Fiber based scaffolds in connective tissue engineering : using the architecture of woven scaffolds to influence and cont...

MISSING IMAGE

Material Information

Title:
Fiber based scaffolds in connective tissue engineering : using the architecture of woven scaffolds to influence and control the formation of organized tissues
Physical Description:
xiv, 150 leaves : ill. ; 29 cm.
Language:
English
Creator:
Seegert, Charles Alan, 1971-
Publication Date:

Subjects

Subjects / Keywords:
Biomedical Engineering thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Biomedical Engineering -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2002.
Bibliography:
Includes bibliographical references (leaves 138-149).
Additional Physical Form:
Also available online.
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by Charles Alan Seegert.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 029235016
oclc - 51024145
System ID:
AA00025801:00001

Full Text










FIBER BASED SCAFFOLDS IN CONNECTIVE
TISSUE ENGINEERING:
USING THE ARCHITECTURE OF WOVEN SCAFFOLDS
TO INFLUENCE AND CONTROL THE FORMATION
OF ORGANIZED TISSUES












By

CHARLES ALAN SEEGERT












A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2002


























Copyright 2002

by

CHARLES ALAN SEEGERT





























To the Pinnacle














ACKNOWLEDGMENTS

I would like to thank everyone who made it possible for me to perform the research in this dissertation. In particular I would like to acknowledge Dr. Colin Sumners and the people in his lab who provided me with the tissues that I used to develop my cell cultures.

I would especially like to thank Dr. Brennan, my major adviser, who allowed me the time and provided me with the resources I needed to develop my ideas.

































iv














TABLE OF CONTENTS
pne

ACKN OW LED GM EN TS ................................................................................................. iv

LIST OF TABLES ........................................................................................................... viii

LIST O F FIGU RES ........................................................................................................... ix

ABSTRA CT ..................................................................................................................... xiii

CHAPTERS

I INTRODU CTION ............................................................................................................ I

Autogeneic Bone Grafting .............................................................................................. 2
Allogeneic Bone Grafts ................................................................................................... 3
Synthetic Bone Replacement Materials and Xenogeneic Bone Grafts ........................... 5
Problem to be Approached .............................................................................................. 6
Tissue Engineering .......................................................................................................... 8
The Case for Hierarchical Organization of Fiber Constructs ......................................... 9


2 BA CK GRO UN D ............................................................................................................ 13

Bone Form ation in Utero .............................................................................................. 13
M SC Differentiation ..................................................................................................... 15
Bone Developm ent and Bone A natom y ....................................................................... 21
Fracture Healing and Ectopic Bone Form ation ............................................................ 25
Porosity and Diffusion Properties of Implant Materials ............................................... 25
Rem odeling and Cellular Orientation ........................................................................... 28
Contact Guidance .......................................................................................................... 29
Architecture of Fiber Based Scaffolds: Development of Cell-Based Tension ............. 33
Autocrine/Paracrine Considerations in Scaffold Design .............................................. 37


3 SIN GLE FIBER STUD IES ............................................................................................. 42

Introduction ................................................................................................................... 42
M aterials ....................................................................................................................... 44
M ethods ......................................................................................................................... 45
Single Fiber Scaffolds ............................................................................................ 45


v








Cell Culture ............................................................................................................ 45
M SC Seeding ......................................................................................................... 46
SEM Imaging (Nylon sutures) ............................................................................... 46
Light M icroscopy (M axon Sutures) ....................................................................... 47
Results ........................................................................................................................... 47
SEM Studies ........................................................................................................... 47
4-0 nylon sutures .............................................................................................. 47
9-0 nylon sutures .............................................................................................. 49
10-0 nylon sutures ............................................................................................ 50
M axon sutures .................................................................................................. 51
Light M icroscopy Studies of M axon ..................................................................... 53
Nuclear form factor .......................................................................................... 55
Nuclear angle analysis ..................................................................................... 58
Discussion ..................................................................................................................... 61
Conclusions ................................................................................................................... 62


4 CELLULAR BRIDGING PHENOM ENA ..................................................................... 64

Introduction ................................................................................................................... 64
M aterials ....................................................................................................................... 66
M ethods ......................................................................................................................... 66
Culture M ethods ..................................................................................................... 66
SEM Imaging ......................................................................................................... 67
Light M icroscopy Studies ...................................................................................... 67
Von Kossa Staining ................................................................................................ 67
Proliferation Studies ............................................................................................... 68
Results ........................................................................................................................... 69
Light Microscopy of Bioactive Glass Fibers After 6 Days in Culture ................... 69
Scanning Electron M icroscopy .............................................................................. 71
Bioactive glass fibers after 6 days in culture ................................................... 71
Polym er fibers .................................................................................................. 72
Light microscopy of Bridge Development on Polystyrene Culture Flask ............. 73
BiTdge Development on Stainless Steel Screens (Light Microscopy and SEM) ... 77 Proliferation Studies of RM SCs on Bioactive Glass .............................................. 81
Discussion ..................................................................................................................... 83
Conclusions ................................................................................................................... 92


5 M ULTI-FIBER STUDIES .............................................................................................. 94

Introduction ................................................................................................................... 94
M aterials ....................................................................................................................... 97
M ethods ......................................................................................................................... 97
Cell Culture ............................................................................................................ 97
Construct Preparation ............................................................................................. 97
M axon Construct Sterilization ............................................................................... 99


vi








Stainless Steel Construct Preparation ................................................................... 100
Transmission Electron M icroscopy ...................................................................... 101
Statistical Analysis ............................................................................................... 101
Results ......................................................................................................................... 104
M axonTm Bridging ................................................................................................ 104
M axonTm bridging day 3 ................................................................................ 104
M axonTm bridging day 6 ................................................................................ 104
M axonTm bridging day 9 ................................................................................ 105
M axonTm bridging day 10 .............................................................................. 105
M axonTm bridging day I I .............................................................................. 105
M ulti-Layer Construct Bridging .......................................................................... 107
Cell Angle and Spacing Distance ......................................................................... 109
Stainless Steel Bridging ....................................................................................... 109
Stainless steel bridging day 3 ......................................................................... 109
Stainless steel bridging day 6 ......................................................................... 110
Stainless steel bridging day 8 ......................................................................... 110
Stainless steel bridging day 10 (mineralization analysis) .............................. 110
Transm ission Electron M icroscopy ...................................................................... III
Transmission electron microscopy of 7-0 single fiber ................................... III
Transmission electron microscopy of 7-0 25 jim spaced parallel array ........ 112
Transmission electron microscopy of contracted 7-0 25 PM parallel
array ......................................................................................................... 114
Discussion ................................................................................................................... 117
Conclusions ................................................................................................................. 123


6 CON CLUSION S AND FUTURE W ORK ................................................................... 125

APPENDIX QUANTITATIVE CONTACT GUIDANCE ANALYSIS METHODS .... 132

Conceptual and Illustrated Review of Nuclear Form Factor (NFF) ........................... 132
Extension of NFF Correction Concept to Nuclear Angle Measurements ................... 134
M athem atical Derivation ............................................................................................ 135


LIST OF REFERENCES ................................................................................................. 138

BIOGRAPHICAL SKETCH ........................................................................................... 150











vii















LIST OF TABLES

Table page

1-1. List of Design Requirements for More Ideal Bone Replacement Material .................. 6

5-1. Summary of Statistical Results from the Mineralization Study Performed on
Stainless Steel Screens ......................................................................................... III






































viii















LIST OF FIGURES

Figure page

1-1 An example of a femoral reconstruction using a segment of allograft bone ....... 5 1-2 Schematic representation of diaphyseal bone replacement...................... 7

1-3 A diagram of the general orientation of lamellae in a segment of bone............. 9

1-4 Schematic representation of collagen fibril orientation within a segment of bone ... 10 1-5 General concept of the hierarchical level of woven fiber based scaffolds ........ 11 2-1 Schematic of the cell cycle ............................................................. 14

2-2 MSC developmental sequence leading to bone producing cells ................... 17

2-3 Lines of force seen in proximal portions of the femur ............................... 21

2-4 Examples of lamellar orientation within a section of bone ......................... 23

2-5 Fibrillar orientation within bone ....................................................... 23

2-6 Rotated plywood model................................................................. 24

2-7 Schematic representation of ectopic bone formation around implanted,
demineralized bone chips............................................................... 26

2-8 Receptor mediated adhesion of osteoblasts via fibronectin ........................ 30

2-9 Diagram of autocrine signaling ........................................................ 38

2-10 Diagram of paracrine signaling ........................................................ 40

3-1 Examples of MaxonTm single fiber constructs........................................ 45

3-2 4-0 nylon fibers with RMSCs after being cultured for 5 days ..................... 48

3-3 Examples of MSC growth on the surface of 4-0 nylon fibers ...................... 49

3-4 Examples of RMSC growth on the surface of 9-0 nylon fibers .................... 50




ix








3-5 Examples of RMSC growth on the surface of 9-0 nylon fibers ............................. 51

3-6 SEM images of 10-0 nylon sutures at various magnifications ............................ 52

3-7 5-0 Maxon sutures exhibiting RMSC adhesion and growth ................................ 53

3-8 6-0 Maxon sutures exhibiting RMSC adhesion and growth . ............................... 54

3-9 7-0 Maxon sutures exhibiting RMSC adhesion and growth ................................ 54

3-9 Continued. 7-0 Maxon sutures exhibiting RMSC adhesion and growth ............. 55

3-10 Example of nuclear form factor measurements ................................................... 55

3-11 Graph showing the independent influence of fiber diameter on NFF .................. 56

3-12 Graph showing the independent influence of time on NFF ................................... 57

3-13 Graph showing the effects of time on NFF for all diameters studied .................. 57

3-14 Example of nuclear angle measurements ............................................................... 59

3-15 Graph showing the effects of diameter independent of time ................................ 59

3-16 Graph showing the effect of time in culture on nuclear orientation as measured via
nuclear angle . ....................................................................................................... 60

3-17 Graph showing the effects of time and diameter together . .................................. 60

4-1 Multilayering on adjacent fibers leading to interactions due to their proximity .... 65

4-2 Example of bioactive glass fiber constructs used for the proliferation study ....... 68

4.3 Montage of light micrographs portraying the cellular interaction with bioactive
glass fibers placed in standard culture well ......................................................... 70

44 Light micrograph of area without fibers in the same culture well as that shown in
Figure 4-3 ............................................................................................................. 70

4-5 Examples of cellular growth on fiber placed in the RMSC system ..................... 71

4-6 Examples of unicellular bridging on bioactive glass over distances of -70 Pm. ..72

4-7 Examples of unicellular bridging between nylon fibers ...................................... 73

4-8 Bridging between Maxon fibers ............................................................................. 73

4-9 Initial RM SC elongation ...................................................................................... 74



x









4-9 continued. Initial RMSC unicellular bridging........................................ 75

4-10 Multicellular bridge at 7 days in the RMSC culture.................................. 76

4-11 Multicellular bridge at 8 days in the RMSC culture.................................. 76

4-12 Multicellular bridge at 10 days in the RMSC culture ................................ 77

4-13 Multicellular bridge at I11 days in the RMSC culture................................ 77

4-14 Development of bridging on stainless steel screens.................................. 79

4-15 SEM of stainless steel screens after 10 days in RMSC culture .................... 80

4-16 SEM of stainless steel screens after 12 days in RMSC culture..................... 80

4-17 SEM of stainless steel screens after 15 days in RMSC culture..................... 81

4-18 SEM of stainless steel screens after 18 days in RMSC culture..................... 81

4-19 SEM of stainless steel screens after 23 days in RMSC culture..................... 82

4-20 RMSC growth curves for each fiber density.......................................... 83

4-21 Schematic generalization of multicellular bridging .................................. 85

5-1 Eye shaped bridging phenomenon caused by the weave of the screen ............ 96

5-2 Pictures of mult-fiber micromanipulator.............................................. 98

5-3 Examples of multi-fiber constructs for RMSC culture .............................. 99

5-4 Diagram of stainless steel construct preparation.................................... 100

5-5 Diagram of stainless steel statistical test............................................. 102

5-6 Examples of single fiber RMSC growth ............................................. 106

5-7 Representative micrographs of 7-0 multi-fiber parallel arrays spaced at 25 im. 107 5-8 Representative micrographs of 7-0 multi-fiber parallel arrays spaced at 55 Pm. 108 5-9 Micrographs of multi-layer parallel arrays........................................... 108

5-10 Examples of bridging between fibers of 25 pm and 55 pm spaced parallel arrays. 109 5-12 Transmission electron micrograph montage of 7-0 single fiber construct in crosssection .................................................................................. 112




xi








5-13 Transmission electron micrograph montage of 7-0 single fiber construct in longsection ................................................................................................................ 113

5-14 Transmission electron micrograph montage of 7-0 multi-fiber 25 pm spaced
parallel array in cross-section ............................................................................ 114

5-15 Transmission electron micrograph montage of contracted 7-0 multi-fiber 25 Pm
spaced parallel array in cross-section ................................................................ 115

5-16 TEM images of longitudinal section of the contracted portion of 7-0 25 Wn
spaced parallel array ........................................................................................... 116

5-17 Characteristic banding pattern of collagen fibrils .............................................. 117

5-18 Schematic representation of flat surface multilayering versus bridging ............ 122

6-1 Surface of a 5-0 maxon fiber imaged with light microscope ............................. 127

6-2 Conceptual diagram of bone replacement .......................................................... 128

6-3 A possible sequence of growth factors for release in a bone replacement system. 131

A-I Diagram of NFF measurement ............................................................................. 132

A-2 Diagram of relationship between nuclear dimensions and fiber geometry .......... 133

A-3 Diagram of trigonometric approximations used to determine nuclear dimensions. 133

A-4 Diagrammatic comparison of NFF and nuclear angle ........................................ 134

A-5 Diagrammatic relationship between NFF measurements and nuclear angle
m easurem ents ....................................................................................................... 135

A-6 Diagrammatic representation of calculated rotational correction ........................ 136

















xii















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

FIBER BASED SCAFFOLDS IN CONNECTIVE TISSUE ENGINEERING:
USING THE ARCHITECTURE OF WOVEN SCAFFOLDS
TO INFLUENCE AND CONTROL THE FORMATION OF ORGANIZED TISSUES

By

Charles Alan Seegert

August 2002



Chair: Anthony B. Brennan Ph.D. Major Department: Biomedical Engineering

The skeleton generates locomotion and provides mechanical support for the human body. It is essential in every aspect of normal living, which is most apparent when it is damaged or rendered useless by disease. The repair of the skeletal system by orthopedists is rather common, occurring millions of times each year. During the 1990s it was estimated that nearly a million cases per year were bone graft cases, which makes them one of the most common forms of transplantation in the United States today.

