Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Université de Sherbrooke
Faculté de génie Département de génie chimique
DEVELOPMENT OF BIOACTIVE, POROUS SCAFFOLDS FOR THE VASCULARIZATION OF
ENGINEERED TISSUES
DÉVELOPPEMENT D’ÉCHAFAUDAGES POREUX, BIOACTIFS
POUR LA VASCULARISATION DE TISSUS ARTIFICIELS
Thesis submitted for the degree of Doctor of Philosophy Chemical Engineering
__________________________________________
Hanna BRAMFELDT
Sherbrooke, Québec, CANADA February, 2008
I
ABSTRACT
The lack of vascular networks to support the growth, function, and survival of three-
dimensional tissues represents one of the greatest challenges of tissue engineering today. This thesis
project was designed to constitute the first part of a work whose final objective is the creation of
functional micro-vessel networks within three-dimensional scaffold constructs for tissue engineering.
Specifically, this part was aimed at creating functional scaffolds, that may be used in the continued
studies towards achieving this long-term goal.
As a first step, three co-polymers on the form poly(ε-caprolactone-co-D,L- lactic acid)-
poly(ethylene glycol)-poly(ε-caprolactone-co-D,L- lactic acid) were synthesized to provide
mechanically stable biomaterials with controllable and tunable material properties. To achieve this, the
monomer content in the side chains was varied and the resulting co-polymers were characterized using 1H-NMR, GPC, DSC, WAXS, and DMA1. Furthermore, the hydrolytic degradation profiles of samples
fabricated from melt-pressed films were studied over six months, as well as the influence of
degradation on tensile strength. Based on obtained results and observations, two of the co-polymers
were judged suitable, in terms of mechanical integrity and degradation profile, for further
investigation.
To improve cell/polymer interactions, co-polymer films were modified by protein
immobilization using imidoester chemistry. The effect of several protein preparations on smooth
muscle cell (SMC) adhesion and proliferation was studied over six days in culture. Cells were counted
at pre-set intervals, and further analyzed by scanning electron microscopy and fluorescent labelling of
the differentiation marker α-actin. Immobilized proteins greatly enhanced adhering cell numbers,
although for stimulation of SMC proliferation covalent immobilization was superior to physisorption.
1 1H-NMR: 1H-nuclear magnetic resonance; GPC: gel permeation chromatography; DSC: differential scanning calorimetry; WAXS: wide-angle x-ray spectrometry; and DMA: dynamic-mechanical analysis
II
RÉSUMÉ La construction d‘un réseau vasculaire pouvant assurer la croissance, la fonction et la survie de
tissus tridimensionnels représente actuellement l’un des défis les plus importants du génie tissulaire.
Cette thèse fait partie d’un vaste projet ayant pour but de développer des méthodes de culture de
micro-vaisseaux sanguins au sein d’un échafaudage tridimensionnel synthétique qui pourra servir en
génie tissulaire en conditions de bioréacteur.
Trois copolymères composés de poly(ε-caprolactone-co-D,L-acide lactique)-PEG-poly(ε-
caprolactone-co-D,L-acide lactique) ont été synthétisés par la polymérisation par ouverture de cycle. La
composition monomérique a été variée et les polymères obtenus ont été caractérisés par 1H-résonance
magnétique nucléaire, chromatographie de perméation de gel, calorimétrie différentielle à balayage,
diffraction des rayons X aux grands angles, et analyseur mécanique dynamique. De plus, la
dégradation hydrolytique des échantillons, fabriqués par moulage à l’état fondu, a été étudiée, ainsi que
l’impact d’une telle dégradation sur l’endurance à la tension mécanique de ces matériaux. Considérant
les résultats observés au niveau des propriétés mécaniques et leur évolution au cours des tests de
dégradation, deux des copolymères ont été sélectionnés pour être étudiés davantage en lien avec leur
possible utilisation en cultures cellulaire et tissulaire.
Dans l’objectif d’améliorer les interactions cellules/matériaux, des films composés de ces deux
copolymères ont été modifiés en immobilisant des protéines à leur surface. L’effet de plusieurs
protéines sur le comportement de cellules musculaires lisses (smooth muscle cells, SMC) a été étudié
au cours de cultures allant jusqu’à six jours. Les cellules ont été comptées à des intervalles fixes, et
caractérisées par la microscopie électronique à balayage ainsi que par le marquage fluorescent du
marqueur de différentiation, α-actine. La présence de protéines a amélioré de façon significative les
interactions cellulaires, notamment au niveau de la prolifération qui était davantage stimulée par
l’utilisation de protéines immobilisées par liaison covalente.
Une technique basée sur des mélanges polymériques co-continus a été appliquée à nos
copolymères. Les échafaudages fabriqués par cette technique démontraient une porosité hautement
interconnectée, dont les tailles de pores peuvent être contrôlées par recuit. La perméabilité et la
résistance à la compression des structures tridimensionnelles ont été déterminées. De plus, l‘adhésion
de SMC à des échafaudages modifiés par protéines a indiqué une dépendance des tailles de pores.
Les résultats de cette thèse démontrent que les échafaudages fabriqués et modifiés tel que
spécifié, pourraient trouver usage dans l’ingénierie de tissus à base de SMC. Ces cultures pourraient
aussi servir à la vascularisation de tissus artificiels par voie de co-culture avec des cellules
endothéliales.
III
GENERAL INTRODUCTION
The in vitro generation of human tissues for transplantation constitutes a long-standing dream
in reparative and reconstructive medicine. Today, the regeneration of tissues such as
cartilage[1], bone[2], and skin[3] is possible with available techniques. However, engineering
(TE) of thick, metabolically demanding tissues is facing an important challenge, as the lack of
functional vascular networks, able to support a growing cell mass, has proven to significantly
limit tissue development[4]. Such tissues, including the pancreas, liver, heart, and kidney,
require intrinsic blood supply for long-term survival and development[5]. To date, the focus of
research has been placed mainly on promoting ingrowth from the pre-existing blood vessels at
the implant site into the TE construct[6]. However, less attention has been given to the
regeneration of viable tissue around pre-constructed vascular networks.
OBJECTIVES. This project was designed to constitute the first part of a work whose final
objective will be the creation of functional vascular beds within three-dimensional scaffold
constructs for tissue engineering. The main objective of the thesis was the development of
three-dimensional, bioactive, porous scaffolds that may be used for advanced tissue culture in
bioreactors. To this end, we wanted to:
1. Synthesize a group of biocompatible and modifiable co-polymers, with
controllable degradation rates and mechanical properties. Characterize and evaluate
these co-polymers for their applicability as scaffold materials.
2. Apply a biochemical surface-modification method pertaining to the optimization of
interactions between the synthesized biomaterials and cells native to blood vessels,
which may eventually serve to improve formation of mature vessel-like structures.
Smooth muscle cells, an essential component of mature vasculature[7,8] was
chosen for this purpose.
3. Fabricate bioactive, three-dimensional, porous scaffolds with highly interconnected
porosity that may withstand long-term culture conditions in a perfusion bioreactor.
To employ the surface modification technique developed to improve cell adhesion
in a three-dimensional environment.
IV
THESIS STRUCTURE. This thesis was written in a manner that substitutes the classical division
into chapters by the scientific publications that are the results of this project. Some of these
articles have been accepted for publication, whereas others are submitted for review. The
details of the publications and their current status are available in the List of Publications.
Below follows a brief description of their content.
Chapter 1, “Scaffold vascularization: a challenge for three-dimensional tissue
engineering”, consists of a literature review of advancements in three-dimensional tissue
engineering with an emphasis on in vitro tissue culture, and the vascularization of porous
scaffolds. Moreover, the effects of co-culture techniques as well as mechanical conditioning
of tissue constructs within the confines of a bioreactor are discussed. Following this
introductory chapter, the articles resulting from the work carried out within the framework of
this project are presented in Chapters 2-5.
In Chapter 2, “Characterisation, degradation and mechanical strength of poly(D,L
lactide-co-ε-caprolactone)-poly(ethylene glycol)-poly(D,L-lactide-co-ε-caprolactone)”, the
detailed characterization of three synthesized co-polymers with varying D,L-lactide and ε-
caprolactone content is described. These co-polymers were chosen due to their
biocompatibility, and the ease with which their material properties can be altered to answer to
the diverse set of demands placed on scaffold materials. The polymers were characterized
using 1H-nuclear magnetic resonance (1H-NMR), gel permeation chromatography (GPC),
differential scanning calorimetry (DSC), wide-angle x-ray spectrometry (WAXS), and
dynamic-mechanical analysis (DMA). The hydrolytic degradation of co-polymer films as well
as the influence of degradation on the tensile strength of these co-polymers is reported. In
addition, blends of these copolymers were prepared and the hydrolytic degradation profiles of
films of different composition and thickness were compared. The results of this study can be
seen in Chapter 3, “Blends as a strategy towards tailored hydrolytic degradation of P(CL-co-
D,L-LA)-PEG-P(CL-co-D,L-LA) copolymers ”.
Chapter 4, “Enhanced smooth muscle cell adhesion and proliferation on protein-
modified polycaprolactone-based copolymers”, describes a two-dimensional cell-culture
study carried out using films from two of the co-polymers that were judged most suitable for
scaffold fabrication. In an effort to create bioactive materials, primary amine groups were
introduced through aminolysis, followed by treatment with dimethyl pimelimidate. The thus
V
activated surfaces were subsequently reacted with several proteins (fibronectin, fibrinogen,
and fibrin layers), which were covalently immobilized. Co-polymer films exhibiting
physisorbed proteins were also included in the study. The effects of this surface modification
on smooth muscle cell adhesion and proliferation were investigated.
In the last phase of this work, three-dimensional scaffolds were fabricated using a
binary polymer blend technique, which resulted in highly interconnected pores and allowed
close control of pore size through static annealing. With the aim of evaluating the applicability
of such scaffolds in tissue engineering, the compressive strength and permeability of non-
annealed and annealed structures were investigated. Moreover, scaffolds of two different pore
sizes were fibronectin-modified using the method used in Chapter 4, and compared to evaluate
cell adhesion and seeding efficiency of smooth muscle cells. The results are presented and
discussed in Chapter 5, “Smooth muscle cell adhesion in surface-modified three-dimensional
co-polymer scaffolds prepared from co-continuous blends”.
CONTRIBUTIONS. During the course of this thesis, I have had important help from several
people, to whom I am very grateful. The co-author of three of the publications making up this
document, Dr Pierre Sarazin, has assisted in the execution and interpretation of results from
DSC, WAXS, and DMA. The DMA analysis itself was carried out by Dr Patrice Cousin,
Department of Civil Engineering, at the Université de Sherbrooke. Pierre Sarazin has
furthermore prepared all the binary polymer blends used for the fabrication of porous
scaffolds, and taught me the techniques to carry out static annealing, and preparation of
samples for SEM analysis. Finally, he contributed the idea to prepare co-polymer blend films
to study the influence of such compositions and of film thickness on degradation.
The testing of compressive mechanical properties was carried out at Industrial
Materials Institute in Boucherville, QC, where Dr Martin Bureau and Manon Plourde assisted
in the use of the equipment. WAXS and 1H-NMR analysis were carried out by the designated
technicians. All other laboratory and and interpretive work was carried out by myself.
VI
References
(1) Stock U, Vacanti JP. Tissue engineering: current state and prospects. Annu. Rev. Med. 2001; 52:443-51.
(2) Middleton JC, Tipton AJ. Synthetic biodegradable polymers as orthopedic devices. Biomaterials 2000; 21:2335-2346.
(3) Balasubramani M, Kumar TR, Babu M. Skin substitutes: a review. Burns 2001; 27:534-544.
(4) Kannan RY, Salacinski HJ, Sales K, Butler P, Seifalian AM. The roles of tissue engineering and vascularisation in the development of micro-vascular networks: a review. Biomaterials 2005; 26:1857-1875.
(5) Borenstein JT, Terai H, King KR, Weinberg EJ, Kaazempur-Mofrad MR, Vacanti JP. Microfabrication technology for vascularized tissue engineering. Biomed. Microdevices 2002; 4:167-175.
(6) Cassell OCS, Hofer S, Morrison WA, Knight KR. Vascularization of tissue-engineered grafts: the regulation of angiogenesis in reconstructive surgery and in disease states. Brit. J. Plastic Surg. 2002; 55:603-610.
(7) Drake CJ, Hungerford JE, Little CD. Morphogenesis of the first blood vessels. Morphogenesis: Cell. Interactions 1998; 857:155-179.
(8) Vailhe B, Vittet D, Feige JJ. In vitro models of vasculogenesis and angiogenesis. Lab. Invest. 2001; 81:439-452.
VII
LIST OF PUBLICATIONS
The thesis contains modified versions of the below publications.
• Bramfeldt, H., Vermette, P. Scaffold vascularization: A challenge for three-dimensional
tissue engineering. In preparation for Frontiers in Biomedicine.
• Bramfeldt, H., Sarazin, P., and Vermette, P. Characterisation, degradation and
mechanical strength of poly(D,L-lactide-co-ε-caprolactone)-poly(ethylene glycol)-
poly(D,L-lactide-co-ε-caprolactone). Journal of Biomedical Materials, Part A, 2007;
83A : 503-511.
• Bramfeldt, H., Sarazin, P., and Vermette, P. Blends as a strategy towards tailored
hydrolytic degradation of P(CL-co-D,L-LA)-PEG-P(CL-co-D,L-LA) copolymers. Submitted to Polymer Stability and Degradation.
• Bramfeldt, H., Vermette, P. Enhanced smooth muscle cell adhesion and proliferation on
protein-modified polycaprolactone-based copolymers. Journal of Biomedical Materials,
Part A, accepted for publication. Reference number: JBMR-A-07-0467.R1
• Bramfeldt, H., Sarazin, P., Vermette, P. Smooth muscle cell adhesion in surface-modified
three-dimensional co-polymer scaffolds prepared from co-continuous blends. Submitted to Biomacromolecules.
VIII
ACKNOWLEDGEMENTS
First of all, I would like to extend a special thank you to my supervisor, Dr. Patrick
Vermette. You have reinforced in me the values of independence and perseverance. You were
never too busy to meet or to discuss. I truly feel I have become a better scientist under your
supervision.
Dr. Pierre Sarazin, a post-doctoral fellow in our group (presently at École
Polytechnique in Montréal). I would like to thank you for everything you have helped me
with: from the technical know-how and the pursuit of perfection to the constant asking of
nagging questions and general merriment. Not only has your knowledge and clear-sightedness
motivated me and helped me to restructure and refocus on many an occasion, but I find your
philosophy on life, your curiosity, and your sense of humour very refreshing!
My very own Matej. What can I say? Without you, what would all this mean anyway?
You have supported me and challenged me, and your good mood makes my good mood. I
thank you also for all the weekends spent at home or outdoors, alone or in the company of
good friends, exploring the most remote corners of Québec and elsewhere. It has been a
constant reminder of how much more there is to life than lab bench tops and evenings spent in
the less stimulating company of primary cell cultures.
To my family: I have always been able to count on you. Thanks for cheering me on,
for pretending to listen when I talk about my work, for the cross-words and the Christmas
decorations, and for never telling me it was a bad idea to come here. Here’s to many more
years of me boring you with fascinating scientific details!
Much appreciation is extended to my friends and colleagues, the patient and fantastic
people who have endured me on a close-to-daily basis for these four years. Without you I
would have given up a long time ago. Good luck to you all in the future, whatever it may
bring!
Finally, I would like to express my gratitude to Drs. Pentti Tengwall and Fredrik Höök
for first inspiring in me the courage to enter into the world of research, and for being the
scientific and personal models to which I shall always aspire.
TABLE OF CONTENTS
CHAPTER 1: SCAFFOLD VASCULARIZATION: A CHALLENGE FOR THREE-
DIMENSIONAL TISSUE ENGINEERING .....................................................................1
1.1 Abstract .............................................................................................................................................................1
1.2 Introduction ......................................................................................................................................................1
1.3 Scaffold materials .............................................................................................................................................3 1.3.1 Biologically derived polymers .........................................................................................................3 1.3.2 Synthetic biomaterials........................................................................................................................5 1.3.3 Scaffold porosity ..................................................................................................................................8
1.4 Vascularization of biomaterials.......................................................................................................................9 1.4.1 Blood vessels and vascularization in vivo....................................................................................9 1.4.2 Angiogenic Factors...........................................................................................................................12 1.4.3 Tube formation in biological scaffolds .......................................................................................13 1.4.4 Tube formation in synthetic scaffolds..........................................................................................18 1.4.5 Co-culture models .............................................................................................................................20
1.5 Bioreactor culture and mechanical conditioning.........................................................................................23 1.5.1 Flow conditioning .............................................................................................................................24 1.5.2 Cyclic strain........................................................................................................................................25
1.6 Conclusions .....................................................................................................................................................27
1.7 References .......................................................................................................................................................29
CHAPTER 2: CHARACTERISATION, DEGRADATION AND MECHANICAL
STRENGTH OF POLY(D,L-LACTIDE-CO-Ε-CAPROLACTONE)-POLY(ETHYLENE
GLYCOL)-POLY(D,L-LACTIDE-CO-Ε-CAPROLACTONE) ...................................... 45
2.1 Abstract ...........................................................................................................................................................45
2.2 Introduction ....................................................................................................................................................45
2.3 Materials and Methods ..................................................................................................................................47 2.3.1 Materials..............................................................................................................................................47 2.3.2 Polymer synthesis..............................................................................................................................47 2.3.3 Film preparation ...............................................................................................................................48 2.3.4 Characterization................................................................................................................................48 2.3.5 Hydrolytic degradation ...................................................................................................................49 2.3.6 Tensile testing.....................................................................................................................................50
2.4 Results and Discussion ...................................................................................................................................50 2.4.1 Characterization of PEG-containing co-polymers..................................................................50
2.4.2 Hydrolytic degradation ...................................................................................................................53 2.4.3 Tensile strength..................................................................................................................................59
2.5 Conclusions .....................................................................................................................................................60
2.6 Acknowledgments...........................................................................................................................................61
2.7 References .......................................................................................................................................................62
CHATPER 3: BLENDS AS A STRATEGY TOWARDS TAILORED HYDROLYTIC
DEGRADATION OF P(CL-CO-D,L-LA)-PEG-P(CL-CO-D,L-LA) COPOLYMERS .......... 66
3.1 Abstract ...........................................................................................................................................................66
3.2 Introduction ....................................................................................................................................................66
3.3 Materials and Methods ..................................................................................................................................68 3.3.1 Materials..............................................................................................................................................68 3.3.2 Copolymers, blend preparation and sample fabrication .......................................................68 3.3.3 Hydrolytic degradation ...................................................................................................................68 3.34 Statistical Analysis.....................................................................................................................................69
3.4 Results and Discussion ...................................................................................................................................70 3.4.1 Mass loss and buffer absorption ...................................................................................................70 3.4.2 Evolution of pH of degradation media........................................................................................71 3.4.3 Morphology.........................................................................................................................................73
3.5 Conclusions .....................................................................................................................................................74
3.6 Acknowledgments...........................................................................................................................................75
3.7 References .......................................................................................................................................................76
CHAPTER 4: ENHANCED SMOOTH MUSCLE CELL ADHESION AND
PROLIFERATION ON PROTEIN-MODIFIED POLYCAPROLACTONE-BASED
COPOLYMERS ...................................................................................................... 79
4.1 Abstract ...........................................................................................................................................................79
4.2 Introduction ....................................................................................................................................................79
4.3 Materials and Methods ..................................................................................................................................81 4.3.1 Materials..............................................................................................................................................81 4.3.2 Co-polymer synthesis .......................................................................................................................81 4.3.3 Film preparation ...............................................................................................................................82 4.3.4 Surface modification ........................................................................................................................82 4.3.5 Characterization of amine content ...............................................................................................83 4.3.6 Cell culture..........................................................................................................................................83 4.3.7 Cytotoxicity .........................................................................................................................................83
4.3.8 Cell adhesion and proliferation ....................................................................................................84 4.3.9 Statistical analysis ............................................................................................................................85 4.4.1 Cytotoxicity .........................................................................................................................................86 4.4.2 Surface modification ........................................................................................................................87 4.4.3 Smooth muscle cell culture .............................................................................................................89
4.5 Conclusions .....................................................................................................................................................96
4.6 Acknowledgements .........................................................................................................................................97
4.7 References .......................................................................................................................................................98
CHAPTER 5: SMOOTH MUSCLE CELL ADHESION IN SURFACE-MODIFIED
THREE-DIMENSIONAL CO-POLYMER SCAFFOLDS PREPARED FROM CO-
CONTINUOUS BLENDS ........................................................................................ 102
5.1 Abstract .........................................................................................................................................................102
5.2 Introduction ..................................................................................................................................................102
5.3 Materials and Methods ................................................................................................................................105 5.3.1 Materials............................................................................................................................................105 5.3.2 Co-polymers......................................................................................................................................105 5.3.3 Preparation of blends ....................................................................................................................105 5.3.4 Annealing and scaffold fabrication............................................................................................105 5.3.5 SEM imaging ....................................................................................................................................106 5.3.6 Mechanical properties ...................................................................................................................106 5.3.7 Permeability of scaffolds...............................................................................................................107 5.3.8 Protein immobilization and smooth muscle cell seeding.....................................................107
5.4 Results and Discussion .................................................................................................................................110 5.4.1 Effects of annealing on pore size ................................................................................................110 5.4.2 Compressive strength .....................................................................................................................113 5.4.3 Permeability of scaffolds...............................................................................................................114 5.4.4 Smooth muscle cell adhesion on protein-modified scaffolds..............................................116
5.5 Conclusions ...................................................................................................................................................119
5.6 Acknowledgements .......................................................................................................................................119
5.7 References .....................................................................................................................................................120
GENERAL CONCLUSION AND SUGGESTED FUTURE WORK ................................ 123
APPENDIX .......................................................................................................... 129
A1. Protocol for synthesis of P(ε-CL-co-D,L-LA)-PEG-P(ε-CL-co-D,L-LA)................................................129
A2. Protocol for surface modification, cell seeding, cell count, and imaging. ...............................................133
A3. Standard curve established for the ninhydrin assay used in Chapter 4. ................................................136
A4. Control α-actin staining of SMC in Chapter 4..........................................................................................137
LIST OF FIGURES
Chapter 1
Figure 1.1 Polymeric structure of poly(glycolic acid), poly(lactic acid), and poly(ε-caprolactone) 6
Figure 1.2 The structure of a large blood vessel, depicting the intima, the media,
and the adventitia 10
Chapter 2
Figure 2.1 WAXD curves of films of PEG-containing copolymers 52
Figure 2.2 Water absorption and mass loss of P(100/0), P(70/30), and P(30/70)
during degradation in PBS (37°C, pH 7,4) for up to 26 weeks 54
Figure 2.3 Gel permeation chromatograms showing the effect of degradation on the polydispersity and molecular weights 55
Figure 2.4 The effect of degradation on the average molecular weight, Mn, as
determined by GPC and the effect of degradation on monomer composition as calculated from 1H-NMR spectra 57
Chapter 3
Figure 3.1 Mass loss (% of initial weight) and water absorption (% of dry weight) presented as a function of degradation time over 26 weeks 71
Figure 3.2 Effect of sample thickness and blend composition on the pH of the
degradation media. 72
Figure 3.3 Scanning electron micrographs of partially degraded P100/P30 copolymer blends. 73
Figure 3.4 Scanning electron micrographs of cross-sections of partially degraded PCL/P70, P100/P70, and P100/P30 copolymer blends. 74
Chapter 4
Figure 4.1 MTT mitochondrial activity assay of the cytocompatibility of degradation medium. 86
Figure 4.2 Absorbance at 570nm of aminolyzed copolymer films. 88
Figure 4.3 Cell adhesion and proliferation on aminolyzed and protein-immobilized P(100/0) films. 90
Figure 4.4 Cell adhesion and proliferation on aminolyzed and protein-immobilized
P(70/30) films. 92
Figure 4.5 Scanning electron micrographs of SMC on modified P(100/0), 8h and 48h after seeding. 93
Figure 4.6 α-actin staining of SMC on modified P(100/0), 8h and 48h after seeding. 94
Figure 4.7 α-actin staining of SMC on modified P(70/30), 8h and 48h after seeding. 95
Chapter 5
Figure 5.1 Schematic drawing of cell-seeding device. 108 Figure 5.2 SEM images confirming continuous porosity in non-annealed samples of
P(100/0), P(70/30), and polystyrene. 110 Figure 5.3 Porous P(100/0): effect of annealing on pore size. 112 Figure 5.4 Porous P(70/30): effect of annealing on pore size. 112 Figure 5.5 Representative stress-strain curves for porous P(100/0) and P(70/30). 114 Figure 5.6 Compressive Young’s moduli and stress at 10% strain of porous P(100/0)
and P(70/30) for different times of annealing. 114 Figure 5.7 A log-linear plot of the permeability constant κ (m2) vs. annealing time
for porous P(100/0) and P(70/30). 115 Figure 5.8 Aminated copolymer scaffold cross-sections stained blue by ninhydrin
reactant. 117 Figure 5.9 Scanning electron micrographs of SMC adhering in a P(100/0) scaffold,
8h after cell seeding. 118 Figure 5.10 Number of SMC adhering on porous P(100/0) scaffolds 7h after cell seeding. 119 Appendix
Figure A.1 Setup for copolymer synthesis. 134 Figure A.2 Standard curve established for ninhydrin assay in Chapter 4. 136 Figure A.3 Control α-actin staining of SMC in Chapter 4. 139
LIST OF TABLES
Chapter 1
Table 1.1 Some angiogenic factors used to promote vascularization of biomaterials. 14 Chapter 2
Table 2.1 Characterization of the three copolymers synthesized by ring- opening polymerization . 50
Table 2.2 Thermodynamic transitions, fusion energy, and % of crystallinity
of P(100/0), P(70/30), and P(30/70) films. 51 Table 2.3 Changes in mechanical properties with time of degradation. 59 Chapter 5
Table 5.1 Pore size and approximate porosity as a function of annealing time and temperature. 111
LIST OF ABBREVIATIONS bFGF basic fibroblast growth factor PDLLA poly(D,L-lactic acid)
DMA dynamic-mechanical analysis PEG poly(ethylene glycol)
DMP dimethyl pimelimidate PEO poly(ethylene oxide)
DSC differential scanning calorimetry PGA poly(glycolic acid)
EC endothelial cells PHEMA poly(2-hydroxyethyl methacrylate)
ECM extracellular matrix PI polydispersity index
EDTA ethylenediamine tetraacetic acid, PLA poly(lactic acid)
ePTFE expanded polytetrafluoroethylene PLGA poly(lactic-co-glycolic acid)
FAP focal adhesion plaques PLLA poly(L-lactic acid)
Fgn fibrinogen PMA 4β-phorbol 12-myristate 13-acetate
FITC fluorescein isothiocyanate PS polystyrene
Fn fibronectin PTFE polytetrafluoroethylene
GAG glucosaminoglycan PVA poly(vinyl alcohol)
GPC gel permeation chromatography RGD arginine-glycine-aspartic acid peptide
HA hyaluronic acid SEM scanning electron microscopy
HMVEC human microvascular endothelial cells
SMC smooth muscle cells
HUVEC human umbilical vein endothelial cells
SnOct stannous octoate
MMP matrix metalloproteinase Tc, Tg, Tm
temperatures of crystallization, glass transition, and melting
MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide
TCPS tissue culture polystyrene
NMR nuclear magnetic resonance TE tissue engineering
PBS phosphate buffered saline TGF- β Transforming growth factor-β
PCL polycaprolactone VEGF vascular endothelial growth factor
PDGF platelet-derived growth factor WAXD wide-angle x-ray diffraction
1
CHAPTER 1:
SCAFFOLD VASCULARIZATION: A CHALLENGE FOR THREE-DIMENSIONAL TISSUE
ENGINEERING
1.1 Abstract
The prevalent challenge facing tissue engineering today is the lack of adequate vascularization
to support the growth, function, and viability of tissues. Researchers rely on the increasing
knowledge of angiogenic and vasculogenic processes to stimulate vascular network formation
within three-dimensional tissue constructs. These processes are mainly endothelial cell-
regulated, although in the context of tissue engineering, specific interactions with scaffold
materials, growth factors, and other cell types may require in vitro vascularization schemes to
be altered accordingly. In order to better mimic the complete in vivo environment, increasing
attention is given to the integration of co-cultures and mechanical conditioning in bioreactors.
Such approaches show great promise for the enhancement of the functionality and clinical
applicability of tissue engineering constructs.
1.2 Introduction
Microvessel beds constitute the infrastructure that provides nutrient delivery and waste
removal in living tissues. In the past few decades many an effort has been made to engineer
functional vascular networks in vitro. Paralleled by progress in tissue-engineered blood
vessels and in the decoding of the complex mechanisms of blood vessel formation, this work
has also generated insight valuable for the repair and reconstruction of three-dimensional
human tissue. There are two principal reasons for promoting the formation of vascular
networks within three-dimensional cell/scaffold constructs. Firstly, in vitro angiogenesis
assays provide useful tools for detailed study of this process by identifying and evaluating the
role of specific biochemical pathways[1]. Secondly, in the absence of a vascular network that
can ensure efficient exchange of metabolites in a growing cell mass, artificial tissue constructs
will remain severely limited in obtainable size, function, and applicability in vivo[2].
2
Vascularization of newly formed tissue in vivo takes place by angiogenesis; the process
of sprouting of new capillaries from pre-existing vasculature[3]. Angiogenesis is essential for
wound healing, tissue repair, and the growth of newly formed tissue. Tissue engineering (TE)
constructs may also become vascularized via a process known as vasculogenesis; the in situ
reassembly of undifferentiated endothelial cells into capillaries and networks[4]. In vivo, this
process normally takes place in developing organs in the embryo[5]. Specific molecules,
called angiogenic factors, that control blood vessel formation have been identified[3] and
some of these are frequently used to enhance vascularization processes. Following the creation
of vascular networks within a TE construct, whether established in vitro or in vivo, these must
ultimately connect to the host vasculature after implantation to fulfill their intended function.
This event is called inosculation and is the process by which the vasculature is rendered
continuous. Finally, a tissue repaired or regenerated with the help of an implanted material,
like any other tissue, will be subject to vascular remodeling with time.
At present, regeneration/repair of cartilage[6], bone[7], and skin[8] – although still
dealing with issues of predictability and long-term stability – can be achieved with available
techniques. However, larger and/or more metabolically demanding tissues such as the
pancreas, liver, heart, and kidney require intrinsic blood supply for long-term survival and
development[9]. For example, it has been suggested that hepatocytes may not be grown
further away than 200-300µm from nutrient supply in order to survive[10]. Vascular networks
are therefore critical for the efficient distribution of oxygen and nutrients to all tissues, as well
as for the removal of waste products. In fact, for any cell mass large enough that it cannot rely
on nutrient supply solely by diffusion, growth, function, and survival post-implantation are
improved by and dependent on ingrowth of blood vessels from the host or, alternatively, by
inosculation of an already existing network[11-14]. The idea of prevascularized TE scaffolds,
as brought forward by Mikos et al.[15], was shown to enhance the performance of forming
tissues compared to non-vascularized constructs[11,16]. However, most attempts to promote
vascularization of scaffolds have been pursued in vivo, by allowing for ingrowth of
vascularized fibrous tissue, which then comes to occupy an important fraction of the available
void space, thus limiting the space available for engraftment of tissue-specific cells[15]. At
present, the development of functional vascular beds in three-dimensional constructs presents
3
a great challenge to researchers, but is no less central to the successful therapeutic application
of tissue engineering principles.
This review treats the background and recent advancements in the vascularization of
porous scaffolds for three-dimensional tissue engineering, with an emphasis on in vitro tissue
culture. Moreover, the effects of co-culture techniques as well as mechanical conditioning of
tissue constructs within the confines of a bioreactor are discussed.
1.3 Scaffold materials
The requirements for three-dimensional TE scaffolds are generally defined in terms of
biocompatibility, degradability, porosity (void volume), pore size, and mechanical properties.
In addition, scaffolds should promote cell-material interactions suitable for a particular
application, either inherently or as a result of surface modification. The choice of a suitable
biomaterial is the first step in the scaffold design process. Polymeric biomaterials can be of
either biological or synthetic origin, and they offer tremendous variation as pure preparations
or as co-polymeric devices. Control over scaffold properties is crucial in the design of
successful scaffolds. Ideally, a scaffold should be highly porous, permeable to flow, and
possess mechanical strength to match the requirements of the tissue to be replaced. On the
same note, it is important that the rate of degradation match an appropriate rate of tissue
growth as it proceeds within the scaffold. Adequate pore size, pore morphology, and a high
degree of interconnectivity are important to achieve efficient cell seeding and growth.
Furthermore, scaffolds should be favorable to fluid flow, ensuring sufficient delivery of
nutrients and removal of waste and scaffold degradation products[17].
1.3.1 Biologically derived polymers
ECM-DERIVED POLYMERS, such as collagens, glycoseaminoglycans (GAGs), and fibrin
gels, are attractive scaffold materials in that they originate from the very tissues that need to
be regenerated. As components of the extracellular matrix (ECM) they offer natural sites for
cell attachment, and signaling that regulate cell phenotype. Traditionally, this group of ECM-
based biomaterials has been the most popular choice for angiogenesis assays and
vascularization studies. They are also used as coating materials to improve cell-material
interactions on synthetic polymeric devices.
4
Collagen is the most abundant protein in mammalian tissues, and is therefore an
interesting scaffold material. Chemical or physical cross-linking may be employed to improve
the mechanical performance of collagen fibers and scaffolds and to moderate enzymatic
degradation of these scaffolds, although cross-link density and porous microstructure are
known to influence cellular interactions with these templates[18,19]. Immunogenic issues,
especially with animal collagen, have been a concern for the application of collagen-based
biomaterials. This effect is primarily due to telopeptides, which may be efficiently removed
enzymatically[20]. However, the full immunogenic potential may not be completely
removable from non-human proteins, and methods for production of recombinant collagen I
and III have been proposed as an alternative source[21].
In response to injury, a fibrin clot forms spontaneously from fibrinogen and thrombin, and
are further cross-linked by factor XIIIa. The implication of fibrin in blood vessel repair makes
it a natural candidate for angiogenic assays[22,23], and has also been used for controlled
release of growth factors in vascular graft engineering and bone regeneration[24-26].
GAGs occur naturally as linear polysaccharide branches of certain protein structures called
proteoglycans. Hyaluronic acid is the simplest of the GAGs, and is present in nearly all
mammalian tissues. Its high molecular weight (50-500kDa) confers mechanical properties that
make it interesting as a scaffold material, and it has been used mainly in ophthalmology and
experimental treatments of joint dysfunction[27]. Pioneering work using collagen-based
materials was carried out by Yannas et al.[28], who developed templates for regeneration of
skin and nerve from co-polymers of collagen and another type of GAG, namely chondroitin 6-
sulfate. Fibrous materials based on fibronectin have also been investigated for similar
applications[29,30].
POLYSACCHARIDES, such as chitosan and alginate, that have been shown to elicit a
relatively mild immunogenic response, have also been investigated for various TE
applications. Chitosan has a molecular structure similar to GAGs, and is susceptible to
enzymatic degradation by the human enzyme lysozyme. The rate of degradation depends on
the degree of crystallinity, which can be easily manipulated by varying the acetyl
content[19,31]. Chitosan has been investigated for such applications as encapsulation,
membrane barriers, and contact lens material[31]. Moreover, DNA complex formation is
favored by its cationic nature, making it an ideal candidate for gene delivery[32]. More
5
recently, hybrid scaffolds containing a synthetic component have also been proposed for drug
release and engineering of cartilage and bone[33,34]. Alginate gels are fabricated under
relatively mild conditions, making it suitable for application in drug delivery and cell
transplantation (e.g. chondrocytes and insulin-producing cells)[19,35]. In particular, growth
factors incorporated into alginate beads have been shown to retain >90% of their
bioactivity[36]. Disadvantages include issues of poorly regulated degradation and the lack of
natural cell adhesion motifs[37].
