Applying side effects of anti fibrotic compounds to promote neovascularization for tissue engineering

Another group used embryonic fibroblasts to maintain VEGF levels in a tri-culture of endothelial cells and myoblasts to engineer vascularized skeletal muscle tissue12 in vitro.. Others h

1 LITERATURE REVIEW

1.1 Neovascularization in Tissue Engineering

Regenerative medicine encompasses a variety of research areas including cell therapy, biomaterials engineering, growth factors and transplantation science Most of these efforts converge into tissue engineering, an exciting field which offers a means of developing biological substitutes for maintaining, restoring and improving tissue function1 One of the major challenges in tissue engineering today is the generation of large vascularized 3D-structures2,3 Cells require a constant supply of essential nutrients and oxygen as well as a constant removal of metabolic waste products Vascularization is

thus essential for maintaining the survival and function of tissue engineered constructs in vitro and vivo, as cellular viability decreases steeply beyond a few hundred μm from a

blood supply4

One of the major challenges in tissue engineering is how accelerate neovascularization of

an implanted tissue construct Current technology limits the survival of implanted tissues

as they depend initially on diffusion and later on neovascularization Diffusion is insufficient for survival of the tissue and limits the thickness of the implants Dependence

on neovascularization can cause fibrovascular ingrowth and hence scarring Presently, there are no established methods to prefabricate a capillary network in tissue constructs that could connect with the host vasculature after implantation5 Currently, strategies to improve vascularization prior to or after implantation involve the use of growth factors, endothelial cells and endothelial progenitor cells They can be summarized into the following four approaches:

i) Pre-vascularization of polymeric scaffolds by host

Prevascularization of polymeric scaffolds can be achieved by implanting the polymeric scaffold without the tissue construct at selected body sites first so that vascularization may occur before subsequent implantation of the tissue construct or cells into the vascularized matrix6 This may be feasible in an animal model, but would require multiple surgeries for a human patient Pre-vascularization using an animal host presents potential host immune reactions and risk of animal disease transmission

ii) Use of angiogenic growth factors

Vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF) are potent inducers of angiogenesis However, such angiogenic growth factors are inherently

unstable in vivo, therefore bolus injections of the recombinant protein has limited use in

neovascularization of implants Sustained release of the exogenous growth factors may be achieved by integration or encapsulation within biodegradable polymers as microspheres7,8 or as part of a scaffold9 Another method of achieving sustained release

of VEGF was developed using microsphere encapsulated cells transfected with VEGF cDNA, but this is dependent on cellular uptake and translation of the cDNA10 There are also issues with the safety of using transfected cells for gene therapy due to the transfection vector Unfortunately, overproduction single angiogenic factors often results

in deformed or non- functional blood vessels11

iii) Seeding endothelial cells on a biodegradable matrix together with VEGF transfected cells

This method ensures sustained release of VEGF to maintain the viability of endothelial cells after implantation2 Another group used embryonic fibroblasts to maintain VEGF levels in a tri-culture of endothelial cells and myoblasts to engineer vascularized skeletal muscle tissue12 in vitro Perfusion, vascularization and survival of the muscle tissue

constructs after transplantation was improved

iv) Co-seeding scaffolds with progenitor cells

The use of stem cells is another avenue that shows promise In the presence of angiogenic factors such as VEGF, they may be differentiated into endothelial progenitor cells13 Co-implantation of vascular endothelial cells with mesenchymal precursor cells

on a three dimensional collagen gel resulted in differentiation of the precursor cells into mural cells Implantation of this construct into mice produced capillary like tubes, some

of which eventually became perfused and joined with the host vasculature, a process called inosculation14 Others have used bone marrow derived endothelial progenitor cells seeded together with other cell types, such as hematopoietic progenitor cells to produce a vascularized matrix15,16,17

1.2 HIF-1α: An Alternative Approach to Neovascularization

Prolyl 4-hydroxylase inhibitors may be an alternative method of neovascularization by chemically stimulating angiogenesis via stabilization of HIF-1 and upregulation of HIF-1 target genes

As angiogenesis is a complex process involving many different factors, the use of exogenous angiogenic factors alone may not produce functional vascular networks Instead of trying to optimize a recipe of cells and growth factors for vascularization to occur, this approach may achieve a more physiological response, similar to hypoxic stress,

by using HIF-1 as a master switch

Targeting HIF-1 has the advantage of bypassing the downstream complexity of multiple cell signaling molecules and angiogenic factors involved in the process of angiogenesis The cells are directed by HIF-1 to produce a number of resulting gene products endogenously, rather than attempting to recreate the series of events leading to angiogenesis by the use of exogenous growth factors and progenitor cell types The amount and timing of these growth factors and signaling molecules are determined and controlled by the cells, therefore more closely resembling physiological stimulation of angiogenesis This method would also remove the need of genetically manipulated cells Alternatively, it may augment the rate of vascularization if used together with engineered tissues and scaffolds containing endothelial cells or endothelial progenitor cells This proposed approach is novel for improving neovascularization in tissue engineering

1.3 Regulation of HIF-1

Hypoxia inducible factor 1 (HIF-1) is a key transcription factor that is upregulated in response to hypoxia Hypoxia is a common physiological stress imposed by low oxygen conditions to cells or tissues HIF-1 is a heterodimer consisting of two subunits, a

hypoxia regulated HIF-1α and an oxygen insensitive subunit HIF-1β18 , 19 HIF-1α consists mainly of a basic-helix-loop-helix (bHLH) region, a Per/Arnt/Sim (PAS) region,

a N-terminal transactivation domain (N-TAD) and a C-terminal transactivation domain (C-TAD) (Figure 1 extracted from Kaelin 200519)

Figure 1 Schematic representation of HIF-1α protein domain structures (Kaelin 2005 19 )

Under normoxia, posttranslational hydroxylation of HIF-1α occurs at conserved proline residues (Pro 402 and Pro 564) within a polypeptide segment known as the oxygen dependent domain (ODD) found at the N-terminal transactivating domain (N-TAD) region19,20,21 These hydroxylated prolyl sites interact with the von-Hippel Lindau tumor suppressor protein (pVHL), which is the recognition protein for the E3 Ubiquitin Ligase complex19,20,21, 22 , 23 The ubiquitin ligase complex subsequently packages HIF-1 for proteosomal degradation via polyubiquitylation, therefore HIF-1 is highly unstable under normoxia19,20,22 In addition, an asparaginyl hydroxylase, Factor Inhibiting HIF-1 (FIH-1), acts on asparaginyl residue Asn 803, found on the C-terminal transactivating domain (C-TAD) region of HIF-1α, blocking the site and preventing its interaction with the transcriptional activator p300 20,24,25 (Figure 2)

Under hypoxic conditions, the oxygen dependent prolyl hydroxylase enzymes is inhibited,

thus HIF-1 is not hydroxylated and accumulates in the cell Hypoxia also prevents

hydroxylation of the C-TAD region, allowing HIF-1α to interact with p300 and promote

transcription of target genes

1.4 HIF-1 Target Genes

HIF-1 target genes (Figure 3) are expressed in hypoxic situations, with many of them

found to be involved in cell metabolism, vascular remodeling and angiogenesis18, 26

