Investigation into the proteolytic activity in
chronic wound fluid and development of a
remediation strategy
By
Erin Alexis Rayment
Bachelor of Biotechnology Innovation (Hons)
Tissue Repair and Regeneration Program
Cells and Tissues Domain
Institute of Health and Biomedical Innovation
A thesis submitted for the degree of Doctor of Philosophy at the Queensland University of Technology
2007
ii.
KEYWORDS
Wound healing, chronic ulcer, hydrogel, matrix metalloproteinase, gelatinase,
zymography, protease inhibition, wound dressing, biomaterials, tissue engineering.
iii.
ABSTRACT
Chronic ulcers are an important and costly medical issue, causing their sufferers a
large amount of pain, immobility and decreased quality of life. The common
pathology in these chronic wounds is often characterised by excessive proteolytic
activity, leading to the degradation of both the extracellular matrix, as well as key
factors critical to the ulcer’s ability to heal. As matrix metalloproteinases (MMPs), a
large family of zinc-dependent endopeptidases, have been shown to have increased
activity in chronic wound fluid (CWF), it was hypothesised that this specific
proteolytic activity was directly related to an ulcer’s chronic nature. Although
previous studies have identified elevated proteases in CWF, many have reported
contradictory results and therefore the precise levels and species of MMPs in CWF are
poorly understood. The studies reported herein demonstrate that MMP activity is
significantly elevated in CWF compared with acute wound fluid (AWF). In particular,
these studies demonstrate that this proteolytic activity can be specifically attributed to
MMPs and not another class of proteases present in wound healing. Furthermore, it is
shown that MMP-9 is the predominant protease responsible for matrix degradation by
CWF and is an indicator of the clinical status of the wound itself. Moreover, MMP-9
can be inhibited with the bisphosphonate alendronate, in the form of a sodium salt, a
functionalised analogue, and also tethered to a synthetic biocompatible hydrogel
compromised of aqueous poly (2-hydroxy methacrylate) PHEMA synthesised in the
presence of poly(ethylene glycol) (PEG). Together, these results highlight the
potential use of a tethered MMP inhibitor as an improved ulcer treatment to inhibit
protease activity in the wound fluid, while still allowing MMPs to remain active in the
wound bed where they perform vital roles in the activation of growth-promoting
agents and immune system regulation.
iv.
TABLE OF CONTENTS
Keywords …ii
Abstract …iii
Table of Contents …iv
List of Figures …viii
List of Tables …x
List of Abbreviations …xi
List of Publications and Presentations …xiv
Statement of Original Authorship …xv
Acknowledgements …xvi
CHAPTER 1: LITERATURE REVIEW
1.1 Introduction …2
1.2 Wound Healing …5
1.2.1 Matrix Metalloproteinases and their Role in Wound
Repair …7
1.2.2 Inhibitors of MMP Activity …12
1.2.2.1 Bisphosphonates …14
1.3 Polymers Suitable for Controlled Drug Release in a Wound
Dressing …18
1.3.1 Hydrogels …20
1.3.2 Crosslinking of Polymers …23
1.3.3 Functionalisation of polymers with bisphosphonates …27
1.4 Project Outline …30
1.4.1 Research Problem …30
1.4.2 Hypothesis …30
1.4.3 Aims …31
1.4.4 Project Design …31
CHAPTER 2: ANALYSIS OF MATRIX METALLOPROTEINASES
IN CHRONIC WOUND FLUID
2.1 Introduction …34
2.2 Materials and Methods …37
v.
2.2.1 Wound fluid sample collection and preparation …37
2.2.2 Identification of protease activity using collagen
zymography …38
2.2.3 Analysis and quantitation of collagen zymography …38
2.2.4 Confirmation of MMP-specific collagen degradation
using GM6001 inhibitor …39
2.2.5 Specific identification of MMPs through
immunoprecipitation …39
2.2.6 Quantitative confirmation of MMP-9 levels through a
direct enzyme-linked immunosorbant assay (ELISA) …40
2.3 Results …41
2.3.1 Elevated proteolytic activity in chronic wound fluid …41
2.3.2 MMPs are responsible for collagen degradation as
shown through a specific inhibitor study …43
2.3.3 Initial immunoprecipitation of MMPs from CWF samples …43
2.3.4 MMP-9 is the predominant protease responsible for
matrix degradation in chronic wound fluid …45
2.3.5 MMP-9 levels present in chronic wound fluid correlate with
the clinical severity of the ulcer …54
2.4 Discussion …56
2.5 Conclusion …60
CHAPTER 3: SYNTHESIS OF HYDROGELS FOR WOUND
DRESSING APPLICATIONS
3.1 Introduction …62
3.2 Materials and Methods …65
3.2.1 Chemicals …65
3.2.2 Synthesis of hydrogels …65
3.2.3 Analysis of polymerisation through NIR FT-Raman
Spectroscopy …66
3.2.4 Swelling of hydrogels …66
3.2.5 Protein loading …67
3.2.6 Protein release …67
3.2.7 Skin collection …67
vi.
3.2.8 Primary keratinocyte cell cultures …68
3.2.9 Toxicity testing using a cell number assay …68
3.2.10 Preparation of a three-dimensional human skin
equivalent …68
3.2.11 Biocompatibility testing of hydrogels using the DED
model …70
3.2.12 Immunohistochemistry …70
3.3 Results …73
3.3.1 Hydrogel preparation …73
3.3.2 Analysis of polymerisation through NIR FT-Raman
Spectroscopy …73
3.3.3 Hydrogel swelling ratios in water …73
3.3.4 HRP release from the hydrogel sheets following
diffusion loading …75
3.3.5 Toxicity testing using a cell number assay …78
3.3.6 Biocompatibility testing using a three-dimensional
human skin equivalent …81
3.4 Discussion …84
3.5 Conclusion …89
CHAPTER 4: SYNTHESIS AND EVALUATION OF
FUNCTIONALISED-ALENDRONATE HYDROGELS
FOR TREATMENT OF EXCESSIVE PROTEASE
ACTIVITY IN CHRONIC WOUND FLUID
4.1 Introduction …92
4.2 Materials and Methods …95
4.2.1 Chemicals …95
4.2.2 Functionalisation of alendronate …95
4.2.3 Characterisation of methacrylated-alendronate through
Fourier Transform Nuclear Magnetic Resonance (FT-NMR)
Spectroscopy …95
4.2.4 Analysis of MMP-inhibition through incubation with
the methacrylated-alendronate …96
4.2.5 Synthesis of hydrogels …96
vii.
4.2.6 Analysis of polymerisation through NIR FT-Raman
Spectroscopy …97
4.2.7 Analysis of MMP-inhibition through incubation with
the alendronate-functionalised hydrogels …97
4.2.8 Biocompatibility testing of hydrogels using the DED
model …97
4.2.9 Immunohistochemistry …98
4.3 Results …99
4.3.1 Synthesis and characterisation of methacrylated-
alendronate …99
4.3.2 Analysis of MMP-inhibitory action of methacrylated-
alendronate …105
4.3.3 Preparation and characterisation of alendronate-
functionalised hydrogels …105
4.3.4 Analysis of MMP-inhibitory action of alendronate-
functionalised hydrogels …108
4.3.5 Biocompatibility testing using a three-dimensional
human skin equivalent …108
4.4 Discussion …112
4.5 Conclusion …116
CHAPTER 5: GENERAL DISCUSSION …117
CHAPTER 6: REFERENCES …129
viii.
LIST OF FIGURES
1.1 The domain structure of MMPs …8
1.2 Overlay of the backbone atoms from 4 separate MMPs. …11
1.3 Structure of bisphosphonates …14
1.4 Chemical structures …22
1.5 Mechanism of gamma induced polymerisation of HEMA …24
1.6 Potential mechanisms for gamma induced random grafting/
termination or polymerisation of HEMA and PEG …25
1.7 Mechanism of gamma induced crosslinking of HEMA and
EDGMA …28
2.1 Collagen Type I zymography demonstrating protease activity
present in wound fluid samples …42
2.2 Collagen Type IV zymography demonstrating protease activity
present in wound fluid samples …42
2.3 GM6001 inhibition of protease activity in CWF-1 confirms the
MMP specific degradation of Collagen Type I as revealed through
zymography …44
2.4 Initial immunoprecipitation of MMP-1, -8, -13 from CWF-1 and
analysis of enriched fractions through Collagen Type I zymography …46
2.5 Immunoprecipitation of MMP-8 from wound fluid samples and
analysis of MMP-8 enriched fractions through Collagen Type I
zymography …47
2.6 Immunoprecipitation of MMP-13 from wound fluid samples and
analysis of MMP-8 enriched fractions through Collagen Type I
zymography …48
2.7 Immunoprecipitation of MMP-9 from wound fluid samples and
analysis of MMP-9-bound fractions through Collagen Type I
zymography …50
2.8 Immunoprecipitation of MMP-9 from wound fluid samples and
analysis of MMP-9-unbound fractions through Collagen Type I
zymography …52
2.9 MMP-9 levels from wound fluid samples analysed through an
ELISA …55
ix.
3.1 Experimental design of the toxicity testing …69
3.2 Experimental timeline of the three-dimensional ex vivo human
skin equivalent model …71
3.3 Hydrogel sample analysis using NIR FT-Raman Spectroscopy …74
3.4 Swelling studies of hydrogels in water …76
3.5 Gravimetric study of the loss of PEG from the HEMA:PEG
hydrogels …77
3.6 Cumulative release of HRP from the hydrogel sheets into a
HEPES buffer at pH 7.4 …79
3.7 Toxicity testing of components leached from the hydrogels in
primary skin keratinocytes using the CyQUANT® NF Cell
Proliferation Assay to determine cell number …80
3.8 Histological and immunohistochemical analysis of an ex vivo
human skin model following exposure to both a commercially
available wound dressing and synthesised hydrogels …82
4.1 Reaction schemes …100
4.2 FT-NMR 1H spectrum of methacrylated-alendronate …101
4.3 FT-NMR 31P spectrum of methacrylated-alendronate …103
4.4 Potential structure of methacrylated-alendronate dimer …104
4.5 Collagen Type I zymography demonstrating inhibition of
protease activity present in wound fluid samples by alendronate
sodium salt and its methacrylated-counterpart …106
4.6 Analysis of alendronate-functionalised hydrogels using NIR
FT-Raman Spectroscopy …107
4.7 Collagen Type I zymography demonstrating inhibition of
protease activity present in wound fluid samples by alendronate-functionalised
hydrogels …109
4.8 Histological and immunohistochemical analysis of an ex vivo
human skin model following exposure to the alendronate-
functionalised hydrogel …110
x.
LIST OF TABLES
1.1 Acute versus chronic wounds …6
1.2 Distribution of MMPs among cell types in cutaneous wound
repair …9
1.3 Relative potencies of bisphosphonates in vitro …15
1.4 Bisphosphonates in clinical studies/development …16
1.5 Description of various polymers available for use in biomedical
applications and their properties …19
1.6 Hydrophilic polymers suitable for hydrogel matrices …21
1.7 The reaction scheme of the radiolysis of neutral water …27
2.1. Patient clinical data from chronic wound fluid samples …37
3.1 Hydrogel formulations prepared for analysis as potential
wound dressings …65
4.1 Functionalised hydrogel formulations prepared for analysis
as potential wound dressings …96
4.2 Features of methacrylated-alendronate FT-NMR 1H spectrum …102
4.3 Features of methacrylated-alendronate FT-NMR 31P spectrum …104
xi.
LIST OF ABBREVIATIONS
ABTS 2,2'-Azinobis [3-ethylbenzothiazoline-6-sulfonic acid]-diammonium
salt
ARC Australian Research Council
AWF acute wound fluid
BSE Bovine Spongiform Encephalopathy
CC-3 cleaved caspase-3
CHCl3 chloroform
CO2 carbon dioxide
CWF chronic wound fluid
DED de-epidermised dermis
DMEM Dulbecco’s Modified Eagle’s Medium
DMF N,N-Dimethylformamide or (CH3)2NC(O)H
D2O deuterated water
ECM extracellular matrix
EDTA ethylenediaminetetraacetic acid
EGDMA ethylene glycol dimethacrylate
EGF epidermal growth factor
ELISA enzyme-linked immunosorbant assay
FCS foetal calf serum
FDA Food and Drug Administration
FG Green’s media + serum
FGF fibroblast growth factor
FT Fourier Transform
GA1X Irradiated H2O:HEMA:PEG:1X methacrylated-alendronate (10 kGy)
GA10X Irradiated H2O:HEMA:PEG:10X methacrylated-alendronate (10 kGy)
G-CSF granulocyte-colony stimulating factor
GM6001 N-[(2R)-2-(hydrox- amidocarbonylmethyl)-4-methylpentanoyl]-L-
tryptophan methylamide
GRAS Generally Regarded as Safe
HCl hydrochloric acid
HEMA 2-hydroxylethyl methacrylate
HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
xii.
H&E haematoxylin and eosin
H2O water
HRP horseradish peroxidise
HS human serum
IGF insulin-like growth factor
IGFBP insulin-like growth factor binding protein
IgG immunoglobulin G
IL-1β interleukin-1β
i3T3s lethally irradiated 3T3 mouse fibroblast feeder cells
K1/10/11 keratin 1/10/11
KGF keratinocyte growth factor
mAb monoclonal antibody
MMP matrix metalloproteinase
MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide
NaAMPS 2-acrylamido-2-methyl- propanesulphonic acid sodium salt
NaCMC sodium carboxymethylcellulose
NaOH sodium hydroxide
NIR Near Infra-Red
NMR Nuclear Magnetic Resonance
nvCJD new variant Creutzfeldt Jakob Disease
PAA poly(acrylic acid)
pAb polyclonal antibody
PBS phosphate buffered saline
PDGF platelet-derived growth factor
PEO poly(ethylene oxide)
PEG poly(ethylene glycol)
PGA poly(glycolic acid)
PH-5 50:20 H2O:HEMA (5kGy)
PH-10 50:20 H2O:HEMA (10 kGy)
PHEMA poly(2-hydroxyethyl methacrylate)
PLA poly(lactic acid)
PLGA copolymer of PGA and PLA
PMA poly(methyl acrylate)
Polymer C 50:20:30 H2O:HEMA:PEG (5 kGy)
xiii.
Polymer G 50:20:30 H2O:HEMA:PEG (10 kGy)
P(PF-co-EG) poly(propylene fumarate-co-ethylene glycol)
PUSH Pressure Ulcer Scale of Healing
PVA poly(vinyl alcohol)
ROSs reactive oxygen species
SDS sodium dodecyl sulfate
SEM standard error of the mean
SFM serum-free medium
SMP skim milk powder
SPA acrylic acid (3-sulphopropyl)ester potassium salt
TBS Tris buffered saline
TBST Tris buffered saline with 0.1% Tween-20
TGF transforming growth factor
TIMP tissue inhibitors of metalloproteinase
TNF-α tumour necrosis factor-α
tPA tissue plasminogen activator
uPA urokinase plasminogen activator
xiv.
LIST OF PUBLICATIONS AND PRESENTATIONS
Rayment, E.A., Upton, Z., Shooter, G.K. ‘Increased matrix metalloproteinase-9 (MMP-9) activity observed in chronic wound fluid is related to the clinical severity of the ulcer’, British Journal of Dermatology, accepted, 4 December, 2007.
Rayment, E.A, Dargaville, T.R., Shooter, G.K., George, G.A., Upton, Z., ‘Attenuation of protease activity in chronic wound fluid with alendronate-functionalised hydrogels’, manuscript under review, 16 July, 2007.
Rayment, E.A, Dargaville, T.R., Upton, Z., ‘Wound repair: composition and methods’, Australian Provisional Patent # 2007903101, filed 8 June, 2007.
Rayment, E.A., Fernandez, M.L. Shooter, G.K., Upton, Z., ‘The clinical severity of a chronic ulcer is associated with elevated levels of matrix metalloproteinase-9 activity observed in chronic wound fluid’, Proc. of Tissue Engineering and Regenerative Medicine International Society – North American Meeting, Toronto, Canada, June 2007.
Rayment, E.A., Shooter, G.K., Upton, Z., ‘Chronic wound fluid exhibits elevated matrix metalloproteinase activity leading to the degradation of collagen in the wound bed’, Proc. of Discovery Science and Biotechnology Meeting, Brisbane, Australia, May 2007. Rayment, E.A., Shooter, G.K., George, G.A., Upton, Z., ‘Elevated matrix-metalloproteinase activity in chronic wound fluid’, Proc. Of European Tissue Repair Society Conference, Pisa, Italy, September 2006.
Rayment, E.A., George, G.A., Upton, Z., ‘Analysis of aqueous PEG-HEMA wound dressing synthesis, protein release and biocompatibility in a three-dimensional skin model’, Proc. of Australasian Society of Medical Research Conference, Brisbane, Australia, May 2006.
Rayment, E.A., George, G.A., ‘Aqueous PEG-HEMA hydrogel synthesis and protein release for wound dressing applications’, Proc. of Australasian Society of Biomaterials Conference, Rotorua, New Zealand, February 2006.
Rayment, E.A., George, K.A., George, G.A., ‘Aqueous PEG-HEMA hydrogel synthesis using gamma irradiation for wound dressing applications’, Proc. of European Society of Biomaterials Conference, Sorrento, Italy, September 2005.
Rayment, E.A., George G.A., ‘Scaffold design for a bioactive wound dressing’, Queensland Polymer Symposium, Brisbane, Australia, February 2005.
xv.
STATEMENT OF ORIGINAL AUTHORSHIP
The work contained in this thesis has not been previously submitted to meet
requirements for an award at this or any other higher education institution. To the best
of my knowledge and belief, the thesis contains no material previously published or
written by another person except where due reference is made.
Signature:
Date:
xvi.
ACKNOWLEDGEMENTS
Firstly, I would like to sincerely thank my principal supervisor, Prof. Zee Upton, for
the opportunity to embark on this project, as well as her continued encouragement,
assistance and direction through the establishment and completion of my PhD studies.
Secondly, I would also like to acknowledge my associate supervisor, Prof. Graeme
George, for his support and direction, especially in terms of the relatively new area of
polymer science for me. I would not have been able to successfully accomplish this
project without their combined help.
In terms of experimental design and practical support, Dr Gary Shooter and, in the
later stages of this project Dr Tim Dargaville, have both been exceptional. From
providing me with fool-proof protocols, to helping me design completely new
methods, I would like to thank you both sincerely. I fully believe that without the both
of you this project would not have been as successful as it has – culminating in the
recent filing of an Australian provisional patent application.
As this was such a diverse project, many people have been instrumental in terms of
specific knowledge and technical expertise. Therefore, in no particular order, I would
like to thank the following people: Rebecca Dawson, Evette Kairuz, Gemma Topping
and Danielle Borg for their help with the DED model and subsequent
immunohistochemistry; Dr Llew Rintoul for the FT-Raman spectroscopy training; Dr
Mark Wellard for his NMR expertise; Mel, James and Shea for their help and
brainstorming in relation to protein work; and the entire Tissue Repair and
Regeneration program here at QUT for their encouragement and assistance over the
years. In addition, the laboratory support team of Sonya, Scott and David have been
instrumental in ensuring everything has run smoothly.
This PhD project would not have been possible without the Australian Government
for their provision of an Australian Postgraduate Award, and the Institute of Health
and Biomedical Innovation for their top-up stipend. I would also like to thank the
numerous funding bodies who gave me money to travel and attend international
conferences and laboratories, namely: the School of Life Sciences, the TRR program,
xvii.
IHBI, QUT’s Grants-in-Aid scheme, the Australasian Society of Biomaterials and The
Australian Federation of University Women SA Inc.
I would also like to thank Mel, Gary and Shea - for their endless coffee breaks, laughs
so loud that you can hear from different levels, Friday drinks that have morphed into
Saturday mornings, and numerous inappropriate jokes – you have made my PhD an
extremely memorable experience. Thanks also to Gemma, Jacqui, Danielle, Brett and
Tony for their friendship and support, as well as everyone who has put up with my
consistently loud voice for the past three and a half years. You have all made my lifer
a richer place.
And an especially huge thanks to Abrona – your organisation of Zee, as well as the
entire TRR team (myself included!), means that things continually get done. Without
you, I can see myself getting stuck in huge piles of forms that nobody knows how to
fill out or who you have to give them to for signatures…
I would also like to thank my family for all of their support and encouragement
through my university years. I would especially like to thank my older sister,
Cassandra for spurring me on. The thought of potentially completing before her has
definitely been a motivator – and most importantly, has made me write faster…
Last, but certainly not least, I would like to thank Dave from the bottom of my heart.
You have been an incredible support and encourager through this entire project, as
well as being a practical helper with the design of numerous objects for my research.
Without you, I don’t think this entire process would have been anywhere near as fun.
And finally, I promise never to try and explain to you my project again.
xviii.
CHAPTER 1
LITERATURE REVIEW
2.
1.1 INTRODUCTION
Chronic wounds are an important and costly medical challenge. Venous insufficiency,
in the form of venous leg ulcers, is estimated to account for 80-90% of all chronic
ulcers (Shai and Halevy 2005). American figures show that more than four million
people (1.3% of the population) each year are affected by these wounds, costing
upwards of US$ 9 billion to treat (Clinicaltrials.gov 2002). Chronic wounds currently
place a significant burden on the Australian healthcare system, with prolonged
hospitalisations (at a cost of AU$ <500 per day for 2-3 months at a time) frequently
required. Current prevalence rates of chronic wounds in Australia are estimated to be
between 1-3% of the population (200,000 – 600,000 people) and cost the Australian
community upwards of AU$500 million per annum (Baker and Stacey 1994; Gruen,
Chang et al. 1996). Currently, there is no one accepted treatment for this costly
medical crisis, which causes the vast majority of lower leg amputations and
significantly compromises the patient’s quality of life. This is primarily due to the
large variety of causes, as well as many co-morbidities associated with this disease,
such as diabetes, osteomyelitis, renal failure and heart disease (MacLellan 2000).
Current treatments of chronic wounds include a number of wound management
strategies. Thus, good nursing skills, including regular washing and debridement,
application of antimicrobials to prevent infection, and daily changing of the dressing,
can play a major role in improving the patient’s outcome. However, in many cases,
the aforementioned procedures are simply not enough to promote wound closure.
Intrinsic health factors, such as diabetes, immune status, and in many cases age, are
often the major inhibitory factors to wound healing (MacLellan 2000).
The development of innovative, cost-effective treatments to increase the healing rates
of leg ulcers is a priority for improved health in the elderly. This was identified as one
of the ‘research gaps’ in a recent workshop addressing “Leg Ulcer Healing in Older
People” (Feb 5-6, 2007, Brisbane) funded by the ARC/NHMRC Research Network in
Ageing Well. Chronic wounds are a common and costly reality for the Australian
community, and have largely been overlooked by researchers in the area of
biotechnology and, specifically, tissue repair. This significant, yet neglected, research
field not only represents a significant challenge to our current understanding of wound
healing, but also involves a healthcare issue that affects our society at all levels. This
3.
is in terms of cost to the patient, the community and the economy. These wounds are a
major cause of pain and anxiety for affected individuals and also contribute
significantly to their lessening mobility, decreased social interactions and overall
diminished quality of life (MacLellan 2000). In addition, patients with an ulcer are
more likely to suffer depression, downfall risk, malnutrition and functional disorders
(Jawien, Szewczyk et al. 2006). Further, a two-year retrospective study published in
1996 identified all patients admitted to a major Australian teaching hospital with a leg
ulcer as the prime admission criteria and found that wound management of these 199
patients cost AU$2.75 million, yet achieved a healing rate of only 16% (McMullin,
Hyde et al. 1998)! Clearly, in many cases appropriate wound management strategies
and good nursing skills are simply not sufficient to promote wound healing
(MacLellan 2000).
Wound dressings currently on the market do not fully address all of the physiological
issues associated with chronic wounds. Many are simply antimicrobials, e.g. Actisorb
Plus (Johnson & Johnson) and Acticoat (Smith & Nephew), which while they control
bacterial growth do not attend to potential growth factor deficiencies (Stewart 2002).
Conversely, interactive dressings, e.g. Alginates (multiple products) and Promogran
(Johnson & Johnson), maintain a moist wound environment and even claim to interact
with cells or matrix proteins in the wound (Stewart 2002). In terms of alginates, the
presence of calcium in the dressing is reported to modify cellular responses (Stashak,
Farstvedt et al. 2004). However, these products are derived from animal and plant
products and are therefore not fully defined and may not be completely pathogen-free.
In terms of animal-derived products this presents unacceptable risks, especially
following the aftermath of the BSE (Bovine Spongiform Encephalopathy or Mad Cow
Disease) and nvCJD (new variant Creutzfeldt Jakob Disease) crises.
In 2000 the advanced wound management market was estimated at £1.2 billion, with a
growth rate of 7.5% over the preceding year (Smith and Nephew 2000). Active wound
management treatments, e.g. growth factors (Regranex, Johnson & Johnson) and
dermal substitutes (Dermagraft, Smith & Nephew), show an accelerated growth of
over 15% per annum with a market value of £200 million (Smith and Nephew 2000),
demonstrating why companies are willing to invest large amounts of money into this
type of product development. Current topical growth factors treatments include
4.
platelet-derived growth factor (PDGF, becaplermin, Regranex®), granulocyte-colony
stimulating factor (G-CSF, Neupogen® Filgrastim), while fibroblast growth factor-2
(FGF-2) is approved for use in Japan. Regranex® gel 0.01% (Ortho-McNeil
Pharmaceutical, Raritan, NJ, USA) is sold under the Johnson & Johnson Wound
Management product label and contains rhPDGF-BB in a sodium
carboxymethylcellulose (NaCMC) gel. This product was the first topically applied
growth factor treatment to show wound closure efficacy in Phase III clinical trials
(LeGrand 1998). In contrast, Filgrastim has demonstrated limited usefulness in
promoting healing in diabetic ulcers (Kästenbauer, Härnlein et al. 2003; Papanas and
Maltezos 2007). The major obstacle for patients obtaining effective growth factor
treatments is cost – Regranex® costs upwards of US$375/15 g tube (Ugarte, Roberts
et al. 2002), which considering the daily treatment regime, can equate to thousands of
dollars per wound. While this has been proven cost-effective by a number of studies
(Kantor and Margolis 2001), this treatment is still out of reach of the typical chronic
ulcer sufferer (Eldor, Raz et al. 2004).
The aging population, along with the increased prevalence of Type II diabetes, has led
many to concentrate on improving wound dressing treatments for these chronic
wounds. However, as this is an incredibly complex issue, many factors need to be
taken into account. With this in mind, this literature review will provide background
information on wound healing; thus will summarise the characteristics of chronic
ulcers, including increased protease activity and possible protease inhibitors; and will
also focus on controlled delivery of active compounds through polymeric materials.
This review will also concentrate on the interplay between these various themes.
5.
1.2 WOUND HEALING
Wound healing in the skin is a dynamic process. It consists of several stages,
including: inflammation, chemotaxis and cell division, neovascularisation, synthesis
of extracellular matrix, and remodelling of the scar tissue (Tarnuzzer and Schultz
1996). Inflammatory cells first release growth factors and cytokines that are essential
in the regulation of effective wound healing (Wagner, Coerper et al. 2003). Many
growth factors are important to the wound healing process. These include: insulin-like
growth factor (IGF), epidermal growth factor (EGF) and keratinocyte growth factor
(KGF); factors that are pivotal to epidermal migration, growth and differentiation
(Soler, Wright et al. 1999; Gibbs, Pinto et al. 2000; Allen, Asnes et al. 2002). There
are many different members of each of these families of growth factors that perform
different functions during wound repair. For example, IGF-I is essential for healing,
but insulin-like growth factor binding protein-3 (IGFBP-3) significantly decreases
repair (Wagner, Coerper et al. 2003). Another important growth factor in wound
healing, transforming growth factor-α (TGF-α), which along with IGF-I, also appears
to induce or enhance the expression of antimicrobial peptides and polypeptides
(Sorensen, Cowland et al. 2003). This demonstrates the multifunctional nature of
growth factors, as well as the complex dynamics of wound healing itself.
There are two main types of wounds – chronic and acute. While acute wounds follow
the wound healing process described above, disruption of any stage in this complex
cascade leads to the formation of chronic wounds (Trengove, Stacey et al. 1999).
Several previous studies have shown that there are significant differences between the
acute and chronic wound environments (Table 1.1) (Tarnuzzer and Schultz 1996;
Stadelmann, Digenis et al. 1998; Trengove, Stacey et al. 1999). In chronic wounds,
healing occurs with the formation of abundant granulation tissue and excessive
fibrosis, both of which lead to scar contraction and loss of function (Stadelmann,
Digenis et al. 1998). In addition, the major difference shown in chronic wound fluid
(CWF) compared with acute wound fluid (AWF) is their significant increase in
proteolytic activity (Tarnuzzer and Schultz 1996; Ladwig, Robson et al. 2002; Li and
Li 2003). Chronic wounds are characterised by elevated levels of pro-inflammatory
cytokines (e.g. tumour necrosis factor-α (TNF-α) and interleukin-1β (IL-1β)),
proteases (e.g. matrix metalloproteinases (MMPs)) and neutrophil elastase (Chen,
Schultz et al. 1999). These wounds also contain reduced levels of tissue inhibitors of
metalloproteinases (TIMPs) (Chen, Schultz et al. 1999; Baker and Leaper 2000).
Furthermore, TNF-α has a demonstrated ability to increase production of MMPs,
which in turn inhibits production of TIMPs. This then causes an extension in pro-
inflammatory cytokine expression, which can change a normal healing wound into an
environment rich in proteases and extremely toxic for connective tissue proteins and
growth factors (Chen, Schultz et al. 1999). Not surprisingly, chronic wounds have
reduced levels of growth factors, such as PDGF, basic fibroblast growth factor
(bFGF), EGF and TGF-β (Stadelmann, Digenis et al. 1998). Therefore, the regulation
of MMPs by their inhibitors might be necessary to promote wound healing, even
before growth factors can be applied (Tarnuzzer and Schultz 1996; Stadelmann,
Digenis et al. 1998; Trengove, Stacey et al. 1999; Cullen, Watt et al. 2002).
