2018
Functional Effects of an
Interpenetrating Polymer Network on
Articular Cartilage Mechanical Properties
Mäkelä, JT
Elsevier BV
Tieteelliset aikakauslehtiartikkelit
© Osteoarthritis Research Society International
CC BY-NC-ND https://creativecommons.org/licenses/by-nc-nd/4.0/
http://dx.doi.org/10.1016/j.joca.2018.01.001
https://erepo.uef.fi/handle/123456789/6319
Downloaded from University of Eastern Finland's eRepository
Functional Effects of an Interpenetrating Polymer Network on Articular Cartilage Mechanical Properties
Janne TA. Mäkelä, Benjamin G. Cooper, Rami K. Korhonen, Mark W. Grinstaff, Brian D. Snyder
PII: S1063-4584(18)30002-5 DOI: 10.1016/j.joca.2018.01.001 Reference: YJOCA 4139
To appear in: Osteoarthritis and Cartilage Received Date: 14 July 2017
Revised Date: 18 December 2017 Accepted Date: 1 January 2018
Please cite this article as: Mäkelä JT, Cooper BG, Korhonen RK, Grinstaff MW, Snyder BD, Functional Effects of an Interpenetrating Polymer Network on Articular Cartilage Mechanical Properties,
Osteoarthritis and Cartilage (2018), doi: 10.1016/j.joca.2018.01.001.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
M AN US CR IP T
AC CE PT ED
Functional Effects of an Interpenetrating Polymer Network on Articular Cartilage Mechanical Properties
Janne TA Mäkelä1, Benjamin G Cooper1,2, Rami K Korhonen3,4, Mark W Grinstaff2,5,6*, Brian D Snyder1,7*
1Center for Advanced Orthopaedic Studies, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, USA
2Department of Chemistry, Boston University, Boston, MA, USA
3Department of Applied Physics, University of Eastern Finland, Kuopio, Finland.
4Diagnostic Imaging Center, Kuopio University Hospital, Kuopio, Finland.
5Department of Biomedical Engineering, Boston University, Boston, MA, USA
6Department of Medicine, Boston University, Boston, MA, USA
7Department of Orthopaedic Surgery, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA
Correspondence:
Professor Brian Snyder, MD, PhD Department of Orthopaedic Surgery Children’s Hospital
Harvard Medical School 300 Longwood Ave Boston, MA 02115 617-355-6021
Brian.Snyder@childrens.harvard.edu
Professor Mark W Grinstaff, PhD Professor Mark W Grinstaff, PhD Department of Chemistry
Department of Biomedical Engineering Department of Medicine
Boston University
Metcalf Center for Science & Engineering 590 Commonwealth Ave
Boston, MA 02215 617-358-3429 mgrin@bu.edu
M AN US CR IP T
AC CE PT ED
Abstract
1
Objective Depletion of glycosaminoglycans (GAGs) and degradation of collagen network are early 2
hallmarks of osteoarthritis. Currently, there are no chondroprotective therapies that mitigate the loss of 3
GAGs or effectively restore the collagen network. Recently, a novel polymeric cartilage supplement 4
was described that forms a charged interpenetrating polymer network (IPN) reconstituting the 5
hydrophilic properties of the extracellular matrix. To investigate the mechanism by which this 6
hydrophilic IPN improves articular cartilage material properties, a finite element (FE) model is used to 7
evaluate the IPN’s effect on the fibrillar collagen network, nonfibrillar matrix, and interstitial fluid 8
flow.
9
Methods Bovine osteochondral plugs were degraded with chondroitinase ABC to selectively decrease 10
GAG content. Samples were mechanically tested before and after IPN treatment using unconfined 11
testing geometry and stress-relaxation protocol. Every measurement was modeled separately using a 12
fibril-reinforced poroviscoelastic finite element model. Measurement replication was achieved by 13
optimizing the following model parameters: initial and strain-dependent fibril network modulus (Ef0
, 14
Ef
ε, respectively), nonfibrillar matrix modulus (Enf), initial permeability (k0) and strain-dependent 15
permeability factor (M).
16
Results Based on the FE model results, treatment of native and GAG depleted cartilage with the 17
hydrophilic IPN increases the ECM stiffness and impedes fluid flow. The IPN did not alter the stiffness 18
of fibrillary network. Cartilage permeability and the strain-dependent permeability factor decreased 19
with increasing IPN w/v%.
20
M AN US CR IP T
AC CE PT ED
Conclusions The IPN reconstitutes cartilage material properties primarily by augmenting the 21
hydrophilic ECM. This reinforcement of the solid phase also affects the fluid phase reestablishing low 22
permeability.
23
Keywords: Articular cartilage; Osteoarthritis; Biomaterials; Biomechanics; Finite element analysis 24
M AN US CR IP T
AC CE PT ED
Introduction
1
Articular cartilage is composed mainly of collagen type II ( 50-80% of dry weight), 2
glycosaminoglycans (GAGs) ( 30% of dry weight) and interstitial fluid ( 60-85% of the tissue 3
weight)1,2. From a biomechanics perspective, the interstitial fluid resists instantaneous loads while the 4
fibrillar network holds the tissue together and resists deformations. Thus, collagen fibers determine 5
primarily the tensile stiffness of the tissue. Under prolonged deformation, fluid flows out of the porous 6
and permeable extracellular matrix (ECM) leaving the hydrophilic GAG matrix responsible for the 7
equilibrium stiffness of the tissue.