Specifically, vascular endothelial factor (VEGF) is well established as an angiogenesis

Figure 2 Regulation of the HIF-1 transcription factor (From Bruick & McKnight

2002 20 ). Under normoxic conditions, the ODD of HIF- 1α is modified by a HIF–

prolyl hydroxylase, triggering HIF-1α recognition by pVHL and subsequent degradation

by the proteasome Similarly,

an asparaginyl hydroxylase modifies the C-TAD of HIF- 1α, blocking its interaction with the transcriptional coactivator p300 Hypoxia blocks both prolyl hydroxylation and asparaginyl hydroxylation,

allowing HIF-1α to accumulate and bind to p300, thereby promoting the transcription of downstream HIF-1 target genes, thus enabling cells and the whole organism to adapt to hypoxia.

differentiation27,28,29 It is often used as a positive control for stimulating angiogenesis VEGF receptor-1 (VEGFR-1), also known as FLT-1, is a tyrosine kinase receptor involved in endothelial cell migration, autocrine signaling and induction of matrix metalloproteinase-9 (MMP-9)27,28,29 Others, such as erythropoietin (EPO), signal the requirement for production of red blood cells, which is important in hypoxia30, and ischemic diseases31

Figure 3 Genes that are directly regulated by HIF-1 18,32 Many of these gene products

of HIF-1 transcription are related to angiogenesis, such as erythropoietin for the production of red blood cells, insulin like growth factor and its binding proteins, VEGF and its receptor FLT-1

Gene product ABCG2

HIF-1α prolyl hydroxylase PHD3 (EGLN3)

HIF-1α prolyl hydroxylase PHD2 (EGLN1)

HGTD-P

ID2 Integrin β2Intestinal trefoil factor Lactate dehydrogenase A (LDHA) Lactase

Leptin Membrane type-1 matrix metalloproteinase Multi-drug resistance 1 (ABCB1) Myeloid cell factor 1 (MNL1) Nitric oxide synthase 2 NIP3

NUR77 p35srj (CITED2) Phosphoglycerate kinase 1 6-Phosphofructo-2-kinase/fructose-2,6- bisphosphatase-3 (PFKFB3)

bisphosphatase-4 (PFKFB4) Plasminogen activator inhibitor 1 Procollagen prolyl-4-hydroxylase α(I) RORα

6-Phosphofructo-2-kinase/fructose-2,6-Stromal-derived factor 1 (SDF-1) Telomerase (TERT)

Transferrin Transferrin receptor Transforming growth factor β3Vascular endothelial growth factor (VEGF)

VEGF receptor-1 (Flt-1)

1.5 HIF-1α Promotes Angiogenesis

There is a considerable amount of evidence published on the involvement of HIF-1α in promoting angiogenesis This has been reviewed extensively18,26,32, therefore only a brief summary will be given here

Ischemia results in cellular exposure to hypoxic microenvironments By stabilization of HIF-1, the cell automatically signals the requirement for growth of blood vessels to the ischemic site by upregulating the local production of angiogenic factors This is also often the case in cancers, where HIF is activated by physiological hypoxia within a rapidly growing tumour cell mass32,33 Overexpression of HIF-1α has been found with an associated increase in microvessel density and VEGF expression in several cancers, such

as bladder, breast, colon, ovarian and pancreatic cancers32

HIF-1 is also critical in angiogenesis and embryogenesis It was reported that HIF-1 deficient mouse embryonic cells produce major vascular defects in the yolk sac and developing embryo, associated with severe hypoxia due to lack of perfusion34 Injection

of these HIF-1 deficient mouse embryonic stem cells into immunocompromised mice resulted in smaller and less vascularized tumors compared to normal embryonic stem cells

Transgenic mice expressing constitutively active HIF-1α in the epidermis exhibited an increase capillary density in the dermis and marked increase in VEGF35 Despite

to transgenic mice overexpressing VEGF cDNA in the skin35,11 Additionally, administration of HIF-1α/VP16 recombinant plasmids constructed by fusion of HIF-1α to the transactivation domain of herpes simplex virus VP16 has improved recovery of local blood flow and vascularization in a rabbit hindlimb ischemic model36 Similar results were obtained using an acute myocardial infarction rat model37 These evidences demonstrate the involvement of HIF-1 in angiogenesis Consequently, several approaches have been undertaken to target HIF-1 for cancer therapy and treatment for ischemic diseases38,39

1.6 HIF Prolyl 4-Hydroxylases

There are three identified human cytoplasmic prolyl 4-hydroxylase isoenzymes that hydroxylate HIF-1α40 , 41 Like collagen prolyl 4-hydroxylases, they are dioxygenases dependent on Fe2+, oxygen, ascorbate and 2-oxoglytarate

1.7 Prolyl 4-Hydroxylase Inhibitors (PHi) Promote Angiogenesis By Stabilizing HIF-1α

Available literature on promotion of angiogenesis using prolyl 4-hydroxylase inhibitors (PHi) is scarce at the moment Gleadle et al (1995)42 was one of the first to look at the effect of prolyl hydroxylase inhibitors on HIF-1 Cobalt ions and iron chelators such as desferrioxamine (also known as deferioxamine, or DFO), upregulated mRNA expression

of some HIF-1 target genes, namely VEGF, platelet derived growth factor A and B (PDGF-A, PDGF-B), erythropoietin (EPO) and transforming growth factor β-1 (TGFβ-1)

The mRNA expression levels were compared to that induced by hypoxia in human tumour cell lines like Hep3B, HepG2 (hepatoma) and HT1080 (fibrosarcoma)

Stabilization of HIF-1α has subsequently been reported using cobalt chloride43, oxoglutarate analogues such as N-oxalylglycine and dimethyloxalyglycine (DMOG)22,43,44, nitric oxide (S-nitroglutathione)45, proprietary substances developed by Fibrogen, Inc46,47,48, Proteosome inhibitors such as MG-132 may also induce nuclear accumulation of HIF-1α44,43 by preventing its degradation In addition, aliphatic and aromatic compounds such as pyridine-2,4-dicarboxylate and 3,4-dihydroxybenzoic acid were found to be competitive inhibitors of HIF-prolyl 4-hydroxylase isoenzymes in comparison to 2-oxoglutarate41

2-Another study made use of recombinant polypeptide constructs bearing either of the two prolyl residues on HIF-1α that are targeted for hydroxylation to induce nuclear accumulation and stabilization of HIF-1α in transfected cell 49 Stimulation of angiogenesis was shown by formation of tubules by microvascular endothelial cells co-cultured with these transfected cells and increased vascularization of subcutaneously implanted polyurethane sponges injected with the fusion protein in a murine model49

