INTRACELLULAR TRAFFICKING AND SECRETION OF MATRIX ... · INTRACELLULAR TRAFFICKING AND SECRETION OF MATRIX METALLOPROTEINASES DURING MACROPHAGE MIGRATION. Joan Rohl . Bachelor of

INTRACELLULAR TRAFFICKING AND SECRETION OF MATRIX

METALLOPROTEINASES DURING MACROPHAGE MIGRATION

Joan Rohl Bachelor of Science, Master of Science

Submitted in fulfilment of the requirements for the degree of

Doctor of Philosophy

School of Biomedical Sciences Faculty of Health

Queensland University of Technology

2016

i

KEYWORDS

Collagen

Chronic wounds

Extracellular matrix

Degradation

Inflammation

Intracellular trafficking

Macrophages

Matrigel

Matrix metalloproteinases

Migration

MMP14

MMP9

Secretion

SNAP23

Soluble NSF attachment protein receptor

Syntaxin 4

Trafficking machinery

Trafficking pathways

VAMP4

VAMP7

VAMP8

ii

ABSTRACT

The skin, the largest organ of the human body, serves as a physical barrier and as

such, has many important biological functions. If compromised, timely repair is vital to

ensure that the skin can fulfil these roles. During the healing process, the wounded area

is cleared from pathogens and debris, a provisional matrix is formed to assist wound

closure, the vascular network is restored and ultimately, the tissue is remodelled

resulting in scar formation. Cutaneous wound healing is characterised by four

overlapping and interdependent phases – haemostasis, inflammation, proliferation,

maturation – and involves multiple different cell types. Any disturbances in this highly

complex process can delay the healing process and lead to chronic wounds. They are

defined as wounds that show no significant healing within four weeks. Chronic wounds

affect around 450,000 Australians every year and treatment of these is estimated to cost

the health care system $3 billion annually.

Persistent and elevated inflammation contributes to the formation of non‐

healing wounds and the number of macrophages within a wound can determine the level

and duration of inflammation. Macrophages are recruited from the bloodstream and

infiltrate the tissue during wound healing. They create a path by degrading the

extracellular matrix (ECM), the acellular component that comprises the dermal tissue

and plays many important roles in wounded and unwounded skin. The ECM acts not only

as a scaffold for the cells, but also regulates cell adhesion, chemotaxis and migration, as

well as storing growth factors and cytokines. Matrix metalloproteinases (MMPs) cleave

the ECM and thus help remodel the tissue during the healing process. During normal

wound healing, MMP activity is tightly controlled. However, in non‐healing wounds

increased levels of certain MMPs, such as MMP9 and MMP14, cause extensive tissue

damage and contribute to sustained inflammation. MMP14 facilitates macrophage tissue

infiltration, which then amplifies the inflammatory response within the wound. Current

treatment strategies to reduce MMP levels in the wound fluid or to inhibit their activity

have not been effective. For these enzymes to access their substrates, they first have to

be secreted as is the case of MMP9 or incorporated into the plasma membrane as is the

case with MM14. Preventing the delivery of MMP9 and MMP14 to the cell surface might

reduce tissue damage and macrophage infiltration leading to an overall reduction of

inflammation. In order to inhibit MMP cell surface delivery, it is necessary to elucidate

their intracellular transport route and identify proteins that regulate these pathways in

macrophages.

iii

At least two major pathways exist in macrophages by which proteins can be

delivered to the cell surface. One is a classical pathway where proteins are synthesised

in the endoplasmic reticulum (ER), shuttled through the Golgi apparatus and transported

to the surface either directly or indirectly through the recycling endosome. One other

major secretory pathway is the lysosomal pathway, in which proteins are delivered to

the cell surface from the cytosol or the Golgi complex through a lysosome or lysosome‐

related organelle. Transport of proteins between these organelles and the cell surface

occurs within membrane‐bound vesicles. Delivery of vesicles to target organelles or the

plasma membrane requires membrane fusion, which is facilitated by soluble N‐

ethylmaleimide‐sensitive‐factor attachment protein receptor (SNARE) proteins at all

points in the trafficking pathway. These proteins form a complex of one R‐SNARE on one

membrane and two or three Q‐SNAREs on the opposing membrane, which not only

brings the membranes into close proximity, but also provides the energy and specificity

required for membrane fusion. Each SNARE protein has a precise subcellular

distribution and regulates a distinct step in a pathway. The intracellular trafficking

pathways and the SNARE machinery responsible for cell surface delivery of MMP9 and

MMP14 proteins are mostly unclear in macrophages. It was therefore the aim of this

thesis to identify the critical SNARE proteins and pathways required for MMP9 secretion

and for MMP14 cell surface delivery in macrophages. Targeting these proteins could

reduce MMP activity, macrophage infiltration and inflammation in chronic wounds.

The work reported herein shows that the expression and cell surface delivery of

MMP9 and MMP14 is upregulated upon activation of macrophages with the bacterial cell

wall component lipopolysaccharide (LPS). This is in contrast to neutrophils, which store

their MMP9 in tertiary granules prior to secretion and to some cancer cell lines that have

been shown to constitutive express MMP14 use a recycling mechanism to regulate the

level at the cell surface. Hence, trafficking pathways for individual MMPs can be cell type

specific. It was found that in macrophages, the newly made enzymes are trafficked via

the Golgi complex prior to MMP9 secretion or incorporation of MMP14 into the plasma

membrane. Immunofluorescence microscopy showed that in addition to their

localisation in the Golgi complex they both were found in late endosomes/lysosomes.

Using targeted siRNA knockdown of SNARE proteins that regulate distinct

pathways in macrophages, the responsible trafficking pathways were further

investigated. Disrupting the key MMP intracellular trafficking pathways by reducing the

levels of a key SNARE involved in their exocytosis should lead to compromised cell

surface delivery of the protein of interest. Surprisingly, for MMP9 multiple SNAREs

iv

involved in the classical or lysosomal pathways were targeted with siRNA but no

reduction in MMP9 secretion was found. However, targeting some of these SNAREs,

including VAMP3, VAMP4, VAMP7, VAMP8, Stx2 and SNAP23, led to an increase in

extracellular MMP9 levels suggesting that an endocytic clearance mechanism might be

influencing extracellular MMP9 levels and masking the exocytic pathways.

The trafficking of MMP14 was also tested. MMP14 trafficking from the Golgi

complex to late endosomes/lysosomes prior to the cell surface was found to be regulated

by Golgi R‐SNARE VAMP4, the late endosomal/lysosomal R‐SNAREs VAMP7/8 and the

surface Q‐SNARE complex Stx4/SNAP23. In macrophages MMP14 was found to be

delivered to podosomes, actin‐rich cell membrane structures implicated in cell migration

and matrix degradation. Targeting any one of these SNARE proteins lead to a reduction

in gelatin matrix degradation in the area of the podosomes. Accordingly, disrupting this

SNARE machinery also significantly reduced the ability of macrophages to effectively

invade into both 3D Matrigel™ and collagen I gels. Thus, SNARE proteins that reduce

surface MMP14 delivery have been identified providing new potential targets to reduce

MMP14 surface activity, attenuate macrophage tissue infiltration, dampen inflammation

and improve wound healing outcomes.

Overall, this thesis advances the knowledge of trafficking pathways for cell

surface delivery of MMP9 and MMP14 in macrophages and identifies key proteins

involved in these pathways. For MMP9, further investigation into its endocytic

mechanisms of clearance might allow better manipulation of MMP9 levels in the

environment to reduce excessive tissue damage in chronic wounds. SNARE machinery

responsible for MMP14 cell surface delivery, matrix degradation and macrophage

invasion was identified and could represent novel therapeutic targets for the treatment

of chronic wounds.

v

TABLE OF CONTENTS

KEYWORDS ......................................................................................................................................... I

ABSTRACT ......................................................................................................................................... II

TABLE OF CONTENTS ..................................................................................................................... V

LIST OF FIGURES ......................................................................................................................... VIII

LIST OF TABLES ................................................................................................................................ X

LIST OF ABBREVIATIONS ............................................................................................................ XI

STATEMENT OF ORIGINAL AUTHORSHIP .......................................................................... XIII

LIST OF PUBLICATIONS ............................................................................................................ XIV

ACKNOWLEDGEMENTS .............................................................................................................. XV

CHAPTER 1: LITERATURE REVIEW .......................................................................................... 1

1.1 INTRODUCTION ........................................................................................................................................ 1 1.2 THE ROLE OF MACROPHAGES IN CUTANEOUS REPAIR ....................................................................... 1 1.2.1 Cutaneous tissue repair ........................................................................................................................... 1 1.2.2 The inflammatory phase ......................................................................................................................... 4 1.2.3 Macrophage phenotypes during cutaneous repair ..................................................................... 5 1.2.4 Macrophages are necessary to promote wound healing ......................................................... 7 1.2.5 Excessive inflammation and increased numbers of macrophages compromise

wound healing .............................................................................................................................................. 8 1.3 MMPS DURING WOUND HEALING ...................................................................................................... 14 1.3.1 The role of the extracellular matrix during wound healing ................................................. 14 1.3.2 MMPs degrade ECM and non-ECM components ........................................................................ 14 1.3.3 Structure, substrates and regulation of MMPs ........................................................................... 15 1.3.4 Pathological roles of MMPs during wound healing .................................................................. 21 1.3.5 Strategies to reduce levels of matrix metalloproteinases in wounds ............................... 23 1.4 INTRACELLULAR TRAFFICKING OF PROTEINS ................................................................................... 25 1.4.1 Endocytic pathways................................................................................................................................. 26 1.4.2 Exocytic/biosynthetic pathways........................................................................................................ 27 1.4.3 Crosstalk between exocytic/biosynthetic and endocytic pathways .................................. 29 1.4.4 Trafficking machinery ............................................................................................................................ 31 1.5 AIMS ........................................................................................................................................................ 40

CHAPTER 2: MATERIALS AND METHODS ........................................................................... 41

2.1 CELL CULTURE AND TREATMENTS ..................................................................................................... 41 2.1.1 Cell culture of RAW264.7 macrophages ........................................................................................ 41 2.1.2 Activation of RAW264.7 macrophages ........................................................................................... 41 2.1.3 Disruption of protein synthesis and intracellular trafficking .............................................. 42 2.1.4 Transient overexpression of proteins .............................................................................................. 42 2.1.5 Targeted knockdown using siRNA .................................................................................................... 43 2.1.6 Fluorescent gelatin degradation assay .......................................................................................... 43

vi

2.1.7 Inverted invasion assay ......................................................................................................................... 45 2.2 BIOCHEMISTRY ASSAYS ........................................................................................................................ 47 2.2.1 Sample preparation ................................................................................................................................ 47 2.2.2 Gelatin Zymography (MMP9) ............................................................................................................. 47 2.2.3 SDS-Page and Immunoblotting ......................................................................................................... 48 2.2.4 Flow cytometry (MMP14) .................................................................................................................... 49 2.2.5 Immunofluorescence labelling ........................................................................................................... 51 2.2.6 ELISA .............................................................................................................................................................. 52 2.3 MICROSCOPY AND IMAGE ANALYSIS .................................................................................................. 53 2.3.1 Fixed cell imaging .................................................................................................................................... 53 2.3.2 Live cell imaging ....................................................................................................................................... 53 2.3.3 Image analysis ........................................................................................................................................... 53 2.4 STATISTICS ............................................................................................................................................ 54

CHAPTER 3: IDENTIFYING INTRACELLULAR TRAFFICKING PATHWAYS FOR SECRETION OF MMP9 ................................................................................................................. 55

3.1 INTRODUCTION ..................................................................................................................................... 55 3.2 RESULTS ................................................................................................................................................. 57 3.2.1 Detecting secreted MMP9 in RAW264.7 cell culture supernatants .................................. 57 3.2.2 MMP9 expression and secretion is upregulated after LPS stimulation in RAW264.7

macrophages .............................................................................................................................................. 67 3.2.3 MMP9 secreted from LPS-activated macrophages is newly made .................................... 69 3.2.4 Expression of mCherry-tagged MMP9 to study intracellular trafficking pathways . 71 3.2.5 MMP9 secretion is not dependent on intact microtubules or actin filaments ............. 73 3.2.6 Newly synthesised MMP9 is trafficked via the Golgi ................................................................ 77 3.2.7 MMP9 localises to late endosomes/lysosomes ............................................................................ 80 3.2.8 Targeting specific SNARE proteins alters MMP9 levels in conditioned medium ........ 82 3.3 DISCUSSION ........................................................................................................................................... 88 3.3.1 Detecting MMP9 ....................................................................................................................................... 88 3.3.2 MMP9 expression and secretion in RAW264.7 macrophages.............................................. 90 3.3.3 The cytoskeleton and MMP9 secretion ........................................................................................... 93 3.3.4 Trafficking pathways for MMP9 in RAW264.7 macrophages ............................................. 94

CHAPTER 4: IDENTIFYING INTRACELLULAR TRAFFICKING PATHWAYS FOR CELL SURFACE DELIVERY OF MMP14 ............................................................................................ 101

4.1 INTRODUCTION ................................................................................................................................... 101 4.2 RESULTS ............................................................................................................................................... 103 4.2.1 MMP14 levels increase upon stimulation of RAW264.7 macrophages ........................ 103 4.2.2 MMP14 is incorporated into the plasma membrane upon stimulation ....................... 106 4.2.3 MMP14 incorporated into the plasma membrane is newly synthesised upon LPS

stimulation ............................................................................................................................................... 109 4.2.4 Live imaging of MMP14 transport in macrophages ............................................................. 111 4.2.5 MMP14 cell surface delivery is not dependent on intact microtubules or actin

filaments .................................................................................................................................................... 113 4.2.6 Newly synthesised MMP14 is trafficked via the Golgi complex........................................ 117 4.2.7 R-SNARE VAMP4 regulates delivery of MMP14 from the Golgi network en route to

the cell surface ........................................................................................................................................ 120

vii

4.2.8 MMP14 is not trafficked via the recycling endosome ........................................................... 122 4.2.9 The R-SNAREs VAMP7 and VAMP8 regulate delivery of MMP14 from late

endosomes/lysosomes en route to the cell surface ................................................................ 125 4.3 DISCUSSION ......................................................................................................................................... 128 4.3.1 MMP14 expression and cell surface delivery in RAW264.7 macrophages .................. 129 4.3.2 The cytoskeleton and MMP14 cell surface delivery ............................................................... 131 4.3.3 Trafficking pathways for MMP14 in RAW264.7 macrophages ........................................ 134

CHAPTER 5: MMP14 AND MACROPHAGE MIGRATION ................................................ 140

5.1 INTRODUCTION .................................................................................................................................. 140 5.2 RESULTS .............................................................................................................................................. 142 5.2.1 The Q-SNAREs Stx4 and SNAP23 regulate delivery of MMP14 to the cell surface and

incorporation into the plasma membrane ................................................................................ 142 5.2.2 MMP14 localises to cell structures implicated in cell migration and matrix

degradation in response to matrix contact ............................................................................... 146 5.2.3 Gelatin degradation by MMP14 can be inhibited through targeting the key SNAREs

implicated in MMP14 surface delivery ........................................................................................ 150 5.2.4 3D migration of macrophages through different matrices ............................................... 155 5.2.5 SNAREs regulate macrophage invasion of Matrigel™ .......................................................... 157 5.2.6 3D migration through Collagen I can be inhibited through targeting SNAREs that

are involved in the cell surface delivery of MMP14 ............................................................... 162 5.3 DISCUSSION ......................................................................................................................................... 166 5.3.1 Trafficking to the cell surface .......................................................................................................... 167 5.3.2 Localisation of MMP14 to structures for migration and degradation ......................... 169 5.3.3 Matrix degradation by macrophages can be inhibited through targeting SNAREs

that are involved in the cell surface delivery of MMP14 ..................................................... 170 5.3.4 3D migration of macrophages can be inhibited through targeting SNAREs that are

involved in the cell surface delivery of MMP14 ........................................................................ 171

CHAPTER 6: GENERAL DISCUSSION .................................................................................... 175

REFERENCES................................................................................................................................. 184

viii

LIST OF FIGURES

Figure 1.1. Cutaneous tissue repair (from Rohl et al., 2014). ....................................................... 3 Figure 1.2. Dynamics of leukocyte extravasation and macrophage phenotypes during

wound healing. ........................................................................................................................ 5 Figure 1.3. Inflammation in wound healing (from Rohl et al., 2014). ....................................... 9 Figure 1.4. Macrophage tissue infiltration occurs through different migration processes

and requires the activity of matrix metalloproteinases. ..................................... 13 Figure 1.5. Classification of MMPs according to their domain structure (from Rohl and

Murray, 2013). ...................................................................................................................... 18 Figure 1.6. Mechanisms of pro‐MMP activation (from Rohl and Murray, 2013). .............. 19 Figure 1.7. Increased MMP levels in wounds contribute to poor wound healing

outcomes. ................................................................................................................................ 22 Figure 1.8. SNARE‐mediated fusion (modified from Murray and Stow, 2014). ................. 33 Figure 1.9. Intracellular trafficking pathways in macrophages and SNARE proteins

associated with the different compartments. .......................................................... 36 Figure 2.1. Inverted Invasion Assay. .................................................................................................... 46 Figure 3.1. The FBS concentration in the cell culture media influences RAW264.7

macrophage cell numbers and morphology. ........................................................... 59 Figure 3.2. Bovine MMP9 found in FBS masks the gelatinase activity of MMP9 secreted

by macrophages. .................................................................................................................. 61 Figure 3.3. The MMP9 ELISA Kit (R&D) is unsuitable for assaying MMP9 in RAW264.7

cell culture supernatants. ................................................................................................ 65 Figure 3.4. The Biosensis MMP9 ELISA Kit is suitable to assay changes in MMP9 levels

in RAW264.7 cell culture supernatants. .................................................................... 66 Figure 3.5. MMP9 secretion and expression is increased in response to LPS stimulation.

..................................................................................................................................................... 68 Figure 3.6. MMP9 is newly synthesised and secreted upon LPS stimulation. .................... 70 Figure 3.7. MMP9‐mCherry localises to the same places as endogenous MMP9 in

RAW264.7 macrophages. ................................................................................................. 73 Figure 3.8. MMP9 secretion is not dependent on microtubules. .............................................. 74 Figure 3.9. Disruption of actin filaments with cytochalasin does not reduce MMP9

secretion. ................................................................................................................................ 76 Figure 3.10. Newly synthesised MMP9 is trafficked to the cell surface via the Golgi

complex. .................................................................................................................................. 78 Figure 3.11. MMP9 localises to late endosomes and lysosomes. ............................................. 81 Figure 3.12. Knockdown of the surface Q‐SNAREs Stx2 and SNAP23 but not Stx3 and

Stx4, leads to an increase in MMP9 levels in cell culture supernatants. ...... 86 Figure 3.13. Knockdown of VAMP2, VAMP3, VAMP4, VAMP7 and VAMP8 increases the

level of MMP9 in cell culture supernatants. ............................................................. 87 Figure 3.14. Hypothetical schematic of LRP‐1‐mediated endocytosis of MMP9. ............ 100 Figure 4.1. MMP14 levels increase upon stimulation of murine macrophages. .............. 105 Figure 4.2. MMP14 is incorporated into the plasma membrane upon stimulation. ...... 109 Figure 4.3. MMP14 incorporated into the plasma membrane upon LPS stimulation is

newly synthesised. ............................................................................................................ 110

ix

Figure 4.4. Fluorescently‐labelled MMP14 can be transiently overexpressed to monitor MMP14 trafficking in live macrophages. ................................................................ 113

Figure 4.5. MMP14 cell surface levels are not dependent on microtubules. .................... 115 Figure 4.6. Disruption of cortical actin does not affect MMP14 cell surface levels. ....... 117 Figure 4.7. Newly synthesised MMP14 is trafficked via the Golgi complex. ..................... 118 Figure 4.8. R‐SNARE VAMP4 regulates delivery of MMP14 from the Golgi network en

route to the cell surface. ................................................................................................. 121 Figure 4.9. MMP14 is not trafficked via the recycling endosome. ......................................... 125 Figure 4.10. R‐SNAREs VAMP7 and VAMP8 regulate delivery of MMP14 from late

endosomes/lysosomes en route to the cell surface. ........................................... 126 Figure 4.11.. Schematic of MMP14 trafficking in macrophages. ............................................ 138 Figure 5.1. The Q‐SNAREs Stx4 and SNAP23 regulate delivery of MMP14 to the cell

surface. .................................................................................................................................. 143 Figure 5.2. The Q‐SNAREs Stx4 and SNAP23 colocalise with MMP14 on the plasma

membrane............................................................................................................................ 145 Figure 5.3. MMP14 localises to cell structures implicated in cell migration in response

to matrix contact. .............................................................................................................. 147 Figure 5.4. MMP14 localises to cell structures implicated matrix degradation in

response to matrix contact. .......................................................................................... 149 Figure 5.5. Gelatin degradation by MMP14 can be inhibited through targeting the Golgi

complex R‐SNAREs VAMP4 or the late endosome/lysosome R‐SNAREs VAMP7 or VAMP8. ............................................................................................................ 151

Figure 5.6. Gelatin degradation by MMP14 can be inhibited through targeting the Q‐SNARE complex Stx4/SNAP23. ................................................................................... 155

Figure 5.7. 3D migration of macrophages through Matrigel™ and Collagen I matrices. .................................................................................................................................................. 157

Figure 5.8. 3D migration through Matrigel™ can be inhibited through targeting SNAREs that are involved in the cell surface delivery of MMP14. ................................. 161

Figure 5.9. 3D macrophage migration through collagen can be inhibited through targeting SNAREs that are involved in cell surface delivery of MMP14. ... 165

Figure 5.10. Schematic of MMP14 trafficking to podosomes in macrophages. ............... 168

x

LIST OF TABLES

Table 1.1. Classification of MMPs according to their substrate specificity (from Rohl and Murray, 2013). ......................................................................................................................... 16

Table 1.2. Alterations in MMP and TIMP levels in chronic venous, diabetic, pressure and mixed ulcers (from Rohl and Murray, 2013). ............................................................. 22

Table 1.3. Q‐ and R‐SNARE family members. ................................................................................... 34 Table 2.1. List of plasmids used for this project. ............................................................................. 42 Table 2.2. siRNA sequences used to knockdown SNARE proteins. ......................................... 43 Table 2.3. Antibodies used for immunoblotting. ............................................................................. 49 Table 2.4. Antibodies used for flow cytometry. ............................................................................... 51 Table 2.5. Antibodies and fluorescent probes used for immunofluorescence labelling. 52

xi

LIST OF ABBREVIATIONS

2D two dimensional 3D three dimensional αSNAP α‐soluble N‐ethylmaleimide‐sensitive‐factor accessory‐protein AP adaptor protein ATCC American type cell collection ATP adenosine triphosphate BFA brefeldin A BSA bovine serum albumin CaCl2 calcium chloride CD cluster of differentiation CHX cycloheximide CLIC clathrin‐independent carrier CLIP class II‐associated invariant chain peptide CO2 Carbon dioxide COPI/II coatamer protein I/II CTL cytotoxic T lymphocyte Cyto cytochalasin DABCO 1,4‐diazabicyclo‐2,2,2‐octane DAMP Damage‐associated molecular pattern DAPI 4’,6’‐diamidino‐2‐phenylindole DMSO dimethyl sulfoxide DNA deoxyribonucleic acid DTT dithiothreitol ECM extracellular matrix EDTA ethylene di‐amine tetra‐acetic acid EEA1 early endosomal antigen 1 eGFP enhanced green fluorescent protein ELISA enzyme‐linked immunosorbent assay ER endoplasmic reticulum ERGIC ER‐Golgi‐intermediate compartment F‐actin filamentous actin FcR Fc receptor FBS foetal bovine serum fMLP N‐formyl‐methionyl‐leucyl‐proline GAP GTPase‐activating proteins GDI GDP dissociation inhibitor GDP guanosine‐5’‐diphosphate GEF guanine nucleotide exchange factors GM130 Golgi marker of 130‐kDa GS27/28 Golgi SNARE of 27/28 kDa GTP guanosine‐5’‐triphosphate HCl Hydrogen chloride HRP horseradish peroxidase IFN interferon Ig immunoglobulin IL interleukin iNOS inducible nitric oxide synthase kDa kilodalton LAMP lysosome‐associated membrane protein

xii

LBPA lysobisphosphatic acid LDL low‐density lipoprotein LPS lipopolysaccharide LTA lipoteichoic acid M6P mannose‐6‐phosphate M6PR mannose‐6‐phosphate receptor MFI median fluorescence intensity MMP matrix metalloproteinase MHCII major histocompatibility complex II MIP‐1 macrophage inflammatory protein‐1 MTOC microtubule organising centre MTT 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium bromide NA numerical aperture NaCl sodium chloride NaOH sodium hydroxide NF‐kB nuclear factor‐kB NK cells natural killer cells Noc nocodazole NSF N‐ethylmaleimide‐sensitive factor PAGE polyacrylamide gel electrophoresis PAMP pathogen‐associated molecular pattern PBS phosphate‐buffered saline PDGF platelet‐derived growth factor PDI protein disulfide isomerase PFA paraformaldehyde pH power of hydrogen PRR pathogen recognition receptor RNA ribonucleic acid ROS reactive oxygen species RPMI Roswell Park Memorial Institute RT room temperature SEM standard error of the mean SDS sodium dodecyl sulphate siRNA short interfering ribonucleic acid SM Sec/Munc SNAP23/25/29/47 synaptosome‐associated protein of 23/25/29/47 kDa SNARE soluble N‐ethylmaleimide‐sensitive‐factor accessory‐protein‐

receptor ssGFP secreted form of GFP Stx syntaxin TACE TNF‐α converting enzyme TAPI TACE protease inhibitor Tfn transferrin TfR transferrin receptor TGF transforming growth factor TGN trans‐Golgi network TLR Toll‐like receptor TNF tumor necrosis factor Tris tris‐(hydroxymethyl)‐aminomethane UV ultraviolet V‐ATPase vesicular adenosine‐5’‐triphosphatase VAMP vesicle‐associated membrane protein VEGF Vascular endothelial growth factor Vti1a/b vesicle transport through interaction with t‐SNARE homolog 1A/B

xiii

STATEMENT OF ORIGINAL AUTHORSHIP

The work contained in this thesis has not been previously submitted to meet

requirements for an award at this or any other higher education institution. To the best

of my knowledge and belief, the thesis contains no material previously published or

written by another person except where due reference is made.

Statement of Contributions by Others to the Thesis as a Whole

Andreea Zaharia performed approximately ⅔ of the immunoblots that provided

data for the densitometry graph in Figure 3.12B, ⅓ of the immunoblots that provided

data for the densitometry graph in Figure 3.13B and the immunoblot in Figure 5.1A.

Sandrine Roy provided the fluorescence microscopy images found in Figure 5.3. Dr

Rachael Murray and Dr Melissa Fernandez supervised the project, contributed to its

conception and design, and revised drafts of this thesis.

Signature: _________________________

Date:

28 November 2016

QUT Verified Signature

xiv

LIST OF PUBLICATIONS

Peer-reviewed journal articles associated with this thesis

Rohl, J., Zaharia, A., Rudolph, M., Murray, R. Z. (2015). The role of inflammation in cutaneous repair. Wound Practice and Research, 23(1), 8‐15.

Rohl, J., Murray, R. Z. (2013). Matrix metalloproteinases during wound healing – a double edged sword. Wound Practice and Research, 21(4), 174‐180.

Conference and meeting presentations

INTERNATIONAL CONFERENCES:

Rohl, J., Zaharia, A., Murray, R. Z. (2015, October). The surface Q-SNARE complex Stx4/SNAP23 regulates delivery of MMP14 to the plasma membrane in macrophages. Joint Meeting of the ETRS & WHS. Copenhagen, Denmark.

NATIONAL CONFERENCES:

Rohl, J., Murray, R. Z. (2014, September). Macrophage infiltration of wound tissue. Cutaneous Biology Meeting, North Stradbroke Island.

Rohl, J., Fernandez, M. L., Murray, R. Z. (2014, May). Intracellular Trafficking And Secretion Of Matrix Metalloproteinases 2, 9 and 14. AWTRS Conference, Gold Coast.

Rohl, J., Fernandez, M. L., Murray, R. Z. (2013, November). Intracellular Trafficking And Secretion Of Matrix Metalloproteinases Up-regulated In Chronic Wounds. IHBI Inspires Postgraduate Student Conference, Brisbane.

INTERNATIONAL SEMINARS:

Rohl, J. (2015, October). Identifying trafficking machinery proteins responsible for Matrix Metalloproteinase release from macrophages to improve wound healing outcomes. Polymorfien Seminar Series, Statens Serum Institut, Copenhagen, Denmark.

POSTER PRESENTATIONS:

Rohl, J., Murray, R. Z. (2015). Intracellular trafficking and secretion of MMP14 from macrophages. Brisbane Cell and Developmental Biology, University of Queensland (UQ), Brisbane.

Rohl, J., Murray, R. Z. (2015). Intracellular trafficking and secretion of MMP14 from macrophages. ASMR Queensland Postgraduate Student Conference, QIMR Berghofer Medical Research Institute, Brisbane.

Rohl, J., Murray, R. Z. (2014). The surface Q-SNARE complex Stx4/SNAP23 regulates delivery of MMP14 to the plasma membrane from the classical transport pathway in macrophages. Brisbane Cell and Developmental Biology Meeting, UQ, Brisbane.

Rohl, J., Murray, R. Z. (2014). The surface Q-SNARE complex Stx4/SNAP23 regulates delivery of MMP14 to the plasma membrane from the classical transport pathway in macrophages. IHBI Inspires Postgraduate Student Conference, Gold Coast.

Rohl, J., Fernandez, M. L., Murray, R. Z. (2014). Intracellular Trafficking And Secretion Of Matrix Metalloproteinases 2, 9 and 14. Australasian Wound & Tissue Repair Society Conference, Gold Coast.

