The Role of Transglutaminase 2 in Cell Motility Processes and Its

The role of transglutaminase 2 in cell motility processes and its

effect on cutaneous wound healing

by Ting Wai Yiu A Thesis Submitted to the Graduate University of New South Wales in fulfillment of the Requirements for the degree of DOCTOR OF PHILOSOPHY

Faculty of Medicine University of New South Wales Sydney December, 2011

ORIGINALITY STATEMENT “I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.”

CONTENTS

CONTENTS ...... iii LIST OF TABLES ...... vii LIST OF FIGURES ...... viii ACKNOWLEDGMENT ...... x ABSTRACT ...... xi 1. INTRODUCTION ...... 1 1.1 Prelude ...... 1 1.2 TG2 structure and biochemical function ...... 1 1.2.1 TG2 protein, expression and its regulation ...... 1 1.2.2 Early work on TG2 three dimensional structure ...... 3 1.2.3 Crystal structure of TG2 ...... 4

1.2.4 The two sides of the same coin – TGase and Gh ...... 6 1.2.5 Other in vitro TG2 biochemical activities ...... 10 1.3 Animal models ...... 11 1.4 Biological roles of TG2? ...... 11 1.4.1 Coeliac disease ...... 12 1.4.2 Ocular diseases ...... 13 1.4.3 Cancer 13 1.4.4 Osteogenesis and osteoarthritis ...... 15 1.4.5 Cardiovascular conditions ...... 16 1.4.6 Neurodegenerative disorders ...... 17 1.4.7 Contribution of TG2 to wound healing? ...... 18 1.4.8 Cutaneous wound healing overview ...... 21 2. MATERIALS AND METHODS ...... 29 2.1 Recombinant rat TG2 protein production ...... 29 2.1.1 Expression vector and host ...... 29 2.1.2 TG2 expression and isolation ...... 29

iii

2.1.3 TG2 purification by glutathione sepharose column affinity chromatography ...... 30 2.1.4 Fast protein liquid chromatography (FPLC) ...... 30 2.1.5 Western blotting for TG2 ...... 31 2.1.6 Transglutaminase assay by 3H-putrescine incorporation ...... 32 2.2 Mouse breeding and housekeeping ...... 32 2.2.1 Mouse genotyping ...... 33 2.3 In vivo wound healing assay ...... 35 2.4 White blood cell counting from wound histology ...... 35 2.4.1 Sample isolation and fixation ...... 35 2.4.2 Sectioning ...... 36 2.4.3 Hematoxylin and eosin staining for granulocytes ...... 36 2.5 Wound analysis ...... 37 2.5.1 Microscopic analysis of wound sections ...... 37 2.5.2 Wound RNA extraction ...... 37 2.5.3 Microarray analysis ...... 38 2.5.4 Reverse transcription of wound total messenger RNA ...... 38 2.5.5 Quantitative PCR analysis of wound RNA ...... 39 2.6 Mouse embryonic fibroblast isolation ...... 39 2.6.1 Freezing, thawing and maintaining MEF cultures ...... 40 2.6.2 MEF growth curves ...... 41 2.7 In vitro cell adhesion assay ...... 41 2.7.1 Preparation of wildtype and point mutant TG2 proteins ...... 41 2.7.2 Detection of Fn-bound TG2 in tissue culture wells ...... 41 2.7.3 Cell adhesion determination using crystal violet staining ...... 42 2.8 In vitro cell spreading assay ...... 43 2.8.1 Cloning of TG2 into a GFP-bicistronic vector ...... 43 2.8.2 TG2 transfection into MEF ...... 45 2.8.3 Detection of cytosolic and membrane TG2 ...... 46 iv

2.8.4 Cell spreading assay ...... 47 2.9 Detection of cell surface TG2 by pull-down assay ...... 47 2.9.1 Immunostaining of cell surface TG2 ...... 48 2.10 RhoA and Rac1 profiling in MEF ...... 49 2.10.1 Expression of GST-Rhotekin Rho-binding or GST-Pak Rac-binding domains from E. coli ...... 49 2.10.2 Glutathione sepharose beads equilibration ...... 49 2.10.3 Pull-down assay of activated RhoA or Rac ...... 50 2.11 In vitro scratch wound assay ...... 51 2.12 Statistical analysis ...... 51 3. THE EFFECT OF TG2 ON IN VIVO CUTANEOUS WOUND HEALING ...... 54 3.1 Introduction ...... 54 3.2 Results ...... 54 3.2.1 In vivo wound healing assay in 129T2-TG2 mouse ...... 54 3.2.2 Exogenous addition of purified TG2 to healing skin wounds ...... 55 3.2.3 Monocytes and neutrophils at the wound site ...... 57 3.2.4 Microarray analysis of 129T2-TG2+/+ and 129T2-TG2-/- wounds ...... 58 3.2.5 Quantitative analysis of wound-related genes using real time (RT)-PCR60 3.3 Discussion ...... 79 4. THE ROLE OF TG2 IN CELL MOTILITY PROCESSES: CELL ADHESION, SPREADING, MIGRATION ...... 83 4.1 Introduction ...... 83 4.1.1 The fibroblast and the extracellular matrix ...... 83 4.1.2 Cell motility: from adhesion to spreading to migration ...... 86 4.1.3 TG2 in cell motility processes ...... 92 4.2 Results ...... 96 4.2.1 TG2+/+ or TG2-/- MEF growth curves ...... 96 4.2.2 Cell adhesion assay ...... 96 4.2.3 Cell spreading assay ...... 100

4.2.4 Activated RhoA and Rac1 levels are different in 129T2-TG2-/- and TG2+/+ MEFs 105 4.2.5 129T2-TG2-/- MEF monolayers re-establish slower from a scratch than 129T2-TG2+/+ MEF monolayers ...... 106 4.3 Discussion ...... 137 5. STRAIN DEPENDENT EFFECTS ON WOUND HEALING ...... 147 5.1 Introduction ...... 147 5.1.1 Knockout studies: a null mutant or background strain effect? ...... 147 5.2 Results ...... 149 5.2.1 Rate of early open wound contraction in B6.Cg-TG2-/- mice is equal to that of B6.Cg-TG2+/+ and 129T2-TG2+/+, but slower than that of 129T2- TG2-/- 149 5.2.2 mRNA expression of TG2 but not of other TG family members, is greater in 129T2-TG2+/+ than in 129T2-TG2-/-, B6.Cg-TG2+/+ and B6.Cg-TG2-/- 150 5.2.3 MEF cell adhesion on Fn is equivalent between 129T2-TG2-/-, B6.Cg- TG2+/+ and B6.Cg-TG2-/- ...... 151 5.2.4 MEF cell spreading on Fn was equivalent between 129T2-TG2-/-, B6.Cg- TG2+/+ and B6.Cg-TG2-/- ...... 151 5.2.5 B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEF monolayers close a scratch wound at equal rates ...... 152 5.2.6 Quantitation of TG2 expression in 129T2-TG2 and B6.Cg-TG2 MEFs152 5.3 Discussion ...... 168 6. CONCLUSION & FUTURE DIRECTIONS...... 172 APPENDIX ...... 177 REFERENCES ...... 199

LIST OF TABLES

Table 1 The transglutaminase family ...... 2 Table 2 Major extracellular matrix and related proteins secreted by fibroblasts ...... 84 Table 3 Some common constituents in adhesion complexes ...... 88 Table 4 Cell adhesion assay summary ...... 130 Table 5 Cell spreading assay summary ...... 130

vii

LIST OF FIGURES

Figure 1 Ribbon representation of TG2 structure ...... 5 Figure 2 Mechanistic scheme of TG-catalysed transamidation reaction ...... 9 Figure 3 Structure of skin ...... 20 Figure 4 Simplified overview of wound healing...... 23 Figure 5 Maintenance of mouse lines for breeding and experimentation ...... 52 Figure 6 TG2 genotyping by PCR schematic ...... 53 Figure 7 TG2 elution peaks from FPLC purification ...... 63 Figure 8 Characterisation of recombinant TG2 protein ...... 65 Figure 9 In vivo wound healing assay on mouse skin ...... 67 Figure 10 In vivo wound healing assay with exogenous TG2 addition ...... 69 Figure 11 Neutrophils and monocytes at the wound edge ...... 71 Figure 12 Analysis of integrity of the RNA from wound samples ...... 73 Figure 13 Microarray comparison of 129T2-TG2+/+ and 129T2-TG2-/- wounds ...... 76 Figure 14 Quantitation of expression of wound related genes in wounded and non- wounded skin from 129T2-TG2+/+ and 129T2-TG2-/- mice ...... 78 Figure 15 129T2-TG2 MEF growth curves ...... 109 Figure 16 Adhesion of 129T2-TG2+/+ and 129T2-TG2-/- on Fn matrix ...... 111 Figure 17 Optimisation of TG2 binding to Fn-coated plates ...... 113 Figure 18 The effect of exogenously added TG2 on adhesion of 129T2-TG2+/+ or 129T2- TG2-/- MEFs to Fn ...... 115 Figure 19 Cell adhesion assay with the transamidation-deficient mutant W241A TG2 or GTP-binding-deficient mutant deficient R579A TG2 ...... 117 Figure 20 Cell adhesion assay with exogenous addition of various TG2 C-terminal truncations of TG2 to the Fn matrix ...... 119 Figure 21 Cell adhesion assay in the presence of RGD and/or heparin (hep) ...... 121 Figure 22 Measurement of the area of phalloidin-stained 129T2-TG2+/+ or 129T2-TG2-/- MEFs adhered on fibronectin ...... 123 Figure 23 Cell spreading of 129T2-TG2+/+ and 129T2-TG2-/- MEFs preincubated with wildtype, GTP-binding-deficient R579A or transamidase-deficient W241A TG2 ... 125 Figure 24 Cell spreading assay with 129T2-TG2-/- MEFs transfected with wildtype, GTP-binding-deficient or transamidase-deficient mutant TG2 cDNAs ...... 127

viii

Figure 25 Transfection of 129T2-TG2-/- MEFs with cDNA-encoding Y274A TG2 ..... 129 Figure 26 Activated Rac1 and RhoA levels in 129T2-TG2+/+, 129T2-TG2-/- MEFs preincubated with TG2, and TG2-transfected 129T2-TG2-/- MEFs ...... 132 Figure 27 In vitro scratch wounds of 129T2- TG2+/+ or TG2-/- MEF monolayer ...... 134 Figure 28 Assay of in vivo wound healing in B6.Cg-TG2 mouse skin ...... 155 Figure 29 Transglutaminase expression in B6.Cg-TG2 and 129T2-TG2 unwounded and wounded skin samples ...... 157 Figure 30 Adhesion of B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEFs on Fn in the presence of exogenous TG2 ...... 161 Figure 32 Cell area measurement of B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEFs on Fn .... 163 Figure 33 In vitro scratch wound assay of B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEF monolayer ...... 165 Figure 34 Detection of TG2 in the membrane and cytosol of 129T2-TG2 and B6.Cg-TG2 MEF ...... 167

ACKNOWLEDGMENT

I am grateful to my supervisors Dr. Siiri Iismaa and Prof. Robert Graham for their concern, patience, guidance, and constructive criticism during the course of this study. I am privileged to have worked under these great minds, without whom this PhD would be nowhere close to fruition. I sincerely thank past and present members of the ‘RMG team’ – Ms. Sara Holman, Dr. Jian Xin Wu, Ms. Kristy Jackson, Dr. Chu Kong Liew, Dr. Marion Mohl, Dr. Mark Aplin, Dr. Ming Li, and Dr. Nicola Smith, who has helped me during my PhD with technical advice, as well as invaluable moral support during times when I was stuck in the rut.

Dr. Gavin Chapman is acknowledged for his advice in confocal microscopy techniques and microscopic imaging, as is Dr. Chris Brownlee and Dr. David Humphreys for his help with flow cytometry, and Dr. Mark Cowley for his expert advice on microarray data analysis. My gratitude is also extend to Mr. Brendan Lee for his infinite I.T. wisdom and patience for fixing software glitches, as every program used to produce this thesis are known to malfunction when I am not looking.

Special thanks are given to Stanley-Troy Artap, a PhD comrade, who has ridden the highs and lows of this journey along side with me. Thanks foddley! My love goes to my family who has always been positive and encouraged me from the very first minute to the last.

Thanks you all, I had fun!

ABSTRACT

Transglutaminase 2 (TG2) is a multi-functional protein that catalyses post-translational protein modification by transamidation, and it functions as a G-protein to α1B- and α1D- adrenergic receptors via GTP-binding and hydrolysis, as well as being an adaptor-like molecule bridging between the extracellular environment and cell surface adhesion receptors. The aim of this work was to investigate the role of TG2 in murine cutaneous wound healing and the relevant cell motility processes of cell adhesion, spreading and migration.

Full-thickness, 5mm-diameter, skin wounds were made on congenic 129T2/SvEmsJ wildtype (129T2-TG2+/+) or TG2 knockout (129T2-TG2-/-) mice by punch biopsy. Wound areas were monitored until closure. 129T2-TG2-/- mice had a slower rate of wound closure during the initial wound-healing phase relative to their wildtype counterpart. This delayed closure rate was rescued with a single application, immediately post-wounding, of purified TG2 to the wounds.

Cell adhesion, spreading, and migration studies were performed in-vitro utilising primary murine embryonic fibroblasts (MEFs), a cell type relevant in the early stages of wound healing. Relative to wildtype, MEFs from 129T2-TG2-/- mice were impaired in cell adhesion, spreading, and migration, and, in concurrence with the in vivo results, TG2 addition rescued these processes in 129T2-TG2-/- and enhanced them in wildtype fibroblasts. TG2 addition also overcame inhibition of integrin cell adhesion receptors mediated by their competitive inhibitor, RGD. Blocking both integrin- and syndecan- mediated adhesion drastically reduced fibroblast adhesion. This was not rescuable in either genotype by TG2 addition and, moreover, 129T2-TG2-/- MEF adhesion remained reduced relative to TG2+/+ MEFs, even with TG2 addition. This suggests an inherent TG2 effect that is independent of TG2 interaction with integrins or syndecans.

Interestingly, congenic C57Bl/6 (B6.Cg-) TG2+/+ and TG2-/- wound closure as well as fibroblast properties were similar to those of 129T2-TG2-/-, and significantly impaired relative to 129T2-TG2+/+. This correlated with lower TG2 expression levels in B6.Cg- TG2+/+ than in 129T2-TG2+/+ mice.

The data presented here demonstrate a role for TG2 in the physiological process of cutaneous wound healing suggest a potential use for TG2 as a therapeutic agent in the treatment of wounds through its effects on fibroblast dynamics.

xii

1. INTRODUCTION

1.1 Prelude

In 1957, N.K. Sarkar, D. D. Clarke, and H. Waelsch published a paper titled: “An enzymically catalyzed incorporation of amines into proteins” [1]. This paper described the discovery of an enzyme from brain and liver extracts that was able to covalently link labeled amines to proteins in the presence of Ca2+ – this enzyme was later designated transglutaminase 2 (TG2; also called tissue transglutaminase; EC 2.3.2.13). TG2 is so named because it catalyses the transamidation reaction: transferral of an amine group to a peptide bound, but not free, recipient glutamine residue [2]. TG2 is a member of the mammalian transglutaminase family, which consists of nine structurally homologous members (Table 1.1), encoded by a group of related genes [3, 4]. This family of isoenzymes, except for the single catalytically-inactive member of the family, protein 4.2, contain a catalytic triad of Cys-His-Asp, that confers transamidation activity (although this activity is only putative in TG6 and 7) [2]. It should be noted that TG activity also exists in invertebrates and in microbes, however the microbial enzymes are structurally unrelated [5]. Perhaps the most intriguing TG is TG2 due to its non-transamidation functions, unique amongst the family. At least six other activities have been attributed to TG2 since its discovery. This intrigues many researchers in the field as its many biochemical facets implicate it in various patho- and/or physiological roles.

1.2 TG2 structure and biochemical function

1.2.1 TG2 protein, expression and its regulation

The human TG2 gene (designated Tgm2) is located on chromosomal 20q11-12. The TG2 gene is comprised of 13 exons and 12 interspersed introns spanning 36.8kb, gives rise to a 3.5kb mRNA product. Tgm2 has several response elements upstream from the initiation site, including those for retinoic acid response element (5’ location: 1.7kb) [6], transforming growth factor β (0.8kb; TGF-β) [7], interleukin 6 (4kb) [8] and two AP-2 like elements (0.6kb and 0.2kb). Human TG2 is monomeric and has a molecular weight of 77.3kD comprising 687 amino acids [9], while the mouse (77.0kD) and rat (76.9kD) homologues have 686 amino acids. The half life of TG2 in culture medium (with 10% fetal calf serum) is about 11hr [10]. TG2 is composed of four domains, which are evolutionarily 1 conserved among all members of the family. The native TG1 and F13a1 proteins are zymogens with a pro-sequence at the N-terminus. The domains in the TG2 protein, from the N-terminus, are: β-sandwich, core, β-barrel 1, and β-barrel 2, and in human TG2 these domains correspond to residues 1-139, 140-454, 479-585, and 586-687 respectively [5]. Beta-sheet structures are predominant in the first, third and fourth domains of TG2, while the core domain has an alpha helical secondary structure [11, 12]. The primary sequence of TG2 was first reported from cDNA cloning of the guinea pig variant [13]; sequences from other species (human, mouse, rat, cow, chimpanzee, zebra fish, rhesus macaque, frogs) have since been deposited into online resources such as NCBI Entrez Gene. TG2 is constitutively expressed throughout the body in cell types such as endothelial cells, smooth muscle cells, and fibroblasts [11]. On a cellular level, TG2 is localised mainly in the cytosol, but is also associated with the plasma and nuclear membrane but not the mitochondria [14-18]. TG2 is also secreted to the extracellular space where it becomes deposited in the extracellular matrix (ECM) [19]. The mechanism for secretion is unclear, however recent data suggest that TG2 is carried in the recycling endosome, which later merges complexed with early integrins in endosomes [20]. Another group showed that at least in renal cell lines, TG2 secretion appeared to depend on a sequence WTATWDQQDCT found within the β-sandwich domain, which is important for fibronectin binding (Tim Johnson [2010], personal communication).

Table 1 The transglutaminase family

Protein Human Gene Expression Function Disease relevance (locus) Factor F13a1 Various blood Blood clotting [21, 22]; Fibrinolysis [24] XIIIa (6p24-25) components wound healing [23] TG1 TGM1 Keratinocyte Maturation of cell Lamellar Ichthyosis [26] (14q11.2) envelop during squamous metaplasia [25] TG2 TGM2 Fibroblast; Various functions (details Implicated in various (20q11-12) endothelial in Section 1.3) pathological setting (details cell; in Section 1.3) Smooth muscle cell; epithelial cells TG3 TGM3 Keratinocyte Terminal epidermis Oncogenic squamous cell (20q11-12) differentiation [27] carcinoma [28]; Darier’s disease; Lamellar Ichthyosis; Ichthyosis Vulgaris; Psoriasis [29] 2

TG4 TGM4 Dorsal Copulatory plugs in Implication in prostate cancer (3q21-22) prostate rodents [30] [31] TG5 TGM5 Keratinocyte Epidermis differentiation Darier’s disease; Lamellar (15q15.2) such as hair follicles [27] Ichthyosis; Ichthyosis Vulgaris; Psoriasis [29] TG6 TGM6 Undetermined Undetermined Gluten ataxia [32] (20q11-15) TG7 TGM7 Testis; Lung Undetermined Undetermined (15q15.2) Erythocyte EPB42 Red blood cells Cytoskeletal structural Moderate hemolytic anemia membrane (15q15) component of red blood [34]; hereditary protein cells [33] spherocytosis [35] band 4.2

1.2.2 Early work on TG2 three dimensional structure

Clues to the three dimensional structure of TG2 initially came from experiments that showed increased susceptibility to protease degradation when TG2 was incubated with Ca2+ but less susceptibility when incubated with GTP, hinting at conformational changes induced by these effectors [36]. Other studies using antibodies [37], fluorescence quenchers [38] and amino acid reactivity tests [39] agreed with there being two conformational states. Investigation using Fourier transform – infrared spectroscopy showed that movement from one state to another did not involve any secondary structure changes [40]. Next, attempts to elucidate TG2 structure utilized computer modeling based on the existing factor XIIIa (F13a1) crystal structures [41] with the rationale that the two TGs have high homologies in their core primary sequences (~85%) [40]. Casadio et al. [9] employed a combination of limited proteolysis, small angle scattering, various computer-modeling techniques, and consideration of known F13a1 structures to shed light on some structure-function features of TG2. TG2 was modeled as a wide prolate ellipsoid with a flat base that has a protease- sensitive hinge loop between the core and β-barrel 1 domains. This allowed dynamic folding of the N-terminal domains over the C-terminal ones (Figure 1), a feature that allows TG2 to catalyse transamidation as well as bind guanine nucleotides [5]. This concurred with the previous notion that movement occurs at the tertiary rather than secondary structure level [40]. The folding over creates a cleft in which the catalytic site is buried agreeing with simulations of the opening and closing of the molecule in the presence of Ca2+ and guanine nucleotide, respectively [9, 42]. The approximate shape of GTP-bound

TG2 was also determined to be a prolate ellipsoid from sedimentation velocity experiments [43].

1.2.3 Crystal structure of TG2

Crystal structures of human TG2 in both the open [44] and closed [12] conformation have now been reported. In the closed form, TG2 binds tightly with guanine nucleotide (kd = 1.6 μM [14]) in a hydrophobic cleft between the core and β-barrel 1 domains at 1:1 molar ratio [45]. In the human TG2 structure, the guanine nucleotide is contacted by critical residues Lys173 and Phe174 from the core as well as R476, R478, V479, M483, R580, and Y583 from the β-strands in the β-barrel 1 domain, with Arg580 forming ionic interactions with the α- and β-phosphates of the guanine nucleotide. Guanine nucleotide binding is further stabilised by anti-parallel β-strand interaction of Ile-416–Ser-419 with Leu-577–Glu-579 [12]. Ca2+ ions are putatively believed to bind these anti-parallel strands to disrupt guanine nucleotide binding. In the guanine nucleotide-bound form, access to the catalytic site is blocked by four β-strands from β-barrel 1 [12, 14, 46]. The open crystal structure (stabilised by the binding of a substrate-like peptide inhibitor) confirmed previous predictions with respect to protein dynamics and transition-state stabilisation [47] and showed that the C- terminal barrels move 120 Å relative to the closed structure in which the β-sandwich and core domains were superimposed. Such movement exposes a region for substrate entry to the catalytic site [48].

Figure 1 Ribbon representation of TG2 structure

This figure is from the review by Iismaa et al (2009) [11], which was adapted from Pinkas et al (2007) [44]. On the left, TG2 is in its closed conformation with GDP bound and on the right TG2 is in its opened conformation kept open by the inhibitor Ac-P(DON)LPF-NH2. The N and C termini are indicated and the four domains are colour coded: β-sandwich (blue); core (green); β-barrel 1 (yellow); and β-barrel 2 (red).

1.2.4 The two sides of the same coin – TGase and Gh

Being a member of the TG family, TG2 is best known for its ability to catalyse the transamidation reaction otherwise known as its TGase activity. This involves the amide group of a glutamine (in a peptide or protein) undergoing a nucleophilic attack from a lone pair of electrons on the nitrogen of an amine group (e.g. lysine). This results in the formation of a Nε(γ-glutamyl)lysine isopeptide bond with ammonia released as a side- product. This ability to form a covalent bond between intra- or inter-molecular residues allows TG2 to cross-link and post-translationally modify proteins. In the absence of an amine substrate, a water or an alcohol molecule acts as the second substrate leading to deamidation or esterification reactions, respectively [2]. Reactivity requires a conserved catalytic triad of: cysteine (Cys) 277, histidine (His) 335 and aspartate (Asp) 358 [2] as well as a tryptophan (Trp) 241 within the core domain, which stabilizes the transition state [47]. A schematic of TG2-mediated catalysis of transamidation is shown in Figure 2. In the presence of Ca2+, the initial step of transamidation involves acylation of Cys277 [12] within the TG2 catalytic site with the release of ammonia. Such acylation requires Trp241 to stabilise the glutamine oxyanion intermediate. Subsequently, the nucleophile His335, consequent upon donation during NH3 formation, attacks the incoming amine substrate, which in turn attacks the acyl carbon of the acylenzyme intermediate resulting in isopeptide bond formation. Asp358 is required to stabilise His335 [12]. In the absence of Ca2+, Cys277 reactivity is reduced through disulphide bonding with Cys335 or through the Cys277 thiol forming a hydrogen bond with the phenol oxygen from Tyr516 (Figure 2) [12, 49]. Absence of Ca2+ favours a conformation that allows guanine nucleotide binding. Some biological cross-linking targets of TG2 includes: actin [50, 51], β-amyloid peptides [52], collagen III [53], and crystallin [54]. If the environment is oxidizing, the Cys-Cys bond formation between Cys370 and Cys377 is favoured and TG2 is held in the open conformation. Conversely, if the environment is reducing, the Cys-Cys is not favoured and TG2 assumes its closed position ready for another round of cross-linking [55].

The transamidase activity of several transglutaminase family members, TG2, 3, 4, and 5, is allosterically inhibited by the binding of guanine nucleotide [56-58]. The ability to hydrolyse GTP to GDP as a signaling molecule, however, is a defining characteristic of TG2 that makes it unique among the transglutaminase family [59] (TG3, TG4, and TG5 can

6 both bind and hydrolyse GTP in vitro but have not been shown to signal [29, 57, 60]). The binding of GTP, as aforementioned, involves several key residues in a pocket between the core domain and β-barrel 1 (Section 2) and regulation of GTP hydrolysis is thought to involve Trp332 [46]. A proposed mechanism for hydrolysis involves positively charged residues in the vicinity of the γ GTP phosphate such as Ser171, Arg476 or Arg478 that are capable of a nucleophilic attack when a water molecule is present. The deliberate placement of these and other positively charged residues also helps to sterically orientate the GTP phosphate groups and potentially stabilise any negatively charged hydrolytic intermediates as observed with the hydrolysis of GTP by other classes of G-proteins [12, 61]. TG2 is referred to as “high molecular weight G-protein” (Gh) in the signaling context and until the purified protein was sequenced [59], it was thought to distinct from TG2 [15, 16, 62]. As a

G-protein, Gh is quite different to either monomeric or heterotrimeric G proteins (Reviewed in [63] and [64]). The most obvious difference is a lack of the four common GTP-binding motifs from the Gh primary sequence: GxxxxGK(S/T), DxxG, NxxD, and (C/S)Axx [65].

The GTP-binding residues in Gh bind GTP through backbone hydrogen bonding with various parts of the GTP molecule. Correspondingly, the conserved GTP-bound tertiary structure (five α-helices and a six-stranded β-sheet) found in typical monomeric and heterotrimeric G-proteins is not found in Gh which, rather, has anti-parallel β-strands from the core and β-barrel domains (Section 1.2.3) [66].

2+ Transamidation and Gh activities are reciprocally regulated by the presence of Ca or GTP respectively (Figure 1.2). The ‘two sides of this coin’ contain separate sets of residues for their functions (catalytic triad & Trp241 for transamidase and Lys173 & barrel 1 Arg580 for GTPase) but are constricted by linked conformational changes, which decide one activity over the other. Spatially, in the GTP-bound state, the GTP-binding site is close to the catalytic site due to the folding of the barrel domains towards the core. Access to the catalytic site is opposite the GTP-binding site [12, 44]. In the GTP-bound form, TG2 is in a compact conformation (gyration radius of 2.96nm and migrates faster on a non-denaturing gel) and keeps the catalytic Cys277 occupied through hydrogen bonding to Tyr516, whilst binding of Ca2+ ions seems to ‘loosen up’ this structure (gyration radius 3.81nm) by binding to regions of β-barrel 1 (Section 1.2.3.), which counteracts some β-sheet structure that is required for the closed form [9, 12, 43]. What does this mean in the physiological setting? The current view of TG2 regulation is that in the normal healthy cell its 7

Figure 2 Mechanistic scheme of TG-catalysed transamidation reaction

The overall transamidation reaction involves the formation of an isopeptide bond, in the presence of Ca2+ cofactor, between a sterically available carboxyl group of a glutamine (Q) residue and an amine group that may be present on a lysine residue or chemical diamines or polyamines with the release of ammonia by- product (top). Transamidation occurs only when access to the catalytic triad Cys (C) 277, His (H) 335, and Asp (D) 358 is open as in the presence of Ca2+ binding. Mechanistically, the Q γ-carboxamide group reacts with C277 with the subsequent acylation of C277, which is stabilised by W241 through a hydrogen bond with the carboxamide oxygen. Consequently NH3 is released with –NH2 group from the substrate 1 carboxamide and a proton from H335. The now nucleophilic N: from H335 attacks the amine nitrogen on substrate 2, which then attacks the acyl carbon of substrate 1 forming an isopeptide bond between the two substrates and releases substrate 1 from the enzyme complex. Finally, the active site is regenerated ready for the next round of reaction. In the GTP-bound form, the active site access is shut as the C-terminal barrels folds back over the N-terminal β-sandwich and core domain; the active site is kept shut by the hydrogen bond (dotted line) between Y516 from the β-barrel (blue) and the catalytic Cys277 (see text for details)

9 transamidase activity is inactive because of the low resting intracellular Ca2+ concentration and high guanine nucleotide concentration, i.e. TG2 is a latent enzyme [67, 68]. However, under pathological conditions when the plasma membrane is damaged or leaky, the resulting influx of extracellular Ca2+ and/or impaired GTP production result in the activation of its transamidase activity [69]. Increased cross-linking activity in dying Swiss 3T3 and ECV 304 cells is thought to mark them for apoptosis [69]. Cross-linking of nuclear DNA-digesting enzymes in the very same cell lines was also detected and is thought to moderate DNA-fragment induced immuno-responses [10, 70]. Upon the release of TG2 to the extracellular environment whether co-exported with fibronectin [19, 71] or as a result of inflammation [72], it appears that TG2 cross-linking activity is transiently activated but this activity becomes attenuated shortly after due to the oxidising environment that keeps TG2 locked in the open formation unable to regenerate itself for another round of catalysis [55].

1.2.5 Other in vitro TG2 biochemical activities

A fascinating aspect of TG2 is the number of biochemical functions ascribed to it. In addition to transamidation and GTP-hydrolysis, TG2 has been reported to be a kinase catalyzing the phosphorylation of insulin-like growth-factor binding protein 3 (IGFBP-3), a protein reported to be active in membranes of breast cancer cells [73]. H1 and H3 histone subunits as well as the oncogenic p53 and retinoblastoma protein have also been shown to be targets of TG2 phosphorylation [74-76]. In these in vitro studies, the kinase activity is thought to have relevance in the regulation of cancer cell survival. It appears that Ca2+ can attenuate TG2 kinase activity, but whether this is due to induced conformational changes and/or increased cross-linking of its targets rendering them unavailable for phosphorylation is not known [73]. In addition to the apparent kinase activity, in vitro studies have also suggested that TG2 has disulphide isomerase activity, catalyzing the formation and breakages of disulphide bonds between cysteine residues. In one study, reduced RNase activity was restored following incubation with TG2, presumably through renaturation of the active RNase conformation [77]. In a later study, a TG assumed to be TG2 (despite TG2 being absent from mitochondria [18]) appeared to modulate ATP generation in mitochondrial extracts and to form prohibitin complexes [78]. However, the kinase and disulphide isomerase activity have yet to be confirmed in in vivo studies. Independent of its enzymatic activities, TG2 can also act as a scaffold/adaptor protein, thereby facilitating cell 10 surface binding to the extracellular matrix. In this regard, TG2 can bind to fibronectin [79, 80], integrins [81-83], and syndecan-4 [84, 85] (details in Section 3.1.3). TG2 has also been shown to interact with the extracellular domain of the orphan G-protein coupled receptor 56, a receptor that inhibits the metastatic activity of certain cancers [86].

1.3 Animal models

In an attempt to elucidate the in vivo functions of TG2, two groups independently and simultaneously developed Tgm2 gene knockout mouse models. The knockout mouse developed by the Graham laboratory was generated by the insertion of LoxP sites into introns 5 and 8, allowing excision of exons 6 to 8 after exposure to Cre recombinase (Figure 4) [87]. The other knockout model was developed by the Melino laboratory, by replacement of part of exon 5 and the whole of exon 6 with the neomycin resistance gene [88]. In both instances, gene targeting was performed in 129/SvJ embryonic stem cells. Positive clones were introduced into C57Bl/6 blastocysts and implanted into C57Bl/6 surrogate mothers. Resultant homozygous knockout progenies from both groups were shown to have no detectable TG2 protein expression. The newly developed animals from these groups were immediately characterised phenotypically. Graham’s group looked at thymocyte apoptosis, heart and lung fibroblast adhesion, and cardiac function, only seeing defects in fibroblast adhesion [87]; whilst Melino’s group also looked at cell death in thymocytes as well as that in embryonic fibroblasts and saw no difference between wildtype and knockout cells. In spite of the ubiquitous expression TG2 in wildtype animals, the knockout animals were born at the expected Medelian ratios and were physiologically normal. A shortened mRNA corresponding to the Cre-deleted Tgm2 mRNA was observed in mice on the C57/B6 background. This was expressed at a very low level and did not lead to protein translation [89].

1.4 Biological roles of TG2?

The diverse range of biochemical activities of TG2 would suggest that it is important in many biological pathways and with the generation of TG2 knockout animals, efforts to decipher the in vivo role(s) of TG2 have greatly increased. Using the knockout animals and primary cells isolates [87, 88], evidence has accumulated for the involvement of TG2 in a 11 variety of (patho) physiological processes, including: coeliac disease, ocular diseases, cancer, osteogenesis, cardiovascular disease, neurodegenerative disease, inflammation, and wound repair [11]. The following sections will highlight the evidence implicating TG2’s involvement in these processes.

1.4.1 Coeliac disease

Coeliac disease (also commonly known as coeliac sprue, gluten-sensitivity enteropathy, or gluten intolerance) is a small intestinal autoimmune disease caused by reaction to the prolamin plant protein gliadin. This protein has a high proline (prol-) and glutamine (- amin) content, and is found in wheat, barley and rye. About 1-2% of the general population is genetically predisposed to this condition [90]. Ninety-five percent of coeliac disease patients have one of two isoforms of the human leukocyte antigen (HLA): DQ2 or DQ8, a receptor that discriminates between self and non-self antigens. These two particular isoforms have increased binding affinity to gliadin antigens [91]. This condition is manifested when gliadin peptides, which are present in food and are resistant to proteolysis because the gut lacks prolyl endopeptidases, are modified by TG2-catalysed deamidation or transamidation in the small intestine [92]. Gliadin deamidation enhances gluten-specific DQ2- and DQ8- restricted CD4+ T-cell recognition of the antigen presented by HLA DQ2 or DQ8 molecules on the surface of antigen-presenting cells. This promotes clonal expansion of these cells, leading to a destructive immune response to the intestinal epithelia and flattening of the villi [93]. TG2 cross-links gluten glutamine residues to itself or to other TG2 proteins [94]. The cross-linked product generates hapten-carrier complexes with novel epitopes that trigger a T-cell activated B-cell response [94, 95]. The resulting intestinal lining damage causes symptoms such as diarrhea, fatigue and malabsorption- related conditions [90]. In most cases, patients develop autoimmune antibodies against TG2 and other self cross-reacting antibodies (e.g. viral VP7 or heat shock protein 60 [96]) [94]. In fact, a standard diagnosis for coeliac disease is serological analysis for anti-TG2 antibodies with a sensitivity of 99% and specificity of >90% [97].

1.4.2 Ocular diseases

TG2 is implicated to have a role in various eye conditions such as cataract [98, 99], allergic conjunctivitis [100] and hyperosmolarity-induced damage [101]. The ocular disease most frequently associated with TG2 is, perhaps, cataract and proliferative vitreoretinopathy [102]. Cataract is the clouding of the eye lens or capsule due to molecular changes in the crystalline medium of the lens leading to vision loss. Knockout mouse studies have shown that lens isolated from wildtype C57BL/6 animals were opaque by day 10 post-culturing in TGF-β2, whereas lens from TG2-/- mixed strain animals remained relatively clear [7]. It appears that transamidation is up-regulated via the TGF-β2 and Smad3 pathway as shown in human lens epithelial cells under oxidative stress (H2O2) [99]. Furthermore, lens crystallins are known substrates of TG2 transamidation activity [98]. In allergic conjunctivitis, a recombinant TG2 competitive inhibitor reduced ocular eosinophil infiltration and epithelial edema in guinea pig by preventing transamidation of phospholipase A2 [100]. Finally, recent work in hyperosmolarity-induced damage showed that TG2 plays a role in inducing corneal epithelial cell apoptosis through mitochondrial stress via the caspase 3/7 and 9 pathway [101]. Taken together, the unregulated cross- linking activity of TG2 in the eye seems to lead to ocular disease states.

1.4.3 Cancer

Cancer is a disease caused by genetic transformation of a cell or a population of cells. Some cancer biology hallmarks include metastases, apoptosis evasion, angiogensis, and insensitivity to inhibitory factors [103]. This results in cancerous growth at expense of surrounding healthy cells. The role of TG2 at different stages of tumour progression is still a contested topic [104]. Thus TG2 expression is decreased during primary (original) tumour formation [105-107], whereas metastasis and secondary tumour propagation seem to involve increased TG2 expression [106, 108]. N-Myc is an oncogenic transcriptional factor that promotes cancer progression [109]. In breast cancer and neuroblastoma cell lines, down-regulation of TG2 expression is mediated by N-Myc through Sp1 and histone deacetylases forming a transrepressor complex on the TGM2 promoter; tumour growth reduction can be induced with N-Myc siRNA expression [110]. In line with this, in vitro intratumour injection of TG2 to CT26 tumours reduced growth post-tumour transplantation [111]. On the other hand, some studies have correlated increased TG2 mRNA and protein 13 levels in different cancer cells lines that were apoptotic [112, 113]. It is thought that cross- linking activity is induced by Ca2+ influx from the loss of membrane homeostasis during apoptosis [69, 114]. Potential targets for cross-linking by TG2 include pro-apoptotic proteins such as pRb and Bax [114, 115]. Conversely, TG2 has been shown to be involved in cancer progression. In a pancreatic adenocarcinoma cell line, TG2 induced nuclear factor κ B (NFκB) activity, a cell growth and survival transcription factor [116]. The “seed and soil” hypothesis for the spread of metastatic tumours involves the dependency of cross talk between cancer cells with the host microenvironment [117]. In this respect, TG2 can modify the host’s ECM, with increased TG2 protein conferring resistance of collagen ECM to matrix metalloprotease digestion, reduced collagen ECM turnover, and prevention of in vitro blood vessel formation [111]. In gliomas and fibrosarcomas, proteolytic cleavage of cell surface TG2 reduced cell adhesion and migration on a fibronectin ECM, but the opposite effect was observed on collagen ECM [118]. In TG2-/- mice, either on a congenic C57Bl/6 or mixed background, subcutaneous injection of a melanoma cell line led to larger tumour size and reduced animal life span compared to the response in TG2+/+ counterparts [111, 119]. TG2 may also encourage malignancy by potentially rescuing cells from anoikis (a form of programmed cell death induced by detachment of anchorage- dependent cells from the surrounding ECM), through its ability to interact with ECM components as well as cell surface integrin receptors [82, 120]. These seemingly conflicting results illustrate the complexity of the factors and exact microenvironment (“seed and soil”), which will either allow or disallow propagation of primary or metastic tumours [117]. TG2 may be a potential target for an anti-tumour strategy. For example, TG2 inhibition via chemical means or antisense (e.g. RNAi, ribozyme) technologies can synergistically sensitise cancer cells to chemotherapeutic treatment by apoptosis or autophagy induction [106]. In neuroblastomas [121] and in Alzheimer’s disease patients’ brains [122], TG2 was shown to have two alternatively spliced isoforms – a short form lacking GTP-binding activity and the full length (long) form of TG2. Transamidation activity, resulting from repression of GTP- binding, induced neuroblastoma differentiation, as shown by cells expressing the GTP- binding deficient TG2 short form or the R580A human TG2 mutant [121]. Differentiation was thought to be under the control of the nuclear transcriptional factor N-Myc, which regulates TG2 expression.

1.4.4 Osteogenesis and osteoarthritis

In vertebrates, bone development, which begins in the prenatal stage continues to adulthood [123]. Osteogenesis is reliant on two similar processes: bone ossification and bone mineralisation [123]. Bone ossification describes actual bone formation from softer hyaline cartilage; in vertebrates this process is initiated by proliferating chondrocytes, which are differentiated mesenchymal cells that migrate to the location where future bone will form. These chondrocytes secrete TG2 and F13a1 as well as cartilage matrix to allow cartilage formation [124, 125]. Once cartilage is formed, proliferation ceases and chondrocytes undergo hypertrophy with an increased expression of Tgm2 and F13a1 [125, 126]. At this stage, these differentiated cells actively secrete collagen type X, which marks hypertrophic and mineralising cartilage [127]. At the same time, vascularisation of the newly-formed matrix is induced. Bone formation is nucleated from cartilage mineralisation and is carried out by osteoblasts, a cell type that is differentiated from perichondrial cells or carried to the locale by newly formed vessels. Calcium ions are required for this process. Osteoblasts secrete both TG2 and F13a1 [128, 129]. In addition, TG2, at least in vitro, catalyses the cross-linking of several bone matrix proteins such as sialoprotein, osteopontin, osteonectin, and fibrillin-1 involved in bone mineralisation [130]. It appears that TG2 secretion by both chondrocytes and osteoblasts is important for sufficient bone mineralisation: the inhibition of chondrocyte TG2 cross-linking leads to reduced osteoblast bone mineralisation [131], and the addition of TG2 and/or F13a1 increases chondrocyte hypertrophy and mineralisation [130]. Such hypertrophy is conducted through the FAK-p38-MAPK pathway as a result of TG2 interaction with α5β1 integrin – a process that apparently does not involve the transamidase activity of TG2 [132]—as well as through an interaction of

F13a1 with α1β1 integrin [133]. In addition TG2 externalisation is stimulated by F13a1 via its interaction with the α1 integrin isoform [134]. In the animal knock-out models of either TG2 or F13a1 skeletal development is normal suggesting that either protein can compensate for the loss of the other , and thus that they have overlapping functions with respect to bone development. On the other hand, excessive chondrial mineralisation can lead to the degenerative condition of osteoarthritis where the cartilage-cushion between bone-ends no longer functions properly. This results in increased joint abrasion leading to pain and subsequent mechanical restrictions [135]. TG2 is implicated in the pathogenesis of this disease. Osteoarthritic molecules such as interleukin β1, GROα, and all-trans retinoic

15 acid can induce matrix calcification, hypertrophy, and collagen type X secretion in chondrocytes from C57Bl/6 TG2+/+ but not TG2-/- mice [124]. In experimentally induced osteoarthritis in mixed strain animals, lack of TG2 led to less cartilage destruction and increased bony outgrowth that was associated with increased TGF-β1 expression [136]. In C57Bl/6 mice, TG2 can also covalently cross-link S100A11, the ligand of the receptor for advanced glycation end products, thereby accelerating osteoarthritis through p38-MAPK pathway signaling [137].

1.4.5 Cardiovascular conditions

TG2 has been implicated in atherosclerosis and cardiac hypertrophy [11]. Atherosclerosis is a condition where fatty deposits in the artery wall trigger chronic inflammation with involvement of macrophages and T-lymphocytes. This ultimately causes plaque formation, which narrows the vascular lumen and disrupts blood flow. TG2 is highly expressed in white blood cells found in the plaque [138]. Its expression also increases in the adjacent smooth muscle cells and matrix and increases as the disease progresses [139]. TG2 expression by macrophages was shown to limit plaque development in irradiated fat-fed low-density lipoprotein receptor knockout mice on a mixed strain background when TG2+/+ but not TG2-/- bone marrow cells were transplanted into these mice [138]. The lack of TG2 is thought to decrease monocyte migration to the plaque as a result of decreased α5β1 integrin association [82]. In addition, lack of TG2 may decrease matrix cross-links such as collagen, which increases plaque-induced inflammation due to descreased plaque stability [138, 140, 141]. Lack of TG2 is also correlated with decreased apoptotic cell clearance [142]. TGF-β is also expressed in lower amounts in TG2-/- macrophages, which affects the stability of collagen matrix formation and thus plaque stability [142]. As a result of osteogenic and chondrogenic differentiation of smooth muscle cells, atherosclerotic lesion may calcify, which then adversely affects vascular elasticity [143]. To this end, TG2-/- smooth muscle cells were protected from arterial chondro-osseous differentiation and calcification by inorganic phosphate or bone morphogenetic protein-2 [144]. In contrast, exogenous addition of TG2 to smooth muscle cells enhanced calcification resulting from low-density lipoprotein receptor related protein-5 signaling and activation of β-catenin signaling [145].

In response to volume or pressure induced workload (or overload), the heart adapts by enlarging as a result of hypertrophic growth of cardiomyocytes and by remodeling of the ventricles. Initially this hypertrophic response is compensatory as it normalizes wall stress, but in time the enlarged heart fails [146]. In a rat model hypertrophy due to either mode of overload was associated with increased steady state TG2 mRNA levels, which were positively correlated with disease severity [146]. Increased TG2 expression was also detected in hearts from patients with dilated cardiomyopathy, but apparent GTPase and cross linking activity were decreased [147]. The effects of over-expressing cardiac TG2 were reported in two separate TG2 transgenic mouse models [148, 149]. In the first, 3-4 month-old animals showed that an increased expression of TG2 correlated with increase in expression of hypertrophic gene markers, such as α-myosin heavy chain and α-skeletal actin [148]. In the second study, 7-10 month-old transgenic animals had significant increased mortality due to age-dependent left ventricular hypertrophy and cardiac decompensation. These phenotypes were caused by increased biosynthesis of the prostanoid thromboxane (the TPα G-protein receptor agonist), and led to downstream

ERK1/2 activation. Sustained activation of phospholipase A2 and the prostanoid biosynthetic enzyme, COX-2, may also have induced hypertrophy, both giving rise to increased prostanoid thromboxane production [149]. It is not known if these cardiac effects are due to the transamidase activitiy of TG2 or its activity as a signaling or adaptor molecule.

1.4.6 Neurodegenerative disorders

Huntington disease is an inherited progressive neurodegenerative disorder that affects muscle coordination and leads to cognitive decline and dementia. It is caused by autosomal dominant mutation of one of the two copies of the gene coding for the cytosolic protein huntingtin. Such mutations give rise to progressive expansion of the number of CAG (coding for glutamine) repeats within the gene, which results in a mutant protein that now has a tract of N-terminal glutamine tract residues and, thus, is susceptible to the formation of insoluble aggregates due to the fact the glutamines can interact non-covalently via a hydrogen bonding network or “polar zipper” [150]. The number of CAG repeats correlates with disease severity. The soluble protein aggregates cause neurotoxicity as well as causing neuronal cell apoptosis in the striatum and cerebral cortex. TG2 is able to cross-link and 17 covalently link huntingtin proteins; it is thought that these cross-links promote increased solubility of huntingtin (as opposed to aggregation formation which would remove these soluble toxic proteins). Recently, it has been suggested that TG2 act on a different level in Huntington disease. TG2 crosslinking was shown to colocalise with crosslinked actin- cofilin in vitro using cells isolated from the Huntington disease mouse model. This suggests that under stress or pathogenic conditions where huntingtin is aggregated and under high calcium ion amounts, TG2 crosslinking activity is up-regulated causing defective nuclear actin stress response leading to synaptic and dendritic dysfunctions [151]. Indeed, removal of TG2 in the Huntington mouse models R6/1 and R6/2 was able to prolong life span and improve motor functions [152, 153].

Aside from Huntington disease, TG2 has also been implicated in other neurodegenerative conditions such as Alzheimer’s disease [154], Parkinson’s disease [155], and amyotrophic lateral sclerosis [156] by virtue of its up-regulation in these diseases. A common observation in these diseases is the presence of insoluble protein aggregate deposits. In Alzheimer’s disease, TG2 is suspected to crosslink tau proteins and amyloid-beta peptides to form cytotoxic aggregates that damage neuronal membranes in brains of Alzheimer’s patients [157, 158]. Interestingly, in Parkinson’s diseases, TG2 seems to play the reverse role by preventing disease progression likely by crosslinking of α-synuclein, which prevents the assembly of structured α-synuclein oligomers [155, 159]. However, whether these proposed models translate to the in vivo setting remains to be seen. Soluble and toxic

1.4.7 Contribution of TG2 to wound healing?

Given the distribution of TG2 in skin tissue milieu and its ability to interact with different ECM components, TG2 has been implicated in different phases of wound healing. For example, TG2 is found consistently at developing dermal-epidermal junctions and TG2- catalysed cross-links have been positively correlated with stability of these junctions [160, 161]. In burns patients, TG2 is detected at the dermal and basement membrane boundary in the healing epidermis, and moreover TG2 expression and activity has been correlated with anchoring skin grafts of skin burn patient dermal-epidermal fibrils [162]. Moreover, in granulation tissue and provisional matrix resulting from punch biopsy wounds in rats, TG2 cross-links are abundantly detected [160]. Before delving further into how TG2 is 18 implicated in skin wound healing, one must first understand the biology of the skin and the processes involved during healing.

1.4.7.1 Structure and function of skin

Skin is the visible and semi-permeable organ that covers the body; it is the body’s first line of defense against external pathogens and damage as well as a barrier that minimises body water loss. The skin is the largest organ for most vertebrates, and comprises 15% of the total body mass in the average human [163]. The skin can broadly be divided into three layers: a keratinocyte-populated stratified epidermis, a collagen-rich connective tissue dermis, and an underlying loose connective tissue containing a variable amount of adipose tissue [164] (described in Figure 3). The epidermis is a dynamic structure with continuous outward migration of keratinocytes that are continuously sloughed off. These cells divide at the basement membrane and move towards the outer-most epidermal layer, the stratum corneum, which is composed of dead and flattened cornified keratinocytes that have undergone apoptosis during their migration [165]. This structure provides a protective barrier that is under constant regeneration. The epidermis is woven through invaginations to the dermis underneath. This elastic connective tissue layer is supplied with blood circulation as well as the lymphatic system for both the dermis and the epidermis. The main cells of this structure are fibrocytes, Langerhans cells, monocytes, eosinophils, histiocytes and lymphocytes [166, 167]. The placement of the two skin structures are adapted for overall function – the ‘dead’ epidermis is fit for external abrasion and is readily replaced by underlying mitotic keratinocytes, whilst delicate circulatory vessels, which nourish the skin and allow access to immuno-surveillance, are tucked away from immediate danger in the dermis [163]. In the event of injury, such as full-thickness wounds, wound healing mechanisms come into play with the aim of restoring tissue integrity.

Figure 3 Structure of skin

This transverse histological section of mouse back skin demonstrates the general structure of skin. Skin is grossly made of three layers: epidermis, dermis, and subcutis. The epidermis consists of a keratinised squamous epithelium. Visible in this section are three epidermal sub-structures: thin dark red stratum corneum (C), light pink stratum granulosum (G), and the pink nucleated stratum spinosum (S). C is a layer of flattened keratin-filled dead cells; G is a layer of basophilic granules containing cells which undergo keratinisation by apoptosing and S is a layer of polyhedral cells bound to adjacent cells with extensive cell contacts called desmosomes. Support and nourishment of the epidermis is sustained by the underlying dermis, a layer of dense fibro-elastic tissue giving skin elasticity. Hair follicles (H) and sebaceous glands (Sb) can be found within this layer. The subcutis beneath is used mainly for storing fats as shown by the predominant white acellular adipose tissue (A). Scale bar = 30μm.