Orthopedic reconstruction restores ftmction, extending and increasing quality of life. Though progress has been made toward reaching this ideal, bone grafting is still fraught with shortcomings. Its weaknesses become most apparent when large skeletal defects like those seen in osteosarcomal resection are treated. Commonly these reconstructions are


xiii








performed using allografts combined with steel rods and other devices. Synthetic materials may also be used, but full skeletal incorporation never occurs and they remain inanimate. Because they are non-living, these materials accumulate fatigue and eventually fail.

Cell-based tissue-engineered replacements are a strong candidate to address these challenges. This treatment system would be composed of a synthetic material, like resorbable polymer fibers, combined with cells that are loaded onto the fibers. After implantation the system would foster development of a living and thus self-repairing replacement tissue.

This study focuses on synthetic scaffolds and how they interact with the cells loaded onto them. Established principles of contact guidance were applied to influence orientation and growth of cells on these scaffolds. Guidance of cells and their extracellular matrix (ECM) products is shown on the level of single fibers. More important, however, is the direction of cells and ECM when fibers were organized into regular 3-D structures. Cellular organization and direction far exceeded what has been seen on flat surfaces, or single fibers. The goal of this work was to use scaffold architectures to organize cells and ECM in a 3-D manner as is seen in normal tissues. The results of this study indicate the success of this concept and represent a large step toward the development of this technology.












xiv














CHAPTER 1
INTRODUCTION

In addition to providing mechanical support for the human body, the skeleton is integral in generating locomotion. Indeed the musculoskeletal system is an essential component in nearly every aspect of a person's life, a fact that is made most apparent when a portion of this system is damaged or rendered useless by disease. Unfortunately the repair of the skeletal system by orthopedists is rather common, occurring millions of times each year alone. During the 1990s it was estimated that nearly a million cases per year were bone graft cases [I], making bone grafting one of the most common forms of transplantation in the United States today.

The intent of orthopedic reconstruction is to restore fiction, thus extending and increasing the quality of life an individual experiences after injury. Clearly much progress has been made toward reaching this goal. However, despite this progress and the frequency with which bone grafting is used, this treatment method is fraught with difficulties and shortcomings. Implantation of a bone graft or a synthetic bone replacement material is performed when a void in the skeleton is created by trauma or resection for pathological treatment. Another major implementation is when injured bones fail to re-join as is seen in non-unions at a fracture site [2,3]. In these situations, emplacement of a bone graft, or a synthetic bone replacement material is then performed to bridge the gap, restoring skeletal integrity.


I







2

When reconstructing skeletal defects, three major graft materials are used: autogeneic bone, allogeneic bone and xenogeneic bone. Synthetic materials like porous hydroxyapatite and bioactive ceramics have been used in some cases, as well. A degree of success has been realized with each of these materials, but there are also significant characteristic limitations to all of them.

Autogeneic Bone Grafting

Autogeneic bone grafting is considered the gold standard and all other materials are compared to it when their efficacy is evaluated [4]. This type of bone graft is taken from a site within the patient's own body and thus it is recognized by the body when implanted elsewhere. Additionally, autograft bone is a living graft, which contains bone producing cells. Because of this feature, cells are deposited into the site where new bone growth is desired, thus stimulating a much more rapid recovery. This type of bone deposition and growth is "osteogenic," a classification that includes grafts containing phenotypically committed osteoblasts, or grafts that stimulate proliferation of committed osteoblasts [4]. Other graft materials depend on the infiltration of cells from the surrounding area; thus recovery takes longer if it occurs at all. This type of healing, which is characterized by materials providing a scaffold to direct bone growth, is called "osteoconductive" [4]. One further classification of bone graft materials that is particularly germane to this study is "osteoinductive" materials. Osteoinductive grafts lead to differentiation of mesenchymal stem cells, or osteoblastic precursors, thus causing them to become fully committed bone producing cells [4].

Though autogeneic bone is considered the best material, there are still significant drawbacks to its use. The quantity of this bone graft material available is very limited.







3

Clearly only so much bone can be removed from one part of the body in order to heal another part; anything more would only create a new defect in skeletal integrity. In addition to this, harvest sites often experience lengthy and painful recovery periods. In many cases the recovery of the harvest site takes longer than the recovery of the original injury the bone graft was used to treat [4].

Allogeneic Bone Grafts

Allogeneic bone grafts are widely used, as well. Unlike autograft bone, allograft bone is not as limited by supply. This material is taken from other human bodies, usually deceased, and then passed through a series of treatments ideally rendering it nonimmunogenic and free of pathogens. Allograft bone is used to repair small skeletal defects with much success. For this use, it is usually "morselized," or broken up into small fragments that are then packed into the fracture site. The volume of most defects treated this way is relatively small, which is related to the success of this application. Allograft bone is acellular and osteoconductive; therefore it must be infiltrated with bone producing cells in order to recover. Additionally, there are a number of growth factors that are utilized by these cells during recovery. Depths of cellular migration into the graft material, as well as the distance that growth factors can diffuse without degradation, are again quite limited [5]. Full integration of allograft bone does occur; however. These distance related factors seem to prevent integration with the body beyond relatively small distances on the order of a few centimeters [6].

In addition to treating smaller skeletal defects with morselized allograft material, large skeletal reconstructions are performed with intact, non-morselized segments of allograft bone. In situations like this the drawbacks of allograft bone become most apparent.







4

Integration by the body ensures the presence of cells and the formation of a living tissue. Living tissues are capable of self-repair and are regenerate in the presence of stresses associated with daily activity. Non-animate replacements, even though of a biological origin, do not repair themselves and normal fatigue processes are continuously at work.

Treatment of osteosarcomas often requires large resections. Historically, when this disease was encountered, a patient's limb was amputated leading to a life long handicap. In the early 1960s, however, limb salvage became the standard of care in these cases. Providing a mechanical means of support via implantation of allograft bone allows the patient to maintain the reconstructed limb and a much higher quality of life [7]. During limb reconstruction, allogeneic bone grafts are frequently augmented with metal fixation devices, like inter-medullary nails, or plates. Figure 1-1 shows an example of just such a construction using allograft bone.

Sometimes, in addition to the fixation devices, bone cements are also used, leading to large static composite structures. At best, these rigid conglomerates of organic and inorganic materials are fractionally incorporated into the patient's skeleton due to the distance limitations mentioned previously. This limited amount of repair occurs at the ends of the allograft, where it comes in contact with the living bone allowing cells to infiltrate [3]. The rest of the composite structure remains a rigid mass of dead bone and inorganic components and even the limited amount of incorporation described above only occurs about 70 % of the time [8].

This remedy is better than the alternative--no limb--but the allograft is never fully integrated and the accumulation of fatigue eventually culminates in failure of the device.







5


















Figure 1-1. An example of a femoral reconstruction using a segment of allograft bone. The white structure in the center of the image is a metal rod holding the composite in place.

This failure tends to occur in areas where a hole has been introduced into the allograft material, sometimes as soon as 1-2 years after implantation [9].

Synthetic Bone Replacement Materials and Xenogeneic Bone Grafts

Synthetic bone replacement materials and xenogeneic bone grafts, like allograft bone, are not limited by supply. Examples of this class of materials are porous hydroxyapatite, choraline hydroxyapatite and bioactive glass materials. Generally these materials are processed into bulk materials with a porous structure leaving relatively poor allowances for diffusion of nutrients. For example, porous hydroxyapatite only possesses a porous volume of approximately 30% [10]. Because of this, when it comes to integration these materials exhibit distance limitations in a way similar to allograft bone. Xenografts, or grafts derived from other species, are rarely used due to the high chances of rejection and the possibility of disease transmission between species [4].






6

Problem to be Approached

Bone grafting, particularly of large diaphyseal segments, is far from ideal. Synthetic materials can only be used to fill small defects, while large defects are repaired using allograft material. Allograft bone is useful in that it is of the correct dimensions and density and does possess some of its original mechanical strength; however, fatigue problems limit its viability over time. For smaller bone defects autograft bone is the material of choice, but it is limited in quantity and its removal often leads to a long painful recovery in the area from which it was taken. A number of synthetic materials have also been used with some limited success, but no truly suitable substitute yet exists, especially for large segmental bone defects [4].

After examining the currently used bone graft materials and methods, a number of deficiencies become apparent that must be overcome in order to develop more viable synthetic bone replacement systems. Opportunities for improvement are summarized in table 1-1.



Table I 1. List of Design Requirements for More Ideal Bone Replacement Material
-Unlimited Quantity

-No Pathogenic Transmission

-No Immunogenicity

-Osteoinductive and Osteoconductive to be Fully Incorporated over Large Volumes.
-Highly Porous to Allow Proper Diffusion and Infiltration of Cells.
-Mechanically Robust to Provide Support Until New Tissue Becomes Self Supporting.







7

Overall, a bone replacement system capable of filling large skeletal defects would be ideal. Using the replacement of a diaphyseal segment of long bone as a goal, or model system presents a somewhat simplified bone replacement problem. Figure 1-2 shows a schematic representation of this idea, which if successfully achieved would pave the way for more complex problems like joint reconstruction, or tendinous attachments. Additionally, many of the problems associated with replacement of bone in other areas of the body would be answered.





. . . . .













XX








Figure 1-2. Schematic representation of diaphyseal bone replacement. A. Damaged, or diseased segment of a long bone. B. Removal of diseased area with margins of healthy bone, followed by emplacement of a graft/ engineered synthetic replacement. C. The reconstructed bone with natural and synthetic components. D. After healing and subsequent resorption of the synthetic replacement material a new segment of living tissue persists and the patient is healed.






8

Tissue Engineering

Tissue engineering is defined as: the application of engineering disciplines to either maintain existing tissue structures, or to enable tissue growth [11]. A tissue engineered construct would solve many of the problems associated with bone grafting. Indeed there are already many studies in the literature that implant scaffold materials loaded with cells taken from the patient, or subject [12-20]. This method, where autogeneic cells are taken from a patient, loaded onto a scaffold and then implanted, is referred to as a "cell based approach" to tissue engineering.

Using this approach requires a scaffold, or support matrix that cells can adhere to and proliferate on prior to implantation. Eventually this scaffold would resort and be excreted firorn the body, leaving behind a living segment of bone tissue that is completely biological. In this model the scaffold is only present to facilitate and direct the growth of cells and the deposition of cell products, while its eventual resorption allows the induced tissue to become completely integrated into a subjects anatomical and physiological systems. Resorbable polymers, particularly in the fiber form meet this requirement, as well as all the needs listed in table I 1. Using appropriate processing methods, individual fibers can be made with remarkably robust tensile strength [12-23]; more importantly, however, these fibers can be organized in a woven construct that is very rigid [11]. Once cells have been induced to adhere and grow on a scaffold they will require adequate nourishment and gas exchange, which presents another major advantage of fiber based constructs: difflusive properties [23,24].

As mentioned previously, other materials commonly used to replace, or reconstruct bone have relatively poor allowances for diffusion of nutrients and thus are limited.







9

Randomly packed fibers, on the other hand, exhibit porosities greater than 95%, allowing for much greater nutrient inflow and waste matter outflow [251.

The Case for Hierarchical Organization of Fiber Constructs

Bone, like many other tissues in the body, exhibits anisotropic mechanical properties. This directional difference in stiffness, which depends on orientation with respect to the bone's long axis, is a characteristic that is directly tied to its layered molecular and cellular organization. Bone's structure is hierarchical [26] with two levels of this hierarchy that are particularly relevant to its mechanical properties: its lamellae, or layered organization and the collagen fibril arrays within each lamella. Lamellar units are approximately 3 gm wide and are oriented in a direction parallel to the long axis of the bone itself (Figure 1-3).











Figure 1-3. A diagram of the general orientation of lamellae in a segment of bone. Red arrow points out the plane the lamellae are oriented in is in the direction of the long axis of the bone. Adapted from Liu et al. [63].

Each of these arrays, though rotated around an axis perpendicular to the bone's long axis, remains parallel to that axis in their other dimensions (Figure 1-4). Each layer represents an oriented collagen fibril array and as these parallel fibrils are stacked each layer is rotated approximately 30 degrees, as is the case in bone [27,28]. The orientations







10

of these collagen fibrils and the presence of many lamellae are responsible for the mechanical anisotropy of bone as will be discussed more thoroughly in the background.

Logically, when engineering a replacement for large segments of bone, it is desirable that these systems introduce a similar anisotropy. Fibers below a diameter of -100 gm have been shown to influence the orientation of cells grown on them, a phenomenon known as contact guidance [29]. Cells that have been oriented using principles of contact guidance have also been shown to deposit their extra-cellular matrix (ECM) products parallel to their orientation [30-32]. Most notably this has been seen in cases where fibers are used to replace tendons [30].














Figure 14. Schematic representation of collagen fibril orientation within a segment of bone. The cylinders that are fanning out with respect to each other (red arrows) represent the orientations of collagen fibrils as they are organized within a single lamellae. Therefore each lamella is composed of collagen fibril layers organized in this pattern. Adapted fi-om Weiner and Wagner [26].

By creating parallel arrays of fibers that are then organized into layers it stands to reason that cells oriented by each lamella would deposit ECM in a manner directed by that layer (Figure 1-5A). Many layers sandwiched together with angles of rotation between adjacent layers conceivably could produce an anisotropic engineered material (Figure 1-5B). This material would be a laminated composite similar to that of normal







11

bone. Weaving scaffolds with desired orientations would then make it possible to design large scale constructs with the architecture necessary for replacing large segments of bone (Figure 1-5Q.


