscoelastic
behavior. Moreover, day 28 of BMSC-derived chondrocytes in our current study
indicated lower Ey (- 0.5 kPa), higher Eo and Eo (-1.2 and 0.8 kPa, respectively), and
much longer T (- 32 sec.) (Figure 6-3), as compared to elastic and viscoelastic properties
of human primary chondrocytes (53). Again, disagreement of mechanical properties
between bovine BMSCs-derived chondrocytes and human primary chondrocytes can be
attributable to the difference in species (bovine vs. human) and/or phenotype of cells
tested (partially differentiated chondrocytes vs. fully differentiated chondrocytes). In our
study, the effect of culture duration is significant on Ey, EoO, Eo, and rl (p<0.05). But
only Ey and E. are significantly correlated with hydroxyproline (for both) and GAG (for
E. only, Table 6-2). Indeed, the trend of the total GAG and collagen contents with
culture duration is analogous to the change of Ey and Eo with culture duration (Figure 6-
137
2(a) vs. 6-3), while the temporal profile of Eo and cl showed a considerable fluctuation
along the time course in culture (Figure 6-3).
The total GAG and collagen contents on day 28 (Figure 6-2(a)) were slightly
higher than those of adipose-derived stem cells cultured in alginate under the similar
conditions (49), but much less than that observed using bovine MSCs in self-assembling
peptide scaffolds (up to 7-fold difference) (131) and that for primary chondrocytes (up to
2.6-fold difference) (4). The implication of these finding are as follows: first, the
difference of the biosynthesis in alginate and peptide hydrogel may reflect the material
environment provided by a tissue engineering scaffold. Cell-aggregation followed by
morphological changes occurred extensively in peptide hydrogel (50), whereas a
spherical morphology persisted throughout the entire culture duration in alginate scaffold
(49). While the seaweed-derived alginate is typically thought to be lack of native ligands
that possibly interact with cells (132), peptide scaffold may provide more favorable
binding site for cell attachment (95). Second, another significant difference still exists in
the tissue repair potential of MSCs and chondrocytes even for the same animal and
scaffold (i.e., bovine BMSCs seeded in alginate scaffold vs. bovine primary chondrocytes
seeded in alginate scaffold) (4, 120, 123). For example, PCM thickness of BMSC-derived
chondrocytes on day 21-29 was comparable to that of primary chondrocytes on day 7-28
(4) (Table 6-1), while less staining intensity was observed for BMSCs (Figure 6-2(b)).
IE*() I of MSCs-derived
chondrocytes was significantly different from that of
primary chondrocytes in 10% FBS culture at 316 Hz for the entire culture duration (day
7-28; one-way ANOVA, p<0.05), while primary chondrocytes cultured with insulin-like
growth factor-I (IGF-1) and osteogenic protein-i (OP-1) treatment showed a significant
138
difference at 1-100 Hz on day 7, at 1, 10, 316 Hz on day 14, and at 100-316 Hz on day
21-28 (one-way ANOVA, p<0.05) (123). The former was clearly demonstrated by -1.53-times higher I E*(f)
at 316 Hz for primary chondrocytes as compared to MSCs-
derived chondrocytes ((123) vs. Figure 6-5(a)) and further supported by the lower total
amount of GAG and collagen of MSCs-derived chondrocytes ((4) vs. Figure 6-2(a)). The
latter was consistent with the difference in the total GAG and collagen amount between
MSCs-derived chondrocytes and primary chondrocytes with the combination of IGF-I
and OP-I treatment as well (4). In fact, the biochemical characterization of primary
chondrocytes resulted in higher GAG amount for GF culture (i.e., the combination of
IGF-1 and OP-1) and higher collagen amount for FBS culture during the entire culture
duration (day 7-28, (4)). Therefore, the difference in ECM synthesis between FBS and
GF culture also influences the distinct difference in
IE*()
I
between primary
chondrocytes and BMSCs. Likewise, IE*() I was higher for primary chondrocytesseeded construct than for BMSCs-seeded constructs by day 32 (120). In addition, the
pericellular environment of BMSCs is composed of constituents, different from than of
primary chondrocytes (120). Nonetheless, similar frequency dependence of dynamic
mechanical behavior was observed for BMSCs at the applied frequency range (1-316 Hz)
at single cell level. By contrast, macroscopic mechanical testing of stem cell-seeded
hydrogel constructs demonstrated little frequency dependence of dynamic shear (49) and
dynamic unconfined compression (120) at the applied frequency range, which suggests
that the macroscopic tissue property might not fully represent the individual cell-matrix
development varying degree of differentiation/maturation.
139
6() of BMSCs-derived chondrocytes was also compared to that of primary
chondrocytes cultured in either FBS culture or GF culture (123); a nonlinear increase of
6() with frequency was more enhanced for BMSCs-derived chondrocytes as compared to
primary chondrocytes cultured in FBS culture. As a result, 6(f) of BMSCs-derived
chondrocytes was higher at 100-316 Hz and lower at 1Hz on day 7-28 than that of
primary chondrocytes in FBS culture (123) (one-way ANOVA, p<0.05). While GF
culture of primary chondrocytes resulted in a decrease of 6() at the minimum (1 Hz) and
maximum (316 Hz) applied frequency in comparison to FBS culture of primary
chondrocytes (123), 6() of MSCs-derived chondrocytes was comparable to that of
primary chondrocytes in GF culture at 1Hz and yet significantly higher at 316 Hz on day
7-28 (one-way ANOVA, p<0.05). While 68() varied considerably with frequency at single
cell level, 6() of stem cell-seeded construct showed a minimal change with a frequency
(49), possibly dominated by the material property of scaffold itself rather than the
contribution of cell-matrix composite. Taken together, the comparison of dynamic
behavior between MSCs-derived chondrocytes and primary chondrocytes suggests that
mechanical behavior of newly developing matrix reflect the unique structural
organization/composition that are influenced by culture medium supplementation and cell
source as observed in the total GAG and collagen amount over time in culture.
Our estimation of -t (92, 104-107) is based on the characteristic length scale, L,
over which fluid flows are approximated as the radius of probe tip (2.51pm), the hydraulic
permeability of the cell-associated matrix, k, from the measured GAG content (Figure 62(a)), and the quasi-static indentation modulus, H, from finite element analysis (FEA)
simulation of the cell-matrix composite (4) (Table 6-4). ,pof the newly developing
140
matrix of BMSCs indicated the [Hk] product, comparable to that of native cartilage (- 109
m 2/s). However, much smaller L of the cell-matrix composite at single cell level leads to
the characteristic poroelastic relaxation frequency,fp, of 90-126 Hz, several orders higher
than that of native cartilage (0.001-0.01 Hz) (78, 98, 111). Again, dynamic stiffness data
for the cell-associated matrix at higher frequencies up to the 10 kHz might be needed to
compare with theoretical predictions (76). The assessment of H and k suggests more
diffusive and porous morphologies of the newborn matrix obtained from BMSCs,
consistent with the lower total GAG content of BMSCs and less staining intensity of
collagen and PG (Figure 6-2), as compared to PCM of primary chondrocytes (4, 9) and
mature chondron (8). In essential, the theoretical consideration of fp proposes that the
fluid flow-dependent poroelasticity is a governing mechanism in the early stages of the
cell-newborn matrix composite, as suggested by the study of primary chondrocytes
cultured over 28 days (123). Accordingly, the possible dissipation source for 6(/) during
dynamic deformation of the newborn matrix is most likely related to the fluid flow and
local hydrostatic pressure within the matrix.
Overall, our study of BMSCs that undergoes chondrogenic differentiation
demonstrated the temporal changes of the cell-matrix composite in elastic and
time/frequency dependent mechanical behavior with culture duration at single cell level,
implying that fundamental differences exist between MSCs-derived chondrocytes and
primary chondrocytes regarding cell-matrix interaction and phenotype-dependent
development/maturation of newborn matrix (e.g., PCM/ECM) (120). Furthermore, it
should be noted that the scaffold material significantly affects the ECM synthesis
141
followed by chondrogenic differentiation of MSCs (49, 120, 131), also being of critical
importance in the success of the cartilage tissue-engineered constructs.
6.5 Conclusions
Stem cell-based tissue engineering holds great potential for the regeneration
and/or replacement of damaged cartilage. Mechanical studies of individual mesenchymal
stem cells (MSCs) cultured in vitro can provide valuable insights into changes in
intracellular morphologies that accompany differentiation to the chondrocyte phenotype,
as well as the synthesis of the cartilage-like neo-tissue during chondrogenesis. Here, we
probe the temporal evolution of the single cell mechanical properties of bovine MSCs
cultured within 3-D alginate scaffolds and stimulated to undergo chondrogenesis using
dexamethasone and TGF-31 over 28 days. Quasistatic indentation, force relaxation, and
dynamic oscillatory compression were applied to each cell-matrix composite being
placed in a microfabricated silicon well using a spherical colloidal probe tip (end radius2.5 jim) in an atomic force microscope in DMEM culture medium. Biochemical assays
indicated that MSCs in alginate scaffolds synthesized and accumulated GAG and
collagen in amounts comparable to MSCs in agarose and yet much lower than MSCs in
self-assembling peptide scaffolds as well as primary chondrocytes in alginate. The
temporal evolution of the elastic modulus (Ey) showed a sharp decrease on day 7 and a
gradual increase with culture duration, indicative of the formation of newborn matrix.