1.4.8 Cutaneous wound healing overview

Wound healing is an important physiological response to injury that ensures the survival of an organism upon a breach of the tissue. Typical full thickness wounds (epidermis and dermis) heal efficiently within 14 days [164]. In broad terms, mammalian wound healing is the speedy gap filling at the wound site with non-functional scar tissue and with some restoration of tissue integrity and aesthetics [168]. Wound healing is dynamic and can roughly be divided into three orderly and overlapping phases: i) inflammation, ii) cell proliferation, and iii) tissue remodeling (Figure 4). Different phases entail distinct physiological processes with the overall goal to restore normal anatomical structure and function [164, 169].

1.4.8.1 Inflammation

Wound healing begins with inflammation almost immediately following injury. This is a physiological response against harmful stimuli, initiating the re-establishment of homeostasis [170]. Inflammation is hallmarked by dolor (pain), calor (heat), tumor (swelling), and rubor (redness) [171]. An injury causes the introduction of pathogenic material and break-down products from surrounding dermal parenchyma triggering the release of an array of signal molecules into the wound milieu [172]. These include: platelet- derived growth factor (PDGF; from platelets, macrophages, epidermal cells), TGF-β (from platelets, macrophages), fibroblast growth factor (FGF; from macrophages and endothelial cells), and epidermal growth factor (EGF; from macrophages, epidermal cells) [170]. An injury may also induce blood coagulation wherein the contact of plasma platelets with tissue collagen starts a cascade of biochemical reactions that results in the formation of a homeostatic blood clot [173] (the TG member F13a1 is involved in this cascade [21]), though bleeding is not a requirement for proper wound healing [174]. The combination of chemotactic signal molecules and changes to endothelial surface molecule expression (e.g. selectins upregulation [175]) promote neutrophils and monocyte infiltration to the wound vicinity [176, 177]. Neutrophils are short-lived cells (up to a few days) that clear pathogenic contaminants through phagocytosis with the release of reactive oxygen species [178]. They are attracted to the wound by the local lymphocytic release of tumour necrosis factor α (TNF-α), which also propagates further production of inflammatory cytokines [179]. Monocytes undergo metamorphosis into macrophages and become adhesion- 21 activated at the wound site. The resultant macrophages accumulate and also participate in pathogenic phagocytosis as well as the breakdown of tissue debris [180]. Neutrophils and macrophages release a battery of protein factors that sets the stage for the next phase of wound healing. In the presence of certain bacterial molecules with certain pathogen associated molecular patterns, pattern recognition receptors of the Toll-like receptor (TLR) family on macrophages are induced to release an important protein called interleukin 6 (IL- 6), which has both pro- and anti-apoptotic activities [181]. Its pro-inflammatory activity sets up further cascades that release chemokines and cytokines that inhibit microbial growth through SOS (“Son of sevenless”, first discovered in Drosophila melanoganster [182]) signaling, SOS is a guanine nucleotide exchange factor that acts on the Ras class of small GTPases regulating survival pathways such as that of ERK/MAPK [181]. On the other hand, it can block the action of several inflammatory cytokines such as TNF-α [183].

1.4.8.2 Cell proliferation

A few hours into injury, chemokines and growth factors (such as TGF-β, FGF, EGF) are continually secreted from local neutrophils and macrophages after much phagocytosis and the wound is prompted for repopulation and reconstruction of new tissue. During this phase, a characteristic change to the epidermal and dermal cells is their surface marker expression. Instead of having anchorages, known as hemidesmosomes, between cell surface integrin α6β4 and basement lamina, keratinocytes lose these structures allowing the neighbouring epidermal and dermal cells to dissociate and migrate laterally into the wound

Figure 4 Simplified overview of wound healing

Mammalian wound healing can be broadly separated into three phases: inflammation, proliferation, and remodeling. This schematic shows the rough timeline of the phases of wound healing with some major physiological processes that define each phase and the extent of the biological response is indicated by the curves (see text for details). Inflammation is marked with processes that try to re-establish homeostasis at the wound site accompanied by immune cells that infiltrate the wound in order to cleanse it of pathogen and debris. Inflammation triggers the release of cytokines and growth factors that mobilise and propagate endogenous keratinocytes and fibroblasts to close the wound with the formation of new blood vessels to nourish newly formed tissue. The production of the ECM is vital to this process, as it provides substratum for cell adhesion and migration peaking at about 5-7 days post-injury. As the wound closes, it is replaced with scar tissue and its function is not fully the same as pre-injury. At the same time, the ECM readjusts as cell migration, which is seen during the proliferation phase, is no longer required. The scar is gradually remodeled to restore some degree of tissue functionality and tensile strength.

23 space [184, 185]. Keratinocytes are also cued to proliferate to provide a critical mass for rebuilding of the wound site. To aid migration, keratinocytes also secrete digestive enzymes, such as collagenase and plasmin, to dissect through the eschar and fibrin clots and rebuild from still viable tissue [186, 187]. The tracks for epidermal migration are provided by the ubiquitous fibroblast. In the presence of PDGF and TGF-β, fibroblasts expand and concurrently migrate into the wound to lay down a provisional ECM from which new epithelium propagates [188, 189]. Such expansion of fibroblasts is known as fibroplasia and requires the expression of cell surface integrins for adhesion and growth on the ECM. This matrix is primarily composed of fibronectin, vitronectin, and collagen III [190, 191]. For efficient re-epithelialisation, keratinocytes increase expression of integrins that recognise separate components of the provisional matrix: α5β1/β3 for fibronectin, αvβ5 for vitronectin, and α2β1 for collagen [190]. At the same time, ECM peptides are released from the pre-existing ECM due to injury and phagocytic activity from inflammation [192]. It is thought that these peptides contain cryptic sites that are normally hidden in healthy tissue and become functionally active in the injury setting. An example of this is the Arg-Gly-Asp (RGD) peptide, a motif that exists within ECM proteins such as fibronectin, collagen and vitronectin [192]. It was shown in vitro that integrin αvβ3 (expressed by macrophages and endothelial cells) could recognise RGD-exposed denatured collagen but not native collagen. On the other hand, α2β1 (expressed by keratinocytes) can recognise native collagen, but less efficiently than denatured collagen [193, 194]. Temporarily, re- epithelisation occurs several hours after the injury, presumably to allow for surface receptor rearrangements and expression of actin filaments (cytoplasmic fibers through which cells transmit mechanical forces) [187]. Migration and proliferation continues until the exposed wound area is covered with a monolayer of keratinocytes and is likely to be stopped through mechanical signals such as cell-cell contact inhibition [195]. Newly stratified epidermal tissue is organized beneath the newly epithelialised surfaces and moves inwards as the wound heals [196]. At the same time, fibroblasts, using the fibronectin IIIA splice variant as a conduit, would have invaded most, if not all, of the wound [197]. Fibroblasts lay down a contractile granulation tissue that is pink-red, moist and granular in appearance. This tissue consists mainly of fibronectin, fibrin, collagen type III, and hyaluronic acid [191, 198]. The granulation tissue is a scaffold for tissue mass generation and is infiltrated primarily by macrophages and fibroblasts: the macrophages continue surveillance and produce chemokines that attract fibroblast migration to and proliferation within this site 24

[170]. Fibroblasts close the wound by forming tight junctions with each other on the underlying matrix and pulling on the wound edge through cytoplasmic α-smooth muscle actin contraction [199]. It is thought that this fibroblast population represents a differentiated population known as myofibroblasts [200]. Neo-vascularisation, stimulated by hypoxia of the wound, also occurs to supply the newly formed tissue with nutrients and oxygen for the anabolic processes of wound healing [170].

1.4.8.3 Remodeling

Besides using specific cell adhesion receptors to migrate through the tissue matrix, fibroblasts also secrete proteases such as matrix metalloproteinases (MMPs) and collagenase to help them migrate efficiently by landscaping the matrix environment [201]. The release of these proteolytic enzymes plays an important part in modifying the granulation tissue as the wound heals. Other cell types such as macrophages, epithelial cells and endothelial cells are also involved in this process, as the granulation tissue is gradually replaced by a matrix made of strong collagen type I [202-205]. Collagen is organized into thicker fibrils, aligning along tension lines, and the overall matrix is more structured than that of granulation tissue. These fibrils, comprised of inter- and intra-molecularly linked collagen peptides, provide the healing wound with a degree of tensile strength [206]. The thicker fibrils are associated with fibroblasts or myo-fibroblasts to provide a strong substratum enabling these cells to generate forces to close the wound [206, 207]. As the synthesis and degradation of collagen turnover reaches equilibrium, fibroblasts stop matrix production and release of MMPs, and undergo apoptosis by an unknown mechanism but one thought to be related to the relative expression of various members of the TGF-β family and mechanical cues resulting from injury [208, 209]. This leaves behind acellular scar tissue that is structurally weaker (about 80%) and functionally different to healthy tissue [170]. At the same time, the basal lamina begins to reassemble with keratinocytes reforming hemidesmosomes [210]. The remodeling phase can take up to one year post- injury.

1.4.8.4 Evidence of TG2 in skin wound healing

An important process early in wound healing is blood clot formation. This curbs blood loss and provides ECM for cell migration and for tissue rebuilding. F13a1, activated by thrombin, is well known for its role during the final phases of the blood coagulation through the transamidation of adjacent alpha and gamma light chains of fibrin monomers to stabilise fibrin clots [21, 211]. TG2 expression in HUVEC endothelial cells is induced by the presence of thrombin locally [212] and erythrocyte TG2 has been shown to generate fibrin cross-links suggesting the possible release of TG2 in vivo as a supporting player alongside F13a1 in the formation of a homeostatic plug. TG2 has also been shown to promote HUVEC cell adhesion and spreading on fibrinogen polymers, suggesting that it may play a role in angiogenesis [213].

During inflammation, vasodilation and increased permeability of blood vessels allow leukocyte infiltration into the wound site. Permeability of vessels is sustained by the activation of phospholipase A2 (PLA2), which in turn regulates vasodilators such as prostaglandins and leukotreines [214]. In vitro TG2 was shown to increase secretory phospholipase A2 activity by altering its conformation via intramolecular isopeptide formation and polyamination of the enzyme [215]. Moreover, in an allergic conjunctivitis model, blockade of TG2 and PLA2 using two recombinant peptides inhibited TG2-induced post-translational modification in PLA2 and decreased inflammation in vivo [100].

As described in previous sections, cell migration is a central process in all phases of wound healing (Section 1.4.8.1 to 1.4.8.3). Some evidence suggests that TG2 is relevant to cell migration in vivo interacting with several matrix proteins to form cell-surface bridging between cells and the ECM (details in Section 4.1.3). Indeed, during inflammation, cytokines such interferon-γ, tumor necrosis factor-α and interleukin-6, separately, were able to up-regulate cell surface TG2 expression [216-218]. In monocytes, increased TG2 expression is correlated with their differentiation into phagocytic competent macrophages [219, 220]. Conversely, anti-sense downregulation of TG2 decreased monocyte adhesion and migration suggesting a role for TG2 in the regulation of monocyte extravasation [81]. Furthermore, lymphocyte diapedesis across an in vitro endothelium was blocked by addition of anti-TG2 monoclonal antibodies [218].

An important modulator in wound healing is TGF-β1, produced by all leukocyte lineages, which controls differentiation, proliferation, and activation state of immune cells [221]. Inactivated TGF-β1 is rendered latent in complex with large latent TGF-β binding protein anchored to the ECM; the release of TGF-β from this complex is required for its biological activity [221, 222]. The cross-linking of the latent TGF-β binding protein to the ECM is thought to modulate stores of TGF-β1 in the ECM indirectly [223]. Mixed strain TG2-/- mice were deficient in TGF-β1 protein expression and this was linked to impaired phagocytosis and prolonged inflammation [142]. This model supports a role for TG2 in facilitation of inflammation. In a study of gouty synovitis, a gout inflammation disease, mixed background TG2-/- animals had difficulty clearing apoptotic neutrophils leading to further recruitment of neutrophils [224]. Cultured macrophages from these animals had their apoptotic activity promoted by the addition of exogenous TG2 through a TGF-β- dependent manner and this did not require transamidation activity, in contrast to previous TGF-β studies [142, 223].

Following on from inflammation is the formation of the provisional matrix and subsequent cell migration into the wound bed (Section 1.4.8.2). Fn is a crucial ECM component of the provisional matrix that influences the rate of provisional matrix maturation and consequent granulation tissue development [225, 226]. TG2 has high affinity, in nM range, towards Fn [79] and is abundantly expressed in skin [227]. TG2 can also transamidate other ECM proteins in vitro – collagen V [228], osteonectin [161], osteopontin [229], nidogen and laminin [230, 231] producing cross-links that are resistant to proteolysis and thereby implicating its involvement in matrix organization in vivo and potentially during wound healing. TG2 from the 3T3 fibroblast cell surface can stabilise Fn by forming non-reducing cross-linked polymers [10]. This process can be repeatedly induced in dermal fibroblasts by exposure to UV damage [232]. When an artificial mechanical wound was generated on a fibroblast monolayer, intracellular TG2 expression was increased [69]. If indeed secreted TG2 is relevant in wound healing, then interaction with Fn appears to protect TG2 from matrix metalloprotease, an enzyme that is required to ‘pave way’ for migrating cells [80, 105]. Endothelial cell migration also appeared to be influenced by TG2; in a rat model TG2 was detected to be upregulated at sites of early angiogenesis. Furthermore, application of

27 recombinant TG2 to these wounds through a skin flap chamber increased blood vessel length and density [160].

In summary, TG2 expression has been observed to be upregulated in vivo in rat and human wounds and TG2 expression and activity has been investigated in response to wounding in vitro in cell cultures using cell lines either over-expressing TG2 or down-regulated for TG2 expression. The work in this thesis directly examines the physiological relevance of TG2 in wound healing. x In vivo by comparing wound healing in TG2 knockout and wild-type mice (Chapters 3 & 5), and x In vitro by comparing adhesion, spreading and migration of primary fibroblasts from TG2 knockout and wild-type mice (Chapter 4).

2. MATERIALS AND METHODS

2.1 Recombinant rat TG2 protein production

2.1.1 Expression vector and host

Recombinant TG2 rat protein was expressed and isolated as previously described [46]. Briefly, full length rat TG2 2.1kb cDNA was amplified and cloned into EcoRI and NotI sites of pGEX-2T expression vector (GE Healthcare, Australia) encoding TG2 as a glutathione-S-transferase-TG2 (GST-TG2) fusion protein with a thrombin cleavage site between the two proteins. The GST protein tag on TG2 allows the fusion protein to interact with reduced glutathione immobilised on a column, aiding in the purification of TG2. Vector was amplified in Escherichia coli DH5α and expressed in E. coli DH5α M15 (pREP4).

2.1.2 TG2 expression and isolation

E. coli transformants were grown in 2x YT medium (16g/l bacto tryptone; 10g/l bacto yeast extract; 5g/l NaCl; adjust to pH7.0) at 37oC to an optical density (600nm) of 0.5-0.6. At this density, expression of the fusion protein, under control of the lac operon was induced with 50mM isopropyl-β-D-thio-galactoside overnight at 30oC. Subsequently, bacteria were collected by centrifugation at 7000g for 20mins at 4oC and pellets were snap-frozen in liquid nitrogen. The following steps were performed at 4oC. Pellets were lysed in a cell lysis buffer (50mM NaCl; 50mM Tris, pH 7.6; 5mM EDTA; 1% Triton-X) with freshly prepared inhibitor cocktail (50μM leupeptin; 7.5μM Pepstatin A; 0.15mM phenylmethanesulfonylfluoride [PMSF]; 5mM DTT) on a vertical mixer for 2hr, then vortexed briefly to dislodge large lumps. Samples were sonicated with a Branson 400 Digital Sonifier (Branson Ultrasonics, CT, USA) on ice at 40% amplitude output for 2mins twice with a 2min break in between. Cell debris was separated by centrifugation at 10,000g for 15mins. The supernatant was retained and subjected to repeated centrifugation until the lysate was clear.

2.1.3 TG2 purification by glutathione sepharose column affinity chromatography

Cleared lysate was diluted to 250ml with cold 5mM EDTA in 1x PBS (PE) with 0.15mM PMSF. This was loaded over a 0.8cm2 x 2cm glutathione sepharose 4B column (GE Healthcare) at 0.16ml/min controlled by an ISCO mechanical pump (Teledyne Isco, NE, USA). The column was washed once with five column volumes of PE with PMSF and then once with PE. TG2 was released by incubating the column with 50U of thrombin in 2ml of PE for 2hr at 37oC with constant mixing. TG2 was eluted and stored at 4oC until purification with fast protein liquid chromatography.

2.1.4 Fast protein liquid chromatography (FPLC)

Eluate collected from the previous step was centrifugated at 16,110g for 20 mins to separate large debris particles. TG2 protein was then loaded onto an anionic chromatography column along with any other contaminating protein by virtue of ionic interactions. Retained proteins were then eluted using a low to high salt gradient, so that protein with the weakest ionic interaction with the column would elute first at lower salt concentrations. The supernatant was loaded onto a Tricorn Mono Q 5/50 GL ion exchange chromatography column (GE Healthcare) and then eluted with a 50-1,000mM sodium ion gradient (50mM HEPES, pH 8.0; 1mM EDTA; 50-1,000mM NaCl; 10% glycerol; 2mM dithiothreitol; 0.15mM PMSF) using the ÄKTA purifier UPC 100 (GE Healthcare) as per manufacturer’s instruction. Fractions were collected at a rate of 1ml/min. Fractions containing TG2 (typically around 20-30% of salt gradient) were pooled and TG2 concentration was quantified against a standard bovine serum albumin (BSA; Thermo Scientific, IL, USA) curve stained with Coomassie plus reagent (Thermo Scientific) and chromatically read at 600nm using an ELISA plate reader. TG2 was stored in aliquots of at least 2mg/ml at –80oC until used. A random aliquot of TG2 was quality tested by Western blotting (Section 2.2.2) and transglutaminase activity (Section 2.2.3). Purified TG2 was split into 50μl aliquots and stored at –80oC until use.

2.1.5 Western blotting for TG2

2.1.5.1 Protein separation

Protein sample was mixed with fresh 6x Laemmli buffer (0.4M SDS; 0.8mM bromophenol blue; 47% glycerol; 0.4M Tris, pH 6.8; 0.6M dithiothreitol) and boiled at 100oC for 5 minutes. Denatured samples and, separately, molecular size markers (Bio-Rad, CA, USA) were subsequently loaded onto 8% reducing polyacrylamide gel (8% acrylamide/ bis solution, 37.5:1, 2.6% C [Biorad]; 0.37M Tris, pH 8.8, 0.01% SDS; 0.005% aluminum persulphate; 0.5M tetramethylethylenediamine) and separated by electrophoresis at 80V until the dye front had just run off the bottom of the gel.

2.1.5.2 Protein transfer and probing

Proteins in the gel were transferred to immobilon-P polyvinylidene difluoride membrane in transfer buffer (24mM Tris base; 0.2M glycine; 10% methanol) in a wet transfer apparatus at 20V, 4oC overnight or in a semi-wet transfer apparatus (Bio-Rad) at 100V, 4oC for 1hr. Membranes were stained briefly with Ponceau S stain to visually confirm transfer of proteins, washed once with distilled water and then blocked for non-specific sites with 5% skimmed milk in TBS buffer (0.05M Tris; 0.15M NaCl; pH 7.5) at room temperature (r.t) for 1.5hr or 4oC overnight. Blocked membranes were washed twice with submerging amounts of TBST (TBS with 0.1% Tween 20) and once with TBS, 10mins each. TG2 was probed with anti-mouse TG2 rat monoclonal antibody (a gift from Gail V.W. Johnson, University of Rochester, NY, USA) in TBS (1:16,000 dilution) at r.t for 1hr. Excess antibody solution was aspirated and membranes were washed twice with TBST and once with TBS 10mins each. Primary antibody was probed using horseradish peroxidase conjugated anti-rat immunoglobulin (Ig) goat antibody (GE Healthcare) in the dark at r.t for 1hr. Excess antibody was washed as per primary antibody. TG2 blots were detected using enhanced chemiluminescence solution (GE Healthcare) as per manufacturer’s instructions. Blots were exposed to chemiluminescent film (Thermo Scientific, IL, USA) and developed using an ALLPRO 100 Plus film processor (ALLPRO, NY, USA).

2.1.6 Transglutaminase assay by 3H-putrescine incorporation

The transamidation activity of purified TG2 can be tested through the detection of radioactivity emenating from the cross-linking of radioactively labelled putrescine to dimethylcasein (DMC). A 30μl “TG2/substrate mix” (400mM Tris-HCl, pH 7.4; 100mM

MgCl2; 200mM DDT; 4mM EDTA; 2mM EGTA; 50% glycerol; 4% DMC; 0.5pmol TG2) was prepared in two triplicate lots: a transamidation ‘positive’ reaction and a transamidation ‘negative’ reaction. Aliquots of 20μl of “Putrescine mix” (5nM of 3H- putrescine [GE Healthcare]; 1,995nM putrescine [GE Healthcare]) were added to all 6 tubes, mixed thoroughly, and kept on ice. One μl of 100mM CaCl2 was added to the ‘positive’ triplicates and 1μl of sterile water was added to the ‘negative’ triplicates, and all tubes were immediately incubated at 37oC for exactly 40mins. The reactions were quenched on ice and incorporated 3H-putrescine was precipitated by the addition of 400μl cold 15% trichloroacetic acid (TCA) following a 15mins incubation at r.t. After centrifugation at 120g for 5mins and the supernatant fraction was aspirated and the pellets were washed three times with 500μl cold 10% TCA with 1min centrifugation and aspiration of supernatant after each wash. Residual supernatant was re-aspirated, the pellet was dissolved in 100μl 0.1M NaOH, transferred to 10ml Ready Safe scintillation fluid (Beckman Coulter, CA, USA) and mixed thoroughly in a scintillation vial. Tritium radioactivity from the ‘positive’ reaction and the ‘negative’ reaction were counted in LS6500 liquid scintillation counter (Beckman Coulter). A 2μl aliquot of the ‘putrescine mix’ (in 10ml scintillation fluid) was also counted, in disintegrations per minute, to calculate the amount of radioactive putrescine in the assay and converting to pmol of putrescine. The amount of radioactivity incorporated into DMC via transamidation was calculated by subtracting the ‘negative’ triplicate average from the ‘positive’ average in dpm. TG2 activity at 2mM CaCl2 was calculated as pmol of putrescine incorporated/pmol of TG2/min.

2.2 Mouse breeding and housekeeping

TG2 knockout mice, designated Tgm2tm1.1Rmgr (TG2-/-), were generated on a 129S1/Sv-ImJ background as previously described [87]. Heterozygous TG2+/- offsprings were backcrossed to wildtype 129T2/SvEmsJ (129T2) or C57Bl/6 (B6.Cg) mice for twelve

32 generations to generate congenic heterozygous TG2+/- mice with ~99.95% 129T2-TG2+/- or B6.Cg-TG2+/- genomic homogeneity, respectively. These heterozygous animals were mated to generate TG2 wildtype (TG2+/+) or TG2-/- breeding pairs to generate TG2+/+ and TG2-/- mice for experimentation, whilst heterozygous offspring were used for maintaining the breeding (detailed in Figure 5). Mice were housed and monitored in a certified physical containment 2 facility (BioCore animal facility, Victor Chang Cardiac Research Institute, Australia) in compliance with local ethics standards. The animal facility enviroment was as follows: temperature of 19-23oC, humidity of 40-60% and 12 hours light/ 12 hours dark cycle. Mice were supplied with 0.2μm-filtered water, pH 2.5-3.0 and fed with Gordon Stockfeeders ad libitum. Cages were equipped with environmental enhancing apparatus and at least two animals were housed together.

2.2.1 Mouse genotyping

2.2.1.1 DNA extraction

Three-millimeter long tail tips or 5mm diameters ear-clippings were collected from weaned mice of unknown genotype. The tissue was dissolved in 500μl TNES buffer (10mM Tris, pH 7.5; 125mM NaCl; 10mM EDTA, pH 7.5; 0.5% SDS) with 50μl of 10mg/ml Proteinase K overnight at 55oC. Protein was precipitated with 150μl of 6M NaCl at 4oC for at least 1 hour. Non-dissolved debris was collected by centrifugation at 10,000g for 20min and the supernatant was retained. DNA was precipitated with 500μl of 90% ethanol, collected by centrifugation at 10,000g for 10min and then a wash with 500μl of 90% ethanol. The pellet was collected again, followed by a wash with 500μl of 70% ethanol and collected by centrifugation. Resultant DNA pellet was semi-dried and dissolved in 100μl of 1x Tris- EDTA buffer (100mM Tris-HCl, pH8; 1mM EDTA). DNA quality and quantity was checked using NanoDrop 2000 spectrophotometer (Thermo Scienfic, DE, USA). DNA 260 260 concentration of at least 100ng/μl, with Abs /280 between 1.7-1.9, and Abs /230 between 1.9-2.2 was deemed suitable for subsequent genotyping steps.

2.2.1.2 DNA amplification and genotype determination by size

Five hundred nanogram of sample DNA was used as template for polymerase chain reaction (PCR) amplification and each sample was subjected to two separate reactions: the first amplifies, with primers P2 and P3, a region between intron 7 and 8 of the Tgm2 gene in TG2+/+ giving a 100bp product, and the second, with primers P1 and P3, amplified a region of intron 4 and 8 in TG2-/- giving a product of 300bp (Figure 4). The primer (Sigma- Aldrich, MO, USA) sequences were:

P1: 5’-GGAGCACACAGGCCTTATGAGCTGAAG-3’ P2: 5’-CAGATAGGGATACAAGAAGCATTGAAG-3’ P3: 5’-GCCCCACAAAGGAGCAAGTGTTACTATGTC-3’

PCR was performed in a 50μl reaction mixture containing: 500ng genomic DNA, 40nM of each primers, 1x GoTaq Flexi buffer with loading dye (Promega, WI, USA), 1U GoTaq

DNA polyermase (Promega), 10mM MgCl2 (Promega), 1mM deoxynucleotides triphosphate (Promega) and sterile water. The PCR was performed using a C1000 Thermal Cycler (BioRad, Hercules, CA, USA) following the program below:

TG2+/+ TG2-/- 94oC for 2 minutes 94oC for 2 minutes 94oC for 30 seconds 94oC for 30 seconds 63oC for 30 seconds 30 cycles 57oC for 30 seconds 30 cycles 72oC for 30 seconds 72oC for 30 seconds 72oC for 2 mintues 72oC for 2 mintues

The products were added to SDS sample loading dye (0.16M Tris-HCl, pH6.8; 2.8M glycerol; 0.14M SDS; 0.02% v/v bromophenyl blue; 0.02% v/v β-mercaptoethanol) and loaded onto a 3% agarose gel in Tris-acetate-EDTA (40mM Tris base; 20mM acetic acid; 50mM EDTA) and separated by electrophoresis at 100V until the loading dye front had just run out of the gel. DNA was visualised under UV light in a gel documentation system (Vilber Lourmat, Cedex, France). A schematic of the target amplicon in genotyping and representative genotype results are shown in Figure 4.

2.3 In vivo wound healing assay

To avoid genetic drift, all animals used were less than two generations away from the backcrossed heterozygous parents used to maintain the line. TG2+/+ or TG2-/- mice of age 12±2 weeks were used for this wound healing assay. One day prior to assay, animal identities were blinded and their backs were gently shaved with an electrical razor. Mice were anaesthetised with 2.5% isofluorane/oxygen. Two circular dorsal full thickness excisional wounds were made 30mm from the neck and 10mm from either side of the spine line using a punch biopsy (Stiefel Laboratory, NC, USA). Due to the looseness of mouse skin, prior to wounding, sterile 0.9% saline was injected beneath the subcutaneous layer creating a bulge at the site of wounding. Mice were allowed to recover in dust-free bedded cages. In some animals, 50μl purified 10mg/ml TG2 in PBS was added to one wound and PBS control to the other at 30 minutes post-injury. Animals were maintained under 1% isofluorane until TG2 solution was absorbed into the wound bed (approx. 40 minutes). Wound closure was monitored daily by digital photography using a Canon IXUS 50 camera (Canon, Japan) at fixed perpendicular distance on a retort stand from day 0. Epidermal wound edges were also traced onto a sterile coverslip. The wound area (expressed as fraction of day 0 wound size) was measured using the software ImageJ. Wounds that had blood clots or signs of infection were excluded from analysis.

2.4 White blood cell counting from wound histology

2.4.1 Sample isolation and fixation

TG2+/+ and TG2-/- 129T2 mice were subjected to wound healing assay (Section 2.3). On each day post-injury, up to day 5, mice (n = 3) were culled by cervical dislocation and wounds were quickly isolated by cutting about 2mm around the wound’s outer edge, the liberated wounds were gently lifted using sterile fine forceps and connective tissue and/or fat were cut with sterile surgical scissors. Wounds were fixed in 4% paraformaldehyde (PFA) in PBS overnight at 4oC. Fixed wounds were washed 5mins in 70% ethanol and then submerged in 70% ethanol overnight at r.t. Samples were wax embedded by the Department of Pathology, St Vincent’s Hospital, Sydney, Australia. Skin samples were embedded perpendicular to the gel cast and processed in a Nutator (TCS Scientific Corp., PA, USA) using the following chemicals, at r.t, in order: 35

CHEMICAL INCUBATION TIME (hr) 75% ethanol 1 95% ethanol 1 95% ethanol 0.5 100% ethanol 1 100% ethanol 0.5 Toulene 1 Toulene 0.5 Infiltration paraffin 1 Infiltration paraffin 1 EM400 embedding medium 1

2.4.2 Sectioning

Wax blocks were trimmed to expose minimal surface area to microtome blade. Wax- embedded samples were sectioned transversely to produce 6μm thick sections using a Leica RM2255 rotary microtome (Leica Microsystems GmbH, Wetzlar, Germany). Section slivers were expanded on 42oC water and caught onto Superfrost Plus micro-slides (Microm Int. GmbH, Walldorf, Germany). Sections were allowed to dry at room temperature and stored at r.t.

2.4.3 Hematoxylin and eosin staining for granulocytes

Slides were submerged in the following fixatives and stains in order:

CHEMICAL INCUBATION TIME (min) Xylene 5 Xylene 5 100% ethanol 1 95% ethanol 1 70% ethanol 1 Distilled water 2 Hematoxylin (Sigma Aldrich) 1 Running distilled water 1 0.01% (v/v) HCl 1

Running distilled water 2 Eosin Y (Sigma Aldrich) 0.5 Running distilled water 2 70% ethanol 1 95% ethanol 1 100% ethanol 1 Xylene 5 Xylene 5

Slides were immediately mounted with 25mm x 50mm cover slips using DPX mountant (EMS, USA). Mountant was allowed to set overnight at r.t.

2.5 Wound analysis

2.5.1 Microscopic analysis of wound sections

Areas in the immediate vicinity of the wound’s edge were examined under a Leica DMRF fluorescent microscope (Leica Microsystems GmBH, Germany); sections were visualised at 20x magnification. For each sample up to day 5 (n = 3), ten random fields were viewed and monocytes and neutrophils were counted. Monocytes were identified as cells with a blue- stained bean-shaped nucleus and neutrophils were identified as cells with a blue-stained multi-lobed nucleus, with thin strands connecting each lobe.

2.5.2 Wound RNA extraction

All equipment and apparatus used in this section were wiped with Rnasezap (Ambion, USA). TG2+/+ and TG2-/- 129T2 mice were subjected to wound healing assay (Section 2.3). On day 2 post-injury, mice (n = 6) were culled by cervical dislocation and wounds were quickly isolated by cutting about 2mm outside of the wound’s outer edge, the liberated wounds were gently lifted using sterile fine forceps and connective tissue and/or fat were removed with sterile surgical scissors. Wounds were quickly placed into an eppendorf tube on ice with 700μl of QIAzol Lysis Reagent from miRNeasy RNA extraction kit (QIAGEN, Germany). The tissue was homogenised on ice by PRO 200 homogeniser (PRO Scientific, USA) until minced. RNA extraction was continued following the kit’s manufacturer instruction. Purified RNA pellets were dissolved in 50μl of RNase-free water at r.t or at 37

60oC for 10mins. RNA concentration was measured using a NanoDrop 2000 spectrometer and subsequently stored at –80oC in 200ng aliquots. An aliquot of RNA from each sample was measured for RNA integrity using Agilent Bioanalzyer 2100 (Agilent Technologies, USA). Samples were loaded onto a RNA 6000 Nano Chip (Agilent Technologies) according to manufacturer’s instructions. If the skin sample’s RNA integrity number was above 7.0, it was deemed good quality and used for subsequent microarray analysis.

2.5.3 Microarray analysis

Five hundred nanograms of total RNA samples from Section 2.5.2 were processed at the Ramaciotti Centre using Agilent QuickAmp Labeling Kit, One colour (Agilent; PN 5190- 0442) and 3’ Gene Expression Hybridisation kit (Agilent; PN 5188-5242). Labelled RNA was hybridised to a one-coloured GeneChip Mouse Exon 1.0 ST Array (Affymetrix, USA), which contains 1.2 million probes, each corresponding to one unique genomic element in the mouse genome. Array scanning and subsequent data extraction were performed by the Ramaciotti Centre (UNSW, Australia) and the Peter Wills Bioinformatics Centre (Garvan Institute for Medical Research, Australia). Microarray was analysed using the software GeneSpring GX 11 (Agilent).

2.5.4 Reverse transcription of wound total messenger RNA

Eleven microlitres of total RNA (Section 2.5.2) was added to 2.5μg of anchored oligo(dT)20 primer, making up a total volume of 12μl. The mixture was heated 65oC for 5mins to melt any secondary structure in the RNA or primer and immediately placed on ice. The mixture was added to 1mM of deoxynucleotide triphosphate (Invitrogen), 1x FS buffer (Invitrogen), 0.05U RNAsin (Promega), and 200U Superscript III reverse transcriptase, making up to a total volume of 20μl and incubated at: r.t for 5mins, 50oC for 60mins, and 75oC for 15mins. The resulting complementary DNA was diluted to 50μl. A PCR reaction was performed to check the newly synthesised cDNA using primers for the housekeeping gene hypoxanthine- guanine phosphoribosyltransferase (HPRT). The forward and reverse sequences for Hprt are 5’-AAGCTTGCTGGTGAAAAGGA-3’ and 5’-TGGCAACATCAACAGGACTC-3’ respectively. The PCR reaction mixture was the same as described in Section 2.2.5.1. except with HPRT primers and the melting temperature was 58oC. A positive control of 38 genomic DNA and two negative controls of total RNA and water were included. The expected amplicon size is 150bp.

2.5.5 Quantitative PCR analysis of wound RNA

Real-time quantitative PCR was performed on cDNA prepared from RNA isolated from TG2+/+ or TG2-/- mouse wounds (day 2; n = 3) or cDNA prepared from RNA isolated from healthy skins (n = 3) using Taqman Gene Expression Assay 384 well format (Applied Biosystems). Quantitative PCR was performed in triplicate to examine the expression levels of Tgm1 (assay ID: Mm00498375_m1*), Tgm2 (Mm00436987_m1*), Tgm3 (Mm00436999_m1*), F13a1 (Mm00472334_m1*), Tgm4 (Mm00626039_m1*), Tgm5 (Mm00551325_m1*), and Tgm6 (Mm00624922_m1*), Tgm7 (Mm03990491_m1*), Il6 (Mm00446190_m1*), Ifng (Mm00497611_m1*), Tnf (Mm00443258_m1*), Egf (Mm00438696_m1*), Fgf2 (Mm00433287_m1*), Tgfb2 (Mm00436955_m1*), Ticam1 (Mm00844508_s1*), Myd88 (Mm00440338_m1*), Nfkb1 (Mm00476361_m1*), Nfkbib (Mm00456849_m1*), Fndc4 (Mm00480765_m1*), Fndc3a (Mm01232694_m1*), Col1a1 (Mm00801666_g1*), Col1a2 (Mm00483888_m1*), Col3a1 (Mm01254476_m1*), Sdc4 (Mm00488527_m1*), with Hprt (Mm00446968_m1*) and Ppia (Mm02342429_g1*) as housekeeping controls. Each reaction was performed in 10μl reaction mix: 5μl LightCycler probe master mix (Applied Bioscience); 0.5μl Taqman probes; 2.5μl water; and 2μl of cDNA. A standard curve using TG2+/+ cDNA diluted to 1, 1:10, 1:100, and 1:1,000 was constructed and a cDNA negative reaction were included for each gene examined. The 384- well plate was sealed and the reaction was incubated in a LightCycler 480 (Roche) with the melting temperature at 60oC for all probe sets. Results were only accepted if the crossing point was between 20-30 cycles and if the standard deviation within one triplicate was less than 0.5.

2.6 Mouse embryonic fibroblast isolation

A homozygous (TG2+/+ or TG2-/-) anestrus female mouse, aged eight to twelve weeks, was caged overnight with male-tainted bedding to synchronise it to estrus. On the following day a male mouse, aged eight to sixteen weeks, of the same genotype was introduced to the cage. Animals no more than one generation removed from the heterozygous parents were 39 used in mating thereby avoid background genetic drift. If a copulatory plug was found on the third day, the female was weighed and separated. At 13.5 days post-coitus, the impregnated female was sacrificed by cervical dislocation (the weight at this stage is roughly 1.4x of the mother’s weight at day zero for a mother carrying six embryos). Using sterile surgical scissors, an incision was made in the abdominal cavity and the embryo- containing uterus was excised into a petri dish containing sterile PBS, pH 7.4 (0.13M NaCl;

2.6mM KCl; 10mM Na2HPO4; 1.7mM KH2PO4). Then in a class II biological safety cabinet (AES Environmental, SA, Australia), the placenta, uterus wall, Reichert’s membrane and visceral yolk sac were removed to expose individual embryos. The head was removed from each embryo and the remaining body was soaked for 16hr in 1x trypsin- EDTA solution (1ml/embryo; Sigma-Aldrich) at 4oC in a 15ml Falcon tube (BD Biosciences, CA, USA). Embryos were subsequently digested at 37oC for 15min and minced by gently pipetting up and down using a 25ml pipette followed by 10ml pipette until large chunks of tissue are broken up. The trypsin-cell solution was neutralised using 5ml of DMEM/10% FCS (Dulbecco’s modified Eagles’ medium high glucose [Invitrogen, CA, USA]; 10% v/v fetal calf serum [Invitrogen]; 400μM L-Glutamine [Invitrogen]; 0.2U/ml penicillin [Invitrogen]; 0.2mg/ml Streptomycin [Invitrogen]) at 37oC. Suspended cells were collected by centrifugation at 120g for 5min and resuspended in DMEM/10% FCS at 1ml/embryo. Cell suspension was plated onto 15cm tissue culture dish (1ml/plate) o containing 15ml of DMEM/10% FCS and incubated at 37 C under 5% CO2, this was marked as passage one. At confluency, the MEF culture was cleared of any large embryo debris by vacuum suction through a Pasteur pipette. Adherent cells were detached by incubating with 1x trypsin-EDTA solution at 37oC for 5mins. Detached cells were confirmed by examination under a Nikon eclipse TS100 microscope (Nikon Coporate, Japan) for floating luminescent spheroid cells. Trypsin was neutralised by addition of at least 1x volume of DMEM/10% FCS media. Cells were collected by centrifugation at 120g for 5mins, supernatant was aspirated and the cell pellet was stored in liquid N2 until use.

2.6.1 Freezing, thawing and maintaining MEF cultures

Confluent cells were detached using 1x trypsin-EDTA solution for 5 minutes at 37oC and then washed twice with PBS. Cells were collected by centrifugation at 120g for 5 minutes and resuspended in ice cold DMEM/40% FCS (0.5ml/plate). The cells were transferred to a 40 cryovial (NALGENE Labware, IL, USA) on ice, and 0.5ml of 20% DMSO (v/v) in DMEM was added drop-wise with constant mixing to minimise heating of the cells. The vials were frozen at a rate of -1oC/min in a freezing container (NALGENE) overnight in -80oC freezer. Subsequently, the vials were transferred to liquid nitrogen and stored until use. To thaw frozen cells, vials were removed from liquid nitrogen and frozen cells were rapidly thawed in a 37oC water bath. Cells were immediately transferred to 10ml of 37oC DMEM/10% FCS, collected by centrifugation at 120g, and washed twice with PBS. The cells were o grown in DMEM/10% FCS at 37 C under 5% CO2. At 70-80% confluency, MEF were detached as previously described (Section 2.6). The cell pellet was resuspended in 10ml and split into new culture dishes with a maximum dilution of 1:4. Cells were cultured up to passage five before being discarded.

2.6.2 MEF growth curves

129T2-TG2+/+ or 129T2-TG2-/- MEF from from passage 2 through 6 were grown separately and maintained on 15cm tissue culture dishes with DMEM as described in Section 2.6.1. On consecutive days post-seeding, cells from each passage were detached (Section 2.6) and total live cell count was determined by counting in a Coulter counter with a gating range of 7μm to 24μm diameter.

2.7 In vitro cell adhesion assay

2.7.1 Preparation of wildtype and point mutant TG2 proteins

Preparation of pure TG2 protein was described in Section 2.1. In addition, TG2 protein containing either point mutants W241A or R579A were expressed and purified from E. coli DH5α transformed with pGEX-2T plasmid cloned with TG2 sequences with these mutations [46].

2.7.2 Detection of Fn-bound TG2 in tissue culture wells

To determine the maximal TG2 saturation levels on Fn matrix, an ELISA-based assay was used in a 96-well template; each condition was repeated in duplicate. Bovine Fn in 50mM Tris-HCl (Invitrogen), pH7.4, was coated onto wells (10μg/cm2) of tissue culture grade 24- 41 well plates for about 16hr at 4oC, subsequently some wells were blocked for non-specific binding with 5% skim milk at 4oC for at least 1 hours. Wells were gently washed once with PBS for 5mins each. Differing amounts of purified rat TG2 was added to Fn-coated wells and allowed to incubate at either 4oC for overnight or 37oC for 30mins. Unbound TG2 was gently washed twice with PBS for 5mins. Primary rabbit polyclonal anti-mouse TG2 antibodies, AB4 (Thermo Scientific; purified with azide) were diluted 1:100 in PBST (1xPBS, 0.1% Tween 20) and incubated in treated wells at r.t for 2hr. Unbound antibodies were gently washed twice with PBST for 5mins each. Secondary horse radish peroxidase- conjugated goat anti-rabbit IgG antibodies (Promega) were diluted 1:500 and incubated with the primary antibodies at r.t for 2hr. Towards the end of this incubation, substrate for peroxidase, SigmaFast o-phenylenediamine dihydrochloride (OPD; Sigma-Aldrich), was prepared according to manufacturer’s instruction. The wells were washed gently twice with PBST for 5mins each, then 200μl of OPD was added to each well and incubated at r.t for

30mins in the dark. The reaction was quenched with 50μl of 3M H2SO4 and the developed colour was read at 492nm using an ELISA plate reader.

2.7.3 Cell adhesion determination using crystal violet staining

Cell adhesion of MEFs on matrix protein was determined using a colorimetric assay where adhered cells are stained with crystal violet stain, which is readily taken up by the cytoplasm. Therefore, the number of cells is proportional to the amount of stain taken up. Bovine Fn in 50mM Tris-HCl, pH7.4, was coated onto wells (10μg/cm2) of tissue culture grade 24-well plates for about 16hr at 4oC, washed once in cold 50mM Tris-HCl and then blocked with 3% BSA in PBS. Subsequently, in some wells purified recombinant rat TG2 (10, 20, 40, 60, or 80μg/cm2) or its N-terminal domains: β-sandwich (βS; 11.4 μg/cm2), core (C; 14.6 μg/cm2), β-sandwich/core (βC; 6.3 μg/cm2), or both βS and C (equal amounts) were added to Fn-coated wells and incubated for 30 minutes at 37oC. N-ternimal domains were added at concentrations that were equimolar to 20μg/cm2 of TG2, which is a concentration that did not saturate cell adhesion during the assay (Section 4.2.2.2). Random MEF batches were serum-starved and cell cycle synchronised in 0.1% FCS DMEM (DMEM media with 400μM L-Glutamine; 0.2U/ml penicillin; 0.2mg/ml Streptomycin) for o about 16 hours at 37 C under 5% CO2. These were detached using enzyme-free dissociation solution (Chemicon, USA); a 10μl sample of the detached cells was diluted in 30μl of 42

Trypan Blue (Invitrogen). Cells were visually confirmed to be alive from Trypan Blue staining. Live unstained cells were counted using a Neubauer hemocytometer. Detached MEF were seeded at 5x104 cells/cm2 into coated wells, this cell concentration was non- saturating for the adhesion assay (Section 4.2.2.1). In some wells, heparin sulphate (42μg/5x104 cells [85]), GRGDTP peptides, or GRADSP peptides (both at 84μg/5x104 cells [120]) were added for competitive inhibition of adhesion pathway(s). Seeded MEF were o allowed to incubate in 0.1% FCS DMEM at 37 C under 5% CO2. Adhered cells were washed once with cold PBS and fixed with 4% PFA in PBS for 15 minutes at r.t. Fixed MEFs were stained with 0.1% crystal violet stain in 25% methanol for 15 minutes at r.t. Stained cells were washed three times with PBS. Crystal violet was eluted with 200μl 70% ethanol and the number of adhered cells was determined from the absorbance at 600nm. To normalise for intra- and inter-plate errors, an internal standard curve was established in duplicate on each plate using 100, 75, 50, and 25% dilutions of total number of cells seeded. Cells were fixed to the substratum using 4% PFA, washed once with PBS, stained with 0.1% crystal violet, and absorbance at 600nm was determined. A negative control of wells with no MEFs was included to quantitate background staining.

2.8 In vitro cell spreading assay

2.8.1 Cloning of TG2 into a GFP-bicistronic vector

To add TG2 endogenously into TG2-/- MEFs, these cells were transfected with a bicistonic vector coding for both TG2 and GFP. TG2 is located upstream to, and transcribed with greater efficiency than, GFP. Therefore any GFP-positive MEF would have expressed transfected TG2 as well.

Plasmid pcDNA3.1 (10μg) previously subcloned with [14] wildtype rat TG2 or point- mutants S171E, W241A, Y274A, C277S, R579A was linearised with 2U of the restriction enzyme NotI (Roche) in the 10x buffer provided by the manufacturer for 2 hr at 37oC. A 1μl sample of the reaction mixture was added to SDS sample loading dye (Section 2.1.5.2), loaded onto a 1% agarose gel in Tris-acetate-EDTA, and separated by electrophoresis at 100V until the loading dye front had run out of the gel to check for linearisation of the plasmid (a band at 7.5 kbp). If linearisation was completed, the reacting mixture is stopped

43 by heating to 65oC. The resulting 3’ recess at the NotI digestion site of the linearised plasmid was filled in 5’ to 3’ by incubating with large fragment Klenow (2U/μl; Roche), 20mM dGTP, and 20mM dCTP for 60 min at 37oC and then stopped by heating for 10 min at 65oC. The 2.1 kbp coding sequence for TG2 or the point mutants was excised from the linearised pcDNA3.1 vector by incubation with 2U of EcoRI in the 10x buffer provided by the manufacturer for 2 hr at 37oC. To check for linearisation of the plasmid, a 1μl sample of the reaction mixture was added to SDS sample loading dye (Section 2.1.5.2), loaded onto a 1% agarose gel in Tris-acetate-EDTA, and separated by electrophoresis at 100V until the loading dye front had just run out of the gel (bands at 5.4kbp and 2.1 kbp corresponding pcDNA3.1 and TG2 or mutant sequence, respectively). If linearisation was completed, the reacting mixture is stopped by heating to 65oC for 5 mins, added to SDS sample loading dye and loaded onto a 1% agarose gel in Tris-acetate-EDTA, and separated by electrophoresis at 100V until the loading dye front had ran out of the gel. The 2.1kbp band was excised from the gel using sterile cutting blade and this DNA fragment was purified using Ultraclean Gelspin DNA purification kit according to manufacturer’s instruction. pIRES-EGFP vector (4μg; Clontech, USA) was linearised with 1U of the restriction enzyme SmaI (Roche) in the 10x buffer provided for at least 2 hrs at 25oC. Linearisation was checked as described previously. Linearised vector (5.2kbp) was purified from the SmaI reaction mix using Bresaspin DNA purification kit according manufacturer’s instructions. The linearised vector was then digested with 1U of EcoRI in 10x buffer provided by the manufacturer for 2 hr at 37oC. The digested vector was then purified using Bresaspin DNA purification kit.

The linearised vector, pIRES-EGFP, was ligated to wildtype TG2 or point mutation sequence at a molar ratio of 8 fmol to 3 fmol in the following reaction mix: T4 ligase/kinase 10x buffer (Roche), 1mM ATP, and 1U T4 DNA ligase (Roche), overnight at 4oC. Two separate control reactions using only the vector with or without the addition of the T4 DNA ligase were included. At the completion of ligation, each reaction mixtures were stopped by heating for 5 min at 65oC. Each mixture was transferred to separate 15ml polystyrene tubes on ice and to this 1/5 volume of Tris EDTA NaCl (100mM Tris, pH 7.5, 5mM EDTA, 150mM NaCl) was added. Two volumes of competent DH5α E. coli was added to this resulting mixture, settled for 30 min on ice and subsequently heat-shocked for 44

1 min at 42oC. The bacteria were allowed to recover in 1ml Luria broth incubated with shaking for 1 hr at 37oC and then plated onto Luria agar plates added with ampicillin (10g/l Tryptone, 5g/l Yeast extract, 10g/l NaCl, 15g/l agar, 100mg/l ampicillin) incubated overnight at 37oC.

Eight colonies for each transformant was picked and inoculated into separate 10ml LB media added with ampicillin (100μg/ml) incubated overnight at 37oC. Plasmid DNA was isolated from 5ml of these cultures using Promega Wizard SV miniprep kit according to manufacturer’s instructions. Recombinants were screened by restriction enzyme digestion using EcoRI for 1 hr at 37oC, which makes one cut in the ligated pIRES-TG2-EGFP plasmid. Digested plasmids (7.4 kbp) were added to DNA loading dye and loaded onto 1% agarose gel in Tris-acetate-EDTA and separated at 100V along with linearised pIRES- EGFP empty vector (5.3 kbp) and TG2 coding DNA fragment (2.1 kbp) as comparisons. Positive recombinants were used to make large-scale plasmid DNA by inoculating a positive single colony into 300ml of Luria broth with ampicillin (100mg/l) overnight at 37oC. 500μl of the overnight culture was added to 500μl of 80% sterile glycerol and frozen at –80oC to make frozen stocks and the remaining culture was lysed for pIRES-TG2-EGFP using the Nucleobond pDNA kit (Clontech) according to manufacturer’s instruction.

2.8.2 TG2 transfection into MEF

MEFs were grown in a 15cm dish until 50-60% confluency (doubling phase of growth curve). The cells were detached as described in Section 2.6.1. Cells were trypsin-neutralised using 10% FCS DMEM supplemented with 2mM GlutaMax (Invitrogen) in place of 400μM L-glutamine. Cells were counted using a Coulter counter and resuspended with media to a concentration of 2x106 cells/ml. Transfection was performed using the Amaxa nucleofection system (Lonza Group Ltd, Switzerland) in accordance with manufacturer’s instruction. Cells were transfected with wildtype TG2 or various point-mutants S171E, W241A, C277S, Y274A, R579A cloned into pIRES-EGFP vector (Section 2.8.1). Transfected MEFs were allowed to recover for 24 hours before further experimentation. Transfection was confirmed by visual inspection under a LSM 700 laser scanning microscope (Zeiss) at an excitation wavelength of 488nm. Transfection efficiency was determined on a proportion of transfected cells by flow analysis and gating of green 45 flouorescent protein (GFP) positive cells (488nm ex.) using FACSCanto II flow cytometer (BD Biosciences). Because the GFP sequence is expressed at reduced effiency with respect to the upstream cloned TG2 (or mutants) sequence(s) in the pIRES vector, GFP-positive cells were taken to be TG2-positive as well.