8
In osteoarthritis (OA) the integrity of the articular cartilage is impaired irreversibly. Before the 9
tissue begins to wear off, changes are already occurring in the properties of the constituents. The very 10
first structural changes include fibrillation of the collagen matrix and diminution in the GAG content3–
11
5. The structural integrity becomes compromised leading to functional tissue changes including:
12
increased permeability, and decreased equilibrium and dynamic stiffness6–12. 13
Currently, there are no chondroprotective therapies used to treat osteoarthritis which mitigate 14
the loss of GAGs or effectively restore the collagen network. Cooper et al., recently describe one such 15
potential therapy where an interpenetrating network (IPN) is formed in situ throughout the articular 16
cartilage tissue (Fig. 1)13. This novel polymeric cartilage supplement is composed of the hydrophilic 17
zwitterionic monomer 2-methacryloyloxyethyl phosphorylcholine (MPC) for hydration and the 18
crosslinker ethylene glycol dimethacrylate (EGDMA) for network formation.
19 20
M AN US CR IP T
AC CE PT ED
It reconstitutes the hydrophilic properties of the extracellular matrix and mimics GAGs. In treated 21
cartilage, the IPN reinforces the compressive properties, decreases coefficient of friction and improves 22
wear resistance13,14. However, the specific mechanism(s) behind the ability of the IPN to reconstitute 23
the functional properties of the cartilage requires further elucidation.
24
To understand the origin of the improved mechanical properties of cartilage after treatment with 25
the IPN, we are using finite element (FE) modeling. FE modeling is an established method to 26
distinguish and separate functional effects of the individual constituents from the mechanical behavior 27
of native and engineered tissue15–17. Moreover, when applied to interpenetrating polymeric materials, 28
such modeling enables elucidation of one network’s mechanical effect over the other18. By modeling 29
articular cartilage as inhomogeneous and anisotropic as well as including many aspects of the real 30
tissue structure and composition, the contribution of individual structural changes to tissue function can 31
be traced using simulations, e.g., specifically how much change in dynamic stiffness is due to 32
degeneration of the collagen network.
33
Herein, we elucidate the mechanism(s) by which the IPN improves articular cartilage functional 34
properties. Specifically using FE modeling, we investigate the effect of the IPN on the fibrillar collagen 35
network, the nonfibrillar extracellular matrix (ECM), and the interstitial fluid flow.
36 37
Materials and methods
38
Osteochondral cylindrical plugs (n=22, 7 mm diameter) were cored from the femoral groove of 39
skeletally mature cows, obtained from a local slaughterhouse. During coring, tissue was irrigated with 40
0.9% saline. Only healthy tissues, i.e., no visible signs of superficial degradation, were harvested.
41
Throughout experimentation, plugs were stored at 4°C in 400 mOsm sodium chloride solution 42
M AN US CR IP T
AC CE PT ED
containing protease inhibitor benzamidine hydrochloride (5 mM), GIBCO Antibiotic/Antimycotic 43
(Invitrogen, Grand Island, NY), and calcium ion chelating agent ethylenediamine tetraacetic acid (5 44
mM).
45
Degradation
46
10 plugs were enzymatically degraded using chondroitinase ABC (Sigma Aldrich, St. Louis, MO) (0.1 47
U/mL in 50 mM tris(hydroxymethyl)aminomethane, 60 mM sodium acetate, 0.02% bovine serum 48
albumin, pH 8.0) at 37°C for 24 h19,20. The degraded plugs were then washed in saline to remove the 49
enzyme and degraded remains from the plug.
50
Biomechanical measurements
51
Thickness measurements for strain calculation were performed via computed tomography at voxel 52
resolution of 36 µm3 (µCT40, Scanco Medical AG, Brüttisellen, Switzerland) using an airtight sample 53
holder to maintain a humid environment to prevent tissue drying. The µCT data were converted to 54
DICOM format and thicknesses were computed using imaging software (Analyze 11.0, Mayo Clinic, 55
Rochester, MN).
56
Biomechanical measurements were conducted using a commercial testing device (BOSE 57
Electroforce 3200, BOSE Corporation, Eden Prairie, MN) with a 150 N load-cell (Honeywell, 58
Columbus, OH). Compressive testing was performed in unconfined geometry by subjecting 59
osteochondral plugs to a four-step stress-relaxation procedure using 5% strain/step at a compression 60
rate of 5 µm/s against a non-porous ultra-high-molecular-weight polyethylene platen. The sample was 61
secured in a saline-filled chamber and a pre-load (5 N) was applied to establish complete contact 62
between the plug’s articular surface and the rigid opposing platen. Following a dwell period (1000 s) to 63
allow relaxation from the pre-load, four incremental compression steps were applied with 45 minute 64
(40 w/v% IPN) and 75 min (others) dwell intervals between successive steps to allow stress 65
M AN US CR IP T
AC CE PT ED
equilibration. The compressive load and displacement data were collected at a sampling frequency of 66