Time-dependent mechanical properties (Eo, E-,, -2) of day 28 also showed a distinctive
change as compared to day 0. Indeed, Ey and Eo are significantly correlated with the
142
total GAG and collagen amounts in culture. IE*(f)I showed a significant, yet significant,
change with culture duration and frequency, being a sharp contrast to that for of primary
chondrocytes. On the other hand, 6(f
was almost constant on day 3-28, while
significantly different from day 0, and comparable to that for primary chondrocytes.
Taken together, mechanical properties for MSCs undergoing chondrogenesis was
distinctive to those of primary chondrocytes in early culture study of 28 days, suggesting
unique differences in intracellular organization and/or PCM that have not reached the
final chondrocyte-like state.
Acknowledgements
Supported by NSF-NIRT 0403903, NSF-CMMI 0758651, NIH Grant AR33236,
and the MIT Institute for Soldier Nanotechnologies, US Army Research Office (contract
number DAAD-19-02-D0002). The content does not necessarily reflect the position of
the government and no official endorsement should be inferred.
Conflicts of interest
None of the authors have a conflict of interest associated with this work.
143
144
Chapter 7
Concluding Remarks
Our study of newly developing PCM using AFM-based nanoindentation first
focused on the evolution of PCM mechanical properties at the single cell level with time
in culture. The newly developing PCM demonstrated a gradual increase in instantaneous
and equilibrium moduli with time in culture using both nano and micron-sized probe tips.
The viscoelastic time constants of the newly developing PCM showed deformation ratedependence. During stress relaxation tests with an applied deformation amplitude of Ijtm,
the effects of tip geometry and culture duration on the cell-matrix mechanical behavior
showed good agreement with previous studies of quasi-static indentation that used the
same deformation amplitude and rate.
Motivated by the dynamic load/deformation behavior of cartilage in vivo, our
experiments then explored the frequency dependence of PCM mechanical properties at
single cell level. The applied frequency range was chosen to bracket the estimated
poroelastic relaxation time constant, based on the characteristic length for fluid flow
underneath the probe tip, as well as the values of the matrix stiffness and hydraulic
permeability. Indeed, the dynamic complex modulus increased with applied frequency, in
a manner strongly suggestive of poroelastic behavior of the cell-associated matrix.
Furthermore, the comparison between native cartilage explants and individual cell-matrix
145
composites confirms the length-scale dependence of the poroelastic time constant (or,
equivalently, the characteristic poroelastic frequency).
Culture duration and growth factor treatment both resulted in an increase of IE*()
I
and a decrease of 6(f). The effect of growth factor treatment on dynamic oscillatory
stiffness was rather modest, while there was a much more pronounced effect on quasistatic indentation stiffness. The latter finding suggests that cell-matrix behavior is
strongly dependent on the deformation amplitude. The effects of both tip geometry and
growth factor treatment on quasi-static indentation stiffness (using I m deformation
amplitude) were significant. On the other hand, the significant effect of growth factor
treatment on dynamic compressive stiffness (using 5nm deformation amplitude) was
significant, though modest.
We then studied the nanomechanical behavior of the cell-associated matrix of
bovine bone marrow stromal cells (BMSCs) during chondrogenesis and with continued
time in culture. After chondrogenic differentiation, these BMSCs exhibited biosynthetic
activity and mechanical behavior that were qualitatively similar to those of primary
chondrocytes. The total accumulation of GAG and collagen using bovine BMSCs
appeared to lag that of primary chondrocytes by 1-2 weeks, suggestive of a more diffuse,
porous matrix. The structural organization of the newly developing matrix of BMSCs via
biochemical characterization was further supported by the mechanical properties of the
cell-matrix composite, which were lower for |E*()I and higher for 6(f) than those of
primary chondrocytes.
Overall, our study of newly developing PCM explored the local mechanical
environment of cell-PCM that undergoes the very initial stage of development. We found
146
that the time-dependent mechanical behavior of the cell-matrix composite at the single
cell level is closely related to the biosynthesis of the major structural macromolecules and
their subsequent organization, which extends previous studies of quasi-static mechanical
behavior. The/
2
-dependence of dynamic properties over the applied frequency range of
interest further suggests that this time-dependent behavior of the newly developing PCM
is governed primarily by fluid flow-dependent poroelasticity, rather than flowindependent viscoelastic processes associated with the solid matrix. While the quasistatic stiffness increased with time in culture and even further with the growth factor
treatment, the dynamic complex modulus changed rather modestly with culture duration
and growth factor stimulation. This finding emphasizes the displacement amplitude
dependence of cell-matrix behavior, and requires more research using a range of
deformation amplitudes to achieve a mechanistic understanding of this finding.
Furthermore, it would be informative to test the cell-matrix dynamic behavior at different
length scales using various probe tip sizes that would provide data over a broader
frequency range and enable scaling to a normalized frequency. Such information would
serve to improve the model prediction of the local mechanical environment of cell-PCM,
and thereby facilitate the success of cell-based tissue engineering approaches in restoring
biological function with a comprehensive, fundamental knowledge base of the temporal
evolution of the structure and properties of the newly synthesized matrix.
147
148
Appendix A
Glossary
AFM atomic force microscopy
BMSC bone marrow stromal cell
DMEM Dulbecco's Modified Eagle Medium
DMMB dimethyl methylene blue
ECM extracellular matrix
FBS fetal bovine serum
GAG glycosaminoglycan
IGF-1 insulin-like growth factor-I
ITS Insulin-Transferrin-Selenium
MSC mesenchymal stem cell
OP-1 osteogenic proteinPCM pericellular matrix
TGF-pl transforming growth factor 31
149
150
Appendix B
Characteristic time constants for time-dependent
behavior of tissue-engineered cell-associated matrix
B.1 Viscoelastic model
Experimental data for the longer-term transient relaxation behavior (>5 sec.) was
well described by viscoelastic model. For example, the relaxation data of the longer time
scale (>5 sec.) of day 14 chondrocytes PCM in Figure SI is successfully described by 3element viscoelastic model with a coefficient of linear regression R2>0.9). However, the
initial fast decaying was not captured by simple viscoelastic model with a few time
constants such as 3- and 5- element model (Figure B-l(b))(5, 84). It is clear that 3element model can get the better capture of the initial fast decaying as it fits the shorter
time scale from 0.3 to 0.03 sec of the initial relaxation data, not the entire relaxation
(Figure B-l(b), the enlarged plot of Figure B-l(a) shown up to 0,01 sec). To further
explore the time constant for the initial transient response, the experimental data was cut
off about the first 0.03, 0.05, and 0.5 second, respectively, and then fitted with the
151
exponential decay function (i.e., Force=Aexp(-time/X)+B, where A and B is fitting
parameter (unit of nN) and k is the characteristic time constant(sec.)). As shown in Table
B-1, the characteristic time constant for each initial transient response was increased as
longer response was included for model prediction. By comparison, poroelastic time
constant for newly developing PCM was calculated using the values from previous
studies to yield -0.01 seconds, much shorter than viscoelastic time constant (2, 4).
(a) 6.1
0.02 0.04 0.06
Time (sec)
5i
Time (sec)
Figure B-i Comparison of initial transient behavior for day 14 chondrocytes with matrix
between experimental data and viscoelastic model prediction: (a) experimental data with
model prediction of 3- and 5- element model (EM) (single time constant for 3-EM and 2
different time constants for 5-EM, respectively), and model prediction of the
experimental data up to 0.03, 0.05, 0.5 seconds (labeled as 0.03s, 0.05s, and 0.3s,
respectively) with 3-element model (b) enlarged plot of Figure B-l(a) up to 0.1 sec. (For
experimental data, a limited number of data points are indicated than the actual number
of raw data points; all data sets are only shown up to either 10 or 0.1 seconds to highlight
the deviation of individual model prediction from experimental data.)
152
Table B-1 Characteristic time constant for transient relaxation behavior of day 14 cellPCM composite from experimental data and model prediction: for k, the exponential
function of force (Force= A exp(-time/X)+B, where A and B is fitting parameter (unit of
nN) and k is the characteristic time constant(sec.) is used; 3-EM is 3-element model with
single time constant and 5-EM is 5-element model with two time constants of z 1(the
initial time constant) and t 2(the longer relaxation constant)
Viscoelastic model of Poroelastic
Characteristic time constant I
(sec.), using exponential function
the entire force
Model
relaxation (~100s)
for the initial force relaxation
Day
-63%
reduction
Data fits up to
0.03s 0.05s 0.3s
5-EM
3-EM
"
(sec.) (sec.)
Vl
(sec.)
T2
(sec.)