Table 1.1. Acute versus chronic wounds (Menke, Ward et al. 2007)
The chronic nature of ulcers is believed to be caused by certain endopeptidases,
namely serine proteases and matrix metalloproteinases as reported above (Grinnell,
Ho et al. 1992; Palohati, Lauharanta et al. 1993). Increased levels of neutrophils in
chronic wounds, along with their secreted proteases, have been implicated mediating
the tissue damage associated with this disease (Edwards and Howley 2007).
Furthermore, excessive inflammation caused by an hyper-stimulated neutrophil
response has also been suggested as a potential cause for a wound’s chronicity (Yager
and Nwomeh 1999). This can be further exacerbated by susceptibility of open wounds
to infection (Kantor and Margolis 2003). Since neutrophils also play a role in a
number of chemotactic, proteolytic and oxidative events in this environment, they
have also been focussed on a potential therapeutic target for these wounds. However,
a previous study by Weckroth et al. (1996) showed that serine proteases of
6.
7.
polymorphonuclear neutrophil origin, e.g. cathepsin G and neutrophil elastase, were
both low in activity in CWF (Weckroth, Vaheri et al. 1996).
Free radicals are another contributing factor to a chronic wound’s inability to heal.
The excess of oxygen free radicals found in a chronic wound environment often leads
to prolonged inflammation and also increased proteolysis as mentioned earlier. These
reactive oxygen species (ROSs) can subsequently cause protein, as well as lipid,
peroxidation, which then leads to cell disruption, tissue damage and cell death
(Cullen, Smith et al. 2002; Cullen, Watt et al. 2002). Therefore, when developing a
wound dressing there are many features to consider, including: the ability to attract
cells to the wound site; promotion of cell proliferation; the biocompatible and
bioresorbable function of the dressing; the limitation of excess proteinase activity;
and, the absorption and neutralisation of free radicals and excess metal ions (Hart,
Silcock et al. 2002).
1.2.1 Matrix Metalloproteinases and their Role in Wound Repair
MMPs, or matrixins, are a subgroup of the much larger metalloproteinase superfamily
(Parks 1999). Currently, there are more than 20 recognised members of this family of
proteases (Parks 1999), which are then divided into five subclasses based on their
substrate specificities: collagenases, gelatinases, stromelysins, membrane-type MMPs,
and other MMPs (Baker and Leaper 2000). This group of proteases also has a highly
conserved domain structure (Figure 1.1), consisting of a catalytic domain and an auto-
inhibitory pro-domain (Page-McCaw, Ewald et al. 2007).
More specifically, proteins classed as MMPs are defined as having the following five
characteristics. Firstly, they must require an active site Zn2+ that binds the conserved
histidine domain sequence, HEXXHXXGXXH, for their catalytic mechanism.
Second, a MMP’s prodomain needs to be approximately 80 amino acids in length and
contain the consensus sequence PRCXXPD. Third, a putative MMP must be
synthesised as an inactive zymogen. Fourth, it needs to be able to degrade or cleave at
least one extracellular matrix (ECM) protein, e.g. fibronectin, vitronectin. Lastly, its
proteolytic activity needs to be inhibited by TIMPs (Parks 1999).
Figure 1.1 The domain structure of MMPs (Somerville, Oblander et al. 2003)
In healthy tissue, MMP production and activation in fibroblasts is regulated by a
number of factors, namely expression of growth factors/cytokines (Ries and Petrides
1995), mechanical properties of the ECM/actin cytoskeleton (Tomasek, Halliday et al.
1997) and cell-matrix interactions (Azzam and Thompson 1992). Of relevance to this
project, urokinase plasminogen activator (uPA) and tissue plasminogen activator
(tPA) have both been associated with excessive matrix degradation and subsequent
activation of MMPs (Wysocki, Kusakabe et al. 1999). Furthermore, only specific
MMPs are present in those cells involving wound repair, and even then, only at certain
stages of this dynamic process (Table 1.2) (Parks 1999). The MMP family is quite
diverse in both its substrate specificity and role in wound healing. The two most
widely known and studied members are the collagenases and the gelatinases. MMP-1
(collagenase-1) and MMP-8 (collagenase-2, neutrophil-collagenase) appear to have a
8.
very limited substrate range, acting only on the fibrillar collagens, types I, II and III
(Parks 1999).
Table 1.2 Distribution of MMPs among cell types in cutaneous wound repair (+ indicates presence and – indicates absence in particular cell type) (Parks 1999).
Indeed, MMP-1 has been shown to be directly related to the ability of keratinocytes in
vivo to migrate from a wound edge (Soler, Wright et al. 1999). Moreover, MMP-8 is a
neutrophil-released collagenase and has substrate preference for type I collagen
(Diegelmann 2003). An important point to note is that while the presence of MMP-13
has been shown in chronic, and not acute, wounds, its contribution to the pathological
wound environment has not been identified (Fray, Dickinson et al. 2003). The
gelatinases (MMP-2, MMP-9) on the other hand cleave collagen more efficiently than
other MMPs (Parks 1999) and further degrade the cleaved type I collagen arising from
MMP-1 activity, as well as act on type IV collagen, the major form of collagen in
basement membranes (Tarnuzzer and Schultz 1996). MMP-2 (gelatinase-A) has a
vital role in ECM turnover and has been shown to have a role in collagen lattice
reorganisation in vitro (Cook, Stephens et al. 2000), as well as being one of only two
MMPs actively expressed in non-injured skin (Parks 1999). Recently, it was
demonstrated that reduced fibroblast MMP-2 production is associated with impaired
dermal healing in vivo in a diabetic mouse model (C57BL/KSJ-db/db) (Cook,
Stephens et al. 2000). MMP-9 (gelatinase-B) has also been shown to have the ability
to degrade Collagen types I, II, III, IV and V into small peptides at 37 ºC (Okada,
Gonoji et al. 1992). In addition, MMP-9 (gelatinase-B) is primarily produced by
9.
10.
inflammatory cells, neutrophils, macrophages and keratinocytes (Cullen, Watt et al.
2002).
Other than collagenases and gelatinases, the MMP family consists of stromelysins,
matrilysins, macrophage metalloelastase and membrane type MMPs. With regard to
stromelysins, MMP-3 (stromelysin-1) has a broad collagen substrate preference and
can also degrade proteoglycans (Tarnuzzer and Schultz 1996). It is also known to
degrade α1-anti-trypsin, an inhibitor of neutrophil elastase, leading to increased levels
of this protease in chronic wounds (Fray, Dickinson et al. 2003). MMP-10
(stromelysin-2) is thought to be an injury-specific protease, involved in desmosome
disassembly during migration (Parks 1999). MMP-7 (matrilysin) on the other hand is
a more potent proteoglycanase than both MMP-3 and MMP-9, and has recently been
implicated in playing an essential role in innate immunity (Wielockx, Libert et al.
2004). In addition, along with MMP-2 (gelatinase-a), matrilysin is found in non-
injured skin, being actively expressed by the ductal epithelium in all sweat and eccrine
glands (Parks 1999). MMP-12, a macrophage metalloelastase, is the most elastolytic
member of this family, i.e. has a strong ability to degrade elastic tissue (Parks 1999).
Furthermore, MMP-14 (membrane-type-1 MMP, MT1-MMP) has been shown to
promote both selective invasion and increased growth of malignant tumour cells in
vivo (Iida, Wilhelmson et al. 2004).
As well as acting to degrade matrix substrates, MMPs also have a wide variety of
growth factor and immune system-related functions (Somerville, Oblander et al.
2003). Indeed, recent mouse knockout models, as well as human diseases, have
indicated many unexpected biological functions of MMPs (Hautamaki, Kobayashi et
al. 1997; Holmbeck, Bianco et al. 1999; Wilson, Ouellette et al. 1999; Bergers,
Brekken et al. 2000; Boulay, Masson et al. 2001; Cataldo, Tournoy et al. 2002),
including many phenotypes that are yet to be fully understood. In terms of growth
factors, MMPs are responsible for both their release from the extracellular matrix
(Imai, Hiramatsu et al. 1997; Suzuki, Raab et al. 1997), as well as their proteolytic
inactivation (McQuibban, Butler et al. 2001; Fujiwara, Matsukawa et al. 2002), and
also have roles in shedding of receptors from the cell membrane (Levi, Fridman et al.
1996). Therefore, in terms of wound healing, this balance between growth factor
activation and degradation has to be carefully controlled to ensure correct progression
towards healing. Furthermore, MMPs have also been implicated in influencing the
immune system, both through activating defensins through cleavage of their pro-
domain (Wilson, Ouellette et al. 1999) and also by their ability to cleave all
immunoglobulin G (IgG) proteins (Gearing, Thorpe et al. 2002). This final point is
extremely important in terms of infection in chronic ulcers, as by cleaving IgG
proteins, MMP-3 and -7 are able to prevent the initiation of the complement cascade, a
key mechanism in clearing pathogens from the host system (Gearing, Thorpe et al.
2002).
There is much debate surrounding the under- and overexpression of various MMPs in
chronic wounds. This can be attributed to the highly conserved nature of this family
(Page-McCaw, Ewald et al. 2007), which can cause incorrect identification when
using antibody-based detection strategies as shown through the superimposition of
four separate MMPs (Figure 1.2).
Figure 1.2 Overlay of the backbone atoms from 4 separate MMPs MMP-1 (green), MMP-9 (purple), MMP-12 (turquoise), MMP-13 (dark green) (Rush III and Powers 2004)
One view is that the overexpression of MMP-3 and MMP-13 correlates with non-
healing wounds, whereas active MMP-1, -2, -9, and -14 are all required for the
progression of normal wound healing (Fray, Dickinson et al. 2003). Ladwig et al.
(2002) and Wysocki et al. (1999) both report that elevated levels of MMP-2 and
MMP-9 are detrimental to wound healing when present for prolonged periods, and
should be inhibited to allow normal repair to occur (Wysocki, Kusakabe et al. 1999;
11.
12.
Ladwig, Robson et al. 2002). Along with MMP-2 and MMP-9, one report has shown
that MMP-8 is likely to act as a debridement enzyme and elevated levels of this have
been shown in chronic pressure ulcer wound fluid (Yager, Zhang et al. 1996). Further
studies indicate that CWF includes high levels of various MMPs, including MMP-1, -
2, -3, -8, and -9, leading to the degradation of key elements in the wound healing
process (Cullen, Watt et al. 2002; Hart, Silcock et al. 2002). Contrary to previous
studies, Cook et al. (2000) reported that significantly decreased levels of active MMP-
2 and MMP-1, corresponding with marked increases in expression of TIMP-1 and
TIMP-2, led to the chronicity of non-healing wounds (Cook, Stephens et al. 2000).
This is in direct contrast with all other previously reported findings. One possible
reason for this is the use of biopsies from the wound bed by Cook et al. (2000),
instead of wound fluids as has been used by other studies (Cook, Stephens et al.
2000). Currently, it is not known if biopsy and fluid samples from chronic wounds
show similar profiles of MMPs and TIMPs; nor is it known which sample type gives a
more accurate assessment of the chronic wound’s molecular environment (Ladwig,
Robson et al. 2002).
1.2.2 Inhibitors of MMP Activity
Current topical treatments of chronic ulcers have proven to be only slightly effective
in treating this pathology (Cullen, Watt et al. 2002). It has been postulated that the
main reason for the unanticipated ineffectiveness of topical growth factor therapy to
promote wound healing is due to the high levels of proteases found in CWF (Cullen,
Watt et al. 2002). CWF also contain reduced levels of TIMPs (Chen, Schultz et al.
1999; Baker and Leaper 2000). Interestingly, TNF-α has a demonstrated ability to
increase production of MMPs, which in turn represses production of TIMPs. This then
prolongs pro-inflammatory cytokine expression, which can change a normal healing
wound into an environment rich in proteases and hence antagonises deposition and
function of connective tissue proteins and growth factors (Chen, Schultz et al. 1999).
Not surprisingly, chronic wounds have reduced levels of growth factors, therefore the
regulation of MMPs by their endogenous inhibitors might be necessary to promote
wound healing, even before topical growth factor treatments can be used (Tarnuzzer
and Schultz 1996; Stadelmann, Digenis et al. 1998; Trengove, Stacey et al. 1999;
Cullen, Watt et al. 2002).
13.
There are three pathways that explain how the excessive levels of proteases contribute
to the chronic ulcer pathology. Firstly, in CWF, growth factors have been shown to
lack biological activity due to proteolytic degradation. Next, growth factor receptors
are also degraded in chronic wounds; thereby decreasing the cell’s ability to recognise
and use growth factors. Lastly, the provisional matrix needed for fibroblasts to migrate
across the ulcer wound bed is degraded by the proteolytic chronic wound environment
(Fray, Dickinson et al. 2003). Therefore, many authors have suggested that the
addition of a protease inhibitor prior to topical treatment of the wound would promote
healing (Tarnuzzer and Schultz 1996; Wysocki, Kusakabe et al. 1999; Yager and
Nwomeh 1999; Cullen, Smith et al. 2002).
There are a number of both biological and chemical inhibitors of MMPs (Massova,
Kotra et al. 1998). Intrinsic MMP inhibitors, TIMPs, are biological inhibitors of MMP
activity commonly found in wound sites (Klein, Anderson et al. 2002). MMPs mainly
form binary non-covalent complexes when inhibited by the TIMPs (Herouy, Trefzer et
al. 2000). In terms of chemical inhibitors, cation chelators, i.e.
ethylenediaminetetraacetic acid (EDTA), have proven successful in in vitro studies
examining growth factor degradation in CWF as they remove the zinc ion required for
MMP catalytic activity (Cullen, Smith et al. 2002). More specific inhibitors include
the following: tetracyclines and their chemically modified derivatives, i.e. doxycycline
(Ramamurthy, Kucine et al. 1998; Ramamurthy, McClain et al. 1999; Yager and
Nwomeh 1999); hydroxamic acids, i.e. (N-[(2R)-2-(hydrox- amidocarbonylmethyl)-4-
methylpentanoyl]-L-tryptophan methylamide), Ilomastat, galardin, GM6001 (Lund,
Romer et al. 1999); and bisphosphonates, i.e. clodronate (Valleala, Hanemaaijer et al.
2003). From these groups, the tetracyclines and their chemically modified derivatives
tend to develop a number of unwanted side-effects, e.g. gastrointestinal disturbances
and the possibility of potential antibiotic resistance (Teronen, Heikkila et al. 1999).
The bisphosphonate family on the other hand, has been shown to exhibit low toxicity
and have been tolerated in human use for several years, hence bisphosphonates seem
to be ideal candidates for MMP-related diseases (Teronen, Heikkila et al. 1999). From
this information, it seems a logical progression for bisphosphonates to be explored as
potential inhibitors of MMPs in chronic ulcers to promote the progression towards a
healing state.
14.
1.2.2.1 Bisphosphonates
Bisphosphonates, sometimes known incorrectly as diphosphonates, are stable
analogues of pyrophosphate characterised by two C―P bonds (Figure 1.3) (Fleisch
1997; Neville-Webbe, Holen et al. 2002). When the two bonds are found on the same
carbon atom, giving a P―C―P structure, they are called geminal bisphosphonates.
However, as only the geminal bisphosphonates have been found to exert strong
clinical activity they are commonly referred to in the literature as simply
“bisphosphonates” (Fleisch 1997). The main difference between bisphosphonates and
pyrophosphates is that pyrophosphate can be cleaved by enzymatic hydrolysis, but the
P―C―P bond found in bisphosphonates is hydrolysis-resistant (Vasikaran 2001).
Figure 1.3 Structure of bisphosphonates (a) General structure of bisphosphonate and pyrophosphate (b) First generation bisphosphonates (c) Second generation bisphosphonates (d) Third generation bisphosphonates (Heymann, Ory et al. 2004)
The bisphosphonates can be divided into two groups with different modes of action –
those that closely resemble pyrophosphate, i.e. clodronate and etidronate; and those
that contain nitrogen, i.e. pamidronate, alendronate, risedronate and ibandronate. The
first group are able to be metabolically incorporated into nonhydrolysable analogues
of ATP that can inhibit ATP-dependent intracellular enzymes. The second more
potent group of nitrogen-containing bisphosphonates do not act in this manner, but
can instead inhibit enzymes of the mevalonate pathway, which can then inhibit the
synthesis of isoprenoid compounds essential for the post-translational modification of
small GTPases (Russell and Rogers 1999). In addition to this grouping,
bisphosphonates can be further divided into first-, second- and third-generation
compounds. The first-generation bisphosphonates have only short alkyl, halide or
hydroxyl side chains, giving them the lowest activity. Second-generation
bisphosphonates, with the one exception of tiludronate, have a nitrogen atom in their
side chain. Finally, the third-generation bisphosphonates, which are by far the most
potent, have a heterocyclic side chain containing nitrogen (Figure 1.3, Table 1.3)
(Heymann, Ory et al. 2004).
Table 1.3 Relative potencies of bisphosphonates in vitro (Vasikaran 2001)
Bisphosphonates have been shown to inhibit MMP-1, -2, -3, -8, -9, -12, -13 and -20 at
both therapeutically obtainable and non-cytotoxic concentrations (Heikkila, Teronen
et al. 2002). For example, alendronate has an IC50 value, i.e. the concentration of
alendronate required to inhibit 50% of MMP activity, of 40-70 µM (Heikkila, Teronen
et al. 2002). Currently, this group of drugs is mainly used in the management of
calcium and bone metabolism disorders (Vasikaran 2001), e.g. osteoporosis, Paget’s
disease, hypercalcaemia and metastatic cancer. These diseases have been effectively
15.
treated with bisphosphonates for a number of years (Table 1.4) (Russell and Rogers
1999; Teronen, Heikkila et al. 1999). However, future indications may include those
that are primarily caused by significant soft tissue destruction (Teronen, Heikkila et al.
1999), such as chronic wounds – an application that has not been previously
mentioned in the current literature.
Table 1.4 Bisphosphonates in clinical studies/development (Russell and Rogers 1999)
In terms of metastases, there have been a number of advances involving the
bisphosphonate-induced inhibition of MMPs. In early studies of patients with breast
cancer, clodronate significantly reduced the incidence of skeletal and visceral
metastases and gave a survival advantage. However, two other similar studies
produced completely different results, with no impact on visceral metastases
(Heymann, Ory et al. 2004). Another early bisphosphonate, alendronate, has proven to
be a potent inhibitor of bone resorption. It was shown that the use of alendronate to
prevent MMP-2 secretion blocks the solubilisation of collagen in metastatic human
prostate PC-3 ML cells (Stearns 1998). One important point to note, however, is that
these trials were performed using first- and second-generation bisphosphonates, the
activities of which vary from their third-generation counterparts (Heymann, Ory et al.
2004). Keeping this in mind, zoledronic acid, a third-generation bisphosphonate, has
been shown to be a potent inhibitor of both prostatic epithelial cells and aggressive
16.
17.
breast cancer cells, preventing invasion and colony formation (Denoyelle, Hong et al.
2003; Montague, Hart et al. 2004). The mechanism responsible for this inhibition is
not dependent on protease modifications, but instead uses a disorganisation of the
actin skeleton due to a RhoA inhibition (Denoyelle, Hong et al. 2003).
These results show that while bisphosphonates appear to be good candidates for
clinical use, much more research still needs to be done to clarify their mechanism of
action and disease specificity. Furthermore, in terms of chronic ulcer treatments, the
delivery, or indeed simply the presentation of bisphosphonates to the site of excessive
MMP activity, has to be carefully considered. For this purpose, various polymers will
be considered for their potential ability to act as a scaffold to present this class of
molecules to the chronic ulcer.
18.
1.3 POLYMERS SUITABLE FOR CONTROLLED DRUG RELEASE IN A
WOUND DRESSING
Controlled drug release via polymers is currently widely studied, with many different
approaches depending on the material and drug of interest. Even more challenging, is
the prospect of manufacturing smart materials that release drugs by responding to a
patient’s individual therapeutic requirements (He, Cao et al. 2004). These delivery
systems would require complete drug protection, local targeting, specific controlled
release, self-regulating action, enzyme inhibition and reporting (He, Cao et al. 2004).
In addition, the biomaterial should be non-immunogenic and have flexible
modification strategies (Schmedlen, Masters et al. 2002; Leach, Bivens et al. 2003;
Ueda and Tabata 2003).
Traditional synthetic polymers include poly(ethylene glycol) (PEG), poly(2-
hydroxyethyl methacrylate) (PHEMA), poly(methyl acrylate) (PMA), poly(glycolic
acid) (PGA), poly(lactic acid) (PLA), and copolymers (PLGA) of these materials
(Hoffman 1998; Lee and Mooney 2001; Smeds and Grinstaff 2001). The major
advantages of these materials are that they can be easily synthesised and factors such
as mechanical, physical and chemical properties can be manipulated (Table 1.5)
(Holland, Tabata et al. 2003). Some limitations of synthetic polymers include:
requirement for inclusion on the Food and Drug Administration (FDA) ‘Generally
Regarded as Safe’ (GRAS) list; inadequate biodegradability; and, many are often
difficult to sterilise (Hoffman 1998).
Current commercially available wound dressings include a number of different types
of polymers, e.g. foams (Lyofoam, ConvaTec), hydrocolloids (Tegasorb, 3M),
hydrogels (IntraSite gel, Smith & Nephew), films (OpSite, Smith & Nephew),
polyurethanes (Allevyn, Smith & Nephew), alginates (Kaltostat, ConvaTec), activated
charcoal (CarboFlex, ConvaTec) and collagen dressings (Promogran, Johnson &
Johnson) (Queen, Orsted et al. 2004; Bouza, Muñoz et al. 2005). The use of these so-
called “modern” dressings has rapidly advanced in the last twenty years, with a large
economic impact on healthcare systems around the world (Cuzzell 1997). While these
“modern” dressings have improved wound treatment, they do not always completely
cover the primary concerns of local wound healing. These include: controlling wound
19.
Material Properties Semicrystalline polymer degraded by micro-organismsHigh solubility and low melting point make it an attractive biomaterial Degrades at a slow pace
Poly(caprolactone) (PCL)
Non-toxic and tissue compatible Highly biocompatible Exhibits versatile physical properties Poly(ethylene glycol) (PEG) Previously used as a porogen in hydrogels
Highly crystalline - high melting point and low solubility in organic solvents
Degradation to non-toxic monomers upon exposure to water Poly(glycolic acid) (PGA)
Tendency to lose mechanical strength rapidly Hydrophobic polymer, less hydrolysis and greater stability than PGA, Led to the development of PGA/PLA copolymers Poly(lactic acid) (PLA)
Choice of application based on crystallinity Combines the advantages of PGA and PLA Biodegradable and biocompatible
Poly(lactic-co-glycolic acid) (PLGA)
Low toxicity Water content similar to living tissue Inert to normal biological processes Shows resistance to degradation Permeable to metabolites Low thrombogenicity Not absorbed by the body Withstands heat sterilization without damage
Poly(2-hydroxyethyl methacrylate) (PHEMA)
Can be prepared in various shapes and sizes Biodegradable and bioabsorbable Good mechanical strength Poly(methylacrylate) (PMA) High tensile strength Good film forming properties High hydrophilicty Poly(vinyl alcohol) (PVA) Biocompatibility and chemical resistance
Table 1.5 Description of various polymers available for use in biomedical applications and their properties. Adapted from various sources (Hubbell 1996; Ramakrishna, Mayer et al. 2001; Hoffman 2002; Shao, Kim et al. 2003).
20.
infection and prolonged inflammation, while still promoting moist interactive wound
healing (Queen, Orsted et al. 2004). In addition, importance needs to be placed on
patient compliance, pain management and quality of life (Queen, Orsted et al. 2004).
Hydrogels are often selected as an ideal wound dressing as they are cool, do not stick
to the wound itself, and minimise pain to the patient (Queen, Orsted et al. 2004; Akita,
Akino et al. 2006). There is also the potential for novel biomaterials to be further
developed as wound dressings, e.g. alginates, but in their current form they are
unsuitable as completely defined wound dressings.
1.3.1 Hydrogels
In recent years, the polymeric structure widely investigated for use in tissue
engineering and controlled release is the hydrogel (Lee and Mooney 2001; Hoffman
2002; Drury and Mooney 2003). Hydrogels are hydrophilic polymer networks which
can absorb from 10-20% up to thousands of times their dry weight in water (Hoffman
2002). They can either be chemically stable or have the ability to degrade and
ultimately disintegrate, a fact that is often exploited in drug delivery (Hoffman 2002).
The structure of hydrogels is often determined by crosslinks formed between polymer
chains via various chemical bonds and physical interactions, e.g. H-bonding and
hydrophobic forces (Drury and Mooney 2003). Hydrogels also have the ability to be
fabricated in various forms, including: solid moulded forms; pressed powder matrices;
microparticles; coatings; membranes or sheets; encapsulated solids; and liquids
(Hoffman 2002). A major advantage with hydrogels is that they closely resemble the
chemical structure of the ECM, which can prove useful for applications in rapid
diffusion of hydrophilic nutrients and metabolites (Landers, Hubner et al. 2002).
There are many options for synthetic materials suitable for use in hydrogels (Table
1.6). These include, but are not restricted to, the following: poly(ethylene oxide)
(PEO); poly(vinyl alcohol) (PVA); poly(acrylic acid) (PAA); and poly(propylene
fumarate-co-ethylene glycol) (P(PF-co-EG)). In terms of naturally derived polymers,
there are also many options, e.g. agarose; alginate; chitosan; collagen; fibrin and
gelatin (Drury and Mooney 2003). However, as these matrices are derived from
animal and plant products that are non-defined, problems with sterility and
transmissible disease have to be considered (Drury and Mooney 2003; Tomizawa
2005). While plant-derived materials may not carry the risk of transmitting diseases,
they are still generally undefined and are often difficult to sterilise effectively before
application to the patient (Drury and Mooney 2003). In addition, some animal-derived
products are unable to be used for medical applications as the risk of disease, e.g.
Bovine Spongiform Encephalopathy (BSE), while low, is still present (Grassi, Farra et
al. 2005; Tomizawa 2005).
Table 1.6 Hydrophilic polymers suitable for hydrogel matrices (Hoffman 2002)
Recent work has suggested that it is possible to achieve controlled drug and protein
release from synthetic polymers (Hubbell 1996; Jeong, Choi et al. 1999; He, Cao et al.
2004). PEG hydrogels (molecular mass 10,000 Da) have been demonstrated to release
proteins from 6,000 to 66,000 Da while still displaying controlled release. This study
also showed that small proteins were released by diffusion without requiring
degradation, whereas larger proteins needed degradation first and then diffusion (West
and Hubbell 1995; Hubbell 1996). In cases of release without degradation, the rate of
permeation of the protein through the gel was inversely and linearly related to the
molecular mass of the protein. Of importance, the photopolymerisation process used
21.
22.
to form the hydrogel has been shown to not adversely affect the incorporated protein’s
activity (West and Hubbell 1995).
In cases where the protein release is Fickian in nature, the diffusion kinetics are
governed by the rate of polymer swelling (Tomic, Micic et al. 2007). More
specifically, the protein release is controlled by the its ability to migrate through the
polymer scaffold, or even by diffusion through micropores of the polymer matrix
(Tomic, Micic et al. 2007). Protein diffusion from monolithic systems can be analysed
using Fick’s second law of diffusion:
αφ = D α2φ
αt αx2
Where: φ is the concentration in dimensions of [(amount of substance) length-3]
t is time [s]
D is the diffusion coefficient in dimensions of [length2 time-1]
x is the position [length] (Fick 1855)
The models of controlled release by diffusion are based on the principle of the
permeability of the polymeric matrix following swelling (Tomic, Micic et al. 2007).
Copolymers of PEG (Figure 1.4a) and 2-hydroxylethyl methacrylate (HEMA) (Figure
1.4b) have also been investigated for use as a hydrogel for protein release (Carenza,
Lora et al. 1993).
(a)
(b)
Figure 1.4 Chemical structures (a) PEG (b) HEMA
H2C
CH3
O
OOH
OH CH2
CH2
O CH2
CH2
OHn
23.
These two materials are advantageous for use in biomaterial applications for various
reasons (Hoffman 2002), including the fact that PHEMA has been used for a number
of years in the production of soft contact lenses and intraocular lenses, as it exhibits
high biocompatibility and low thrombogenecity (Caliceti, Veronese et al. 1992;
Carenza and Veronese 1994). In addition, PHEMA has good strength characteristics
for these applications, as well as the imbibed water content properties desirable in
hydrogels (George, Wentrup-Byrne et al. 2004). PEG is also highly biocompatible and
exhibits versatile physical properties, which are highly dependent on its weight
percentage, molecular chain length and crosslinking density (Almany and Seliktar
2005).
Furthermore, gamma irradiation is a very convenient tool for polymerisation when
considering hydrogels (Park, Nho et al. 2004), and has previously been used to form
PHEMA hydrogels, with the mechanism described in Figure 1.4. Unlike conventional
chemical initiation methods, including photo-polymerisation, radiation polymerisation
does not need the addition of catalysts or additives and also gives good control over
the degree of crosslinking (Park, Nho et al. 2004). By using gamma irradiation,
Carenza et al. claimed to crosslink these two materials, PEG and the HEMA
monomer, and obtained a complete polymerization conversion at a total dose of 25
kGy (Carenza, Lora et al. 1993), potentially using the mechanism described in Figure
1.5. Therefore, the crosslinking of polymers is extremely important when examining
their physical properties.
1.3.2 Crosslinking of polymers
Crosslinked hydrogels based on polymers or copolymers of HEMA have a long
history of clinical use in contact lenses and drug delivery systems. Primarily this is
because crosslinked hydrogels are inert, insoluble and easy to manufacture – and
therefore cannot become absorbed into the body like their water-soluble linear
analogues (Zahedi and Lee 2007). There are a number of ways to crosslink a polymer
namely, irreversible chemical crosslinking, crosslinking through ionic interactions,
electron bombardment, or gamma irradiation (Berger, Reist et al. 2004; Rouif 2005)
24.