There are three reports on stimulation of angiogenesis using pharmaceutically induced prolyl hydroxylase inhibition to stabilize HIF-1α, all published shortly before this research study was begun50,51,52 Warnecke et al (2003) showed accumulation of HIF-1α

in western blots of lysates from human HT1080 fibrosarcoma cells, rat PC12W

pheochromocytoma cells and mouse 3T3L1 embryonic fibroblast like cells treated Mimosine, ethyl 3,4-dihydroxybensoate (3,4-DHB) and 6-chloro-2-hydroxychinolin-2-carbonic acid-N-carboxymethylamid (S956711) Their study also presented some evidence of angiogenesis simulation by these prolyl hydroxylase inhibitors using a rat sponge model, where implanted sponges repeatedly injected with the inhibitors had increased invasion of vascularized tissue into the sponge centers50

L-Around the same time, two independent groups examined the angiogenic effects of ciclopirox olamine (CPX) and published completely opposing conclusions Linden et al (2003) demonstrated HIF-1 induction by CPX using a HIF-1 dependent reporter gene activity in transfected HRCHO5 cells and Western blot of HepG2 hepatoma cells They also found that CPX increased VEGF expression in HepG2 cells using northern blot and ELISA CPX appeared to promote angiogenesis using the chick embryo chorioallantoic membrane assay (CAM)51

The other group (Clement et al, 2002) suggested that CPX inhibits endothelial cell growth and angiogenesis due to the down-regulation of collagen production via inhibition

of prolyl hydroxylase and deoxyhypusine53 They did not look at HIF-1 expression or HIF-1 target genes However, they showed that HUVEC proliferation and tube formation

in Matrigel was inhibited by CPX They also reported that CPX inhibited chick aortic ring sprouting compared to controls Surprisingly, this group had also used DFO and L-mimosine for comparison with CPX DFO and L-mimosine showed no stimulation of tube formation and aortic ring sprouting at lower concentrations and inhibition at higher

concentrations of 100-200 µM DFO and 400 µM L-mimosine53, which had been reported

by other groups to induce HIF-1α42,50

In the third report, Knowles et al (2004)52 described the angiogenic effects of hydralazine hydrochloride (HDZ) HDZ induced an endothelial cell specific increase in proliferation and stabilization of HIF-1α Endothelin-1 levels were also elevated in the conditioned medium from HUVEC culture Although they were unable to detect significant amounts

of VEGF in the HUVEC culture supernatant, VEGF levels were increased in the culture supernatant of human vascular smooth muscle cells (HVSMC) and MDA468 cells (breast carcinoma cell line) Intravenous administration of HDZ in mice induced an increase in blood vessel density in subcutaneously implanted polyether sponge There was also increased stabilization of HIF and VEGF expression in various organ tissues

As the studies published on CPX and mimosine were controversial and in vitro

angiogenesis assays used in the other studies were limited to endothelial cell proliferation alone, it was important to further investigate and compare the potential angiogenic effects

of some of these PHi in vitro

1.8 In vitro Angiogenesis Assays

Existing in vitro angiogenic assays have been mainly developed and used for testing

effect of anti-angiogenic compounds Although some of these assays are able to show pro-angiogenesis induced by growth factors such as VEGF, they pose some limitations

for testing the potential angiogenic effects prolyl hydroxylase inhibitors One of the most

commonly used in vitro angiogenesis assays is the Matrigel tube formation assay

Matrigel is an extracellular matrix extract from Engelbrecht-Holm-Swarm mouse tumour cells with abundant amounts of laminin, type IV collagen, perlecan (or heparan sulfate proteoglycan) and entactin (or nidogen)54 Basal levels of growth factors such as basic fibroblast growth factor (bFGF), platelet-derived growth factor (PDGF), insulin-like growth factor-1 (IGF-1) and transforming growth factor-β (TGF-β) are present in various quantities and contribute to spontaneous induction of angiogenesis This is useful for testing anti-angiogenic drug effects, but presents difficulties in showing augmentation of angiogenic effects Although growth factor-depleted formulations of Matrigel are available to allow some room for stimulation of angiogenesis, angiogenesis still proceeds rapidly Extensive and time-consuming analysis is required to measure the length of tubule formation and the number of nodes to quantify angiogeneis, although some software algorithms have been developed for this purpose Furthermore, there is the issue

of cost as Matrigel is expensive and significant quantities are required to test a number of substances in a range of concentrations

In addition, only one cell type (endothelial cells) is involved in most in vitro angiogenesis

assays, such those focusing on endothelial cell migration, proliferation and tubule formation in three dimensional matrices (such as Matrigel or fibrillar collagen gels) As angiogenesis is a complex process involving pro-angiogenic factors, receptors and cellular signaling, other surrounding cell types may be required This is particularly

important when angiogenesis is stimulated via drug stabilization of HIF-1α and subsequent upregulation of its gene targets, as the cells responsible for manufacturing

these target proteins must be present in order for in vitro angiogenesis to occur For

example, endothelial cells do not produce much VEGF, but express VEGF receptors 1 and 2 for ligand binding27

Currently, there is no in vitro assay that can accurately model the complexity of

angiogenesis While the aortic ring sprouting assay or such similar organ culture includes

surrounding cells and extracellular matrix and thus more closely mimic the in vivo

situation, it is difficult to quantify and growth requirements may differ between the explant and cellular outgrowth55

We therefore explored the use of co-culture of endothelial cells with stromal cells such as fibroblasts56,57 as our in vitro angiogenesis model for testing PHi The presence of

fibroblasts would support any PHi induced upregulation of VEGF production The simplicity of the assay also offered potential manipulations to ensure that angiogenesis was inducible rather than spontaneous within the system

1.9 In vivo Angiogenesis Models

Several in vivo angiogenesis models are available for drug screening, each varying in the technical skills required for accuracy and repeatability of the assay results As with the in vitro assays, some of the in vivo models are more effective for testing anti-angiogenic

substances because of naturally occurring angiogenesis or as a result of non-specific immune responses

i) Subcutaneously Implanted Sponges

As described earlier, the effect of systemic administration of 5 mg/kg HDZ has been investigated on neovascularization of subcutaneously implanted sponges Matrigel plugs may also be used in place of more solid polymer sponges

While there was higher vessel density in the sponges from HDZ treated mice compared to controls after 6 days, the vessel density decreased over time (at 14 and 21 days) with HDZ treatment In contrast, vessel density was increased at 14 and 21 days in controls At these time points, there were no significant differences between HDZ treatment and controls, but the mean vessel density in the control group had slightly exceeded that of the HDZ group This was attributed to the decreasing volume of the leading edge deeper into the sponge, as this was where the highest blood vessel concentration occurred Such difficulties in comparison between the control and treatment group limits the accuracy in the assessment of angiogenesis, particularly if effects of different substances are to be compared A local release of the substances rather than systemic administration would also be more applicable for neovascularization of tissue engineered constructs

ii) Chick Chorioallantoic Membrane (CAM)

The chick chorioallantoic membrane (CAM) assay was used in the CPX study Inert polymer discs containing solvent (Control) or 50 mM CPX were placed onto the CAM on