C



B e "











Figure 1-5. General concept of the hierarchical level of woven fiber based scaffolds. A.) represents the effects of the first level, which is composed of the fibers themselves. B.) Shows the second level of the hierarchy, which is weaving of the individual fibers to form large woven sheets responsible for cellular orientation on a lamellar basis. C.) Finally combining these many lamellae together and putting them in to a 3-D structure provides the highest level of the hierarchy and even more mechanical integrity.

A first step toward developing this technology is understanding how mesenchymal stem cells, the osteoblastic precursors, interact and are oriented by fibers over time. Additionally, understanding how MSCs are influenced by the spacing and organization of fibers in their multi-layer configurations is a key requirement.

The specific aims of this work are designed to elucidate the effects on cellular activity resulting from some of the levels in the hierarchy described above. Studying and extending what is known about the effects of fiber diameter on cellular and ECM orientation will be first. Following the single fiber studies will be multi-fiber studies, which focus on the effects of fiber spacing within a construct composed of fibers







12

arranged in parallel array. Finally, multi-layered constructs with parallel arrays arranged as shown in Figure 1-513 will be examined.













CHAPTER 2
BACKGROUND

To engineer any synthetic tissue replacement material, it is essential to have an understanding of the normal physiology of that tissue. Once a familiarity with the systems involved and the interplay of these systems is achieved, likely sites of manipulation become evident. Controlling these critical sites and thus the development of the desired organ system is the goal of a true tissue engineer.

Bone Formation in Utero

Bone formation begins early on in fetal development and progresses rapidly

throughout normal gestation. The cartilaginous beginnings of the skeletal system are seen as early as the first month of fetal growth when mnesenchymal cells begin to lay down different forms of collagen in an organized manner. These cells separate into two layers, the outer layer, which is composed of differentiated fibroblasts, and the inner layer, which remains undifferentiated mesenchymal cells. These layers together are referred to as the perichondrium, a structure that later becomes the periosteum. [33].

Ossification, or the mineralization of the developing skeleton, begins in the second month of pregnancy and proceeds via two mechanisms. The first mechanism of mineralization to occur is known as intra-membranous ossification and it takes place in sites like the calvarium and the clavicles. Intra-membranous ossification is a direct mineralization in the connective tissues of the fetus, which contrasts with the other mechanism of bone formation: endochondral ossification.




13






14


Endochondral ossification occurs after a scaffold of cartilage is laid down. This

scaffold of cartilage is then mineralized and eventually replaced with bone that is guided by its presence. The only difference between intra-membranous ossification and endochondral ossification is the latter's requirement of a cartilaginous scaffold, which must be deposited first [33].

After the initial mineralization of the cartilaginous scaffold, the development of true bone begins. This formation of true bone begins following the infiltration of vasculature and subsequent supply of oxygen and nutrients [33-34]. MSCs are the developmental precursors to nearly all the connective tissues in the body as schematically shown in Figure 2- 1. The presence of oxygen as supplied by the vasculature plays a significant role in controlling the level of differentiation achieved by MSCs [33-35].








-Bone
-Cartilage
Interh~se-Ligament
-Tendon
-Marrow
-Connective Fibroblast
C, S G2 M-Dermal Fibroblast






Figure 2-1. Schematic of the cell cycle. Top left corresponds to the diagram of cell division below it; outlining the proliferative stage of MSC development. With the advent of correct environmental cues MSCs undergo differentiation and enter the Go phase. Depending on the cues sensed by MSCs they are capable of becoming many different tissue types as listed on the right.






15


MSC Differentiation

Transformation of MSCs into mature bone producing osteoblasts is a multi-step

process involving a number of environmental and cell based signals. This differentiation occurs in the Go phase of the cell cycle (Figure 2-1), which is also known as the quiescent phase since cells are no longer multiplying [3 6-391. The diagram also shows that differentiation at this point leads to different tissue types depending on the cues provided by the environment.

It is hypothesized that bone development along this pathway depends in part on one major environmental cue: the supply of oxygen by newly formed vasculature [33-35]. Indeed it appears that MSCs become fibroblasts if there is a relatively low supply of 02, but osteoblasts if 02 is readily available. Once neovascularization has occurred and this signal has been received, passage of MSCs to fully developed bone producing cells proceeds along an orderly and well-defined path, which is reviewed below. This development is governed by a cascade of cellular based signals, or cytokines, which act in close concert with the cellular events seen in the developing tissue.

The differentiation of MSCs that leads to bone occurs in the GO phase of the cell cycle and is schemnatized below in Figure 2-2. Osteogenic cells develop in a three stage process with two point in between called restriction points [36-39]. Differentiation and advancement through each point are restricted until certain conditions are met; then development may proceed [36,37].

It is generally held that this progression of cellular events is regulated by an intricate system of feedback mechanisms and chemical cues among the developing cells themselves and between the cells and their environment [40-421. The proteinaceous chemical cues, or cytokines, involved in osteogenic cell development are composed of a






16

family of growth factors called bone morphogenetic proteins (BMPs). This family of proteins is quite large and includes the TGF-P isoforms. Each protein is numbered; for example the protein involved in the first stages of MSC differentiation is labeled BMP-2.

The role of the BMPs in bone development has been examined from many different viewpoints. One type of study involves following bone formation, while examining the locations and temporal sequence of the individual components of the skeletal system that are produced [43]. Other types of studies follow the production of growth factors and BMPs spatially and temporally in vivo. The first type of study shows each cell product is characteristic of a particular stage of cellular development. The second type of study shows where and when the growth factors and BMPs are released, allowing the two to be inter-related by shared temporal and spatial relationships [40,411]

In addition, there are other studies that focus on the function of individual growth

factors and BMPs, which provide corroborating evidence for the conclusions drawn from the inter-related temporal and spatial studies [44]. Essentially knowledge of the BMP location in the temporal and spatial sequence can be ascertained by the cellular responses they have been seen to induce. The validity of correlating specific protein production with a level of bone cell differentiation, as is described above, has been examined in the past and found to be acceptable in analogous situations [44]. Using this method, the systemic influences that guide the development of the cells and the location of these BMPs in the cycle can be understood.

The development of osteogenic cells proceeds in a three stage process with two points in between called restriction points (Figure 2-2). Differentiation and advancement






17

through each point are prevented until certain conditions are met; then development may proceed.

Stage one is proliferation. Bone morphogenetic protein-2 (BMP-2) and transforming growth factor-P3 (TGF-P) are active during this first and earliest phase of cellular development. BMP-2 is responsible for the initial levels of differentiation [45]. In fact BMP-2 is specifically located to MSCs, which are considered its target cell [41]. TGF-P is largely responsible for proliferation, or expansion, and it also stimulates production of extra-cellular matrix (ECM) components [35,46]. BMP-2 and TGF-P3 each have an influence on MSC development when they are present individually, but when they are both present concurrently, they act in a synergistic manner [47].



1PROLIFE oN

NEGATIVE FEEDBACK COLLAGEN/
DOWN REGULATING FIBRONECTIN
PROLIFERATION SYNTHESIS


ECMVMATURA1O ,*,
NEGATIVE FEEDBACK
DOWN REGULATING
ECM MATURATION
ECM MINERLIZATION I-


Figure 2-2. MSC developmental sequence leading to bone producing cells. This sequence occurs in the Go phase of the cell cycle as seen in Figure 2-1. Adapted from Stein etal. [36].

In addition to proliferation, the production of ECM is also an important aspect of the first stage in bone development [48]. TGF-13 stimulates production of collagen I, but collagen II is also produced in significant portions during this time. Collagen II is integrally associated with endochondral ossification [44] and a cell product that has been






18

associated with BMP-4 [49,50]. BMP-4 appears to be responsible for the number of chondrocytes that are recruited into the bone producing pathway [51 ], and it has been linked to the production of alkaline phosphatase [52,53]. Alkaline phosphatase indicates that the cells have progressed past the first restriction point, which implicates the value of BMP-4 in stage one with its activity beginning after TGF-P3 and BMP-2 have started to produce their effects.

The first restriction point resides between stage 1 and stage 2 of MSC differentiation. Proliferation and production of ECM occur simultaneously. During stage one, cells continue to proliferate until they are closely associated with each other, forcing each other into a less flattened shape. This change in cell density [36] and shape [54] signals the cells to further differentiate, forming products that lead to ECM maturation, the next stage in cell development.

The ECM and developing osteogenic cells also interact enhancing differentiation and cessation of proliferation, as BMP-4 comes into play. This occurs in a feed forward mechanism where the ECM influences the cells, which in turn influences the ECM [55]. So cell shape (which depends on proliferation) and the ECM largely dictate passage into the second stage.

The second stage of differentiation is the maturation and modification of ECM, thus preparing it for mineralization. The Vgr-1 gene is produced at this time by osteogenic cells and localized into the ECM surrounding hypertrophic chondrocytes [43]. This specifically happens around cells that are implicated in mineralization later. Vgr-1 has been associated with vascularization as bone development continues [56], as well as the further differentiation of osteogenic cells. Interestingly, Vgr-1 must be integrated into the






19

ECM for it to be active, which suggests as conformational change, or cleavage of a portion of the molecule by the ECM.

The Vgr-1 gene leads to the production of BMP-6; that is BMP-6 is the gene product of Vgr-1. BMP-6 leads to the production of the LMP-1 protein, a cell product which is important and necessary for the final differentiation of MSCs. LMP-1 is not regulated by BMP-2, or BMP-4, only BMP-6 has influence on this protein [57], which makes it the main growth factor involved in the second stage of MSC differentiation.

Another aspect of the second developmental stage is the preparation of the ECM for mineralization. It is thought that mineralization of bone requires nucleation sites for the hydroxyapatite crystals that compose bone. Bone Sialoprotein, (BSP) a non-collagenous ECM protein, is expressed in high levels in areas of bone that first begin mineralization, showing its probable importance as a nucleation [58]. This evidence is further supported by in vitro studies, which show BSP specifically causes nucleation of hydroxyapatite crystals, where other non-collagenous ECM proteins do not [59].

The second restriction point is after the ECM maturation of the second stage. BMP-6 and the presence of BSP prepare the cellular environment for this transition, which occurs when that environment is adequately supplied with the necessary quantity of ECM (BSP, collagen, etc...) and level of differentiation [36,37].

Mineralization is the last stage in bone development and MSC differentiation. Mature osteoblasts lead to the production of a number of non-collagenous protein, which are integral for mineralization. This production largely occurs during the second stage of development with the activity of BMP-6 and BSP. There are proteins, however, that are






20

produced during this last stage of development. Osteocalcin is one of these proteins and is a calcium binding protein necessary for mineralization [46].

OP- I (BMP-7) has been purported to be responsible for mineralization and, in conjunction with the ECM changes, cause terminal stages of osteogenic cell differentiation [60]. OP-i also leads to the up-regulation of BMP-6 and the downregulation of BMP-2 and -4, which implies that its activity normally occurs in the later stages of cell development [61].

OP- I has been shown to induce differentiation of osteoblasts and production of bone in a number of studies. Its role, however, appears to be in this later stage of bone development. Using OP- I on cells that have varying levels of development, causes many of them to differentiate before they normally would without its stimulation [44]. Many more cells can be stimulated to begin mineralization, inducing bone formation. OP-i1, however, does not lead to ECM production [44]. Essentially, if OP- I is present too soon it forces cells to differentiate before they can produce the necessary ECM for vascular development and other features of filly formed bone.

Presence of OP- I can lead to fully developed bone when it is administered alone

[44,45], but important aspects of development that are dependent on ECM and the noncollagenous proteins are attenuated. This implicates OP- I as the signal protein that is most important later in development, whose action is to stimulate final differentiation and aid mineralization. After the ECM has been developed by the earlier stages of growth, cells that are growing exist in population with varied levels of differentiation depending on the specific local environment. Terminal differentiation stimulated by OP- I seems to influence all cells at their given stage of development, speeding them to final






21


differentiation and mineralization, bringing the whole tissue to a final differentiated whole.

Bone Development and Bone Anatomy

Growth occurs rapidly at the ends, or epiphyses of developing bones and more slowly toward the centers of the shafts, or the diaphyses. Rapid growth primarily leads to spongy bone, while the relatively slow build up of bone that occurs in the diaphysis is more compact and dense. Spongy bone is composed of trabeculae, a porous honeycomb of interconnected bony processes. Trabeculae initially develop with a random orientation, but with the application of the stresses associated with living the orientation of these processes assume a pattern (Figure 2-3). In a probabilistic manner, trabeculae oriented at


Force












Figure 2-3. Lines of force seen in proximal portions of the femur. Weight bearing pressures as represented by the arrow are responsible for development of trabecular organization Adapted from Sinclair [331.

odd angles with the lines of stress imposed on the bone are broken down, while those that are aligned are relatively unchanged [33].

Bone tissue has been optimized through evolution in ways that are very specific to its function. During development, as mentioned above, it is remodeled through the stresses associated with weight bearing and muscular tension, so that it exhibits properties of






22

mechanical anisotropy. The source of this anisotropy is traced to the macromolecular and cellular level of bone.

There are a many different types of bone to be found in nature, but one of the most important is larnellar bone. This bone type, named for its organizational structure, is the most common bone type found in humans and is primarily responsible for the load bearing fimction of the skeleton. Lamellar bone consists of many mineralized layers, a characteristic which was noticed as early as 1906 [27]. The elucidation of the lamellar structure has taken some time, but now it is generally held that a "rotated plywood" motif best describes the organization of this bone type [26-28, 62,63].

The structure of bone is hierarchical [27]. Two levels of this hierarchy that are

particularly relevant to its mechanical properties its the larnellae and the collagen fibril arrays within each lammela. Lamellar units are approximately 3 Pm wide and are oriented in a direction parallel to the long axis of the bone itself (Figure 2-4). The next step down the hierarchy is the collagen fibril. Each lamella is composed of a number of collagen fibril arrays rotated at angles of about 30 degrees with each other. That is, each subsequent collagen array is rotated with respect to the one before it as one imagines, passing from one lamellar boundary to the next. Each array, though rotated around an axis perpendicular to the bone's long axis, remains parallel to that axis in their other dimension (Figure 2-5). When the rotated plywood structures, which compose larnella, are viewed in cross section they produce microscopic patterns referred to as "nested arcs" (Figure 2-613) [28,62].