Rohl, J., Fernandez, M. L., Murray, R. Z. (2013). Intracellular Trafficking and Secretion of Matrix Metalloproteinases. Brisbane Cell and Developmental Biology Meeting, UQ, Brisbane.

xv

ACKNOWLEDGEMENTS

This PhD journey is coming to an end and I would not have made it to this point

without the support of dedicated scientists and invaluable friends and I would like to

take the opportunity to express my gratitude towards all of them:

First and foremost, I would like to thank my Principal Supervisor, Dr Rachael

Murray, for believing in me and putting everything necessary, including my scholarship,

in motion so I could start the project here at QUT, for making me fully appreciate the

beauty of macrophages as well as teaching me all these great cell biology concepts and

methods. You were always available when I needed advice and I appreciate the trust you

put into my work. I am also deeply grateful for all the other opportunities you

encouraged me to pursue throughout my candidature, such as attending conferences,

partaking in the postgraduate student conference committee or working as a sessional

academic. My sincere thanks also go to my Associate Supervisor, Dr Melissa Fernandez,

for her ongoing contributions even after her relocation to another institute. I would also

like to extend my appreciation to Dr Eliza Whiteside for her support in the very beginning

of my PhD. Many thanks also to my internal panel members, Associate Prof Terry Walsh,

Prof Rik Thompson, Dr Annalese Semmler and Dr Danica Hickey, as well as my external

reviewers for taking the time to read my thesis and providing me with their feedback. I

furthermore owe a great debt of gratitude to Cell Imaging Facility Manager Dr Leonore

de Boer as well as the rest of the lab services team for their help throughout the years

here at IHBI.

Special thanks go to Andreea for being the best possible lab member I could ever

ask for as well as being an incredible friend – the last few months without you while you

have been following your snowboard dreams have not quite been the same. I am also

very grateful to all my other IHBI peeps that always kept me smiling, Chris, Dan, Jacqui,

James, Lucas, Pat, Simon – thank you guys for always believing in me and all the great

times we had inside and outside of IHBI.

I would also like to thank my family and friends back in Berlin for continuously

supporting me in all aspects of my life no matter what I do or how far away I might be.

Last but not least I want to thank Brent for sticking with me through good times

and bad times in the past five years. To many more adventures to come!

xvi

Chapter 1: Literature review 1

Chapter 1: Literature review

1.1 Introduction

Chronic wounds are wounds that show no significant healing after four weeks of

receiving evidence‐based care [Frykberg and Banks, 2015]. These include venous,

diabetic and pressure ulcers and are debilitating and often very painful. Approximately

450,000 Australians develop a chronic wound annually and wound care is estimated to

cost the Australian health system an estimated $3 billion a year [Wound CRC, 2014].

Inflammation is part of the normal wound healing process and occurs even in the absence

of infection. However, prolonged and excessive inflammation is a major contributing

factor to the recalcitrant nature of chronic non‐healing wounds [Bannon et al., 2013]. The

number of macrophages infiltrating a wound can determine the level of inflammation

[Martin and Leibovich, 2005]. The influx of macrophages into the skin is partly dependent

on the exocytosis of matrix metalloproteinases (MMPs). These proteases promote

macrophage mobility by degrading the surrounding extracellular matrix (ECM). They

also remove damaged ECM, and play a role in angiogenesis and tissue remodelling during

wound healing [Stechmiller et al., 2010]. In order to inhibit MMP cell surface delivery it

is necessary to elucidate their intracellular transport route and identify proteins that

regulate these pathways in macrophages. The polarised delivery of proteins within

membrane‐bound vesicles to the cell surface for exocytosis requires membrane fusion

mediated by soluble N‐ethylmaleimide‐sensitive‐factor attachment protein receptor

(SNARE) proteins at all points in the trafficking pathway [Jahn and Scheller, 2006] (see

section 1.4). Identifying the critical SNARE proteins and pathways for MMP secretion is

therefore important to identify potential targets to reduce MMP activity as a therapeutic

treatment of chronic wounds.

1.2 The role of macrophages in cutaneous repair

1.2.1 Cutaneous tissue repair

The skin is the biggest organ of the human body and serves as physical barrier to

maintain internal homeostasis. This barrier is crucial for water retention, as well as

regulation of the body’s temperature, and provides protection from chemical or

mechanical assault, UV radiation, allergens and harmful pathogens. Playing such an

important role, it is critical that the integrity of the skin is conserved. If compromised,

timely repair is vital to ensure that the skin can fulfil its functions. The intricate process

2 Chapter 1: Literature review

of wound healing is characterised by four overlapping and interdependent phases that

involves multiple different cell types, such as fibroblasts, keratinocytes, neutrophils and

macrophages, from which some are resident, while others have to be recruited to the

wound site [Darby et al., 2014; Koh and DiPietro, 2011; Martin and Leibovich, 2005]. The

four phases are haemostasis, inflammation, proliferation and maturation and occur

irrespective of the type of injury [Martin and Leibovich, 2005].

Immediately upon wounding, haemostasis and coagulation are initiated to halt

blood loss [Martin and Leibovich, 2005]. As a result of the injury, collagen fibres are

exposed and clotting factors released, which induces platelets to aggregate and secrete

chemical stimuli. Together with cross‐linked fibrin and fibronectin the platelets form a

plug that acts to stop blood loss, as well as forms a provisional wound matrix. Next,

chemokines, cytokines and growth factors, secreted by platelets and resident

inflammatory cells, initiate the inflammatory phase through the recruitment of

phagocytic cells from the blood stream to the wounded tissue (Figure 1.1). Neutrophils

are the first to arrive at the site of injury only a few hours post‐wounding, while

macrophages appear slightly later [Gray et al., 2011]. Together, they are responsible for

the removal of tissue debris, as well as the phagocytosis and killing of pathogens to

protect the wound from infection. Normally, the inflammatory phase lasts for about 5–7

days [Martin and Leibovich, 2005]. During this time, macrophages also release growth

factors and other chemical stimuli that are important for the proceeding phases of the

wound healing process. These stimuli will initiate the proliferation phase, which starts

around day 3 post‐wounding and lasts for about a week. During this phase, fibroblasts

arrive at the wound site to produce collagen III and fibronectin, which is organised into

loose bundles, acting as a scaffold for the wound healing process [Darby et al., 2014].

This aids re‐epithelialisation of the wound as keratinocytes proliferate and migrate

across the new matrix material [Lau et al., 2009]. Wound contraction through the activity

of myofibroblasts helps to reduce the wound diameter and speeds up the healing process

[Darby et al., 2014]. To restore blood supply in the new tissue, blood capillaries are

formed through sprouting from pre‐existing vessels in a process called angiogenesis.

Factors secreted by macrophages stimulate endothelial cells to migrate into the

wounded area to build the vasculature [Johnson and Wilgus, 2014]. Macrophages,

fibroblasts, the provisional matrix and new blood vessels constitute the granulation

tissue and remodelling of it is initiated through the proteolytic activity of MMPs

[Stechmiller et al., 2010]. During the remodelling phase of the wound healing process

collagen III is replaced with collagen I, laid down in parallel bundles along tension lines

to complete wound contraction and increase the strength of the new tissue [Stechmiller


et al., 2010]. As this structure is different from the basket‐weave collagen present in

uninjured skin, the healing process ultimately leads to the formation of a scar.

Elimination of the hypoxic environment and the apoptosis of cells within the granulation

tissue results in avascular and acellular tissue at the end of the maturation process

[Greenhalgh, 1998].

Figure 1.1. An overview of cutaneous tissue repair (from Rohl et al., 2015). Together with cross-linked fibrin and fibronectin the platelets form a plug that acts to stop blood loss, as well as forms a provisional wound matrix. The inflammatory phase is initiated through the recruitment of phagoctytic cells from the blood stream to the wounded tissue. Neutrophils are the first to arrive at the site of injury, while macrophages follow a bit later. Together, they are responsible for the removal of tissue debris, as well as the phagocytosis and killing of pathogens to protect the wound from infection. Macrophages also release growth factors and other chemical stimuli that are important for the proliferation phase. Fibroblasts arrive at the wound site to produce collagen III and fibronectin, acting as a scaffold for the wound healing process. This aids re-epithelialisation of the wound as keratinocytes proliferate and migrate across the new matrix material. Factors secreted by macrophages stimulate endothelial cells to migrate into the wounded area to rebuild the vasculature. Macrophages, fibroblasts, the provisional matrix and new blood vessels constitute the granulation tissue and remodelling of it is initiated through the proteolytic activity of MMPs.


1.2.2 The inflammatory phase

Inflammation is a normal part of the cutaneous wound healing process even in

uninfected wounds and is mediated by resident mast cells, as well as neutrophils,

macrophages and T cells that are recruited from the blood stream upon injury [Eming et

al., 2007; Koh and DiPietro, 2011]. Upon injury, resident mast cells degranulate to release

factors, such as vasoactive histamines. This results in vascular dilation and increased

vessel permeability and, together with the secretion of chemotactic factors, allows

extravasation of leukocytes. Neutrophils and macrophages are the main leukocytes to

enter the tissue after wounding, as revealed by in vivo imaging techniques in mice skin

wounds [Rodero et al., 2014]. Yet, they show a different dynamic behaviour in their

migration and recruitment (Figure 1.2). In mice, neutrophil recruitment can be observed

from 6‐12 h with numbers peaking around one day after wounding, whereas

macrophages arrive slightly later but continue to actively infiltrate the tissue until about

48 h post‐injury [Ellett et al., 2011; Gray et al., 2011; Kim et al, 2008; Rodero et al., 2013].

Recruited neutrophils secrete proteases and antimicrobial products, such as ROS, to

breakdown damaged tissue and kill pathogens. Clearance of debris and microbes occurs

through phagocytosis. Recruitment of neutrophils stops around day 2 and cell numbers

decline due to apoptosis and then the phagocytosis of spent neutrophils by macrophages

[Kim et al., 2008]. Wound associated macrophages are predominantly recruited from

circulating monocytes and resident tissue macrophages play only a minor role in the

outcome of skin repair [Macdonald et al., 2010; Rodero et al., 2013; Rodero et al., 2014].

Infiltration of the wound tissue by macrophages occurs in two waves [Rodero et al.,

2014]. The first small pool of monocytes enters the wound shortly after injury when the

blood vessels are dilated and porous as a result of mast cell degranulation [Rodero et al.,

2014]. However, this increased vascular permeability is only transient. Upon repair of

the leakage monocytes extravasation must occur actively following their adhesion to the

endothelium [Rodero et al., 2013]. Within the wound, monocytes differentiate into

wound‐associated macrophages through signals from the local environment, such as

matrix contact, cytokines and pathogens [Brown et al., 2014; Mosser and Edwards,

2008]. Once activated, they play many important roles in the wound healing process.

Apart from the phagocytosis of spent neutrophils, macrophages assist in the pathogen

clearance and wound debridement. The breakdown of devitalised tissue is facilitated

through the activity of proteases, such as MMPs, and is important for proper

incorporation of new matrix material. Macrophages do not only secrete large quantities

of MMPs but they also release many pro‐inflammatory cytokines, such as TNFα, IL‐1 and

IL‐6, to mediate the inflammatory response [Daley et al., 2010; Rodero et al., 2013.


Additionally, they secrete multiple growth factors, including transforming growth factor

(TGF), vascular endothelial growth factor (VEGF) and platelet‐derived growth factor

(PDGF), which actively stimulate fibroblast proliferation/differentiation and collagen

production, as well as angiogenesis [Mosser et al., 2008]. Macrophage numbers are

stable until four to five days after wounding and they then decrease to pre‐wounding

levels by day 14 [Rodero et al., 2013; Kim et al, 2008; Ellett et al., 2011].

Figure 1.2. Dynamics of leukocyte extravasation and macrophage phenotypes during wound healing (from Rohl et al., 2015). Neutrophil recruitment can be observed within a few hours with numbers peaking around one day after wounding, whereas macrophages arrive slightly later but continue to actively infiltrate the tissue for longer. Recruitment of neutrophils stops around day two and cell numbers decline due to apoptosis and then the phagocytosis of spent neutrophils by macrophages. The first small pool of monocytes enters the wound shortly after injury when the blood vessels are dilated and porous as a result of mast cell degranulation. Upon repair of the leakage monocytes extravasation must occur actively following their adhesion to the endothelium. Within the wound, monocytes differentiate into wound-associated macrophages. Once activated, they play many important roles in the wound healing process. During the early phases of wound healing, pro-inflammatory M1 macrophages predominate. They express high levels of the cell surface markers Ly6c and CCR2 and secrete large quantities of the pro-inflammatory cytokines TNFα, IL-6 and IL-1 as well as MMP9. Macrophages of the M2 type are found later in the repair process in response to changes in the environment. They express only low levels of Ly6c but express arignase-1, mannose receptor (CD206) and IL-4 receptor. They secrete anti-inflammatory cytokines, such as IL-10 and TGFβ, as well as ECM components.

1.2.3 Macrophage phenotypes during cutaneous repair

Macrophages play many different and important roles during wound healing,

such as pathogen defence, wound debridement as well as stimulation of fibroblast

proliferation and angiogenesis [Mosser et al., 2008]. As such they mitigate wound


infection and facilitate functional tissue restoration. To accommodate their diverse

functions during the wound healing process macrophages can adapt various phenotypes

[Mosser and Edwards, 2008]. Accordingly, different types of macrophages can be found

in acute murine wounds that change over time. These macrophage phenotypes are

aligned along a continuous spectrum between two extremes that are called ‘M1’ and ‘M2’

where M1 macrophages are considered pro‐inflammatory while macrophages of the M2

type are anti‐inflammatory and promote tissue repair. According to this concept, M1

macrophages can be ‘classically activated’ by IFNγ, pathogen‐associated molecular

patterns (PAMPs), such as the bacterial cell wall component lipopolysaccharide (LPS),

and/or damage‐associated molecular patterns (DAMPs), such as degraded matrix

components [Bianchi, 2007; Mosser et al., 2008]. M1 macrophages secrete proteolytic

enzymes, such as MMPs, that would allow them to infiltrate and remodel the wound

[Daley et al., 2010]. These macrophages have strong microbicidal and phagocytic

capacity and produce pro‐inflammatory cytokines (TNFα, IL‐1, IL‐6), superoxide anions,

as well as oxygen and nitrogen radicals [Daley et al., 2010; Rodero et al., 2013;

Willenborg et al., 2012; Mirza et al., 2013]. They are responsible for pathogen control

and removal of necrotic cells and tissue debris, but also recruit other cells to the wound.

The removal of microbes, as well as apoptotic cells and debris, reduces the inflammatory

potential of the wound environment. Polarisation towards the anti‐inflammatory ‘M2’

type is facilitated by phagocytosis of spent neutrophils and ‘alternative activation’

through IL‐4 and IL‐23 [Brown et al., 2014]. M2 macrophages also produce effector

molecules, such as VEGF, PDGFβ, and TGFβ [Mosser et al., 2008]. This recruits

endothelial cells and fibroblasts, promotes the deposition of ECM material as well as

myofibroblast differentiation and has pro‐angiogenic properties.

During the early phases of wound healing (day 1 post‐wounding), pro‐

inflammatory M1 macrophages predominate [Daly et al., 2005; Mirza et al., 2013; Rodero

et al., 2013; Willenborg et al., 2012]. They express high levels of the cell surface markers

Ly6c and CCR2 and secrete large quantities of the pro‐inflammatory cytokines TNFα, IL‐

6 and IL‐1 as well as MMP9 [Daley et al., 2010; Mirza et al., 2013; Rodero et al., 2013;

Willenborg et al., 2012]. Macrophages of the M2 type are found later in the repair process

(day 7 post‐wounding) [Brown et al., 2014; Daley et al., 2005; Rodero et al., 2013]. They

express only low levels of Ly6c but express arignase‐1, mannose receptor (CD206) and

IL‐4 receptor. They secrete anti‐inflammatory cytokines, such as IL‐10 and TGFβ, as well

as ECM components [Daley et al., 2010; Rodero et al., 2013].


1.2.4 Macrophages are necessary to promote wound healing

Immune cell mediated inflammation is a normal part of tissue repair and

necessary to trigger subsequent healing stages. Yet, the role of neutrophils for this is

conflicting within the literature. Neutrophil depletion using anti‐neutrophil serum had

been shown to have no impact on skin repair in a guinea pig model [Simpson and Ross

1972] but accelerated wound closure in mice [Dovi et al., 2003]. However, there is

evidence that defective neutrophil recruitment due to CXCR2 (receptor for IL‐8)

knockout in mice leads to altered monocyte recruitment, delayed re‐epithelialisation and

decreased neovascularisation [Devalaraja et al., 2000]. Another study using integrin β2

(CD18) knockout mice suggests that the failure to recruit neutrophils deprives

macrophages from phagocytosing apoptotic neutrophils necessary for macrophage

differentiation into the M2 type causing wound healing defects [Peters et al., 2005].

Early experiments depleting macrophages using anti‐macrophage serum in

guinea pig skin healing models showed impaired clearance of wound tissue, reduced

fibroblast accumulation and disturbance in the healing process [Leibovich and Ross

1975]. However, pleiotropic anti‐inflammatory effects due to the hydrocortisone used in

the study cannot be excluded. Wound healing studies in mice, which have been depleted

of the monocyte chemoattractant MIP‐1α (also known as CCL3), showed not only

reduced collagen deposition but also impaired angiogenesis [DiPietro et al., 1998].

However, CCL3 does not only mediate macrophage but also neutrophil recruitment

[Reichel et al., 2009]. Furthermore, knockout of MIP‐1α in mice had no effect on wound

healing outcomes [Low et al., 2001]. Alternatively, in another murine wound healing

model with reduced infiltration of neutrophils and macrophages, due to simultaneous

knockout of the adhesion molecules ICAM‐1 and L‐selectin, there was a reported delay

in wound repair [Nagaoka et al., 2000].

Advancement in our understanding of the diverse roles that macrophages play in

wound repair has been achieved through temporal, rather than complete, depletion of

this cell type during the different phases of the wound healing cascade [Lucas et al.,

2010]. In a murine liver injury model, depletion of macrophages during injury reduces

collagen deposition and improves fibrosis [Duffield et al., 2005]. Depletion of

macrophages during recovery on the other hand leads to a persistent fibrotic response

[Duffield et al., 2005]. Similar observations have been made in a murine cutaneous

wound healing model where macrophages are temporally depleted. Depletion of

macrophages early during the inflammatory phase of skin repair (days 0‐5) resulted in

a minimised scar but impaired the rate of wound closure [Lucas et al., 2010]. However,


macrophage influx subsequent to early depletion rescues the wound closure rates

showing that depletion of macrophages in the early stages improves wound healing

quality without compromising wound closure [Rodero et al., 2013]. In contrast,

depletion of macrophages during tissue formation (days 4‐9) had negative effects on

wound closure and vascularisation due to decreased levels of TGFβ and VEGF, whereas

depletion of macrophages during the late stage of tissue maturation (days 9‐14) had no

significant impact [Lucas et al., 2010]. These studies show that macrophages have strong

impact on different aspects of cutaneous repair, such as scar formation and wound

closure, but that this is strongly time dependent.

1.2.5 Excessive inflammation and increased numbers of macrophages compromise wound healing

After neutrophils have apoptosed, macrophages are the predominant cell type to

affect the level of inflammation. The kinetics of macrophage recruitment upon tissue

damage, as described above, are representative of the normal wound healing processes.

Alterations in this process can result in excessive inflammation and compromised wound

healing outcomes, such as chronic non‐healing wounds or increased fibrosis [Rohl et al.,

2015].

1.2.5.1 Inflammation and macrophages regulate scar formation

In order to restore the integrity of the skin in a timely manner, wound healing

processes have evolved to meet functional restoration over aesthetic or pre‐injury skin

appearance. The result of this is the formation of a scar, which is comprised of parallel

aligned collagen fibres, rather than the basket weave structures found in uninjured skin.

Although reforming the barrier and closing the wound, the scarred tissue lacks the

tensile strength of the surrounding normal skin. Dependant on localisation of the scar

and the degree of scar formation, the functionality of the new tissue can also be

compromised, additionally to the cosmetic problems they may cause. If present close to

or over joints, scars can restrict and inhibit movement, especially in growing children.

Excessive scar tissue deposition resulting in hypertrophic or keloid formations can cause

further discomfort, tenderness, and possible irritation, ranging from itchy to painful.

Evidence suggests that macrophage‐mediated inflammation determines the level

of scar formation [Stramer et al., 2007] (Figure 1.3). Although macrophages are recruited

to both adult and foetal wound sites, the numbers and persistence of these cells are lower

in the foetus than in the adult. This results in scar‐free healing in the foetal mice (until

the second trimester when the foetus begins to develop a full immune system) compared


to scar formation in the adult mice [Cowin et al., 1998]. Likewise, the healing process

without excessive inflammation and scarring in mouse embryos perseveres as long as

the monocyte lineage has not fully developed while the artificial induction of

inflammation in foetal wounds will cause scar formation [Hopkinson‐Woolley et al.,

1994]. Furthermore, the repair process of the oral mucosa is characterised by low levels

of macrophages resulting in faster healing, reduced inflammation and less scarring when

compared to cutaneous wounds [Szpaderska et al., 2003]. In further support of this, PU.1

transcription factor knockout mice (Sfp1‐/‐), which are lacking macrophages,

neutrophils, B cells, mast cells and eosinophils, show scar‐free healing that is

independent of inflammation [Martin et al., 2003].

Together with the results from the temporal macrophage depletion studies, this

suggests that if we could reduce early macrophage‐mediated inflammation (days 0–5) it

might prevent the formation of hypertrophic scars.

Figure 1.3. Inflammation in wound healing (from Rohl et al., 2015). The time period and extend of the inflammatory phase of tissue repair under different wound healing conditions.

1.2.5.2 High levels of inflammation and macrophages contribute to the chronicity of

wounds

Excessive macrophage‐mediated inflammation also plays a role in pathological

tissue repair, as seen in chronic wounds, where proteolytic enzymes and pro‐


inflammatory cytokines impede the healing process (Figure 1.3) [Bannon et al., 2013;

Pierce, 2001]. While normal wound healing is characterised by transient macrophage

infiltration of the tissue, macrophage numbers are highly elevated and their levels do not

reduce over time in impaired wound healing models of obese (ob/ob) and diabetic

(db/db) mice [Goren et al., 2003; Wetzler et al., 2000], as well as in chronic wounds of

human venous and diabetic leg ulcers [Rosner et al., 1995; Loots et al., 1998]. Although

neutrophil numbers in acute wounds are high, biopsies of human chronic wounds have

shown that 80% of the cells in the wound margin are macrophages [Sindrilaru et al.,

2011]. Interestingly, macrophages found within chronic wounds can be characterised as

persistent pro‐inflammatory M1 macrophages rather than switching towards the

healing‐promoting M2 type seen in acute wounds 5 days following wounding [Sindrilaru

et al., 2011]. The phenotype of these M1 macrophages can be characterised through their

expression of high levels of the pro‐inflammatory M1 markers TNFα, IL‐12p40 and CCR2

while only moderate levels of the M2 markers arginase, CD206, Dectin‐1, IL‐10, IL‐4Rα

and CD36 [Sindrilaru et al., 2011]. The enhanced release of TNFα and hydroxyl radicals

from these pro‐inflammatory M1 activated macrophages was shown to be responsible

for compromised healing outcomes [Sindrilaru et al., 2011]. Together, this suggests that

elevated numbers of macrophages with an unrestrained pro‐inflammatory response

play a major role in the inflammation driving the chronicity in wounds. Thus, reducing

macrophage numbers in chronically inflamed wounds is expected to improve healing

outcomes. Accordingly, dampening inflammation by reduction of macrophages using

antibodies has been shown to reinstate tissue repair in healing impaired obese mice

[Goren et al., 2007]. Understanding inflammation in the context of wound healing has

therefore become an important focus in the treatment of chronic wounds. The

development of strategies to reduce tissue infiltration by macrophages, as well as

limiting the amount of macrophage‐derived molecules that exacerbate inflammation,

could help to break the vicious cycle of inflammation in non‐healing wounds.

1.2.5.3 Strategies to reduce macrophage tissue infiltration and inflammation

Excessive inflammation in wounds could potentially be reduced via a number of

different approaches. These may include reducing macrophage infiltration of the wound

tissue and mitigating or inhibiting the release of pro‐inflammatory cytokines and other

factors that actively damage the wound environment and exacerbate inflammation.


Reducing macrophage tissue infiltration

Monocytes are recruited from the blood to the wound site where they become

activated and exhibit their inflammatory behaviour [Koh and DiPietro, 2011; Eming et

al., 2007]. The initial step in this process is for monocytes to adhere to the endothelium

of the inner blood vessel wall. They then move along the vessel wall in a rolling

movement until they cross the endothelium and basement membrane in a process called

diapedesis. Once in the tissue, macrophages have to migrate through the 3D environment

of the interstitial tissue, which is composed predominantly of collagen I and other

extracellular matrix (ECM) components.

Adhesion of monocytes to endothelial cells of the blood vessel is facilitated via

contact of adhesion molecules on the endothelium with integrins located on the cell

surface of monocytes [Muller, 2003]. This interaction can be disrupted either by

targeting selectins, which are adhesion molecules on the endothelium, or integrins,

which are the cognate adhesion molecules on the cell surface of immune cells [Muller,

2003]. In a mouse model lacking P‐ and E‐selectins macrophage recruitment is

significantly reduced [Subramaniam et al., 1997]. However, as macrophage infiltration

in these acute wounds is disturbed throughout the entire healing process beneficial

factors released by macrophages are subsequently unavailable and as a result, the

overall healing is impaired. Alternatively, macrophage‐mediated inflammation was

reduced in a rabbit model of burn injury by treatment with an anti‐integrin alpha L

antibody during the first 24 h post‐injury to target the adhesion molecules on the

immune cell [Bucky et al., 1994]. Moreover, in a rabbit burn model, targeting ICAM‐1 and

integrin β2 has, likewise, shown improved healing [Bucky et al., 1994; Fuchs et al., 2006].

Treatment with ICAM‐1 antibodies also led to accelerated healing in human burn injuries

when administered within 6 h of injury [Mileski et al., 2003]. Whether treatments such

as these have a positive impact on scar formation or their impact on chronic wounds has

not been determined and requires further investigation.

Following successful adhesion of monocytes to the endothelium, these cells then

migrate along the inner blood vessel wall in a 2D migration mode [Verollet et al., 2011].

For this, the macrophage polarises into an elongated shape with a flattened migration

front at the leading edge, called a lamellipodium. Establishment of the lamellipodium is

achieved through actin filament assembly. This process of migration involves the

continual formation of adhesive contacts, with the substratum at the leading edge and

the disassembly of these adhesion contacts at the rear of the cell (Figure 1.4). Integrins

are the major cell surface receptors mediating the attachment of macrophages to the


underlying matrixle. Proteases, namely MMP14, are delivered to the leading edge to

regulate the turnover of these adhesion sites [Muller, 2003]. Both, integrins and MMPs

are localised to actin‐rich structures at the leading edge of macrophages, which are called

podosomes, and spontaneously form upon matrix contact [Cornfine et al., 2010; Gawden‐

Bone et al., 2010; Linder, 2007; Wiesner et al., 2010]. Ultimately, monocytes cross the

endothelium in a process called diapedesis. The emigration of leukocytes from the blood

stream involves the loosening of the endothelium, followed by the degradation of the

basement membrane, which is partly achieved by the catalytic activity of MMPs, mainly

MMP2 and 9 (Figure 1.4) [Le et al., 2007]. Once in the interstitial tissue, leukocytes

migrate through the 3D ECM towards the wound site. Lymphocytes and neutrophils have

been known to use the amoeboid migration mode, which is characterised by a rounded

to ellipsoid cell shape, as these leukocytes can squeeze through pores of the ECM [Friedl

and Weigelin, 2008]. However, macrophages are much larger than neutrophils and T

cells and manoeuvre through the tissue using the protease‐dependent mesenchymal

migration mode to create a path through the dense matrix material [Friedl and Weigelin,

2008]. Analogous to 2D migrating cells, macrophages form podosomes at the leading cell

protrusion in 3D environments. These podosomes are the site for MMP secretion and

help establish the migration path by localised matrix degradation (Figure 1.4) [Bjorklund

and Koivunen, 2005; Verollet et al., 2011]. Therefore, targeting the activity or secretion

of proteases, such as the MMPs, and MMP14 in particular, could regulate tissue

infiltration by macrophages. Inhibiting MMP activity has had limited success so far,

suggesting that reducing surface delivery might be an alternative strategy (section 1.3.4).


Figure 1.4. Macrophage tissue infiltration occurs through different migration processes and requires the activity of matrix metalloproteinases. Once attached to the endothelium, macrophages polarise to exhibit a leading edge where MMPs secreted at podosomes facilitate adhesion complex-turnover while detachment takes place at the trailing edge. During diapedesis, macrophages use MMPs to loosen the endothelium and degrade the basement membrane. To migrate through the dense interstitial tissue, macrophages form 3D podosomes where they secrete MMPs to create a path by matrix degradation.

Reducing molecules in the wound that exacerbate inflammation

As increased levels of inflammation has a deleterious effect on wound healing,

targeting pro‐inflammatory cytokines has been considered a promising therapeutic

approach to speed up the healing process and improve overall outcomes. Progress has

been made in targeting the potent pro‐inflammatory cytokine TNFα. Deficiency of the

inflammation‐mediating TNF receptor has been shown to reduce the number of wound

macrophages and accelerate repair, but is not translatable as it involved the use of

transgenic mice [Mori et al., 2002]. A TNF antibody (Infleximab) has been successful in

clinical studies to reduce inflammation and improve healing outcomes of chronic

wounds [Streit et al., 2006]. When applying antibodies against the pro‐inflammatory

cytokines IL‐1β and IL‐17, wound healing is enhanced in normal mice as well as in mice

with compromised healing [Rodero et al., 2013; Mirza et al., 2013]. Moreover, the

inhibition of the IL‐1β pathway in wounds of diabetic mice, using a neutralising antibody,

improved healing of these wounds [Mirza et al., 2013]. The treatment induced the shift

from the pro‐inflammatory M1 phenotype towards the M2 type, which is associated with

repair [Mirza et al., 2013]. This also confirms that macrophage maturation is critical for

proper wound healing and can be polarised.

Some of the chronic inflammation observed in non‐healing wounds can be

attributed to the persistent and elevated levels of certain MMPs in the wound. The

unrestrained activity of these MMPs causes extensive tissue damage by breaking down

too much of the ECM. The damaged ECM subsequently activates macrophages to secrete

more pro‐inflammatory cytokines and MMPs, which leads to a vicious cycle of

inflammation. Hence, reducing MMP levels has long been a focus in the development of

therapeutics to improve wound healing outcomes but has had limited success so far

(section 1.3.4).