10 Hz.
67
Synthesis of interpenetrating network
68
After biomechanical measurements, all plugs were incubated for 24 hrs in the dark at 25°C in 400 69
mOsm saline containing a three-component photoinitiating mixture of eosin Y (0.1 mM), 70
triethanolamine (115 mM), N-vinylpyrrolidone (94 mM) and 2-methacryloyloxyethyl 71
phosphorylcholine (MPC) at concentrations: 20 w/v% (N=8, 3 degraded), 40 w/v% (N=6, 3 degraded), 72
and 60 w/v% (N=8, 4 degraded) along with crosslinker ethylene glycol dimethacrylate (EGDMA) at 73
concentrations of 1.3 µl/ml, 2.6 µl/ml, and 3.8 µl/ml, respectively13. The plugs were removed from 74
solution, irradiated with green laser light (514 nm, 500 mW/cm2, 10 mins) to create the IPN, and rinsed 75
in saline to wash out residual non-reacted monomer. Uniaxial stress-relaxation compression tests were 76
repeated to evaluate the mechanical effect of IPN.
77
Finite element analysis
78
An axisymmetric, homogeneous, biphasic, poroviscoelastic FE model consisting of a viscoelastic 79
fibrillar collagen network and a nonfibrillar ECM specific to each plug geometry was created using 80
Abaqus 6.13 (Dassault Systèmes Simulia Corp., Providence, RI). No other information as to cartilage 81
structure and composition was included to keep the modeling analysis independent of the structural 82
analysis. Details of the FE model are presented in the supplementary material 1.
83
The FE meshes consisted of 300 linear axisymmetric pore pressure continuum elements.
84
Sensitivity of the element size was tested using 4 measurements: element drop from 4000 to 300 caused 85
<1% difference in reaction force while computation time dropped an average by 93.6% from 460s to 86
29s. The following boundary conditions were applied: the cartilage-bone interface was fixed in all 87
directions, the cartilage edges were assumed to be fully permeable (zero pore pressure) and the contact 88
M AN US CR IP T
AC CE PT ED
between the platen and cartilage was assumed to be impermeable. At the axis of symmetry, lateral 89
displacements were prevented and fluid was not allowed to penetrate this boundary.
90
In order to obtain unique material parameters from the model fits, the damping coefficient of 91
the viscoelastic fibrils (η) was set to constant 947 MPa s, Poisson’s ratio of the nonfibrillar matrix was 92
fixed to 0.48 and initial water fraction was set to constant 78%21–24. Five parameters were optimized by 93
minimizing the mean squared error between the experimental and the FE model derived stress- 94
relaxation curves using Matlab's (R2016b, The MathWorks, Inc., Natick, MA) built-in minimum search 95
algorithm (fminsearch) in combination with Abaqus (Fig. 2): initial and strain-dependent fibril network 96
modulus (Ef0
, Efε
, respectively), nonfibrillar ECM modulus (Enf), initial permeability (k0) and strain- 97
dependent permeability factor (M). Using the optimized material parameters, the 4 step stress- 98
relaxation tests for each plug were replicated using constant displacement step times (10 s) and 99
constant 2700 s relaxation time periods. The time-dependent and average contact pressures (CPress) 100
and pore pressures (POR) were extracted, and the interstitial fluid load support (IFLS) was calculated.
101
IFLS is the ratio of the load supported by the fluid pressure inside the pores, i.e. POR, to the total 102
Cpress25,26. Pre- versus post-treatment model parameters were compared using 2-tailed Student’s t test.
103
Results are presented with 95 % confidence intervals.
104 105
Results
106
The FE model derived stress-relaxation curves replicated well the experimental data (Fig. 2; R² >0.99 107
for all). IPN treatment increased contact and pore pressures, and decreased maximum IFLS (Figs. 3 &
108
4, Table 1). Significant differences in the optimized model parameters (P<0.05) from pre-treatment to 109
post-treatment were seen in both normal and degraded samples (Fig. 5, Supplementary material 2).
110
M AN US CR IP T
AC CE PT ED
With the 60 w/v% IPN treatment, the relative changes in model parameters were larger for the 111
degenerated chondroitinase treated samples compared to healthy non-treated.
112
For the cartilage samples treated with the 20, 40, and 60 w/v% IPN, nonfibrillar matrix stiffness 113
Enf increased by 102% (P=0.003), 100% (P=0.00005), and 222 % (P=0.0003), and the initial 114
permeability k0 decreased by 33% (P=0.03), 33% (P=0.001), and 51% (P=0.001), respectively. The 115
strain-dependent permeability factor M increased by 56% (P=0.0004) and 187% (P=0.007) for the 40 116
and 60 w/v% groups respectively. Treatment with the IPN did not statistically significantly alter the 117
fibrillar network parameters Ef0
and Efε
. 118
In FE simulations, IPN treatment significantly increased contact pressures 76% (P=0.0006), 119
70% (P=0.000004), and 140% (P=0.00008) for the 20, 40, and 60 w/v% treatment groups, respectively 120
(Table 1). Also, average pore pressure increased significantly 42% (P=0.005), 58% (P=0.005), and 121
210% (P=0.03) for the 20, 40, and 60 w/v% treatment groups, respectively. IPN treatment decreased 122
significantly the maximum IFLS after displacement steps: the maximum IFLS was significantly lower 123
by 2% (P=0.04), 5% (P=0.000006), and 10% (P=0.004), respectively, after the 20, 40, and 60 w/v%
124
IPN treatment. The Average IFLS value did not significantly change after any treatment for the entire 125
measurement period.