Figure 1.5 Mechanism of gamma induced polymerisation of HEMA
CH2 C
CH3
C O
O
CH2
CH2
OH
CH2
C CH3
C O
O
CH2
CH2
OH
I
CH2
C CH3
C O
O
CH2
CH2
OH
CH2 C
CH3
C O
O
CH2
CH2
OH
I
C O
O
CH2
CH2
OH
CCH2
CH3
CH2
C CH3
C O
O
CH2
CH2
OH
C O
O
CH2
CH2
OH
C CH2
CH3
CO
O
CH2
CH2
OH
C CH2
CH3
C O
O
CH2
CH2
OH
CCH2
CH3
C
CH3
CO
O
CH2
CH2
OH
I I
Inititation
Propagation
+
Termination
.
γ
25.
Figure 1.6 Potential mechanisms for gamma induced random grafting/
termination or polymerisation of HEMA and PEG
C O
O
CH2
CH2
OH
C CH2
CH3
O PEG OH
C O
O
CH2
CH2
OH
CCH2
CH3
O PEG OH
CO
O
CH2
CH2
OH
C CH2
CH3
C O
O
CH2
CH2
OH
CCH2
CH3
O PEG O CH2 CO
O
CH2
CH2
OH
C
CH3
O
C O
O
CH2
CH2
OH
C
CH3
O PEG CH2
.
Random grafting / Termination
.
PEG copolymerisation
γ
26.
To synthesise a chemically crosslinked hydrogel, more than two functional sites are
required in each polymer chain with at least one reaction partner (Rouif 2005). This
then creates chains of polymers interconnected by crosslinkers, i.e. difunctional
crosslinking agents, resulting in the creation of a three-dimensional network (Berger,
Reist et al. 2004). However, as with any crosslinking agent, residual amounts are
always present in the ensuing polymer, often causing problems with biocompatibility
(Hassan and Peppas 2000). In contrast, crosslinking using irradiative methods
removes the need for potentially toxic additives, e.g. residual chemical initiators,
allowing a smoother transition to commercial use for products for biomedical
applications (Hassan and Peppas 2000).
It is also important to note that when irradiating aqueous polymer solutions, the
radiolysis properties of water need to be considered. Primarily, gamma irradiation of
aqueous solutions produces oxygen radicals. Free radicals, e.g. .OH and HO2.
(Leguéné, Clavère et al. 2001), are typically scavenged by the solutions during
irradiation (Namkyu, Seunghee et al. 2004). Furthermore, molecular products (H2O2,
H2) are formed by the combination of these radicals (e–aq, H., OH.) in the spurs
(Wasselin-Trupin, Baldacchino et al. 2002). The mechanism of water radiolysis is
quite complex and involves several other reactions, as shown by Table 1.7 (Pastina,
Isabey et al. 1999). In addition, these reactions are influenced by the radiation
characteristics, the temperature, and the chemical composition of the water (Pastina,
Isabey et al. 1999). For these reasons, it is very difficult to foresee the exact effect of
these parameters.
The radiolysis of water forms both molecular (H2, O2, H2O2) and radical (H, OH, e−aq,
HO2) species as shown by the following equation:
H2O H2, H2O2, .OH, H., e−aq, HO2, H3O
+ (Pastina, Isabey et al. 1999)
In terms of HEMA specifically, the monomer is synthesised through either the
transesterification of ethylene glycol, or by reacting ethylene oxide and methacrylic
acid (Montheard, Chatzopoulos et al. 1992). Both of these methods can cause
impurities in the resulting monomer, e.g. methacrylic acid from the hydrolysis of
HEMA, and ethylene glycol dimethacrylate (EGDMA) from the esterification
between methacrylic acid or HEMA and ethylene glycol (Montheard, Chatzopoulos et
al. 1992). As EGDMA is a difunctional crosslinking agent, the resulting copolymer
hydrogels (Figure 1.6) are crosslinked and swell in water (instead of dissolving) and,
even at very low percentages of the impurity (Montheard, Chatzopoulos et al. 1992).
Table 1.7 The reaction scheme of the radiolysis of neutral water. A, Arrhenius factor frequencies (A in moles-1 dm3 s-1); Ea, activation energies (Ea in kJ moles-1); Rate constants (moles-1 dm3 s-1) (Pastina, Isabey et al. 1999).
1.3.3 Functionalisation of polymers with bisphosphonates
The functionalisation of polymers has been widely used for a number of years to
improve various properties of the original polymer, e.g. cell adhesion (Buchmeiser
2007). Theoretically, functionalisation can be accomplished via three different
methods: copolymerisation of functional monomers; functionalising the material post-
polymerisation, commonly referred to as grafting; and imprinting (Buchmeiser 2007).
In terms of modifying bisphosphonates for clinical delivery, a number of bone
27.
28.
Figure 1.7 Mechanism of gamma induced crosslinking of HEMA and
EDGMA
CH2
C CH3
R
CH2 C
CH3
C
O
CH2
CH2
O
O
C
C
CH3
CH2
O
CCH2
CH3
R
CH2
C
CH3
C
O
CH2
CH2
O
O
C
C
CH3
CH2
O
C O
O
CH2
CH2
OH
CCH2
CH3
R
CH2
C
CH3
C
O
CH2
CH2
O
O
C O
C
CH3
CH2
R =
EDGMA crosslinking
29.
specific applications have been developed. These include: conjugation of an amino-
bisphosphonate to bovine serum albumin to improve delivery to mineralised tissue
(Uludag and Yang 2002); copolymerisation of alendronate with PEG to act as a bone-
targeting polymeric drug delivery system (Wang, Miller et al. 2003); and, loading of
specific bisphosphonates into siliceous ordered mesoporous materials to enhance bone
implants (Balas, Manzano et al. 2006). Furthermore, alendronate is a useful model
bisphosphonate as it possesses a primary amine, which is convenient for conjugation
(Wang, Miller et al. 2003) using a number of already-established chemical reactions.
Nucleophilic acyl substitution reactions describe the substitution reaction of
nucleophiles, e.g. amines, and acyl compounds, e.g. acid halides (Otera 1993). In the
case of amines, they form a condensation reaction to then create amides. Therefore, it
seems a logical next step that this technique can be used to functionalise the
bisphosphonate, alendronate, to allow copolymerisation into a hydrogel.
30.
1.4 PROJECT OUTLINE
1.4.1 Research Problem
As outlined above, the issue of chronic non-healing ulcers is a major medical concern.
They are not only a major cause of pain and anxiety for those aged over sixty, but also
largely contribute to their lessening mobility, decreased social interactions and overall
diminished quality of life. Previous literature has shown that the chronic wound
environment is a very complex and toxic place for new cells to grow (Chen, Schultz et
al. 1999). Furthermore, many authors have suggested that inhibition of these proteases
is critical to enable correct progression of healing for these ulcers (Tarnuzzer and
Schultz 1996; Wysocki, Kusakabe et al. 1999; Yager and Nwomeh 1999; Cullen,
Smith et al. 2002). However, there is other evidence suggesting that levels of
proteases in the wound bed itself vary greatly from those levels found in the CWF
(Cook, Stephens et al. 2000). Therefore, it appears that it is important to inhibit the
proteases present in CWF, while not affecting those that show decreased activity in
the wound bed, to provide an effective treatment for these compromised wounds.
Once that inhibition has taken place, it is then possible to either use the body’s own
growth factors to heal the wound, or to release a synthetic growth factor complex, e.g.
VitroGro®, comprised of vitronectin, insulin-like growth factor and an insulin-like
growth factor binding protein, into the neutralised environment. This approach should
also allow for a more cost-effective solution, as minimal amounts of expensive growth
factors would be used, if at all, in the dressing. Therefore, this wound dressing would
be available as a viable treatment option to the majority of chronic leg ulcers sufferers,
regardless of their socioeconomic position. In conclusion, by treating this problem
quickly and effectively, improved holistic patient outcomes will be achieved.
1.4.2 Hypothesis
The overall goal of this applied research project was to examine the following
hypothesis:
Synthesis of a bioactive wound dressing that acts by inhibiting proteases in the CWF
will lead to the development of a therapy that can more effectively heal the
compromised ulcer.
31.
1.4.3 Aims
Thus, the specific aims of this project were to:
Aim 1: Use collagen zymography and a specific enzyme-linked immunosorbant assay
to analyse the increased activity of proteases in CWF, and then use this information to
identify the specific proteases responsible for the majority of collagen degradation in
the wound bed;
Aim 2: Identify a suitable synthetic and biocompatible polymer for the controlled
release of a biologically active compound into a chronic wound site; and
Aim 3: Apply both of these findings to the generation of a bioactive wound dressing
that would inhibit proteases in the wound fluid with a protease inhibitor tethered to the
polymer. This will then potentially still allow those proteases in the wound bed to
perform their vital functions in wound healing.
1.4.4 Project Design
This project has been designed to address both the biological and chemical issues that
need to be considered in the development of a bioactive wound dressing for chronic
leg ulcers. In addition, this project has deliberately blended the two disciplines of
wound healing biology and polymer chemistry to create a thesis that is
interdisciplinary in nature. Importantly, the key focus of this project was centred on
the development toward a tangible product as the key outcome, rather than a project
centred on generation of fundamental new mechanistic knowledge.
32.
CHAPTER 2
ANALYSIS OF MATRIX METALLOPROTEINASES
IN CHRONIC WOUND FLUID
34.
2.1 INTRODUCTION
As outlined in Chapter 1, wound healing is a highly complex process that can lead to
chronic ulcers if any stage in this complex series of events is interrupted (Trengove,
Stacey et al. 1999). While there is no single reported cause of these ulcers, many
factors are considered important when distinguishing chronic wounds from their acute
counterparts. Primarily, these wounds are characterised by excessive granulation
tissue and increased fibrosis (Stadelmann, Digenis et al. 1998), along with their
significant increase in proteolytic activity (Tarnuzzer and Schultz 1996; Ladwig,
Robson et al. 2002; Li and Li 2003). While it is understood that there are many
potential causes of these wounds, for the purposes of this thesis, the role of proteolytic
degradation and its relationship to the wound’s chronicity is the core focus. It is this
protease activity, primarily caused by a specific group of proteases called matrix
metalloproteinases (MMPs), that is believed to be responsible for the increased
extracellular matrix destruction in these chronic wounds (Chen, Schultz et al. 1999).
However, there are large amounts of conflicting information in the literature
concerning both the under- and over-expression of specific MMPs in chronic wounds.
Therefore, the purpose of this work was to clarify existing discrepancies in the
literature, and use multiple techniques to confirm individual MMP presence and
activity.
Gelatinases, also referred to as MMP-2 (gelatinase-A) and MMP-9 (gelatinase-B), are
thought to have an extremely important function in normal wound healing during both
the remodelling and re-epithelialisation phases, and would therefore impair wound
healing if inhibited (Fray, Dickinson et al. 2003). However, other reports have
suggested that both MMP-2 and MMP-9, when present at the significantly higher
activity levels as reported in these chronic wounds, need to be inhibited to allow
normal healing to occur (Wysocki, Kusakabe et al. 1999; Ladwig, Robson et al. 2002).
The main reason why gelatinases are so important is that they are able to break down
collagen more effectively than other MMPs (Parks 1999), as well as cleave the major
constituent of the basement membrane – collagen type IV (Tarnuzzer and Schultz
1996). In a comprehensive study, MMP-9 was shown to be able to cleave collagen
types I-V into small peptides at the physiological temperature of 37 ºC (Okada, Gonoji
et al. 1992). In addition, MMP-2 and MMP-9 differ from other MMPs in that they
35.
contain three tandem fibronectin type II repeats within the amino-terminus of the
catalytic domain, which allows for binding to gelatin (Somerville, Oblander et al.
2003). Furthermore, MMP-9 contains an additional Collagen Type V-like insert in its
hinge region, with its function still to be determined (Somerville, Oblander et al.
2003).
MMP-9 is produced by a number of different cell types, namely inflammatory cells,
neutrophils, macrophages, monocytes and keratinocytes (Wysocki, Kusakabe et al.
1999; Cullen, Smith et al. 2002). As the chronic ulcers are classified as nonhealing,
the MMP-9 present in CWF is thought to be either neutrophil- or macrophage-derived
(Cullen, Smith et al. 2002). In terms of MMP-9 expression in chronic ulcers, one
study demonstrated that MMP-9 is temporally expressed during the healing of chronic
wounds (Wysocki, Kusakabe et al. 1999). This also correlates with another report that
showed MMP-9 activity levels in CWF were 25-fold elevated over a normal control
(Yager, Zhang et al. 1996). In contrast, Cullen et al. (2002) reported a lack of MMP-2
in CWF, which, as MMP-2 is fibroblast-derived, suggests that there must a lack of
fibroblasts in these chronic wounds, which may be associated with a lack of wound
closure (Cullen, Smith et al. 2002).
The ratio of MMPs and their natural tissue inhibitors, TIMPs, are also thought to be
important in the normal wound healing cascade (Cook, Stephens et al. 2000). Indeed,
a recent study, Ladwig et al. (2002) report that the ratio of MMP-9 to TIMP-1 is
inversely correlated with pressure ulcer healing (Ladwig, Robson et al. 2002). While
this study was performed in a different ulcer aetiology, there are still some similarities
between pressure ulcers and venous ulcers. Furthermore, it is interesting to note that
Cook et al. reported in 2000 that significantly decreased levels of active MMPs, along
with marked increases in expression of TIMPs, led to the chronicity of non-healing
wounds (Cook, Stephens et al. 2000) – directly contrasting all previous reports. The
most probable explanation for this is that Cook et al. (2000) used biopsies from the
wound bed, instead of wound fluids as used by other studies, including the one
reported in this chapter (Cook, Stephens et al. 2000). However, recent reports have
suggested that the MMPs present in the chronic wound bed may also be involved in
important biological processes, such as growth factor activation and immune system
regulation, rather than simply matrix and growth factor degradation (Levi, Fridman et
36.
al. 1996; Suzuki, Raab et al. 1997; McQuibban, Butler et al. 2001; Gearing, Thorpe et
al. 2002). Therefore, other MMP-interactions in the wound bed also need to be
considered when discussing chronic wounds (Somerville, Oblander et al. 2003).
At this time, there have been no studies that directly compare wound biopsies and
wound fluids, so therefore it is not know if they are indeed similar. Furthermore, it is
still unknown as to which sample provides a better indicator of the chronic wound’s
status (Ladwig, Robson et al. 2002). For the study reported herein, CWF was chosen
as the clinical sample to provide insight into the status of the chronic wound
environment, as the work described in this thesis is focussed on wound dressing
applications that could potentially modulate the proteolytic environment in CWF.
CWF is reported to be derived from plasma and, in a healing situation, delivers all of
the necessary components to the wound bed (Cutting 2003). Trengrove et al. (1996)
also describe similarities between human serum and CWF, indicating that CWF
displays the biochemical components expected in an extracellular fluid (Trengove,
Langton et al. 1996). Therefore, the aims of the study reported in this chapter were to:
identify proteolytic differences between AWF and CWF; to relate these to specific
proteases; and then in turn, identify protease inhibitors that would specifically inhibit
the major proteases with the goal of restoring the chronic wound towards a healing
state.
37.
2.2 MATERIALS AND METHODS
2.2.1 Wound fluid sample collection and preparation
Chronic wound fluid (CWF) samples were obtained from consenting patients of the
St. Luke’s Nursing Services (Brisbane, QLD, Australia), suffering from chronic
venous ulcers undergoing compression therapy (Table 2.1). Ethical approval to collect
these samples was obtained from both the Queensland University of Technology
(QUT) and St. Luke’s Nursing Services. No specific instructions were given to
patients prior to collection, e.g. no fasting requirements. A standard wound fluid
collection technique has been established and carried out at the clinical site. Briefly,
ulcers were washed with sterile water prior to collecting wound fluid, followed by the
application of an occlusive dressing over the wound. Exudate accumulated under the
dressing after 30 min to 1 hour was recovered by washing with 1 mL of saline.
Patient Age Sex Sample Code
Ulcer Duration (Weeks)
Ulcer Size (cm²)
% Reduction in wound area (from baseline)
Compression (mmHg)
PUSH Score
1 54 F 33 102 44 0 <10 15 2 79 F 34 102 15 -25 30 - 35 12 3 76 F 57 162 84 64 30 - 35 15 4 79 F 89 261 16 0 30 - 40 14 5 88 M 95 320 10.7 83.7 30 - 35 13 6 67 F 96 37 16 0 30 - 40 10 7 77 M 131 550 8 0 20 - 25 10 8 86 F 142 100 7.1 0 30 - 35 9 9 75 M 173 20 7 0 30 - 40 10
Table 2.1. Patient clinical data from chronic wound fluid samples.
Acute wound fluid (AWF) samples were collected from naturally occurring sub-
epidermal blisters on the feet of volunteers. The fluid was removed with 26G x 0.5”
needle and syringes (Terumo Medical, Somerset, NJ, USA) and collected in 1.5 mL
Protein LoBind tubes (Eppendorf, Hamburg, Germany). The wound fluid samples
were centrifuged at 14,000 g for 10 min, then the supernatant filtered using 0.45 µm
cellulose acetate filters (Agilent Technologies, Wilmington, DE, USA). This wound
fluid collection technique was chosen as being the most suitable approach, primarily
38.
because it does not rely on the absorption of wound fluid on to a porous, hydrophilic
wound dressing. In particular, the use of absorption techniques has been reported to
result in lower sample volume and protein amount (Moseley, Hilton et al. 2004). In
addition, other techniques that rely on this absorption approach do not take into
account that proteins may differentially adhere to these dressings, which will then
influence the CWF sample in later analyses (Cook, Stephens et al. 2000). A pooled
human serum (HS) sample was purchased commercially from Sigma-Aldrich (St
Louis, MO, USA). Protein content for all samples was quantitated and standardised
using the BCA Protein assay kit (Pierce Biotechnology, Rockford, IL, USA). The
samples were then sub-aliquoted and stored at -80 ºC until further analysis.
2.2.2 Identification of protease activity using collagen zymography
Collagen zymography was performed as previously described (Gogly, Groult et al.
1998) using Collagen Type I and Collagen Type IV (Sigma-Aldrich) at a final
concentration of 0.5 mg/mL in 10% total acrylamide gels under non-reducing
conditions. MMP-8 (Calbiochem, San Diego, CA, USA) and MMP-9 (Calbiochem)
were run alongside the fluid samples as standards. Briefly, electrophoresis was
performed at 4 ºC under Laemmli conditions (Laemmli 1970). The gels were then
washed in 2.5% Triton X-100 for 30 minutes, then a further 60 minutes, prior to
incubation in 50 mM Tris-HCl, 10 mM CaCl2, 50 mM NaCl at pH 7.6 for 24 hours at
37 ºC. The gels were then stained using 0.25% Coomassie brilliant blue R-250 (Bio-
rad Laboratories, Hercules, CA, USA) (40% methanol, 10% acetic acid) and destained
appropriately (40% methanol, 10% acetic acid). Protease activity was visualised as
clear (unstained) bands.
2.2.3 Analysis and quantitation of collagen zymography
Gels were scanned using GeneSnap version 6.07 (SynGene, Cambridge, UK) and
analysed using GeneTools version 3.07 to determine the molecular weight and
quantitate the clear bands (SynGene). All quantitations were performed using the
analysis of 3 separate gels per treatment and values were expressed as a relative % of
the MMP standard, similar to previously published reports (Cook, Stephens et al.
2000; Lerman, Galiano et al. 2003). The data was also analysed in terms of grouping
39.
the samples based on the “PUSH” score of the wounds from which the CWF was
collected. Clinicians use the PUSH (Pressure Ulcer Scale of Healing) score to grade
the wounds (1 = best, 17 = worst) with this value taking into account a number of
factors to score the ulcer (Ratliff and Rodeheaver 2005). While this measurement
technique was originally designed by the National Pressure Ulcer Advisory Panel to
grade pressure ulcers, it has also been validated as an effective tool to monitor healing
trends in venous ulcers (Ratliff and Rodeheaver 2005). Thus, Ratliff et al. (2005)
noted that this system was a simple, valid and reliable tool for monitoring venous
ulcer clinical status (Ratliff and Rodeheaver 2005). In view of this, the data in our
study was organised in terms of this PUSH score, with five samples (CWF-1, -2, -3, -
4, -5) exhibiting a high PUSH score (≥12) and four samples (CWF-6, -7, -8, -9)
showing a lower PUSH score (≤11). While we ideally would have preferred to assess
a greater number of samples, CWF sample numbers were restricted due to patient
availability, as well as the amount of protein required from individual samples to
complete the full spectrum of experiments reported here. Statistical analysis was
performed using Tukey’s test (all group’s comparison), with statistical significance
determined as either * (p<0.05) or # (p<0.01).
2.2.4 Confirmation of MMP-specific collagen degradation using GM6001
inhibitor
To confirm that the collagen degrading activity visualised through zymography was
due to MMPs, a specific MMP inhibitor, GM6001 (Ilomastat or N-[(2R)-2-(hydrox-
amidocarbonylmethyl)-4-methylpentanoyl]-L-tryptophan methylamide) (Chemicon,
Temecula, CA, USA) was used. Increasing concentrations of GM6001 (1, 3, 10, 30,
100, 300 and 1000 µM) were incubated with CWF-1 samples at 37 ºC for 24 hours.
These samples were then run on Collagen Type I zymograms and analysed as
described above.
2.2.5 Specific identification of MMPs through immunoprecipitation
Further confirmation of MMP activity was established through the
immunoprecipitation of specific MMPs followed by collagen zymography. UltraLink
Immobilized Protein A/G resin (Pierce Biotechnology) was incubated with individual
40.
MMP antibodies in a Tris-buffered saline solution with 0.1% Tween-20 (TBST) and
rotated for 5 hours at 4 ºC in 0.5 mL Protein LoBind tubes (Eppendorf). The
antibodies were: a mouse mAb for MMP-9 (Ab-7) (Calbiochem, San Diego, CA,
USA); a rabbit pAb for MMP-1 (Chemicon); a rabbit pAb for MMP-8 (Chemicon);
and a rabbit pAb for MMP-13 N-terminal (Chemicon). Following incubation, the
solution was centrifuged at 2,000 rpm for 2 minutes, the supernatant removed, and the
resin then washed twice with TBST. This “charged” resin was then incubated with the
human serum, AWF and CWF samples for 18 hours at 4 ºC, constantly rotating. As
before, following incubation the solution was centrifuged at 2,000 rpm for 2 minutes,
the supernatant removed, and the resin was washed twice with TBST. The resin was
then mixed with a 2X SDS sample buffer and left at room temperature for 1 hour. To
examine the bound proteins, these resin samples were run on Collagen Type I
zymograms and analysed as described above.
2.2.6 Quantitative confirmation of MMP-9 levels through a direct enzyme-
linked immunosorbant assay (ELISA)
Quantitative MMP-9 levels in HS, AWF and CWF samples were obtained using a
direct ELISA. Briefly, 96-well plates (Nunc) were coated with 40 ng of the
sample/well and left overnight at 4 ºC. Next, the wells were washed three times with
TBST and blocked with 5% skim milk powder (SMP) in TBST for 1 hour at room
temperature. Wells were again washed with TBST, and then probed with a rabbit
polyclonal antibody (1:1000) raised against the MMP-9 whole molecule (Abcam,
Cambridge, UK) for 1 hour at room temperature. Following washing, the wells were
probed with polyclonal goat anti-rabbit immunoglobulins/HRP (1:1000) (Dako
Denmark A/S, Glostrup, Denmark) for 1 hour at room temperature. Then, wells were
again washed and rinsed a final time with TBS before adding the substrate solution,
2,2′-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) (Sigma-Aldrich) for
30 minutes at room temperature. The optical density of the plate was then read at 405
nm on a standard plate reader (Bio-rad Laboratories). The amount of MMP-9 in the
samples was calculated from a standard curve (MMP-9, Calbiochem) and then
converted to MMP-9 (ng)/total sample protein (μg). Data was organised in terms of
the sample’s PUSH score.
41.
2.3 RESULTS
2.3.1 Elevated proteolytic activity in chronic wound fluid
In order to analyse the proteolytic activity in wound fluid samples, CWFs from nine
separate patients were analysed using Collagen Type I and Collagen Type IV
zymography to identify protease activity. Collagens were used as they are the major
extracellular matrix constituents of the skin, contributing to approximately 70% of its
dry weight. Furthermore, Collagen Type I accounts for approximately 80% of the skin
collagens, while Collagen Type IV forms the basement membrane (Schacke, Docke et
al. 2002). Collagen degrading activity was identified as clear bands on a dark
background in the zymograms, with some smearing evident in the standards, as seen
previously by others (Stuart, Evans et al. 2007). This activity was then quantitated
through densitometric analysis. Ten times more total protein of both the HS and AWF
samples had to be loaded on the gels in order to adequately visualise the collagenase-
like activity on both substrates. As shown in Figure 2.1, the degradation profiles vary
between the patient samples in terms of both intensity and molecular weight of the
visualised bands. However, the bands appear to be consistent between the two
different substrates.
More specifically, Collagen Type I zymography revealed a high level of protease
activity in the CWF samples, especially considering the lower amount of the sample
loaded (Figure 2.1). In addition to the main band at approximately 102 kDa, there are
also smaller and potentially more active forms visible at 75 kDa, 66 kDa, 54 kDa and
48 kDa across the samples. Compared to CWF, less protease activity was evident in
both the HS and AWF samples, especially considering the increased protein loaded
for these samples. The main band in the AWF samples correlates with a mass of 66
kDa, similar to the main band seen in the HS. CWF-6 appears to have less proteolytic
activity than the other CWF samples, although, the activity present was greater than
that found with the HS and AWF samples.
Similar to the results found with the Collagen Type I zymography, Collagen Type IV
zymography also revealed increased levels of protease activity in CWF as compared
with HS and AWF samples (Figure 2.2). The main band in CWF is visualised at 102
42.
Figure 2.1 Collagen Type I zymography demonstrating protease activity present in wound fluid samples. Lane 1 is the positive control of MMP-8 (100 ng). Lane 2 is HS and lanes 3-5 are AWF samples 1-3 with 5 µg of total protein loaded per sample. Lanes 6-14 are CWF samples 1-9 with 500 ng of total protein loaded per sample.
Figure 2.2 Collagen Type IV zymography demonstrating protease activity present in wound fluid samples. Lane 1 is the positive control of MMP-9 (10 ng). Lane 2 is human serum and lanes 3-5 are AWF samples 1-3 with 10 µg of total protein loaded per sample. Lanes 6-14 are CWF samples 1-9 with 1 µg of total protein loaded per sample.
43.
kDa, suggesting that this is the same protease as identified with Collagen Type I
zymography. The main differences observed with Collagen Type IV zymography, as
compared to Collagen Type I zymography, are that firstly, there seem to be less bands
of lower molecular weight activity with this substrate, and also, the amount of
proteolytic activity is reduced when taking into account the fact that twice the sample
amount was loaded.
2.3.2 MMPs are responsible for collagen degradation as shown through a
specific inhibitor study
The MMP-specific inhibitor, GM6001, was used to confirm that the collagen
degradation revealed by zymography was indeed due to MMPs and not another group
of proteases present in the wound fluid samples. Through incubation with increasing
concentrations of GM6001 at 37 ºC over 24 hours, the proteolytic activity in CWF-1
gradually decreased as revealed by Collagen Type I zymography. Indeed, no visible
degradation of collagen was apparent at 1 mM GM6001 (Figure 2.3A). Densitometry
was used to quantify the reduction in proteolytic activity revealed by zymography, and
the levels are represented graphically in Figure 2.3B. Both the total amount of
protease activity, plus the protease activity in individual molecular weight species,
were followed across the increasing concentrations of 1, 10, 100 and 1,000 µM
concentrations of GM6001 and the data is expressed as relative % of the untreated
CWF-1 ± SEM. In all cases, the collagen degrading activity decreases with the
increasing concentrations of the MMP-specific inhibitor, confirming that the majority
of the proteases responsible for collagenase-like activity in these CWF samples are
from the MMP family.
2.3.3 Initial immunoprecipitation of MMPs from CWF samples
Following confirmation that the majority of collagen-degrading activity seen in CWF
samples was due to MMP-specific activity, individual MMPs were
immunoprecipitated from the CWF samples. Individual MMP antibodies were linked
to a protein A/G resin and then incubated with the CWF samples. In the first instance,
antibodies for MMP-1, -8 and -13 were used to identify particular species in terms of
relative abundance when compared to the complete samples (Figure 2.1). Specific
44.
(A)
(B)
Figure 2.3 GM6001 inhibition of protease activity in CWF-1 confirms the MMP specific degradation of Collagen Type I as revealed through zymography. (A) Lane 1 is the positive control of MMP-8 (100 ng). Lane 2 is the untreated sample of CWF-1 with 500 ng of total protein loaded. Lanes 3-9 are CWF-1 samples (500 ng) treated with increasing concentrations of the specific MMP-inhibitor, GM6001, at 1, 3, 10, 30, 100, 300, and 1000 µM at 37 ºC for 24 hours. (B) Relative levels of protease activity in CWF-1 samples treated with increasing concentrations of GM6001. The MMP-specific inhibition of collagen degrading activity was represented quantitatively through densitometric analysis. Levels are shown as the % collagen degrading activity, of both the total and also individual molecular weight bands, as compared to the untreated sample, ± SEM (n=3).
45.
MMPs were isolated from CWF-1 and then analysed through Collagen Type I
zymography (Figure 2.4). It appeared that there was minimal activity in all three
MMP-specific immunoprecipitations, as shown through the decreased degradation of
Collagen Type I when compared with the complete samples (Figure 2.1).