Day 9 of embryonic development It was reported that in CPX treated CAMs, there were numerous newly formed vessels after 2 days, in a radial arrangement from the center of the disc However, no scoring of these vessels were reported and the representative images provided only showed 2 CAMs each from the control (n=7) and CPX group (n=10) at Day 11 The corresponding image of the same CAM at Day 9 when the discs were first introduced was also not shown, thus it was difficult to assess the extent of the angiogenic effects As a comparatively high concentration of 50 mM CPX was used

(versus 0-20 µM CPX in vitro) and CPX is fairly insoluble, turbidity of the polymer discs

were observed The actual concentrations of CPX in CAMs may differ, making comparisons difficult and the results may be less convincing

The CAM assay is widely used for studying angiogenesis as it is relatively simple and inexpensive As it has a well developed vascular network, it is difficult to distinguish new capillaries from existing ones Corresponding images of the same CAM should be made

at the start and end of the treatment to enable more accurate assessment if the angiogenesis is induced by treatment or potential inflammatory responses to the procedure of cutting out a window in the egg shell and placing the disc on the membrane

iii) Zebrafish Embryo Angiogenesis

The zebrafish, or danio rerio, is evolving as an in vivo pharmaceutical screening

model58,59,60 As the zebrafish embryo is transparent, internal structures can be easily visualized under the microscope without requiring dissections, making it an ideal model for studies in developmental biology While embryo development is rapid, they remain

small enough to depend solely on the yolk and oxygen diffusion for survival during the first 6 days of embryogenesis60 The embryos are thus easy to maintain in the early stages

of development, requiring only salt water and can survive at room temperatures (optimally around 28ºC) Introduction of chemicals or drugs may be done easily by simply adding them to the zebrafish embryo culture water, as the embryos are readily permeable by diffusion59 The development of a variety of vascular-specific mutants and transgenic lines has further propagated the zebrafish as a useful vertebrate model for investigating angiogenesis59,60,66

The TG(fli1:EGFP) was a transgenic line developed by Lawson and Weinstein from NIH

(National Institutes of Health)61 The enhanced green fluorescent protein (EGFP) was transfected into the zebrafish and driven by the fli1 promoter, the earliest known marker

in zebrafish endothelial cells Embryos expressing the EGFP were grown to adulthood and crossed to identify germ line founders exhibiting high fluorescence levels This transgenic line enables real time, live observation of vasculature development in the zebrafish embryo using a fluorescence microscope

Current research has mainly focused on intersegmental vessels (ISVs) that run between each pair of somites from the dorsal aorta (DA) to the dorsal longitudinal anastomotic vessel (DLAV) These ISVs have an extremely regular arrangement and are formed by angiogenic sprouting around 23 hpf Angioblasts migrate from the DA to the DLAV and each ISV consists of 3-4 cells that form one T-junction at the DLAV and another at the

DA, joined by 1-2 connector endothelial cell, resembling the letter “Z”62

The out of bounds (obd) mutant zebrafish embryos have vessel patterning mutations,

where ISVs are highly disorganized and tortuous Angiogenic sprouting was also observed to occur earlier, around the 17-18 somite stage, compared to the 24 somite stage for wild type embryos62 When treated with anti-angiogenic substances such as SU5416, flavopiridol, genistein64 and paclitaxel, the ISVs were disrupted with several vessels missing between somites60,63,64

While anti-angiogenic studies were readily available, reports on pharmaceutical stimulation of angiogenesis in zebrafish embryos were not easily found The only relevant research published involved injection of human recombinant VEGF peptide or DNA constructs (to induce VEGF overexpession) into the yolk sac of zebrafish embryos and these induced changes in the subintestinal vessels (SIVs) 65,66 The SIVs develop in the ventral trunk of the zebrafish embryo, underneath the somites, and grow over the yolk These vessels are outgrowths of the posterior cardinal vein62, therefore they may similarly be used for assessment of angiogenesis in zebrafish embryos

In the context of this study, we wanted to determine the local angiogenic effect, therefore microinjection of PHi would be difficult to control compared to placing these substances into the water in which the embryos were raised While it was probable that the ISVs and SIVs would be the regions of interest where ectopic vessels might occur, there was no established criteria for assessing pro-angiogenesis Hence, an important implication from the study was demonstrating that transgenic zebrafish could be a suitable small vertebrate

model for screening potential drugs for inducing neovascularization in tissue engineering and advanced wound healing technology This model would have the advantage of being simple, rapid and cheap, compared to mouse or rat implantation models, making it ideal

as an in vivo model for drug screening, although it cannot fully replace the need to verify

neovascularization in implantation experiments in small mammals

1.10 Fibrosis in Tissue Engineering – A Result of Foreign Body Reaction and Wound Healing

Regenerative medicine and tissue engineering often inevitably require surgery for implantation of tissue engineered constructs Surgery creates wounds, therefore provoking a wound healing response at the surgical site

Wound healing involves phases of inflammation, proliferation and remodeling In the inflammatory phase, clotting leads to hemostasis, followed by the release of plasma like fluid containing enzymes, proteins and antibodies into the tissue space This stimulates neutrophils and monocytes to migrate to the wound site Neutrophils remove bacteria and foreign material, while monocytes become phagocytic macrophages that resorb necrotic tissue, phagocytose debris and clean up the wound site Macrophages also release growth factors that mediate tissue repair and regeneration, characterized by fibroblast, smooth muscle cell and endothelial cell recruitment and proliferation, stimulation of angiogenesis and synthesis of collagen, proteoglycans and other extracellular matrix proteins to form granulation tissue 67,68,69

In addition to the wound healing response, implantation of biomaterials invokes a host immune response, also known as the foreign body reaction Macrophages adhere to the surface of the biomaterial and spread as they try to phagocytose the foreign body However, as they are unable to digest or engulf a large mass of biomaterial, they fuse to form multinucleated giant cells These giant cells signal for increased fibroblast migration and collagen production, resulting in fibrous encapsulation of the implanted biomaterial67, 70 This avascular and collagenous barrier is usually 50-200 µM thick Excessive amount of collagen production and deposition results in fibrosis This uncontrolled biological encapsulation can impede the survival and success of the implant

as vascularization may be compromised, leading to necrosis and cellular death.70

Foreign body reactions occur with nearly all implanted materials with varying degrees This includes materials that are supposedly deemed biocompatible, although some of them promote wound repair and regeneration more than foreign body encapsulation70 This is a major limiting step in ensuring the survival and function of implanted tissue constructs, controlled drug delivery devices and sensors To tackle this problem, some researchers have explored the use of scaffold free tissue constructs, while many more are involved with surface modifications of the biomaterials to limit foreign body encapsulation The use of anti-fibrotic substances to minimize and reduce the formation

of a fibrotic sheath around an implanted tissue engineered construct would therefore be of great benefit in tissue engineering

1.11 Collagen Biosynthesis

Collagen forms the major protein component of the extracellular matrix The collagen molecule comprises of three polypeptides that fold to form triple helical domains Each polypeptide is determined by the high glycine and amino acid contents in specific repeating triplets of Gly-X-Y, where X is often proline and Y is often hydroxyproline71,72