126
Discussion
127
The objective of this study is to understanding how the IPN affects specific mechanical properties of 128
articular cartilage and to propose a mechanism of action. For this, FE modeling is used to deconvolute 129
the effects of the treatment on the tissue’s solid phases of fibrillar matrix and nonfibrillar ECM, and on 130
fluid phase. IPN treatment increases the nonfibrillar stiffness Enf as a function of IPN concentration.
131
This was anticipated as Enf is a major contributor to the equilibrium stiffness in the model, and the 132
M AN US CR IP T
AC CE PT ED
elastic equilibrium stiffness had been earlier shown to increase after the IPN treatment . The increase 133
in Enf is significant for all three IPN treatment concentrations (20, 40, and 60 w/v%). Similarly, the IPN 134
also affects the fluid flow for all IPN concentrations; cartilage initial permeability k0 decreases and the 135
strain-dependency factor M increases with increasing IPN w/v%. Lower permeability allows less fluid 136
flow, and high M causes more obstruction as a function of compression, thus prolonging the relaxation 137
periods in the stress-curves (Fig. 3). These observed mechanical effects are a consequence of the 138
chemical structure of the interpenetrating polymer being zwitterionic and hydrophilic. We propose that 139
the IPN functionally mimics GAGs by imparting fixed charges that attract and retard the motion of 140
water molecules, and by decreasing the effective porosity of the tissue by sterically occupying pores 141
inside the tissue.
142
Treatment of cartilage with the IPN does not affect fibrillar stiffness. This result suggests two 143
conclusions. First, the IPN does not possess sufficient tensile properties to reinforce the fibrillar 144
collagen network; indeed, single-network hydrogels of poly(MPC) are weaker gels than cartilage and 145
do not possess strong intramolecular attractive interactions. These materials typically require tensile 146
reinforcement with a second interpenetrating hydrogel to provide the cumulative double network 147
material for biomedical applications in corneal implantation27,28, contact lenses29, and vascular 148
devices30. Second, the IPN acts as a separate secondary network within the tissue volume and is not 149
bound to the cartilage solid matrix. A caveat to this interpretation is that while unconfined mechanical 150
testing is sensitive to identifying the predominantly GAG-defined equilibrium properties, indentation 151
testing is more sensitivity at revealing the changes in the predominantly collagen-driven mechanical 152
properties of the superficial cartilage zone6,11,31. 153
Studies are ongoing to further understand the mechanism of mechanical reinforcement and the 154
role of IPN physicochemical properties on performance. Cooper et al., using Fourier Transform 155
M AN US CR IP T
AC CE PT ED
Infrared Imaging (FTIRI), fluorescence microscopy and T2-weighted MRI, showed the IPN to form 156
throughout the cartilage tissue13. When IPN treated cartilage extracellular tissue was digested with 157
papain, a clear and cohesive homogeneous hydrogel was present. If the gel was present only as a 158
discontiguous phase, then it would not have yielded a cohesive hydrogel upon this treatment but rather 159
would crumble into many small gel particulates. The particular structure of the polymer was designed 160
to mimic the GAG’s via the coordination of water molecules to the MPC. It will be fundamentally 161
important to further examine the interaction of the hydrogel with the solid matrix in greater detail, to 162
determine the structure-function relationships of an isolated hydrogel, and to relate the 163
physicochemical properties and chemical composition to function. These studies will be the subject of 164
a separate future study.
165
IPN treatment stiffens the tissue and produces higher contact pressures during measurements 166
(Fig. 3, Fig. 4, Table 1). Average contact and pore pressure values, derived from the models, are 167
significantly higher post-treatment in all IPN concentrations. However, the effect to the maximum IFLS 168
is the opposite. This is a result of the larger relative increase in the nonfibrillar matrix stiffness 169
compared to the decrease in the permeability, increasing the load support of the solid matrix. As the 170
decreased permeability prolongs the pressurization of the interstitial fluid, no change in overall average 171
IFLS is observed.