1.3 MMPs during wound healing

1.3.1 The role of the extracellular matrix during wound healing

The ECM is the acellular component that comprises the dermal tissue. It is a gel‐

like matrix that is made up of fibrous structural proteins, such as collagens and

proteoglycans, produced by the surrounding cells [Rozario and DeSimone, 2010]. The

combination of these two components allows the skin to have both elasticity and

compressibility. The ECM acts not only as a scaffold for the cells, but also regulates cell

adhesion, chemotaxis and migration, as well as storing growth factors and cytokines

[Rozario and DeSimone, 2010; Schultz and Wysocki, 2009]. The production of ECM

proteins is also crucial to the wound healing process. During haemostasis, ECM proteins,

such as fibrin, fibronectin and collagen, build a clot that stops the bleeding. It also acts as

provisional matrix for cells infiltrating the wounded area, and contains factors to

stimulate fibroblasts, epithelial cells and endothelial cells [Rozario and DeSimone, 2010;

Schultz et al., 2011; Schultz and Wysocki, 2009]. Immune cells, fibroblasts and

endothelial cells release enzymes, including MMPs, which can degrade this provisional

matrix in order to facilitate cellular recruitment into the wound space and promote

angiogenesis and remodelling of the tissue. Breakdown of the ECM, release of growth

factors and activation of cytokines by proteases, such as MMPs, during tissue repair has

to be appropriately regulated to allow efficient healing [Schultz and Wysocki, 2009].

While unrestrained synthesis of ECM and reduced degradation of the same can lead to

increased scarring, uncontrolled degradation of the ECM through the activity of MMPs

such as MMP9 and MMP14 leads to excessive tissue damage and delays the healing

process [Schultz et al., 2011; Xue et al., 2006]. It is therefore important to understand

how MMP activity is (dys)regulated during the wound healing process.

1.3.2 MMPs degrade ECM and non-ECM components

In their active form MMPs can cleave both ECM proteins, such as collagen, and

certain non‐ECM proteins found within the tissue [Martins et al., 2013]. The proteolytic

cleavage of ECM proteins by MMPs is very important during the wound healing process

on multiple levels. During the inflammatory phase, the secretion of MMPs enables

immune cells to infiltrate the wounded area, where these cells can release more MMPs

to remove broken ECM. As broken collagen is unable to interact with newly produced

fibres, MMP activity debrides the wounded area clearing it of damaged ECM to allow for

an organised and strong new matrix [Schultz et al., 2011]. MMPs also facilitate the

migration of various other cell types through the tissue by breaking down the ECM thus


creating a path through the dense material [Quaranta, 2000]. The degradation of

different ECM components also alters their ratio within the wound environment, which

in turn affects cell adhesion and migration [Vu and Werb, 2000]. Throughout the

proliferative phase fibroblasts, endothelial cells and keratinocytes also utilise MMP

activity to migrate to the wound site where they regulate the formation of granulation

tissue, and re‐epithelialisation [Quaranta, 2000]. Furthermore MMPs can change the

availability or function of signalling molecules like receptors, cytokines and chemokines

[Vu and Werb, 2000]. The latter may occur through mechanisms, such as the release of

these molecules out of the ECM network making them biologically available, reducing

their levels through degradation or transformation of inactive precursors [Stamenkovic,

2003]. MMPs assist in angiogenesis by degrading the basement membrane of existing

blood vessels in order to allow the sprouting of new capillaries [Rundhaug, 2005].

Remodelling of the provisional matrix is also dependent on the action of MMPs and they

aid wound contraction as the scar tissue forms [Schultz et al., 2003; Stechmiller et al.,

2010]. Together with the fact that active MMPs can also induce the conversion of

zymogen MMPs to their active state, MMPs can have very diverse effects on cell

behaviour and the microenvironment throughout the wound healing process [Loffek et

al., 2011].

1.3.3 Structure, substrates and regulation of MMPs

MMPs form a family of Zn2+‐ and Ca2+‐dependent endopeptidases produced as

inactive precursor proteins, called zymogens, which share similar mechanisms of

activation [Vu and Werb, 2006]. Activated MMPs are able to cleave many ECM

components including collagen, elastin, fibronectin and laminin [Martins et al., 2013].

This is why MMPs play an important role in tissue remodelling. They can also cleave non‐

ECM proteins to affect growth‐factor activation and cell‐cell and cell‐matrix signalling

[Martins et al., 2013] (Table 1.1). Due to their proteolytic activity they are fundamental

for many physiological processes like immune cell infiltration during wound repair and

in the regulation of inflammation [Le et al., 2007; Loffek et al., 2011; Stamenkovic, 2003].

16

Chapter 1: Literature review

Table 1.1. Classification of MMPs according to their substrate specificity (from Rohl and Murray, 2013). Agg, aggrecan; Dec, decorin; EL, elastin; Fib, fibrillin; FBN, fibrin; FN, fibronectin; Gel, gelatin; IGFBPs, insulin-like growth factor binding protein; LN, laminin; PG, proteoglycan-linked protein; TGF, transforming growth factor; TNFa, tumor necrosis factor alpha; VN, vitronectin.

Class Common Name(s) MMP# ECM Substrates Non-ECM Substrates (examples) Collagenases Collagenase 1, Interstitial,

Fibroblast, Tissue C MMP1 Col I, II, III, VI, VII, X, Agg, FN, Gel, LN, PG, VN IGFBPs, CXCL12, pro-TNFα

Collagenase 2, Neutrophil C MMP8 Col I, II, III, V, VII, VIII, Agg, EL, FN, Gel, LN CXCL5 Collagenase 3 MMP13 Col I, II, III, IV, VII, IX, X, XIV, Agg, FN, Gel, Fib CXCL12

Gelatinases Gelatinase A, 72-kDa gelatinase

MMP2 Col I, III, IV, V, VII, X, XI, XIV, Agg, EL, FN, Gel, LN, PG, VN, Dec, Fib,

CCL7, CXCL12, pro-TNFα, IGFBPs, pro-TGFβ, pro-IL-1β

Gelatinase B, 92-kDa gelatinase

MMP9 Col IV, V, VII, X, XIV, Agg, EL, Gel, LN, PG, VN, Dec, Fib, FBN,

pro-TNFα, pro-IL-1β, IL-8, IL-2Rα, CXCL7, CXCL8, CXCL1, CXCL12

Stromelysins Stromelysin 1, Transin 1 MMP3 Col III, IV, IX, X, XI, Agg, EL, FN, Gel, LN, PG, VN, Dec, Fib

CXCL12, E-cadherin, pro-TGFβ1, pro-TNFα, IGFBPs, pro-IL-1β

Stromelysin 2, Transin 2 MMP10 Col III, IV, V, IX, X , Agg, EL, FN, Gel, LN, PG Stromelysin 3 MMP11 IGFBPs

Matrilysins Matrylisin, Matrin, PUMP-1 MMP7 Col IV, X, Agg, EL, FN, Gel, LN, PG, VN, Dec, Fib pro-TNFα, E-cadherin, β4-Integrin Matrylisin-2, Endometase MMP26 Col IV, FN, Gel

Membrane-type MT1-MMP MMP14 Col I, II, III, Agg, FN, gel CXCL12, CD44, pro-TNFα/TGFβ MT2-MMP MMP15 Agg, FN, Gel, LN, FBN MT3-MMP MMP16 Col III, FN, FBN, VN, LN, Gel MT4-MMP MMP17 Fib, FBN Pro-TNFα MT5-MMP MMP24 Gel MT6-MMP, Leukolysin MMP25 Col IV, Gel, FN

Other Enamelysin MMP20 Agg Epilysin MMP28 Metalloelastase MMP12 Col I, IV, Agg, EL, FN, Gel, LN, PG, LN, Fib Latent TNF RASI 1 MMP19 Col IV, I, FN, Gel, LN Xenopus MMP MMP21 MMP23b/CA-MMP MMP23 Gel MMP27


There are 23 human MMP members that show similarities in their domain

organisation (Figure 1.5). All MMPs, but one (MMP23), have an N‐terminal signal (pre‐

domain), which allows the protein to traffic to the endoplasmic reticulum (ER) and

subsequently also the Golgi apparatus. This allows the protein to be either secreted or

incorporated into the plasma membrane. Those MMPs that are not destined for secretion

have a either a transmembrane domain, a glycosylphosphatidylinositol (GPI) anchor or

a signal anchor that links them to cell‐membranes [Egeblad and Werb, 2002]. No matter

if they are secreted or reside as transmembrane proteins, all MMPs contain a catalytic

domain with three conserved histidines (HEXGHXXGXXH) coordinating a Zn2+‐ion

involved in the hydrolytic mechanism [Martins et al., 2013]. In their zymogen form, a

pro‐domain of about 80 amino acids containing a free thiol group (PRCKKPD) that can

further complex the Zn2+‐ion blocks via a linker region the entry of the catalytically

important water molecule into the active site [Martins et al., 2013] (Figure 1.6). To

convert the zymogen form into the active protein, this interaction has to be disturbed

and occurs through a ‘cysteine switch’ (Figure 1.6). In vivo, this can be achieved by

proteolytic cleavage of the pro‐domain through the action of serine proteases or other

MMPs [Bjorklund and Koivunen, 205; Loffek et al., 2011; Martins et al., 2013; Vu and

Werb, 2000]. Additionally, zymogen activation can occur through conformational

perturbation induced by chaotropic agents, such as SDS, or ROS, which chemically

modify the cysteine within the pro‐domain, in vitro [Bjorklund and Koivunen, 2005;

Loffek et al., 2011; Martins et al., 2013; Vu and Werb, 2000]. Cleavage‐mediated zymogen

activation usually occurs after secretion of the pro‐MMP, as is the case for MMP9

[Stamenkovic, 2003]. However, some MMPs have a furin‐cleavage site within the pro‐

domain, such as MMP14, which allows for them to be activated intracellularly before

their delivery to the cell surface [Stamenkovic, 2003]. Once activated, substrate

specificity for the individual MMPs is partly determined by small variances within the

catalytic domain [Bjorklund and Koivunen]. Additional coordination of the substrate is

accomplished through interactions mediated by a hemopexin‐like domain, which can be

found in most MMPs [Bjorklund and Koivunen, 2005] (Figure 1.5).


Figure 1.5. Classification of MMPs according to their domain structure (from Rohl and Murray, 2013). Pro, pro-domain; Pre, N-terminal signal sequence, Zn2+, zinc; Fi, a gelatin binding domain that resembles collagen-binding repeats in fibronectin; Fu, furin; TM, transmembrane domain; Vn, vitronectin-like; SA, signal anchor; GPI, glycosylphophatidylinositol anchoring sequence; CA, cysteine array; Ig, immunoglobulin.


Figure 1.6. Mechanisms of pro-MMP activation (from Rohl and Murray, 2013). All MMPs contain a catalytic domain with three conserved histidines coordinating a Zn2+-ion involved in the hydrolytic mechanism. In their zymogen form, a pro-domain of about 80 amino acids containing a free thiol group that can further complex the Zn2+-ion blocks via a linker region the entry of the catalytically important water molecule into the active site. To convert the zymogen form into active protein this interaction has to be disturbed through a ‘cysteine switch’. In vivo, this can be achieved by proteolytic cleavage of the pro-domain through the action of serine proteases or other MMPs. Additionally, zymogen activation can occur through conformational perturbation induced by chaotropic agents, such as SDS, or ROS, which chemically modify the cysteine within the pro-domain, in vitro.

Once activated, MMPs are highly efficient in the degradation of ECM material. To

prevent excess tissue damage, it is important that MMP activity is controlled on multiple

levels such as gene expression, cell surface delivery, localisation, zymogen activation and

through specific endogenous inhibitors [Bjorklund and Koivunen, 2005]. Generally, the

expression of MMPs, such as MMP9, is absent or minimal in normal tissues, but is found

to be greatly enhanced during inflammation and wound healing [Opendakker et al.,

2001]. Within the wound environment, ECM contact and cytokines can induce

expression of MMP genes in keratinocytes, fibroblasts, endothelial cells and immune

cells. In response to stimuli, the levels of MMPs produced by these cells can be further

controlled through epigenetic modifications or mRNA (de)stabilisation [Martins et al.,

2013]. After MMP synthesis, the enzymes need to be delivered to the cell surface in order

to access their substrates, like the ECM components or molecules housed within. This

occurs either through the secretion of soluble MMPs, as is the case for MMP9, or


incorporation of transmembrane MMPs, such as MMP14, into the plasma membrane.

Intracellular trafficking machinery, such as specific SNARE and Rab proteins, regulate

these discrete trafficking processes [Murray and Stow, 2014]. This also allows the

polarised delivery of MMPs and limits their activity to areas of the cell where it is

required, including structures necessary for migration and invasion through the ECM.

MMP9 and MMP14 are delivered to membrane protrusions, such as filopodia, actin‐rich

rod‐like structures at the leading edge of migrating cells. Particularly in macrophages,

they are also found at podosomes, which are actin‐rich structures that give the cells the

ability to invade dense material [Gialeli et al., 2011].

Association of membrane‐anchored MMPs like MMP14 to the plasma membrane

restricts their activity to the immediate pericellular space. Soluble MMPs are also known

to bind back to the cell surface via adhesion molecules or proteoglycans to focalise their

enzymatic activity. MMP9, for example, binds to CD44, as well as the integrins α4β1,

α5β1 and αvβ5 [Deryugina and Quigley, 2011]. Unless conversion of the zymogen form

to the active enzyme has occurred intracellularly, as is the case for MMP14, activation of

MMPs without a furin‐cleavage site, such as MMP9, occurs after secretion [Stamenkovic,

2003]. Interestingly, activation of MMP2 occurs through cell surface docking of the pro‐

enzyme to TIMP2 and together they bind to the catalytic domain of MMP14, which

results in its activation [Deryugina and Quigley, 2011]. Once activated this complex is

responsible for the conversion of MMP9 to its active form [Itoh et al., 2001; Itoh and Seiki,

2004; Toth, 2003; Vandooren et al., 2013]. Although additional MMPs, as well as other

proteases, have been implicated in the activation of MMP9 this docking mechanism,

involving the MMP14/MMP2/TIMP complex, could increase efficiency of zymogen

conversion and help concentrating MMP activity to the pericellular space.

Once activated, MMP activity can be restricted through the presence of non‐

specific endogenous inhibitors, such as α2‐macroglobulin, or specific ones, which are the

tissue inhibitors of metalloproteinases (TIMPs). Four TIMPs (TIMP1, TIMP2, TIMP3,

TIMP4) are known and they are cysteine‐rich proteins of approximately 21‐28 kDa in

size that reversibly associate to MMPs in a 1:1 ratio [Bjorklund and Koivunen, 2005;

Stamenkovic, 2003]. The levels and ratios of MMPs and TIMPs within a wound has a

strong influence on the wound healing outcome and alterations in this balance can

influence wound chronicity (section 1.3.3). Interestingly, MMP9 levels can also be

modulated through an endocytic clearance mechanism. Binding of MMP9 to low density

lipoprotein receptor‐related protein 1 (LRP‐1) triggers the uptake of the complex, which


ultimately leads to its degradation within the lysosome, thereby reducing extracellular

MMP9 [Van den Steen et al., 2006].

1.3.4 Pathological roles of MMPs during wound healing

Through their activity of cleaving ECM and non‐ECM proteins, MMPs play many

important roles during different phases of wound healing. However, when dysregulated,

MMPs can cause excessive damage to the provisional matrix, inhibiting re‐

epithelialisation in chronic wounds (Figure 1.7) [Beidler et al., 2008; Ladwig et al., 2002;

Lobmann et al., 2002; Mwaura et al., 2006; Pirila et al., 2007]. These non‐healing wounds

are understood to be locked in a vicious cycle of inflammation with elevated levels of

neutrophils and macrophages and unable to progress to the granulation phase of healing

[Gethin, 2012; Raffetto, 2013]. Wounds with little or no macrophages on the other hand

heal faster, with reduced scarring [Martin and Leibovich, 2005]. It is believed that

elevated MMP levels, as well as other pro‐inflammatory molecules released from

macrophages and neutrophils, contribute to the persistent inflammation seen in these

wounds [Le et al., 2007]. An imbalance of particular MMPs and TIMPs, with increased

MMP1, 2, 8 and 9 expression and decreased TIMP1 and 2 levels, have been found in

chronic ulcers and are considered accountable for unrestrained ECM breakdown,

inflammation, altered cell growth and increased immune cell infiltration in wound tissue

[Loffek et al., 2011; Martins et al., 2013] (Table 1.2; Figure 1.7). Proteolytic cleavage of

the ECM by MMPs creates damage‐associated molecular patterns that cause activation

of macrophages to secrete pro‐inflammatory cytokines as well as more MMPs further

perpetuating the inflammatory state of non‐healing wounds [Olczyk et al., 2014].


Figure 1.7. Increased MMP levels in wounds contribute to poor wound healing outcomes. MMPs in chronic wounds cause uncontrolled matrix degradation, depletion of growth factors, and activation of pro-inflammatory cytokines, which prolongs inflammation and prevents re-epithelialisation by keratinocytes. Macrophages are major secretors of MMPs during the inflammatory phase of wound healing.

Table 1.2. Alterations in MMP and TIMP levels in chronic venous, diabetic, pressure and mixed ulcers (from Rohl and Murray, 2013).

Elevated levels Reduced levels Reference

Chronic venous leg ulcers

MMP1, MMP2, MMP3, MMP8, MMP9, MMP12,

MMP13 TIMP1, TIMP2

[Beidler et al., 2008; Mwaura et al., 2006]

Diabetic ulcers

MMP1, MMP2, MMP8, MMP9 TIMP2 [Lobmann et

al., 2002] Pressure ulcers

MMP9 TIMP1 [Ladwig et al., 2002]

Ulcers (mixed group)

MMP8, MMP26 [Pirila et al.,

2007]

In particular, macrophage‐derived MMP9 has been considered the main protease

responsible for unwanted matrix degradation as its levels correlate with chronic wound

severity [Rayment et al., 2008; Tarlton et al., 1999]. It has been hypothesised that the

negative impact of MMP9 levels on wound healing is due to the fact that MMP9 is highly

efficient at degrading collagen [Mackay et al., 1990]. Additionally, MMP9’s substrates

include non‐ECM molecules, such as cytokines and chemokines, which could have

immunomodulatory potential and further perpetuate the inflammation and immune cell


recruitment [Marom et al., 2007; Okamura et al., 2001]. Hence, MMP9 levels, also in

relation to its inhibitor TIMP1, within wound fluid are considered to be an indicator of

the competence of wound closure [Ladwig et al., 2002]. This is confirmed through studies

showing that adding high levels of MMP9 to acute wounds actually delays the wound

healing process [Reiss et al., 2010]. This means that although MMP9 expression is

temporally upregulated to assist in the wound healing process, persistent and elevated

levels of this enzyme can delay healing [Reiss et al., 2010].

1.3.5 Strategies to reduce levels of matrix metalloproteinases in wounds

As MMPs play a significant role in inflammatory diseases, like non‐healing

wounds, they have been regarded important therapeutic targets. One approach to

reducing the levels of MMPs in the wound environment is the use of highly absorbent

foam dressings [Eming et al., 2008; Lobmann et al., 2006]. These dressings can soak up

the wound fluid and thereby draw away detrimental MMPs, such as MMP9, as well as

other pro‐inflammatory factors but also beneficial factors that could be present. They

also have no influence on membrane anchored MMPs, such as MMP14. Their overall

clinical effect on the microenvironment of non‐healing wounds remains to be tested. The

application of pressure has been shown to reduce the levels of MMP3, MMP8 and MMP9,

in chronic wound fluid either through compression bandages or by topical negative

pressure therapy [Beidler et al., 2008; Moues et al., 2008]. However, the reduction of

MMP3, MMP8 and MMP9 was not correlated with overall wound healing outcomes.

Excessive MMP activity in chronic wounds has also been targeted through the use

of broad‐spectrum MMP inhibitors like GM6001 (Galardin). This small molecule

inhibitor works by chelating the zinc ion necessary for the catalytic activity of MMPs.

Unfortunately, this inhibitor was found to negatively affect keratinocyte migration,

myofibroblast differentiation and wound contraction [Barletta et al., 1996; Miratschijski

et al., 2004; Schultz et al., 1992]. Being a broad spectrum inhibitor, the limited efficacy of

this inhibitor on the overall wound healing outcome is owed to the fact that not all MMPs

are detrimental to the wound healing process. Some MMPs, like MMP8, have important

and beneficial roles during wound healing [Gooyit et al., 2014; Martins et al., 2013].

Fibroblasts, endothelial cells as well as keratinocytes use MMPs for their cell migration

and as a result generally inhibiting MMPs will negatively affect important wound healing

processes such as re‐epithelialisation and angiogenesis. Similar observations have been

made when applying TIMPs onto chronic wounds [Miyoshi et al., 2005]. As such, broad‐

spectrum inhibitors should be avoided and individual members of the MMP family that

are thought to play a harmful role in non‐healing wounds should be targeted instead [Xue


et al., 2006]. Interestingly, doxycycline, which chelates the zinc or calcium ions in the

catalytic domain of MMPs, appears to be an exception to this. When doxycycline is used

at sub‐antimicrobial doses, improvement in wound healing for patients with diabetic or

venous leg ulcers can be observed [Chin et al., 2003; Serra et al., 2013].

The design of specific inhibitors has proven difficult as the structures of the

catalytic domain is similar amongst the individual members of the MMP family [Loffek et

al., 2011]. Nonetheless, some more selective MMP inhibitors have been developed but

have not yet been tested on non‐healing wounds [Dahl et al., 2012; Devy et al., 2005; Sina

et al., 2009]. Functional blocking antibodies, such as one targeting MMP14, also have the

potential to improve chronic wounds but this has also not been tested yet [Devy et al.,

2009]. It is important to consider that treatments involving antibodies are very costly.

Thus, there are currently no viable options available to inhibit MMP‐mediated tissue

damage in chronic wounds. Therefore, identifying alternative targets to regulate MMP

activity might lead to the development of better and more cost‐effective targeted

treatments. This could be achieved by understanding the specific trafficking mechanisms

involved in MMP cell surface delivery or secretion.


1.4 Intracellular trafficking of proteins

To establish a border to the environment, all cells are surrounded by a

membrane. In addition to this plasma membrane, eukaryotic cells are

compartmentalised as they contain multiple membrane‐bound organelles with different

shape, specialised function, and composition. This allows larger cells to be more efficient

but requires a refined system that enables exchange between these organelles [Tokarev

et al., 2009]. Between organelles, soluble and membrane‐bound cargo, such as proteins

as well as lipids, are transported in membrane‐bound carriers called vesicles, which are

often of tubular or pleiomorphic character [Lacy, 2015]. Directional transport between

compartments requires the fusion of two separate lipid bilayers [Martens and McMahon,

2008].

Retrograde or ‘inwards’ transport is called endocytosis (section 1.4.1) and allows

the uptake of nutrients and pathogens but also regulates cell signalling. By budding from

the plasma membrane and subsequent inwards transport cell surface receptors with or

without ligands can be internalised and fluids or particles be brought into the cell

[Doherty and McMahon, 2009]. Anterograde or ‘outwards’ delivery of cargo to the cell

surface on the other hand is achieved by exocytosis or secretion (section 1.4.2).

Ultimately, vesicles fuse with the plasma membrane to release soluble cargo and

incorporate membrane‐associated molecules into the cell membrane. This allows cell

communication and also for cells to modulate the environment and their contact with

other cells or the ECM [Stow et al., 2006]. Both pathways are connected by bi‐directional

transport (section 1.4.3) and requires sophisticated machinery (section 1.4.4). Vesicle

transport from one organelle or membrane to another involves four major steps

[Stenmark, 2009]. Initially, a vesicle has to bud off from a donor membrane and this

budding is facilitated by specific cytosolic coat proteins like COPI, COPII, or clathrin and

their adaptors. The membrane‐bound cargo is then transported to the target membrane

with the help of motor proteins that can move vesicles along microtubules and actin

filaments. Tethering factors, such as the Rab proteins (see section 1.4.4.1), aid docking of

the vesicle at the target membrane while other trafficking machinery, in particular

SNARE proteins (see section 1.4.4.2), in conjunction with other regulators, ultimately

enable membrane fusion [Stow et al., 2006]. Specific machinery and pathways have been

identified in macrophages and other cell types and can be used to manipulate

intracellular trafficking pathways of various effector molecules (see section 1.4.4.3).

However, the pathways responsible for the cell surface delivery of important MMPs in


macrophages remain unclear (see section 1.4.4.4) and investigating these might allow

the development of novel therapeutics.

1.4.1 Endocytic pathways

Endocytosis describes the retrograde transport of vesicles, which allows cargo to

be internalised from the cell membrane and environment [Doherty and McMahon, 2009].

Phagocytosis, the engulfment of large particles, is one form of endocytosis that is limited

to professional phagocytes such as macrophages and is a crucial mechanism for immune

function and protection from pathogens [Greenberg and Grinstein, 2002]. Non‐specific

bulk‐phase uptake of extracellular fluid as well as soluble macromolecules is

accomplished by all cells when the plasma membrane ruffles fold back to internalise

materials and is called micropinocytosis [Lim and Gleeson, 2011]. The actin cytoskeleton

plays a major role for both, phagocytosis and macropinocytosis. Other endocytic

mechanisms include clathrin‐mediated, caveolae‐mediated and alternative clathrin‐ and

caveolae‐independent pathways. [Lopez and Arias, 2010]. These endocytic pathways are

often receptor‐mediated to internalise specific extracellular macromolecules upon

binding to their respective cell surface receptors, e.g. transferrin (Tfn) and transferrin

receptor (TfR) or low density lipoprotein (LDL) and its receptor [Van der Goot and

Gruenberg, 2006].

Following the endocytosis from the cell surface, membrane‐bound cargo is

delivered to the sorting or early endosome. This membrane‐bound compartment is

mildly acidic with a pH of about 6.2, which sometimes induces the release of ligands from

their receptors (e.g. LDL) while other ligand‐receptor complexes remain associated (e.g.

Tfn or TLR4 and LPS) [Luzio et al., 2005]. Detachment of ligand and receptor allows both

of them to be trafficked separately in case the receptor is recycled back to the cell surface

either directly or via other compartments [Van der Goot and Gruenberg, 2006]. The

tether molecule Rab5 as well as the early endosomal antigen‐1 (EEA1) surround the

early endosome and play important roles in endocytic trafficking.

From the early endosome, vesicles can be trafficked to the late endosomes. The

late endosome compartment is characterised by a growing number of intraluminal

vesicles as well as an increasingly acidic pH (~5.5‐5.0) [Luzio et al., 2007; Van der Goot

and Gruenberg, 2006]. These intraluminal membranes are enriched in

lysobisphosphatidic acid (LBPA), while the limiting membrane contains lysosomal‐

associated membrane protein 1 (LAMP1). The presence of tethering factor Rab5 in these

organelles is lost while Rab7 is recruited instead.


Next in the endocytic pathway are lysosomes, which are the most acidic of these

organelles (pH ~4.6‐5.0) and hold acidic hydrolases, such as Cathepsin D. Within the

endocytic pathway, it is often the terminal degradation compartment. While lysosomes

are also positive for LAMP1, they do not contain LBPA found in late endosomes [Luzio et

al., 2007]. Transfer of endocytosed material from late endosomes to lysosomes involves

heterotypic fusion of both organelles to form a transient hybrid organelle. Late

endosomes and lysosomes re‐establish by selective retrieval of late endosome

components [Luzio et al., 2007]. The reformation of late endosomes appears to be

facilitated, in part, by LBPA [Van der Goot and Gruenberg, 2006]. Both, late endosomes

and lysosomes often localise in clusters in the perinuclear region near the microtubule‐

organising centre [Luzio et al., 2007; Van der Goot and Gruenberg, 2006].

1.4.2 Exocytic/biosynthetic pathways

Exocytosis describes the anterograde transport of vesicles towards the cell

surface, which ultimately allows soluble cargo to be secreted into the environment and

membrane‐associated molecules, such as transmembrane proteins and lipids, to be

incorporated into the plasma membrane [Lippincott‐Schwartz et al., 1998]. Exocytosis

allows reestablishment of the plasma membrane with the appropriate lipids and cell‐

surface proteins following endocytic events, expulsion of waste products but also

enables transport and discharge of newly synthesised material into the environment to

fulfil cell functions, such as the release of cytokines by macrophages to exert immune

properties [Stow et al., 2006]. Classically, newly made proteins destined to be secreted

are usually trafficked from the ER to the Golgi complex and then to the cell surface.

However, it has to be noted that proteins, such as IL‐1, exist that are secreted by non‐

classical pathways independent of ER or Golgi complex [Stanley and Lacy, 2010]. Also,

not all proteins are produced for the release into the environment but remain for use

inside of the cell.

One third of all the newly synthesised proteins – secreted proteins,

transmembrane proteins and proteins that will remain within intracellular organelles –

are translocated into the ER [Ghaemmaghami et al., 2003]. The translocation process is

initiated at the ribosome as detection of an N‐terminal signal peptide by the signal

recognition particle halts translation [Rapoport, 1992]. The ribosome then attaches to

the ER and translation of the emerging peptide chain through a translocon pore of the

ER membrane is initiated. Once translation is finalised, signal peptidases will remove the

signal sequence from the new protein [Rapoport, 1992]. The entry into the ER also marks


the start of the co‐ and post‐translational modification of the new protein. If the relevant

consensus sequence (NXS/T) is present, N‐linked glycans are added en bloc to asparagine

residues within the protein by oligosaccharyltransferase [Rapoport, 1992]. The

oxidative environment inside the ER allows for the formation of disulphide bonds with

the help of the oxidoreductase protein disulphide isomerase (PDI) [Kleizen and

Braakman, 2004]. Chaperones aid in the folding process and quality control mechanisms

ensure that only properly folded proteins leave the ER [Tokarev et al., 2009].

Correctly folded proteins leave the ER in vesicles at one of 100‐200 ER exit sites

and are transported to the ER‐Golgi‐intermediate compartment (ERGIC) by COPII

vesicles [Appenzeller‐Herzog and Hauri, 2006]. This process is regulated by activation of

the Sar1 GTPase by ER‐associated guanine nucleotide exchange factor (GEF) sev12,

guanine nucleotide activating protein GAP Sec23/24 and its regulator Sec13/31 complex

[Beraud‐Dufour and Balch, 2002; De Matteis and Luini, 2008]. ER‐resident proteins,

recognised by a KDEL motif, as well as misfolded proteins are shuttled back from the

Golgi complex to the ER. Vesicles with cargo destined for retrograde transport to the ER

are assembled by COPI coat proteins, a process that is regulated by ARF1 GTPase [De

Matteis and Luini, 2008].