1.3.2 Synthetic biomaterials
Despite the obvious advantages of using biologically derived materials, important
drawbacks limit their use as scaffolds. These include poor mechanical properties, and inherent
batch-to-batch variability. Synthetic polymers therefore represent an attractive alternative,
especially for the fabrication of large and/or load-bearing scaffolds. These materials offer a
high level of control, as well as seemingly endless variations, through the manipulation of
monomer composition, molecular weight, degree of cross-linking and/or branching, co-
polymerization, blending, etc. These and other polymer processing techniques may be used as
tools to manipulate such material properties as degradability, mechanical strength, and surface
chemistry.
Many classes of resorbable synthetic polymers are being studied for use in tissue
engineering[38]. Degradable polyesters, such as poly(lactic acid) (PLA), poly(glycolic acid)
(PGA), poly(ε-caprolactone) (PCL), and their co-polymers have been approved by the FDA
for a number of medical and drug delivery devices, and are thus very attractive candidates for
scaffold fabrication. These biocompatible polymers exhibit interesting and tunable mechanical
properties and predictable degradation kinetics. PLA, PGA and their copolymers (PLGA)
were utilized in early applications for sutures and orthopedic fixations[39]. PGA is an inelastic
and highly crystalline material. Poly(lactic acid) exists in two practically employed isoforms,
namely the highly crystalline poly(L-lactic acid) (PLLA) and the amorphous poly(D,L-lactic
acid) (PDLLA). Although less crystalline than PGA, PLA is more hydrophobic, and less
susceptible to hydrolysis[2]. PCL possesses properties that set it apart somewhat from the
other polyesters mentioned. It has a very low glass transition temperature (Tg = -60ºC), low
melting temperature (Tm = 55-60ºC), and shows high thermal stability. Its decomposition
temperature is approximately 100ºC higher (Td = 350ºC) than PLA and PGA, which makes it
6
better suited for some processing techniques. Due to its low Tg, PCL is rubbery at room and
physiological temperatures, a temperature that contributes to its high permeability for many
drugs[31]. PCL is furthermore a flexible plastic that degrades at a much slower rate than either
PLA or PGA.
H [ O (CH2) C ] OH H [ O CH2 C ] OH
O
n
H [ O CH C ] OH
O
n
CH3
O
n5
a) b) c)
H [ O (CH2) C ] OH H [ O CH2 C ] OH
OO
n
H [ O CH C ] OH
OO
n
CH3
O
n5
a) b) c)
Figure 1.1: Polymeric structure of a) poly(glycolic acid), b) poly(lactic acid), and c) poly(ε-
caprolactone).
Polyesters can be easily co-polymerized, to achieve materials with intermediate
properties. For example, block co-polymers of PCL and PLA combines the rapid degradation
of PLA with the flexibility of PCL[40]. Incorporation of poly(ethylene glycol) (PEG) has also
been shown to improve water uptake, and speed up degradation of PCL[41]. Vert and
coworkers synthesized a series of block co-polymers containing different combinations of
PLA/PCL/PEG to achieve materials exhibiting a wide range of mechanical properties and
degradation rates[42-45]. This work clearly shows that such properties may be efficiently
tailored to meet the demands of particular applications. In addition, PLA-PGA co-polymers
(PLGA) are commonly occurring biomaterials, currently in clinical use for surgical sutures,
orthopedic implants, and as drug delivery systems[46].
The main concern of these materials is related to the hydrolytic degradation mechanism,
which propagates by ester bond cleavage and generates acidic by-products. These lower the
local pH, increasing the rate of bulk degradation via autocatalysis, limiting cell viability in the
proximity of the implant, and in some cases eliciting inflammatory reactions[47].
Furthermore, PGA and PLA have a tendency to crumble during degradation, creating large
particles that may remain in the site for years. Such uncontrolled breaking of implants may
also cause damage to nascent tissue[48].
Beside the classic group of aliphatic polyesters, a large quantity of biodegradable
polymer and co-polymer systems have emerged for biomaterial systems. These include
7
polyanhydrides, polyorthoesters, polyphosphazenes, and polyurethanes. Polyanhydrides have
been extensively investigated for use in drug delivery for over 25 years[49-51]. Initial
limitation due to extensive degradation was addressed by researchers and resulted in materials
with degradation rates ranging from weeks to years[7]. Vascularization of polyanhydride
implants was observed after 4 weeks in vivo[52]. Polyorthoesters are hydrophobic polymers
that are stable at physiological pH, but degrade rapidly at pH 5.5 via hydrolysis[38,53]. These
two families of biodegradable polymers have found principal use in drug-delivery systems, as
they are mechanically unsuited for scaffolding purposes[48,54]. Polyphosphazene-based
biomaterials range in stiffness from soft to hard, show great backbone flexibility, and are
being investigated for their applicability in engineering tissue[38]. Laurencin and coworkers
studied the applicability of these materials for bone regeneration, and reported
osteocompatibility as well as non-toxicity of the degradation products[55]. Composite
materials containing hydroxyapatite – a natural component of bone – were investigated for the
same application[56]. Furthermore, blending polyphosphazenes with PLGA was shown to
moderate PLGA degradation, even as the polyphophazene degradation products had a
buffering effect on the degradation milieu[57].
Polyurethanes constitute yet another class of polymers that have long been in use as
implant materials[58]. They consist of alternating “soft” segments (polyether or polyesters)
and “hard” segments (diisocyanate), and chain extenders that may be either diols or diamines.
Great flexibility in chemical composition is possible through the choice of the soft and hard
segments, and by varying the ratios of these components a wide range of mechanical
properties can be obtained. Earlier concerns regarding the putative carcinogenity of the
degradation products have been attended to mainly by replacing diisocyanate by e.g. 1,4-
diisocyanatobutane[59] and lysine diisocyanate[60], which degrade into biocompatible
components. The latter was polymerized with glucose, which produced glucose and lysine as
the principal degradation products[60].
Synthetic polymers may also be processed into hydrogels, which by definition have high
water content, ranging from 10-20 wt% and up. Synthetic hydrogels structurally resemble the
ECM of many tissues[19] and present inherently good mass transport properties[61]. The
structural stability of synthetic hydrogels can be controlled by the degree of cross-linking
between polymer chains. Typically, hydrogels used for scaffold applications can be processed
8
under relatively mild conditions, allowing for the incorporation of bioactive molecules or even
cells. The disadvantages include issues of sterilization and low mechanical strength[61,62].
Synthetic polymers used to make hydrogels include poly(ethylene oxide) (PEO) and the
chemically similar poly(ethylene glycol) (PEG), poly(vinyl alcohol) (PVA), poly(acrylic
acid), and polypeptides[19]. Photo-polymerisable PVA hydrogels are injectible and found to
be elastic and strong. Methods to improve the inherently poor cell adherence in PVA and PEG
gels by modification with an RGD cell adhesion peptide have been developed[63,64].
Hydrogels have also been fabricated from PEO-containing block co-polymers with PLLA[65]
and poly(propylene fumarate)[66], although the latter exhibited certain cytotoxicity in vitro .
Finally, newly synthesized materials have also been proposed for use in TE applications.
Matrices based on PCL copolymerized with hydrophilic poly(ether ester amide)s have been
investigated for drug delivery[67], and polypyrroles, possessing modifiable electrical
properties have been suggested for nerve regeneration[68].
1.3.3 Scaffold porosity
The optimal pore size of a scaffold varies with application[46], but needs to be large
enough that cells can easily penetrate into the scaffold bulk, while at the same time optimizing
surface area, which usually increases with decreasing pore size. Highly porous scaffolds are
usually preferred, as the surface area and the capacity for large cell mass increases with void
volume. On the other hand, excessively porous structures become fragile and may not fulfill
the mechanical requirements. For the regeneration of bone, pore sizes of 100-400µm are
usually recommended[69,70], whereas they should be 20-150µm for skin regeneration[71],
and approximately 20µm for ingrowth of hepatocytes and fibroblasts[46]. In the case of
neovascularization, data must be interpreted in the context they are generated. For example,
independent microvessel invasion of membranes or thin porous scaffolds have been reported
for pore sizes of 5-15µm[72,73]. However, neovascularization of a TE scaffold (especially for
larger structures) may also take place in conjunction with tissue ingrowth, and would in such a
case be more dependent on a pore size that promotes successful invasion of the tissue in
question. For example, van Tienen and coworkers[74] observed ingrowth of vascularized
fibrous tissue in pores that measured at least 30µm. It should be noted that reported results
often depend on scaffold design and on the material chemistry that often differs from one
study to another. Induction of bone regeneration has been reported in biodegradable scaffolds
9
with pore sizes as small as 16-32µm[73], while in a general manner, much larger pore sizes
are found in the literature, as mentioned.
Pore morphology and pore interconnectivity also play a significant role in regulating cell
attachment, proliferation, and matrix deposition – e.g. for bone ingrowth[70] and in the culture
of smooth muscle cells[75,76]. Subcutaneous PVA sponges, with well defined pores were
found to induce better vascularized tissue ingrowth than did PTFE implants of similar pore
size[77]. Furthermore, pore microstructure and interconnectivity strongly influence mass
transport (permeability and diffusion) and mechanical properties, as well as degradation
profiles[46,78-80]. Numerous techniques exist for the fabrication of porous scaffolds, and
these have been reviewed elsewhere[81].
1.4 Vascularization of biomaterials
A variety of in vitro assays for angiogenesis exists. Classical testing schemes such as the
chicken chorioallantoic membrane (CAM) assay, pouch assays or the study of
neovascularization of the avascular cornea are still in use, but have to some extent been
replaced by reconstituted biological materials (for reviews, see e.g. Cockerill et al.[82,83]).
These assays provide a simplified but controlled environment in which the processes and
mechanisms of tubular network formation can be studied in great detail. Given the proper 3D
environment, endothelial cells will migrate into the matrix bulk, where they will rearrange and
form a network of capillaries, which may or may not contain lumina[84]. Developments
within the field tend to consist of substrates, growth factor delivery, and culture techniques
that increasingly mimic the complex physiological environment of the cells. However,
obtaining patent networks (i.e. that are capable of withstanding the strain of a circulating
medium) still presents a challenge for researchers. Encouragingly enough, it was recently
shown that engineered endothelial networks of tube-like structures indeed have the potential
to evolve to functional blood vessels in vivo[85]. The presence of such patent networks within
TE constructs would greatly facilitate the invasion of tissue-specific cells and enhance the
viability and function of the forming tissue.
1.4.1 Blood vessels and vascularization in vivo
The structure of all veins and arteries in the body consist of three distinct layers: the
intima, the media, and the adventitia[86] (see Figure 1.2). The adventitia is the outermost
10
layer of the blood vessels, and consists primarily of connective tissue and collagen fibers,
which allows the vessel wall to return to its initial shape following dilatation or contraction.
The media is composed mainly of elastic fibers and mural cells. Mural cells is the common
name referring to either smooth muscle cells, as in the case of larger vessels, or pericytes,
which are rather undifferentiated and present in newly formed and small blood vessels. The
role of these cells is to stabilize the vessel structure. Smooth muscle cells also provide the
mechanically dynamic properties of larger blood vessels. Finally, the intima makes up the part
of the blood vessel which is in actual contact with the blood stream. This layer is composed of
a monolayer of endothelial cells, supported by the basal lamina2, and thin layer of connective
tissue. Capillaries are much smaller, and consist only of a layer of endothelium, lacking both
the media and adventitia. The cells on which the focus will be placed in this text are
endothelial cells (EC) and smooth muscle cells (SMC), due to their implication in mature
vessel structures in vivo.
Figure 1.2: The structure of a large blood vessel, depicting the intima, the media, and the adventitia.
(From original image http://en.wikipedia.org/wiki/Image:Anatomy_artery.png#file)
2 A layer rich in collagens and laminin, deposited by cells.
11
New blood vessels may form by one of the two following processes: 1) by angiogenesis,
which is the sprouting of new blood vessels from preexisting ones, or 2) by vasculogenesis,
which is the assembly of undifferentiated EC in situ.
Vasculogenesis proceeds through five basic steps[4]: 1) generation of differentiated EC
from bone marrow–derived precursor cells; 2) EC form aggregates and establish cell-cell
contacts; 3) EC polarize and rearrange into nascent lumen-containing tubes; 4) an immature
vessel network is formed; and 5) pericytes and SMC are recruited, which stabilizes the newly
formed vessels.
Angiogenesis involves a growth phase and a stabilization phase[87]. Four steps define
the growth phase, which is initiated by 1) cell-mediated, proteolytic degradation of the
basement membrane, allowing for endothelial cells to invade the surrounding ECM; 2)
Activated EC then proliferate and migrate into the created space, where they 3) organize into
aligned, lumen-containing cords. 4) Lastly, newly formed sprouts connect to form loops of
immature vasculature. The immature vasculature is subsequently stabilized through the
recruitment of smooth muscle cells or pericytes, accompanied by a halt in EC proliferation,
and the deposition of a basement membrane (reinforced basal lamina) around the cords.
The first reports of in vitro tube formation by capillary endothelial cells was made in
1980 by Folkman and Haudenschild[84], who observed capillary networks developing on a
gelatine substrate after several weeks of culture. Since then, methods have been refined, as the
understanding of angiogenic processes has evolved. Today it is largely recognized that the
interplay between EC and insoluble molecules in the ECM is critical in many events during
angiogenesis and vasculogenesis, including the regulation of cell growth, migration,
differentiation and tube formation[88]. Vessel growth is also influenced by cell-cell
interactions and soluble species, such as growth factors. In addition to the choice of substrate,
the source of cells will influence the outcome of a study. To some extent, endothelial cells
from rat or bovine aorta behave differently than ECs from human tissue. The latter have
higher requirements of growth factors, and may degenerate in some basic culture conditions
where animal cells spontaneously differentiate into capillary structures[89]. Further
distinctions can be made between human ECs of different origin. The commonly used human
umbilical vein ECs (HUVEC) are of macrovascular origin, whereas the cells most involved in
inflammation, wound healing and vascularization in vivo are microvascular ECs. However,
12
any difference in the capacities of these two cell types to form tube-like structures is not
readily apparent[90]. Some differences may also be found in the ability to exert contractive
forces on malleable substrates, where e.g. human blood outgrowth ECs are stronger compared
to HUVEC[91]. In addition, recent publications have demonstrated capillary formation using
endothelial progenitor cells, who are implicated in post-natal neovascularization[92,93].
1.4.2 Angiogenic Factors
A number of growth factors and other angiogenic promoters are used to induce and
enhance EC sprouting and capillary network formation in cell/polymer constructs. These
factors may act directly on EC or indirectly, inducing an angiogenic response in vivo, without
eliciting a direct response in cultured EC, such as an altered migration or proliferation[4].
Vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF) are
examples of highly potent EC mitogens, stimulating angiogenesis via direct interactions with
EC. Both growth factors are widely distributed in various tissues in the body, but while bFGF
is known to act on various cell-types, VEGF is EC-specific[4]. Other angiogenic factors, such
as platelet-derived growth factor (PDGF) and transforming growth factor-beta (TGF-beta), act
indirectly to induce blood vessel formation and/or stabilization[4,94]. Table 1.1 lists these
molecules and others, their roles in angiogenesis in vivo, and examples of their use to induce
angiogenesis in biomaterials.
Many groups have explored the controlled delivery of angiogenic growth factors as a
means to improve the local vessel recruitment in implants and cell transplants[11,16,103]. The
modulation of growth factor delivery is highly important since growth factors such as VEGF
and bFGF are rapidly cleared and inactivated following administration in an unprotected
formulation[16]. Another aspect is that bolus delivery may cause a toxic or otherwise
unwanted response in the studied (or surrounding) cell population, as such administration does
not correspond to any normal situation in vivo[101,111]. Moreover, if too high concentrations
of e.g. VEGF are sustained within a tissue, the developing vasculature may become deformed
and nonfunctional[112]. Control of the release rate is therefore of utmost importance. To deal
with these concerns, the encapsulation of drugs within degradable polymers from which they
can be liberated in a controlled manner has come forth as a successful method. Biologically
derived hydrogels, including fibrin, alginate, hyaluronic acid, chitosan, and gelatine have
proved very useful as encapsulation vehicles due to their biocompatibility and inertness
13
towards bioactive molecules[16,19,25]. Moreover, encouraging results have been reported
from trials with synthetic materials such as PLGA[11,36,102,103] and PEG hydrogels[98].
Growth factors may be released from encapsulating matrices via diffusion, mechanical
stimulation, or matrix degradation[19]. Control of the release rate may be obtained by varying
the rate of degradation of the capsule material, and this may in turn be achieved by altering the
molecular weights or degree of polymer cross-linking[102]. Another very interesting approach
consists of pacing the release of growth factor from the matrix to the demand of the growing
cell popluation[98,101,113]. Such designs aim at more closely mimicking the way in which
naturally sequestered growth factor is released from the ECM by proteolytic activity, and have
been shown to induce a more controlled growth of capillary networks in fibrin[101].
1.4.3 Tube formation in biological scaffolds
A few methods dominate the three-dimensional in vitro angiogenesis assays in current
use. In so-called in-gel seeding cells are mixed with the matrix solution prior to gelation. Cell-
seeded collagen type I gels prepared in this manner were not able to induce EC tube formation
in the absence of endothelial mitogen[114]. Alternatively, ECs can be cultured to confluence
or near-confluence on top of a gel (e.g. fibrin, collagen, Matrigel3), at which time a second
layer is placed on top of the cells, creating a “sandwich” consisting of two layers of gel
matrix. Trapped within the sandwich, the cells may subsequently differentiate, migrate and
form tubes, an event which is largely dependent on the interaction between the overlaying
matrix layer and the apical cell surface[116, 117]. A third approach involves the use of intact
ring sections of aortas that are embedded in collagen, fibrin, or Matrigel gels, and the
angiogenic activity is observed as radial sprouting of endothelial cells[118]. This so-called
aortic explant model has traditionally employed segments of rat aorta[119], but can be
extended to include tissues from other animals[120]. The advantage of this method compared
to cell-based assays is the realistic representation of in vivo structural cell arrangements[120].
In addition to the models mentioned, in vivo experiments are conducted in which seeded
or acellular matrices are implanted subcutaneously in mice, where a capillary network starts to
form. Subcutaneous implantation of fibrin and Matrigel lead to blood vessel ingrowth[22,
121], while collagen type I did not[22,121]. Vascularization of the implant may be evaluated
after surgical recovery[122] or, alternatively, in vivo through transparent “windows” in the 3 Matrigel: a commercial, multi-molecular matrix material extracted from Englebreth-Holm-Swarm tumors[115]
14
skin of the animal[85]. (Note: Several approaches are currently being used to quantify
growing and established 3D tubular networks within a biomaterial. The lack of a standardized
method for quantification makes the direct comparison between different assays and results
complicated.)
Table 1.1: Some angiogenic factors used to promote vascularization of biomaterials.
*Coll I: collagen type I; CAM: chorioallantoic membrane; PLGA: poly(lactic-co-glycolic
acid).
Angiogenic agent Roles in vivo Examples of use in
biomaterials in vitro*
Vascular endothelial growth factor (VEGF)
Potent angiogen, required for the formation of
immature vasculature[95], promotes ECM
degradation[94]. Exists in five isoforms of different
potency. Must be well regulated. Reinforces the
angiogenic potential of bFGF[96].
Coll I[1,97], fibrin[98-
100] ,
fibrin and CAM[101],
PLGA[102,103],
PEG[98]
Basic fibroblast growth factor (bFGF)
Stimulates expression of VEGF and its receptors in
EC. Promotes ECM degradation[94], EC migration
and pericyte attraction[104]. Reinforces the
angiogenic potential of VEGF[96].
Coll I[1,97,105]
Fibrin[99,100]
PLGA in alginate[11]
Tissue growth factor-beta (TGF- beta)
Induces mural cell differentiation[3], inhibits ECM
degradation[94]. Dose-dependent stimulation of
tube formation in vitro [106]. EC proliferation and
migration are inhibited at higher
concentrations[107].
Matrigel[106]
Platelet-derived growth factor (PDGF)
Recruitment of pericytes/SMC[83]. PLGA[102]
Angiopoietins (Ang-1, Ang-2)
Ang-1 stabilizes vessels, while its antagonist Ang-2
acts as a destabilizer, promoting EC migration and
angiogenesis in presence of other factors, such as
VEGF[108,109].
Co-culture in Coll
I[109]
4 beta-phorbol 12-myristate 13-acetate
(PMA)
Known tumor promoter. Coll I[1,97,110]
15
The various matrices used for endothelial cell culture each represent different
characteristics of the basement membrane. Therefore it is not surprising that the mechanism of
tubulogenesis in vitro in some aspects is matrix-specific. It has even been suggested that the
nature of the supporting material influences the angiogenic potential of growth factors[123].
The specific interactions between EC and the matrix are mediated by integrin receptors on the
cell surface[124]. Dallabrida et al.[125] observed matrix-specific tubulogenesis in a
publication where down-regulation of the expression of the
alpha v beta 3 integrin (a receptor expressed by angiogenic EC, binding both VEGF[126] and
PDGF-B[127]) in ECs reduced tubulogenesis in fibrin gels, whereas no such effect was seen
on Matrigel. Moreover, Hall et al. were able to induce angiogenesis on fibrin substrates,
covalently modified with synthetic binding sites for the alpha v beta 3 integrin[124].
Chalupowicz et al.[117] identified the interaction between HUVEC and the beta15-41
sequence of fibrinopeptide B as necessary for tube formation in fibrin sandwich.
Soluble species and matrix constituents work in concert with many cell membrane-
associated adhesion molecules during the angiogenesis process. Among these are the cell-cell
junction proteins VE-cadherin (CD144) and PECAM-1 (CD31), which have been
demonstrated to regulate lumen formation and intercellular association during EC network
formation, respectively[128]. In addition, there is evidence to suggest that VE-cadherin is
necessary for the maintenance of tubular networks in fibrin and collagen type I gels[1, 129].
Antibodies against VE-cadherin not only inhibited tubulogenesis but also disrupted already
developed networks[129].
The controlled and localized release of the growth factor bFGF has been shown to
significantly enhance vascularization of injured sites[16], and is furthermore suggested to
increase the size of growing vessels[11]. Sakakibara and coworkers[16] reported that slow
release of bFGF from gelatin microspheres in infarction areas improved cardiac function and
survival of transplanted myocytes in rats. Endothelial cells grown on gelatin-coated micro-
carriers dispersed in a fibrin gel exhibited important increases in their capacity to sprout when
bFGF or VEGF was added to the gel[100]. Ehrbar and coworkers[101] used an engineered
version of VEGF121 that was covalently incorporated into fibrin gels and subsequently
evaluated in the CAM assay. Grafted VEGF was gradually released as the matrix was
enzymatically degraded by EC, and enhanced arterial and venous network development was
16
observed. In a control experiment, non-grafted soluble VEGF121 was released rapidly and
triggering chaotic vessel growth. Other examples of VEGF encapsulation include the use of
alginate[19,36] or combined alginate/PLGA matrices[102,103]. Furthermore, attempts have
been made to coordinate the controlled delivery of multiple growth factors[102]. In an
approach combining bFGF and VEGF delivery, Nillesen et al.[130] used collagen-heparin
sponges with 100µm pores to stimulate vascularization subcutaneously in rats. (Heparin
increases the amount of bFGF that may be incorporated into cross-linked collagen[105] and
alginate[131] gels.) They found that while the two growth factors alone supported vessel
formation, the combined delivery of bFGF and VEGF was superior in promoting the
formation of mature, pericyte-lined, vasculature within the scaffolds. In addition, there was a
dramatic decrease in the number of hypoxic cells in the scaffold centre when both growth
factors were present[130].
INFLUENCE OF SUBSTRATE STIFFNESS
It has long been claimed that endothelial migration and elongation take place along lines
of reorganized matrix – a result of cell forces acting on the matrix components[132]. These
forces are communicated via cell adhesion sites called focal adhesion plaques (FAP) that
consist of assemblies of structural and signaling proteins, such as integrins[133]. These sites
enable mechanical tension from the ECM to be transferred into the cell, and conversely from
the cytoskeleton to the ECM[134]. Relatively rigid substrates have been found to better
support the formation of stable focal contacts than more compliant substrates, by promoting
stronger integrin-cytoskeletal linkages[135]. However, they also bring on a more spread cell
morphology, which in turn leads to decreased cell elongation and cells organized in isolated
clusters or monolayers rather than network formations[91,132,136,137]. More malleable
substrates promote a rounded morphology that is associated with a more differentiated
phenotype in endothelial cells, although fully rounded cells will not form capillary tubes[136].
Appropriate matrix rigidity is not associated with any absolute magnitude of “stiffness”,
but needs to be put into a more elaborate picture, incorporating matrix type, matrix geometry,
conditions of constraint, and the cell type involved. The force generated by cells cultured in or
on a gel construct depends both on the cell type and on the flexibility of the matrix[91].
However, it has been clearly established that matrix contraction and tube formation within
17
reconstituted gels require an active cytoskeleton[137-139]. During gel contraction the number
of stress fibers (actin filaments) present in cells can be related to the stress being imposed on
the cytoskeleton, which in turn is a reflection of the matrix resistance to deformation[91,137].
Moreover, stress fibers become oriented along the tubule axis following EC rearrangement
into capillaries[137]. Analysis of the concentration and organization of actin filaments and
proteins associated in focal adhesion plaques reveals EC-specific morphology changes in
response to changes in substrate flexibility[140].
Substrate stiffness can be altered by changing the polymer concentration, degree of
cross-linking or polymerization conditions. Ingber and Folkman[136] reported on the role of
ECM concentration on rigid surfaces, and found that intermediate coating densities supported
EC rearrangement into tubular structures. Several groups have later repeated these results in
thick substrates concluding that the onset of tubulogenesis can be controlled by altering the
stiffness of the cell support[132,137]. For example, on Matrigel or polyacrylamide gels that
were prepared with varying degree of cross-linking, HUVEC reorganized more quickly into
tube-like structures on more compliant substrates[137]. Similar results were obtained by
decreasing the density of collagen gels[141]. Furthermore, in a recent publication Sieminski
and coworkers[91] demonstrated the possibility of controlling lumen size by varying matrix
concentration. By doubling the concentration of type I collagen from 1.5mg/ml to 3mg/ml
lumen size was increased by 57% and 40% in floating and constrained gels, respectively[91].
This increase was attributed to a resistance to contraction (increasing with the density of the
substrate), bringing on a more spread cell morphology.
In order to study the role of cellular adhesion in tube formation, the inherent ability of
cells to adhere to the support material can be manipulated. Many cell types, including
endothelial cells, are dependent on the activity of the beta 1 integrin unit for adhesion on a
variety of ECM-derived substrates. Tubule formation in Matrigel is, for instance, inhibited
when the beta 1 unit is blocked by antibodies[125]. Indeed, the alpha 2 beta 1-integrin is the
main laminin and collagen receptor in EC[137], and varying the adhesion strength using anti
alpha 2 beta 1-antibodies, influences both the number and dimension of forming capillary
tubes[110]. Deroanne et al.[137] found that as the rigidity of their collagen type I substrate
decreased, the expression of the alpha 2 subunit was reduced, and EC more easily switched to
a tube-like pattern. Furthermore, they established that parallel to these effects, increased
18
substrate malleability also resulted in reduced expression of actin and focal adhesion plaques
proteins, as determined by Western blot[137]. Thus, there seems to be a “window” of
favorable adhesion strength for tube formation, where cells are allowed to adhere to the
substrate without being hindered to differentiate, migrate and rearrange into network patterns.
1.4.4 Tube formation in synthetic scaffolds
Cell scaffolds fabricated from synthetic polymers present a different set of challenges for
vascularization compared to reconstituted gels. The identification of parameters crucial to
capillary formation within synthetic materials is still a work in progress, and these may prove
difficult to ascertain in general terms, given the variation in material properties. For “dry”
synthetic polymer scaffolds, the concensus is that sufficient porosity is a prerequisite for
uniform cellular invasion and long-term survival. Pore morphology may also be crucial to the
ability of synthetic scaffolds to induce cellular invasion[142]. Related to more efficient cell
invasion is the observation that porous implanted matrices become surrounded by loose and
highly vascularized connective tissue, whereas non-porous scaffolds become enveloped by a
dense fibrous capsule[143]. Vascularization around and into the implant has been observed to
be increasingly facilitated with increasing pore size, reaching a maximum at pore sizes on the
order of cellular dimensions[143,144]. Moreover, it is believed that in the absence of natural
ECM components, EC must first secrete and organize their own basement membrane in order
to trigger tube formation[84]. Alternatively, surface modification of synthetic polymers may
improve cell-material interactions and facilitate the process. However, as in the case of
biological ECM-based scaffolds, it is believed that intermediate cell adhesion is necessary for
implanted EC to be able to reorganize and actively participate in the vascularization
process[145].
Expanded polytetrafluoroethylene (ePTFE), although not a degradable biomaterial, has
been extensively studied for use in vascular grafts. Investigations of implanted ePTFE arterial
grafts, carried out by Clowes et al.[146], showed that internodal distances of 60µm provoked
the invasion of microvessels through the graft, ultimately leading to complete endothelial cell
coverage of the luminal side of the implants. Internodal distances of 30µm did not allow for
vessel invasion to the same extent, and endothelial coverage was significantly lower.
Supporting results were reported by Saltzmann et al.[147], again indicating an optimum pore
size of 60µm. Interestingly, by modifying the surface with ECM components prior to
19
implantation in adipose and subcutaneous tissue, extensive vascularization could be stimulated
in similar ePTFE grafts with 30µm internodal distance[72,148]. The implants were modified
either by covalent modification of laminin type I[148] or by exposing the implants to a
preconditioning culture with a cancer cell line, which produced a coating of complex and
hypothetically angiogenic ECM[72]. No vascular infiltration was observed in unmodified
controls.
Degradable polymers are preferred for applications in tissue regeneration. Degradation
mechanisms are sometimes enzymatic[149], but in the case of the commonly used polyesters,
degradation takes place mainly through hydrolytic scission of the ester bonds[150]. This
hydrolytic action results in the production of acidic by-products which, if clearance is not
swift enough, will accumulate and lower the local pH. A degradation-related pH drop has
significant negative influence on cell survival and invasion into the matrix, as was recently
confirmed by Sung and coworkers[47] in a study comparing PLGA and PCL porous scaffolds.
PLGA degrades at a higher rate than PCL, and both polymers degrade faster in vivo than in
vitro. In vivo cell survival and migration into the scaffolds, and thus the extent of
vascularization, was directly affected by increasing acidity of the implants with degradation.
Consequently, PCL implants supported a significantly higher number of newly formed
microvessels[47]. In contrast, upon implantation of porous PLLA and PLGA, Mikos et al.[15]
found that the faster-degrading PLGA was better suited for prevascularization schemes. In this
case, PLGA void volume was filled more rapidly with vascularized fibrous tissue, and not to
same extent as in the PLLA scaffold, and was for both those reasons judged as more suitable
for subsequent cell grafting.
Material design and the method used for scaffold fabrication influence pore size,
morphology and interconnectivity. These characteristics strongly affect cell invasion and must
be considered for the engineering of hard[70]and soft tissue[75,76]as well as for scaffold
vascularization[142]. For example, in PEG-grafted PHEMA (poly(2-hydroxyethyl
methacrylate)) hydrogel sponges seeded with human micro-vascular endothelial cells
(HMVEC), Dziubla et al. observed the formation of tube-like structures with highly variable
lengths and diameters[142]. These highly porous scaffolds were extremely fragile, but
nevertheless allowed for the formation of tubules that were 102 µm long on average, with a
maximum length of 680µm and diameters of 5-10µm, although there was little evidence of
20
HMVEC beyond the moderate distance of 180µm from the surface[142]. Matrices containing
larger pores (up to 15µm) and higher porosity (up to 90%) supported increasing cell invasion
and tubule formation. However, when the same experiment was conducted with ungrafted
PHEMA sponges, it did not to the same extent result in tube formation. This was proposed to
be the effect of unsuitable pore shapes and lower pore interconnectivity, as well as increased
cell adhesion in comparison with the PEG-modified sponges[142].
Two-dimensional models of micro-patterned polymers have been developed in attempts
to quantify the extent of EC spreading necessary for tube formation. EC seeded on
micropatterned, fibronectin-coated strips, 10-30µm wide, formed cell-cell contacts, spread to
an intermediate degree, and became elongated and aligned in the direction of the strips.
However, cells grown in the 10µm grooves formed lumen-containing tubes, while cells
cultured on strips that were 30µm wide continued to proliferate without any tube formation
taking place[151]. In a related PLGA model, Borenstein et al.[9] used a combination of micro-
machining and replica molding technology to fabricate surfaces containing intricate channel
networks. The resulting materials provide seeded cell populations with an elaborate “map” for
migration and capillary network formation throughout the tissue construct. Specifically, with
the aim of eventually creating patterns that mimic biological vasculature, the same group
reported that seeded EC were successfully grown within these channels for up to four weeks.
However, this technique was limited to the fabrication of channels no smaller than 30µm
wide. The optimization of cell spreading on synthetic substrates may be associated with the
importance of malleability reported for ECM gels. In support of this notion, Ingber and
Folkman found that FGF-stimulated tube formation occurred when coating concentrations of
fibronectin, gelatin, or collagen type IV were adapted to stimulate intermediate EC
spreading[136].
1.4.5 Co-culture models
Endothelial cells in blood vessels are generally surrounded and supported by either
pericytes (small vessels) or SMC (in larger vessels). Paracrine interactions in vivo between
these cell types are highly regulatory and affect the differentiation of ECs and mural
cells[109] as well as the stability and functionality of blood vessels[85]. Engineered blood
vessels often are immature and unstable[152] due to delays in the recruitment of mural cells
21
from surrounding host tissue. Therefore, researchers have tried to recreate the stabilizing
effects of mural cells in vitro, resulting in the development of several 3D co-culture assays.
These assays include aortic explant models[153], and EC/mural cell co-culture models in
collagen type I[85,114], Matrigel[106], and agarose[109,122]. Trans-membrane culture has
also been used to study the influence of co-culture on cell migration, proliferation, and protein
expression[154-156].
It has been clearly established that co-culturing SMC with EC stimulates SMC to adopt a
more spindle-shaped morphology, typical of a contractile phenotype[154]. Conversely, mural
cells influence EC phenotype, but the effects of are dependent on the matrix. Co-cultures of
EC and SMC rapidly reorganized into tubular networks on Matrigel-coated dishes, whereas
the same experiment conducted on fibrin- or collagen-coated surfaces merely produced
monolayers[119]. Controversial results have been reported for ECs co-seeded with SMC in
3D collagen gels. Asakawa and Kobayashio observed co-culture-dependent EC tube formation
when cultured without external mitogens for EC in the medium[114]. Korff et al. saw that EC
take on a mature, quiescent phenotype in SMC co-culture, as verified by a down-regulation of
PDGF-B expression, a diminished responsiveness to VEGF stimulation, and reduced
apoptosis within the EC population[109]. Simultaneous stimulation with both VEGF and Ang-
2 was nevertheless able to induce endothelial sprouting. However, this model was based on
co-seeded cellulose microspheres embedded in collagen, which involved different cellular
distribution and direct cell-cell contacts[109]. In a study clearly demonstrating the potential of
co-cultures, Koike and coworkers[85] implanted fibronectin-modified collagen gels, co-
seeded with HUVEC and 10T1/2 precursor cells, and observed formation of functional,
stabilized vasculature. These successfully engineered blood vessels were able to connect to
host circulation and to withstand blood flow[85]. Implantation of gels seeded with HUVEC
alone, however, resulted in formation of noticeably fewer and regressing vessels[85].