2.8.3 Detection of cytosolic and membrane TG2

Expression of TG2 protein was confirmed by lysing transfected MEFs on ice using 500μl RIPA buffer (50mM Tris-HCl, pH 7.4; 150mM NaCl; 0.1% sodium dodecyl sulphate; 0.5% sodium deoxycholate; 1% Tergitol-type NP-40 [Sigma Aldrich]) with fresh inhibitor cocktail (1mM phenylmethanesulfonylfluoride; 0.4mM Aprotinin; 0.4mM Leupeptin; 0.4mM Pepstatin) per 10cm dish. Cells were scraped with a rubber policeman and collected at the bottom of the plate. The lysed cells in RIPA were transferred to 1.5ml eppendorf tube on ice and passed through a 1ml insulin syringe at least 3 times. The lysate was sonicated with a Branson 400 Digital Sonifier on ice at 40% amplitude output for 2mins twice with a 2min break in between. Samples were centrifuged at 1000g for 20mins at 4oC, transferred to a new eppendorf tube and centrifuged at 16,100g for 1hr at 4oC. The supernatant was transfered into a new tube, the centrifugation step was repeated and the pellet was resuspended in 50μl of RIPA buffer. The supernatant was transferred to a new eppendorf tube and the pellet, if any, was combined with that from the previous centrifugation. The supernatant was the cytoplasmic fraction and the resuspended pellet was the membrane fraction. Both of these were subjected to Western blotting (Section 2.1.5) and TG2 was detected using anti-mouse TG2 rat monoclonal antibody (1:16,000). GAPDH and pan- cadherin served as loading controls for cytoplasmic and membrane fractions, respectively. GAPDH was detected using rabbit anti-mouse GAPDH polyclonal antibody (1:5,000; Abcam, USA) and pan-Cadherin was detected using rabbit anti-mouse pan-Cadherin polyclonal antibody (1:2,500; Abcam). Expression was measured by densitometric analysis using the software ImageJ. TG2 was normalised for loading using the relevant loading control blots.

2.8.4 Cell spreading assay

A sterile square coverslip (15x15mm) was placed in each well of a 6-well tissue culture plate and coated with bovine Fn in 50mM Tris-HCl pH7.4 (10.5μg/cm2) overnight at 4oC. Coverslips were washed with cold 50mM Tris-HCl and blocked with 3% BSA in PBS. Excess BSA was aspirated and washed once with PBS. Immediately, cells for the spreading assay were prepared as follows: random batch of TG2+/+ or TG2-/- MEF was grown to sub- confluence, serum-starved, and detached as described in Section 2.7.2. MEFs were diluted to 5,000 cells/ml and 1,000 cells (200μl) were carefully seeded onto a coverslip such that a meniscus was formed with spilling over. On some coverslips, MEFs were treated with 20μg of exogenously added wildtype or mutant TG2 (Section 2.1 & 2.7.1), which was the amount used in the cell adhesion assays, or had been previously transfected (Section 2.8.2). The cells were allowed to settle onto the coverslip for 30, 60, or 90mins in 0.1% FCS o DMEM at 37 C under 5% CO2. At a set time point, unattached cells were carefully removed by aspiration from one side of the meniscus and the coverslip was washed once with cold 1x PBS. Cells were fixed with 4% PFA at r.t for 15mins and washed once with PBS. Cells were permeabilised with 1% Triton-X in PBS at r.t for 30mins, washed once with PBS and stained with 7nM TRITC-conjugated phalloidin at r.t for 1hr. Excess stain was aspirated and coverslips were submerged in 1x PBS. Coverslips were mounted onto slides using Prolong Gold with DAPI mounting medium (Invitrogen). Medium was allowed to set overnight at 4oC in the dark. Samples were blinded and ten random fields of view were recorded for each MEF and cell area was measured for all non-overlapping cells within each field. Area pixels were calculated using the software ImageJ by tracing the cell perimeter using the polygonal shape tool. Areas were normalised to TG2+/+ cells at 30mins post-seeding.

2.9 Detection of cell surface TG2 by pull-down assay

Serum-starved MEFs of interest were washed three times using 1xPBS and immediately kept on ice. Adherent cells were labeled with 0.5mg/ml of sulfo-NHS-LC-biotin (Thermo Scientific) in 1xPBS for 15mins at 4oC on a rocking platform and the reaction was quenched and gently washed twice with 50mM Tris-HCl, pH8.0 buffer. Excess buffer was aspirated, and cells were scraped using a rubber policeman and passed through an insulin

47 syringe at least three times. Biotinylated cell lysate was added to 50μl of PBS pre-washed streptavidin-agarose beads (Thermo Scientific) and incubated overnight at 4oC on a vertical rotating platform. Bound agarose beads were collected by centrifugation at 5000g for 5mins and the flow-through supernatant was transferred to a new eppendorf tube. Pelleted beads were washed three times with 1xPBS and centrifugation at 5000g. Six-times Laemmli loading buffer (Section 2.1.5.1) was added to the beads and 50μl of the flow-through. Both were boiled at 100oC for 5mins, and Western blotting was performed as described in Section 2.1.5.2 using GAPDH as a loading control for the flow-through and pan-cadherin as a loading control for the membrane fraction. A separate sample without biotinylation treatment was used as a negative control for the pull-down assay.

2.9.1 Immunostaining of cell surface TG2

Serum-starved MEFs of interest (5x103 cells) were seeded onto Fn-coated (10μg/cm2) glass coverslips in each well of a 6-well plate and were allowed to settle overnight at 37oC under

5% CO2. All following steps were performed on ice unless otherwise stated. Media was aspirated and cold fresh DMEM/10% FCS was added to block the cells for 30mins. To detect cell surface TG2, rabbit anti-mouse TG2 polyclonal antibody (Ab4, Thermo Scientific) diluted 1:1000 in DMEM/10% FCS was incubated for 2.5hrs at 4oC. Cells were gently rinsed three times with cold DMEM/10% FCS and washed once for 10mins. Cells were fixed with 4% PFA in PBS++ (1xPBS with 1mM MgCl2 and 0.1mM CaCl2) for 15mins. PFA was aspirated, and cells were rinsed once with 150mM glycine in PBS++ was rinsed once to neutralise residual PFA, followed by a wash with 150mM glycine in PBS++ for 15mins at r.t, and then two washes with PBS++ at r.t. Fixed cells were then treated with 2% IgG-free BSA (Jackson Labs, USA) in PBS++/0.1% Triton-X for 30mins. This solution was replaced with goat anti-rabbit IgG conjugated with AlexaFluor 488 dye (Invitrogen) diluted 1:5000 with 2% BSA with in PBS++/0.1% Triton-X for 1hr at r.t in the dark. Subsequently, the samples were washed three times with PBS++ for 10mins each. The cells were stained using 7nM TRITC-conjugated phalloidin as well as DAPI (Invitrogen) in PBS++ (1:10,000) at r.t for 1hr in the dark and then washed once with PBS++. The coverlips were lifted from wells using fine forceps and excess liquid was removed by gently blotting. Coverlips were mounted onto a microscope slide with Prolong Gold anti- fade reagent (Invitrogen) and allowed to set overnight. A control using IgG instead of anti- 48

TG2 antibodies was used to take into account background fluorsecence. Cells were examined under a Zeiss LSM 700 laser scanning microscope with excitation at 358nm (blue), 488nm (green), and 600nm (red) for DAPI, Alexafluor 488, and phalloidin, respectively.

2.10 RhoA and Rac1 profiling in MEF

The small GTPases, RhoA and Rac1 are dynamic regulators of the cell spreading process (details in Section 4.1.2.2), which are active when bound to GTP. Because of this, activity of these small GTPases is reflected by the GTP-bound form and not total protein expression. To detect the active form of these proteins were detected using Rhotekin Rho- binding domain and Pak Rac-binding bacterial proteins, for RhoA and Rac1 respectively.

2.10.1 Expression of GST-Rhotekin Rho-binding or GST-Pak Rac-binding domains from E. coli

E. coli rosetta harbouring either GST-Rhotekin Rho-binding domain or GST-Pak Rac- binding domain both in pGEX2T (a gift from Marion Mohl, VCCRI, Australia) in YT medium (2x; 5ml) was inoculated and agitated at 250rpm overnight at 37oC. To this was added 150ml 2x YT medium with ampicillin (250μg/ml) until an optical density of 0.6-0.8 at 600nm was reached. Protein expression was induced with 1mM IPTG for 1hr at 37oC for with agitation at 250rpm. E. coli was collected by centrifugation at 2800g for 10mins at 4oC. Supernatant was aspirated and the pellet was resuspended in 10ml ice-cold 1x PBS buffer. Cells were sonicated with a Branson 400 Digital Sonifier on ice at 40% amplitude output for 2mins twice with a 2min break in between. After centrifugation at 16100g for 10mins at r.t., the supernatant fraction was aspirated and loaded onto glutathione sepharose beads (Section 2.1.1.3).

2.10.2 Glutathione sepharose beads equilibration

Glutathione sepharose beads (500μl) were washed twice with 10ml of 1xPBS buffer. Wash buffer was carefully aspirated without removing any beads, and supernatant from Section 2.1.14.1, containing either GST-Rho-binding domain or GST-Rac-binding domain, was

49 incubated with the beads, for at least 45mins at 4oC on a rotating rack. Beads were washed twice with 10ml of cold 1xPBS buffer and centrifuged at 1000g for 5mins at 4oC. Beads were resuspended in 600μl of 1xPBS buffer and a 1μl sample was quantitated against a BSA standard curve developed with Coomassie plus reagent and the 600nm absorbance determined using an ELISA plate reader. Protein-bound beads were divided into 50μg aliquots in 100μl of GST-Fish buffer (10% v/v glycerol, 50mM Tris-HCl buffer pH 7.4, 100mM NaCl, 1% Tergitol-type NP-40, and 2mM MgCl) and stored at –80oC.

2.10.3 Pull-down assay of activated RhoA or Rac

Serum-starved MEF (1x106 cells) were allowed to settle on plates coated with Fn (10μg/cm2) for 0, 30, 60, or 90mins. All subsequent steps in this procedure were performed at 4oC unless otherwise stated. At the desired time point, plates containing MEFs were placed on ice, media was aspirated and cells were scraped in 500μl of GST-Fish buffer using a rubber policeman. Cells were lysed by pipetting up and down, transferred to a fresh eppendorf tube and centrifuged at 16100g for 5mins. A portion of the lysate (100μl) was transferred to a new tube with 6x Laemmli buffer (Section 2.1.2.1), and the remainder was added to 50μg of either GST-Rhotekin Rho-binding domain or GST-Pak Rac-binding domain and incubated for 1hr on a vertical rotating platform. Incubated beads were collected by centrifugation at 1000g for 5mins and washed twice with 1ml GST-Fish buffer. After the last wash, all supernatant was discarded and 25μl of GST-Fish plus 5μl of Laemmli buffer was added. Both bead samples and pre-pull down lysate were boiled at 100oC for 5mins. Denatured samples and, separately, molecular size markers were subsequently loaded onto a 15% reducing polyacrylamide gel (15% acrylamide/ bis solution, 37.5:1, 2.6% cross-linker monomer [Biorad]; 0.37M Tris, pH 8.8, 0.01% SDS; 0.005% aluminum persulphate; 0.5M tetramethylethylenediamine). Samples were subjected to Western blotting (Section 2.1.2.2) using rabbit anti-mouse RhoA polyclonal antibody (Abcam) to detect RhoA and rabbit anti-mouse Rac1 polyclonal antibody (Abcam). A positive control was included where TG2+/+ MEF lysate was incubated with 5mM GTPγS for 30mins at 37oC prior to pull-down with beads.

2.11 In vitro scratch wound assay

TG2+/+ or TG2-/- MEFs were grown to confluency in wells of a 6-well plate in DMEM/10% FCS. Confluent MEF monolayers were synchronised in 0.1% FCS DMEM for at least 16 o hrs at 37 C under 5% CO2. Just prior to experiment, 0.1% FCS DMEM was aspirated from MEF monolayers and replaced with fresh 37oC 10% FCS DMEM. Using a sterile 200μl tip, a scratch was made to the fibroblast monolayer. The border of the denuded area was immediately marked under a Nikon eclipse TS100 microscope at 10x magnification. The denuded area was recorded at 0hrs and 48hrs post-scratching. In some experiments, purified TG2 (500μg/ml) was added immediately to wells containing TG2-/- MEF. Denuded areas were evaluated using the software ImageJ and expressed as a fraction of day zero area.

2.12 Statistical analysis

All results are shown as mean ± standard error of the mean (SEM). In vivo wound healing assays were compared between genotypes using repeated measures two-way ANOVA with post-hoc Bonferonni correction. Areas under the curve for wound healing assays were compared between genotypes using one way ANOVA or a two-tailed Student t test. Microarray data was compared between genotypes using two-way ANOVA with post-hoc Bonferonni correction. Quantitative PCR for an examined gene was compared between healthy and wounded tissue using a two-tailed Student t test. Statisitcal significance was reached if p-value is less than 0.05. Cell adhesion assays performed for varying times, seeding concentrations, Fn amounts, and TG2 or mutant amounts were compared between genotypes using two-way ANOVA with post-hoc Bonferonni correction. Cell adhesion assays with the addition of TG2 were compared between untreated and treated MEF using a two-tailed Student t test. Cell spreading assays were compared between untreated and treated MEFs or between time points within the same genotype using a two-tailed Student t test. Rho and Rac expression was compared between time points using a two-tailed Student t test. Scratch assays were compared between untreated and treated MEF at the same time point using a two-tailed Student t test. Statistical significance was considered to have been reached with a p-value less than 0.05.

Figure 5 Maintenance of mouse lines for breeding and experimentation

The use of a stem and expansion colony ensures that breeding continues to propagate the line (Stem) while the production of suitable study mice can be optimised (Expansion). An example is given here of a small colony depicted over 5 generations (G0-G4), where two pairs, a heterozygous x wildtype (het x wt) heterozygous x heterozygous (het x het), populate the stem colony. Note that one pair only contributes future stem breeders. Mates used in this example are unrelated wt, as in a backcross, but brother-sister pairs are recommended if an inbred background is to be maintained. Het animals supply the expansion breeding, where a sufficient number of mated pairs (N) are set up to produce the required animals (wt or homozygous knockout [hom]) for study. Expansion colony breeding animals should be refreshed from the stem colony at each generation to prevent disparity developing between the two colonies due to genetic drift.

Figure 6 TG2 genotyping by PCR schematic

This is a schematic showing the target amplicon for the wildtype Tgm2 gene (TG2+/+) using primers, P2 and P3 (top), and the deleted Tgm2 gene (TG2-/-), using primers, P1 and P3 (bottom). The location of primers is shown in red. Horizontal lines represent a portion of the TG2 gene, and vertical lines represent exons with numbering above. The region deleted in TG2-/- (bottom) is part of intron 5 and intron 8, with the residual loxP site shown in blue [87]. Below each schematic is a typical photograph of the PCR products of 13 samples (M=100bp ladder markers) run on a 3% agarose gel, the wildtype product is 100bp (top) and knockout product is 300bp (bottom). A homozygous wildtype sample has the wildtype band only (i.e. sample 4, 5, 7 and 9); a homozygous knockout sample has the knockout band 1only (i.e. sample 1, 3, and 8); and a heterozygous sample has bands from both reactions (i.e. sample 2, 6, 10, 11, 12, and 13)

3. THE EFFECT OF TG2 ON IN VIVO CUTANEOUS WOUND HEALING

3.1 Introduction

There are many reports in the literature implicating TG2 in the wound healing process [11, 233, 234], based on its ability to interact with and affect inflammatory cells [142, 224], cell surface receptors [82, 83, 85] and matrix proteins [82, 83, 118]. Fibroblasts, a central player in skin wound healing, isolated from TG2-/- mice was shown to be defective in cell adhesion [87]. However, as yet there is no direct evidence for the involvement of TG2 in wound healing in vivo. Using one founder line of TG2-/- mice on a mixed background (129SJ/SvImJ/C57Bl/6), a few years ago our laboratory obtained preliminary evidence that these animals had delayed cutaneous wound healing relative to the TG2+/+ mice of the same mixed background strain [235]. This initial observation has been followed up in this work with the TG2-/- mice on a pure 129T2 genetic background (129T2-TG2-/-).

3.2 Results

3.2.1 In vivo wound healing assay in 129T2-TG2 mouse

Age-matched male animals were subjected to the wound healing assay using a circular wound model (Figure 9). The wounds did not bleed or expose hypodermal structures such as blood vessels and underlying muscles. To satisfy animal welfare concerns, wound healing experiments were initially performed using local injection of the anesthetic bupivacaine (8mg/kg) prior to skin incision (Figure 9a & b). In both genotypes, from day 1 to 3 post-injury the wounds increased in size by up to 1.4 times (peaks: 1.4 times in 129T2- TG2-/- at day 1 and 1.2 times in 129T2-TG2+/+ at day 2) relative to the original wound created by the punch biopsy. Wounds began to close at day 4 in both genotypes and the areas under the curves for the wounds in anesthetised 129T2-TG2+/+ and 129T2-TG2-/- mice were no different from each other. The total number of days taken for complete wound closure was 13 days (Figure 9a).

This result was disappointing. Two things were not immediately clear: 1) why the initial observation of delayed wound healing was not confirmed in vivo, particlarly since early experiments undertaken in parallel in vitro (Section 4) indicated an adhesion defect in 54 primary fibroblasts from 129T2-TG2-/- relative to 129T2-TG2+/+ mice, and 2) why the wound size increased on day 1 to 3 relative to day 0, when in the previous preliminary results where no bupivacaine had been administered, there was no increase, Since bupivacaine has anti-inflammatory properties and inflammation is an important process immediately post-wounding, animal ethics approval was sought to omit bupivacaine administration to the wounds.

When the wound biopsy was repeated in the absence of bupivacaine, there was no initial increase in wound are and the wounds of 129T2-TG2-/- animals were now retarded in their initial rate of closure. This was most evident in the initial 5 days of wound healing, with no significant difference in wound closure rate being observed thereafter (Figure 9c). Thus, in 129T2-TG2+/+ mice, on the first day the wound area was reduced to ~0.5 of day 0 compared to about 0.7 of the day 0 area in 129T2-TG2-/- mice. By day 5, the wounds in TG2+/+ mice had contracted to ~0.3 of day 0 area relative to about 0.5 in TG2-/- mice. From day 6 onwards, there was no difference between the two groups in terms of wound size as a fraction of day zero. Complete wound closure in both genotypes required 10 or 11 days. The integrated wound closure time was calculated by measuring the total area under the curve of wound area fractions versus time (Figure 9a & c). This measure is an indication of the total time the wound was exposed. The integrated wound closure times were not different between TG2+/+ and TG2-/- mice that were administered with bupivacaine (Figure 9b). In TG2+/+ mice was: 1.86±0.13 vs 4.11±0.33 in TG2-/- mice (p<0.001, two-tailed t-test) (Figure 9d). Further more, the integrated wound closure time for the first 5 days of wound healing was significantly different in wounds from TG2+/+ vs TG2-/- mice (2.284±0.157 in TG2+/+ mice vs 3.024±0.093 in TG2-/- mice; p<0.001, two-tailed t-test), but was not different from day 6 onwards (0.366±0.094 in TG2+/+ mice vs 0.392±0.093 in TG2-/- mice; statistically not different two-tailed t-test). The omission of bupivacaine from the wound healing experiments solved the apparent increases in wound size.

3.2.2 Exogenous addition of purified TG2 to healing skin wounds

Since wounds from 129T2-TG2 mice lacking TG2 expression showed a retarded rate of initial wound closure, it was of interest to investigate the potential effect(s) of exogenous application of purified TG2. 55

The expression of rat TG2 protein, as a GST fusion protein, was induced with IPTG under the control of the lac operon. Expressed protein was adsorbed to a glutathione sepharose column and TG2 released by thrombin digestion, which cleaves the fusion protein at a site between TG2 and GST. TG2 was further purified by FPLC using anion exchange chromatography. TG2 typically eluted as a sharp peak, as indicated by the UV spectrometery at 280nm, at a sodium gradient of around 25-30% (Figure 7), with an average yield of about 2mg of purified TG2 per 1L of E.coli culture. A sample of the collected protein was identified as the 75kDa TG2 protein by Western blotting with anti- TG2 antibodies (Figure 8). TG2 catalytic function was tested by 3H-putrescine incorporation into DMC (Section 2.1.3). The activity of freshly isolated TG2 was compared to TG2 that was stored for a day or a week at 4oC (Figure 8). The assay revealed that in the presence of 2mM Ca2+ TG2 has a specific activity of 247pmol cross-linked product per pmol of TG2 per minute. The activity was not affected by storage at 4oC overnight but was markedly reduced to less than 50% of the fresh protein’s activity with storage at 4oC for a week. Purified TG2 thawed from –80oC stock or from stock stored for up to one night at 4oC was used for subsequent experiments.

Age-matched male animals were subjected to the same punch biopsy wounding as in Section 2.1.6. Genotypes were blinded and wounds were randomly chosen for one of the following two treatments: either with 0.5mg of TG2 (in PBS) applied to the surface of a freshly created wound on day zero, or, as a control PBS alone. The PBS or TG2 in PBS droplet was allowed to absorb into the wound while the animals were still anesthesised. The wounds in 129T2-TG2+/+ and 129T2-TG2-/- animals, had similar healing profiles to the previous experiment, and in both genotypes complete wound closure was observed by 10±1 days (Figure 9c). This shows that PBS had no confounding effect on wound healing. A single application of TG2 on day zero to wounds in 129T2-TG2+/+ mice did not accelerate wound healing at any stage nor did it speed up wound closure time relative to saline alone, i.e. the rate of wound closure was identical (Figure 10a). The overall wound closure in 129T2-TG2+/+ mice, calculated from the area under the curves, was no different between wounds treated with or without TG2 (Figure 10c). In the case of wounds in 129T2-TG2-/- mice, a single exogenous application of TG2 on day zero decreased the wound area at day 1, 3, 7, and 8 days post-injury (Figure 10b). As early as day 1 post-injury, the effect of 56 exogenous TG2 decreased wound size from 0.76 to 0.67 of that on day 0 (p < 0.05). The efficiency of wound healing as measured by the area under the curve was increased, relative to saline-treated controls, a result of TG2 treatment (Figure 10c). The efficiency of healing of TG2-treated wounds in 129T2-TG2-/- mice was statistically not significantly different from that of saline or TG2-treated wounds in 129T2-TG2+/+ mice (Figure 10c).

3.2.3 Monocytes and neutrophils at the wound site

Next, given the significant impairment of early wound healing in the 129T2-TG2-/- animals, we investigated the reasons underlying this phenotype. Inflammation is the earliest phase in wound healing. Given that the significant retardation in wound closure coincided with this phase, it was of particular interest to study two predominant inflammatory cells: monocytes and neutrophils. Lack of TG2 expression in these two cell types has been reported to affect the functions of these cell types in both C57Bl/6J mice [219, 220, 236] and in humans [219, 237] and consequently wound healing. Transversely sectioned wounds were examined at the wound’s edge next to the wound void as indicated by Figure 10a. In wounds from TG2+/+ mice, the neutrophil and monocyte profiles were as anticipated at the site of injury. Neutrophils and monocytes were identified based on H&E stained morphology from the sections (Figure 11b). During day 1 post-injury, neutrophil infiltration was high, at around 70 cells per field, and over the course of the next four days this population slowly declined (Figure 11c-g). Monocytes, on the other hand, followed a reverse trend, with 20 cells found per field on day 1 post-injury, presumably derived from the resident population of monocytes within the tissue, and slowly, as healing progressed, monocyte numbers doubled to 40 cells per field. The profiles of neutrophil and monocyte infiltration into wounds of TG2-/- mice were the same as their wildtype counterparts and no significant differences were observed between TG2+/+ and TG2-/- mice at any time point for either population (Figure 11c-g). These findings differ from the observations that macrophage infiltration was positively correlated with TG2 protein expression in a skin injury study [238] or decreased neutrophil recruitment to kidneys with endotoxic shock in TG2-deficient mice [239].

3.2.4 Microarray analysis of 129T2-TG2+/+ and 129T2-TG2-/- wounds

Wounds from 129T2-TG2+/+ or 129T2-TG2-/- animals were collected at day 2 post- wounding (n = 6) for microarray analysis. At this time the difference in wound size between 129T2-TG2+/+ and 129T2-TG2-/- mice was statistically different (Figure 8a) and since results from Section 3.3.4 indicated neutrophil and monocyte infiltration of the wound was unaffected immediately after wounding (day 1), wounds were sampled at day 2 as this represents the late inflammation/early cell proliferation phase. Because microarray experiments require high quality RNA that is pure and intact, samples were subjected to RNA integrity testing (Figure 11a & b). The readout from the Bioanalyzer is an RNA integrity number (RIN) that is calculated based on the 28s and 18s rRNA peaks and their rate of degradation over the course of capillary electrophoresis [240]. Samples with RIN greater than 7 are considered fit for microarray analysis and all twelve samples had RIN greater than 7 (Figure 11b).

For the single time-point comparison between the two genotypes, a one-colour array was used. This array contains probes that correpThis detects sample mRNA that is labeled with a single-colour dye by hybridization to a given array of probes, and each sample is subjected to separate single arrays. A microarray experiment using one-colour arrays yields signal intensity data that reflects the abundance of detected mRNA, and relative gene expression levels can be compared between arrays hybridised to different samples (e.g. control vs treatment). This is in contrast to a two-colour array where control samples and treated samples are differentially labeled and hybridised to the same array. Relative expression levels between the different samples are scanned and relative intensity of each fluorophore can be compared for each gene/probe on the one array. To subject RNA prepared from day 2-post injury wounds of TG2+/+ or TG2-/- mice for microarray analysis, either a one-colour or two-colour array would suffice. A one-colour array was selected for this investigation because it offers the convenience to extend gene expression analysis to other time points or treatments, if needed. For example, utilisation of the one-colour array would only require additional array hybridisation of samples from another time point(s) in order to be able to compare it with the existing day-2 data. Whereas in the case of the two- colour array, new arrays would be required for each unique combination of comparisons (e.g. day 2 129T2-TG2+/+ vs day 3 129T2-TG2+/+; day 2 129T2-TG2-/- vs day 3 129T2-

TG2-/-; day 3 129T2-TG2+/+ vs day 3 129T2-TG2-/- … etc). Thus the overall cost efficiency of the experiments would fall if more arrays and sampling was needed, and the number of animals required would increase.

From the microarray results, ten of the samples had consistent spread of data with similar inter-quartile ranges, while the remaining two samples, 13568 and 13571, had a larger spread that could impact on analysis outcome (Figure 12a). The normalised expression signals (normalised to control spike-in RNA to take into account non-biological or technical variations [241]) of all 1.2 million probes are shown in Figure 12b with one line representing a single detected probe. Quality controls included internal grid controls and RNA spike-ins. The array has control spots to varied amounts of prokaryotic genes from the biotin synthesis pathway: bioB, bioC, and bioD from E.coli and Cre recombinase from P1 bacteriophage and each sample was spiked with the complementary RNA for these genes (Figure 12c). One sample (12515) had a reduced score for bioB and bioC relative to all other samples, which otherwise were all consistent in all other samples. To prevent bias, this sample was omitted from subsequent analysis. The quality and quantity were robust and fulfilled all criteria for validation and control. The relative expression (signal to noise ratio) of genes in wounds from 129T2-TG2-/- mice compared to 129T2-TG2+/+ mice was plotted in a volcano plot (Figure 12d). This plot displays the p-value (two-way ANOVA with Bonferroni correction) versus relative expression changes, allowing quick visual inspection for individual elements that have significant changes. The threshold values were set at p < 0.05 on the y-axis and a relative fold change of at least 0.5. On this 129T2-TG2+/+ vs 129T2-TG2-/- plot, the Tgm2 gene was the only gene (red box, right top quadrant) that was significantly different between the two groups indicated by a two-fold increased expression in 129T2-TG2+/+ relative to 129T2-TG2-/- samples. Similarly, rainbow plots of the whole wound RNA repertoire showed one outstanding line representing Tgm2, changing from –1 normalised expression intensity in wounds from 129T2-TG2-/- mice to >+1 normalised expression intensity in wounds from 129T2-TG2+/+ mice (Figure 12e). Lastly, principle component analysis, a statistical method that breaks down each sample into a set of influencing genes and attempts to correlate samples with treatments [242], showed that there was no correlation between the two genotypes (Figure 12f). Thus even though a phenotypic difference between 129T2-TG2+/+ and 129T2-TG2-/-was observed, microarray analysis did not detect any unique expression patterns between the two groups; 59 a finding that was as expected since TG2 was the only gene that differed significantly between the two in the analysis. A full list of genes that had changes of at least 0.1 fold is listed in the Appendix.

3.2.5 Quantitative analysis of wound-related genes using real time (RT)-PCR

As with all microarrays used to examine gene expression, the data yielded from microarrays must be validated by means of mRNA quantitation techniques such as RT- PCR [243]. The first group we were particularly keen to look at included all other catalytic members of the TG family to discount any potential compensatory effects from these other TGs. The second group comprised components of the important pattern recognition receptors, toll-like receptors (TLRs) involved in innate immunity. TLRs signals through two distinct pathways: MyD88 dependent and the MyD88 independent pathways. MyD88 (Myd88 as labeled in Figure 14) binds to a Toll/Interleukin-1 receptor (TIR) domain of TLRs leading to expression of proinflammatory cytokines such as IL6 and TNF-α. In the MyD88-independent pathway, TICAM1 (Ticam1), a classical mediator that also binds directly to TLR4, results in the induction of apoptotic cytokines such as IFN-γ [244]. Both pathways transmit signals through nuclear factor κB (NfkB) [245], a potent transcription factor regulating inflammation, and its inhibitor IκB (Nfkbib). TG2 has been shown to induce the activation of NFκB by transamidation of IκB in microglial cells [246]. IκB transamidation sequesters it and prevents its interacting with NFκB, hence allowing the nuclear translocation of NFκB. Because the Tgm2 promoter is a target of NFκB, TG2 expression is consequently upregulated with increased NFκB activity [247]. The third group included genes of some early inflammatory cytokines and growth factors such as INF-γ (Ifng), TNF-α (Tnf), and IL-6 (Il6), which are involved in early innate immunity, and TGF-β-2 (Tgfb2), EGF (Egf) and FGF-2 (Fgf2), which are secreted in wounds to promote reepithelialisation. The fourth group consisted of cellular adhesion and migration proteins: collagen type I and III (gene: Col1a1, Col1a2, Col3a1), fibronectin (Fndc4) and its wound- related splice variant fibronectin extra-domain A (Fnd3c2), and the cell surface receptor syndecan 4 (Sdc4).

Quantitative RT-PCR (qPCR) was performed on RNA isolated from excised wounds, and as a control RNA isolated from non-wounded skin. Expression levels of investigated genes 60 were calculated as a ratio to the housekeeping gene hypoxanthine-guanine phosphoribosyltransferase (Hprt), the expression of which has been shown not to change in skin wounds [248]. Of the TG group of genes (Figure 13a), Tgm2 expression was absent from both 129T2-TG2-/- healthy and wounded skin as expected (p<0.001). There was no difference in expression of any of the other TGs between the genotypes within either the healthy or the wounded skin group. With wounding, the expression of three TGs was upregulated relative to the unwounded group: Tgm1 by ~1.5 folds in both genotypes (p<0.05, two-tailed Student t-test), Tgm2 by ~2 fold (p<0.001) in 129T2-TG2+/+ only, and F13a1 by ~1.5 fold (p<0.001) in both genotypes. Components downstream of the TLR4 pathway showed increased expression following wounding (Figure 13b) including, Ticam1, ~1.5 fold (129T2-TG2+/+ p<0.01, 129T2-TG2-/- p<0.001); MyD88, ~2 fold (129T2-TG2+/+ p<0.05, 129T2-TG2-/- p<0.01); Nfkb1, ~1.4 fold (p<0.01); and Nfkbib, ~1.6 fold (129T2- TG2+/+ p<0.05, 129T2-TG2-/- p<0.001). Cytokines and growth factors that were upregulated in expression following wounding (Figure 13c) were: Il6, ~6 fold (p<0.001); Ifng, ~5.1 fold (p<0.001); Tnf, ~6.5 fold (p<0.001); Egf, ~9.5 fold (p<0.001); Fgf2, ~13.2 fold (p<0.001); and Tgfb2, ~2 fold (p<0.001). Various ECM components that were upregulated in expression (Figure 13d) were: Fndc4, ~2 fold (p<0.001); Fndc3a, ~4 fold (p<0.001); Col1a1, ~1.6 fold (p<0.001); Col1a2, ~3 fold (p<0.001); Col3a1, ~1.25 fold (129T2-TG2+/+ p<0.001, 129T2-TG2-/- p<0.05); and Sdc4, ~3 fold (p<0.001). However, there were no significant differences in the increases in these genes between the 129T2- TG2+/+ and 129T2-TG2-/- wounds.

Figure 7 TG2 elution peaks from FPLC purification

This shows a typical peak recorded for TG2 elution from an FPLC (Mono Q 5/50 ion exchange column, GE Healthcare). The blue trace represents UV spectrometer readings at 280nm as protein is eluted as a single narrow peak from the column. The protein was later identified as TG2 by Western blotting. The red diagonal line represents the NaCl gradient, in percentage, of eluting buffer as marked by the values in the middle.

Figure 8 Characterisation of recombinant TG2 protein

Protein concentration was determined by a colorimeteric protein determination quantification assay using Coomassie Plus reagent (Section 2.1.4, data not shown). A typical yield from six litres of E.coli DH5α M15 culture ranges from 8 to 10mg of TG2. (a) TG2 quality was tested by checking its cross-linking activity via the incorporation of 3H-putrescine into DMC. The displayed graph shows mean specific cross-linking activity caculated as pmol of product/pmol TG2/min (± SEM) for fresh TG2, TG2 stored for 1 day at 4oC and TG2 stored for one week at 4oC.(b) TG2 protein was run on an 8% denaturing polyacrylamide gel and identified by Western blotting using an anti-TG2 antibody.

Figure 9 In vivo wound healing assay on mouse skin

Wound healing was monitored in 129T2-TG2 mice after wounding with a 5mm punch biopsy. (a) Wound healing was monitored with 129T2-TG2+/+ animals (solid squares) and 129T2-TG2-/- animals (hollow squares) receiving the local anesthetic 8mg/kg Bupivacaine and wounds were measured and expressed as a fraction of day 0 area (mean ± SEM) until closure (n=10). (b) The corresponding total area under curve (mean ± SEM). (c) Wound areas from 129T2-TG2+/+ (solid squares) and 129T2-TG2-/- (hollow squares) were measured and expressed as a fraction of day 0 area (mean ± SEM) until closure (n=16). (d) Total area under the curves calculated from (c). (e) Mouse wounds were normalised between measurements using a metric ruler. (f) Representative photographs of mouse wounds from (c) of both genotypes at day 0, 2, 4, 6, 8, and 10 are shown; scale bar = 5mm. *** denotes p<0.001 for a repeated measure with two-way ANOVA with Bonferroni correction, and ### denotes p<0.001 for a two tailed Student t-test.

Figure 10 In vivo wound healing assay with exogenous TG2 addition

129T2-TG2 mouse wounds were treated with the application of exogenous TG2. (a) 129T2-TG2+/+ (n = 10) and (b) 129T2-TG2-/- mouse wounds were treated with PBS (square) or with 50μl of 10mg/ml of purified TG2 (diamond) and expressed as the fraction of day 0 (mean ± SEM) area until closure. (c) The overall wound area exposed during healing was measured by taking the total area under each curve (mean ± SEM). * denotes p<0.05 and ** denotes p<0.01 for a repeated measure with two-way ANOVA with Bonferroni correction, and ### denotes p<0.001 one way ANOVA with multiple group comparison, ns denotes ‘no significance’.

Figure 11 Neutrophils and monocytes at the wound edge

129T2-TG2 mouse wounds from in vivo wound healing assays were collected and histologically examined for monocytes and neutrophils. Wounds were transversely sectioned and stained with hematoxylin and eosin. Tissue immediate to the wound’s void was examined by taking ten random fields. (a) An example wound micrograph (10X magnification) indicates area from which random fields were taken as enclosed by dashed line, the stained area to the left is the wound void, bar = 100μm. (b) A typical random field (60X magnification) showing an example of a bean-shaped nucleus of a monocyte (white arrow) and an example of a multi-lobed nucleus of a neutrophil (black arrow), bar = 10μm. (c-g) Monocytes (yellow bar) and neutrophils (pink bar) (± SEM) were counted (n = 3) from ten different random fields for each genotype from day 1 to 5 post-injury.

Figure 12 Analysis of integrity of the RNA from wound samples

RNA isolated from 129T2-TG2 wounds, from 6 TG2+/+ and TG2-/- animals each, was checked for integrity before subjecting to microarray analysis. Samples were run on a nanochip 600. Laser-induced fluorescence from dye that was added to the samples was detected and a computer-generated (a) gel image and (b) electropherograms were obtained. The gel image displays molecular weight markers (Ladder; 25, 200, 500, 1000, 2000, & 4000bp) and twelve sample lanes. Each sample showed two prominent bands corresponding to 28s rRNA (top) and 18s rRNA (bottom) with varying degree of breakdown products. Green bands at the bottom represent the dye front. Good quality samples are indicated by the crispness of these bands and low background tinge. Electropherograms show fluorescence peaks detected from each sample and an RNA integrity number (RIN) number out of ten. Skin samples with RIN above 7 were considered sufficiently good quality for subsequent microarray analysis.

, 74

Figure 13 Microarray comparison of 129T2-TG2+/+ and 129T2-TG2-/- wounds

+/+ -/- Wound RNA from total 129T2-TG2 and 129T2-TG2 mice (n = 6) was hybridised to GeneChip Mouse Exon 1.0 ST Array using single colour probes. (a) The distribution of the normalised signal intensity is plotted in a box and whisker box for all twelve samples against normalised log fold changes. Red boxes represent signal intensities that were beyond 1.5 times the interquartile range. (b) The profile plot of the 44,000 3’ expressed genes detected in each sample is shown as normalised log fold changes against the twelve mouse RNA samples. Each coloured line represents one of the 44,000 genes. (c) Quality control of microarray was checked by the efficiency of probe hybridization for each sample. There are four controls (see text) used for this experiment. One of the samples, 12515 (indicated by the arrow), had a score for the “BioC” control that deviated from the other samples and was excluded from subsequent study. (d) A volcano +/+ -/- plot shows gene expression between 129T2-TG2 vs 129T2-TG2 , y-axis represents –log10 Bonferonni- corrected p-values (comparing gene expression between the genotypes) and x-axis represents the differential normalised expression between the genotypes. Threshold values were set at p<0.05 for y-axis and ± 0.5 fold for x-axis, grey boxes represent genes that are below both thresholds and red boxes represent genes that are above both thresholds. (e) The samples were segregated into their respective genotype, TG2-/- (“KO”) left and TG2+/+ (“WT”) right. Each gene examined in the microarray analysis is represented by a line, which show their relative expressions in either genotype. Tgm2 (blue line indicated by arrow) was the sole gene that was significantly downregulated in the 129T2-TG2-/- relative to 129T2-TG2+/+ for two-way ANOVA with Bonferroni correction (*** p < 0.001). (f) A three-dimensional principle component analysis (PCA) plot shows that there are no pattern clusters between the samples based on unqiue gene expression patterns between the compared groups.

Figure 14 Quantitation of expression of wound related genes in wounded and non-wounded skin from 129T2-TG2+/+ and 129T2-TG2-/- mice

Total cDNA was generated from RNA isolated from wounded (or, as a control, from non-wounded ‘normal’) skin) from 129T2-TG2+/+ and 129T2-TG2-/- mice on day 2 post-injury (n = 3). These samples were examined for the expression of genes relevant to wound healing: TG family of genes (a); TLR pathway genes (b), inflammatory cytokines and growth factors (c); and ECM-related genes (d). Mean expression (± SEM) is normalised against the housekeeping gene Hprt. ### denotes p<0.001 for a two-tailed Student’s t-test comparing between genotypes within the same treatment. * denotes p<0.05, ** denotes p<0.01, *** denotes p<0.001 for a two-tailed Student’s t-test comparing between normal and wounded samples within genotypes.

3.3 Discussion

Excisional wound healing experiments were initially performed using bupivacaine as a local anaestetic (Figure 9a). Bupivacaine has anti-inflammatory properties, which intereferes with the production of cytokines such as TNF-α, IL-1β, and IL-6 through the inhibition of COX-2, an enzyme required for the biosynthesis of prostanoids that are subsequently required for the induction of pro-inflammatory cytokine production and mitogenesis [249]. This is why these wounds initially increased in size and showed no evidence of closure until day 4 post-injury. Interestingly, there was no difference in wound healing between 129T2-TG2+/+ and 129T2-TG2-/- mice. Thus, bupivacaine completely obscured any effect of TG2 ablation and was omitted from subsequent wound healing assays.

The results presented show that lack of TG2 expression in skin leads to retardation of wound closure during the first five days following wound healing, however, the time taken for complete wound closure was the same in 129T2-TG2+/+ and 129T2-TG2-/- mice i.e. 10±1 days. Nevertheless, overall 129T2-TG2+/+ animals had a significantly smaller wound area exposed indicating faster healing than 129-T2-TG2-/- animals. These observations suggest that TG2 is a moderator of wound healing but not an essential player in the healing process. This effect on wound healing of a lack of TG2 is reminiscent of the effects observed with the lack of various matrix metalloprotease (MMP) family members during wound healing [250]. Thus, as in our studies of TG2 knockout models (all on a 129T2- C57Bl/6 mixed background) of MMP 1, 3 and 8 are physiologically normal. However, when subjected to wound healing assays, wound closure is delayed compared to the wildtype, but complete closure was achieved at the same time post-injury in both wildtype and knockout animals [251-253]. Unlike those studies, however, our 129T2-TG2-/- mice only showed retarded wound closure during early wound healing, whereas wound healing was retarded in the early as well as the mid-wound healing phases in the MMP8 knockout [252], or MMP13 knockout studies [253], or in wildtype mouse treated with the broad spectrum MMP inhibitor, galardin [251]. For the 129T2-TG2-/- to show retarded early wound healing, but complete closure by the same day as the 129T2-TG2+/+ counterpart means that the rates of wound closure changed throughout the course of healing. The healing curve appeared to be tri-phasic: the first phase occurring from days 0 to 1 post- 79 injury with closure rates of 50% per day for 129T2-TG2+/+ and 30% per day for 129T2- TG2-/- (Figure 8). The second phase occurred from days 1 to 6, overlapping with significant overall retarded 129T2-TG2-/- healing (in terms of area closure). The rate of closure was however faster in 129T2-TG2-/- than 129T2-TG2+/+ with a closure rate of 4%/day for 129T2-TG2+/+ and 5%/day for 129T2-TG2-/-. The last phase occurs from day 6 until complete wound closure, which was day 11 for all samples, and the overall rates were 1%/day for 129T2-TG2+/+ and 1.25%/day for 129T2-TG2-/-. The lack of TG2 had an immediate effect on slowing the initial healing rate but there was an apparent compensatory increase in the subsequent rates of wound healing (Figure 8). The period of impaired wound healing in 129T2-TG2-/- mice coincides with the inflammatory and part of the early cell proliferation phase of wound healing. In vitro, TG2 expression is induced by thrombin, and cytokines such as IL-6, TNFα, INF-γ, all of which are found during early wound healing [124, 133, 212, 254]. On the other hand, TG2 can target an array of ECM proteins (Section 1.4.7) as well as non-ECM components such as TGF-β latent binding protein, MMPs and cell surface receptors like integrins and syndecans [85, 118, 255]. TG2 has also been shown to act as a strong adaptor molecule that augments adhesion of the ECM receptor integrin β1/β3 to Fn substrate in WI-38 fibroblasts [83]. In vivo studies in other knock out models that demonstrate defective wound healing without affecting total time to full wound closure, such as MMP8 [252] and peroxisome proliferator-activated receptor (PPAR) β [256], have attributed the defective healing to disturbances during the inflammatory phase, which impacts on healing progression.

To address the underlying mechanism of the initial wound healing defect in 129T2-TG2-/- mice, a global expression analysis approach was used with the aim of elucidating candidates genes whose expression is significantly different between the two genotypes. In turn, it was hoped these candidates would potentially map out certain pathways in which TG2 may act to give rise to the wound healing phenotype. For this purpose, the Affymetrix microarray GeneChip Mouse Exon 1.0 ST Array containing 1.2 million probe sets was chosen for this experiment. On the chip, there are on average four probes that hybridise to the mRNA of each exon; this was suitable because it enables detection of potential splice variants. For example, an extra doman A variant of Fn is expressed in wounds [257]. Samples were collected at day 2 post-injury because that is roughly when both inflammation and cell proliferation takes place and at this time there is a significant 80 difference in terms of wound size in the 129T2-TG2+/+ and 129T2-TG2-/- mice. One wound was collected per animal in either genotype, and the global expression in the two genotypes was compared directly. The microarray analyses revealed that only Tgm2 mRNA levels were significantly different between 129T2-TG2+/+ and 129T2-TG2-/- (Figure 13). This in itself was a good validation control for the microarray experiment and indicated that the experiment had worked since it showed that 129T2-TG2+/+ wounds were expressing TG2 mRNA and that 129T2-TG2-/- wounds were expressing essentially none i.e. 500-fold less compared with TG2+/+.

The microarray findings indicate that at day 2 post-wounding there was no difference in terms of mRNA expression between the two sample types. Similary, qPCR of a plethora of important wound-related genes (Figure 14) failed to show any differences between the genotypes, and we also did not find differences in neutrophil or monocyte wound infiltration (Figure 11). Taken together our studies provide compelling evidence that the predominant mechanism for the early delay in wound healing in 129T2-TG2-/- relative to 129T2-TG2+/+ mice is an effect restricted directly to a lack of TG2 protein and/or activity, rather than to indirect effects of TG2 on neutrophil or monocyte function, or on the expression of TG2 regulated genes; a conclusion supported by the finding that application of even a single dose of TG2 exogenously rescued the wound healing defect in 129T2-TG2- /- mice. In addition, our studies provide evidence that other members of the TG family fail to compensate for a lack of TG2 that is there is no evidence of redundancy in the TG family.

Of course, we cannot exclude that such indirect effects of TG2 may have been operative earlier after wound healing, in the first 24 hours of injury when there is a profound upregulation of macrophages, granulocytes, T lymphocytes, natural killer cells, and B lymphocytes markers [258, 259]. Other potential confounding factors are that i) the wound sample harvested may have contained too much healthy tissue, which may have diluted the “injury” transcriptome, or ii) transcripts within the wound environment may have been rapidly degraded. With regards to the former, this is unlikely to be a factor, despite the fact that it can be calculated (Appendix II) that only 0.3% of the sample was wound tissue itself (rather than adjacent normal skin), since all of the many candidate genes evaluated by qPCR, which are well described to be involved in inflammation and wound healing, were 81 found to be significantly altered in RNA prepared from wound samples as compared to normal skin sampled from a site distal to the wounds. It has been shown that oligonucleotide microarray and qPCR detection of gene expression can vary by 13 to 16% [260], and in other studies correlations between the two methods can generate R-values that ranged from –0.48 to +0.93 [261]. Furthermore, microarrays in general, including whole genome arrays used in this study, are documented to reach a limit of detection (a point where data cannot be differentiated from noise) with starting total RNA sample of about 100ng [262]. If only 0.3% of the injury sample were indeed injured tissue, then the current microarray protocol may not have detected genuine differences between 129T2-TG2+/+ and 129T2-TG2-/-.

Wound closure involves the physical contraction of the wound edge, a process that relies on fibroblasts [200]. Fibroblasts are integrators of wound healing and perform two major tasks that facilitate proper wound healing. They produce the provisional matrix, the ground substance for new epithelium formation [188, 189] and differentiate into myofibroblasts, which contract the wound edge [199]. Fibroblast motility is dependent on the expression of receptors that recognise the extracellular matrix. In the wound setting, fibroblasts rely primarily on integrin β1 and β3 to migrate on wound substrates such as Fn [263]. TG2 has been shown to interact with these different components [83]. Considering the importance of fibroblast migration on Fn in wound healing and TG2’s strong affinity for integrins and Fn, it is possible that the lack of TG2 expression in wounds could lead to compromised wound healing, as fibroblast migration could be affected.

Wound healing is dynamic and requires the active migration of leukocytes, keratinocytes, and fibroblasts to restore skin integrity. However, it is evident from the wound healing assay result as well as the mRNA expression data that defective wound healing was neither due to defective leukocyte nor compensation by other TGs. This would rather suggest that it is due to the presence or absence of TG2 protein expression in fibroblast. In this sense, TG2 would be important in early wound healing as a rapid dimunition in wound area can be expected to reduce pathogens exposure and, thus, infection. Since TG2 is highly expressed in fibroblasts [264], the activity of these cells may be importantly influenced by TG2 through an autocrine mechanism. To this end, we undertook a careful in vitro evaluation of the role of TG2 in fibroblasts cell motility processes, as detailed in Chapter 4. 82

4. THE ROLE OF TG2 IN CELL MOTILITY PROCESSES: CELL ADHESION, SPREADING, MIGRATION

4.1 Introduction

4.1.1 The fibroblast and the extracellular matrix

Fibroblasts (from fibro- meaning fibrous tissue, -blast denoting a metabolically active cell or stem cells) are a heterogeneous population of cells that reside in connective tissues. Typically they are long spindle-shaped cells that adhere and spread readily on tissue culture substrates [265, 266]. They originate from multiple sources during fetal development: the dermomyotome gives rise to dorsal fibroblasts, the somatopleure provide fibroblasts to the ventral dermis, and the neural crest provides fibroblasts for the scalp and facial region [267- 269]. Fibroblasts are found throughout the body residing in connective tissue stromas. In skin, this ubiquitous cell is integral to maintaining tissue integrity by secreting various ECM proteins, serving as the “scaffolding” for tissue. Fibroblasts can lay down both structural and non-structural ECM proteins that includes the carbohydrate-containing glucosaminoglycan as well as modifiers of the ECM (Table 2) [263]. The exact composition of the ECM produced is dependent on tissue location, which in turn is determined by other cell types present and/or functional requirements. For example, at dermal-epidermal junctions, strong anchoring is required to prevent detachment of the epithelia in response to external applied forces [270]. To support this functional requirement, fibroblasts at this site secrete collagen type IV and VII and laminin 1 with neighbouring keratinocytes releasing the same collagen isoforms and other laminin types [271]. Such collateral deposition of ECM protein appears to be coordinated by keratinocyte-induced production of TGFβ by fibroblasts, which in turn up-regulates the production of collagen and laminin by keratinocytes [272, 273]. Other well-documented interactions with fibroblasts include various cell such as vascular endothelial cells [274], T- cells [275, 276], and macrophages [277]. There is no one unique biological marker that identifies the fibroblast cell. Fibroblast identification has traditionally been done through visual confirmation of the spindle cell shape. Staining for some suitable cytoplasmic markers are also used for identifying fibroblast subpopulations, these include vimentin [278], fibroblast-

Table 2 Major extracellular matrix and related proteins secreted by fibroblasts

ECM protein Function Fibronectin A glycoprotein with binding sites to other ECM proteins such as collagen, heparin, and fibrin as well as cell surface receptors such as various integrin isoforms. An important ground substance in tissue that promotes cell migration, growth, and differentiation.