14
3.92
0.500
0.629
2.58
8.18
0.766
12.0
0.016
28
4.22
0.676
0.874
2.45
3.95
0.801
14.0
0.031
B.2 Poroelastic model
IE*I in the entire
1-316Hz range of Figures 4-2a, 4-3a, 4-5a, 4-6a does, indeed,
show an f ' dependence (Figure B-2, Table B-2; R2>0.94, p<0.001). As described in the
main text (See 4.4. Discussion), we estimated the characteristic poroelastic relaxation
time, z,p [L2/(Hk)] to understand the mechanisms underlying the frequency deoendence
of the measured dynamic stiffness of the cell-associated matrix. The geometric
configuration comprised of the cell-associated matrix sandwiched in-between the
impermeable spherical probe tip and the cell membrane (Figure 4-1(a)) is qualitatively
approximated by an uniaxial unconfined compression geometry; the characteristic length
scale, L, over which fluid flows is then approximated by the radius of the probe tip, 2.5
jLm. The equilibrium modulus H can be approximated by the quasi-static indentation
153
stiffness (-0.12 kPa for day 14 and 0.17 kPa for day 28 in FBS culture (4). To estimate
the hydraulic permeability, k, we hypothesized that matrix GAG chains are the most
critical determinant of the permeability of the newly developing matrix, based on
previous measurements and models for the permeability of native cartilage (129). We
therefore used the unit cell model to calculate k (108, 109); the unit cell is comprised of a
central charged GAG rod of radius, a, and its surrounding concentric cylindrical fluid
shell having outer radius, b (i.e., b is the average center-to-center GAG spacing within the
matrix (109)):
k
-3
8b2
(4b
4
Inb 3b 4 +4a2b 2
a
4
2
2
34)
Inb
a
b4
2b
4
-a
4
(B)
where q is the fluid viscosity (10- 3 Pa-s). Outer radius b was independently calculated to
be -40-70 nm using the Poisson-Boltzmann cell model (130) along with the measured
total GAG content of the cell-associated matrix (CGAG = 40-130 tg/mL for days 7-28,
(4)):
b = 2TBN CGAG
MCS
(B2)
where B is the distance between charge groups along the GAG chain (0.64nm), N is the
Avogadro's number (6.02 x 1023), and M, is the molecular weight of a dissociated
chondroitin sulfate disaccharide (458g/mol) (130). With the estimated value of b,
equation BI yields values of permeability, k, of the newly developing matrix to be
154
1.8-6.0x10 - 12 m4/N-s (Table B-3). This range of values for k is much higher than that of
the mature chondron (10'- 7 m 4 /N's (16)), suggesting that the newly developed cell-
associated matrix is much less resistant to the fluid flow. Based on the above estimates
for L, H, and k, we calculated values for t, to be in the range 0.011 - 0.021 sec for day
14-28 cells in FBS culture (Table B-3). This corresponds to characteristic frequencies
fr[ 1/,] of 50-90 Hz (Table B-3).
(a) 30
10% FBS
-.- Day 7 (n=5)
-- Day 14 (n=5)
a. 20' -0- Day 21 (n=4)
-K- Day 28 (n=5)
W
10.
0
10u
Frequency(Hz)1/ 2
10
Frequency(Hz)
Frequency(Hz) 1/2
Frequency(Hz) 1/2
2
(C) 30
o 20
1
Figure B-2 Plot I E*(f) I vs./2 for newly developing PCM during 1-month culture: (a)
from FBS medium (10% FBS) and (b) from GF medium (IGF- l +OP- 1) using the micronsized spherical probe tip, and (c) from FBS medium (10% FBS) and (d) from GF medium
(IGF-I+OP-1) using the nanosized pyramidal probe tip (n=number of cells, mean±SEM)
155
Table B-2 Linear regression correlation between I E*() I and /f2(Figure B-2) of the
cell-matrix composite in both FBS and GF culture using both micron-sized spherical
probe tip and nanosized pyramidal probe tip
FBS culture
Tip geometry
Day
Micron-sized
spherical tip
Nanosized
pyramidal tip
GF culture
2
0
7
Slope
R
0.635 0.988
0.675 0.969
p
Tip geometry
0.000 Micron-sized
0.000 spherical tip
Day
0
7
Slope
0.738
0.573
R2
0.985
0.964
14
21
0.971
1.198
0.982
0.958
0.000
0.001
14
21
1.090
1.297
0.982 0.000
0.999 0.000
28
1.182 0.945
0.001
28
1.262
0.985 0.000
0
7
0.459 0.957 0.001
0.433 0.975 0.000
0
0.642
7
0.719
0.992 0.000
0.950 0.001
14
0.621
0.943 0.001
14
0.879
0.989 0.000
21
0.732 0.943 0.001
21
1.118
0.984 0.000
28
0.815 0.986 0.000
28
1.036
0.965 0.000
Nanosized
pyramidal tp
p
0.000
0.001
Table B-3 Estimation of the hydraulic permeability, k, based on the total GAG amount
synthesized over 28 days and the characteristic poroelastic time constant,
P[-l/f,],
using chondrocytes in FBS culture (b: GAG spacing, H: elastic modulus determined by
FEA simulation)
Primary chondrocytes in FBS culture
Day
GAG (Ig/10 Cells)
b (nm)
k (x 0-12 m'lNs)
H (kPa)
p (sec)
f, (Hz)
7
43.9
65.65
5.97
0.1
0.010
95.6
14
54.8
58.76
4.66
0.12
0.011
89.4
21
80.1
48.60
3.04
0.15
0.014
72.9
28
129.7
38.19
1.76
0.17
0.021
47.8
156
Appendix C
Time-dependent mechanical properties of single
chondrocytes
C.1 Introduction
Chondrocytes are the cells that are solely responsible for the synthesis, retention,
and maintenance of the extracellular matrix (ECM) in articular cartilage in order that it
may properly function as a load-bearing tissue. Previous studies have shown that the
mechanical stimuli (e.g., loading or deformation) are transmitted to chondrocytes and
significantly influence the metabolism of chondrocytes and the gene expression of nuclei
in chondrocytes, i.e. mechanotransduction, where the mechanism is largely unknown (77,
113, 133). Previously, quasistatic indentation of chondrocytes showed the deformation
rate-dependent hysteresis, suggesting the time dependence of mechanical properties (4).
Indeed, stress relaxation transients revealed significantly higher viscoelastic properties
for chondrocytes in superficial zone as compared to middle/deep zone using the standard
linear solid model (5). Chondrocytes are exposed to time-varying loads and deformations
157
during joint movement and undergo cyclic and impact loading that includes frequency
components as high as I kHz (45). However, little information exists on the
physiologically relevant dynamic oscillatory properties, and furthermore the expected
poro-viscoelasticity of single chondrocytes (76). Consequently, understanding of the
mechanical properties of chondrocytes would provide valuable insight on the biochemical
and biophysical mechanism of chondrocytes. In an attempt to examine the dynamic
mechanical properties of single chondrocytes, we applied small sinusoidal amplitude (-5
nm) to single chondrocytes with varying frequencies (f = 1-316 Hz) and quantified their
mechanical response to AFM-based dynamic compression.
C.2 Methods
Chondrocytes were obtained from femoral condyles of 1-2 week old bovine
calves via the enzymatic digestion as described previously (4). Suspended with highglucose Dulbecco's Modified Eagle Medium (DMEM), each chondrocyte was positioned
into an inverted pyramidal well of Si substrate (Fig. la). A multimode Nanoscope IV
AFM with PicoForce piezo (Veeco, Santa Barbara, CA) was then used to perform
compression measurements on each cell within its well with both pyramidal (end radius,
Rtip -50 nm) and spherical colloidal probe tips (Rtip -2.5 tm) for each cell to investigate
whether different length scales might reveal different PCM properties. Script mode
(Nanoscope IV software, V613rl) was first programmed to apply a l-pm ramp
indentation into each chondrocyte at 1- tm/s; force-relaxation at constant displacement
was then measured during the subsequent 180-second interval (Fig. lb). After force-
158
relaxation reached quasi-equilibrium within -100 seconds (5), a sinusoidal displacement
(5nm amplitude) was superimposed on the 1- tm static offset displacement using an
external wave generator (Rockland 5100) connected to the AFM piezo. The magnitude
and phase of the resulting force were measured at discrete frequencies of 1, 3.16, 10, 31.6,
100, and 316 Hz (Fig. Ic). Measurements were repeated 3 times at the same location of
each cell with both probe tips and 5 minutes are given for cell recovery between
successive measurements. Dynamic force and displacement data for both probe tips were
analyzed using Fast Fourier Transforms to obtain the magnitude IE*(f) and phase
angle 9() of the complex modulus, E*(f), measured at each frequency. 8(f) was directly
calculated from dynamic force and displacement data and also corrected for the
hydrodynamic drag (101). For each probe tip, IE*(f was obtained from a Taylor series
expansion of the Hertzian contact model (99, 102):
E*spherical(f)=-2V
Epyramidal(f )
2
F(f
I-v 2
F(f)
2rd0 tan a d( f )
where do is indentation depth at zero frequency, Poisson's ratio, v, was assumed to be
0.4 (103), R is the effective size of the cantilever (i.e., I/R = 1/Rspherical tip
±l/Rchondrocytes),
pyramidal tip opening angle a is 350, and F() is the magnitude of the force at frequencyf,
and d() is the amplitude of deformation at frequency f (4). The effects of frequency and
tip geometry on IE*(f) and (f) were evaluated using two-way ANOVA. A Tukey post
hoc test was performed when significant effects were observed (p<0.05).
159
(a)
(I ) 3
Si substrate
in culture medium
AFM
2.
lever
lll
: p Indentation(-ipm)l
(Cell(R
,- 3.8mn )
)
U.
compressio0
O
square pyramidal well
*f
C
--_
1
125.0
SO
100
Time(sec)
i
200
%..
1.005
1.17
(C)
a-
0.5
Force
r
relaxation
9------**1
,
Dynamic .0
Si
oscillatory
amlitud
*I111
.',
a o%
1 F IE*
.. oF II
1
9
1 6!
'g
125.1
125.2
Time(sec)
-
P
Figure C-1 Dynamic oscillatory compression of freshly isolated chondrocyte (day 0)
probed by spherical tip: (a) Experimental setup: Chondrocyte in a microfabricated silicon
well, (b) Force and Indentation depth curve (c) Enlarged plot of the square area in (b)
C.3 Results
The magnitude of the complex modulus I E*(f) and phase angle. (f) varied with
applied displacement frequency for each probe tip (ANOVA (p<0.05)) (Fig. 2). With the
micron-sized tip, IE*() Iof day 0 chondrocytes increased from -1.26 kPa at 1 Hz to
-13.01 kPa at 316 Hz, -20-40% higher as compared to nanosized tip. 6(f) varied from
160
-11.1 deg at 1 Hz to -67.9 deg at 316 Hz with the micron-sized tip (Fig. 2b,3b), again
10-40% higher as compared to nanosized tip. However, tip geometry effect was not
significant, possibly due to approximately the same contact area of each probe tip at 5nm
oscillatory compression once chondrocytes are fully relaxed at 1 p~m pre-indentation.