Co-cultures have helped produce evidence that mural cell precursor recruitment is
stimulated by EC-derived PDGF-BB and that their subsequent differentiation into SMCs or
pericytes is mediated by TGF-β[106,122]. Levels of TGF-β have moreover been found to be
significantly higher in human arterial EC/SMC co-cultures than monocultures of each cell
type[155]. Thus, these heterotypic cell-cell interactions seem to mutually direct the
phenotypes of ECs and mural cells as well as favor the survival and stabilization of newly
22
formed blood vessels. Additionally, a related phenomenon was observed in a representation of
the dermis in a composite skin substitute where the presence of ECM-producing fibroblasts
improved EC survival, morphogenesis and tube formation[89].
Co-culture of EC with other cell-types is also the dominating strategy employed in the
engineering of small-diameter vascular grafts, from both biological and synthetic scaffolding
materials[157]. In these constructs, SMC are employed to provide mechanical integrity,
something which is achieved in vivo by the cells themselves, as well as the ECM components
they synthesize (mainly collagens). EC are subsequently seeded onto the luminal side of the
tubular constructs to create a confluent non-thrombogenic monolayer corresponding. Auger
and collegues[158] have gone one step further in their development of an entirely biological
vessel substitute. Working with cell sheets wrapped around a porous mandrel, an SMC sheet
representing the media was co-cultured in direct contact with an outer layer of fibroblasts. The
two cell layers eventually fused following several weeks of culture, and were subsequently
seeded with EC. This resulted in a well-defined multilayer organization similar to native tissue
although circumferential ECM fiber orientation was not observed, and the constructs were less
compliant[157,158].
With respect to co-culture in synthetic matrices, increasing knowledge is building of
such techniques and their potential in biomaterial vascularization. For instance, Wu and
coworkers[92] observed microvascular network formation of EPC-derived EC on PGA/PLLA
only when they were co-seeded with SMC. In another study comparing the performance of
two implanted PCL or PLGA porous membranes (10µm), it was shown that the PCL
membrane better supported survival of seeded SMC[47]. This result was accredited in part to
slower rate of degradation and in part to a higher degree of neovascularization within this
scaffold. In more clinically relevant studies, co- and multi-cellular strategies for
prevascularization of three-dimensional scaffolds have been proposed for the engineering of
liver and muscular tissue[14,159]. Co-seeding of hepatocytes with endothelial cells on porous
PLGA with a tube-like pore morphology (inner diameter: 500µm), resulted in endothelial cells
lining the channels and the hepatocytes attached to the polymer walls[160]. However, these
scaffolds were fabricated by a 3D printing method, and it was later stated that techniques
enabling the incorporation of much smaller features are necessary for the engineering of
capillary networks[10]. In a recent publication by Levenberg et al., the successful in vitro
23
vascularization of PLLA/PLGA sponges for regeneration of skeletal muscle was reported[14].
The scaffolds, which were fabricated using particulate leaching (pore size: 225-500µm), were
seeded with myoblasts, HUVEC, and embryonic fibroblasts in vitro and maintained in culture
to allow for prevascularization to occur prior to subcutaneous and intra-muscular implantation
in mice. The presence of embryonic fibroblasts strongly promoted vascularization of the
constructs, and these cells were found to differentiate into smooth muscle cells in the co-
culture milieu, confirming their stabilizing role. Following implantation, prevascularized
scaffolds better supported further vascularization, blood perfusion and survival of the
engineered muscle tissue[14].
It can be concluded that a large number of factors influence cell behavior in scaffolds,
many of which are related to intrinsic properties of the matrix material itself. In general,
biologically derived matrices present ECM-mimicking motifs to optimize cell-material
interactions, whereas synthetic materials offer enhanced control of material properties. Still,
despite considerable advances and increased insights into the processes involved in EC
network formation, we are still some distance away from successfully engineering vascular
beds to support 3D tissue growth on a larger scale. Furthermore, the examples discussed
clearly illustrate the relevance of co-culture strategies for tissue engineering. Co-cultures may
not only be used for comprehensive studies of paracrine interactions between different cell
types, but they offer the possibility of more autonomous and self-regulating cell populations in
TE constructs. The co-culture approach may help to overcome various difficulties encountered
when dealing solely with ECs. Cues provided, e.g. by direct or indirect cell-cell contact, a
more balanced supply of growth factors, as well as additional and more qualitatively diverse
ECM deposition are all factors that may allow for stable and functional vascular networks to
form with greater ease, thus meeting a crucial requirement of engineering of functional thick
tissue.
1.5 Bioreactor culture and mechanical conditioning
The term “bioreactor” in the context of tissue engineering is used to describe a dynamic
environment specifically designed to enhance culture conditions for a developing tissue.
Within the confines of the bioreactor, the exchange of gases and nutrients (e.g. CO2, O2, and
glucose) may be continuously monitored by sensors, and augmented by a constant medium
24
turnover. Moreover, tissue-specific mechanical forces may be applied, such as shear stress,
stretching, and compression. The controlled application of external forces has emerged as an
efficient means of regulating cell phenotype, and may therefore be used to enhance
mechanical strength of TE constructs and to more successfully mimic biological
function[157].
1.5.1 Flow conditioning
Native blood vessel endothelium is constantly subject to fluid shear stress. These flow-
induced forces affect many levels of cell function, including the mechanical stiffness and
orientation of the cell[134,161], cell proliferation, and synthesis and secretion of
proteins[162]. Shear stress affects gene transcription and expression of several signaling
molecules, such as TGF-beta, as well as integrins and adhesion molecules[108]. The native
vessel wall responds to changes in flow in a manner that regulates shear stress and maintains
homeostasis, and it has been proposed that these changes are endothelial-dependent[162,163],
and mediated by bFGF[164]. Consequently, vessels will respond to a decrease in blood flow
rate by decreasing the luminal diameter and increasing wall thickness, and conversely for a
flow rate increase[165,166]. Repeated studies of EC monolayers and endothelium lining the
vessel lumen have shown that endothelial orientation reflects the direction of the
flow[93,162,167]. In regions of high shear stress, endothelial cells are highly elongated owing
to transmission of shear forces through the cytoskeleton, resulting in an alignment of actin
stress fibers[162] and reorganization of focal adhesion contacts[168]. Finally, it is now known
that in many cases the same second messengers that are activated by chemical agonists can
also be stimulated by fluid shear stresses[162].
The role of laminar shear stress in vascular remodeling, is reflected in its enhancing
effects on endothelial network formation in vitro[169]. These effects include an integrin-
mediated increase in cytoskeleton stiffness in reply to increased applied stress[161], and may
be compared to high-stiffness substrates giving rise to additional stress fibers. In a study of
bovine EC cultures on collagen gels, Ueda et al.[169] showed the effects of fluid shear stress
(3 dyne/cm2) on EC migration and the morphology of the forming networks. Following
angiogenic stimulation with bFGF, it could be seen that shear-conditioned cultures exhibited
significantly faster rates of elongation and bifurcation in forming networks as compared to
static cultures.
25
Laminar shear forces influence both endothelial release of some growth factors, and the
responsiveness to these factors. In fact, some studies suggest the importance of flow-induced
shear stress as being more related to vascular development than the chemical stimulation of
certain growth factors. For example, increased VEGF expression in ECs has been reported
both at high (“4-fold increase”)[170] and low (1.5 dyne/cm2) [171] shear stresses relative to
physiological levels. Endothelial release of PDGF is likewise modulated by flow-induced
stress[108,165,172]. Palumbo and coworkers[173] investigated the chemotactic effect of EC-
conditioned culture media on SMC migration. PDGF-BB expression was enhanced under flow
conditions compared to static culture, but its effects were dependent on the magnitude of
laminar flow. At the higher shear stress level (15 dyne/cm2) EC produced predominantly the
PDGF-BB isoform, whereas low shear stress (5 dyne/cm2) also induced significant PDGF-AA
release. These two isomers bind the same SMC membrane receptor, and whereas PDGF-BB
alone exerted a strong chemotactic effect on SMC, the presence of PDGF-AA inhibited this
effect[173]. These observations support the notion that higher shear stress corresponds to
increased vessel wall thickness, achieved by recruitment of mural cells. Conversely, Redmond
et al.[174] demonstrated that EC subjected to shear stress produced plasminogen activator
inhibitor-1, which is inhibitory to flow-induced SMC migration.
Additionally, shear stress also functions as a regulator of the permeability of the vessel
wall to fluid and macromolecules, mediated via complex endothelial cell signaling pathways.
In vitro studies have shown that permeability can be reversibly increased 4- to 10-fold by
applying shear stresses of 1 dyne/cm2 and 10 dyne/cm2, respectively[175].
1.5.2 Cyclic strain
Cyclic strain is mainly applied to condition cells that play a role in the mechanical
integrity of tissues, such as smooth muscle cells[176] and cardiac fibroblasts[177]. This type
of mechanical stimulation has also been used extensively in the culture of artificial blood
vessels[157,178-180].
In addition to shear stress, native blood vessels are also subject to circumferential and
longitudinal stretch through the distension and relaxation of the vessel during the cardiac
cycle. These forces constitute the key mechanical stimuli of smooth muscle cells in the media,
since these cells are not normally exposed to the shear stresses exerted by blood flow. Studies
based on SMC cultured on elastic substrata or in reconstituted gels, and subjected to cyclic
26
stretching have demonstrated effects on such diverse cellular functions as SMC
phenotype[181], orientation[178,181,182], proliferation[176], growth-factor release[183,184],
and ECM production and remodelling[176,179,185]. As this research has been focused mainly
on SMC behavior under stretch, much less is known about stretch-induced effects in
endothelial cells. However, stress fiber formation[186] and increased PDGF-B
expression[172] have been reported.
Kanda et al. observed that several cell types tend to organize in a direction perpendicular
to the axis of applied cyclic stretch[182]. The same group also found increased contractile
elements, such as myofilaments, in bovine SMC cultured in a collagen-based scaffold and
subjected to 10% strain amplitude at 1Hz for up to 4 weeks[187]. Niklason et al. observed
corresponding upregulation of contractile markers in SMC grown on tubular PGA mesh[188].
These results indicate the induction of a more contractile SMC phenotype in response to
cyclic mechanical stretching.
Kim et al.[176] studied the effects of 7% stretch at 1Hz on SMC cultured on PGA/PLLA
or collagen sponges, and found strain-stimulated increase in elastin and collagen production as
well as in cell proliferation and alignment. Similar results were found for SMCs grown on
P(LA-co-CL) following 8 weeks of culture in pulsatile flow[189]. However, Kim et al.
noticed that collagen sponges did not evoke these responses when cultured in serum-free
medium, indicating that adhesion molecules such as fibronectin and vitronectin, which may
adhere from serum during culture, were necessary for the translation of stretch of the
underlying substrate into a cellular response[176].
An important aspect in tissue engineering is the common lack of mechanical strength of
the TE construct in comparison to the native tissue being replaced. Cyclic stretching has been
shown to significantly increase the mechanical strength of reconstructed artificial vessels
fabricated from both biologically derived[176] and synthetic materials[188]. It seems likely
that increasing mechanical strength is tightly connected to proliferation, cellular alignment,
phenotype (contractile vs. non-contractile), and ECM production. SMC grown on collagen
type I sponges (7% strain, 1Hz) showed dramatically higher tensile modulus and ultimate
strength after 20 weeks of culture compared to static controls[176]. At this time a clear
increase in cellular alignment on the scaffold surface could also be observed. In a series of
investigations carried out by Seliktar et al., SMC-seeded collagen gels placed under 10%
27
strain at 1Hz exhibited increased gel contraction and mechanical strength as compared to
static controls over 8 days[178]. Furthermore, the capacity of cells to remodel collagen
scaffolds was shown to be affected by mechanical stretching, through increasing expression of
matrix metalloproteinases (MMP)[179,180]. Modulation of MMP expression and/or cellular
responsiveness is important, especially in ECM–based scaffolds, to avoid excessive
degradation of the scaffolds and subsequent loss of mechanical integrity. These investigations,
which were performed using aortic SMC from rat or human, and human dermal fibroblasts,
showed great inconsistency in cell response between cell types, with rat SMC responding
more clearly to cyclic stretch and mediating the greatest increase in scaffold strength[180].
In conclusion, different mechanical stimuli, such as shear and cyclic stretching, are
capable of evoking a wide range of responses from endothelial and smooth muscle cells. The
results generated through mechanical conditioning are very promising, and these methods
represent tools that should be further explored for the promotion of scaffold vsacularization.
1.6 Conclusions
As knowledge of cellular interactions with various scaffolds, soluble factors, and other
cells, the complexity of artificial systems is developing to slowly approach conditions that will
one day allow for the successful engineering of three-dimensional tissues. However, the
promotion of vascularization in 3D tissue constructs remains the most important obstacle for
the generation of large tissue mass. The notion of prevascularized TE scaffolds seems a highly
promising approach, although such attempts have had little real success.
Currently, biodegradable synthetic scaffolds are preferred to ECM-based ones, mainly
due to the superior mechanical properties and controllable aspects of the former. Nevertheless,
more work towards the generation of blood vessel-like networks in porous synthetic materials
that mimic circulation conditions in vivo is required. For the efficient and functional
vascularization of biomaterials, the incorporation and synchronization of several factors
known to greatly influence vessel formation in vitro must be achieved. As discussed in this
review, these factors are to be found in the inherent characteristics of the chosen biomaterial;
the methods by which the scaffold may be modified; the choice of the endothelial cell source,
and whether these are cultured alone or in co-culture; the stimulation by regulatory molecules;
and the mechanical culture conditions to which the cell/polymer construct is subjected.
28
Meeting the complex requirements of scaffolds and culture conditions will be critical for the
generation of functional vascular networks, critical to successful tissue generation.
29
1.7 References
(1) Ilan N, Mahooti S, Madri JA. Distinct signal transduction pathways are utilized during
the tube formation and survival phases of in vitro angiogenesis. J. Cell. Sci. 1998; 111:3621-3631.
(2) Kannan RY, Salacinski HJ, Sales K, Butler P, Seifalian AM. The roles of tissue engineering and vascularisation in the development of micro-vascular networks: a review. Biomaterials 2005; 26:1857-1875.
(3) Folkman J, Damore PA. Blood vessel formation: What is its molecular basis? Cell 1996; 87:1153-1155.
(4) Soker S, Machado M, Atala A. Systems for therapeutic angiogenesis in tissue engineering. World J. Urol. 2000; 18:10-18.
(5) Risau W, Flamme I. Vasculogenesis. Annu. Rev. Cell Dev. Biol. 1995; 11:73-91.
(6) Stock U, Vacanti JP. Tissue engineering: current state and prospects. Annu. Rev. Med. 2001; 52:443-451.
(7) Middleton JC, Tipton AJ. Synthetic biodegradable polymers as orthopedic devices. Biomaterials 2000; 21:2335-2346.
(8) Balasubramani M, Kumar TR, Babu M. Skin substitutes: a review. Burns 2001; 27:534-544.
(9) Borenstein JT, Terai H, King KR, Weinberg EJ, Kaazempur-Mofrad MR, Vacanti JP. Microfabrication technology for vascularized tissue engineering. Biomed. Microdev. 2002; 4:167-175.
(10) Kulig KM, Vacanti JR. Hepatic tissue engineering. Transpl. Immunol. 2004; 12:303-310.
(11) Perets A, Baruch Y, Weisbuch F, Shoshany G, Nuefeld G, Smadar C. Enhancing the vascularization of three-dimensional porous alginate scaffolds by incorporating controlled release basic fibroblast growth factor microspheres. J. Biomed. Mat. Res., Part A 2003; 65A:489-497.
(12) Young D, Greulich K, Weier H. Species-specific in situ hybridation with fluorochrome-labeled DNA probes to study vascularization of human skin grafts on athymic mice. J. Burn Care Rehab. 1996; 17:305-310.
(13) Colton CK. Implantable biohybrid artificial organs. Cell Transplant. 1995; 4:415-436.
(14) Levenberg S, Rouwkema J, Macdonald M, Garfein ES, Kohane DS, Darland DC, et al. Engineering vascularized skeletal muscle tissue. Nat. Biotechnol. 2005; 23:879-884.
30
(15) Mikos AG, Sarakinos G, Lyman MD, Ingber DE, Vacanti JP, Langer R. Prevascularization of porous biodegradable polymers. Biotechnol. Bioeng. 1993; 42:716-723.
(16) Sakakibara Y, Nishimura K, Tambara K, Yamamoto M, Lu FL, Tabata Y, et al. Prevascularization with gelatin microspheres containing basic fibroblast growth factor enhances the benefits of cardiomyocyte transplantation. J. Thorac. Cardiovasc. Surg. 2002; 124:50-56.
(17) Martin Y, Vermette P. Bioreactors for tissue mass culture: Design, characterization, and recent advances. Biomaterials 2005; 26:7481-7503.
(18) Yannas IV. Tissue regeneration templates based on collagen-glycosaminoglycan copolymers. In Biopolymers II: Advances in biopolymers. Heidelberg, Germany: Springer-Verlag; 1995. p.219-244.
(19) Drury JL, Mooney DJ. Hydrogels for tissue engineering: scaffold design variables and applications. Biomaterials 2003; 24:4337-4351.
(20) Hubbell JA. Materials as morphogenetic guides in tissue engineering. Curr. Opin. Biotechnol. 2003; 14:551-558.
(21) Toman PD, Chisholm G, McMullin H, Gieren LM, Olsen DR, Kovach RJ, et al. Production of recombinant human type I procollagen trimers using a four-gene expression system in the yeast Saccharomyces cerevisiae. J. Biol. Chem. 2000; 275:23303-23309.
(22) Dvorak H, Harvey V, Estrella P, Brown L, McDonaugh J, Dvorak A. Fibrin containing gels induce angiogenesis: implications for tumor stroma generation and wound healing. Lab. Invest. 1987; 57:673-686.
(23) Dvorak H, Nagy J, Berse B, Brown L, Yeu K, Yeo T, et al. Vascular permeability factor, fibrin, and the pathogenesis of tumor stroma formation. Ann. N.Y. Acad. Sci. 1992; 667:101-111.
(24) Xue L, Greisler HP. Angiogenic effect of fibroblast growth factor-1 and vascular endothelial growth factor and their synergism in a novel in vitro quantitative fibrin-based 3-dimensional angiogenesis system. Surgery 2002; 132:259-267.
(25) Mackenzie DJ, Sipe R, Buck D, Burgess W, Hollinger J. Recombinant human acidic fibroblast growth factor and fibrin carrier regenerates bone. Plast. Reconstr. Surg. 2001; 107:989-996.
(26) Tassiopoulos AK, Greisler H. Angiogenic mechanisms of endothelialization of cardiovascular implants: a review of recent investigative strategies. J. Biomat. Sci., Pol. Ed. 2000; 11:1275-1284.
31
(27) Alexander H, Brunski JB, Cooper S, Hench L, Hergenrother R, Hoffmann A, et al. Classes of materials used in medicine. In: Ratner B, Hoffmann A, Schoen F, Lemons J, editors. Biomaterials Science: an introduction to materials in medicine.San Diego: Academic Press; 1996. p. 37-130.
(28) Yannas IV. Biologically active analogues of the extracellular matrix: Artificial akin and nerves. Angew. Chem. Int. Ed. Engl. 1990; 29:20-35.
(29) Whitworth IH, Brown RA, Dore C, Green CJ, Terenghi G. Orientated mats of fibronectin as a conduit material for use in peripheral-nerve repair. J. Hand Surg.:Brit. Eur. Vol. 1995; 20B:429-436.
(30) Harding SI, Afoke A, Brown RA, MacLeod A, Shamlou PA, Dunnill P. Engineering and cell attachment properties of human fibronectin-fibrinogen scaffolds for use in tissue engineered blood vessels. Bioprocess Biosyst. Eng 2002; 25:53-59.
(31) Pachence J, Kohn J. Biodegradable polymers for tissue engineering. In: Lanza R, Langer R, Chick W, editors. Principles of tissue engineering. 1 ed. Austin, Texas, USA: R.G. Landes Company; 1997. p. 273-293.
(32) Di Martino A, Sittinger M, Risbud MV. Chitosan: A versatile biopolymer for orthopaedic tissue-engineering. Biomaterials 2005; 26:5983-5990.
(33) Prabaharan M, Rodriguez-Perez MA, de Saja JA, Mano JF. Preparation and characterization of poly(L-lactic acid)-chitosan hybrid scaffolds with drug release capability. J. Biomed. Mat. Res., Part B 2007; 81:427-434.
(34) Zhang YF, Cheng XR, Chen Y, Shi B, Chen XH, Xu DX, et al. Three-dimensional nanohydroxyapatite/chitosan scaffolds as potential tissue engineered periodontal tissue. J. Biomat. Appl. 2007; 21:333-349.
(35) Smidsröd O, Skjåk-Braek G. Alginate as immobilization matrix for cells. Trends Biotechnol. 1990; 8:71-78.
(36) Sheridan MH, Shea LD, Peters MC, Mooney DJ. Bioadsorbable polymer scaffolds for tissue engineering capable of sustained growth factor delivery. J. Controlled Release 2000; 64:91-102.
(37) Augst AD, Kong HJ, Mooney DJ. Alginate hydrogels as biomaterials. Macromol. Biosci. 2006; 6:623-633.
(38) Martina M, Hutmacher DW. Biodegradable polymers applied in tissue engineering research: a review. Polym. Int. 2007; 56:145-157.
(39) Lee JW, Gardella JA. Surface perspectives in the biomedical applications of poly(alpha-hydroxy acid)s and their associated copolymers. Analyt. Bioanalyt. Chem. 2002; 373:526-537.
32
(40) Domb A, Kumar N, Sheskin T, Bentolila A, Slager J, Teomim D. Biodegradable polymers as drug carrier systems. In: Dumitriu S, editor. Polymeric biomaterials. 2 ed. Marcel Dekker, Inc.; 2002. p. 91-121.
(41) Cohn D, Stern T, Gonzalez M, Epstein J. Biodegradable poly(ethylene oxide)/poly(ε-caprolactone) multiblock copolymers. J. Biomed. Mat. Res. 2002; 59:273-281.
(42) Huang MH, Li SM, Vert M. Synthesis and degradation of PLA-PCL-PLA triblock copolymer prepared by successive polymerization of ε-caprolactone and (DL)-lactide. Polymer 2004; 45:8675-8681.
(43) Huang MH, Li SM, Hutmacher DW, Schantz JT, Vacanti CA, Braud C, et al. Degradation and cell culture studies on block copolymers prepared by ring opening polymerization of ε-caprolactone in the presence of poly(ethylene glycol). J. Biomed. Mat. Res., Part A 2004; 69:417-427.
(44) Li SM, Garreau H, Vert M, Petrova T, Manolova N, Rashkov I. Hydrolytic degradation of poly(oxyethylene)-poly-(ε-caprolactone) multiblock copolymers. J. Appl. Pol. Sci. 1998; 68:989-998.
(45) Huang MH, Suming LM, Coudane J, Vert M. Synthesis and characterization of block copolymers of ε-caprolactone and DL-lactide initiated by ethylene glycol or poly(ethylene glycol). Macromol. Chem. Phys. 2003; 204:1994-2001.
(46) Maquet V, Jerome R. Design of macroporous biodegradable polymer scaffolds for cell transplantation. Mat. Sci. Forum 1997; 250:15-42.
(47) Sung HJ, Meredith C, Johnson C, Galis ZS. The effect of scaffold degradation rate on three-dimensional cell growth and angiogenesis. Biomaterials 2004; 25:5735-5742.
(48) Griffith L. Emerging design principles in biomaterials and scaffolds for tissue engineering. Ann. N.Y. Acad. Sci. 2002; 961:83-95.
(49) Kumar N, Langer RS, Domb AJ. Polyanhydrides: an overview. Adv. Drug Delivery Rev. 2002; 54:889-910.
(50) Domb AJ. Synthesis and characterization of biodegradable aromatic anhydride copolymers. Macromolecules 1992; 25:12-17.
(51) Shieh L, Tamada J, Tabata Y, Domb A, Langer R. Drug-release from a new family of biodegradable polyanhydrides. J. Controlled Release 1994; 29:73-82.
(52) Laurencin CT, Pierrejacques HM, Langer R. Toxicology and biocompatibility considerations in the evaluation of polymeric materials for biomedical applications. Clin. Lab. Med. 1990; 10:549-570.
(53) Heller J, Maa YF, Wuthrich P, Ng SY, Duncan R. Recent developments in the synthesis and utilization of poly (ortho esters). J. Controlled Release 1991; 16:3-13.
33
(54) Wang C, Ge Q, Ting D, Nguyen D, Shen HR, Chen JZ, et al. Molecularly engineered poly(ortho ester) microspheres for enhanced delivery of DNA vaccines. Nat. Mater. 2004; 3:190-196.
(55) Laurencin CT, Norman ME, Elgendy HM, Elamin SF, Allcock HR, Pucher SR, et al. Use of polyphosphazenes for skeletal tissue regeneration. J. Biomed. Mat. Res. 1993; 27:963-973.
(56) Greish YE, Bender JD, Lakshmi S, Brown PW, Allcock HR, Laurencin CT. Composite formation from hydroxyapatite with sodium and potassium salts of polyphosphazene. J. Mater. Sci.: Mater. Med. 2005; 16:613-620.
(57) Ambrosio AMA, Allcock HR, Katti DS, Laurencin CT. Degradable polyphosphazene/poly(alpha-hydroxyester) blends: degradation studies. Biomaterials 2002; 23:1667-1672.
(58) Vermette P, Griesser H, Laroche G, Guidoin R. Biomedical applications of polyurethanes. Georgetown, TX, USA: Landes Biosciences, Eurekah.com; 2001.
(59) Guan JJ, Fujimoto KL, Sacks MS, Wagner WR. Preparation and characterization of highly porous, biodegradable polyurethane scaffolds for soft tissue applications. Biomaterials 2005; 26:3961-3971.
(60) Zhang JY, Beckman EJ, Hu J, Yang GG, Agarwal S, Hollinger JO. Synthesis, biodegradability, and biocompatibility of lysine diisocyanate-glucose polymers. Tissue Eng. 2002; 8:771-785.
(61) Hoffmann A. Hydrogels for biomedical applications. Adv. Drug Delivery Rev. 2002; 43:3-12.
(62) Hutmacher DW. Scaffold design and fabrication technologies for engineering tissues - state of the art and future perspectives. J. Biomater. Sci.: Polym. Ed. 2001; 12:107-124.
(63) Mann BK, Schmedlen RH, West JL. Tethered-TGF-beta increases extracellular matrix production of vascular smooth muscle cells. Biomaterials 2001; 22:439-444.
(64) Schmedlen KH, Masters KS, West JL. Photocrosslinkable polyvinyl alcohol hydrogels that can be modified with cell adhesion peptides for use in tissue engineering. Biomaterials 2002; 23:4325-4332.
(65) Huh KM, Bae YH. Synthesis and characterization of poly(ethylene glycol)/poly(L-lactic acid) alternating multiblock copolymers. Polymer 1999; 40:6147-6155.
(66) Suggs LJ, Shive MS, Garcia CA, Anderson JM, Mikos AG. In vitro cytotoxicity and in vivo biocompatibility of poly(propylene fumarate-co-ethylene glycol) hydrogels. J. Biomed. Mat. Res. 1999; 46:22-32.
34
(67) Barbato F, La Rotonda MI, Maglio G, Palumbo R, Quaglia F. Biodegradable microspheres of novel segmented poly(ether-ester-amide)s based on poly(ε-caprolactone) for the delivery of bioactive compounds. Biomaterials 2001; 22:1371-1378.
(68) George PM, Lyckman AW, Lavan DA, Hegde A, Leung Y, Avasare R, et al. Fabrication and biocompatibility of polypyrrole implants suitable for neural prosthetics. Biomaterials 2005; 26:3511-3519.
(69) Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaffolds and osteogenesis. Biomaterials 2005; 26:5474-5491.
(70) Otsuki B, Takemoto M, Fujibayashi S, Neo M, Kokubo T, Nakamura T. Pore throat size and connectivity determine bone and tissue ingrowth into porous implants: Three-dimensional micro-CT based structural analyses of porous bioactive titanium implants. Biomaterials 2006; 27:5892-5900.
(71) Yannas IV, Lee E, Orgill DP, Skrabut EM, Murphy GF. Synthesis and characterization of a model extracellular-matrix that induces partial regeneration of adult mammalian skin. Proc. Natl. Acad. Sci. U.S.A. 1989; 86:933-937.
(72) Kidd KR, Dal Ponte DB, Kellar RS, Williams SK. A comparative evaluation of the tissue responses associated with polymeric implants in the rat and mouse. J. Biomed. Mater. Res. 2002; 59:682-689.
(73) Whang K, Healy KE, Elenz DR, Nam EK, Tsai DC, Thomas CH, et al. Engineering bone regeneration with bioabsorbable scaffolds with novel microarchitecture. Tissue Eng. 1999;.5:35-51.
(74) van Tienen TG, Heijkants RGJC, Buma P, de Groot JH, Pennings AJ, Veth RPH. Tissue ingrowth polymers and degradation of two biodegradable porous with different porosities and pore sizes. Biomaterials 2002 Apr;23(8):1731-1738.
(75) Zeltinger J, Sherwood JK, Graham DA, Mueller R, Griffith LG. Effect of pore size and void fraction on cellular adhesion, proliferation, and matrix deposition. Tissue Eng. 2001; 7:557-572.
(76) McGlohorn JB, Holder WD, Grimes LW, Thomas CB, Burg KJL. Evaluation of smooth muscle cell response using two types of porous polylactide scaffolds with differing pore topography. Tissue Eng. 2004; 10:505-514.
(77) Sharkawy A, Klitzman B, Truskey G, Reichert M. Engineering the tissue which encapsulates subcutaneous implants. III. Effective tissue response times. J. Biomed. Mater. Res. 1997; 40:598-605.
(78) Li SH, de Wijn JR, Li JP, Layrolle P, de Groot K. Macroporous biphasic calcium phosphate scaffold with high permeability/porosity ratio. Tissue Eng. 2003; 9:535-548.
35
(79) Hollister SJ. Porous scaffold design for tissue engineering. Nat. Mater. 2005; 4:518-524.
(80) Sarazin P, Virgilio N, Favis BD. Influence of the porous morphology on the in vitro degradation and mechanical properties of poly(L-lactide) disks. J. Appl. Polym. Sci. 2006 15; 100:1039-1047.
(81) Sachlos E, Czernuszka J. Making tissue engineering scaffolds work. Review on the application of solid freeform fabrication technology to the production of tissue engineering scaffolds. Eur. Cells Mater. 2003;5:29-40.
(82) Auerbach R, Auerbach W, Polakowski I. Assays for angiogenesis: A Review. Pharmac. Ther. 1991;51:1-11.
(83) Cockerill GW, Gamble J, Vadas M. Angiogenesis: models and modulators. Intl. Rev. Cytology 1995;159:113-60.
(84) Folkman J, Haudenschild C. Angiogenesis in vitro. Nature 1980;288:551-6.
(85) Koike N, Fukumura D, Gralla O, Au P, Schechner J, Jain R. Creation of long-lasting blood vessels. Nature 2004;428:138-9.
(86) Campbell NA, Mathieu R. Circulation et échange gazeux. In: Édition du renouveau pédagogique Inc., editor. Biologie. 3 ed. St-Laurent, QC: The Benjamin/Cummings Publishing Company; 1995. p. 818-49.
(87) Vailhe B, Vittet D, Feige JJ. In vitro models of vasculogenesis and angiogenesis. Lab. Invest. 2001; 81:439-452.
(88) Sottile J. Regulation of angiogenesis by extracellular matrix. Biochim. Biophys. Acta, Rev. Cancer 2004; 1654:13-22.
(89) Black AF, Berthod F, L'Heureux N, Germain L, Auger FA. In vitro reconstruction of a human capillary-like network in a tissue-engineered skin equivalent. FASEB J. 1998; 12:1331-1340.
(90) Unger R, Peters K, Wolf M, Motta A, Migliaresi C, Kirkpatrick C. Endothelialization of a non-woven silk fibroin net for use in tissue engineering: growth and gene reulation of human endothelial cells. Biomaterials 2004; 25:5137-5146.
(91) Siegbahn A, Hebbel RP, Gooch KJ. The relative magnitudes of endothelial force generation and matrix stiffness modulate capillary morphogenesis in vitro. Exp. Cell Res. 2004; 297:574-584.
(92) Wu X, Rabkin-Aikawa E, Guleserian K, Perry T, Masuda Y, Sutherland F, et al. Tissue-engingeering microvessels on three-dimensional biodegradable scaffolds using human endothelial progenitor cells. Am. J. Physiol. Heart Circ. Physiol. 2004; 287:H480-H487.
36
(93) Yamamoto K, Takahashi T, Asahara T, Ohura N, Sokabe T, Kamiya A, et al. Proliferation, differentiation, and tube formation by endothelial progenitor cells in response to shear stress. J. Appl. Phys. 2003; 95:2081-2088.
(94) Saksela O, Moscatelli D, Rifkin DB. The Opposing Effects of Basic Fibroblast Growth-Factor and Transforming Growth-Factor-Beta on the Regulation of Plasminogen-Activator Activity in Capillary Endothelial Cells. J. Cell Biol. 1987; 105:957-963.
(95) Yancopoulos GD, Davis S, Gale NW, Rudge JS, Wiegand SJ, Holash J. Vascular-specific growth factors and blood vessel formation. Nature 2000; 407:242-248.
(96) Neufeld G, Cohen S, Gengrinovitch S, Poltorak Z. Vascular endothelial growth factor (VEGF) and its receptors. FASEB J. 1999; 13:9-22.
(97) Satake S, Kuzuya M, Ramos MA, Kanda S, Iguchi A. Angiogenic stimuli are essential for survival of vascular endothelial cells in three-dimensional collagen lattice. Biochem. Biophys. Res. Commun. 1998; 244:642-646.
(98) Zisch AH, Lutolf MP, Ehrbar M, Raeber GP, Rizzi SC, Davies N, et al. Cell-demanded release of VEGF from synthetic, biointeractive cell-ingrowth matrices for vascularized tissue growth. FASEB J. 2003; 17.
(99) Lafleur MA, Handsley MM, Knauper V, Murphy G, Edwards DR. Endothelial tubulogenesis within fibrin gels specifically requires the activity of membrane-type-matrix metalloproteinases (MT-MMPs). J. Cell Sci. 2002; 115:3427-3438.
(100) Nehls V, Drenckhahn D. A novel, microcarrier-based in-vitro assay for rapid and reliable quantification of 3-dimensional cell-migration and angiogenesis. Microvasc. Res. 1995; 50:311-322.
(101) Ehrbar M, Djonov VG, Schnell C, Tschanz SA, Martiny-Baron G, Schenk U, et al. Cell-demanded liberation of VEGF(121) from fibrin implants induces local and controlled blood vessel growth. Circ. Res. 2004; 94:1124-1132.
(102) Richardson TP, Peters MC, Ennett AB, Mooney DJ. Polymeric system for dual growth factor delivery. Nat. Biotechnol. 2001; 19:1029-1034.
(103) Peters MC, Polverini PJ, Mooney DJ. Engineering vascular networks in porous polymer matrices. J. Biomed. Mater. Res. 2002; 60:668-678.
(104) Presta M, Dell'Era P, Mitola S, Moroni E, Ronca R, Rusnati M. Fibroblast growth factor/fibroblast growth factor receptor system in angiogenesis. Cytokine Growth Factor Rev. 2005; 16:159-178.