23













Figure 2-4. Examples of lamellar orientation within a section of bone Adapted firom. Liu et al. [63].

.... ..........











Figure 2-5. Fibrillar orientation within bone. Fibrils are oriented in one plane which parallels that of the bone's long axis, while rotating at 30 degree increments with respect to each other. Adapted from Weiner [27].

Figure 2-6A shows how the rotated layers, when stacked, produce this effect. Each layer represents an oriented collagen fibril layer and as they are stacked each layer is rotated approximately 30 degrees, as is the case in bone [27,281.

One can envision how the orientations of these collagen fibrils and the presence of many lamellae make bone so mechanically anisotropic. The anisotropic nature of bone has been studied for some time, particularly on the macroscopic level. Directionally oriented bone specimens, which are very large compared to the larnellar sub-units, have been subjected to stress-strain measurements. Tensile and compressive examinations of






24

P/2




. . . . . . . . . . . 2 A[A]











Figure 2-6. Rotated plywood model. A.) Drawing of many parallel fibril arrays each rotated 30 degrees with respect to the adjacent layers. The blue arrows highlight the centers of the nested arcs. B.)Notice how the rotation leads to a nested arc motif as seen in this cross-section. Adapted from Giraud-Guille [28]. these samples have shown that lamellar bone possesses markedly higher modulus values when loaded parallel to its long axis than in any other direction [27,641. Extending these measurements down to the lamellar and fibrillar subunit scale has been difficult, however, some results have been forthcoming supporting the relationship between anisotropy and the rotated plywood model.

Using microhardness instruments, the presence of anisotropy on a very small scale was related to the orientation of the mineralized fibrils [65]. Similarly, stress-strain data using very small scale bending specimens (-160 Lm diameter) supports the relationship between the lamellar structure and its function [63].

Overall, the numerous layers present in lamellar bone confer its highest strength in the longitudinal direction. In addition to this quality, the multi-angular orientation of the rotated plywood model makes the bone very resistant to fracture with the application of lateral stresses [26]. These examples demonstrate a relationship similar to what has been






25


described in aortic leaflets and arterial tissue, namely: the lamellar nature of bone, which is derived on the macromolecular and cellular level, is a quality integral to its physiologic function.

Fracture Healing and Ectopic Bone Formation Skeletal repair after fracture follows a sequence of events that is virtually analogous to that seen during development [35]. MSCs gather at the fracture site and form a repair blastema, or fracture callus. If the fracture site is stable and not subject to micro-motion, the MSCs will directly differentiate and become bone producing cells. If, however, there is instability that allows small amounts of motion the MSCs will form a more cartilaginous callus that will stabilize the break. With stability the ability for vasculature to successfully infiltrate increases, a condition that leads to endochondral ossification [35].

Ectopic bone formation, or the formation of bone in sites well away from the skeleton proper, proceeds very similarly to that of fetal bone formation as well. Use of demineralized bone chips, or polymer carriers loaded with BMPs leads to the development of ectopic bone [ 15,3 5,66]. The initial inflammation associated with implantation of the bone chips, or carrier is responsible for delivering MSCs to the ectopic site, which can be subcutaneous, or within a muscle. Figure 2-7 schematically represents this process, which again is dependent on the infiltration of vasculature.

Porosity and Diffusion Properties of Implant Materials

Proper porosity is an integral quality in a scaffold material as it controls the influx of nutrients to the cells. Initially the nutrients must be able to diffuse into the scaffold until such time as new vasculature has established itself. In order for vascularization to occur






26



Encysted
Demineralized by MSCS Cartilage Differentiation
Bone Chips


00









Bone Formation I Vascular Cartilage Hypertrophy/
Marrowization Invasion First Bone Formation



Figure 2-7. Schematic representation of ectopic bone formation around implanted, demineralized bone chips. It is thought that demineralized bone chips lead to ectopic bone formation because they contain BMPs. Adapted from Caplan [35]. the pore structures must be on the order of 200-500 pm in diameter [14]. This requirement is easily manipulated by varying the weave, tightness and diameter of the fibers used in the fabric.

There are many studies in the literature that attempt to use osteoconductive scaffolds to act as a synthetic bone replacement material [4,6,10,67]. Osteoconduction is defined as a material that allows vascular ingress, cellular infiltration, cartilage formation and mineral deposition [4]. These properties are indeed required in a bone replacement material, though they are by no means inclusive. Loading scaffolds with MSCs has also been used with some success in studies examining bone replacement [ 10,16,67,681. Osteoblastic precursors, which differentiate and become fully functioning osteoblasts are encompassed in a phenomena defined as osteoinduction. Any method or material that






27


induces differentiation of precursor cells into adult osteoblasts is included in this definition.

Large mechanical supports like allografi material have been shown to be

osteoconductive, but only at the ends of the allograft where it contacts the living bone [3]. Allograft bone has been fully incorporated in some instances where it has been used to fill areas around a collapsed acetabulum [6]. This was limited to only a few centimeters of material that was surrounded by living bone on all sides. Tricalcium phosphate scaffolds with pores ranging from 100 gim to 300 gim and a 36% porous volume show similar results, with vascular and cellular invasion of only 0.75 cm into the synthetic [10]. Replacing large volumes of bone has never been successful in this respect. Scaffold materials when used alone; seem to limit the nutrient supply, which prevents continued ingrowth.

Fibers have also been effective as a substrate in fixed bed bioreactors. Fiber type beds, or substrates, possess porosity much higher than other types of fixed beds (> 90% porosity). This porosity facilitates nutrient exchange by increasing the volume of medium allowed in and out of the scaffolds structure [24,25]. Because of the fiber architecture versus beads, or some other substrate, the cells in bioreactors with fiber beds produced

0. 15 IU of interferon per cell greater than an order of magnitude higher than cells on other substrates. A lack of diffusive ability has been a major drawback of the porous scaffolds mentioned previously.

A fixed bed is basically a solid support matrix that is used in bioreactor technology to maintain contact dependent cells in the culture environment. Its 3-D nature increases the surface area for cell adhesion beyond that of a 2-D culture dish. Some systems use glass






28


beads to pack the fixed bed, but their geometry limits the void fraction available for cell growth and diffusion [24]. Using 3 mm glass beads the void fraction is as low as 35 %, but using fibers the void fractions increase to > 90%, providing much better flow medium flow and thus greater cell growth [25].

Remodeling and Cellular Orientation

During development, as well as fracture healing reorganization of cells and

remodeling of their ECM products occurs. This remodeling and organization is in response to mechanical stresses and is responsible for the development of anisotropic strength characteristic of bone [33]. This portion of the healing sequence can be quite lengthy and logically, it depends on how much the cells are organized when remodeling begins. It stands to reason that cells that start the remodeling process in a more organized state will have to undergo less change than randomly oriented cells and ECM.

A number of resorbable polymer scaffolds have been examined for use as scaffolds. Martin et al. used porous polyglycolic acid and polyethylene glycol scaffolds, which showed the ability for MSCs to differentiate when loaded on the scaffold [ 17]. Similarly, Holy et al. used porous polylactic acid scaffolds in vitro, demonstrating normal cell activity as well [12]. In vivo a number of studies have been performed using porous ticalcium phosphate, or hydroxyapatite ceramics [ 14,16,18-20]. Using MSCs, these implants show a development of bone tissue that incorporates the degradable biomnaterial. Both polymer and ceramic systems, however, show cells that appear to be randomly oriented. No effort has been made to examine cell orientation in any of these studies.

Though the problem of fatigue and tissue replacement is addressed with the

development of cell based tissue engineering technology, so far there has been no attempt to organize the implanted cells, or direct the ECM and mineralization products they lay






29

down. As reviewed next in the contact guidance section, there have been many studies performed in vitro that indicate the efficacy of this type of cellular and ECM organization. Applying this knowledge to cell based tissue engineering is the next step in this area of research and the focus of this work.

As we have seen, cellular and molecular organization is primarily responsible for the mechanical properties many tissues possess and so accomplishing this organization seems to be a worthwhile goal. A self-healing, or living tissue engineered replacement possessing anisotropic: mechanical properties similar to bone, may be the solution to many bone replacement problems.

Contact Guidance

Cell systems and tissues are influenced by a number of factors during their

development and the course of their existence within an organism. Two of the most prominent factors in vivo appear to be chemistry and topography. Chemical cues have been shown to effect cell activity during the phenomenon of chemotaxis, a cellular movement toward a gradient of a chemo-attractant molecule [391. Cells guided by this type of stimulus will migrate in the direction of a released cell product thus localizing themselves to the site of injury, or need, as in the migration of leukocytes to damaged tissue [69]. Upon fracture, it is likely that bone repair is instigated in an analogous fashion by the release of chemical agents like TGF-0 and other proteinaceous cell products, thus attracting osteoblastic pre-cursors to the site [41,46].

Another form of influence on cells, which may be considered chemical is cell

binding to specific receptors, or arrangements of chemical molecules that are bound to a surface (Figure 2-8). Protein mediated receptor binding to substrates is an event that






30




4W)





Receptors
Fibronectin (Fn) in ECM



Figure 2-8. Receptor mediated adhesion of osteoblasts via fibronectin. Other ECM proteins perform similar functions, though Fn is one of the most prominent. Adapted from Alberts etal. [39].

occurs in numerous cell types within the body including those of endothelial (vascular endothelial cells) [70-72]1, mesenchymal (mesenchymal stem cells and fibroblasts) [20,73,74] and epithelial cells (neural) [75] origin.

Attachment of cells to substrate surfaces is almost always mediated by protein

adsorption [76]. These proteins are components of the serum used for cell culture and include fibronectin and vitronectin among others [76,77]. Adhesion and attachment by cells to these proteins depends on the ability for the proteins to adsorb to the surface, which in turn is dependent on the surface energy of the substrate. These proteins must also interact with the substrate in a way that does not change their conformation, thus remaining recognizable to the cell. Adhesion and attachment have been shown to influence rate of proliferation, or growth [71,72,76-79].

Modulus and stiffness of the substrate also have an effect on cell function and growth. Studies performed using collagen gels as scaffolds for cells of mesenchymal origin show that increased stiffness of the scaffold leads to an increased rate of proliferation. In addition to this increased rate the duration of proliferation persisted longer than that seen





31

in scaffolds with less stiffness [801. In similar studies collagen gels were anchored on one axis leading to a preferential tension development in the direction of this axis by fibroblasts. Cell numbers increased by five times in the anchored gels and ECM generation was greater than that seen in the un-anchored gels, which in contrast demonstrated a five times decrease in cell number [8 1 ].

Another major form of cellular influence: topography, has been shown to direct the function of cells in vivo, as well. Neural cells migrate along the lengdi of radial glial cell fibers until they reach their destination, thus being guided to different regions of the central nervous system [82,83]. Basement membranes, the thin layer that many cell types grow directly on, also have an inherent topography that has been suggested to influence corneal epithelial cells [841 and renal endothelial cells [85).

Introduction of synthetic devices into the body disrupts many of these cell and tissue systems, which then try to recover and re-establish stability in the presence of the implant. Implant design has come to focus on minimization of these disturbances and optimization of the cellular response to these devices once they have been placed in vivo [86]. Ideally cells could be induced to respond to devices in an optimal manner by using their inherent cellular mechanisms and machinery, which allow them to exist and live in their normal environment. Inducing migration of a given cell type, or using microtopography to direct a cell's function are examples of this concept.

Toward this end, many studies have been performed that examine how topography and chemistry influence cell shape [70,87,88], migration [75, 89-911 and function [72, 91,92]. This cellular direction, which is known as contact guidance, has been widely examined and found to occur in many different cell types. Similarly, studies have been performed






32


that use niicropatterned, chemical cues to spatially guide cell growth, in addition to influencing cell function [70,72,77]. There have also been a number of studies that examine the effects of receptors and receptor-like molecules on cell activity.

Contact guidance, the topographical, or chemical control of cellular orientation and

activity, is well supported by many studies in the literature [75,82-91,64,79,93,94]. Some of the earliest experiments studied the influence of topographical cues by examining how cells oriented themselves on glass fibers [85]. Since then the electronics industry has developed viable micro-fabrication methods and a number of studies have shown how topographical cues with specific sizes and patterns can invoke cells to behave in predictable ways [64,79,90,91,93,94].

Using these methods, many cells of mesenchymal origin have been shown to orient

themselves parallel to micron scale topographical features, like ridges [64,86,93,94]. This is also true of osteoblasts, the primary bone forming cells in mammals. Once oriented, osteoblasts and osteoblast-like cells will mineralize and lay down extra-cellular matrix parallel to microtopographical features on a culture substrate [31,32,91]. In fact in vitro mineralization of bone has been directed on a macro-scale merely by scratching the surface of a culture dish with sand paper [3 1].

Since the aforementioned development of micro-fabrication in the electronics industry a number of studies have shown the specific effects that topographical cues can have on cell shape and activity. Cues with specific sizes and patterns, which were created using the micro-machining technology, can invoke cells to behave in predictable ways [64,79,82-91,93-95]. A number of theories have been put forth regarding the mechanisms of contact guidance. The direction of a colloidal exudate from the cell by the grooves






33

[76], avoidance by cells of discontinuities in their paths [29], as well as the thought that focal contacts may only form on the tops of the ridges [32]. All of these theories enjoy some degree of experimental support, though which has the most dominant effect is not clear at present. There does, however, appear to be some correlation of feature size with the influence on cell shape. Groove dimensions that seem most effective across many cell types, are grooves with dimensions on the order of magnitude of the cell [76]. This is not always the case, however. In a recent study by Nealy et al., which used nano-scale features, contact guidance was exhibited by corneal epithelia] cells [96].

In addition to glass fibers, early contact guidance studies often used spider webs [32] to orient fibroblasts. This effect was achieved below a critical diameter of 100 lim. Above 100 jim, the cells were no longer oriented on the fiber, which led to the hypothesis that cells were unable to bend around a certain sharpness of curvature. That is, the cells could not bend around the circumference of a certain fiber once its diameter had gone below a certain point, leaving them no recourse, but to elongate in the direction of the fiber's long axis [29,97]. Fibers have also been shown to orient many other cells types including neurons, schwann cells, macrophages and transformed BHK fibroblasts [75,98]. In these studies the contact guidance effect was shown on other fiber types too, namely carbon filaments and synthesized fibronectin filaments.