As the smallest amino acid, glycine at every third position in the repeating sequence allows close packing of the three helical chains coiled around a common axis72

Figure 4 Collagen synthesis, processing and assembly

The biosynthesis of the collagen molecule begins with the transcription of collagen genes for formation of procollagen fibrils71 Post-translational modifications of procollagen include proline hydroxylation, which renders it thermally stable71,72 Procollagen is then secreted into the extracellular space, where N- and C-proteinases cleave off the N- and C-terminal propeptides to form triple helical collagen molecules Subsequently, these collagen molecules align in quarter staggered arrays and form collagen fibers71 (Figure 4)

1.12 Inhibition of Collagen Biosynthesis Using Prolyl Hydroxylase Inhibitors

Prolyl-4-hydroxylase is the key enzyme involved in the proline hydroxylation step in collagen biosynthesis Prolyl residues on procollagen are hydroxylated, giving rise to hydroxyprolines found in the Gly-X-Y sequences of collagen fibers71,72

Chemical inhibition of collagen prolyl hydroxylases by prolyl hydroxylase inhibitors such as pyridine dicarboxylates and its derivatives73,74, DFO53, 2,2’-dipiridyl53, CPX53, deferiprone53, mimosine53,75, HDZ76,77,78, and phenanthrolinone derivatives79 have been shown to inhibit collagen deposition by disrupting this hydroxylation step As a result, the procollagen molecules do not assemble into triple helical chains and are thus retained intracellularly instead of being secreted into the extracellular matrix73 This function of prolyl hydroxylase inhibitors has been exploited towards the design of antifibrotic drugs80,81, most of which are currently being developed and patented by Fibrogen, Inc (South San Francisco)79,82

1.13 Selection of Prolyl 4-Hydroxylase Inhibitors

Three PHi have been selected for investigation in this research study They are hydralazine hydrochloride (HDZ), ciclopirox olamine (CPX) and pyridine-2,4-dicarboxylate (PDCA)

HDZ and CPX were chosen because they are both FDA approved substances HDZ is approved for systemic administration as a vasodilator to treat severe hypertension52,83, while CPX, a structural analogue of mimosine, is used topically for treatment of fungal and yeast infection of the skin or mucosa51,53,84

PDCA is a well established and effective inhibitor of collagen prolyl 4-hydroxylase Several analogues of PDCA were developed and patented by the former Hoechst AG These patents were subsequently purchased by Fibrogen, Inc (South San Francisco) and some of these compounds have been found to prevent HIF-1 degradation46 FG-4592 and FG-2216, for example, are oral compounds currently tested in phase I and phase II clinical trials for treatment of anaemia and are thought to stimulate erythropoiesis by stabilizing HIF85,86

As the use of PHi to improve local vascularization for tissue engineering applications has not been reported, the results of this research study will be regarded as a novel approach towards this problem

2 MATERIALS & METHODS

2.1 Cell Culture

Normal human fetal lung fibroblasts (IMR-90) (American Tissue Culture Collection, VA, USA) were routinely cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and Penicillin Streptomycin (PS), all obtained from Invitrogen, Carlsbad, CA at 37°C in a humidified atmosphere of 5% CO2

Human umbilical vein endothelial cells (HUVEC; Lonza, Walkersville, MD, USA) were similarly cultured for expansion at 37°C, 5% CO2 using EGM™-2MV Single Quots Bulletkit (Lonza, Walkersville, MD, USA) which comprised of endothelial cell basal medium-2 (EBM-2) supplemented with 5% FBS, epidermal growth factor (EGF), hydrocortisone, gentamicin sulfate/amphotericin-B (GA-100), vascular endothelial growth factor (VEGF), fibroblasts growth factor-B (FGF-B), insulin-like growth factor-1 (R3-IGF-1) and ascorbic acid

When HUVECs were co-cultured with fibroblasts, a reduced serum medium was used (EBM-2, 0.5% FBS, PS) without any other growth factors or supplements apart from the test substances Trypsin-EDTA (Invitrogen, Carlsbad, CA, USA or Lonza, Walkersville,

MD, USA) was used to dissociate cells from the culture flasks for cell seeding Trypsin Neutralizing Solution (TNS, Lonza, Walkersville, MD, USA) or DMEM (10% FBS) was used to neutralize trypsin for HUVECs and fibroblasts respectively

2.2 Chemicals and Reagents

Hydralazine hydrochloride (HDZ), ciclopirox olamine (CPX) and 4,dicarboxylic acid (PDCA), phosphate buffered saline (PBS) and bovine serum albumin (BSA) were obtained from Sigma Aldrich, St Louis, MO, USA Human recombinant vascular endothelial growth factor (VEGF) was obtained from Chemicon, Temecular, CA, USA and used as a positive control Porcine collagen type I and human placenta collagen type V standards were obtained from Koken (Tokyo, Japan) and Sigma Aldrich (St Louis,

pyridine,2-MO, USA) respectively

2.4 Preparation of Prolyl Hydroxylase Inhibitors and Recombinant Human VEGF

HDZ (Molecular Weight of 196.64 g) is readily soluble in water A 100 mM stock solution was prepared by dissolving 20mg per ml of ultrapure water A primary stock solution of 0.5 M CPX (Molecular Weight of 268.35 g) was prepared by dissolving 67

mg of CPX in 500 µl methanol and stored at -20°C in aliquots A secondary stock solution of 1 mM CPX was prepared prior to use, by (1:500) dilution of 0.5 M CPX with

* k denotes 1,000 For example, 150k is 150,000 and 10k is 10,000

culture medium PDCA (Molecular Weight of 185 g) was solubilized by and acid-base reaction using 110 mg of PDCA powder with 120mg sodium bicarbonate (NaHCO3) and

2 ml of ultrapure water to form a 300 mM solution of pyridine-2,4-dicarboxylate (for simplicity, this salt solution will be used synonymously with pyridine-2,4-dicarboxylic acid and will be referred to similarly as PDCA throughout this report.)

2.5 Sequential Co-cultures

Fibroblasts were seeded in 24-well plates (LumoxTM, Greiner Bio-One, Frickenhausen, Germany) maintained in DMEM (10% FBS) to proliferate to confluence After 3 days, the medium was changed to EBM-2 (0.5% FBS) ± PHi or VEGF HUVECs were then seeded on top of the fibroblast layer and the co-culture was propagated for 3 days HUVEC morphology was identified by von Willebrand Factor (vWF) immunostaining

2.6 Admixed Co-cultures

Fibroblasts and HUVECs were admixed, seeded in 24-well plates and cultured in DMEM (10% FBS) with or without 30 µg/ml L-ascorbic acid (phosphate salt; Wako Pure Chemical Industries, Osaka, Japan) for 3 days The medium was changed to EBM-2 (0.5% FBS) ± PHi or VEGF and the co-culture was propagated for a further 3 days HUVEC morphology was identified by von Willebrand Factor (vWF) immunostaining

2.7 Immunohistochemistry

Immunohistochemistry was performed on cells in 24-well plates (Cellstar or LumoxTM, Greiner Bio-One, Germany), Slide Flasks (NUNC, Roskilde, Denmark) or 8-well Labtek