(105) Wissink M, Beernink R, Pieper A, Poot A, Engbers G, Beugeling T, et al. Binding and release of basic fibroblast growth factor from heparinized collagen matrices. Biomaterials 2001; 22:2291-2299.
37
(106) Darland DC, D'Amore PA. TGF-beta is required for the formation of capillary-like structures in three-dimensional cocultures of 10T1/2 and endothelial cells. Angiogenesis 2001; 4:11-20.
(107) Szekanecz Z, Halloran M, Haskell C, Shah M, Polverini PJ, Koch A. Mediators of angiogenesis: The role of cellular adhesion molecules. Trends Glycosci. Glycotechnol. 1999; 11:73-93.
(108) Conway EM, Collen D, Carmeliet P. Molecular mechanisms of blood vessel growth. Cardiovasc. Res. 2001; 49:507-521.
(109) Korff T, Kimmina S, Martiny-Baron G, Augustin HG. Blood vessel maturation in a 3-dimensional spheroidal coculture model: direct contact with smooth muscle cells regulates endothelial cell quiescence and abrogates VEGF responsiveness. FASEB J. 2001; 15:447-457.
(110) Gamble JR, Matthias LJ, Meyer G, Kaur P, Russ G, Faull R, et al. Regulation of in vitro capillary-tube formation by anti-integrin antibodies. J. Cell Biol. 1993; 121:931-943.
(111) Mazue G, Bertolero F, Garofano L, Brughera M, Carminati P. Experience with the preclinical assessment of basic fibroblast growth factor (bFGF). Toxicol. Lett. 1992; 64-65:329-338.
(112) Springer ML, Chen AS, Kraft PE, Bednarski M, Blau HM. VEGF gene delivery to muscle: Potential role for vasculogenesis in adults. Mol. Cell 1998;2:549-558.
(113) Zisch AH, Schenk U, Schense JC, Sakiyama-Elbert SE, Hubbell JA. Covalently conjugated VEGF-fibrin matrices for endothelialization. J. Controlled Rel. 2001;72:101-113.
(114) Asakawa H, Kobayashi T. The effect of coculture with human smooth muscle cells on the proliferation, the IL-1 beta secretion, the PDGF production and tube formation of human aortic endothelial cells. Cell Biochem. Funct. 1999; 17:123-130.
(115) Kleinman HK, McGarvey ML, Liotta LA, Robey PG, Tryggvasson K, Martin GR. Isolation and characterisation of type IV procollagen, laminin, and heparin sulfate proteoglycan from the EHS sarcoma. Biochemistry. 1982; 21:6188-6193.
(116) Jackson CJ, Giles I, Knop A, Nethery A, Schrieber L. Sulfated polysaccharides are required for collagen-induced vascular tube formation. Exp. Cell Res. 1994; 215:294-302.
(117) Chalupowicz DG, Chowdhury ZA, Bach TL, Barsigian C, Martinez J. Fibrin-II induces endothelial-cell capillary-tube formation. J. Cell Biol. 1995; 130:207-215.
38
(118) Hoying JB, Boswell C, Williams SK. Angiogenic potential of microvessel fragments established in three-dimensional collagen gels. In Vitro Cell. Dev. Biol. 1996; 32:409-419.
(119) Nicosia RF, Ottinetti A. Modulation of microvascular growth and morphogenesis by reconstituted basement-membrane gel in 3-dimensional cultures of rat aorta - A comparative study of angiogenesis in matrigel, collagen, fibrin, and plasma clot. In Vitro Cell. Dev. Biol. 1990; 26:119-128.
(120) Stiffey-Wilusz J, Boice J, Ronan J, Fletcher A, Anderson M. An ex vivo angiogenesis assay utilizing commercial porcine carotidartery: Modification of the rat aortic ring assay. Angiogenesis 2001; 4:3-9.
(121) Haralabopoulos GC, Grant DS, Kleinman HK, Maragoudakis ME. Thrombin promotes endothelial cell alignment in Matrigel in vitro and angiogenesis in vivo. Am. J. Physiol.: Cell Physiol. 1997; 273:C239-C245.
(122) Hirschi KK, Rohovsky SA, D'Amore PA. PDGF, TGF-beta, and heterotypic cell-cell interactions mediate endothelial cell-induced recruitment of 10T1/2 cells and their differentiation to a smooth muscle fate. J. Cell Biol. 1998; 141:805-814.
(123) Fournier N, Doillon CJ. In-vitro effects of extracellular-matrix and growth-factors on endothelial-cell migration and vessel formation. Cells Mater. 1994; 4:399-408.
(124) Hall H, Baechi T, Hubbell JA. Molecular properties of fibrin-based matrices for promotion of angiogenesis in vitro. Microvasc. Res. 2001;62(3):315-26.
(125) Dallabrida SM, De Sousa MA, Farrell DH. Expression of antisense to integrin subunit beta(3) inhibits microvascular endothelial cell capillary tube formation in fibrin. J. Biol. Chem. 2000; 275:32281-32288.
(126) Schneller M, Vuori K, Ruoslahti E. alpha v beta 3 integrin associates with activated insulin and PDGF beta receptors and potentiates the biological activity of PDGF. Embo J. 1997; 16:5600-5607.
(127) Soldi R, Mitola S, Strasly M, Defilippi P, Tarone G, Bussolino F. Role of alpha(v)beta(3) integrin in the activation of vascular endothelial growth factor receptor-2. Embo J. 1999; 18:882-892.
(128) Yang S, Graham J, Kahn JW, Schwartz EA, Gerritsen ME. Functional roles for PECAM-1 (CD31) and VE-cadherin (CD144) in tube assembly and lumen formation in three-dimensional collagen gels. Am. J. Pathol. 1999; 155:887-895.
(129) Bach TL, Barsigian C, Chalupowicz DG, Busler D, Yaen CH, Grant DS, et al. VE-cadherin mediates endothelial cell capillary tube formation in fibrin and collagen gels. Exp. Cell Res. 1998; 238:324-334.
39
(130) Nillesen STM, Geutjes PJ, Wismans R, Schalkwijk J, Daamen WF, Van Kuppevelt TH. Increased angiogenesis and blood vessel maturation in acellular collagen-heparin scaffolds containing both FGF2 and VEGF. Biomaterials 2007; 28:1123-1131.
(131) Tanihara M, Suzuki Y, Yamamoto E, Noguchi A, Mizushima Y. Sustained release of basic fibroblast growth factor and angiogenesis in a novel covalently crosslinked gel of heparin and alginate. J. Biomed. Mater. Res. 2001; 56:216-221.
(132) Vernon R, Angello J, Iruela-Arispe L, Lane T, Sage H. Reorganization of basement membrane matrices by cellular traction promotes the formation of cellular networks in
vitro. Lab. Invest. 1992; 66:536-547.
(133) Yamada KM, Geiger B. Molecular interactions in cell adhesion complexes. Curr. Opin. Cell Biol. 1997; 9:76-85.
(134) Wang N, Butler J, Ingber D. Mechanotransduction across the cell surface and through the cytoskeleton. Science 1993; 260:1124-1127.
(135) Maheswari G, Brown G, Lauffenburger D, Wells A, Griffith L. Cell adhesion and motility depend on nanoscale RGD clustering. J. Cell Sci. 2000; 113:1677-1686.
(136) Ingber DE, Folkman J. Mechanochemical switching between growth and differentiation during fibroblast growth factor-stimulated angiogenesis in vitro: Role of extracellular matrix. J. Cell Biol. 1989; 109:317-330.
(137) Deroanne CF, Lapiere CM, Nusgens BV. In vitro tubulogenesis of endothelial cells by relaxation of the coupling extracellular matrix-cytoskeleton. Cardiovasc. Res. 2001; 49:647-658.
(138) Davis GE, Allen KAP, Salazar R, Maxwell SA. Matrix metalloproteinase-1 and -9 activation by plasmin regulates a novel endothelial cell-mediated mechanism of collagen gel contraction and capillary tube regression in three-dimensional collagen matrices. J. Cell Sci. 2001; 114:917-930.
(139) Numaguchi Y, Huang S, Polte T, Eichler G, Wang N, Ingber D. Caldesmoson-dependent switching between capillary endothelial cell growth and apoptosis through modulation of cell shape and contractility. Angiogenesis 2003; 6:55-64.
(140) Bayless KJ, Davis GE. Microtubule depolymerization rapidly collapses capillary tube networks in vitro and angiogenic vessels in vivo through the small GTPase Rho. J. Biol. Chem. 2004; 279:11686-11695.
(141) Kanzawa S, Endo H, Shioya N. Improved in vitro angiogenesis model by collagen density reduction and the use of type-III collagen. Ann. Plast. Surg. 1993; 30:244-251.
(142) Dziubla TD, Lowman AM. Vascularization of PEG-grafted macroporous hydrogel sponges: A three-dimensional in vitro angiogenesis model using human microvascular endothelial cells. J. Biomed. Mater. Res., Part A 2004; 68:603-614.
40
(143) Sharkawy A, Klitzman B, Truskey G, Reichert M. Engineering the tissue which encapsulates subcutaneous implants. II. Plasma-tissue exchange properties. J. Biomed. Mater. Res. 1997; 40:586-597.
(144) Sharkawy A, Klitzman B, Truskey G, Reichert M. Engineering the tissue which encapsulates subcutaneous implants. I. Diffusion properties. J. Biomed. Mater. Res. 1996; 37:401-412.
(145) Sieminski A, Gooch KJ. Biomaterials-microvascular interactions. Biomaterials 2000; 21:2233-2241.
(146) Clowes AW, Kirkman TR, Reidy MA. Mechanisms of arterial graft healing - rapid transmural capillary ingrowth provides a source of intimal endothelium and smooth-muscle in porous PTFE prostheses. Am. J. Pathol. 1986; 123:220-230.
(147) Salzmann DL, Kleinert LB, Berman SS, Williams SK. The effects of porosity on endothelialization of ePTFE implanted in subcutaneous and adipose tissue. Journal J. Biomed. Mater. Res. 1997; 34:463-476.
(148) Williams SK, Kleinert LB, Hagen KM, Clapper DL. Covalent modification of porous implants using extracellular matrix proteins to accelerate neovascularization. J. Biomed. Mater. Res., Part A 2006; 78:59-65.
(149) Li SM, Molina I, Martinez MB, Vert M. Hydrolytic and enzymatic degradations of physically crosslinked hydrogels prepared from PLA/PEO/PLA triblock copolymers. J. Mater. Sci.: Mater. Med. 2002; 13:81-86.
(150) Li S, Anjard S, Rashkov I, Vert M. Hydrolytic degradation of PLA/PEO/PLA triblock copolymers prepared in the presence of Zn metal or CaH2. Polymer 1997; 39:5421-5430.
(151) Dike LE, Chen CS, Mrksich M, Tien J, Whitesides GM, Ingber DE. Geometric control of switching between growth, apoptosis, and differentiation during angiogenesis using micropatterned substrates. In Vitro Cell. Dev. Biol. Anim. 1999; 35:441-448.
(152) Jain R. Molecular regulation of vessel maturation. Nat. Med. 2003;.9:685-693.
(153) Nicosia RF, Ottinetti A. Growth of microvessels in serum-free matrix culture of rat aorta. Lab. Invest. 1990; 63:115-122.
(154) Powell RJ, Cronenwett JL, Fillinger MF, Wagner RJ, Sampson LN. Endothelial cell modulation of smooth muscle cell morphology and organizational growth pattern. Ann. Vasc. Surg. 1996; 10:4-10.
(155) Axel D, Riessen R, Athanasiadis A, Runge H, Koveker G, Karsch K. Growth factor expression of human arterial smooth muscle cells and endothelial cells in a transfilter coculture system. J. Mol. Cell Cardiol. 1997; 29:2967-2978.
41
(156) Axel DI, Brehm BR, WolburgBuchholz K, Betz EL, Koveker G, Karsch KR. Induction of cell-rich and lipid-rich plaques in a transfilter coculture system with human vascular cells. J. Vasc. Res. 1996;.33:327-339.
(157) Isenberg BC, Williams C, Tranquillo RT. Small-diameter artificial arteries engineered in vitro. Circ. Res. 2006; 98:25-35.
(158) L'Heureux N, Pâquet S, Labré R, Germain L, Auger FA. A completely biological tissue-engineering human blood vessel. FASEB J. 2004; 12:47-56.
(159) Marler JJ, Upton J, Langer R, Vacanti JP. Transplantation of cells in matrices for tissue regeneration. Adv. Drug Delivery Rev. 1998; 33:165-182.
(160) Chaignaud B, Wu B, Lee HB, Utsunomiya H, Cima LG, Vacanti JP. Complex three-dimensional tissues with a vascular network using a new technology of solid free form fabrication. Am. Ped. Surgical Assoc. Annual Meeting, San Diego. 1996.
(161) Ingber D. Tensegrity: The architectural basis of cellular mechanotransduction. Annu. Rev. Physiol. 1997; 59:575-599.
(162) Ziegler T, Nerem RM. Tissue engineering of a blood vessel: Regulation of vascular biology by mechanical stress. J. Cell. Biochem. 1994; 56:204-209.
(163) Langille B. Remodelling of developing and mature arteries: endothelium, smooth muscle cells, and matrix. J. Cardiovasc. Pharmacol. 1993;21:S11-S17.
(164) Shane R, Bjercke R, Erichsen D, Rege A, Lindner V. Vascuar remodeling in response to altered blood flow is mediated by fibroblast growth factor-2. Circ. Res. 1999; 84:323-328.
(165) Aromatario C, Sterpetti A, Palumbo R, Patrizi A, Di Carlo A, Proietti P, et al. Fluid shear stress increases the release of platelet derived growth factor BB (PDGF BB) by aortic endothelial cells. Minerva Cardioangiol. 1997; 45:1-7.
(166) Greenwald SE, Berry CL. Improving vascular grafts: the importance of mechanical and haemodynamic properties. J. Pathol. 2000; 190:292-299.
(167) Nerem RM, Levesque MJ, Cornhill JF. Vascular endothelial morphology as an indicator of blood flow. ASME J. Biomech. Eng. 2004; 103:172-176.
(168) Chicurel M, Chen C, Ingber D. Cellular Control lies in the balance of forces. Curr. Opin. Cell Biol. 1998; 10:232-239.
(169) Ueda A, Koga M, Ikeda M, Kudo S, Tanishita K. Effect of shear stress on microvessel network formation of endothelial cells with in vitro three-dimensional model. Am. J. Physiol. Heart. Circ. Physiol. 2004; 287:H994-H1002.
42
(170) Milkiewicz M, Brown MD, Egginton S, Hudlicka O. Association between shear stress, angiogenesis, and VEGF in skeletal muscles in vivo. Microcirculation 2001; 8:229-241.
(171) Conklin BS, Zhong DS, Zhao WD, Lin PH, Chen CY. Shear stress regulates occludin and VEGF expression in porcine arterial endothelial cells. J. Surg. Res. 2002; 102:13-21.
(172) Sumpio B, Du W, Galagher G, Wang X, Khashigian M, Collins T, et al. Regulation of PDGF-B in endothelial cells exposed to cyclic strain. Arterioscler.,Thromb., Vasc. Biol. 1998; 18:349-355.
(173) Palumbo R, Gaetano C, Antonini A, Pompilio G, Bracco E, Rönnstrand L, et al. Different effects of high and low shear stress on platelet-derived growth factor isoform release by endothelial cells. Arterioscler.,Thromb., Vasc. Biol. 2002; 22:405-411.
(174) Redmond E, Cullen J, Cahill P, Sitzmann J, Stefansson S, Lawrence D, et al. Endothelial cells inhibit flow-induced smooth muscle cell migration: Role of plasminogen activator inhibitor-1. Circulation 2001; 103:597-603.
(175) Jo H, Dull RO, Hollis TM, Tarbell JM. Endothelial albumin permeability is shear dependent, time-dependent, and reversible. Am. J. Physiol. 1991; 260:H1992-H1996.
(176) Kim BS, Nikolovski J, Bonadio J, Mooney DJ. Cyclic mechanical strain regulates the development of engineered smooth muscle tissue. Nat. Biotechnol. 1999; 17:979-983.
(177) Wang J, Chen H, Seth A, McCulloch CA. Mechanical force regulation of myofibroblast differentiation in cardiac fibroblasts. Am. J. Physiol. Heart. Circ. Physiol. 2003; 285:H1871-H1881.
(178) Seliktar D, Black RA, Vito RP, Nerem RM. Dynamic mechanical conditioning of collagen-gel blood vessel constructs induces remodeling in vitro. Annals Ann. Biomed. Eng. 2000; 28:351-362.
(179) Seliktar D, Nerem RM, Galis ZS. The role of matrix metalloproteinase-2 in the remodeling of cell-seeded vascular constructs subjected to cyclic strain. Ann. Biomed. Eng. 2001; 29:923-934.
(180) Seliktar D, Nerem RM, Galis ZS. Mechanical strain-stimulated remodeling of tissue-engineered blood vessel constructs. Tissue Eng. 2003; 9:657-666.
(181) Cha JM, Park TN, Noh TH, Suh T. Time-dependent modulation of alignment and differentiation of smooth muscle cells seeded on a porous substrate undergoing cyclic mechanical strain. Artificial Organs 2006; 30:250-258.
(182) Kanda K, Matsuda T. Behavior of arterial-wall cells cultured on periodically stretched substrates. Cell Transplant. 1993; 2:475-484.
43
(183) Cheng GC, Briggs WH, Gerson DS, Libby P, Grodzinsky AJ, Gray ML, et al. Mechanical strain tightly controls fibroblast growth factor-2 release from cultured human vascular smooth muscle cells. Circ. Res. 199; 80:28-36.
(184) Wilson E, Mai Q, Sudhir K, Weiss RH, Ives HE. Mechanical strain induces growth of vascular smooth-muscle cells via autocrine action of PDGF. J. Cell Biol. 1993; 123:741-747.
(185) Sumpio BE, Banes AJ, Link WG, Johnson G. Enhanced collagen production by smooth-muscle cells during repetitive mechanical stretching. Arch. Surg. 1988; 123:1233-1236.
(186) Luo SS, Sugimoto K, Fujii S, Takemasa T, Fu SB, Yamashita K. Role of heat shock protein 70 in induction of stress fiber formation in rat arterial endothelial cells in response to stretch stress. Acta Histochem. Cytochem. 2007; 40:9-17.
(187) Kanda K, Matsuda T, Oka T. Mechanical stress induced cellular orientation and phenotypic modulation of 3-D cultured smooth muscle cells. ASAIO J 1993; 39:M690.
(188) Niklason LE, Gao J, Abbott WM, Hirschi KK, Houser S, Marini R, et al. Functional arteries grown in vitro. Science 1999; 284:489-493.
(189) Jeong SI, Kwon JH, Lim JI, Cho SW, Jung YM, Sung WJ, et al. Mechano-active tissue engineering of vascular smooth muscle using pulsatile perfusion bioreactors and elastic PLCL scaffolds. Biomaterials 2005; 26:1405-1411.
44
PREFACE TO CHAPTER 2. As a first step towards the fabrication of a three-
dimensional scaffold, it was judged important to have a series of biocompatible
polymers available, whose material properties could be fine-tuned to respond to
the demands of a specific application. Among these properties are mechanical
properties and the rate of degradation. The compositions chosen, the synthesis
and characterization of those polymeric materials are described in this chapter.
45
CHAPTER 2:
CHARACTERISATION, DEGRADATION AND MECHANICAL STRENGTH OF POLY(D,L-LACTIDE-CO-ε-CAPROLACTONE)-POLY(ETHYLENE GLYCOL)-POLY(D,L-LACTIDE-
CO-ε-CAPROLACTONE)
2.1 Abstract
A series of three biocompatible P(CL-co-LA)-PEG-P(CL-co-LA) copolymers were
synthesized using ring-opening polymerization and characterized by 1H-NMR, GPC, DSC,
DMA, and X-ray diffraction. The number of monomer units was kept constant, while the D,L-
LA fraction was varied so as to constitute 0, 30, or 70% of the end segments. The molecular
weights were sufficiently high to eventually permit 3D scaffold preparation. A degradation
study was carried out over 26 weeks, and the effect of monomer composition on the rate of
degradation as well as on changes in mechanical strength, was investigated. Pure PCL-PEG-
PCL copolymer, P(100/0), was a crystalline material displaying no measurable mass loss, a
30% reduction in mean molecular weight (Mn), and only very slight changes in tensile
strength. The random incorporation of 30% and 70% D,L-LA, into the end sections of the
polymer chain, produced more and more amorphous materials, exhibiting increasingly high
rates of degradation, mass loss and loss of tensile strength. Compared to random P(CL-co-
LA), the presence of the PEG block was found to improve hydrophilicity and thus the rate of
degradation. The tested copolymers range from materials exhibiting low mechanical strength
and high rate of degradation to slow-degrading materials with high mechanical strength
suitable, e.g., for three-dimensional scaffolding.
2.2 Introduction
Among the polymers most intensively studied for biomedical applications in the past
decade are polycaprolactone (PCL) and poly(lactic acids) (PLA). These resorbable polymers
belong to the group of aliphatic polyesters, and their copolymers present a wide range of
tuneable physicochemical properties depending on molecular weight, molecular structure and
monomer composition[1-3]. Thus, material properties may be tailored to suit a particular
application, with respect to e.g., mechanical strength, degradability, and drug-release profile.
Some characteristics set PCL apart from other aliphatic polyesters. Its low glass-transition
46
temperature (-60°C), makes it flexible at physiological temperature and contributes to its high
permeability for many therapeutic drugs[4]. Poly(lactic acid) is available in two isoforms,
namely the crystalline PLLA and the amorphous PDLLA. Whereas the hydrolysis of PLLA is
a very slow process, PDLLA will degrade in a matter of weeks, and even the inclusion of
very small amounts of D,L-LA in the polymer chain will dramatically accelerate
degradation[5,6]. In a similar manner it is possible to control the rate of degradation of the
highly crystalline and hydrophobic PCL by varying the CL-to-D,L-LA monomer ratio of the
corresponding co-polymer. Studied separately as well as in various copolymer formulations
and blends, CL/LA copolymers have proved useful in a number of biomedical applications
including surgical sutures[7], controlled-release drug delivery[8], and nerve guide
channels[9]. However, PDLLA-PCL-PDLLA triblock copolymers have been known to
exhibit burst-like degradation behaviour during hydrolysis[10], due to poor initial water
absorption followed by removal of the rapidly degrading PDLLA blocks. Such features
would be highly undesirable in a polymer implant that is designed to degrade in a controlled
manner in order to be gradually replaced by invading tissues. Random P(CL-co-DLLA)
copolymers, however, show stable but slow water uptake and degradation profiles[11].
Poly(ethylene glycol) (PEG) is a polymer well known for its low-fowling properties,
biocompatibility and solubility both in water and organic solvents. PEG is not susceptible to
hydrolysis, although its incorporation into the polymer backbone has been shown to enhance
the rate of degradation by increasing hydrophilicity and, thus, water uptake[12]. The effect of
PEG on the degradation of amorphous PDLLA blocks is reportedly modest[13]. In contrast,
increasing the weight fraction of PEG from 23% to 62% in PEG-PCL multiblock polymers
increased mass loss due to hydrolysis from 20% to 96%, respectively, in a study conducted
over 70 weeks[14]. PEG-containing co-polymers have been proposed for such applications as
tissue engineering[15] and drug-delivery systems[16,17]. For example, successful bone
formation was observed in vivo during a three-week study of a cytokine-liberating PDLLA-
PEG-PDLLA scaffold[16].
In the present study we chose to combine these two measures to generate P(CL-co-
DLLA)-PEG-P(CL-co-DLLA) copolymers of sufficiently high molecular weights for these
materials to be considered for scaffold fabrication. In such a ternary copolymer,
hydrophilicity is improved by including a PEG block in the polymer backbone, while at the
47
same time using a random distribution of CL and DLLA units rather than a pure block
structure, producing a more amorphous structure, whose degradation profile should be easier
to control. Importantly, although some previous reports on this type of copolymer can be
found in literature[18,19], previously investigated materials all had modest molecular weights
(MW = 2 500 – 13 500 Da) and were designed for drug-delivery applications[19]. These
materials would therefore be unsuitable for use in large, three-dimensional tissue engineering
scaffolds, which require good mechanical integrity – not only initially but also during the
course of degradation. Here, we describe and evaluate the influence of the monomer
composition (CL-to-DLLA ratio) on various aspects of hydrolytic degradation and
crystallinity. Additionally, the influence of degradation on the tensile strength of the three
test materials are investigated, which has not yet been done for this type of polymer.
2.3 Materials and Methods
2.3.1 Materials
PEG diols (MW=7500, cat. # 06103) and D,L-lactide (D,L-LA, cat. # 16640) were
purchased from Polysciences Inc. (Warrington, PA, USA). PEG diols were vacuum-dried at
least 48h at 50°C and D,L-LA was recrystallized from ethyl acetate prior to use. Stannous
octoate (SnOct, cat. # S3252) and ε-caprolactone (ε-CL, cat. # 21510) were purchased from
SigmaAldrich (Oakville, ON, Canada) and used as received. Penicillin and streptomycin were
purchased from Invitrogen (Burlington, ON, Canada; cat. # 15140-122).
2.3.2 Polymer synthesis
All copolymers were synthesized according to Bogdanov et al.[20] by SnOct-catalyzed
ring-opening polymerization of ε-CL and D,L-LA in the presence of PEG diols. The polymers
were named P(x/y) according to the molar percent content of ε-CL (x) and D,L-LA (y) in the
end segments. Briefly, for the synthesis of P(100/0), PEG diols and ε-CL were introduced
into a three-neck round-bottom flask and melted at 140°C. The molar ratio of EG-to-
monomer was 1:4.2. The temperature was then lowered to 120°C and the melted mixture
degassed and purged with N2. Stannous octoate was added at a monomer-to-catalyst ratio of
1000:1, and the polymerization was allowed to continue for 24h at 120°C under constant
stirring.
48
To obtain P(70/30) and P(30/70) copolymers, PEG diols and monomers were introduced
into a three-neck round-bottom flask at a molar ratio of monomer:EG of 4.2:1. The molar
ratio of ε-caprolactone:D,L-lactide was either 70:30 or 30:70. Following degassing and N2-
purging at 90°C, the temperature was briefly raised to 140°C to complete the melting of D,L-
LA and subsequently lowered to 120°C at which temperature polymerization was carried out
for 24h under constant stirring.
All polymers were dissolved in dichloromethane and precipitated in cold ethanol,
followed by filtration and drying in a vacuum oven at 30°C until constant weight.
2.3.3 Film preparation
Films were fabricated between Teflon sheets in a Carver Laboratory hot melt press at
150°C, using a stainless steel mold. The entire mold was subsequently cooled under running
water. The obtained film thickness was either 0.25mm or 0.75mm. Following drying of the
films, 0.75-mm thick samples were punched out using a 1-cm diameter punch, to be used for
the degradation study. Rectangular shapes (10x60mm) were cut from the 0.25-mm films for
tensile testing.
2.3.4 Characterization
The polymers were characterized by 1H-NMR (Brüker AC-300F, 300.13MHz) using
deuterated chloroform as solvent.
Gel permeation chromatography (GPC) was performed with THF as the mobile phase and
0.5ml/min flow rate, using a Waters system equipped with a differential refractometer. The
system was calibrated with polystyrene standards.
Differential scanning calorimetry (Perkin-Elmer DSC7) measurements were carried out
on 0.75-mm film samples over a temperature range of -80°C to 100°C at a rate of 10°C/min.
The apparatus was calibrated with both Hg and In. Thermal transitions were identified as the
peak values of the enthalpy changes. The percentage of CL that was able to crystallize within
the copolymer films was calculated from the thermograms, using the weight fractions of PCL
based on 1H-NMR results. The crystallinity (Xc) of the CL fraction was then calculated as:
Xc = (∆Hexo + ∆Hendo)/∆HPCL (2.1)
49
where ∆Hexo and ∆Hendo, respectively, are sample melting and crystallizing enthalpies of
the same heating run, and ∆HPCL is the reference heat of melting of pure PCL. As glass
transition temperatures (Tg) were not clearly visible in the DSC thermograms, they were
identified by dynamic-mechanical analysis (DMA), performed on a Q800 model from TA
Instruments. Measurements were carried out from -80°C to 10°C at a heating rate of
5°C/min, using the three-point flexion clamp at a frequency of 1Hz.
Wide-angle X-ray diffraction (WAXD) spectra were recorded in a Panalytical X’pert Pro
MPD (Almelo, The Netherlands) apparatus using a Cu Kα radiation source (45 kV, 40 mA)
and a scan rate of 0.44°/min.
2.3.5 Hydrolytic degradation
Polymer degradation by hydrolysis was studied by incubating the 0.75-mm-thick samples
(1-cm diameter) in a buffer solution for up to 26 weeks. Weighed samples were sterilized by
soaking in 70% ethanol for 30 minutes. Following rinsing in sterile water, samples were
immersed in 2mL PBS at pH 7.4 and 37°C. The buffer was supplemented with 100U/mL
penicillin and 100mg/mL streptomycin to prevent bacterial growth. The medium was
changed every 3 weeks for the first 6 weeks, every 2 weeks for the following three months,
and from week 18 the medium was changed every week until the end of the study. At
predetermined intervals (week 3, 6, 10, 14, 18, and 26; n=3) samples were removed from the
degradation medium, washed 3x5min in deionized water and gently blotted on filter paper
before being weighed to determine PBS absorption. Samples were then dried under vacuum,
weighed again, and prepared for analysis by 1H-NMR and GPC. Mass loss due to degradation
was determined from Equation 2.2, and the absorption of degradation medium from Equation
2.3:
Mass loss (%) = 100*(w0-wt)/w0 (2.2)
Absorption (%) = 100*(wwet-wt)/wt , (2.3)
where w0 is the initial weight of the sample, wt is the mass of the degraded sample after
drying, and wwet is the wet weight of a retrieved sample.
50
2.3.6 Tensile testing
Mechanical testing to determine tensile strength of initial (dry) and partially degraded
(wet) films was carried out according to the ASTM 882-02 standard at room temperature.
Five samples were evaluated for each material and condition. All measurements were carried
out on an Instron 5565 using a 50-N load cell and a cross-head speed of 40 mm/min. Partially
degraded films were tested while wet. Pure PCL films (CAPA 6500, Solvay Caprolactones,
Houston, USA; MW=50 kDa) were used as a reference material.
2.4 Results and Discussion
2.4.1 Characterization of PEG-containing co-polymers
Polymers prepared by the method described were characterized by 1H-NMR and GPC to
obtain the monomer composition, the molecular weight and the polydispersity index. The
results of these analyses showed a good agreement between the monomer ratios in the feed
and in the final product (Table 2.1). However, some loss of monomer in the P(30/70) material
was observed during synthesis, although the CL-to-LA ratio remained precisely the same as
in the feed. Additionally, there was a small quantity of unreacted PEG and D,L-LA during the
polymerization of P(100/0) and P(70/30), respectively. The incomplete incorporation of D,L-
LA has already been reported[1]. Moreover, we may note that the deviations between
monomer ratios in the feed and in the final polymers are well in line with the differences
between the theoretical molecular weights and those derived from 1H-NMR spectroscopy.
Finally, polydispersity indices (PI) of 1.7 were found for all of the synthesized polymers.
Table 2.1: Characterization of the three copolymers synthesized by ring-opening
polymerization.
Monomer content Molecular weight (Mn)
Polymer Feed
(CL+LA):EG Finala
Feed CL:LA
Finala Theoretical 1H-NMR GPC PIb
P(100/0) 4.2:1 4.5:1 88 900 94 800 60 900 1.7
P(70/30) 4.2:1 4.2:1 70:30 74:26 79 900 81 100 46 900 1.7
P(30/70) 4.2:1 3.1:1 30:70 30:70 67 900 52 400 35 500 1.7
Values derived from a) 1H-NMR spectra and b) GPC measurements. The thermodynamic properties were characterized by DSC and DMA, and the values of
the transition temperatures for the copolymer films are presented in Table 2.2. DMA
51
measurements revealed significantly increasing glass transition temperatures (Tg) with
increasing D,L-LA content. The single melting peak (Tm) of the first heating of P(100/0) was
found at 57°C, close to the Tm of a pure PCL. A single crystallization peak was also recorded
at 26°C, showing that P(100/0) crystallizes at temperatures close to room temperature.
P(30/70) only exhibited one melting transition, during the first heating (Tm = 40°C), and was
then unable to recrystallize under the experimental conditions. A similar absence of
crystallinity in P(CL-co-D,L-LA) of comparable composition has been previously
reported[21]. Nevertheless, the clearly semi-crystalline state of the melt- pressed film
signifies that this material crystallized either during the cooling of the mold or during storage
at room temperature.
Table 2.2: Thermodynamic transitions, fusion energy, and % of crystallinity
of P(100/0), P(70/30), and P(30/70) films. 1st heatinga 1st coolinga 2nd heatinga
Polymer Tg
b
(°C) Tm
(°C)
∆Hm
(J/g)
Tc
(°C)
∆Hc
(J/g)
Tc
(°C)
∆Hc
(J/g)
Tm
(°C)
∆Hm
(J/g)
Xc
(%CL)
P(100/0) -59 57 73 26 51 53 46 43
P(70/30) -45 40/43 39 -9 26 -22 5 40/44 39 38
P(30/70) -21 40 18 0
Data obtained from: a: DSC; b: DMA, using the peak of the loss modulus (E’’). Xc was calculated from the
2nd heating.
Furthermore, in P(70/30) but not in P(30/70), phase separation took place as indicated by
the presence of a bimodal melting peak (Tm = 40/43°C) in the DSC thermogram of the
former. Although PCL is a highly crystalline polymer, PEG blocks are sometimes able to
crystallize within copolymers provided that the block length is sufficiently large, and that
surrounding components do not hinder their crystallization[20,22]. In this case, however, the
weight fraction of PEG blocks for either material does not exceed 15%, which, according to
Gan et al.[23], is insufficient for PEG to crystallize in PEG-PCL diblock copolymers. Thus, it
was hypothesized that the bimodal melting endotherm reflected different crystallization
conditions of the PCL fraction, as previously described[23], according to e.g. their position
relative to surrounding blocks and the presence of impurities. P(70/30) exhibited
crystallization peaks (Tc) both during heating (Tc = -22°C) and cooling (Tc = -9°C), where the
52
latter is in agreement with values reported for short CL oligomers[18]. P(70/30) and P(30/70)
both exhibited lower Tm and melting enthalpies (∆Hm) compared to P(100/0). This was
expected seeing as the CL fraction in the polymer chains is smaller, but may also reflect the
influence from the amorphous D,L-LA, which strongly increase the mobility of the chains and
thereby their ability to crystallize[2].
To further investigate the nature of copolymer crystallinity, samples were analyzed by
wide-angle X-ray diffraction (WAXD). The WAXD results showed two dominant peaks for
crystalline PCL at 2θ = 21.4 and 23.7°, and two peaks for PEG at 2θ = 19.1 and 23.3° (Fig.
2.1). The spectra of P(100/0) and P(70/30) revealed that crystallinity was due entirely to the
PCL blocks. Since P(70/30) contains relatively less CL than P(100/0), this explains the lower
intensity observed, although the presence of D,L-LA in P(70/30) may also play a role in
reducing crystallinity. Finally, the P(30/70) spectrum reflects a largely amorphous material,
where the intensity of the peak (at 2θ = 21.4°) corresponding to crystalline CL-rich regions is
clearly diminished. It is very interesting to note that PEG is able to crystallize in this material.
2 Thetao16 18 20 22 24 26
Inte
nsi
ty
PEG
PCL
P(70/30)
P(100/0)
P(30/70)
Figure 2.1: WAXD curves of films of PEG-containing copolymers compared to the homo-polymers PEG (powder) and PCL (film).