The Golgi complex is made up of four to eight tightly stacked disk‐like cisternae

and takes up a peri‐centrosomal position within the cell. Proteins move from the

cisternae nearest to the ER (cis‐Golgi) towards the opposite site (trans‐Golgi). While in

the Golgi cisternae, proteins may be modified by O‐linked glycosylation to serine and

threonine side chains without any known consensus sequences required for this process

[De Matteis and Luini, 2008]. The trans‐Golgi network (TGN) is a sorting station that

allows the exit from the trans‐Golgi cisternae where cargo is organised into different

vesicles to be transported to various endosomes [De Matteis and Luini, 2008]. The TGN

is also the place for final modifications of the newly synthesised protein, such as

sialylation and fucosylation of N‑linked and O‑linked glycans and sulphation of

glycosaminoglycans, as well as conversion of zymogen to mature protein forms [De

Matteis and Luini, 2008]. As an example, processing of the inactive MMP14 pro‐form

occurs in the TGN and is a prerequisite for the delivery of mature MMP14 to the cell

surface as well as its function upon incorporation into the plasma membrane [Frittoli et

al., 2011]. The sorting of cargo at the TGN is influenced by multiple mechanisms;

recognition of tyrosine (NPXY or YXXO) or leucine ([DE]XXXL[LI] or DXXLL) consensus

sequences within the protein, post‐translational modifications (glycosylation,

ubiquitinylation, phosphorylation), affinity to micro‐domains of the TGN membrane and


sorting machinery proteins [De Matteis and Luini, 2008]. Sorted cargo is transported

from the TGN to other organelles or the plasma membrane along microtubules [Luini et

al., 2008]. Plus‐end‐directed movement of vesicles is mediated by kinesin motor proteins

that use ATP‐hydrolysis for a conformational change that translates into movement

along the microtubules [Caviston and Holzbaur, 2006; Soldati and Schliwa, 2006]. The

interaction between these motor proteins and cargo can be influenced by further

regulators like the Rabs and affect the destination of the vesicle [De Matteis and Luini,

2008].

1.4.3 Crosstalk between exocytic/biosynthetic and endocytic pathways

While endocytosis and exocytosis are concepts to describe retrograde versus

anterograde trafficking events, they do not actually occur independently of each other

but are connected by bi‐directional transport. Endosomes act as mediating

compartments that do not only facilitate degradation of internalised material but also

the recycling of cargo and even the delivery of newly synthesised proteins to the cell

surface [Stow et al., 2009]. For instance, following endocytosis of Tfn and TfR and after

iron bound to Tfn has been released within the cell, receptor and ligand are recycled back

to the cell surface via recycling endosomes for reuse [Maxfield and McGraw, 2004].

However, newly synthesised TfR also passes through the recycling endosome before

being transported to the plasma membrane as shown in epithelial cells [Futter et al.,

1995]. More recently, it was shown that the biosynthetic route of other proteins, such as

TNF and IL‐6, also occurs via the recycling endosome in macrophages [Manderson et al.,

2007; Murray et al., 2005]. The recycling endosome also participates in the retrograde

transport of cargo back to the TGN [Mallard et al., 1998; Tran et al., 2007]. To allow this

level of complexity of cargo sorting the recycling endosome is sub‐compartmentalised

into multiple domains, which are differently enriched in the respective trafficking

machinery.

A further example for the connection of endocytic and exocytic pathways are the

mechanisms that allow antigen presentation. The MHCII complex is assembled within

the ER and Golgi complex before being transported to late endosomes/lysosome [Chow

and Mellman, 2005; Geuze, 1998]. In the late endosomes/lysosomes the MHCII complex

binds peptides, which originate from internalised pathogens as the phagosome fused

with late endosomes/lysosome [van den Hoorn et al., 2011]. The peptide‐MHCII complex

is then shuttled to the cell surface where antigen presentation can occur [Bertho et al.,

2003; Boes et al., 2002; Chow et al., 2002; Kleijmeer et al., 2001; Vyas et al., 2007]. Thus


late endosomes/lysosomes are not exclusively for degradation of endocytosed material

but cargo can traffic through them en route to the cell surface.

In many haematopoietic cells, and some other cell types, such as melanocytes,

secretory lysosomes or lysosome‐related organelles enable storage of pre‐produced

proteins derived from the TGN and efficient release of the cargo from these

compartments in response to a stimulus [Blott and Griffiths, 2002; Dell’Angelica et al.,

2000]. This process is also called regulated exocytosis and is integral for important

effector functions of different immune cells. For example, upon T cell receptor stimulus

CTLs and NK cells release cytotoxic molecules, such as perforin and granzymes, to kill

infected or transformed cells [Griffiths, 2003]. Mast cells and basophils secrete large

amounts of histamines and other pro‐inflammatory molecules following Fc receptor

stimulus, while the same stimulus triggers neutrophils to release chemoattractants

[Blott and Griffiths, 2002; Clark and Griffiths, 2003; Stanley and Lacy, 2010]. Although

also of haematopoietic lineage, macrophages do not possess this type of granules or

secretory lysosomes but they are nonetheless able to export cargo to the cell surface via

lysosomes in a constitutive fashion [Stanley and Lacy, 2010]. Macrophages can

upregulate the required machinery and pathways in response to stimulation to increase

the rate at which important inflammatory mediators are released into the environment

[Murray et al., 2005; Pagan et al., 2003].

Proteins that are transported to the lysosomes for secretion or degradation may

be derived from biosynthetic or endocytic pathways and sorting is influenced by a

variety of signals. Proteins that are membrane‐associated often display a tyrosine‐based

signal within their cytoplasmic tail, while soluble proteins may receive a mannose‐6‐

phosphate moiety while in the Golgi complex, which binds to the mannose‐6‐phosphate

receptor (M6PR) and targets the protein to the late endosome and then lysosome. Other

cell‐specific pathways may exists to influence sorting of cargo to be delivered to these

compartments [Blott and Griffiths, 2002; Clark and Griffiths, 2003; Luzio et al., 2007].


1.4.4 Trafficking machinery

Vesicular transport requires budding, movement, docking and fusion of

membrane‐bound cargo at every step of the trafficking pathway [Lacy and Stow, 2011].

To aid with vesicular trafficking eukaryotic cells are equipped with a range of trafficking

machinery, which includes lipids and proteins. Initially, newly made proteins need to be

sorted into the appropriate carriers, while cytosolic coat and adaptor proteins facilitate

budding of the vesicle. Motor proteins drive the movement of the cargo to its target

compartment along the cytoskeleton. The family of Rho GTPases, which are

predominantly cytoplasmic, can regulate aspects of the intracellular actin network while

the family of membrane‐associated Rab GTPases control vesicle docking at the target

membrane [Stenmark, 2009]. Fusion of the vesicle and target membrane is dependent

on the action of a family of membrane‐associated proteins called SNAP (soluble NSF

attachment protein) receptors (SNAREs) [Stanley and Lacy, 2010].

1.4.4.1 Rabs

Rab GTPases influence budding, movement and docking of vesicles during

intracellular trafficking events [Stenmark, 2009]. Rab proteins are a family of small

GTPases of the Ras superfamily and there are 70 known members in mammals [Bhuin

and Roy, 2014]. Rabs switch between two conformational states as typical of GTPases.

When inactive, Rabs are bound to GDP and found in the cytosol. Guanine nucleotide

dissociation inhibitor (GDI) converts the Rab into its active, GTP‐bound state, which then

becomes membrane‐associated through two hydrophobic geranylgeranyl groups that

are attached to cysteine residues within the protein [Bhuin and Roy, 2014; Stenmark,

2009]. Individual members of the Rab family localise to distinct membranes within the

cell [Stenmark, 2009]. While in the active state, Rabs recruit and/or activate effector

molecules, such as sorting adaptors, kinases, phosphatases, motor proteins and other

tethering factors [Stenmark, 2009].

Sorting of cargo is initiated through recognition of the cargo by adaptor proteins

and assembly of coat proteins, such as clathrin or COPI and COPII, recruited from the

cytosol [Cai et al., 2007]. Rabs can assist in cargo‐specific vesicle sorting by recruiting

effector molecules to distinct intracellular membranes. For example, Rab9 recruits the

adaptor protein TIP47 to late endosome, which enables the recycling of M6PR by

transporting M6PR‐positive carriers back to TGN [Bhuin and Roy, 2014; Carroll et al.,

2001; Stenmark et al., 2009]. Rab5 participates in the assembly of clathrin‐coated pits at

the plasma membrane for the endocytosis of TfR [Bhuin and Roy, 2014; Liu and Storrie,


2012; McLauchlan et al., 1998; Stenmark, 2009]. Once a vesicle is established and budded

from the donor membrane, Rabs can also help to remove coat proteins. This coat

shedding is important as the presence of coat protein complexes inhibits membrane

fusion. Rab5 is known to coordinate disassembly of the adaptor protein complex AP2

following endocytosis [McLauchlan et al., 1998]. Rabs also influence the motility of

vesicles as they control interactions of the cargo with motor proteins that facilitate long‐

range transport along elements of the cytoskeleton. Rab27 connects melanosomes,

which are organelles that store melanin, e.g. in melanocytes, to the actin motor protein

myosin Va [Bhuin and Roy, 2014; Liu and Storrie, 2012; Stenmark, 2009].

Once in proximity to the target membrane, Rabs facilitate vesicle docking often

together with other tethering factors. Rab5 tethers endocytosed vesicles to the

membrane of early endosomes in cooperation with EEA1 [Bhuin and Roy, 2014; Nielsen

et al., 2000; Simonsen et al., 1998; Stenmark, 2009]. At the ERGIC, Rab1 first recruits the

tether molecule p115 during COPII vesicle budding at the ERGIC, and then also binds to

the Golgi matrix protein GM130 [Beraud‐Dufour and Balch, 2002; Stenmark, 2009].

Rabs specifically locate to particular membranes of different organelles; for

instance, Rabs 5, 7, and 11 associate with early, late, and recycling endosomes,

respectively, while Rab6 is found at the Golgi complex and Rab27a is known to regulate

granule release in certain cell types [Lacy and Stow, 2011; Liu and Storrie, 2012; Stow

and Murray, 2013; Stanley and Lacy, 2010]. If various Rab members associate with the

same organelle, they are often found spatially segregated and occupy distinct membrane

microdomains, which are sometimes referred to as Rab domains. These domains remain

dynamic but do not significantly intermix to allow for their function in cargo sorting

[Stenmark, 2009; Zerial and McBride, 2001].

While Rabs play many important roles for vesicular trafficking by influencing

vesicle formation, movement and docking, fusion is ultimately achieved through the

specific pairing of cognate SNARE proteins on the opposing membranes [Jahn and

Scheller, 2006]. Rabs have been suggested to further influence this process as Rab

effector proteins interact with SNARE machinery, SNARE priming machinery, and SNARE

proteins themselves to increase efficiency as well as specificity of cargo transport [Bhuin

and Roy, 2014; Zerial and McBride, 2001]. However, the exact mechanisms through

which Rabs might regulate membrane fusion are not yet known [Bhuin and Roy, 2014].


1.4.4.2 SNAREs

Ultimately, the delivery of membrane‐bound vesicles to the cell surface for

secretion or incorporation into the plasma membrane requires membrane fusion

between vesicles and organelles or the cell surface. A family of proteins known as the

SNAREs (soluble N‐ethylmaleimide‐sensitive factor attachment protein receptor)

regulate this process at multiple points along the trafficking pathway by providing the

energy and specificity required [Jahn and Scheller, 2006].

SNAREs are small proteins of around 100–300 amino acids in length. Each

member contains at least one evolutionally‐conserved SNARE motif of about 60‐70

residues, which participates in the formation of a helical SNARE complex important for

the membrane fusion mechanisms [Hong, 2005]. 36 members of the SNARE family exist

in humans to date and most of them possess a C‐terminal hydrophobic type II

transmembrane anchor, while some of them associate to the cytoplasmic side of the

membrane via prenylation, palmitoylation or interaction with another anchored SNARE

member [Hong, 2005; Malsam et al., 2008; Murray and Stow, 2014]. Members of the

SNARE family are in general ubiquitously expressed, with the exception of two neuronal

specific SNAREs, Stx1, and SNAP25, and the immune specific SNARE Stx11 [Murray and

Stow, 2014]. SNAREs are found on all organelle membranes, as well as the plasma

membrane, and, similar to the Rabs, they are characterised by a precise subcellular

distribution. The selective interaction of a SNARE on a donor membrane and two to three

cognates SNAREs on a target membrane forms a trans‐SNARE complex that drives the

fusion process (Figure 1.8) [Fasshauer, 2003].

Figure 1.8. SNARE‐mediated fusion (modified from Murray and Stow, 2014). The R‐SNARE (pink) together with the Q‐SNARE complex (aqua), pull vesicle and target membrane close together and potentially generate the force required for the fusion of the two lipid bilayers. After fusion the SNARE complex is located on the target membrane and called cis‐SNARE complex. It will then be disassembled.


In yeast, of about 300 possible combinations of SNARE proteins tested in an in

vitro liposome fusion assays, only 9 of them were found to be fusogenic [Malsam et al.,

2008]. This is because SNAREs are not all the same. According to the composition of their

SNARE motif, SNAREs can be classified as either R‐ or Q‐SNAREs. R‐SNAREs, such as the

VAMPs, carry an arginine (R) as central functional residue within SNARE motif and are

usually found on the vesicle membrane [Jahn and Scheller, 2006]. Q‐SNAREs, such as the

syntaxins (Stx), are characterised by a glutamine (Q) as the central functional residue of

the SNARE motif and are typically on the target membrane. These conserved SNARE

motifs facilitate the protein interactions in the trans‐SNARE complex, which forms a

highly stable coiled‐coil superstructure [Jahn and Scheller, 2006]. Depending on their

position in the trans‐complex, Q‐SNAREs can be further classified as Qa, Qb, Qc or Qbc

SNAREs (Table 1.3). All R‐SNAREs, as well as most Q‐SNAREs (Qa, Qb, Qc), have one

SNARE motif near their C‐terminal anchor to contribute to the trans‐SNARE complex.

However, three Q‐SNAREs SNAREs, namely SNAP23, SNAP25 and SNAP29, contain two

SNARE motifs (Qbc). Hence, the formation of functional trans‐complex with four SNARE

motifs is established either through the interaction of four (R‐Qa‐Qb‐Qc) or three (R‐Qa‐

Qbc) SNAREs [Stow et al., 2006].

Table 1.3. Q‐ and R‐SNARE family members.

R-SNAREs Q-SNAREs Qa Qb Qc Qbc

VAMP1, VAMP2, VAMP3, VAMP4, VAMP5, VAMP7, VAMP8,

ERS24, YKT6

STX1, STX2, STX3, STX4, STX5, STX7,

STX11, STX13, STX16, STX17,

STX18

GS27, GS28, Vti1a, Vti1b

STX6, STX8, STX10, GS15, BET1, SLT1

SNAP23, SNAP25, SNAP29, SNAP47

The formation of the four‐chain helix that forms the trans‐complex is mediated

by the hydrophobic side chain residues, which are positioned perpendicular to the

helical axis [Beraud‐Dufour and Balch, 2002; Hong, 2005; Stanley and Lacy, 2010]. The

zippering of this helix is initiated at the N‐terminus of the SNARE proteins and

progresses toward the C‐terminus, which is closer to the membrane. This conformational

change brings the two membranes in close proximity, deforms the membrane and

provides the energy required to overcome repulsion of the opposing membranes due to

the negative charges of phospholipid headgroups of the lipid bilayers [Hong, 2005]. This

allows the formation of an aqueous fusion pore through which cargo can be discharged

while transmembrane proteins and lipids of the donor membrane are incorporated into


the target membrane [Murray and Stow, 2014]. Alternatively, in a process called “kiss

and run”, only some of the cargo is released and the pore might subsequently close again

[Murray and Stow, 2014]. It is estimated that 3 to 15 SNARE complexes are required to

catalyse a single fusion event [Jahn and Scheller, 2006]. As a result of the fusion event, all

participating SNAREs are located on the same membrane, which is called a cis‐SNARE

complex. This cis‐complex has to be disassembled to allow further fusion events and

requires conformational rearrangement regulated by the NSF and α‐SNAP molecular

chaperones [Jahn and Scheller, 2006].

As SNARE members have a precise distribution within the cell (Figure 1.9), and

SNARE come together in specific combinations to form distinct SNARE complex, the

process of membrane fusion for vesicular trafficking is highly specific [Murray and Stow,

2014]. Additionally, plasma‐membrane SNAREs are known to concentrate in nano

domains rather than being randomly distributed within the membrane to allow

polarised release of cargo at certain areas of the cell surface [Jahn and Scheller, 2006].

Further spatial and temporal regulation of SNARE‐mediated fusion is achieved through

post‐translational modifications of SNARE proteins, as well as binding of regulatory

proteins [Stow et al., 2006]. Sec1/Munc18 and Rab proteins are regulatory machinery

that can influence assembly of the SNARE fusion‐complex [Beraud‐Dufour and Balch,

2002; Martens and McMahon, 2008; Stanley and Lacy, 2010].

As each type of organelle has distinct SNAREs sets that can only bind their

equivalent SNARE partner on the respective target membrane, the absence or

malfunction of just one of the participating SNAREs inhibits the fusion process [Stow et

al., 2006]. Therefore, mutant forms or depletion of individual SNAREs block specific

fusion steps and have been used to map transport pathways in macrophages [Stow and

Murray, 2013]. Pathologically, SNAREs are targets of the zinc‐dependent endopeptidase

activity of the clostridial neurotoxins [Pickett and Perrow, 2011]. These toxins enter the

cell by endocytosis after binding to the cell surface of neuronal cells. Once within the cell,

they can cleave certain members of the SNARE family and thereby inhibit membrane

fusion. The defect in the release of the neurotransmitter acetylcholine causes the

diseases tetanus and botulism and can be fatal. However, the toxin may be re‐engineered

to target specific trafficking pathways in other cell types, such as macrophages [Pickett

and Perrow, 2011].


Figure 1.9. Intracellular trafficking pathways in macrophages and SNARE proteins associated with the different compartments. Endocytosed material transported to the early endosome (orange) is either targeted to the lysosome (green) or recycled back to the cell surface. There are at least two major exocytic pathways by which proteins can be delivered to the cell surface in macrophages. The classical pathway takes cargo from the Golgi complex (blue) to the plasma membrane (black) either directly or via the recycling endosome (purple). In the lysosomal secretory pathway, cargo is transported via lysosomes (green) or lysosome‐related organelles (brown) en route to the cell surface. Both pathways require cargo to be transported in membrane bound carriers, which must at each step fuse with the plasma membrane and is mediated by SNARE proteins. Individual SNARE members locate to distinct membranes (boxes) and the selective interaction of one R-SNARE (pink) on a donor membrane and two to three cognate Q-SNAREs (aqua) on a target membrane forms a trans‐SNARE complex that drives the fusion process.

1.4.4.3 Trafficking pathways and machinery in macrophages

To fulfil their various tasks during the immune response, macrophages

extensively use their endocytic and exocytic pathways, which allows phagocytosis and

the release of large amount of newly made inflammatory mediators into the extracellular

space, as well as other important processes. Hence, SNARE proteins and other associated

trafficking machinery have a great impact on macrophage function and the overall

inflammatory response within tissues [Stow et al., 2006].

In unactivated macrophages, low levels of constitutive trafficking occurs to

maintain basic functions of protein transport and cell homeostasis. Some proteins, such

as cytokines, may be transported through to the cell surface for secretion, however the

protein levels of certain SNAREs are rate‐limiting for these trafficking events [Murray et


al., 2005]. Upon activation, macrophages upregulate the expression of important SNAREs

and other machinery to increase vesicular trafficking [Stow et al., 2006]. This is

important to permit secretion of large quantities of cytokines within minutes to hours

following activation as macrophages lack the ability to pre‐pack proteins to be released

from granules as is the case in some other immune cells [Murray and Stow, 2014]. LPS‐

stimulation for example up‐regulates the SNAREs Stx6‐Stx7‐Vti1b to accommodate for

increased TNF transport [Murray et al., 2005]. The levels of specific SNAREs can also be

changed experimentally in order to manipulate particular trafficking pathways.

At least two major pathways are known in macrophages by which proteins can

be delivered to the cell surface [Murray and Stow, 2014]. Proteins that have an N‐

terminal signal sequence are targeted for synthesis in the ER. Proteins can then be

shuttled through the Golgi apparatus and transported to the surface in membrane bound

compartments or vesicles either directly or through the recycling endosome (Figure 1.9).

This is known as the classical transport pathway. In the second major secretory pathway,

the lysosomal secretory pathway, lysosomes or lysosome‐related organelles target

proteins to the surface either directly from the cytosol or from the Golgi complex (Figure

1.9). Both pathways require cargo to be transported between organelles and to the cell

surface in membrane bound carriers, which must at each step fuse with the organelles

en route or the plasma membrane mediated by the appropriate trafficking machinery

proteins [Stow and Murray, 2013].

Secretory mechanisms in macrophages are not fully elucidated yet but mapping

of exocytic pathways have begun and are best understood for certain cytokines. Tumour

necrosis factor α (TNFα) is trafficked from the Golgi to the recycling endosome through

interaction of the Golgi SNARE complex Stx7/Vti1b/Stx6 with the Recycling Endosome

SNARE VAMP3 and its secretion then mediated through association of VAMP3 with

Stx4/SNAP23 at the plasma membrane in macrophages [Murray et al., 2005a; Murray et

al., 2005b; Pagan et al., 2003]. In macrophages, VAMP3 is also required for secretion of

IL‐10 and IL‐6. For the release of IL‐6, VAMP3 cooperates with Vti1b and Stx6

[Manderson et al., 2007; Stanley et al., 2012]. On the other hand, Stx11 has been shown

to regulate the secretion of a protein called Flightless (Flii) via a late

endocytic/lysosomal compartment in macrophages [Lei et al., 2012]. Generally,

knowledge about other SNARE proteins still relies mostly on information obtained from

other cell types and requires further investigation in macrophages.


1.4.4.4 Intracellular trafficking of MMP9 and MMP14 in macrophages

There are currently no effective drugs that clinically alter MMP levels in wounds

and as such there is a need for a greater understanding of how these enzymes are

secreted in order to provide new therapeutic targets to treat chronic wounds. MMP9 is

highly elevated in chronic wounds and linked to poor wound healing outcomes as they

lead to excessive ECM degradation [Rohl and Murray, 2013]. Membrane‐associated

MMP14 is believed to play a role in macrophage migration and infiltration of wound

tissue, where they further contribute to the level of MMPs at the wound site [Linder,

2007]. The intracellular trafficking pathways responsible for secretion of MMP9 and the

delivery of MMP14 to the cell surface are unknown in macrophages and unclear in other

cell types.

Intracellular trafficking of MMP9

In neuroblastoma cells, MMP9 has been found to traffic through a Golgi

dependant pathway along microtubules and filaments to the cell surface in an

anterograde and retrograde manner [Sbai et al., 2010]. In astrocytes, MMP9 was further

demonstrated to co‐localise with the lysosome marker LAMP2 [Sbai et al., 2010] while

in breast carcinoma cells [Bobrie et al., 2012] and neutrophils [Brzezisnka et al., 2008],

a lysosome‐related GTPase, Rab27a, is involved in MMP9 secretion, suggesting MMP9 is

stored in and trafficked via a lysosomal‐related organelle in these cell types. It is

noteworthy that trafficking of MMP9 is cell type dependent [Vandooren et al., 2013] and

this may depend on the pathways utilised although this has yet to be tested. Upon

stimulation, mature neutrophils, which have MMP9 pre‐stored in granules, can release

MMP9 within minutes [Chakrabarti et al., 2006]. Macrophages on the other hand, rely on

de novo synthesis prior to the secretion of MMP9, a process that takes at least several

hours [Opendakker et al., 2001]. In murine macrophages MMP9 has been shown to be

trafficked in Golgi‐derived vesicles that are microtubule‐associated but do not co‐localise

with markers for lysosomal degradation such as LAMP1, suggesting MMP9 is trafficked

through a classical pathway [Hanania et al., 2012]. Furthermore, the SNARE VAMP3

(recycling endosome associated) together with the plasma membrane SNARE SNAP23

are required for secretion of MMP9 in fibrosarcoma cells, suggesting the classical

pathways is used in these cells [Kean et al., 2009]. These contradictive findings suggest

that MMP9 could be trafficked through the classical pathway via the recycling endosome

in some cell types and secreted via a lysosomal pathway in other cells. The responsible


pathways for secretion of MMP9 in macrophages remain unclear and the machinery that

regulates MMP9 secretion pathways is unknown.

Intracellular trafficking of MMP14

Many studies have been undertaken to investigate the trafficking of MMP14. As

MMP14 is believed to play a major role in cancer invasion, most of these studies have

been conducted using breast adenocarcinoma and fibrosarcinoma cell lines and not

macrophages [Poincloux et al., 2009]. Studies found that pro‐MMP14 processing occurs

in the Golgi complex and is required for the delivery of mature MMP14 to the plasma

membrane [Frittoli et al., 2011]. In resting fibrosarcoma cells, MMP14 is also

continuously internalised and trafficked to early and late endosomes before being

recycled back to the cell surface [Poincloux et al., 2009; Remacle et al., 2003]. As a result,

a substantial amount of MMP14 is localised inside the cell with cell surface expression

being weak in most cell types and thus, polarised secretion of pre‐existing MMP14 is

highly regulated [Poincloux et al., 2009]. There are conflicting results about whether

MMP14 could be trafficked to the cell surface via a classic pathway or through the

lysosome. MMP14‐positive vesicles were found to localise in Rab7‐positive late

endosomes [Williams and Coppolino, 2011; Yu et al., 2012] and to co‐localise with the

lysosome‐related SNARE VAMP7 [Steffen et al., 2008; Williams and Coppolino, 2011] in

breast adenocarcinoma and fibrocarcinoma cells. Controversially, it was found that

polarised MMP14 exocytosis from the Golgi to the cell surface involves the recycling

endosome‐associated GTPases Rab4 and Rab11 in fibrosarcoma cells [Remacle et al.,

2005], suggesting it traffics through these organelles. Additionally, the endosomal

SNAREs VAMP3 and Stx13 together with the plasma membrane SNARE SNAP23 mediate

trafficking of MMP14 to the cell surface in fibrosarcoma cells [Kean et al., 2009] as does

the cell surface SNARE Stx4 in gastric epithelia cells [Miyata et al., 2004]. This shows that

there are contradictory results about whether MMP14 could be trafficked via a classical

or lysosomal pathway for these cell types. In primary human macrophages, MMP14

endocytosis is regulated by Rab5a and recycled by Rab14 (early endosome to cell

surface) and Rab22 (early endosome to recycling endosome), whereas newly

synthesised MMP14 is proposed to be trafficked by a Rab8a‐dependant pathway from

the Golgi apparatus to the cell surface [Wiesner et al., 2013]. But whether the transport

in exocytic vesicles to the cell surface occurs directly or via the recycling endosome is not

clear. It would be advantageous to expand the current knowledge of pathways for

MMP14 surface delivery in macrophages and its influence on macrophage invasion.


1.5 Aims

MMP14 may play a key role in macrophage migration and infiltration of wound

tissue and therefore influences the level of inflammation within a wound. MMP9 may

also promote macrophage mobility by degrading the extracellular matrix. In addition,

MMP9 and MMP14 are highly elevated in chronic wounds further contributing to tissue

damage and inflammation at the wound site. Their increased levels have been linked to

poor wound healing outcomes and chronic wounds. Exactly how MMP9 and MMP14 are

delivered to the cell surface in macrophages and whether they have an influence on

macrophage mobility is mostly unknown. At present, there are no clinically viable drugs

that reduce MMPs in wounds and as such there needs to be greater understanding of

how these enzymes are secreted in order to provide new therapeutic targets to dampen

inflammation and speed up the repair process in chronic wounds. Based on this, the

following research questions were identified:

1. Which trafficking routes are used to deliver MMP9 to the cell surface

and which are the key trafficking proteins involved in this process?

2. Which trafficking routes are used to deliver MMP14 to the cell surface

and which are the key trafficking proteins involved in this process?

3. What influence does MMP14 and its surface delivery trafficking

pathways have in the migration of macrophages through the

extracellular matrix?

This body of research aims to answer these questions and thereby elucidate the

processes of MMP trafficking in macrophages, their influence on macrophage infiltration

into wounded tissue and identify potential therapeutic targets to reduce MMP levels in

chronic wounds. To investigate these research questions, the following aims were

established:

Aim 1: Determine the specific intracellular trafficking pathways

responsible for trafficking of MMP9 to the cell surface for secretion.

Aim 2: Determine the specific intracellular trafficking pathways

responsible for trafficking MMP14 to the cell surface for incorporation

into the plasma membrane.

Aim 3: Characterising MMP‐mediated macrophage migration in 3D.

Chapter 2: Materials and methods 41

Chapter 2: Materials and methods

2.1 Cell culture and treatments

2.1.1 Cell culture of RAW264.7 macrophages

The mouse macrophage cell line RAW264.7 (American Type Culture Collection

(ATCC); ATCC Number TIB‐71™) was maintained as semi‐adherent cultures on

bacteriological Petri dishes (BD Biosciences) in RPMI1640 medium (Gibco®)

supplemented with 10% (v/v) heat‐inactivated Foetal Bovine Serum (FBS; Sigma‐

Aldrich) and 2 mM GlutaMAX™ (Gibco®). Cells were maintained in a humidified

incubator at 37 °C and 5% CO2. Cells were passaged every 2‐3 days, up to passage 17,

using a 10 ml syringe with an 18 gauge needle to resuspend the cells. The cells were then

pelleted at 300 x g for 3 min and resuspended in 10 ml media for subculturing at a ratio

of 1:5 to 1:20. For most experiments, cells were seeded at 1:10 onto tissue culture‐

treated plates (Corning Incorporated), where they grow as adherent cells. For

fluorescence microscopy analysis, cells were seeded onto round glass coverslips

(diameter 10 mm, thickness 0.13‐0.16 mm; Menzel‐Gläser). In relevant experiments cells

were plated onto coverslips that had been incubated for 2 h at 37 °C with 50 μg poly‐L‐

lysine (Sigma‐Aldrich) in PBS and for a further 4 h at 37 °C with 10 μg fibronectin (Sigma‐

Aldrich) in PBS. To record the dynamic movement of MMP‐containing vesicles by live cell

imaging, cells were seeded into Matek dishes (Matek), which contain a 12‐mm diameter

round glass coverslip inserted into the bottom of the plate. Prior to live imaging, the

medium was changed to pre‐warmed CO2‐independent medium (Gibco®) supplemented

with 10% FBS and 2 mM GlutaMAX™.

2.1.2 Activation of RAW264.7 macrophages

To classically activate macrophages, RAW264.7 cells were cultured with

lipopolysaccharide (LPS), a component of the outer bacterial membrane of Gram‐

negative bacteria, derived from Salmonella enterica serotype Minnesota (Sigma‐Aldrich),

at a final concentration of 100 ng/ml for up to 18 h. For migration experiments

macrophages were stimulated with the chemoattractant N‐formyl‐methionyl‐leucyl‐

proline (fMLP; Sigma‐Aldrich) at a final concentration of 100 nM.