Pair-wise comparison of
E*(fi) and. ~f) for each probe tip showed that higher
frequencies (100 and 316 Hz) resulted in significantly higher IE*() Iand (f), and the
frequency effect on (f) was more clearly visible in micron-sized tip. The trend was
similar, but rather modest in nanosized tip. (Tukey-post hoc test followed by ANOVA,
p<0.05)
U'
II
(a
(b) 90
go
r4
110. (mean±SEM; n=10)
r5
30
0
1
Freq10 ncy
Frequency (Hz) -
10
10
100
1io
Frequency (Hz)
Figure C-2 Dynamic properties of single chondrocytes with both probe tips: (a) IE(f ,
(b.)5(f) (n=number of cells, mean±SEM)
161
(a) 15
*p<0.05
(b) 90
p*p<o.oos
(mean+SEM;Pn=10)
360
10
1
(c) 15
3
10
32 100 316
1
Frequency (Hz)
S
(d) 90
*p<o.os
3
10
32 100 316
Frequency (Hz)
]
*p<os
(mean+SEM;n=10)
30
0.
030
5
0
0
1
3
10 32 100 316
1
Frequency (Hz)
3
10 32 100 316
Frequency (Hz)
Figure C-3 Dynamic properties of single chondrocytes with both probe tips: (a,b)
micron-sized spherical tip, (c,d) nanosized pyramidal tip(n=number of cells, mean+SEM;
*,**vs. lower frequencies)
C.4 Discussion
In our study, the frequency dependence of mechanical behavior of single
chondrocytes was observed in the range of 1-316 Hz, being qualitatively similar to other
cell types (41, 99).
Our testing method enabled to apply the constant deformation
amplitude over the entire frequency range, which was shown to be frequency dependent
(41). Since dynamic oscillatory compression was followed by the stress relaxation after
reaching the plateau, dynamic properties here come from the cell at equilibrium, while
162
others are not well defined (41, 99). The experimental configuration of an inverted
pyramidal well minimized the cell-substrate adhesion and avoid the issue of thin areas
observed in other cell types (99). Mechanical response of chondrocytes at higher
frequency is distinctively different from that at lower frequencies for both probe tips,
suggesting that load-bearing components or mechanism might alter with frequency.
Meanwhile, different tip geometry did not reveal any significant change in dynamic
properties. This is a sharp contrast to the previous studies, where the elastic properties of
chondrocytes are dependent on the probe tip geometry (4). The disparity in the tip
geometry dependence is likely implicated with approximately the same contact area of
probe tips at a small amplitude deformation under dynamic oscillatory compression and
the equilibrium status of chondrocytes.
The viscoelastic properties of chondrocytes are attributable to flow-dependent
(fluid-solid interactions and fluid viscosity) and flow independent mechanisms
(viscoelasticity
of the cytoskeleton) (2). The theoretical
model study recently
demonstrated that the presence of semi-permeable membrane surrounding isolated
chondrocytes results in a very negligible transmembrane fluid outflow; as a result,
isolated chondrocytes behave as an incompressible solid under mechanical loading with
the duration of minutes or hours (134). Nonetheless, the hydrostatic pressure can be
heterogeneous inside the cell (128), implying the flow-dependent mechanism might still
be dominant inside the isolated chondrocytes as well. Indeed, the poroelastic behavior
was inferred from the observation of IE*(f
I linearly
proportional to /2 (R2 > 0.99 for
linear regression for both probe tips; data not shown) (110). Consequently, we
hypothesized that dissipation mechanism of isolated chondrocytes are contributed by the
163
solid viscoelasticity of cytoskeleton as well as by the poroelasticity of solid-fluid
interaction inside the cell.
To account for the observed dynamic response of chondrocytes, the viscoelastic
model initially was used to fit the force relaxation part of the experimental data; using 3element viscoelastic model (5), the viscoelastic time constant, tz is - 9.64 sec. (R2 > 0.93;
data not shown). The viscoelastic time constant varies with the type of applied
mechanical force and the experimental configuration (-,- 31 sec. using micropipette
aspiration (14), -c - 5 sec. AFM-based nanoindentation (5). Furthermore, cytoskeletal
disruption resulted in lower viscoelastic modulus, lower apparent viscosity and higher
viscoelastic time constant of chondrocytes via micropipette aspiration (87).
By comparison, the poroelastic model obtained the poroelastic relaxation time, tu,
- 0.023 sec. (,= L /Hk (92), where L is the sample thickness (7.65 pm for the diameter
of chondrocytes diameter (4)), H is the equilibrium elastic modulus (lkPa for freshly
isolated chondrocytes (4)), and k is the hydraulic permeability (2.6x10 -12 m4/N-s for
chondrocytes (33)). The correspondingfp [-l1/p] (i.e., an inflection frequency of E*() I
and a peak frequency of 6()) is
-44 Hz, while none of
E*(f) and 8(f) in our
measurements exhibited such behavior. It is noted that the estimation of f, is largely
dependent on the value of hydraulic permeability (134). For example, the permeability on
the same order of magnitude as that of the ECM (e.g., 4.2x10 -" m4/N's for chondrocytes
(135)) leads tof of - 0.1 Hz, whereas the permeability -3 orders magnitude smaller than
in the ECM (e.g., lxl0 -17 m 4/N-s for chondrocytes (136)) results in the breakpoint
frequency of - 0.0002 Hz. To fully confirm fp, the application of more broad range of
164
frequencies (e.g., 0.01-1000 Hz) would be needed, as yet currently limited by AFM
instrumentation.
Overall, our dynamic testing has successfully demonstrated the dynamic behavior
of chondrocytes at cellular level, the load-sustaining and energy dissipation varying with
frequency. The response of isolated chondrocytes to dynamic loading is also distinct from
ECM of articular cartilage; being qualitatively similar, isolated chondrocytes are
responsive to higher frequency range and the change in dynamic properties at given
frequency are larger for chondrocytes as compared to ECM (98). These findings can be
implicated with the modulation of frequency as a potential means to influence the
biosynthetic activity and gene expression of single chondrocytes at early phase of the cell
culture (113). Future study on the theoretical and experimental characterization of
dynamic properties at cellular level might elaborate the biomechanical comprehension of
cell microenvironments in relation to structural properties, ultimately for the tissueengineered constructs. With this regard, stem cell-based research is ongoing to
investigate the poro-viscoelastic nanomechanical behavior of cell microenvironment
during in vitro culture and to test whether this approach improves the engineered tissue
constructs with better mechanical functionality as well as biochemical integrity.
C.5 Conclusion
To explore the dynamic properties of chondrocytes in the broad range of time
scale, our study performed the dynamic oscillatory compression (amplitude - 5 nm) on
individual chondrocytes as a function of deformation frequency (f= 1-316 Hz) and probe
165
tip geometry (2.5 pam radius sphere and 50 nm radius square pyramid) using Atomic
Force Microscopy (AFM)-based nanoindentation. Based on our findings using stress
relaxation followed by dynamic compression, we hypothesized that the time-dependent
mechanical behavior of chondrocytes is governed by solid viscoelasticity in the longerterm relaxation and significantly affected by poroelasticity in the initial fraction of the
transient behavior. The frequency-dependence of chondrocytes was qualitatively similar
to that of the mature cartilage tissue. However, dynamic properties of chondrocytes
exhibit lower dynamic modulus and higher permeability than those of the mature
cartilage, possibly due to the contribution of different molecular components
(cytoskeletons vs. collagen and aggrecan) due to its unique structural feature. This
experimental study with theoretical model serves to elaborate the biomechanical
comprehension of the chondrocytes, critically important in the design of the tissue
engineered constructs for cartilage tissue engineering.
Acknowledgements
Supported by NSF-NIRT 0403903, NSF-CMMI 0758651, NIH Grant AR33236,
and the MIT Institute for Soldier Nanotechnologies, US Army Research Office (contract
number DAAD-19-02-D0002). The content does not necessarily reflect the position of
the government and no official endorsement should be inferred.
Conflicts of interest
None of the authors have a conflict of interest associated with this work.
166
Appendix D
Cell Fixation Protocol
Paraformaldehyde
D.1
Fixation
(preferred
when
planning to
do
immunostaining of beads)
1. Place beads in a 4% paraformaldehyde solution containing 0.1M cacodvlate,
10mM CaC12, 10% CPC (cetylpyridinium chloride; to precipitate PGs) (Solution
1) at pH 7.4 for 4 hrs at RT. For each bead, Add 1.5 mL of Solution 1 to each
bead in the sample vial.
2.
Wash in 0.1M cacodvlate, 50mM BaCl2 (Solution 2) twice briefly at RT and then
overnight at 4C, changing the wash once if possible.
3. NOTE: if fixation occurs longer that 4 hours, wash time should also be increased.
4. Wash beads with water for 15 min prior to dehydration
5. Add 70% EtOH and Keep the sample vials at 40 C.
167
168
Appendix E
Histology Staining Protocol
E.1 Fixing Cells
Use a 1:1 ratio of cell suspension to fixative (SafefixII, Cat No. 042-600, Fisher
Scientific; this is much safer than paraformaldehyde). The fixed cells can be used
immediately.