The percentage of CL that was able to crystallize within the copolymer films was
calculated, using the weight fractions of PCL based on 1H-NMR results and a value of the
fusion enthalpy of 100% crystalline PCL of 135 J/g[25]. For the second heating, Xc of PCL
53
were 43, 38, and 0% in P(100/0), P(70/30), and P(30/70), respectively. Thus, we can
conclude that the presence of D,L-LA indeed has a reducing effect on PCL crystallinity,
especially in the case of non-isothermal crystallization of P(30/70). Additionally, the value of
PCL crystallinity calculated for the first heating of P(100/0) is somewhat higher, 57%, which
indicates that significant isothermal crystallization of P(100/0) films seems to take place
during storage at room temperature.
2.4.2 Hydrolytic degradation
Polymer samples were exposed to hydrolytic conditions for up to 26 weeks. Samples
were recovered at set time intervals and water absorption (Fig. 2.2a) and mass loss (Fig. 2.2b)
were determined along with changes in molecular weight and monomer composition (Fig.
2.3). Several factors come into play during degradation, including the presence of PEG
blocks that facilitate water uptake into the bulk, as does the increasingly amorphous nature
caused by higher D,L-LA content[2]. The degradation process involves two stages. It
proceeds, first, through random scission at the ester bonds along the polymer chain of the
amorphous regions. Once the amorphous regions are almost entirely degraded, the second
stage begins, involving stepwise scission of chain segments in the crystalline zones[26]. This
reaction produces carboxyl acids that lower the pH locally and are known to catalyze the
degradation process[27].
54
Degradation time (weeks)
0 5 10 15 20 25
Abs
orpt
ion
(wt%
)
0
10
20
30
40
50P(100/0)P(70/30)P(30/70)
0 5 10 15 20 25
Mas
s lo
ss (
%)
0
5
10
15
20
P(100/0)P(70/30)P(30/70)
a) b)Degradation time (weeks)
Figure 2.2: a) Water absorption of P(100/0), P(70/30), and P(30/70) during degradation in PBS (37°C, pH 7,4) for up to 26 weeks. b) Mass loss of P(100/0), P(70/30), and P(30/70) samples as degradation in PBS (37°C, pH 7,4) proceeds for up to 26 weeks. (P(30/70) were not possible to handle correctly after 3 weeks of degradation.)
55
10 12 14 16 18 20
a)0 w
10 w
26 w PI=1.9
PI=1.8
PI=1.7
10 12 14 16 18 20
b)
Elution time (min)
10 12 14 16 18 20
c)
26 w
10 w
0 w
0 w
26 w
10 w
PI=2.2
PI=2.1
PI=1.7
PI=1.4*
PI=2.1*
PI=1.7
Figure 2.3: Gel permeation chromatograms (signal intensity vs. elution time) showing the effect of degradation on the polydispersity and molecular weights of a) P(100/0), b) P(70/30), and c) P(30/70) samples after 0 (0w), 10 (10w) and 26 (26w) weeks of degradation. *Note that baseline separation was not obtained for P(30/70) samples after 10 and 26 weeks of degradation.
56
The rate of degradation was strongly influenced by D,L-LA content. P(100/0) remained
seemingly unaffected by hydrolysis within the time frame of this study, retaining its initial
shape and appearance. In sharp contrast, the P(30/70) samples, which initially were
transparent, turned white and sticky, and barely retained their shape past the 3rd week of
degradation. Beyond this time point, these samples were too difficult to handle in order to
correctly determine mass loss and water absorption. In addition to the initial molecular
weight of this copolymer being lower, 1H-NMR and GPC measurements as well as early data
(Fig. 2.2) clearly point out P(30/70) as the fastest degrading material, owing to the larger
amount of D,L-LA. Furthermore, we must take into account the physical state of the LA-
containing copolymers, which may be more prone to degradation at 37°C. In fact, as the Tm
onset is observed at 35°C (data not shown), the material is slightly more amorphous at 37°C
and should therefore be more susceptible to hydrolysis.
The water absorption of the comparatively crystalline P(100/0) remained at a constant
level of approximately 5% of the sample dry weight, and this material seemed practically
unchanged throughout the investigated time period, with the exception of a 31% total drop in
mean molecular weight (Fig. 2.4a). The water content in P(70/30) samples, immediately
exceeded 10% and began to increase steadily after 6 weeks of immersion until a level of 50%
of the dry weight was reached at 26 weeks. From Figure 2.2b, it is also clear that the polymer
had not yet reached saturation at this time point. From our results, higher LA-content was
consistent with higher water absorption (Fig. 2.2a), and increasing mass loss from the
material (Fig. 2.2b). Cho et al.[18], although studying copolymers of considerably lower
molecular weights, reported similar trends of mass loss, while they observed lower rates of
water absorption, possibly due to the shorter PEG segment. Furthermore, in a study of the
hydrolytic degradation of random P(CL-co-LA) of comparable molecular weights, Malin et
al.[11] reported drastically lower water absorption and mass loss. This further emphasizes the
significant role played by the central PEG segment.
57
Degradation time (weeks)
0 5 10 15 20 25
Mn
0
10000
20000
30000
40000
50000
60000
70000
P(100/0)P(70/30)P(30/70)
Degradation time (weeks)
0 5 10 15 20 25
(CL
+L
A):
EG
rat
io
0
2
4
6
8
P(100/0)P(70/30)P(30/70)
Degradation time (weeks)
0 5 10 15 20 25
% C
L in
end
seg
men
ts
20
30
40
50
60
70
80
90
100
P(70/30)P(30/70)
a)
b)
c)
Figure 2.4: a) The effect of degradation on the average molecular weight, Mn, as determined by GPC (n=3). Past week 10, baseline separation was no longer achieved for P(30/70) samples. The effect of degradation on monomer composition as calculated from 1H-NMR spectra: b) the (CL+LA):EG ratio; and c) the CL content in the end segments of P(70/30) and P(30/70) as compared to initial percentage. Insufficient amounts of P(30/70) remained to perform GPC at the last time point (26 weeks).
58
The changes in polydispersity and molecular weight were followed with the help of gel
permeation chromatography (Fig. 2.3) and 1H-NMR (Fig. 2.4a). The results indicate several
differences in the way degradation occurs in the three polymers. Firstly, while P(100/0) is
subject to random hydrolytic scission in the PCL chain, the polydispersity index (PI) did not
change a great deal, and the total decrease in mean molecular weight (Mn) was
approximately 30% (see Fig. 2.4a). Meanwhile, the hydrolysis of P(70/30) generates CL-LA
blocks of varying sizes depending on the distribution of D,L-LA along the macromolecular
chain. A clear attenuation in Mn was noticed (86% over 26 weeks) along with an increase in
PI as degradation proceeds. In P(30/70), Mn diminishes drastically (96% overall). However,
the changes in polydispersity were difficult to quantify as baseline separation for P(30/70)
was lost after 10 weeks. It seemed, however, that all materials showed a more or less
pronounced trend of increasing polydispersity until 14 weeks (10 weeks for P(30/70)) (data
not shown), after which it gradually decreased. Similar behaviour has been previously
reported for PDLLA-PCL-PDLLA[10] and PLLA-PEG-PLLA[12] block copolymers.
Contrary to those two reports, our polymers displayed an initially bimodal molar mass
distribution, which for the LA-containing materials eventually disappeared with increased
degradation.
After 26 weeks of degradation of P(70/30), the ratio of (CL+LA)-to-EG increased
markedly (Fig. 2.4b). This abrupt change suggests the loss of PEG blocks from the
copolymer, as reported previously by Li et al.[14]. This may be explained by the fact that the
PEG unit, although in itself unsusceptible to hydrolysis, is water soluble, and so, may be
leached from the sample once it has been separated from surrounding P(CL-co-LA)
components as the result of extensive chain cleavage. No such event was observed in
P(30/70). This fact is most likely due to extensive mass loss from the sample, including PEG
blocks as well as D,L-LA and CL; hardly anything remained of P(30/70) at the end of the
degradation period.
Additional analysis of 1H-NMR spectra revealed that the fraction of CL rose from 30 to
53% in the CL-LA portion of P(30/70) between weeks 10 and 18 of degradation, and from 74
to 87% in P(70/30) samples between weeks 14 and 26 (Fig. 2.4c). However, clear evidence
of mass loss appeared much earlier during the study (weeks 3 and 10, respectively). As the
degrading films were of sufficient thickness (0.75mm), it is suggested that this may be a
59
consequence of heterogeneous degradation mechanisms[28]. The results imply that
solubilised mono- and oligomers on the surface of the films immediately leach from the
sample, resulting in equal rates of loss of CL and LA (and PEG) due to random scission.
Meanwhile, degradation in the bulk can proceed under the additional action of autocatalysis
before the cleaved species may leave the sample[26]. The onset of leaching from the bulk
depends on the copolymer composition, and our results suggest that this appears earlier with
larger D,L-LA fraction (Fig. 2.4c). The D,L-LA units in the bulk also seem to be cleaved to a
greater extent than the CL fraction in the bulk, and are thus lost at a higher rate from the
samples at later stages of the degradation process (Fig. 2.4c).
2.4.3 Tensile strength
Tensile testing of P(100/0) and P(70/30) was performed on non-degraded samples as well
as partially degraded samples after 2-10 weeks of degradation. However, for P(30/70), given
the virtual absence of mechanical resistance, and the rapid loss of structural integrity during
degradation, only non-degraded specimens were tested. We noted that the tensile properties
of P(100/0) (Mn=95kDa) matched the tensile properties of pure PCL (Mw=50kDa). Thus, to
offer a point of reference, this data is included in Table 2.3.
Table 2.3: Changes in mechanical properties with time of degradation.
Degradation
Time (weeks)
Stress at Yield
(MPa)
Stress at Break
(MPa)
Elongation at Break
(%)
Young’s Modulus
(MPa)
PCL 0 11.1±0.6 14.6±1.7 750±56 184±7
0 13.9±1.2 18.3±2.1 747±44 254±21
2 13.0±0.8 14.7±0.8 665±68 211±11
4 14.1±1.0 15.2±1.0 601±80 223±4 P(100/0)
10 14.4±0.8 16.9±2.6 731±64 220±11
0 4.1±0.3 6.6±0.9 938±155 74±4
2 3.5±0.4 4.7±0.2 378±106 51±4
4 3.5±0.1 3.7±0.3 90±61 51±2 P(70/30)
10 - 1.8±0.2 5±1 38±3
P(30/70) 0 0.7±0.1 - > 1000 14±1
Partially degraded samples (2, 4, and 10 weeks) were tested while wet. No values for the maximum extension
were obtained for P(30/70), which did not break within the limits of the testing apparatus.
60
When comparing the non-degraded co-polymers (Table 2.3), the mechanical strength
dropped quickly with increasing LA-content. Higher LA-content also coincides with lower
molecular weight and lower degree of crystallinity, which are both factors that lower the
mechanical resistance in a polymeric material. For example, the Young’s modulus (or initial
modulus) for P(100/0) was found to be 3.5 and 18 times higher than those of P(70/30) and
P(30/70), respectively. Conversely, elongation at break increased significantly with
increasing LA-content, whereas for P(100/0) and P(70/30), the corresponding values were
747% and 938%. Clearly, out of the synthesized co-polymers, P(100/0) was the material
possessing the highest mechanical strength, both in the dry state and after 10 weeks of
incubation in aqueous solution at 37°C.
In full agreement with the above reported observations of degradation, P(100/0) did not
appear to be significantly affected by 10 weeks of hydrolytic degradation, except a slightly
decreased Young’s Modulus, which falls from an average of 254±21 MPa for dry samples to
a rather centred medium value of 218±10 MPa for all partially degraded P(100/0) samples
(weeks 2, 6, and 10 inclusively). This may be the result of a certain degree of water
absorption in the material bulk (samples were wet when tested), and/or a small drop in mean
molecular weight. In comparison, the same period of hydrolysis resulted in a steady decrease
in mechanical strength of P(70/30). At the end of the study, a decrease in ultimate stress from
6.6±0.9MPa to 1.8±0.2MPa was observed, as well as a 51% decrease in Young’s modulus.
Notably, at 10 weeks, at which time the mean molecular weight of P(70/30) had dropped
below 20kDa (Fig. 2.4a), it became brittle and no yield could be observed. Drastically
decreasing values of the elongation at break were observed, beginning with a 60% relative
drop already after 2 weeks of degradation and finally yielding a poor value of 5% for
maximum elongation at 10 weeks. Thus, within the 10-week timeframe of this study, the loss
of tensile strength due to hydrolytic degradation may be principally attributed to the presence
of LA.
2.5 Conclusions
In this study we characterize three PEG-containing copolymers with varying ratios of D,L-
lactide (D,L-LA) and ε-caprolactone (ε-CL). As crystallinity in all samples was due mainly to
the CL fraction, the pure block copolymer P(100/0) was, not surprisingly, a highly crystalline
material, and hydrolytic degradation of this material was very slow. As a consequence, the
61
tensile strength of this material was marginally affected by degradation. The incorporation of
increasing amounts of D,L-LA into the chain, rendered the copolymer increasingly
amorphous and thus, more apt to water uptake and subsequent degradation. Furthermore,
higher D,L-LA content did not only reduce the toughness of non-degraded copolymers, but
also resulted in a faster loss of mechanical resistance and integrity with degradation time.
We conclude that these copolymers are materials whose mechanical properties and
stability in vitro may be effectively altered by changing the content of D,L-LA. Based on the
findings of this study, we propose copolymers such as P(100/0) and P(70/30) as suitable
materials for large, three-dimensional scaffolds for tissue engineering.
2.6 Acknowledgments
This work was supported by the NSERC through a Discovery grant (PV) and a Strategic
grant (PV) and a post-doctorate fellowship (PS).
62
2.7 References
(1) Helminen AO, Korhonen H, Seppälä JV. Cross-linked poly(ε-caprolactone/D,L-lactide) copolymers with elastic properties. Macromol. Chem. Phys. 2002; 203:2630-2639.
(2) Huang MH, Suming LM, Coudane J, Vert M. Synthesis and characterization of block copolymers of ε-caprolactone and DL-lactide initiated by ethylene glycol or poly(ethylene glycol). Macromol. Chem. Phys. 2003; 204:1994-2001.
(3) Hiljanen-Vainio MP, Orava PA, Seppälä V. Properties of ε-caprolactone/D,L-LA (ε-CL/D,L-LA) copolymers with a minor ε-CL content. J. Biomed. Mater. Res. 1997; 34:39-46.
(4) Pachence J, Kohn J. Biodegradable polymers for tissue engineering. In: Lanza R, Langer R, Chick W, editors. Principles of tissue engineering. Austin, Texas, USA: R.G. Landes Company, 1997: p.273-293.
(5) Bergsma JE, Rozema FR, Bos RRM, Boering G, Joziasse CAP, Pennings AJ. In vitro predegradation at elevated temperatures of poly(lactide). J. Mat Sci.: Mat. Med. 1995; 6:642-646.
(6) Pitt CG. Poly-ε-caprolactone and its copolymers. In: Chasin M, Langer R, editors. Biodegradable Polymers as Drug Delivery Systems. New York: Marcel Dekker, 1990: p.71.
(7) Baimark Y, Molloy R, Molloy N, Siripitayananon J, Punyodom W, Sriyai M. Synthesis, characterization and melt spinning of a block copolymer of L-lactide and ε-caprolactone for potential use as an absorbable monofilament surgical suture. J. Mat Sci.: Mat. Med. 2005; 16:699-707.
(8) Ge H, Hu H, Yang S, Jiang X, Yang C. Preparation, characterization, and drug release behaviors of drug-loaded ε-caprolactone/L-lactide copolymer nanoparticles. J. Appl. Polym. Sci. 1999; 75:874-882.
(9) Meek M, Jansen K, Steendam R, van Oeveren W, van Wachem P, Luyn M. In vitro degradation and biocompatibiliaty of poly(D,L-lactide-ε-caprolactone) nerve guides. J. Biomed. Mater. Res. 2004; 68:43-51.
(10) Huang MH, Li SM, Vert M. Synthesis and degradation of PLA-PCL-PLA triblock copolymer prepared by successive polymerization of ε-caprolactone and (DL)-lactide. Polymer 2004; 45:8675-8681.
(11) Malin M, Hiljanen-Vainio M, Karjalainen T, Seppala J. Biodegradable Lactone Copolymers. II. Hydrolytic Study of ε-Caprolactone and Lactide Copolymers. J. Appl. Pol. Sci. 1996; 59:1289-1298.
63
(12) Li S, Anjard S, Rashkov I, Vert M. Hydrolytic degradation of PLA/PEO/PLA triblock copolymers prepared in the presence of Zn metal or CaH2. Polymer 1997; 39:5421-5430.
(13) von Burkersroda F, Gref R, Gopferich A. Erosion of biodegradable block copolymers made of poly(D,L-lactic acid) and poly(ethylene glycol). Biomaterials 1997; 18:1599-1607.
(14) Li SM, Garreau H, Vert M, Petrova T, Manolova N, Rashkov I. Hydrolytic degradation of poly(oxyethylene)-poly-(ε-caprolactone) multiblock copolymers. J. Appl. Polym. Sci. 1998; 68:989-998.
(15) Gorlatchouk A, Freier T, Shoichet M. Synthesis of degradable poly(L-lactide-co-ethylene glycol) porous tubes by liquid-liquid centrifugal casting for use as nerve guidance channels. Biomaterials 2005; 26:7555-7563.
(16) Saito N, Okada T, Horiuchi H, Murakami N, Takahashi J, Nawata M et al. A biodegradable polymer as a cytokine delivery system for inducing bone formation. Nat. Biotechnol. 2001; 19:332-335.
(17) Hu Y, Jiang X, Ding Y, Zhang L, Yang C, Zhang J et al. Preparation and drug release behaviors of nimodipine-loaded poly(caprolactone)-poly(ethylene oxide)-polylactide amphiphilic copolymer nanoparticles. Biomaterials 2003; 24:2395-2404.
(18) Cho H, An J. The effect of ε-caproyl/D,L-lactyl unit composition on the hydrolytic degradation of poly(D,L-lactide-ran-ε-caprolactone)-poly(ethylene glycol)-poly(D,L-lactide-ran-ε-caprolactone). Biomaterials 2006; 27:544-552.
(19) Cho H, Chung D, Jeongho A. Poly(D,L-lactide-ran-ε-caprolactone)-poly(ethylene glycol)-poly(D,L-lactide-ran- ε-caprolactone) as parenteral drug delivery systems. Biomaterials 2004; 25:3733-3742.
(20) Bogdanov B, Vidts A, Van Den Bulcke A, Verbeeck R, Schacht E. Synthesis and thermal properties of poly(ethylene glycol)-poly(ε-caprolactone) copolymers. Polymer 1998; 39:1631-1636.
(21) Rich J, Karjalainen T, Ahjopalo L, Seppala J. Model compound release from DL-lactide/ε-caprolactone copolymers and evaluation of specific interactions by molecular modeling. J. Appl. Polym. Sci. 2002; 86:1-9.
(22) Piao L, Dai Z, Deng M, Chen X, Jing X. Synthesis and characterization of PCL/PEG/PCL triblock copolymers by using calcium catalyst. Polymer 2003; 44:2025-2031.
(23) Gan ZH, Zhang J, Jiang BZ. Poly(ε-capralactone)/poly(ethylene oxide) diblock copolymer. 2. Nonisothermal crystallization and melting behavior. J. Appl. Polym. Sci. 1997; 63:1793-1804.
64
(24) Sosnik A, Cohn D. Poly(ethylene glycol)-poly(ε-caprolactone) block oligomers as injectable materials. Polymer 2003; 44:7033-7042.
(25) Cresceni V, Manzini G, Calzolari G, Borri C. Thermodynamics of fusion of poly-β-propiolactone and poly-ε-caprolactone. Comparative analysis of the melting of aliphatic polylactone and polyester chains. Eur. Polym. J. 1987; 8:449-463.
(26) Li S. Hydrolytic degradation characteristics of aliphatic polyesters derived from lactic and glycolic acids. J. Biomed. Mater. Res., Part B 1999; 48:342-353.
(27) Pitt CG, Gratzel M, Kimmel G, Surles J, Schindler A. Aliphatic polyesters. 2. The degradation of poly(DL-lactide), poly(ε-caprolactone) and their copolymers in vivo. Biomaterials 1981; 2:215-220.
(28) Grizzi I, Garreau H, Li S. Hydrolytic degradation of devices based on poly(DL-lactic acid) size-dependence. Biomaterials 1995; 16:305-311.
65
PREFACE TO CHAPTER 3. Following the reults of Chapter 2, it was decided –out
of pure curiosity – to look into wether blending of the co-polymers could be an
easy and efficient way to obtain intermediate rates of degradation. It was also
hypothesized that blending of slow- and fast-degrading co-polymers may result
in porous materials during the course of degradation.
66
CHAPTER 3:
BLENDS AS A STRATEGY TOWARDS TAILORED HYDROLYTIC DEGRADATION OF
P(CL-CO-D,L-LA)-PEG-P(CL-CO-D,L-LA) COPOLYMERS
3.1 Abstract
Binary blends were prepared from poly(ε-caprolactone) (PCL), and P(CL-co-D,L-lactic
acid)-P(ethylene glycol)-P(CL-co-D,L-lactic acid) copolymers, where the D,L-LA content in
the side segments varied from 0 to 70 mol%. Blend discs were fabricated by melt-molding,
and the effect of blend composition on hydrolytic degradation was studied. Variations in
medium pH were monitored, and morphological changes were observed using scanning
electron microscopy. Blending of these copolymers was found to constitute a simple means by
which intermediate rates of water absorption and mass loss were obtained, compared to those
observed in pure copolymer preparations. In one of the blends, prepared from the two
components containing 70 or 0 mol% D,L-LA in the side segments, and thereby exhibiting
large differences in degradation rate, hydrolysis resulted in the formation of a porous material
over time. Furthermore, all blend samples maintained their initial shape throughout the study.
Such materials may be interesting for further investigations for applications in cellular therapy
and controlled release.
3.2 Introduction
Polymer blends offer a simple means of combining the properties of two materials to
achieve a third. Its simplicity makes it preferable, in some cases, to copolymerisation, which
is laborious and time consuming. In the field of tissue engineering, blending two polymers
with different degradation profiles may serve to tune physical and mechanical properties[1],
or to achieve a desirable rate of degradation[2]. Parameters affecting the controlled
degradation of blends, apart from the properties of the components, include the composition,
preparation method, compatibility and miscibility of the components[3,4]. Several blend pairs
of biodegradable homo- and co-polymers have been studied, including materials composed of
poly(ε-caprolactone) (PCL), polyanhydrides, poly(ethylene glycol) (PEG), and poly(D,L-lactic
67
acid) (PDLLA)[1,2,4]. Morphological and thermodynamic analyses of polymer blends have
shown that PCL are immiscible but compatible with PEG[2,5] and PDLLA[6], whereas
PDLLA and PEG have been reported to be miscible[7]. However, in two co-
polymer/homopolymer blends, random P(CL-co-D,L-LA) co-polymers were found to be
partially compatible with PDLLA as well as with the amorphous part of the semicrystalline
PCL[4]. Furthermore, blends of PCL and PEG-PCL-PLA co-polymers exhibited properties
indicative of immiscibility of the PEG blocks with PCL[2].
The mechanism of hydrolytic degradation in large devices fabricated from aliphatic
polyesters has been described by Pitt et al.[8] and Grizzi et al.[9]. According to Grizzi et
al.[9], in the case of materials in the poly(glycolic acid)/polylactide family, initial water
uptake is very rapid and essentially homogeneous. However, after the onset of hydrolysis,
degradation proceeds at a faster rate in the sample interior due to the autocatalytic effect of
generated carboxylic acids and the difficulty of soluble degradation products to diffuse out of
the sample. As a result, a thick sample will eventually be completely degraded internally,
while a “skin” layer of less degraded polymer forms around the device, a phenomenon that
has been observed both in vitro and in vivo[10,11]. For semi-crystalline polymers, such as
PCL, the way a sample may be expected to degrade over time is thus thickness dependent.
In a recent publication, P(CL-co-D,L-LA)-PEG-P(CL-co-D,L-LA) co-polymers, containing
either 100, 70 or 30 mol% of CL in the side segments, were characterized and their
degradation profiles studied[12]. The resulting materials are referred to here as P100, P70, and
P30, respectively. Using melt-molded films, it was found that by varying the LA-content in
the co-polymers, mechanical properties and the rate of hydrolytic degradation could be
controlled. However, the degradation profiles of the three tested materials covered a very
large range, and in the case of P30, rapid degradation resulted in early loss of sample
integrity[12]. In view of previous results, this study was designed to answer two main
questions: 1) Can intermediate degradation profiles be obtained by way of co-polymer
blending? And 2) does the blending of a slow-degrading material with a fast-degrading one
yield a porous matrix during the course of degradation? In order to answer these questions
three blend pairs were prepared; PCL/P70, P100/P70, and P100/P30. Based on previous
reports on the morphology and miscibility of similar co-polymers and blends, it is reasonable
to assume that phase separation would result in PCL/P70 blends, due to the reported
68
immiscibility of PCL with both PDLLA and PEG. The same would be true for P100/P30 and
P100/P70, due to the D,L-LA content of P30 and P70, although these two blends may be
expected to exhibit a higher degree of compatibility.
3.3 Materials and Methods
3.3.1 Materials
PEG diols (MW 7 500, cat. # 06103) and D,L-LA dimers (cat. # 16640) were purchased
from Polysciences Inc (Warrington, PA, USA). SnOct (cat. # S3252) and ε-CL (cat. # 21510)
were obtained from Sigma-Aldrich (Oakville, ON, Canada). PCL pellets (CAPA® 6500, MW
50 000) were provided by Solvay Caprolactones (Houston, TX, USA).
3.3.2 Copolymers, blend preparation and sample fabrication
All copolymers were synthesized according to Bogdanov et al.[13] as described
earlier[12]. The resulting co-polymers were composed of a central PEG block with two side
segments in which the molar ratio of ε-CL:D,L-LA was either 100:0, 70:30 or 30:70. The
corresponding copolymers are referred to as P100, P70, and P30, respectively, according to
the ε-CL-content in the side segments. The obtained molecular weights for P100, P70, and
P30 were 94 800, 81 100, and 52 400 Da, respectively, as previously determined[12].
Copolymer blends were prepared from a 10% solution (wt/wt) of two (co)polymers
(weight ratio of 1:1) in dichloromethane in a round-bottom flask, where they were stirred for
one hour. Polymer blends were recovered by precipitation in cold ethanol, filtration and
drying to constant weight. The blends thus prepared were PCL/P70, P100/P70, and P100/P30.
Films were fabricated using a hot melt press at 120°C, using a stainless steel mold. From each
blend film, discs-shaped specimens of two thicknesses (1.25mm and 0.75mm) were prepared.
Samples were punched out using a 1-cm diameter biopsy punch and dried under vaccum prior
to use.
3.3.3 Hydrolytic degradation
MASS LOSS, BUFFER ABSORPTION, AND PH OF THE DEGRADATION MEDIUM
Polymer degradation by hydrolysis was studied by incubating the circular discs in a
phosphate buffer solution for up to 26 weeks. Samples were dried in vacuum overnight,
weighed, and sterilized by soaking in 70% ethanol for 30 minutes. (The average weight of
thin and thick samples was 97 and 126mg, respectively.) Each sample was then immersed in
69
2mL PBS, pH 7.4, and incubated at 37°C. The buffer was supplemented with 100U/mL
penicillin and 100mg/mL streptomycin to prevent bacterial growth. The medium was initially
changed every 3 weeks for the first 6 weeks, every 2 weeks for the following three months,
and finally every week from week 18 on. At predetermined intervals (weeks 1, 3, 6, 10/12, 14,
18, and 26; n=3), samples were removed from the degradation medium, washed 3x5min in
deionized water and gently blotted on filter paper before being weighed to determine PBS
absorption. Samples were then dried under vacuum, weighed again and the mass loss and PBS
absorption were determined from Equation 3.1 and 3.2, respectively:
Mass loss (%) = 0
0100w
ww t−× (3.1)
Absorbtion (%) = t
twet
w
ww −×100 (3.2)
where w0 is the initial weight of the sample, wt is the mass of the degraded sample after
drying, and wwet is the wet weight of a retrieved sample. In addition, changes in the pH of the
degradation buffer were monitored throughout the 26-week study.
MORPHOLOGY
Scanning electron microscopy (SEM) was used to follow any morphological changes, at
the surface and in cross-sections of thick samples, that might take place during the course of
degradation. Cross-sections were obtained by breaking the discs following quenching at
-80°C. Discs were sputter-coated with Pd/Au, and imaged at 5kV using a JEOL JSM-840-A
scanning electron microscope.
3.34 Statistical Analysis
Experimental data were evaluated for statistical significance using the Student’s t-test. A
value of p < 0.05 was considered significant.
70
3.4 Results and Discussion
3.4.1 Mass loss and buffer absorption
Figure 3.1A shows the water absorption in thin samples as a function of time and blend
composition. Absorption curves for pure co-polymer films of identical geometry were also
included for comparison [12]. At the initial absorption level (reached after one week),
absorption increased in the following order: PCL/P70 < P100/P70 < P100/P30. This order was
then conserved throughout the duration of the study. For P100/P30 blends it was observed that
the rate of absorption in thin samples increased until week 18, whereas a marked decrease was
observed in this rate between weeks 18 and 26. Similar profiles around the six-month mark
were previously reported by Huang et al.[2] for thinner (0.4mm) films fabricated from PLA-
PCL-PLA and PLA-PCL-PEG block co-polymers, and this was accredited to behaviour
“typical” of PEG and PLA blocks.
Regarding mass loss in the three blends, similar trends were observed, as may be seen in
Figure 3.1B. Twenty-six weeks of degradation resulted in 4.5-7% total mass loss in PCL/P70
and P100/P70. As expected, P100/P30 exhibited significantly faster mass loss compared to the
other blend compositions. As in the case of absorption, the rate of mass loss in thin P100/P30
samples decreased towards the end of the study. However, in this binary blend, composed of
slow-degrading P100 and fast-degrading P30, it would be expected to see the slower
degradation profile of P100 appearing towards the end of the degradation of the P30 phase.
Similar mass loss and absorption profiles were obtained for thick samples (data not shown).
Compared to pure P100, P70, and P30, intermediate rates of water absorption and mass
loss were indeed achieved for blends (see Fig. 3.1). The large differences found for pure P70
and P30 films, due to disparities in hydrophilicity, molecular weight and crystallinity[12],
induced considerable differences also in this study, as shown by elevated rates of mass loss in
P100/P30 blends compared to P100/P70. It is interesting to observe that pure P30 samples,
which were impossible to handle past three weeks of hydrolysis, contributed to faster rates of
sample degradation in this study, while the integrity of the discs was maintained by the P100
phase in the blend preparation. In fact, all blend samples studied preserved their initial shape
throughout the study. In addition, we may conclude that the PEG block in P100 facilitates
water uptake in P100/P70 blends compared to PCL/P70 (p<0.05 from weeks 1 and 6 in thin
71
and thick samples, respectively). The effect of the PEG block on mass loss was less
pronounced, but significantly higher mass loss was observed after 26 weeks (p<0.05).
Fig. 3.1: Mass loss (% of initial weight) and water absorption (% of dry weight) presented as a function of degradation time over 26 weeks. The graphs show water absorption (A) and mass loss (B) in thin samples as a function of blend composition. The profiles of pure co-polymers discs of the same thickness have been incorporated for comparison [9]. The pure co-polymers are represented by filled symbols and the blends by open symbols.
3.4.2 Evolution of pH of degradation media
The changes in pH over time of degradation are shown in Figure 3.2. The blends
composed of relatively slow-degrading (co)polymers, i.e. PCL/P100 and P100/P70, did not
induce any large fluctuation in pH over 26 weeks. P100/P30 samples provoked significantly
lower pH between eight and 20 weeks of degradation, with thin samples reaching a minimum
of 3.8 at 12 weeks and thick P100/P30 samples reaching a minimum of 3.0 at 14 weeks. The
decrease in pH observed for P100/P30 blend discs corresponds to the degradation and loss of
amorphous regions[14]. Thus, at the end of the study, the more resistant crystalline regions
remain and pH is maintained stable. The pH for thick samples behaved similarly, but was
slightly shifted towards the end of the study. Moreover, while the pH suddenly increased at 18
weeks, mass loss in thick samples continued at a nearly constant rate from week 10 to 26.
This suggests that the acidic degradation products were retained longer within the bulk of the
thicker samples, thus continuing to catalyze the hydrolytic process, and resulting in sustained
mass loss, in agreement with previous reports on thinner films[15].
Time of degradation (weeks)
0 3 6 9 12 15 18 21 24 27
Ab
s (%
)
0
20
40
60
80
100
120
P100PCL/P70P70P100/P70P30P100/P30
Time of degradation (weeks)
0 3 6 9 12 15 18 21 24 27
Mass
lo
ss (
%)
0
5
10
15
20
25
30
35
40
A) B)
72
We may also note that the pH for degrading thick samples exhibited a sharp increase after
week 18, at the same time as medium changes become more frequent. This is hardly a
coincidence, and may serve to illustrate the importance of frequent medium changes for fast-
degrading polyesters, in support of previous findings[16].
Time of degradation (weeks)
0 4 8 12 16 20 24
pH
2
3
4
5
6
7
8
Fig. 3.2: Effect of sample thickness and blend composition on the pH of the degradation media. Changes in pH of the degradation media are presented for (○) thick PCL/P70, (●) thin PCL/P70, ( ∇ ) thick P100/P70, (▼) thin P100/P70, (□) thick P100/P30, and (■) thin P100/P30. The vertical lines indicate the times at which the frequency of the change of media was changed from every three weeks to every two weeks (left line), and from every two weeks to every week (right line).
73
3.4.3 Morphology
Surfaces of P100/P30 changed considerably with continuing degradation, as can be seen
in Figure 3.3, while no changes in PCL/P70 and P100/P70 blend discs were observed, apart
from a slight cracking of the dried sample surfaces. Dispersed, micrometer-sized pores could
be seen already in non-degraded samples, and pores of a slightly larger size were visible by 10
weeks of degradation. However, rather than the initial pores simply enlarging, there seems to
be a top layer containing tiny pores, which is gradually removed, revealing an underlying
layer exhibiting fewer pores.
Fig 3.3: Scanning electron micrographs of partially degraded P100/P30 copolymer blends. Surfaces of thick samples (1,25mm) are shown un-degraded (A), and after 10 weeks (B), and 26 weeks (C). The white scale bars indicate 10µm.
Cross-sections of thick discs showed gradual changes in all blends as a result of
progressing hydrolysis (Figure 3.4). For P100/P30, signs of degradation were clearly visible
already at 6 weeks (not shown), and samples were evidently porous at weeks 18 and 26, with
homogenously dispersed pores throughout the bulk in the size range of 1-10µm. These pores
are likely the result of the degradation and solubilisation of the P30-rich phase, and their
appearnce coincide with the lower pH levels observed in the middle of this study.
74
Fig 3.4: Scanning electron micrographs of partially degraded PCL/P70, P100/P70, and P100/P30 co-polymer blends. Images of cross-sections of thick samples (1.25mm) were taken in the region close to the surface of the discs, and shown un-degraded, and after 18 weeks and 26 weeks. The white scale bars indicate 10µm.
3.5 Conclusions
The present study illustrates how the use of co-polymer blends may constitute a simple
and straightforward technique to fine tune rates of hydrolytic degradation in these materials.