Collagen Primary constituent of dermal tissue. Type I (~80%) and III (~10%) are the most dominant of the isoforms responsible for providing tissue tensile strength and elasticity. Type IV collagen associates with other ECM proteins to form a structure known as basal lamina on which cells attach. Thrombospondins A functional ECM component that regulates cell migration in the ECM by inhibiting cell adhesion and spreading Elastin Elastin is organised into elastic fibres to provide tissue with resistance to recoil after stretching. Heparan sulphate These proteins contain charged carbohydrate chains (heparan sulphates) that retain proteoglycans water and form a gel-matrix through which hormones and signaling molecules travel. The heparan sulphate chains extend into the extracellular space and and are able to bind secreted ligands in the vicinity. Heparan sulphate shedding, with or without bound ligand, appears to carry out long-distant signaling Hyaluronic acid Anionic component of the ECM that regulates osmotic content of tissue through its ionic interactions with water. Its release during wounding leads to up-regulation of cytokine production by macrophage. Matrix metallo- These are a family of enzymes that degrade specific ECM proteins requiring zinc as a proteases cofactor. Important in the remodeling of tissue.

84 specific protein-1 (in patho-fibrosis setting) [279], and α-smooth muscle actin in differentiated myofibroblasts [199]. For live cell imaging or separation, it is pivotal to utilise cell surface markers. Some accepted fibroblast surface markers include: CD90 and prolyl hydroxylase in skin fibroblasts [280, 281] and CD13 at epithelial/mesenchymal sites [282, 283].

4.1.1.1 Fibronectin

Fibronectin (Fn) is an ECM protein that crucially supports cell adhesion, migration, growth, and differentiation [284]. This protein is important in vertebrate development as inactivation of the gene leads to embryonic lethality [285]. Fibronectin exists as a dimer of two 250kDa monomers, each comprised of three repeating units: type I (40 residues), II (60 residues), and III (90 residues) [286]. The Fn gene is subject to alternate splicing of the pre- mRNA, giving rise to 20 variants in humans. The most common are the two extra domain A or B variants (extra insertion of type III repeats through exon usage); a variable region with differing length of type III repeats; and a variant that lacks the variable region altogether. Fn can be divided into two classes based on solubility: plasma Fn and cellular Fn. Plasma Fn is abundantly available in blood plasma (300μg/ml). It is produced by hepatocytes with simple splice variants of V+ or V0. Cellular Fn, on the other hand, is heterogeneous depending on the ECM tissue requirement for the specific Fn variant needed [287]. Fn can recognise several members of the heterodimeric integrin cell surface receptors and binding propagates downstream cell signaling. The best characterised interaction is between integrin α5β1 and its binding to the triplet Fn motif, Arg-Gly-Asp. Such apparently simple interaction is dependent on other complex features such as flanking regions, three-dimensional presentation of Arg-Gly-Asp, and specific interactions with the integrin binding pocket. Other known integrin-binding domains are also described that bind to integrins α4β1, α4β7, and α9β1 [288]. Fn functional activity is not limited to integrin interactions. It also binds to other proteins that include: collagen, fibrin, heparin, glycosaminoglycan and chondroitin sulfate, which allows complex ECM formation supporting a wide variety of cell types [284]. Fn incoporation into ECM requires cellular involvement and is under tight regulation by integrins [289]. Soluble (both cellular and plasma) protomeric Fn binds to the cell surface via its N-terminus, which then becomes disulphide-bonded with other Fn molecules into multimers [290, 291]. These then become 85 incoporated into fibrils of large apparent molecular mass (LAMM) on the surfaces of integrin clusters, which are exposed through the generation of tension across the cell surface [292]. This process is carried out by fibroblast, endothelial cells, and vascular smooth muscle cells [293].

4.1.2 Cell motility: from adhesion to spreading to migration

The fibroblast is central to the skin wound healing process (Chapter 3). To fulfill this action these cells must be dynamic over a range of substrates that are present in the wound bed. Cell adhesion, spreading and migration are continual processes but can be viewed separately with specific spatial and temporal cues. The dynamics of these processes are highly dependent on the cell type and context in question. For example fibroblasts, under normal physiological conditions, are widely spread and have large adhesions whereas migrating fibroblasts in wounds would be highly motile with weak adhesions and prominent protruding cytoplasmic processes.

4.1.2.1 Cell adhesion

Cell adhesion is the binding of a cell to the substratum through cell surface molecules, leading to downstream signaling for cell survival. In essence, adhesion in mammalian cells is mediated primarily by integrins. Integrins are heteromeric proteins consisting of an α and β subunit. Each subunit has a large extracellular domain, a single pass transmembrane, and a relatively small cytoplasmic tail [294]. To date, there are eighteen α subunits and eight β subunits described, giving rise to 24 biologically relevant combinations [295]. β subunits are responsible for integrin activation and signal propagation, and α subunits are integrin regulatory units [296]. Each different integrin can recognise different ECM proteins, and different ECM proteins can be recognised by different integrins. In fibroblasts and myofibroblasts, the integrins that are highly expressed are α5β1 (which recognise fibronectin; detailed in Section 4.1.1.1) and α1β1 (which recognise collagen) [263]. For cells to adhere to the ECM, integrin must bind to these molecules with high affinity and this requires inside-out activation of integrin. In the resting state, the cytoplasmic tails of the α and β subunits are engaged in a weakly-bonded clasp rendering the whole complex inactive [297]. It is proposed that if a cell was to receive signals through survival or growth 86 pathways such as: Akt [298], G-protein [299], mitogen-activated protein kinase [300], or protein kinase C [301, 302], this would activate the cytoskeletal protein talin [303]. Talin is thought to be a key protein in integrin activation; it binds β subunits 1A, 1D, 2, and 3 [304], colocalises with activated integrins [305], and has been shown to functionally activate integrin αIIbβ3 [306] and αvβ3 [307]. Latent talin homodimers are activated by calpain digestion [308], PIP2-induced conformational changes [309], and/or phosphorylation [310] to expose a domain that binds the cytoplasmic tail of the β subunit, disrupting the α-β subunit cytoplasmic interaction [297]. Talin binds to Tyr747 of integrin β subunit through its phosphotyrosine-binding domain [311]. Consequently, the β subunit switches from a bent inactive conformation to an extended active form, allowing ligand occupancy and integrin clustering and thereby propagating integrin outside-in signaling [312]. Active integrins recognise ligands based on motif sequences. For example Arg-Gly-Asp in fibronectin and vitronectin is recognised by α5β1, αIIbβ3, αvβ integrins; whilst Asp-Gly-Glu-

Ala in collagen type-I is recognised by α1β1 and α2β1 [313]. In some integrins, the binding of divalent cations such as Ca2+ is required for α-β interaction in the extended conformation and hence proper ligand binding [295, 314].

Outside-in signaling leads to the formation of adhesion complexes at the integrin cytoplasmic tail. Adhesion complexes are macromolecular complexes (up to 5 μm2) through which mechanical force and biochemical signaling is transmitted. They anchor the cell membrane to the ECM by virtue of linking the ECM to the cell’s actin filaments, as well as sensing changes in the ECM, which in turn dictates cytoplasmic movement [315]. Adhesion complexes can be categorised based on spatial and temporal appearance – focal adhesion (elongated ovoid contacts at the cell periphery); focal complex (small dot contacts at cell protrusions); and fibrillar adhesions (fibrillar or beaded contacts at cell central regions) [316]. These structures are variations of each other with varying amounts of proteins such as tensin and phosphotyrosine proteins rather than distinct classes of proteins [317]. Assembly of focal adhesions follows from integrin activation and clustering. At the cytoplasmic tail, different proteins (more than 100 identified to date, exemplifying both functional diversity and complexity) interact and aggregate with each other to form a focal adhesion complex (some common players are listed in Table 3), with actin microfilament bundles associating at the other end [318]. Cell contractility, and hence sustained

Table 3 Some common constituents in adhesion complexes

Focal adhesion Function protein Cytoskeletal and tensin Actin filament cross-linking and contains Src hology 2 domain related proteins vinculin Links integrins to actin filament through binding to either talin or α-actinin paxillin Signal transduction adaptor protein that has binding site to other structural and enzymatic proteins, and is tension sensitive. α-actinin Binds actin filament to the membrane talin Integrin activation by binding to β subunit cytoplasmic tail, also binds other focal adhesion components such as vinculin Tyrosine kinases Src Phosphorylates focal adhesion components (such as FAK, vinculin, paxillin) to facilitate focal adhesion signal transduction Focal adhesion Important in cell migration, its activity phosphorylates kinase downstream targets that promote focal adhesion turnover PYK2 Regulates FAK through FAK GTPase regulator Csk Contains SH2, SH3, and tyrosine kinase, phosphorylated Src Serine/threonine Integrin linked Interacts with cytoplasmic portion of β1 integrin, functions as kinase kinase proximal receptor kinase regulating integrin mediated signaling Protein kinase C Various downstream signaling p21 activated Target of small GTP binding protein such as Cdc42 and Rac kinase 1 and regulates cell motility and spreading Small GTPase ASAP-1 Interacts with Src PSGAP Interacts with PYK2, increases GTPase activity of Cdc42 and RhoA GRAF Sculpts the membrane locale at the focal adhesion complex Tyrosine SHP-2 Binds to various SH2 domains, involved in signaling of cell phosphatase growth LAR PTP Involved in signaling of cell growth and cell to cell contact

88 adhesion and spreading, is generated and maintained through the myosin II machinery [319]. In this complex, common binding motifs include Src-homology 2 (SH2) and Src- homology 3 (SH3) domains that recognise phosphotyrosines and proline-rich sequences, respectively [318].

The focal adhesions are complex not only because of the number of undiscovered components, but also due to conformational changes, alternative splicing, post-translational modification and proteolytic processing of its participants [318]. For example, vinculin, the abundant linking protein between integrin-bound talin and actin filament-bound α-actinin, can be conformationally changed by PIP2 binding, which induces a change in its affinity for its binding partners [320]. There are several ways in which focal adhesion components come together, but chiefly the initiation of focal adhesion is dependent on local mechanical forces. Isometric tension generated between the cell surface and solid substrate (ECM or other cell surface) induces subsequent integrin clustering and occupancy [321]. The first structures formed immediately following integrin activation are called focal complexes. These complexes are small, of size ~1μm2, can sustain weak adhesion, and defining players include paxillin, vinculin, and tyrosine-phosphorylation proteins [322]. These evolve to form focal adhesions if continual force is applied externally or internally. For example, gentle external mechanical stretching [315] or serum stimulation of myosin II-driven contractility [323] of serum-starved cells activates focal adhesion activation. Common constituents in focal adhesions are αv-integrins, paxillin, vinculin, α-actinin, talin, focal adhesion kinase, and tyrosine-phosphorylated proteins [316]. Focal adhesions are found primarily on the cell’s periphery as the cell spreads and/or migrates to an uncharted substratum. As new processes are formed, existing focal adhesions are turned over supplying pre-material for fibrillar adhesion formation [324]. These structures are translocated directionally towards the central regions of the cell, coinciding with the formation and organization of fibronectin fibrils [317]. Not all cell types display all three classes of adhesion complexes; myeloid cells such as macrophages and neutrophils develop mainly focal complexes, and migrating fibroblasts display focal complexes and stable focal adhesions [325].

Adhesion complexes effect cytoskeletal actin reorganisation by targeting downstream small GTPases, which notably include the Rho family: Rac, RhoA, and Cdc42. These GTPases 89 are active when bound to GTP, which is exchanged for GDP with the aid of guanine nucleotide exchange factors. Deactivation of these small GTPases is under the control of the GTPase activating proteins that stimulate GTP hydrolysis. Another level of regulation exists in the form of GDP dissociation inhibitors, which prevent GTPases from being activated [326]. In fibroblasts and other cells, small GTPase activation results in the formation of actin stress fibers, focal adhesion lamellipodia, and membrane ruffles [327, 328]. Hence, the adhesion complex and Rho family GTPases are involved in a signaling feedback loop [329]. Rho-kinase 1 and Rho-kinase 2 are key effectors of Rho-family kinases. These have multiple downstream substrates, some are: myosin phosphatase binding-subunit that inhibits the dephosphorylation of myosin light chains leading to tension generation [330]; Lin11-Isl-1-Mec-3 (LIM)-kinase, which phosphorylates cofilin to release globular actin (G actin) for new actin filament formation [331]; and ezrin-radixin- moesin family proteins that act as a linker between the plasma membrane and actin filaments [332]. The end products are actin polymerisation and cytoskeletal organization.

4.1.2.2 Cell spreading

Flowing from initial adhesion to ECM substrate, the cell begins to probe around, extending its cytoplasmic processes in the form of broad lamellipodia and/or spike-like filopodia, both of which are driven by actin polymerisation [333]. Cell spreading is reliant on adhesion dynamics – its formation, maturation, and disassembly – which is in turn dictated by actin filament turnover and actomyosin contraction. Cell spreading begins with protrusion of the leading edge lamellipodia from the rest of the cell lamellum via actin polymerisation and branching just beneath the membrane [333]. Lamellipodial actin growth is catalysed by actin-related protein-2 and -3 (Arp 2/3) complex, which nucleates filamentous actin (F- actin) formation. The small GTPases Rac and Cdc42 regulate Arp2/3 through the Wiskott- Aldrich syndrome protein (WASP) and WASP-family verprolin homologue (WAVE) families of proteins [334]. As new ends are added to the filaments, older parts are quickly displaced backwards, in a retrograde manner, to the cell centre due to resistance of the membrane [335]. In contrast, the actin filaments near the cell centre and rear move slowly because of myosin II contraction, which tethers parallel actin filaments and reorganizes them into thicker stress actin fibres under the regulation of RhoA-ROCK-mDia1 pathway [336, 337]. Regulation by these Rho GTPases is antagonistic. RhoA activates ROCK, 90 which inhibits dephosphorylation of myosin light chain phosphatase (MLC) by an upstream phosphatase, encouraging cell contractility. Conversely, Rac activates downstream p21- activated kinase that phophorylates MLC, decreasing contractility and promoting cell spreading [338].

Adhesion complexes (Section 4.1.2.1) are formed behind the leading edge of a lamellipodia, simultaneously promoting nearby actin polymerisation through local Rho GTPase downstream activation. Nascent adhesion complexes may elongate and grow into stable focal adhesions with myosin II-generated tension at the lamellipodia or filopodia to grip onto the substrate [339, 340]. Subsequent integrin clustering and adhesion complex aggregation recruits actin filaments [341]. These adhesions, decorated along the actin filaments, act as traction points slowing retrograde movement of actin fibres and, thus, ‘clutching’ cell lamellipodia and filopodia in place allowing further expansion of protrusions [342]. The final step for cell spreading is disassembly of adhesion to allow redistribution of membrane and cytoplasmic matter for new protusion formation [339, 340]. Disassembly in a spreading cell is most prominent at the lamellipodia-lamellum border where constant actin filament turnover is needed. In a retracting cell, adhesion complexes may also dissolve as a result of a Rho GTPase- and myosin II- dependent treadmilling mechanism at the cell edge. A key protease important in the mediation of adhesion complex disassembly is the Ca2+ activated calpain, which targets talin and the cytoplasmic domain of β3 for digestion [343].

4.1.2.3 Cell migration

Cell migration is necessary for tissue organisation, organogenesis, and homeostasis [344]. Migration is stimulated by the presence of a directional cue, such as a chemical (e.g. chemokine gradient), environmental (e.g. fibronectin ECM) or mechanical (e.g. cell stretch) signals [345]. Under favourable conditions (e.g. serum activation of receptor tyrosine kinase-Akt pathway in fibroblasts [346]), integrins are activated from inside-out allowing the cell to recognise the ECM readily. Binding onto the ECM propagates outside-in signaling leading to increased cell adhesion through adhesion complex formation and increased cell spreading through actin fiber polymerization (Section 4.1.2.1 & 4.1.2.2). Cell migration is a dynamic process resulting from tight temporal and spatial regulation of these 91 two processes. In fibroblastic migration, the fibroblast is polarised with a distinct leading and trailing edge; the leading edge points in the direction of movement. During migration, adhesions are assembled for leading edge crawling and disassembled at the trailing end for retraction [338]. Cell polarity aside from being defined by the leading and trailing edge, is also dependent on nucleus positioning, and reorientation of the Golgi apparatus and microtubule organisation centre (MTOC) [347]. In a migrating cell, these organelles follow the translocating cell body by positioning themselves towards the cell front. Regulating the movement of these is primarily attributed to the small GTPase Cdc42 through the Par/PKC pathway for MTOC and the myotonic dystrophy-related Cdc42 binding kinase (MRCK) pathway for nuclear positioning [348, 349]. Adhesions at the front can also undergo ‘turnovers’ in which complex disassembly frees up protein components allowing nascent complexes to form at the leading edge as the cell grips onto the ECM [350]. At the trailing end, adhesions are disassembled, otherwise increased tension across the leading and trailing end would rip the cell apart. There are several ways disassembly occurs: 1) endocytosis of adhesions mediated by microtubules; 2) RhoA and Rac regulation of MLC phosphorylation; 3) proteolysis of adhesion proteins by calpain; and 4) affinity modulation of adhesion components towards each other by phosphatases such as calcineurin [338]. If a cell were to remain stationary, adhesion complexes would mature into strong focal adhesions or fibrillar adhesions instead. The molecular basis for deciding whether a cell will continue to migrate (adhesion disassembly) or stay in place (adhesion assembly) is still poorly characterised, but, at least in part, this ‘decision-making’ is dependent on a number of factors: integrin density and patterning [351], the nature and order of the integrin ligand (ECM) on a nanometre to micrometre scale [352], focal adhesion distribution [353], and actin cytoskeleton organisation and structure [354].

4.1.3 TG2 in cell motility processes

The earliest tentative work to suggest a role for TG2 in cell adhesion and growth was described in 1987 by Bryd and colleagues using PC12 pheochromocytoma cells [355]. In 1992, Gentile and colleagues definitively characterised effects of TG2 on cell adhesion and spreading. In this study, overexpression of TG2 by transfection resulted in increased cell spreading and resistance to trypsin-induced detachment of Balb-C 3T3 fibroblasts [356]. Similarly, anti-TG2 antibodies added on the outside were able to reduce both cell adhesion 92 and spreading on Swiss 3T3 fibroblasts [10], and similar observations were extended to ECV304 endothelial cells [71]. Furthermore, melanoma cells treated with TG2 were protected against dislodgement by laminar flow [357]. A mechanism for these observations was unknown until a landmark paper by Belkin’s laboratory showed by means of co- immunoprecipitation and cell assays that TG2 interacts in a ternary structure with Fn and integrin subunits β1 and β3, augmenting integrin-mediated cell adhesion [83]. TG2 binds to

Fn in a stochiometric ratio of 2:1 with a relatively high affinity, Kd 8-10nM [79, 358]. Within Fn, TG2 interacts with the N-terminal 42 kDa collagen interaction domain of Fn

(consisting of modules I6II1.2I7-9), which does not contain the classical Arg-Gly-Asp motif and does not support any known integrin isoforms. The Fn-binding site on TG2 was localised to a 28kDa N-terminal fragment within the β-sandwich domain [359]. The sequence has been honed to 88WTATVVDQQDCTLSLQLTT106, and forms an anti-parallel hairpin structure that extends out to interact with Fn [360]. TG2 interacts non-covalently with integrin β-subunits 1, 3, and 5 but not 2 [83, 361] with a 1:1 stoichiometry that requires neither of the two enzymatic activities of TG2, transamidase or GTPase [83]. Functionally, the presence of TG2 was able to support stable adhesions, some degree of spreading, and formation of focal adhesions [82, 83, 118].

Syndecans are a family of four membrane proteoglycans consisting of a protein core covalently linked to heparan sulphate or chondroitin sulphate glycosaminoglycan sugar chains forming heparan sulphate proteoglycan (HSPG) or chondroitin sulphate proteoglycan. Syndecans are receptors for the ECM (e.g. Fn) and growth factors (e.g. VEGF) utilising these long flexible sugar chains to engage ligands [362]. Syndecan-4 (Synd4) is the chief proteoglycan receptor in general cell adhesion and migration [362]. Interestingly ablation of Synd4 leads to animals with delayed wound healing [363]. Synd4 signals through protein kinase C α (PKCα) upon ligand binding, which is followed by focal adhesion formation [364]. Other syndecans (1-3) have tissue-specific distribution and interact with specific cytoplasmic partners upon activation. Syndecans are known for cooperative signaling with integrins. For example, the extracellular domain of the core protein, independent of the sugar chains, can act as a ligand for integrins, thus initiating integrin signaling [365, 366]. Another example is the recognition of ECM proteins, such as Fn or laminin, by integrins and syndecans at separate binding site to promote cell adhesion and/or cell spreading [362, 367]. TG2 is known to bind to heparin, an analogue of HSPG, 93 as is evident from the fact that heparin-sepharose can be used as an efficient affinity purification step for TG2 extraction from erythrocytes [368] and, it was later shown that this bind occurs with nM affinity [369]. The first investigation of TG2’s interaction with HSPG in a cell system was performed in a human osteoblast cell line [120]. The cleavage of cell surface HSPG using heparitinase diminished RGD-independent cell adhesion and spreading on Fn-TG2 matrix. These observations led to investigations of the biological implications of syndecan-TG2 interactions. Thus, whereas TG2-/- mouse embryonic fibroblasts (MEF) isolated from animals originially generated by Melino’s group [88] were unaffected in cell adhesion compared to their TG2+/+ counterparts, competing heparin or heparan sulphate reduced adhesion of TG2-/- MEFs but not TG2+/+ MEFs [369]. As shown by immunofluorescence experiments [369], direct interaction of cell surface TG2 proteins with Synd4 receptor involves the HSPG, rather than an interaction with Fn. Externalised TG2 in the ECM provides extra binding sites for Synd4. This may be important in vivo at least in wound healing where the presence of RGD peptides resulting from tissue degradation products may act as competitive inhibitors and prevent cell survival. In vitro,

TG2 activates PKCα signaling as a result of its Synd4 binding and subsequent β1 integrin outside-in signaling in an RGD-independent manner [85]. In addition to a functional role in cell adhesion, it appears that Synd4 may also be involved in TG2 trafficking. Knocking down Synd4 in primary fibroblasts reduced the amount of TG2 on the cell surface without affecting overall expression [369]. Working in the other direction, TG2 endocytosis may also involve HSPG together with low-density lipoprotein receptor related protein 1 [370].

Being a matrix-associated protein, TG2 abundance is regulated by a class of digestive enzymes known as matrix metalloproteases (MMP). These enzymes are endopeptidases requiring metal ions as cofactors (e.g. zinc). TG2 is a target of MMP2, a protease active in the breakdown of the ECM during embryonic development or tissue remodeling processes

[105]. In vitro results showed that MMP2 targets the TG2 core domain with an affinity (Kd) of ~380nM. Conversely, TG2 interacts with and cross links the activating intermediate of MMP2 [105]. Thus, given that TG2 binds the ECM and adhesion receptors, this would synergistically slow down MMP2 degradation of the ECM allowing for cell adhesion and spreading. Such delicate interaction between the two allows tight control over cell adhesion dynamics on the ECM [105]. TG2 has also been shown to interact with other cell surface molecules that have roles in moderating cell motility. Calreticulin is a calcium ion binding 94 protein involved in calcium homeostasis and has been shown to induce thrombospondin- mediated focal adhesion disassembly via the low-density lipoprotein receptor protein [371]. TG2 was shown to interact with both proteins in silico, although whether it is relevant in vivo has not yet been determined [371, 372].

Fn supports a variety of cell motility processes such as embryonic development, wound repair, inflammation, and vascular remodeling [284, 373]. There is mounting in vitro evidence that TG2 plays a role in cell migration. The most prominent work comes from investigations looking at TG2 in metastasis. TG2 was shown to be highly expressed in transformed 3T3 NIH fibroblasts [374], with a positive correlation with malignancy and cancer stage [375] as well as increased migration over a 3D matrix [106]. At least in the MDA231 breast cancer line and the MCF10A epithelial line, TG2 expression could be used as a prognostic marker for malignancy in situ and metastastic potential in vivo [106]. Cleavage of TG2 by MMP2 suppressed adhesion and migration of fibrosarcoma and glioma cells, but increased TG2 expression prevented apoptosis through anoikis [120]; findings that support the notion that TG2 plays a direct role in metastatic tumour spreads [118]. The interaction of Fn with TG2 seems to protect TG2 proteolysis and hence aid migration [118]. Besides cancer, TG2 has also been shown to be involved in monocyte migration on Fn [81]. Monocyte differentiation into macrophages correlates with increased TG2-Fn complex formation [81]. In CD8+ T-lymphocytes, the presence of anti-TG2 antibodies reduced their translocation across the endothelium but not their adhesion ability [218]. In pigment epithelial cells, TGF-β2 induction of surface TG2 increased both adhesion and migration [376].

Fibroblasts are important players in wound healing; they lay down a new Fn-predominant matrix as the ground substance in the proper formation of a new substratum as well as being active players in the contraction of wound edges. Because TG2 facilitates motility processes in fibroblast cell lines and the work in the previous chapter has shown 129T2- TG2-/- animals have delayed initial wound closure compared to 129T2-TG2+/+ animals (Chapter 3), this chapter explores the hypothesis that fibroblasts isolated from 129T2-TG2- /- animals are defective in cell adhesion, spreading, and migration on Fn.

4.2 Results

4.2.1 TG2+/+ or TG2-/- MEF growth curves

TG2+/+ or TG2-/- MEFs were isolated from 129T2-TG2 embryos and the first passage (P1) was used to propagate cells on 15cm tissue culture dishes for the subsequent studies. The growth rates of subsequent passages up to passage six, were monitored in terms of cell number (Figure 15). Equal numbers of cells were seeded for each genotype. After cell counting, the number of P2 (Figure 15a) and P3 (Figure 15b) 129T2-TG2+/+ MEFs was greater than those of 129T2-TG2-/- MEFs (>2x105 cells in 129T2-TG2+/+ vs <2x105 cells in 129T2-TG2-/-, [Figure 15, enlargement of boxed region]). By day 2, a pronounced difference was observed in P2 MEF numbers, and a small difference was observed for P3. This difference in cell number was sustained for P2 MEFs on day 3. As populations of MEFs of both genotypes reached saturation of >2x106, the cells were split at 1:4 dilution onto new dishes. No difference in MEF growth rates was observed between P4 (Figure 15c) and P5 (Figure 15d). The proliferation rate of both genotypes fell with subsequent passage numbers. At P6 (Figure 15e), the cells seemed to have gone into senescence and did not proliferate as quickly; a phenomenon called replicative senescence where cells taken from in vivo to in vitro stop dividing but remain metabolically active [377]. In all experiments only passage 2-5 cells were used, and those at greater than P5 were discarded.

4.2.2 Cell adhesion assay

4.2.2.1 129T2-TG2-/- MEFs are defective in adhesion

In an attempt to dissect the observations from the in vivo wound healing experiments (Chapter 3), fibroblasts, an integral player in the wound healing process, were isolated from 129-TG2-TG2+/+ or TG2-/- mice and examined in vitro for their cell motility dynamics. Because of TG2’s ability to interact with integrins and Fn [48], 129T2-TG2-/- versus 129T2-TG2+/+ primary MEFs, were examined in terms of cell adhesion to Fn matrix. The cell adhesion assay involved seeding MEFs onto a Fn matrix that is coated with TG2; assay optimisation required that the culture dishes were coated with sufficient Fn, to prevent no non-specific binding of MEFs to the plates (Figure 16). In these optimisation experiments, 129T2-TG2+/+ or TG2-/- MEFs were seeded for 30mins before the number of adherent cells was determined by crystal violet staining (Figure 16a & b). Cell adhesion was quantitated 96 as a “% of total number of cells seeded”. This percentage was determined by comparing to a standard curve constructed by seeding various numbers of cells and then cross-linking them to the coated dishes by incubation with paraformaldehyde. The results for all cell adhesion assay data are summarised in Table 4. First, it was evident that when 10μg/cm2 of Fn was used to coat (each well of a 24-well plate for 16hr at 4oC) adhesion of cells from either genotype had not saturated, and at this Fn concentration there was a significant difference between the number of adherent 129T2-TG2+/+ and TG2-/- MEFs, with less 129T2-TG2-/- MEFs adherent than TG2+/+ MEFs (Figure 16a). The largest difference between the two genotypes was observed with at cell-seeding density of 50,000 cells (per well) (Figure 16b). Cell adhesion was quantitated at 10, 20, 30, 40, 60, 90, and 120 minutes after seeding on Fn (Figure 16c). The number of adherent cells increased until 20mins post- seeding, after which the number of adherent cells plateaued. A significant difference in adhesion between 129T2-TG2+/+ and 129T2-TG2-/- was observed and maintained from 20mins post-seeding. By 30mins, compared to 129T2-TG2+/+ MEFs, where ~60% of the total number of cells seeded had adhered, only ~40% of the total number of 129T2-TG2-/- cells seeded were adherent.

4.2.2.2 TG2 addition rescues adhesion defect of 129T2-TG2-/- MEFs

Next, the effect on 129T2-TG2-/- MEF adhesion of TG2 addition to the Fn matrix was examined. First, conditions for maximising saturation binding of TG2 to Fn-coated tissue culture plastic (TCP) were established. Binding of TG2 to Fn was measured by ELISA and increased in a dose-dependent manner until saturation, at equal to or greater than 80 μg/cm2 (Figure 17). More TG2 was bound to Fn after overnight coating at 4oC then after 30mins at 37o. On uncoated TCPs, non-specific TG2 binding of MEFs saturated at 5μg/well. From these various assays, the optimal conditions to see maximum difference between 129T2- TG2+/+ and TG2-/- MEFs in the cell adhesion assay were: Fn-coating at 10μg/cm2 for 16hr at 4oC; 5x104 cells per well for a 24-well plate, with or without 20μg/cm2 TG2 incubation overnight at 4oC.

Next, increasing amounts of purified TG2 were added to the Fn matrix before cell seeding and cell adhesion to the Fn-TG2 matrix was quantitated 30 mins post-seeding, the first time point at which the number of cells that had adhered started to plateau (Figure 16c). 97

Increasing the amounts of exogenously added TG2 resulted in a dose-dependent increase in the number of adherent cells. Addition of 10μg/cm2 of TG2 to the Fn matrix increased the number of adherent 129T2-TG2-/- MEFs from ~50% to ~60% of the total number of seeded cells. Addition of 20μg/cm2 of TG2 rescued the number of adherent 129T2-TG2-/- to that of the TG2+/+ MEFs plated on Fn with no exogenous TG2 added. The number of adherent cells was further increased with addition of more TG2, peaking at ~90% of the total number of cells seeded with ≥ 60μg/cm2 of TG2 added. In the case of 129T2-TG2+/+ MEFs, cell adhesion was also enhanced with exogenous TG2 addition. Comparable to 129T2-TG2-/-, 10μg/cm2 exogenous TG2 increased adhesion by 10% from 66% to ~70%. The increase in adhesion peaked at 90% of the total number of 129T2-TG2+/+ cells seeded with 20μg/cm2 exogenous TG2 compared with 60μg/cm2 of TG2 for 129T2-TG2-/- cells.

4.2.2.3 TG2 point mutants rescued 129T2-TG2-/- and enhanced 129T2-TG2+/+ MEF adhesion

TG2 is a protein with both cross-linking activity and GTPase activity. To examine whether these activities are involved in adhesion of primary MEFs, the transamidation-deficient mutant W241A TG2 lacking catalytic activity [48] and the GTP-binding-deficient mutant R579A TG2 [48] were expressed, purified and added to Fn-coated plates prior to cell plating (Figure 19). These mutants were chosen because they were easier to express in large quantities compared to other suitable point mutants, e.g. C277S for crosslinking mutation or S171E for GTP-binding mutant. Cell adhesion was quantitated with increasing concentrations of the mutant proteins (0, 10, 20, and 40 μg/cm2). The adhesion profiles of both 129T2-TG2+/+ and 129T2-TG2-/- MEFs, in the presence of mutant proteins were similar to those observed in the presence of wildtype TG2 protein, indicating that neither the TGase nor the GTP-binding activity of TG2 are required for cell adhesion.

4.2.2.4 The β-sandwich and core domains together enhance adhesion

TG2 was reported to interact with Fn via its N-terminal β-sandwich domain (βS) [360], however, the location of the integrin-binding site is not known. In an attempt to dissect the integrin-binding site at the domain level, various domains of TG2 were purified and added exogenously to the cell adhesion assay. The domains used were β-sandwich, core, and β-

98 sandwich-core. These were added to the cell adhesion assay in equal molarity (e.g. 14.6 μg of the core equivalent in moles to 20μg of TG2). This assay was performed using wells precoated with 10μg/cm2 of Fn +/- various domains of TG2; 5x104 cells were allowed to settle on the matrix for 30min. The addition of the core (C) or βS domain proteins, either alone or together, did not increase the number of adherent 129T2-TG2+/+ or 129T2-TG2-/- MEFs (Figure 20a, b, &d). Maximal cell adhesion with the addition of these domains to the Fn matrix was ~75% of the total number of cells seeded, which is the same as that observed with seeding on Fn matrix alone. As a control, the full length TG2 protein was used and enhanced the number of adherent 129T2-TG2+/+ MEF adhesion from ~75% to ~90% and 129T2-TG2-/- MEFs from ~55% to ~75% of the total number of cells seeded, which is comparable to our previous findings with the full length protein (Figure 18). In contrast, when the first two domains of TG2 were expressed together as a single fusion protein (βC) and added exogenously to the assay, the number of adherent cells was enhanced to the same level observed with full length TG2 (Figure 20d).

4.2.2.5 TG2 enhances adhesion in the presence of Itg or Synd inhibitor but not both

Cellular adhesion on Fn requires crosstalk between members of the integrin receptor family

(α4β1, α5β1, β3) and of the syndecan co-receptor family (Synd4 in fibroblasts) [48]. Given the effect of the lack of TG2 on cell adhesion, it is of interest to inhibit the integrin or syndecan pathways through the use of competitive inhibitors and observe the resulting effect on MEF adhesion to Fn.

First, the integrin-signaling pathway was blocked using the competitive inhibitor GRGDTP peptides (RGD), which compete with Fn for integrin on the surface of MEFs (Figure 21). The addition of TG2 to Fn matrix, in the absence of RGD peptides, resulted in a dose- dependent increase in both 129T2-TG2+/+ and TG2-/- MEF adhesion to Fn, similar to previous observations. RGD peptide addition to TG2-Fn matrix, however, reduced cell adhesion (Figure 21a). With zero TG2 protein, the number of adherent 129T2-TG2+/+ and 129T2-TG2-/- cells was reduced by ~25 and 50%, respectively (p<0.05). Addition of TG2 (20μg to 80μg/cm2) to the Fn matrix increased the number of adherent 129T2-TG2+/+ and TG2-/- MEFs in the presence of RGD inhibitors. About 50% of the total number of 129T2- TG2+/+ MEFs were adherent in the presence of RGD peptides and when 20μg/cm2 of TG2 99 was added this increased to ~60%. Adhesion saturated at ~65% with the addition of ≥40μg/cm2 of TG2 per well. Likewise, the addition of TG2 was able to rescue RGD- inhibited adhesion of 129T2-TG2-/- MEFs. With these cells, ~25% of the total number of cells seeded were adherent in the presence of RGD and with the addition of 20μg/cm2 of TG2, cell adhesion increased to ~35%. With further increments of exogenously added TG2, cell adhesion increased to ~45% with 40μg/cm2 of TG2, and then a plateau was reached at ~60% of the total number of cells seeded with addition of 60μg/cm2. The non-specific GRADSP (RAD peptides) were used as a negative control and had no effect on cell adhesion relative to MEFs seeded on FN plates treated with TG2 addition (Figure 21a).

We next evaluated the effect of blocking the syndecan pathway on cell adhesion. The syndecan pathway was then blocked either alone or together with blockade of the integrin pathway using heparin (hep) and RGD peptides, respectively (Figure 21b). In the presence of 42μg/cm2 of hep, which competes with analogue of heparan sulphate carbohydrate chains on Synd4, there was no reduction in the number of adherent 129T2-TG2+/+ or TG2-/- MEFs relative to no heparin treatment and addition of increasing concentrations of TG2 to the Fn matrix resulted in similar increases in the number of adherent MEFs of both genotypes (Figure 21a, top left and right Tables). Addition of hep in conjunction with RGD peptides, however, resulted in markedly reduced adhesion. Whereas about 70% of the total number of 129T2-TG2+/+ MEFs adhered to Fn alone (with no additional TG2 in the matrix), in the presence of both inhibitors, only ~40% of the total number of 129T2-TG2+/+ cells were adherent. This compared to adhesion of ~50% of the total number of 129T2- TG2-/- MEFs on Fn being reduced to ~20% in the presence of both inhibitors. Unlike treatment with hep or RGD alone, in the presence of both hep and RGD, addition of TG2 to the Fn matrix was not able to enhance adhesion of MEFs of either genotype (Figure 21b, bottom Table).

4.2.3 Cell spreading assay

4.2.3.1 129T2-TG2-/- MEFs are delayed in cell spreading

Cell spreading continues from cell adhesion, as cells grip onto the substratum and begin to extend their cytoplasmic processes. Here, cell spreading was investigated by examining the area of individual phalloidin-stained cells. 129T2-TG2+/+ and 129T2-TG2-/- MEFs were 100 seeded onto Fn and the areas of phalloidin-stained cells were recorded at 30, 60, and 90 mins post-seeding. Areas were normalised to the mean area of 129T2-TG2+/+ cells at 30mins post-seeding (Figure 22a & b). The results for all cell spreading assay data are summarised in Table 5. At all time points there were significant differences in the extent of spreading of 129T2-TG2+/+ vs 129T2-TG2-/- MEFs. At 30 mins, the mean 129T2-TG2+/+ cell area was twice that of 129T2-TG2-/- MEFs. At this time, the cells of both genotypes were rounded and showed diffuse actin staining with no visible actin fibres. By 60 mins, the cell area of both genotypes increased, yet 129T2-TG2+/+ cells were significantly more spread by about 35% than 129T2-TG2-/-. The 129T2-TG2+/+ cell area was 1.2 times greater than the starting cell area and the 129T2-TG2-/-area increased by to 1.5 fold compared to its area at 30mins, and was equivalent to that of the 129T2-TG2+/+ cell area at 30 mins. Compared to 129T2-TG2-/- MEFs, 129T2-TG2+/+ MEFs displayed very thick actin fibres as indicated by the intense phalloidin-stained lines in the cytoplasm (Figure 22b, 60mins). At 90 mins, the cell areas of both genotypes were again increased; 129T2-TG2-/- MEF area was about 80% of that of 129T2-TG2+/+. The 129T2-TG2+/+ MEF cell area was 1.5 times greater than at 30 mins and 129T2-TG2-/- MEF cell area was 2.4 times greater than at 30 mins. Both genotypes were observed to have developed thick actin stress fibres across the cell body, and to have adopted a semistar-like fibroblastic shape (Figure 22b).

4.2.3.2 Exogenous addition of wildtype, GTP-binding- or transamidase-deficient TG2 proteins enhances cell spreading

Given a) that the 129T2-TG2-/- MEFs were defective in cell adhesion and spreading relative to wildtype cells, and b) that the cell adhesion defect could be rescued by incubation of MEFs with exogenously added wildtype TG2, the GTP-binding-deficient mutant TG2, R579A, or the transmidase-deficient mutant TG2, W241A, prior to seeding, the effect of exogenously added wildtype or mutant TG2 on cell spreading was investigated. As before, cell spreading was recorded at three time points: 30, 60, and 90 mins post-seeding (Figure 23). At 30 mins, 129T2-TG2-/- MEF cell area was about half that of 129T2-TG2+/+ MEFs; at 60mins 129T2-TG2-/- MEF cell area was about two-thirds that of 129T2-TG2+/+ MEFs; and by 90mins 129T2-TG2-/- MEF cell area was about 80% of that of 129T2-TG2+/+ MEFs. When 20μg/cm2 of wildtype, R579A mutant, or W241A mutant protein was pre-incubated with cells of either genotype, the 129T2-TG2-/- MEF cell area was equivalent to that of

101

TG2+/+ MEFs, whereas that of TG2+/+ MEF cell area increased relative to the untreated 129T2-TG2+/+ MEFs. Thirty minutes after pre-incubation of 129T2-TG2+/+ MEF with TG2, the extent of cell spreading was increased 1.5 fold relative to untreated 129T2-TG2+/+ MEFs. Preincubation of 129T2-TG2-/- MEFs with either of the mutant proteins increased the extent of cell spreading to that observed with wildtype TG2 preincubation. Sixty minutes after preincubation of 129T2-TG2+/+ MEFs with wildtype or mutant TG2 protein the extent of spreading increased above that observed 30 mins after incubation. Sixty minutes after pre-incubation of 129T2-TG2-/- MEFs with wildtype or mutant proteins, the extent of cell spreading increased to that of untreated 129T2-TG2+/+ MEFs. Finally, 90 mins after preincubation of 129T2-TG2+/+ MEFs with wildtype or mutant TG2 protein, the extent of cell spreading was similar to 60 mins preincubation, indicating that at this time cells were already maximally spread. Ninety minutes after preincubation of 129T2-TG2-/- MEFs with wildtype or mutant TG2 proteins the extent of cell spreading was equal to that of treated and untreated 129T TG2+/+ MEFs.

4.2.3.3 Transfection of wildtype, or GTP-binding- deficient or transamidase-deficient mutant TG2 cDNA enhances cell spreading

On Fn, 129T2-TG2+/+ MEFs were more spread than TG2-/- MEFs (Section 4.2.3.1). and exogenous addition of 129T2-TG2 enhanced cell spreading of both TG2+/+ and TG2-/- MEFs (Section 4.2.3.2). To investigate the effect of endogenous supplementation of TG2, wildtype and various mutant TG2 cDNAs were transfected into MEFs using the vector pIRES-EGFP, in which the protein of interest is cloned under the human cytomegalovirus (CMV) early promoter and upstream of an internal ribosome entry site to allow downstream expression of enhanced green fluorescent protein (EGFP). Because the EGFP molecule is expressed at a lower efficiency than the cloned protein under the CMV promoter, any cell that fluoresces at an exitation wavelength of 488nm should also have TG2 expressed. Transfectants were therefore identified as any cells that fluoresces. Transfection efficiency was initially estimated in MEFs transfected with TG2-pIRES-EGFP (Figure 24a & b). By fluorescence microscopy visualisation, a transfection efficiency of roughly 20-30% of the total number of cells transfected, was determined (Figure 24a). Moreover, using flow cytometry, fluorescence signals were first gated based on forward scatter and side scatter to filter out small cell debris or large cell clusters, then fluorescence

102 signals were gated based on fluorescence at 488nm excitation (n = 3). Transfection efficiency of TG2-/- MEFs was 32 ± 4.3%. Following transfection with cDNA encoding wildtype TG2, the GTP-binding-deficient mutants S171E and R579A, or the transamidase- deficient mutants C277S and W241A, live cells were sorted by FACS and immediately lysed for analysis of TG2 expression by Western blot analysis. This indicated that some mutants were less efficiently expressed than others (data not shown) Accordingly expression was equalized by varying the amount of plasmid DNA per well; the final amounts of plasmid DNA used were wildtype TG2: 10μg, S171E: 11.5μg, C277S: 11.5μg, W241A: 9.6μg, and R579A: 10μg. The cytoplasmic (Figure 24c) and membrane (Figure 24d) fractions of transfected and FACS-sorted MEFs were evaluated for TG2 expression. Cytoplasmic TG2 was normalised to GAPDH expression (about 9.5:1) and membrane TG2 was normalised to pan-cadherin expression (1.4:1). Using this information, 129T2-TG2-/- MEFs were transfected and subjected to a cell spreading assay. Negative controls included 129T2-TG2-/- MEFs transfected with empty pIRES-EGFP plamid, and untreated 129T2- TG2+/+ and 129T2-TG2-/- MEFs. Transfection of wildtype TG2 or any of the various mutant TG2 cDNAs into 129T2-TG2-/- MEFs increased the extent of cell spreading to that of 129T2-TG2+/+ MEFs at each time point examined. Mock-transfected cells spread to the same extent as untreated 129T2-TG2-/- MEFs indicating that neither the plasmid nor the transfection procedure affected cell spreading.

The results obtained thus far do not distinguish between an effect of TG2 intracellularly and/or extracellularly (on the cell surface and/or in the matrix) to rescue the delayed cell spreading of 129T2-TG2-/- MEFs. To investigate this issue in more detail, a TG2 mutant was used that has been reported to be externalised to the cell surface but not secreted into the media. To date, three such externalisation mutants of TG2 have been described: the GTP-binding defective K173L [48], the transamidation activity mutant C277S, and Y274A [133, 378], based on the ability to detect TG2 on the cell surface but not its secretion into the media of, transfected CH8 cells or Swiss 3T3 fibroblasts. Human Y274A TG2 has been reported to be transamidation competent [133] or transamidation defective [378]. In our hands, rat Y274A TG2 is transamidation active (Iismaa and Holman, unpublished results). The rat mutant Y274A was used in this study. 129T2-TG2-/- MEF were transfected with 10μg/well of pIRES-Y274A-EGFP and FACS sorted. The expression of TG2 in the cytoplasmic and membrane fractions was examined by Western blotting with wildtype TG2 103 cDNA transfection serving as a positive control and mock-transfected cells as a negative control (Figure 25a). The relative expression of Y274A TG2 in each fraction was similar to wildtype TG2, with the mock-transfected cells expressing no TG2. The Y274A TG2 cDNA transfectants were seeded onto Fn-coated plates and the extent of cell spreading was monitored, along with that of the controls: non-transfected 129T2-TG2+/+ or 129T2-TG2-/- MEFs. Consistent with findings shown in Figure 23, spreading of 129T2-TG2+/+ MEFs was greater than that of 129T2-TG2-/- MEFs at each time point and transfection of 129T2-TG2-/- MEFs with wildtype TG2 cDNA enhanced spreading to that observed for 129T2-TG2+/+ MEFs. Transfection of Y274A TG2 cDNA into 129T2-TG2-/- MEFs also enhanced the area of cell spreading to that of TG2+/+ MEFs at each time point. To check that Y274A TG2 was localised to the cell surface in these transfectants, cell surface TG2 was detected using a biotinylation pull-down assay (Figure 25c). Biotinylation of non-permeabilised cells, followed by cell lysis and streptavidin pull-down, detected TG2 on the surface of 129T2- TG2+/+ MEFs and on the surface of both wildtype TG2- and Y274A TG2-transfected 129T2-TG2-/- MEFs, but not on the surface of untransfected 129T2-TG2-/- MEFs (Figure 25c). There was no contamination of cytoplasmic proteins in the biotinylated membrane fraction as indicated by the lack of GAPDH detection in this fraction. The presence of cell surface TG2 was also confirmed with visualisation of surface-stained TG2 in Y274A- transfected 129T2-TG2-/- MEFs, but not untransfected MEFs, using anti-TG2 antibody. Individual MEFs were identified based on their nuclear (blue) and actin (red) staining, with only the Y274A TG2-transfected 129T2 TG2-/- MEFs staining positive (green) for cell surface TG2 (Figure 25d). Taken together these findings indicate that spreading of 129T2- TG2-/- MEFs is impaired relative to that of 129T2-TG2+/+ MEFs, but can be rescued with either extracellular addition of TG2 or intracellular transfection of TG2 cDNA (regardless of whether the TG2 is wildtype or defective for GTP-binding or transamidation activity). This rescue in spreading by TG2 could in part be due to TG2 aiding fibroblasts to attach quicker to the Fn-matrix as shown in Section 4.2.2. but TG2 may also affect downstream cell spreading pathway proteins.

104

4.2.4 Activated RhoA and Rac1 levels are different in 129T2-TG2-/- and TG2+/+ MEFs

Two crucial signaling molecules involved in cell adhesion and spreading processes are the small GTPases of the Rho family: RhoA and Rac1 [328]. In general, activation of these GTPases is thought to have opposing cellular actions on actin filaments: activated RhoA discourages actin turnover increasing static cell adhesion, whilst activated Rac1 encourages actin depolymerisaton allowing actin redistribution to the cell periphery for cell spreading. Because of the importance of these GTPases, their relative levels of activation were examined. Since TG2 rescued the delayed cell adhesion and spreading of 129T2-TG2-/- MEFs with a cDNA transfection or with exogenous addition of TG2, 129T2-TG2+/+ and TG2-/- MEFs as well as TG2 transfected 129T2-TG2-/- MEFs and 129T2-TG2-/- MEFs to which TG2 was added exogenously, were examined. The activated or GTP-bound forms of RhoA and Rac1 are quantitated by their respective binding to Rhotekin (which binds specifically to GTP-RhoA and not to GDP-RhoA) and Pak (which binds specifically to GTP-Rac1 and not GDP-Rac1) (Figure 26a & b). As a positive control for this assay, we used GTPγS treated lysates. GTPγS is slowly hydrolysable and was incubated with lysates to keep the GTPase population in the activated GTP-bound form.

These Rhotekin and Pak pulldown assays revealed that 129T2-TG2+/+ MEFs had different activated Rac1 and RhoA profiles at 10, 30, 60, and 90mins compared to TG2-/- MEFs. In terms of Rac1, 129T2-TG2+/+ MEFs had little activated Rac1 levels (0.02 relative to the total Rac1 expression level [REL]) at 10mins post-seeding. This increased to 0.1 REL by 30mins, peaked at about 0.25 REL at 60mins and dropped to 0.2 REL by 90mins. In contrast, 129T2-TG2-/- MEFs had relatively highly activated Rac1 levels at just below 0.1 REL, which peaked at 30mins (c.f. TG2+/+) at just below 0.3 REL. At 60mins post-seeding, activated Rac1 levels dropped in 129T2-TG2-/- to above 0.2 REL, at which time there was no statistically significant difference in activated Rac1 levels between the two genotypes. At 90mins, the REL of Rac1 in TG2-/- had dropped to 0.15 REL. Overall, the level of active Rac1 in both 129T2-TG2+/+ and TG2-/- MEFs followed a bell-shaped profile, except that the peaks occurred at different times: in 129T2-TG2+/+ MEFs activated Rac1 levels peaked at 60mins post-seeding, whereas in 129T2-TG2-/- the peak was at 30mins. With the transfection of TG2 cDNA into 129T2-TG2-/- MEFs, the activated Rac1 profile over time resembled that of 129T2-TG2+/+ cells. Activated Rac1 levels were low in MEFs (0.02 REL) 105 and increased over time, with maximal levels of activated Rac1 at 90mins and no statistically significantly difference between 129T2-TG2+/+ and transfected TG2-/- MEFs. With exogenous addition of TG2 to the 129T2-TG2-/- MEFs the profile of activated Rac1 levels was similar to the transfected 129T2-TG2-/- MEFs. Interestingly, at 90mins, 129T2- TG2-/- MEFs treated exgenously with TG2 had elevated activated Rac1 levels compared to 129T2-TG2+/+ MEFs.

Next, activated RhoA levels were examined (Figure 26c & d). The 129T2-TG2+/+ MEFs had 0.1 REL of activated RhoA at 10mins post-seeding and this increased to 0.38 REL at 30min, 0.6 REL at 60mins and 0.84 REL at 90mins. On the other hand, 129T2-TG2-/- MEFs had undetectable levels of activated RhoA at 10mins and even at 30mins the REL was relatively low at about 0.02. Thereafter, activated RhoA levels increased to 0.2 REL at 60mins and to 0.7 REL at 90mins. Both transfection of TG2 cDNA into and exogenous addition of TG2 to 129T2-TG2-/- MEFs restored the activated RhoA profile to that of wildtype cells. In conclusion, the findings here suggest that the lack of TG2 caused a delay of the cell spreading process in 129T2-TG2-/- MEFs as these cells were deficient in cell adhesion. Once TG2 was introduced to these TG2-/- cells, these were able to display wildtype cycling of RhoA and Rac1 expressions as well as wildtype cell spreading (Section 4.2.3.). These data suggest that the lack of TG2 resulted in delayed cell spreading on Fn as a consequence of delayed cell adhesion.