E.2 Mounting Cells onto a Slide
1. Cytospin (Cyto-Tek Miles, Diagnostic Division, Mishawaka, IN; Massachusetts
General Hospital, contact Han-Hwa Hung)
2.
Blotting paper can be used for the insert gasket between the fluid chamber and
the slide.
169
3. Add 3mL of 6% hetastarch (in lactated electrolyte injection) to two drops of cell
solution (lxl0 6 cells/mL)
4.
Spin cells (14x100 rpm) for 10 min.
5. Wipe away excess fluid around the cell mounted area
6. Air dry
E.3 Toluidine Blue
Toluidine Blue is used for staining negatively charged molecules and is typically
used to stain for proteoglycans.
1. Make 0.1% Tol Blue Stain solution by dissolving ig Toluidine Blue O (SigmaAldrich, Cat No. 198161, FW 305.83) in 1000g DI water
2.
Filter the solution through filter paper to remove larger particles
3. Gently pipette -I 001tL of Tol Blue onto the mounted cells on glass slide.
4.
Incubate for 5 minutes
5. Shake off excess dye
6.
Gently dip the slide a few times (3-4) in DI water. Note: Do not rinse too well, or
the stain will wash away
170
E.4 Aniline Blue
Aniline Blue is typically used to stain for collagen. The phosphomolybdic acid is
taken up by connective tissue (e.g., collagen I and II). A substitution reaction occurs for
phosphomolybdic with aniline blue.
* Phosphomolybdic acid (PMA), 1% aq. [Rowley Biochemical Inc., Danver, MA;
Cat No. F-362-7]
* Aniline Blue Solution [Rowley Biochemical Inc., Danver, MA; Cat No. F-362-4]
1. Gently pipette -150iL of PMA onto the mounted cells on glass slide.
2. Incubate for 2 minutes.
3. Shake off excess PMA.
4. Pipette -150 VtL of Aniline Blue.
5. Incubate 5-8 minutes.
6. Shake off excess dye.
7. Gently dip the slide 2-3 times in DI water (can also try 95% ethanol). Note: This
dye is rinsed off very easily, so do not dip the slide too many times.
171
E.5 Mounting Coverslip
Once the slides are dry, the coverslip should be put on the sample for protection.
Using a blunted glass probe, drip one drop of Toluene (Note: one drop is enough since it
will spread well; Permount, Fischer Scientific, Cat No. UN1294) onto cell area. Quickly
place a coverslip over the drop. Allow to dry in hood overnight. Use xylene to remove
any toluene adhered to the glass probe.
E.6 Taking Microscopy images
Contact Division of Comparative Medicine to use a microscopy that can capture
the images to a computer (Arlin B. Rogers abr(dmit.edu, 617-253-9441).
172
Appendix F
Casting Alginate Beads with Cells
F.1 Materials
1. Autocl ave:
50mL Beaker,
3 x 500mL Bottles (for Solutions below)
1 spatula (to move beads)
2. Soluti )ns:
2% Alginate in 0.9% NaCI
0.9% NaCI: 9g in 1000mL DI water
150mM NaCl (FW 58.44): 4.38g in 500ml DI water
102mM CaCl 2 (FW 111) in 0.9% NaCI:
5.66g in 500mL of 150 mM Na(=l
55mM Na Citrate (FW 294.10) in 150mM NaCl
*Filter all solutions through a 200gm filter and into the sterile bottles
173
F.2 Formation of alginate beads
1. Fill the sterile 50mL beaker with 30mL of 102mM CaCl2 solution.
2. Get a cell count of cells using the hematocytometer. Spin cells down(if needed).
1800rpm, 10min. Suction off supernatant (being careful not to suck up any cells).
Gently tap bottom of vial to break up cells. May need to add a ImL sterile PBS(no
Mg 2+ and Ca 2 + , pH7) to help break up the cells.
3. Add appropriate amount of alginate. (ideal conc.: 10-15x106 cells/mL; can be as
high as 20x10 6 cells/mL)
4. Use 16G needle and 3mL syringe to suction up the alginate-chondrocyte solution.
Remove as many air bubbles as possible from the syringe.
5. Remove 16G needle and replace with 22G needle. Hold the syringe at a 450 angle
with the hole facing downward.
6. With steady pressure, drop even-sized droplets into the CaCI 2 solution. Halfwaythrough you may want to exchange the CaCl 2 solution with fresh solution so the
Ca 2+ does not get depleted.
7. After 10min, suction off the CaCl2 and rinse twice with 20mL sterile PBS(no Mg2+
and Ca ). Let each rinse go for Imin.
8. Replace PBS with culture medium.
9. Fill a 12-well plate with 2mL culture medium in each well. Using a sterile spatula,
transfer 6-8beads/well.
10. Do a cell viability assay on one of the beads.
11. Keep alginate-cell beads in an incubator.
174
F.3 Dissolution of alginate beads
1. Put 4 beads into a 1.5mL eppendorf tube. Add 1.2mL NaCitrate. Place in a 370 C
water bath or incubator for 10 min. Shake gently to mix. Do not vortex.
2. Spin 3000-4000rpm (-1000rcf) in Eppendorf centrifuge 5415C for 5min.
3. Remove supernatant. Add ImL 0.9% NaCl. Centrifuge 3000-4000rpm, 5min.
Repeat.
4. Resuspend final pellet in PBS (if doing a cell viability assay) or culture medium
at desired concentration (-0.5mL).
175
176
Appendix G
Bovine Bone Marrow Stem Cells Isolation
*Adapted from protocols developed by Connelly et al(120), J Mouw, and Kisiday et al (50)
G.1 Bone Marrow Harvest and Processing
1. Prepare 1 50mL conical with 25 mL PBS+PSA for each bone to be harvested.
2. Remove all connective tissue and muscle around mid-diaphysus of each bone.
3. Clean bone with 70% ethanol.
4. Cut the bone with a sterilized hand-saw blade and transfer to the tissue culture hood.
5. Carefully removed bone marrow with a forceps and small scoop and transfer to
50mL conical tubes with PBS. Avoid scraping the interior surface of the bone.
6. Break up bone marrow by vigorous pipetting with a 25mL followed by 5mL pipette.
*May also need to use a syringe without a needle attached to apply enough pressure.
7. Continue to homogenize bone marrow by passing the material through a 16 gauge
needle
8. Centrifuge homogenized bone marrow 1000x g for 15 minutes.
177
9. Carefully aspirate supernatant and the fatty layer.
10. Resuspend in fresh PBS, pass material through an 18 gauge needle and then through
a 70ptm cell strainer
G.2 Nucleated Cell Count, Red Blood Cell Lysis, and Cell
Plating
1. Dilute cell suspension in Ammonium Chloride-Tris Base (7.6mg/mL AmCI,
2.4mg/mL Tris Base). * JC suggested 1:1 dilution in 4% acetic acid.
2. 1:20 dilution in AmCI/Tris for nucleated cell count via Trypan Blue (1:2000
dilution in PBS for red blood cell count). * Dilutions may need to be adjusted for
each harvest.
3. Obtain approximate nucleated cell count.
4. Aspirate supematant and add low glucose DMEM + 10% ES-FBS, PSA, HEPES, &
Ing/mL b-FGF.
5. Pre-plate approximately 20x10 6 cells in 5mL on 100mm tissue culture treated petri
dishes.
6. Incubate for 30 min to allow rapidly adhering cells to attach.
7. Transfer media and non-adherent cells to a T-flask and add another 10mL
expansion medium (approximately 250k cells/cm 2 in 15 mL media; ~20 million/T75).
178
8. After 2 days in culture, remove expansion media, (optional. wash cells thoroughly
with PBS) and add 15 mL fresh expansion medium.
9. Continue to culture flasks until cells are nearly (75-80%) confluent [PO].
10. Remove cells with 0.05% trypsin-EDTA, centrifuge and resuspend in PBS, and
count via Trypan Blue.
11.
Either freeze in 10 million cells/vial aliquots or re-plate at approximately 6,000/cm2 .
12. Culture until nearly confluent, passage [P1], and reseed again at 6,000/cm2 .
13. Culture again until nearly confluent, passage [P2], and seed into 3-D constructs for
chondrogenesis.
179
180
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Dudhia J. 2005. Aggrecan, aging and assembly in articular cartilage. Cell Mol
Life Sci. 62:2241-2256.
Guilak F., Mow, V.C. 2000. The mechanical environment of the chondrocyte: a
biphasic finite element model of cell-matrix interactions in articular cartilage. J
Biomech. 33:1663-1673.
Lee G. M., Johnstone, B., Jacobson, K., Caterson, B. 1993. The dynamic structure
of the pericellular matrix on living cells. J Cell Biol. 123:1899-1907.
Ng L., Hung H-H, Sprunt A., Chubinskaya S., Ortiz C., Grodzinsky A. J. 2007.
Nanomechanical properties of individual chondrocytes and their developing
growth factor-stimulated pericellular matrix. J Biomech. 40:1011-1023.
Darling E. M., Zauscher S., Guilak F. 2006. Viscoelastic properties of zonal
articular chondrocytes measured by atomic force microscopy. Osteoarthr
Cartilage. 14:571-579.
Knudson W., Knudson, C. B. 1991. Assembly of a chondrocyte-like pericellular
matrix on non-chondrogenic cells J. cell Sci. 99:
Lee G. M., Poole, C. A., Kelley, S. S., Chang, J., Caterson, B. 1997. Isolated
chondrons: a viable alternative for studies of chondrocyte metabolism in vitro.
Osteoarth Cartilage. 5:261-274.
Lee G. M., Loeser, R. F. 1998. Interactions of the chondrocyte with its
pericellular matrix Cell Mater. 8:135-149.