The presence of a PEG block in the P100 component significantly increased water uptake in
P100/P70 blends as compared to PCL/P70. P100/P30 exhibited the highest water uptake and
mass loss, due to the LA-rich P30 component, and degradation resulted in a porous material
over time. The pH in the degradation media remained stable for PCL/P100 and P100/P70
blends, while rapid degradation in P100/P30 samples resulted in very low pH levels (pH 3-5)
2
PCL/P70
P100/P70
P100/P30
Undegraded 18 weeks 26 weeks
2
PCL/P70
P100/P70
P100/P30
Undegraded 18 weeks 26 weeks
75
in the middle of the study, reflecting the release of acidic degradation products from the
samples.
It would be of interest to optimize blending conditions, possibly using other blending
techniques, in order to investigate the development of the morphology in these degrading
blends in more detail. The objective of such a study would be to fabricate porous materials
from initially non-porous ones, as such materials may find interesting use in cellular therapy
and controlled release applications.
3.6 Acknowledgments
This work was supported by the NSERC through a Discovery grant (PV) and a Strategic
grant (PV) and a post-doctorate fellowship (PS).
76
3.7 References
(1) Tsuji H, Ikada Y. Blends of aliphatic polyesters .1. Physical properties and
morphologies of solution-cast blends from poly(DL-lactide) and poly(ε-
caprolactone). J. Appl. Polym. Sci. 1996; 60:2367-2375.
(2) Huang MH, Li SM, Hutmacher DW, Coudane J, Vert M. Degradation characteristics of poly(ε-caprolactone)-based copolymers and blends. J. Appl. Polym. Sci. 2006; 102:1681-1687.
(3) Domb AJ. Degradable polymer blends .1. Screening of miscible polymers. J. Polym. Sci., Part A: Polym. Chem. 1993; 31:1973-1981.
(4) Shen YQ, Sun WL, Zhu KJ, Shen ZQ. Regulation of biodegradability and drug release behavior of aliphatic polyesters by blending. J. Biomed. Mater. Res. 2000; 50:528-535.
(5) Tsuji H, Smith R, Bonfield W, Ikada Y. Porous biodegradable polyesters. I. Preparation of porous poly(L-lactide) films by extraction of poly(ethylene oxide) from their blends. J Appl Polym Sci 2000; 75:629-637.
(6) Aslan S, Calandrelli L, Laurienzo P, Malinconico M, Migliaresi C. Poly(D,L-lactic acid)/poly(e-caprolactone) blend membranes: preparation and morphological characterisation. J Mater Sci 2000; 35: 1615-1622.
(7) Na YH, He Y, Shuai X, Kikkawa Y, Doi Y, Inoue Y. Compatibilization effect of poly(e-caprolactone)-b-poly(ethylene glycol) block copolymers and phase morphology analysis in immiscible poly(lactide)/poly(e-caprolactone) blends.
(8) Pitt CG, Gratzel M, Kimmel G, Surles J, Schindler A. Aliphatic polyesters II. The degradation of poly(DL-lactide), poly(e-caprolactone), and their copolymers in vivo. Biomaterials 1981; 2(4): 215-220.
(9) Grizzi I, Garreau H, Li S, Vert M. Hydrolytic degradation of devices based on poly(DL-lactic acid) size-dependence. Biomaterials 1995; 16:305-311.
(10) Li SM, Garreau H, Vert M. Structure property relationships in the case of the degradation of massive poly(alpha-hydroxy acids) in aqueous-media .2. Degradation of lactide-glycolide copolymers - PLA37.5GA25 and PLA75GA25. J. Mater. Sci.: Mater. Med. 1990; 1:131-139.
(11) Therin M, Christel P, Li SM, Garreau H, Vert M. In vivo degradation of massive poly(Alpha-hydroxy acids) - validation of in vitro findings. Biomaterials 1992; 13:594-600.
(12) Bramfeldt H, Sarazin P, Vermette P. Characterization, degradation, and mechanical strength of poly(D,L-lactide-co-ε-caprolactone)-poly(ethylene glycol)-poly(D,L-lactide-co-ε-caprolactone). J. Biomed. Mater. Res., Part A 2007; 83(2): 503-511.
77
(13) Bogdanov B, Vidts A, Van Den Bulcke A, Verbeeck R, Schacht E. Synthesis and thermal properties of poly(ethylene glycol)-poly(ε-caprolactone) copolymers. Polymer 1998; 39:1631-1636.
(14) Hu D, Luh J. Role of crystallin emorphology in degradation kinetics of poly(D,L-lactide), poly(L-lactide, and poly(ε-caprolactone) for medical implants. Biomed Eng Applic Basis Commun 1994; 6:88-93.
(15) Lu L, Garcia CA, Mikos AG. In vitro degradation of thin poly(DL-lactic-co-glycolic acid) films. J. Biomed. Mater. Res. 1999; 46:236-244.
(16) Grayson ACR, Cima MJ, Langer R. Size and temperature effects on poly(lactic-co-glycolic acid) degradation and microreservoir device performance. Biomaterials 2005; 26:2137-2145.
78
PREFACE TO CHAPTER 4. As stated in the conclusion of Chapter 2, two of the synthesized co-
polymers were considered suitable for use in large, porous scaffolds. However, initial tests
showed very poor cell adhesion on films fabricated from these co-polymers, which indicated a
need for surface modification.
Co-cultures of smooth muscle cells and endothelial cells are being used as a technique to
generate stable and functional vascular networks in biomaterials, and such strategies was also
discussed in Chapter 1. As these cells constitute a natural and essential part of blood vessels, it
was chosen to optimize cell-material interactions using smooth muscle cells. In Chapter 4, the
effect of a protein-based surface modification technique on smooth muscle cell adhesion and
proliferation is evaluated.
79
CHAPTER 4:
ENHANCED SMOOTH MUSCLE CELL ADHESION AND PROLIFERATION ON
PROTEIN-MODIFIED POLYCAPROLACTONE-BASED COPOLYMERS
4.1 Abstract
Smooth muscle cells (SMC) were cultured for up to six days on copolymer films
fabricated from a PCL-PEG-PCL block co-polymer or P(ε-CL-co-D,L-LA)-PEG-P(ε-CL-co-
D,L-LA), named P(100/0) and P(70/30), respectively. The films were modified by aminolysis
using 1,6-hexanediamine, and fibronectin, fibrinogen, or fibrin layers were subsequently
immobilized by physisorption or by covalent coupling using imidoester chemistry.
Immobilization of all the tested proteins resulted in significantly enhanced cell adhesion on
these polymers. Moreover, we found that covalently immobilized proteins supported
significantly greater cell proliferation than physisorbed proteins over six days. SMC cultured
on P(100/0) films modified by covalently attached fibronectin or fibrin layers proliferated at a
rate comparable to that observed on control tissue culture polystyrene. The proposed surface
modification schemes were much less efficient in improving cell attachment and proliferation
on P(70/30) films. However, pre-wetting P(70/30) with a phosphate buffer prior to aminolysis
significantly improved cell numbers following immobilization of fibronectin. Immuno-
staining of smooth muscle-specific α-actin of SMC grown on protein-modified P(100/0) 8h
and 48h after cell seeding, confirmed preserved SMC phenotype on all modified surfaces.
4.2 Introduction
The performance of scaffolds for tissue engineering applications is governed not only by
material properties, biocompatibility and degradability, but also by the way they facilitate and
guide cellular interactions with their surfaces. Scaffold surfaces need to be conceived in a
manner that permits adequate cell adhesion, growth, morphology and viability over time in
order to meet the demands of a given application. Synthetic scaffold materials in particular,
although presenting superior mechanical properties and offering a greater degree of control
over surface chemistry and material properties, lack the biochemical motifs provided by
biologically derived materials, such as collagen and fibrin gels. Chemical and/or biochemical
80
modification of synthetic materials therefore becomes necessary in order to enhance cell-
material interactions. Numerous publications can be found describing the detailed effects of
modified synthetic-surface properties on cellular function e.g., in endothelial cells[1-3],
fibroblasts[4], smooth muscle cells[3,5], osteoblasts[6], and adipose stem cells[7].
Smooth muscle cells (SMC) are present in gastrointestinal, urological, and cardiovascular
tissues (e.g., in the lining of larger blood vessels), and represent a cell type of high interest for
tissue engineering applications. The above mentioned tissues are all examples of mechanically
dynamic environments, which explain the dependence of SMC phenotype on mechanical
stimuli, as previously demonstrated by several research groups[8-11]. Moreover, studies
comparing various protein coatings have shown that attachment strength and migration in
SMC may be controlled both by the type and concentration of plated proteins[5,12]. On the
same note, covalent modification of biomaterials with adhesion peptides and growth factors
has been shown to enhance SMC adhesion[13] as well as extracellular matrix production[14].
However, some researchers argue that isolated peptides sequences, although efficient in
promoting a particular cell response (such as the RGD motif supporting cell adhesion), are
rather poor at reproducing the diverse functions of key proteins, such as fibronectin[15].
Consequently, it becomes clear that optimizing the strength and nature of SMC adhesion to a
biomaterial would be crucial to controlling phenotype while achieving effective culture
conditions under flow and/or post-implantation.
Polyesters, including polycaprolactone (PCL), polyglycolide, polylactides (PLA) and their
co-polymers, can be activated through the covalent attachment of amino groups to PCL films
by reacting the ester groups with a diamine, such as 1,6-hexanediamine. The aminolyzed PCL
films, may then be further modified by the immobilization of proteins or adhesion peptides
using, e.g., glutaraldehyde[16] or carbodiimide chemistry[7]. These approaches were both
shown to effectively improve cell attachment and proliferation compared to unmodified and
aminolyzed PCL[7,16]. Dimethyl pimelimidate (DMP) is a homobifunctional, cross-linking
agent that may be used to couple proteins to aminated surfaces via reaction with primary
amines on proteins, while introducing a 7-carbon spacer. This water soluble molecule is
readily available and holds potential for biomaterial applications, as shown by previous
studies in which it was used for the effective immobilization of enzymes, while preserving
enzyme activity and stability[17,18].
81
The present study reports the adhesion and proliferation of smooth muscle cells grown on
surface-modified co-polymers consisting of PEG, D,L-LA, and/or ε-CL (for detailed
characterization, see Bramfeldt et al.[19]). Following aminolysis, DMP chemistry was used to
investigate the effects of immobilized proteins and thin fibrin layers on SMC adhesion and
growth. It is demonstrated that this immobilization method represents a highly efficient means
of creating bioactive materials from polyether-esters. Cell adhesion and proliferation were
both significantly improved compared to unmodified polymer films, while preserving smooth
muscle cell phenotype, as shown by smooth muscle α-actin labeling.
4.3 Materials and Methods
4.3.1 Materials
PEG diols (MW 7500, cat. no. 06103) and D,L-LA monomers (cat. no. 16640) were
purchased from Polysciences Inc. Stannous octoate (SnOct, cat. no. 3252) and ε-caprolactone
(ε-CL, cat. no. 21510) were obtained from Sigma-Aldrich. For surface modification, bovine
plasma thrombin (cat. no. T4265) and fibrinogen (cat. no. F4753), as well as fibronectin (cat.
no. F4759), 1,6-hexanediamine (cat. no.H11696), and dimethyl pimelimidate (DMP, cat. no.
D8388) were purchased from Sigma-Aldrich.
4.3.2 Co-polymer synthesis
Two co-polymers were synthesized as described earlier[19]. Briefly, macroinitiator (PEG
diols) and monomers (ε-CL or a mix of 70mol% ε-CL and 30mol% D,L-LA) were introduced
into a round-bottom flask, where they were melted and degassed under constant stirring. The
monomer:EG ratio was kept at 4.2:1 The ring-opening polymerization reaction took place at
120˚C for 24h, using SnOct as a catalyst (monomer:catalyst = 1:1000). The co-polymers were
recovered by dissolution in dichloromethane and precipitation in cold ethanol. The resulting
co-polymers were a pure block co-polymer, PCL-PEG-PCL, and a poly(ε-caprolactone-co-
D,L-lactide)-poly(ethylene glycol)-poly(ε-caprolactone-co- D,L-lactide), referred to here as
P(100/0) and P(70/30), respectively. The molecular weights of P(100/0) and P(70/30) were
95kDa and 81kDa , respectively, with polydispersity indices of 1.7. The glass transition
temperatures, Tg, of P(100/0) and P(70/30) were -59˚C and -45˚C, respectivley as previously
determined19.
82
4.3.3 Film preparation
Films were fabricated between Teflon sheets in a Carver Laboratory hot melt press at
120°C, using a stainless steel mold. The entire mold was subsequently cooled under cold,
running water. Following drying of the films, samples were punched out using a 1-cm
diameter biopsy punch. For this study, films of 0.25-mm thickness were prepared.
4.3.4 Surface modification
AMINOLYSIS
Samples (prepared as described above) were immersed in a 10wt% solution of 1,6-
hexanediamine in isopropanol, and the incubation time varied from 0-2h. The samples were
subsequently washed three times in Milli-Q water to remove excess of 1,6-hexanediamine.
P(100/0) samples were further treated in 5mM HCl to protonize the amine groups. Three
additional rinsing steps were carried out and the aminolyzed films were then dried under
vacuum and stored in a desiccator until use. P(70/30) samples were always used freshly
aminolyzed (i.e., same day).
PROTEIN IMMOBILIZATION
Aminolyzed samples were incubated in 10mM DMP in PBS, pH 8, for 1h at room
temperature. Following a rinsing step (PBS, pH 8), fibronectin and fibrinogen solutions of
predetermined concentration (10µg/mL) were added and allowed to react for 4h at 4˚C
followed by 1h at room temperature. Fibrin-mimicking surfaces were prepared by adsorbing
fibrinogen from a 100µg/ml fibrinogen (Fgn) solution onto aminated polymer films for 1h at
room temperature without previous treatment with DMP. The excess Fgn was aspirated and a
fresh mixture of 25 NIH units/ml thrombin and 10mM DMP (pH 8) was immediately added.
An immobilized fibrin-like gel layer was allowed to form for 3h at 37˚C. Control surfaces
carrying physisorbed proteins were prepared in a similar fashion, but without DMP. Following
the completed immobilization or adsorption of proteins, all modified surfaces were rinsed in
PBS, and sterilized overnight at 4˚C in PBS containing 500U/mL penicillin and 500mg/mL
streptomycin. Sterilization using antibiotics was previously shown to be an efficient and mild
alternative for aliphatic polyesters20. Finally, the modified polymer films were thoroughly
rinsed in sterile PBS containing no antibiotics and transferred to new, sterile 48-well plates,
after which they were immediately used for cell culture experiments.
83
4.3.5 Characterization of amine content
Polymer films were prepared by melt press moulding and aminolyzed, as described
above. The presence of amine groups on the films was confirmed and quantified by a
ninhydrin assay. Ninhydrin reacts with available amino groups on the surface, forming a
purple compound, which can be detected by a spectrophotometer. A 1.0M ninhydrin solution
was prepared in ethanol and the aminolyzed co-polymer discs were immersed for one minute.
The discs were then separately transferred to clean glass vials and heated in a water bath at
70˚C for 15min to accelerate the reaction. The reacted samples were dissolved in 3ml THF
and 2ml isopropanol was added to the solution. Finally, the absorbance at 570nm was
measured in a spectrophotometer. Each condition was analyzed in triplicate and identical, but
untreated polymer discs were used as controls. Pure PCL was included in the experiment as a
reference material. A standard curve was prepared using known concentrations of 1,6-
hexanediamine.
4.3.6 Cell culture
Human Umbilical Artery Smooth Muscle Cells (SMC) (Promocell, Heidelberg, Germany)
were cultured in basic growth medium recommended by the manufacturer, supplemented with
basic fibroblast growth factor (2ng/mL), epithelial growth factor (0.5ng/mL), insulin
(5µg/mL), 5% fetal bovine serum, and antibiotics (100U/mL penicillin and 100mg/mL
streptomycin). Medium was changed every two to three days, and cells were used for
experiments between passages 4-6.
4.3.7 Cytotoxicity
The cytotoxicity of the degradation media collected from partially degraded co-polymer
films was tested using a mitochondrial activity MTT assay. P(100/0) and P(70/30) film
samples (0.75 mm thick, 10 mm Ø, 100mg medium sample weight) were allowed to degrade
individually in 2mL phosphate buffer, pH 7.4, at 37˚C, and medium from the degrading
samples was collected at 2 weeks, 6 weeks, and 10 weeks and stored at 4˚C until use.
SMC were seeded into 24-well plates and allowed to proliferate. When the cells reached
near-confluence, growth medium containing 10% (v./v.) filter-sterilized degradation medium
was added to each well. Following a 24h incubation period, the medium was aspirated and the
wells rinsed in PBS. Medium containing MTT-formazan substrate (Thiazolyl Blue
Tetrazolium Bromide, cat. no. M5565, Sigma-Aldrich) at a final concentration of 5mg/ml was
84
added and allowed to react for 6h at 37˚C, whereupon the medium was removed. The formed
formazan crystals were subsequently solubilized in DMSO and a glycine buffer was added to
stabilize the product. Each condition was tested in 4 separate wells and absorption was
measured in a plate reader at λ=560nm. In addition, in order to quantify any Sn content in the
degradation media, atomic absorption spectroscopy was carried out using a Varian SpectrAA
55 apparatus at an absorption wavelength λ=235.5nm.
4.3.8 Cell adhesion and proliferation
Human umbilical artery smooth muscle cells (SMC) were detached using trypsin/EDTA
and seeded at 10,000 cells/well (7,800 cells/cm2) onto modified and unmodified polymer
samples, prepared as described in Section 4.3.4.
Cellular proliferation was followed by stopping the culture at given time intervals (24h, 3
days, and 6 days). The cells were trypsinized, stained in trypan blue, and the live cells were
subsequently counted in a haemacytometer. Prior to trypsinization, the films were carefully
transferred to new multiwell plates and rinsed in PBS to avoid counting loosely adhering cells
and cells adhering to underlying tissue culture polystyrene (TCPS). Three discs of each
surface type were counted together (considered as n=1), and the cell number/cm2 was
calculated. Cell numbers were then normalized to the number of adhering cells on TCPS after
24h. For each experiment, n=3-4 (i.e., 9-12 discs). Aminolyzed and unmodified co-polymers
were used as controls. In addition, SMC cultured on Fn-immobilized films as well as on TCPS
were trypsinized at the end of the experiment, reseeded, and labeled for α-actin to confirm
differentiated phenotype.
The morphology of cells cultured on modified co-polymers was studied using scanning
electron microscopy (SEM) and fluorescent staining of α-actin. As P(70/30) did not support
the fixation procedure, only P(100/0) films could be analyzed. For SEM imaging, cells were
fixed in glutaraldehyde (1,4% for 30min at room temperature and 2,5% overnight at 4˚C ),
post-fixed in OsO4 (1h at room temperature), followed by dehydration in serial dilutions of
ethanol and drying under vacuum. (Neither co-polymer resists critical point drying due to low
Tm[19]). Samples were sputter-coated in Au/Pd and imaging was carried out in a JSM-840-A
JEOL microscope (V=5kV, I=6×10-8A). P(70/30) samples did not resist the chemical
treatment used to fix the cells for SEM imaging and therefore only P(100/0) films could be
analyzed.
85
Further evaluation of the cell shape during the adhesion and the subsequent spreading of
the cells on various surfaces, was carried out by fluorescent labeling of smooth muscle cell-
specific α-actin using a mouse monoclonal antibody. A secondary FITC-conjugated goat-anti-
mouse antibody was used to visualize the labeling. Cells were observed using a Nikon TE200-
S inverted microscope equipped with a high power mercury lamp.
4.3.9 Statistical analysis
Experimental data from MTT and ninhydrin assays were evaluated for statistical
significance using the Student’s t-test, and significance was considered for p<0.05. Cell
adhesion and proliferation data were analyzed using ANOVA. When overall significance
(P<0.05) was indicated, data pairs within a group were compared using calculated values for
critical difference (CD). Data pairs between groups were compared using Student’s t-test.
86
4.4 Results and Discussion
4.4.1 Cytotoxicity
The cytocompatibility of the two co-polymers P(100/0) and P(70/30) was evaluated using
the MTT assay. (For the results of the degradation study, the reader is referred to Chapter 2.)
The cell viability after 24h incubation in culture medium containing 10% degradation medium
is shown in Figure 4.1. No significant difference in viability compared to TCPS cultures was
observed for degrading P(100/0) or for P(70/30) at two and six weeks of degradation.
However, at 10 weeks of degradation, the P(70/30) degradation medium caused a significant
decrease in cell viability (64±10%) compared to the TCPS control (p<0.05).
Fig 4.1: MTT mitochondrial activity assay of the cytocompatibility of degradation medium. The degradation medium was collected after 2, 6, and 10 weeks of hydrolytic degradation.
In an attempt to identify a possible cause of the observed cytotoxic effect of the 10-week
P(70/30) sample, the pH of the degradation media was monitored. The pH of the P(70/30)
degradation medium was observed to decrease from an initial value of 7.4 to 6.3 following 10
weeks of degradation (data not shown). However, at the concentration tested (10% v./v.), the
pH did not deviate more than 0.1 units from 7.4 when measured immediately following
mixing, and had in all cases returned to physiological pH after 30 minutes in room
temperature. It was also thought that residual Sn from the co-polymer synthesis may affect
cell viability, and a verification of the Sn content in the degradation media was made using
Control P(100/0) P(70/30)
Cel
l v
iab
ilit
y (
%)
0
20
40
60
80
100
120
Control2 weeks6 weeks10 weeks
87
atomic absorption spectroscopy. All samples were tested and they were all found to contain 2-
3ppm Sn. This is approximately 10 times smaller than the IC50 of Sn previously reported for
fibroblasts[22]. Therefore, in the final concentration (10%) used for the MTT assay, it is
unlikely that Sn would have caused the observed decrease in cell viability.
4.4.2 Surface modification
Co-polymer films, prepared as described, were aminolyzed for 0, 0.5, 1, or 2h and
analyzed for free amino group content using a colorimetric ninhydrin assay. The results were
compared to a reference pure PCL film, and can be seen in Figure 4.2. No significant increase
in amine groups between 0.5 and 2h was observed on PCL or P(70/30). The comparable
amine content on PCL and P(100/0) films following 0.5h and 1h of aminolysis, indicates that
the presence of unreactive PEG-blocks does not interfere with the reaction. Seeing as the PEG
content of P(100/0) decreases the degree of crystallization compared to PCL[19], the
amorphous regions would possibly allow for the diamine molecules to more effectively
penetrate the film surface with longer incubation times, which may be a possible explanation
for the rapid and significant increase in amine content between 1h and 2h (p<0.05). Moreover,
it has been reported that amorphous regions more aptly react with diamines as they increase
surface wettability[22]. This was supported by the observations made of amine group
incorporation onto P(70/30). This material is even more amorphous in nature[19], and rapidly
reached amine densities higher than those of both P(100/0) and PCL (Fig. 4.2). However, only
small increases in amine group content were achieved with prolonged times of incubation.
The incubation times of the two co-polymers were selected to achieve a similar degree of
surface activation, while minimizing the negative impacts of prolonged incubation times on
mechanical strength[16,22]. Therefore, for protein immobilization schemes, P(100/0) films
were incubated for 2h, while P(70/30) aminolysis was limited to 30 minutes.
In view of the poor results obtained in previous studies[16] as well as in our laboratory,
using surface characterization techniques such as X-ray photoelectron spectroscopy (XPS) and
Fourier-transform infrared (FT-IR) spectroscopy, no additional characterization is reported
here. Nonetheless, cell behavior was considered to constitute a reliable, although non-
quantitative, indication of successful surface modification.
88
Fig 4.2: (A) Absorbance at 570nm of aminolyzed copolymer films, showing the dependence of incubation time on the amount of 1,6-hexanediamine incorporated in the copolymer film surfaces. A pure PCL reference is included for comparison.
The significance towards a final engineered tissue composite of the three protein
compositions chosen for surface modification here may be illustrated by the integrin
interactions they trigger. These integrins, in turn, activate distinct intracellular signaling
pathways that play an important role in regulating cell phenotype[23-26]. SMC adhesion to Fn
stimulates expression of several integrins, including the main Fn receptor α5β1, as well as
integrins α8β1, and αvβ3[26]. The α8β1 integrin is also a known differentiation marker in
vascular smooth muscle cells, and it was recently shown that expression of the α8 subunit is
required to maintain a differentiated phenotype[24]. Fibronectin is, in fact, the only adhesion
protein recognized by the α5β1 integrin, which plays a key role in the formation of fibrillar
adhesions[15]. Moreover, both α5β1 and αvβ3 receptors are implicated in the regulation of
angiogenesis[15] and Fn has been suggested as a suitable matrix component in scaffolds
designed to regenerate blood vessels[27,28]. The αvβ3 integrin is likewise expressed in cells
adhering to fibrin and fibrinogen[26], and is believed to enhance SMC migration and
proliferation in vivo[23]. And although these are not always desirable events in vivo, they may
be profitable when promoting cellular scaffold invasion in vitro.
A) Time of aminolysis (h)
0,0 0,5 1,0 1,5 2,0
A.U
. ( λλ λλ
=5
75
nm
)
0
30
60
90
120
150
180
NH
2 g
rou
ps
( µµ µµm
ol/
cm2)
0
2
4
6
8
10
PCLP(100/0)P(70/30)
*
A) Time of aminolysis (h)
0,0 0,5 1,0 1,5 2,0
A.U
. ( λλ λλ
=5
75
nm
)
0
30
60
90
120
150
180
NH
2 g
rou
ps
( µµ µµm
ol/
cm2)
0
2
4
6
8
10
PCLP(100/0)P(70/30)
*
89
4.4.3 Smooth muscle cell culture
SMC ADHESION AND PROLIFERATION ON P(100/0)
Cell adhesion was evaluated 24h after seeding, and cell proliferation was assessed
following three and six days in culture. Cell adhesion was improved on all protein-modified
surfaces, compared to unmodified and aminolyzed P(100/0) (see Fig. 4.3). No significant
difference in cell adhesion on P(100/0) was found between the two latter groups. Cell
proliferation showed a stronger dependence on protein type than did cell adhesion. Notably,
surfaces bearing covalently immobilized Fn and fibrin showed cell numbers comparable to
those on TCPS surfaces after six days. Fibrinogen alone stimulated strong proliferative
activity in SMC, although cell numbers were clearly lower on Fgn (229 ± 27%) than on fibrin
(304 ± 16%) on day six.
In addition, covalently immobilized proteins evoked different cell responses than
physisorbed proteins, indicating that the degree of substrate anchorage was a determining
factor. Adsorbed Fn promoted more efficient SMC adhesion (78±15%) than immobilized Fn
(59±10%) (p<0.01). Moreover, a large difference was found in proliferation on physisorbed
versus covalently coupled Fn, where cell numbers on physisorbed Fn were significantly lower
after six days in culture (p<0.01). A similar but more pronounced trend was observed for
fibrin-mimicking surfaces. Despite good initial cell adhesion, the physisorbed fibrin did not
support any significant proliferation over six days. The dependence of early cell adhesion
mechanisms on Fn surface anchorage was previously shown[2], and such mechanisms may
play a role in subsequent proliferation rates. However, as for the differences observed between
the adsorbed fibrin and Fn, the displacement of proteins from the surface may play an
important role. It can be hypothesized that the lower cell numbers observed on physisorbed
fibrin compared to that on Fn, may be explained by a more rapid and complete detachment of
the gel-like fibrin layer compared to individually adsorbed Fn molecules. Further
investigations will be needed to address this hypothesis.
Finally, in order to verify SMC phenotype, cells that were grown on Fn-immobilized
P(100/0) and TCPS were trypsinized after 6 days of culture, plated separately on TCPS and
stained for α-actin. Both cell groups exhibited strong α-actin labeling (see Appendix, Fig A3),
90
and it was concluded that differentiated phenotype was conserved in these samples throughout
the duration of this experiment.
Fig 4.3 : Cell adhesion and proliferation on aminolyzed and protein-immobilized P(100/0) films. The P(100/0) surfaces were aminolyzed (NH2), and modified with covalently immobilized (“imm”) fibronectin (Fn), fibrin (Fib), and fibrinogen (Fgn) or with adsorbed (“ads”) Fn and Fib. Unmodified P(100/0) and TCPS were included for comparison. All cell numbers are normalized to TCPS at 24h. * indicates significantly higher cell adhesion numbers after 24h compared to the aminated surface (NH2) at 24h. ** indicates significant increase in cell number from 24h to 6 days within one group, p<0.05.
SMC ADHESION AND PROLIFERATION ON P(70/30)
Cell-surface interactions on P(70/30) clearly were not optimized using the surface
modification method described here. The observed cell numbers (Fig. 4.4) indicated
unsuccessful protein immobilization, in turn signifying different surface chemistry as
compared to P(100/0). This was somewhat unexpected seeing as both co-polymers
investigated were poly(ether-esters) with much the same monomer content. However,
although partial crystallization takes place in both materials, P(70/30) is a much more
amorphous material compared to the highly crystalline P(100/0)[19]. Additionally, the onset
of water absorption and decreasing molecular weights in P(70/30) was observed to occur
much earlier during hydrolytic degradation. It was thus hypothesized that unsuccessful protein
TCPS P(100/0) NH Fn ads Fn imm Fib ads Fib imm Fgn imm
Cel
l n
um
ber
(%
TC
PS
at
24
h)
0
100
200
300
400
24h3 days6 days
*
* **
*
**
****
**
2TCPS P(100/0) NH Fn ads Fn imm Fib ads Fib imm Fgn imm
Cel
l n
um
ber
(%
TC
PS
at
24
h)
0
100
200
300
400
24h3 days6 days
*
* **
*
**
****
**
TCPS P(100/0) NH Fn ads Fn imm Fib ads Fib imm Fgn imm
Cel
l n
um
ber
(%
TC
PS
at
24
h)
0
100
200
300
400
24h3 days6 days
*
* **
*
**
****
**
2
91
immobilization may have been the result of polymer chain reorganization. In both cases, the
“hiding” of incorporated amine groups from the surface layer, would drastically diminish the
capacity for protein immobilization using DMP. In order to investigate the plausibility of this
explanation, dry samples as well as P(70/30) samples that were pre-wet for 24h in PBS (pH
7.4) at room temperature were used for aminolysis and subsequent protein modifications, and
the cell response was compared. The results can be seen in Figure 4.4.
From Figure 4.4, it is clear that pre-wet samples were superior to dry-treated films in terms
of cell numbers. Aminated surfaces without proteins did not improve cell adhesion, regardless
of whether samples were pre-wet or not. However, at six days, the pre-wet aminolyzed
P(70/30) supported significantly higher cell numbers (p<0.05) than their dry-treated
counterparts. Among the surfaces tested, only Fn immobilized on pre-wet P(70/30) lead to a
significant increase in adhering cell numbers compared to unmodified P(70/30) (p<0.01).
Moreover, the surfaces tested, only pre-wet P(70/30) bearing covalently coupled Fn supported
weak, but non-significant cell proliferation between days 1 and 6 (27±6% and 44±13%,
respectively). This stands in sharp contrast to the poor cell adhesion observed on its dry-
treated counterpart. No significant changes in cell numbers between days 1 and 3 on any
surface were observed. Adsorbed Fn on pre-wet surfaces also lead to markedly improved cell
adhesion compared to unmodified and aminated co-polymers, although in contrast to P(100/0)
results, this surface treatment did not support cell proliferation. The more hydrophilic nature
of P(70/30) compared to P(100/0), while as yet unmeasured, is clearly visible to the eye.
Displacement of physisorbed Fn was previously shown to take place to a greater extent from
hydrophilic polymer surfaces, whereas hydrophobic polymer substrates are better able to
retain the proteins[2]. Differences in Fn displacement from the two surfaces may thus offer a
partial explanation as to why cell numbers continue to multiply over six days of culture on
physisorbed Fn on P(100/0), but not on the corresponding P(70/30) surface.
92
Fig 4.4 : SMC growth was compared on Fn-modified P(70/30), previously aminated from either pre-wet (marked with ♦ in group labels) or dry films. The effect of amination alone, covalently immobilised Fn (“imm”) on dry and pre-wet films as well as absorbed Fn (“abs”) were investigated. Unmodified P(70/30) was included for comparison. The 100% level corresponds to the cell number on TCPS after 24h of culture. ** in the graph indicates significantly higher cell adhesion numbers after 24h compared to the NH2* surface at 24h, p<0.05.
SMC MORPHOLOGY
SMC in vivo are characterized by phenotypic plasticity. Highly responsive to
environmental cues, divergent populations exhibiting different phenotypes may be present
within one tissue at the same time[29]. Via reversible mechanisms, SMC may switch from an
immature, proliferative, synthetic phenotype to a quiescent, contractile state and back
again[29,30]. These phenotypic conditions are characterized by specific cell shapes. Synthetic
SMC have a more hypertrophic and rectangular morphology while the contractile phenotype
appears bipolar and spindle-shaped[31].
Cell morphology and spreading were observed using scanning electron microscopy (SEM)
and fluorescent α-actin labeling. Unfortunately, the poor resistance of P(70/30) samples to the
fixation procedures limited SEM imaging to include P(100/0) surfaces only (Fig. 4.5). Firstly,
it can be seen that SMC attached on aminolyzed P(100/0) (Fig. 4.5D) are much smaller (60
µm) than cells on protein-modified surfaces (100 µm) (Fig. 4.5A-C). All cells showed distinct
filopodia 8h after seeding. SMC grown on protein-modified surfaces became more elongated
with time, with some extremely elongated features (100 µm) appearing at 48h. Additionally,
parallel grooves, resulting from irregularities in the Teflon sheets used during melt press film
fabrication, are clearly visible in the SEM images. However, there was no preferential cellular
alignment along these surface features.
P(70/30) NH NH Fn ads Fn imm Fn imm*
Cel
l n
um
ber
(%
of
TC
PS
at
24h
)
0
10
20
30
40
50
60
24h3 days6 days
2 2
**
**
♦ ♦ ♦P(70/30) NH NH Fn ads Fn imm Fn imm*
Cel
l n
um
ber
(%
of
TC
PS
at
24h
)
0
10
20
30
40
50
60
24h3 days6 days
2 2
**
**
♦ ♦ ♦♦ ♦ ♦
93
8h 48h
A)
B)
C)
D)
8h 48h
A)
B)
C)
D)
Fig 4.5: α-actin staining of SMC on modified P(100/0), 8h and 48h after seeding. Surfaces were treated by aminolysis (A), followed by physisorption of Fn (B), or by covalent immobilization of Fn (C) and Fgn (D). The white bar indicates 100µm in all images but for (A) 8h, where the white bar indicates 10µm.
94
8h 48h
A)
B)
C)
D)
E)
8h 48h
A)
B)
C)
D)
E)
Fig 4.6 : Alpha-actin staining of SMC on modified P(100/0), 8h and 48h after seeding. Surfaces were treated by aminolysis (A), followed by physisorption of Fn (B), or by covalent immobilization of Fn (C), Fgn (D), and fibrin (E). Original magnification: 40x.
95
Fig 4.7 : Alpha-actin staining of SMC on modified P(70/30), 8h and 48h after seeding. Surfaces were treated by aminolysis on pre-wet films (A), followed by adsorption of Fn on pre-wet films (B), and covalent immobilization of Fn on dry (C) or pre-wet (D) films. Original magnification: 10x. Where higher magnification shows interesting features, inserts at 60x magnification have been included.
8h 48h
A)
B)
C)
D)
8h 48h
A)
B)
C)
D)
8h 48h
A)
B)
C)
D)
96
Staining of the SMC-specific cytoskeleton component α-actin further supported the
morphological observations made in the SEM. In addition, positive α-actin immuno-staining
confirmed SMC phenotype. Cells grown on aminated surfaces were irregular and retracted.