42 Chapter 2: Material and methods

2.1.3 Disruption of protein synthesis and intracellular trafficking

To stop protein synthesis, cycloheximide (Sigma‐Aldrich) was added to the

medium of LPS‐stimulated cells (9 or 12 h depending on the experiment) at a final

concentration of 1 μg/ml for the last 4 h, to block protein synthesis. Brefeldin A (Sigma‐

Aldrich), which inhibits protein transport from the ER to the Golgi and leads to collapse

of the Golgi stacks, was added to LPS‐stimulated (9 or 12 h depending on the experiment)

cells at a final concentration of 5 μg/ml for the last 6 h. To disrupt microtubules,

nocodazole (Sigma‐Aldrich) was added to LPS‐stimulated (9 or 12 h depending on the

experiment) RAW264.7 cells at a final concentration of 5 μM for the last 3 h. Cytochalasin

D (Sigma‐Aldrich), which disrupts filamentous actin, was added to LPS‐stimulated cells

(9 or 12 h depending on the experiment) at a final concentration of 1 μM for the last 2 h.

2.1.4 Transient overexpression of proteins

The cationic lipid based transfection reagent Lipofectamine® 2000

(Invitrogen™) was used to transiently overexpress proteins (fluorescence‐labelled

MMPs) in macrophages that had been plated onto tissue culture dishes with or without

coverslips on the previous day. Macrophages were transfected with 0.1 μg DNA and

0.4 μl Lipofectamine® 2000 reagent diluted in 25 μl of Opti‐MEM® per cm2 of tissue

culture surface area. The cells were incubated with the final transfection mix for 3 h at

37 °C and then with normal media at 37 °C for 3 h and were either imaged live (section

2.3.2), fixed and further immunofluorescence labelled (section 2.2.5) or lysed and

subjected to SDS‐PAGE and immunoblotting analysis (section 2.2). The mammalian

expression vectors used in this project are listed in Table 2.1.

Table 2.1. List of plasmids used for this project.

Protein Vector Obtained from

MMP14-GFP peGFP Prof Jim Norman, The Beatson Institute for Cancer Research, Glasgow, UK

MMP14-mCherry pcDNA3.1(+) Prof Jim Norman, The Beatson Institute for Cancer Research, Glasgow, UK

MMP9-mCherry pCI-neo Dr Tyler Duellman, University of Wisconsin-Madison, USA

ssT-Cad*-eGFP pCI-neo Dr Tyler Duellman, University of Wisconsin-Madison, USA

* T-cadherin signal sequence


2.1.5 Targeted knockdown using siRNA

To reduce the expression levels of SNARE proteins, double‐stranded short

interfering RNA (siRNA) molecules were used (Table 2.2) according to the

manufacturer’s instructions. In brief, RAW264.7 macrophages were transfected using a

mix of 22.5 pmol siRNA and 0.45 µl Lipofectamine® 2000 (Invitrogen™) in 25 µl Opti‐

MEM® (Gibco®) per cm2 tissue culture surface area. The transfection mixture was

prepared by incubation of the Lipofectamine® 2000 in Opti‐MEM® for 5 min before the

addition of the respective siRNA and incubating for a further 20 min. The cells were then

incubated with the transfection mix for 3 h at 37 °C, the media was replaced with fresh

RAW cell culture media and the cells cultured overnight. This process was repeated the

following day. After the second transfection, cells were either incubated overnight or, if

equal cell numbers were necessary, removed from the tissue culture plates 5 h after

transfection. Detachment of cells was achieved by incubating them in 0.5 mM EDTA/PBS

for 5 min at 37 °C, followed by gently scraping the cells off the plate using a rubber cell

scraper. Cells were then counted and reseeded at equal densities.

Table 2.2. siRNA sequences used to knockdown SNARE proteins.

Target ID* Sequence Negative Control

Catalogue number: AM4635

scrambled siRNA; no sequence information available for Silencer® Negative Control No. 1

SNAP23 64778 5’-GGCAUGGACCAAAUAAAUAtt-3’ 3’-UAUUUAUUUGGUCCAUGCCtt-5’

Stx2 157352 5’-GGCGGAAAAAGUGGAUAAUtt-3’ 3-‘AUUAUCCACUUUUUCCGCCtg5’

Stx3 151904 5’-CGGCAGCUUGAAAUUACUGtt-3’ 3-‘CAGUAAUUUCAAGCUGCCGtt5’

Stx4 151036 5’-GGAAGCUGAUGAGAAUUACtt-3’ 3’-ctCCUUCGACUACUCUUAAUG-5’

VAMP2 186987 5’-GGAAGGAUCUAAUCUUUUUtt-3’ 3-‘AAAAAGAUUAGAUCCUUCCtc5’

VAMP3 186988 5’-CCUGCCGUGUUAUCGAGCUtt-3’ 3’-ttGGACGGCACAAUAGCUCGA-5’

VAMP4 184543 5’-GCUUCGAAGGCAAAUGUGGtt-3’ 3-‘CCACAUUUGCCUUCGAAGCtg5’

VAMP7 261085 5’-GGCACAAGUCUCCUUGUAAtt-3’ 3-‘UUACAAGGAGACUUGUGCCtt5’

VAMP8 188662 5’-GCCACGUCUGAACACUUCAtt-3’ 3-‘UGAAGUGUUCAGACGUGGCtt5’

* All purchased from Ambion®, Thermo Fisher Scientific

2.1.6 Fluorescent gelatin degradation assay

To perform fluorescent gelatin degradation assays, coverslips were prepared

as follows. First, the coverslips were cleaned by incubation in 1 M HCl overnight,

followed by rinsing in distilled water twice and 95% ethanol twice. To aid attachment of

the Oregon Green® 488 conjugated gelatin (Molecular Probes™), the coverslips were


first coated with 50 μg poly‐L‐lysine (Sigma‐Aldrich) for 1 h at room temperature (RT).

After removing the poly‐L‐lysine, the coverslips were washed three times with PBS. Next,

0.5% glutaraldehyde in PBS was added for 15 minutes at 4 °C to activate the poly‐L‐

lysine surface for further protein attachment. Following removal of the glutaraldehyde,

each coverslip was rinsed three times with PBS. Finally, coverslips were coated with 0.2

mg/ml gelatin in 2% sucrose/PBS and allowed to set for 15 minutes at RT, followed by

two rinses in PBS. All steps including, and subsequent to, fluorescent‐gelatin coating

were performed in low lighting or in the dark so as to protect the fluorophore from

photobleaching. Any free aldehydes were quenched by incubation with amino‐acid‐

containing growth medium at RT for at least 20 minutes. To prepare for cell plating, the

coated coverslips were sterilised with 95% ethanol for 15 minutes at RT. After this step,

coverslips were either used immediately or stored at 4 °C for up to two weeks.

Macrophages were seeded on to the gelatin coated coverslips and incubated for

18 h in a humidified incubator at 37 °C. In relevant experiments MMP activity was

inhibited with a broad spectrum MMP inhibitor, GM6001 (Sigma‐Aldrich) at a final

concentration of 25 μM in DMSO and DMSO alone was added to control wells. MMP14

activity was also inhibited with a function blocking antibody [LEM‐2/63.1] specific to the

catalytic domain of MMP14 (Abcam, ab78738). A matched isotype control was added to

some wells at a final concentration of 10 μg/ml. Approximately 18 h after plating, the

growth medium was removed and the cells on the coverslips were fixed and stained as

described in section 2.2.5.

Using ImageJ 1.46r software (National Institutes of Health, Bethesda, MD, USA),

the mean grey value for the gelatin fluorescent signal (Oregon Green® 488 conjugated

to gelatin) was determined for areas that were positive for phalloidin staining.

Degradation was determined per cell by subtracting this mean grey value from the mean

grey value of the background adjacent to the individual cell.


2.1.7 Inverted invasion assay

Inverted invasion assays [Hennigan et al., 1994] were performed using either

undiluted Matrigel (12 mg/ml) or gelled collagen plugs. To obtain gelled collagen, Bovine

collagen I (Gibco®) was mixed with PBS (final concentration: 1 X), NaOH (final

concentration: 0.02 N) and distilled water to a final concentration of 4 mg/ml. 100 μl of

either Matrigel or collagen matrix was then transferred into a Transwell® insert

(Corning) and allowed to set for at least 30 minutes at 37 °C. Inserts were then inverted,

5×105 cells were seeded directly onto the bottom side of the membrane and allowed to

attach for at least 2 h. Inserts were reinverted, more growth medium was added to the

bottom chamber, and medium supplemented with the chemoattractant 100 nM fMLP

was added on top of the matrix (Figure 2.1). In some experiments the broad spectrum

MMP inhibitor GM6001 (25 μM in DMSO) or an antibody specific for the catalytic domain

of MMP14 [LEM‐2/63.1] (10 μg/ml) was added to the bottom chamber. Invading cells

were fixed 72 h after seeding in 4% paraformaldehyde in PBS for 1 h at RT, washed once

with PBS and were stored in PBS at 4 °C until analysis.

Permeabilisation of fixed cells within gel plugs was performed using 0.1% Triton

X‐100 in PBS for 10 min at RT. Cells were washed twice and blocked with 0.5% (w/v)

Bovine Serum Albumin (BSA, Sigma) in PBS for 10 min at RT. The cells were then

incubated with phalloidin‐TRITC diluted in 0.5% BSA in PBS (see Table 2.3 for dilutions

used) for 1 h at RT in the dark. After three washes, cells were stored in PBS at 4 °C until

they were imaged.

Imaging was conducted on a Leica SP5 confocal laser scanning microscope (see

2.3.1). After imaging the cells that did not cross the membrane by confocal microscopy,

the membrane was removed to visualise those cells that had invaded the gel. Serial

optical sections of the collagen gel were captured at 3‐15 µm intervals. The cell area for

each section was measured by setting a high intensity threshold for the phalloidin signal

and using the particle analysis function of ImageJ. The sum of all area measurements for

each stack was determined and the mean of at least three technical replicates calculated.

This index of invasion was normalised by comparing the treatment groups to a control.

At least three independent experiments were performed for each sample unless stated

otherwise.


Figure 2.1. Inverted Invasion Assay. Matrix material (Matrigel or collagen I) was allowed to set in a Transwell® insert (Corning) before cells were seeded onto the bottom side of the membrane and placed in growth medium. Medium supplemented with fMLP as the chemoattractant was added on top of the matrix.


2.2 Biochemistry assays

2.2.1 Sample preparation

To collect conditioned media, cell culture supernatant was collected and

centrifuged at 2500 x g for 5 min to remove any cells and cell debris. The cell culture

supernatant was transferred to a new tube and either used immediately or snap‐frozen

in liquid nitrogen and stored at ‐20 °C until use. To prepare whole cell lysates, cells grown

on tissue culture plates were placed on ice, washed once with ice‐cold PBS and scraped

into ice‐cold lysis buffer (20 mM Tris/HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton

X‐100 and Complete™ protease inhibitors (Roche Applied Sciences)). Cells were lysed by

sonication in a standard sonicating water bath for 2 x 5 min at 4 °C, with a 5 min

incubation on ice. The lysate was then centrifuged for 20 min at 17,000 x g and the

supernatant collected. Protein concentration was determined using the Bio‐Rad

Bradford protein assay (Bio‐Rad Laboratories) according to the manufacturer’s

instructions. Samples were either used immediately or stored at ‐20 °C.

2.2.2 Gelatin Zymography (MMP9)

Samples were mixed with 5 x Zymography sample buffer (5 x buffer: 200 mM

Tris/HCl pH 6.7, 10% SDS, 50% glycerol, 0.025% bromophenol blue). Equal volumes

(conditioned media) or equal amounts of protein (30‐100 µg lysates) were loaded onto

10% Ready Gel® Zymogram Gels (Bio‐Rad), along with a lane each of Benchmark™

prestained protein ladder (Invitrogen), and PagerRuler™ prestained protein ladder

(Thermo Scientific™) markers to enable the visualisation of protein separation and to

determine the size of protein bands. Gels were run at 125 V in buffer containing 25 mM

Tris, 192 mM glycine and 0.1% SDS until the bromophenol blue tracking dye ran out of

the bottom of the gel. After incubation with Renaturing Buffer (2.5% Triton) for 30 min

with gentle agitation at RT the gels were equilibrated in Developing Buffer (50 mM Tris

pH 7.6, 10 mM CaCl2, 50 mM NaCl) for 30 min at RT and then incubated in fresh

Developing Buffer overnight at 37 °C. Gels were stained with 0.1% Coomassie Brilliant

Blue in 50% methanol and 10% acetic acid overnight and destained in water or 50%

methanol and 10% acetic acid. Gelatinase activity is revealed as clear bands in the blue

gel, and the zymogen (105 kDa) or mature form of murine MMP9 (95 kDa) can be

discriminated by molecular weight. FBS‐free conditioned media from the fibroblast cell

line baby hamster kidney (BHK) cells overexpressing the human zymogen form of MMP9

(kindly provided by Dr Eliza Whiteside, QUT) was used as a positive control for the assay.

Images of the gels were taken using the Bio‐Rad ChemiDoc™ XRS system.


2.2.3 SDS-Page and Immunoblotting

Protein samples were denatured in Laemmli sample buffer (5 x buffer: 200 mM

Tris/HCl pH 6.7 containing 10% SDS, 50% glycerol, 0.025% bromophenol blue, and 250

mM DTT) by boiling for 5 min at 95°C. Up to 100 μg of protein per lane was loaded onto

SDS‐PAGE gels. Depending on the size of the proteins of interest, gels ranged from 10‐

15% acrylamide (Bio‐Rad Laboratories). Gels placed in the Bio‐Rad Mini‐Protean gel

apparatus (Bio‐Rad Laboratories) were electrophoresed in 25 mM Tris buffer containing

192 mM glycine and 0.1% SDS, at 90 V until the proteins had moved from the stacking

gel into the separation gel, at which point the voltage was increased to 180 V.

Benchmark™ prestained protein ladder (Invitrogen), and PagerRuler™ prestained

protein ladder (Thermo Scientific™) markers were also loaded onto each gel. Gels were

transferred onto methanol‐hydrated Immuno‐Blot® poly‐vinyl‐D‐fluoride (PVDF)

membranes (0.2‐μm pore size; Bio‐Rad Laboratories). Proteins were transferred at 100

V for 90 min at 4 °C using the Bio‐Rad Mini‐Protean wet transfer system (Bio‐Rad

Laboratories) and a standard transfer buffer (15 mM Tris buffer containing 120 mM

glycine and 10% or 20% (v/v) methanol depending on protein size). Protein transfer

was confirmed by staining the membrane with 0.1% Coomassie Brilliant Blue 50%

methanol, and 10% acetic acid followed by destaining (50% methanol and 10% acetic

acid) to visualise protein bands. Membranes were then dried and stored for later use or

rehydrated and then blocked with 15% (w/v) skim milk powder (Woolworths

Homebrand) in PBS containing 0.05% Tween‐20 (PBST) for 15 min at RT. Membranes

were then incubated in primary antibody at an antibody dependant concentration (see

Table 2.3 for individual dilutions) in 3% (w/v) skim milk in PBST for 1 h at RT or

overnight at 4 °C depending on the antibody. Subsequently, membranes were washed

three times in PBST, followed by incubation with HRP‐conjugated secondary antibody

(Table 2.3) in 3% (w/v) skim milk in PBST for 1 h at RT. Membranes were then washed

five times with PBST for 5 minutes each and incubated with a 1:1 mix of ECL solution 1

(100mM Tris pH 8.5, 2.5mM luminol, 0.4mM p‐coumaric acid) and ECL solution 2

(100mM Tris pH 8.5 containing 0.02% H2O2) for 5 min. Protein bands were detected by

exposure of X‐ray film to the membrane for various times and the X‐ray film developed

and scanned. Densitometry was performed using ImageJ 1.46r software. Protein bands

of interest were normalised to the respective actin loading control band. To quantify

changes in protein levels over time, the first time point was let equal 1 and the fold

change was calculated for the following time points. To quantify changes in protein levels

following treatment, the control was set to 1 and the respective fold change was

calculated for the treatment group.


Table 2.3. Antibodies used for immunoblotting.

Antibody Host Dilution Source

Primary

Actin [C4] Mouse 1:20000 Millipore (MAB1501) MMP14 [EP1264Y] Rabbit 1:2000 Abcam (ab51074)

MMP9* Rabbit 1:1000 Abcam (ab38898) MMP9 [E-11]† Mouse 1:500 Santa Cruz (sc-393859)

SNAP23 Rabbit 1:2000 Synaptic Systems (111202) Stx2 Rabbit 1:1000 Synaptic Systems (110123) Stx3 Rabbit 1:1000 Synaptic Systems (110033) Stx4 Rabbit 1:1000 Synaptic Systems (110042)

VAMP2 Mouse 1:1000 Synaptic Systems (104211) VAMP3 Rabbit 1:500 Synaptic Systems (104103) VAMP4 Rabbit 1:1000 Abcam (ab3348) VAMP7 Mouse 1:1000 Abcam (ab36198) VAMP8 Rabbit 1:1000 Synaptic Systems (104302)

Secondary Anti-mouse-HRP Goat 1:3000 Novex™ (A16072) Anti-rabbit-HRP Goat 1:3000 Novex™ (G21234)

* Only a limited number of experiments were performed with this polyclonal rabbit antibody as antibody from a new lot was no longer able to detect MMP9 in conditioned media or lysates from RAW264.7 macrophages. †Some later experiments were performed using this antibody as it became commercially available and was positively tested to detect MMP9 in lysates from RAW264.7 macrophages.

2.2.4 Flow cytometry (MMP14)

Cells were harvested by incubating them in 0.5 mM EDTA in PBS for 5 min at

37 °C, followed by gently scraping with a rubber cell scraper. Cells were pelleted at 300

x g for 3 min, washed twice in ice‐cold 2% FBS in PBS for 10 min and incubated for 5 min

on ice with 10 µg/ml Fc Block (BD Biosciences). Cells were labelled for 30 min on ice

with 5 µg/ml MMP14 antibody, isotype control antibody (


2.4) in 2% FBS in PBS or 2% FBS in PBS alone (unstained control). Cells were

washed 3 x with 2% FBS in PBS, followed by 30 min incubation on ice with AlexaFluor®

647‐conjugated anti‐rabbit IgG in 2% FBS in PBS, or buffer alone (unstained control).

Cells were then washed 2 x with 2% FBS in PBS followed by PBS. For analysis, cells were

resuspended in PBS. Flow cytometry analysis was performed using the FACSAria cell

sorter with the FACSDiva acquisition software (BD Biosciences). FlowJo software

(FlowJo, LLC) version 10 was used for data analysis. Cell populations were gated based

on forward and sideward scatter characteristics. Fluorescence intensity of each sample

was assessed based on the mean fluorescence intensity (MFI) of the cell population,

which was measured in arbitrary units and corrected for by subtracting the MFI of the

isotype control.


Table 2.4. Antibodies used for flow cytometry.

Antibody Host Dilution Source

Primary MMP14 [EP1264Y] Rabbit 1:100 Abcam (ab51074) Rabbit Isotype [DA1E] Rabbit 1:1000 CST (3900)

Secondary Anti-rabbit AlexaFluor® 647 Goat 1:100 Novex™ (A21245)

2.2.5 Immunofluorescence labelling

2.2.5.1 Standard protocol

Cells on glass coverslips were fixed for 30 min at RT in 4% paraformaldehyde

solution in PBS, washed once with PBS and stored in PBS at 4 °C until required.

Permeabilisation of fixed cells was performed using 0.1% Triton X‐100 in PBS for 5 min

at RT. Cells were washed twice and blocked with 0.5% (w/v) BSA in PBS for 10 min at

RT. The cells were then incubated with primary antibodies diluted in 0.5% BSA in PBS

(Table 2.5) for 1 hour at RT. After three washes with 0.5% BSA in PBS, cells were

incubated with secondary antibodies and non‐antibody fluorescent probes (Table 2.5)

diluted in 0.5% BSA in PBS for 1 hour at RT in the dark, followed by another five washes

with 0.5% BSA in PBS. The coverslips were mounted on glass slides in mounting media

(PBS containing 25 mg/ml 1,4‐diazabiccyclo‐2,2,2‐octane (DABCO, Sigma‐Aldrich) and

80% glycerol), sealed with nail polish and stored at 4 °C.

2.2.5.2 Cell surface labelling (MMP14)

To label MMP14 on the surface of live non‐permeabilised cells, coverslips were

placed on ice. Cells were washed twice, blocked with ice‐cold 0.5% BSA in PBS for 10 min.

Cells were then incubated with MMP14 antibody ([EP1264Y], Abcam) diluted in 0.5%

BSA in PBS for 30 min on ice. Cells were washed three times with 0.5% BSA in PBS and

fixed in 4% paraformaldehyde solution in PBS for 30 min at RT. The remaining labelling

procedure was performed as described in section 2.2.5.1.


Table 2.5. Antibodies and fluorescent probes used for immunofluorescence labelling.

Antibody/Probe Host Dilution Source

Primary

EEA1 Rabbit 1:100 Pierce™ (PA1-063A) GM130 [35/GM130] Mouse 1:100 BD Biosciences (610822)

LAMP1 [1D4B] Rat 1:100 Abcam (ab25245)

LBPA Mouse 1:100 Jean Gruenberg,

University of Geneva, Switzerland

MMP9 Rabbit 1:1000 Abcam (ab38898) MMP14 [LEM-2/63.1] Mouse 1:100 Abcam (ab78738) MMP14 [EP1264Y] Rabbit 1:100 Abcam (ab51074)

TfR [H68.4] Mouse 1:100 Novex™ (13-6800) PDI [1D3] Mouse 1:100 Abcam (ab190883)

SNAP23 Rabbit 1:100 Synaptic Systems (111202)

Stx4 Rabbit 1:100 Synaptic Systems (110042)

α-tubulin Mouse 1:100 Invitrogen

Secondary

Anti-mouse AlexaFluor® 488 Goat 1:100 Novex™ (A11029)

Anti-mouse AlexaFluor® 647 Goat 1:100 Novex™ (A21236)

Anti-rabbit AlexaFluor® 488 Donkey 1:100 Novex™ (A21206)

Anti-rabbit AlexaFluor® 647 Goat 1:100 Novex™ (A21245)

Anti-rat AlexaFluor® 488 Goat 1:100 Novex™ (A11006)

Anti-rat AlexaFluor® 647 Goat 1:100 Novex™ (A21247)

Fluorescent probes

DAPI 1:1000 Sigma-Aldrich (D9542) Phalloidin-FITC 1:200 Sigma-Aldrich (P5282)

Phalloidin-TRITC 1:200 Sigma-Aldrich (P1951) Phalloidin-

AlexaFluor® 647 1:100 Molecular Probes™ (A-22287)

2.2.6 ELISA

The commercially available mouse Total MMP9 DuoSet® ELISA (R&D Systems)

and MMP9 ELISA Kit (biosensis) were used according to the manufacturer’s instructions.

In brief, standards and undiluted conditioned media samples were added to wells of the

pre‐coated plates and incubated overnight at 4 °C. The next day, the plate was washed

five times with PBS before adding the biotinylated antibody diluted in the provided

dilution buffer for 3 h. After washing the plate three times with PBS, the Avidin‐Biotin‐

Peroxidase Complex solution provided was added for 1 h followed by five washing steps

with PBS. For colour development, 3,3′,5,5′‐tetramethylbenzidine substrate was added,

and after 20 min the reaction was terminated using the provided stop solution. The plate

was then read immediately at 450 nm with a microplate reader.


2.3 Microscopy and image analysis

2.3.1 Fixed cell imaging

Fixed cells were imaged using one of the following microscopes:

Nikon® Eclipse Ti‐S epifluorescence microscope fitted with a CFI S Plan Fluor

ELWD ADM 20x (MA 0.45), 40x (NA 0.60) dry objective lens, with a CFI Plan Fluor 60XS

Oil (with iris) and a CFI Plan 100X Oil immersion lens. The system was operated using

the NIS‐Elements BR software (Version 4.11.00). Pictures were captured using a

DigitalSight DS‐Qi1Mc colour camera.

Leica TCS SP5 laser scanning confocal microscope (Leica Microsystems GmbH,

Wetzlar, Germany) fitted with a HCX PL APO CS 20x (NA 0.70), 40x (NA 1.25) and 60x

(NA 1.4) oil immersion lens. The system was operated using the Leica Application Suite

Advanced Fluorescence software (Version 2.6.3.8173).

Olympus BX63F motorised upright fluorescence microscope (Olympus,

Hamburg, Germany) fitted with a UPLSAPO100XO (NA 1.40) oil immersion lens. The

system was operated using the cellSens software. Pictures were captured using a Dual‐

CCD DP80 Microscope Digital Camera.

2.3.2 Live cell imaging

Live cell imaging was carried out using a Leica AF6000LX microscope fitted with

a HCX PL APO CS 63x (NA 1.30) glycerol immersion lens. The system was operated using

the Leica Application Suite Advanced Fluorescence software (Version 2.6.3.8173).

Pictures were captured using a DFC 350 FX digital camera. To record the dynamic

movement of MMP‐containing vesicles, images were acquired at rates of one frame every

5 s for 5 min. Time‐lapse series were constructed, cropped, and adjusted using ImageJ

1.46r.

2.3.3 Image analysis

Images were merged, analysed, cropped and optimised using ImageJ 1.46r, and

Adobe Photoshop CC 2015 (Adobe Systems, San Jose, CA, USA). In some instances images

were pseudo‐coloured using an intensity‐coded palette in ImageJ. Intensity profiles were

produced by measuring the fluorescence intensity along a line scan within an image and

displaying the values in a scatter plot that was graphed in Microsoft Excel.


2.4 Statistics

All graphs present means of at least three independent experiments unless stated

otherwise. Error bars show the standard error of the mean (SEM).

One‐way ANOVA with either Tukey’s or Dunnett’s post‐hoc test was performed

for multiple comparison of data with more than two groups while unpaired, two‐tailed

Student’s t‐test was carried out for data with two groups using a confidence interval of

95%. P‐values <0.05 were considered significant (*P<0.05, **P≤0.01, ***P≤0.001,

****P≤0.0001).

Chapter 3: Identifying intracellular trafficking pathways for secretion of MMP9 55

Chapter 3: Identifying intracellular trafficking pathways for secretion of MMP9

3.1 Introduction

MMPs are a family of enzymes that in concert are responsible for the degradation

of most extracellular matrix proteins. Their proteolytic activity during the repair process

has to be tightly controlled as increased levels of certain MMPs lead to excessive tissue

damage and impairs the wound healing process. MMP9 is highly elevated in chronic

wound fluid and elevated levels correlate with poor wound healing outcomes [Beidler et

al., 2008; Lobmann et al., 2002; Liu et al., 2009; Mwaura et al., 2006]. Macrophage‐

secreted MMP9 is thought to be one of the main enzymes responsible for ECM

degradation in chronic wounds [Rayment et al., 2008; Wall et al., 2002]. Thus, being able

to reduce MMP9 secretion from macrophages may lead to a new therapeutic to treat

chronic wounds. However, the secretory pathways and intracellular trafficking steps

involved in the cell surface delivery and secretion of MMP9 in macrophages are still

poorly understood.

Multiple pathways exist in macrophages for the exocytosis of proteins. Most

secreted proteins, as is the case for MMP9, have an N‐terminal signal sequence that

targets them for import into the ER during biosynthesis, from which they are then

shuttled through to the Golgi apparatus. From the Golgi apparatus, proteins may be

directly transported to the cell surface in membrane bound compartments. However,

more recently it has become clear that transport en route to the cell surface can also

occur via other organelles, such as recycling endosomes. For example, TNFα is secreted

via these endosomes in macrophages [Murray et al., 2005a; Murray et al., 2005b].

Proteins destined for secretion can also be trafficked through late

endosomes/lysosomes, which can function as secretory compartments releasing their

contents upon fusion with the cell surface. For example, IL‐1β is secreted via lysosomes

[Andrei et al., 1999]. Delivery of MMP9 within membrane‐bound vesicles to the cell

surface for secretion requires membrane fusion that is mediated by the SNARE proteins

at all points in its trafficking pathway. There are currently 38 mammalian SNARE

proteins, with some of them being differentially expressed in different cell types. These

SNARE proteins are associated with distinct membranes within the cell, for example,

56 Chapter 3: Identifying intracellular trafficking pathways for secretion of MMP9

VAMP3 is found on recycling endosomes and VAMP4 is found in the Golgi complex in

macrophages. It is the selective pairing of three to four SNARE proteins that mediates

membrane fusion between vesicles, organelles and/or the plasma membrane to regulate

discrete transport pathways. For this reason, they can be used to map intracellular

transport pathways. Identifying the specific SNAREs that facilitate MMP9 secretion in

macrophages will identify the intracellular transport pathways and also provide targets

to prevent secretion of the enzyme from macrophages that could be useful for the

treatment of chronic wounds.

Therefore, this chapter begins with the optimisation of methods to detect

intracellular MMP9 and MMP9 secreted from macrophages*. Next, the subcellular

localisation of MMP9 in macrophages is determined and intracellular compartments that

might form part of its secretory route in these cells characterised. Finally, levels of

SNARE proteins are reduced by siRNA knockdown to test for their involvement in MMP9

secretion and to identify trafficking pathways and machinery that might regulate MMP9

secretion.

* Note, a number of immunoblots throughout this chapter were performed using a polyclonal rabbit antibody to MMP9 that was able to detect MMP9 at the correct size. Along the way, the same antibody was repurchased but the received product was from a different batch with a new lot number and antibody from this batch no longer recognised MMP9. The old lot no longer exits. Data obtained from using the original batch of this antibody is shown but in some cases there was only 1 immunoblot (conditioned media samples). Where this is the case the results have been repeated using a second method (ELISA) with at least three biological replicates.