In view of this, six CWF samples were analysed using specific immunoprecipitation
of MMP-8, followed by analysis through Collagen Type I zymography as described
previously. In these preliminary optimisation studies, only six CWFs were used due to
limited sample volume. Again, there appeared to be minimal MMP-8 activity across
all 6 CWF samples containing the MMP-8 enriched fractions (Figure 2.5), confirming
the initial observation with CWF-1 (Figure 2.4). MMP-13 was also analysed across
the same 6 CWF samples and revealed similar results (Figure 2.6). Therefore, from
both the initial observations in CWF-1, followed by the analysis of 6 separate CWF
samples, it appeared that the levels of MMP-1, -8 and -13 were insufficient to cause
the majority of Collagen Type I degrading activity as revealed in the complete CWF
samples through zymography (Figure 2.1).
2.3.4 MMP-9 is the predominant protease responsible for matrix degradation
in chronic wound fluid
In light of the minimal levels of MMP-1, -8 and -13 activity detected in the initial
studies, MMP-9 was chosen due to its ability to degrade a number of substrates,
namely collagen types I-V (Okada, Gonoji et al. 1992), as well as being reported in
excessive levels in CWF in previous studies (Yager, Zhang et al. 1996). Through
incubation with a MMP-9 monoclonal antibody linked to an A/G protein resin, this
particular MMP species was able to be analysed in terms of relative abundance within
HS, AWF and CWF samples. Both the bound and unbound fractions were analysed
through Collagen Type I zymography and then quantitated through densitometric
analysis. The degradation profiles obtained via zymography varied in terms of both
abundance and molecular weight of the proteolytic activity.
The MMP-9 enriched fractions of CWF showed high levels of Collagen Type I
degrading activity (Figure 2.7A). However, there was minimal activity in the enriched
46.
Figure 2.4 Initial immunoprecipitation of MMP-1, -8, -13 from CWF-1 and analysis of enriched fractions through Collagen Type I zymography. Lanes 1-3 contains MMP-1-bound fraction, MMP-8-bound fraction and MMP-13-bound fraction respectively with 500 ng of total protein loaded per sample.
47.
Figure 2.5 Immunoprecipitation of MMP-8 from wound fluid samples and analysis of MMP-8 enriched fractions through Collagen Type I zymography. Lane 1 contains the positive control of MMP-8 (100 ng). Lanes 2 and 3 are the negative controls of Protein A/G and Protein A/G + MMP-8 pAb. Lanes 4 -9 are the bound fractions of CWF samples 1-6 with 500 ng of total protein loaded per sample
48.
Figure 2.6 Immunoprecipitation of MMP-13 from wound fluid samples and analysis of MMP-8 enriched fractions through Collagen Type I zymography. Lane 1 contains the positive control of a 27 kDa truncated version of MMP-13 (100 ng). Lanes 2 and 3 are the negative controls of Protein A/G and Protein A/G + MMP-13 pAb. Lanes 4 -9 are the bound fractions of CWF samples 1-6 with 500 ng of total protein loaded per sample
49.
fractions of human serum and AWF samples, except for AWF-2. The molecular
weight profile in these bound fractions appeared quite similar to the Collagen Type I
zymograms of the complete samples (Figure 2.1), possibly indicating that MMP-9
was indeed the major constituent of the collagen degrading activity in the wound
fluids. Quantitatively, there were significant differences between HS and AWF-2
(p<0.05), and with all CWF samples (p<0.01), except for CWF-6 (Figure 2.7B). The
main clinical difference in the patient from whom CWF-6 was collected is that the
patient had had the ulcer for a shorter duration, which may provide further insight into
wound’s clinical status. In terms of grouped samples, the CWF samples showed
statistically significant increased collagenase-like activity (p<0.01) when compared to
both HS and AWF samples (Figure 2.7C). Further detailed analysis of the samples in
terms of the PUSH scores revealed that both the lower PUSH score (PUSH ≤11)
grouped samples and the higher PUSH score (PUSH ≥12) grouped samples both had
significantly increased MMP-9 activity compared to both HS and AWF samples
(p<0.01) (Figure 2.7D).
MMP-9 depleted fractions showed lower levels of Collagen Type I degrading activity
(Figure 2.8A) than both the unfractionated samples (Figure 2.1) and enriched
fractions (Figure 2.7A). Importantly, the total protein amount loaded on the
zymograms with these depleted fractions was identical to that loaded in the
unfractionated samples described above (Figure 2.1). Furthermore, there were
increased levels of collagenase-like activity in the depleted fractions of CWF samples
1-5 and 7-9, compared with HS, AWF samples and CWF-6. Quantitatively, the levels
of MMP-9 activity were significantly increased in CWF samples 1-3 and 8-9 (p<0.01)
and CWF-7 (p<0.05), with similar trends evident in CWF samples 4 and 5, although
they were not significant (Figure 2.8B). When the samples were grouped into their
relative types, the CWF samples displayed highly significant Collagen Type I
degrading activity when compared with HS and AWF samples (p<0.01) (Figure
2.8C). In addition, analysis of the samples in terms of PUSH scores revealed that the
higher PUSH score (PUSH ≥12) group had significantly increased MMP-9 activity
compared to HS and AWF samples (p<0.01), as well as the lower PUSH score (PUSH
≤11) group (p<0.05). In addition, the lower PUSH score (PUSH ≤11) group was not
significantly different from either HS or AWF samples. The main band differences in
the zymography profiles of the bound and unbound fractions were at an approximate
50.
(A)
(B)
Figure 2.7 Immunoprecipitation of MMP-9 from wound fluid samples and analysis of MMP-9 enriched fractions through Collagen Type I zymography. (A) Lane 1 contains the positive control of MMP-9 (500 pg). Lanes 2 and 3 are the negative controls of Protein A/G and Protein A/G + MMP-9 mAb. Lanes 4 -16 are the bound fractions of HS, AWF samples 1-3 and CWF samples 1-9 with 500 ng of total protein loaded per sample (B) Relative levels of MMP-9-bound fractions in individual samples represented quantitatively through densitometric analysis. Levels are shown as the mean activity ± SEM (n=3). Statistical significance is relative to HS and shown as either * (p<0.05) or # (p<0.01) as determined by Tukey’s test.
51.
(C)
(D)
Relative levels of MMP-9 enriched fractions in samples represented quantitatively through densitometric analysis. Levels are shown as the mean activity ± SEM (n=3).(C) Data from (B) is separated into HS, AWF and CWF pooled samples. Statistical significance is relative to both HS and AWF and shown as # (p<0.01) as determined by Tukey’s test. (D) Data from (B) is separated into samples with a higher PUSH score (CWF samples 1-5, PUSH ≥12) and lower PUSH scores (CWF samples 6-9, PUSH ≤11). A higher PUSH score indicates a clinically worse ulcer. Statistical significance is relative to both HS and AWF and shown as # (p<0.01) as determined by Tukey’s test.
52.
(A)
(B)
Figure 2.8 Immunoprecipitation of MMP-9 from wound fluid samples and analysis of MMP-9 depleted factions through Collagen Type I zymography. (A) Lane 1 contains the positive control of MMP-9 (500 pg). Lanes 2 -14 are the unbound fractions of HS, AWF samples 1-3 and CWF samples 1-9 with 500 ng of total protein loaded per sample (B) Relative levels of MMP-9-unbound fractions in individual samples represented quantitatively through densitometric analysis. Levels are shown as the mean activity ± SEM (n=3). Statistical significance is relative to HS and shown as either * (p<0.05) or # (p<0.01) as determined by Tukey’s test.
53.
(C)
(D)
Relative levels of MMP-9 depleted fractions in samples represented quantitatively through densitometric analysis. Levels are shown as the mean activity ± SEM (n=3).(C) Data from (B) is separated into HS, AWF and CWF pooled samples. Statistical significance is relative to both HS and AWF and shown as # (p<0.01) as determined by Tukey’s test. (D) Data from (B) is separated into samples with a higher PUSH score (CWF samples 1-5, PUSH ≥12) and lower PUSH scores (CWF samples 6-9, PUSH ≤11). A higher PUSH score indicates a clinically worse ulcer. Statistical significance is relative to both HS and AWF and shown as # (p<0.01) as determined by Tukey’s test.
54.
molecular weight of 65 and 55 kDa respectively. Therefore, from these results it
appears that increased levels of MMP-9 activity are significantly different in CWF, as
compared with both HS and its acute counterpart, in terms of both activity and
molecular weight species. In addition, the MMP-9-unbound fractions, which had been
depleted of MMP-9, appear to display a trend towards a higher level of Collagen Type
I degrading activity correlating with a higher PUSH score, i.e. a clinically worse ulcer.
2.3.5 MMP-9 levels present in chronic wound fluid correlate with the clinical
severity of the ulcer
To confirm the findings found by immunoprecipitation, a direct ELISA was used with
a different MMP-9 antibody to ensure accuracy in antibody specificity. MMP-9 levels
for all samples ranged from approximately 70 ng/µg total protein to 370 ng/µg total
protein. In all AWF samples, there were no statistically significant differences to HS.
Concerning the CWF samples, CWF-8 displayed significantly increased levels of
MMP-9 compared with HS (p<0.05), with CWF samples 1-3, 5 and 9 also showing
statistically increased MMP-9 levels (p<0.01) (Figure 2.9A). When grouped samples
were analysed, CWF showed a statistically significant increase in MMP-9 levels
compared with both AWF and HS (p<0.01) (Figure 2.9B). In addition, when the
samples were grouped according to their clinical PUSH score, the lower PUSH score
(PUSH ≤11) group was not significantly different from HS, but did have statistically
significant differences from both AWF and the higher PUSH score (PUSH ≥12) group
(p<0.01) (Figure 2.9C). Furthermore, the higher PUSH score (PUSH ≥12) group
showed a significant increase in MMP-9 levels compared with both HS and AWF
(p<0.01) (Figure 2.9C). From these results it appears that there is a trend towards
increased levels of MMP-9 correlating with an increased PUSH score, and hence a
clinically worse ulcer.
55.
(A)
(B) (C)
Figure 2.9 MMP-9 levels from wound fluid samples analysed through an ELISA. Levels are shown as MMP-9 (ng)/total protein (µg) ± SEM (measured in triplicate and repeated).(A) Data is represented as individual samples. Statistical significance is relative to HS and shown as either * (p<0.05) or # (p<0.01) as determined by Tukey’s test. (B) Data from (A) is separated into HS, AWF and CWF pooled samples. Statistical significance is relative to both HS and AWF and shown as # (p<0.01) as determined by Tukey’s test. (C) Data from (A) is separated into samples with a higher PUSH score (CWF samples 1-5, PUSH ≥12) and lower PUSH scores (CWF samples 6-9, PUSH ≤11). A higher PUSH score indicates a clinically worse ulcer. Statistical significance is relative to both HS and AWF and shown as # (p<0.01) as determined by Tukey’s test.
56.
2.4 DISCUSSION
A major factor responsible for the chronic nature of ulcers is believed to be the tissue-
destructive events mediated by certain endopeptidases (Grinnell, Ho et al. 1992;
Palohati, Lauharanta et al. 1993). These neutral endopeptidases are able to degrade the
extracellular matrix at the ulcer’s physiological pH and can be classified into two
groups: serine proteases and matrix metalloproteinases (Palohati, Lauharanta et al.
1993). A previous study by Weckroth et al. (1996) showed that serine proteases of
polymorphonuclear neutrophil origin, e.g. cathepsin G and elastase, were both low in
activity in CWF (Weckroth, Vaheri et al. 1996). For this reason, the studies reported
in this chapter focussed on MMPs.
In terms of MMPs, there have been a number of studies measuring MMP activity in
CWF, but many of these studies have only used general techniques to identify a broad
range of MMPs e.g. the Azocoll assay (Tarnuzzer and Schultz 1996; Trengove, Stacey
et al. 1999), gelatin zymography (Wysocki, Staianocoico et al. 1993; Yager, Zhang et
al. 1996; Trengove, Stacey et al. 1999; Wysocki, Kusakabe et al. 1999), and antibody
detection (Valleala, Hanemaaijer et al. 2003). As most members of the MMP family
are structured into three essential, characteristic and highly conserved domains
(Massova, Kotra et al. 1998), it is often quite difficult to differentiate MMP species
based on molecular weight alone. In addition, antibody assays that are based on
specific MMPs often find it difficult to differentiate between closely related species,
as there is a very close evolutionary relationship within the family, as is reflected by
their highly conserved domains (Somerville, Oblander et al. 2003). Indeed, this high
degree of structural similarity is likely to be the main reason underlying the
controversy surrounding levels of specific MMPs in chronic wounds. In view of this,
the studies in this chapter used multiple techniques to identify specific MMPs species.
Together, the combination of an antibody-based identification assay with an activity
assay has allowed more conclusive and robust results. Indeed, the findings reported
herein indicate that there are minimal levels of MMP-1, -8 and -13 activity in CWF,
with MMP-9 being the most likely candidate responsible for Collagen Type I
degradation in CWF. This was further confirmed by using a different MMP-9
antibody for the ELISA as compared with immunoprecipitation.
57.
As reported in this chapter, all nine CWF samples exhibit higher levels of MMP
activity than the AWF and HS samples. CWF-6 appears to have less MMP activity
than the other CWF samples; however, it still contains more MMP activity than the
HS and AWF samples. These differences within the sample groups highlight the
difficulties of dealing with CWF as a diagnostic sample; due to the heterogeneity of
these samples, and the wounds from which they are obtained, it is often quite difficult
to achieve statistically conclusive results. However, in this study elevated levels of
Collagen Type I and IV degradation by CWF samples compared with those found
with HS and AWF samples has been reported. In addition, the collagen degradation
can be specifically attributed to MMPs as shown through the targeted inhibitor study
using GM6001, a specific MMP inhibitor (Levy, Lapierre et al. 1998). Thus, GM6001
was shown to inhibit the CWF-induced collagen degradation in a dose-dependent
manner. Together, these data confirm the hypothesis that elevated MMP levels in
CWF are indicative of a clinically worse chronic ulcer and furthermore, suggest that
inhibition of these proteases holds promise as a therapeutic treatment for these
debilitating chronic wounds.
More specifically, these experiments suggest that MMP-9 appears to be the major
protease responsible for matrix degradation in CWF. This is similar to previous
reports by Ladwig et al. (2002), which suggested that there was a statistically
significant correlation between poor healing and elevated levels of MMP-9 in patients,
regardless of the treatment regime (Ladwig, Robson et al. 2002). Moreover, Wysocki
et al. (1999) state that as healing proceeds, the levels of MMP-9 decrease in CWF and
reach those found in acute wounds (Wysocki, Kusakabe et al. 1999). Further evidence
suggests that the MMP-9 present in CWF is derived from macrophages infiltrating the
wound bed (Moses, Marikovsky et al. 1996), which considering the excessive levels
of MMP-9 in CWF compared with AWF, could further confirm the hypothesis that
chronic wounds are partly caused by disproportionate inflammation and the inability
to proceed into the normal remodelling stage of wound healing (Chen, Schultz et al.
1999). Therefore, the clinically worse ulcers are those that show prolonged
inflammation, presenting as increased levels of macrophage-derived MMP-9 in the
resultant CWF.
58.
In the results reported herein, the lower and higher PUSH score groups were
indistinguishable through immunoprecipitation, as both groups showed a significant
increase in MMP-9 activity compared with HS and AWF samples (p<0.01). Similar
results have been obtained in previous studies, with MMP-9 in CWF showing a
minimum of a 2-fold increase when compared with AWF and HS (Wysocki,
Kusakabe et al. 1999; Ladwig, Robson et al. 2002). When considering the MMP-9
depleted fractions, there were significant MMP-9 activity differences between the
lower PUSH score (PUSH ≤11) group and the higher PUSH score (PUSH ≥12) group
(p<0.05), with the higher PUSH score (PUSH ≥12) group displaying a significant
increase in Collagen Type I degrading activity as compared with both HS and AWF
(p<0.01). This could be due to the large levels of MMP-9 present in these samples,
which in turn suggests that the zymograms may well have been overloaded. If indeed
this did occur, this may explain why statistically significant differences between the
two clinical groupings were not obtained. Nevertheless, Tarlton et al. (1999) were
able to use gelatin zymography to show a statistical correlation between pro-MMP-9
and the increasing severity of the ulcer (p=0.006) (Tarlton, Bailey et al. 1999).
In comparison, the Tarlton et al. (1999) study used wound site specific collection
techniques, i.e. they collected CWF from the advancing wound margin in a
deteriorating ulcer and not the ulcer it its entirety; so the results are not directly
comparable (Tarlton, Bailey et al. 1999). In this study, we specifically chose to use a
standardised collection technique that recovered the wound fluid in its entirety without
any added protein selection processes, i.e. differential protein absorption and retrieval
from wound dressings. Furthermore, while the samples were not collected from
clinically infected wounds, it is important to recognise that there would still be a
bacterial presence in the wound, which may then contribute to the wound fluid’s
proteolytic activity. The extent to which bacterial proteases contribute to the overall
amount of MMP activity in CWF will be a focus of future studies. Another interesting
point to note is that while some MMPs have been known to degrade IgG proteins,
namely MMP-3 and MMP-7 (Gearing, Thorpe et al. 2002), this was not observed to
contribute to cleavage of the antibody from the resin in these experiments. This could
be due to lower levels of these specific MMPs, i.e. MMP-3 and -7, in the CWF
samples, or indeed, competitive binding of MMP-9 to the antibody, which could then
potentially block any further interactions.
59.
To further investigate any potential differences in CWF collected from wounds with
different clinical status, a direct ELISA was used with a different MMP-9 antibody to
show quantitative levels of MMP-9 in these samples. This technique showed clear
differences between the levels of MMP-9 present in the two separate groupings. The
lower PUSH score (PUSH ≤11) group displayed significantly higher levels of MMP-9
than AWF (p<0.01), as well as significantly lower levels of MMP-9 than the higher
PUSH score (PUSH ≥12) group (p<0.01). As a result, it appears that there is a strong
correlation between increased MMP-9 levels in CWF from wounds in the higher
PUSH score (PUSH ≥12) group compared to AWF samples (p<0.01). This therefore
suggests that MMP-9 is the predominant protease involved in degrading the
extracellular matrix in the chronic wound environment and healing could potentially
be promoted by its attenuation. In contrast, Fray et al. (2003) propose that MMP-2 and
-9 are both essential for wound healing, a process that would therefore be impaired if
these two proteases were inhibited in chronic ulcer treatments (Fray, Dickinson et al.
2003). However, their hypothesis is based on the fact that gelatinases are required for
tissue remodelling and for the onset on re-epithelialisation – stages of wound healing
that normally occur when inflammation has subsided (Trengove, Stacey et al. 1999).
Therefore, if significantly increased levels of MMP-2 and -9 are present before
inflammation has abated, an excessively proteolytic environment will continually
degrade key growth promoting agents and thus will not allow normal wound healing
to occur.
60.
2.5 CONCLUSION
In this study, the levels of MMP-9 present in the CWF samples showed a statistically
significant increase compared with both HS and AWF. When these CWF samples
were separated by their clinical score, the fluids from the clinically worse ulcers, i.e.
those with the higher PUSH scores (PUSH ≥12), displayed significantly higher levels
of MMP-9 than those from the lower PUSH score (PUSH ≤11) group (p<0.01). In
addition, it appears that high levels of MMP-9 activity in CWF are a better indicator
of a wound's chronicity than the total protease activity levels alone. Further, these data
suggest it is not surprising that current topical treatments of chronic ulcers with
bioactives, e.g. growth factors, have proven to be only slightly effective in treating
chronic wounds since high levels of proteases can rapidly degrade these growth
promoting agents (Cullen, Watt et al. 2002).
As demonstrated in this chapter, it is likely that these bioactives have been ineffective
as therapeutics in the clinical setting due to the high levels of proteases found in CWF.
Therefore, many authors have suggested that the addition of a protease inhibitor prior
to topical treatment of the wound would promote healing (Tarnuzzer and Schultz
1996; Wysocki, Kusakabe et al. 1999; Yager and Nwomeh 1999; Cullen, Smith et al.
2002). In view of these previous studies, along with the results reported herein, it is
suggested that a specific inhibitor of MMP-9 could potentially be more therapeutically
effective than general MMP inhibitors in modulating chronic ulcers towards a healing
state. This hypothesis is further explored in Chapter 4.
61.
CHAPTER 3
SYNTHESIS OF HYDROGELS FOR WOUND
DRESSING APPLICATIONS
62.
3.1 INTRODUCTION
As outlined in Chapter 1, the hydrogel is of particular interest in many human
therapeutic applications as it possesses a high water content, good biocompatibility,
and also often a texture similar to that of the body’s own soft tissue (Schmedlen,
Masters et al. 2002). In addition, many human biomedical applications have been
suggested as suitable for hydrogels, including: drug delivery, contact lenses, corneal
implants, as well as substitutes for skin, tendons, and cartilage (Schmedlen, Masters et
al. 2002). More recently, hydrogels have also been investigated for wound dressing
applications. As a large amount of importance needs to be placed on patient
compliance, pain management and quality of life; the hydrogel is often chosen as it is
cool, non-sticking and comfortable for the patient (Queen, Orsted et al. 2004; Akita,
Akino et al. 2006). Current commercially available hydrogel wound dressings include:
IntraSite™ gel (Smith & Nephew); Tegagel™ (3M); Curasol® gel (Healthpoint); Nu-
Gel® (Johnson & Johnson); as well as various others from a large range of wound
care companies. Therefore, hydrogels appear to be a logical choice for the basis of
further wound dressing developments and modifications.
The choice of starting materials in hydrogel preparation is extremely important when
considering specific tissue engineering and medical applications (Lee and Mooney
2001). PHEMA based hydrogels have been widely studied for human biomedical
applications and are one of the most commonly used materials in biomedical
hydrogels e.g. soft contact lenses (Carenza and Veronese 1994; George, Wentrup-
Byrne et al. 2004). Therefore, PHEMA was chosen for this study as it not only
possesses good biocompatibility and low thrombogenicity (Caliceti, Salmaso et al.
2001; George, Wentrup-Byrne et al. 2004), but also is resistant to enzymatic
degradation and extremes in pH (Prashantha, Rashmi et al. 2006). This last point is
critical to chronic wound dressing applications as there is a marked difference
between acute and chronic wounds in terms of proteolytic activity: thus chronic
wounds exhibit excessive protease activity and a wide variability in pH (Tarnuzzer
and Schultz 1996; Ladwig, Robson et al. 2002; Li and Li 2003). A scaffold that is able
to degrade in this toxic environment may potentially exacerbate the inflammation
burden, i.e. the excessive neutrophil response, already present in these slow healing
wounds (Yager and Nwomeh 1999). In addition, PHEMA has extensively-
63.
documented resistance to non-specific protein adsorption, which will then promote
low levels of cell adhesion – another critical factor for wound dressing changes
(Martins, Wang et al. 2003). Furthermore, PHEMA is extremely inert and has not
been shown to change the pH of solutions (Montheard, Chatzopoulos et al. 1992),
which indicates that it will not adversely interfere with the biological system that it
comes into contact with when applied topically.
Gamma irradiation has been shown to be an favourable tool for hydrogel
polymerisation (Park, Nho et al. 2004). Primarily, this is because it eliminates the
need for potentially toxic catalysts or additives, as is required with chemical initiation
methods (Park, Nho et al. 2004). In terms of crosslinking, irradiative methods are
again advantageous as they do not use potentially toxic crosslinking agents, which
then allows for a easier transition through the commercial development of a human
biomedical product (Hassan and Peppas 2000). Furthermore, crosslinked hydrogels
are preferred over their water-soluble polymers, as they are inert and insoluble, i.e.
they cannot be absorbed into the body (Zahedi and Lee 2007).
The HEMA monomer itself can also contain impurities that are active crosslinking
agents when placed under gamma irradiation (Montheard, Chatzopoulos et al. 1992).
For example, as HEMA is synthesised through the transesterification of ethylene
glycol, often the monomer can contain small amounts of ethylene glycol
dimethacrylate (EGDMA) – even after distillation (Montheard, Chatzopoulos et al.
1992). In addition, as EGDMA contain difunctional reactive groups, it is a common
crosslinking agent; making resulting copolymers that are able to form hydrogels, even
at very low percentages of the impurity (Montheard, Chatzopoulos et al. 1992).
Therefore, while Carenza et al. (1993) claimed to crosslink the HEMA monomer with
PEG at a total dose of 25 kGy to create a completely polymerised crosslinked
hydrogel (Carenza, Lora et al. 1993), it is also possible that small amounts of
EDGMA contributed to the polymerisation (Figure 1.6).
PEG itself has been extensively shown to act as a porogen in the preparation of porous
polymers (Xi, Wu et al. 2006). This therefore assumes that it does not actually
covalently react into the polymer, potentially conflicting with Carenza et al.’s (1993)
work (Carenza, Lora et al. 1993). However, in terms of physical properties, PEG is
64.
highly biocompatible and is also able to exhibit versatile physical properties, which
are highly dependent on its weight percentage, molecular chain length and
crosslinking density (Almany and Seliktar 2005). Therefore, the idea of a potential
copolymerisation of these two materials certainly provides scope for an intriguing
novel material. Indeed, the specific aims of the experiments reported in this chapter
examine different PHEMA hydrogels synthesised both with and without the presence
of PEG. This was assessed in terms of properties desirable in a wound dressing, i.e.
ease of application, comfort to the patient, as well as looking at skin irritation
potential. Furthermore, the differing characteristics of the hydrogels were examined in
terms of polymer conversion, water uptake, and protein release, while cell based
assays examined the dressing’s biocompatibility.
65.
3.2 MATERIALS AND METHODS
3.2.1 Chemicals
2-Hydroxyethyl methacrylate (HEMA, assay ≥ 99% GC, Sigma-Aldrich, St Louis,
MO, USA) was distilled under reduced pressure immediately prior to use.
Polyethylene glycol 20,000 (PEG, Sigma-Aldrich), peroxidase from horseradish
(Sigma-Aldrich) and 2,2'-Azinobis [3-ethylbenzothiazoline-6-sulfonic acid]-
diammonium salt (ABTS, Pierce Biotechnology, Rockford, IL, USA) were used as
supplied under the manufacturers’ instructions. All H2O used was double deionised by
ion exchange (MilliQ, Millipore, Billerica, MA, USA).
3.2.2 Synthesis of hydrogels
Aqueous solutions of distilled HEMA were prepared in two separate ratios, both with
PEG 20,000 (50% water, 30% PEG and 20% HEMA) and without PEG 20,000
(71.4% water, 28.6% HEMA). Initial optimisation studies were performed to obtain
these ratios, as hydrogels synthesised with a lower HEMA ratio were not structurally
sound. In addition, the PHEMA alone ratio was obtained by simply preparing the
same solutions in the absence of PEG to ensure comparability. Solutions were placed
between two glass plates separated by a silicone gasket and purged with argon. Each
mould was approximately 75 mm x 50 mm x 3 mm and filled with 10 mL of monomer
solution. The moulds were then gamma irradiated by a Gamma-cell 220 (Atomic
Energy of Canada Ltd, Ottawa, Canada) using a Co60 source to give a total dose of 5
and 10 kGy, respectively (Table 3.1).
Sample ID H2O (g) HEMA (g) PEG 20,000 (g) Total Dose (kGy)
C (50:20:30 H2O:HEMA:PEG) 5.0 2.0 3.0 5
G (50:20:30 H2O:HEMA:PEG) 5.0 2.0 3.0 10
PH-5 (71:29 H2O:HEMA) 5.0 2.0 - 5
PH-10 (71:29 H2O:HEMA) 5.0 2.0 - 10
Table 3.1 Hydrogel formulations prepared for analysis as potential wound dressings
66.
Hydrogels were removed from their moulds and cut into 1 cm x 1 cm square pieces
for further analysis.
3.2.3 Analysis of polymerisation through NIR FT-Raman Spectroscopy
Un-irradiated and irradiated samples were analysed by a Perkin Elmer System 2000
NIR FT-Raman spectrophotometer (Perkin Elmer, Waltham, MA, USA). The
hydrogel samples were placed into glass vials for spectroscopic analysis and the
spectra were analysed using Grams/AI (Thermo Electron Corporation, Waltham, MA,
USA).
3.2.4 Swelling of hydrogels
Hydrogels were then placed into 12-well tissue culture plastic plates (Nunc A/S,
Roskilde, Denmark) with 3 mL of H2O and swollen at room temperature for a 2-week
period. The hydrogels were then dried in a vacuum desiccator and rehydrated under
similar conditions for one week. The hydrogel samples were both weighed and
measured at appropriate intervals to provide an indication of water uptake into the
system. Swelling ratios were then measured by the following equations:
St = Mt
Mp
where Mt is the mass of the swollen polymer at time t and Mp is the mass of the
original polymer (George, Wentrup-Byrne et al. 2004). Similarly,
Vt = Vt
Vp
where Vt is the volume of the swollen polymer at time t and Vp is the volume of the
original polymer. Furthermore, to gravimetrically analyse the loss of PEG in the
hydrogels, the theoretical amount of PEG was calculated from the weight of the
hydrogel and this was then compared to the percentage of the actual weight lost.
67.
These swelling experiments were performed to gain an appreciation of the hydrogel’s
ability to take up wound exudate as would be required in its final application.
3.2.5 Protein loading
Peroxidase from horseradish (HRP ~ 40 kDa) was prepared in solutions of ddH2O at 2
µg/mL. All solutions were filtered sterilised (0.22 μm) before use. The hydrogels were
placed in 24-well plates (Nunc) that had been siliconised with Sigmacote (Sigma-
Aldrich) as per the manufacturer’s directions, along with 1 mL of HRP solution and
then placed on a rotary shaker at room temperature for 24 hours.
3.2.6 Protein release
Hydrogels that had been loaded with HRP were first briefly dried for approximately 1
hour under vacuum at room temperature. Next, the hydrogels were placed in
siliconised 24-well plates (as per protein loading) and immersed in 1 mL of 50 mM 4-
(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES – Roche, Basel,
Switzerland) at pH 7.4 and placed on a rotary shaker at room temperature. Sub-
samples of the HEPES solutions were taken at intervals of 1, 2, 6, 24, 48, 72 and 168
hours and measured for HRP concentration. Briefly, this was achieved by adding the
HRP sample to 150 µl ABTS in a 96-well plate (Nunc) and incubating this at room
temperature for 30 minutes. The optical densities of the samples were then read at 410
nm in a Benchmark Plus Microplate Spectrophotometer (Bio-Rad Laboratories,
Hercules, CA, USA) and compared to a standard curve to measure HRP
concentrations. This concentration was then used to calculate the cumulative HRP
released from the hydrogels (ng).