Architecture of Fiber Based Scaffolds: Development of Cell-Based Tension

Using the phenomena of contact guidance, which orients not only cells, but also their ECM products, advanced scaffold materials can be designed. Commercial fibers can be woven into textiles of various 3-1) configurations possessing strong mechanical properties. Adjusting the tightness and geometry of the weave, as well as the diameter of






34


the fibers allows a number of control features. Fiber based scaffolds can be designed to allow for initial nutrient exchange, as well as longer term vascular infiltration. In addition the direction of cellular growth and ECM deposition can be influenced so that a decreased remodeling requirement is present after the newly created tissue is formed. That is cells can be aligned to a degree that is close to the alignment and orientation they will exhibit after remodeling and so will be less distant from their equilibrium state.

An extensive review of the literature indicates that one of the key components of connective tissue cellular physiology is the achievement of proper cell spreading, or tension. Cell spreading is a phenomenon influenced by topography, chemistry and stiffness of the substrate the cell is bound to. These are the elements of contact guidance as reviewed earlier. As we saw, surface properties led to proper adhesion through the protein mediated receptor binding (Figure 2-9). Topography and chemical micropattems also influence adhesion and the direction, which cells are able to achieve stretch. Indeed spreading and the achievement of proper cell tension appear to be one of the main requirements for normal cell function.

Without adhesion, contact dependent cells, like those of mesenchymal origin remain rounded without inducing tension and fail to differentiate [54,88,92,99]. On another front inhibition of stretch mediated chloride receptors attenuate response of connective tissue cells to topographical guidance [100], strongly implicating stretch as a requirement in topographical guidance. Mechanical stretching of the substrate cells reside on has also been shown to direct cellular orientation and deposition of ECM proteins in a MSC system [ 10 1 ]. Indeed a translational strain of 10% applied concurrently with 25% rotational strain was responsible for increasing the alignment of these cells 2.5 times






35


when compared to controls, which received no stimulus [ 10 1]. In addition to cellular alignment, collagen fibrils were found aligned as well, when no fibrils were even seen in the control constructs. Interestingly 0.2 grams of tensile stress has also been shown to increase BMP-4 by two times that of control as part of suture development in the mineralization of rat calvaria [102].

So in addition to the development of cellular tension, which is necessary for normal and optimal cell function, cellular stretch plays a role in developmental process. Indeed stretch mediated receptors may be the signaling mechanism responsible for translating cell based tension generation into enhanced cellular function and differentiation. Therefore scaffolds developed for application in connective tissue engineering should include consideration of these phenomena in their design.

In tissue engineering applications, scaffolds loaded with cells are often seen to

contract through the development of cellular based mechanical force [80,81,103-105]. This phenomenon has been strongly associated with the contractile protein "Smooth Muscle Actin" (SMA). SMA is so named because it was first discovered in smooth muscle cells, but it is also found in other cells of mesenchymal origin like fibroblasts and osteoblasts. In fact it has been shown that MSCs and osteoblasts, will contract tissueengineered scaffolds when they are loaded onto them [103,104,106].

Contraction by osteoblasts and other cells of mesenchymal origin occurs in close

relation to the numbers of cells loaded on these scaffolds [ 104,106]. Once loaded onto the scaffold, the force generation is linearly associated with the numbers of cells, which indicates a collective effort by the cells leading a total force generation [105]. This relationship between cell number and contraction also indicates that cellular






36


communication is occurring through gap junctions [ 104]. Gap junctions exist on the physical level and are essentially pores that interconnect adjacent cells, or cells that are in actual contact with each other. Gap junction interactions have been implicated in macroscopic phenomenon of scaffold contraction [ 104], which indicates these cells are acting in a multi-cellular manner. In fact these findings suggest that osteoblasts and connective tissue cells are acting in an aggregate manner with many physically associated cells generating force cooperatively.

The importance of tension generation by connective tissue cells becomes apparent

when its effects on cell function are considered. When seeded on collagen and collagen glycosaminoglycan scaffolds proliferation of cells is markedly enhanced, but only when the gels are secured to the culture dish. Free-floating gels/scaffolds experienced a regression in the number of cells, a fact attributed to their inability to generate tension [81,103]. In a similar manner, when a series of collagen scaffolds with increasing degrees of cross-linking leading to increasing amounts of stiffness are used, the stiffest formulations led to the greatest proliferation rate and duration [80].

In addition to enhanced proliferation, the ability of cells to generate tension led to increased production of collagen [8 1 ], as well as increased production of proteins, calcium and alkaline phosphatase [103]. All these responses indicate increased levels of differentiation by cells, which also has been attributed to the ability to generate tension.

Overall it appears that connective tissue cells lead to contraction of scaffold material as each tries to achieve the an optimal tension, or stretch necessary for proper differentiation and optimal fiction. The sum of these many individual cell tensions leads to a cooperative force causing the whole construct to contract. It also becomes apparent






37


that these cells are not acting individually, but are physically linked to each other in an aggregate manner, allowing cell communication through gap junctions.

This contraction is commonly seen in wound healing and it can lead to scar formation [107]. Similarly it is part of the normal healing cycle seen in tendon healing [ 108]. The same proteins that cause this in other closely related cell types are found in Osteoblasts and MSCs, therefore it seems likely that this phenomenon is a part of normal fracture healing. In fact it follows logically that contraction of cells healing a fr-acture would be useful in bringing separated fragments of bone back together. Given the importance of this cellular activity, it must be considered in the design of a fiber-based scaffold's architecture and spacing.

Proper spacing of fibers is necessary to allow optimal cell tension and stretch, but also it must be optimized for the aggregate activity, or cooperative bridging of cells occurring across gaps in the scaffold. During connective tissue healing, a fibrin clot acts as scaffold allowing cells to migrate from areas of healthy tissue to the area of injury [108]. Without this scaffold, or a synthetic replacement, in vitro studies show cells are unable to bridge gaps as small as 50 pm [ 108].

Autocrine/Paracrine Considerations in Scaffold Design

As mentioned previously, tensile stress increases BMP-4 by a factor of two as part of the developmental cycle [102]. In addition, this stimulus leads to an increased responsiveness of osteoblasts to morphogens and vitamin D [102]. Applying exogenous TGF-3 on the other hand increases the tension produced per cell by a factor of two [ 105]. Increased TGF-3 and BMP-4 precede increased production of collagen, alkaline phosphate and increased MSC differentiation. Production of these growth factors as





38

stimulated by tension occurs in such a way that they have their effects in an autocrine and paracrine manner, thus tying this form of cellular communication to tensile stimulus. Additionally it is seen that exogenously administered TGF-P increases multi-layering of cells within the confines of pores in collagen scaffolds and causes an increase in macroscale contraction of these scaffolds [105].

Inter-cellular communication is a phenomenon that takes on many forms. Autocrine signaling is the ability of a cell to produce cytokines that are then sensed by receptors on the same cell (Figure 2-9). This method of communication was first elucidated in cancer cells, which were seen to produce growth factors independently of environmental stimulus, thus freeing them to grow uncontrollably and form tumors [39]. Since then autocrine signaling has been established as part of the normal physiology of many tissues including bone [69,85].

Paracrine signaling is the intercellular communication that occurs between adjacent cells, confining its action to a local area (Figure 2-1OA)[69]. Some paracrine signaling




0
4 0

0








Figure 2-9. Diagram of autocnine signaling whereby a cell produces a certain cytokine and then releases it. Cell product is then bound by receptors on the cell's own surface. Adapted from McCance and Heuther [69].






39


occurs by growth factors being released in small quantities to the local extra-cellular milieu. In the case of TGF-P and BMP-4, however, the signaling molecules are bound in the matrix and non-soluble. Signaling occurs by actual physical contact between the cells creating the stimulus as cells nearby extend filopod-like extensions that contact the stimulating cells directly (Figure 2-1013)[5].

In order to optimize the cellular microenvironent, paracrine signaling distances must be considered in addition to fiber spacing leading to the most biologically optimal cellular tension. Indeed the two phenomena are inter-related as fiber spacing is essentially a topographical stimulus and surface topography has been shown to stimulate autocrine and paracrine growth factors [ 109]. But how far paracrine signals released into the extra-cellular milieu able to travel and still remain effective is also a highly relevant question. It seems logical that there is a fiber spacing distance that will allow intercellular paracrine communication between adjacent fibers and that it should be optimized in a 3-D scaffold.

Work in the area of paracrine signaling distance is rather sparse, though it is generally accepted that this is a major contributor to osteoblastic development and how it occurs [85,109]. However, some studies have been performed in culture using various cell types. Generally it is seen in culture models using pituitary cells, as well as parathyroid cells that increasing distance between secretory cells leads to a decreased response, or communication between cells [1I10, 111 ]. An apparent critical distance beyond which interaction becomes negligible in pituitary cells is approximately 75 j tm [110]. This distance of course is for diffusible cell products released into the local environment.






40





















Figure 2-10. Diagram of paracrine signaling. A.) shows a cell releasing a diffusible cytokine (square nucleus) that acts over short distances creating a gradient of responses in surrounding cells (note color scale of oval nuclei). B.) shows similar paracrine phenomena, but the cytokine is passed by filopod-like extensions, or actual physical connections. Adapted from Christian [5].

Matrix bound BMPs on the other hand may possess paracrine activity with a more limited distance of interaction, perhaps as low as a few cell diameters, or tens of microns [5].

Distance and spacing between fibers is therefore likely to be important in the

development of proper cell adhesion and tension. This phenomena, as we have seen, is closely related to autocrine and paracrine signaling distance, which must also be considered in the design of fiber based scaffolds.

Overall, it seems apparent that fiber based scaffolds possess strong potential for use in development of hierarchical bone replacement materials. Their design flexibility allows for control of cellular organization, as well as ECM organization, two of the most integral components leading to the mechanical integrity of bone. In addition, weaving and organization of fibers allows for higher level 3-D construct design that allows greater





41

nutrient diffusion and facilitation of vascular ingrowth, while optimizing cell adhesion and induction of tension, as well as intercellular autocrine and paracrine interactions.














CHAPTER 3
SINGLE FIBER STUDIES

Introduction

As mentioned previously, fibers have been used for a number of scaffold-based tissue engineering experiments. Ahmed and Brown used synthetic Fibronectin (Fn) fibers for the direction of Schwann cells, as well as dermal fibroblasts, macrophages and epitenon fibroblasts. These fibers, with diameters of 0.5-7.0 Wi, were organized into a mat leading to a surface composed of fibers oriented in the same direction [75].

Carbon and polymer filaments have also studied for use as scaffold materials in the area of connective tissue engineering. In vitro Kevlar-49, nylon and carbon allow centimeter scale migration of mesenchymal originating cells from tendon explants along their surfaces. In addition to undergoing extensive outgrowth, these cells were oriented with the long axes of these synthetic fibers. Kevlar-49, nylon and carbon fibers in this study were 12 jim, 7.5 p~m and 22 pWm respectively [ 112].

Other tendon outgrowth studies performed using carbon (8 P~m diameter), Dacron (I11 jim diameter), polyethylene (20 p~m diameter) and nylon (102, 52 and 22 g~m diameters) show cells of mesenchymnal origin orient in the direction of their long axes primarily in response to the diameter of the fibers they are grown on. This is true of all fibers, though there is an interesting and important qualification to this trend: Larger diameter fibers on the order of 100 p~m did not orient cells until after they had become confluent on the fiber






42






43

surface [113]. This implies some socially based mechanical interaction is at work as cells become confluent and come into contact with each other.

In vivo, carbon bundles composed of 8 pm thick fibers have been used as a ligament replacement material and exhibit the ability to organize fibroblasts and tendon cells. More than orienting these cells they appear to orient the collagen produced by the oriented cells [301. However, these tows were under biomechanical stress as the subjects used them in the course of everyday living, so it is unclear if the collagen orientation was a result of topographical cellular orientation. It is conceivable that collagen was oriented by the dynamic stresses of walking and muscular contraction that were placed on it in its role as a ligament replacement.

Though these studies provide much qualitative data about the orientation of cells on

the surfaces of fibers, they are lacking in hard quantitative evidence. Overall it is apparent that diameter has an effect and that time, which allows the achievement of confluence, are important factors in contact guidance of mesenchymal cells on fibers.

Some quantitative work has been done on the fiber based contact guidance of

connective tissue cells. Dunn and Heath performed alignment measurements of chick heart fibroblasts on the surfaces of soda glass fibers at 48 hours, showing an order of magnitude increase in NFF on fibers of 40 Lrn diameter compared to fibers of 100 pm [29]. Above 100 gm there was no significant contact guidance. Similarly, Fischer and Tickle quantitatively showed normal BHK rat fibroblasts elongated on the surface of glass fibers, but at only 24 hours [97]. Both studies use quantitative measures of cellular elongation, but are limited to very short time periods. Indeed





44

measures are taken to prevent social interaction between cells even for these short times in culture.

In this study mesenchymal stem cells are studied quantitatively and for time periods that extend beyond those needed to achieve confluence. In fact the time periods used were lengthy enough to observe the characteristic multilayering observed in this cell type. The object of this study was to examine the effects of diameter and time on MSC contact guidance and to do this in a quantitative manner. It is hypothesized that decreasing diameter will lead to an increased level of contact guidance, as will increasing culture times. Knowledge and understanding of this cellular response to fibers will prove invaluable in fiber-based scaffold design.

This section's main hypothesis is: MSCs will have a significantly (alpha = 0.05)

different elongation and cellular activity when grown on -140 pm, 100 pm and 79 pm diameter fibers. This difference in cell activity will exist between each fiber diameter group and inducing contact guidance of these cells will be the primary goal.

Materials

Black Nylon (EthilonTM, Ethicon) sutures of 4-0 (- 140 pm diameter), 9-0 (-79 pm diameter) and 10-0 (- 39 pam diameter) sizes were used for initial SEM studies. Clear polyglactin (MaxonTM, Davis and Geck) fibers of 5-0 (-140 jn diameter), 6-0 (- 99 Ptm diameter) and 7-0 (- 79 pm diameter) were used for light microscopy studies.