3.2 Results

3.2.1 Detecting secreted MMP9 in RAW264.7 cell culture supernatants

The influence of FBS concentration on RAW264.7 macrophage cell growth

To identify the specific intracellular trafficking pathways that are responsible for

the surface delivery and secretion of MMP9, an assay was needed that could detect

changes in intracellular and secreted MMP9. MMP levels are typically determined using

zymography, where their proteolytic activity is detected on the basis of the degradation

of a substrate incorporated into an SDS‐PAGE gel [Vandooren et al., 2013]. In the case of

MMP9 that substrate is gelatin. Typically, MMP levels in conditioned media are measured

by zymography in the absence of FBS, which has been found to contain high levels of

gelatinolytic activity [del Zoppo et al., 2012]. RAW264.7 macrophages are usually

maintained in the presence of 10% FBS in their cell culture media. Thus, the effect of

removing or reducing FBS levels was first tested to determine whether it would have any

detrimental effects on macrophage growth. For this, RAW264.7 cells were cultured in

five different concentrations, ranging from 0‐10%, of FBS (in RPMI) and the effect on cell

numbers was initially evaluated using phase contrast microscopy under low

magnification (Figure 3.1A). The growth of RAW264.7 macrophages was found to be

considerably retarded under low FBS conditions relative to control cells grown in the

presence of 10% FBS (Figure 3.1A). To quantify the reduction in growth, cells were

counted using a haemocytometer (Figure 3.1B). A significant reduction (31±6%) in

macrophage cell numbers was observed for cells cultured in serum‐free medium,

compared to control cells grown in 10% FBS‐containing media (*P<0.05). Macrophages

grown in low‐serum conditions (1.25, 2.5 or 5% FBS) had approximately 15% less cell

numbers than those grown in 10% FBS‐containing media. These results reflect the

changes in cell confluency observed by microscopy (Figure 3.1A). To determine whether

reduced FBS levels altered cell morphology, macrophages grown in normal (10%) or low

serum (1.25%) conditions were carefully examined by microscopy. The RAW264.7

macrophages grown under low serum conditions had large vacuoles typically not

present in control cells grown in 10% serum (Figure 3.1C). Additionally, in the presence

of low serum many cells presented with two or more nuclei, which was typically not the

case for macrophages grown in normal serum conditions (Figure 3.1C). Thus, reduced

serum conditions appear to have a deleterious effect on macrophages and as such are

not compatible with RAW264.7 cells.



Figure 3.1. The FBS concentration in the cell culture media influences RAW264.7 macrophage cell numbers and morphology. Equal numbers of RAW264.7 cells were seeded in media containing 10% FBS and allowed to attach to the dish for 2 h. Media was then replaced with media containing different FBS concentrations (0, 1.25, 2.5, 5 or 10%) and cells were incubated for a further 18 h. (A) Representative images from the cell cultures grown in the presence of different FBS concentrations taken with low-magnification phase contrast microscopy. Scale bar is 50 μm. (B) Cell counts of macrophages grown in reduced FBS conditions were calculated as a percentage of cells grown in 10% FBS-containing media and represented in a bar graph; mean ± SEM (n=3), *P<0.05 (One-way ANOVA with Tukey’s post-hoc test). (C) Cells grown on coverslips in 10% (control) or 1.25% FBS (low serum) conditions were fixed and their nuclei labelled with DAPI. Representative overlaid images (top row) from Bright-field (grey) and DAPI fluorescence (blue) channel, as well as single-channel images for the insets, are shown (bottom row). Scale bar is 20 μm.

Zymography of RAW264.7 macrophage conditioned media

As reducing the serum concentration in the growth medium of RAW264.7

macrophages negatively influences cell numbers and morphology, it was then

determined if gelatin zymography could detect MMP9 secreted from RAW264.7cells into

serum‐containing media (conditioned media). MMP levels in serum can vary so it was

determined whether the FBS used in RAW264.7 cell culture media had detectable

proteolytic activity in gelatin zymography assays (Figure 3.2A). Media containing 0‐10%

FBS was assayed using zymography, along with control conditioned FBS‐free media from

BHK cells overexpressing human pro‐MMP9 and endogenously expressing pro‐MMP2.

The lanes loaded with the BHK conditioned media control displayed bands of the

predicted size for pro‐MMP9 (92 kDa) and pro‐MMP2 (72 kDa). Media without FBS (0%)

showed no proteolytic activity in the zymography assays, as expected (Figure 3.2A).

However, media supplemented with 1.25‐10% serum showed a concentration‐

dependant increase in gelatinase activity for both pro‐MMP9 and pro‐MMP2,

respectively (Figure 3.2A). As bovine and murine MMP9 are a similar molecular weight,

bovine MMP9 is most likely masking detection of mouse MMP9 secreted in the

conditioned media. A method to upregulate MMP9 expression was needed to test

whether the mouse MMP9 activity could be upregulated and detected above that of

bovine MMP9. MMP expression has been found to be upregulated upon stimulation with

the strong Toll‐like receptor (TLR) agonist lipopolysaccharide (LPS) [Newby, 2008] and

MMP9 secretion from LPS‐activated RAW264.7 macrophages can be detected in

conditioned media from as early as 8 h post‐stimulation [Yang et al., 2015]. Thus, it was

examined whether LPS activation of macrophages leads to a noticeable upregulation in

MMP9 levels that would allow discrimination between bovine and murine MMP9 by

zymography. Comparing the gelatinase activity in FBS‐containing media and FBS‐

containing conditioned media from macrophages stimulated in the absence or presence

of LPS for 9 h shows that the level of FBS‐originated bovine gelatinolytic activity in media


masks any detectable changes in macrophage‐secreted MMP9 levels after LPS

stimulation (Figure 3.2B). Next, it was tested whether increasing the length of

stimulation could lead to the secretion of MMP9 levels above those found within the FBS

by incubating RAW264.7 cells with LPS over an 18 h time course (Figure 3.2C). However,

after stimulating the cells for 18 h the gelatinase activity of conditioned media did not

reach an intensity level significantly above that found in medium containing 10% FBS

(Figure 3.2C). Overall, these results show that high bovine MMP activity in FBS‐

containing medium masks any macrophage‐secreted MMP9 activity, making it very

difficult to measure gelatinase levels in FBS‐containing conditioned media by

zymography. This method is still suitable to measure intracellular MMP9 as only mouse

MMP9 will be present in the cells. However, a different approach was needed to detect

changes in secretion of MMP9 into the FBS‐containing cell culture supernatant of

RAW264.7 macrophages.


Figure 3.2. Bovine MMP9 found in FBS masks the gelatinase activity of MMP9 secreted by macrophages. (A) RAW264.7 cell culture media containing increasing concentrations of FBS (0, 1.25, 2.5, 5 or 10%) were analysed by gelatin zymography. Conditioned media (without FBS) from human pro-MMP9-secreting BHK cells was used as a positive control (PC) and shows bands for pro-MMP9 (92 kDa) and pro-MMP2 (72 kDa). (B) RAW264.7 cell culture media containing 10% FBS and FBS-supplemented conditioned media (CM) from cells that were stimulated with 100 ng/ml LPS (+LPS) for 0 h and 9 h were analysed by gelatin zymography. Conditioned media from human pro-MMP9-secreting BHK cells was used as a positive control (PC) (C) RAW264.7 cell culture media with and without 10% FBS and FBS-supplemented conditioned media from cells that were stimulated with 100 ng/ml LPS 0-18 h were analysed by gelatin zymography. Conditioned media from human pro-MMP9-secreting BHK cells was used as a positive control (PC).


MMP9 ELISAs and RAW264.7 macrophage conditioned media

Since the serum‐derived and macrophage‐secreted MMPs originate from

different species (bovine and mouse, respectively), it was tested whether a commercially

available ELISA kit against mouse MMP9 (mouse Total MMP9 DuoSet® ELISA, R&D

Systems) could specifically detect murine RAW264.7 macrophage‐originated MMP9 in

the presence of FBS. To first assess whether the sample itself might interfere with the

ability of the assay to detect the target protein (matrix effect), known amounts of the

recombinant mouse MMP9 standard (1000 pg/ml) were spiked into i) dilution buffer

provided in the kit, ii) culture media with and without 10% FBS and iii) conditioned

media from cells incubated in the presence or absence of LPS for 9 h. The spiked samples

were then analysed along with standards using the mouse MMP9 ELISA kit and the

percentage of the expected concentration (recovery) was calculated from a standard

curve. As expected, almost all (97%) of the spiked 1000 pg/ml of MMP9 standard was

recovered when made up in the kit dilution buffer (Figure 3.3A). However, the recovery

of MMP9 spiked in media that lacked FBS was reduced to 75% of that spiked in the kit

dilution buffer. The recovery of MMP9 in media containing 10% FBS was lower again at

44% of that spiked in the kit dilution buffer. The low recovery of murine MMP9 spiked

into media containing FBS suggests that the assay is specific for murine MMP9 but the

sample matrix, both RPMI media and FBS, individually have a negative impact on the

assay itself. Additionally, the recovery of MMP9 spiked in conditioned media from LPS

stimulated and unstimulated macrophages was with 59% and 55%, respectively. This is

less than the recovery of MMP9 from buffer or media without FBS but was higher than

the recovery from media with FBS that had not been conditioned by cells. This could be

due to macrophage‐derived MMP9 being present in the conditioned media, which is

detected by the ELISA and therefore increases the apparent recovery. Nevertheless, the

recovery in these samples remained reduced due to the above mentioned matrix effects.

One way to reduce matrix effects is to dilute samples in dilution buffer to greatly

reduce the amount of media and FBS and thus their effect during the assay. To test

whether this would reduce the matrix effect, buffer and media with and without FBS

were spiked with 5000 pg/ml MMP9 and serially diluted with the kit dilution buffer

(Figure 3.3B). The matrix effect of both RPMI media and FBS decreased as the samples

were increasingly diluted and disappeared completely at 8‐fold or higher dilutions of the

matrix (Figure 3.3B). However, when samples of LPS‐conditioned media were similarly

diluted there was no detectable secreted MMP9 signal within the range of the standard

curve (not shown). This suggested that when the conditioned media is diluted to remove


the matrix effects then the concentration of secreted MMP9 is below the sensitivity range

of the kit and/or there is a factor(s), secreted by the cells, that is further interfering with

the assay.

Since the dilution of sample in the buffer provided with the kit was no longer an

option, it was next tested whether MMP9 recovery could be improved by diluting both

the MMP9 standards and samples in media containing FBS (Figure 3.3C). Murine MMP9

was detected in undiluted conditioned media samples at levels equivalent to the FBS‐

containing media spiked with 5000 pg/ml of the MMP9 (Figure 3.3C). However, when

the sample was serially diluted in FBS media it can be seen that the recovery of the MMP9

in the conditioned media samples is much lower. This suggests that a factor(s) in the

conditioned media might be interfering with the assay. Based on these results, it was

decided to discontinue the optimisation of this particular assay for measuring mouse

MMP9 in RAW264.7 conditioned media samples.



Figure 3.3. The MMP9 ELISA Kit (R&D) is unsuitable for assaying MMP9 in RAW264.7 cell culture supernatants. (A) MMP9 (1000 pg/ml) was spiked into ELISA kit dilution buffer, into RPMI media with and without 10% FBS, and into conditioned media (CM) from RAW264.7 macrophages stimulated with or without 100 ng/ml LPS for 9 h. MMP9 levels were then measured using the mouse total MMP9 DuoSet® ELISA (R&D Systems), according to the manufacturer’s instructions. The amount of MMP9 recovered was calculated from the standard curve as a percentage of the expected concentration and graphed. The graph shows the mean of two technical replicates, n=1. (B) A MMP9 standard (5000 ng/ml in dilution buffer) was prepared in dilution buffer (), media without () and media with 10% FBS () and subsequently serially diluted with dilution buffer and assayed using the mouse total MMP9 DuoSet® ELISA according to the manufacturer’s instructions. The graph shows the standard curves as a function of the Optical density (OD) dependent on the concentration of the standard. Dilution of the top standard in diluents other than buffer results in deviation from the curve. The graph shows the mean of two technical replicates, n=1. (C) MMP9 levels in cell culture supernatant from cells that had been conditioned (CM) with 100 ng/ml LPS for 18 h (undiluted) were compared to a standard curve that was produced through dilution of the standard in media with 10% FBS. The sample was further diluted in media with 10% FBS. MMP9 concentrations in the diluted samples were compared to the standard curve and relative recovery of diluted samples was expressed as a percentage of that and graphed. The graph shows the mean of two technical replicates, n=1.

A second commercially available ELISA kit (Biosensis) with a broader

concentration range (156 pg/ml ‐ 10,000 pg/ml) was subsequently trialled. To test

whether the new ELISA kit had any matrix effects, media with and without FBS, and cell

culture supernatants (conditioned media) from LPS‐stimulated macrophages were

spiked with known amounts of the MMP9 standard and the recovery determined as a

percentage of the expected concentration. For all sample matrices tested, recovery of

MMP9 was similar but significantly lower than the expected value, with an average

MMP9 recovery of ≈63% (Figure 3.4A), which was slightly higher than the R&D ELISA

kit. This suggested that there may again be some sort of matrix effect occurring.

However, unlike the R&D ELISA kit, the matrix effects found with the Biosensis kit

predominantly come from the media alone, with FBS and factors in the conditioned

media having little to no additional effect (Figure 3.4A). The same kit was further tested

for linearity by spiking media containing FBS and conditioned media with 1250 pg/ml

MMP9, then serially diluting this in dilution buffer. It can be seen in Figure 3.4B that it is

linear in this range. Hence, despite the non‐optimal recovery of spiked MMP9 in cell

culture media, this assay was considered suitable to detect fold changes in the amount of

MMP9 secreted by macrophages and was used for subsequent experiments.


Figure 3.4. The Biosensis MMP9 ELISA Kit is suitable to assay changes in MMP9 levels in RAW264.7 cell culture supernatants. (A) Media with and without 10% FBS , and conditioned media (CM) containing 10 % FBS from cells stimulated with 100 ng/ml LPS for 18 h were spiked with 1250 pg/ml recombinant MMP9 standard protein provided in the kit and MMP9 was measured using the Biosensis mouse MMP9 ELISA Kit according to the manufacturer’s instructions. The graph shows the MMP9 recovered as a percentage of the expected concentration calculated using the MMP9 standard curve. The graph shows the mean of four technical replicates, n=1. (B) Media containing 10% FBS and conditioned media (CM) containing 10 % FBS from cells stimulated with 100 ng/ml LPS for 18 h were each spiked with 1250 pg/ml recombinant MMP9 standard protein and subsequently serially diluted with dilution buffer. Recovery determined as a percentage of the expected concentration was graphed. The graph shows the mean of two technical replicates, n=1.


3.2.2 MMP9 expression and secretion is upregulated after LPS stimulation in RAW264.7 macrophages

In order to be able to determine any changes in MMP9 secretion following the

disruption of trafficking pathways, it was first necessary to establish the ideal time points

for such experiments. To investigate the kinetics of LPS‐induced MMP9 synthesis and

secretion from RAW264.7 macrophages, cells were stimulated with 100 ng/ml LPS over

an 18 h time course, and their conditioned media and cell lysates were analysed at 3‐

hourly time points. MMP9 was measured in conditioned media using the Biosensis ELISA

kit (Figure 3.5). Mouse MMP9 was absent in unstimulated samples (0 h) but could be

detected by ELISA in the conditioned media of LPS‐stimulated samples as early as 3 h

post‐stimulation and continued to increase over the 18 h time course (Figure 3.5A).

Figure 3.5A). MMP9 levels 18 h post‐stimulation were significantly increased from 0 h

and 3 h post‐stimulation (*P<0.05). MMP9 secretion did not plateau in the timeframe

tested. To confirm the results obtained from the ELISA, an MMP9 antibody was sourced

and tested for detection of mouse MMP9 by immunoblotting in RAW264.7 macrophages

conditioned media. This antibody did not react with bovine MMP9, as is apparent by the

lack of bands in the unstimulated samples (Figure 3.5B). It was able to detect one band

in conditioned medium, which most likely is the fully glycosylated zymogen form of

MMP9 and runs at 105 kDa (Figure 3.5B). Secreted MMP9 was detected by

immunoblotting as early as 6 h post‐stimulation and the levels continue to increase over

the time course in a similar manner as observed by ELISA analysis (Figure 3.5B).

Intracellular MMP9 levels were also assessed over the same 18 h time course.

Pro‐MMP9 was seen as a single band around 105 kDa by zymography (Figure 3.5C) while

two MMP9 bands were detected by immunoblotting (Figure 3.5D). The higher one of the

two bands is most likely representing fully glycosylated MMP9 (105 kDa) and the lower

molecular weight band, detected only in activated samples (3 to 18 h), is possibly an

underglycosylated form of MMP9 (98 kDa) that occurs during its de novo synthesis

[Olson et al., 2000]. Figure 3.5C and 3.5D show MMP9 in lysates from unstimulated

macrophages (0 h), suggesting there is a low basal MMP9 level present in unstimulated

RAW264.7 macrophages. As with the secreted form, intracellular MMP9 levels increase

after LPS stimulation (Figure 3.5C and D). Together these results show that MMP9

secretion greatly increases in response to LPS stimulation and that time points between

9 h and 18 h were the most suitable for measuring changes in MMP9 levels.


Figure 3.5. MMP9 secretion and expression is increased in response to LPS stimulation. (A) RAW264.7 macrophages were stimulated with 100 ng/ml LPS over an 18 h time course and the conditioned media collected. MMP9 was quantified by ELISA. The graph shows the mean ± SEM (n=3); *P<0.05 (One-way ANOVA with Tukey’s post-hoc test). (B) The same conditioned media was analysed by immunoblotting for mouse MMP9 (105 kDa) confirming this result (C) Cell lysates taken over the same time course were collected and analysed by gelatin zymography. A representative gel image shows the increase in MMP9 activity over time (D) Cell lysates taken over the same 18 h time course were collected and analysed by immunoblotting for MMP9. Actin was used as a loading control. A representative immunoblot is shown.


3.2.3 MMP9 secreted from LPS-activated macrophages is newly made

The increase in secreted MMP9 from LPS‐stimulated macrophages could

predominantly be due to MMP9 de novo synthesis, as suggested in Figure 3.5, rather than

it being due to the release of intracellular pools, as would be expected with neutrophils,

or by the inhibition of the degradation of intracellular MMP9. To confirm this, RAW264.7

macrophages were LPS‐stimulated for 12 h and treated with 1 μg/ml cycloheximide to

inhibit protein synthesis during the final 4 h. Treatment with cycloheximide over the

whole 12 h led to abnormal cells and eventually cell death, so 4 h treatment was chosen

as a sound compromise. Cells on coverslips were stained for F‐actin to define the cell

surface and with DAPI to label nuclei. Activated RAW264.7 macrophages cultured in the

absence of cycloheximide showed the typical macrophage morphology (Figure 3.6A),

where cells are spread out, with what appears to be a leading edge on one side in many

cells. In the presence of cycloheximide, macrophages are elongated as previously

observed [Liu et al., 2010] and the leading edge is less visible (Figure 3.6A). Conditioned

media and cell lysates from the same cycloheximide treatment were analysed for MMP9.

The ELISA results for conditioned media revealed a significant decrease (44±6%;

**P≤0.01) in the level of secreted MMP9 when compared to media from the control cells

(Figure 3.6B). A similar reduction in MMP9 was found when the same samples were

immunoblotted for MMP9 (Figure 3.6C). Inside the cell, MMP9 proteolytic activity was

also reduced in the cycloheximide‐treated cell lysates as shown by zymography (Figure

3.6D), confirming that the reduction in MMP9 secretion was due to reduced intracellular

MMP9 levels. The same lysates were also immunoblotted for MMP9 and it was found that

the underglycosylated nascent precursor form of MMP9 typically found in the ER is

almost absent in lysates from cells treated with cycloheximide (Figure 3.6E). Actin

remained unaffected over the time course (Figure 3.6E). Together, these results suggest

that LPS‐stimulation strongly induces the de novo synthesis of MMP9 and that the MMP9

secreted from activated macrophages is predominantly newly synthesised rather than a

stored form.


Figure 3.6. MMP9 is newly synthesised and secreted upon LPS stimulation. (A) RAW264.7 macrophages grown on coverslips were stimulated with 100 ng/ml LPS for 8 h, then the cells were incubated in the absence (control) or presence of 1 μg/ml cycloheximide (CHX) for a further 4 h. Cells were fixed, permeabilised and stained for F-actin (red) and nuclei were labelled using DAPI (blue). Overlay images of typical macrophage morphologies are shown. Scale bar is 20 μm. (B) Macrophages were stimulated with 100 ng/ml LPS for 8 h, then the cells were incubated in the absence (control) or presence of 1 μg/ml cycloheximide (CHX) for a further 4 h. Conditioned media was collected and the total secreted MMP9 levels were quantified using an ELISA kit (Biosensis). Bar graph of the mean value ± SEM shows the changes in MMP9 levels (n=3). **P≤0.01 (unpaired, two-tailed Student’s t-test). (C) The same conditioned media was collected and MMP9 levels were analysed by immunoblotting (n=1) (D) Cell lysates were collected and intracellular MMP9 levels were analysed by gelatin zymography (n=1) (E) Cell lysates were collected and intracellular MMP9 levels were analysed by immunoblotting for MMP9 and Actin (n=3). A representative image is shown.


3.2.4 Expression of mCherry-tagged MMP9 to study intracellular trafficking pathways

Localisation of a protein in a cell can give an indication of what pathways might

be used for its secretion. Four MMP9 antibodies (one rabbit polyclonal, two different

goat polyclonal and one mouse monoclonal) were tested for use in immunofluorescence

microscopy, of which only one (rabbit polyclonal) was able to detect MMP9. However, a

second batch of this antibody was unfortunately no longer able to detect MMP9. Hence,

RAW264.7 cells were transfected with a fluorescently (mCherry)‐tagged MMP9 as an

alternative approach to immuno‐labelling endogenous MMP9 with an antibody. Cells

were transiently transfected for 6 h with the MMP9‐mCherry construct and lysates from

these cells were immunoblotted for MMP9 (Figure 3.7A). Both the control and MMP9‐

mCherry overexpressing cells show a band at around 105 kDa, representing endogenous

mouse MMP9 (Figure 3.7A). The MMP9‐mCherry expressing cells had an additional

higher molecular weight band at about 120 kDa, representing human MMP9 (92 kDa)

plus the mCherry tag, which is 28 kDa (Figure 3.7A).

To determine whether the mCherry tagged form of MMP9 behaves similarly to

endogenous MMP9, macrophages were grown on coverslips and transiently transfected

with MMP9‐mCherry. Cells were fixed at 3, 6, 9, 12, 15 and 18 h after transfection and

imaged by fluorescence microscopy (Figure 3.7B). MMP9‐mCherry can be observed from

as early as 6 h post‐transfection with the peak fluorescence intensity occurring between

9 and 12 h (Figure 3.7B). At 6 h post transfection MMP9‐mCherry locates predominantly

to the perinuclear region but over time more MMP9‐mCherry is seen in peripheral

structures (Figure 3.7B). A pseudo‐coloured image using an intensity‐coded palette,

where blue is low and red/pink is high, shows MMP9 in the perinuclear region and a

punctate staining pattern in periphery of the cell (Figure 3.7B right hand panel). Co‐

immunostaining for endogenous MMP9 shows both endogenous and overexpressed

MMP9 are colocated in the perinuclear region and in punctate structures in the periphery

of the cell (Figure 3.7C). This suggests that MMP9‐mCherry is being trafficked in the same

vesicles as endogenous MMP9 and the punctate structure at the periphery of the cells

suggest that MMP9 might not be trafficked directly from the Golgi apparatus to the cell

surface for secretion. All subsequent localisation studies were performed using MMP9‐

mCherry transiently transfected cells.



Figure 3.7. MMP9-mCherry localises to the same places as endogenous MMP9 in RAW264.7 macrophages. RAW264.7 cells were transfected to transiently express MMP9-mCherry. (A) Cell lysates from control macrophages (mock) and macrophages transiently expressing MMP9-mCherry were collected and analysed by immunoblotting for MMP9 and Actin. (B) Cells transiently expressing MMP9-mCherry grown on coverslips were fixed at the appropriate time points and MMP9-mCherry fluorescence was imaged (grey scale or pseudo-coloured using an intensity-coded palette). Scale bar is 20 μm. (C) Cells transiently expressing MMP9-mCherry grown on coverslips expressing MMP9-mCherry (red) were fixed, permeabilised and immunolabelled for endogenous MMP9 (green). Scale bar is 20 μm.

3.2.5 MMP9 secretion is not dependent on intact microtubules or actin filaments

Microtubules can assist with the ‘long‐distance’ bi‐directional movement of

vesicles to and from the plasma membrane. Association of protein cargo to microtubules

can be an indication of microtubule‐dependent vesicle transport. Hence, RAW264.7

macrophages grown on coverslips were transiently transfected with MMP9‐mCherry

and immunostained for α‐tubulin to label microtubules (Figure 3.8A). In the periphery

of macrophages, fluorescently‐labelled MMP9 locates in punctate structures, some of

which appear to localise along microtubules (see arrows in Figure 3.8A). To examine

whether the secretion of MMP9 is dependent on intact microtubules, RAW264.7 cells

were stimulated with LPS for 12 h and nocodazole was added for the last 3 h to block

microtubule polymerisation and disrupt the microtubule network. Successful

microtubule destabilisation was confirmed through immunofluorescence staining of α‐

tubulin (Figure 3.8B). In untreated macrophages, microtubule networks are observed

emerging from the microtubule‐organising centre near the nucleus radiating towards

the cell surface (Figure 3.8B). In nocodazole treated cells tubulin‐levels appear to remain

unaffected with equal fluorescence intensity. However, the tubule network is disrupted

and tubulin staining is punctate and dispersed throughout the cell body (Figure 3.8B

right hand panel). Disruption of the microtubule network affects cell shape and

macrophages appear unorganised with multiple protrusions (Figure 3.8B right hand

panel). MMP9 levels in RAW264.7 macrophage conditioned media and cell lysates were

analysed in these cells following nocodazole treatment to determine whether the

microtubule network might regulate MMP9 secretion. The level of secreted MMP9 was

determined by ELISA (Figure 3.8C) and confirmed using immunoblotting (Figure 3.8D)

of the conditioned media. MMP9 secretion was unaffected after disruption of

microtubules (Figure 3.8C and D) (P>0.05). Similarly, intracellular levels of MMP9

measured from lysates of the same treated cells remained unchanged following

nocodazole treatment, as seen by gelatin zymography (Figure 3.8E) and by


immunoblotting for MMP9 (Figure 3.8 F). These results suggest that MMP9 secretion

under such conditions is not dependent on transport along microtubules.

Vesicle transport towards the cell surface not only occurs along microtubules but

can also track along actin filaments. Furthermore, actin‐rich podosomes found on the

outer surface of the plasma membrane of macrophages are believed to play a role in MMP

secretion and migration [Linder 2007]. Thus, to determine whether MMP9 secretion

occurs via actin based structures, LPS‐stimulated macrophages (12 h) were treated with

actin‐disrupting drug cytochalasin for the last 2 h. In general, RAW264.7 cells treated

with cytochalasin exhibited a less pronounced actin cortex and shortened actin filaments

than the control cells (Figure 3.9A). Treated cells also appeared more spread out and

flatter in comparison (Figure 3.9A) as published previously [Otto et al., 2015]. To

determine whether disrupting actin altered MMP9 secretion, levels of MMP9 in

conditioned media and the corresponding cell lysates were analysed. Surprisingly,

cytochalasin treatment resulted in a small increase in secreted MMP9 as determined by

ELISA (Figure 3.9B) and by immunoblotting (Figure 3.9D) of the same samples.

Intracellular MMP9 was also analysed from the same samples by zymography (Figure

3.9D) and immunoblotting (Figure 3.9E). The intracellular MMP9 appeared to remain

unaffected at the 12 h time point. Thus, MMP9 secretion does not require microtubules

or actin for secretion under the conditions analysed.

Figure 3.8. MMP9 secretion is not dependent on microtubules. (A) Cells grown on coverslips expressing MMP9-mCherry (red) were fixed, permeabilised and co-labelled for α-tubulin (green). Lower panels show zoom in regions of upper panels and the arrows indicate MMP9 staining along microtubules. Scale bar is 10 μm. (B-F) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 12 h. After 9 h, the cells were incubated in the absence (control) or presence of 5 μM/ml nocodazole (Noc) for further 3 h. (B) Cells on coverslips were fixed and stained for α-tubulin (green), F-actin (red) and the nuclei (blue). Scale bar is 20 μm. Lower panels show zoomed in regions of upper panels to allow comparison of microtubule structures. (C) Total MMP9 levels in cell culture supernatants were quantified by ELISA and the fold change in secretion is shown in the graph; mean ± SEM (n=3), n.s. P>0.05 (unpaired, two-tailed Student’s t-test). (D) Conditioned media (secreted fraction) was collected and analysed by immunoblotting for MMP9 to confirm the ELISA data (n=1). (E) Cell lysates (cellular fraction) were collected and analysed by gelatin zymography (n=1). (F) Cell lysates were also analysed by immunoblotting for MMP9 using actin as a loading control. A representative blot is shown (n=3).



Figure 3.9. Disruption of actin filaments with cytochalasin does not reduce MMP9 secretion. RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 12 h. For the last 2 h, the cells were incubated in the absence (Control) or presence of 1 μg/ml cytochalasin (Cyto). (A) Cells grown on coverslips were fixed and stained with phalloidin to label F-actin (red) and nuclei were labelled using DAPI (blue). Scale bar is 20 μm. (B) Total MMP9 levels in conditioned media were quantified by ELISA. The graph shows the fold change in MMP9 levels; mean ± SEM (n=3), n.s. P>0.05 (unpaired, two-tailed Student’s t-test). (C) Conditioned media (secreted fraction) was collected and analysed by immunoblotting for MMP9 (n=1). (D) Cell lysates (cellular fraction) were collected and analysed by gelatin zymography (n=1) and (E) by immunoblotting for MMP9 and with actin as a load control (n=3).


3.2.6 Newly synthesised MMP9 is trafficked via the Golgi

MMP9 has an N‐terminal signal sequence that targets it for import into the ER

during biosynthesis [Egeblad and Werb, 2002]. Most secreted proteins are then shuttled

through the Golgi apparatus from where they can be transported to the surface in

membrane bound compartments or vesicles either directly or through other organelles

such as recycling endosomes or late endosomes [Stow and Murray, 2013]. However,

some proteins can bypass the Golgi complex and are shuttled directly from the ER to the

plasma membrane [Stow and Murray, 2013]. To determine whether MMP9 localises to

the Golgi complex, macrophages transiently transfected with MMP9‐mCherry for a

period of 6 h were fixed, permeabilised and immuno‐labelled for the cis‐Golgi marker

GM130. Figure 3.10A shows that some MMP9 localises to the Golgi complex.

