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