On modified P(100/0) surfaces (Fig. 4.6), a well-defined, rather homogeneously distributed
cytoskeleton was observed in all cells following 8h of adhesion. The adhered cells exhibited
rounded, dendritic cell shapes, characteristic of the synthetic phenotype. Cells were clearly
elongated on all protein-modified surfaces after 48h, although cells growing on physisorbed
fibronectin (Fig. 4.6B) seemed to rearrange in a more pronounced aligned fashion than cells
on covalently immobilized proteins.
SMC cytoskeleton staining on P(70/30) films modified with fibronectin again revealed
large differences in morphology as compared to that of cells on P(100/0). Partial and extensive
spreading was observed already after 8h on pre-wet aminated or Fn-modified films (Fig. 4.7).
With the exception of dry-treated P(70/30), cells were more bipolar and slightly more
elongated following 48h incubation, although the populations were heterogeneous in this
regard. Strong α-actin staining appeared around the nucleus after 8h on prewet, aminolyzed
P(70/30) with or without subsequent Fn physisorption, but this feature disappeared following
48h in culture (7A, B). Films that had been aminolyzed from the dry state and subsequently
Fn-modified (Fig. 7C) did not support cell spreading, which is most probably a result of
unsuccessful fibronectin immobilization as discussed above, but this requires further
investigation.
It should be noted that all cell cultures were carried out in medium containing fetal bovine
serum (FBS). It was deemed necessary to perform cultures with FBS to promote proliferation
and avoid cell death within the six days of culture. Although this implies that all surfaces may
have been modified by adsorbing proteins from the medium, which may have aided cell
adhesion, significant differences in cell numbers were still observed between unmodified,
aminolyzed, and protein-modified polymers.
4.5 Conclusions
In conclusion, we observed that immobilization of fibronectin, fibrinogen and fibrin using
imidoester chemistry greatly enhanced the ability of P(100/0) and P(70/30) to support SMC
adhesion and proliferation. Cell adhesion was equally improved on physisorbed fibronectin
97
and fibrin, although proliferation rates on these surfaces were hampered, demonstrating a clear
dependence on the degree of protein anchorage. Improved cell adhesion and proliferation were
obtained on all protein-modified surfaces. On modified P(100/0) covalently immobilized
fibronectin and fibrin supported SMC proliferation comparable to that on TCPS control
surfaces following six days in culture. Protein immobilization on P(70/30) was not as efficient
in improving cell-material interactions. Pre-wetting this co-polymer in PBS prior to
aminolysis and subsequent modification steps greatly improved cell numbers, in all
probability as a result of more efficient protein immobilization. However, cell numbers on
protein-modified surfaces were constantly and considerably lower on P(70/30) than on
P(100/0), and no significant cell proliferation was observed on any of the P(70/30)-based
surfaces. Further investigation will be required to establish why.
Morphological observations of cells grown on aminated copolymer films showed irregular
and retracted cell shapes. SMC grown on protein-modified P(100/0) revealed rectangular,
dendritic cell morphology 8h after seeding, while cells became bipolar and elongated after
48h. A less evident elongation with time of culture was observed on modified P(70/30),
although cells were comparatively more spread already at 8h.
4.6 Acknowledgements
This work was supported by the NSERC through a Discovery grant (PV) and a Strategic
grant (PV).
98
4.7 References
(1) Baguneid M, Murray D, Salacinski HJ, Fuller B, Hamilton G, Walker M, et al. Shear-
stress preconditioning and tissue-engineering-based paradigms for generating arterial substitutes. Biotechnol. Appl. Biochem. 2004;39:151-157.
(2) Pompe T, Kobe F, Salchert K, Jorgensen B, Oswald J, Werner C. Fibronectin anchorage to polymer substrates controls the initial phase of endothelial cell adhesion. J. Biomed. Mat. Res., Part A 2003;67A:647-657.
(3) Serrano MC, Pagani R, Manzano M, Comas JV, Portoles MT. Mitochondrial membrane potential and reactive oxygen species content of endothelial and smooth muscle cells cultured on poly(ε-caprolactone) films. Biomaterials 2006;27:4706-4714.
(4) Faucheux N, Schweiss R, Lutzow K, Werner C, Groth T. Self-assembled monolayers with different terminating groups as model substrates for cell adhesion studies. Biomaterials 2004;25:2721-2730.
(5) Dimilla PA, Stone JA, Quinn JA, Albelda SM, Lauffenburger DA. Maximal migration of human smooth-muscle cells on fibronectin and type-IV collagen occurs at an intermediate attachment strength. J. Cell Biol. 1993;122:729-737.
(6) Lieb E, Hacker M, Tessmar J, Kunz-Schughart LA, Fiedler J, Dahmen C, et al. Mediating specific cell adhesion to low-adhesive diblock copolymers by instant modification with cyclic RGD peptides. Biomaterials 2005;26:2333-2341.
(7) Santiago LY, Nowak RW, Rubin JP, Marra KG. Peptide-surface modification of poly(caprolactone) with laminin-derived sequences for adipose-derived stem cell applications. Biomaterials 2006;27:2962-2969.
(8) Costa KA, Sumpio BE, Cerreta JM. Increased elastin synthesis by cultured bovine aortic smooth-muscle cells subjected to repetitive mechanical stretching. FASEB J. 1991;5:A1609.
(9) Kulik TJ, Alvarado SP. Effect of stretch on growth and collagen-synthesis in cultured rat and lamb pulmonary arterial smooth-muscle cells. J.Cell. Physiol. 1993;157:615-624.
(10) Kim BS, Nikolovski J, Bonadio J, Mooney DJ. Cyclic mechanical strain regulates the development of engineered smooth muscle tissue. Nat. Biotechnol. 1999; 17:979-983.
(11) Riha GM, Wang XW, Wang H, Chai H, Mu H, Lin PH, et al. Cyclic strain induces vascular smooth muscle cell differentiation from murine embryonic mesenchymal progenitor cells. Surgery 2007;141:394-402.
99
(12) Goodman LV, Majack RA. Vascular smooth-muscle cells express distinct transforming growth factor-beta receptor phenotypes as a function of cell-density in culture. J. Biol. Chem. 1989;264:5241-5244.
(13) Mann B, Tsai A, Scott-Burden T, West J. Modification of surfaces with cell adhesion peptides alters extracellular matrix deposition. Biomaterials 1999;20:2281-2286.
(14) Mann BK, Schmedlen RH, West JL. Tethered-TGF-beta increases extracellular matrix production of vascular smooth muscle cells. Biomaterials 2001;22:439-444.
(15) Vogel V, Baneyx G. The tissue engineering puzzle: A molecular perspective. Annu. Rev. Biomed. Eng. 2003;5:441-463.
(16) Zhu YB, Gao CY, Liu XY, Shen JC. Surface modification of polycaprolactone membrane via aminolysis and biomacromolecule immobilization for promoting cytocompatibility of human endothelial cells. Biomacromolecules 2002;3:1312-1319.
(17) Chernukhin IV, Klenova EM. A method of immobilization on the solid support of complex and simple enzymes retaining their activity. Anal. Biochem. 2000;280:178-181.
(18) Moeschel K, Nouaimi M, Steinbrenner C, Bisswanger H. Immobilization of thermolysin to polyamide nonwoven materials. Biotechnol. Bioeng. 2003;82:190-199.
(19) Bramfeldt H, Sarazin P, Vermette P. Characterization, degradation, and mechanical strength of poly(D,L-lactide-co-ε-caprolactone)-poly(ethylene glycol)-poly(D,L-lactide-co-ε-caprolactone). J. Biomed. Mater. Res., Part A 2007;83:502-511
(20) Shearer H, Ellis MJ, Perera SP, Chaudhuri JB. Effects of common sterilization methods on the structure and properties of poly(D,L-lactic-co-glycolic acid) scaffolds. Tissue Eng. 2006;12:2717-2727.
(21) Tanzi MC, Verderio P, Lampugnani MG, Resnati M, Dejana E, Sturani E. Cytotoxicity of some catalysts commonly used in the synthesis of copolymers for biomedical use. J. Mat. Sci.: Mat. Med. 1994;5:393-396.
(22) Fukatsu K. Mechanical-Properties of Poly(Ethylene-Terephthalate) Fibers Imparted Hydrophilicity with Aminolysis. J. Appl. Pol. Sci. 1992;45:2037-2042.
(23) Clemmons DR, Horvitz G, Engleman W, Nichols T, Moralez A, Nickols GA. Synthetic alpha V beta 3 antagonists inhibit insulin-like growth factor-I-stimulated smooth muscle cell migration and replication. Endocrinology 1999;140:4616-4621.
(24) Zargham R, Thibault G. alpha 8 Integrin expression is required for maintenance of the smooth muscle cell differentiated phenotype. Cardiovasc. Res. 2006;71:170-178.
100
(25) Zargham R, Thibault G. alpha 8 integrin overexpression in the less differentiated vascular smooth muscle cells restores the differentiation properties and attenuates migratory activity. Arterioscler., Thromb., Vasc. Biol. 2006;26:E70.
(26) Moiseeva EP. Adhesion receptors of vascular smooth muscle cells and their functions. Cardiovasc. Res. 2001;52:372-386.
(27) Underwood S, Afoke A, Brown RA, MacLeod AJ, Dunnill P. The physical properties of a fibrillar fibronectin-fibrinogen material with potential use in tissue engineering. Bioprocess Eng. 1999;20:239-248.
(28) Harding SI, Afoke A, Brown RA, MacLeod A, Shamlou PA, Dunnill P. Engineering and cell attachment properties of human fibronectin-fibrinogen scaffolds for use in tissue engineered blood vessels. Bioprocess Biosyst. Eng. 2002;25:53-59.
(29) Halayko AJ, Solway J. Molecular mechanisms of phenotypic plasticity in smooth muscle cells. J. Appl. Physiol. 2001;90:358-368.
(30) Halayko AJ, Camoretti-Mercado B, Forsythe SM, Vieira JE, Mitchell RW, Wylam ME, et al. Divergent differentiation paths in airway smooth muscle culture: induction of functionally contractile myocytes. Am. J. Physiol.: Lung Cell. Mol. Physiol. 1999;276:L197-L206.
(31) Kocher O, Gabbiani F, Gabbiani G, Reidy MA, Cokay MS, Peters H, et al. Phenotypic features of smooth-muscle cells during the evolution of experimental carotid-artery intimal thickening - biochemical and morphological studies. Lab. Invest. 1991;65:459-470.
101
PREFACE TO CHAPTER 5. In this final chapter, the co-polymers that were
characterized and surface-modified in previous chapters are used to fabricate
three-dimensional, porous scaffolds with controllable pore size. The surface
modification technique employed in Chapter 4, is translated to the three-
dimensional structures and an initial smooth muscle cell adhesion test is carried
out.
102
CHAPTER 5 :
SMOOTH MUSCLE CELL ADHESION IN SURFACE-MODIFIED THREE-DIMENSIONAL
CO-POLYMER SCAFFOLDS PREPARED FROM CO-CONTINUOUS BLENDS
5.1 Abstract
In this article, we report on the applicability of the co-continuous blend technique for
the fabrication of tissue-engineering scaffolds from co-polymers, using polystyrene as the
porogen phase. The co-polymers consisted of P(ε-CL-co-D,L-LA)-PEG-P(ε-CL-co-D,L-LA),
referred to as P(100/0) and P(70/30), respectively, reflecting the %monomer content (CL/LA)
in the side chains. By means of static annealing of the blends for 30-120min followed by
extraction of the porogen phase, pore sizes roughly in the range of 15-350µm could be
obtained. The efficiency of in vitro seeding of smooth muscle cells was investigated in three-
dimensional, fibronectin-modified P(100/0) scaffolds of two different pore sizes. Considerably
higher numbers of adhering cells were found in the scaffolds with larger pore sizes, indicating
the importance of this parameter to promote effective cell intrusion into bulk materials.
Furthermore, the influence of scaffold pore size on compressive strength, permeability was
investigated. Compressive moduli ranged from 2.3±0.3 to 67±15 MPa and decreased with
increasing pore size. The reverse trend was found for scaffold permeability, which generated
values of κ from 8.5 × 10-16 to 6.7 × 10-11 m2. This level of permeability was comparable to
values reported for other scaffolds with higher pore sizes and void volumes, but with less open
pore morphologies. Taken together, the results obtained in this study strongly motivate further
investigation for possible future application in tissue engineering.
5.2 Introduction
The fabrication of porous polymer scaffolds has attracted much attention in the past 15
years. For tissue engineering, the importance of matching the properties of the scaffold to that
of the tissue being replaced has been repeatedly underlined by many research teams. These
properties include mechanical characteristics, permeability, pore size and porosity, and
resemblance of the cell-scaffold interface to the extracellular matrix surrounding cells in
natural tissues[1-3]. It has also been proposed that the rate of degradation in vitro and/or in
vivo should match an appropriate rate of tissue growth as it proceeds within the scaffolds[2].
103
Numerous techniques have been described in the literature for the fabrication of porous
polymer structures for tissue engineering, including the solvent casting/particulate leaching
method[4-6], gas foaming[7], emulsion freeze drying[8], phase separation[9], and solid free-
form fabrication[10]. For further details on these and other techniques, the reader is referred to
the numerous reviews on the subject[11-14]. The majority of scaffolds used for the
engineering of smooth muscle-based tissues are either fabricated from non-woven
mesh[15,16] or particulate leaching[5,17] methods using members of the poly(lactic acid)
(PLA)/ poly(glycolic acid) (PGA) family. Methods combining particulate leaching with free-
form fabrication[18] and gas foaming[19] have also been reported. The drawbacks of particle
leaching methods are the use of organic solvents, and the commonly occurring heterogeneous
pore morphology (principal pores connected by smaller fenestrations), which may complicate
cell invasion. However, results indicate that synthetic biomaterials are suitable for the
engineering of smooth muscle-based tissues, as smooth muscle cells (SMC) generally show
higher proliferative potential and elastin production when seeded in synthetic scaffolds
compared to collagen-based constructs[15,19]. Furthermore, to enable effective SMC
invasion, pore sizes of at least 100µm have been recommended[18].
The use of immiscible binary polymer blends for potential applications in tissue
engineering has attracted increasing attention lately[20-23]. In immiscible blends, the minor
phase forms dispersed droplets in the main matrix. However, upon increasing the
concentration of the minor component, the dispersed phase will grow and coalesce until it
becomes the major phase. Close to this phase inversion point, both phases become fully
interconnected, and the blend morphology can be described as co-continuous. This technique
not only offers control over parameters such as porosity and pore size[20,23], but due to the
continuous nature of the blend phases, it also ensures an entirely interconnected porosity once
the porogen phase has been leached out. However, since the fully co-continuous region does
not allow for very large variations in blend composition, the porosity generated is generally
not higher than 60-70%[21,23]. Pore size has been effectively altered for several blend pairs
by static annealing of the blends at temperatures above the Tm of both components[24].
Among the blends that have been studied, and which may be relevant to tissue engineering,
are P(L-LA)/polystyrene (PS), poly(ε-caprolactone) (PCL) /poly(ethylene glycol), PLLA/PCL,
P(D,L-LA)/PCL, and P(L-LA)/PGA[20-23,25]. Pore size and porosity may also influence such
properties as permeability and mechanical strength, which may thus be tuned to meet the
104
demands of a specific scaffold application. These qualities imply the suitability of these
scaffolds towards the study of cell interactions with three-dimensional porous structure. This
was recently illustrated by Washburn and co-workers[26], who showed the feasibility of using
such structures for regeneration of bone tissue.
It is generally accepted that surface modification methods are key elements in implant or
tissue engineering to improve and control the interaction of cells with synthetic polymer
materials[29-33]. However, to our knowledge, while many detailed studies of two-
dimensional models can be found in the literature, very few results of cell culture in truly
three-dimensional, covalently surface-modified, synthetic tissue-engineering scaffolds have
been reported. Schemes involving coating techniques based on physisorption of proteins from
aqueous solution may improve cell adhesion and proliferation within porous scaffolds, but it is
not clear whether such coatings would resist long-term culture or withstand mechanically
demanding environments, such as e.g. the shear stresses occurring within a perfusion
bioreactor.
In a recent publication we characterized three flexible (ε-CL-co-D,L-LA)-PEG-P(ε-CL-co-
D,L-LA) co-polymers with varying ε-CL: D,L-LA ratios in the side chains[27]. Based on the
obtained results, co-polymers containing 100 and 70 monomer% ε-CL in the side chains, were
judged to possess mechanical properties and degradation profiles suitable for scaffold
fabrication. These materials, referred to here as P(100/0) and P(70/30), respectively, were
subsequently evaluated for protein-modification and two-dimensional smooth muscle cell
culture[28]. In this work, we attempted to prepare porous scaffolds from these two co-
polymers using immiscible binary blends with polystyrene as the porogen phase. Following
selective extraction of the polystyrene phase and verification of the co-continuous nature of
the resulting porosity, the efficiency of annealing as a tool for increasing and controlling pore
size of the scaffolds was studied. In addition, the changes in compressive stiffness and
permeability with varying pore size were followed. Finally, three-dimensional scaffolds were
covalently modified with fibronectin and seeded with SMC using a dynamic cell seeding
method, to examine whether these structures may be suitable for the engineering of smooth-
muscle based tissues.
105
5.3 Materials and Methods
5.3.1 Materials
PEG diols (MW 7500, cat. no. 06103) and D,L-LA monomers (cat. no. 16640) were
purchased from Polysciences Inc (Warrington, PA, USA). Stannous octoate (SnOct, cat. no.
3252) and ε-caprolactone (ε-CL, cat. no. 21510) were obtained from Sigma-Aldrich (Oakville,
ON, Canada), while polystyrene (PS, MW 192 000) was obtained from Dow (Midland, MI,
USA). For surface modification, fibronectin (cat. no. F4759), 1,6-hexanediamine (cat.
no.H11696), and dimethyl pimelimidate (DMP, cat. no. D8388) were all purchased from
Sigma-Aldrich.
5.3.2 Co-polymers
Two copolymers were synthesized as described earlier[27] using ring-opening
polymerization, stannous octoate as a catalyst, and a PEG diol as macro-initiator. By varying
the percent composition of the monomer mix, the side chains contained either 100% ε-CL, or
70% ε-CL and 30% D,L-LA in random distribution. The resulting two copolymers were thus a
pure block copolymer, PCL-PEG-PCL, MW 95 kDa, and a poly(D,L-lactide-co-ε-
caprolactone)-poly(ethylene glycol)-poly(D,L-lactide-co-ε-caprolactone), MW 81 kDa,
referred to here as P(100/0) and P(70/30), respectively.
5.3.3 Preparation of blends
Binary melt blends were prepared by melt mixing in a Brabender internal mixer (Plasti-
corder Digi-System, C.W. Brabender Instruments Inc., NJ 07606). Vacuum-dried polystyrene
and co-polymer were introduced to the mixing chamber at a 1:1 weight ratio. The temperature
was set to 160°C for both blends. However, to enable blending of P(70/30)/PS, the
temperature was increased to reach an effective temperature of 165°C, which was maintained
for five minutes before returning to the initial level. The time of blending was 15 minutes for
all blends, and the mixing speed was kept constant at 50rpm. The torque remained nearly
stable, indicating that no significant degradation took place in either blend.
5.3.4 Annealing and scaffold fabrication
For this study, polystyrene (PS) was used as the porogen phase. P(100/0)/PS and
P(70/30)/PS blends were rapped in Teflon sheets and aluminium foil and placed between
metal slabs (without applied pressure) in an oven and annealed at 200˚C and 180˚C,
106
respectively. Additionally, a set of P(100/0)/PS blends was annealed at 160˚C to verify
whether the lower annealing temperature would produce smaller pore sizes. The time of
annealing for all blends was 0, 30, 60, or 120 minutes, after which time they were quenched in
liquid nitrogen. The annealed blends were machined into cylinders (Ø=4mm) and further cut
into pieces, approximately 5mm in length. The PS phase was then extracted by selective
dissolution in cyclohexane. The extractions were performed initially in Falcon tubes and then
continued in glass vials at room temperature until the weight of the dried samples was
constant. The resulting porous, cylindrical samples were used for the experiments described in
Sections 3.6-3.8.
As a direct consequence of the co-continuity of the blends and the extraction method
employed to remove the porogen phase, the resulting porosity can be considered to be fully
interconnected. Approximate values for the porosity of the scaffolds were calculated based on
the mass change due to PS extraction:
%100% ×−
=initial
finalinitial
m
mmporosity (5.1)
where minitial and mfinal are the mass of the specimen before and after PS extraction,
respectively.
5.3.5 SEM imaging
For SEM imaging, annealed samples were cut to appropriate size using a sharp razor blade
and then microtomed to achieve a smooth, flat surface. The microtome cuts were performed
with a Leica RM 2155 in a liquid nitrogen cell (LN21) using glass knives. Samples were
sputter coated in Au/Pd and SEM imaging was carried out in a JSM-840-A JEOL (V=5kV,
unless otherwise indicated). Pore diameters were measured by image analysis using Sigma
Scan Pro software. The surface regions of cell-seeded porous scaffolds were imaged in the
same fashion.
5.3.6 Mechanical properties
Cylindrical samples of porous P(100/0) and P(70/30) were prepared as detailed above and
tested in the wet state at room temperature, following 48h immersion in distilled water. The
compression strength of the porous samples was measured using an Instron apparatus. The
107
loading rate was set to 1mm/min using a 1000N load cell. For each condition, 3-5 samples
were tested. Results are presented as mean values ± standard deviation.
5.3.7 Permeability of scaffolds
The permeability of P(100/0) and P(70/30) scaffolds resulting from the annealing of the
blends at 200˚C and 180˚C, respectively, was determined by measuring the flow of distilled
water through cylindrical samples. The samples were fixed in a flexible silicone tube (inner
diameter < 4mm) connected to a water reservoir. All samples were immersed in distilled water
24h prior to the experiment to limit the effects of swelling and water absorption, which may
disturb the flow profile. Once mounted in the experimental set up, water was allowed to flow
through the samples until a uniform flow was obtained. Finally, Darcy’s law was applied to
calculate the permeability21:
PA
Ql
∆=
µκ (5.2)
where Q is the volumetric flow rate (m3/s), l is the sample length (m), µ is the fluid viscosity
(1mPas for water), ∆P is the pressure drop across the specimen (Pa) and A is the cross-
sectional area (m2) of the specimen. Three samples per copolymer and annealing condition
were used in this experiment, and the flow through each sample measured at two different ∆P.
The data is presented as mean values of κ ± standard deviation.
5.3.8 Protein immobilization and smooth muscle cell seeding
AMINOLYSIS AND PROTEIN IMMOBILIZATION
Cylindrical, porous P(100/0) was used to evaluate surface modification and cell adhesion
using a simple apparatus constructed in-house to carry out chemical activation and cell
seeding of the produced scaffolds. The device (see Fig. 5.1) consisted of a 1mL syringe
(i.d.=5mm), in which a scaffold was fixed using two tight-fitting Teflon rings (o.d.=5mm,
i.d.=3mm). Two pistons were used in each syringe to enable and to control flow through the
scaffolds. For each rinsing step and change of solution, the new solution was added and
flushed several times through the fixed samples (from side to side in the syringes), aspirated to
remove traces of old solution, and fresh solution was then added a second time. In order to
facilitate cell penetration into the scaffold interior, a small hole (approximately 2mm deep)
108
was carved into one end of the cylindrical samples using a hypodermic needle. Each sample
was dried and weighed prior to modification and cell seeding.
Fig 5.1: Schematic drawing of cell-seeding device, consisting of a 1mL syringe (i.d.=5mm) and two Teflon rings (o.d.=5mm, i.d.=3mm). The scaffold is shown in gray between the two Teflon rings. The small cavity in the seeding end (large arrow) of the scaffold serves to facilitate the distribution of suspended cells into the interior of the scaffold. For clarity, the second piston has been left out.
The scaffolds were initially treated with 10mg/ml 1,6-hexanediamine in isopropanol for 2h
to introduce primary amines on the external as well as internal surfaces. The aminolyzed
samples were rinsed in large amounts of distilled, sterile-filtered water, and subsequently
incubated in 10mM dimethylpimelimidate (DMP) in PBS, pH 8, for 1h at room temperature.
Following a rinsing step in PBS, pH 8, a fibronectin solution (10µg/mL) was added and
thoroughly flushed through the samples to ensure thorough dispersion, and allowed to react
for 4h at 4˚C and 1h at room temperature. Finally, the scaffolds were sterilized in a 500U/ml
penicillin/streptomycin overnight.
Amine-reactive ninhydrin was used as a colorimetric indicator of the presence of amine
groups on exterior and interior surfaces of the scaffolds. Aminated scaffolds were cut into
slices using a sharp razor blade, and allowed to react for 1min in a 1M ninhydrin solution. The
slices were then placed in separate glass vials and heated to 70˚C in a water bath to accelerate
the reaction. The appearance and distribution of blue color on the samples were considered
qualitative proof of successful aminolysis.
CELL CULTURE AND ADHESION TEST
Human Umbilical Artery Smooth Muscle Cells (SMC) (Promocell, Heidelberg,
Geramany) were cultured in basic growth medium recommended by the manufacturer,
supplemented with basic fibroblast growth factor, epithelial growth factor, insulin, 5% fetal
Teflon rings Piston
Scaffold
109
bovine serum, and antibiotics (100U/mL penicillin and 100mg/mL streptomycin). Medium
was changed every two to three days, and cells were used for experiments in passage 6.
Using the same chamber, sterilized scaffolds were rinsed twice in PBS, immersed in
complete medium, and equilibrated for 30 minutes at 37˚C. Subconfluent SMC were detached
using trypsin/EDTA and seeded at 1.2×106 cells/mL by carefully moving the cell suspension
several times from end to end in the syringe. The syringes containing the scaffolds were
incubated for 5h in 37˚C, at which time the cell suspension was carefully replaced by fresh
medium.
The number of adhering cells was verified 7h after seeding using a haemacytometer. Prior
to cell detachment, the scaffolds were transferred to new syringes and rinsed twice in PBS.
Cells were subsequently detached with trypsin/EDTA and vigorously flushed to extract all
adherent cells. The total cell number was determined from the medium value of eight counted
fields, and three scaffolds of each pore size were analyzed.
In a separate set of identical scaffolds, adhering cells were visualized by SEM imaging.
Cells were fixed in 1.4% glutaraldehyde for 30 minutes and in 2.5% glutaraldehyde at 4˚C
overnight. Following post-fixing in OsO4, the cell-scaffold constructs were dried in a vacuum
oven and finally sputter-coated in Au/Pd prior to imaging.
110
5.4 Results and Discussion
5.4.1 Effects of annealing on pore size
In this study, the use of a co-continuous polymer blend technique was investigated as the
first step in the fabrication of macroporous scaffolds from two co-polymers. The resulting
materials can be seen in Figure 5.2, showing SEM micrographs of porous P(100/0), P(70/30),
and PS (with the P(100/0) phase extracted). These micrographs show interconnected pores,
regardless of which phase is extracted, signifying that these blends are immiscible.
Fig 5.2: SEM images confirming continuous porosity in non-annealed samples of P(100/0) (A) and P(70/30) (B) with the polystyrene phase extracted. In (C) is shown porous polystyrene after extraction of P(100/0) using acetic acid. The white scale bars indicate 10 µm.
Following the verification of the applicability of the melt blend technique, the effect of
static annealing on the pore sizes of the two co-polymer/PS blends was studied to determine
whether pore size could be efficiently controlled through this method. Since in these
structures, the pores may be better described as channels, the mention of “pore size” in the
discussion below will correspond to the diameter of these channels. During annealing,
coalescence of each of the phases results in larger pore sizes following the subsequent
dissolution of the porogen phase. This change in morphology during annealing is influenced
by parameters such as temperature, time, interfacial tension, and viscosity of the
components[23,24]. Yuan and Favis proposed that the driving force for this phase coarsening
during static annealing is driven by capillary instability[24]. Considering the interconnected
networks of each of the components to be composed of thin and thick, cylindrical rods, they
further described this process as one where the thinner rods merges/flows into thicker ones,
eventually resulting in the retraction or break-up of the thin rods. The coarsening of one
phase due to coalescence reduces the interfacial area, and hence the free energy of the
system[24].
111
SEM micrographs of the P(100/0) scaffolds annealed for 30-120min at 160˚C and 200˚C
can be seen in Figure 5.3. The resulting pore sizes as determined by image analysis can be
found in Table 5.1. As expected, the higher temperature induced a more rapid coalescence of
each of the two phases, resulting in important pore growth over two hours. The pore size
found in non-annealed porous P(100/0) was 1.3±0.4µm, and this increased to reach
354±202µm after two hours of annealing at 200˚C. Lowering the annealing temperature to
160˚C, which is just above the critical flow temperature of PS, resulted in a mean pore
diameter of 17±7µm after two hours annealing. The pore size of non-annealed porous
P(70/30) was determined to be 3.2±1.2µm. Upon annealing for up to 120min at 180˚C, mean
pore sizes grew as shown in Fig. 5.4. Due to the method employed, only a medium value of
the pore size is obtained. However, it has been previously shown that porous structures
obtained from co-continuous blends have unimodal pore-size distribution[34,35]. Altogether,
these results show that by varying the time and temperature of annealing, efficient control of
pore size was accomplished in both copolymer/PS systems studied.
Table 5.1: Pore size and approximate porosity as a function of annealing time and temperature.
Pore diameters, obtained via SEM image analysis, are expressed as the mean value ± standard
deviation. The number of measurements made is indicated in the N0 column for each condition.
Co-polymer
Temperature (˚C)
P(70/30)
180
P(100/0)
200
P(100/0)
160
Annealing
time (min)
Pore
diameter
(µm)
No Porosity
(%)
Pore
diameter
(µm)
No Porosity
(%)
Pore
diameter
(µm)
No
0 3.2±1.2 110 49 1.3±0.4 201 48 - -
30 23±11 200 53 85±36 113 54 7±3 201
60 37±18 135 54 137±51 112 54 12±5 202
120 47±22 101 59 354±202 36 52 17±7 204
112
A B C
D E F
A B CA B C
D E FD E F
Fig 5.3: Porous P(100/0): effect of annealing on pore size. The samples were annealed at 160˚C (A-C) or at 200˚C (D-F) for 30min (A, D), 60min (B, E), or 120min (C, F). The white scale bars indicate 100µm.
Fig 5.4: Porous P(70/30): effect of annealing on pore size. The samples were annealed at 180˚C for 30min (A), 60min (B), 120min (C). The white scale bars indicate 100µm.
It may be noted that compared to several other scaffold fabrication methods, co-
continuous blend technology leads to a porosity with more curved pore walls and unimodal
size distribution[34,35]. The pore morphology of the resulting scaffolds is thus qualitatively
different, and may potentially play a role in the performance of the respective scaffolds, e.g.
with respect to cell-material interactions.
113
5.4.2 Compressive strength
Figure 5.5 shows representative compression curves of porous P(100/0) and P(70/30) as a
function of annealing time (annealing temperatures were 200˚C and 180˚C, respectively). The
compressive Young’s moduli were determined from the linear region and are presented in
Figure 5.6 together with the compressive strength at 10% strain. Both graphs reveal similar
trends: the compressive strength of both copolymers decreased significantly as a result of
annealing. The Young’s modulus of P(100/0) decreased from 69±15 MPa for non-annealed
samples to 28 ± 8 MPa following 30min annealing. For the corresponding interval in P(70/30)
scaffolds, the modulus dropped from 12±3 to 5.9±1.0 MPa. No significant change in
compressive modulus was observed between 30 and 120min annealing in P(100/0) scaffolds,
whereas P(70/30) showed a decrease in modulus from 5.9±1.0 to 2.3±0.3 MPa.
The observed decrease in compressive strength with increasing annealing time/channel
diameters is somewhat confusing. In contrast to most other methods used in the fabrication of
porous scaffolds, the “wall thickness” increases with increasing pore size as a result of the
coarsening of the two phases during annealing. Normally, such a change in morphology would
cause an increase in mechanical strength22. However, there are several other factors that may
come into play and explain our results. For example, following their preparation in the internal
mixer the polymer blends were allowed to slowly cool down to room temperature, while all
annealed samples were quenched in liquid N2 at the end of the treatment. It is therefore quite
possible that the annealed samples were less crystalline than their non-annealed counterparts,
and therefore exhibited lower compressive strength. This would also explain the small
differences in compressive modulus observed for the varying times of annealing (t=30, 60,
and 120min). In addition, increasing annealing time caused a certain increase in the porosity
of the scaffolds (see Table 1), which is consistent with decreasing mechanical strength21.
Furthermore, any degradation of the co-polymer during annealing would similarly have
resulted in a decrease in mechanical resistance.
It was also noted that P(70/30) samples showed a great capacity to partially regain their
initial shape following 60% compression. Furthermore, this capacity increased with increasing
time of annealing. Non-annealed samples were found to recover to approximately 80% of
their initial length, whereas samples that had been annealed for 120min showed up to 95%
recovery. P(100/0) samples showed a similar trend with an approximate recovery of 50% and
70% of the initial length for non-annealed and annealed samples, respectively.
114
Strain
0,0 0,2 0,4 0,6 0,8 1,0
Str
ess
(MP
a)
0
5
10
15
20
25
30
P(100/0) 0minP(100/0) 120minP(70/30) 0minP(70/30) 120min
Fig 5.5: Representative stress-strain curves for porous P(100/0) and P(70/30). One curve is shown for each co-polymer for annealing times of 0 and 120 minutes.
Annealing time (min)
0 30 60 90 120
Str
ess
at
10
% s
tra
in (
MP
a)
0
1
2
3
4
5
6
P(100/0)P(70/30)
Annealing time (min)
0 30 60 90 120
Com
pre
ssiv
e m
od
ulu
s (M
Pa)
0
20
40
60
80
100
P(100/0)P(70/30)
A) B)
Fig 5.6: Compressive Young’s moduli (A) and stress at 10% strain (B) of porous P(100/0) and P(70/30) for different times of annealing.
5.4.3 Permeability of scaffolds
The intrinsic permeability of a porous material is a measurement of the ease with which a
fluid flows through it. In most tissue engineering applications, scaffolds should be designed to
achieve the most efficient mass transport properties[36]. This property can be determined
using Darcy’s law[1,37,38], which states that the flow rate through a sample is proportional to
115
the applied pressure. In an earlier study, Hui et al.[39] found a threshold level of fluid
conductance below which bone xenografts in rabbit models failed to connect to the host tissue.
(Fluid conductance is the ratio of induced flow to pressure drop across the specimen). In fact,
the permeability is an indicative measure of several parameters, including porosity, pore size,
interconnectivity, and orientation of pores1, and may thus offer a simple means to obtain
relevant information related to specific scaffolds. For this study, the permeability of porous
P(100/0) and P(70/30) was measured for different pore sizes. The values of the permeability
constant, κ, as a function of annealing time (i.e. pore size) are presented in Figure 5.7.
Annealing time (min)
0 30 60 90 120
log (
κκ κκ)
-18
-16
-14
-12
-10
-8
P(100/0)P(70/30)
Fig 5.7: A log-linear plot of the permeability constant κ (m2) vs. annealing time for porous P(100/0) and P(70/30).
From Figure 5.7, it may be seen that the intrinsic permeability of the specimens tested
spanned seven orders of magnitude, ranging from 8.5 × 10-16 to 6.7 × 10-11 m2. P(100/0)
scaffolds were consistently more permeable than P(70/30) for the corresponding time of
annealing, which may be expected in view of the larger pore sizes of the former. Increasing
the annealing time from 30 to 120min, induced an 18-fold increase in P(70/30) scaffolds,
whereas the same interval caused a 70-fold increase in P(100/0) samples. For comparison, the
permeability of P(100/0) samples, annealed 120min, falls within the range of values (2.7 × 10-
11 to 2.0 × 10-8 m2) reported in the literature for human vertebral and femoral trabecular
bone[40].