To confirm that newly synthesised MMP9 transits through the Golgi complex,

macrophages were stimulated with 100 ng/ml LPS for 12 h with the final 6 h being in the

presence of brefeldin A. Brefeldin A inhibits protein transport from the ER to the cis‐

Golgi leading to the collapse of the Golgi complex stacks [Lippincott‐Schwartz et al.,

1989]. Immunofluorescence staining of brefeldin A treated RAW264.7 macrophages

using the cis‐Golgi marker GM130 shows that, as expected, treatment with brefeldin A

disrupts the Golgi stacks in RAW264.7 macrophages (Figure 3.10B). Next, RAW264.7

cells were co‐transfected with MMP9‐mCherry and a secreted form of GFP (ssGFP), and

were cultured with LPS for 6 h in the absence or presence of brefeldin A. Cells were

analysed by flow cytometry and gated for double‐positive cells and the intracellular

median fluorescence intensities (MFI) for MMP9‐mCherry and ssGFP compared (Figure

3.10C). The secreted form of GFP (ssGFP) does not traffic through the Golgi complex and

thus should not be affected by brefeldin A (Duellman et al., 2014). Figure 3.10C confirms

that this is the case for ssGFP, but the level of intracellular MMP9‐mCherry increased

over 4‐fold compared to untreated cells. Thus, the disruption of anterograde ER‐Golgi

transport leads to a substantial accumulation in the mCherry tagged MMP9 within the

cell suggesting MMP9 must traffic through the Golgi complex en route to the cell surface.

To confirm endogenous MMP9 also traffics through the Golgi complex, the level

of endogenously secreted MMP9 was measured by ELISA in conditioned media from

macrophages that had been stimulated with LPS for 12 h and treated in the presence or

absence of brefeldin A for the final 6 h (Figure 3.10D). The secretion of endogenous

MMP9 was significantly decreased (84%; *0**P≤0.0001) in the presence of brefeldin A

compared to control cells (Figure 3.10D). Cell lysates from the same experiment were

analysed by zymography (Figure 3.10E) and by immunoblotting (Figure 3.10F). Both,


intracellular proteolytic activity (Figure 3.10E) and MMP9 protein levels (Figure 3.10F)

were strongly increased after BFA treatment macrophages compared to the control

suggesting a build‐up in the cell. Levels of non‐secreted proteins, such as actin, remain

unaffected by the BFA treatment (Figure 3.10F). These results, along with the localisation

of MMP9 to the Golgi complex, confirm newly synthesised MMP9 traffics through a Golgi

complex en route to the cell surface.

Figure 3.10. Newly synthesised MMP9 is trafficked to the cell surface via the Golgi complex. (A) RAW264.7 cells transiently expressing MMP9-mCherry (red) were fixed on coverslips, permeabilised and immunostained for the cis-Golgi marker GM130 (green). The lower panels show a zoomed in region where MMP9 and GM130 co-localise. Scale bar is 10 μm. (B) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 12 h. For the last 6 h the cells were incubated in the absence (control) or presence of 5 μg/ml brefeldin A (BFA). Cells on coverslips were fixed and stained for GM130 (green), F-actin (red) and the nuclei (blue). Scale bar is 20 μm. (C) RAW264.7 cells were co-transfected with MMP9-mCherry and a secreted form of GFP (ssGFP). Cells were stimulated with 100 ng/ml LPS for 6 h in the absence (control) or presence of 5 μg/ml brefeldin A (BFA). After BFA treatment macrophages were analysed by flow cytometry and the macrophages were gated for double-positive cells. The graph shows the mean fluorescence intensity (MFI) of ssGFP and MMP9-mCherry in the double positive cells following brefeldin A treatment relative to the control (n=1). (D) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 12 h. For the last 6 h the cells were incubated in the absence (control) or presence of 5 μg/ml brefeldin A (BFA). Total MMP9 levels in cell culture supernatants were quantified by ELISA; mean ± SEM (n=3), ****P≤0.0001 (unpaired, two-tailed Student’s t-test). (E) Cell lysates were collected and analysed by gelatin zymography (n=1) (F) Cell lysates were collected and analysed by immunoblotting for MMP9 and actin. A representative image is shown (n=3).



3.2.7 MMP9 localises to late endosomes/lysosomes

Both the endogenous and overexpressed MMP9 are located in the perinuclear

Golgi complex and as punctate structures in the periphery of the cell (Figure 3.7). Protein

cargo that is to be secreted but not directly transported to the plasma membrane can be

trafficked via other organelle structures, such as the late endosome to the cell surface

[Stow and Murray, 2013]. To test whether MMP9 is trafficked via late

endosome/lysosomal compartments en route to the cell surface in RAW264.7 cells,

macrophages were transfected with MMP9‐mCherry for 6 h and co‐stained with known

organelle markers (Figure 3.11). Lysosome‐associated membrane protein 1 (LAMP1) is

a structural glycoprotein integrated into the limiting membranes of late endosomes and

lysosomes and is a commonly used marker for these compartments [Chen et al., 1985;

Eskelinen et al., 2003]. Immunostaining of LAMP1 in cells expressing MMP9‐mCherry

show MMP9 and LAMP1 are found on the same structures (Figure 3.11). LAMP1 appears

as a punctate ring surrounding MMP9, suggesting that MMP9 is enclosed in LAMP1‐

positive membranes (see arrows in Figure 3.11A). As LAMP1 is a marker for both late

endosomes and lysosomes, cells were next stained for a late endosome marker to

determine whether MMP9 was present in both structures. Lysobisphosphatic acid

(LBPA) is a lipid that accumulates within intraluminal membranes of late endosomes but

not lysosomes [Kobayashi et al., 1999; Matsuo et al., 2004] and is commonly used as a

late endosome marker. MMP9‐mCherry is located in enlarged vesicles in the periphery

of the cell and partially colocalised with LBPA‐positive endosomes (Figure 3.11B). Again

the LBPA‐positive stained membranes often appeared to surround the MMP9 (arrows in

Figure 3.11B). These results suggest that MMP9 could be trafficked from the Golgi

complex to the late endosome and lysosome prior to secretion.


Figure 3.11. MMP9 localises to late endosomes and lysosomes. Cells expressing MMP9-mCherry (red) grown on coverslips were fixed, permeabilised and immunostained for either (A) late endosome/lysosome marker LAMP1 (green) or (B) the late endosome marker LBPA (green). The lower panels show zoomed in regions of the upper panels. Arrows indicate LAMP1- or LBPA-positive membranes surrounding the MMP9 cargo. Scale bar is 20 μm.


3.2.8 Targeting specific SNARE proteins alters MMP9 levels in conditioned medium

Having shown that MMP9 is transported through the Golgi complex but can be

found in late endosomes/lysosomes, possibly en route to the cell surface, it was next

determined which SNARE proteins might regulate this transport to the cell surface.

SNARE proteins will regulate every point of membrane fusion in the transport of MMP9

to the cell surface and can be used to map transport pathways. Each of the 38 mammalian

SNARE proteins has a precise subcellular distribution and as such regulates distinct

fusion events. These proteins regulate fusion events such as late endosome/lysosome

membrane fusion with the cell surface and thus identifying which SNAREs alter secretion

will help identify the transport pathways. Identification of the trafficking machinery

proteins involved in the exocytosis of MMP9 would not only further pinpoint the

intracellular trafficking pathways for this protein, but might also provide novel

therapeutic targets to reduce proteolytic degradation and inflammation in chronic

wounds.

Targeting surface Q-SNAREs

As each type of organelle has distinct SNARE sets that can only bind their

equivalent SNARE partner on the respective target membrane, the absence or

malfunction of just one of the participating SNAREs inhibits the fusion process.

Therefore, mutant forms or depletion of individual SNAREs block specific fusion steps

and are commonly used to map transport pathways [Stow and Murray, 2013]. To identify

possible regulators of MMP9 secretion from macrophages the four main Q‐SNAREs (Stx2,

Stx3, Stx4 and SNAP23) found on the macrophage cell surface, and most likely to mediate

secretion, were targeted for knockdown with specific siRNA (Figure 3.12). RAW264.7

macrophages were transfected with siRNA to these SNARE proteins on day 1, this was

repeated on day 2 and then cells stimulated with LPS for 15 h. Knockdown was successful

for all four tested surface SNARE proteins and it can be seen that the individual

knockdowns were specific for the targeted SNARE and did not lead to changes in the

levels of other surface non‐targeted SNAREs (Figure 3.12A). To measure the

effectiveness of the protein knockdown for each individual SNARE, the levels of the

remaining protein were quantified by densitometry of immunoblots from multiple

experiments (n=3) and compared to the amount of that particular protein in control cells

transfected with scrambled siRNA (Figure 3.12B). Stx2, Stx3 and Stx4 were significantly

reduced following knockdown, with knockdown rates of 74% (**P≤0.01), 50% (*P<0.05)


and 55% (*P<0.5) respectively. Protein levels for SNAP23 were reduced by 20%

(P>0.05).

It is expected that if a specific SNARE protein is essential for MMP9 delivery to

the cell surface then knockdown of that SNARE would lead to a reduction in secretion

and MMP9 might accumulate within the cell. Reductions in SNAREs not involved with

MMP9 delivery would be expected to have no effect on secretion. Conditioned media was

collected from cells treated with siRNA to Stx2, Stx3, Stx4 or SNAP23, as well as from

cells treated with control scrambled siRNA. The fold change in MMP9 levels were then

determined by ELISA (Figure 3.12C). None of the surface SNARE knockdowns led to a

decrease in MMP9 in the conditioned media compared to the control cells. On the

contrary, knockdown of surface SNAREs Stx2 and SNAP23 lead to about strong

enrichment of MMP9 in cell culture supernatants (*P<0.05), while knockdown of Stx3

and Stx4 had only minor effects on MMP9 levels (P>0.05) (Figure 3.13C). Partial

knockdown of SNAP23 had a very striking effect, increasing MMP9 levels in conditioned

media by almost three‐fold, although SNAP23 protein levels where only reduced by 20%

(*P<0.05). Secreted MMP9 levels were more than three times as high as the control when

Stx2 levels where reduced to a quarter of the levels in control cells (*P<0.05). Levels of

intracellular MMP9 were also measured in the same cells (Figure 3.12A) and quantified

by densitometry (Figure 3.12C). MMP9 levels remained mostly unaffected following

knockdown of Stx2, Stx3 and Stx4 with changes of less than 10% (P>0.05). However,

intracellular MMP9 levels decreased by 14% in macrophages that had reduced SNAP23.

These results together suggest that Stx2, Stx3, Stx4 and SNAP23 do not directly mediate

delivery of MMP9 to the cell surface but Stx2 and SNAP23 might regulate another process

that can alter the level of MMP9 in the media.

Targeting endosome and Golgi associated R-SNAREs

Q‐SNAREs must form complexes with R‐SNAREs to regulate fusion at each point

in the intracellular transport pathways. Five R‐SNARE proteins, VAMP2, VAMP3, VAMP4,

VAMP7 and VAMP8, which mediate fusion of distinct intracellular transport pathways,

were targeted using siRNA (Figure 3.13). RAW264.7 macrophages were transfected with

siRNA to these SNARE proteins on day 1, this was repeated on day 2 and then cells

stimulated with LPS for 15 h. As with the surface SNAREs, knockdown was successful

(38‐71%) and specific for all four tested R‐SNARE proteins (Figure 3.13A).

VAMP4 is located in the Golgi complex in RAW264.7 macrophages [Lei et al.,

2012] and known to regulate fusion both to and from the Golgi in other cell types. As


newly synthesised MMP9 destined for secretion was found to be trafficked via the Golgi

(Figure 3.10) VAMP4 was considered a potential regulator for MMP9 trafficking from the

Golgi complex. Surprisingly, siRNA knockdown of VAMP4 by 38% (**P≤0.01) increased

MMP9 levels in cell culture supernatants by 150% while reducing intracellular levels by

42% (P>0.05) (Figure 3.13C).

Apart from the Golgi complex, MMP9 was also found to localise to late

endosomal/lysosomal compartments (Figure 3.11). VAMP8 regulates homotypic fusion

of late endosomes and VAMP7 regulates late endosome to lysosome fusion [Pryor et al.,

2004]. In macrophages VAMP7 has also been shown to deliver late endosome/lysosome

membrane to the cell surface during phagocytosis [Braun et al., 2004]. In other cells

VAMP7 has been shown to regulate the delivery of exosomes from late endocytic

compartments (multivesicular bodies) to the cell surface for secretion [Fader et al.,

2009]. After siRNA treatment, VAMP7 and VAMP8 protein levels were reduced by 58%

(****P<0.0001) and 71% (****P<0.0001), respectively. Again, an enrichment of MMP9

levels, rather than a reduction, was seen following knockdown of VAMP7 or VAMP8

(Figure 3.13D). After VAMP7 knockdown, MMP9 levels in the cell culture supernatants

increased by 170% (n.s. P>0.5) and this was accompanied by a significant reduction in

intracellular MMP9 (46% (*P<0.05)). The same effect, albeit to a lesser extent, was

observed in macrophages with reduced VAMP8, where MMP9 levels in the conditioned

media increased by 78% and decreased by 37% intracellularly (P>0.5).

The reduction in secreted MMP9 seen when key SNAREs are reduced and the

localisation of MMP9 to late endosomal/lysosomal compartments could also be a result

of the internalisation of secreted MMP9, thus R‐SNAREs involved in endocytic pathways

were also investigated for a role in altering MMP9 levels. VAMP2 is located on early

endosomes in macrophages [Veale et al., 2010] and has been found to regulate endocytic

events in neuronal and HeLa cells [Miller et al., 2011; Koo et al., 2011]. Reduction of

VAMP2 levels by about 38% led to a small enrichment of MMP9 in the conditioned media

(38% (p>0.05)), while intracellular levels remained mostly unaffected. Newly made

MMP9 could also be trafficked via the recycling endosome to then be delivered to the

plasma membrane for secretion. VAMP3 is located on the recycling endosome and has

been shown to regulate the delivery of protein cargo, such as TNF and IL‐6 to cell surface

in macrophages [Manderson et al., 2007; Murray et al., 2005]. It can also regulate the

recycling of internalised cargo, such as the integrins, back to the cell surface [Veale et al.,

2010]. With only 30% VAMP3 remaining (****P<0.0001), MMP9 levels in the cell culture

supernatants increased by more than 220% and were reduced by 24% within the cell.


In summary, none of the tested SNAREs can conclusively be attributed to mediate

secretion of MMP9 in macrophages. Surprisingly, knockdown of Stx2, SNAP23, VAMP2,

VAMP3, VAMP4, VAMP7, VAMP8 but not Stx3 or Stx4 lead to an increased release of

MMP9 levels in the extracellular environment rather than reducing its secretion. The

most substantial enrichment of MMP9 (more than 2.5‐fold) in conditioned media was

detected in macrophages with reduced levels of VAMP7, Stx2 and SNAP23. However,

most of the changes observed only represent a trend and a significant effect could only

be determined for VAMP7 knockdown. This might reflect the fact that the level of SNARE

protein could not be completely ablated. Yet, the unanticipated results still suggest that

there might be other mechanisms influencing MMP9 levels in cell culture supernatants

apart from MMP9 secretion itself and that this mechanism could be mediated by these

SNARE proteins.


Figure 3.12. Knockdown of the surface Q-SNAREs Stx2 and SNAP23 but not Stx3 and Stx4, leads to an increase in MMP9 levels in cell culture supernatants. RAW264.7 macrophages were transfected with scrambled (scr), Stx2, Stx3, Stx4 or SNAP23 siRNA, re-transfected 24 h later and then stimulated with 100 ng/ml LPS for 15 h. (A) Whole cell lysates were immunoblotted for Stx2, Stx3, Stx4, SNAP23, MMP9 and also for actin as a loading control. Representative images are shown (n=3). (B) Bar graph showing the remaining SNARE protein levels relative to that present in control cells (scrambled siRNA) after SNARE knockdown as determined through densitometry of immunoblots; mean ± SEM (n=3), *P<0.05, **P≤0.01 (one-way ANOVA with Dunnett’s post-hoc test). (C) Bar graphs showing fold change in MMP9 levels in conditioned media (n=3) as determined by ELISA (red) and intracellular MMP9 levels as determined through densitometry of immunoblots (n=3) from lysates (grey) following SNARE knockdown; mean ± SEM, *P<0.05, (one-way ANOVA with Dunnett’s post-hoc test).


Figure 3.13. Knockdown of VAMP2, VAMP3, VAMP4, VAMP7 and VAMP8 increases the level of MMP9 in cell culture supernatants. RAW264.7 macrophages were transfected with scrambled (scr), VAMP2, VAMP3, VAMP4, VAMP7 or VAMP8 siRNA, re-transfected 24 h later and then stimulated with 100 ng/ml LPS for 15 h. (A) Whole cell lysates were immunoblotted for VAMP2, VAMP3, VAMP4, VAMP7, VAMP8, MMP9 and actin, representative images shown. (B) Bar graph showing the remaining SNARE protein levels relative to control (scrambled siRNA) after SNARE knockdown as determined through densitometry of immunoblots; mean ± SEM (n=3 with exception of VAMP2 where n=1), **P≤0.01, ****P≤0.0001 (one-way ANOVA with Dunnett’s post-hoc test). (C-E) Bar graphs showing fold change in MMP9 levels in conditioned media as determined by ELISA (red) and intracellular MMP9 levels as determined through densitometry of immunoblots from lysates (grey) following SNARE knockdown; mean ± SEM (n=3), *P<0.05, otherwise n.s. P>0.05 (one-way ANOVA with Dunnett’s post-hoc test).


3.3 Discussion

MMP9 has a detrimental role in non‐healing wounds but specific inhibition of the

enzyme’s activity as a therapy has been mostly unsuccessful so far, most likely due to

their broader specificity. Preventing secretion of this enzyme from macrophages, which

are major producers of MMP9 in chronic wounds, could represent a different approach

to reduce proteolytic activity in those wounds that appear to be stuck in an inflammatory

state. Thus, this chapter aimed to identify key organelles and machinery involved in the

intracellular transport and secretion of MMP9. Results showed that in macrophages

secretion of MMP9 is upregulated upon stimulation with LPS and that unlike neutrophils,

where MMP9 is stored in tertiary granules prior to secretion, the released MMP9 is newly

synthesised as needed. MMP9 was located in the Golgi complex, in late endosomes and

lysosomes in macrophages. Brefeldin A disruption of the Golgi complex confirmed MMP9

is transported through the Golgi compartment but other drug treatments showed that

microtubules and actin were not necessary for MMP9 surface delivery and secretion

under the conditions used. SiRNA knockdown of R‐SNARE proteins involved in both,

classical and non‐classical pathways, did not reveal which pathways were involved in

MMP9 secretion but instead suggest that MMP9 might also be regulated by an endocytic

mechanism making it difficult to distinguish the exocytic pathways. Inhibiting the

function of the four major surface Q‐SNARE proteins did not reduce MMP9 secretion.

This suggests that another SNARE at the cell surface is likely to be responsible for the

final delivery of membrane bound MMP9 to the cell surface or that alterations in

endocytic pathways are masking its identity.

3.3.1 Detecting MMP9

To monitor any changes in the intracellular levels and secretion of MMP9, it was

necessary to optimise assays to detect MMP9 in RAW264.7 whole cell lysates and cell

culture supernatants. Many publications use gelatin zymography to determine MMP9

levels but the cells are usually cultured in the absence of FBS due to the intrinsic

gelatinolytic activity from the bovine MMP9 in the serum. Unfortunately, removal of FBS

from the cell culture media had detrimental effects on RAW264.7 macrophages cell

growth and morphology. As FBS is used as a growth supplement in cell culture media

because of its high content of growth promoting factors, the negative influence on

macrophage cell numbers grown in the absence of FBS is not unexpected. RAW264.7

macrophages have a doubling time of about 11 h and the exhibited reduced growth rates

and cell death upon starvation has been shown previously [Sakagami et al., 2009].

Importantly, cells growing in the low serum conditions showed distinct morphological


changes. Starvation has previously been shown to induce bi‐nucleation in HeLa cells due

to cytokinesis failure [Nishimura et al., 2015] and accordingly increased numbers of bi‐

nucleated macrophages were observed when serum concentration were reduced.

Macrophages grown in low‐serum media also contained large vacuoles, most likely

autophagic vacuoles, which are another known marker for starvation [Mizushima 2007].

Immunostaining for specific autophagy markers could determine whether these

observed structures are in fact autophagic vacuoles or denote other compartments. As a

consequence, reducing FBS in the culture medium of RAW264.7 macrophages was not

an option. Unfortunately, the levels of bovine MMP9 in the FBS‐containing media were

high enough to mask any macrophage‐originated MMP9 activity in zymography assays

despite stimulation with the strong TLR agonist LPS over extended time periods to

increase MMP9 production. Thus, while MMP9 could be measured in cells lysates from

cells grown in 10% FBS, another method for detecting MMP9 in media was required.

Commercially available ELISA kits and an antibody raised against mouse MMP9

for immunoblotting were tested and were able to specifically detect the murine form of

the protein in culture media. Two ELISA kits were tested but one (R & D) was ruled out

due to what appeared to be a factor(s) secreted by the RAW264.7 cells that interfered

with the assay. This specific ‘factor’ was not an issue with the second ELISA kit

(Biosensis). However, the recovery of the MMP9 levels (Biosensis) was not 100% due to

matrix effects from the RPMI media, which means that the MMP9 concentrations

observed in cell culture supernatants do not represent an accurate quantification of

absolute MMP9. But the absolute concentrations are not necessary to determine changes

in MMP9 levels and this assay was linear across the concentration range used. It is

important to note that the ELISA kits do not allow discrimination between zymogen and

active MMP9, as it detects both of these forms. From the immunoblots of MMP9 in the

conditioned media it became clear that only the zymogen form of MMP9 is present in cell

culture supernatants. However, the ELISA kit was the more sensitive technique of the

two as MMP9 could be detected in cell culture supernatants from as early as 3 h post‐

stimulation, whereas MMP9 was only seen from 6 h post‐stimulation by immunoblotting.

Immunoblotting of cell lysates allowed for differentiation between the different

forms of MMP9. Zymography of cell lysates was also utilised in some instances to detect

intracellular MMP9 since the FBS, which was shown to mask macrophage‐originated

MMP9 activity in cell culture supernatants, is absent in those samples. However, this

technique appeared to be less sensitive and also only detected one form of MMP9. Hence,

immunoblotting for MMP9 in cell lysates and immuno‐detection of murine MMP9 in


conditioned media using a commercially available ELISA kit were considered most

suitable for the investigation of intracellular trafficking and secretion of MMP9.

Unfortunately, only a limited amount of experiments were performed with the anti‐

MMP9 polyclonal rabbit antibody as antibody from a newly purchased batch with a

different lot number was subsequently unable to detect MMP9 in conditioned media or

lysates from RAW264.7 macrophages. The batches from the old lot number were no

longer available. Some later experiments were performed using a monoclonal mouse

MMP9‐antibody as it became commercially available towards the end of the PhD and was

positively tested to detect mouse MMP9 in lysates from RAW264.7 macrophages.

However, detection of MMP9 in fixed, permeabilised cells for fluorescence microscopy

analysis using this antibody was poor. Instead, fluorescently‐tagged MMP9 protein,

which appears to locate to the same vesicles as endogenous MMP9 in RAW264.7

macrophages and allows for better detection in microscopy assays, was used for these

experiments.

3.3.2 MMP9 expression and secretion in RAW264.7 macrophages

MMP9 levels are typically low in normal tissue but greatly enhanced during the

wound healing process [Opendakker et al., 2001]. Spatio‐temporal control of MMP9

activity is crucial in order to avoid excessive tissue damage as seen in chronic wounds.

Endotoxins, tissue debris and other signalling molecules can activate macrophages to

produce cytokines, proteases and other pro‐inflammatory factors in vivo. Macrophages

are major producers of MMP9 during the inflammatory phase of the wound healing

process and studies have shown that the MMP9 gene is strongly up‐regulated in LPS‐

activated human blood monocytes and RAW264.7 macrophages in vitro leading to a

temporal upregulation of MMP9 secretion into the environment [Huang et al., 2012; Rhee

et al., 2007; Yang et al., 2015]. Accordingly, unstimulated RAW264.7 macrophages

showed only basal levels of intracellular MMP9 but levels where strongly upregulated

after stimulation with LPS, and secreted MMP9 was only detectable post‐stimulation.

The time‐dependant upregulation of MMP9 expression and secretion observed was

generally in agreement with previous studies where MMP9 secretion from LPS

stimulated RAW264.7 macrophages was detected from 8‐12 h post‐stimulation [Rhee et

al., 2007; Woo et al 2004; Yang et al., 2015]. However, MMP9 was also detected in

conditioned media from as early as 3 h post stimulation. This may be due to a higher

temporal resolution of this experiment as the first time point considered in the above

mentioned previous studies was not until 6, 8 or 12 h post‐stimulation, respectively for

the individual studies. In addition MMP9 levels were assessed using zymography or


immunoblotting techniques in these studies, whereas the analysis by ELISA is much

more sensitive and could therefore detect MMP9 at 3 h post stimulation. In fact, MMP9

secretion was only detectable from 6 h post‐stimulation when using immunoblotting

techniques for this time series confirming that detection limits can vary for different

techniques. Furthermore, the temporal patterns for MMP9 upregulation do not only

depend on the resolution of the time series and sensitivity of the assay that is used to

measure MMP9 levels but are also affected by the concentration of the LPS used to

stimulate MMP9 secretion [Yang et al., 2015]. MMP9 secretion from RAW264.7

macrophages can also be instigated through stimulation with IL‐1β or LPS in

combination with IFNγ [Hanania et al., 2012; Yoo et al., 2003]. MMP9 secretion in LPS‐

stimulated RAW264.7 has been shown to peak around 24 h post‐stimulation [Rhee et al.,

2007; Yang et al., 2015]. Future work could determine whether the upregulation of

MMP9 follows a similar temporal pattern after stimulation with these factors. A plateau

for MMP9 secretion was not observed in this study as the time series was only assessed

over an 18 h time frame but could be expected at later time points due to timely

downregulation of MMP9 expression. The later time points were not considered in this

study as sufficient levels of MMP9 were achieved for 9‐18 h LPS stimulation to be able to

measure changes in MMP9 transport and secretion.

MMP9 is synthesised as an inactive precursor (pro‐MMP9) and activation of the

zymogen can occur through cleavage of the pro‐peptide upon rebinding at the cell

surface following secretion. Hence, non‐cleaved pro‐MMP9 (zymogen) as well as

processed MMP9 (active) could be expected to be present in the conditioned media. The

ELISA results represent total MMP9 levels as the assay does not discriminate between

pro‐MMP9 and the cleaved form. The active and zymogen form are, however,

distinguishable by molecular weight. Immunoblots of conditioned media detected only

one band for MMP9, most likely, from its size, the non‐cleaved zymogen form of MMP9

(105 kDa). This indicates that processing of pro‐MMP9 to active MMP9 might not occur

under these in vitro conditions. It is known that in vitro pro‐MMP9 rapidly dissociates

again following its rebinding to the surface and this dissociating would prevent it to be

activated [Fridman et al., 2003]. Additionally, the culture media may also dilute any

soluble activators, such as plasmin/MMP3, that might be present in the cell culture

supernatant [Fridman et al., 2003]. This, together, might explain why the active form (98

kDa) cannot be detected. This is probably not representative of in vivo conditions, where

activators and MMP9 molecules are expected to be much more concentrated, particularly

within chronic wound fluid. In vivo regulation of MMP9 activity is not only limited to a


transcriptional regulation, secretion and activation but is also achieved through the

presence of endogenous activators and inhibitors, such as the TIMPs.

While short‐lived neutrophils are known to produce and prepack MMP9 in

tertiary granules for rapid release [Van den Steen et al., 2002], macrophages do not have

granules for the storage and release of pre‐made proteins. Hence, the upregulation of

MMP9 secretion might be expected to stem mostly from newly synthesised protein or

from the inhibition of MMP9 degradation. As with the majority of proteins destined to be

secreted, MMP9 has a signal‐peptide targeting it to the ER during translation [Egeblad

and Werb, 2002]. Indeed, a previous study was able to show that MMP9 co‐localises with

the ER markers protein disulfide‐isomerase (PDI) and calreticulin in classical activated

RAW264.7 macrophages [Hanania et al., 2012]. MMP9 is heavily glycosylated with both

N‐ and O‐glycosylation sites. N‐glycosylation of fully unglycosylated MMP9 (98 kDa) to a

short‐lived intermediate form (102 kDa) occurs in the ER, whereas O‐glycosylation of

the rapidly processed 102 kDa‐sized intermediate form to the fully glycosylated protein

with a molecular weight of 105 kDa form takes place in the Golgi complex [Duellman et

al., 2015; Hanania et al., 2012; Olson et al., 2000]. By immunoblotting two bands can be

seen in stimulated samples; one just below 100 kDa, most likely representing the

underglycosylated (98 kDa) form of MMP9, and a band above 100 kDa detectable in

stimulated, as well as unstimulated samples, which is likely the fully glycosylated

zymogen form of MMP9 (105 kDa) [Olson et al., 2000]. The increasing presence of the

underglycosylated form in LPS‐stimulated samples suggests that increased levels of

MMP9 in cell culture supernatants of LPS‐stimulated RAW264.7 macrophages originate

from newly made protein. This was confirmed when macrophages were treated with

cycloheximide to stop protein synthesis where reduced levels of MMP9 were seen in the

cellular and the secreted fraction. Correspondingly, the underglycosylated precursor (98

kDa) form of MMP9 that occurs during the de novo synthesis of the protein becomes

almost absent in lysates from cells treated with cycloheximide. The reduction of MMP9

levels observed for cycloheximide treated macrophages is quite substantial considering

protein synthesis was only inhibited for the last 4 h of LPS stimulation. This might be due

to the fact that MMP9 secretion over time is not linear but appears to continually

increase. It is likely that some MMP9 might be internalised and regulated by degradation,

as implied in the SNARE knockdown experiments, as the time course progresses leading

to a greater effect that anticipated when MMP9 production is inhibited at these later time

points.


3.3.3 The cytoskeleton and MMP9 secretion

The cytoskeleton can assist in the anterograde transport of cargo‐containing

vesicles for their secretion [Anitei and Hoflack, 2012]. Microtubules have previously

been reported to facilitate MMP9 secretion in RAW264.7 macrophages stimulated with

LPS and IFNγ [Hanania et al., 2012]. In this study, MMP9 localised to stable microtubules

in immunofluorescence microscopy as well as kinesin isoforms KIF5B and KIF3B

[Hanania et al., 2012]. Disruption of microtubules with 10 μM nocodazole reduced

secretion of MMP9 into cell culture supernatants. Concurrently, stabilisation of

microtubules with taxol or low‐doses of nocodazole (0.1 μM) increased secretion of

MMP9 [Hanania et al., 2012]. In RAW264.7 macrophages stimulated with LPS alone,

MMP9 was also found to distribute along microtubules. Yet, disrupting microtubule

networks with 5 μM nocodazole did not change intracellular or secreted levels of MMP9.