Chamber Slides (NUNC, Rochester, NY, USA) For 8-well Labtek Chamber Slides, approximately 10k cells were seeded per well

at 4°C Slides were placed in a humidified chamber to prevent fluids from drying out Unbound primary antibodies were removed by washing 3 times with PBS, 5 minutes per wash Appropriate fluorescent dye conjugated secondary antibodies were then incubated for 30 minutes followed by 3 x PBS washes and the cells were imaged using a fluorescence microscope (Olympus IX31 or Nikon TE600)

Primary antibodies used were polyclonal rabbit anti human vWF (Dako Cytomation, Glostrup, Denmark), monoclonal mouse anti human CD31 (Dako Cytomation, Glostrup, Denmark) monoclonal mouse anti-human HIF-1α (BD Biosciences, Franklin Lakes, NJ, USA), monoclonal mouse anti protein disulfide isomerase (PDI; Affinity Bioreagents, Golden, CO, USA) and polyclonal rabbit anti human collagen I (Chemicon, Temecular,

CA, USA) Secondary antibodies used were Alexa Fluor 488 chicken anti mouse IgG, Alexa Fluor 488 chicken anti rabbit IgG, Alexa Fluor 546 goat anti mouse IgG, Alexa Fluor 546 Alexa Fluor 546 goat anti rabbit IgG or 594 goat anti rabbit IgG (all from Molecular Probes, Invitrogen, Eugene, Oregon, USA) All antibodies were diluted in 3%

BSA/PBS 0.5 µg/ml 4',6-diamidino-2-phenylindole, dihydrochloride (DAPI) was used for nuclear staining and was added together with secondary antibodies

For HIF-1α staining, the Dako EnVisionTM Kit was also used for visualization of the staining This kit consisted of a labeled polymer-horse radish peroxidase (HRP) goat-anti-mouse IgG secondary antibody and a diaminobenzidine (DAB) chromogenic substrate that turns brown with peroxidase activity Stained cells were imaged using light microscopy (Olympus IX31 or Nikon TE600)

Basic image processing such as background subtraction, adjustment of brightness and contrast, was applied to all images where necessary to improve visualization of the detected signal Fluorescence microscopy images were taken using the same exposure times within each set of immunostained experiments to allow comparisons Care was taken to ensure that the images being compared were image processed in a similar manner to prevent biased enhancements

2.8 Quantification of Angiogenesis by Image J and Metamorph

The immunostained co-cultures in 24-well plates were imaged on a Bioimaging Station comprising a motorized Nikon TE600 microscope, a Xenon illuminator with Ludl stage and shutter, and equipped with a camera (Photometrics CoolSNAP HQ, Roper Scientific) Images were acquired at 2x magnification using the Metamorph ® Imaging System software (Molecular Devices) 9 images (in a 3 x 3 format) were captured per well using the “Screen Acquisition” application

For each well, 4 images at sites 2, 4, 6 and 8 were chosen for image processing Image J 1.36b (Rasband, W.S., ImageJ, U S National Institutes of Health, Bethesda, Maryland, USA, http://rsb.info.nih.gov/ij/, 1997-2006) was used to manually trace out the capillary like structures formed by HUVEC and measure the length of the structures The average total tube length (measured in pixels) per image was computed as a measurement of the angiogenic effect A capillary like structure was defined as having a thickness of one cell width and a length of 2 cells or more The “Measure and Label” function in Image J was used to compute the total number and length of capillary like structures in each of the 4 images For each treatment, triplicate wells were analyzed and the means ± standard deviations obtained were computed from a total of 12 images per treatment for each experiment (4 images x 3)

2.9 VEGF Production

VEGF secreted into the culture medium of monocultures or co-cultures of fibroblasts and HUVEC were analyzed using a commercial human VEGF ELISA (enzyme linked immunosorbent assay) kit according to the manufacturer’s instructions (R&D Systems, Minneapolis, USA) Human recombinant VEGF standards or culture medium samples were placed into the 96-well microplate that had been pre-coated with a monoclonal VEGF specific antibody The VEGF present in the sample or standards were captured by the immobilized antibody while unbound substances were washed away An enzyme-linked polyclonal antibody specific for VEGF was then added to the wells After washing away any unbound antibody-enzyme reagent, a substrate solution was added to the wells

A color change developed in proportion to the amount of enzyme and hence, amount of VEGF bound in the initial step After 20 minutes, the reaction was stopped and the intensity of the color was analyzed by measuring the absorbance at 450 nm using a microplate reader (PheraStar, BMG Labtechnologies GmbH, Offenburg, Germany)

2.10 Gel Electrophoresis and Western Blotting

Cells treated with PHi for 4 hours were washed twice with cold HBSS and directly lysed with 30 µl of LDS sample buffer (Invitrogen, Carlsbad, CA) containing 5 mM DTT (US Biological, Swampscott, MA, USA) and protease inhibitors (Complete, Roche, Mannheim, Germany) on ice The cell layer was scraped after 5 min, transferred into microcentrifuge tubes and heated to 95°C for 10 min Cell lysates and molecular weight markers were loaded into pre-cast NuPAGE® Novex 3-8% Tris-Acetate gels (Invitrogen, Carlsbad, CA) and separated by gel electrophoresis The proteins were then transferred onto a nitrocellulose membrane (0.45 µm, Biorad, Hercules, CA, USA) using the XCell SureLockTM Mini-Cell (Invitrogen, Carlsbad, CA) Running buffers and transfer buffers were prepared according to the manufacturer’s instructions (Invitrogen, NuPAGE Tris Acetate SDS Running Buffer and NuPAGE Transfer Buffer with 10% methanol) Molecular weight markers used were Precision Plus Kaleidoscope Standards (Biorad, Hercules, CA) or Chemiluminescent Blueranger Marker Mix (Pierce Biotechnology, Rockford, IL, USA)

After blocking non-specific binding sites with 5% non fat dry milk/TBST (TBS: 50 mM Tris, 150 mM NaCl; TBST: TBS with 0.05% Tween-20), the blot was cut into two near

the 75 kDa molecular weight marker The two halves of the blot were separately probed with primary antibodies diluted in 1% milk/TBS for 90 minutes at room temperature or 4°C overnight, with gentle agitation The top half corresponding to proteins larger than

75 kDa was incubated with monoclonal mouse anti human HIF-1α (BD Biosciences) antibody while the bottom half corresponding to proteins below 75 kDa was incubated with either monoclonal mouse β-tubulin (Santa Cruz, Santa Cruz, CA, USA) or polyclonal rabbit β-actin antibody (Delta Biolabs, Gilroy, CA, USA) Unbound antibodies were removed by three washes of TBST, 5 minutes each The blots were then incubated for 1 hour at room temperature with HRP-conjugated goat anti-mouse or goat anti-rabbit secondary antibody (Pierce Biotechnology, Rockford, IL, USA) and subsequently washed again with TBST (3 x 5 min) to remove excess unbound antibodies Super Signal® West Pico substrate (Pierce Biotechnology, Rockford, IL, USA) was added and chemiluminescent signal from the reconstructed blot was captured using the VersaDoc™ MP 5000 Imaging System (BioRad, Hercules, CA, USA) A corresponding photograph of the blot was also captured to visualize the molecular weight markers in relation to the immunostained bands