3.2.7 Skin collection
Skin samples were collected from consenting patients undergoing breast or abdomen
reductions at the St. Andrews and Wesley Hospitals, Brisbane, QLD, Australia.
Human ethical approval was obtained from both the hospitals and the Queensland
University of Technology. The skin samples were collected in sterile jars containing
antibiotic/antimycotic solution containing 10,000 units of penicillin/mL, 10,000 µg of
68.
streptomycin/mL and 25 µg of amphotericin B/mL, with penicillin G, streptomycin
sulfate and amphotericin B as antimycotic in 0.85% saline (Invitrogen, Carlsbad, CA,
USA). These samples were processed within 12 hours following storage at 4 ºC in
Dulbecco’s Modified Eagle’s Medium (DMEM) containing antibiotics.
3.2.8 Primary keratinocyte cell cultures
Primary skin keratinocytes were isolated as previously described (Topping, Malda et
al. 2006) and expanded on a feeder layer of lethally irradiated 3T3 mouse fibroblast
feeder cells (i3T3s) in Green’s Media (Chakrabarty, Dawson et al. 1999) that
contained 10% foetal calf serum (FCS – Hyclone, Logan, UT, USA). Keratinocytes
were grown for 7 days with the medium replaced every 3-4 days.
3.2.9 Toxicity testing using a cell number assay
Polymers were synthesised as above and soaked in 3 mL Green’s media for 3 days in
12-well plates (Nunc) along with samples of Green’s media and serum-free media
(SFM) alone at 37 ºC/5% CO2. These solutions are subsequently referred to as
“treated media”. Primary keratinocytes at p0 were passaged and seeded into black 96-
well plates (5x103 cells/well) that had been preseeded 24 hours earlier with i3T3s at
the same concentration. The keratinocytes were allowed to adhere for 24 hours before
the treatments were applied. Briefly, the media was aspirated, the cells were washed
with a buffered salt solution and then the “treated” media replacements were added at
final “treated” media concentrations of 10, 50 and 100% respectively (Figure 3.1).
Cells were then incubated for 48 hours at 37 ºC/5% CO2. Cell number was determined
using the CyQUANT® NF Cell Proliferation Assay (Invitrogen) as per the
manufacturer’s instructions. Fluorescence was read at λex485 nm λem530 nm by a
POLARstar Optima Microplate Reader (BMG LABTECH GmbH, Offenburg,
Germany). Results were expressed as a % of the SFM negative control.
3.2.10 Preparation of a three-dimensional human skin equivalent
A de-epidermised dermis (DED) model was used as previously described
69.
Figure 3.1 Experimental design of the toxicity testing
70.
(Topping, Malda et al. 2006) with the following modifications. Following
keratinocyte addition to the sterile stainless steel rings on top of the DEDs, the
composites were incubated for 24 hours at 37 ºC/5% CO2. After this period, the rings
were removed and the composites were raised to the air-liquid interface by moving
them onto stainless steel grids in 6-well plates (Nunc). Cultures were maintained for 5
days at 37 ºC/5% CO2 before the hydrogel treatments were applied.
3.2.11 Biocompatibility testing of hydrogels using the DED model
The DED model was prepared as described above. At five days post air-liquid
interface culture, the hydrogels were placed on top of the composite, along with no
treatment and Allevyn™ dressing (Smith & Nephew, London, United Kingdom), a
commercially available wound dressing, as controls. These wound dressings were
exposed to the composite for 7 days at 37 ºC/ 5% CO2, with the media being replaced
at 3 days (Figure 3.2). At completion of the treatment, a 3-(4,5-dimethylthiazol-2-yl)-
2,5-diphenyltetrazolium bromide conversion (MTT) assay (Denizot and Lang 1986)
was used to analyse one of each treatment group. Briefly, the DED was submerged in
6 mL of 0.5 mg/mL MTT dye (Sigma-Aldrich) and incubated at 37 ºC for 2 hours.
Following incubation, the metabolically active cells could be visualised as a purple
colour and the DED were photographed. Next, both the MTT-stained and un-stained
samples were fixed and paraffin-embedded using standard protocols for haematoxylin
and eosin (H&E) histological analysis (Sharpey-Schafer 1954).
3.2.12 Immunohistochemistry
Paraffin sections were cut (3 μm sections) and then deparaffinised in ethanol and
xylene. Briefly, this involved sequential incubations with solutions of 100% xylene,
100% ethanol, 95% ethanol, 70% ethanol and distilled water. Sections were then
probed separately for: keratin 1/10/11 (K1/10/11), a marker for cornification and
squamous cell differentiation; p63 (RDI Research Diagnostics, Concord, MA, USA), a
p53 analogue that identifies normal basal cells as opposed to malignant tumours; and
cleaved caspase-3 (Cell Signaling Technology, Danvers, MA, USA), a critical
mediator of apoptosis in mammalian cells. For slides that were probed for cleaved
71.
Figure 3.2 Experimental timeline of the three-dimensional ex vivo human skin equivalent model
72.
caspase-3, antigen unmasking was required before the blocking step. Briefly, the
slides were incubated at 37 ºC in 10 mM citrate buffer (pH 3.0) for 30 min, before
proceeding as normal. After incubation with primary antibodies for K1/10/11 (1:400),
p63 (1:100) and cleaved caspase-3 (1:100), sections were stained using a Dako
Envision kit (Dako Denmark A/S, Glostrup, Denmark) as per the manufacturer’s
instructions, with the exception that phosphate buffered saline (PBS) was used instead
of TBS. After antibody development of the labelled secondary antibody with the 3,3’
diaminobenzidine (DAB) chromogen solution, all sections were counterstained with
haematoxylin for 30 seconds and analysed using light microscopy.
73.
3.3 RESULTS
3.3.1 Hydrogel preparation
Four sets of hydrogels were synthesised according to the conditions outlined earlier
(Table 3.1). There were clear differences in the PHEMA hydrogels polymerised in the
presence of PEG compared to those without PEG, namely in terms of transparency
and the apparent ease of handling and topical application. The PHEMA alone
hydrogels appeared as opaque sheets for both doses of gamma irradiation. The sheets
were then characterised using a number of techniques, including NIR FT-Raman
Spectroscopy, ability to swell with water, protein uptake and release, and two- and
three-dimensional cell culture biocompatibility studies.
3.3.2 Analysis of polymerisation through NIR FT-Raman Spectroscopy
NIR FT-Raman Spectroscopy demonstrated that a minimal dose of gamma irradiation
induced almost compete polymerisation of the HEMA monomer to its polymer form.
This was displayed by the characteristic C=C peak at 1639 cm-1 becoming reduced in
the non-irradiated samples as compared to the irradiated samples at a total dose of 5
kGy. This peak completely disappeared in the samples irradiated to a dose of 10 kGy
(Figure 3.3B). Polymerisation was observed in samples synthesised both with and
without PEG. However, for the 5 kGy samples containing PEG, it appears that there is
a small amount of residual monomer still present, as shown from the slight peak
present at 1639 cm-1 (Figure 3.3B). Therefore, it appears that the PEG interferes to a
degree with the HEMA polymerisation, but this is overcome by the increased gamma
irradiation dose as demonstrated with the 10 kGy sample.
3.3.3 Hydrogel swelling ratios in water
To further characterise the hydrogel sheets, swelling studies were performed in water.
Small squares of each hydrogel (1 cm x 1 cm) were immersed in water for a period of
time and analysed according to certain variables. The weight and volume of the
hydrogel at various time-points were compared with the original weight and volume
74.
Figure 3.3 Hydrogel sample analysis using NIR FT-Raman Spectroscopy. Spectra indicate intensity/ Raman shift in cm-1.(A) 1 – Unirradiated H2O:HEMA, 2 – Irradiated H2O:HEMA (5 kGy) (B) 1 – Unirradiated H2O:HEMA:PEG, 2, 3 – Irradiated H2O:HEMA:PEG (5, 10 kGy)
75.
of the hydrogel and expressed as this ratio ± SEM. Furthermore, after the initial
swelling, samples were dried and then rehydrated to gain information on the capacity
of the “washed” hydrogels to swell. The swelling ratios, as determined from the initial
swelling profile compared to the rehydration profiles, varied greatly (Figure 3.4). In
addition, the hydrogels that were polymerised without the presence of PEG showed
very different initial swelling profiles to those polymerised in the presence of PEG.
However, both groups of samples appeared quite similar in the rehydration study.
In the initial swelling study, the hydrogels polymerised in the presence of PEG
displayed a quick burst of swelling in the period up to six hours, and then decreased
down to below their initial weight and volume from approximately 48 hours onwards
(Figure 3.4A-B). In contrast, those hydrogels polymerised without the presence of
PEG displayed a continuous steady increase in swelling until the conclusion of the
experiment at 336 hours (2 weeks). The ratios were quite similar for both the 5 and 10
kGy samples in each group. Overall, there were minimal differences between
polymers PH-5 and PH-10 as shown by initial and rehydration swelling in terms of
both weight and volume (Figure 3.4). For the hydrogels polymerised in the presence
of PEG, polymer G consistently appeared to be capable of swelling with a higher ratio
than polymer C. Furthermore, the percentage of PEG released into the water compared
to the amount of PEG initially in the monomer solution was calculated to be
approximately 80% (Figure 3.5), indicating a large amount of PEG is not covalently
linked to the hydrogel.
3.3.4 HRP release from the hydrogel sheets following diffusion loading
Release studies were performed with HRP as a model protein (40 kDa) to analyse the
capacity of the hydrogel to release protein in a controlled manner, while still
maintaining protein function. In all cases, the release profiles were quite similar,
averaging between approximately 400 ng and 500 ng of enzymatically active HRP
being released over the one-week period (Figure 3.6a).
The HRP release profiles for all hydrogels correlated well with the Fickian diffusion
model (Figure 3.6b), which was described in more detail in Section 1.3.1. Briefly, in
this type of diffusion the rate of polymer chain relaxation is increased compared with
76.
Figure 3.4 Swelling studies of hydrogels in water. Levels are shown as a ratio of the measurement at a specific time-point compared with the initial measurement ± SEM (n=3). (A) Initial weight (B) Initial volume (C) Rehydration weight (D) Rehydration volume
77.
Figure 3.5 Gravimetric study of the loss of PEG from the HEMA:PEG hydrogels. Level is calculated from the initial swelling studies in water and is shown as the percentage of PEG released from the hydrogel compared with the amount of PEG in the monomer solution ± SEM (n=3).
78.
the rate of solute diffusion (Singh, Sharma et al. 2007). This means that as the
hydrogel swells, more protein is then able to be released from the hydrogel into the
surrounding buffer solution. All of the R2 values are approximately 1, indicating
Fickian diffusion. In the hydrogels that had been synthesised in the presence of PEG,
the cumulative amount of HRP released appeared slightly lower than that obtained
with the samples that had been synthesised without the presence of PEG (Figure 3.6).
However, the total dose of gamma irradiation did not appear to affect the amount of
HRP released over the one-week period in both the hydrogels synthesised with and
without the presence of PEG at 5 and 10 kGy. Finally, it is important to note that these
HRP measurements are based on the enzymatic conversion of a substrate to its
coloured product and, therefore, will only detect levels of functional protein.
3.3.5 Toxicity testing using a cell number assay
Primary skin keratinocytes that had been isolated from patient skin samples were used
to identify any potentially toxic effects of leached components from the hydrogel
sheets. Hydrogels were soaked in cell culture media for 3 days. This media was then
applied to cultured cells at various concentrations and cell numbers determined using
the CyQUANT® NF Cell Proliferation Assay. Toxicity of the hydrogel-conditioned
media was expressed as the relative % of the negative (SFM) control ± SEM. In this
particular case, treatments were compared to Green’s media + serum (FG), the current
“gold standard” media for culture of skin keratinocytes, that had also spent 3 days at
37 ºC/5% CO2, but in the absence of hydrogels.
There were no significant differences between the control (FG) and all three
concentrations of polymer G “conditioned” media (Figure 3.7). In contrast, and as
expected due to the lack of serum proteins that are essential for keratinocytes to
survive in culture, the highest concentration of SFM (100%) displayed significant
toxicity to the primary keratinocytes (p<0.05). Further, the highest concentration of
PH-10 “conditioned” media appeared to be the most toxic of the samples applied to
these primary skin cells, with a significant decrease in cell numbers (p<0.01). Indeed,
there appears to be a trend in the PH-10 “conditioned” media samples with a decrease
in cell number as higher concentrations of the “conditioned” media were exposed to
the cells. However, this effect is not apparent when PEG is present in the synthesis of
79.
Figure 3.6 Cumulative release of HRP from the hydrogel sheets into a HEPES buffer at pH 7.4. Hydrogels were first immersed in an aqueous HRP solution for 24 hours and then dried under vacuum. (A) Levels are shown as cumulative HRP released (ng) as determined by ABTS conversion to a coloured product at 410 nm ± SEM (n=3). (B) HRP diffusion is shown through plotting Mt/M∞ versus time1/2. All of the R2
values are between 0.9890 and 1.000 indicating Fickian diffusion.
80.
Figure 3.7 Toxicity testing of components leached from the hydrogels in primary skin keratinocytes using the CyQUANT® NF Cell Proliferation Assay to determine cell number. Hydrogels were immersed in FG media for 3 days and then this “conditioned” media was added at a final “conditioned” media concentration of 10, 50 and 100% respectively (shown from left to right). FG control and SFM control were fresh media samples; compared with FG and SFM replacements, which were media samples that were stored at 37 ºC/5% CO2 for the same 3-day period. Levels are shown as % of the SFM control ± SEM (n=3).
81.
the polymer G hydrogels, as shown by the minimal differences in cell numbers
observed with keratinocytes exposed to polymer G “conditioned” media as compared
to the positive (FG) control.
3.3.6 Biocompatibility testing using a three-dimensional human skin equivalent
A three-dimensional ex vivo human skin model was used to compare potential toxicity
effects from the synthesised hydrogels with a commercially available wound dressing,
namely Allevyn® (Smith & Nephew). The wound dressings were applied to the skin
models for a seven-day period, followed by histological and immunochemical
staining. As shown through the H&E staining, all treatments consist of a
haematoxylin-stained basal layer, with an eosin-stained cornified layer of relatively
the same thickness throughout (Figure 3.8). The Allevyn® treatment displayed
minimal basal cells, along with large numbers of nucleated cells in the cornified layer.
The other two hydrogel treatments also show some nucleated cells in the cornified
layer, but these appear to be present at a reduced level to the Allevyn® treatment. In
addition, the three wound dressing treatments all appear to result in the generation of a
“looser” weave in the cornified cells. However, this could potentially also be due to
artefacts from the paraffin embedding process. Artefacts can be introduced by several
common histological processes, including: the initial fixing in formalin; the presence
of methanol, formic acid, as well as the lack of control over osmolarity and pH;
mechanically induced artefacts involved in handling the tissue; and heating induced
tissue damage during the paraffin embedding process (Melo, Rosa et al. 2007).
In terms of specific skin markers, both the keratin 1/10/11 and p63 appear to be
relatively uniform in all samples, with the main difference being that the control
treatment shows more intense immunoreactivity in both cases (Figure 3.8). In terms of
the apoptotic marker, cleaved caspase-3, there are strong differences between the
Allevyn® treatment and all other samples. In the control treatment, there is no visible
immunoreactivity with apoptotic cells identified and defined by immunoreactivity to
the cleaved caspase-3 antibody, which is quite similar to the two other synthetic
hydrogel treatments. However, in the commercially available dressing Allevyn®,
82.
Figure 3.8 Histological and immunohistochemical analysis of an ex vivo human skin model following exposure to both a commercially available wound dressing and synthesised hydrogels. From left to right the staining or immunoreactivity is: haematoxylin and eosin (H&E), keratin 1/10/11 (K1/10/11), p63 and cleaved caspase-3 (CC-3). The arrow heads point to positively stained cells and/or regions. Scale bar = 20 µm.
83.
there is strong immunoreactivity in the cornified layer, thus revealing increased
apoptosis in cells in this layer. In terms of the specific markers tested, this is the most
marked difference between treatments (Figure 3.8).
84.
3.4 DISCUSSION
Hydrogels have been widely used for medical and tissue engineering applications as
they possess a number of advantageous properties. These include, but are not
restricted to: a high water content; good biocompatibility; patient compliance in
wound dressing applications as they are cool and comfortable for the wearer
(Schmedlen, Masters et al. 2002; Queen, Orsted et al. 2004). In addition, when
hydrogels are used as dressings they generally discourage the formation of capillary
loops in the dressing structure, i.e. the dressing is able to prevent the inward migration
of endothelial cells and subsequent division-forming capillaries, which then allows for
easy dressing changes (Queen, Orsted et al. 2004). Furthermore, as there are already a
number of non-interactive hydrogel wound dressings currently on the market, this
polymer category appears to be favourable for further wound dressing development.
In this study, synthesis of a hydrogel comprised of aqueous PHEMA has been
successfully demonstrated through gamma irradiation in the presence of PEG. Most
importantly, this dressing appears to have the ability to take up water, shows
controlled release of a model protein, and does not demonstrate toxicity effects when
assessed in both two-dimensional and three-dimensional primary skin cell culture
systems.
All four sets of hydrogel solutions effectively created crosslinked hydrogels that
possessed the ability to swell with water. In terms of appearance, the hydrogels
synthesised in the presence of PEG were both transparent and easier to manoeuvre, i.e.
would be easily applied to an uneven surface, than those synthesised as a simple
aqueous HEMA solution. Furthermore, the PHEMA alone hydrogels appeared to
show some phase separation during irradiation as they emerged opaque following
polymerisation. This could be due to the fact that the presence of PEG in the
water:HEMA solution increased the solubility of the PHEMA as it forms under
gamma irradiation. Furthermore, while the FT-Raman results suggest close to a
complete conversion, in the PHEMA alone hydrogels, it is possible that with the phase
separation some local gel particles may be created with less than 100% conversion. In
terms of the instrument, FT-Raman spectroscopy suffers from quite low sensitivity
compared with other characterisation methods (Tian, Zhang et al. 2007) and generally
cannot detect <1% impurity. In addition, while the PEG does not appear to be
85.
covalently linked to the hydrogel, it does interfere slightly with the conversion
process. However, this phenomenon was overcome when the total dose of irradiation
exceeded 10 kGy. This could primarily be due to the increase in energy required for
the HEMA polymerisation to occur when in the presence of an interfering molecule,
even though the dose does not appear to be sufficient for the PEG to react into the
polymer.
Previous studies have shown that PEG is highly biocompatible and has also been
investigated in terms of minimising biofouling (Courtois, Bystrom et al. 2006). Of
importance to this work, PEG has been demonstrated to act as a porogen in the
preparation of various porous polymers (Courtois, Bystrom et al. 2006; Xi, Wu et al.
2006). From the initial swelling studies examined here, it was apparent that the PEG
present in the gamma irradiated hydrogels C and G allowed for an initial burst of
swelling, followed by a steady decrease at approximately 48 hours to their initial
weight and volume. This can be explained by the high solubility of PEG in water, thus
the PEG started to dissolve rapidly in the first 6 hours, and thereby reduced the weight
and volume of the sheet once the hydrogels had reached their saturation point. Further
confirmation of this was obtained from the gravimetric study of PEG’s removal from
the polymer during water uptake. While it was initially hypothesised by Carenza et al.
(1993) that the PEG formed physical crosslinks between the HEMA backbone
(Carenza, Lora et al. 1993), it appears from this work, along with the proposed
gamma-induced reaction schemes (Figures 1.4-1.6) that the crosslinking occurs
through small amounts of EDGMA still present in the monomer after distillation, and
not through PEG acting as a crosslinking agent. In addition, the doses used in this
study, 5 and 10 kGy respectively, do not appear to be sufficient to create the alkoxy
radical on the end of the PEG chain. Previous studies that have used gamma
irradiation to form PEG-copolymers have ranged in total irradiation doses of 50-200
kGy (Abd Alla, Said et al. 2004). In terms of rehydration, there were minimal
differences between the PHEMA alone hydrogels. However, for the PEG-containing
hydrogels, polymer G showed a higher swelling ratio than its 5 kGy counterpart,
Polymer C. This is possibly due to a more complete polymerisation as shown by the
FT-Raman spectroscopy, thereby ensuring that no weight was lost due to possible
small percentages of unreacted HEMA monomer being soluble in the water.
86.
Both sets of hydrogels appeared to have the ability to release physiologically relevant,
i.e. similar amounts of protein as would be found in an in vivo environment, amounts
of protein, in this case HRP, over a seven-day period. Recent studies in our laboratory
have shown that growth factors present in nanogram levels are required to elicit a
beneficial wound healing response in vivo. Therefore, while many studies describe
protein release using milligram amounts (Caliceti, Veronese et al. 1992), it is
hypothesised that these supra-physiological levels of growth factors may contribute to
the down-regulation of growth factor receptors, and in turn down-regulation of
intracellular signalling – well documented outcomes of prolonged exposure to growth
factors (Rubin, Gur et al. 2005). Furthermore, HRP was used as a model protein as it
is detected using the functional conversion of a substrate, namely ABTS, to its
coloured product. Therefore, if the hydrogels themselves were to interact in a negative
manner with the protein, e.g. cause detrimental conformational changes, only the
functionally active proteins would be measured. This is extremely important in
biological applications as minor conformational changes in proteins caused by
scaffold interactions may produce major effects on the protein’s functional properties
in vivo. Furthermore, this protein release data is a good indicator that proteins are able
to access immobilised compounds on the polymer backbone.
Next, the toxicity of extracts from both sets of hydrogels was tested. As indicated
previously, the addition of PEG in the water:HEMA solution may allow for increased
solubility of the PHEMA upon synthesis. This increased solubility should also aid in
limiting phase separation between the water and HEMA, thereby increasing the
conversion from monomer to polymer and preventing monomer “pockets” from
forming. This is extremely important in terms of cell biocompatibility as the HEMA
monomer itself has demonstrated cytotoxicity in several studies, primarily through
inducing apoptosis in vitro by the differential activation of mitogen-activated protein
(MAP) kinases leading to initiation of the apoptotic cascade (Samuelsen, Dahl et al.
2007). In addition, many of the leachable components of polymers have been shown
to activate caspases and produce reactive oxygen species (ROS) – two processes that
can initiate progression into an apoptotic cascade and cause cell apoptosis and
necrosis (Samuelsen, Dahl et al. 2007). From our two-dimensional in vitro cell culture
results, it appears that the presence of PEG during aqueous PHEMA synthesis
minimises the presence of toxic leachable components. In contrast, the highest
87.
concentration of PH-10 “conditioned” media had a highly toxic effect on the
keratinocytes as shown by the strong decrease in cell numbers (p<0.01). As the PH-10
polymer is more likely to have small pockets of residual HEMA monomer, which
were undetectable by the 1% impurity limit of FT-Raman spectroscopy, leaching into
the model; and since HEMA has been associated with increasing levels of active
caspase-9-inducing apoptosis (Samuelsen, Dahl et al. 2007), this provides an
explanation for the differences between the polymer G and PH-10 treatments. Another
important point to note in this experiment is that the observed cytotoxicity can be
attributed to components leaching out from the polymers, and not the polymers
binding proteins from the media and thus depleting the media of essential
components. This conclusion can be reached since the low protein-fouling properties
of PHEMA itself is extensively documented (Martins, Wang et al. 2003). Therefore,
while the PEG itself does not polymerise into the polymer system, it is still very
important in terms of in vitro cell biocompatibility.
Three-dimensional ex vivo skin models are a useful tool for evaluating topical
products, as well as overcoming a number of ethical and practical issues associated
with animal studies (Tornier, Rosdy et al. 2006). It is important to note that this
composite does not contain inflammatory cells or a blood supply, which means that
there are some limitations in correlating this data to an in vivo system. However, 3D-
stratified layers of human epidermal cells grown in culture systems in vitro have been
shown to be useful in assessing skin irritation potential (Lawrence 1997). In this
study, the DED model showed that there were minimal differences between the
commercially available wound dressing, Allevyn®, and the two chemically
synthesised hydrogels in terms of visible structure and experimentally tested skin cell
markers. Allevyn® was selected as the commercial wound dressing as it is commonly
used in clinical practice, and is recognised as a non-interactive dressing. An
interesting point to note is that with all three wound dressing treatments, nucleated
cells became visible in the previously anuclear cornified layer. Furthermore, the
Allevyn® treatment showed a marked reduction in the number of cells in the basal
layer. These differences, compared with the control of no wound dressing treatment,
could be due to a number of causes. Primarily, the application of a topical dressing
results in culture medium being absorbed into the dressing, and therefore converts the
air-liquid interface of the 3D skin model to a liquid-liquid interface. This may then
88.
stimulate the keratinocytes to migrate faster through the cornified layers and produce a
more immature phenotype. This particular behaviour has also been described in
previous reports, where cells in a similar three-dimensional skin model did not
completely differentiate when still covered with culture medium (Ohsawa, Maruyama
et al. 1999).
In terms of skin markers, the four treatments appeared relatively similar when probed
with the skin differentiation marker keratins 1, 10 and 11 and the basal cell marker,
p63. Interestingly, there were clear differences between the treatments when the
sections where analysed in regards to their levels of apoptotic cells. Apoptosis in the
keratinocytes was analysed by staining with a marker for cleaved caspase-3 – a major
effector caspase that is specific for keratinocytes undergoing apoptosis and not just
those undergoing normal terminal differentiation (Xue, Campbell et al. 2007). The
Allevyn® treatment showed strong immunoreactivity for cleaved caspase-3
throughout the cornified layer of the skin model, thus indicating consistent apoptosis
in the cells in this stratified layer. This was not apparent in the other three treatments,
with only minimal immunoreactivity present in the skin models with the two
synthesised hydrogels. In the polymer G treated three-dimensional ex vivo skin model
there was some immunoreactivity in the outermost section of the cornified layer, and
with the PH-10 treatment there was minimal staining below the basal layer.
Concerning Allevyn®, Shuster et al. (2001) reported induction of a CD-95-dependent
apoptosis pathway in human T cells following exposure to polyurethane based
biomaterials (Schuster, Ankersmit et al. 2001). As Allevyn® is composed of a
polyurethane foam core, it appears likely that this material, or a leachable by-product
of its processing, is responsible for the apoptosis seen in this skin model. This is a
particularly interesting finding, as this product is regularly used in a clinical setting
without apparent adverse effects. This may indicate that the HSE model used in this
study is more sensitive to subtle variations that may not be visible in an in vivo clinical
situation.
89.
3.5 CONCLUSION
From the results reported above, these hydrogels, in particular polymer G, appear
promising for potential wound dressing applications. This is due to a number of
experimentally tested properties, including: the ability to swell with water; uptake and
controlled release of a biologically active protein; primary keratinocyte compatibility
in a two-dimensional in vitro cell culture system; and, biocompatibility in a three-
dimensional ex vivo skin model. Importantly, polymer G appears to demonstrate
similar, if not improved, properties to that of the commercially available wound
dressing, Allevyn®, which was tested in this study. In conclusion, this hydrogel will
be further explored in Chapter 4 with the possibility of improving it from a passive
polymer simply placed on the wound bed to that of an interactive dressing that is able
to aid in modulating the chronic wound environment such that it is conducive to
healing.
90.
CHAPTER 4
SYNTHESIS AND EVALUATION OF
FUNCTIONALISED-ALENDRONATE HYDROGELS
FOR TREATMENT OF EXCESSIVE PROTEASE
ACTIVITY IN CHRONIC WOUND FLUID
92.
4.1 INTRODUCTION
Many current topical treatments for chronic ulcers have been designed to rectify the
decreased levels of growth-factors in these wounds, e.g. Regranex® (Johnson &
Johnson), but have only been minimally effective (Cullen, Watt et al. 2002). Many
authors have postulated that this is due to the excessively high levels of proteolytic
activity found in CWF (Tarnuzzer and Schultz 1996; Stadelmann, Digenis et al. 1998;
Trengove, Stacey et al. 1999; Cullen, Watt et al. 2002), as described in Chapter 2. As
outlined in the literature review, there are a number of specific synthetic inhibitors of
MMP, namely tetracyclines and their chemically modified derivatives, e.g.
doxycycline (Ramamurthy, Kucine et al. 1998; Ramamurthy, McClain et al. 1999;
Yager and Nwomeh 1999); hydroxamic acids, e.g. GM6001 (Lund, Romer et al.
1999); and bisphosphonates, e.g. clodronate (Valleala, Hanemaaijer et al. 2003).
However, recent reports have suggested that the MMPs present in the wound bed may
also be involved in other biological processes important to wound healing, such as
growth factor activation and immune system regulation, rather than simply degrading
the wound bed matrix and growth factors (Levi, Fridman et al. 1996; Suzuki, Raab et
al. 1997; McQuibban, Butler et al. 2001; Gearing, Thorpe et al. 2002). Therefore,
while many researchers have suggested that the addition of a protease inhibitor prior
to topical treatment of the wound with growth factors would promote healing
(Tarnuzzer and Schultz 1996; Wysocki, Kusakabe et al. 1999; Yager and Nwomeh
1999; Cullen, Smith et al. 2002), other MMP-interactions in the wound healing
process also need to be considered (Somerville, Oblander et al. 2003). The goal of this
approach, which is significantly different from other proteases-inhibitory strategies
proposed by others, is to inactivate the proteases in the wound fluid after the fluid is
absorbed away from the actual wound bed. In this situation, the proteases required for
healing-associated functions within the upper cellular layers of the wound bed remain
available.
The bisphosphonates are particularly appealing for use in chronic ulcers as they have
been used clinically for a number of years to treat MMP-related disorders.