DiMicco M. A., Kisiday J. D., Gong H., Grodzinsky A. J. 2007. Structure of
pericellular matrix around agarose-embedded chondrocytes. Osteoarthr Cartilage.
15:1207-1216.
Durrant L. A., Archer, C. W., Benjamin, M., Ralphs, J. R. 1999. Organisation of
the chondrocyte cytoskeleton and its response to changing mechanical conditions
in organ culture. J Anat. 194:343-353.
Knight M. M., Bomzon, Z., Kimmel, E., Sharma, A. M., Lee, D. A., Bader, D. L.
2006. Chondrocyte deformation induces mitochondrial distortion and
heterogeneous intracellular strain fields. Biomechan Model Mechanobiol. 5:180191.
Szafranski J. D., Grodzinsky, A. J., Burger, E., Gaschen, V., Hung, H.-H.,
Hunziker, E. B. 2004. Chondrocyte mechanotransduction: effects of compression
on deformation of intracellular organelles and relevance to cellular biosynthesis.
Osteoarth Cartilage. 12:937-946.
Darling E. M., Hu, J. C., Athanasiou, K .A. 2004. Zonal and topographical
differences in articular cartilage gene expression. J Orthop Res. 22:1182-1187.
181
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
Trickey W. R., Vail, T.P., Guilak, F. 2004. The role of the cytoskeleton in the
viscoelastic properties of human articular chondrocytes. J Orthop Res. 22:131-139.
Alexopoulos L. G., Haider, M. A., Vail, T. P., Guilak, F. 2003. Alterations in the
mechanical properties of the human chondrocyte pericellular matrix with
osteoarthritis. J Biomech Eng. 125:323-333.
Alexopoulos L. G., Williams, G. M., Upton, M. L., Setton, L. A., Guilak, F. 2005.
Osteoarthritic changes in the biphasic mechanical properties of the chondrocyte
pericellular matrix in articular cartilage. J Biomech. 38:509-517.
Sato H., Ohshuma, N., Nerem, R. M. 1996. Viscoelastic properties of cultured
porcine aortic endothelial cells exposed to shear stress. J Biomech. 29:461-467.
Trickey W. R., Baaijens, F. P. T., Laursen, T. A., Alexopoulos, L. G., Guilak, F.
2006. Determination of the Poisson's ratio of the cell: recovery properties of
chondrocytes after release from complete micropipette aspiration. J Biomech.
39:78-87.
Collinsworth A. M., Zhang, S., Kraus, W. E., Truskey, G. A. 2002. Apparent
elastic modulus and hysteresis of skeletal muscle cells through differentiation.
Am J Physiol Cell Physiol. 283:C1219-C1227.
A-Hassan E., Heinz, W F, Antonik, M D, D'Costa, N P, Nageswaran, S,
Schoenenberger, C A, Hoh, J H. 1998. Relative microelastic mapping of living
cells by atomic force microscopy. Biophys J. 74:1564-1578.
Mathur A. B., Collinsworth, A.M., Reichert, W.M., Kraus, W.E., Truskey, G.A.
2001. Endothelial, cardiac muscle and skeletal muscle exhibit different viscous
and elastic properties as determined by atomic force microscopy. J Biomech.
34:1545-53.
Charras G. T., Horton, M. A. 2002. Determination of cellular strains by combined
atomic force microscopy and finite element modeling. Biophys J. 83:858-79.
Stolz M., Raiteri, R., Daniels, A. U., VanLandingham, M. R., Baschong, W., Aebi,
U. 2004. Dynamic elastic modulus of porcine articular cartilage determined at two
different levels of tissue organization by indentation-type atomic force
microscopy. Biophys J. 86:3269-3283.
Rotsch C., Jacobson, K., Radmacher, M. 1999. Dimensional and mechanical
dynamics of active and stable edges in motile fibroblasts investigated by using
atomic force microscopy. Proc. Natl. Acad. Sci. USA. 96:921-926.
Bausch A. R., Ziemann, F., Boulbitch, A. A., Jacobson, K., Sackmann, E. 1998.
Local measurements of viscoelastic parameters of adherent cell surfaces by
magnetic bead microrheometry. Biophys J. 75:2038-2049.
Fabry B., Maksym, G. N., Butler, J. P., Glogauer, M., Navajas, D., Fredberg, J. J.
2001. Scaling the microrheology of living cells. Phys rev let. 87:148102-1-4.
Yamada S., Wirtz, D., Kuo, S.C. 2000. Mechanics of living cells measured by
laser tracking microrheology. Biophys J. 78:1736-1747.
Guilak F., Erickson, G.R., Ting-Beall, H.P. 2002. The effects of osmotic stress on
the viscoelastic and physical properties of articular chondrocytes. Biophys J.
82:720-727.
Chao P. G., Tang, Z., Angelini, E., West, A. C., Costa, K. D., Hung, C. T. 2005.
Dynamic osmotic loading of chondrocytes using a novel microfluidic device. J
Biomech. 38:1273-1281.
182
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
Peeters E. A., Oomens, C.W., Bouten, C.V., Bader, D.L., Baaijens, F.P. 2005.
Viscoelastic properties of single attached cells under compression. J Biomech Eng.
127:237-43.
Koay E. J., Shieh, A.C., Athanasiou, K.A. 2003. Creep indentation of single cells.
J Biomech Eng. 125:334-341.
Shieh A. C., Athanasiou K. A. 2006. Biomechanics of single zonal chondrocytes.
J Biomech. 39:1595-1602.
Leipzig N. D., Athanasiou, K.A. 2005. Unconfined creep compression of
chondrocytes. J Biomech. 38:77-85.
Bremmell K. E., Evans, A., Prestidge, C. A. . 2006. Deformation and nanorheology of red blood cells: An AFM investigation. Colloids Surfaces B. 50:43-48.
Thoumine O., Ott, A. 1997. Time scale dependent viscoelastic and contractile
regimes in fibroblast probed by microplate manipulation. J. Cell. Sci. 110:2109.
Desprat N., Richert, A., Simeon, J., Asnacios, A. 2005. Creep function of a single
living cell. Biophys J. 88:2224-2233.
Freeman P. M., Natarajan, R.N., Kimura, J.H., Andriacchi, T.P. 1994.
Chondrocyte cells respond mechanically to compressive loads. J Orthopaedic Res.
12:311-320.
Trepat X., Grabulosa, M., Puig, F., Maksym, G.N., Navajas, D., Farre, R. 2004.
Viscoelasticity of human alveolar epithelial cells subjected to stretch. Am J
Physiol Lung Cell Mol Physiol. 287:L1025-34.
Trickey W. R., Lee, G. M., Guilak, F. . 2000. Viscoelastic properties of
chondrocytes from normal and osteoarthritic human cartilage J Orthop Res.
18:891-898.
Smith B., Tolloczko, B., Martin, J., Grutter, P. 2005. Probing the viscoelastic
behavior of cultured airway smooth muscle cells with atomic force microscopy:
stiffening induced by contractile agonist. Biophys J. 88:2994-3007.
Alcaraz J., Buscemi, L., Grabulosa, M., Trepat, X., Fabry, B., Farre, R., Navajas,
D. 2003. Microrheology of human lung epithelial cells measured by atomic force
microscopy. Biophys J. 84:2071-2079.
Mahaffy R. E., Park, S., Gerde, E., Kis, J., Shih, C.K. 2004. Quantitative analysis
of the viscoelastic properties of thin regions of fibroblasts using atomic force
microscopy. Biophys J. 86:1777-1793.
Rico F., Roca-Cusachs, P., Gavara, N., Farre, R., Rotger, M., Navajas, D. 2005.
Probing mechanical properties of living cells by atomic force microscopy with
blunted pyramidal cantilever tips. Phys Rev E. 72:1-10.
Park S., Koch D., Cardenas R., Kis J., Shih, C. K. 2005. Cell motility and local
viscoelasticity of fibroblasts. Biophys J. 89:4330-4342.
Shepherd D. E. T., Seedhom, B.B. 1997. A technique for measuring the
compressive modulus of articular cartilage under physiological loading rates with
preliminary results. Pro Instn Mech Engrs Part H. 211:155-165.
Mehlhorn A. T., Schmal, H., Kaiser, S., Lepski, G., Finkenzeller, G., Stark, G. B.,
Sidkamp, N. P. 2006. Mesenchymal stem cells maintain TGF-P-mediated
chondrogenic phenotype in alginate bead culture Tissue Eng. 12:1393-1403.
Mehlhorn A. T., Niemeyer, P., Kaiser, S., Finkenzeller, G., Stark, G. B., Siidkamp,
N. P., Schmal, H. 2006. Differential expression pattern of extracellular matrix
183
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
molecules during chondrogenesis of mesenchymal stem cells from bone marrow
and adipose tissue. Tissue Eng. 12:2853-2862.
Ma H.-L., Hung, S.-C., Lin, S.-Y., Chen, Y.-L., Lo, W.-H. 2003. Chondrogenesis
of human mesenchymal stem cells encapsulated in alginate beads. J Biomed
Mater Res. 64A:273-281.
Awad H. A., Wickham, M. Q., Leddy, H. A., Gimble, J. M., Guilak, F. 2004.
Chondrogenic differentiation of adipose-derived adult stem cells in agarose,
alginate, and gelatin scaffolds. Biomaterials. 25:3211-3222.
Kisiday J. D., Kopesky, P. W., Evans, C. H., Grodzinsky, A. J., Mcllwraith, C. W.,
Frisbie, D. D. 2008. Evaluation of adult equine bone marrow- and adiposederived progenitor cell chondrogenesis in hydrogel cultures. J Orthop Res.