2.11 Growth Factor Reduced Matrigel Assay

Growth factor reduced Matrigel™ (BD Biosciences) was thawed at 4°C 200 μl Matrigel™ was added to each well of a 24-well plate and allowed to polymerize at 37°C for approximately 1 hour 100k HUVEC were seeded into each well in 0.5 ml EBM-2 (0.5% FBS) ± PHi or VEGF and incubated at 37°C, 5% CO2 After 24 hours, the cells were imaged using phase contrast microscopy

2.12 Co-cultures in PLLA Scaffolds

PLLA felt scaffolds (poly-L-lactide, 45 mg/cc, 2 mm thick, Transome, Inc, Florida, USA) were cut into 8 mm diameter circles using an 8 mm biopsy punch The scaffolds were sterilized by soaking in 70% ethanol for 10 minutes, followed by two washes of HBSS and 10 minutes of exposure to ultraviolet light in the biosafety cabinet The scaffolds were then placed in 6-well plates and soaked in DMEM (10% FBS) for at least 1 hour prior to cell seeding

Each scaffold was seeded with an admixture of 1 million (1 x 106) fibroblasts with 200k HUVEC suspended in a small volume of approximately 30-50 µl culture medium Just prior to cell seeding, the media in and surrounding the scaffold was removed as much as was possible, so that the scaffold would hold all the fluid of the cell suspension without cells spilling out The cells were left to attach for approximately 2-3 hours before the media was carefully topped up with DMEM (10% FBS) supplemented with 30 µg/ml of L-ascorbic acid (phosphate salt; Wako Pure Chemical Industries, Osaka, Japan) After 3 days of co-culture, the medium was changed to 0.5% FBS EBM2 ± PHi or VEGF and the co-culture was further propagated for 3 days

The cells in the scaffolds were gently fixed using a stepwise methanol fixation (25%, 50%, 75% and 100% methanol mixed with PBS, each with gentle agitation for 10 minutes each) and rehydrated in a similar stepwise manner before immunostaining (75%, 50%, 25% methanol mixed with PBS, followed by one PBS wash, 10 minutes each) Immunostaining of vWF and DAPI was performed as described above, but the incubation

time for both primary and secondary antibodies was increased to 3 hours at room temperature with gentle agitation to ensure permeation of the antibody into the scaffold The morphology and arrangement of the cells within the scaffold were visualized using a confocal laser scanning microscope (Olympus FluoViewTM FV1000)

2.13 Collagen Biosynthesis in Cell Culture

Fibroblasts were seeded in 24-well plates (~50k cells/well) and allowed to attach overnight in DMEM (10% FBS) The next day, the media was replaced with 500 µl of 0.5% FBS DMEM supplemented with 30 μg/ml of L-ascorbic acid (phosphate salt; Wako Pure Chemical Industries, Osaka, Japan), with or without PHi After 48 hours, the culture medium was transferred to separately labeled tubes and the cells were washed with HBSS The cell layer and culture medium were then separately subjected to peptic digest under acidic conditions to extract collagen87 Briefly, 50 µl of 1 mg/ml of pepsin (from porcine gastric mucosa, Roche, Mannheim, Germany) dissolved in 1N HCl was added to each tube of approximately 500 µl of conditioned medium and vigorously mixed for 2 hours at room temperature using a rotational mixer (final concentration: 0.1 mg/ml pepsin) 100 µl of 0.25 mg/ml pepsin with 0.5% Triton-X100 (BioRad, Hercules, CA, USA) and Phenol Red was prepared from the 1 mg/ml pepsin solution, added to the cell layer in each well and incubated for 2 hours on a rotating orbital shaker The reaction was the stopped by adding 30 µl and 60 µl of 1N NaOH to each sample of digested cell layer and conditioned medium respectively The digested cellular components were transferred into separately labeled tubes Both cellular and medium components of samples may be stored at -20°C

Cellular and medium samples were thawed and mixed with appropriate amounts of 4 x LDS sample buffer (usually 15 µl sample + 5 µl sample buffer), heated to 95ºC for 5 minutes and separated using NuPAGE® Novex 3-8% Tris-Acetate gels as described above Protein bands were stained with the SilverQuest™ Silver Staining Kit (Invitrogen, Carlsbad, CA) according to the manufacturer’s microwave protocol This kit is based on

a highly sensitive chemical reaction that involves reduction of silver ions to metallic silver on protein bands and has a detection threshold of approximately 0.3 ng Stained gels were scanned while wet, using a GS-800™ Calibrated Densitometer (Bio-Rad, Hercules, CA), and densitometric analysis was performed using the Quantity One v4.5.2 image analysis software (Bio-Rad, Hercules, CA) Collagen bands were quantitated by defining each band with the rectangular tool and local background subtraction was applied Similar cultures were prepared in parallel and immunostained for collagen and PDI as described above

2.14 Zebrafish Embryo Collection and Drug Treatment

Zebrafish embryos were generated by natural pair-wise mating and raised at 28.5°C in embryo water (0.2 g/l of Instant Ocean Salt in distilled water) as described in the zebrafish book88 Approval had been given by the NUS-Institutional Review Board (IRB) and NUS Institutional Animal Care and Use Committee (IACUC 022/06) to conduct

these animal experiments For each mating 4-5 pairs of TG(fli1:EGFP) mature adult

males and females (Zebrafish International Resource Center, University of Oregon) were set up The male and female fish were separated by dividers to control the time of mating

and ensure that embryos produced were at similar stages of development At 6 hours post fertilization (hpf), the embryos were sorted for viability and developmental stage (shield stage) Approximately 100-200 healthy embryos were placed in 100 mm Petri dishes containing 30 ml embryo water ± PHi or VEGF The embryo water was replaced with fresh embryo water ± PHi or VEGF at 24 hpf, 48 hpf and 72 hpf The embryos were examined daily for viability, gross morphological defects and blood vessel development using a fluorescent microscope Dead or unhealthy-looking embryos were removed

2.15 Screening Ectopic SIV in Zebrafish Embryos

At 72 hpf, the embryos were anesthetized using 0.05% 2-phenoxyethanol (Sigma Aldrich) mixed with embryo water 1-3 embryos were placed into each well of a 96 well plate and each embryo was examined for presence of ectopic subintestinal vessels (SIV) as a indication of a pro-angiogenic effect using the following criteria: (a) presence of vessels which terminate as spikes (instead of looping to join another vessel) and/or (b) extension

of the SIV basket into the yolk extension region with more than 7 vertical branches within the basket Embryos not expressing green fluorescent protein (GFP) were excluded, as were the embryos in which the SIV is not clearly visible due to weak fluorescence