4.2.5 129T2-TG2-/- MEF monolayers re-establish slower from a scratch than 129T2- TG2+/+ MEF monolayers

Cell migration is the directional movement of cells towards a stimulus, and occurs following cell adhesion and spreading on the ECM, which in turn occurs upon cell settlement [344]. It was therefore of interest to determine if 129T2-TG2-/- MEFs, which are defective in adhesion and spreading, would also demonstrate a defect in cell migration on a Fn substrate. As an in vitro model of wounding 129T2-TG2+/+ or TG2-/- monolayers were scratched with a 200μl pipette tip and the migration of cells into the denuded area was recorded at 24 and 48hrs after scratching (Figure 27). The denuded area of 129T2-TG2+/+ MEF monolayers was 60% at 24hrs and 10% at 48hrs, whilst that of 129T2-TG2-/- MEF monolayers was 70% at 24hrs and 50% at 48hrs. Addition of TG2 immediately post-

106 scratching resulted in denuded areas that were reduced in size at both 24 and 48hr. Indeed, the denuded area of the 129T2-TG2+/+ MEF monolayer was reduced by more than half at 24hrs and was completely closed by 48hrs and in the case of the 129T2-TG2-/- MEFs, the scratch ‘wound’ in the monolayer was reduced by nearly half at 24hrs and to about 15% of the denuded area at 48hrs. At both 24 and 48hrs post-wounding, addition of TG2 increased the rate of closure of the scratch in the 129T2-TG2-/- MEF monolayers to that of the rate of closure of the scratch in the 129T2-TG2+/+ monolayers (p < 0.05 at 24 and 48 hours post- wounding for 129T2-TG2+/+ MEF; p < 0.05 at 24 hours and p < 0.001 at 48 hours post- wound for 129T2-TG2-/- MEF). In conclusion, in line with the cell adhesion and cell spreading results, 129T2-TG2-/- MEFs exhibited retarded migration on Fn matrix compared to 129T2-TG2+/+ MEFs, and the addition of exogenous TG2 was able to speed up MEF migration in both genotypes, and in the case of TG2-/- MEF, the migration rate was rescued to the rate of TG2+/+ MEF.

107

108

Figure 15 129T2-TG2 MEF growth curves

Random batches of 129T2-TG2+/+ and 129T2- TG2-/- MEFs (n = 3) were monitored for growth. rate at (a) P2, (b) P3, (c) P4, (d) P5, and (e) P6. Cell proliferation was drastically reduced with increasing passages. At P2, cells reached confluence on day 4 and for subsequent passages, up to P5, confluence was reached on day 5. Cells were split on the day they reached confluency. MEFs did not propagate beyond P6.

109

110

Figure 16 Adhesion of 129T2-TG2+/+ and 129T2-TG2-/- on Fn matrix

The adhesion assay was performed using 129T2-TG2+/+ (solid squares) and 129T2-TG2-/- (hollow squares) MEFs, varying (a) amount of Fn coated, (b) cell density per well, and (c) amount of time for adhesion. The adhesion assay was performed in a 24-well plate format. (a) Fifty thousand TG2+/+ or TG2-/- MEFs were seeded onto wells that were coated with 0 to 40μg of Fn and adhesion was monitored after 30min in triplicate (n = 3). (b) Increasing cell numbers were seeded onto 10μg of Fn for 30mins in triplicate (n = 3). (c) 129T2- TG2+/+ or 129T2-TG2-/- MEFs were allowed to settle in triplicate on Fn coated wells and adhesion was measured at time points ranging from 10 to 120 mins (n = 4). (d) Representative microscopic photographs of fixed crystal violet-stained 129T2-TG2+/+ and 129T2-TG2-/- MEFs dhered to Fn after early (10mins, 20mins) and late (120mins) time points. Magnification is at 10x objective, scale bar = 30μm. ** is p<0.01 and *** is p <0.005 for a two way ANOVA analysis with post-hoc Bonferroni correction.

111

112

Figure 17 Optimisation of TG2 binding to Fn-coated plates

To optimise conditions for maximal saturation binding of TG2 to Fn coated-plates, varying amounts of TG2 (0 to 100μg per well) were incubated either overnight at 4oC (solid circles) or for 30mins at 37oC (hollow circles) in triplicate (n = 2). TG2 binding to uncoated tissue culture plastic (TCP; solid triangles) and binding to BSA-blocked TCP with no Fn coating (squares) were also observed to test for non-specific binding. A negative control of no TG2 (crosses) was included to take into account background from TCP.

113

114

Figure 18 The effect of exogenously added TG2 on adhesion of 129T2-TG2+/+ or 129T2-TG2-/- MEFs to Fn

To examine the effect of TG2 on MEF adhesion, 129T2-TG2+/+ and 129T2-TG2-/- MEFs were allowed to settle for 20 mins on TG2-Fn matrix prepared with increasing amounts of exogenously added TG2 (0 to 80μg per well, in triplicate, n = 4). ** denotes p<0.01 and *** denotes p <0.001 for a two way ANOVA analysis with post-hoc Bonferroni correction.

115

116

Figure 19 Cell adhesion assay with the transamidation-deficient mutant W241A TG2 or GTP- binding-deficient mutant deficient R579A TG2

Cell adhesion assay was performed with exogenous addition of W241A (a) or R579A (b) to the Fn matrix. 129T2-TG2+/+ (solid squares) or 129T2-TG2-/- (hollow squares) MEFs were allowed to settle on Fn coated wells in the presence of varying amounts of mutant TG2 protein (0 to 40μg per 5x104 cells) and adhesion was measured at 30mins (in triplicate, n = 4). *** denotes p <0.001 for a two way ANOVA analysis with post-hoc Bonferroni correction.

117

118

Figure 20 Cell adhesion assay with exogenous addition of various TG2 C-terminal truncations of TG2 to the Fn matrix

129T2-TG2+/+ and TG2-/- MEFs were allowed to settle for 30 mins on Fn precoated with various N-terminal domains of TG2. The domains added were (a) β-sandwich (βS); (b) core (C); and (c) β-sandwich-core (βC). 129T2-TG2+/+ (solid bars) and 129T2-TG2-/- MEF (hollow bars) cell adhesion was expressed as % of total number of cells seeded. # denotes p<0.001 for a two tailed Student t-test compared with MEFs seeded on Fn only.

119

120

Figure 21 Cell adhesion assay in the presence of RGD and/or heparin (hep)

Adhesion of 129T2-TG2 MEFs to Fn matrix in the presence of RGD, hep, or both RGD and hep. MEF adhesion was quantitated as % of total number of cells seeded. (a) To block integrin-mediated cell adhesion, a process that is propagated through the interaction of integrins with the RGD motif on Fn, competitive RGD,or RAD as a control, (84μg/well) peptides were added to MEFs as they were allowed to settle on the Fn matrix. In some wells, various concentrations of TG2 (20, 40, 60, 80μg/well) were bound to the Fn prior to cell seeding. The table on the left shows statistical analysis of 129T2-TG2+/+ or 129T2-TG2-/- cell adhesion in the presence of exogenous TG2 compared to cell adhesion at 0μg of exogenous TG2 and 129T2-TG2+/+ compared to 129T2-TG2-/- designated as “B/W groups”. The table on the right displays statistical analysis of 129T2-TG2+/+ and 129T2-TG2-/- MEF adhesion on Fn +/- TG2 in the presence of RGD peptides. (b) Synd4- m\ediated cell adhesion was blocked with addition of heparin (42μg/well) to Fn matrix or both integrin- and synd4-mediated adhesion was blocked with both RGD (84μg/well) and heparin (42μg/well) to Fn matrix as MEFs were seeded. The top left table in (b) displays statistical analysis of 129T2-TG2+/+ and TG2-/- MEFs on Fn with the indicated amount of exogenous TG2. The top right Table displays statistical analysis of 129T2- TG2+/+ and TG2-/- MEFs on Fn in the presence of heparin. The bottom Table displays statistical analysis of TG2+/+ and TG2-/- MEFs on Fn in the presence of heparin and RGD peptides. Statistical analysis was done using two way ANOVA with Bonferroni post-hoc test, * denotes p<0.05; ** denotes p<0.01; *** denotes p<0.001; “ns” denotes no significant difference.

121

122

Figure 22 Measurement of the area of phalloidin-stained 129T2-TG2+/+ or 129T2-TG2-/- MEFs adhered on fibronectin

Extent of MEF cell spreading on Fn was measured by evaluating the stained actin area (n = 5 experiments, each with 10 random single cells). (a) The mean area (± SEM) of 129T2-TG2+/+MEFs (solid columns) and 129T2-TG2-/- MEFs (hollow columns) was measured at 30, 60, and 90 mins. Cell areas were normalised to mean 129T2-TG2+/+ cell area at 30mins post-seeding. (b) Representative phalloidin-stained MEFs at each time point are shown. Scale bar = 10μm. # denotes p<0.001 for a two-tailed Student t-test comparing 129T2- TG2+/+ with 129T2-TG2-/- area at a given time point. ‡ denotes p<0.001 for a two-tailed Student t-test comparing cell area of that genotype to the respective area at 30mins.

123

124

Figure 23 Cell spreading of 129T2-TG2+/+ and 129T2-TG2-/- MEFs preincubated with wildtype, GTP- binding-deficient R579A or transamidase-deficient W241A TG2

Extent of MEF cell spreading was measured at 30, 60, and 90 mins by evaluating the stained actin area (± SEM) (n = 5 experiments of 10 random single cells) of 129T2-TG2+/+MEFs (solid columns) and 129T2-TG2-/- MEFs (hollow columns) with or without preincubation using wildtype TG2, GTP-binding deficient mutant R579A, or transamidase-deficient mutant W241A, as indicated. Cell areas were normalised to the mean of 129T2-TG2+/+ at 30mins post-seeding. # denotes p<0.001 for a two-tailed Student’s t-test comparing the cell area of MEFs without TG2 protein preincubation vs that of MEFs with TG2 protein preincubation..

125

126

Figure 24 Cell spreading assay with 129T2-TG2-/- MEFs transfected with wildtype, GTP-binding- deficient or transamidase-deficient mutant TG2 cDNAs

The extent of cell spreading on Fn of 129T2-TG2-/- MEFs following transfection with wildtype or various mutant TG2 cDNAs or empty vector (mock), was measured by evaluating the stained actin area (5 experiments of 10 random single cells). Mutants included transamidase-deficient mutants W241A and C277S, and GTP-binding-deficient mutants S171E and R579A. Transfection efficiency was evaluated following transfection by examination of EGFP fluorescence under a fluorescence microscope. (a) A representative micrograph of fluorescent cells (488nm ex.) from a confluent population. (b) During flow analysis (n = 3) the initial cell population was analysed based on forward scatter and side scatter to identify single live cell populations (left panel, circled). This sub-population was analysed for GFP negative cells (middle panel, 78.2% ± 2.6% of the sub-population) and GFP positive cells (right panel, 32% ± 4.3%). FACS-sorted transfectants were lysed and (c) cytosolic TG2 and (d) membrane TG2 was quantitated (mean ± SEM) relative to GAPDH and pan-cadherin respectively using Western blots (n = 3). The mean area (± SEM) of 129T2-TG2+/+MEFs (solid bars) and TG2-/- MEFs (hollow bars) and of transfected 129T2-TG2-/- MEFs (hollow bar with labels) was determined at 30, 60, and 90 mins. Cell areas were normalised to the cell area of 129T2-TG2+/+ MEFs at 30mins post-seeding. # denotes p<0.001 for a two-tailed Student t-test comparing the area between transfected MEF vs untreated MEFs at a given time point.

127

128

Figure 25 Transfection of 129T2-TG2-/- MEFs with cDNA-encoding Y274A TG2

Cell spreading area on Fn of 129T2-TG2-/- MEFs transfected with cDNAs encoding wildtype or Y274A mutant TG2 was measured by evaluating the stained actin area (5 experiments of 10 random single cells). FACS- sorted transfectants were lysed and cytosolic TG2 and membrane TG2 were quantitated (mean ± SEM) relative to GAPDH and pan-cadherin respectively using Western blots (n = 3). The mean area (± SEM) of 129T2-TG2+/+MEFs (solid bars) and 129T2-TG2-/- MEFs (hollow bars) and transfected 129T2-TG2-/- MEFs (hollow bar with mutant cDNA mutants) was measured at 30, 60, and 90mins (b). The cell areas were normalised to the mean TG2+/+ cell area at 30mins post-seeding. # denotes p<0.001 for a one-way ANOVA with Bonferroni correction comparing the area between transfected MEFs to untreated MEFs at a given time point. MEFs used in the cell spreading assay from (b) were subjected to cell surface biotinylation (+bt) and biotin was pulled-down using streptavidin beads (n = 3), a representative blot is shown (c). Western blotting for TG2 was performed on the flow through and the pulled down cell surface proteins, using GAPDH and pan-cadherin as controls, respectively. A negative control of 129T2-TG2+/+ MEFs without biotinylation (-bt) was included. Immunostaining for cell surface TG2 was performed to check for expression of Y274A (n = 2). (d) 129T2-TG2-/- MEFs and Y274A-transfected 129T2-TG2-/- MEFs were stained at 30 mins post-seeding for the nucleus (top left panel), actin (bottom left panel) and TG2 (top right panel), and the merged images of the triple stained cells is shown in the bottom right panel.

129

Table 4 Cell adhesion assay summary

Parameter TG2+/+ TG2-/- ++++ ++ +TG2 ++++++ ++++ +Crosslinking mutant ++++++ ++++ +GTP-binding mutant ++++++ ++++ +RGD ++ + +RGD +TG2 +++ ++ +RGD +Heparin + (+) +RGD +Heparin +TG2 + (+)

Table 5 Cell spreading assay summary

Parameter TG2+/+ TG2-/- ++++ ++ +TG2 ++++++ ++++ +Crosslinking mutant ++++++ ++++ +GTP-binding mutant ++++++ ++++ +Y274A ++++++ ++++ TG2 transfection ++++++ ++++ Crosslinking mutant transfection ++++++ ++++ GTP-binding mutant transfection ++++++ ++++ Y274A transfection ++++++ ++++

130

131

Figure 26 Activated Rac1 and RhoA levels in 129T2-TG2+/+, 129T2-TG2-/- MEFs preincubated with TG2, and TG2-transfected 129T2-TG2-/- MEFs

129T2-TG2+/+ MEFs (solid bars), 129T2-TG2-/- MEFs (hollow bars), 129T2-TG2-/- MEFs transfected with TG2 cDNA (dotted bars) or 129T2-TG2-/- MEFs with TG2 exogenously added (striped bars) were seeded onto Fn for 10, 30, 60, and 90mins, and were harvested to probe for activated GTP-bound (a,b) Rac1 and (c,d) RhoA (n = 3). Mean active GTP-bound (a) Rac1 or (c) RhoA levels (± SEM) were quantitated densiometrically as a fraction of the total small GTPase population. Representative Western blots of (b) activated Rac1 and total Rac1 or (d) activated RhoA and total RhoA from these MEFs are shown, with GTPγS-incubated 129T2-TG2+/+ MEF lysate serving as positive control (+ve). # denotes p<0.001 for two- tailed Student’s t-test comparing levels of activated GTPases in TG2-/- versus TG2+/+ MEFs at the respective time points.

132

133

Figure 27 In vitro scratch wounds of 129T2- TG2+/+ or TG2-/- MEF monolayer

129T2-TG2+/+or TG2-/- MEFs were scratched using a 200μl pipette tip and were monitored at 0, 24 and 48hrs post-scratching (n = 3). (a) Representative micrographs shown at 0 and 48h. (b) The area of the scratch was expressed as a fraction of day 0 area. 129T2-TG2+/+ (solid columns), 129T2-TG2-/- (hollow columns), 129T2- TG2+/+ or 129T2-TG2-/- to which 500μg/ml TG2 was added exogenously immediately post-wounding (dotted columns). * denotes p<0.05 and *** denotes p<0.001 for a two-tailed Student’s t-test comparing between MEFs of the same genotype with or without addition of TG2.

134

135

Figure 28 The role of TG2 in wound healing

A schematic cartoon showing the migration of a wildtype fibroblast from wound edge to the wounded area, highlighting fibroblast cell surface receptor interactions with the ECM that involve TG2 (bottom dotted box). From the results thus far, a model emerges where TG2 expression in wildtype fibroblasts leads to efficient cell adhesion, spreading and migration. Such dynamics are brought about by interaction of Fn, a predominant ECM protein in the wound bed, with cell surface receptors Itg (yellow and red heterodimers) isotype β1 or β3 and/or Synd4 (orange receptor marked “Synd”) and these interactions are augmented by the presence of TG2. In the wound environment, tissue degradation products resulting from injury and inflammation, some of which contain RGD motif will competitively inhibit Itg outside-in engagement. As proposed by Telci et al. (2008), Synd4 engagement allows for an alternative cell adhesion pathway that would permit cell survival on nearby ECM [85]. The results from cell dynamics experiments (Section 4.2), as well as those from in vivo wound healing experiments (Section 3.2) support a role for TG2 as a mediator in wound healing, enabling efficient fibroblast migration to the wound bed laden with wound ECM setting a foundation for subsequent physiological events in healing.

136

4.3 Discussion

To test if the defective wound healing observed in vivo in 129T-TG2-/- mice is attributable to altered fibroblast dynamics, MEFs were isolated from 129T2-TG2+/+ or 129T2-TG2-/- animals and studied in vitro for adhesion, spreading, and migration, all of which are cellular processes central to the fibroblast’s role in physiological wound healing. The processes were assayed by seeding MEFs on tissue culture plastic pre-coated with plasma Fn (from bovine sources). Non-specific binding was initially blocked with BSA. Thereafter, MEFs were allowed to adhere on the substratum (cell adhesion or spreading assays) or the confluent monolayer was mechanically injured (cell migration assay). These assays represent a two-dimensional cell model [379]. There are advantages and disadvantages to utilising such a model. The major advantage of this model is that it allows a few individual chosen components to be directly examined without interference from other heterogenous tissue component(s). However, the lack of complexity also limits extrapolation of the findings to the in vivo environment. Given that the focus of these studies was on the interaction between fibroblasts and Fn, the two-dimensional single cell system used here was thought to be appropriate. A criticism of this model may be the use of plasma Fn which might seem inappropriate as it is different from cellular Fn in nature – based on its electrophoresis migration rate, plasma Fn is smaller in size than cellular Fn [48] and functionally, plasma fibronectin is both 50 times less active in the restoration of transformed fibroblast cell lines as well as 150 times less active in agglutinating formalin- fixed sheep erythrocytes [380]. However, plasma Fn remains highly relevant in wound healing as it becomes incorporated into blood clots during coagulation [381] and plasma Fn has also been shown to incorporate into tissue ECM in vivo [382, 383]. For these reasons, plasma Fn was used to evaluate fibroblast dynamics during wound healing.

As shown by the cell-based assays (Section 4.2) 129T2-TG2-/- MEFs were defective in cell adhesion, and delayed in spreading and migration relative to 129T2-TG2+/+ MEFs. The ability of TG2 to augment cell adhesion is well described in the literature. Improved adhesion and/or enhanced signaling through the adhesion pathway, with the exogenous addition of TG2, has been demonstrated in several fibroblast cell lines including BalbC 3T3 [48], Swiss 3T3 [48], NIH 3T3 [48], and WI-38 cells [48]. The cells used in the current study were primary cells isolated from 129T2-TG2+/+ or 129T2-TG2-/- animals, cells that 137 more closely resemble in vivo cells than do cell lines. Indeed, in one microarray study, a comparison of mouse primary hepatocytes to an immortalised hepatoma Hepa1-6 cell line [384] found that Hepa1-6 cells had lost or downregulated many liver-specific genes compared to primary hepatocytes; and many signaling components of proliferative pathways were upregulated in Hepa1-6 cells [384]. Given that equal numbers of cells of both genotypes were seeded, our studies show that at non-saturating Fn concentrations, fewer 129T2-TG2-/- MEFs adhered than 129T2-TG2+/+ MEFs (Figure 15c). These findings here differ from those of Scarpellini and colleagues [369] who saw no difference in cell adhesion [88], although they used primary MEFs isolated from C57Bl/6 rather than 129T2 mice.

The inability of 129T2-TG2-/- MEFs to eventually ‘catch up’ to 129T2-TG2+/+ MEFs in terms of the total number of adherent cells over time, suggests that unattached cells had died. A small proportion of the 129T2-TG2+/+ MEF population also died as less than 100% of these cells were adherent. Cell death due to inadequate or inappropriate cell-matrix contacts is known as anoikis [385]. Survival signaling is mediated by β1 or β3 activation upon ECM binding through FAK, Shc and/or integrin-linked kinase [48]. Cell adhesion and presumably adhesion-mediated cell survival, was increased in both genotypes in a dose- dependent manner by the exogenous addition of TG2, until saturation binding of TG2 to the Fn matrix was reached. At this point, cell adhesion plateaued (Figure 18). Thus, it would appear the ability of TG2 to enhance adhesion is limited. Interestingly, in a study by Verderio et al (2003) of either human osteoblasts or Swiss 3T3 fibroblast, TG2 addition did not increase cell adhesion on Fn or on a 42kD fragment of Fn in either [120]. This could be due to the smaller assay size (96-well plates, 5μg/ml Fn) where the amount of Fn may be saturating or the cell line used may display different properties to the MEFs used in the present experiments.

TG2 has been reported to prevent RGD-mediated anoikis [85, 120], through interactions with the gelatin binding domain of Fn [83] and with HSPG of syndecan 4 receptors [85]. The effect of TG2 on enhancing cell adhesion to the matrix is independent of the cross- linking or GTP-binding activity of TG2, rather it is due to its role as a Fn adhesion coreceptor for integrins. In the context of cell adhesion, it is evident that cross-linking activity is not essential as TG2 becomes catalytically inactive when in a complex with Fn [10], and 138 the transfection of wildtype [48], or the catalytically-inactive C277S TG2 mutant [48], increases cell adhesion to levels comparable to that of wildtype TG2 in a fibroblast cell line that is deficient in TG2. There is also evidence that extracellular GTP-binding and/or GTPase activity is not required for cell adhesion [44, 72]. In line with these reports, exogenous addition of either wildtype TG2, the cross-linking deficient mutant W241A [48] or the GTP-binding-deficient mutant R579A [48], to our cell adhesion assay enhanced both 129T2-TG2+/+ and TG2-/- MEF adhesion onto Fn. Since the adaptor function of TG2 is important for cell adhesion, TG2 must be in a conformation that can contact both integrins and Fn. Despite the evidence of TG2 functioning mainly as an adaptor molecule in the open conformation in cell adhesion and related processes, there are studies that show cross- linking to be required as well for this. A study by Verderio et al (2003) suggested that TG2 binding to Fn requires the presence of calcium, presumably to drive the protein into the open conformation, and the addition of GTPγS, which induces the closed conformation of TG2, compromised cell adhesion [120]. In support of this, other studies showed that TG2 cross-linking in the ECM appears to be required for proper cell adhesion in endothelial [71] and fibroblast [10] cell lines. However, such cross-linking activity appear to be short-lived since in a small intestinal injury model secreted TG2 was observed to be transiently activated under injury conditions but was rapidly inactivated [72] due to disulphide bond formation between C370 and C371 locking TG2 into an open conformation that is unable to participte in another round of catalysis [55]. Thus, in addition to TG2 having a role in cell adhesion, spreading and migration as a result of its adaptor function, it may also contribute to these processes by stabilizing the matrix through its crosslinking activity, as shown by in vitro experiments [232, 386]. Such a crosslinking-mediated effect was not evident in this study, as the use of cross-linking mutants in adhesion and spreading assays yielded results that were no different from those observed with the wildtype TG2 protein. This indicates that at least under the current in vitro assay conditions cross-linking might not be a critical factor. In future studies, a potential crosslinking effect could be probed by using a removable platform in the tissue culture plates to create an artificial wound rather than scratch wounding used here [387]. This system is the same as growing cells on tissue culture plastic, except that a temporary platform usually made of poly(dimethylsiloxane) is used that can be conveniently detached, hence creating a scratch like void. In this system, by completely removing the old substratum containing any pre-secreted/processed ECM

139 material, cell migration is now dependent on newly synthesised ECM. This way, any newly catalysed TG2 crosslinks resulting from the wounding procedure, could be studied.

As mentioned, TG2 binds to the 42kD gelatin binding N-terminal domain of Fn [83]. Adhesion to this site alone without the rest of the Fn molecule does support adhesion with focal adhesion formation and limited cell spreading in WI 38 fibroblasts and human monocytes [81, 83, 118]. However, in Swiss 3T3 albino or human osteoblast cell lines, it was shown that stable cell adhesion was only evident with full-length Fn molecules, which supports TG2 binding interaction with integrins and/or other cryptic receptors [120]. The Fn recognition motif is located in a heparin structure within the latter half of the β- sandwich domain [360]. In contrast, although it has been shown that TG2 can interact with integrin heterodimers α4β1 α9β1 and α5β1 [388-390], the site(s) on TG2 that interact with these heterodimers is still not well characterised. TG2 is known to to interact with β1, β3 or β5 integrin [81, 83, 118, 361] at a 1:1 stoichiometry [83]. The present studies have located this integrin-binding site(s) to the β-sandwich-core domain of TG2. Thus, like full-length TG2, exogenous addition of a truncated form of its β-sandwich and core domains, but lacking the barrel domains, enhanced MEF adhesion on Fn. Although, the β-sandwich- core protein rescued MEF adhesion, the inability of either domain alone to rescue cell adhesion (Figure 20 a-c), suggests that the domains must be linked to produce the active conformation required to promote adhesion, likely because in the absence of such linkage they cannot individually associate with the ECM. Given that the truncated β-sandwich-core TG2 protein cannot bind GTP as it lacks crucial GTP-binding residues within the C- terminal barrel domains [391], this is also further confirmation that involvement of TG2 in cell adhesion is independent of its GTP-binding activity. To further elucidate the structural interaction between TG2 and integrins, one could enlist the help of both experimental and/or computational methods. Examples of experimental methods include crystallography; electron microscopy; surface plasmon resonance, and mutation studies, whereas examples of computational methods include protein surface determinant modeling and molecular dynamics studies.

One of the disadvantages of the adhesion assay used here is that it does not provide ‘readout’ of the regulation on TG2 by the in vivo micro-environment of the ECM. For example in a small intestinal injury model, it was shown that Ca2+ was released as a burst into the 140

ECM [72]. This resulted in TG2 adopting its open conformation, which temporarily activated its TG2 cross-linking activity. Thus, in addition to TG2 having a role in cell adhesion, spreading and migration as a result of its adaptor function, it may also contribute to these processes by stabilizing the matrix through its cross-linking activity, as shown by in vitro experiments [232, 386]. Such a cross-linking-mediated effect was not investigated in this study, but the use of cross-linking mutants in adhesion and spreading assays yielded results that were no different from those observed with the wildtype TG2 protein. This indicated that at least under the current in vitro conditions cross-linking might not be relevant in the assays. In future studies, a potential cross-linking effect could be probed by using a removable platform in the tissue culture plates to create an artificial wound rather than scratch wounding used here [387]. This system is the same as growing cells on tissue culture plastic, except that a temporary platform usually made of poly(dimethylsiloxane) is used that can be conveniently detached, hence creating a scratch like void. In this system, by completely removing the old substratum containing any pre-secreted/processed ECM material, cell migration is now dependent on newly synthesised ECM. This way, any newly catalysed TG2 cross-links resulting from the wounding procedure, could be studied.

Wound fluid collected from the injury site contains a plethora of soluble factors, which are elaborated as a result of degradation of the ECM within the wound [170]. Two such TG2- interacting molecules are syndecans which shed heparan sulphate [392] and ECM proteins (e.g vitronectin, collagen, and fibronectin), which release RGD peptides [393]. Synd4 is the only syndecan that participates in focal adhesion complexes [392]. Thus, fibroblast adhesion and spreading on Fn does not evoke the development of focal adhesions or thick actin stress fibres unless syndecans engage the heparin binding site on Fn [394]. Interestingly, during a twelve-day skin wound healing assay, synd4-null animals showed retarded skin wound closure on three to six days post-injury [363]. During wounding, thrombin and EGF induce the shedding of the synd4 protein ectodomain (containing heparan sulphate chains). This is thought to promote the protease activity of elastase and cathepsin G, resulting in breakdown of damaged tissues [395]. RGD-peptides are another type of molecule released within the wound as local ECM proteins become degraded [396, 397]. The release of these peptides seems to be a ‘two-edged sword’: whilst these peptides are required for matrix remodeling and have biological actions such as arteriolar vasodilation [398], they also interfere with cell adhesion by acting as competitive inhibitors 141 for integrin binding sites [396]. TG2 interacts with heparan sulphate chains on synd4 with high affinity [399], and, thus, may provide an alternate adhesion pathway in the event of integrin binding being blocked by released RGD peptides [85]. Consistent with the study of Telci et al (2008) [85], inhibition of integrin binding by competitive RGD peptides reduced 129T2-TG2+/+ and TG2-/- MEF adhesion on Fn, and adhesion was rescued by exogenous addition of TG2 protein. Also consistent was the observation that competitive heparin blocking of Synd4 had no effect on cell adhesion of either MEF genotype. Intriguingly, extending from the work of Telci et al. (2008) [85], blocking both integrin and syndecan drastically reduced adhesion of MEFs of both genotypes but this adhesion was not rescuable in either genotype by exogenous addition of TG2 at the range used in the study. Moreover, the adhesion of 129T2-TG2-/- MEFs was reduced relative to TG2+/+ MEFs in the presence of both inhibitors. These observations suggest that there is an inherent TG2 effect that is independent of its interaction with integrins and syndecans. This novel interaction of TG2 with an undetermined protein target(s) warrants further investigation.

While cell adhesion was quantified by examining MEFs as a population, cell spreading, a process that follows intimately from adhesion, was examined by looking at individual cells was next examined. Decreased cell spreading area as effected by the downregulation of TG2 has been observed in HCA2 fibroblast [400], ECV304 endothelial cells [71], Swiss 3T3 fibroblasts [10], WI38 fibroblasts [83], rat embryonic fibroblast [83] and human erytholeukemia cells [83]. In keeping with these observations the extent of cell spreading of our primary 129T2-TG2-/- MEFs was reduced compared to 129T2-TG2+/+ MEFs. Nor was this difference due to 129T2-TG2-/- MEFs being smaller, since the volume of the MEFs did not differ between the genotypes (data not shown). In agreement with the cell adhesion data, cell spreading was enhanced with exogenous addition or transfection of wildtype TG2 or various TG2 transamidase- or GTP-binding deficient point-mutants. In the transfection studies, an additional GTP-binding deficient mutant, S171E, and cross-linking deficient mutant, C277S, were also used, and these had the same effect as wildtype TG2. Although it is likely that the effects of TG2 on cell adhesion, speading and migration are due to an extracellular action, we cannot exclude the possibility that it is secreted extracellularly after cDNA transfection, or translocates intracellularly across the plasma membrane after exogenous addition. TG2 is constitutively exported with integrin β1 in recycling endosomes and becomes secreted into the ECM [20, 83, 401, 402], but TG2 is only known 142 to internalise through LRP-1, which results in its lysosomal degradation [370]. Therefore enhanced cell spreading from the exogenously added TG2 pool should only be a direct exogenous effect of the added protein, as any internalised protein will be targeted for degradation. Conversely, since TG2 cDNA transfection enhanced cell spreading to the same degree observed with TG2 exogenous addition the effect of TG2 transfection is likely the result of TG2 export from the cell.

Based on the above data we propose a model (Figure 28) whereby increased amounts of cell surface wildtype TG2, transamidase-, or GTP-binding-deficient TG2, either through exogenous addition or cDNA transfection prepares the cell for effective adhesion onto the Fn substratum through interactions with β1/3 integrins [83] and/or interaction with heparan sulphate chains on syndecan 4 [85, 369]. To examine the downstream signaling from these receptors in adhering 129T2-TG2+/+ and 129T2-TG2-/- MEFs, active levels of the common downstream effectors RhoA and Rac1 were measured [362]. The levels of activated RhoA and Rac1 were different in 129T2-TG2+/+ relative to 129T2-TG2-/- MEFs but upon TG2 addition to or cDNA transfection into 129T2-TG2-/- MEFs activated RhoA and Rac1 levels were restored to those of 129T2-TG2+/+ MEFs. The activated RhoA levels in 129T2-TG2-/- MEFs were low during early adhesion at 10 and 30mins post-seeding relative to TG2+/+ MEFs. This indicates that formation of mature or strong focal adhesions, and hence cell adhesion, is lacking in 129T2-TG2-/- MEFs early on. The limited RhoA activity early in the adhesion process of 129T2-TG2-/- MEFs is counterbalanced by increased activity of the antagonistic Rac1 [328], relative to 129T2-TG2+/+ MEFs. Cell spreading was reduced in 129T2-TG2-/- MEFs, yet GTP-Rac1 expression was up-regulated compared to the wildtype counterpart. These data may seem to contradict with the dogma that active Rac1 leads to cell protrusions [403]. However, Rac1 is not the only small GTPase that influence cell spreading, as RhoA has also been reported in NIH3T3 fibroblast to interact with molecules downstream to Rac1, such as mDia, and contribute to actin polymerization and cell protrusions [404]. Indeed, increased active RhoA expression was observed in 129T2- TG2+/+ MEFs (Figure 26 c & d). In addition, Rac1 cycling between off and on states is important for cell dynamics [326], it is plausible that pre-30 minutes 129T2-TG2+/+ MEFs had higher GTP-Rac1 expression, this certainly warrants further investigation. As activated RhoA increased in 129T2-TG2-/- MEFs activated Rac1 levels fell. In keeping with our findings, another cell spreading study using TG2 anti-sense technology in HCA2 143 fibroblasts suggested that Src activity in the FAK/Src-Cas/Crk pathway may be dysregulated in the absence of TG2 [400]. Since Rho GTPases are downstream effectors of this pathway, TG2-induced restoration of activated RhoA and Rac1 levels in 129T2-TG2-/- MEFs to those of 129T2-TG2+/+ would suggest that the upstream FAK pathway, which is induced by integrin engagement with the matrix is functioning normally in the restored cells.

The work from this chaper has demonstrated that 129T2-TG2-/- MEFs are defective in cell adhesion and delayed in cell spreading. The addition of TG2 can enhance both these processes and rescue 129T2-TG2-/- MEFs to wildtype levels. A question being addressed here is whether TG2 protein in the extracellular environment can promote cell migration and hence explain the in vivo wound healing phenotype in the 129T2-TG2-/- mice (Chapter 3). The release of TG2 to the extracellular environment under artificial injury models has been shown in several other in vitro studies. One of the earlier studies was done by Upchurch et al (1991), where puncture to a fibroblast monolayer triggered the release of endogenous sources of TG2 [405]. In a UV damage model, TG2 was secreted to the ECM upon injury, and was shown to stabilise the Fn ECM, accompanied with its reduced turnover [232]. This was also observed in a renal scarring model [406]. Moreover, cells treated with the anti-TG2 antibody CUB7402 led to a reduction in cell adhesion of various cell types [10, 71, 128]. In keeping with these studies, the closure rate of a scratch wound was slower in our 129T2-TG2-/- MEF monolayers compared to TG2+/+ MEF monolayers. This phenotype is also in line with our cell adhesion and spreading data, and correlates positively with our in vivo wounding results. Some in vitro wound assay experiments by others had shown that TG2 cross-linking may be relevant in wound closure, as transamidation was detected in scratch wounds of cultured cell lines [72] and TG2 does indeed catalyse the cross-linking of coated Fn in culture [10]. The possibility of TG2 cross- linking the matrix and aiding in cell migration cannot be ruled out from the current work as any formed Nε(γ-glutamyl)lysine isopeptide bonds were not investigated in the Fn-coated tissue culture dishes and neither was the effect of exogenous cross-linking TG2 mutant(s) on cell migration examined. However, the addition of TG2 cross-linking mutants in both adhesion (Section 4.2.2.3) and spreading assays (Section 4.2.3.2) yielded phenotypes that were no different to MEFs treated with wildtype TG2. Supporting this, the overexpression

144 of the cross-linking mutant C277S TG2 in HCA2 fibroblasts resulted in migration rates that were comparable to those of wildtype fibroblasts [400].

A study by Balklava et al (2002) [378] showed increased TG2 in the matrix led to increased cell adhesion but reduced cell migration. From our data, exogenous addition of TG2 aided closure of the monolayer wound in both genotypes and remarkably in 129T2-TG2+/+ it achieved near full wound closure in 48hrs. The discrepancy in migration could be due to the fact that an agarose outgrowth model was used rather than a scratch assay in the work of Balklava and colleagues. Such a model represents a three-dimensional migration model that poses different cell dynamics. A striking example would be that canonical FAK signaling, in 2D matrices, upon adhesion in a 3D matrix may not be necessary for downstream MAPK pathway activation [407]. Another point of note is that the isometric force generated between rigid 2D matrices is different to soft 3D matrices in which the matrix itself can become stressed when adhering cells pull onto such matrices [408]. A relevant example is that fibroblasts becomes quiescent when seeded onto an unstressed 3D matrix but proliferate when they adhere to a stressed one [409]. So the scratch assay presented here cannot be directly compared to the results observed by Balklava et al. Along the same vein, the use of a 2D matrix in this study, although revealing some differences in cell dynamics between primary 129T2-TG2+/+ and TG2-/- fibroblasts, was a simplified model of the in vivo wound. Cell dynamics in 3D differs from 2D and may represent an environment that better approximates the in vivo environment [410]. A better model for study may be the use of a decellularised cadaver skin with immunogenic cells removed leaving behind an intact ECM [411].

In summary, the results here demonstrate that 129T2-TG2-/- MEFs are defective in cell adhesion, spreading, and migration on Fn compared to their wildtype counterpart. Exogenous addition of TG2 rescued cell adhesion, spreading, and migration of 129T2- TG2-/- MEFs and enhanced these processes in 129T2-TG2+/+ MEFs. Transfection of wildtype TG2 and various transamidation- or GTP-binding-deficient mutants also enhanced cell spreading indicating that TG2 acts as an adaptor molecule in the ECM independent of cross-linking activity. 129-TG2+/+ MEFs are able to express and secrete TG2 protein and are better prepared to adhere to Fn through their cell surface receptor-matrix interactions as well as via TG2. TG2 through its ternary interaction with Fn and integrins β1 and β3 is able 145 to propagate outside-in integrin signaling [83], and in the event of RGD blocking such as in wounding, TG2 is able to salvage the signaling by interacting with syndecan 4 heparan sulphate chains [85] or the 42kD gelatin domain of Fn [83], both of which lead to activation of PKCα and subsequent focal adhesion formation and downstream small GTPase activation. During early in vivo wound healing, the migration of fibroblasts into the wound is crucial for subsequent processes [199]. For this to happen factor XIIIa must coat fibrin clots with plasma fibronectin [381]; fibronectin being a chemoattractant to fibroblasts also provides a signal for fibroblast migration to the wound [412, 413]. The in vitro results here showed that TG2 by virtue of its adaptor function was able to increase cell motility. This notion is supported by the in vivo wound healing data. An experiment that will further strengthen this working hypothesis is to look for Fn in wound beds in vivo and observe whether organised Fn [184] and/or collagen [414] fibrils are more established in the wound matrix 129T2-TG2+/+ animals, as would be expected if fibroblasts had already migrated to these sites. The addition of TG2 (50μl of 10mg/ml) to 129T2-TG2+/+ in vivo wounds, however, did not result in improved wound healing unlike isolated fibroblastic scratch wounds. This could be due to wildtype fibroblasts in wounds already migrating at a maximal rate. Alternatively, the effect of increasing doses of TG2 on wound closure could be tested. It should be noted that in common animal models (including mouse, rat, pig, hamster, rabbit), skin wound healing occurs at the wound edge and is also aided by the contraction of the subcutaneous panniculus carnosus muscle, which is absent from humans [379]. Therefore, the enhancement of cell motility processes by TG2 may likely be more relevant in human wounds where wound closure is dictated predominantly by cell migration from the wound edges. Regardless, TG2 is a facilitator early in wound healing where minimisation of contact to the external environment is crucial to prevent potential infection of an open wound.

146

5. STRAIN DEPENDENT EFFECTS ON WOUND HEALING

5.1 Introduction

5.1.1 Knockout studies: a null mutant or background strain effect?

Knockout animals are extremely useful for studying the roles of genes in the biology of whole organisms. Generation of a gene knockout, or a null mutant, is achieved by gene targeting via homologous recombination to remove the gene of interest from blastocyst stem cells and reintroducing these cells into a developing blastocyst. This results in offspring with one allele inactivated, which can then be crossbred to produce homozygous null mutant knockout animals [415-417]. Gene function can be inferred by observing phenotypic differences between wildtype and knockout animals. For easy identification of the modified offspring, blastocyst stem cells are often of a different background strain to the recipient blastocyst, resulting in chimeric offspring that contain both coat colours of the two parent background strains involved. For example, for the TG2-/- mice engineered by our laboratory, gene targeting was performed in W9.5/W95 (agouti) stem cells from the 129S1/Sv-Oca2+Tyr+Kitl+ strain and these cells were introduced into C57Bl/6 (black) blastocysts [87]. This resulted in easily identified chimeras, whose fur had both agouti and black patches.

Due to the way knockout animals are generated, there are caveats when using them for investigations of the gene in question. For example, if the gene of interest is epistatic i.e. the effects of the gene are modified by other modifier genes, it may be unclear if an observed phenotype is due to the null mutation, to the influence of background genes, or both. The low probability of a recombination event occurring close to the null locus when coupled with selection for the null locus, means that animals that carry the null locus are also likely to have surrounding loci that belong to the original strain from the blastocyst stem cells. To exclude, as best as possible, contributing background strain influences on the phenotype, it is now recommended that congenic knockout lines be generated, whereby knockout mice are backcrossed to wildtype inbred animals. Progeny of the knockout mice that have been backcrossed for 10 generations are considered to be a congenic line and would have on average at least 99.8% loci homology to their wildtype counterparts [418, 419]. Background strain effects including variability in genetic make up, flanking genes, polymorphism, gene penetrance, and gene expressivity, can all have far reaching effects 147 and phenotypic outcomes [420] and it is hard to quantify the contribution of these when using knockouts generated on mixed backgrounds.

The influence of background strain on phenotypes is widely recognised and appreciated across many fields [421]. Even at the stage of generating genetically modified animals, choosing a suitable background strain is a key consideration. For examples, FVB mice provide a large pro-nucleus convenient for injection of DNA for gene targeting; 129/SvJ embryonic stem cells are more likely to result in successful germline transmission [422], and BALB/c and C3H mice are particularly amenable to mutagenesis study due to their sensitivity to the mutagen ethyl nitrosourea [418]. In neurobiology studies, background strain dependence has been described to affect many aspects of animal behaviour including: open field locomotion, learning, memory, aggression, reproductive behaviours, acoustic startle, and prepulse inhibition as well as differences in the pharmacological response to drugs such as ethanol, nicotine, cocaine, opiates, antipsychotic and anxiolytics (reviewed in [423]). In a developmental biology study, knockout of the fibronectin-encoding gene in the C57Bl/6 embryo led to the formation of a defective bulbous heart, but on a 129/SvJ background the heart does not even form [424]! In a study of the tendency of different strains to develop prostate cancer, wildtype C57Bl/6 mice with normal prostates expressed 124 genes that were at least 1.3 fold different from those expressed in the prostates of wildtype 129/SvJ mice. Some of these genes were prostate-relevant and were involved in prostate development, androgen regulation, tumorigenesis, and metastasis development. In a study that monitored healing of ear punch wounds, complete wound closure was observed in MRL/Mpj mice, but not in either C57/Bl6 nor BALB/c mice [425]. In some instances, background modifier genes have been identified that resolve the discrepancy between strains, for example, in studies of the embryonic lethality of TGF-β1 null animals on C57Bl/6-J/Ola and NIH/Ola backgrounds, the former strain produced virtually no viable offsrping [426]. It was determined through marker linkages and genotypic risk ratio analyses that background genetic determinants on chromosome 5 accounted for higher embryonic survival in the NIH/Ola group (TGF-β1 is located on chromosome 7) [426], which was later found to be affected by the survival to birth locus on chromosome 17 [427]. Mentioned here are only a few examples of background strain influences on phenotype, highlighting the care one must take when choosing and mixing animal strains for in vivo studies 148

The issues of background strain influences on phenotype discussed above, relate directly to our laboratory’s experience with respect to the delayed wound healing phenotype of TG2-/- mice. As mentioned in Chapter 3, our laboratory had observed preliminary evidence a few years ago for a delayed healing phenotypes in the mixed strain (129S1/SvImJ/C57Bl/6) TG2-/- mice relative to wildtype mice. However, subsequent crossing of this founder line with another founder line resulted in 129S1/SvImJ/C57Bl/6 TG2-/- progeny, whose MEFs no longer exhibited an adhesion defect (Bryony Mearns thesis, unpublished data). To remove any contributing background strain effects that might potentially account for this sudden loss of phenotype, we then backcrossed the mixed strain TG2-/- animals onto 129T2- or C57Bl/6J backgrounds for >12 generations to generate congenic 129T2-TG2 or B6.Cg-TG2 lines that are >99.98% homologous to the respective parent strains.

To prevent genetic drift, which may occur when a small population becomes isolated by chance and experiences a change in the frequency of an allele, congenic lines are maintained using heterozygous crosses (het x het), with heterozygotes being outcrossed to the parent strain after at most three to five generations, and expansion colonies of homozygous x homozygous and wildtype x wildtype only being propagated from het x het crosses (Figure 3). Given that previous chapters described delayed wound healing in the 129T2-TG2 mice (Chapter 3) and defective adhesion, spreading, and migration of MEFs from 129T2-TG2 mice (Chapter 4), the aim of this section of the work was to compare TG2+/+ and TG2-/- genotypes of the 129T2-TG2 and B6.Cg TG2 lines with respect to in vivo and in vitro wound healing and in vitro cell adhesion.

5.2 Results

5.2.1 Rate of early open wound contraction in B6.Cg-TG2-/- mice is equal to that of B6.Cg-TG2+/+ and 129T2-TG2+/+, but slower than that of 129T2-TG2-/-

An in vivo wound healing assay using B6.Cg-TG2+/+ or B6.Cg-TG2-/- animals showed that unlike the 129T2 line, there was no difference between genotypes, with the overall area under the curve being no different (Figure 29a, b & c; B6.Cg-TG2+/+: 2.96±0.22 vs B6.Cg- TG2-/-: 2.78±0.26; p = n.s). These areas under the B6.Cg-TG2+/+ and B6.Cg-TG2-/- curves were between those of 129T2-TG2+/+ (1.86 ± 0.13) and 129T2-TG2-/- (4.11 ± 0.33). Wound 149 closure in the B6.Cg lines was complete after 10 or 11 days, similar to that for the 129T2- TG2 line.

5.2.2 mRNA expression of TG2 but not of other TG family members, is greater in 129T2-TG2+/+ than in 129T2-TG2-/-, B6.Cg-TG2+/+ and B6.Cg-TG2-/-

There is evidence of compensation in bone development between F13a1 and TG2 [428], and there are suggestions of functional compensation of F13a1 by other TG members in coated-platelet formation [429, 430]. In animals with F13a1 gene inactivation, transamidation activity from knee chondrocytes was not completely abolished [124] and was not different between bone extracts from F13a1 wildtype and knockout mice [431]. Thus, no gross skeletal defects were observed [428], and vice versa for TG2 knockout animals [87]. In the case of coated-platelet formation, transamidation activity in platelet lysates from F13a1 knockout animals was no different to their heterozygous counterparts, therefore pointing to a probable compensation by other TG members [428]. To determine if the expression of other non-TG2 members of the TG family is altered in the B6.Cg-TG2 line relative to 129T2-TG2, RNA from healthy skin and 2-day old wounds from both wildtype and null 129T2-TG2 and wildtype and null B6.Cg-TG2 lines was subjected to quantitative RT-PCR to evaluate mRNA expression of the eight catalytic members of the transglutaminase family: Tgm1, Tgm2, Tgm3, Tgm4, Tgm5, Tgm6, Tgm7, and F13a1 (Figure 30). Expression of Tgm1, 3, 4, 5 and F13a1 mRNA was detected in all samples, whilst Tgm2 mRNA was only detected in TG2+/+ wounds, as expected. Intriguingly, steady state levels of TG2 mRNA were different in the two background strains, with 129T2- TG2+/+ expressing significantly more TG2 mRNA than B6.Cg-TG2+/+ in both unwounded (2 fold vs 1.6 fold) and wounded (1.2 fold vs 0.75 fold) skin samples. Moreover, in response to wound injury Tgm1, Tgm3, and F13a1 mRNA levels increased similarly in both the wild type and null 129T2 and B6.Cg.-TG2 mice. This revealed that during a response to wounding, expression of the aforementioned TGs were not different between the backgrounds, and hence there was no compensation by these TGs, despite lower Tgm2 mRNA in the B6.Cg-TG2 mice.

150

5.2.3 MEF cell adhesion on Fn is equivalent between 129T2-TG2-/-, B6.Cg-TG2+/+ and B6.Cg-TG2-/-

TG2+/+ or TG2-/- MEFs were isolated from B6.Cg embryos and growth curves were established using the same method as Section 4.2.1 (Figure 31). After counting, no difference was observed between B6.Cg-TG2+/+ MEFs and B6.Cg-TG2-/- MEFs. Interestinglt, B6.Cg MEFs of both genotypes had similar growth rates compared to 129T2- TG2-/- MEFs at all passages (Figure 15a-e, hollow squares). The proliferation rate of both genotypes fell with subsequent passage numbers. Cell adhesion assays were performed where B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEFs were allowed to settle on Fn matrix for 10mins to 120mins. Equivalent adhesion was observed for both genotypes, with ~20% of the total number of cells adhered after 10mins of seeding and ~55% of the total number of cells adherent after 120mins (Figure 32a). This adhesion profile was similar to that of 129T2-TG2-/- MEFs but was different from 129T2-TG2+/+ MEFs (two way ANOVA analysis with post-hoc Bonferroni correction, p<0.05) in that the total number of adherent cells was less than that observed for 129T2-TG2+/+ MEFs (Figure 32a). Next, the effect of exogenous TG2 addition on B6.Cg-TG2 MEF cell adhesion was examined 30 mins after seeding. TG2 dose-dependently enhanced cell adhesion in both genotypes (Figure 32b). This dose-dependent response peaked at 60μg exogenous TG2 with ~90% of the total number of cells adhering. This pattern of B6.Cg MEF adhesion was, again, similar to that observed with 129T2-TG2-/- MEFs but different from 129T2-TG2+/+ MEFs in that maximal adhesion of 129T2-TG2+/+ MEFs required less exogenous TG2 (Figure 32b). Thus, similar to the in vivo wound healing assay the efficiency of adhesion of 129T2-TG2-/-, B6.Cg- TG2+/+ and B6.Cg-TG2-/- MEFs is equivalent, but less than that of 129T2-TG2+/+. Moreover, exogenous addition of TG2 improved adhesion of all cells.

5.2.4 MEF cell spreading on Fn was equivalent between 129T2-TG2-/-, B6.Cg-TG2+/+ and B6.Cg-TG2-/-

Next, B6.Cg-TG2 fibroblasts were individually inspected for cell spreading on Fn. Here, 129T2-TG2+/+ MEFs were used as a control in that the cell spreading area of both B6.Cg- TG2+/+ and B6.Cg TG2-/- MEFs was normalised to that of 129T2-TG2+/+ MEFs (Figure 33). At 30mins post-seeding, both B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEFs were significantly less spread ~40% than 129T2-TG2+/+ MEFs (about 70% of that 129T2-TG2+/+

151 in MEFs). With the addition of 20μg exogenous TG2, cell spreading of both B6.Cg-TG2+/+ and B6.Cg-TG2-/- was enhanced by ~25% compared to untreated B6.Cg MEFs and was now equivalent to that of 129T2-TG2+/+ MEFs. After 60mins, the area of spreading of both B6.Cg-TG2+/+ and B6.Cg-TG2-/- cells was similar to that of 129T2-TG2+/+ MEFs at 30 mins, and B6.Cg-TG2 MEFs that had been treated exogenously with TG2 for 30min. At this time point, the addition of exogenous TG2 increased the area of spreading of both B6.Cg-TG2+/+ and TG2-/- MEFs by ~30%. At 90mins, the area of cell spreading of both B6.Cg-TG2+/+ and B6.Cg-TG2-/- was increased relative to the 60 min time point and the addition of exogenous TG2 further enhanced cell spreading in both B6.Cg-TG2 genotypes. The cell spreading profiles of B6.Cg-TG2, plus or minus TG2 addition, were similar to that of 129T2-TG2-/- MEFs as seen in Figure 14.