172
During rapid loading, the fibrillar network dominates the dynamic stiffness, relative to the 173
nonfibrillar network24,32,33. Under these conditions, the tensile stiffness of the network resists the fluid 174
pressurization. The relatively slow strain-rate in the present study (ca. 0.3% of tissue thickness / s, 175
compared with potential strain-rates in indentation testing of ca. 100% of tissue thickness / s) may have 176
not caused enough pressurization and tension in fibrillar network so that the fibrillar model parameters 177
could have been uniquely determined. Based on the outcomes from our modeling, in the current 178
M AN US CR IP T
AC CE PT ED
mechanical tests the nonfibrillar stiffness and fluid flow are the main contributors to the dynamic 179
response peak forces. Elastic dynamic moduli (initial E0 and strain-dependent Eε, Edyn = Eεε + E0)34, 180
calculated for comparison from the mechanical data, are significantly increased in the 40 and 60 w/v%
181
treatment groups. As no change is observed in the fibrillar moduli of the model, the ECM is the major 182
contributor to the dynamic responses during measurements. On the other hand, IFLS does approach 183
close to 100% after the first displacement steps representing normal pressurization26,35. 184
The nonfibrillar stiffness of the chondroitinase treated samples is significantly lower while the 185
effects of the degradation to fluid flow are ambiguous (Supplementary data). Enf values are about the 186
half of the non-treated samples. No distinct differences are observed in fibrillar moduli. This was 187
expectable as chondroitinase is known to mainly degrade only the GAGs22. Compared to the 24 h 188
chondroitinase incubation herein, Korhonen et al.22 used 44 h digestion time and saw clear changes also 189
in permeability. It is likely that a longer chondroitinase treatment time would afford a more dramatic 190
degenerative effect, and that the degeneration caused by the current incubation protocol compares to 191
early OA.
192
In this study, structurally homogeneous models are used between samples to ensure that the 193
structural properties would not influence the functional results. Furthermore, depth-dependent 194
properties are not implemented. Meng et al.36 reported that using a similar homogeneous fibrillar 195
structure inside a model, such as used here, produces comparable IFLS overall results compared to 196
depth-dependent arcade like structure. We did not measure structural properties for the samples so 197
estimating and implementing sample specific or depth-dependent collagen orientation, collagen content 198
and/or GAG content in the models would have led to deceptive estimations also making the modeling 199
analysis more dependent on the structure. For the analysis of the mechanical response of the IPN solely 200
based on tissue structure and composition, implementation of the collagen and GAG distributions and 201
M AN US CR IP T
AC CE PT ED
amounts would become necessary. The possible change in water content in the samples was also not 202
studied. Sensitivity tests showed the initial void ratio to not play a significant role on the results.
203
Achieved parameter values represent similar values to those measured earlier using fibril- 204
reinforced poroviscoelastic model for bovine tibial cartilage21. Based on the measured force curves, 205
even with the 60 w/v% IPN treatment, cartilage does not lose its biphasic time-dependent viscoelastic 206
properties and functions naturally. Importantly, the highest mechanical parameters are below those 207
measured for the naturally stiffer patella21,23,37. We acknowledge the limitation of using a rigid material 208
to describe the bone in the model. Venäläinen et al. reported that using a solid elastic material 209
describing bone produced lower contact and pore pressures, compared to rigid material38. However, 210
maximum contact pressure in their results was less than 5% lower using the elastic compared to a rigid 211
one. Based on these results, it is unlikely that using an elastic bone model here would have changed the 212
overall conclusions.
213
For the models of contact and pore pressure analysis, we used constant relative velocity instead 214
of the absolute velocity as in the experiments. This enabled us to determine identical pressurization 215
periods for each sample. The resulting 10s step displacement time was faster than the experimentally 216
produced average 14.5 (13.7-15.3) s. The resulting increase in pressurization created greater differences 217
between the groups, while not significantly changing the mechanics of the test from the original 218
experimental measurement. On average, the faster relative velocity produced 6.3% higher maximum 219
IFLS values compared to the experimental protocol. For the 40 w/v% group that originally used the 220
same 2700s relaxation time in experiments, using the faster displacement velocity in the models did not 221
change the average IFLS values of the groups.
222
The results are of clinical relevance as potential therapies to mitigate or restore GAG depletion 223
or its effect on tissue properties are of significant interest. GAG loss and increased permeability of the 224
M AN US CR IP T
AC CE PT ED
articular cartilage are indicators of early OA . Higher permeability increases fluid flow and 225
decreases interstitial fluid pressure, contributing to decreased stiffness and faster tissue relaxation.
226
These alterations accelerate the susceptibility of cartilage to further degeneration and decreased 227
functional performance. Treatment with the 60 w/v% IPN produces a clearly greater effect on the 228
chondroitinase degenerated (OA) compared to non-degraded (healthy) samples (Table 1, 229
Supplementary material 2)13. In degenerated and control samples, the contact pressure increases by 230
201% and 93%, pore pressure increases by 288% and 77%, Enf increases by 280% and 186%, and k0
231
decreases by 67% and 38%, respectively. The increase in M and the decrease in maximum IFLS are 232
significant only for the degenerated samples. These results suggest that the more degraded cartilage 233
regions will receive the greatest benefit while the healthier regions will receive a small benefit upon a 234
single IPN treatment. Overall, this capability to homogenize the properties of the joint may be 235
beneficial, i.e., it would bring the properties of degraded areas up to the same level as the healthy areas 236
and not generate adjacent regions with significantly different mechanical properties and gradients. This 237
approach is in contrast to most scaffold-based cartilage replacement strategies were a single defect is 238
site-specifically repaired with resultant mismatched mechanical properties at the tissue-scaffold 239
interface and surrounding tissue areas39,40. 240
From a biomechanical perspective, treatment with the IPN reconstitutes cartilage mechanical 241
properties primarily by augmenting the hydrophilic ECM, thereby decreasing tissue permeability while 242
at the same time still maintaining the viscoelasticity of the tissue. From a clinical perspective, treatment 243
of the cartilage with the IPN stabilizes the degradative changes occurring during early and mid-stage 244
OA and confers normal mechanical function to the tissue. Continued investigation into material based 245
treatment strategies for OA, such as the one described herein, will provide new insights into 246
M AN US CR IP T
AC CE PT ED
mechanisms of action as well as will spur the development of new material solutions where mechanics 247
play a central role.