Furthermore, they exhibit low toxicity and are therefore generally well-tolerated
(Teronen, Heikkila et al. 1999). Bisphosphonates are small molecules, characterised
93.
by two C—P bonds (Heymann, Ory et al. 2004). They are also more specifically
referred to as geminal bisphosphonates when the two bonds are found on the same
carbon atom (Fleisch 1997). Quite similar to endogenous pyrophosphates,
bisphosphonates replace the O—P with a C—P, thereby allowing two additional
functional groups (Heymann, Ory et al. 2004), as well as creating a hydrolysis-
resistant P—C—P bond (Vasikaran 2001) (Figure 1.2). In terms of MMP-inhibition,
they are capable of binding to divalent metal ions, e.g. Zn2+, Ca2+ (Heikkila, Teronen
et al. 2002), through a three-dimensional structure that allows the coordination of one
oxygen from the phosphate group with the cation (Heymann, Ory et al. 2004). This
affinity for divalent cations can be increased even further if one of the functional
groups is either a hydroxyl (OH) or a primary amine (NH2), as this then facilitates the
formation of a tridentate conformation with the cation (Heymann, Ory et al. 2004). In
addition, the nitrogen atom in the side chain must be in a particular spatial
arrangement, as well as being a critical distance from the cation, to be the most potent
(Russell and Rogers 1999). Therefore, there has been a lot of work aimed at designing
new and improved bisphosphonates in recent years, with current numbers being
synthesised in the hundreds (Russell and Rogers 1999).
As mentioned previously, bisphosphonates have been shown to inhibit MMP-1, -2, -3,
-8, -9, -12, -13 and -20 at both therapeutically obtainable and, most importantly, non-
cytotoxic concentrations (Heikkila, Teronen et al. 2002). Current clinical applications
generally include the management of calcium and bone metabolism disorders, e.g.
osteoporosis, Paget’s disease, hypercalcaemia and metastatic cancer (Vasikaran 2001).
For a number of years now, these diseases have been effectively treated with
bisphosphonates. Indeed, Fosamax® (sodium alendronate, Merck & Co Inc.) has
recently been listed on the Australian Pharmaceutical Benefits Scheme (Australian
Government 2007). Therefore, it appears that for future indications, including those
primarily caused by significant soft tissue destruction, e.g. chronic ulcers,
bisphosphonates may be a promising treatment (Teronen, Heikkila et al. 1999). In
terms of chronic ulcer treatments, the delivery of specific bisphosphonates, or indeed
simply the presentation of this particular chemical to the site of excessive protease
activity, has to be carefully contemplated.
94.
Clinically, bisphosphonates can be classified into at least two groups with different
modes of action. Firstly, there are those that closely resemble pyrophosphate, e.g.
clodronate, which can be metabolically incorporated into non-hydrolysable analogues
of ATP, thereby altering the structure and potentially inhibiting ATP-dependent
intracellular enzymes. Next, there are the nitrogen-containing bisphosphonates, e.g.
alendronate, which are able to interfere with the mevalonate pathway, thus prevent the
synthesis of isoprenoid compounds, which can then lead to cellular apoptosis (Russell
and Rogers 1999). In addition, bisphosphonates also have the ability to chelate
divalent cations, which can also inhibit MMP activity by removing the zinc ions
required for catalytic activity and the calcium ions required for protein stability
(Boissier, Ferreras et al. 2000; Heikkila, Teronen et al. 2002; Neville-Webbe, Holen et
al. 2002). In this particular application, it is the bisphosphonate’s chelating ability that
is thought to decrease MMP activity in CWF.
Bisphosphonates have previously been modified for clinical delivery, especially in
terms of bone-specific applications (Uludag and Yang 2002; Wang, Miller et al. 2003;
Balas, Manzano et al. 2006). In addition, alendronate is a prime candidate
bisphosphonate for functionalisation as it contains a primary amine – a convenient
chemical bond that can be easily manipulated for conjugation into a polymer using
already-established methods (Wang, Miller et al. 2003). Through a nucleophilic acyl
substitution reaction, an amine, present in the alendronate sodium salt form, can be
reacted with an acid halide (Otera 1993) to produce an alendronate tethered to an
unsaturated carbon group. This alkene bond then allows for further polymerisation
into an established hydrogel system, such as that reported in Chapter 3. The
experiments in this chapter focus on the details surrounding this preliminary
alendronate condensation reaction, and then describe how this methacrylated-
alendronate can be copolymerised into the PHEMA/PEG hydrogels described in
Chapter 3. Furthermore, this bioactive wound dressing is tested for efficacy against
protease activity in CWF, and is assessed in further cell-based assays to confirm
biocompatibility with a human ex vivo skin model. Therefore, the specific aim of the
experiments in this chapter were to analyse whether a functionalised alendronate
hydrogel can be successfully synthesised, and in turn whether it can still inhibit MMPs
in CWF.
95.
4.2 MATERIALS AND METHODS
4.2.1 Chemicals
2-Hydroxyethyl methacrylate (HEMA, assay ≥ 99% GC, Sigma-Aldrich, St Louis,
MO, USA) and methacryloyl chloride (Sigma-Aldrich) were distilled under reduced
pressure immediately prior to use. Polyethylene glycol 20,000 (PEG, Sigma-Aldrich),
alendronate sodium salt (Merck, San Diego, CA, USA) and 4-methoxy phenol
(Sigma-Aldrich) were used as supplied under the manufacturers’ instructions. All H2O
used was double deionised by ion exchange (MilliQ, Millipore, Billerica, MA, USA).
4.2.2 Functionalisation of alendronate
In a round-bottom flask, alendronate sodium salt (300 mg, 0.92 mmol) was dissolved
in aqueous NaOH (148 mg, 3.70 mmol in 7.4 mL H2O) with 4-methoxy phenol to
inhibit polymerisation. The solution was cooled in an ice-salt bath then freshly
distilled methacryloyl chloride (120 mg, 1.15 mmol) and NaOH (111 mg, 2.768 mmol
in 5.5 mL H2O) was added step-wise, making sure that the pH remained above 11.
The reaction was stirred vigorously for 20 hours, then acidified with HCl to pH 7. The
resulting product was evaporated to dryness using a rotary evaporator and extracted
with chloroform (CHCl3) to remove any impurities. The water layer was then
precipitated with N,N-Dimethylformamide [DMF or (CH3)2NC(O)H] to obtain
methacrylated-alendronate.
4.2.3 Characterisation of methacrylated-alendronate through Fourier
Transform Nuclear Magnetic Resonance (FT-NMR) spectroscopy
1H and 31P spectra were recorded using a 400 MHz NMR Spectrometer (Bruker,
Ettlingen, Germany) at room temperature in deuterated water (D2O). Spectra were
analysed using MestReC Version 4.9.9.6 software (Mestrelab Research, Santiago de
Compostela, Spain).
96.
4.2.4 Analysis of MMP-inhibition through incubation with the methacrylated-
alendronate
A pooled sample of CWF was run on Collagen Type I zymograms as described
previously in Section 3.2.2, with the following exception. After the electrophoresis
had been performed, the zymograms were washed in triton X-100. Next, alendronate
(2 mM) and methacrylated-alendronate (2 mM of the equivalent bisphosphonate
segment of methacrylated-alendronate) were included separately in the zymogram
incubation buffer and the zymograms were incubated at 37 ºC for 24 hours. Collagen
Type I zymograms were then analysed as described previously in Section 2.2.3.
4.2.5 Synthesis of hydrogels
Aqueous solutions of distilled HEMA were prepared as previously described (50%
water, 30% PEG and 20% HEMA) with a vehicle control (no methacrylated-
alendronate) and two concentrations of methacrylated-alendronate (2 mM and 20 mM
of the equivalent bisphosphonate segment of methacrylated-alendronate in the
monomer solution). Solutions were then placed between two glass plates separated by
a silicone gasket and purged with argon. Each mould was approximately 75 mm x 50
mm x 3 mm and filled with 10 mL of monomer solution. The moulds were then
gamma-irradiated in a Gamma-cell 220 (Atomic Energy of Canada Ltd, Ottawa,
Canada) using a Co60 source at a rate of 3.25 kGy/h to give a total dose of 10 kGy
(Table 4.1). Hydrogels were removed from their moulds and cut into 1 cm x 1 cm
square pieces for further analysis.
Sample ID H2O (g) HEMA (g) PEG 20,000
(g)
Methacrylated-
alendronate (mg)
G 5.0 2.0 3.0 -
GA1X 5.0 2.0 3.0 7.68
GA10X 5.0 2.0 3.0 76.80
Table 4.1 Functionalised hydrogel formulations prepared for analysis as potential wound dressings
4.2.6 Analysis of polymerisation through NIR FT-Raman Spectroscopy
97.
Irradiated samples were analysed by a Perkin Elmer System 2000 NIR FT-Raman
spectrophotometer (Perkin Elmer, Waltham, MA, USA). The hydrogel samples were
placed into glass vials for spectroscopic analysis and the spectra were analysed using
Grams/AI (Thermo Electron Corporation, Waltham, MA, USA).
4.2.7 Analysis of MMP-inhibition through incubation with the alendronate-
functionalised hydrogels
Similar to that described in Section 3.2.2, a pooled sample of CWF was analysed
through Collagen Type I zymography with the following alteration. Alendronate-
functionalised hydrogels (as prepared in Section 4.2.5) were cut into small pieces and
included in the incubation buffer, along with a blank and a non-functionalised
hydrogel to act as a vehicle control, and then incubated at 37 ºC for 24 hours.
Collagen Type I zymograms were then analysed as described previously in Section
3.2.3.
4.2.8 Biocompatibility testing of hydrogels using the DED model
The DED model was prepared as described previously in Section 2.2.10 – 2.2.11. At
five days post air-liquid interface culture, the alendronate-functionalised hydrogels
were placed on top of the composite, along with “no treatment” and “non-
functionalised” hydrogels as controls. These wound dressings were exposed to the
composite for 7 days at 37 ºC/5% CO2, with the media being replaced at 3 days. At
completion of the treatment, a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide conversion (MTT) assay (Denizot and Lang 1986) was used to analyse one of
each treatment group. Briefly, the DED was submerged in 6 mL of 0.5 mg/mL MTT
dye (Sigma-Aldrich) and incubated at 37 ºC for 2 hours. Following incubation, the
metabolically active cells could be visualised by a purple colour and the DED were
photographed. Next, both the MTT-stained and un-stained samples were fixed and
paraffin-embedded using standard protocols for haematoxylin and eosin (H&E)
histological analysis.
4.2.9 Immunohistochemistry
98.
As described previously in Section 2.2.12, paraffin sections were cut and
deparaffinised in ethanol and xylene. Sections were probed separately for keratin
1/10/11 (K1/10/11), a marker for cornification and squamous cell differentiation; p63
(RDI Research Diagnostics, Concord, MA, USA), a p63 analogue that identifies
normal basal cells as opposed to malignant tumours; and cleaved caspase-3 (Cell
Signaling Technology, Danvers, MA, USA), a critical mediator of apoptosis in
mammalian cells. For slides that were probed for cleaved caspase-3, antigen
unmasking was required before the blocking step. Briefly, the slides were incubated at
37 ºC in 10 mM citrate buffer (pH 3.0) for 30 min, before proceeding as normal. After
incubation with primary antibodies for K1/10/11 (1:400), p63 (1:100) and cleaved
caspase-3 (1:100), sections were stained using a Dako Envision kit (Dako Denmark
A/S, Glostrup, Denmark) as per the manufacturer’s instructions, with the exception
that phosphate buffered saline (PBS) was used instead of Tris-buffered saline (TBS).
After antibody development, all sections were counterstained with haematoxylin and
analysed using light microscopy.
99.
4.3 RESULTS
4.3.1 Synthesis and characterisation of methacrylated-alendronate
Following the identification of excessive MMP levels in CWF as described in Chapter
3, potential MMP inhibitors were considered for incorporation into a wound dressing.
The bisphosphonate alendronate was chosen as it has been used for a number of years
in clinical practice (Vasikaran 2001) and possessed an easily modified primary amine.
This was important as the approach to be followed differed significantly from
previous approaches in that it was proposed to inactivate the proteases in the wound
fluid after the fluid is absorbed away from the actual wound bed. This strategy was
chosen so that proteases required for functions within the upper cellular layers of the
wound bed remain available. In view of this, the alendronate had to be modified to
allow eventual polymerisation into a hydrogel system. The primary amine in
alendronate appeared a logical place to perform the modification, and was reacted
with methacryloyl chloride to form the amide along with a pendant alkene bond
(Figure 4.1A). This unsaturated carbon then allowed for further polymerisation
(Figure 4.1A) into the already established aqueous PHEMA:PEG hydrogel system
outlined in Chapter 3.
Following the reaction between alendronate and methacryloyl chloride, the resulting
product required structural analysis. The two 1H NMR spectra reveal the original
alendronate, along with its methacrylated counterpart, which also contains some
contaminants from the synthesis protocol (Figure 4.2). The peaks have been assigned
and further details of the spectrum are outlined in Table 4.2. The four equivalent
protons on the phenol group of the 4-methoxy phenol used in the reaction as a
polymerisation inhibitor appear as a single peak at 7.84 ppm. In addition, the spectrum
reveals that there some reaction impurities, namely, methacrylic acid and
poly(methacrylic acid), which are still present even after multiple chloroform washes
and DMF precipitation. Two separate 31P spectra of the methacrylated products, along
with the original alendronate, are shown in Figure 4.3 to demonstrate the differences
following reaction work up and medium term storage. Again, the peaks have been
assigned and further details of the spectrum are outlined in Table 4.3. In Figure 4.3C a
second triplet is seen upfield of the first, which can be assigned to a dimer of the
100.
(A)
O
NH PO
OH OH
OHP
O OHONa
O
Cl NH2PO
OH OH
OHP
O OHONa . 3 H2O
NaOHIce-salt bathStir vigorously20 h
+ HCl . 3 H2O
+
(B)
O
O OH
O
NH PO
OH OH
OHP
O OHONa O
O
O
O OH
O
NH
OOH
O OH
O
OOH
P
PO OH
ONa
OOH OH
OHgamma-irradiation
+
Figure 4.1 Reaction schemes. (A) Condensation reaction of methacryloyl chloride and alendronate sodium salt. (B) Gamma induced polymerisation of HEMA and methacrylated-alendronate.
101.
(A)
(B)
(C)
Figure 4.2 FT-NMR 1H spectrum of methacrylated-alendronate. (A) Labelled schematic of the methacrylated-alendronate. (B) 1H spectrum of alendronate. (C) 1H spectrum of methacrylated-alendronate.
102.
Peak ppm Multiplicity, coupling constant (Hz)
Norm. Int Identity
1 7.84 s 0.15 CH on 4-methoxy phenol
c 5.60 m, * 1.00 Alkenyl H-trans to CH3 on product
5.562
5.56m, * 0.83
Alkenyl H-trans to CH3 on methacrylic acid
b 5.34 m, * 1.01 Alkenyl H-cis to CH3
on product
3 5.25 m, * 0.88 Alkenyl H-cis to CH3
on methacrylic acid
# 4.70 s 18.04 Water 3.213.20d,e 3.18
t, 6.8 2.01 CH2 on product adjacent to N
4,5 2.93 s 0.73 CH2 on polymer
4,5 2.77 s 0.56 CH2 on polymer 1.901.881.871.841.781.771.761.751.74
a 6
f,g h,i
1.73
m, * 10.41
Various CH3 on product (a), methacrylic acid, and other polymer forms
CH2 on product
* Coupling constants were unresolved
Table 4.2 Features of methacrylated-alendronate FT-NMR 1H spectrum. Columns are from left to right: peak number; shift (ppm); multiplicity (s, d, t, m) and coupling constant (Hz); normalised integral; and identity.
103.
(A)
(B)
(C)
(D)
Figure 4.3 FT-NMR 31P spectrum of methacrylated-alendronate. (A) Labelled schematic of the methacrylated-alendronate. (B) 31P spectrum of alendronate (C) 31P spectrum of product (D) 31P spectrum of worked up product with longer sample storage
104.
NH2PO
OH OH
OHP
O OHO
NH2P O
OHOH
OHP
OOHO
Figure 4.4 Potential structure of methacrylated-alendronate dimer
Peak ppm Multiplicity,
coupling constant (Hz)
Norm. Int Identity
21.5121.43a,b
21.35
t, 12.7 1.00 P on methacrylated- alendronate, split by CH2
21.1621.091 21.01
t, 11.8 0.35 P on methacrylated-alendronate dimer (Figure 4.4), split by CH2
Table 4.3 Features of methacrylated-alendronate FT-NMR 31P spectrum. Columns are from left to right: peak number; shift (ppm); multiplicity (s, d, t, m) and coupling constant (Hz); normalised integral; and identity.
105.
original methacrylated-alendronate formed through P—O—P bonds of the
bisphosphonate groups (Figure 4.4), and not the starting product as shown by the shift
upfield. This becomes more apparent after the work-up of the reaction, teamed with
longer term storage of the product (Figure 4.3B-C).
4.3.2 Analysis of MMP-inhibitory action of methacrylated-alendronate
To analyse the methacrylated-alendronate, an activity assay was used to compare its
MMP-inhibitory activity to that of the original alendronate. Through incubation of
CWF with increasing concentrations of either the original or the methacrylated-
alendronate, the proteolytic activity was revealed by the use of Collagen Type I
zymography (Figure 4.5A-B). This demonstrated that both forms of the alendronate
are able to inhibit CWF samples 1-6 to varying degrees. Densitometry was used to
quantify the reduction in proteolytic activity revealed by zymography and are
represented graphically in Figure 4.5C. Quantitatively, both inhibitors were able to
decrease the amount of Collagen Type I degradation, as compared to the untreated
CWF samples (p<0.01). From this functional assay, it appears that the addition of the
methacrylate group still allows for inhibition of MMPs at a physiological temperature
over 24 hours, although at a reduced level of function.
4.3.3 Preparation and characterisation of alendronate-functionalised hydrogels
Following successful inhibition of MMPs in CWF with the methacrylated-
alendronate, further polymerisation was required to form the functionalised wound
dressing. Three sets of hydrogels were successfully synthesised according to the
conditions outlined earlier (Table 4.1, Figure 4.1B). NIR FT-Raman Spectroscopy
demonstrated that a total dose of 10 kGy of gamma irradiation induced almost
compete polymerisation of the methacrylated-alendronate and HEMA monomer to its
copolymer form (Figure 4.6). This was evident by the absence of a characteristic C=C
band at 1639 cm-1 in the irradiated samples when compared with the monomer
mixture as described in Figure 3.1B. The hydrogel sheets were then characterised
using both a functional protease assay, along with exposure to a three-dimensional ex
vivo skin model to determine biocompatibility.
106.
(A) (C)
(B)
Figure 4.5 Collagen Type I zymography demonstrating inhibition of protease activity present in wound fluid samples by alendronate sodium salt and its methacrylated-counterpart. (A) Lanes 1-6 are CWF samples 1-6 (500 ng); Lanes 7-12 are CWF samples 1-6 (500 ng) and 2 mM of alendronate present in the incubation buffer at 37 ºC for 24 hours. (B) Lanes 1-6 are CWF samples 1-6 (500 ng); Lanes 7-12 are CWF samples 1-6 (500 ng) and 2 mM of methacrylated-alendronate present in the incubation buffer at 37 ºC for 24 hours. (C) Relative levels of protease activity in pooled CWF samples treated with respective inhibitors. The MMP-specific inhibition of collagen degrading activity was represented quantitatively through densitometric analysis. Levels are shown as the % collagen degrading activity as compared to the untreated samples, ± SEM (n=3). Statistical significance is relative to the untreated samples and shown as # (p<0.01) as determined by Tukey’s test.
107.
Figure 4.6 Analysis of alendronate-functionalised hydrogels using NIR FT-Raman Spectroscopy. Spectra indicate intensity/ Raman shift in cm-1. (A) Unirradiated H2O:HEMA:PEG (B) GA1X Irradiated H2O:HEMA:PEG:1X methacrylated-alendronate (10 kGy) (C) GA10X Irradiated H2O:HEMA:PEG:10X methacrylated-alendronate (10 kGy)
108.
4.3.4 Analysis of MMP-inhibitory action of alendronate-functionalised hydrogels
Following the demonstrated MMP-inhibitory activity of the methacrylated-
alendronate, this molecule was then co-polymerised into a hydrogel-based support to
examine whether the tethered form had similar properties. The alendronate-
functionalised hydrogels were analysed using Type I Collagen zymography to
determine if the tethered alendronate still demonstrated MMP-inhibitory action. The
hydrogels were cut into small pieces and incubated with the zymograms containing
the CWF, along with an untreated control. Collagen Type I zymography of the pooled
CWF samples showed a large amount of protease activity in the untreated control
(Figure 4.7A). However, when the hydrogel pieces were present, all three treatments
showed a visible decrease in the amount of Collagen Type I degradation (Figure
4.7A). When the treatments were analysed using densitometry to quantitate the
relative decrease in protease activity, the GA10X hydrogel, i.e. the hydrogel
containing 20 mM of methacrylated-alendronate, displayed the ability to significantly
reduce Collagen Type I degradation as compared with the non-hydrogel treated
control (p<0.01) (Figure 4.7B). These results suggest the tethered alendronate is still
able to inhibit MMPs as shown through a functional assay.
4.3.5 Biocompatibility testing using a three-dimensional human skin equivalent
A three-dimensional de-epidermised skin model was used to assess potential toxicity
effects from the alendronate-functionalised hydrogels as compared to the vehicle
hydrogel. The various hydrogels were applied to the stratified skin models for a seven-
day period, with histological and immunochemical staining carried out upon
completion of the experiment. Treatments included the vehicle control of hydrogel G,
along with the two alendronate-functionalised hydrogels, GA1X and GA10X.
Histological analysis showed a haematoxylin-stained basal layer, with an eosin-
stained cornified layer of relatively the same thickness throughout for all treatments
(Figure 4.8). For the GA10X hydrogel there appears to be minimal basal cells, along
with increased numbers of nucleated cells in the cornified layer. Hydrogel G and
GA1X also show some nucleated cells in the cornified layer, however, they do not
appear to be as numerous. When the sections were probed for specific skin markers,
keratin 1/10/11 and p63, and the apoptotic marker, cleaved caspase-3,
109.
(A) (B)
Figure 4.7 Collagen Type I zymography demonstrating inhibition of protease activity present in wound fluid samples by alendronate-functionalised hydrogels.(A) Lane 1 is the pooled CWF sample with no treatment in the incubation buffer. Lanes 2-4 are the same pooled CWF sample with polymer G, GA1X and GA10X respectively cut into small pieces and present in the incubation buffer at 37 ºC for 24 hours. (B) Relative levels of protease activity in pooled CWF samples treated with respective polymers. The MMP-specific inhibition of collagen degrading activity was represented quantitatively through densitometric analysis. Levels are shown as the % collagen degrading activity as compared to the control sample, ± SEM (n=3). Statistical significance is relative to the control sample and shown as # (p<0.01) as determined by Tukey’s test.
110.
Figure 4.8 Histological and immunohistochemical analysis of an ex vivo human skin model following exposure to the alendronate-functionalised hydrogel. From left to right the staining is: haematoxylin and eosin (H&E), keratin 1/10/11 (K1/10/11), p63 and cleaved caspase-3 (CC-3). The scale bar measures 20 µm.
111.
hydrogel revealed a minimal amount of p63 immunoreactive basal cells, confirming
the histological analysis. In the control and hydrogel G treatments, there is no visible
immunoreactivity of apoptotic cells, with minimal cell surface marker identification in
its functionalised counterparts, GA1X and GA10X. This indicates that the synthesised
hydrogels appear promising for potential wound dressing treatments, especially when
compared with the commercial wound dressing tested in Chapter 3 (Figure 3.8).
112.
4.4 DISCUSSION
From the results reported in Chapter 2, MMP-9 was identified as the predominant
protease present in CWF. Hence it is likely to have a key role in degrading the
extracellular matrix in the chronic wound environment. In addition, it appears that
high levels of MMP-9 activity in CWF are a stronger indicator of a wound's chronicity
than the total protease activity levels alone. Hence, it is not surprising that current
topical treatments of chronic ulcers with bioactives, i.e. growth factors, have proven to
be only slightly effective in treating this condition (Cullen, Watt et al. 2002). In view
of this, a specific MMP inhibitor, namely the bisphosphonate alendronate, was chosen
as a clinical treatment to aid in neutralising the aggressive proteolytic chronic wound
environment; the ultimate goal of this strategy being to modulate the ulcer towards a
healing state. The bisphosphonate family, which exhibits low toxicity and has been
well tolerated for several years of human use, seems to be the ideal candidate for
MMP-related diseases (Teronen, Heikkila et al. 1999). The wound treatment explored
in this study examined the use of bisphosphonates to inhibit MMP activity in CWF.
Importantly, the approach described differs significantly from previous approaches in
that the proteases are inactivated in the wound fluid, after the fluid is absorbed away
from the actual wound bed, so that proteases required for functions within the upper
cellular layers of the wound bed remain available.
Alendronate was chosen as the MMP-inhibitor for inclusion into the hydrogel,
primarily due to the fact that it contains both amine and hydroxyl functional groups,
which together can enhance the bisphosphonate’s affinity for divalent cations
(Heymann, Ory et al. 2004). However, it first has to be modified to allow further
polymerisation into the synthetic wound dressing. Because alendronate is insoluble in
common organic solvents, a two phase reaction called the Shotten-Baumann reaction
was chosen to react an acid chloride with the amine of alendronate. This reaction is
well characterised and allows high yields of amines, mainly due to the fact that they
are able to successfully compete with other reactive species present (Kemp and
Vellaccis 1980). Briefly, this is a two phase reaction with the acid chloride suspended
in the water phase containing the amine compound. At the boundary of the two
phases, the amine reacts with the acid chloride, thereby generating the desired
product. The acid (HCl) generated from the reaction of the acid chloride must be
113.
neutralised by a base (in this case, NaOH) in order to prevent the conversion of the
amine to the non-nucleophillic ammonium salt. Unfortunately, one of the unavoidable
by-products of the Schotten Baumann reaction is the hydrolysis of the acid chloride to
generate the corresponding acid, in this case, methacrylic acid. Typically, this can be
removed during work-up. However, due to the small scale of the reaction, not all of
this impurity could be removed in the reaction reported here. In addition, the 31P
spectra showed an additional phosphorus-containing species in the resulting product,
possibly due to the dimerisation of two functionalised alendronate molecules through
the –OH group on the phosphate. Similar dimerisation has been shown with a similar
phosphate-containing molecule, i.e. ethylene glycol methacrylate phosphate (MOEP)
(Suzuki, Whittaker et al. 2007). Following the reaction to modify alendronate, the
resulting product was then copolymerised into the PHEMA hydrogel system
previously outlined in Chapter 3. FT-Raman spectroscopy revealed a complete
polymerisation, thereby ensuring that no functionalised alendronate could be released
from the hydrogel.
Critically, it is demonstrated that the alendronate tethered to a hydrogel, along with its
methacrylated-counterpart, were able to significantly decrease the levels of Type I
Collagen degradation in CWF as compared to the untreated control (p<0.01).
Interestingly, the vehicle hydrogel, i.e. the hydrogel without the methacrylated-
alendronate, was also able to inhibit MMP activity as revealed through the Collagen
Type I zymography, but to a lesser degree than the most concentrated alendronate-
functionalised hydrogel, GA10X. A possible explanation for this is the chemical
composition of PHEMA itself, as it contains 3 oxygen atoms in its unit structure that
are all available for chelation (Zainuddin, Hill et al. 2006). Further evidence for this
hypothesis is that Ca2+ itself has a strong tendency to chelate to oxygen atoms (Levine
and Williams 1982). Therefore, through the use of a “self-chelating” hydrogel support,
the vehicle treatment (hydrogel G) has the ability to chelate the cations required by
MMPs for their stability and catalytic activity, e.g. Ca2+, Zn2+. This chelating ability is
then disrupted by copolymerisation with the methacrylated-alendronate at the lowest
concentration (hydrogel GA1X).
Zainuddin et al. (2006) previously postulated that by reducing the number of available
oxygen atoms, the chelating mechanism is thereby reduced, and may then prevent the
114.
chelation of metal ions with PHEMA. Furthermore, Chirila et al. (2007) report
massive calcification of PHEMA, both in vitro in simulated body fluid and in vivo as
subcutaneous implants (Chirila, Zainuddin et al. 2007). However, when the highest
concentration of methacrylated-alendronate is copolymerised into the system
(GA10X), the chelating ability of alendronate alone is stronger than the PHEMA
support and overcomes the less MMP-inhibitory response seen previously. Therefore,
taken together, these results indicate that the alendronate is still available to inhibit the
MMPs in the CWF, while not being released into the wound bed. This is a significant
advance over topically applied inhibitors as it allows MMPs to remain active in the
wound bed where they perform vital roles in growth factor activation and immune
system regulation (Levi, Fridman et al. 1996; Suzuki, Raab et al. 1997; McQuibban,
Butler et al. 2001; Gearing, Thorpe et al. 2002), yet at the same time the unwanted
MMPs, those in the CWF, are absorbed into the hydrogel where they are inactivated.
Furthermore, in this chapter, the experiments with the DED model showed that there
were minimal differences between the three treatments, i.e. the vehicle and two
alendronate-functionalised hydrogels, in terms of impact on visible structure and
effect on a range of skin cell surface expression markers. An interesting point to note
is that with both wound dressing treatments, nucleated cells became visible in the
previously anuclear cornified layer. As discussed previously, this is due to the
application of a topical dressing inducing an increased amount of culture medium into
the dressing, and therefore converting the air-liquid interface to a liquid-liquid
interface. Indeed, this may well underlie the well accepted evidence that a moist
wound environment is essential for optimal wound healing (Winter 1962). In terms of
skin cell surface expression markers, the three treatments appeared relatively similar
when probed with the skin differentiation marker keratins 1, 10 and 11 and the basal
cell marker, p63. The main difference observed between the three treatments being the
minimal immunoreactivity seen with p63 in the GA10X hydrogel. This is likely to be
alleviated by first washing the alendronate-functionalised hydrogels, thereby
minimising any potential “stickiness” of the phosphonate polymer component.