5.2.5 B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEF monolayers close a scratch wound at equal rates

Confluent fibroblast monolayers of B6.Cg-TG2+/+ or B6.Cg-TG2-/- MEFs were subjected to a scratch wound with or without the addition of 500μg (in 0.5μl) of TG2 (Figure 34). Consistent with the in vivo and in vitro assays, the response of both genotypes was similar and statistically not significantly different from each other. After 24 hr, the denuded area was 75% of the original and after 48 hr it was just below half at ~45%. Addition of TG2 enhanced wound closure: after 24h, the denuded area of both was 60% of the original and after 48 hr it was ~20%.

5.2.6 Quantitation of TG2 expression in 129T2-TG2 and B6.Cg-TG2 MEFs

Given the difference in transcriptional expression of Tgm2 observed between 129T2-TG2+/+ and B6.Cg-TG2+/+ mice and the increase in adhesion and cell spreading of B6.Cg-TG2+/+, B6.Cg-TG2-/- and 129T2-TG2-/- MEFs in the presence of exogenous TG2 addition, it was of interest to compare TG2 protein expression in the MEFs. MEFs were fractionated into cytosolic and membrane fractions and quantitated against housekeeping genes in the respective fractions (Figure 34). Relative to 129T2-TG2+/+ MEF, TG2 expression was signifcantly lower in the B6.Cg-TG2+/+ MEF in both fractions: in the membrane fraction, TG2 in B6.Cg-TG2+/+ was only 0.25 time the TG2 protein expression in the 129T2-TG2+/+

152 counterpart, whilst the B6.Cg-TG2+/+ cytosolic fraction was also only 0.25 times that of the 129T2-TG2+/+ fraction.

153

154

Figure 29 Assay of in vivo wound healing in B6.Cg-TG2 mouse skin

Wound healing was monitored in B6.Cg-TG2 mice after wounding with a 5mm punch biopsy. (a) Wound area in B6.Cg-TG2+/+ (solid circles) and TG2-/- (hollow circles) was measured and expressed as a fraction of the mean of day 0 area ± SEM until closure (n=10). (b) The rate of wound closure in B6.Cg-TG2 mice is compared to that of 129T2-TG2 mice. The graphs of wound healing in B6.Cg-TG2 mice (black lines) from (a) is overlaid with that from 129T2-TG2 (grey lines) from Figure 8c. (c) Total area under the curves from (a) were taken to compare overall wound.exposure (d) Representative photographs of wounds from (a) of both genotypes at day 0, 2, 4, 6, 8, and 10 are shown; scale bar = 5mm. Wound closure was statistically analysed using a repeated measure with two-way ANOVA with Bonferroni correction, and area under the curve was analysed using a two tailed Student t-test. No statistical significance was detected.

155

156

Figure 30 Transglutaminase expression in B6.Cg-TG2 and 129T2-TG2 unwounded and wounded skin samples

Total cDNA was generated from TG2+/+ and TG2-/- unwounded and day 2 post-injury skin samples from both B6.Cg-TG2 and 129T2-TG2 mice (n = 3). Tgm1, 2, 3, 4, 5, 6, 7, and F13a1 mRNA levels were examined in healthy (a) and 2 day post-injury skin samples (b). Mean expression (± SEM) in B6.Cg- TG2+/+ (black columns) and B6.Cg-TG2-/- (white column) and 129T2-TG2+/+ (dark grey columns) and 129T2-TG2-/- (light grey columns) are normalised against the housekeeper gene Hprt.* denotes p<0.05 and *** denotes p<0.001 comparing between B6.Cg-TG2 and 129T2-TG2 of the same genotype for a two way ANOVA with post-hoc Bonferroni correction.

157

158

Figure 31 B6.Cg-TG2 MEF growth curves

Random batches of B6.Cg-TG2+/+ and B6.Cg- TG2-/- MEFs (n = 3) were monitored for growth. rate at (a) P2, (b) P3, (c) P4, (d) P5, and (e) P6. Cell proliferation was drastically reduced with increasing passages. At P2, cells reached confluence on day 4 and for subsequent passages, up to P5, confluence was reached on day 5. Cells were split on the day they reached confluency. MEFs did not propagate beyond P6.

159

160

Figure 32 Adhesion of B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEFs on Fn in the presence of exogenous TG2

Adhesion of B6.Cg-TG2+/+ and TG2-/- MEF was tested on Fn matrix (a) and on TG2-Fn matrix (c). (a) TG2+/+ (solid circles) or TG2-/- (hollow circles) MEFs were allowed to settle on Fn coated wells in triplicate and adhesion was measured at time points ranging 10 to 120 mins (n = 3). Adhesion of B6.Cg-TG2 MEFs (black lines) is overlaid with that of 129T2-TG2 MEFs (grey lines) from Figure 16c. (b) Effect of exogenous TG2 on MEF adhesion was examined. MEFs were seeded for thirty minutes in the presence of varying amounts of TG2. Adhesion of B6.Cg-TG2 MEFs (black lines) is overlaid with that of 129T2-TG2 MEFs (grey lines) from Figure 18. Cell adhesion was compared using two way ANOVA analysis with post-hoc Bonferroni correction. Statistical signficance was observed between 129T2-TG2+/+ MEF adhesion and B6.Cg-TG2-/- or B6.Cg- TG2+/+ MEF adhesion. (c) Representative photographs of fixed crystal violet stained TG2+/+ and TG2-/- MEFs on Fn at 0mins, 20mins (early) and 120mins (late) are shown under 10x magnification, scale bar = 30μm.

161

162

Figure 33 Cell area measurement of B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEFs on Fn

B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEF cell spreading on Fn was measured by evaluating the phalloidin- stained actin area (n = 4) after the MEFs had settled on Fn for a certain period of time. (a) The mean area (± SEM) of B6.Cg-TG2+/+MEF (dark grey columns) and B6.Cg-TG2-/- MEF (light grey columns) was measured at 30, 60, and 90mins and normalised to the area of 129T2-TG2+/+ MEF area at 30mins post-seeding (solid column). In some experiments, 20μg of exogenous TG2 was added to B6.Cg-TG2 MEF of both genotypes whilst settling (labeled “+TG2”) ### denotes p<0.001 relative to 129T2-TG2+/+ MEF cell area using a two- tailed Student t-test. (b) Representative phalloidin-stained MEF at each time point is shown. Scale bar = 10μm.

163

164

Figure 34 In vitro scratch wound assay of B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEF monolayer

Confluent monolayers of B6.Cg-TG2+/+or TG2-/- MEFs were scratched using a 200μl pipette tip and monitored at 0, 24 and 48hrs post-scratching (n = 3). (a) Representative micrographs are shown at 0 and 48 h. (b) Quantitation of the scratch assay TG2+/+(dark grey columns) or TG2-/- (light grey columns). In some experiments 500μg of TG2 was added exogenously to the wells (stippled). The area of scratch was expressed as a fraction of day 0 area. * denotes p<0.05 and ** denotes p<0.01 for a two-tailed Student t-test comparing MEFs of the same genotype with and without addition of TG2.

165

166

Figure 35 Detection of TG2 in the membrane and cytosol of 129T2-TG2 and B6.Cg-TG2 MEF

129T2-TG2 or B6.Cg-TG2 MEFs were lysed and separated into cytosolic and membrane fractions (n = 3). (a) Membrane TG expression was normalised to the housekeeping protein pan-cadherin and representative Western blots are shown in (b). (c) Cytosolic TG expression was normalised against the housekeeping protein GAPDH and and representative Western blots are shown in (d). # denotes p<0.001 for a two-tailed Student t- test comparing TG2 expression in 129T2-TG2+/+ and B6.Cg-TG2+/+.

167

5.3 Discussion

The results of this work explain the initial observations in the laboratory of delayed wound healing in one mixed strain founder line but not in another. Thus backcrossing of the mixed strain mice to their respective inbred strains allowed segregation of the background coincident with the phenotype such that it was observed in one inbred strain but not the other. Thus, in B6.Cg-TG2 mice, skin wound healing was no different between TG2+/+ and TG2-/- animals, whereas in 129T2, early wound closure was delayed in 129T2-TG2-/- relative to 129T2-TG2+/+. Interestingly, when compared to 129T2-TG2 mice, wound closure in both B6.Cg-TG2+/+ and B6 TG2-/- was similar to that in 129T2-TG2-/- and was significantly slower early on than in 129T2-TG2+/+ mice. Consistent with the skin wound healing phenotypes, both B6.Cg-TG2+/+ and B6.Cg-TG2-/- MEFs exhibited similar cell adhesion, spreading, and migration characteristics to 129T2-TG2-/- MEFs, and, in turn, were significantly less adherent were smaller in cell spreading area, and migrated more slowly than 129T2-TG2+/+ MEFs. The wound healing (Figure 9) and cellular adhesion data (Figure 16 & 18) obtained using 129T2-TG2 mice show that the phenotypes observed are definitely an effect of TG2, but this TG2 effect were apparently ‘lost’ on the B6.Cg background. This demonstrates the influence of the background strain on phenotypic outcomes stemming from the presence of unknown background determinant(s) or modifier gene(s). When considering potential effects of background modifier gene(s), they could have a direct effect on TG2 expression, steady state amounts, turnover, and/or its exportation to the extracellular environment. Alternatively, these background genes could influence the phenotype indirectly via other pathways involved in wound healing or cell motility. In the current study, the strain differences could be due to TG2 expression level differences between the B6.Cg-TG2+/+ and 129T2-TG2+/+ animals. qPCR and Western blotting which examine steady-state mRNA and protein levels in the different strains, revealed that 129T2-TG2+/+ mice have an inherently higher expression of TG2 mRNA, cytosolic TG2 and membrane TG2 than B6.Cg-TG2+/+ mice. There is no evidence from this study to suggest compensation from other TG family members. It appears that MEFs from B6.Cg-TG2+/+ mice are naturally less adherent, spread less and have a slower migration rate than their 129T2-TG2+/+ counterpart because of lower amounts of steady state TG2, and that exogenous TG2 addition can enhance adhesion, spreading, and migration. Since the difference was observed at both the mRNA and protein level, it is possible that there is a 168 difference between the two TG2+/+ strains in terms of the biological machinery that regulates Tgm2 transcription or stabilises Tgm2 mRNA. This is not unprecedented, for example it has been shown that B6.Cg mice were more suspectible to asbestos-induced pulmonary damage compared with mice of a 129 strain [432]. It was detected that TGF-β1 expression was higher in B6.Cg mice, than 129 mice, leading to high rates of TGF-β1- induced apoptosis leading to increased lung lesions [432]. An experiment with controlled knock-down (e.g. anti-sense technology) of TG2 expression in 129T2-TG2 MEFs should yield B6.Cg-TG2 MEF-like phenotypes in terms of cell adhesion, spreading, and migration. In terms of in vivo wound healing, based on the 129T2-TG2 data, one would predict that exogenous application of TG2 to B6.Cg-TG2 wounds should increase the rate of wounding healing to that observed with 129T2-TG2+/+ mice.

To tease out the B6.Cg strain effect on wound healing and cell motility processes, one could perform global expression analysis (e.g. microarray analysis). Such an approach may reveal changes elsewhere in the genome, which may shed light on how TG2 is differentially expressed between the B6.Cg and 129T2 background strains. There are examples of studies where strain-dependent biology was investigated, two such examples include investigation of the cataract mechanism by Tang et al [420] and lung fibrosis mechanism by Brody et al [433]. Tang et al. [420] looked at the phenotype outcome of connexin wildtype and knockout mice on SV and C57Bl/6 background strains. Connexin is a protein required for proper formations of intercellular gap junctions in the eye and prevents the formation of cataracts [434]. It was found from the study that wildtype SV animals displayed advanced cataracts formation than wildtype C57Bl/6 animals, whilst the wildtype of both strains developed cataracts more slowly compared to their knockout counterparts. Microarray data revealed that the SV strain had lower expression of proteins involved in protein synthesis, protein metabolism and protein catabolism. One candidate indentified was heat shock protein 25 (Hsp25), which was upregulated in C57Bl/6 by both microarray and real-time PCR analysis [420]. Hsp25 is important in maintaining cell homeostasis in response to cellular stress and it is likely to be responsible for maintaining the redox potential of the lens milieu during the onset of cataract formation where there is protein breakdown and aggregation. These findings suggest that cataract formation in the SV animals was dependent on post-translational events rather than the connexin protein expression levels. Similarly, in the current study, improved wound healing in 129T2-TG2+/+ 169 correlated with higher extracellular TG2, whereas in B6.Cg animals, physiological TG2 expression does not seem to be involved in wound healing or cell motility processes. A global expression comparison approach here using a four-way comparison of 129T2-TG2+/+ healthy skin and wounded skin and B6.Cg-TG2+/+ healthy skin and wounded skin should reveal differences in the healthy and wounded genome state of the two strains in terms of wound healing. The same can be repeated on actively growing/migrating MEFs to investigate global expression differences associated with cell motility.

Wound healing defect was observed in 129T2-TG2-/- animals compared to 129T2-TG2+/+ animals, in parallel, cell migration defects were observed in 129T2-TG2-/- MEFs compared to 129T2-TG2+/+ MEFs in the scratch wound assay. On the other hand, B6.Cg-TG2-/- animals displayed no wound healing defects compared to the wildtype counterpart, and consistently there were no cell migration defects between B6.Cg-TG2-/- and B6.Cg-TG2+/+ MEFs in the scratch wound assay. Taken together, these results show that in vitro wound models using MEFs are a good representation for the in vivo wound situation, as in vitro phenotypes are mirrored with the in vivo phenotypes.

Despite the importance of using background strain controlled animals, some studies in the TG2 field have used mixed strain animals for phenotypic characterisation leading to conflicting or unclear results between independent studies. This is illustrated in liver studies where TG2 has been shown to be both protective and apoptotic in response to liver damage induced by various chemical. In one, Nardacci et al (2003) found that the liver toxin CCl4 in TG2-/- mixed strain animals had higher mortality compared to wildtype animals [435] Tatsukawa et al (2009), on the other hand, showed that treatment of TG2-/- mixed strain animals with ethanol or with the pro-apoptotic antibody Jo2 resulted in reduced hepatocyte apoptosis compared to TG2+/+ [436]. In yet another twist, TG2+/+ FVB mice treated with Jo2 had higher survival rates compared to TG2-/- animals on a mixed background of B6 and 129 [437], however FVB animals are well known to have higher tolerance towards Jo2 than B6 animals. Although the results from the latter described Jo2 experiments support the conclusion from Nardacci’s work as opposed to Tatsukawa’s work, the means to the end are flawed as non-identical animal strains were used as controls. Therefore the role of TG2 in liver injury is still unclear as it appears that background genetics play a significant role in the outcome of the experiments performed. In terms of cardiology, there is also conflicting 170 data. Our group showed that TG2+/+ or TG2-/- on a mixed strain were no different in cardiac function [87], whilst another group showed on an unspecificed background that cardiac parameters such as coronary flow, aortic flow, aortic pressure, heart rate, and myocardial ATP were lower in TG2-/- resting animals compared with TG2+/+ B6 animals [438, 439], despite earlier reports by the same group showing no difference in ATP levels in resting TG2-/- animals compared to TG2+/+ [78]. In two recent reports, the role of TG2 in the formation of atherosclerotic plaques was investigated in the apolipoprotein E knockout mouse model, which develop advanced and complex atherosclerotic plaques upon a high fat diet [140, 141]. In both, TG2 and apolipoprotein E double knockout animals were generated by crossing the apolipoprotein E knockout mice with mixed strain TG2-/- mice. In the study by Williams et al. [141], which used our mixed strain TG2-/- mice [87]. TG2 was not found to affect plaque deposition and composition, but conflictingly, in the study by Van Harck et al. [140], which used mixed strain TG2-/- mice from the Melino laboratory [88], the lack of TG2 was found to disrupt plaque stability through reduced collagen incoporation into the plaques and hence leading to advanced atherosclerosis. Clearly, there are many studies in the field that have used mixed strain animals that may have genetically drifted and/or used improper controls. As it is widely acknowledged that background strains could have significant effects on phenotypes, and this is further supported by the data of the current study, any animal data should accompany a description of the background strains and breeding strategy.

171

6. CONCLUSION & FUTURE DIRECTIONS

In this thesis, it is demonstrated for the first time that the lack of TG2 protein expression leads to defective wound healing in vivo. The mechanism for this phenotype is through the action of TG2 as an adaptor molecule interactions that augments cell surface receptor molecules with ECM proteins, thus allowing efficient cell migration to minimise wound site exposure. Data from this thesis showed that neither the transamidation activity or GTP- binding activity of TG2 is required for proper wound healing. The slower rate of 129T2.TG2-/- MEF adhesion, and subsequent cell spreading and migration, on Fn matrix strongly attributes defective wound healing to defective cell dynamics in the absence of TG2. Further cementing this, were the background strain-dependent observations that in vivo wound healing was defective in the B6.Cg-TG2+/+ animals compared with 129T2- TG2+/+ animals (and similar to TG2-/- animals of either strain), and consistently, the lack of phenotypic differences was extended to cell adhesion, spreading and migration. Moreover, compared with skin tissue (wounded or healthy) and MEFs isolated from 129T2-TG2+/+ mice, B6.Cg-TG2+/+ Tgm2 mRNA expression was lower and further more steady state TG2 protein expression was reduced in the B6.Cg-TG2+/+ MEFs. The effect of TG2 on wound healing is, therefore, dependent on expression of the protein in the local ECM environment of the wound, a phenotype that can be restored to normal wildtype levels with exogenous addition of TG2 in the TG2-/- animals.

Human wound healing involves primarily cell migration over the wound epithelium, whereas mouse wound healing requires both cell migration and myofibroblast contraction for wound closure [164, 440]. Given that migration is the central process in the human scenario, TG2 may have more relevance to human than to mouse wound healing.

Acute wounds, such as puncture wounds from this study, heal naturally going through the orderly phases of inflammation, cell proliferation, and remodeling (Chapter 3.1.2). The overall goal of healing is to restore tissue integrity. Since each phase is intimately linked, it should therefore be beneficial for overall healing if any of its entailing processes can be facilitated. A punctured wound like the one created by punch biopsy in this study represents a secondary-intention wound: a wound where wound edges are allowed to close naturally without any manipulation [441]. This is in contrast with a primary intention wound where

172 the edges are closed by suture, immediately reducing void space between wound edges to minimise scar formation. There are situations where wound closure by suture is not suitable: wounds that are soiled and need constant dressing; wounds that are too large to close by suture; or wounds with excessive swelling where suturing may cause ischemia [441]. The ability of TG2 to facilitate early wound closure was demonstrated in this study. This role of TG2 fits with the aim of early wound healing— to close off the wound and re- establish homeostasis [170]. Because of this, TG2 has the potential to serve as an active ingredient in healing agents such as topical medication or wound dressing. Delivery method of TG2 is a critical aspect to consider, as it would greatly influence the kinetics of the protein at the site of action. For therapeutic delivery, proteins are usually encapsulated in carriers or vehicles. An ideal carrier should protect its content from denaturation and retain its biological activity, facilitate its optimal release at the site of action, and allowed to be absorbed into the tissue [442, 443]. The kinetics and halflife of TG2 in wounds have not been examined here. In the current study, TG2 was delivered in a bolus of PBS that leaked from the wound site as the droplet was delivered. Investigations should be made to determine a suitable solvent or carrier for TG2 delivery to the wound site, a chemical that should improve TG2 protein halflife or kinetics and is inert to the wound. In some pilot experiments methylcellulose, a sturdy inert gel was trialed with limited success. Another possible candidate are niosomes, a non-ionic surfactant carrier that encapsulates an aquaeous core, may be a possible carrier for TG2 [444]. Niosomes display some attractive characteristics: increase content absorption into the tissue [445], increase retention in skin tissue [446], and can be applied topically as shown by the application of recombinant human granulocyte-macrophage colony stimulating factor in full thickness wounds [447].

TG2 can also be considered for application in healing chronic wounds – wounds that do not follow through with the healing phases properly and hence do not heal as quickly as an acute wound. Chronic wound treatment poses a burden on healthcare systems in the US and UK, it was estimated to cost the US healthcare system $5 billion to $9 billion annually [448] and the UK nation health services £1 billion per year [449]. The majority of chronic wounds are ulcers of various types: pressure, venous, and diabetic [449]. The cause of chronic wounds differ in each wound type, but on a cellular level, these tend to be attributed to any combination of the following: perpetuated inflammation [450], reduced expression and or response to various growth factors [451], imbalance of proteinase activity 173 and their inhibitors [452], and cell senescence [453]. The end result in most chronic wounds is that wound closure is significantly delayed accompanied by changes to underlying wound bed histology. Therefore therapeutic aim to treat chronic wounds involves overcoming these defects. Currently, the major focus on chronic wound treatment has been testing the topical application of various wound-related growth factors in an attempt to restore normal healing. These factors include TGF-β1 (re-epithelialisation, granulation tissue formation and neovascularisation), FGF (re-epithelialisation and neovascularisation), IL-β1 (infection clearance), and granulocyte macrophage-colony stimulating factor (acute wound improvement) all of the above are undergoing clinical trial [449]. PDGF, commercially named as Regranex, is the only growth factor approved by the Food and Drug Administration (USA) for the treatment of chronic wounds in diabetic patients [454, 455]. Therapeutic application of individual growth factors has met with limited success. Even the FDA-approved Regranex only improves wound closure by 10% [450]. Therefore, application of these factors is not as simple as adding them straight to the wound. It has been suggested that growth factors should be applied in combination and sequentially at timed intervals to better emulate physiological healing [449, 456]. The potential therapeutic use of TG2 here has not been suggested despite being implicated in different wound healing processes (reviewed in [457]). Considering its role in wound closure and cell motility processes, TG2 could serve as a useful active agent in chronic wound treatment. Topical TG2 could be applied in conjunction with growth factors that complement each other, with TG2 active to promote cell migration to the wound bed and other growth factors serving as chemoattractants or regulators to recruit various cell types for wound healing. In one report, cultured venous ulcer dermal fibroblasts were induced to proliferate using the growth factors EGF and FGF [453]. In the same way, TG2 application might be used to promote fibroblast growth under ulcerated conditions. TG2 biological activity may be compromised in chronic wounds, however, as TG2 is a target of MMP-2 [105] and MMP-2 has been shown to be upregulated in wounds such as leg ulcers [458] . It is thus possible that the degradation of TG2 at these sites may prevent proper wound healing by preventing TG2’s action to promote proper cell motilities within the wound milieu. In addition, since TGF-β signaling is dependent on TG2 expression [7], efficiency in wound healing could further be reduced. In addition to aiding of wound closure, moderation of inflammation [69] and stabilisation of wound ECM [10, 232] both brought about by cross-linking in vitro also appear to be helpful againt chronic wounds. 174

Another promising approach to chronic wound treatment is through tissue bioengineering [449]. Tissue engineering is the technology of regenerating new tissue by means of scaffolds, cells, and growth factors where the introduction of these biocompatible materials would provide a 3D ECM that would stimulate endogenous tissue growth at injury sites [443]. These scaffolds could be populated with cells of the appropriate tissue type in vitro or in situ [443]. In the in vitro setting, cells of a single source are allowed to infiltrate the scaffold producing an engineered tissue within bioreactors, and in the in situ setting the scaffold is transplanted onto the patient and is infiltrated with the patient’s own cells. The common material for skin scaffolds is collagen, which supports cell adhesion for skin cell types such as endothelial cells, keratinocytes and fibroblasts [459] and has the advantage of low antigenicity so that the scaffold does not get rejected [460]. Currently collagen-based scaffolds that are available for dermal treatment are Alloderm (allogeic collagen matrix) [461], Integra (allogeneic collagen matrix and chondroitin sulphate) [462], and DermalTC (neonatal dermal fibroblasts on polyglactin mesh), which are applied mainly to burn and diabetic ulcer wounds [449]. Despite these advancements, there is still a demand for improvements on current products [463]. A few characteristics are required for an ideal scaffold. Firstly they must support sufficient cell adhesion and growth chemically and physically such as having collagen on which cells would adhere rapidly and have micropores to increase the number of cells per scaffold volume. In addition, cell migration or proliferation does not equate to functional restoration. It was found in one study that collagen scaffolding blocked wound contraction in vitro as fibroblasts migrated into it [464]. It was suggested that migrated fibroblasts were disoriented, as determined by examining alignment of cell fronts, and were not able contract the wound [465]. Secondly, the scaffold must be mechanically and chemically resilent to forces and enzymatic degradation. Thirdly, the scaffold must be able to host physiological processes such as wound contraction and vascularisation [443]. The investigation of TG2 cross-linking of collagen by Chau et al (2005) showed that increased protease-resistance in collagen matrices supported adhesion and cell motility of both human foreskin dermal fibroblasts and human osteoblasts [399], which satisfies the characteristic requirements of a suitable scaffold. This consequently led to the patent US20080305517 in the application of TG2 in the synthesis of a biomaterial for scaffold. As described in Section 3.1.2, wound healing requires more than just collagen as the ground substance. Other important matrix materials 175 include fibronectin, vitronectin, and heparan sulphate proteoglycans [190, 191]. The addition of TG2 and Fn components, which has been shown in this study to aid wound healing and cell motility processes, could prove to be useful for developing scaffolds for wound healing. Interestingly, addition of soluble Fn to Fn-null MEFs adhered to collagen I was able to trigger spontaneous cell proliferation and the self-formation of structured microtissues [466]. Utilisation of TG2 in such composite tissue scaffolds both in its synthesis and as a component of the scaffold may prove to be suitable, if not better than just collagen alone. Firstly, both collagen and fibronectin are able to support cell growth and proliferation, and as demonstrated TG2 addition is able to improve cell adhesion, spreading and migration onto Fn matrix. Secondly, TG2 cross-links both collagen [399] and Fn [82] giving rise to covalently linked protease-resistant structures ideal for scaffolds. Thirdly, TG2-ECM matrices, drawing from cellular assays, at least seem to enhance wound contraction (Section 3.2.2 & 4.2.5), and TG2 has been observed at sites of neovacularisation [160], but whether it will support or enhance neovascularisation into the tissue scaffold remains to be tested.

In conclusion, this work has demonstrated that TG2 is involved in cell motility processes and contributes to murine cutaneous wound healing. Since the primary mechanism of human wound healing is cell migration over the wound epithelium the clinical relevance of TG2 in human skin wound healing deserves further investigation.

176

APPENDIX

geneSymbol logFC P.Value Tgm2 -2.25266 3.61E-08 Mup2 -0.99513 0.044244 Rrad -0.94954 0.002651 Klhl31 -0.89374 0.028146 Apobec2 -0.86925 0.008187 Xirp2 -0.85486 0.029905 Smpx -0.84001 0.008905 Kbtbd10 -0.83936 0.014641 Myoz2 -0.76154 0.028233 Asb5 -0.71936 0.035307 Nrap -0.71351 0.0372 Lrrc2 -0.69317 0.032042 Des -0.68684 0.024143 Tnnc2 -0.6663 0.042856 Tnni2 -0.62209 0.046102 Mb -0.61904 0.030163 Srl -0.61187 0.032999 Mustn1 -0.58975 0.012266 Pkia -0.58651 0.031822 Mylk2 -0.58112 0.037744 Mmp13 -0.55315 0.023628 Usp13 -0.54959 0.018649 EG382931 -0.54584 0.004871 Acta1 -0.54158 0.040607 Tspan8 -0.53317 0.043318 EG667441 -0.52767 0.003546 100039528 -0.5065 0.010237 Gadl1 -0.48945 0.027436 Eef1a2 -0.48938 0.030035 Asb14 -0.48806 0.028284 Rragd -0.47859 0.016296 Cox6a2 -0.46071 0.038904 OTTMUSG00000017746 -0.45821 0.001451 Tas2r135 -0.45336 0.006623 Cryab -0.45185 0.044414 Prkg1 -0.44182 0.048171 Trim72 -0.44149 0.022915 Lmod2 -0.43139 0.036902 Olfr1148 -0.41777 0.002768 Myo18b -0.41563 0.039337 AA416372 -0.41311 0.011337 LOC100045610 -0.41165 0.025075 Serpinb1a -0.40274 0.031873

177

Dusp27 -0.3998 0.025337 Myo18b -0.39751 0.010706 OTTMUSG00000007808 -0.37206 0.022392 ENSMUSG00000074845 -0.3579 0.004762 Cnksr1 -0.35501 0.036583 Kcnc4 -0.35324 0.006386 Ptgs2 -0.35217 0.047123 Khdc1a -0.34369 0.002499 Unc45b -0.3436 0.04265 Lrrn1 -0.3406 0.045767 Tgm7 -0.33956 0.015476 Smtnl1 -0.3395 0.003997 Olfr310 -0.33731 0.002365 OTTMUSG00000007474 -0.33113 0.032497 Tas2r143 -0.32702 0.016844 Myo18b -0.32498 0.042229 Fbxo40 -0.32312 0.009443 Atp1b1 -0.32139 0.012909 Kcna7 -0.32006 0.002281 Sync -0.31903 0.041394 Olfr373 -0.31845 0.025584 Kcnq5 -0.31399 0.032714 Apcs -0.31358 0.009282 Krtap5-1 -0.31254 0.002391 Musk -0.30698 0.03447 Abcb4 -0.30549 0.020088 Ihpk3 -0.30523 0.025864 Olfr605 -0.30324 0.026712 Klk1b24 -0.30161 0.022554 C1qtnf3 -0.29754 0.005602 Ak1 -0.29567 0.020939 Olfr239 -0.29538 0.005394 Sh3bgr -0.29536 0.038447 F2rl3 -0.28792 0.003993 Cfl2 -0.28769 0.045883 ENSMUSG00000073569 -0.28152 0.010332 Gm996 -0.27902 0.021061 EG218444 -0.27871 0.029302 Kir3dl1 -0.27604 0.032181 Synpo2l -0.27513 0.037758 Ociad2 -0.26812 0.017785 ENSMUSG00000072690 -0.26491 0.019088 Olfr1031 -0.2636 0.010467 OTTMUSG00000008822 -0.25881 0.001785 LOC627626 -0.25802 0.04568 Vmn2r87 -0.25161 0.032898

178

EG620119 -0.25148 0.038811 Olfr957 -0.24796 0.024398 Avpr2 -0.24754 0.008686 A130082M07Rik -0.24699 0.032031 Crhr2 -0.24566 0.01745 Dnajc5b -0.24429 0.04422 Srms -0.24408 0.004572 Sema6c -0.24207 0.010198 Hormad1 -0.24167 0.01298 Rprm -0.24065 0.018237 LOC676914 -0.24041 0.035641 Mkrn3 -0.24009 0.002989 Shisa2 -0.23957 0.031507 Mmp10 -0.23896 0.032533 Stac3 -0.23867 0.017624 OTTMUSG00000005131 -0.2384 0.016192 Gpr25 -0.23838 0.005103 100040113 -0.238 0.005613 LOC100048255 -0.2379 0.026365 Myh6 -0.23537 0.017544 EG628893 -0.23427 0.020705 Olfr1364 -0.23272 0.012392 Itgb1bp3 -0.23242 0.044022 Lmbr1 -0.23241 0.041136 6430517E21Rik -0.2322 0.000818 Kcnj11 -0.23219 0.010002 Lrrtm3 -0.23208 0.015281 Speer2 -0.23139 0.021965 Neurl -0.23124 0.038216 Sema4g -0.23079 0.004594 ENSMUSG00000074555 -0.22825 0.014433 Cyp2d10 -0.22825 0.016883 ENSMUSG00000074072 -0.22796 0.008609 Igk-V1 -0.22514 0.047424 EG666123 -0.22489 0.036902 Tgfb2 -0.22473 0.021239 Mpo -0.22439 0.047303 1110030E23Rik -0.22397 0.042011 Slco5a1 -0.22337 0.024639 Olfr350 -0.22327 0.037706 Hspb2 -0.22282 0.046337 Tas2r129 -0.22235 0.02447 9430060I03Rik -0.22189 0.019093 Htr5b -0.22179 0.005622 EG546708 -0.22176 0.017049 Tnnt2 -0.22158 0.036705

179

Kcnj5 -0.22149 0.019185 621968 -0.2203 0.017832 EG624023 -0.22 0.048371 ENSMUSG00000071392 -0.21964 0.017485 A430072C10Rik -0.21912 0.02293 LOC622659 -0.21763 0.02644 Atp8b3 -0.21726 0.035614 B4galnt2 -0.21711 0.038741 Syt10 -0.21669 0.008492 Il28a -0.21598 0.041862 Hspb3 -0.21593 0.044028 Slc23a1 -0.21577 0.009136 V1rd15 -0.21545 0.04634 EG668156 -0.21545 0.037941 Fgf20 -0.21528 0.010173 Olfr464 -0.2139 0.04477 Nlrp14 -0.21383 0.013414 Sox8 -0.21327 0.011115 4930471G03Rik -0.21295 0.015523 Olfr557 -0.21231 0.019395 Ovol3 -0.21086 0.042808 V1rg6 -0.21053 0.039398 Ctsg -0.21018 0.031171 Taar5 -0.21009 0.032926 Isl1 -0.20826 0.01206 V1rg8 -0.20754 0.039637 Nefl -0.20715 0.010202 Camp -0.20697 0.023225 Glyctk -0.20654 0.032992 Skap1 -0.20633 0.021563 H2-M10.2 -0.20491 0.012051 Insm2 -0.20431 0.018205 Col4a6 -0.20428 0.012816 Tgif2 -0.20409 0.035834 Ppifos -0.20385 0.005462 Tff2 -0.20356 0.010665 Tmem139 -0.20348 0.029328 Rdh19 -0.20301 0.012137 Pnma3 -0.20294 0.008392 Gpr158 -0.20282 0.010118 EG382156 -0.2027 0.042222 D430041D05Rik -0.20182 0.019249 A930035D04Rik -0.20025 0.039357 2610021K21Rik -0.20007 0.032767 Pdc -0.19977 0.015245 Slc13a4 -0.19975 0.034007

180

9530053H05Rik -0.19967 0.018619 1700013G24Rik -0.19941 0.016036 OTTMUSG00000023126 -0.19927 0.04429 1110001D15Rik -0.19908 0.034463 LOC546695 -0.19805 0.043672 Adamts3 -0.19781 0.039381 Cnih3 -0.19755 0.041114 3322402L07Rik -0.19744 0.020407 Megf6 -0.19718 0.038448 LOC100044855 -0.19691 0.006297 Nhedc2 -0.19635 0.025083 Wdr16 -0.19615 0.007324 4921530L21Rik -0.196 0.026632 Olfr1272 -0.1959 0.039068 Tram1l1 -0.19529 0.015511 LOC100045615 -0.19484 0.029621 Kif12 -0.19365 0.008537 Mycs -0.19298 0.032547 Prhoxnb -0.19294 0.0344 Gast -0.19274 0.016107 V1ra2 -0.1926 0.041092 Hsf2 -0.19221 0.039006 Kcnh7 -0.1918 0.02181 Pcsk9 -0.19055 0.020233 Olfr611 -0.19046 0.035995 Ctsr -0.19029 0.009526 4921509A18Rik -0.19007 0.009605 Dab1 -0.18971 0.031283 EG639530 -0.18961 0.011469 Slc6a15 -0.18943 0.038577 Lefty2 -0.18915 0.039255 Cd209e -0.18902 0.035945 Egfl6 -0.18884 0.023113 Rit2 -0.1873 0.038794 Olfr1511 -0.18713 0.035411 Gm568 -0.18684 0.019182 Glis1 -0.18679 0.0267 Amhr2 -0.18668 0.013659 Wdr31 -0.18623 0.016842 Cacna1g -0.18563 0.039921 Crygn -0.18554 0.01444 Ica1l -0.18553 0.007614 Plb1 -0.18552 0.029057 Vgll1 -0.1855 0.022826 Tcp11 -0.18545 0.014324 Pga5 -0.18539 0.017111

181

Sh3rf3 -0.18413 0.006472 Slc12a5 -0.18403 0.023018 Olfr180 -0.1836 0.01275 Chrna9 -0.18355 0.042787 Gpr20 -0.18354 0.044477 Col9a3 -0.18292 0.020421 EG218997 -0.1823 0.042421 Olfr390 -0.18181 0.046352 Apol8 -0.18176 0.02744 Sstr2 -0.18175 0.029653 Sphkap -0.18147 0.014015 Zfp651 -0.1814 0.029486 Prf1 -0.18136 0.028853 Olfr1305 -0.18133 0.040487 Nr2e3 -0.18084 0.013504 Dcakd -0.18079 0.019358 A4gnt -0.18067 0.009931 Stmn3 -0.18059 0.016294 Dnajc6 -0.18034 0.023619 Zfp264 -0.18011 0.01628 Cntnap5c -0.17983 0.021066 Mos -0.17924 0.016024 Banf2 -0.17861 0.010025 2810051F02Rik -0.17849 0.049897 Igf2as -0.17833 0.011878 Ern2 -0.17748 0.021993 Cryba4 -0.17684 0.038092 Cst11 -0.17671 0.042793 Slc6a3 -0.17656 0.018306 Ina -0.17641 0.032588 Olfr1231 -0.1762 0.044051 Nlgn1 -0.1759 0.021283 EG330305 -0.17566 0.036242 Il27 -0.17562 0.038789 Vax2 -0.17526 0.014942 2900005J15Rik -0.17507 0.041133 Myo18b -0.1738 0.015721 Olfr157 -0.1732 0.033746 Tchhl1 -0.17295 0.033893 Rhcg -0.17188 0.022235 Lrrc50 -0.17176 0.047765 Slc25a29 -0.17147 0.026488 Bmp8b -0.17092 0.032086 Prdm13 -0.17076 0.018772 1600012P17Rik -0.17036 0.019166 Lrfn1 -0.1701 0.046124

182

EG629583 -0.17007 0.033131 Impg2 -0.16989 0.0315 Mug1 -0.16977 0.025881 Ppm1e -0.16964 0.015225 Aqp12 -0.16946 0.018625 1700018M17Rik -0.16859 0.03948 Htr2b -0.16841 0.040263 ENSMUSG00000072619 -0.16716 0.047356 Clec4a4 -0.16655 0.015616 Tmem91 -0.1663 0.016442 Ank1 -0.16568 0.018855 Serpina5 -0.16544 0.047745 Kifc5c -0.16531 0.011217 Fgf12 -0.16442 0.023106 Slc1a4 -0.16403 0.027653 Otos -0.16399 0.03192 Fstl4 -0.16383 0.02757 Ak5 -0.16355 0.046123 Gpr3 -0.16352 0.031405 Spink4 -0.16272 0.047458 Rbm11 -0.16256 0.048175 Adra2c -0.16253 0.033032 Slc13a5 -0.16207 0.007239 EG546638 -0.16161 0.035673 Gm1323 -0.16143 0.031838 2810055G20Rik -0.16082 0.049596 Dyx1c1 -0.16031 0.013626 Olfr1030 -0.16031 0.037131 Phxr1 -0.16016 0.03811 Ccdc136 -0.16002 0.014545 Nkx6-1 -0.16001 0.038953 Ncf2 -0.15962 0.049651 Onecut1 -0.15961 0.049195 D930048N14Rik -0.15922 0.047464 Abcc8 -0.15775 0.034015 Traf4 -0.15747 0.039724 Stmn4 -0.15637 0.024327 4930550C14Rik -0.15611 0.042949 Cd276 -0.15584 0.038204 Grm8 -0.15563 0.030321 ENSMUSG00000052323 -0.15496 0.030708 Rhox6 -0.15493 0.029318 Acpt -0.15452 0.043986 Artn -0.15373 0.01941 Olfr1038 -0.15273 0.031516 Slc17a3 -0.1516 0.038747

183

Zbtb8b -0.15149 0.027241 Icam5 -0.15137 0.023742 2810453I06Rik -0.15102 0.031768 Nhs -0.1507 0.047212 Thsd3 -0.15032 0.046399 Pramel7 -0.15016 0.044338 Zfp689 -0.14997 0.019808 EG620899 -0.14978 0.021945 Cck -0.14973 0.031675 Bfsp2 -0.14951 0.038984 BC038167 -0.14941 0.036212 Slamf8 -0.14918 0.049948 Gpr142 -0.14902 0.02518 Pou2f2 -0.14802 0.018496 Wnk2 -0.14797 0.01935 Prss29 -0.14794 0.029931 Hydin -0.14739 0.037273 EG622432 -0.14721 0.046542 Olfr791 -0.14695 0.045341 Rgn -0.14643 0.029192 Dak -0.14491 0.047288 5830411N06Rik -0.14479 0.035918 Nrl -0.14419 0.037786 Dusp9 -0.14418 0.041152 Hes5 -0.14398 0.027345 EG629678 -0.14387 0.045344 OTTMUSG00000005162 -0.14352 0.028509 Wipf3 -0.14295 0.027382 EG665887 -0.14186 0.046253 Tcea1 -0.14162 0.024619 Dpf1 -0.14142 0.045677 Spinlw1 -0.14136 0.04584 Slitrk3 -0.14058 0.022219 Arhgdig -0.14054 0.035177 LOC100048615 -0.14043 0.030577 Tex16 -0.1396 0.044328 Rnf186 -0.13857 0.039831 BC043934 -0.13773 0.04458 Col19a1 -0.13695 0.023405 Crisp3 -0.13678 0.041669 Fbxl7 -0.13608 0.047813 Rpap1 -0.13509 0.0289 Zfp358 -0.13493 0.048833 Slc1a6 -0.13428 0.047418 ENSMUSG00000049982 -0.13358 0.029368 Atad4 -0.13325 0.040191

184

Adamtsl5 -0.13252 0.030539 Arl4c -0.13226 0.04505 V1rc16 -0.13158 0.037491 Serpinb1c -0.13072 0.049047 Barhl2 -0.12939 0.037136 Prss32 -0.12618 0.027336 Olfr66 -0.12356 0.04886 Sstr1 -0.11816 0.04312 EG631906 -0.11723 0.038677 V1rh2 -0.11359 0.049585 Lrrc29 -0.11157 0.038377 1200015N20Rik 0.105208 0.04804 Kctd20 0.111652 0.040691 Rars2 0.114474 0.041733 Usp21 0.114958 0.047303 Cnot10 0.115514 0.047523 Lsg1 0.117324 0.047156 Pbx2 0.117646 0.047072 Glg1 0.117736 0.049863 Phc3 0.117804 0.037574 BC059842 0.118284 0.035235 Recql 0.118483 0.044592 Arl3 0.118555 0.045619 Plekhm2 0.119989 0.033429 Fastkd1 0.121105 0.04413 Syne2 0.122441 0.04908 App 0.122855 0.048423 Zfp276 0.122915 0.046392 Sap130 0.12365 0.030179 Mllt1 0.124156 0.043126 Dapk1 0.124513 0.047912 Klhl18 0.124718 0.04767 Pik3ca 0.124719 0.02558 Fastkd5 0.125223 0.030588 Nme1 0.12541 0.041183 1700011F03Rik 0.125413 0.033099 Rint1 0.125734 0.038098 Nxt1 0.125739 0.036926 Zfp294 0.12579 0.04059 Hmg20a 0.125886 0.035696 Them4 0.126169 0.022123 Prrg2 0.126269 0.046693 Shq1 0.126456 0.029047 Qrsl1 0.127063 0.044299 4932442K08Rik 0.127259 0.036141 Hspa9 0.127758 0.037511

185

Tsen34 0.128146 0.029161 Gmcl1 0.128941 0.035155 Lrrc14 0.129037 0.019641 Dhdds 0.129202 0.039461 Ap1gbp1 0.129365 0.037453 Nucb1 0.13063 0.036566 1500001M20Rik 0.130995 0.044226 Ercc3 0.131045 0.031274 Ccdc95 0.131078 0.043305 Tmod3 0.131546 0.042366 Tsfm 0.131547 0.045148 Gtf3c1 0.131584 0.039774 Kptn 0.131589 0.040549 Elp3 0.131826 0.044968 1700055N04Rik 0.132112 0.039167 Git2 0.13214 0.040988 Upf1 0.132318 0.049411 Thop1 0.133129 0.049438 Nudt16l1 0.13363 0.043024 Pi4kb 0.134159 0.046515 Fbl 0.134202 0.03727 Ccdc100 0.13424 0.049503 1810048J11Rik 0.134806 0.043728 Sh3bp5l 0.134844 0.043399 Centg2 0.134989 0.021942 Atp6v0a2 0.135077 0.03152 Rhod 0.135522 0.028938 Paf1 0.135794 0.028069 Zbtb7b 0.136074 0.048259 Eno1 0.136147 0.036161 D2Ertd391e 0.136286 0.033326 Cln3 0.13651 0.04254 Slc39a9 0.136768 0.049257 Yrdc 0.137142 0.038871 Spna2 0.137577 0.032885 Polr1e 0.13782 0.041291 Mett10d 0.137835 0.042412 Nol1 0.137869 0.024167 Csrp1 0.137982 0.046134 Gabpb2 0.138058 0.049216 Arhgef3 0.138089 0.033982 Sart1 0.13809 0.038761 Rexo2 0.138126 0.044614 Tnpo2 0.138605 0.017683 OTTMUSG00000014597 0.138823 0.04688 Snd1 0.138938 0.047877

186

Gak 0.139014 0.040527 Cdk5rap3 0.139181 0.047276 LOC626152 0.1392 0.020343 Gnptab 0.139236 0.042174 Nlk 0.139256 0.033774 Cdk4 0.139525 0.044831 2010204K13Rik 0.139802 0.044242 Mid1 0.139804 0.035844 Usp19 0.139812 0.033029 Ttll12 0.139827 0.03694 Arfgef2 0.14061 0.043868 Gfm1 0.141221 0.039332 Ccdc94 0.141817 0.04651 Gdap2 0.141874 0.030173 Frap1 0.141998 0.032449 Ext2 0.142562 0.04967 Ube2f 0.142669 0.043528 Harbi1 0.142808 0.04493 Spg11 0.142907 0.033682 Acot8 0.142999 0.013835 Cacybp 0.143023 0.043232 Tpp1 0.143109 0.031864 Dgcr2 0.143232 0.046134 Pdcd11 0.143309 0.018141 Mis12 0.143392 0.042237 Zfp592 0.144017 0.021354 Wdr18 0.144506 0.031796 Mta1 0.144638 0.025914 Ddx23 0.145099 0.035146 Sart3 0.145131 0.040823 Prcc 0.145448 0.045594 Mrpl12 0.145523 0.041594 Pi4ka 0.14573 0.039049 Cnpy2 0.145985 0.027466 Ccbl1 0.146702 0.048894 Eno1 0.146968 0.027711 Ddx54 0.147261 0.037068 Cldn12 0.147468 0.038897 Afg3l1 0.147709 0.020574 Erlin2 0.147777 0.0274 Ercc4 0.147908 0.041356 Txndc15 0.1484 0.037277 Tgfbrap1 0.148425 0.040823 6430548M08Rik 0.14848 0.042683 Gtpbp1 0.148534 0.024486 Rae1 0.148554 0.041487

187

Tbl1x 0.14861 0.034635 Cherp 0.148682 0.045564 Sertad3 0.148817 0.030332 Parvg 0.14891 0.039012 Timm44 0.148984 0.012696 Psmf1 0.149019 0.038718 100040879 0.149145 0.030285 C530043G21Rik 0.149188 0.038766 A630047E20Rik 0.149325 0.031199 Pold2 0.14938 0.040878 Zdhhc14 0.14969 0.036583 BC004004 0.149693 0.039818 Tm9sf4 0.149855 0.041851 Hmgcl 0.149949 0.035708 Tmem39b 0.15075 0.046603 Ranbp6 0.150953 0.021311 Clcn7 0.150983 0.042559 Ppan 0.151016 0.04712 Slbp 0.15139 0.031353 Sfmbt1 0.151401 0.037051 Wwc2 0.151619 0.044254 Mrps5 0.151685 0.025912 Pprc1 0.151835 0.049398 Spg7 0.151886 0.02422 Cdc6 0.152054 0.032011 U2af2 0.152239 0.024535 Parl 0.152326 0.041392 3930401K13Rik 0.152436 0.026263 Nat11 0.15256 0.024748 Uap1l1 0.152635 0.042928 Slc10a7 0.152855 0.044949 Sys1 0.152863 0.02729 Phb2 0.15288 0.045827 Heatr6 0.152981 0.021326 Sars 0.153143 0.024215 Noc3l 0.153163 0.047926 Slc7a5 0.153248 0.041601 Ascc3l1 0.153293 0.027426 Taf12 0.153563 0.037847 Eif2ak3 0.153766 0.02107 Cyp2r1 0.15404 0.04991 Prmt2 0.154103 0.045132 Ift172 0.154673 0.039101 Mxd4 0.154737 0.04093 Mrs2 0.155089 0.04074 Golgb1 0.155107 0.031654

188

BC057893 0.15549 0.036855 Gart 0.155734 0.019435 Rab35 0.155826 0.026213 Prelid2 0.155927 0.047611 Telo2 0.156058 0.032122 Mbtps1 0.156208 0.020969 Pgm1 0.156375 0.018483 Ctdspl 0.156454 0.029635 Pgam1 0.156672 0.042057 Wdsub1 0.156699 0.028709 Syf2 0.156832 0.028794 Prkd3 0.156923 0.015734 Syne2 0.156959 0.015794 Mlx 0.157009 0.043867 Man1b1 0.157035 0.013429 Jagn1 0.157061 0.042839 Ecd 0.157207 0.047009 Gpr180 0.157214 0.021657 Asna1 0.157632 0.04204 Prpf3 0.157746 0.017167 Tars2 0.157983 0.025168 Rragb 0.157986 0.017641 4933426G20Rik 0.158295 0.029172 Dcun1d3 0.159224 0.039988 Dsn1 0.159309 0.04338 Mettl1 0.159758 0.042011 Uprt 0.160212 0.027912 Ruvbl2 0.160239 0.027662 Prep 0.160353 0.049224 Abt1 0.160487 0.043915 EG627967 0.160734 0.030778 Mocs1 0.16096 0.032809 Htf9c 0.161037 0.03535 Ppig 0.161206 0.033893 Alg2 0.161331 0.035439 Pde8a 0.161436 0.032357 Pkp4 0.161474 0.043771 Dguok 0.161588 0.045048 Nudt9 0.161693 0.034849 Mtrr 0.161759 0.03142 Plscr1 0.162493 0.024197 Zc3h13 0.162594 0.028174 Myo1b 0.162867 0.041135 Qtrtd1 0.162958 0.026354 Adck1 0.162972 0.030937 Sap30bp 0.163011 0.048301