248
Acknowledgements
249
Finnish Cultural Foundation (Mäkelä JTA), National Science Foundation Graduate Research 250
Fellowship Program (Cooper BG; NSF DGE-1247312). Petri K. Tanska, PhD, and Mika E. Mononen, 251
PhD, (Department of Applied Physics, University of Eastern Finland, Kuopio, Finland) are 252
acknowledged for providing consultancy in computational methods.
253
Contributions
254
All authors contributed to the conception and design of the study, the data acquisition, analysis, and 255
interpretation of the results. The manuscript was drafted, revised and finally approved by all authors.
256
Mäkelä JTA (jmakela@bidmc.harvard.edu) takes responsibility for the integrity of the work.
257
Role of the funding source
258
Funding sources did not have any involvement in the study design, collection, analysis and 259
interpretation of data, in the writing of the manuscript or in the decision to submit the manuscript for 260
publication.
261
Competing interests
262
Authors have no competing interests.
263
Bibliography
264
M AN US CR IP T
AC CE PT ED
1. Mow, Van C; Zhu, Wenbo; Ratcliffe A. Structure and Function of Articular Cartilage and 265
Meniscus. In: Basic Orthopaedic Biomechanics. Raven Press, Ltd, New York, 1991; 1991.
266
2. Sophia Fox AJ, Bedi A, Rodeo SA. The Basic Science of Articular Cartilage: Structure, 267
Composition, and Function. Sports Health. 2009;1(6):461-468. doi:10.1177/1941738109350438.
268
3. Buckwalter JA, Mankin HJ. Articular cartilage: degeneration and osteoarthritis, repair, 269
regeneration, and transplantation. Instr Course Lect. 1998;47:487-504. doi:9571450.
270
4. Buckwalter JA, Mankin HJ. Articular Cartilage, Part II: Degeneration and osteoarthrosis, repair, 271
regeneration, and transplantation. JBone Jt SurgAm. 1997;79:612-632.
272
5. Mäkelä JTA, Rezaeian ZS, Mikkonen S, et al. Site-dependent changes in structure and function 273
of lapine articular cartilage 4 weeks after anterior cruciate ligament transection. Osteoarthr 274
Cartil. 2014;22(6):869-878. doi:10.1016/j.joca.2014.04.010.
275
6. Saarakkala S, Julkunen P, Kiviranta P, Mäkitalo J, Jurvelin JS, Korhonen RK. Depth-wise 276
progression of osteoarthritis in human articular cartilage: investigation of composition, structure 277
and biomechanics. Osteoarthr Cartil. 2010;18(1):73-81. doi:10.1016/j.joca.2009.08.003.
278
7. Mankin HJ, Lippiello L, A Z. Biochemical and metabolic cartilage abnormalities in articular 279
cartilage. J Bone Jt Surg. 1971;53(3):533-537. http://www.ncbi.nlm.nih.gov/pubmed/4246573.
280
8. Stoop R, Buma P, Van Der Kraan PM, et al. Type II collagen degradation in articular cartilage 281
fibrillation after anterior cruciate ligament transection in rats. Osteoarthr Cartil. 2001;9(4):308- 282
315. doi:10.1053/joca.2000.0390.
283
9. Sah RL, Yang AS, Chen AC, et al. Physical properties of rabbit articular cartilage after 284
transection of the anterior cruciate ligament. J Orthop Res. 1997;15(2):197-203.
285
doi:10.1002/jor.1100150207.
286
10. Turunen SM, Han SK, Herzog W, Korhonen RK. Cell deformation behavior in mechanically 287
loaded rabbit articular cartilage 4 weeks after anterior cruciate ligament transection. Osteoarthr 288
Cartil. 2013;21(3):505-513. doi:10.1016/j.joca.2012.12.001.
289
11. Guilak F, Ratcliffe A, Lane N, Rosenwasser MP, Mow VC. Mechanical and biochemical changes 290
in the superficial zone of articular cartilage in canine experimental osteoarthritis. J Orthop Res.
291
1994;12(4):474-484. doi:10.1002/jor.1100120404.
292
12. Panula HE, Hyttinen MM, Arokoski JP, et al. Articular cartilage superficial zone collagen 293
birefringence reduced and cartilage thickness increased before surface fibrillation in 294
experimental osteoarthritis. Ann Rheum Dis. 1998;57(4):237-245. doi:10.1136/ard.57.4.237.