26:322-331.
Kolambkar Y. M., Peister, A., Soker, S., Atala, A., Guldberg, R. E. 2007.
Chondrogenic differentiation of amniotic fluid-derived stem cells. J Mol Hist.
38:405-413.
Yourek G., Hussain, M. A., Mao, J. J. 2007. Cytoskeletal changes of
mesenchymal stem cells during differentiation. ASAIO. 53:219-228.
Darling E. M., Topel, M., Zaucher, S., Vail, T. P., Guilak, F. 2008. Viscoelastic
properties of human mesenchymally-derived stem cells and primary osteoblasts,
chondrocytes, and adipocytes. J Biomech. 41:454-464.
Poole C. A. 1997. Articular cartilage chondrons: form, function and failure. J
Anat. 191:1-13.
Schinagl R. M., Gurskis, D., Chen, A. C., Sah, R. L. 1997. Depth-dependent
confined compression modulus of full-thickness bovine articular cartilage. J
Orthop Res. 15:499-506.
Wang C. C., Guo, X. E., Sun, D., Mow, V. C., Ateshian, G. A., Hung, C. T. 2002.
The functional environment of chondrocytes within cartilage subjected to
compressive loading: a theoretical and experimental approach. Biorheology.
39:11-25.
Guilak F. 2000. The deformation behavior and viscoelastic properties of
chondrocytes in articular cartilage. Biorheology. 37:27-44.
Poole C. A., Flint M. H., Beaumont B. W. 1984. Morphological and functional
interrelationships of articular cartilage matrices. J Anat. 138:113-138.
Hunziker E. B., Kapfinger, E., Geiss, J. 2007. The structural architecture of adult
mammalian articular cartilage evolves by a synchronized process of tissue
resorption and neoformaion during postnatal development. OA&C. 15:403-413.
Guilak F., Alexopoulos, L. G., Upton, M. L., Youn, I., Choi, J. B., Cao, L., Setton,
L. A., Haider, M. A. 2006. The pericellular matrix as a transducer of
biomechanical and biochemical signals in articular cartilage. Ann NY Acad Sci.
1068:498-512.
Poole C. A., Ayad, A., Gilbert, R. T. 1992. Chondrons from articular cartilage: V.
Immunohistochemical evaluation of type VI collagen organization in isolated
chondrons by light, confocal and electron microscopy. J Cell Sci. 103:1101-1110.
Knudson W., Aguiar, D. J., Hua, Q., Knudson, C.B. 1996. CD44-anchored
hyaluronal-rich pericellular matrices: an ultrastructural and biochemical analysis.
Exp Cell Res. 228:216-228.
184
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
Poole C. A., Flint M. H., Beaumont B. W. 1987. Chondrons in cartilage:
ultrastructural analysis of the pericellular microenvironment in adult human
articular cartilages. J. Orthop. Res. 5:509-522.
Fraser S. A., Crawford, A., Frazer, A., Dickinson, S., Hollander, A. P., Brook, I.
M., Hatton, P. V. 2006. Localization of type VI collagen in tissue-engineered
cartilage on polymer scaffolds. Tissue Eng. 12:569-577.
Mauck R. L., Soltz, M. A., Wang, C. C. B., Wong, D. D., Chao, P.-H. G., Valhmu,
W. B., Hung, C. T., Ateshian, G. A.,. 2000. Functional tissue engineering of
articular cartilage through dynamic loading of chondrocyte-seeded agarose gels. J
Biomech Eng. 122:252-260.
Villanueva I., Bishop, N. L., Bryant, S. J. (in press). Medium osmolarity and
PCM development improves chondrocyte survival when photoencapsulated in
PEG hydrogels at low densities. Tissue Eng.
Masuda K., Pfister, B. E., Sah, R. L., Thonar, E. J-M. A. 2006. Osteogenic
protein-i promotes the formation of tissue-engineered cartilage using the alginaterecovered-chondrocyte method. Osteoarth Cartilage. 14:384-391.
Evanko S. P., Tammi, M. I., Tammi, R. H., Wight, T. N. 2007. Hyaluronandependent pericellular matrix. Adv Drug Deliv Rev. 59:1351-1365.
Athanasiou K. A., Rosenwasser, M.P., Buckwalter, J. A., Malinin, T. I., Mow, V.
C. . 1991. Interspecies comparisons of in situ intrinsic mechanical properties of
distal femoral cartilage. J Orthop Res. 9:330-340.
Athanasiou K. A., Agarwal, A., Muffoletto, A., Dzida, F. J., Constantinides, G.,
Clem, M. 1995. Biomechanical properties of hip cartilage on experimental animal
models. Clin Orthop. 316:254-266.
Athanasiou K. A., Agarwal, A., Dzida, F. J. 1994. Comparative study of the
intrinsic mechanical properties of the human acetabular and femoral head
cartilage. J Orthop Res. 12:340-349.
Athanasiou K. A., Liu, G. T., Lavery, L. A., Lanctot, D. R., Schenck Jr., R. C.
1998. Biomechanical topography of human articular cartilage in the first
metatarsophaloangeal joint. Clin Orthop. 348:269-281.
Langelier E., Suetterlin, R., Hoemann, C. D., Aebi, U., Buschmann, M. D. 2000.
The chondrocyte cytoskeleton in mature articular cartilage: structure and
distribution of actin, tubulin, and vimentin filaments. J Histochem Cytochem.
48:1307-1320.
Alexopoulos L. G., Setton, L. A., Guilak, F. 2005. The biomechanical role of the
chondrocyte pericellular matrix in articular cartilage. Acta Biomater. 1:317-325.
Grodzinsky A. J., Lipshitz, H., Glimcher, M. J. 1978. Electromechanical
properties of articular cartilage during compression and stress relaxation. Nature.
275:448-450.
Kim E., Guilak, F., Haider, M. A. 2008. The dynamic mechanical environment of
the chondrocyte: a biphasic finite element model of cell-matrix interactions under
cyclic compressive loading. J Biomech Eng. 130:061009-1-10.
Buschmann M. D., Gluzband Y. A., Grodzinsky A. J., Hunziker E. B. 1995.
Mechanical compression modulates matrix biosynthesis in chondrocyte/agarose
culture. J Cell Sci. 108:1497-1508.
185
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
Buschmann M. D., Kim, Y.-J.,Wong, M., Frank, E., Hunziker, E. B., Grodzinsky,
A. J. . 1999. Stimulation of aggrecan synthesis in cartilage explants by cyclic
loading is localized to regions of high interstitial fluid flow. Arch Biochem
Biophys 366:1-7.
Leipzig N. D., Athanasiou, K.A. 2008. Static compression of single chondrocytes
catabolically modifies single-cell gene expression. Biophys J. 94:2412-2422.
Chahine N. O., Hung C. T., Ateshian G. A. 2007. In-situ Measurements of
chodnrocytes deformation under transient loading. Eur Cells Materials. 13:100111.
Frangonas E., Valente M., Pozzi-Mucelli M., Toffanin R., Rizzo R., Silvestri F.,
et al. 2000. Articular cartilage repair in rabbits by using suspensions of allogenic
chondrocytes in alginate. Biomaterials. 21:795-801.
Masuda K., Sah, R.L., Hejna, M.J., Thonar, E. 2003. A novel two-step method for
the formation of tissue-engineered cartilage by mature bovine chondrocytes: the
alginate-recovered chondrocyte (ARC) method. J Orthop Res. 21:139-148.
Kovacs G. T. A., Maluf, N. I., Petersen, K. E. 1998. Bulk Micromachining of
Silicon. Proceedings of the IEEE. 86:1536-155 1.
Krupitka A., Johansson, M., Hult, A. 2003. Viscoelasticity in polymer films on
rigid substrates Macromol Mater Eng. 288:108-116.
Johnson K. L., Greenwood, J.A. 1997. An adhesion map for the contact of elastic
spheres. J Colloid Interface Sci. 192:326-333.
Rutland M. W., Tyrrell, J. W. G., Attard, P. . 2004. Analysis of atomic force
microscopy data for deformable materials J. Adhesion Sci. Technol. 18:11991215.
Trickey W. R., Lee, M., Guilak, F. 2000. Viscoelastic properties of chondrocytes
from normal and osteoarthritic human cartilage. J Orthop Res. 18:891-898.
Papagiannopoulos A., Waigh, T. A., Hardingham, T. E., Heinrich, M. 2006.
Solution structure and dynamics of cartilage aggrecan. Biomacromolecules.
7:2162-2172.
Papagiannopoulos A., Waigh, T. A., Hardingham, T. E. 2008. The viscoelasticity
of self-assembled proteoglycan combs. Faraday Discuss. 139:337-357.
Pins G. D., Huang, E. K., Christiansen, D. L., Silver, F. H. 1997. Effects of static
axial strain on the tensile properties and failure mechanisms of self-assembled
collagen fibers. J Appl Polym Sic. 63:1429-1440.
Han L., Dean D., Mao P., Ortiz C., Grodzinsky A. J. 2007. Nanoscale shear
deformation mechanisms of opposing cartilage aggrecan macromolecules.
Biophys J. 93:L23-L25.
Mow V. C., Kuei, S.C., Lai, W.M., Armstrong, C.G. 1980. Biphasic creep and
stress relaxation of articular cartilage in compression: theory and experiments. J
Biomech Eng. 102:73-84.
Lin D. C., Horkay, F. 2008. Nanomechanics of polymer gels and biological
tissues: A critical review of analytical approaches in the Hertzian regime and
beyond. Soft Matter. 4:669-682.