189

Ptrh2 0.163035 0.033961 Bfar 0.163191 0.02658 Letm1 0.163192 0.026863 Nbea 0.163377 0.021442 Tspan3 0.163397 0.03377 Ferd3l 0.163446 0.019131 Fcgr3 0.163462 0.019515 Eya3 0.163521 0.025625 Rfng 0.163548 0.020963 B4galnt1 0.163597 0.046726 Las1l 0.163629 0.039491 Tomm40l 0.163643 0.017323 Copg 0.163662 0.045051 Ptcd1 0.163696 0.045709 Rpp14 0.163881 0.046831 Diablo 0.163957 0.01016 Scamp4 0.164244 0.034631 Pak1ip1 0.16429 0.014029 Cntln 0.164712 0.019014 Dlst 0.165076 0.030032 Man2c1 0.165323 0.034068 Ap1s1 0.165402 0.025535 Ddx41 0.165417 0.030851 Slc5a6 0.165485 0.021729 Mynn 0.165654 0.035678 Lrsam1 0.165741 0.043691 Snx17 0.1658 0.046885 Sec14l2 0.165838 0.033057 Trmt12 0.166112 0.021584 2410089E03Rik 0.166141 0.018635 Edil3 0.166186 0.028447 OTTMUSG00000000657 0.166286 0.030957 Mrps34 0.166422 0.039373 Stx2 0.166612 0.021123 Mtap9 0.166679 0.03819 Ndnl2 0.166844 0.042542 D430028G21Rik 0.166864 0.039686 Tyw1 0.166954 0.027514 Dcps 0.167223 0.030852 Dnase1l3 0.167266 0.022199 Cdk10 0.167456 0.013717 Rpl7l1 0.167863 0.031678 Slc35b1 0.168002 0.043606 Lsm1 0.168058 0.043144 Sf3b4 0.168137 0.015351 Irf3 0.168192 0.027997

190

Slc41a1 0.168431 0.026539 Rangap1 0.16864 0.046945 Smarcb1 0.169435 0.041601 Zfp187 0.169487 0.049664 Agpat6 0.169571 0.036819 Wbp11 0.169572 0.018009 Ppp1r15b 0.16976 0.028112 Nr1h2 0.169761 0.014771 Alg1 0.170515 0.021831 Tsta3 0.170536 0.032247 Srm 0.170593 0.02691 Stom 0.170639 0.026045 ENSMUSG00000002791 0.1708 0.024111 OTTMUSG00000011448 0.171189 0.028061 Rbm14 0.171321 0.01617 Axl 0.171338 0.023961 Dohh 0.1714 0.04782 1110005A23Rik 0.171516 0.048555 Mrps7 0.171524 0.018134 Cdk10 0.17154 0.027311 Rnps1 0.171767 0.041419 Mall 0.171798 0.015388 Pak4 0.172075 0.011002 Myg1 0.172163 0.009687 D0HXS9928E 0.17224 0.025099 Clp1 0.172279 0.0248 Ccdc124 0.172873 0.029088 Slc35b4 0.173074 0.035356 Tubgcp4 0.173086 0.034285 Rapgef3 0.173114 0.027856 Mepce 0.17344 0.017564 Arpc4 0.173561 0.048851 Cyp4f13 0.173726 0.042788 Ndufa8 0.173852 0.043976 Pdia4 0.173875 0.028064 Gbp1 0.174241 0.035586 Samm50 0.174432 0.028002 Pop4 0.174551 0.037171 Ddx39 0.175007 0.01558 Fen1 0.175231 0.014628 Grwd1 0.175365 0.02675 Pot1a 0.175763 0.008593 Sipa1l1 0.176002 0.032916 Wdr60 0.176285 0.008975 Gars 0.176363 0.028814 Iqcf3 0.176491 0.012639

191

Cirh1a 0.176685 0.044937 Slc25a32 0.177442 0.025697 Chst12 0.177612 0.042645 Ppt2 0.177748 0.039483 Myl6 0.178049 0.043146 R3hcc1 0.17842 0.035288 Shkbp1 0.178554 0.0249 Bop1 0.178584 0.02047 Grpel1 0.178888 0.008574 Erp29 0.179383 0.040432 Tmem208 0.179422 0.043684 Rpl23 0.179905 0.025503 Chid1 0.180064 0.028734 Entpd6 0.18051 0.043709 St13 0.18056 0.041706 LOC100048603 0.18059 0.026119 Sap30l 0.180774 0.043282 Tut1 0.180873 0.014672 Tbc1d17 0.180873 0.020908 Slc38a7 0.181229 0.010068 Slc39a7 0.181284 0.015453 Adck4 0.181544 0.018258 Txndc5 0.182035 0.045309 Nagk 0.182219 0.029789 Gpr108 0.182229 0.033352 Ascc1 0.182629 0.036007 Nr2f6 0.183035 0.011741 Gpatch1 0.183106 0.007236 Exosc2 0.183309 0.017315 Cbr4 0.183379 0.043625 Mcm2 0.183678 0.029282 Mrpl10 0.183927 0.015367 Zbtb2 0.18428 0.031981 Rpl23 0.184748 0.025368 Gbp4 0.185008 0.019104 Nkap 0.185207 0.047763 Ldhb 0.185397 0.049263 3110040N11Rik 0.185492 0.008709 Myl9 0.186011 0.047443 Tmed9 0.186606 0.012084 AA673488 0.186954 0.006646 Jag1 0.187315 0.009638 Med31 0.187602 0.017537 1190005F20Rik 0.187761 0.014675 Sema4b 0.188019 0.021609 Adcy3 0.188038 0.003842

192

Prpf31 0.188154 0.02608 2510003E04Rik 0.188165 0.034753 Cope 0.188216 0.043806 Ncbp2 0.188538 0.020158 Dnaja3 0.18868 0.015687 Nrbp2 0.189344 0.025754 Phb 0.189654 0.019305 Mrps33 0.189903 0.036163 Creld2 0.189923 0.036598 Psmb10 0.190089 0.02757 Farsa 0.190113 0.02121 L3mbtl2 0.190608 0.022813 Zfp598 0.190634 0.007162 Slc19a1 0.190787 0.032634 Afap1 0.190852 0.046272 Pira3 0.191264 0.029933 Rrp1 0.191278 0.011801 Srprb 0.191573 0.013897 1810032O08Rik 0.191622 0.030746 Vac14 0.191673 0.036986 Pop5 0.19216 0.00899 Bag4 0.192436 0.027739 Stard8 0.19257 0.028681 Ggcx 0.193989 0.02121 Tbcb 0.194172 0.036798 Fancm 0.194972 0.02813 Terf2 0.195002 0.005099 Rai12 0.195155 0.026523 Gm1040 0.195402 0.020261 Shb 0.195844 0.021164 Tufm 0.195881 0.017073 Ppp4c 0.19673 0.034257 Aarsd1 0.196744 0.023321 Flot2 0.197162 0.040653 Usp5 0.197391 0.048784 Prmt6 0.197502 0.008074 Suclg1 0.197601 0.039126 Tmed3 0.197646 0.027516 Coq6 0.197881 0.037213 Lgi4 0.198667 0.024689 Prkcsh 0.198773 0.034519 Cotl1 0.199348 0.038495 Lrrc8c 0.199537 0.040733 Frrs1 0.200056 0.018751 5830443L24Rik 0.200302 0.036475 Faim 0.20038 0.043968

193

Upf3a 0.200386 0.007606 Dnpep 0.200513 0.021256 Ogfod1 0.201113 0.00799 0610031J06Rik 0.201319 0.047228 Mafg 0.201418 0.025585 Ssr2 0.202124 0.025697 Rpusd4 0.202474 0.022083 Cyp2j6 0.202597 0.047677 Hyou1 0.202612 0.045371 Il10 0.202723 0.021215 Phb 0.202861 0.022167 Aldh18a1 0.203108 0.016617 Imp4 0.20323 0.03771 Rad50 0.203656 0.00771 Syne2 0.203666 0.002704 Mrps33 0.203887 0.03079 2310011J03Rik 0.20411 0.014979 Nudt22 0.204336 0.033481 Angptl3 0.205283 0.042336 Oprs1 0.20552 0.028432 Lama4 0.205781 0.048876 Hadha 0.206074 0.046703 Prkaca 0.206075 0.035234 Taf5 0.206122 0.011102 Crip2 0.206462 0.043496 1110049F12Rik 0.207157 0.038472 Sf3a1 0.207168 0.019473 Wars 0.207391 0.013553 Tufm 0.207917 0.00845 4930504E06Rik 0.207996 0.007314 Hoxc8 0.208493 0.022119 Usp49 0.208552 0.016967 Pecr 0.209243 0.026803 9030617O03Rik 0.21009 0.031575 Mmaa 0.210159 0.027865 Slc39a10 0.210292 0.022427 Gtf2e1 0.210478 0.02468 Mtmr9 0.210951 0.013508 Capg 0.210993 0.012135 Gss 0.21106 0.047225 Qdpr 0.211114 0.028929 Crtap 0.211444 0.036781 Tagln 0.211552 0.037706 Npr2 0.211656 0.039991 Pla2g2d 0.21222 0.028729 Zfat 0.213478 0.008062

194

Ctnnbl1 0.213904 0.021964 Smpd1 0.214137 0.026484 Cgnl1 0.214221 0.033916 Lman2 0.215493 0.019381 Plod3 0.21662 0.010742 Kank3 0.216853 0.026816 Xpnpep3 0.217114 0.040996 Ddhd2 0.217159 0.048344 666761 0.217169 0.022746 Slc26a4 0.217243 0.007917 Katnal1 0.218036 0.010587 Psat1 0.218865 0.02825 N6amt2 0.218876 0.048091 Dus2l 0.219671 0.027533 A230067G21Rik 0.221068 0.005492 Crls1 0.221202 0.0458 Cyyr1 0.221216 0.035881 Mcm3 0.221294 0.02291 Spnb2 0.22147 0.033885 Taf11 0.222148 0.00541 Nat9 0.223007 0.012913 Gstm4 0.223024 0.039692 Gsto2 0.223216 0.036548 Zfp294 0.223403 0.006569 Prdm9 0.223682 0.010702 Mrpl22 0.22417 0.041541 Tubb2b 0.224257 0.028309 Phlda3 0.224282 0.039321 Slc22a5 0.225063 0.022274 Dock6 0.225138 0.013651 Itpk1 0.225147 0.048174 Zc3h12a 0.225451 0.007051 Cul7 0.225461 0.013604 Rnpep 0.225708 0.028374 Mut 0.225863 0.046398 Hsdl2 0.226774 0.045452 Gnb5 0.227771 0.016616 4932441K18Rik 0.229172 0.044844 Wrb 0.22968 0.008522 Bphl 0.229775 0.038775 Cth 0.23 0.014211 1500003O03Rik 0.23005 0.033848 ENSMUSG00000071392 0.231512 0.005138 Mtus1 0.233195 0.005718 Igfbp4 0.23433 0.046121 Cry2 0.234594 0.027204

195

Taldo1 0.23501 0.03101 Pomt1 0.235077 0.035157 Alas1 0.235864 0.026406 Cspp1 0.236014 0.009682 Clpb 0.237032 0.018523 Slc10a6 0.23715 0.036713 Nme7 0.238413 0.011987 Cxx1b 0.239926 0.04381 Mertk 0.241155 0.043101 Hrsp12 0.241477 0.04444 Fkbp2 0.242594 0.016428 Apex2 0.243413 0.008526 Has3 0.243589 0.013628 Pmm1 0.243678 0.011723 Ttpa 0.245695 0.007304 Creg1 0.246031 0.031509 Ttc27 0.246698 0.002841 Ablim3 0.247028 0.004043 Casp4 0.247236 0.019611 Xdh 0.247535 0.030437 Coasy 0.247776 0.024888 Ear1 0.247997 0.023517 Akr1e1 0.24816 0.009973 Drg2 0.248449 0.013033 Adra1a 0.248569 0.018977 Ampd2 0.248809 0.006669 Ppa1 0.249021 0.025931 Pcdhb22 0.250766 0.023919 Triap1 0.251814 0.014319 Pex5 0.251816 0.023299 Rpl13 0.252796 0.04254 D2hgdh 0.252819 0.015113 Sprr2j 0.253216 0.034419 Rpp40 0.253663 0.025597 Mtg1 0.255209 0.017277 Yipf3 0.25774 0.008746 Rnf5 0.259822 0.03606 Ptprg 0.260072 0.008184 Eral1 0.260981 0.006463 2310061C15Rik 0.261279 0.020172 Abhd5 0.262318 0.036199 Tmeff1 0.263592 0.04673 Atp10d 0.26418 0.008021 Degs2 0.264851 0.027302 Acadsb 0.265701 0.03014 Uchl5ip 0.267318 0.017758

196

Lonp1 0.267888 0.002086 Srpx 0.268348 0.024573 Pygl 0.268452 0.008866 Polr2i 0.269423 0.012731 Hint2 0.273091 0.014557 Agpat3 0.273309 0.04483 Snord35b 0.273385 0.03712 Tlr8 0.274775 0.049221 Cbr3 0.27629 0.018108 Fancl 0.277242 0.000845 Hsd17b4 0.279452 0.029691 Cyp4b1 0.280836 0.030512 Ptplb 0.281459 0.014965 Irak1bp1 0.283615 0.033624 Sntb1 0.283697 0.047064 Snrk 0.28382 0.00163 Trap1 0.286188 0.026225 Zfp637 0.286492 0.01544 Scarb1 0.286877 0.029164 Pold4 0.287706 0.043349 Atp9a 0.288252 0.036294 Gulp1 0.288254 0.000319 Ak2 0.291075 0.016629 4932441K18Rik 0.291305 0.01574 1810046J19Rik 0.292891 0.002116 Mccc2 0.293173 0.017009 Entpd5 0.294107 0.024939 Anxa3 0.295564 0.031017 Asns 0.295883 0.030919 Insr 0.296112 0.041332 Adamts9 0.296884 0.037056 Csad 0.297939 0.0257 G6pdx 0.298027 0.017837 Iltifb 0.300343 0.002404 Acaa2 0.300516 0.042505 F10 0.303798 0.048093 Mfsd8 0.304012 0.03293 Gnpat 0.306033 0.046253 Tspan4 0.306507 0.01808 Pp11r 0.308659 0.040583 Hadh 0.310518 0.043749 H6pd 0.316618 0.036383 Ntrk2 0.317723 0.044723 Comt 0.318351 0.041044 Leprel1 0.320726 0.008414 Palmd 0.322173 0.024766

197

2310042E22Rik 0.322243 0.026161 Steap4 0.323654 0.046589 Slc9a6 0.330506 0.047944 Tenc1 0.331184 0.037679 Bcar3 0.359238 0.045436 Mecr 0.360201 0.014395 Gbp2 0.363605 0.028334 Bscl2 0.365325 0.011735 Pex11a 0.379716 0.024288 Mgll 0.387841 0.038459 1100001G20Rik 0.390513 0.021786 Trf 0.390788 0.024652 Ppp2r1b 0.391846 0.032701 Pfkl 0.396348 0.00208 Gnai1 0.396788 0.016571 Ehhadh 0.400019 0.047034 Cpt2 0.412258 0.034615 EG626785 0.440782 0.001434 Gstt2 0.448686 0.045386 Id3 0.457744 0.024473 Sfxn1 0.459008 0.028413 Asah3l 0.47918 0.001688 Acaca 0.48718 0.025686 Acad11 0.500459 0.024214 Sfrp4 0.514222 0.035332 Mgl2 0.535517 0.018523 Dlc1 0.554076 0.044425 Slc1a3 0.560558 0.028196 Chi3l3 0.58156 0.022181 Aspa 0.584456 0.035218 Abca8a 0.607574 0.046606 Rbp7 0.692221 0.048274 Sdpr 0.737569 0.045736 Acsl1 0.740428 0.036117 Lipe 0.770016 0.042404 D430015B01Rik 0.795781 0.034967 Slc36a2 0.842079 0.046337 Retsat 0.899542 0.037957 Sycp3 0.909188 0.047779 Ucp1 1.118152 0.038291 Pck1 1.537555 0.033318

198

REFERENCES

1. Sarkar, N.K., D.D. Clarke, and H. Waelsch, An enzymically catalyzed incorporation of amines into proteins. Biochim Biophys Acta, 1957. 25(2): p. 451-2. 2. Lorand, L. and R.M. Graham, Transglutaminases: crosslinking enzymes with pleiotropic functions. Nat Rev Mol Cell Biol, 2003. 4(2): p. 140-56. 3. Aeschlimann, D., et al., Isolation of a cDNA encoding a novel member of the transglutaminase gene family from human keratinocytes. Detection and identification of transglutaminase gene products based on reverse transcription-polymerase chain reaction with degenerate primers. J Biol Chem, 1998. 273(6): p. 3452-60. 4. Grenard, P., M.K. Bates, and D. Aeschlimann, Evolution of transglutaminase genes: identification of a transglutaminase gene cluster on human chromosome 15q15. Structure of the gene encoding transglutaminase X and a novel gene family member, transglutaminase Z. J Biol Chem, 2001. 276(35): p. 33066-78. 5. Griffin, M., R. Casadio, and C.M. Bergamini, Transglutaminases: nature's biological glues. Biochem J, 2002. 368(Pt 2): p. 377-96. 6. Ohtake, Y., et al., Effect of retinoic acid-induced transglutaminase on cell growth in regenerating liver. Biomed Res, 2006. 27(2): p. 75-80. 7. Shin, D.M., et al., TGFbeta mediates activation of transglutaminase 2 in response to oxidative stress that leads to protein aggregation. FASEB J, 2008. 22(7): p. 2498-507. 8. Suto, N., et al., Identification of promoter region of guinea pig liver transglutaminase gene. Biochim Biophys Acta, 1993. 1172(3): p. 319-22. 9. Casadio, R., et al., The structural basis for the regulation of tissue transglutaminase by calcium ions. Eur J Biochem, 1999. 262(3): p. 672-9. 10. Verderio, E., et al., Regulated expression of tissue transglutaminase in Swiss 3T3 fibroblasts: effects on the processing of fibronectin, cell attachment, and cell death. Exp Cell Res, 1998. 239(1): p. 119- 38. 11. Iismaa, S.E., et al., Transglutaminases and disease: lessons from genetically engineered mouse models and inherited disorders. Physiol Rev, 2009. 89(3): p. 991-1023. 12. Liu, S., R.A. Cerione, and J. Clardy, Structural basis for the guanine nucleotide-binding activity of tissue transglutaminase and its regulation of transamidation activity. Proc Natl Acad Sci U S A, 2002. 99 (5): p. 2743-7. 13. Ikura, K., et al., Amino acid sequence of guinea pig liver transglutaminase from its cDNA sequence. Biochemistry, 1988. 27(8): p. 2898-905. 14. Begg, G.E., et al., Mutation of a critical arginine in the GTP-binding site of transglutaminase 2 disinhibits intracellular cross-linking activity. J Biol Chem, 2006. 281(18): p. 12603-9. 15. Im, M.J., R.P. Riek, and R.M. Graham, A novel guanine nucleotide-binding protein coupled to the alpha 1-adrenergic receptor. II. Purification, characterization, and reconstitution. J Biol Chem, 1990. 265(31): p. 18952-60. 16. Im, M.J. and R.M. Graham, A novel guanine nucleotide-binding protein coupled to the alpha 1- adrenergic receptor. I. Identification by photolabeling or membrane and ternary complex preparation. J Biol Chem, 1990. 265(31): p. 18944-51. 17. Singh, U.S., J.W. Erickson, and R.A. Cerione, Identification and biochemical characterization of an 80 kilodalton GTP-binding/transglutaminase from rabbit liver nuclei. Biochemistry, 1995. 34(48): p. 15863-71. 18. Krasnikov, B.F., et al., Transglutaminase activity is present in highly purified nonsynaptosomal mouse brain and liver mitochondria. Biochemistry, 2005. 44(21): p. 7830-43. 19. Gaudry, C.A., et al., Cell surface localization of tissue transglutaminase is dependent on a fibronectin-binding site in its N-terminal beta-sandwich domain. J Biol Chem, 1999. 274(43): p. 30707-14. 20. Zemskov, E.A., et al., Unconventional secretion of tissue transglutaminase involves phospholipid- dependent delivery into recycling endosomes. PLoS One. 6(4): p. e19414. 21. Laki, K. and L. Lorand, On the Solubility of Fibrin Clots. Science, 1948. 108(2802): p. 280. 22. Macfarlane, R.G., An Enzyme Cascade in the Blood Clotting Mechanism, and Its Function as a Biochemical Amplifier. Nature, 1964. 202: p. 498-9.

199

23. Nahrendorf, M., et al., Factor XIII deficiency causes cardiac rupture, impairs wound healing, and aggravates cardiac remodeling in mice with myocardial infarction. Circulation, 2006. 113(9): p. 1196-202. 24. Sakata, Y., J. Mimuro, and N. Aoki, Differential binding of plasminogen to crosslinked and noncrosslinked fibrins: its significance in hemostatic defect in factor XIII deficiency. Blood, 1984. 63(6): p. 1393-401. 25. Jessen, B.A., Q. Qin, and R.H. Rice, Functional AP1 and CRE response elements in the human keratinocyte transglutaminase promoter mediating Whn suppression. Gene, 2000. 254(1-2): p. 77- 85. 26. Russell, L.J., et al., Mutations in the gene for transglutaminase 1 in autosomal recessive lamellar ichthyosis. Nat Genet, 1995. 9(3): p. 279-83. 27. Lee, J.H., et al., The proximal promoter of the human transglutaminase 3 gene. Stratified squamous epithelial-specific expression in cultured cells is mediated by binding of Sp1 and ets transcription factors to a proximal promoter element. J Biol Chem, 1996. 271(8): p. 4561-8. 28. Liu, W., et al., Functional studies of a novel oncogene TGM3 in human esophageal squamous cell carcinoma. World J Gastroenterol, 2006. 12(24): p. 3929-32. 29. Candi, E., et al., Expression of transglutaminase 5 in normal and pathologic human epidermis. J Invest Dermatol, 2002. 119(3): p. 670-7. 30. Williams-Ashman, H.G., Transglutaminases and the clotting of mammalian seminal fluids. Mol Cell Biochem, 1984. 58(1-2): p. 51-61. 31. Cho, S.Y., et al., Differential alternative splicing of human transglutaminase 4 in benign prostate hyperplasia and prostate cancer. Exp Mol Med. 42(4): p. 310-8. 32. Hadjivassiliou, M., et al., Autoantibodies in gluten ataxia recognize a novel neuronal transglutaminase. Ann Neurol, 2008. 64(3): p. 332-43. 33. Rybicki, A.C., et al., Increased rotational mobility and extractability of band 3 from protein 4.2- deficient erythrocyte membranes: evidence of a role for protein 4.2 in strengthening the band 3- cytoskeleton linkage. Blood, 1996. 88(7): p. 2745-53. 34. Najfeld, V., et al., The gene for human erythrocyte protein 4.2 maps to chromosome 15q15. Am J Hum Genet, 1992. 50(1): p. 71-5. 35. Takaoka, Y., et al., A novel mutation in the erythrocyte protein 4.2 gene of Japanese patients with hereditary spherocytosis (protein 4.2 Fukuoka). Br J Haematol, 1994. 88(3): p. 527-33. 36. Bergamini, C.M., GTP modulates calcium binding and cation-induced conformational changes in erythrocyte transglutaminase. FEBS Lett, 1988. 239(2): p. 255-8. 37. Fesus, L. and K. Laki, Two antigenic sites of tissue transglutaminase. Biochemistry, 1977. 16(18): p. 4061-6. 38. Signorini, M., F. Dallocchio, and C.M. Bergamini, Inhibition of activity and quenching of intrinsic fluorescence of transglutaminase by acrylamide are independent events. Biochim Biophys Acta, 1988. 957(1): p. 168-71. 39. Bergamini, C.M. and M. Signorini, Calcium dependent reversible inactivation of erythrocyte transglutaminase by acrylamide. Biochem Int, 1988. 17(5): p. 855-62. 40. Tanfani, F., et al., Structural investigation of transglutaminase by Fourier transform infrared spectroscopy. Eur J Biochem, 1993. 218(2): p. 499-505. 41. Yee, V.C., et al., Structure and function studies of factor XIIIa by x-ray crystallography. Semin Thromb Hemost, 1996. 22(5): p. 377-84. 42. Mariani, P., et al., Ligand-induced conformational changes in tissue transglutaminase: Monte Carlo analysis of small-angle scattering data. Biophys J, 2000. 78(6): p. 3240-51. 43. Begg, G.E., et al., Mechanism of allosteric regulation of transglutaminase 2 by GTP. Proc Natl Acad Sci U S A, 2006. 103(52): p. 19683-8. 44. Pinkas, D.M., et al., Transglutaminase 2 undergoes a large conformational change upon activation. PLoS Biol, 2007. 5(12): p. e327. 45. Bergamini, C.M. and M. Signorini, Studies on tissue transglutaminases: interaction of erythrocyte type-2 transglutaminase with GTP. Biochem J, 1993. 291 ( Pt 1): p. 37-9. 46. Murthy, S.N., et al., Conserved tryptophan in the core domain of transglutaminase is essential for catalytic activity. Proc Natl Acad Sci U S A, 2002. 99(5): p. 2738-42. 47. Iismaa, S.E., et al., Evolutionary specialization of a tryptophan indole group for transition-state stabilization by eukaryotic transglutaminases. Proc Natl Acad Sci U S A, 2003. 100(22): p. 12636- 41. 48. !!! INVALID CITATION !!!

200

49. Noguchi, K., et al., Crystal structure of red sea bream transglutaminase. J Biol Chem, 2001. 276(15): p. 12055-9. 50. Safer, D., T.R. Sosnick, and M. Elzinga, Thymosin beta 4 binds actin in an extended conformation and contacts both the barbed and pointed ends. Biochemistry, 1997. 36(19): p. 5806-16. 51. Nemes, Z., Jr., et al., Identification of cytoplasmic actin as an abundant glutaminyl substrate for tissue transglutaminase in HL-60 and U937 cells undergoing apoptosis. J Biol Chem, 1997. 272(33): p. 20577-83. 52. Rasmussen, L.K., et al., Identification of glutamine and lysine residues in Alzheimer amyloid beta A4 peptide responsible for transglutaminase-catalysed homopolymerization and cross-linking to alpha 2M receptor. FEBS Lett, 1994. 338(2): p. 161-6. 53. Orban, J.M., et al., Crosslinking of collagen gels by transglutaminase. J Biomed Mater Res A, 2004. 68(4): p. 756-62. 54. Groenen, P.J., H. Bloemendal, and W.W. de Jong, The carboxy-terminal lysine of alpha B-crystallin is an amine-donor substrate for tissue transglutaminase. Eur J Biochem, 1992. 205(2): p. 671-4. 55. Stamnaes, J., et al., Redox regulation of transglutaminase 2 activity. J Biol Chem. 285(33): p. 25402- 9. 56. Hitomi, K., et al., Characterization of recombinant mouse epidermal-type transglutaminase (TGase 3): regulation of its activity by proteolysis and guanine nucleotides. J Biochem, 1999. 125(6): p. 1048-54. 57. Spina, A.M., et al., GTPase and transglutaminase are associated in the secretion of the rat anterior prostate. Biochem Biophys Res Commun, 1999. 260(2): p. 351-6. 58. Kim, H.C., et al., Protransglutaminase E from guinea pig skin. Isolation and partial characterization. J Biol Chem, 1990. 265(35): p. 21971-8. 59. Nakaoka, H., et al., Gh: a GTP-binding protein with transglutaminase activity and receptor signaling function. Science, 1994. 264(5165): p. 1593-6. 60. Ahvazi, B., et al., Structural basis for the coordinated regulation of transglutaminase 3 by guanine nucleotides and calcium/magnesium. J Biol Chem, 2004. 279(8): p. 7180-92. 61. Scheffzek, K., et al., The Ras-RasGAP complex: structural basis for GTPase activation and its loss in oncogenic Ras mutants. Science, 1997. 277(5324): p. 333-8. 62. Takeuchi, Y., et al., Putative nucleotide binding sites of guinea pig liver transglutaminase. FEBS Lett, 1992. 307(2): p. 177-80. 63. Chabre, M. and M. le Maire, Monomeric G-protein-coupled receptor as a functional unit. Biochemistry, 2005. 44(27): p. 9395-403. 64. McCudden, C.R., et al., G-protein signaling: back to the future. Cell Mol Life Sci, 2005. 62(5): p. 551-77. 65. Mhaouty-Kodja, S., Ghalpha/tissue transglutaminase 2: an emerging G protein in signal transduction. Biol Cell, 2004. 96(5): p. 363-7. 66. Iismaa, S.E., et al., The core domain of the tissue transglutaminase Gh hydrolyzes GTP and ATP. Biochemistry, 1997. 36(39): p. 11655-64. 67. Zhang, J., et al., Modulation of the in situ activity of tissue transglutaminase by calcium and GTP. J Biol Chem, 1998. 273(4): p. 2288-95. 68. Smethurst, P.A. and M. Griffin, Measurement of tissue transglutaminase activity in a permeabilized cell system: its regulation by Ca2+ and nucleotides. Biochem J, 1996. 313 ( Pt 3): p. 803-8. 69. Nicholas, B., et al., Cross-linking of cellular proteins by tissue transglutaminase during necrotic cell death: a mechanism for maintaining tissue integrity. Biochem J, 2003. 371(Pt 2): p. 413-22. 70. Ballestar, E. and L. Franco, Use of the transglutaminase reaction to study the dissociation of histone N-terminal tails from DNA in nucleosome core particles. Biochemistry, 1997. 36(20): p. 5963-9. 71. Jones, R.A., et al., Reduced expression of tissue transglutaminase in a human endothelial cell line leads to changes in cell spreading, cell adhesion and reduced polymerisation of fibronectin. J Cell Sci, 1997. 110 ( Pt 19): p. 2461-72. 72. Siegel, M., et al., Extracellular transglutaminase 2 is catalytically inactive, but is transiently activated upon tissue injury. PLoS One, 2008. 3(3): p. e1861. 73. Mishra, S. and L.J. Murphy, Tissue transglutaminase has intrinsic kinase activity: identification of transglutaminase 2 as an insulin-like growth factor-binding protein-3 kinase. J Biol Chem, 2004. 279(23): p. 23863-8. 74. Mishra, S., G. Melino, and L.J. Murphy, Transglutaminase 2 kinase activity facilitates protein kinase A-induced phosphorylation of retinoblastoma protein. J Biol Chem, 2007. 282(25): p. 18108-15. 75. Mishra, S. and L.J. Murphy, The p53 oncoprotein is a substrate for tissue transglutaminase kinase activity. Biochem Biophys Res Commun, 2006. 339(2): p. 726-30. 201

76. Mishra, S., et al., Phosphorylation of histones by tissue transglutaminase. J Biol Chem, 2006. 281(9): p. 5532-8. 77. Hasegawa, G., et al., A novel function of tissue-type transglutaminase: protein disulphide isomerase. Biochem J, 2003. 373(Pt 3): p. 793-803. 78. Mastroberardino, P.G., et al., "Tissue" transglutaminase contributes to the formation of disulphide bridges in proteins of mitochondrial respiratory complexes. Biochim Biophys Acta, 2006. 1757(9- 10): p. 1357-65. 79. Turner, P.M. and L. Lorand, Complexation of fibronectin with tissue transglutaminase. Biochemistry, 1989. 28(2): p. 628-35. 80. LeMosy, E.K., et al., Visualization of purified fibronectin-transglutaminase complexes. J Biol Chem, 1992. 267(11): p. 7880-5. 81. Akimov, S.S. and A.M. Belkin, Cell surface tissue transglutaminase is involved in adhesion and migration of monocytic cells on fibronectin. Blood, 2001. 98(5): p. 1567-76. 82. Akimov, S.S. and A.M. Belkin, Cell-surface transglutaminase promotes fibronectin assembly via interaction with the gelatin-binding domain of fibronectin: a role in TGFbeta-dependent matrix deposition. J Cell Sci, 2001. 114(Pt 16): p. 2989-3000. 83. Akimov, S.S., et al., Tissue transglutaminase is an integrin-binding adhesion coreceptor for fibronectin. J Cell Biol, 2000. 148(4): p. 825-38. 84. Bass, M.D., M.R. Morgan, and M.J. Humphries, Integrins and syndecan-4 make distinct, but critical, contributions to adhesion contact formation. Soft Matter, 2007. 3(3): p. 372-376. 85. Telci, D., et al., Fibronectin-tissue transglutaminase matrix rescues RGD-impaired cell adhesion through syndecan-4 and beta1 integrin co-signaling. J Biol Chem, 2008. 283(30): p. 20937-47. 86. Xu, L., et al., GPR56, an atypical G protein-coupled receptor, binds tissue transglutaminase, TG2, and inhibits melanoma tumor growth and metastasis. Proc Natl Acad Sci U S A, 2006. 103(24): p. 9023-8. 87. Nanda, N., et al., Targeted inactivation of Gh/tissue transglutaminase II. J Biol Chem, 2001. 276(23): p. 20673-8. 88. De Laurenzi, V. and G. Melino, Gene disruption of tissue transglutaminase. Mol Cell Biol, 2001. 21(1): p. 148-55. 89. Bailey, C.D. and G.V. Johnson, Developmental regulation of tissue transglutaminase in the mouse forebrain. J Neurochem, 2004. 91(6): p. 1369-79. 90. Rewers, M., Epidemiology of celiac disease: what are the prevalence, incidence, and progression of celiac disease? Gastroenterology, 2005. 128(4 Suppl 1): p. S47-51. 91. van Heel, D.A. and J. West, Recent advances in coeliac disease. Gut, 2006. 55(7): p. 1037-46. 92. Skovbjerg, H., et al., Gliadin is a good substrate of several transglutaminases: possible implication in the pathogenesis of coeliac disease. Scand J Gastroenterol, 2002. 37(7): p. 812-7. 93. Fleckenstein, B., et al., Gliadin T cell epitope selection by tissue transglutaminase in celiac disease. Role of enzyme specificity and pH influence on the transamidation versus deamidation process. J Biol Chem, 2002. 277(37): p. 34109-16. 94. Koning, F., et al., Pathomechanisms in celiac disease. Best Pract Res Clin Gastroenterol, 2005. 19(3): p. 373-87. 95. Sollid, L.M., et al., Autoantibodies in coeliac disease: tissue transglutaminase--guilt by association? Gut, 1997. 41(6): p. 851-2. 96. Zanoni, G., et al., In celiac disease, a subset of autoantibodies against transglutaminase binds toll- like receptor 4 and induces activation of monocytes. PLoS Med, 2006. 3(9): p. e358. 97. van der Windt, D.A., et al., Diagnostic testing for celiac disease among patients with abdominal symptoms: a systematic review. JAMA. 303(17): p. 1738-46. 98. Boros, S., et al., Tissue transglutaminase catalyzes the deamidation of glutamines in lens betaB(2)- and betaB(3)-crystallins. Exp Eye Res, 2008. 86(2): p. 383-93. 99. Shin, D.M., et al., Cell type-specific activation of intracellular transglutaminase 2 by oxidative stress or ultraviolet irradiation: implications of transglutaminase 2 in age-related cataractogenesis. J Biol Chem, 2004. 279(15): p. 15032-9. 100. Sohn, J., et al., Novel transglutaminase inhibitors reverse the inflammation of allergic conjunctivitis. J Clin Invest, 2003. 111(1): p. 121-8. 101. Png, E., et al., Hyperosmolarity mediated mitochondrial dysfunction requires transglutaminase-2 in human corneal epithelial cells. J Cell Physiol. 102. Barathi, V.A., et al., Transglutaminases (TGs) in ocular and periocular tissues: effect of muscarinic agents on TGs in scleral fibroblasts. PLoS One. 6(4): p. e18326. 103. Hanahan, D. and R.A. Weinberg, The hallmarks of cancer. Cell, 2000. 100(1): p. 57-70. 202

104. Kotsakis, P. and M. Griffin, Tissue transglutaminase in tumour progression: friend or foe? Amino Acids, 2007. 33(2): p. 373-84. 105. Belkin, A.M., et al., Cell-surface-associated tissue transglutaminase is a target of MMP-2 proteolysis. Biochemistry, 2004. 43(37): p. 11760-9. 106. Mehta, K., et al., Prognostic significance of tissue transglutaminase in drug resistant and metastatic breast cancer. Clin Cancer Res, 2004. 10(23): p. 8068-76. 107. Hunter, I., et al., The cell adhesion molecule C-CAM is a substrate for tissue transglutaminase. FEBS Lett, 1998. 425(1): p. 141-4. 108. Jiang, D., et al., Identification of metastasis-associated proteins by proteomic analysis and functional exploration of interleukin-18 in metastasis. Proteomics, 2003. 3(5): p. 724-37. 109. Otto, T., et al., Stabilization of N-Myc is a critical function of Aurora A in human neuroblastoma. Cancer Cell, 2009. 15(1): p. 67-78. 110. Liu, T., et al., Activation of tissue transglutaminase transcription by histone deacetylase inhibition as a therapeutic approach for Myc oncogenesis. Proc Natl Acad Sci U S A, 2007. 104(47): p. 18682-7. 111. Jones, R.A., et al., Matrix changes induced by transglutaminase 2 lead to inhibition of angiogenesis and tumor growth. Cell Death Differ, 2006. 13(9): p. 1442-53. 112. Grigoriev, M.Y., et al., Tissue transglutaminase expression in breast carcinomas. J Exp Clin Cancer Res, 2001. 20(2): p. 265-8. 113. Piacentini, M., et al., Transglutaminase overexpression sensitizes neuronal cell lines to apoptosis by increasing mitochondrial membrane potential and cellular oxidative stress. J Neurochem, 2002. 81(5): p. 1061-72. 114. Fesus, L. and Z. Szondy, Transglutaminase 2 in the balance of cell death and survival. FEBS Lett, 2005. 579(15): p. 3297-302. 115. Oliverio, S., et al., Tissue transglutaminase-dependent posttranslational modification of the retinoblastoma gene product in promonocytic cells undergoing apoptosis. Mol Cell Biol, 1997. 17(10): p. 6040-8. 116. Mann, A.P., et al., Overexpression of tissue transglutaminase leads to constitutive activation of nuclear factor-kappaB in cancer cells: delineation of a novel pathway. Cancer Res, 2006. 66(17): p. 8788-95. 117. Weiss, L., Metastasis of cancer: a conceptual history from antiquity to the 1990s. Cancer Metastasis Rev, 2000. 19(3-4): p. I-XI, 193-383. 118. Belkin, A.M., et al., Matrix-dependent proteolysis of surface transglutaminase by membrane-type metalloproteinase regulates cancer cell adhesion and locomotion. J Biol Chem, 2001. 276(21): p. 18415-22. 119. Di Giacomo, G., et al., In vivo evaluation of type 2 transglutaminase contribution to the metastasis formation in melanoma. Amino Acids, 2009. 36(4): p. 717-24. 120. Verderio, E.A., et al., A novel RGD-independent cel adhesion pathway mediated by fibronectin- bound tissue transglutaminase rescues cells from anoikis. J Biol Chem, 2003. 278(43): p. 42604-14. 121. Tee, A.E., et al., Opposing effects of two tissue transglutaminase protein isoforms in neuroblastoma cell differentiation. J Biol Chem. 285(6): p. 3561-7. 122. Citron, B.A., et al., Intron-exon swapping of transglutaminase mRNA and neuronal Tau aggregation in Alzheimer's disease. J Biol Chem, 2001. 276(5): p. 3295-301. 123. Olsen, B.R., A.M. Reginato, and W. Wang, Bone development. Annu Rev Cell Dev Biol, 2000. 16: p. 191-220. 124. Johnson, K.A., et al., Distinct transglutaminase 2-independent and transglutaminase 2-dependent pathways mediate articular chondrocyte hypertrophy. J Biol Chem, 2003. 278(21): p. 18824-32. 125. Aeschlimann, D., et al., Expression of tissue transglutaminase in skeletal tissues correlates with events of terminal differentiation of chondrocytes. J Cell Biol, 1993. 120(6): p. 1461-70. 126. Nurminskaya, M., et al., Plasma transglutaminase in hypertrophic chondrocytes: expression and cell-specific intracellular activation produce cell death and externalization. J Cell Biol, 1998. 142(4): p. 1135-44. 127. Triffitt, J.T., Initiation and enhancement of bone formation. A review. Acta Orthop Scand, 1987. 58(6): p. 673-84. 128. Heath, D.J., et al., Characterization of tissue transglutaminase in human osteoblast-like cells. J Bone Miner Res, 2001. 16(8): p. 1477-85. 129. Nurminskaya, M., et al., Chondrocyte-derived transglutaminase promotes maturation of preosteoblasts in periosteal bone. Dev Biol, 2003. 263(1): p. 139-52. 130. Nurminskaya, M. and M.T. Kaartinen, Transglutaminases in mineralized tissues. Front Biosci, 2006. 11: p. 1591-606. 203

131. Al-Jallad, H.F., et al., Transglutaminase activity regulates osteoblast differentiation and matrix mineralization in MC3T3-E1 osteoblast cultures. Matrix Biol, 2006. 25(3): p. 135-48. 132. Tanaka, K., et al., The integrin alpha5beta1 regulates chondrocyte hypertrophic differentiation induced by GTP-bound transglutaminase 2. Matrix Biol, 2007. 26(6): p. 409-18. 133. Johnson, K.A. and R.A. Terkeltaub, External GTP-bound transglutaminase 2 is a molecular switch for chondrocyte hypertrophic differentiation and calcification. J Biol Chem, 2005. 280(15): p. 15004-12. 134. Johnson, K.A., D.M. Rose, and R.A. Terkeltaub, Factor XIIIA mobilizes transglutaminase 2 to induce chondrocyte hypertrophic differentiation. J Cell Sci, 2008. 121(Pt 13): p. 2256-64. 135. Poole, A.R., An introduction to the pathophysiology of osteoarthritis. Front Biosci, 1999. 4: p. D662- 70. 136. Orlandi, A., et al., Transglutaminase-2 differently regulates cartilage destruction and osteophyte formation in a surgical model of osteoarthritis. Amino Acids, 2009. 36(4): p. 755-63. 137. Cecil, D.L. and R. Terkeltaub, Transamidation by transglutaminase 2 transforms S100A11 calgranulin into a procatabolic cytokine for chondrocytes. J Immunol, 2008. 180(12): p. 8378-85. 138. Boisvert, W.A., et al., Leukocyte transglutaminase 2 expression limits atherosclerotic lesion size. Arterioscler Thromb Vasc Biol, 2006. 26(3): p. 563-9. 139. Sumi, Y., et al., Expression of tissue transglutaminase and elafin in human coronary artery: implication for plaque instability. Atherosclerosis, 2002. 160(1): p. 31-9. 140. Van Herck, J.L., et al., Transglutaminase 2 deficiency decreases plaque fibrosis and increases plaque inflammation in apolipoprotein-E-deficient mice. J Vasc Res. 47(3): p. 231-40. 141. Williams, H., et al., Effect of transglutaminase 2 (TG2) deficiency on atherosclerotic plaque stability in the apolipoprotein E deficient mouse. Atherosclerosis. 210(1): p. 94-9. 142. Szondy, Z., et al., Transglutaminase 2-/- mice reveal a phagocytosis-associated crosstalk between macrophages and apoptotic cells. Proc Natl Acad Sci U S A, 2003. 100(13): p. 7812-7. 143. Abedin, M., Y. Tintut, and L.L. Demer, Vascular calcification: mechanisms and clinical ramifications. Arterioscler Thromb Vasc Biol, 2004. 24(7): p. 1161-70. 144. Johnson, K.A., M. Polewski, and R.A. Terkeltaub, Transglutaminase 2 is central to induction of the arterial calcification program by smooth muscle cells. Circ Res, 2008. 102(5): p. 529-37. 145. Faverman, L., et al., Extracellular transglutaminase 2 activates beta-catenin signaling in calcifying vascular smooth muscle cells. FEBS Lett, 2008. 582(10): p. 1552-7. 146. Iwai, N., H. Shimoike, and M. Kinoshita, Genes up-regulated in hypertrophied ventricle. Biochem Biophys Res Commun, 1995. 209(2): p. 527-34. 147. Hwang, K.C., et al., Alpha 1-adrenergic receptor coupling with Gh in the failing human heart. Circulation, 1996. 94(4): p. 718-26. 148. Small, K., et al., Cardiac specific overexpression of transglutaminase II (G(h)) results in a unique hypertrophy phenotype independent of phospholipase C activation. J Biol Chem, 1999. 274(30): p. 21291-6. 149. Zhang, Z., et al., COX-2-dependent cardiac failure in Gh/tTG transgenic mice. Circ Res, 2003. 92(10): p. 1153-61. 150. A novel gene containing a trinucleotide repeat that is expanded and unstable on Huntington's disease chromosomes. The Huntington's Disease Collaborative Research Group. Cell, 1993. 72(6): p. 971-83. 151. Munsie, L., et al., Mutant huntingtin causes defective actin remodeling during stress: defining a new role for transglutaminase 2 in neurodegenerative disease. Hum Mol Genet. 20(10): p. 1937-51. 152. Mastroberardino, P.G., et al., 'Tissue' transglutaminase ablation reduces neuronal death and prolongs survival in a mouse model of Huntington's disease. Cell Death Differ, 2002. 9(9): p. 873- 80. 153. Bailey, C.D. and G.V. Johnson, Tissue transglutaminase contributes to disease progression in the R6/2 Huntington's disease mouse model via aggregate-independent mechanisms. J Neurochem, 2005. 92(1): p. 83-92. 154. Wang, D.S., D.W. Dickson, and J.S. Malter, Tissue transglutaminase, protein cross-linking and Alzheimer's disease: review and views. Int J Clin Exp Pathol, 2008. 1(1): p. 5-18. 155. Junn, E., et al., Tissue transglutaminase-induced aggregation of alpha-synuclein: Implications for Lewy body formation in Parkinson's disease and dementia with Lewy bodies. Proc Natl Acad Sci U S A, 2003. 100(4): p. 2047-52. 156. Fujita, K., et al., Transglutaminase activity in serum and cerebrospinal fluid in sporadic amyotrophic lateral sclerosis: a possible use as an indicator of extent of the motor neuron loss. J Neurol Sci, 1998. 158(1): p. 53-7. 204

157. Johnson, G.V., et al., Transglutaminase activity is increased in Alzheimer's disease brain. Brain Res, 1997. 751(2): p. 323-9. 158. Wilhelmus, M.M., et al., Transglutaminases and transglutaminase-catalyzed cross-links colocalize with the pathological lesions in Alzheimer's disease brain. Brain Pathol, 2009. 19(4): p. 612-22. 159. Andringa, G., et al., Tissue transglutaminase catalyzes the formation of alpha-synuclein crosslinks in Parkinson's disease. FASEB J, 2004. 18(7): p. 932-4. 160. Haroon, Z.A., et al., Tissue transglutaminase is expressed, active, and directly involved in rat dermal wound healing and angiogenesis. FASEB J, 1999. 13(13): p. 1787-95. 161. Aeschlimann, D., O. Kaupp, and M. Paulsson, Transglutaminase-catalyzed matrix cross-linking in differentiating cartilage: identification of osteonectin as a major glutaminyl substrate. J Cell Biol, 1995. 129(3): p. 881-92. 162. Raghunath, M., et al., Cross-linking of the dermo-epidermal junction of skin regenerating from keratinocyte autografts. Anchoring fibrils are a target for tissue transglutaminase. J Clin Invest, 1996. 98(5): p. 1174-84. 163. Madison, K.C., Barrier function of the skin: "la raison d'etre" of the epidermis. J Invest Dermatol, 2003. 121(2): p. 231-41. 164. Martin, P., Wound healing--aiming for perfect skin regeneration. Science, 1997. 276(5309): p. 75- 81. 165. Proksch, E., J.M. Brandner, and J.M. Jensen, The skin: an indispensable barrier. Exp Dermatol, 2008. 17(12): p. 1063-72. 166. Braverman, I.M., The cutaneous microcirculation. J Investig Dermatol Symp Proc, 2000. 5(1): p. 3- 9. 167. Skobe, M. and M. Detmar, Structure, function, and molecular control of the skin lymphatic system. J Investig Dermatol Symp Proc, 2000. 5(1): p. 14-9. 168. Clark, L.D., R.K. Clark, and E. Heber-Katz, A new murine model for mammalian wound repair and regeneration. Clin Immunol Immunopathol, 1998. 88(1): p. 35-45. 169. Eming, S.A., T. Krieg, and J.M. Davidson, Inflammation in wound repair: molecular and cellular mechanisms. J Invest Dermatol, 2007. 127(3): p. 514-25. 170. Singer, A.J. and R.A. Clark, Cutaneous wound healing. N Engl J Med, 1999. 341(10): p. 738-46. 171. Ryan, G.B. and G. Majno, Acute inflammation. A review. Am J Pathol, 1977. 86(1): p. 183-276. 172. Diegelmann, R.F. and M.C. Evans, Wound healing: an overview of acute, fibrotic and delayed healing. Front Biosci, 2004. 9: p. 283-9. 173. Heldin, C.H. and B. Westermark, Mechanism of action and in vivo role of platelet-derived growth factor. Physiol Rev, 1999. 79(4): p. 1283-316. 174. Clark, R.A., et al., Fibronectin and fibrin provide a provisional matrix for epidermal cell migration during wound reepithelialization. J Invest Dermatol, 1982. 79(5): p. 264-9. 175. Labow, M.A., et al., Characterization of E-selectin-deficient mice: demonstration of overlapping function of the endothelial selectins. Immunity, 1994. 1(8): p. 709-20. 176. Leibovich, S.J. and R. Ross, The role of the macrophage in wound repair. A study with hydrocortisone and antimacrophage serum. Am J Pathol, 1975. 78(1): p. 71-100. 177. Mayadas, T.N., et al., Leukocyte rolling and extravasation are severely compromised in P selectin- deficient mice. Cell, 1993. 74(3): p. 541-54. 178. Witko-Sarsat, V., et al., Neutrophils: molecules, functions and pathophysiological aspects. Lab Invest, 2000. 80(5): p. 617-53. 179. Malaviya, R., et al., Mast cell modulation of neutrophil influx and bacterial clearance at sites of infection through TNF-alpha. Nature, 1996. 381(6577): p. 77-80. 180. Brown, E.J., Phagocytosis. Bioessays, 1995. 17(2): p. 109-17. 181. Kopf, M., et al., Impaired immune and acute-phase responses in interleukin-6-deficient mice. Nature, 1994. 368(6469): p. 339-42. 182. Rogge, R.D., C.A. Karlovich, and U. Banerjee, Genetic dissection of a neurodevelopmental pathway: Son of sevenless functions downstream of the sevenless and EGF receptor tyrosine kinases. Cell, 1991. 64(1): p. 39-48. 183. Jones, S.A., Directing transition from innate to acquired immunity: defining a role for IL-6. J Immunol, 2005. 175(6): p. 3463-8. 184. Clark, R.A., Fibronectin matrix deposition and fibronectin receptor expression in healing and normal skin. J Invest Dermatol, 1990. 94(6 Suppl): p. 128S-134S. 185. Larjava, H., et al., Expression of integrins and basement membrane components by wound keratinocytes. J Clin Invest, 1993. 92(3): p. 1425-35.