295
13. Cooper BG, Stewart RC, Burstein D, Snyder BD, Grinstaff MW. A Tissue-Penetrating Double 296
Network Restores the Mechanical Properties of Degenerated Articular Cartilage. Angew Chemie 297
- Int Ed. 2016;55(13):4226-4230. doi:10.1002/anie.201511767.
298
M AN US CR IP T
AC CE PT ED
14. Cooper BG, Lawson T, Snyder BD, Grinstaff MW. Reinforcement of articular cartilage with a 299
tissue-interpenetrating polymer network reduces friction and modulates interstitial fluid load 300
support. Osteoarthr Cartil. 2017:1-7. doi:10.1016/j.joca.2017.03.001.
301
15. Julkunen P, Wilson W, Isaksson H, Jurvelin JS, Herzog W, Korhonen RK. A review of the 302
combination of experimental measurements and fibril-reinforced modeling for investigation of 303
articular cartilage and chondrocyte response to loading. Comput Math Methods Med.
304
2013;2013:23 pages. doi:10.1155/2013/326150.
305
16. Meloni GR, Stoeckl BD, Fisher MB, Dodge GR, Mauck RL. Biphasic Finite Element Modeling 306
Reconciles Mechanical Properties of Engineered Cartilage Constructs Derived from Different 307
Testing Modalities. 2017;0(0):4-6. doi:10.1089/ten.tea.2016.0191.
308
17. Mäkelä JTA, Han S-KK, Herzog W, et al. Very early osteoarthritis changes sensitively fluid flow 309
properties of articular cartilage. J Biomech. 2015;48(12):3369-3376.
310
doi:10.1016/j.jbiomech.2015.06.010.
311
18. Schmidt A, Bergamini A, Kovacs G, Mazza E. Multiaxial Mechanical Characterization of 312
Interpenetrating Polymer Network Reinforced Acrylic Elastomer. Exp Mech. 2011;51(8):1421- 313
1433. doi:10.1007/s11340-010-9462-z.
314
19. Basalo IM, Chahine NO, Kaplun M, Chen FH, Hung CT, Ateshian GA. Chondroitin sulfate 315
reduces the friction coefficient of articular cartilage. J Biomech. 2007;40(8):1847-1854.
316
http://www.ncbi.nlm.nih.gov/pubmed/17084404.
317
20. Basalo IM, Raj D, Krishnan R, Chen FH, Hung CT, Ateshian GA. Effects of enzymatic 318
degradation on the frictional response of articular cartilage in stress relaxation. J Biomech.
319
2005;38(6):1343-1349. http://www.ncbi.nlm.nih.gov/pubmed/15863119.
320
21. Julkunen P, Kiviranta P, Wilson W, Jurvelin JS, Korhonen RK. Characterization of articular 321
cartilage by combining microscopic analysis with a fibril-reinforced finite-element model. J 322
Biomech. 2007;40(8):1862-1870. doi:10.1016/j.jbiomech.2006.07.026.
323
22. Korhonen RK, Laasanen MS, Töyräs J, Lappalainen R, Helminen HJ, Jurvelin JS. Fibril 324
reinforced poroelastic model predicts specifically mechanical behavior of normal, proteoglycan 325
depleted and collagen degraded articular cartilage. J Biomech. 2003;36(9):1373-1379.
326
doi:10.1016/S0021-9290(03)00069-1.
327
23. Wilson W, Van Donkelaar CC, Van Rietbergen B, Ito K, Huiskes R. Stresses in the local collagen 328
network of articular cartilage: A poroviscoelastic fibril-reinforced finite element study. J 329
Biomech. 2004;37(3):357-366. doi:10.1016/S0021-9290(03)00267-7.
330
24. Mäkelä JTA, Korhonen RK. Highly nonlinear stress-relaxation response of articular cartilage in 331
indentation: Importance of collagen nonlinearity. J Biomech. 2016;49(9):1734-1741.
332
doi:10.1016/j.jbiomech.2016.04.002.
333
M AN US CR IP T
AC CE PT ED
25. Ateshian GA. The role of interstitial fluid pressurization in articular cartilage lubrication. J 334
Biomech. 2009;42(9):1163-1176. doi:10.1016/j.jbiomech.2009.04.040.
335
26. Krishnan R, Kopacz M, Ateshian G a. Experimental verification of the role of interstitial fluid 336
pressurization in cartilage lubrication. J Orthop Res Off Publ Orthop Res Soc. 2004;22(3):565- 337
570. doi:10.1016/j.orthres.2003.07.002.
338
27. Liu W, Deng C, McLaughlin CR, et al. Collagen-phosphorylcholine interpenetrating network 339
hydrogels as corneal substitutes. Biomaterials. 2009;30(8):1551-1559.
340
doi:10.1016/j.biomaterials.2008.11.022.
341
28. Hackett JM, Lagali N, Merrett K, et al. Biosynthetic corneal implants for replacement of 342
pathologic corneal tissue: Performance in a controlled rabbit alkali burn model. Investig 343
Ophthalmol Vis Sci. 2011;52(2):651-657. doi:10.1167/iovs.10-5224.