Culture materials included a-minimal essential medium (Sigma, M0894) with 15% fetal bovine serum (Sigma, F4135), 50 mg/ml ascorbic acid (Sigma, A4034), 10 mM bglycerophosphate (Sigma, G9891), antibiotics (0.1 mg/ml penicillin G, 0.05 mg/ml gentamicin and 0.3 mg/ml fungizone) and 10-8 M dexamethasone (Sigma, D2915).






45

Bovine Fibronectin (Fn) -25 ug/mI (Sigma, F 114 1) was used to soak constructs prior to cell seeding, thus making fiber surfaces more amenable to cell adhesion. After soaking, En solution was pipetted off and cells were seeded directly onto constructs.

MSCs were collected from both femora of grown, Sprague Dawley rats (-1 50-300 g) that were provided courtesy of Dr. Colin Sumner's lab at the UF brain institute.

Methods

Single Fiber Scaffolds

Single fiber scaffolds were prepared by stringing Maxon and Nylon sutures with the diameters mentioned previously across polystyrene support rings (Figure 3-1). Constructs were designed to suspend the fiber completely above the floor of the culture dish so all cellular response was caused by the topography of the fiber alone.

















Figure 3-1. Examples of Maxon TM single fiber constructs. 5-0 is on the left, 6-0 center and 7-0 on the right.

Cell Culture

After dissecting each femur out, the epiphyses were cut off and the marrow plugs were flushed out of both diaphyses using aliquots of the fully supplemented medium(FSM) to






46

be described next. Plugs were flushed into 30 ml of FSM, which will be: a-minimal essential medium (Sigma, M0894) with 15% fetal bovine serum (Sigma, F4135), 50 mg/ml ascorbic acid (Sigma, A4034), 10 mM b-glycerophosphate (Sigma, G989 1), antibiotics (0.1 mg/ml penicillin G, 0.05 mg/ml gentamicin and 0.3 mg/ml fungizone).

After the primary culture passage of --6 days, cells were trypsinized and passaged using a 0.01% trypsin and 10 M EDTA mixture (Sigma, T3924) in phosphate buffered saline. Released cells were reseeded into three 75 mm2 culture flasks. MSC Seeding

MSC Seeding onto the fibers was done at extremely high concentration by taking a whole flask of trypsinized MSCs and adding -10ml of medium to it. Addition of this small amount of medium served to quench the trypsin reaction, while leaving the cells at a very high density. After EtOH and UV sterilization (24 hours in absolute EtOH followed by drying under 256 nm UV light), fibers were placed in 24 well plates and secured to the walls via melting of the polystyrene support ring to the polystyrene of the dish (heated implement was a small soldering iron). Fibronectin (25 ug/ml) was used to soak constructs (except for 9-0 constructs) for one hour then each construct was covered with a large drop if high density cells in suspension and allowed to incubate for about 2-4 hours while cell adhesion occurred. At this point the wells were flooded with culture medium to the normal level.

SEM Imaging (Nylon sutures)

Cells were fixed in a phosphate buffered saline with 10% formalin at pH 7.4. Each of the constructs were dehydrated by serial ethanol incubation at concentrations of 30%, 50%, 70%, 90% and 100% over periods of about two days (longer dehydration time






47


minimizes shrinkage of cellular material). Following ethanolic dehydration, constructs were critical point dried, coated with -20 A of Pd/Au and examined via SEM. Light Microscopy (Maxon Sutures)

Constructs were removed at 3, 6, 9 and 12 days and passed through serial EtOH

dehydration of 30%, 50%, 70%, 90% and 100%. Upon reaching 100% EtOH, constructs were left in 100% EtOll until all constructs were removed from culture then all were stained simultaneously using Hematoxylin and Eosin stains. Nuclear Form Factor (NFF) and the angle were measured with respect to the fiber's long axis as described in Appendix 1.

Results

SEM Studies

4-0 nylon sutures

Constructs initially showed a robust amount of cell adhesion and coverage under the light microscope during culture. By the last day, every fiber was completely covered with cells and some even appeared to have begun multi-layering.

Cells liberally covered the surface of the fibers, but an interesting phenomena was that these cells were elongated though not parallel to the long axis of the fibers (Figure 32). Instead they were spirally oriented around the surface of the fiber, much like a barber's pole. A similar result was seen in work by Ricci et al. when larger diameter fibers were used [113].

In addition to the spiraling effect of cell orientation, what appeared to be

multilayering was present. Light microscopy during culture demonstrated what appeared to be multilayers, or areas of cell buildup at the periphery of the fibers. Because the fibers were dyed black and used primarily to develop protocols, light microscopy staining was






48










4 i









i ji






Figure 3-2. 4-0 nylon fibers with RMSCs after being cultured for 5 days. A and B) Note elongated and flattened cells that spiral around the long axis of the fiber. C.) is a close up. Arrows indicate some of the most prominent cells. Flattened cells cover nearly the entire not possible and cells were only visible in a x-section-like view where they were not superimposed on the fiber.

Examination of 4-0 fibers showed a definite striation-like appearance (Figure 3-3), which appeared to proceed in two directions (red arrow and green arrow indicate approximate directions). In some areas it appeared that each of these oriented striae were composed of cells and that the cells in one direction were overlapping the cells that were growing in the other spiral direction.






49









'JA
V




Figure 3-3. Examples of MSC growth on the surface of 4-0 nylon fibers. These cells exhibit spiraling and what may be alternating spiral orientations between layers.

Though this result requires further support, there is some indication in the literature

that may explain it. Commonly osteoblast-like cells multilayer during culture, depositing collagen at each level in a manner that is situated in orthogonal directions between each layer. That is cells in one layer are separated by collagen laid down by the next cell in the layer above, this collagen in turn is at a 90 degree angle from the collagen laid down by the cell below (both chicken and rat cell culture models show this) [114-1161. 9-0 nylon sutures

9-0 Nylon sutures shown in Figure 34 exhibit less cell adhesion than the other nylon fibers examined, as they were not pre-soaked in Fn. This illustrates the importance of surface properties and receptor mediated adhesion in MSC settling and growth. Despite the fewer numbers of cells, every cell observed was indeed elongated in the direction of the fiber's long axis.

Images of much more flattened cells that are also extending in the same direction as the fiber are reproduced in Figure 3-5. These cells may be similar to "sheath cells" found in tendon experiments on fibers [30,114];. These studies that found three populations of






50



20
60 JLM








A


E








Figure 34. Examples of RMSC growth on the surface of 9-0 nylon fibers. A.) shows a cell with two long processes essentially parallel to the fiber. B and Q show a cell and a close up of the same cell respectively. D) shows a cell that is much flatter. E.) a close up of the same cell. It is elongated in the direction of the fiber as were all the cells seen on the fiber.

cells grew on their fibers, one a flat sheath like cell, similar to these, that covered the surface of the fiber, another spherical cell type like the one seen in Figure 3-5A and a spindle shaped cell similar to that seen in Figure 3-6. These cells may be very similar as they are derived from the same cells in vivo as the MSCs used in this study. Morphologically the cells in the current study appear very similar. 10-0 nylon sutures

In contrast to the 4-0 fibers, 10-0 fibers show cells that follow the long axis of the

fiber very closely (Figure 3-6). Figures 3-6A, 3-613 and 3-6C show more flattened cells, while 3-613, 3-6E and 3-6F show cells with a more spindle shaped morphology.






51


vow,












............


A FB]


Figure 3-5. Examples of RMSC growth on the surface of 9-0 nylon fibers. This set of images shows flattened cells that are at least qualitatively extended in the direction of the 9-0 fiber they are growing on. On the right is a close up of the cell material showing many filopodia.

Maxon sutures

Images of 5-0, 6-0 and 7-0 sutures were obtained on samples cultured for 12 days under conditions described above. 5-0 fibers were fully covered with MSCs that show spiraling pattern (Figure 3-7). Cells on 5-0 fibers were flattened in a manner that was much more pronounced than that seen on the 6-0 (Figure 3-8) and 7-0 fibers (Figure 3-9). Additionally, cells on the 6-0 and 7-0 fibers were much more elongated than those seen on the 5-0 fibers. 7-0 fibers appeared to exhibit the greatest amount of multi-layering and elongation. These results are supported by the quantitative data taken to measure cellular orientation and elongation, though SEM of these fibers was performed only on the twelfth day.






52






































E A/ F


Figure 3-6. SEM images of 10-0 nylon sutures at various magnifications. A, B and Q show highly flattened and multi-layered cells that are elongated in the direction of the fibers. D, E and F) show a more spindle shaped cell that also is elongated in the same orientation as the fiber.






53






















Figure 3-7. 5-0 Maxon sutures exhibiting RMSC adhesion and growth. A.) Shows MSCs covering the entire surface of the fiber and marked spiraling, or angling with respect to the fiber's long axis. Note cells are rather flattened and not highly elongated. 500x. Arrow indicates general trend of the spiraling behavior exhibited by cells. B) Again shows flattened and angled MSCs, but at I 000x. Q Flattened and angled MSCs at 1500x. Light Microscopy Studies of Maxon

In general the larger diameter 5-0 and 6-0 fibers showed the barber pole spiraling similar to that seen in the SEM images of 4-0 nylon (Figures 3-2 & 3-3), as well as SEM images of 5-0 Maxon (Figure 3-7). 7-0 Maxon however did not show noticeable evidence of this spiraling behavior. Elongation of RMSCs on the surfaces of the fibers was evident and contact guidance of cells was clearly occurring. Differences in nuclear orientation between fiber diameters were visible and confirmed by statistical analysis. Multilayering occurred on all three fiber diameters by day 12. These findings are consistent with SEM data, which is presented below.






54
























Figure 3-8. 6-0 Maxon sutures exhibiting RMSC adhesion and growth. A) shows MSCs covering the entire surface of the fiber. Many flattened cells are present and much more elongated than those seen on the 5-0 fibers. Spindle shaped cells are present as well (blue arrowheads) 500x. B) Shows flattened and elongated MSCs at l000x. C) Shows flattened and elongated MSCs at I 500x. An example of multilayering is clearly evident (red arrowhead) as cells overlap.














Figure 3-9. 7-0 Maxon sutures exhibiting RMSC adhesion and growth. A.) Fibers are completely covered by flattened and elongated MSCs 500x. 13.) shows flattened and elongated MSCs at 1 000x. Examples of multilayering are evident as cells overlap each other (red arrowheads).






55













Figure 3-9 Continued. 7-0 Maxon sutures exhibiting RMSC adhesion and growth. Again flattened and elongated MSCs are evident. Multilayering and overlap are evident (red arrowheads) 1500x.

Nuclear form factor

Figure 3-10 shows a nucleus with all NFF measurements as they were performed in Adobe Photoshop. Originally three day data was slated to be collected, however, there was not sufficient cell growth for any kind of meaningful analysis.


















Figure 3-10. Example of Nuclear form factor measurements. The image shows a representative 5-0 fiber after 6 days in culture.

Figure 3-l1 is a graph of the effects of diameter on NFF independent of time, while Figure 3-12 shows effects of time independent of diameter. Clearly both factors, which were the factors used in the 2 way analysis of variance, show that NFF leads to increased






56


orientation. All differences resulting from diameter were statistically significant with a power P = <0.05. When comparing time factors of 6 days in culture vs. 12 days, there was a statistically significant difference with a power of P<0.05, while the other differences were significant with P = <0.1I (i.e. 6 days vs. 9 days, and 9 days vs. 12 days).


UaTeter Effects on N~iewn (Oien~aon (F

(InepndrXW of lime)
0.35


0.30- lg0T


0.25





0.15


0.10


0.05


0.00
5-0 6-0 7-0

&u'osize(LIsp. sz

Figure 3-11. Graph showing the independent influence of fiber diameter on NFF. Bars are standard errors.

All of the data in Figure 3-13 has been shown to be statistically different from each other at a power of P = <0. 1. There appear to be effects on cellular elongation that are related to social interactions of cells, or how they squeeze together when the layers become confluent. 6-0 and 5-0 fibers show what seem to be fluctuations in the degree of







57


Effect of Culture Time on Nuclear Orientation (NFF)
(Independent of Diameter)
0.4




0.3




0.2



OAN N
0.1





0.0
4 6 10 12 14
Tim. in culture Idavsl

Figure 3-12. Graph showing the independent influence of time on NFF. Bars are standard errors.





Effects of Culture Time on 0.4- Nuclear Orientation (NFF)

5-0 NFF MM6-0 NFF 7-0 NFF
0.3




0.2




0.1




0.0
Dav 6 Dav 9 Dav 12
Time in Culture (Days)


Figure 3-13. Graph showing the effects of time on NFF for all diameters studied. Note 70 fibers are the only ones showing a gradual and continuous increase in elongation. Both 5-0 and 6-0 fibers fluctuate with time probably due to the effects of social interactions. Bars are standard errors.





58

nuclear elongation. Multilayering was apparent on all the constructs and imaging of nuclei for analysis was primarily done on what appeared to be the top layer cell coverage. Deeper layers suffered from a loss of contrast thus making it virtually impossible to collect data from them. Fluctuation in nuclear eloRgation as seen in Figure 3-13 may indicate how social interaction is important in cell orientation. Basically each layer of cells, as it becomes confluent on top of the layer before it, may elongate and orient as it nears full confluence. This is followed by the accumulation of the next layer, which is less elongated at first. This is analogous to the phenomena seen by Ricci et al. who observed that tendon outgrowth cells did not elongate until confluence was reached on fibers above 100 pm [ 113].

In the similar vein, 7-0 fibers showed increasing elongation throughout the culture

times examined. These fibers were less than 100 un, the diameter that seems established in the literature as the critical diameter for continuous contact guidance. Nuclear angle analysis

Figure 3-14 shows the same nucleus as seen in figure 3-9. This time the image is

demonstrating the method used to measure nuclear angle. Figures 3-15 and 3-16 show the main effects of the 2 way ANOVA (time and diameter). Both factors exhibit a significant effect with p = <0.00 1, while the interaction between the main effects is significant with p = <0.056. Overall both time and diameter influence the angle of the nucleus with respect to the long axis of the fiber. Also there is an interaction between time and diameter leading to an influence on nuclear angle that is greater than each factor acting alone.