Apoptosis in the keratinocytes was also analysed by immunoprobing for cleaved
caspase-3 – a major effector caspase that is specific for keratinocytes undergoing
apoptosis and not just those undergoing terminal differentiation (Xue, Campbell et al.
2007). None of the three treatments showed obvious levels of immunoreactivity,
115.
which appears promising for future clinical treatments, especially in terms of
developing MMP-inhibiting, CWF-absorbing functionalised hydrogels. In the future,
efficacy of the hydrogel wound dressing in promoting wound healing will be assessed
using a HSE model that has artificially created wounds, e.g. punch biopsies, and that
exhibits impaired healing in the presence of exogenous CWF. This would provide
further information to confirm both the safety and efficacy of the dressings before
proceeding to an initial human clinical trial.
116.
4.5 CONCLUSION
Taken together, the results reported herein suggest this novel alendronate-
functionalised hydrogel holds promise as a dressing treatment for chronic wounds.
This is due to its ability to inhibit the excessive levels of MMPs in CWF, as well as its
biocompatibility when exposed to a human ex vivo skin model. Of importance to the
development of this wound dressing was the ability to attach a functional group to the
alendronate sodium salt, which then allowed polymerisation into the already-
established hydrogel system. This was achieved through a Shotten-Baumann reaction,
which permitted high yields of the amide in a single-step reaction. Furthermore, it has
been demonstrated that the MMPs in CWF can be inhibited by the original
bisphosphonate alendronate, a methacrylated-analogue, and in a tethered state, i.e.
attached to a hydrogel support. This last point is critical to the development of a
successful chronic ulcer wound dressing as it inhibits the MMPs in CWF, while still
allowing those on the wound bed itself to perform their essential functions in wound
healing, namely, activation of growth-promoting agents and immune system
modulation.
CHAPTER 5
GENERAL DISCUSSION
118.
Chronic ulcers are a very controversial topic in terms of both pathophysiology and
optimal treatment practices. Considering the significant burden that they place on
many healthcare systems around the world, it is surprising that more is not known
about this costly medical challenge. Current prevalence rates of chronic wounds in
Australia are estimated to be between 13-3% of the population/year and cost the
Australian community upwards of AU$500 million per annum to treat (Baker and
Stacey 1994; Gruen, Chang et al. 1996). These wounds are a major cause of grief and
concern for affected individuals and also contribute significantly to their overall
diminished quality of life (MacLellan 2000). However, at the commencement of the
studies described herein, very little research was occurring in Australia to investigate
new treatments, or even at a more general level, few groups were interested in
identifying potential causes of these ulcers themselves.
As a result of this situation, the aim of this thesis was to identify potential proteolytic
markers of chronic ulcers and, from this information, design a bioactive wound
dressing to modulate these ulcers towards a healing state. This thesis focussed on
proteases as their excessive levels have been identified by a number of studies as
being the main difference between chronic and acute wounds (Tarnuzzer and Schultz
1996; Ladwig, Robson et al. 2002; Li and Li 2003). Therefore, this dissertation
involved the synthesis and development of a novel hydrogel wound dressing that
would interact with the chronic wound environment to promote healing.
Characterisation of this hydrogel employed numerous techniques including, FT-
Raman spectroscopy, swelling studies, protein uptake and release, along with cell
based assays in both two- and three-dimensional environments. The main objective
was to also identify specific protease inhibitors that may be applied to chronic wounds
in order to attenuate excessive proteolytic activity and restore normal healing. As
reported herein, it has been ascertained that MMP-9 is indeed the most abundant
protease activity in non-healing wounds, revealed through Collagen Type I
zymography, immunoprecipitation and a direct ELISA. In addition, after identifying
bisphosphonates, namely alendronate, as MMP inhibitors, an alendronate-containing
functionalised wound dressing was created that holds significant promise as a novel
strategy to modulate and enhance the healing of chronic wounds. Therefore, this thesis
contains a number of novel results that confirm the hypotheses that were outlined in
Chapter 1.
119.
Chronic wound healing is an extremely complex issue due to the lack of knowledge
surrounding the cascade of events leading to a wound’s nonhealing status. Venous
hypertension itself has been indicated to lead to capillary bed distension, resulting in
the leakage of large molecules, namely fibrinogen, from the blood into the dermis
(Falanga 1993). This can then lead to the formation of fibrin cuffs in the surrounding
dermis, which in turn cause ulceration due to the impaired delivery of oxygen and
other nutrients to the surrounding tissue (Falanga 1993). In addition, more recent
studies have shown that the in vivo capillary bed distension is extremely complicated,
with some fibrin cuffs being shown to contain highly organised ECM molecules.
These are also often found in other inflammatory diseases, e.g. ulcers of the oral
cavity (Chen and Rogers 2007). Therefore, while this thesis focuses on proteolytic
causes of an ulcer’s chronicity, it is important to note that there are many processes
involved in chronic wounds and proteases are simply one of the major contributors.
That being said, chronic wounds are commonly characterised by elevated levels of
pro-inflammatory cytokines and proteases (Chen, Schultz et al. 1999). These wounds
also contain reduced levels of TIMPs (Chen, Schultz et al. 1999; Baker and Leaper
2000). The majority of previous studies involved with analysing the chronic wound
environment use CWF as their clinical sample, due to the difficulties in obtaining a
wound bed biopsy. However, the one study that did use cells isolated from a wound
bed biopsy, instead of the fluid itself, reported significantly different results. Cook et
al. (2000) reported that significantly decreased levels of active MMPs, along with
marked increases in expression of TIMPs, led to the chronicity of non-healing wounds
(Cook, Stephens et al. 2000). This is in direct contrast with all other previously
reported findings, and indeed with the results reported in Chapter 2. Therefore, while
these differences were taken into account, for the study reported herein, CWF was
chosen for exploration as a possible indicator of the chronic wound environment. This
was due to the goal of this thesis project being to explore non-invasive diagnostic
approaches and also, develop potential future wound dressings that could modulate
this proteolytic environment.
There is much controversy surrounding MMP activity in CWF. A number of studies
suggest that elevated MMP-2 and -9 levels are detrimental to the wound healing
process (Wysocki, Kusakabe et al. 1999; Ladwig, Robson et al. 2002), while others
120.
recommend that they are essential for normal wound healing to occur (Cook, Stephens
et al. 2000). In particular, Ladwig et al. (2002) reported that levels of activated MMP-
9 tended to decrease as healing proceeded (Ladwig, Robson et al. 2002). Furthermore,
Wysocki et al. (1999) demonstrated that expression of MMP-9 in CWF was 32-fold
higher than in blood-derived serum, and activity was also increased in early collection
times (Wysocki, Kusakabe et al. 1999). As most members of the MMP family are
structured into three essential, characteristic and highly conserved domains (Massova,
Kotra et al. 1998), it is often quite difficult to differentiate MMP species based on
molecular weight alone. In addition, antibodies that are based on specific MMPs often
find it difficult to differentiate between closely related species, as there is a very close
evolutionary relationship within the family further reflected by their highly conserved
domains (Somerville, Oblander et al. 2003). Indeed, this is likely to be the main
reason for the controversy surrounding levels of specific MMPs in chronic wounds.
In view of these previously conflicting results, the aim of Chapter 2 was to use
multiple techniques to identify specific MMPs species in CWF across a number of
samples. Initially, total protease levels were analysed using both Collagen Type I and
Collagen Type IV zymography. All CWF samples revealed increased levels of
collagen-degrading activity compared to that found in both HS and AWF samples. In
addition, this activity could be attenuated through the use of a MMP-specific inhibitor,
namely GM6001. Having obtained these results, specific MMPs were
immunoprecipitated from the wound fluid samples and then analysed using a
functional assay to minimise any non-specific immunoreactivity. MMP-9 was
subsequently identified as the major protease responsible for collagen degradation by
CWF. Next, the clinical status of the ulcer was considered; with the CWF samples
separated into both higher and lower PUSH score groups. Through
immunoprecipitation, CWF samples in the lower and higher PUSH score groups were
indistinguishable, as both groups showed a significant increase in MMP-9 activity
compared with HS and AWF samples (p<0.01).
To discern if there were potential differences in clinical groupings, a direct ELISA
was used with a different MMP-9 antibody to determine quantitative levels of MMP-9
in these samples. This technique revealed clear difference between the levels of
MMP-9 present in the two separate PUSH score groupings. The lower PUSH score
121.
(PUSH ≤11) group displayed significantly higher levels of MMP-9 than AWF
(p<0.01), as well as significantly lower levels of MMP-9 than the higher PUSH score
(PUSH ≥12) group (p<0.01). As a result, it appears that there is a strong correlation
between increased MMP-9 levels in CWF from wounds in the higher PUSH score
(PUSH ≥12) group compared to AWF samples (p<0.01). Therefore, high levels of
MMP-9 activity in CWF may be a good indicator of a wound's chronicity.
A further characteristic of chronic wounds is reduced expression of growth-factors
(Stadelmann, Digenis et al. 1998). Many current topical treatments for chronic ulcers
attempt to rectify the decreased levels of growth-factors, but are only minimally
effective (Cullen, Watt et al. 2002). Evidence from previous reports, combined with
data from Chapter 2, indicates that this is due to the excessive levels of protease
activity, i.e. MMP-9, found in CWF (Tarnuzzer and Schultz 1996; Stadelmann,
Digenis et al. 1998; Trengove, Stacey et al. 1999; Cullen, Watt et al. 2002). There are
a number of specific synthetic inhibitors of MMP, namely tetracyclines and their
chemically modified derivatives (Ramamurthy, Kucine et al. 1998; Ramamurthy,
McClain et al. 1999; Yager and Nwomeh 1999); hydroxamic acids (Lund, Romer et
al. 1999); and bisphosphonates (Valleala, Hanemaaijer et al. 2003). The
bisphosphonates were chosen for potential use in modulating the wound environment
as they have been used in clinical practice for a number of years in conditions such as
Paget’s disease and hypercalcaemia, both of which are linked to excessive MMP
activity (Vasikaran 2001). Due to their low toxicity, bisphosphonates are generally
well-tolerated (Teronen, Heikkila et al. 1999), and have been shown to inhibit various
MMPs at therapeutically obtainable and non-cytotoxic concentrations (Heikkila,
Teronen et al. 2002).
The second aim investigated in this project was that a synthetic hydrogel, in particular
one created using PHEMA, represented an ideal wound dressing for chronic ulcers.
This hypothesis was supported by the findings in Chapter 3 revealing that an aqueous
PHEMA hydrogel synthesised in the presence of PEG was able to demonstrate
similar, if not improved, properties compared with a commercially available wound
dressing in both two- and three-dimensional cell culture systems. Furthermore, this
synthetic hydrogel showed complete polymerisation at a total gamma irradiative dose
of 10 kGy, displayed the ability to swell with water, and, could take up and release
122.
physiologically relevant amounts of a biologically active protein. Interestingly, initial
reports from Carenza et al. (1993) proved to be inaccurate in their claims that PEG
acted as a crosslinker between the PHEMA backbone chains when synthesised using
gamma irradiation (Carenza, Lora et al. 1993). Instead, it appears that the crosslinked
hydrogels are formed from small amounts of residual EDGMA present in the HEMA
monomer. However, at higher total doses of irradiation, e.g. 100 kGy, there is the
potential for the alkoxy radical to form on the end of the PEG molecule to allow
copolymerisation. Indeed, Abd Alla et al. (2004) were able to form copolymers of
PVA and PEG with doses ranging up to 200 kGy (Abd Alla, Said et al. 2004).
A three-dimensional ex vivo skin equivalent model was also used to evaluate the
synthetic hydrogels against a readily available commercial wound dressing. An
interesting point to note is that with all three wound dressing treatments, nucleated
cells became visible in the previously anuclear cornified layer. A potential explanation
for this is that the application of a topical dressing converts the air-liquid interface of
the three-dimensional skin model to a liquid-liquid interface caused by culture
medium being absorbed through the dermis and into the wound dressing. This has
previously been shown by others in a similar model in which the cells did not
completely differentiate when the model remained covered with culture medium
(Ohsawa, Maruyama et al. 1999).
Commercially, there are several hydrogel wound dressings on the market. Some
examples of these include: IntraSite™ gel (Smith & Nephew, Hull, UK); Nu-Gel®
(Johnson & Johnson, New Brunswick, NJ, USA); Curasol® (Healthpoint, Fort Worth,
TX, USA); and Tegaderm™ (3M, St Paul, MN, USA). Hydrogels are often selected as
an ideal wound dressing as they are quite cooling on the skin and do not stick to the
wound itself – factors that aid in minimising pain to the patient (Queen, Orsted et al.
2004; Akita, Akino et al. 2006). Therefore, it appears that hydrogels address the
important issues of patient compliance, pain management and the individual’s quality
of life (Queen, Orsted et al. 2004).
More recently, First Water (Wiltshire, UK) has developed a novel range of hydrogels
to overcome initial performance issues with hydrogels in a clinical setting (First-
Water 2004). In terms of chronic leg ulcers, their product, Hydroform Cool, aims to
123.
have finely tuneable properties to allow optimal water transport properties (First-
Water 2004). This means that they are able to create a favourable chronic wound
environment by maintaining a fine balance between absorbing the wound exudate, and
donating moisture to ensure the wound bed itself does not dry out and delay wound
healing. When considering this together with the results reported in Chapter 2, the
application of a topical hydrogel may even stimulate the keratinocytes to migrate
faster through the cornified layers. Indeed, this may well underlie the well accepted
evidence that a moist wound environment is essential for optimal wound healing
(Winter 1962) and explain why hydrogels appear beneficial as wound dressings.
First Water have also developed a proprietary hydrogel technology that claims to
provide antimicrobial properties against a range of microbes, namely, Candida
albicans, Aspergillus niger, Pseudomonas aeruginosa, Staphylococcus aureus and
Escherichia coli (Sparrow 2005). Furthermore, they assert that their patented
technology, i.e. a copolymer of 2-acrylamido-2-methyl- propanesulphonic acid
sodium salt (NaAMPS) and acrylic acid (3-sulphopropyl)ester potassium salt (SPA)
allows for improved overall patient outcomes. This is mainly due to their claims that it
is able to selectively adhere to the patient’s skin, i.e. it is able stick to the wound’s
peri-skin, but not to the wound bed itself (Munro and Yasin 2002). Such advances in
patient-focussed wound dressing treatments have not only enhanced wound dressing
compliance, but have also improved the patient’s overall quality of life.
More sophisticated ulcer therapies include the recently developed active wound
management treatments, e.g. growth factors (Regranex, Johnson & Johnson) and
dermal substitutes (Dermagraft, Smith & Nephew). However, the major obstacle for
patients obtaining effective growth factor treatments is price – Regranex® costs
upwards of US$375/15 g tube (Ugarte, Roberts et al. 2002), which considering the
daily treatment regime, can equate to thousands of dollars per wound. Moreover, even
though the cost of these treatments is extremely prohibitive, they have only proven to
be slightly effective in treating chronic wounds since high levels of proteases in CWF
can rapidly degrade these topically-applied growth-promoting agents (Cullen, Watt et
al. 2002). Therefore, the toxic chronic wound environment has to be analysed further
before designing an effective and affordable interactive dressing to modulate the
chronic wound environment to that of a healing ulcer.
124.
In view of previous reports, as well as the evidence displayed in Chapter 2, a specific
MMP inhibitor, namely the bisphosphonate alendronate, was chosen as a treatment to
aid in neutralising the aggressive proteolytic chronic wound environment. Of note, the
wound treatment explored in the study described herein differs significantly from
previous approaches in that the goal was to inactivate proteases in the wound fluid,
after the fluid is absorbed away from the actual wound bed, so that proteases required
for functions within the upper cellular layers of the wound bed remain available.
Indeed, it is demonstrated in Chapter 4 that alendronate tethered to a hydrogel, along
with its methacrylated counterpart, were able to significantly decrease the levels of
Type I Collagen degrading activity in CWF as compared to the untreated control
(p<0.01). This indicates that the alendronate is still available to inhibit the MMPs in
the CWF, while not being released into the wound bed. This approach represents a
significant advance over topically applied inhibitors as it allows MMPs to remain
active in the wound bed where they perform vital roles such as activation of select
growth promoting factors and in immune system regulation (Levi, Fridman et al.
1996; Suzuki, Raab et al. 1997; McQuibban, Butler et al. 2001; Gearing, Thorpe et al.
2002), yet inactivates the abundant proteases in the wound fluid.
Pharmacologically, bisphosphonates act in a number of ways to inhibit MMP activity.
When used in vivo as an injectable or through oral administration, MMPs are able to
interact with two separate metabolic pathways depending on their structure (Russell
and Rogers 1999). Nitrogen-containing bisphosphonates, e.g. alendronate, are able to
inhibit enzymes of the mevalonate pathway, which can then affect further downstream
processing events, e.g. post-translational modifications of small GTPases (Russell and
Rogers 1999). Mevalonate is a precursor for isoprenoid intermediates and therefore is
responsible for a number of end-products, namely, cholesterol, farnesylated proteins
and geranylgeranylated proteins. Alterations in the end products, e.g. mutations in
GTP that make it resistance to hydrolysis by GTPases, can then constitutively activate
both the PI3K/AKT and Raf/MEK/MAPK/ERK pathways and cause disruption in
both apoptosis and cell cycle progression, both of which are commonly seen in a
variety of cancers (Swanson and Hohl 2006). In terms of this particular application,
due to the fact that the bisphosphonate is not intended to be released into the wound
bed, the main mode of action for MMP-inhibition will be through cation chelation. As
mentioned previously, MMPs are zinc-dependent endopeptidases, which also contain
125.
a number of calcium ions for stability (Parks 1999). Bisphosphonates are able to bind
to divalent metal ions, e.g. Zn2+, Ca2+ (Heikkila, Teronen et al. 2002), through the two
phosphate groups to form a three-dimensional structure between one oxygen from the
phosphate group and the cation itself (Heymann, Ory et al. 2004). Furthermore,
previous studies have also shown that the addition of high levels of either Zn2+or Ca2+
to in vitro cell culture systems can reduce the inhibitory activity of bisphosphonates
towards MMPs (Boissier, Ferreras et al. 2000; Heikkila, Teronen et al. 2002; Neville-
Webbe, Holen et al. 2002). Moreover, when bisphosphonates are administered orally,
taking EDTA, a cation chelator, at the same time can increase bisphosphonate
absorption (Vasikaran 2001).
As mentioned previously, one of the major obstacles to the average patient receiving
the optimal level of wound care is cost (Eldor, Raz et al. 2004). Even though many
studies have shown that these advanced treatments are cost-effective in the longer-
term (Kantor and Margolis 2001), governments are still slow to subsidise these
treatment regimes. Therefore, for a wound dressing to be effective in the individual
patient, as well as the community, its development has to be focussed on minimising
expensive costs to the user. The wound dressing outlined in this thesis has paid close
attention to this important issue. Alendronate itself has been used for a number of
years in clinical settings to attenuate MMP activity (Russell and Rogers 1999;
Teronen, Heikkila et al. 1999; Vasikaran 2001), as well as recently being added to the
Australian Pharmaceutical Benefits Scheme (Australian Government 2007) in the
form of its sodium salt, Fosamax® (Merck & Co, Inc, Whitehouse Station, NJ, USA).
Furthermore, this compound reaches the end of its life of patent protection in 2008,
which will then encourage an increased number of generic products. This recent
development will ensure that price and availability of this compound is within the
reach of the ordinary chronic ulcer sufferer. In addition, PHEMA itself has been used
for specific biomedical applications in the field of contact lenses for over forty years
(Alvarez-Lorenzo, Yanez et al. 2006), thus this hydrogel support is also widely
available and cost-effective. Furthermore, as this is a completely synthetic wound
dressing, it minimises transport and storage issues associated with biological
compounds, which are also responsible for increased consumer prices.
126.
It is clear that the studies outlined in this thesis have led to a range of novel findings in
the fields of chronic wound healing and wound dressing development. However, it has
also raised many issues that need further clarification. Most importantly, there is a
need to further investigate the chronic wound environment in its entirety, i.e. in both
the chronic wound fluid and the wound bed. This is needed as while many studies
focus on CWF as an indicator of the wound’s clinical status (Ladwig, Robson et al.
2002), little is known about the actual wound bed environment. The main reason for
this is that ulcer biopsies are extremely invasive, as they require the creation of a new
wound in an already non-healing ulcer. Therefore, care needs to be taken to optimise
the information generated from these smaller patient studies. Furthermore, it has to be
noted that both of these samples, i.e. CWF and wound bed biopsies, are likely to be
extremely heterogeneous diagnostic samples. Clearly, multiple patient samples need
to be analysed to gain statistically significant results. In addition, improved models
need to be developed – to allow initial studies to be performed in in vitro or ex vivo
environments, while still providing relevant information that can be correlated to
human in vivo situations, e.g. a three-dimensional ex vivo skin model with impaired
wound healing properties. In terms of proteases specifically, further investigation is
required into the other MMPs present in CWF samples. In Chapter 3, MMP-9 has
been identified as the predominant protease in CWF and has also been shown to
provide a good indicator of the wound’s clinical status. However, other MMPs are still
present in these depleted samples. Therefore, complete MMP-9 depletion of the
samples through a series of immunoprecipitation steps could reveal less abundant
MMPs, and these may well provide improved biomarkers of ulcer prognosis.
Once the complete protease profile of chronic wounds has been investigated, further
development of the wound dressing itself is required. Hence, future studies will
investigate the reaction between methacryloyl chloride and alendronate in more detail,
along with examining other potential chemical linkages between a synthetic hydrogel
and the chosen bisphosphonate. In terms of future in vitro work, these functionalised
alendronate hydrogels will be analysed with a particular focus on potential extracts, in
a similar manner to the initial primary keratinocyte toxicity testing. For the purpose of
this study, they were only examined in a three-dimensional ex vivo skin model as this
a more representative of their intended future human biomedical application.
Furthermore, longer-term stability testing of this chemical bond needs to be performed
127.
in a simulated wound environment to ensure that no alendronate will be released into
the wound bed. This point is critical as the only wound bed-based patient samples that
have been analysed thus far revealed that there were decreased levels of MMPs in the
wound bed compared with the CWF (Cook, Stephens et al. 2000). Therefore, the
release of MMP inhibitors in the wound bed could potentially impede the wound
healing process even further. In addition, recent research conducted by others in our
Program has been directed at linking the alendronate to a synthetic hydrogel through
custom-made PEG molecules. This approach may provide better flexibility and an
enhanced method for controlling the location of the tethered alendronate within the
dressing, thereby improving its potential clinical applications.
The results outlined in this thesis, which examined the relationship between the
clinical status of an ulcer and its resultant wound fluid, along with improved wound
dressing developments, have provided further insight into the chronic wound healing
environment and also provided the stimulus for further research within the Tissue
Repair and Regeneration Program at QUT. Of most significance, an Australian
Provisional Patent (2007903101) was recently filed by Tissue Therapies Limited on
behalf of QUT to protect the invention of a tethered bisphosphonate wound dressing
to promote chronic ulcer healing. Moreover, along with the IHBI ECR grant
mentioned previously, work has also been undertaken in a joint Australian Research
Council (ARC) Linkage project with the University of Queensland and Tissue
Therapies Limited to further investigate the concept of a bioactive wound dressing
that can deliver growth factors and, at the same time, inhibit protease activity in CWF.
Furthermore, Prof Graeme George is also co-supervising a more recent PhD student,
Jason Yeh, who is investigating the use of different molecular weight PEG molecules
to act as porogens in PHEMA hydrogels synthesised through gamma irradiation. The
main differences in this project is the deliberate addition of EDGMA to act as a
crosslinker, which then allows for further investigation into the crosslinking itself.
It is interesting to note that there are indeed potential other applications for this
technology. For example, there are many situations where excessive protease activity
may be detrimental to medical implants and biomaterials, a common case being
aseptic prosthesis loosening in total hip replacements (Kido, Pap et al. 2007). For this
particular application, a protease inhibitor that is tethered to the implant or biomaterial
128.
itself may provide improved performance of the material and minimise implant
failures by attenuating local proteolytic activity. Alternatively, an MMP-inhibitor
tethered to non-absorbable surgical sutures could also provide patients with an
improved healing outcome. In this situation, by inhibiting MMPs around the sutures
the overall wound breaking strength may potentially be improved, thereby reducing
the time for the wound to heal completely. In fact, Witte et al. (1998) previously
performed a similar MMP-inhibitor study on Sprague-Dawley rats and showed that
mechanical strength of the surgical incision was improved with inhibition of MMPs,
while not causing an increase in collagen deposition (Witte, Thornton et al. 1998).
From this information, it is clear that there are many other potential biomedical
applications for this technology.
In summary, the in vitro results reported in this thesis suggest that this novel
functionalised hydrogel holds promise as a dressing treatment for chronic wounds. As
demonstrated herein, MMPs are the major class of proteases responsible for the
degradation of collagen in the wound bed. Furthermore, this activity can be largely
attributed to the high levels of MMP-9 present in CWF. In addition, levels of MMP-9
correlate with a wound’s clinical status. Importantly, MMP-9 can be inhibited with the
bisphosphonate alendronate, in the form of a sodium salt, a methacrylated analogue,
and in a tethered state. This last point is critical in terms of development of an optimal
wound dressing treatment that can inhibit MMPs in the wound fluid, but will still
allow MMPs to perform biological functions important to healing in the wound bed
itself, such as activation of growth promoting factors. In conclusion, the emphasis on
minimising potential adverse side-effects by tethering the MMP inhibitor, rather than
releasing, or topically applying, the inhibitor to the wound bed, allows for an
improved, more device-orientated wound dressing treatment for these chronic ulcers.
CHAPTER 6
REFERENCES
130.
Abd Alla, S. G., H. M. Said, et al. (2004). "Structural properties of gamma-irradiated poly(vinyl alcohol)/poly(ethylene glycol) polymer blends." Journal of Applied Polymer Science 94(1): 167-176.
Akita, S., K. Akino, et al. (2006). "A polyurethane dressing is beneficial for split-thickness skin-graft donor wound healing." Burns 32(4): 447-451.
Allen, F. D., C. F. Asnes, et al. (2002). "Epidermal growth factor induces acute matrix contraction and subsequent calpain-modulated relaxation." Wound Repair and Regeneration 10(1): 67-76.
Almany, L. and D. Seliktar (2005). "Biosynthetic hydrogel scaffolds made from fibrinogen and polyethylene glycol for 3D cell cultures." Biomaterials 26(15): 2467-2477.
Alvarez-Lorenzo, C., F. Yanez, et al. (2006). "Imprinted soft contact lenses as norfloxacin delivery systems." Journal of Controlled Release 113(3): 236-244.
Australian Government. (2007). "Alendronate Sodium." Retrieved 24 April 2007, 2007, from http://pbs.gov.au/html/consumer/search/results?medicine=Fosamax%2040%C2%A0mg&publication=GE.
Azzam, H. S. and E. W. Thompson (1992). "Collagen-induced activation of the Mr 72,000 type-IV collagenase in normal and malignant fibroblastoid cells." Cancer Research 52: 4540-4544.
Baker, E. A. and D. J. Leaper (2000). "Proteinases, their inhibitors, and cytokine profiles in acute wound fluid." Wound Repair and Regeneration 8(5): 392-398.
Baker, S. R. and M. C. Stacey (1994). "Epidemiology of chronic leg ulcers in Australia." Australia and New Zealand Journal of Surgery 64: 258-261.
Balas, F., M. Manzano, et al. (2006). "Confinement and Controlled Release of Bisphosphonates on Ordered Mesoporous Silica-Based Materials." Journal of the American Chemical Society 128(25): 8116-8117.
Berger, J., M. Reist, et al. (2004). "Structure and interactions in covalently and ionically crosslinked chitosan hydrogels for biomedical applications." European Journal of Pharmaceutics and Biopharmaceutics 57(1): 19-34.
Bergers, G., R. Brekken, et al. (2000). "Matrix metalloproteinase-9 triggers the angiogenic switch during carcinogenesis." Nature Cell Biology 2: 737 - 744.
Boissier, S., M. Ferreras, et al. (2000). "Bisphosphonates inhibit breast and prostate carcinoma cell invasion, an early event in the formation of bone metastases." Cancer Research 60: 2949-2954.
Boulay, A., R. Masson, et al. (2001). "High cancer cell death in syngeneic tumors developed in host mice deficient for the stromelysin-3 matrix metalloproteinase." Cancer Research 61: 2189 - 2193.
131.
Bouza, C., A. Muñoz, et al. (2005). "Efficacy of modern dressings in the treatment of leg ulcers: a systematic review." Wound Repair And Regeneration 13(3): 218-229.
Buchmeiser, M. R. (2007). "Polymeric monolithic materials: Syntheses, properties, functionalization and applications." Polymer 48(8): 2187-2198.
Caliceti, P., S. Salmaso, et al. (2001). "Controlled release of biomolecules from temperature-sensitive hydrogels prepared by radiation polymerization." Journal of Controlled Release 75(1-2): 173-181.
Caliceti, P., F. M. Veronese, et al. (1992). "Controlled release of proteins and peptides from hydrogels synthesized by gamma-ray induced polymerization." Il Farmaco 47(3): 275-286.
Carenza, M., S. Lora, et al. (1993). "Hydrogels obtained by radiation-induced polymerization as delivery systems for peptide and protein drugs." Radiation Physics and Chemistry 42(4-6): 897-901.
Carenza, M. and F. M. Veronese (1994). "Entrapment of biomolecules into hydrogels obtained by radiation-induced polymerization." Journal of Controlled Release29: 187-193.
Cataldo, D., K. Tournoy, et al. (2002). "Matrix metalloproteinase-9 deficiency impairs cellular infiltration and bronchial hyperresponsiveness during allergen-induced airway inflammation." American Journal of Pathology 161: 491 - 498.
Chakrabarty, K. H., R. A. Dawson, et al. (1999). "Development of autologous human dermal-epidermal composites based on sterilized human allodermis for clinical use." British Journal of Dermatology 141(5): 811-823.