205

186. Pedersen, H., et al., Prognostic impact of urokinase, urokinase receptor, and type 1 plasminogen activator inhibitor in squamous and large cell lung cancer tissue. Cancer Res, 1994. 54(17): p. 4671- 5. 187. Mitchison, T.J. and L.P. Cramer, Actin-based cell motility and cell locomotion. Cell, 1996. 84(3): p. 371-9. 188. Xu, J. and R.A. Clark, Extracellular matrix alters PDGF regulation of fibroblast integrins. J Cell Biol, 1996. 132(1-2): p. 239-49. 189. Werner, S., et al., Large induction of keratinocyte growth factor expression in the dermis during wound healing. Proc Natl Acad Sci U S A, 1992. 89(15): p. 6896-900. 190. Breuss, J.M., et al., Expression of the beta 6 integrin subunit in development, neoplasia and tissue repair suggests a role in epithelial remodeling. J Cell Sci, 1995. 108 ( Pt 6): p. 2241-51. 191. Greiling, D. and R.A. Clark, Fibronectin provides a conduit for fibroblast transmigration from collagenous stroma into fibrin clot provisional matrix. J Cell Sci, 1997. 110 ( Pt 7): p. 861-70. 192. Davis, G.E., et al., Regulation of tissue injury responses by the exposure of matricryptic sites within extracellular matrix molecules. Am J Pathol, 2000. 156(5): p. 1489-98. 193. Davis, G.E., Affinity of integrins for damaged extracellular matrix: alpha v beta 3 binds to denatured collagen type I through RGD sites. Biochem Biophys Res Commun, 1992. 182(3): p. 1025-31. 194. Santoro, S.A., Identification of a 160,000 dalton platelet membrane protein that mediates the initial divalent cation-dependent adhesion of platelets to collagen. Cell, 1986. 46(6): p. 913-20. 195. Gipson, I.K., S.J. Spurr-Michaud, and A.S. Tisdale, Hemidesmosomes and anchoring fibril collagen appear synchronously during development and wound healing. Dev Biol, 1988. 126(2): p. 253-62. 196. Grinnell, F., Wound repair, keratinocyte activation and integrin modulation. J Cell Sci, 1992. 101 ( Pt 1): p. 1-5. 197. Ffrench-Constant, C., et al., Reappearance of an embryonic pattern of fibronectin splicing during wound healing in the adult rat. J Cell Biol, 1989. 109(2): p. 903-14. 198. Oksala, O., et al., Expression of proteoglycans and hyaluronan during wound healing. J Histochem Cytochem, 1995. 43(2): p. 125-35. 199. Amadeu, T.P., et al., Cutaneous wound healing: myofibroblastic differentiation and in vitro models. Int J Low Extrem Wounds, 2003. 2(2): p. 60-8. 200. Grinnell, F., Fibroblasts, myofibroblasts, and wound contraction. J Cell Biol, 1994. 124(4): p. 401-4. 201. Madlener, M., W.C. Parks, and S. Werner, Matrix metalloproteinases (MMPs) and their physiological inhibitors (TIMPs) are differentially expressed during excisional skin wound repair. Exp Cell Res, 1998. 242(1): p. 201-10. 202. Barnes, M.J., L.F. Morton, and C.I. Levene, Synthesis of collagens types I and III by pig medial smooth muscle cells in culture. Biochem Biophys Res Commun, 1976. 70(2): p. 339-47. 203. Barnes, M.J., et al., Presence of type III collagen in guinea-pig dermal scar. Biochem J, 1976. 157(1): p. 263-6. 204. Fleischmajer, R., et al., Type I and type III collagen interactions during fibrillogenesis. Ann N Y Acad Sci, 1990. 580: p. 161-75. 205. Horslev-Petersen, K., et al., Collagen type IV and procollagen type III during granulation tissue formation: a serological, biochemical, immunohistochemical and morphometrical study on the viscose cellulose sponge rat model. Eur J Clin Invest, 1988. 18(4): p. 352-9. 206. White, J.F., et al., Collagen fibril formation in a wound healing model. J Struct Biol, 2002. 137(1-2): p. 23-30. 207. Berry, D.P., et al., Human wound contraction: collagen organization, fibroblasts, and myofibroblasts. Plast Reconstr Surg, 1998. 102(1): p. 124-31; discussion 132-4. 208. Shah, M., D.M. Foreman, and M.W. Ferguson, Neutralising antibody to TGF-beta 1,2 reduces cutaneous scarring in adult rodents. J Cell Sci, 1994. 107 ( Pt 5): p. 1137-57. 209. Shah, M., D.M. Foreman, and M.W. Ferguson, Neutralisation of TGF-beta 1 and TGF-beta 2 or exogenous addition of TGF-beta 3 to cutaneous rat wounds reduces scarring. J Cell Sci, 1995. 108 ( Pt 3): p. 985-1002. 210. Borradori, L. and A. Sonnenberg, Structure and function of hemidesmosomes: more than simple adhesion complexes. J Invest Dermatol, 1999. 112(4): p. 411-8. 211. Muszbek, L., V.C. Yee, and Z. Hevessy, Blood coagulation factor XIII: structure and function. Thromb Res, 1999. 94(5): p. 271-305. 212. Auld, G.C., et al., Thrombin upregulates tissue transglutaminase in endothelial cells: a potential role for tissue transglutaminase in stability of atherosclerotic plaque. Arterioscler Thromb Vasc Biol, 2001. 21(10): p. 1689-94.

206

213. Belkin, A.M., et al., Transglutaminase-mediated oligomerization of the fibrin(ogen) alphaC domains promotes integrin-dependent cell adhesion and signaling. Blood, 2005. 105(9): p. 3561-8. 214. Diaz, B.L. and J.P. Arm, Phospholipase A(2). Prostaglandins Leukot Essent Fatty Acids, 2003. 69(2- 3): p. 87-97. 215. Cordella-Miele, E., et al., Transglutaminase-catalyzed incorporation of polyamines into phospholipase A2. J Biochem, 1993. 113(2): p. 164-73. 216. Kim, S.Y., E.J. Jeong, and P.M. Steinert, IFN-gamma induces transglutaminase 2 expression in rat small intestinal cells. J Interferon Cytokine Res, 2002. 22(6): p. 677-82. 217. Suto, N., K. Ikura, and R. Sasaki, Expression induced by interleukin-6 of tissue-type transglutaminase in human hepatoblastoma HepG2 cells. J Biol Chem, 1993. 268(10): p. 7469-73. 218. Mohan, K., D. Pinto, and T.B. Issekutz, Identification of tissue transglutaminase as a novel molecule involved in human CD8+ T cell transendothelial migration. J Immunol, 2003. 171(6): p. 3179-86. 219. Seiving, B., et al., Transglutaminase differentiation during maturation of human blood monocytes to macrophages. Eur J Haematol, 1991. 46(5): p. 263-71. 220. Murtaugh, M.P., et al., Induction of tissue transglutaminase in mouse peritoneal macrophages. J Biol Chem, 1983. 258(18): p. 11074-81. 221. Letterio, J.J. and A.B. Roberts, Regulation of immune responses by TGF-beta. Annu Rev Immunol, 1998. 16: p. 137-61. 222. Telci, D., et al., Increased TG2 expression can result in induction of transforming growth factor beta1, causing increased synthesis and deposition of matrix proteins, which can be regulated by nitric oxide. J Biol Chem, 2009. 284(43): p. 29547-58. 223. Verderio, E., et al., Regulation of cell surface tissue transglutaminase: effects on matrix storage of latent transforming growth factor-beta binding protein-1. J Histochem Cytochem, 1999. 47(11): p. 1417-32. 224. Rose, D.M., et al., Transglutaminase 2 limits murine peritoneal acute gout-like inflammation by regulating macrophage clearance of apoptotic neutrophils. Arthritis Rheum, 2006. 54(10): p. 3363- 71. 225. Corbett, S.A. and J.E. Schwarzbauer, Fibronectin-fibrin cross-linking: a regulator of cell behavior. Trends Cardiovasc Med, 1998. 8(8): p. 357-62. 226. McClain, S.A., et al., Mesenchymal cell activation is the rate-limiting step of granulation tissue induction. Am J Pathol, 1996. 149(4): p. 1257-70. 227. Collighan, R.J. and M. Griffin, Transglutaminase 2 cross-linking of matrix proteins: biological significance and medical applications. Amino Acids, 2009. 36(4): p. 659-70. 228. Kleman, J.P., et al., Transglutaminase-catalyzed cross-linking of fibrils of collagen V/XI in A204 rhabdomyosarcoma cells. Biochemistry, 1995. 34(42): p. 13768-75. 229. Kaartinen, M.T., et al., Cross-linking of osteopontin by tissue transglutaminase increases its collagen binding properties. J Biol Chem, 1999. 274(3): p. 1729-35. 230. Aeschlimann, D. and M. Paulsson, Cross-linking of laminin-nidogen complexes by tissue transglutaminase. A novel mechanism for basement membrane stabilization. J Biol Chem, 1991. 266(23): p. 15308-17. 231. Aeschlimann, D., M. Paulsson, and K. Mann, Identification of Gln726 in nidogen as the amine acceptor in transglutaminase-catalyzed cross-linking of laminin-nidogen complexes. J Biol Chem, 1992. 267(16): p. 11316-21. 232. Gross, S.R., Z. Balklava, and M. Griffin, Importance of tissue transglutaminase in repair of extracellular matrices and cell death of dermal fibroblasts after exposure to a solarium ultraviolet A source. J Invest Dermatol, 2003. 121(2): p. 412-23. 233. Telci, D. and M. Griffin, Tissue transglutaminase (TG2)--a wound response enzyme. Front Biosci, 2006. 11: p. 867-82. 234. Verderio, E.A., T.S. Johnson, and M. Griffin, Transglutaminases in wound healing and inflammation. Prog Exp Tumor Res, 2005. 38: p. 89-114. 235. Mearns, B.M., et al., Impaired wound healing and altered fibroblast cytoskeletal dynamics in Gh knockout mice. Minerva Biotechnologica, 2002. 14(2): p. 1. 236. Sarang, Z., et al., Tissue transglutaminase (TG2) facilitates phosphatidylserine exposure and calpain activity in calcium-induced death of erythrocytes. Cell Death Differ, 2007. 14(10): p. 1842-4. 237. Murtaugh, M.P., W.P. Arend, and P.J. Davies, Induction of tissue transglutaminase in human peripheral blood monocytes. J Exp Med, 1984. 159(1): p. 114-25. 238. Bowness, J.M., A.H. Tarr, and T. Wong, Increased transglutaminase activity during skin wound healing in rats. Biochim Biophys Acta, 1988. 967(2): p. 234-40.

207

239. Falasca, L., et al., Transglutaminase type II is involved in the pathogenesis of endotoxic shock. J Immunol, 2008. 180(4): p. 2616-24. 240. Schroeder, A., et al., The RIN: an RNA integrity number for assigning integrity values to RNA measurements. BMC Mol Biol, 2006. 7: p. 3. 241. Boes, T. and M. Neuhauser, Normalization for Affymetrix GeneChips. Methods Inf Med, 2005. 44(3): p. 414-7. 242. Raychaudhuri, S., J.M. Stuart, and R.B. Altman, Principal components analysis to summarize microarray experiments: application to sporulation time series. Pac Symp Biocomput, 2000: p. 455- 66. 243. Chuaqui, R.F., et al., Post-analysis follow-up and validation of microarray experiments. Nat Genet, 2002. 32 Suppl: p. 509-14. 244. Yamamoto, M., K. Takeda, and S. Akira, TIR domain-containing adaptors define the specificity of TLR signaling. Mol Immunol, 2004. 40(12): p. 861-8. 245. Hornef, M.W., B. Henriques-Normark, and S. Normark, The function and biological role of toll-like receptors in infectious diseases: an update. Curr Opin Infect Dis, 2008. 21(3): p. 304-12. 246. Lee, J., et al., Transglutaminase 2 induces nuclear factor-kappaB activation via a novel pathway in BV-2 microglia. J Biol Chem, 2004. 279(51): p. 53725-35. 247. Kuncio, G.S., et al., TNF-alpha modulates expression of the tissue transglutaminase gene in liver cells. Am J Physiol, 1998. 274(2 Pt 1): p. G240-5. 248. Ashcroft, G.S. and S.J. Mills, Androgen receptor-mediated inhibition of cutaneous wound healing. J Clin Invest, 2002. 110(5): p. 615-24. 249. Huang, Y.H., P.S. Tsai, and C.J. Huang, Bupivacaine inhibits COX-2 expression, PGE2, and cytokine production in endotoxin-activated macrophages. Acta Anaesthesiol Scand, 2008. 52(4): p. 530-5. 250. Parks, W.C., C.L. Wilson, and Y.S. Lopez-Boado, Matrix metalloproteinases as modulators of inflammation and innate immunity. Nat Rev Immunol, 2004. 4(8): p. 617-29. 251. Lund, L.R., et al., Functional overlap between two classes of matrix-degrading proteases in wound healing. EMBO J, 1999. 18(17): p. 4645-56. 252. Gutierrez-Fernandez, A., et al., Increased inflammation delays wound healing in mice deficient in collagenase-2 (MMP-8). FASEB J, 2007. 21(10): p. 2580-91. 253. Hattori, N., et al., MMP-13 plays a role in keratinocyte migration, angiogenesis, and contraction in mouse skin wound healing. Am J Pathol, 2009. 175(2): p. 533-46. 254. Merz, D., et al., IL-8/CXCL8 and growth-related oncogene alpha/CXCL1 induce chondrocyte hypertrophic differentiation. J Immunol, 2003. 171(8): p. 4406-15. 255. Verderio, E. and A. Scarpellini, Significance of the syndecan-4-transglutaminase-2 interaction. ScientificWorldJournal. 10: p. 1073-7. 256. Tan, N.S., et al., Critical roles of the nuclear receptor PPARbeta (peroxisome-proliferator-activated receptor beta) in skin wound healing. Biochem Soc Trans, 2004. 32(Pt 1): p. 97-102. 257. Muro, A.F., et al., Regulated splicing of the fibronectin EDA exon is essential for proper skin wound healing and normal lifespan. J Cell Biol, 2003. 162(1): p. 149-60. 258. Cole, J., et al., Early gene expression profile of human skin to injury using high-density cDNA microarrays. Wound Repair Regen, 2001. 9(5): p. 360-70. 259. Deonarine, K., et al., Gene expression profiling of cutaneous wound healing. J Transl Med, 2007. 5: p. 11. 260. Dallas, P.B., et al., Gene expression levels assessed by oligonucleotide microarray analysis and quantitative real-time RT-PCR -- how well do they correlate? BMC Genomics, 2005. 6(1): p. 59. 261. Etienne, W., et al., Comparison of mRNA gene expression by RT-PCR and DNA microarray. Biotechniques, 2004. 36(4): p. 618-20, 622, 624-6. 262. Duggan, D.J., et al., Expression profiling using cDNA microarrays. Nat Genet, 1999. 21(1 Suppl): p. 10-4. 263. Sorrell, J.M. and A.I. Caplan, Fibroblasts-a diverse population at the center of it all. Int Rev Cell Mol Biol, 2009. 276: p. 161-214. 264. Thomazy, V. and L. Fesus, Differential expression of tissue transglutaminase in human cells. An immunohistochemical study. Cell Tissue Res, 1989. 255(1): p. 215-24. 265. Conrad, G.W., G.W. Hart, and Y. Chen, Differences in vitro between fibroblast-like cells from cornea, heart, and skin of embryonic chicks. J Cell Sci, 1977. 26: p. 119-37. 266. Garrett, D.M. and G.W. Conrad, Fibroblast-like cells from embryonic chick cornea, heart, and skin are antigenically distinct. Dev Biol, 1979. 70(1): p. 50-70.

208

267. Houzelstein, D., et al., The expression of the homeobox gene Msx1 reveals two populations of dermal progenitor cells originating from the somites. Development, 2000. 127(10): p. 2155-64. 268. Scaal, M. and B. Christ, Formation and differentiation of the avian dermomyotome. Anat Embryol (Berl), 2004. 208(6): p. 411-24. 269. Le Lievre, C.S. and N.M. Le Douarin, Mesenchymal derivatives of the neural crest: analysis of chimaeric quail and chick embryos. J Embryol Exp Morphol, 1975. 34(1): p. 125-54. 270. Burgeson, R.E. and A.M. Christiano, The dermal-epidermal junction. Curr Opin Cell Biol, 1997. 9(5): p. 651-8. 271. Fleischmajer, R., et al., Skin fibroblasts are the only source of nidogen during early basal lamina formation in vitro. J Invest Dermatol, 1995. 105(4): p. 597-601. 272. Smola, H., et al., Dynamics of basement membrane formation by keratinocyte-fibroblast interactions in organotypic skin culture. Exp Cell Res, 1998. 239(2): p. 399-410. 273. Monical, P.L. and N.A. Kefalides, Coculture modulates laminin synthesis and mRNA levels in epidermal keratinocytes and dermal fibroblasts. Exp Cell Res, 1994. 210(2): p. 154-9. 274. Bishop, E.T., et al., An in vitro model of angiogenesis: basic features. Angiogenesis, 1999. 3(4): p. 335-44. 275. Parsonage, G., et al., A stromal address code defined by fibroblasts. Trends Immunol, 2005. 26(3): p. 150-6. 276. Clark, R.A. and T.S. Kupper, IL-15 and dermal fibroblasts induce proliferation of natural regulatory T cells isolated from human skin. Blood, 2007. 109(1): p. 194-202. 277. Steinhauser, M.L., et al., Macrophage/fibroblast coculture induces macrophage inflammatory protein-1alpha production mediated by intercellular adhesion molecule-1 and oxygen radicals. J Leukoc Biol, 1998. 64(5): p. 636-41. 278. Sorrell, J.M., M.A. Baber, and A.I. Caplan, Clonal characterization of fibroblasts in the superficial layer of the adult human dermis. Cell Tissue Res, 2007. 327(3): p. 499-510. 279. Strutz, F., The fibroblast--a (trans-) differentiated cell? Nephrol Dial Transplant, 1995. 10(9): p. 1504-6. 280. Fries, K.M., et al., Evidence of fibroblast heterogeneity and the role of fibroblast subpopulations in fibrosis. Clin Immunol Immunopathol, 1994. 72(3): p. 283-92. 281. Murad, S., A. Sivarajah, and S.R. Pinnell, Prolyl and lysyl hydroxylase activities of human skin fibroblasts: effect of donor age and ascorbate. J Invest Dermatol, 1980. 75(5): p. 404-7. 282. Atherton, A.J., et al., Immunocytochemical localization of the ectoenzyme aminopeptidase N in the human breast. J Histochem Cytochem, 1992. 40(5): p. 705-10. 283. Sorrell, J.M., et al., Cytokines and glucocorticoids differentially regulate APN/CD13 and DPPIV/CD26 enzyme activities in cultured human dermal fibroblasts. Arch Dermatol Res, 2003. 295(4): p. 160-8. 284. Pankov, R. and K.M. Yamada, Fibronectin at a glance. J Cell Sci, 2002. 115(Pt 20): p. 3861-3. 285. George, E.L., et al., Defects in mesoderm, neural tube and vascular development in mouse embryos lacking fibronectin. Development, 1993. 119(4): p. 1079-91. 286. Patel, R.S., et al., Organization of the fibronectin gene provides evidence for exon shuffling during evolution. EMBO J, 1987. 6(9): p. 2565-72. 287. Pearlstein, E., L.I. Gold, and A. Garcia-Pardo, Fibronectin: a review of its structure and biological activity. Mol Cell Biochem, 1980. 29(2): p. 103-28. 288. Liao, Y.F., et al., The EIIIA segment of fibronectin is a ligand for integrins alpha 9beta 1 and alpha 4beta 1 providing a novel mechanism for regulating cell adhesion by alternative splicing. J Biol Chem, 2002. 277(17): p. 14467-74. 289. Magnusson, M.K. and D.F. Mosher, Fibronectin - Structure, assembly, and cardiovascular implications. Arteriosclerosis Thrombosis and Vascular Biology, 1998. 18(9): p. 1363-1370. 290. Mckeownlongo, P.J. and D.F. Mosher, Binding of Plasma Fibronectin to Cell-Layers of Human-Skin Fibroblasts. Journal of Cell Biology, 1983. 97(2): p. 466-472. 291. Mckeownlongo, P.J. and D.F. Mosher, Mechanism of Formation of Disulfide-Bonded Multimers of Plasma Fibronectin in Cell-Layers of Cultured Human-Fibroblasts. Journal of Biological Chemistry, 1984. 259(19): p. 2210-2215. 292. Peters, D.M. and D.F. Mosher, Localization of cell surface sites involved in fibronectin fibrillogenesis. J Cell Biol, 1987. 104(1): p. 121-30. 293. Chen, F. and D.F. Mosher, Formation of sodium dodecyl sulfate-stable fibronectin multimers - Failure to detect products of thiol-disulfide exchange in cyanogen bromide or limited acid digests of stabilized matrix fibronectin. Journal of Biological Chemistry, 1996. 271(15): p. 9084-9089.

209

294. Ginsberg, M.H., A. Partridge, and S.J. Shattil, Integrin regulation. Curr Opin Cell Biol, 2005. 17(5): p. 509-16. 295. Hynes, R.O., Integrins: versatility, modulation, and signaling in cell adhesion. Cell, 1992. 69(1): p. 11-25. 296. Shimaoka, M., J. Takagi, and T.A. Springer, Conformational regulation of integrin structure and function. Annu Rev Biophys Biomol Struct, 2002. 31: p. 485-516. 297. Vinogradova, O., et al., A structural mechanism of integrin alpha(IIb)beta(3) "inside-out" activation as regulated by its cytoplasmic face. Cell, 2002. 110(5): p. 587-97. 298. Barry, F.A. and J.M. Gibbins, Protein kinase B is regulated in platelets by the collagen receptor glycoprotein VI. J Biol Chem, 2002. 277(15): p. 12874-8. 299. Li, Z., et al., A stimulatory role for cGMP-dependent protein kinase in platelet activation. Cell, 2003. 112(1): p. 77-86. 300. Li, Z., X. Xi, and X. Du, A mitogen-activated protein kinase-dependent signaling pathway in the activation of platelet integrin alpha IIbbeta3. J Biol Chem, 2001. 276(45): p. 42226-32. 301. Kovacsovics, T.J., et al., Phosphoinositide 3-kinase inhibition spares actin assembly in activating platelets but reverses platelet aggregation. J Biol Chem, 1995. 270(19): p. 11358-66. 302. Quinton, T.M., et al., Protein kinase C- and calcium-regulated pathways independently synergize with Gi pathways in agonist-induced fibrinogen receptor activation. Biochem J, 2002. 368(Pt 2): p. 535-43. 303. Ratnikov, B.I., A.W. Partridge, and M.H. Ginsberg, Integrin activation by talin. J Thromb Haemost, 2005. 3(8): p. 1783-90. 304. Liu, S., D.A. Calderwood, and M.H. Ginsberg, Integrin cytoplasmic domain-binding proteins. J Cell Sci, 2000. 113 ( Pt 20): p. 3563-71. 305. O'Halloran, T., M.C. Beckerle, and K. Burridge, Identification of talin as a major cytoplasmic protein implicated in platelet activation. Nature, 1985. 317(6036): p. 449-51. 306. Fang, J., et al., Therapeutic expression of the platelet-specific integrin, alphaIIbbeta3, in a murine model for Glanzmann thrombasthenia. Blood, 2005. 106(8): p. 2671-9. 307. Xiong, J.P., et al., Crystal structure of the extracellular segment of integrin alpha Vbeta3. Science, 2001. 294(5541): p. 339-45. 308. Yan, B., et al., Calpain cleavage promotes talin binding to the beta 3 integrin cytoplasmic domain. J Biol Chem, 2001. 276(30): p. 28164-70. 309. Martel, V., et al., Conformation, localization, and integrin binding of talin depend on its interaction with phosphoinositides. J Biol Chem, 2001. 276(24): p. 21217-27. 310. Bertagnolli, M.E., et al., Talin distribution and phosphorylation in thrombin-activated platelets. J Cell Sci, 1993. 106 ( Pt 4): p. 1189-99. 311. Tadokoro, S., et al., Talin binding to integrin beta tails: a final common step in integrin activation. Science, 2003. 302(5642): p. 103-6. 312. Qin, J., O. Vinogradova, and E.F. Plow, Integrin bidirectional signaling: a molecular view. PLoS Biol, 2004. 2(6): p. e169. 313. Loike, J.D., et al., CD11c/CD18 on neutrophils recognizes a domain at the N terminus of the A alpha chain of fibrinogen. Proc Natl Acad Sci U S A, 1991. 88(3): p. 1044-8. 314. Gailit, J. and E. Ruoslahti, Regulation of the fibronectin receptor affinity by divalent cations. J Biol Chem, 1988. 263(26): p. 12927-32. 315. Riveline, D., et al., Focal contacts as mechanosensors: externally applied local mechanical force induces growth of focal contacts by an mDia1-dependent and ROCK-independent mechanism. J Cell Biol, 2001. 153(6): p. 1175-86. 316. Geiger, B., et al., Transmembrane crosstalk between the extracellular matrix--cytoskeleton crosstalk. Nat Rev Mol Cell Biol, 2001. 2(11): p. 793-805. 317. Zamir, E., et al., Molecular diversity of cell-matrix adhesions. J Cell Sci, 1999. 112 ( Pt 11): p. 1655- 69. 318. Zamir, E. and B. Geiger, Components of cell-matrix adhesions. J Cell Sci, 2001. 114(Pt 20): p. 3577- 9. 319. Vicente-Manzanares, M., et al., Non-muscle myosin II takes centre stage in cell adhesion and migration. Nat Rev Mol Cell Biol, 2009. 10(11): p. 778-90. 320. Ziegler, W.H., R.C. Liddington, and D.R. Critchley, The structure and regulation of vinculin. Trends Cell Biol, 2006. 16(9): p. 453-60. 321. Miyamoto, S., S.K. Akiyama, and K.M. Yamada, Synergistic roles for receptor occupancy and aggregation in integrin transmembrane function. Science, 1995. 267(5199): p. 883-5.

210

322. Laukaitis, C.M., et al., Differential dynamics of alpha 5 integrin, paxillin, and alpha-actinin during formation and disassembly of adhesions in migrating cells. J Cell Biol, 2001. 153(7): p. 1427-40. 323. Chrzanowska-Wodnicka, M. and K. Burridge, Rho-stimulated contractility drives the formation of stress fibers and focal adhesions. J Cell Biol, 1996. 133(6): p. 1403-15. 324. Volberg, T., et al., pp60(c-src) and related tyrosine kinases: a role in the assembly and reorganization of matrix adhesions. J Cell Sci, 2001. 114(Pt 12): p. 2279-89. 325. Parsons, J.T., A.R. Horwitz, and M.A. Schwartz, Cell adhesion: integrating cytoskeletal dynamics and cellular tension. Nat Rev Mol Cell Biol. 11(9): p. 633-43. 326. Brown, J.H., D.P. Del Re, and M.A. Sussman, The Rac and Rho hall of fame: a decade of hypertrophic signaling hits. Circ Res, 2006. 98(6): p. 730-42. 327. Ridley, A.J., Rho GTPases and cell migration. J Cell Sci, 2001. 114(Pt 15): p. 2713-22. 328. Burridge, K. and K. Wennerberg, Rho and Rac take center stage. Cell, 2004. 116(2): p. 167-79. 329. Schwartz, M., Rho signalling at a glance. J Cell Sci, 2004. 117(Pt 23): p. 5457-8. 330. Kimura, K., et al., Regulation of myosin phosphatase by Rho and Rho-associated kinase (Rho- kinase). Science, 1996. 273(5272): p. 245-8. 331. Sumi, T., et al., Cofilin phosphorylation and actin cytoskeletal dynamics regulated by rho- and Cdc42-activated LIM-kinase 2. J Cell Biol, 1999. 147(7): p. 1519-32. 332. Tsukita, S. and S. Yonemura, ERM proteins: head-to-tail regulation of actin-plasma membrane interaction. Trends Biochem Sci, 1997. 22(2): p. 53-8. 333. Pollard, T.D. and G.G. Borisy, Cellular motility driven by assembly and disassembly of actin filaments. Cell, 2003. 112(4): p. 453-65. 334. Nicholson-Dykstra, S., H.N. Higgs, and E.S. Harris, Actin dynamics: growth from dendritic branches. Curr Biol, 2005. 15(9): p. R346-57. 335. Alexandrova, A.Y., et al., Comparative dynamics of retrograde actin flow and focal adhesions: formation of nascent adhesions triggers transition from fast to slow flow. PLoS One, 2008. 3(9): p. e3234. 336. Narumiya, S., M. Tanji, and T. Ishizaki, Rho signaling, ROCK and mDia1, in transformation, metastasis and invasion. Cancer Metastasis Rev, 2009. 28(1-2): p. 65-76. 337. Naumanen, P., P. Lappalainen, and P. Hotulainen, Mechanisms of actin stress fibre assembly. J Microsc, 2008. 231(3): p. 446-54. 338. Vicente-Manzanares, M., D.J. Webb, and A.R. Horwitz, Cell migration at a glance. J Cell Sci, 2005. 118(Pt 21): p. 4917-9. 339. Vicente-Manzanares, M., et al., Segregation and activation of myosin IIB creates a rear in migrating cells. J Cell Biol, 2008. 183(3): p. 543-54. 340. Choi, C.K., et al., Actin and alpha-actinin orchestrate the assembly and maturation of nascent adhesions in a myosin II motor-independent manner. Nat Cell Biol, 2008. 10(9): p. 1039-50. 341. Miyamoto, S., et al., Integrin function: molecular hierarchies of cytoskeletal and signaling molecules. J Cell Biol, 1995. 131(3): p. 791-805. 342. Hu, K., et al., Differential transmission of actin motion within focal adhesions. Science, 2007. 315(5808): p. 111-5. 343. Franco, S.J., et al., Calpain-mediated proteolysis of talin regulates adhesion dynamics. Nat Cell Biol, 2004. 6(10): p. 977-83. 344. Ridley, A.J., et al., Cell migration: integrating signals from front to back. Science, 2003. 302(5651): p. 1704-9. 345. Geiger, B., J.P. Spatz, and A.D. Bershadsky, Environmental sensing through focal adhesions. Nat Rev Mol Cell Biol, 2009. 10(1): p. 21-33. 346. Somanath, P.R., et al., Akt1 signaling regulates integrin activation, matrix recognition, and fibronectin assembly. J Biol Chem, 2007. 282(31): p. 22964-76. 347. Magdalena, J., T.H. Millard, and L.M. Machesky, Microtubule involvement in NIH 3T3 Golgi and MTOC polarity establishment. J Cell Sci, 2003. 116(Pt 4): p. 743-56. 348. Jaffe, A.B. and A. Hall, Rho GTPases: biochemistry and biology. Annu Rev Cell Dev Biol, 2005. 21: p. 247-69. 349. Gomes, E.R., S. Jani, and G.G. Gundersen, Nuclear movement regulated by Cdc42, MRCK, myosin, and actin flow establishes MTOC polarization in migrating cells. Cell, 2005. 121(3): p. 451-63. 350. Webb, D.J., et al., FAK-Src signalling through paxillin, ERK and MLCK regulates adhesion disassembly. Nat Cell Biol, 2004. 6(2): p. 154-61. 351. Arnold, M., et al., Induction of cell polarization and migration by a gradient of nanoscale variations in adhesive ligand spacing. Nano Lett, 2008. 8(7): p. 2063-9.

211

352. Stevens, M.M. and J.H. George, Exploring and engineering the cell surface interface. Science, 2005. 310(5751): p. 1135-8. 353. Cavalcanti-Adam, E.A., et al., Cell spreading and focal adhesion dynamics are regulated by spacing of integrin ligands. Biophys J, 2007. 92(8): p. 2964-74. 354. Hirschfeld-Warneken, V.C., et al., Cell adhesion and polarisation on molecularly defined spacing gradient surfaces of cyclic RGDfK peptide patches. Eur J Cell Biol, 2008. 87(8-9): p. 743-50. 355. Byrd, J.C. and U. Lichti, Two types of transglutaminase in the PC12 pheochromocytoma cell line. Stimulation by sodium butyrate. J Biol Chem, 1987. 262(24): p. 11699-705. 356. Gentile, V., et al., Expression of tissue transglutaminase in Balb-C 3T3 fibroblasts: effects on cellular morphology and adhesion. J Cell Biol, 1992. 119(2): p. 463-74. 357. Menter, D.G., et al., Transglutaminase stabilizes melanoma adhesion under laminar flow. Cell Biophys, 1991. 18(2): p. 123-43. 358. Lorand, L., J.E. Dailey, and P.M. Turner, Fibronectin as a carrier for the transglutaminase from human erythrocytes. Proc Natl Acad Sci U S A, 1988. 85(4): p. 1057-9. 359. Jeong, J.M., et al., The fibronectin-binding domain of transglutaminase. J Biol Chem, 1995. 270(10): p. 5654-8. 360. Hang, J., et al., Identification of a novel recognition sequence for fibronectin within the NH2- terminal beta-sandwich domain of tissue transglutaminase. J Biol Chem, 2005. 280(25): p. 23675- 83. 361. Zemskov, E.A., et al., The role of tissue transglutaminase in cell-matrix interactions. Front Biosci, 2006. 11: p. 1057-76. 362. Morgan, M.R., M.J. Humphries, and M.D. Bass, Synergistic control of cell adhesion by integrins and syndecans. Nat Rev Mol Cell Biol, 2007. 8(12): p. 957-69. 363. Echtermeyer, F., et al., Delayed wound repair and impaired angiogenesis in mice lacking syndecan- 4. J Clin Invest, 2001. 107(2): p. R9-R14. 364. Keum, E., et al., Syndecan-4 regulates localization, activity and stability of protein kinase C-alpha. Biochem J, 2004. 378(Pt 3): p. 1007-14. 365. Beauvais, D.M., B.J. Burbach, and A.C. Rapraeger, The syndecan-1 ectodomain regulates alphavbeta3 integrin activity in human mammary carcinoma cells. J Cell Biol, 2004. 167(1): p. 171- 81. 366. McQuade, K.J., et al., Syndecan-1 regulates alphavbeta5 integrin activity in B82L fibroblasts. J Cell Sci, 2006. 119(Pt 12): p. 2445-56. 367. Hozumi, K., et al., Laminin alpha1 chain LG4 module promotes cell attachment through syndecans and cell spreading through integrin alpha2beta1. J Biol Chem, 2006. 281(43): p. 32929-40. 368. Signorini, M., et al., Human erythrocyte transglutaminase: purification and preliminary characterisation. Biol Chem Hoppe Seyler, 1988. 369(4): p. 275-81. 369. Scarpellini, A., et al., Heparan sulfate proteoglycans are receptors for the cell-surface trafficking and biological activity of transglutaminase-2. J Biol Chem, 2009. 284(27): p. 18411-23. 370. Zemskov, E.A., et al., Cell-surface transglutaminase undergoes internalization and lysosomal degradation: an essential role for LRP1. J Cell Sci, 2007. 120(Pt 18): p. 3188-99. 371. Orr, A.W., et al., Low density lipoprotein receptor-related protein is a calreticulin coreceptor that signals focal adhesion disassembly. J Cell Biol, 2003. 161(6): p. 1179-89. 372. Boucher, P., et al., LRP: role in vascular wall integrity and protection from atherosclerosis. Science, 2003. 300(5617): p. 329-32. 373. Midwood, K.S., et al., Modulation of cell-fibronectin matrix interactions during tissue repair. J Investig Dermatol Symp Proc, 2006. 11(1): p. 73-8. 374. Akimov, S.S. and A.M. Belkin, Opposing roles of Ras/Raf oncogenes and the MEK1/ERK signaling module in regulation of expression and adhesive function of surface transglutaminase. J Biol Chem, 2003. 278(37): p. 35609-19. 375. Mangala, L.S. and K. Mehta, Tissue transglutaminase (TG2) in cancer biology. Prog Exp Tumor Res, 2005. 38: p. 125-38. 376. Priglinger, S.G., et al., TGF-beta2-induced cell surface tissue transglutaminase increases adhesion and migration of RPE cells on fibronectin through the gelatin-binding domain. Invest Ophthalmol Vis Sci, 2004. 45(3): p. 955-63. 377. Smith, J.R. and O.M. Pereira-Smith, Replicative senescence: implications for in vivo aging and tumor suppression. Science, 1996. 273(5271): p. 63-7. 378. Balklava, Z., et al., Analysis of tissue transglutaminase function in the migration of Swiss 3T3 fibroblasts: the active-state conformation of the enzyme does not affect cell motility but is important for its secretion. J Biol Chem, 2002. 277(19): p. 16567-75. 212

379. Gottrup, F., M.S. Agren, and T. Karlsmark, Models for use in wound healing research: a survey focusing on in vitro and in vivo adult soft tissue. Wound Repair Regen, 2000. 8(2): p. 83-96. 380. Yamada, K.M. and D.W. Kennedy, Fibroblast cellular and plasma fibronectins are similar but not identical. J Cell Biol, 1979. 80(2): p. 492-8. 381. Grinnell, F., M. Feld, and D. Minter, Fibroblast adhesion to fibrinogen and fibrin substrata: requirement for cold-insoluble globulin (plasma fibronectin). Cell, 1980. 19(2): p. 517-25. 382. Oh, E., M. Pierschbacher, and E. Ruoslahti, Deposition of plasma fibronectin in tissues. Proc Natl Acad Sci U S A, 1981. 78(5): p. 3218-21. 383. Grinnell, F. and T.V. Phan, Deposition of fibronectin on material surfaces exposed to plasma: quantitative and biological studies. J Cell Physiol, 1983. 116(3): p. 289-96. 384. Pan, C., et al., Comparative proteomic phenotyping of cell lines and primary cells to assess preservation of cell type-specific functions. Mol Cell Proteomics, 2009. 8(3): p. 443-50. 385. Frisch, S.M. and R.A. Screaton, Anoikis mechanisms. Curr Opin Cell Biol, 2001. 13(5): p. 555-62. 386. Aeschlimann, D. and V. Thomazy, Protein crosslinking in assembly and remodelling of extracellular matrices: the role of transglutaminases. Connect Tissue Res, 2000. 41(1): p. 1-27. 387. Nikolic, D.L., et al., Role of boundary conditions in an experimental model of epithelial wound healing. Am J Physiol Cell Physiol, 2006. 291(1): p. C68-75. 388. Isobe, T., et al., Activity-independent cell adhesion to tissue-type transglutaminase is mediated by alpha4beta1 integrin. Eur J Cell Biol, 1999. 78(12): p. 876-83. 389. Ueki, S., J. Takagi, and Y. Saito, Dual functions of transglutaminase in novel cell adhesion. J Cell Sci, 1996. 109 ( Pt 11): p. 2727-35. 390. Takahashi, H., et al., Tissue transglutaminase, coagulation factor XIII, and the pro-polypeptide of von Willebrand factor are all ligands for the integrins alpha 9beta 1 and alpha 4beta 1. J Biol Chem, 2000. 275(31): p. 23589-95. 391. Iismaa, S.E., et al., GTP binding and signaling by Gh/transglutaminase II involves distinct residues in a unique GTP-binding pocket. J Biol Chem, 2000. 275(24): p. 18259-65. 392. Alexopoulou, A.N., H.A. Multhaupt, and J.R. Couchman, Syndecans in wound healing, inflammation and vascular biology. Int J Biochem Cell Biol, 2007. 39(3): p. 505-28. 393. Midwood, K.S., L.V. Williams, and J.E. Schwarzbauer, Tissue repair and the dynamics of the extracellular matrix. Int J Biochem Cell Biol, 2004. 36(6): p. 1031-7. 394. Woods, A., et al., Adhesion and cytoskeletal organisation of fibroblasts in response to fibronectin fragments. EMBO J, 1986. 5(4): p. 665-70. 395. Subramanian, S.V., M.L. Fitzgerald, and M. Bernfield, Regulated shedding of syndecan-1 and -4 ectodomains by thrombin and growth factor receptor activation. J Biol Chem, 1997. 272(23): p. 14713-20. 396. Wysocki, A.B. and F. Grinnell, Fibronectin profiles in normal and chronic wound fluid. Lab Invest, 1990. 63(6): p. 825-31. 397. Felding-Habermann, B. and D.A. Cheresh, Vitronectin and its receptors. Curr Opin Cell Biol, 1993. 5(5): p. 864-8. 398. Mogford, J.E., et al., Vascular smooth muscle alpha v beta 3 integrin mediates arteriolar vasodilation in response to RGD peptides. Circ Res, 1996. 79(4): p. 821-6. 399. Chau, D.Y., et al., The cellular response to transglutaminase-cross-linked collagen. Biomaterials, 2005. 26(33): p. 6518-29. 400. Stephens, P., et al., Crosslinking and G-protein functions of transglutaminase 2 contribute differentially to fibroblast wound healing responses. J Cell Sci, 2004. 117(Pt 15): p. 3389-403. 401. Gaudry, C.A., et al., Tissue transglutaminase is an important player at the surface of human endothelial cells: evidence for its externalization and its colocalization with the beta(1) integrin. Exp Cell Res, 1999. 252(1): p. 104-13. 402. Barsigian, C., A.M. Stern, and J. Martinez, Tissue (type II) transglutaminase covalently incorporates itself, fibrinogen, or fibronectin into high molecular weight complexes on the extracellular surface of isolated hepatocytes. Use of 2-[(2-oxopropyl)thio] imidazolium derivatives as cellular transglutaminase inactivators. J Biol Chem, 1991. 266(33): p. 22501-9. 403. Burridge, K. and R. Doughman, Front and back by Rho and Rac. Nat Cell Biol, 2006. 8(8): p. 781-2. 404. Kurokawa, K. and M. Matsuda, Localized RhoA activation as a requirement for the induction of membrane ruffling. Mol Biol Cell, 2005. 16(9): p. 4294-303. 405. Upchurch, H.F., et al., Localization of cellular transglutaminase on the extracellular matrix after wounding: characteristics of the matrix bound enzyme. J Cell Physiol, 1991. 149(3): p. 375-82. 406. Johnson, T.S., et al., Transglutaminase transcription and antigen translocation in experimental renal scarring. J Am Soc Nephrol, 1999. 10(10): p. 2146-57. 213

407. Cukierman, E., et al., Taking cell-matrix adhesions to the third dimension. Science, 2001. 294(5547): p. 1708-12. 408. Halliday, N.L. and J.J. Tomasek, Mechanical properties of the extracellular matrix influence fibronectin fibril assembly in vitro. Exp Cell Res, 1995. 217(1): p. 109-17. 409. Fringer, J. and F. Grinnell, Fibroblast quiescence in floating or released collagen matrices: contribution of the ERK signaling pathway and actin cytoskeletal organization. J Biol Chem, 2001. 276(33): p. 31047-52. 410. Even-Ram, S. and K.M. Yamada, Cell migration in 3D matrix. Curr Opin Cell Biol, 2005. 17(5): p. 524-32. 411. Livesey, S.A., et al., Transplanted acellular allograft dermal matrix. Potential as a template for the reconstruction of viable dermis. Transplantation, 1995. 60(1): p. 1-9. 412. Tsukamoto, Y., W.E. Helsel, and S.M. Wahl, Macrophage production of fibronectin, a chemoattractant for fibroblasts. J Immunol, 1981. 127(2): p. 673-8. 413. Postlethwaite, A.E., et al., Induction of fibroblast chemotaxis by fibronectin. Localization of the chemotactic region to a 140,000-molecular weight non-gelatin-binding fragment. J Exp Med, 1981. 153(2): p. 494-9. 414. Nakagawa, S., P. Pawelek, and F. Grinnell, Extracellular matrix organization modulates fibroblast growth and growth factor responsiveness. Exp Cell Res, 1989. 182(2): p. 572-82. 415. Capecchi, M.R., The new mouse genetics: altering the genome by gene targeting. Trends Genet, 1989. 5(3): p. 70-6. 416. Evans, M.J., Potential for genetic manipulation of mammals. Mol Biol Med, 1989. 6(6): p. 557-65. 417. Koller, B.H. and O. Smithies, Inactivating the beta 2-microglobulin locus in mouse embryonic stem cells by homologous recombination. Proc Natl Acad Sci U S A, 1989. 86(22): p. 8932-5. 418. Beck, J.A., et al., Genealogies of mouse inbred strains. Nat Genet, 2000. 24(1): p. 23-5. 419. Silver, L.M., Mouse Genetics Concepts and Applications, 2001, Oxford University Press. 420. Tang, Y., T.E. Crowley, and N.M. Kumar, Global gene expression analysis of lenses from different mouse strains and in the alpha3Cx46 knockout mouse. Mol Vis. 16: p. 113-21. 421. Gerlai, R., Gene-targeting studies of mammalian behavior: is it the mutation or the background genotype? Trends Neurosci, 1996. 19(5): p. 177-81. 422. Taketo, M., et al., FVB/N: an inbred mouse strain preferable for transgenic analyses. Proc Natl Acad Sci U S A, 1991. 88(6): p. 2065-9. 423. Crawley, J.N., et al., Behavioral phenotypes of inbred mouse strains: implications and recommendations for molecular studies. Psychopharmacology (Berl), 1997. 132(2): p. 107-24. 424. George, E.L., H.S. Baldwin, and R.O. Hynes, Fibronectins are essential for heart and blood vessel morphogenesis but are dispensable for initial specification of precursor cells. Blood, 1997. 90(8): p. 3073-81. 425. Rajnoch, C., et al., Regeneration of the ear after wounding in different mouse strains is dependent on the severity of wound trauma. Dev Dyn, 2003. 226(2): p. 388-97. 426. Bonyadi, M., et al., Mapping of a major genetic modifier of embryonic lethality in TGF beta 1 knockout mice. Nat Genet, 1997. 15(2): p. 207-11. 427. Tang, Y., et al., Genetic modifiers interact with maternal determinants in vascular development of Tgfb1(-/-) mice. Hum Mol Genet, 2003. 12(13): p. 1579-89. 428. Lauer, P., et al., Targeted inactivation of the mouse locus encoding coagulation factor XIII-A: hemostatic abnormalities in mutant mice and characterization of the coagulation deficit. Thromb Haemost, 2002. 88(6): p. 967-74. 429. Jobe, S.M., et al., Role of FcRgamma and factor XIIIA in coated platelet formation. Blood, 2005. 106(13): p. 4146-51. 430. Dale, G.L., et al., Stimulated platelets use serotonin to enhance their retention of procoagulant proteins on the cell surface. Nature, 2002. 415(6868): p. 175-9. 431. Nakano, Y., et al., Expression and localization of plasma transglutaminase factor XIIIA in bone. J Histochem Cytochem, 2007. 55(7): p. 675-85. 432. Brody, A.R., et al., Analyzing the genes and peptide growth factors expressed in lung cells in vivo consequent to asbestos exposure and in vitro. Environ Health Perspect, 1997. 105 Suppl 5: p. 1165- 71. 433. Brody, A.R., et al., Identifying fibrosis susceptibility genes in two strains of inbred mice. Chest, 2002. 121(3 Suppl): p. 31S. 434. Li, L., et al., Connexin mediated cataract prevention in mice. PLoS One. 5(9). 435. Nardacci, R., et al., Transglutaminase type II plays a protective role in hepatic injury. Am J Pathol, 2003. 162(4): p. 1293-303. 214

436. Tatsukawa, H., et al., Role of transglutaminase 2 in liver injury via cross-linking and silencing of transcription factor Sp1. Gastroenterology, 2009. 136(5): p. 1783-95 e10. 437. Sarang, Z., et al., Tissue transglutaminase (TG2) acting as G protein protects hepatocytes against Fas-mediated cell death in mice. Hepatology, 2005. 42(3): p. 578-87. 438. Szondy, Z., et al., Tissue transglutaminase (TG2) protects cardiomyocytes against ischemia/reperfusion injury by regulating ATP synthesis. Cell Death Differ, 2006. 13(10): p. 1827-9. 439. Sarang, Z., et al., Some lessons from the tissue transglutaminase knockout mouse. Amino Acids, 2009. 36(4): p. 625-31. 440. Sullivan, T.P., et al., The pig as a model for human wound healing. Wound Repair Regen, 2001. 9(2): p. 66-76. 441. Zeitani, J., et al., Superficial wound dehiscence after median sternotomy: surgical treatment versus secondary wound healing. Ann Thorac Surg, 2004. 77(2): p. 672-5. 442. Tiwari, A.K., et al., Carrier mediated protein and peptide stabilization. Drug Deliv. 17(8): p. 605- 16. 443. Ikada, Y., Challenges in tissue engineering. J R Soc Interface, 2006. 3(10): p. 589-601. 444. Baillie, A.J., et al., The preparation and properties of niosomes--non-ionic surfactant vesicles. J Pharm Pharmacol, 1985. 37(12): p. 863-8. 445. Brewer, J.M. and J. Alexander, The adjuvant activity of non-ionic surfactant vesicles (niosomes) on the BALB/c humoral response to bovine serum albumin. Immunology, 1992. 75(4): p. 570-5. 446. Brewer, J.M., et al., Lipid vesicle size determines the Th1 or Th2 response to entrapped antigen. J Immunol, 1998. 161(8): p. 4000-7. 447. Memisoglu, E., et al., In vivo evaluation for rhGM-CSF wound-healing efficacy in topical vehicles. Pharm Dev Technol, 1997. 2(2): p. 171-80. 448. Wysocki, A.B., Wound fluids and the pathogenesis of chronic wounds. J Wound Ostomy Continence Nurs, 1996. 23(6): p. 283-90. 449. Harding, K.G., H.L. Morris, and G.K. Patel, Science, medicine and the future: healing chronic wounds. BMJ, 2002. 324(7330): p. 160-3. 450. Mustoe, T., Understanding chronic wounds: a unifying hypothesis on their pathogenesis and implications for therapy. Am J Surg, 2004. 187(5A): p. 65S-70S. 451. Higley, H.R., et al., Extravasation of macromolecules and possible trapping of transforming growth factor-beta in venous ulceration. Br J Dermatol, 1995. 132(1): p. 79-85. 452. Barrick, B., E.J. Campbell, and C.A. Owen, Leukocyte proteinases in wound healing: roles in physiologic and pathologic processes. Wound Repair Regen, 1999. 7(6): p. 410-22. 453. Stanley, A.C., et al., Reduced growth of dermal fibroblasts from chronic venous ulcers can be stimulated with growth factors. J Vasc Surg, 1997. 26(6): p. 994-9; discussion 999-1001. 454. Steed, D.L., Clinical evaluation of recombinant human platelet-derived growth factor for the treatment of lower extremity diabetic ulcers. Diabetic Ulcer Study Group. J Vasc Surg, 1995. 21(1): p. 71-8; discussion 79-81. 455. Wieman, T.J., J.M. Smiell, and Y. Su, Efficacy and safety of a topical gel formulation of recombinant human platelet-derived growth factor-BB (becaplermin) in patients with chronic neuropathic diabetic ulcers. A phase III randomized placebo-controlled double-blind study. Diabetes Care, 1998. 21(5): p. 822-7. 456. Pierce, G.F. and T.A. Mustoe, Pharmacologic enhancement of wound healing. Annu Rev Med, 1995. 46: p. 467-81. 457. Verderio, E.A., T. Johnson, and M. Griffin, Tissue transglutaminase in normal and abnormal wound healing: review article. Amino Acids, 2004. 26(4): p. 387-404. 458. Wysocki, A.B., L. Staiano-Coico, and F. Grinnell, Wound fluid from chronic leg ulcers contains elevated levels of metalloproteinases MMP-2 and MMP-9. J Invest Dermatol, 1993. 101(1): p. 64-8. 459. Silver, F.H. and G. Pins, Cell growth on collagen: a review of tissue engineering using scaffolds containing extracellular matrix. J Long Term Eff Med Implants, 1992. 2(1): p. 67-80. 460. Goo, H.C., et al., Development of collagenase-resistant collagen and its interaction with adult human dermal fibroblasts. Biomaterials, 2003. 24(28): p. 5099-113. 461. Jung, S.N., et al., One-stage skin grafting of the exposed skull with acellular human dermis (AlloDerm). J Craniofac Surg, 2008. 19(6): p. 1660-2. 462. Kremer, M., E. Lang, and A.C. Berger, Evaluation of dermal-epidermal skin equivalents ('composite-skin') of human keratinocytes in a collagen-glycosaminoglycan matrix(Integra artificial skin). Br J Plast Surg, 2000. 53(6): p. 459-65. 463. MacNeil, S., Progress and opportunities for tissue-engineered skin. Nature, 2007. 445(7130): p. 874- 80. 215

464. Ferdman, A.G. and I.V. Yannas, Scattering of light from histologic sections: a new method for the analysis of connective tissue. J Invest Dermatol, 1993. 100(5): p. 710-6. 465. Yannas, I.V., Facts and theories of induced organ regeneration. Adv Biochem Eng Biotechnol, 2005. 93: p. 1-38. 466. Sevilla, C.A., D. Dalecki, and D.C. Hocking, Extracellular matrix fibronectin stimulates the self- assembly of microtissues on native collagen gels. Tissue Eng Part A. 16(12): p. 3805-19.

216