Tuli R., Li, W.-J.,Tuan, R. S. 2003. Current state of cartilage tissue engineering.
Arthritis Res Ther. 5:235-238.
186
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
Kisiday J., Jin, M., Kurz, B., Hung, H., Semino, S., Zhang, S., Grodzinsky, A. J.
2002 Self-assembling peptide hydrogel fosters chondrocyte extracellular matrix
production and cell division: implications for cartilage tissue repair. Proc Natl
Acad Sci. 99:9996-10001.
Millward-Sadler S. J., Salter, D. M. 2004. Integrin-dependent signal cascades in
chondrocyte mechanotransduction. Ann Biomed Eng. 32:435-446.
Millward-Sadler S. J., Wright, M.O., Davies, L.W., Nuki, G., Salter, D.M. 2000.
Mechanotransduction via integrins and interleukin-4 results in altered aggrecan
and matrix metalloproteinase 3 gene expression in normal, but not osteoarthritic,
human articular chondrocytes. Arthritis Rheum. 43:2091-2099.
Kim Y.-J., Bobassar, L. J., Grodzinsky, A. J. 1995. The role of cartilage streaming
potential, fluid flow and pressure in the stimulation of chondrocyte biosynthesis
during dynamic compression J Biomech. 28:1055-1066.
Mahaffy R. E., Shih, C.K., MacKintosh, F.C., Ks, J. 2000. Scanning probe-based
frequency-dependent microrheology of polymer gels and biological cells. Phys
Rev Lett. 85:880-883.
Loeser R. F., Pacione, C. A., Chubinskaya, S. 2003. The combination of insulinlike growth factor I and osteogenic protein 1 promotes increased survival of and
matrix synthesis by normal and osteoarthritic human articular chondrocytes
Arthritis Rheum. 48:2188-2196.
Alcaraz J., Busceme, L., Puig-de-Morales, M., Colchero, J., Bar6, A., Navajas, D.
2002. Correction of microrheological measurements of soft samples with atomic
force microscopy for the hydrodynamic drag on the cantilever. Langmuir. 18:716721.
Radmacher M. 1997. Measuring the elastic properties of biological samples with
the AFM. IEEE Eng Med Biol. 16:47-57.
Freeman P. M., Natarajan, R.N., Kimura, J.H.,Andriacchi, T.P. 1994.
Chondrocyte cells respond mechanically to compressive loads. Journal of
Orthopaedic Research 12:311-320.
Soltz M. A., Ateshian, G. A. 2000. Interstitial fluid pressurization during confined
compression cyclical loading of articular cartilage. Ann Biomed Eng. 28:150-159.
Morel V., Quinn, T. M. 2004. Cartilage injury by ramp compression near the gel
diffusion rate. J Orthopaedic Res. 22:145-151.
Lee R. C., Frank, E. H., Grodzinsky, A. J., Roylance, D. K. 1981. Oscillatory
compressional behavior of articular cartilage and its associated electromechanical
properties J Biomech Eng. 103:280-292.
Grodzinsky A. J., Roth, V., Myers, E., Grossman, W. D., Mow, V. C. 1981. The
significance of electromechanical and osmotic forces in the nonequilibrium
swelling behavior of articular cartilage in tension. J Biomech Eng. 103:221-23 1.
Happel J. 1959. Viscous flow relative to arrays of cylinders. AIChE J. 5:174-177.
Eisenberg S. R., Grodzinsky, A. J. 1988. Electrokinetic micromodel of
extracellular matrix and other polyelectrolyte networks. Physicochem Hydrodyn.
10:517-539.
Frank E. H., Grodzinsky, A.J. 1987. Cartilage Electromechanics - II. Continuum
model of cartilage electrokinetics and correlation with experiments. J Biomech.
20:pp.629-639.
187
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
Huang C.-Y., Soltz, M. A., Kopacz, M., Ateshian, G. A. 2003. Experimental
verification of the roles of intrinsic matrix compression nonlinearity in the
biphasic response of cartilage. J Biomech Eng. 125:84-93.
Buschmann M. D., Gluzband, Y. A., Grodzinsky, A. J., Kimura, J. H., Hunziker,
E. B. 1992. Chondrocytes in agarose culture synthesize a mechanically functional
extracellular matrix. J Orthopaedic Res. 10:745-758.
Shieh A. C., Athanasiou, K.A. 2006. Dynamic compression of single cells.
Osteoarthr Cartilage. 15:328-324.
Han L., Dean D., Daher, L. A., Grodzinsky A. J., Ortiz, C. 2008. Cartilage
aggrecan undergoes self-adhesion. Biophys J. 95:4862-4870.
Flechtenmacher J., Huch, K., Thonar, E. J.-M. A., Mollenhauer, J. A., Davies, S.
R., Schmid, T. M., Puhl, W., Sampath, T. K., Aydelotte, M. B., Kuettner, K. E. .
1996. Recombinant human osteogenic protein 1 is a potent stimulator of the
synthesis of cartilage proteoglycans and collagens by human articular
chondrocytes Arthritis Rheum. 39:1896-1904.
Nishida Y., Knudson, C. B., Eger, W., Kuettner, K. E., Knudson, W. 2000.
Osteogenic protein I stimulates cell-associated matrix assembly by normal human
articular chondrocytes Arthritis Rheum. 43:206-214.
Ng L., Grodzinsky, A. J., Patwari, P., Sandy, J., Plaas, A., Ortiz, C. 2003.
Individual cartilage aggrecan macromolecules and their constituent
glycosaminoglycans visualized via atomic force microscopy. Journal of Structural
Biology. 143:242-257.
Dean D., Han L., Ortiz C., and Grodzinsky A. J. 2005. Nanoscale conformation
and compressibility of cartilage aggrecan using microcontact printing and atomic
force microscopy. Macromol. 38:4047-4049.
Tuan R. S. 2006. Stemming cartilage degeneration: adult mesenchymal stem cells
as a cell source for articular cartilage tissue engineering. ARTHRITIS &
RHEUMATISM. 54:3075-3078.
Connelly J. T., Wilson, C. G., Levenston, M. E. 2008. Characterization of
proteoglycan production and processing by chondrocytes and BMSCs in tissue
engineered constructs. Osteoarth Cartilage. 16:1092-1100.
Danti S., D'Acunto, M., Trombi, L., Berrettini, S., Pietrabissa, A. 2007. A
micro/nanoscale surface mechanical study on morpho-functional changes in
multilineage-differentiated human mesenchymal stem cells. Macromol Biosci.
7:589-598.
Docheva D., Radula, D., Popov, C., Mutschler, W., Clausen-Schaumann, H.,
Schieker, M. 2008. Researching into the cellular shape, volume and elasticity of
mesenchymal stem cells, osteoblasts and osteosarcoma cells by atomic force
microscopy. J cell Mol Med 12:537-552.
Lee B., Han, L., Frank, E. H., Chubinskaya, S., Ortiz, C., Grodzinsky, A. J. 2009
(submitted). Dynamic mechanical properties of the tissue-engineered matrix
associated with individual chondrocytes J Biomech.
Luna E. G., Manual of histologicalstainingmethods. 1968, New York: McGrawHill Publication.
188
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
Famdale R. W., Buttle, D. J., Barrett, A. J. 1986. Improved quantitation and
discrimination of sulphated glycosaminoglycans by use of dimethylmethlyene
blue Biochim Biophys Acta. 883:173-177.
Woessner Jr. J. F. 1961. The determination of hydroxyproline in tissue and
protein samples containing small amounts of this imino acid Arch Biochem
Biophys. 93:440-447.
Kim Y.-J., Sah, R. L., Doong, J. Y., Grodzinsky, A. J. 1998. Fluorometric assay
of DNA in cartilage explants using Hoechst 33258. Anal Biochem. 174:168-176.
Charras G. T., Yarrow, J. C., Horton, M. A., Mahadevan, L., Mitchison, T. J.
2005. Non-equilibration of hydrostatic pressure in blebbing cells. Nature.
435:365-369.
Maroudas A., Mizrahi, J., Ben Haim, E., Ziv, I. 1987. Swelling pressure in
cartilage. Adv Microcirc. 13:203-212.
Buschmann M. D., Grodzinsky, A. J. 1995. A molecular model of proteoglycanassociated electrostatic forces in cartilage mechanics. J Biomech Eng. 117:179192.
Kopesky P. W., Vanderploeg E. J., Sandy J. S., Kurz B., and Grodzinsky A. J.
2009 (submitted). Self-assembling peptide hydrogels modulate in vitro
chondrogenesis of bovine bone marrow stromal cells. Tissue Eng A.
Rowley J. A., Madlambayan, G., Mooney, D. J. 1999. Alginate hydrogels as
synthetic extracellular matrix materials. Biomaterials. 20:45-53.
Bachrach N. M., Valhmu, W. B., Stazzone, E., Ratcliffe, A., Lai, W. M. 1995.
Changes in proteoglycan synthesis of chondrocytes in articular cartilage are
associated with the time-dependent changes in their mechanical environment. J
Biomech. 28:1561-1569.
Ateshian G. A., Costa, K. D., Hung, C. T. 2007. A theoretical analysis of water
transport through chondrocytes. Biochem Model Mechanobiol. 6:91-101.
Armstrong C. G., Mow, V. C. 1982. Variations in the intrinsic mechanical
properties of human articular cartilage with age, degeneration and water contents
J Bone Joint Surg. 64:88-94.
Wu J. Z., Herzog, W., Epstein, M. 1999. Modelling of location- and timedependent deformation of chondrocytes during cartilage loading. J Biomech.
32:563-572.
189