In an attempt to identify key scaffold properties useful in the prediction of scaffold
performance in vivo, permeability alone and the permeability/porosity ratio has been
116
suggested to be a more suitable property than pore size distribution or pore size alone[1]. We
made observations in support of this proposal, while comparing permeability values obtained
for porous structures fabricated with different techniques. For example, Shimko and
Nauman[37] reported the permeability of poly(methyl methacrylate) scaffolds fabricated using
the solvent casting/particle leaching method. The constructs were prepared with salt porogens
in the size range of 200-700µm and a final porosity of 70%. And although both pore size and
porosity were significantly higher for all but the P(100/0) 120min samples, the permeability
range obtained (6.6 × 10-16 to 1.4 × 10-10 m2) was comparable to that found for the scaffolds
tested here. Conversely, calcium phosphate scaffolds for bone reconstruction that exhibited
pore morphologies very similar to ours (pore sizes 200-700µm, porosity >70%), showed
permeability on the order of 10-10 m2.[1] Altogether, these results may serve to illustrate the
importance of the high degree of interconnectivity achieved by binary blends, resulting in a
microstructure that is more conducive to fluid flow.
5.4.4 Smooth muscle cell adhesion on protein-modified scaffolds
Covalent immobilization of fibronectin (Fn) was shown in a previous study to drastically
improve smooth muscle cell (SMC) adhesion and proliferation on P(100/0) films[28].
Compared to unmodified films and films exhibiting amine groups only, covalently
immobilized Fn improved cell adhesion 3.5 and 1.9 times, respectively. Fn-modified surfaces
were furthermore able to support 5-fold increase in cell numbers over six days in culture,
whereas no proliferation was observed on the two control surfaces. Adsorbed Fn was
similarly found to enhance cell adhesion on this material, but did not to the same extent
support proliferation (2-fold increase over six days)[28].
In this study, the same surface modification scheme was applied to promote cell
attachment in 3D matrices. Utilizing a custom-made reaction chamber (see Fig. 5.1), designed
to facilitate homogeneous surface modification and subsequent cell seeding, cylindrical
P(100/0) scaffolds were aminated through a reaction with 1,6-hexanediamine. Ninhydrin was
used as a colorimetric indicator of the presence of amine groups on the scaffold exterior and
interior surfaces. The successful and thorough aminolysis of the samples was confirmed
visually by the homogeneous, blue staining of several cross-sections of the two scaffold types
(Fig. 5.8). Following aminolysis, the scaffold surfaces were activated with dimethyl
pimelimidate (an amino-reactive cross-linker) to couple Fn covalently to the modified
surfaces. In addition, scaffolds of two different pore sizes (30 vs. 60min, annealing
117
temperature = 200˚C) were used to observe the effect of pore size on the efficiency of cell
seeding.
Fig 5.8: (A) Aminated P(100/0) scaffold cross-sections stained blue by ninhydrin reactant. (B) Un-modified scaffolds were not stained by the ninhydrin treatment. Each scaffold was cut with a razor blade and numbered as indicated in (C).
SEM images of adhering cells bridging a pore gap and spreading out on a pore wall close
to the scaffold surface can be seen in Figure 5.9. Figure 5.10 shows the number of adhering
cells in the two types of scaffold 7h after seeding. Specifically, increasing the pore size from
85 to 137µm improved cell numbers from approximately 400 to 1500 cells/mg of scaffold.
Given the similar geometry of the samples, these results indicate a more efficient penetration
into the bulk of the scaffolds exhibiting larger pores. Seeing as no difference in porosity (void
volume) was observed for these two preparations (Table 5.1), the increase in cell number was
a direct effect of enhanced diffusion caused by larger pore sizes. This effect becomes ever
clearer when considering the smaller surface area available for cell adhesion in samples of
larger pore sizes. The obtained cell seeding efficiency was 43±4% and 17±10% for the
“60min” and “30min” samples, respectively, whereas it was a mere 4±2% for aminated
0123
C)
0123
C)
118
“30min” samples. Keeping in mind that rather large quantities of cells were visible in the
space outside the scaffolds in the reaction chambers after 7h incubation, these results can be
considered as good. It is expected that by automating the chamber setup so as to achieve flow
of the cell-seeding suspension across the scaffolds throughout the adhesion period, seeding
efficiency should be significantly improved.
Fig 5.9: Scanning electron micrographs of SMC adhering in a P(100/0) scaffold, 8h after cell seeding.
Cells spanning a pore void (A) and cells spread out on the Fn-modified copolymer surface (B) are
indicated by white arrows.
Cel
l num
ber/
mg
scaf
fold
0
500
1000
1500
2000
P(100/0) 60minP(100/0) 30minP(100/0) 30min, no Fn
Fig 5.10: Number of SMC adhering on surface-modified porous P(100/0) scaffolds 7h after cell
seeding. The scaffolds were previously annealed for 30min or 1h (n=3).
Our results show that biochemical surface modification within three-dimensional is
possible using the methods described. Furthermore, as reported elsewhere[18], pore size was
found an important parameter that influences the efficiency of dynamic in vitro cell seeding
119
techniques. The data presented here are certainly in support of these findings, although further
investigations will be needed to more accurately establish optimal pore diameters for cell
seeding and culture.
5.5 Conclusions
Scaffolds with entirely interconnected porosity were prepared from binary co-
polymer/polystyrene blends and investigated for their applicability in tissue engineering. The
original mean pore sizes of P(100/0) and P(70/30) were 1.3 and 3.2µm, respectively. Pore
sizes could be efficiently controlled by means of static annealing for 30-120min, resulting in
pore sizes roughly in the range of 15-350µm. The efficiency of in vitro seeding of smooth
muscle cells was investigated in fibronectin-modified P(100/0) scaffolds of two different pore
sizes. Considerably higher numbers of adhering cells were found in the scaffolds with larger
pore sizes, indicating the importance of this parameter to promote effective cell intrusion into
bulk materials. The number of adhering cells was also indicative of successful protein
immobilization.
Furthermore, the influence of scaffold pore size on compressive strength, permeability
was investigated. Compressive moduli ranged from 2.3±0.3 to 67±15 MPa and decreased with
increasing pore size. The reverse trend was found for scaffold permeability, which generated
values of κ from 8.5 × 10-16 to 6.7 × 10-11 m2. Importantly, in comparison with other scaffolds
that were prepared by particulate leaching and showing larger pore size and higher porosity,
the scaffolds studies here showed similar permeability, clearly indicating the pivotal role of
pore morphology.
Especially, with regard to the permeability and successful cell adhesion tests, we
propose that these scaffolds may be well suited for future studies of three-dimensional tissue
culture in vitro.
5.6 Acknowledgements
This work was supported by the NSERC through a Discovery grant (PV) and a
Strategic grant (PV) and a post-doctorate fellowship (PS). Appreciation is extended to Dr.
Martin Bureau and Manon Plourde, at Industial Materials Institute (NRC), Boucherville, QC,
for the use of and assistance with testing of mechanical properties.
120
5.7 References
(1) Li SH, de Wijn JR, Li JP, Layrolle P, de Groot K. Macroporous biphasic calcium phosphate scaffold with high permeability/porosity ratio. Tissue Eng. 2003; 9:535-548.
(2) Maquet V, Jerome R. Design of macroporous biodegradable polymer scaffolds for cell transplantation. Mater. Sci. Forum 1997; 250:15-42.
(3) Griffith L. Emerging design principles in biomaterials and scaffolds for tissue engineering. Ann. N. Y. Acad. Sci. 2002; 961:83-95.
(4) Mikos AG, Thorsen AJ, Czerwonka LA, Bao Y, Langer R, Winslow DN, Vacanti JP. Preparation and characterization of poly(L-lactic acid) foams. Polymer 1994; 35:1068-1077.
(5) McGlohorn JB, Holder WD, Grimes LW, Thomas CB, Burg KJL. Evaluation of smooth muscle cell response using two types of porous polylactide scaffolds with differing pore topography. Tissue Eng. 2004; 10:505-514.
(6) Draghi L, Resta S, Pirozzolo MG, Tanzi MC. Microspheres leaching for scaffold porosity control. J. Mater. Sci.: Mater. Med. 2005; 16:1093-1097.
(7) Harris LD, Kim BS, Mooney DJ. Open pore biodegradable matrices formed with gas foaming. J. Biomed. Mater. Res. 1998; 42:396-402.
(8) Whang K, Thomas CH, Healy KE, Nuber G. A Novel Method to Fabricate Bioabsorbable Scaffolds. Polymer 1995; 36:837-842.
(9) Mooney DJ, Baldwin DF, Suh NP, Vacanti LP, Langer R. Novel approach to fabricate porous sponges of poly(D,L-lactic-co-glycolic acid) without the use of organic solvents. Biomaterials 1996; 17:1417-1422.
(10) Park A, Wu B, Griffith LG. Integration of surface modification and 3D fabrication techniques to prepare patterned poly(L-lactide) substrates allowing regionally selective cell adhesion. J. Biomater. Sci.: Polym. Ed. 1998; 9:89-110.
(11) Liu XH, Ma PX. Polymeric scaffolds for bone tissue engineering. Ann. Biomed. Eng. 2004; 32:477-486.
(12) Mikos AG, Temenoff JS. Formation of highly porous biodegradable scaffolds for tissue engineering. Electr. J. Biotechnol. 2000; 3:114-119.
(13) Sachlos E, Czernuszka J. Making tissue engineering scaffolds work. Review on the application of solid freeform fabrication technology to the production of tissue engineering scaffolds. Eur. Cells Mater. 2003; 5:29-40.
(14) Yang SF, Leong KF, Du ZH, Chua CK. The design of scaffolds for use in tissue engineering. Part 1. Traditional factors. Tissue Eng. 2001; 7:679-689.
121
(15) Kim BS, Nikolovski J, Bonadio J, Mooney DJ. Cyclic mechanical strain regulates the development of engineered smooth muscle tissue. Nat. Biotechnol. 1999; 17:979-983.
(16) Oberpenning F, Meng J, Yoo JJ, Atala A. De novo reconstitution of a functional mammalian urinary bladder by tissue engineering. Nat. Biotechnol. 1999; 17:149-155.
(17) Lee SH, Kim BS, Kim SH, Choi SW, Jeong SI, Kwon IK, Kang SW, Nikolovski J, Mooney DJ, Han YK, Kim YH. Elastic biodegradable poly(glycolide-co-caprolactone) scaffold for tissue engineering. J. Biomed. Mater. Res., Part A 2003; 66A:29-37.
(18) Zeltinger J, Sherwood JK, Graham DA, Mueller R, Griffith LG. Effect of pore size and void fraction on cellular adhesion, proliferation, and matrix deposition. Tissue Eng. 2001; 7:557-572.
(19) Kim BS, Nikolovski J, Bonadio J, Smiley E, Mooney DJ. Engineered smooth muscle tissues: Regulating cell phenotype with the scaffold. Exp. Cell Res. 1999; 251:318-328.
(20) Washburn NR, Simon CG, Tona A, Elgendy HM, Karim A, Amis EJ. Co-extrusion of biocompatible polymers for scaffolds with co-continuous morphology. J. Biomed. Mater. Res. 2002; 60:20-29.
(21) Reignier J, Huneault MA. Preparation of interconnected poly(ε-caprolactone) porous scaffolds by a combination of polymer and salt particulate leaching. Polymer 2006; 47:4703-4717.
(22) Sarazin P, Roy X, Favis BD. Controlled preparation and properties of porous poly(L-lactide) obtained from a co-continuous blend of two biodegradable polymers. Biomaterials 2004; 25:5965-5978.
(23) Sarazin P, Favis BD. Morphology control in co-continuous poly(L-lactide)/polystyrene blends: A route towards highly structured and interconnected porosity in poly(L-lactide) materials. Biomacromolecules 2003; 4:1669-1679.
(24) Yuan ZH, Favis BD. Coarsening of immiscible co-continuous blends during quiescent annealing. AIChE Journal 2005; 51:271-280.
(25) Na YH, He Y, Shuai X, Kikkawa Y, Doi Y, Inoue Y. Compatibilization effect of poly(ε-caprolactone)-b-poly(ethylene glycol) block copolymers and phase morphology analysis in immiscible poly(lactide)/poly(ε-caprolactone) blends. Biomacromolecules 2002; 3:1179-1186.
(26) Washburn NR, Weir M, Anderson P, Potter K. Bone formation in polymeric scaffolds evaluated by proton magnetic resonance microscopy and X-ray microtomography. J. Biomed. Mater. Res., Part A 2004; 69A:738-747.
(27) Bramfeldt H, Sarazin P, Vermette P. Characterization, degradation, and mechanical strength of poly(D,L-lactide-co-ε-caprolactone)-poly(ethylene glycol)-poly(D,L-lactide-co-ε-caprolactone). J. Biomed. Mater. Res., Part A 2007; 83A: 503-511.
122
(28) Bramfeldt H, Vermette P. Enhanced smooth muscle cell adhesion and proliferation on protein-modified caprolactone-based copolymers. J. Biomed. Mater. Res., Part A, accepted.
(29) Lebaron RG, Athanasiou KA. Extracellular matrix cell adhesion peptides: Functional applications in orthopedic materials. Tissue Eng. 2000; 6:85-103.
(30) Hu YH, Winn SR, Krajbich I, Hollinger JO. Porous polymer scaffolds surface-modified with arginine-glycine-aspartic acid enhance bone cell attachment and differentiation in vitro. J. Biomed. Mater. Res., Part A 2003; 64A:583-590.
(31) Pompe T, Kobe F, Salchert K, Jorgensen B, Oswald J, Werner C. Fibronectin anchorage to polymer substrates controls the initial phase of endothelial cell adhesion. J. Biomed. Mater. Res., Part A 2003; 67A:647-657.
(32) Serrano MC, Pagani R, Manzano M, Comas JV, Portoles MT. Mitochondrial membrane potential and reactive oxygen species content of endothelial and smooth muscle cells cultured on poly(ε-caprolactone) films. Biomaterials 2006; 27:4706-4714.
(33) Lieb E, Hacker M, Tessmar J, Kunz-Schughart LA, Fiedler J, Dahmen C, Hersel U, Kessler H, Schulz MB, Gopferich A. Mediating specific cell adhesion to low-adhesive diblock copolymers by instant modification with cyclic RGD peptides. Biomaterials 2005; 26:2333-2341.
(34) Sarazin P, Favis BD. Stability of the co-continuous morphology during melt mixing for poly(epsilon-caprolactone)/polystyrene blends. J. Polym. Sci.: Polym. Phys. 2007; 45:864-872.
(35) Yuan ZH, Favis BD. Macroporous poly(L-lactide) of controlled pore size derived from the annealing of co-continuous polystyrene/poly(L-lactide) blends. Biomaterials 2004; 25:2161-2170.
(36) Martin Y, Vermette P. Bioreactors for tissue mass culture: Design, characterization, and recent advances. Biomaterials 2005; 26:7481-7503.
(37) Shimko DA, Nauman EA. Development and characterization of a porous poly(methyl methacrylate) scaffold with controllable modulus and permeability. J. Biomed. Mater. Res.: Appl. Biomater. 2007; 80B:360-369.
(38) Grimm MJ, Williams JL. Measurements of permeability in human calcaneal trabecular bone. J. Biomechanics 1997; 30:743-745.
(39) Hui PW, Leung PC, Sher A. Fluid conductance of cancellous bone graft as a predictor for graft-host interface healing. J. Biomechanics 1996; 29:123-132.
(40) Nauman EA, Fong KE, Keaveny TM. Dependence of intertrabecular permeability on flow direction and anatomic site. Ann. Biomed. Eng. 1999; 27:517-524.
123
GENERAL CONCLUSION AND SUGGESTED FUTURE WORK
GENERAL CONCLUSION
This thesis project was designed to constitute the first part of a work whose final
objective is the creation of functional vascular beds within three-dimensional scaffold
constructs for tissue engineering. It is the author’s expectation, that the methods and results of
this work may constitute a helpful knowledge base and direct future work towards the ultimate
objective.
Briefly, this thesis has resulted in the synthesis of a series of biocompatible co-polymers,
whose material properties vary from highly crystalline, mechanically strong, and slow-
degrading to highly amorphous, malleable and fast-degrading, as specified in Chapter 2.
These co-polymers may be further manipulated with relative ease to obtain materials
possessing characteristics appropriate for a specific application. This can be achieved either
by changing the monomer composition and/or molecular weight of the copolymers, or, as
shown in Chapter 3, by simple blending strategies. In addition, by preparing binary blends
composed of one fast-degrading and one slow-degrading phase, porous materials may
develop as hydrolysis proceeds.
Two of the characterized co-polymers (P(100/0) and P(70/30)) were considered
interesting candidates for scaffold fabrication. In order to improve cell-material interactions
on these materials, an aminolysis technique that was previously applied to pure PCL films,
was here applied to the described co-polymers, thereby introducing reactive amine groups on
the surface. Following aminolysis, films were further modified by covalent protein
immobilization using dimethylpimelimidate. This method strongly enhanced material
interactions with smooth muscle cells (SMC); specifically adhesion and proliferation, while
preserving a smooth muscle cell phenotype, although very different results were obtained for
the two co-polymers. It was found necessary for the LA-conainting P(70/30) to be pre-wet
prior to aminolysis and protein immobilization in order to increase cell numbers on these
surfaces. However, cell numbers were still not comparable to the ones obtained for P(100/0).
As for the different proteins preparations of this study, results implied that covalently
immobilized fibronectin and fibrin layers were the most suited for medium- to long-term
culture.
124
The same two co-polymers were investigated for the applicability of immiscible binary
blends for the fabrication of porous scaffolds with highly interconnected porosity. Static
annealing was found to constitute an efficient means to increase pore sizes. Cylindrical
scaffolds were then tested for compressive strength and permeability, and were both affected
by pore size. Encouraging preliminary results of smooth muscle cell adhesion in cylindrical
scaffolds, suggested the successful translation of the technique used for two-dimensional
surface modification to this three-dimensional scaffold. The results for permeability and SMC
adhesion within these three-dimensional co-polymer scaffolds indicated the pivotal role of
sufficient pore size. Taken together, the porous scaffolds developed in this thesis were judged
as promising for further tissue-culture studies in a bioreactor environment.
SUGGESTED FUTURE WORK
Based on the results of the studies described in this thesis and the many question marks
and ideas that have appeared along the way, the following work is proposed for the further
advancement of the project:
• Carry out a detailed surface study to identify and solve the problems encountered
during attempts to immobilize proteins on P(70/30) to ameliorate cell adhesion and
proliferation on these films.
o The ninhydrin assay needs to be repeated on aminated pre-wet films.
o The ninhydrin assay may also be useful to quantify free NH2 goups on film
surfaces following covalent protein immobilization, both on P(100/0) and
P(70/30). This method may then be used to assess the success rate of protein
immobilization, which may be interesting to compare between materials, and
between dry and pre-wet P(70/30).
o An alternative route may be to detach immobilized Fn by enzymatic cleavage
(e.g. by trypsination) and quantify the protein content by a Lowry or total
protein assay. The digestion would of course reduce the polymers into peptides
and amino acids depending on the time of digestion. In the Lowry assay, the
Folin-Ciocalteu’s phenol reagent reacts with mainly tyrosine residues in the
protein, and results in an absorbance maximum at 750nm. Yet another
125
technique is the quartz-crystal microbalance with dissipation monitoring.
Although coating a crystal with co-polymer may be rather difficult, amine-
terminated monolayers can be generated on gold-coated crystals, on which the
subsequent surface-modification steps (DMP-activation and protein-
immobilization) can be carried out. The conformational state may then easily
be evaluated using specific antibodies.
o One hypothesis is that the P(70/30) films have PEG chains oriented towards the
surface to a higher extent than P(100/0). If that is the case the P(70/30) films
are less reactive than P(100/0). High resolution XPS may perhaps be useful on
thinner and flatter films, prepared e.g. by spin-coating. Solid state NMR may
also be used to determine molecular structure and intermolecular packing
patterns.
o The conformation of the immobilized proteins (adsorbed and covalently
coupled) may be investigated using AFM. Try a trial and error-approach using
different monoclonal antibodies towards Fn may turn the previously
unsuccessful attempt to visualize the protein using immunofluorescence
around.
• Further characterization of the phenotype of smooth muscle cells grown on protein-
modified surfaces. This may also be further developed to promote dedifferentiation/
differentiation by growth factor delivery. Of interest would be the stimulation of 1) a
migrating phenotype in order to populate the scaffold bulk with SMC, and 2) a
contractile phenotype, as found in native blood vessels, and other mature SMC-based
tissues. This study should first be carried out in copolymer films, and later on in 3D
scaffolds. Suggested growth factors are: TGF-alpha, which acts to increase SM alpha-
actin. (Riha et al., Surgery 2007; 141, p.394-402): A combination of PDGF-BB and
IL-1 beta, which together act to lower levels of the contractile markers alpha-actin,
myosin heavy chain, and calponin, thereby possibly stimulating a migrating and ECM-
producing phenotype. (Chen et al., Proceedings Natl Acad Sciences 2006; 103, p.
2665-2670). VEGF and bFGF stimulate migration in SMC (Couper et al. Circulation
Res 1997; 81, p.932-939; Wang and Keiser, Circulation Res 1998; 83, p.832-840).
126
• Characterization and promotion of ECM deposition by SMC cultured on prepared
surfaces and scaffolds. Suitable techniques are e.g. Western blots (protein) and
Northern blots (gene expression, RNA). Suitable ECM proteins are Carry out a
comparative study of immobilized adhesion peptides vs. whole proteins to identify
effects on cell adhesion, proliferation, and ECM production. Interesting growth factors
to stimulate protein production may be: TGF-beta (Collagen type I, fibronectin)
(Schmidt et al. Eur J Clin Invest 2006; 36, p. 473-482);
• Polystyrene was used as a model porogen in this work but must be replaced by a
biocompatible polymer to avoid inflammatory response to any residual PS trapped in
the 3D constructs after extraction.
• Maximize the void volume obtainable by the binary co-continuous blend technique.
This may be done by varying the volume fraction of the porogen phase. Also worth
looking into would be the parameters during melt mixing. Lee and Han (Polymer
1999; 40, p. 6277-6296) were also able to increase the range of co-continuity to 30-
70% of the porogen phase by working close to / below the critical flow temperature of
the two components.
• A study of P(100/0) and P(70/30) material properties (crystalline content, degradation,
mechanical strength) following annealing as described in Chapter 5. A degradation
study should also be carried out of 3D scaffolds, and parallel measurements of the
mechanical properties of these structures as a function of degradation time.
• Apply melt blending as a technique for preparation of the co-polymer blends described
in Chapter 3, in order to obtain co-continuous porosity as a result of differentiated
degradation. Investigate whether viable long-term cell culture may be accomplished on
such materials, and if so, if cells will invade the forming void space.
127
• Combine the co-polymer blend preparation of Chapter 3 with the melt-
mixing/annealing method described in Chapter 5, in order to fabricate scaffolds of
bimodal porosity.
• In vitro: develop culture protocols and optimize conditions for tissue culture in a
perfusion bioreactor. Specifically, the effects of shear stress and soluble growth factors
on SMC phenotype should be evaluated to establish in what way smooth muscle cells
(or other cell types) may be controlled in this environment.
• Investigate the applicability of techniques like high resolution magnetic resonance
imaginng and µ-CT (microcomputed tomography) to image the interior of the
scaffolds. The µ-CT technique has been used to measure bone thickness in
regenerating bone (see e.g. Sell et al. Calcified Tissues Int 2005; 76, p.355-364) and
has also been proposed as a non-invasive monitoring system for vascularization (ref:
www.medicalnewstoday.com/articles/20243.php). It may therefore be interesting to
follow the vascularization and tissue growth within the 3D scaffolds using this
technique.
• Additionally, it would be interesting to study the possibility of growth factor
immobilization, possibly in combination with whole adhesive proteins, on scaffold
surfaces to promote particular cellular behavior. Such strategies would, if effective,
help to decrease the quantitative requirements of growth factors, and thus keep costs
down.
• In vitro: use protein-modified scaffolds, such as described in Chapter 5, to pre-culture
smooth muscle cells for a time sufficient to allow thorough cellular invasion of the
scaffold interior as well as the deposition of ECM. Subsequently, the scaffolds may be
seeded with endothelial or a mix of EC and tissue-specific cells, and cultured in a
bioreactor. Objective: to achieve initial differentiation of EC into tubular networks,
possibly induced by growth factor delivery, followed by EC-controlled recruitment of
SMC for vessel stabilization.
128
• In vivo: subcutaneous implantation of surface-modified, P(100/0) and P(70/30)
sponges seeded with SMC alone or EC+SMC, to verify the biocompatibility of the
scaffolds and to find out whether vascular ingrowth may be promoted.
129
APPENDIX
A1. Protocol for synthesis of P(ε-CL-co-D,L-LA)-PEG-P(ε-CL-co-D,L-LA)
Co-polymer synthesis using simultaneous, ring-opening co-polymerization of ε-CL and D,L-
LA in the presence of PEG diols (MW=7500), catalyzed by stannous octoate (SnOct).
MATERIAL
ChemGlass two- or three-neck round bottom flask (RBF), max. 250mL, 24/40 neck.
Rubber septum
Stirring shaft and blade of appropriate size
Air tight glass stop cock (with Teflon fittings and o-ring) to fix the stirring shaft
Motor with adjustable rotation speed
Robust stand and clips
Oil bath with temperature control equipped with magnetic stirrer
Equipment for degassing and purging with either N2 or Ar
STEP 1: MONOMER PREPARATION
• PEG diols: dry under vacuum at 50°C for 48h.
• ε-CL monomers: use as received
• D,L-LA monomers: recrystallize from ethyl acetate
1. Dissolve monomers in ethyl acetate in a crystallizing dish. Heat under stirring
to approximately 70°C. Carefully add solvent drop by drop until dissolution. It
is important that the minimal amount of solvent necessary to dissolve the
monomer is used.
2. Take the crystallizing dish off the hot plate and allow the solution to cool
down, and some of the solvent to evaporate (might need to be partially
covered to slow down the rate of evaporation). D,L-LA crystals will start to
form on the solvent/air interface.
3. When the recrystallization is finished, filter the crystals and wash with ethyl
acetate. Dry the crystals in room temperature and finally under vacuum.
130
STEP 2: REACTION CHAMBER ASSEMBLY
1. Clean all glassware, metal, and Teflon parts, and dry under vacuum.
2. Assemble the stirrer components. The threaded glass stopper must be put in place
before attaching the stirrer blade. Insert the stirring shaft into the middle neck of the
RBF and screw the rod tight. The blade should be close to the bottom, without the
metal rod touching the bottom of the flask.
3. Close the remaining neck(s) with a rubber septum.
4. Firmly attach the top of the metal rod to the motor.
5. Place the heating/stirring plate with the oil bath on the base of the stand and mount
the motor while adjusting its height so that the RBF is in a good position in the oil
bath. Further adjustments must be made to assure that the rod is perfectly vertical, so
as not to be moving the RBF while rotating.
STEP 3: SYNTHESIS
For the synthesis of P(100/0), PEG diols and ε-CL were introduced into an RBF and melted
at 140°C. The molar ratio of ethylene glycol-to-monomer was 1:4.2. The temperature was
then lowered to 120°C and the melted mixture degassed and purged with N2. SnOct was
added at a monomer-to-catalyst ratio of 1000:1, and the polymerization was allowed to
continue for 24h at 120°C under constant stirring.
To obtain P(70/30) and P(30/70) copolymers, PEG diols and monomers were introduced into
a three-neck round-bottom flask at a molar ratio of monomer:ethylene glycol of 4.2:1. The
molar ratio of ε-caprolactone:D,L-lactide was either 70:30 or 30:70. Following melting,
degassing and N2-purging at 90°C, the temperature was briefly raised to 140°C to complete
the melting of D,L-LA and subsequently lowered to 120°C at which temperature
polymerization was carried out for 24h under constant stirring. The RBF was immersed in the
oil bath as completely as possible to avoid condensation of D,L-LA on the walls of the flask.
1. Start heating the oil bath under stirring.
2. Remove the septum and add PEG diols and monomer to the flask. Start stirring
(appropriate stirring speed is slow, approximately 60 rpm) and seal the RBF with the
septum.
131
3. When all components are melted, carry out 3-4 cycles of degassing/N2 purging
through a gauge needle inserted through the septum. Stop when no more gas is being
removed from the mix, and finish by purging one last time with N2.
4. When the oil bath is stable at 120°C, inject SnOct catalyst through the septum using a
glass/steel syringe (25-100µl), making sure no air is injected. Avoid SnOct dripping
onto the flask wall. To avoid air bubbles, mount the syringe with a clip to keep it still,
and slowly aspire and eject SnOct a few times. Each time the syringe is filled, move
the piston little by little (5-10µl steps) and wait for the viscous liquid to fill the
syringe (several minutes).
5. Once the catalyst has been added to the melted monomer/diol mix, allow the synthesis
to proceed for 24h under constant stirring. As the polymerization proceeds the
reaction mixture will become increasingly viscous. It is a good idea to check on the
set up now and then. Towards the end of the reaction time it may become necessary to
stop and start the stirring, in order to let the polymer mass slide back down to the
bottom of the flask where it may be better stirred, or to stop the stirring completely to
avoid breaking the RBF.
6. At 24h, turn off the oil bath and remove it. Allow the polymer mass to cool down
(several hours or overnight).
STEP 4: RECOVERY
Polymers can be dissolved in dichloromethane in the reaction bottle, and precipitated in cold
ethanol, followed by filtration (coarse glass and standard filter paper) and drying in a vacuum
oven at 30°C until constant weight.
132
Figure A.1: Setup for copolymer synthesis. Pieces for stirring are shown separately (A) and assembled (B). (C) and (D) show the setup attached to the motor unit and mounted on a stand.
A B
C D
133
A2. Protocol for surface modification, cell seeding, cell count, and imaging.
Films were prepared in a hot melt press. Samples that fit into a 48-well plate can be punched
out using a 1cm diameter biopsy punch. P(100/0) is a tough and resistant material, while
P(70/30) is softer and can brake during manipulation and chemical/biochemical treatment,
especially if thin films are used (0.25mm). A gentler version of the aminolysis treatment is
performed on P(70/30) to avoid massive cracking of the surface. In the treatment of P(70/30),
the HCl protonizing step is excluded, and it is therefore recommended that any further surface
modification be carried out immediately following aminolysis.
Care has to be taken when transferring samples from one plate to another to avoid breaking
them. Make sure that films do not flip over in their wells at any time during the experiment.
This can easily occur during rinsing if glass rings are not used and also upon transfer between
wells. In the following, approximate time for each step is given in parentheses. RT = room
temperature.
STEP 1: AMINOLYSIS (5H/2H)
Instructions for P(100/0) and (P(70/30)).
Prepare: 10% (wt/wt) 1,6-hexanediamine in isopropanol (heat slightly and stir to aid
dissolution)
5mM HCl
1. Incubation in 10% (wt/wt) 1,6-hexanediamine for 2h (30 min) at RT. Rinse 3×mQ.
2. Rinse in 5mM HCl (mQ) for 1h (30 min) to protonize the aminogoups
3. Rinse 3×mQ
4. Dry in vacuum oven (pass directly to protein immobilization)
STEP 2: PROTEIN IMMOBILIZATION (2H + 0,5H+INCUBATION)
Prepare: Sterile filtration of the DMP solution (10mM, in PBS pH 8).
Sterile 48-well plates
Sterile tweezers (2)
Sterile PBS pH 7.4 d pH 8
Marked, sterile tubes for protein solutions
134
Filter 0.45um and syringe for Fgn filtration
1. Incubate the films in 10mM DMP, 1h in RT. (If control surfaces are included,
remember to leave these without!) In the meantime, prepare the protein solutions (as
late as possible).
2. Aspire the liquid, and rinse twice with PBS pH 8.
3. Add the protein solutions, 300µl/well (wells with glass rings need less). Carefully push
floating films to the bottom (they will stay there).
4. Incubate 4h at 4°C and 1h in room temperature. Leave fibrinogen surfaces (for fibrin)
in RT.
5. After 2h in room temperature, aspire the fibrinogen solution (don’t rinse), add
thrombin solution (or thrombin + DMP), and incubate 3h at 37°C.
6. Rinse all surfaces in sterile PBS×2, and sterilize (step 3).
STEP 3: STERILIZATION (OVERNIGHT + 45 MIN)
1. Incubate the surfaces in PBS with a 5X concentration of antibiotics overnight at 4°C
(to conserve proteins).
2. Rinse the surfaces 2x10min PBS
3. Transfer to new sterile 48-well plates.
STEP 4: CELL SEEDING (1,5H)
Prepare: Use freshly prepared surfaces if possible.
At the suggested seeding density, have one T75 of SMC for each 48-well plate
ready for passage.
Sterile glass rings
1. Equilibrate the films in 200µl of medium for 1h at 37°C. Place sterile glass rings
where necessary.
2. Passage the cells.
3. Seed SMC at 10,000 cells/well for counting experiments, and at 5,000 cells/well for
imaging. A small Petri dish may be seeded for control fluorescent staining.
4. Change medium after 8h, and then every two-three days for longer experiments.
135
STEP 5: COUNTING THE CELLS (TIME DEPENDS ON SIZE OF EXPERIMENT) OR FIXING THEM FOR
SEM IMAGING (45MIN + OVERNIGHT).
Prepare:
For counting: Centrifuge tubes (15ml) identifying the surface type and batch number
Identically marked 1,5ml tubes for Trypan Blue staining
Warm medium (or DMEM, cheaper), Trypsin-EDTA, serum
For fixing: Glutaraldehyde solutions (1,4% and 2,5%)
Cell count:
1. Add 100µl of warm serum to each of the identified centrifuge tubes.
2. Rinse the surfaces once in PBS (max 4 pools) and carefully transfer them without glass
rings to a new 48-well plate (to avoid counting cells adhering to the TCPS below).
3. Add 200µl of warm trypsin-EDTA to the samples (max. 4 pools) and incubate for 5
minutes at RT.
4. Transfer the trypsin solution of identical surfaces to the corresponding 15ml centrifuge
tube. Add 300µl of warm DMEM (or medium) to the wells to recover the remaining
cells.
5. Centrifuge SMC at 1600rpm for 4 min
6. Carefully remove the supernatant (you will probably not see the pellet) and resuspend
in 100µl of DMEM. Keep the remaining suspension in its tube.
7. Transfer the 100µl to a 1.5ml Eppendorf, add 50µl of Trypan Blue, and incubate for 7
minutes.
8. Count with hemacytometer (at least 6 fields)
9. Measure the volume of the remaining suspension in the Falcon tubes and use this to
calculate the total number of cells/surface condition.
10. Normalize the cell number to cells/area unit.
136
Fixing and imaging in SEM
1. Fix the cells in 1,4% glutaraldehyde, 30 min at RT
2. Aspire the liquid and add glutaraldehyde 2,5% to fix overnight in the fridge (can be
left up to a week).
3. CHUS: Post-fixing with OsO4 and dehydration in an EtOH gradient followed by
drying in a critical point dryer.
4. Metallize the samples
5. Image in SEM, using high contrast and low current.
Fixing and fluorescent staining
Follow laboratory SOPs.
A3. Standard curve established for the ninhydrin assay used in Chapter 4.
Ninhydrin assay: standard curve
y = 5578xr2 = 0,9997
0
100
200
300
400
500
600
0,00 0,02 0,04 0,06 0,08 0,10
C (mM)
AU
(57
5nm
)
Fig. A.2: Standard curve established for ninhydrin assay.
137
A4. Control α-actin staining of SMC following culture on TCPS and Fn-immobilized
P(100/0) in Chapter 4.
Fig. A.3: Alpha-actin staining of cells reseeded on TCPS following six days of culture on (A) Fn-
immobilized P(100/0) and (B) TCPS.
A) B)