2.16 Collagen Analysis of Zebrafish Embryos by Peptic Digest and SDS-PAGE

After screening, the anesthetized embryos were placed into microcentrifuge tubes (30 embryos per tube) and quickly washed with distilled water The volume of water within each tube was reduced to approximately 0.1 ml before the embryos were frozen and

stored at -20°C For collagen extraction, the embryos were thawed and digested by adding 20 µl of 1mg/ml pepsin dissolved in 1N HCl (volumes were normalized between samples and the final concentration was 200 µg/ml pepsin) for 2 hrs at room temperature The reaction was neutralized with 30 µl of 1N NaOH and cleared by centrifugation for 15 min at 13,000 rpm 100 µl of the supernatant from each sample was extracted, mixed with 36 µl of 4x LDS sample buffer and heated to 95°C for 5 minutes 10 µl of sample (representing approximately 1.5 zebrafish embryos) was loaded per lane in NuPAGE®Novex 3-8% Tris-Acetate gels and analyzed by gel electrophoresis and analyzed silver staining as described earlier

2.17 Cytotoxicity Assay

Fibroblasts and HUVECs were seeded in 96 well plates and allowed to attach over night

in 10% FBS DMEM The cells were then treated with PHi in 0.5% FBS DMEM or 0.5% FBS EBM-2 for 16 hours Cytotoxicity was assayed using the Vybrant Cytotoxicity Assay Kit (Molecular Probes, Eugene, Oregon, USA) according to the manufacturer’s instructions The cytotoxicity assay is based on the release of cytosolic glucose 6-phosphate dehydrogenase (G6PD) from damaged cells into the surrounding culture medium Detection of G6PD is by a two-step enzymatic process that leads to the reduction of resazurin into the red-fluorescent resorufin (absorption: 563 nm, emission:

587 nm) The fluorescent signal was recorded using a microplate reader at 530-560 nm excitation and 580-600 nm emission (BMG Labtechnologies GmbH, Offenburg, Germany) and is proportional to the amount of G6PD released into the cell medium, which in turn correlates with the number of dead cells in the sample

3 RESULTS

3.1 In vitro Angiogenesis

3.1.1 CD31 and von Willebrand Factor: Markers for Endothelial Cells

CD31 and von Willebrand factor (vWF) are common markers used to identify endothelial cells (Figure 5) The optimal antibody dilutions required for staining HUVECs were found to be 1:100 for monoclonal mouse anti-human CD31 (green) and 1:1000 for polyclonal rabbit anti-human vWF (red) Negative conjugate controls (no primary antibodies) demonstrated that there was no non-specific binding of the secondary antibodies used (Alexa Fluor 488 chicken anti-mouse and Alexa Fluor 546 goat anti-rabbit), therefore the positive staining observed was due to the specific primary antibody used

Figure 5 HUVECs were immunostained with either monoclonal mouse anti-human

CD31 (green) or polyclonal rabbit anti-human vWF (red), with DAPI nuclear staining Alexa Fluor 488 chicken anti-mouse (AF488) and Alexa Fluor 546 (AF546) goat anti- rabbit secondary antibodies were used Negative conjugate controls (CC) showed absence of non-specific binding of secondary antibodies These images were obtained by overlapping corresponding pictures captured using appropriate filters for emission of fluorescence signal and DAPI 40x magnification

CD31 CD31 + DAPI

vWF vWF 25μm + DAPI 25μm 25μm

25μm 25μm

AF546 CC

CD31 is an endothelial cell surface marker, as shown by the stronger staining of the outlines of the cells vWF is a protein synthesized by endothelial cells and is stored in the Weibel-Palade bodies, as seen by granular staining of the cytoplasm with dark unstained elliptical region within the cells corresponding to the nuclei

Fibroblasts generally do not stain positively for either CD31 or vWF We tested this in IMR-90, the fibroblast cell line used in our research study and found that this was the case in monocultures of fibroblasts (data not shown) This allowed HUVECs to be easily distinguished from fibroblasts when co-cultured Figure 6 shows an example of such a sequential co-culture that had been double stained for CD31 and vWF The vWF (red) and CD31 (green) staining were co-localized on the HUVECs in the co-culture, enabling visualization of their orientation and morphology In this instance, the HUVECs had formed elliptical clusters (yellow rectangles) instead of assuming a cobblestone morphology that is typical of monocultures DAPI staining, however, revealed a confluent layer of cells that were evenly spread The red vWF staining reveals the orientation and morphology of the HUVECs in the co-culture, while numerous DAPI stained nuclei not associated with vWF staining indicated the presence of the confluent underlying layer of fibroblasts

As there was no significant difference in the information obtained with either vWF or CD31, the use of either one staining alone would be sufficient to visualize HUVEC morphology vWF was preferred as the staining appeared stronger and it was more economical, given the optimal antibody dilution of the polyclonal rabbit antibody from

Dako Cytomation was 1:1000, while the monoclonal mouse CD31 antibody from the same company required an antibody dilution of 1:100

Figure 6 HUVECs were sequentially co-cultured on top of a confluent layer of

fibroblasts for 3 days and double stained for vWF (red) and CD31 (green) vWF and CD31 were found to be co-localized on HUVECs, which had assembled into elliptical clusters (two such elliptical clusters marked by yellow rectangles) DAPI nuclear staining indicates the presence of an evenly spread, confluent layer of cells, revealing the fibroblasts which do not stain positively for CD31 and vWF

From Figure 6, it was also evident that images taken at 4x magnification covered a larger area within the well and would be a better representative of the general orientation and assembly of HUVECs in the co-culture without significant loss of information, compared

to higher magnifications of 10x or 20x As a confluent layer of fibroblasts is required and

fewer HUVEC are seeded within each well, the DAPI staining may overwhelm the immunofluorescence signal from the HUVECs Therefore, for the remainder of this report, vWF staining of co-cultures will primarily be shown without DAPI overlay at 4x magnification, to indicate HUVEC morphology and arrangement

3.1.2 Development of Sequential Co-culture for Pro-angiogenesis

In the development of a suitable co-culture assay for assessment of pro-angiogenesis, the aim was to obtain an inducible system which produces distinct differences between untreated controls and angiogenesis stimulants The sequential co-culture protocol was adapted from a study published by Friis et al57 as it was more appropriate to use low serum (0.5% FBS) EBM-2 instead of EGM-2MV to eliminate simulatory growth factors and FBS inherently present in the culture medium The reported cell seeding densities for fibroblasts did not produce a confluent layer after 3 days of culture, thus optimization of cell seeding was necessary

It was determined that when approximately 50k IMR90 fibroblasts per well were seeded into 24-well plates and allowed to proliferate in DMEM (10% FBS), a confluent monolayer would be obtained in each well after 3 days 2.5k, 5k, 10k or 20k HUVECs per well were then seeded on top of this confluent fibroblast layer and co-cultured for 4 days in 0.5% FBS EBM-2 with or without 100 µM HDZ It was observed that when lower seeding densities such as 2.5k and 5k per well were used, the HUVECs were confined mainly to one section of the well, with sparsely scattered or no cells in the remainder of the well (Figure 7) However, the arrangement and morphology of the