344
29. Shimizu T, Goda T, Minoura N, Takai M, Ishihara K. Super-hydrophilic silicone hydrogels with 345
interpenetrating poly(2-methacryloyloxyethyl phosphorylcholine) networks. Biomaterials.
346
2010;31(12):3274-3280. doi:10.1016/j.biomaterials.2010.01.026.
347
30. Morimoto N, Iwasaki Y, Nakabayashi N, Ishihara K. Physical properties and blood compatibility 348
of surface-modified segmented polyurethane by semi-interpenetrating polymer networks with a 349
phospholipid polymer. Biomaterials. 2002;23(24):4881-4887. doi:10.1016/S0142- 350
9612(02)00246-6.
351
31. Korhonen RK, Wong M, Arokoski J, et al. Importance of the superficial tissue layer for the 352
indentation stiffness of articular cartilage. Med Eng Phys. 2002;24(2):99-108.
353
doi:10.1016/S1350-4533(01)00123-0.
354
32. Párraga Quiroga JM, Wilson W, Ito K, van Donkelaar CC. The effect of loading rate on the 355
development of early damage in articular cartilage. Biomech Model Mechanobiol.
356
2017;16(1):263-273. doi:10.1007/s10237-016-0815-0.
357
33. Quiroga JMP, Wilson W, Ito K, van Donkelaar CC. Relative contribution of articular cartilage’s 358
constitutive components to load support depending on strain rate. Biomech Model Mechanobiol.
359
2017;16(1):151-158. doi:10.1007/s10237-016-0807-0.
360
34. Huttu MRJ, Puhakka J, Mäkelä JTA, et al. Cell–tissue interactions in osteoarthritic human hip 361
joint articular cartilage. Connect Tissue Res. 2014;8207(16):1-10.
362
doi:10.3109/03008207.2014.912645.
363
35. Basalo IM, Mauck RL, Kelly T-AN, et al. Cartilage Interstitial Fluid Load Support in 364
Unconfined Compression Following Enzymatic Digestion. J Biomech Eng. 2004;126(6):779.
365
doi:10.1115/1.1824123.
366
M AN US CR IP T
AC CE PT ED
36. Meng Q, An S, Damion RA, et al. The effect of collagen fibril orientation on the biphasic 367
mechanics of articular cartilage. J Mech Behav Biomed Mater. 2017;65:439-453.
368
doi:10.1016/j.jmbbm.2016.09.001.
369
37. Wilson W, van Donkelaar CC, van Rietbergen B, Ito K, Huiskes R. Erratum to “Stresses in the 370
local collagen network of articular cartilage: a poroviscoelastic fibril-reinforced finite element 371
study” [Journal of Biomechanics 37 (2004) 357–366] and “A fibril-reinforced poroviscoelastic 372
swelling model for articular cartil. J Biomech. 2005;38(10):2138-2140.
373
doi:10.1016/j.jbiomech.2005.04.024.
374
38. Venäläinen MS, Mononen ME, Jurvelin JS, Töyräs J, Virén T, Korhonen RK. Importance of 375
Material Properties and Porosity of Bone on Mechanical Response of Articular Cartilage in 376
Human Knee Joint—A Two-Dimensional Finite Element Study. J Biomech Eng.
377
2014;136(12):121005. doi:10.1115/1.4028801.
378
39. Temenoff JS, Mikos AG. Review: tissue engineering for regeneration of articular cartilage.
379
Biomaterials. 2000;21(5):431-440. doi:10.1016/S0142-9612(99)00213-6.
380
40. Kuo CK, Li W-J, Mauck RL, Tuan RS. Cartilage tissue engineering: its potential and uses. Curr 381
Opin Rheumatol. 2006;18(1):64-73. doi:10.1097/01.bor.0000198005.88568.df.
382
M AN US CR IP T
AC CE PT ED
M AN US CR IP T
AC CE PT ED
Fig. 2: Experimental and FEM derived stress–relaxation for representative sample Pre- and Post- treatment 60w/v% IPN.
M AN US CR IP T
AC CE PT ED
Fig. 3: Pretreatment (Pre) and post-treatment (Post) average contact pressure (CPress) and pore pressure (POR) for the sample in Fig. 2.
M AN US CR IP T
AC CE PT ED
Fig. 4: Pretreatment (Pre) and post-treatment (Post) axisymmetric pore pressure (MPa) distributions for the sample in Fig. 2 at the beginning of relaxation period after 5%, 10%, 15%, and 20% displacements.
M AN US CR IP T
AC CE PT ED
Fig. 5: Mean values of pre-treatment (Pre) and post-treatment (Post) model parameters with 95% confidence intervals and statistical
M AN US CR IP T
AC CE PT ED
Table 1: Mean values of preoperative (Pre) versus postoperative (Post) model parameters with 95% confidence intervals with healthy and chondroitinase degraded groups separated. Statistical differences in bold. CPress is the contact pressure, POR the pore pressure, and IFLS
the interstitial fluid load support (Average = average over time, Maximum = average maximum IFLS after displacement).