59




















Figure 3-14. Example of nuclear angle measurements. Image of the same nucleus depicted in Figure 3- 10. This micrograph demonstrates the method used to measure the nuclear angle with respect to the fiber's long axis.

Nuclear Angle Effects of Fiber Diameter 100 (independent of Time)


0
1060
I
0
Zu 20

0



5-0 6-0 7-0
Diameter (U.S.P. suture size)

Figure 3-15. Graph showing the effects of diameter independent of time. Bars are standard errors.

Figure 3-17 shows all fiber sizes vs. time in culture. With the exception of the 6-0, 6 day data, nuclear angle behaved exactly as hypothesized by increasing (becoming closer to alignment with the fiber) with diameter and time. When comparing figure 3- 1OC to






60


Nuclear Angle Effects of Time 100 (independent of Diameter)



80
0


40


0


Day 6 Day 9 Dayl12
Time (days) Figure 3-16. Graph showing the effect of time in culture on nuclear orientation as measured via nuclear angle. Bars are standard errors.

Corrected Nuclear Angle Effects of Diameter and Time
100 5-0 fibers

6-0fier
7- ibr
LO
"a 60
2
~40

Z 20


Day 6 Day 9 Day 12

Time (Days)


Figure 3-17. Graph showing the effects of time and diameter together. Bars are standard errors. Figure 3-17 it becomes apparent that NFF and nuclear angle are not directly correlated, which is interesting, as these values were calculated from the very same nuclear images.






61

So nuclear orientation appears to be influenced by the diameter of the fiber in a way that is independent of cellular elongation. Again we see that 7-0 fibers exhibit the greatest influence on cellular activity and lead to elongation that is more robust than either the 60, or 5-0 diameter fibers.

Overall it seems that nuclear orientation measured by the angle of the nucleus is a more predictable method of examining cellular contact guidance. Nuclear orientation looks as if it will require further study and increased understanding.

Discussion

It is abundantly clear that decreasing diameter and increasing culture time led to

greater contact guidance of MSCs in this culture system. These responses by MSCs were exactly as hypothesized. 5-0 and 6-0 diameter fibers showed what is likely to be the effect of the social interaction of cells on their nuclear elongation. 7-0 fibers on the other hand are of small enough diameters that their topographical influence overpowers the need for social interactions. All fiber sizes show a nuclear orientation, however, this orientation appears to be independent of nuclear elongation.

This behavior has implications for fiber-based scaffold design. By quantifying the behavior of these MSCs on various diameters and for various times, the fiber candidate that is most optimal for a given system can now be chosen. Though similar studies have been performed with similar cell types [29,97] an extensive literature review shows they have never been done using the MSC culture system. In addition the level of measurement in this study, using quantitative methods has not been matched for time, or numbers of quantitative measures (i.e. both NFF and nuclear angle).

Though the presence of social interaction between cells has been observed on a qualitative level, it has never been observed and measured as closely as this study.






62


Fluctuations in the nuclear elongation, independent of the angle suggest that as confluence is reached each layer of cells in the multilayered system becomes more and more elongated. This is to say that layers of cells are built up by a period of oriented, but un-elongated cellular morphologies followed by elongation in the direction of the fiber's long axis. After the achievement of confluence, the process begins again on the next level of multilayer.

Larger size fiber diameters show spiraling and as seen in figures 5, 6 and 8. This

spiraling may occur at equal and opposite directions within each layer. ECM~collagen has been linked to cellular orientation and in this cell type it is seen that collagen layers are deposited in orthogonal directions to each other [114-116]. Perhaps this mechanism of cellular layering is related to the mechanism that lays down collagen in alternating layers. As previously noted orientation of ECM is strongly correlated to the orientation of the cells depositing it [31,76,117-120]. Indeed the deposition and orientation of many individual ECM proteins including collagen has been shown to parallel the oriented layers of osteoblasts laying them down [ 12 1].



Conclusions

This study quantitatively shows that RMSCs elongate and orient in response to diameter, as well as time in culture. In the case of both time and diameter contact guidance occurred in an increasing manner when measured via nuclear angle. A critical fiber diameter existed at approximately 100 p.m, however, for the cellular elongation measured by NFF. Below this critical diameter, cells were increasingly oriented in a gradual manner over time. Though hen grown on fibers above the critical diameter, cells






63

were oriented by nuclear angle, but elongation seemed to occur in response to social interactions between cells as each layer achieved confluence.

Overall, given the nature of this cell type and the way it has been shown to orient ECM in a manner that parallels cellular orientation, this form of contact guidance will allow the control of structural ECM proteins. As mentioned in chapter 2, control of ECM proteins; namely collagen, is what creates the anisotropic mechanical properties found in natural bone. This study shows, through rigorous scientific methods coupled to strong statistical support, that controlling cellular orientation and thus ECM orientation is now a capability to be included in the arsenal of design characteristics available to connective tissue engineers.













CHAPTER 4
CELLULAR BRIDGING PHENOMENA

Introduction

The overall goal of this work is to develop an understanding of cellular organization and ECM deposition, then use this knowledge to organize cells and ECM in a desired manner. Developing tissue constructs with multiple layers of cellular and ECM orientation appears to be a promising method of increasing the construct's anisotropic mechanical strength. As reviewed in the background, mechanical anisotropy is a common strategy employed by tissues as part of their physiologic function. This is particularly true of bone and connective tissues, though this fact appears to have been overlooked in the tissue engineering literature when scaffold materials are chosen. Typically randomly organized scaffold materials have been employed despite the fact that substrate topography has been shown to strongly influence cellular orientation and ECM deposition. These induced forms of cellular and ECM organization is very similar to the structure seen in vivo, which is the very same structure and organization responsible for mechanical anisotropy.

In 2-1) MSC systems, cells multi-layer and act cooperatively to produce a collagen

rich ECM, which is organized into orthogonal layers. Orthogonal lamellae of collagen are separated by cellular layers, which are responsible for collagen deposition. Originally this cellular multilayering provided the rationale for the multi-fiber construct experiments discussed in the 5th chapter using 30 ttn as a spacing distance between fibers.




64





65

Multilayering phenomena in MSC culture, by estimates and projection from data in the literature, achieve heights of approximately 15-20 pLm, so 30 pm would be proper for cells on adjacent fibers to grow and eventually join in the space between fibers (Figure 4-1).


















Figure 4-1. Multilayering on adjacent fibers leading to interactions due to their proximity. Possibly these interactions are of a paracrine nature. The correct spacing would allow cells to span the distance between fibers, while using the fibers as a stimulus for cellular and ECM organization.

In the course of the experimental work it became apparent that there was at least one more phenomenon involved in the 3-D cellular organization process beyond that of mere multi-layering, which appears to be the primary process in 2-D systems. The single fiber constructs studied in the last chapter primarily demonstrated MSC multilayering, but when more than one physical surface, or topographical feature is present, another cellular process becomes evident. Cells begin to span distances, or "bridge" between surfaces when these surfaces are found within certain ranges and under the right conditions of proximity.





66

This bridging phenomenon is obviously a prominent behavior contributing to the

cellular interaction with 3-D scaffolds. Furthermore this behavior does not appear to be explicitly addressed in the literature as it currently stands. For both these reasons cellular bridging merits study here, not only to elucidate and characterize it as a previously unobserved phenomenon, but also to understand how it influences cellular development and function.

Materials

Bridging was induced and observed on many different materials including: standard polystyrene culture flasks (CorningTM9, 4-0 monofilament nylon sutures (Ethilon TM), 7-0 monofilament polyglyconate sutures (Maxon TM ), stainless steel screen and bioactive glass fibers synthesized as previously described by Dominguez et al. [ 122]. Culture materials were identical to those used in chapter 3.

VonKossa staining was performed using Silver nitrate 5%- 5 gmn in 100 ml DI water =5 % solution. Sodium thiosulfate 5% 5 gm in 100 ml DI water = 5 % solution nuclear fast red- 5gm aluminum sulfate, 100 ml DI water, nuclear fast red 0. 1 gin.

Methods

Culture Methods

Culture and MSC seeding was performed as described in chapter 3. Nylon sutures were cultured for 12 days, Maxon TM for 13 days and Bioactive glass fibers 6 days before data collection.. Polystyrene studies were performed for 11I days total with data collected on days 7, 8, 10 & 11. Stainless steel screens were followed for 23 days with data collected on days 10, 12, 15, 18 &23.





67


SEM Imaging

SEM was performed on Nylon, Maxon TM, stainless steel and bioactive glass fibers with preparation for microscopy as described in chapter 3. Light Microscopy Studies

Polystyrene imaging was performed on live cells in situ as they grew in the culture dish. No staining was performed so viable cells could be maintained allowing the development and documentation of individual bridges. Von Kossa Staining

Von Kossa staining was performed on stainless steel screens to follow the

development of ECM and its subsequent mineralization. This method demonstrates salts of calcium and it works by replacing Ca +2 ions with silver from a silver nitrate solution (5% in DI water). The silver salt that is produced by this is then photosensitive (silver nitrate is the same chemical used in the production of a black and white photographic negative). Exposing the sample to light then causes a photic reduction of the silver salts leading to a black precipitation of silver in the locations where calcium is present.

Von Kossa staining procedure[118,120,1231

1.) Absolute EtOH dehydration, 2 times for 2 minutes each.

2.) 95% EtOH dehydration, 2 times for 2 minutes each.

3.) DI water rinse 2 times.

4.) Silver nitrate 5% incubate 30 minutes under high power light (- 100-150 Watt

bulb leading to photic reduction of silver).

5.) DI water rinse

6.) Sodium thiosulfate 5% incubate 3 minutes (washes away residual silver nitrate)

7.) DI water rinse






68

8.) Nuclear Fast Red incubate 5 minutes (stains nuclei dark red and cytoplasm

pink).

9.) 95% EtOH dehydration, 2 times for 2 minutes each

10.) Absolute EtOH dehydration, 2 times for 2 minutes each. Proliferation Studies

Bridging of MSCs between two topographical features is logically dependent on the distance between them. In an effort to understand how spacing between fibers effects



9.17 E0.05 mm






















Figure 4-2. Example of bioactive glass fiber constructs used for the proliferation study. Dimensions of these cylindrical containers allowed the calculation of fiber density and provided a constant volume necessary for the study. cell function we performed proliferation studies using bioactive glass fibers with various packing densities. Various masses of bioactive glass fibers (10, 20, 30 & 40 mg) were gently packed into constructs with a given volume (Figure 4-2). Increasing quantities of





69


fibers within this given volume was assumed to decrease the distance between fibers and thus influence the ability for these cells to perform bridging.

Cells were seeded onto all constructs at an estimated density of 1.10 +!- 0.07 x10 cells/mi and allowed to incubate for 3, 7, 10, 14 and 21 days. As each time point came due, cells were removed from culture, trypsinized and counted using a Beckman-Coulter Multi-sizer III cell counter. Sample sizes were 500 mL and were diluted with Isoton 11 diluent.

Cell counts were analyzed using a two way analysis of variance with time (3, 7, 10, 14 and 21 days) and fiber density (10, 20, 30 & 40 mg) as factors. Post hoc testing was performed using Tukey's multiple comparison.

Results

Light Microscopy of Bioactive Glass Fibers After 6 Days in Culture

Preliminary studies on bioactive glass fibers placed directly into the RMSC culture system for 6 days showed a marked degree of multi-layering in the vicinity of the fibers where they rested on the floor of the dish (Figure 4-3). This degree of multi-layering was particularly evident when comparing cellular behavior on flat areas of the dish where fibers were not present (Figure 4-4). In fact multi-layering was so extensive that four micrographs were necessary to include the entirety of the representative image (Figure 4-3 montage).

In addition to multi-layering in the vicinity of the fibers, RMSCs were oriented toward the fibers as if using it as an anchor for the development of tension.






70



















Figure 4-3. Montage of light micrographs portraying the cellular interaction with bioactive glass fibers placed in standard culture well. Note orientation of cells with respect to the fibers, they appear to be pointed in toward the fibers and multilayering is extensive.










_4






Figure 4-4. Light micrograph of area without fibers in the same culture well as that shown in Figure 4-3. Note the lack of multi-layering particularly when compared to Figure 4-3.






71

Scanning Electron Microscopy

Bioactive glass fibers after 6 days in culture

Figure 4-5 shows SEM images of RMSCs grown on bioactive glass fibers. Figure 45A shows cells covering the fibers and growing in a robust manner on the fiber surfaces. In some cases cells grew over fibers, securing them to the bottom of the petri-dish. Figures 4-5 shows a close up example of cellular interactions with the fibers and the dish. Figure 4-6 shows examples of fiber to fiber bridging, a phenomenon that was commonly seen throughout the extent of the specimen's surface. This bridging was often multi-cellular as seen in Figure 4-5, but there were many instances of unicellular bridging as well (Figure 4-6).




















Figure 4-5. Examples of cellular growth on fiber placed in the RMSC system taken with SEM. A.) shows a large aggregate of cells and fibers with the same type of orientation seen under light microscopy. B.) shows a large multi-cellular bridge that spans the distance from the bottom of the dish to the fiber.

Figure 4-513 demonstrates bridging from the dish floor to fibers, another commonly seen cellular behavior. Again the bridge demonstrated is multi-cellular, though there






72

were unicellular examples of this type of bridging as well. This multi-cellular aggregate demonstrates a more columnar, or cylindrical bridging motif, in contrast the the flattened, sheet-like bridging was also seen. Both structures were clearly multi-cellular with their respective morphologies determined by what appears to be the topography, or the shapes presented by the fibers in their vicinity.


I
A










Figure 4-6. Examples of unicellular bridging on bioactive glass over distances of -70 gm. In addition to being unicellular these are also fiber to fiber bridges. Polymer l1bers

Figure 4-7 shows micrographs of RMSCs bridging distances between two adjacent Nylon fibers. These constructs were incubated for 12 days prior to SEM examination and demonstrate what appears to be a multi-cellular bridge (left most bridge) and a unicellular bridge (right most bridge).