Chen, C., G. S. Schultz, et al. (1999). "Molecular and mechanistic validation of delayed healing rat wounds as a model for human chronic wounds." Wound Repair and Regeneration 7(6): 486-494.
Chen, W. Y. J. and A. A. Rogers (2007). "Recent insights into the causes of chronic leg ulceration in venous diseases and implications on other types of chronic wounds." Wound Repair and Regeneration 15(4): 434-449.
Chirila, T. V., Zainuddin, et al. (2007). "Effect of phosphate functional groups on the calcification capacity of acrylic hydrogels." Acta Biomaterialia 3(1): 95-102.
Clinicaltrials.gov. (2002). "Pathophysiology of chronic wounds." Retrieved 15 April 2004, 2004, from http://www.clinicaltrials.gov/ct/show/NCT00006437?order=2.
Cook, H., P. Stephens, et al. (2000). "Defective extracellular matrix reorganization mby chronic wound fibroblasts is associated with alterations in TIMP-1, TIMP-2 and MMP-2 activity." Journal of Investigative Dermatology 115(2): 225-233.
132.
Cook, H., P. Stephens, et al. (2000). "Defective extracellular matrix reorganization by chronic wound fibroblasts is associated with alterations in TIMP-1, TIMP-2, and MMP-2 activity." Journal of Investigative Dermatology 115(2): 225-233.
Courtois, J., E. Bystrom, et al. (2006). "Novel monolithic materials using poly(ethylene glycol) as porogen for protein separation." Polymer 47(8): 2603-2611.
Cullen, B., R. Smith, et al. (2002). "Mechanism of action of PROMOGRAN, a protease modulating matrix, for the treatment of diabetic foot ulcers." Wound Repair and Regeneration 10(1): 16-25.
Cullen, B., P. W. Watt, et al. (2002). "The role of oxidised regenerated cellulose/collagen in chronic wound repair and its potential mechanism of action." The International Journal of Biochemistry & Cell Biology 34(12): 1544-1556.
Cutting, K. F. (2003). "Wound exudate: composition and function." British Journal of Dermatology 8(9): 4-9.
Cuzzell, J. (1997). "Choosing a wound dressing." Geriatric Nursing 18(6): 260-265.
Denizot, F. and R. Lang (1986). "Rapid colorimetric assay for cell growth and survival: modifications to the tetrazolium dye procedure giving improved sensitivity and reliability." Journal of Immunological Methods 89: 271-277.
Denoyelle, C., L. Hong, et al. (2003). "New insights into the actions of bisphosphonate zoledronic acid in breast cancer cells by dual RhoA-dependent and -independent effects." British Journal of Cancer 88(10): 1631.
Diegelmann, R. F. (2003). "Excessive neutrophils characterize chronic pressure ulcers." Wound Repair and Regeneration 11(6): 490-495.
Drury, J. L. and D. J. Mooney (2003). "Hydrogels for tissue engineering: scaffold design variables and applications." Biomaterials 24(24): 4337-4351.
Edwards, J. and P. Howley (2007). "Human neutrophil elastase and collagenase sequestration with phosphorylated cotton wound dressings." Journal of Biomedical Materials Research Part A 83A(2): 446-454.
Eldor, R., I. Raz, et al. (2004). "New and experimental approaches to treatment of diabetic foot ulcers: a comprehensive review of emerging treatment strategies." Diabetic Medicine 21(11): 1161-1173.
Falanga, V. (1993). "Chronic wounds: pathophysiologic and experimental considerations." Journal of Investigative Dermatology 100(5): 721-725.
Fick, A. (1855). "On liquid diffusion." Annals of Physics 170: 59.
First-Water (2004). "News and views." International Wound Journal 1(3): 159-161.
133.
Fleisch, H. (1997). Bisphosphonates in bone disease: from the laboratory to the patient. New York, Parthenon Publishing Group.
Fray, J. M., R. P. Dickinson, et al. (2003). "A potent, selective inhibitor of matrix metalloproteinase-3 for the topical treatment of chronic dermal ulcers." Journal of Medicinal Chemistry 46: 3514-3525.
Fujiwara, K., A. Matsukawa, et al. (2002). "Functional distinction between CXC chemokines, interleukin-8 (IL-8), and growth related oncogene (GRO)alpha in neutrophil infiltration." Laboratory Investigation 82: 15 - 23.
Gearing, A., S. Thorpe, et al. (2002). "Selective cleavage of human IgG by the matrix metalloproteinases, matrilysin and stromelysin." Immunology Letters 81: 41 - 48.
George, K. A., E. Wentrup-Byrne, et al. (2004). "Investigation into the diffusion of water into HEMA-co-MOEP hydrogels." Biomacromolecules 5: 1194-1199.
Gibbs, S., A. N. S. Pinto, et al. (2000). "Epidermal growth factor and keratinocyte growth factor differentially regulate epidermal migration, growth, and differentiation." Wound Repair and Regeneration 8(3): 192-203.
Gogly, B., N. Groult, et al. (1998). "Collagen zymography as a sensitive and specific technique for the determination of subpicogram levels of interstitial collagenase." Analytical Biochemistry 255(2): 211-216.
Grassi, G., R. Farra, et al. (2005). "Temperature-sensitive hydrogels: potential therapeutic applications." American Journal of Drug Delivery 3(4): 239-251.
Grinnell, F., C.-H. Ho, et al. (1992). "Degradation of fibronectin and vitronectin in chronic wound fluid: analysis by cell blotting, immunoblotting, and cell adhesion assays." Journal of Investigative Dermatology 98(4): 410-416.
Gruen, R. L., S. Chang, et al. (1996). "Optimizing the hospital management of leg ulcers." Australia and New Zealand Journal of Surgery 66: 171-174.
Hart, J., D. Silcock, et al. (2002). "The role of oxidised regenerated cellulose/collagen in wound repair: effects in vitro on fibroblast biology and in vivo in a model of compromised healing." The International Journal of Biochemistry & Cell Biology 34: 1557-1570.
Hassan, C. M. and N. A. Peppas (2000). Structure and applications of poly(vinyl alcohol) hydrogels produced by conventional crosslinking or by freezing/thawing methods. Advances in Polymer Science: Biopolymers, PVA Hydrogels, Anionic Polymerisation Nanocomposites. Berlin, SpringerLink. 153: 37-65.
Hautamaki, R., D. Kobayashi, et al. (1997). "Requirement for macrophage elastase for cigarette smoke-induced emphysema in mice." Science 277: 2002 - 2004.
134.
He, H., X. Cao, et al. (2004). "Design of a novel hydrogel-based intelligent system for controlled drug release." Journal of Controlled Release 95(3): 391-402.
Heikkila, P., O. Teronen, et al. (2002). "Bisphosphonates inhibit stromelysin-1 (MMP-3), matrix metalloelastase (MMP-12), collagenase-3 (MMP-13) and enamelysin (MMP-20), but not urokinase-type plasminogen activator, and diminish invasion and migration of human malignant and endothelial cell lines." Anti-Cancer Drugs 13: 245-254.
Herouy, Y., D. Trefzer, et al. (2000). "Matrix metalloproteinases and venous leg ulceration." European Journal of Dermatology 9: 173-180.
Heymann, D., B. Ory, et al. (2004). "Bisphosphonates: new therapeutic agents for the treatment of bone tumors." Trends in Molecular Medicine 10(7): 337-343.
Hoffman, A. S. (1998). A Commentary on the Advantages and Limitations of Synthetic Polymer-Biomolecule Conjugates. Biorelated Polymers and Gels. T. Okano. Boston, Academic Press: 231-248.
Hoffman, A. S. (2002). "Hydrogels for biomedical applications." Advanced Drug Delivery Reviews 54(1): 3-12.
Holland, T. A., Y. Tabata, et al. (2003). "In vitro release of transforming growth factor-[beta]1 from gelatin microparticles encapsulated in biodegradable, injectable oligo(poly(ethylene glycol) fumarate) hydrogels." Journal of Controlled Release 91(3): 299-313.
Holmbeck, K., P. Bianco, et al. (1999). "MT1-MMP-deficient mice develop dwarfism, osteopenia, arthritis, and connective tissue disease due to inadequate collagen turnover." Cell 99: 81 - 92.
Hubbell, J. A. (1996). "Hydrogel systems for barriers and local drug delivery in the control of wound healing." Journal of Controlled Release 39(2-3): 305-313.
Iida, J., K. L. Wilhelmson, et al. (2004). "Membrane Type-1 Matrix Metalloproteinase Promotes Human Melanoma Invasion and Growth." Journal of Investigative Dermatology 122(1): 167-176.
Imai, K., A. Hiramatsu, et al. (1997). "Degradation of decorin by matrix metalloproteinases: identification of the cleavage sites, kinetic analyses and transforming growth factor-beta1 release." Biochemical Journal 322: 809 - 814.
Jawien, A., M. T. Szewczyk, et al. (2006). "Functional and biopsychosocial restrictions among patients with a venous ulcer." Archives of Medical Science2(1): 36-41.
Jeong, B., Y. K. Choi, et al. (1999). "New biodegradable polymers for injectable drug delivery systems." Journal of Controlled Release 62(1-2): 109-114.
135.
Kantor, J. and D. J. Margolis (2001). "Treatment Options for Diabetic Neuropathic Foot Ulcers: A Cost-Effectiveness Analysis." Dermatologic Surgery 27(4): 347-351.
Kantor, J. and D. J. Margolis (2003). "Management of leg ulcers." Seminars in Cutaneous Medicine and Surgery 22(3): 212-221.
Kästenbauer, T., B. Härnlein, et al. (2003). "Evaluation of granulocyte-colony stimulating factor (Filgrastim) in infected diabetic foot ulcers." Diabetologia46(1): 27-30.
Kemp, D. and F. Vellaccis (1980). Acyl Derivatives. Organic Chemistry. New York, Worth Publishers Inc.: 368-369.
Kido, A., G. Pap, et al. (2007). "Disease-specific expression patterns of proteases in synovial tissues." Pathology - Research and Practice 203(6): 451-456.
Klein, S. A., G. L. Anderson, et al. (2002). "The Effects of a Broad-Spectrum Matrix Metalloproteinase Inhibitor on Characteristics of Wound Healing." Journal of Investigative Surgery 15(4): 199-207.
Ladwig, G. P., M. C. Robson, et al. (2002). "Ratios of activated matrix metalloproteinase-9 to tissue inhibitor of matrix metalloproteinase-1 in wound fluids are inversely correlated with healing of pressure ulcers." Wound Repair and Regeneration 10(1): 26-37.
Laemmli, U. K. (1970). "Cleavage of structural proteins during the assembly of the head of bacteriophage T4." Nature 227: 680-685.
Landers, R., U. Hubner, et al. (2002). "Rapid prototyping of scaffolds derived from thermoreversible hydrogels and tailored for applications in tissue engineering." Biomaterials 23(23): 4437-4447.
Lawrence, J. N. (1997). "Application of in vitro human skin models to dermal irritancy: a brief overview and future prospects." Toxicology in Vitro 11(3): 305-312.
Leach, J. B., K. A. Bivens, et al. (2003). "Photocrosslinked Hyaluronic Acid Hydrogels: Natural, Biodegradable Tissue Engineering Scaffolds." Biotechnology and Bioengineering 82: 578-589.
Lee, K. Y. and D. J. Mooney (2001). "Hydrogels for Tissue Engineering." Chemical Reviews 101(7): 1869-1879.
LeGrand, E. K. (1998). "Preclinical promise of becaplermin (rhPDGF-BB) in wound healing." The American Journal of Surgery 176(2, Supplement 1): 48S-54S.
Leguéné, C., P. Clavère, et al. (2001). "Oxydation radiolytique du tamoxifène par les radicaux libres ·OH et (ou) HO·2." Canadian Journal of Physiology & Pharmacology 79(2): 184-188.
136.
Lerman, O. Z., R. D. Galiano, et al. (2003). "Cellular dysfunction in the diabetic fibroblast: impairment in migration, vascular endothelial growth factor production, and response to hypoxia." American Journal of Pathology 162(1): 303-312.
Levi, E., R. Fridman, et al. (1996). "Matrix metalloproteinase 2 releases active soluble ectodomain of fibroblast growth factor receptor 1." Proceedings of the National Academy of Sciences of the United States of America 93: 7069 - 7074.
Levine, B. and R. Williams (1982). The Role of Calcium in Biological Systems. Boca Raton, CRC Press.
Levy, D. E., F. Lapierre, et al. (1998). "Matrix metalloproteinase inhibitors: a structure-activity study." Journal of Medicinal Chemistry 41(2): 199-223.
Li, W. W. and V. W. Li (2003). "Therapeutic angiogenesis for wound healing." Wounds - A Compendium of Clinical Research and Practice 15(9): 3S-12S.
Lund, L. R., J. Romer, et al. (1999). "Functional overlap between two classes of matrix-degrading proteases in wound healing." European Molecular Biology Organization Journal 18(17): 4645-4656.
MacLellan, D. G. (2000). "Chronic wound management." Australian Prescriber 23: 6-9.
Martins, M. C. L., D. Wang, et al. (2003). "Albumin and fibrinogen adsorption on PU-PHEMA surfaces." Biomaterials 24(12): 2067-2076.
Massova, I., L. P. Kotra, et al. (1998). "Matrix metalloproteinases: structures, evolution, and diversification." The Federation of American Societies for Experimental Biology Journal 12(12): 1075-1095.
McMullin, G., C. Hyde, et al. (1998). "Cost of leg ulcer management in the central Sydney health area." Australia and New Zealand Journal of Surgery 68: A68-69.
McQuibban, G., G. Butler, et al. (2001). "Matrix metalloproteinase activity inactivates the CXC chemokine stromal cell-derived factor-1." Journal of Biological Chemistry 276: 43503 - 43508.
Melo, R. C. N., P. G. Rosa, et al. (2007). "Histological approaches for high-quality imaging of zooplanktonic organisms." Micron 38(7): 714-721.
Menke, N. B., K. R. Ward, et al. (2007). "Impaired wound healing." Clinics in Dermatology 25(1): 19-25.
Montague, R., C. A. Hart, et al. (2004). "Differential Inhibition of Invasion and Proliferation by Bisphosphonates: Anti-Metastatic Potential of Zoledronic Acid in Prostate Cancer." European Urology 46(3): 389-402.
137.
Montheard, J. P., M. Chatzopoulos, et al. (1992). "2-Hydroxyethyl Methacrylate (HEMA) - Chemical Properties and Applications in Biomedical Fields." Journal of Macromolecular Science - Reviews in Macromolecular Chemistry and Physics C32(1): 1-34.
Moseley, R., J. R. Hilton, et al. (2004). "Comparison of oxidative stress biomarker profiles between acute and chronic wound environments." Wound Repair Regen 12(4): 419-429.
Moses, M., M. Marikovsky, et al. (1996). "Temporal study of the activity of matrix metalloproteinases and their endogenous inhibitors during wound healing." Journal of Cellular Biochemistry 60(3): 379-386.
Munro, H. S. and M. Yasin (2002). Bioadhesive compositions and wound dressings containing them. United Kingdom, First Water, Ltd.
Namkyu, S., L. Seunghee, et al. (2004). "Characterization of a carotenoid-hyperproducing yeast mutant isolated by low-dose gamma irradiation." International Journal of Food Microbiology 94(3): 263-267.
Neville-Webbe, H. L., I. Holen, et al. (2002). "The anti-tumour activity of bisphosphonates." Cancer Treatment Reviews 28(6): 305-319.
Ohsawa, T., I. Maruyama, et al. (1999). "Collateral occurrence of deimination of keratins with differentiation of an immortalized newborn rat keratinocyte cell line cultured at air-liquid interface." Journal of Dermatological Science 19(1): 68-73.
Okada, Y., Y. Gonoji, et al. (1992). "Matrix metalloproteinase-9 (92-kDa gelatinase type-IV collagenase) from Ht-1080 human fibrosarcoma cells - purification and activation of the precursor and enzymatic properties." Journal of Biological Chemistry 267(30): 21712-21719.
Otera, J. (1993). "Transesterification." Chemical Reviews 93(4): 1449-1470.
Page-McCaw, A., A. J. Ewald, et al. (2007). "Matrix metalloproteinases and the regulation of tissue remodelling." Nature Reviews: Molecular Cell Biology8(3): 221-233.
Palohati, M., L. Lauharanta, et al. (1993). "Proteolytic activity in leg ulcer exudate." Experimental Dermatology 2: 29-37.
Papanas, N. and E. Maltezos (2007). "Growth Factors in the Treatment of Diabetic Foot Ulcers: New Technologies, Any Promises?" International Journal of Lower Extremity Wounds 6(1): 37-53.
Park, S.-E., Y.-C. Nho, et al. (2004). "Preparation of poly(polyethylene glycol methacrylate-co-acrylic acid) hydrogels by radiation and their physical properties." Radiation Physics and Chemistry 69(3): 221-227.
138.
Parks, W. C. (1999). "Matrix metalloproteinases in repair." Wound Repair and Regeneration 7(6): 423-432.
Pastina, B., J. Isabey, et al. (1999). "The influence of water chemistry on the radiolysis of the primary coolant water in pressurized water reactors." Journal of Nuclear Materials 264(3): 309-318.
Pastina, B., J. Isabey, et al. (1999). "The influence of water chemistry on the radiolysis of the primary coolant water in pressurized water reactors." Journal of Nuclear Materials 264(3): 309-318.
Prashantha, K., B. J. Rashmi, et al. (2006). "Spectral characterization of apatite formation on poly(2-hydroxyethylmethacrylate)-TiO2 nanocomposite film prepared by sol-gel process." Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 65(2): 340-344.
Queen, D., H. Orsted, et al. (2004). "A dressing history." International Wound Journal1(1): 59-77.
Ramakrishna, S., J. Mayer, et al. (2001). "Biomedical applications of polymer-composite materials: a review." Composites Science and Technology 61(9): 1189-1224.
Ramamurthy, N. S., A. J. Kucine, et al. (1998). "Topically applied CMT-2 enhances wound healing in streptozotocin diabetic rat skin." Advanced Dental Research12: 144-148.
Ramamurthy, N. S., S. A. McClain, et al. (1999). "Wound healing in aged normal and ovariectomized rats: effects of chemically modified doxycycline (CMT-8) on MMP expression and collagen synthesis." Annals New York Academy of Sciences 878: 720-723.
Ratliff, C. R. and G. T. Rodeheaver (2005). "Use of the PUSH tool to measure venous ulcer healing." Ostomy Wound Management 51(5): 58-63.
Ries, C. and P. E. Petrides (1995). "Cytokine regulation of matrix metalloproteinase activity and its regulatory dysfunction in disease." Journal of Biological Chemistry 376: 345-355.
Rouif, S. (2005). "Radiation cross-linked polymers: Recent developments and new applications." Nuclear Instruments & Methods in Physics Research Section B-Beam Interactions with Materials and Atoms 236: 68-72.
Rubin, C., G. Gur, et al. (2005). "Negative regulation of receptor tyrosine kinases: unexpected links to c-Cbl and receptor ubiquitylation." Cell Research 15: 66-71.
Rush III, T. and R. Powers (2004). "The application of X-ray, NMR, and molecular modelling in the design of MMP inhibitors." Current Topics in Medicinal Chemistry 4: 1311-1327.
139.
Russell, R. G. G. and M. J. Rogers (1999). "Bisphosphonates: from the laboratory to the clinic and back again." Bone 25(1): 97-106.
Samuelsen, J. T., J. E. Dahl, et al. (2007). "Apoptosis induced by the monomers HEMA and TEGDMA involves formation of ROS and differential activation of the MAP-kinases p38, JNK and ERK." Dental Materials 23(1): 34-39.
Schacke, H., W.-D. Docke, et al. (2002). "Mechanisms involved in the side effects of glucocorticoids." Pharmacology & Therapeutics 96(1): 23-43.
Schmedlen, R. H., K. S. Masters, et al. (2002). "Photocrosslinkable polyvinyl alcohol hydrogels that can be modified with cell adhesion peptides for use in tissue engineering." Biomaterials 23(22): 4325-4332.
Schuster, M., J. Ankersmit, et al. (2001). "Induction of T-cell apoptosis by polyurethane biomaterials used in left ventricular assist devices is dependent on calcineurin activation." Transplantation Proceedings 33(1-2): 1958-1959.
Shai, A. and S. Halevy (2005). "Direct triggers for ulceration in pateints with venous insufficiency." International Journal of Dermatology 44: 1006-1009.
Shao, C., H.-Y. Kim, et al. (2003). "Fiber mats of poly(vinyl alcohol)/silica composite via electrospinning." Materials Letters 57(9-10): 1579-1584.
Sharpey-Schafer, E. (1954). Schafer's essentials of histology: description and practical. London, Longmans.
Singh, B., N. Sharma, et al. (2007). "Synthesis, characterization and swelling studies of pH responsive psyllium and methacrylamide based hydrogels for the use in colon specific drug delivery." Carbohydrate Polymers 69(4): 631-643.
Smeds, K. A. and M. W. Grinstaff (2001). "Photocrosslinkable polysaccharides for in situ hydrogel formation." Journal of Biomedical Materials Research 54: 115-121.
Smith and Nephew. (2000). "Results Summary." Retrieved 21 March, 2007, from http://www.smith-nephew.com/investors/interim00/aug00/00interims.ppt.
Soler, P. M., T. E. Wright, et al. (1999). "In vivo characterization of keratinocyte growth factor-2 as a potential wound healing agent." Wound Repair and Regeneration 7(3): 172-178.
Somerville, R., S. Oblander, et al. (2003). "Matrix metalloproteinases: old dogs with new tricks." Genome Biology 4(6): 216.
Sorensen, O. E., J. B. Cowland, et al. (2003). "Wound healing and expression of antimicrobial peptides/polypeptides in human keratinocytes, a consequence of common growth factors." Journal of Immunology 170(11): 5583-5589.
Sparrow, N. (2005). "Wound-care technology." Retrieved 23 July, 2007, from http://www.devicelink.com/emdm/archive/05/10/marketwatch.html.
140.
Stadelmann, W. K., A. G. Digenis, et al. (1998). "Physiology and healing dynamics of chronic cutaneous wounds." The American Journal of Surgery 126(Suppl 2A): 26S-38S.
Stashak, T. S., E. Farstvedt, et al. (2004). "Update on wound dressings: Indications and best use." Clinical Techniques in Equine Practice 3(2): 148-163.
Stearns , M. E. (1998). "Alendronate blocks TGF-β1 stimulated collagen 1 degradation by human prostate PC-3 ML cells." Clinical & Experimental Metastasis 16(4): 332-339.
Stewart, J. (2002). "Next generation products for wound management." World Wide Wounds.
Stuart, E. L., G. S. Evans, et al. (2007). "The proteolytic profile of prelabour ruptured amnion at term: A case-control study." European Journal of Obstetrics & Gynecology and Reproductive Biology 131(2): 158-162.
Suzuki, M., G. Raab, et al. (1997). "Matrix metalloproteinase-3 releases active heparin-binding EGF-like growth factor by cleavage at a specific juxtamembrane site." Journal of Biological Chemistry 272: 31730 - 31737.
Suzuki, S., M. Whittaker, et al. (2007). In vitro mineralization of soluble and cross linked phosphate polymers. International Symposium of Polymeric Materials for Regenerative Medecine, Montreal, Canada.
Swanson, K. M. and R. J. Hohl (2006). "Anti-Cancer Therapy: Targeting the Mevalonate Pathway." Current Cancer Drug Targets 6(1): 15-37.
Tarlton, J. F., A. J. Bailey, et al. (1999). "Prognostic value of markers of collagen remodeling in venous ulcers." Wound Repair and Regeneration 7(5): 347-355.
Tarnuzzer, R. W. and G. S. Schultz (1996). "Biochemical analysis of acute and chronic wound environments." Wound Repair and Regeneration 4(3): 321-325.
Teronen, O., P. Heikkila, et al. (1999). "MMP inhibition and downregulation by bisphosphonates." Annals New York Academy of Sciences 878: 453-465.
Tian, Y. J., L. Y. Zhang, et al. (2007). "Raman sensitivity enhancement for aqueous absorbing sample using Teflon-AF 2400 liquid core optical fibre cell." Analytica Chimica Acta 581(1): 154-158.
Tomasek, J. J., N. L. Halliday, et al. (1997). "Gelatinase A activation is regulated by the organization of the polymerized actin cytoskeleton." Journal of Biological Chemistry 272: 7482-7487.
Tomic, S. L., M. M. Micic, et al. (2007). "Swelling and drug release behavior of poly(2-hydroxyethyl methacrylate/itaconic acid) copolymeric hydrogels
141.
obtained by gamma irradiation." Radiation Physics and Chemistry 76(5): 801-810.
Tomizawa, Y. (2005). "Clinical benefits and risk analysis of topical hemostats: a review." Journal of Artificial Organs 8(3): 137-142.
Topping, G., J. Malda, et al. (2006). "Development and characterisation of human skin equivalents and their potential application as a burn wound model." Primary Intention 14(1): 14-21.
Tornier, C., M. Rosdy, et al. (2006). "In vitro skin irritation testing on reconstituted human epidermis: Reproducibility for 50 chemicals tested with two protocols." Toxicology in Vitro 20(4): 401-416.
Trengove, N. J., S. R. Langton, et al. (1996). "Biochemical analysis of wound fluid from nonhealing and healing chronic leg ulcers." Wound Repair and Regeneration 4(2): 234-239.
Trengove, N. J., M. C. Stacey, et al. (1999). "Analysis of the acute and chronic wound environments: the role of proteases and their inhibitors." Wound Repair and Regeneration 7: 442-452.
Ueda, H. and Y. Tabata (2003). "Polyhydroxyalkanonate derivatives in current clinical applications and trials." Advanced Drug Delivery Reviews 55(4): 501-518.
Ugarte, D. A. D., R. L. Roberts, et al. (2002). "Treatment of chronic wounds by local delivery of granulocyte-macrophage colony-stimulating factor in patients with neutrophil dysfunction." Pediatric Surgery International 18(5): 517-520.
Uludag, H. and J. Yang (2002). "Targeting Systemically Administered Proteins to Bone by Bisphosphonate Conjugation." Biotechnology Progress 18(3): 604-611.
Valleala, H., R. Hanemaaijer, et al. (2003). "Regulation of MMP-9 (gelatinase B) in activated human monocyte/macrophages by two different types of bisphosphonates." Life Sciences 73(19): 2413-2420.
Vasikaran, S. D. (2001). "Bisphosphonates: an overview with special reference to alendronate." Annals of Clinical Biochemistry 38: 608-623.
Wagner, S., S. Coerper, et al. (2003). "Comparison of inflammatory and systemic sources of growth factors in acute and chronic human wounds." Wound Repair and Regeneration 11(4): 253-260.
Wang, D., S. Miller, et al. (2003). "Synthesis and Evaluation of Water-Soluble Polymeric Bone-Targeted Drug Delivery Systems." Bioconjugate Chemistry14(5): 853-859.
142.
Wasselin-Trupin, V., G. Baldacchino, et al. (2002). "Hydrogen peroxide yields in water radiolysis by high-energy ion beams at constant LET." Radiation Physics and Chemistry 65(1): 53-61.
Weckroth, M., A. Vaheri, et al. (1996). "Matrix metalloproteinases, gelatinase and collagenase, in chronic leg ulcers." Journal of Investigative Dermatology106(5): 1119-1124.
West, J. L. and J. A. Hubbell (1995). "Photopolymerized hydrogel materials for drug delivery applications." Reactive Polymers 25(2-3): 139-147.
Wielockx, B., C. Libert, et al. (2004). "Matrilysin (matrix metalloproteinase-7): a new promising drug target in cancer and inflammation?" Cytokine & Growth Factor Reviews 15(2-3): 111-115.
Wilson, C., A. Ouellette, et al. (1999). "Regulation of intestinal alpha-defensin activation by the metalloproteinase matrilysin in innate host defense." Science286: 113 - 117.
Winter, G. D. (1962). "Formation of the Scab and the Rate of Epithelization of Superficial Wounds in the Skin of the Young Domestic Pig." Nature193(4812): 293-294.
Witte, M. B., F. J. Thornton, et al. (1998). "Metalloproteinase inhibitors and wound healing: A novel enhancer of wound strength." Surgery 124(2): 464-470.
Wysocki, A. B., A. O. Kusakabe, et al. (1999). "Temporal expression of urokinase plasminogen activator, plasminogen activator inhibitor and gelatinase-B in chronic wound fluid switches from a chronic to acute wound profile with progression to healing." Wound Repair and Regeneration 7(3): 154-165.
Wysocki, A. B., L. Staianocoico, et al. (1993). "Wound fluid from chronic leg ulcers contains elevated levels of metalloproteinases MMP-2 and MMP-9." Journal of Investigative Dermatology 101(1): 64-68.
Xi, F., J. Wu, et al. (2006). "Novel nylon-supported organic-inorganic hybrid membrane with hierarchical pores as a potential immobilized metal affinity adsorbent." Journal of Chromatography A 1125(1): 38-51.
Xue, M., D. Campbell, et al. (2007). "Protein C is an autocrine growth factor for human skin keratinocytes." Journal of Biological Chemistry 282(18): 13610-13616.
Yager, D. R. and B. C. Nwomeh (1999). "The proteolytic environment of chronic wounds." Wound Repair and Regeneration 7(6): 433-441.
Yager, D. R., L. Y. Zhang, et al. (1996). "Wound fluids from human pressure ulcers contain elevated matrix metalloproteinase levels and activity compared to surgical wound fluids." Journal of Investigative Dermatology 107: 743-748.
143.
Zahedi, P. and P. I. Lee (2007). "Solid molecular dispersions of poorly water-soluble drugs in poly(2-hydroxyethyl methacrylate) hydrogels." European Journal of Pharmaceutics and Biopharmaceutics 65(3): 320-328.
Zainuddin, D. Hill, et al. (2006). "In-vitro study of the spontaneous calcification of PHEMA-based hydrogels in simulated body fluid." Journal of Materials Science: Materials in Medicine 17(12): 1245-1254.