Figure 4-8 is a micrograph of RMSC bridging on adjacent Maxon fibers. Though this particular bridge is shorter than that seen on the bioactive glass fibers and the Nylon fibers, bridges were not limited to shorter lengths. This construct was incubated for 13 days in RMSC culture.






73


















Figure 4-7. Examples of unicellular bridging between nylon fibers. This image demonstrates what appears to be a multi-cellular bridge (left most bridge) and a unicellular bridge (right most bridge).

















Figure 4-8. Bridging between Maxon fibers. Light microscopy of Bridge Development on Polystyrene Culture Flask

In order to classify, or identify the sequence of events involved in the development of MSC bridging, RMSCs were seeded in standard polystyrene culture flasks and individual bridges were observed and documented serially over time.






74

Initially after seeding there were many individual cells that had elongated along the flat surface of the dish, sometimes for hundreds of im. An example of this elongated cellular morphology is shown in Figure 4-9A (green arrow). These highly elongated cells were interspersed with cells that had assumed more flattened morphologies. In addition to elongation on flat surfaces, there were also cells that elongated and had associated with the topography of the polystyrene wall similar to the cellular behavior shown in Figure 4-913 (red arrow). Initial stages of bridging, as demonstrated by single cell bridging, indicated that bridging was possible only when there was an actual physical pathway for bridge formation. It appeared that elongation of cells, which can occur on the order of

















Figure 4-9. Initial RMSC elongation. A.) Representative micrograph of highly elongated cells that seemed to be responsible for initial unicellular bridging. hundreds of microns facilitated bridge formation by traveling the distance initially. In other words it did not seem possible for cells to reach across gaps without there being a pathway for them to elongate on while constantly maintaining adhesion and contact (Figures 4-6 and 4-9).






75












4X








Figure 4-9 continued. Initial RMSC unicellular bridging. B.) shows the formation of a unicellular bridge between the floor of the culture dish and the wall of the dish (red arrow).

Bridge development progressed between the floor of the culture flask and the wall. By day 7 large multi-cellular bridges were apparent similar to the one shown in Figure 4-10. Bridges extended for hundreds if not thousands of microns and were clearly multicellular in nature. The areas underneath the bridge, that is the areas of the floor between the site of attachment at the wall and where the bridge was secured to the floor of the dish exhibited few cells. It seemed as if the cells that had once inhabited that region of the dish

had released and become part of the bridge as it pulled off the floor and suspended itself in the medium. This reduced number of cells is apparent in Figure 4- 10 as well. With continued time in culture the cells composing the bridge showed evidence of proliferating within the bridge proper. This was apparent in a thickening in the bridge near its site of attachment to the flask wall (Figure 4-11, red arrow).






76















Flask Wall Flask Floor

Figure 4-10. Multicellular bridge at 7 days in the RMSC culture. Bridge is extending from the floor of a standard polystyrene culture flask to the wall. On the 10th day in culture, there were instances of ECM nodule formation within the body of the bridge in the location where the thickening of the bridge had occurred (Figure 4-12). These nodules were very similar morphologically to those seen on the flat surface of other areas of the culture dish. On the 11Ith day in culture the bridge contracted back from the wall of the dish and formed a large multi-cellular nodule that appeared to be contain a large fraction ECM as part of its composition. Again this












Figure 4-11. Multicellular bridge at 8 days in the RMSC culture. Bridge shows proliferation and thickening within its body. This is the same bridge shown in figure 410.






77











Al
Figure 4-12. Multicellular bridge at 10 days in the RMSC culture. Nodules appear to be forming within the body of the bridge at the site of proliferation. nodule was morphologically very similar to those seen in flat areas of the dish, however, this nodulewas much larger, on the order of hundreds of microns versus approximately fifty microns for nodules on the flat surface. Bridge Development on Stainless Steel Screens (Light Microscopy and SEM)

After classifying, or identifying the sequence of events involved in the development of MSC bridging, RMSCs were seeded onto stainless steel screens and the general















Figure 4-13. Multicellular bridge at I11 days in the RMSC culture. Bridge contracted back from the wall of the dish and formed a large aggregate of cells and ECM. This nodule was morphologically identical to others seen in the center of the dish where no bridging occurred, only much larger.





78

development of bridges was observed and documented serially over time. This study was performed in order to introduce a controlled topography into the RMSC system and examine the bridging phenomenon associated with it. The screens used had a weave that led to square holes with dimensions of approximately 170 tin on a side. The fact that it was a woven fabric was also important in the overall study of fibers and their use as scaffold materials. Using screens, therefore allowed insight into fiber based scaffold applications, though the fibers composing these particular woven screens were steel.

To study the effect of bridging on the development of ECM and mineralization, constructs were stained via the Von Kossa Method. Additionally, to ensure that the bridging effects were the result of the fibers incorporated in the weave of the screen, the screens were suspended above the floor of the petri-dish.

After seeding cells were apparent on the lengths of the fibers and could be visualized at the periphery of the fibers, which, like the Nylon fibers used, were opaque. Numbers of cells increased until by the 10th day then began to form unicellular bridges, or bridges composed of a few cells (Figure 4-14). This bridging is also shown in a 3-D formation of even larger aggregates of cells that assumed flattened bridges and in some cases more cylindrical bridges (Day 12 and Day 15 images Figures 4-14, 4-16A, 4-16B, 4-17A and 4-17B respectively).

By the 18th day entire holes of the screen were covered by flat sheets of cells acting cooperatively and there was even the presence of occasional non-mineralized nodules of ECM (green arrowhead Day 18 image figure 11, also see SEM representation in Figure 4-15A and 4-15B. Continued time in culture led to the cylindrical bridges progressively






79

spanned distances across the square holes until they had completely bridged the diagonals of the square holes.




ay 10 Day 12 Day 15








Bar
-15OPM

Day I ay 23






Figure 4-14. Development of bridging on stainless steel screens. Day 10 shows unicellular bridging, or bridging with only a few cells. As time passes cells proliferate and form more complete sheet-like bridges until, by Day 18 most holes in the screen are filled with cells. At this time early nodule formation has begun, though it is unmineralized (green arrowhead). By Day 23 mineralization is highly apparent as evidenced by nodules stained black with the Von Kossa stain. By the last day of the study, mineralization was seen within the bodies of bridges that were spanning distances across the holes of the screen. Though the presence of flattened sheet-like bridges and cylindrical bridges were both apparent during the course of the study, prior to the 23rd day, flattened sheets appeared to be most dominant in numbers. Later in the study, especially day 23, cylindrical bridges became moreprevalent, which indicated that some of the flattened bridges were becoming cylindrical (Figures 4-14, 419A and 4-19B).






80








-A B







Figure 4-15. SEM of stainless steel screens after 10 days in RMSC culture. Unicellular bridging is prominent.














AB

Figure 4-16. SEM of stainless steel screens after 12 days in RMSC culture. Bridging becomes more noticeably multicellular with cells that are remarkably flattened.

This behavior may be related to the contraction of cells as was seen later in the

bridging sequence documented in the polystyrene flasks. An example of a cylindrical bridge is shown in Figures 4-14 and 4-14B. At the base of the cylindrical bridge seen in Figure 4-14B is a mineralized nodule (red arrow). This was verified visually by light microscopy prior to SEM preparation (data not shown).






81














A B

Figure 4-17. SEM of stainless steel screens after 15 days in RMSC culture.












A

Figure 4-18. SEM of stainless steel screens after 18 days in RMSC culture. Proliferation Studies of RMSCs on Bioactive Glass

Two way ANOVA showed both main factors were significant with P< 0.001.

Growth curves for 20 mg, 30 mg and 40 mg constructs displayed a very similar growth kinetic, which was not a strict sigmoidal curve (Figure 4-20). The 10 mg construct on the other hand displayed traditional sigmoidal growth, in fact this fiber density was the only density that showed a statistical difference between cell counts taken on day 7 and day 10 (P<0.05).






82


















Figure 4-19. SEM of stainless steel screens after 23 days in RMSC culture.

Cell counts for the 10 mg construct were very close to those from the 20 mg construct for much of the time period measured. In fact the differences in cell concentration between these two fiber densities did not become statistically significant until the 2 1 st day (P<0.05). That there are differences between the shapes of the two curves is apparent, however, indicating some sort of critical difference, or cut off point between 10 mg and 20 mg. Of all the curves only the 60 mg construct showed a distinct plateau at the end of the 21 day study indicating it was the only one that had finished proliferating (i.e. day 14 data was not significantly different from day 2 1). In addition to this, the 60 mg curve was markedly higher than any of the other experimental conditions indicating not only that the cells finished their replicative stage sooner, but also that the proliferation had been more robust on this fiber density. On the final day of this study the 20 mg construct displayed a cell count very close to that of the 40 mg construct the two were not significantly different at this point, and it appeared to still be proliferating. Overall the 60 mg curve provided the most robust example of the characteristic growth kinetic and had the highest cell counts as well.





83




RMSC Growth on 77s Bloactive Glass Fibers of Various Densities
2500
--10 mg of Fibers
-0- 20 mg of Fiber's
Z 2000 -w 40 mg of Fibers
--v-- 60 mg of Fibers 0U 0mg of Fibers

1500
0

C 1000
0

=500

0

0 510 15 20 25
Time (days)

Figure 4-20. RMSC growth curves for each fiber density. Star indicates site of 1 st plateau where Schmidt et al. show maximal osteocalcin concentration, the characteristic late differentiation marker. # indicates time point where Schmidt et al. found alkaline phosphatase, the early differentiation marker was highest.

Discussion

The potential of fiber based scaffolds in bone replacement is extensive. Flexible mechanical properties, high porous volume, designable woven architectures, all contribute to the attractiveness of fiber based scaffolds in connective tissue engineering. Single fibers influence cell orientation and elongation as shown in chapter 2. This behavior is very similar to that seen on flat 2-D surfaces with microtopographies of given dimensions [29,31,32,89]. This cell type leads to multi-layering on flat surfaces, a cellular behavior that is essential to the formation of ECM nodules and the





84


mineralization of these nodules [80,115,116]. Similar multi-layering behavior is seen on single fibers leading to similar nodule formation and mineralization.

Multi-layering and nodule formation are functions of many cells, or aggregations of cells, which communicate with each other [104,124] and cooperatively perform. How these cells grow on single fibers, as well as flat micropattemed surfaces is similar in that there is only one substrate for them to interact with, either a fiber, or a flat surface. With the introduction of more than one surface, or macrotopography the cellular response becomes more complex. In this context macrotopographies can be thought of as those created by multiple surface with features on the size order of the multi-cellular aggregates, while microtopographies are those that act at the level of the individual cell.

Bridging acts at the macrotopographical level, the level that 3-D scaffolds necessarily exhibit. Single fibers lead to a microtopographical response, but where two fibers contact, a macrotopographical feature exists. In this sense, 3-D fiber-based scaffolds are hierarchical, combining not only microtopography where cells are in contact with only a single fiber surface, but also macrotopography features where two fibers interact and bridging occurs. Therefore, an understanding of single fiber effects on cell function (i.e. multilayering), as well as multi-fiber effects (bridging) is essential for future fiber based scaffold design, or any other hierarchical 3-D scaffold design. This is particularly so in light of the fact that tissues are 3-D and these interactions, or mechanisms are obviously part of the tissue level machinery necessary for the development of bone, or ligaments. Bridging seems to progress in a generalizable developmental sequence of events as is shown in Figure 4-21. Single cell bridging occurs first (Figure 4-21 A) as a cell encounters macrotopographies like the wall of a culture dish, or the intersection of two





85


fibers. An example of this is shown in Figure 4-9 of the results. Multi-cellular bridging follows as more cells become involved either as they are lifted from their surface in a sheet-like manner, or proliferation that occurs within the body of the bridge itself. Both behaviors appear to be present as seen by the sparsity of cells below the bridge in Figure 4-10, as well as the thickening of the same bridge over time. Eventually multi-cellular



A]








FD F]






Figure 4-2 1. Schematic generalization of multicellular bridging. structure exhibits ECM nodules within its body (Figure 4-2 1 C) that eventually mineralize as seen when MSCs were grown on the stainless steel screens. In the case of the polystyrene culture dish the bridge contracted until it had pulled from the side of the dish and formed a large nodule of ECM and cells (Figure 4-2 1 D).

The context of this bridging phenomenon is to be found in many areas of the literature. Cells of mesenchymal origin grown on 2-D surfaces with [31,32,73,74,89,94,125], show that proliferation and differentiation are strongly influence by the ability for these cells to adhere to their substrate. Adhesion is





86

influenced by the factors of surface roughness [90,121,126,127], or topography [31,32], surface chemistry [128] and surface energy [129].

In addition to the relationship between adhesion and the cellular ability to develop tension, there is the effect of mechanical stretch on cellular activity. Stretch applied externally to MSC leads to increased differentiation, as well as an orientation of ECM in the direction of the force [101]. Similarly, inhibition of chloride sensitive stretch receptors causes a decreased response to topography, drawing a link between the topography and the cellular need for stretch, or tension [100]. It is interesting to note how the strength with which a cell adheres and develops tension, as well as an applied stretch stimulus both result in similar cellular responses in the form of proliferation and differentiation. Overall it seems tension generation is key for proper proliferation and differentiation of mesenchymally derived cells on a 2-D surface.

Though 2-D culture systems provide a glimpse of the cellular machinery involved in proliferation and differentiation, tissues are 3-D structures composed of cells. It seems therefore that the cellular process of tension generation is designed to operate in the 3-D environment, a theory that is bom out by many examples in the literature. Studies performed using collagen gels as scaffolds for cells of mesenchymal origin show that increased stiffness of the scaffold leads to an increased rate of proliferation. In addition to this increased rate the duration of proliferation persisted longer than that seen in scaffolds with less stiffness [80]. In similar studies collagen gels were anchored on one axis leading to a preferential tension development in the direction of this axis by fibroblasts. Cell numbers increased by 5 times in the anchored gels and ECM generation was greater than that seen in the un-anchored gels, which in contrast demonstrated a 5