A lack of an effect could be due to the fact that nocodazole also causes the separation of

the Golgi complex from its central position resulting in the formation of scattered mini‐

stacks, which are nevertheless competent for maintaining protein transport and

secretion [Cole et al., 1996]. To add to this, microtubules only increase the probability

for a cargo vesicle to encounter the target membrane, which means that the secretion in

nocodazole‐treated samples is only slowed and not fully blocked [Wacker et al., 1997].

The treatment with nocodazole was present for the last 3 h of 12 h LPS stimulation

whereas treatment in the above mentioned study was for the last 3 h of only 9 h LPS

stimulation so any effect could have been too small to be detected given the large

amounts of MMP9 being secreted during the absence of nocodazole. Using a lower

nocodazole concentration might have added to this.

The actin skeleton has been reported to have both supporting as well as

hindering roles during the secretion process [Porat‐Shliom et al., 2013]. On the one hand,

actin filaments can assist in the guidance of exocytic vesicles to fusion sites. On the other

hand, actin can also act as physical barrier for any cargo trying to access the plasma

membrane [Porat‐Shliom et al., 2013]. Disruption of cortical F‐actin in RAW264.7

macrophages led to an increase of secreted MMP9 levels. This suggests that MMP9

secretion may not be dependent on transportation along actin filaments but that the

cytochalasin‐induced breakdown of cortical actin filaments allows for an easier access of

MMP9‐containing vesicles to the cell surface for release, thus explaining the slight

increase in secreted MMP9 after cytochalasin treatment detected by ELISA. Yet, there

were no detectable changes by immunoblot. Cytochalasin treatment of macrophages was

only performed for the last 2 h of the 12 h LPS‐stimulation period so changes in


intracellular MMP9 levels might be too small to be detected with less sensitive

techniques applied as immunoblotting and zymography.

3.3.4 Trafficking pathways for MMP9 in RAW264.7 macrophages

MMP9 in the Golgi

Intracellular trafficking pathways responsible for the release of newly made

MMP9 were also investigated. Following translation into the ER, most secreted proteins

are then shuttled through the Golgi apparatus. Co‐localisation with GM130, a known

Golgi organelle marker, indicated that newly made MMP9 is trafficked via the Golgi.

Treatment of macrophages with brefeldin A causes blockage of anterograde traffic from

the ER to the Golgi complex [Lippincott‐Schwartz et al., 1989]. The significantly reduced

secretion seen after drug treatment verified that MMP9 is trafficked via this organelle en

route to the cell surface. This blockage led to the accumulation of intracellular MMP9,

seen by both immunoblotting and zymography. The intracellular accumulated MMP9

was approximately 100 kDa, slightly lower than the fully glycosylated MMP9 (105 kDa)

and higher than the unglycosylated form of MMP9 (98 kDa) seen in control samples. This

intermediate form of MMP9 has previously been reported as a short‐lived

underglycosylated MMP9 version of the size of 102 kDa. MMP9 (98 kDa) is N‐

glycosylated in the ER to produce a 102 kDa form which is usually short‐lived and

undetectable, as its transport to the Golgi complex leads to MMP9 being rapidly O‐

glycosylated to produce the 105 kDa MMP9 form [Duellman et al., 2015; Hanania et al.,

2012; Olson et al., 2000]. The accumulation of this band is consistent with the disruption

of the anterograde transport from the ER to the Golgi complex. Interestingly, the

accumulated underglycosylated 102 kDa form of MMP9 but not the completely

unglycosylated form of MMP9 (98 kDa) is able to degrade gelatin suggesting that N‐

glycosylation alone is necessary for the proteolytic mechanism. N‐glycosylation of MMP9

is also known to be crucial for efficient secretion [Duellman et al., 2015a; Duellman et al.,

2015b].

MMP9 in late endosomes/lysosomes

Secretion of MMP9 has previously been demonstrated to be dependent on

Rab3D, a trafficking machinery protein regulating exocytosis from the TGN, in

macrophages [Hanania et al., 2012; Millar et al., 2002; Pavlos et al., 2005]. But the exact

pathway Rab3D regulates in macrophages is unclear. Proteins destined for secretion can

be trafficked via other organelles such as late endosomes/lysosomes, which can function

as secretory compartments releasing their contents upon fusion with the cell surface.


However, in osteoclasts, which can be differentiated from macrophages, the Rab3D

pathway is non‐endosomal/lysosomal [Pavlos et al., 2005]. To determine if MMP9

traffics through lysosome it was tested whether MMP9 would co‐localise with known

markers for late endosomes/lysosomes in RAW264.7 macrophages. Surprisingly, MMP9

was found to co‐localise with both, late endosomal marker LBPA and late

endosome/lysosome maker LAMP1. While most LAMP1‐positive vesicles contained

MMP9 some punctate non‐perinuclear MMP9 staining could be seen that did not co‐

localise with LPBA. These staining patterns therefore suggest that MMP9 locates to both,

lysosomal and late endosomal compartments. The localisation of MMP9 to late

endosomal/lysosomal compartments in macrophages could suggest that newly made

MMP9 is trafficked via these organelles for secretion. The lysosomal trafficking

machinery protein Rab27a does regulate MMP9 secretion in neutrophils [Brezinska et

al., 2005]. However, this represents the release of stored pools of MMP9 from granules,

which do not exist in macrophages. It has been shown using RAW264.7 macrophages

stimulated with LPS and IFNγ that there is very low recruitment of endogenous MMP9

to LAMP1‐positive vesicles [Hanania et al., 2012]. This might be due to the fact that

quantification of the co‐localisation between MMP9 and LAMP1 might only lead to a very

small overlap as LAMP1 is only located in the limiting membrane of lysosomes. Another

reason why MMP9 would not have been located to LAMP1‐positive compartments in that

study could be that activation of RAW264.7 macrophages was achieved through

simultaneous stimulation with LPS and IFNγ. Intracellular MMP9 levels were found to

peak at 9 h following this stimulation regime [Hanania et al., 2012]. When RAW264.7

macrophages are stimulated with LPS alone intracellular MMP9 levels do not decrease

until after 18 h of treatment [Woo et al., 2004]. This represents a significant difference

in the temporal expression profile of MMP9. Co‐localisation studies in LPS/IFNγ‐

activated macrophages were performed in cells that were fixed 9 h post‐stimulation

when expression levels begin to decline, which could affect the distribution of the protein

within the cell [Hanania et al., 2012]. Although the co‐localisation studies herein were

achieved through transient overexpression of fluorescently labelled MMP9, alterations

to trafficking pathways as a result of protein overexpression is not likely to be an

explanation for any observed inconsistencies as MMP9‐mCherry localises to the same

vesicles as the endogenous protein. Overexpression of proteins can sometimes lead to

aggregation and clearance of these ‘aggresomes’ is facilitated via lysosomal mediated

degradation [Ovádi and Orosz, 2008]. However, these structures would show a much

brighter fluorescence signal and would expect to be LBPA‐negative. Together, this


supports the data herein that MMP9 can indeed locate to late endosomal/lysosomal

structures in macrophages.

MMP9 in late endosomal compartments could also represent pools of previously

released MMP9 that had been endocytosed again for clearance. Endocytosis of secreted

MMP9 would represent another level for the regulation of its activity and it is known that

MMP9 is actively endocytosed. MMP9 is a ligand of low density lipoprotein receptor‐

related protein 1 (LRP‐1) and binding of MMP9 to LRP‐1 on the plasma membrane

mediates its endocytosis leading to degradation of the protease [Hahn‐Dantona et al.,

2001; Van den Steen et al., 2006]. Degradation of MMP9 as a consequence of this process

would take place within the lysosome and trafficking to the lysosomal compartment

from the early endosome after endocytosis would occur via the late endosome.

Therefore, some or all of the MMP9 found in any of these structures could be a result of

this clearance mechanism. The increase in MMP9 in the media when specific SNARE

proteins are inhibited might suggest that not only do they regulate this degradation

pathway but also that MMP9 downregulation can occur by this pathway in macrophages.

More work would be needed to determine whether trafficking of MMP9 via these

organelles is necessary for its secretion or its degradation.

Targeting SNARE proteins

The delivery of MMP9 within membrane‐bound compartments to the cell surface

for secretion requires membrane fusion mediated by SNARE proteins at all points in its

trafficking pathway. To identify trafficking machinery proteins that could be involved in

the trafficking of MMP9 to the cell surface for secretion, individual protein levels of a

range of SNARE proteins were reduced through targeted siRNA knockdowns. Differences

in knockdown efficiency between the individual tested SNAREs were seen and might be

due to altered stability or half‐life for the individual SNARE protein. For some SNAREs,

such as SNAP23, the average knockdown efficiency was as low as 20%. This is owed to

the fact that SNAREs play a crucial role for many cellular and biological processes. This

is also the reason why knockout mice for most SNAREs, e.g. SNAP23, are lethal and hence

unavailable [Suh et al., 2011]. It was also found that the level of knockdown of any one

SNARE was variable between biological replicates, occasionally to a large extent, which

ultimately also had an effect on the extent of any occurring changes in MMP9 levels. The

major surface Q‐SNAREs, Qa‐SNAREs Stx2, Stx3, Stx4 and Qbc‐SNARE SNAP23, which

could potentially mediate the final secretion trafficking step at the plasma membrane

were tested by siRNA knockdown for their involvement in MMP9 secretion. Disrupting

intracellular trafficking pathways by reducing the levels of a key SNAREs involved in


fusion of vesicles with the cell surface should lead to a reduction in secreted MMP9 and

a build‐up of intracellular MMP9. Surprisingly, a three‐fold increase in secreted MMP9

was observed for macrophages treated with siRNA to Stx2 or SNAP23. This unexpected

result suggests that perhaps an additional mechanisms, such as endocytosis, might also

influence the levels of extracellular MMP9. Other SNARE proteins that might be found at

the plasma membrane should in future be tested for a role in MMP9 secretion.

Q‐SNAREs must form complexes with R‐SNAREs to regulate fusion at each point

in the intracellular transport pathways. Multiple R‐SNAREs from different pathways in

macrophages were also tested for their involvement in MMP9 trafficking. At first, VAMP4

was considered a potential regulator for MMP9 trafficking. Trafficking of MMP9 via the

Golgi complex was shown to be essential for MMP9 secretion (Figure 3.10) and VAMP4

can be found in the Golgi complex in RAW264.7 macrophages [Lei et al., 2012]. MMP9 is

also known to be sorted into vesicles positive for VAMP4 in breast cancer cells [Jacob et

al., 2013]. Surprisingly, siRNA knockdown of VAMP4 increased MMP9 levels in cell

culture supernatants. This suggests that VAMP4 is not necessary for MMP9 secretion.

Yet, it does seem to affect MMP9 levels through another mechanism. VAMP4 can regulate

retrograde transport as well as anterograde transport from the Golgi complex and it can

be found on early endosomes [Mallard et al., 2002]. Thus, the knockdown of this SNARE

could potentially have disrupted endocytosis of MMP9 rather than its secretion, leading

to the increase MMP9 seen in the media. The requirement of other Golgi complex SNARE

machinery, such as Stx6, in MMP9 release should in future be investigated.

Apart from the Golgi complex, MMP9 was also found to localise to late

endosomal/lysosomal compartments. As these compartments can have secretory

functions in macrophages this could suggest that MMP9 is trafficked to the cell surface

via these organelles or degraded via these organelles. To further characterise the role of

these compartments for intracellular trafficking of MMP9 the function of the late

endosome/lysosome SNAREs VAMP7 and VAMP8 were attenuated in RAW264.7

macrophages using siRNA. Knockdown of VAMP7 and VAMP8 led to the enrichment of

MMP9 in cell culture supernatants with a more significant effect observed in cells with

reduced VAMP7. This suggests that neither of these two SNAREs is involved in secretion

of MMP9 from a late endosomal/lysosomal compartment but nonetheless appear to

regulate extracellular MMP9 levels. VAMP7 regulates late endosome to lysosome fusion

[Pryor et al., 2004] and could hypothetically mediate the delivery of MMP9 cargo to the

lysosome for its degradation. The localisation of MMP9 to late endosomal/lysosomal

compartments could be indicative of a pool of MMP9 to be targeted for degradation. It is


known that MMP9 levels can regulated through endocytic events, which would also lead

to accumulation of this protein within late endosomes and lysosomes.

To allow for a more comprehensive investigation of the intracellular trafficking

routes of MMP9 in macrophages machinery involved in recycling pathways were also

considered. If a substantial amount of MMP9 is in fact endocytosed in macrophages pools

of the protein would locate to early endosomes. Co‐localisation studies with an early

endosome marker such as Early Endosome Antigen 1 (EEA1) could add valuable insights.

VAMP2 locates to early endosomes in macrophages [Veale et al., 2010] and has been

found to regulate endocytosis in other cell types [Miller et al., 2011; Koo et al., 2011]. Yet,

changes in MMP9 levels following VAMP2 knockdown were only minimal in comparison

to other tested SNARE proteins. Other SNAREs that could mediate endocytosis in

macrophages should be further investigated.

Endocytosed surface cargo that is not targeted for degradation can be delivered

to the cell surface via the recycling endosome, which is known to be the case for integrin

α5β1 [Caswell et al., 2008; Jones et al., 2006; Powelka et al., 2004]. The recycling of

integrin α5β1 to the cell surface is regulated by VAMP3 in macrophages [Veale et al.,

2010; Veale et al., 2011]. At the same time, VAMP3 has also been shown to regulate the

delivery of newly made cargo such as TNF and IL‐6 from the Golgi complex to the cell

surface via the recycling endosome in macrophages and so VAMP3 can function in

endocytic and exocytic pathways [Manderson et al., 2007; Murray et al., 2005].

Knockdown of VAMP3 led to increased MMP9 levels in cell culture supernatants,

suggesting that it is most likely not regulating delivery of newly made or recycled MMP9

to the cell surface.

Extracellular MMP9 levels can be regulated through endocytosis by binding to

LRP‐1 at the cell surface [Hahn‐Dantona et al., 2001; Van den Steen et al., 2006]. Based

on LRP‐1‐dependant uptake of other proteins, MMP9 is expected to remain associated

with LRP‐1 in early endosomes facilitating the transit of MMP9 from early to late

endosomes where the complex dissociates [Laatsch et al., 2012]. MMP9 would then be

targeted to the lysosome for degradation while LRP‐1 may be recycled back from late

endosomes to the plasma membrane [Laatsch et al., 2012]. The machinery involved in

the recycling of LRP‐1 is unknown but it is possible that some or all of the SNAREs that

were tested and led to an increase in extracellular MMP9 levels following knockdown are

involved in the LRP‐1 recycling process and hence influencing MMP9 levels through

altering the cells ability to recycle the receptor involved in MMP9 endocytosis


(Figure 3.14). VAMP4 and VAMP3 could be involved in the cell surface delivery of

unloaded LRP‐1, which could be either newly made originating from the Golgi or recycled

from the plasma membrane. If this were true then knockdown of VAMP4 or VAMP3 could

potentially lead to less LRP‐1 being available at the cell surface for the binding and

uptake of secreted MMP9, resulting in more MMP9 being present in cell culture

supernatants. This is one potential explanation for the results seen when the level of

certain SNARE proteins are reduced. When normal levels of LRP‐1 exist at the cell

surface, endocytosis of MMP9 associated with LRP‐1 could be mediated by the early

endosome VAMP2, whose knockdown lead to increased extracellular MMP9. After

internalisation of the LRP‐1‐MMP9‐complex MMP9 would then be delivered to late

endosomes and lysosome resulting in MMP9 degradation and the delivery of LRP‐1 to

the cell surface to repeat the process. These final stages could be mediated by VAMP7

and/or VAMP8 and the results provided herein are consistent with this, although further

experimentation would be required to confirm this theory.

In summary, knockdown of the SNAREs VAMP3, VAMP4, VAMP7, VAMP8, Stx2

and SNAP23 increased the level of extracellular MMP9. This suggests rather than

mediating secretion of newly made MMP9 these SNAREs appear to influence other

pathways that can alter the level of MMP9 in the cell culture supernatant. Of course, other

SNAREs that have not been investigated could also be tested for their involvement in

MMP9 trafficking and for their potential to reduce MMP9 secretion. If MMP9 activity

cannot be reduced through the inhibition of the enzyme’s secretion it is possible that

regulating MMP9 activation through inhibiting surface delivery of components of the

MMP14/MMP2/TIMP2 complex, which activates the zymogen at the cell surface, could

be used to target its extracellular activity [Itoh et al., 2001; Itoh and Seiki, 2004; Toth,

2003]. The intracellular trafficking and protein machinery involved in the cell surface

delivery of MMP14 is the subject of investigation of the next chapter.


Figure 3.14. Hypothetical schematic of LRP-1-mediated endocytosis of MMP9. MMP9 is secreted after transiting the Golgi complex via an unidentified pathway. VAMP4 and VAMP3 could be involved in the cell surface delivery of LRP-1, which could be either newly made originating from the Golgi (blue arrows) or recycled from the plasma membrane (purple arrows). At the cell surface LRP-1 binds secreted MMP9. Endocytosis of MMP9 associated to LRP-1 could be mediated by VAMP2 or VAMP4. Upon successful endocytosis of the LRP-1-MMP9-complex the subsequent delivery to late endosomes and disassociation of the complex therein occurs. Targeting of MMP9 cargo to the lysosome would result in its degradation while now unloaded LRP-1 is recycled back to the cell surface. Transport through these latter stages could be mediated by VAMP7 and/or VAMP8. A hypothetical VAMP7/Stx2/SNAP23 complex could possibly be a regulator of recycling of unloaded LRP-1 from the late endosome to the cell surface but this requires further investigation.

Chapter 4: Identifying intracellular trafficking pathways for cell surface delivery of MMP14 101

Chapter 4: Identifying intracellular trafficking pathways for cell surface delivery of MMP14

4.1 Introduction

MMP14 belongs to the group of membrane anchored MMPs and is active at the

cell surface. As such, MMP14 is not found in chronic wound fluid but its expression

appears to be highly elevated in biopsies of chronic wounds, where they contribute to

enhanced matrix turnover [Norgauer et al., 2002]. As part of a complex with TIMP2 and

MMP2, MMP14 also regulates the activation of MMP9, which has a detrimental role in

non‐healing wounds as it further facilitates collagen breakdown [Itoh et al., 2001; Itoh

and Seiki, 2004; Toth 2003]. MMP14 is further thought to play a role in macrophage

infiltration of wound tissue [Matias‐Roman et al., 2006; Verollet et al., 2001; Wiesner et

al., 2014]. As high numbers of macrophages contribute to prolonged inflammation and

wound chronicity it would be advantageous to elucidate the pathways for MMP14

surface delivery in macrophages and its influence on macrophage invasion.

Understanding how MMP14 is delivered to the cell surface might provide new

therapeutic targets to reduce inflammation and improve wound healing.

In cancer cells, two pathways have been reported for MMP14 transport from the

Golgi complex to the cell surface: one pathway is via the recycling endosome [Bravo‐

Cordero et al., 2007; Kean et al., 2015 Remacle et al., 2005] and the other via a late

endosome/lysosome pathway [Loskutov et al., 2014; Macpherson et al., 2014; Marchesin

et al., 2015; Monteiro et al., 2013; Steffen et al., 2008; Williams & Coppolino, 2011;

Williams et al., 2014; Yu et al., 2012]. However, these studies represent recycling

pathways for MMP14 that has been delivered to the surface, endocytosed and recycled

back to the cell surface. This recycling mechanism is the major pathway for controlling

surface MMP14 levels in these cell types. In macrophages MMP14 has been shown to be

upregulated 13 fold at the RNA level after LPS activation. Thus, it is expected that the

majority of MMP14 that is delivered to the cell surface is newly made in macrophages.

Like other transmembrane proteins possessing a signal peptide, MMP14 is

translated in the ER and it is then transported to the Golgi complex. Transport through

the Golgi complex has been confirmed as essential for processing of pro‐MMP14 to the

102 Chapter 4: Identifying intracellular trafficking pathways for cell surface delivery of MMP14

mature form in the TGN, which is a requisite for successful incorporation into the plasma

membrane [Mazzone et al., 2004]. From the Golgi complex, it is unclear what pathway is

used to transport MMP14 to the cell surface in macrophages for them to be able to

regulate inflammation. To date there is only one study that investigated trafficking

pathways of MMP14 in macrophages. This study looked at Rab proteins that regulate

MMP14 trafficking to the cell surface [Wiesner et al., 2013]. Rab proteins, like the

SNAREs, typically regulate distinct transport steps within the cell. This macrophage

study, mainly using overexpressed MMP14, suggested that MMP14 delivery to the cell

surface from the Golgi complex occurs in a Rab8 dependent manner. Although the exact

route is unclear, the authors suggest MMP14 might be transported directly in Rab8a

positive vesicles or via the recycling endosome, in a Rab8a dependent manner, to the cell

surface [Wiesner et al., 2013]. However, no overlap of MMP14 with Rab11, which is

typically located on the recycling endosome, could be seen [Wiesner et al., 2013]. Rabs

that regulate the late endosome/lysosome (Rab7 and Rab9) were not tested in this study.

Typically, monocytes/macrophages entering sites of injury and inflammation

become activated en route to the site or on entering it. This activation alters the function

of the macrophage, such as it might switch on the production and secretion of distinct

cytokines and enzymes. Therefore, this chapter begins to establish MMP14 expression

and cell surface levels in response to macrophage activation. Next, the subcellular

localisation of MMP14 in macrophages is defined to pinpoint and characterise

intracellular compartments that form part of the route for MMP14 exocytosis in these

cells. Finally, levels of candidate SNARE proteins are reduced by siRNA to identify

trafficking pathways and machinery that regulate incorporation of MMP14 into the

plasma membrane.


4.2 Results

4.2.1 MMP14 levels increase upon stimulation of RAW264.7 macrophages

MMP14 expression has been shown to be upregulated at the RNA level in human

and mouse monocytes/macrophages following 2, 6, 18 h or 24 h of LPS stimulation [Hald

et al., 2012; Ming et al., 2012; Murray et al., 2013; Reel et al., 2011] and is upregulated at

the protein level in human macrophages after 24 h of IFN and LPS activation [Johnson et

al., 2014]. In order to determine the route by which MMP14 is delivered to the cell

surface after activation the timing of its upregulation and delivery first needed to be

characterised. Hence, to determine the time course for MMP14 protein upregulation,

RAW264.7 macrophages were stimulated with 100 ng/ml LPS over an 18 h time course

and cell lysates were analysed for MMP14 protein by immunoblotting (Figure 4.1A).

MMP14 is initially synthesised as a zymogen (66 kDa) with the pro‐peptide covering the

active site that then requires conversion to the mature active form (54 kDa). This process

occurs intracellularly through the cleavage of the pro‐peptide by furin after exit from the

Golgi complex but before reaching the cell surface [Mazzone et al., 2004]. Figure 4.1

shows that an MMP14 band around 54 kDa was observed in cell lysates of both

stimulated and unstimulated macrophages, which most likely represents the mature

processed form of MMP14. Post LPS activation the level of MMP14 was found to

continually increase over time (Figure 4.1A). Quantification of MMP14 levels from

multiple experiments (n=5) revealed that its protein expression is increased 2.7‐fold

(±0.63) at its peak around 15 h post‐stimulation. Thus, LPS upregulates MMP14

expression in RAW264.7 macrophages, with levels peaking around 15 h after LPS

activation.

Unlike MMP9, MMP14 is not secreted but is a membrane‐bound MMP and

contains a trans‐membrane domain. Therefore, the protein levels detected in whole cell

lysates represent total MMP14, which includes the intracellular pools, as well as any

MMP14 anchored in the plasma membrane. As such, discriminating between these two

fractions is not possible by immunoblotting of whole cell lysates. Hence, the distribution

of MMP14 in stimulated and unstimulated RAW264.7 macrophages was assessed by

immunofluorescence microscopy to compare surface and internal MMP14 (Figure 4.1B).

Macrophages were stimulated with LPS, fixed at the indicated time points, permeabilised

and immunostained for MMP14. In agreement with the immunoblotting results,

unstimulated macrophages (0 h) showed low basal levels of MMP14 expression and the

levels were markedly upregulated following LPS stimulation (Figure 4.1B). The MMP14

fluorescence signal was most intense in regions around the nucleus and some surface


MMP14 was also apparent in LPS stimulated macrophages, although the ratio of the level

of MMP14 in the perinuclear region reduced in comparison to that in the periphery over

the time course (Figure 4.1B). This suggests that LPS stimulates the production of

nascent MMP14 and its delivery to the cell surface.


Figure 4.1. MMP14 levels increase upon stimulation of murine macrophages. (A) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for the indicated times. Following LPS stimulation, cell lysates were collected and analysed by immunoblotting for MMP14 and actin. A representative immunoblot is shown. Bar graph shows the increase of total MMP14 levels relative to the loading control (actin) over time; n=5, mean ± SEM, n.s. P.0.05 (one-way ANOVA with Tukey’s post-hoc test). (B) RAW264.7 macrophages grown on coverslips were LPS stimulated for 0, 9 and 18 h, fixed at the indicated time points and immunostained for total MMP14 (permeabilised cells) following LPS stimulation. Scale bar is 20 μm.


4.2.2 MMP14 is incorporated into the plasma membrane upon stimulation

Due to the intensity of MMP14 intracellular staining it is difficult to clearly

distinguish the cell surface MMP14 pool by fluorescence microscopy, as the surface

levels are typically lower. In order to better differentiate this surface pool an assay was

developed to look at only surface MMP14, where cells were immunostained live on ice

for MMP14 prior to fixation. With the membrane integrity remaining intact only the

extracellular cell surface MMP14 is then accessible for antibody labelling, which

recognises the extracellular domain, and cells can then be fixed and co‐stained as

appropriate. Performing the assay on ice prevents any MMP14‐antibody complexes

being internalised as endocytosis is halted at 4°C. Firstly, RAW264.7 macrophages grown

on coverslips were stimulated with LPS over an 18 h time course and cells were

immunostained live for surface MMP14, as described above. After immunofluorescence

labelling of this surface MMP14 pool, these cells can be analysed by microscopy and flow

cytometry (Figure 4.2). Only very low levels of surface staining can be seen on

unstimulated cells but upon stimulation with LPS there is an increase in surface MMP14

that can be observed from around 6 h post‐stimulation (Figure 4.2A). Surface MMP14

continued to increase over the remaining 12 h. It is important to note that fluorescence

intensities varied between cells. This is not unusual for macrophages, for example, when

looking at TNF post LPS stimulation, typically around 70‐80% of the cells express TNF at

2 h and levels vary in those cells that do express TNF [Murray et al., 2005b]. In addition,

not every cell appeared to have increased levels of MMP14 at the cell surface following

LPS stimulation.

Next, LPS stimulated macrophages were immunostained live for MMP14 and the

level of MMP14 at the cell surface quantified over an 18 h time course using flow

cytometry. Analysis showed a strong upregulation of MMP14 cell surface delivery in

response to LPS activation (Figure 4.2B). Again, it is clear from the histogram that

MMP14 surface levels after LPS activation varied from cell to cell (Figure 4.2B), which is

consistent with the observations made by immunofluorescence microscopy. Less than

20% of macrophages (0 h) were positive for surface MMP14 in the absence of LPS. This

number increased dramatically to around 70% of cells being positive for surface MMP14

following LPS stimulation (9 or 18 h), representing a significant increase of 3.5‐fold

(**P<0.01) (Figure 4.2C). More strikingly, median fluorescence intensities (MFI) of

surface MMP14 for the total cell population were increased by 8.8‐fold on macrophages

stimulated with LPS for 9 h and 10.3‐fold on those stimulated for 18 h (Figure 4.2D).

These results confirm that although basal levels of MMP14 are present in unstimulated


macrophages, and most cells do not have MMP14 at the cell surface, their activation with

LPS strongly induces the delivery of MMP14 to the cell surface for incorporation into the

plasma membrane. This together with the results in Figure 4.1 showing total MMP14

levels are increased suggesting that MMP14 at the cell surface might stem mainly from

newly made protein.

Levels of MMP14 are not only influenced by the presence of endotoxins or

cytokines. While low levels of surface MMP14 due to increased MMP14 recycling have

been observed for breast adenocarcinoma cells that had been seeded on glass, seeding

on collagen abolished endocytosis of MMP14 and increased its levels at the cell surface

[Bravo‐Cordero et al., 2007; Lafleur et al., 2006]. A general increase in MMP14

expression was also reported for human peripheral blood mononuclear cells seeded onto

fibronectin but surface levels of the protein were not assessed [Jacob et al., 2002; Reel et

al., 2011]. To investigate the effect of fibronectin on MMP14 surface levels, RAW264.7

macrophages were seeded onto coverslips that had been coated with 0, 5 or 10 μg

fibronectin and unpermeabilised cells were immunostained for MMP14 after 18 h of

culture (Figure 4.2E). Only very low levels of MMP14 were detected at the cell surface of

macrophages that were grown in the absence of fibronectin. When macrophages were

grown on fibronectin, MMP14 surface levels were greatly enhanced in comparison to

macrophages grown on uncoated glass coverslips. MMP14 surface levels of macrophages

grown on 10 μg fibronectin appeared slightly higher than cells that were seeded onto

lower (5 μg) concentrations of fibronectin. The effect of growing RAW264.7

macrophages on coverslips coated with 200 μg gelatin, the hydrolysed form of collagen,

for 18 h on MMP14 surface levels was also investigated. Figure 4.2F shows that contact

with gelatin in the absence of LPS stimulus leads to MMP14 surface levels that are equal

to those on macrophages that had been stimulated with LPS. Together these results

suggest that matrix components will also affect the surface levels of MMP14.



Figure 4.2. MMP14 is incorporated into the plasma membrane upon stimulation. (A) RAW264.7 macrophages grown on coverslips were stimulated with 100 ng/ml LPS for the indicated times. Macrophages were immunostained live on ice for surface MMP14 (unpermeabilised cells) and then fixed. Scale bar is 20 μm. (B) Macrophages stimulated for 0, 9 and 18 h with LPS were immunostained live (unpermeabilised cells) with anti-MMP14 antibody or with an isotype control antibody, fixed and then analysed using the FACSAriaII flow cytometer. The overlaid histogram shows surface MMP14-immunostained cell populations over time following LPS stimulation in comparison to the control population immunostained with an isotype antibody. (C) Bar graph shows percentage of cells positive for surface MMP14 or (D) fold change of relative MMP14 surface levels (MFI), n=3; mean± SEM, **P≤0.01 (one-way ANOVA with Tukey’s post-hoc test). (D) RAW264.7 macrophages grown on coverslips coated with different concentrations of fibronectin for 18 h were immunostained live (unpermeabilised cells) for surface MMP14. (E) RAW264.7 macrophages grown on gelatin-coated coverslips (200 μg) for 18 h in the absence or presence of LPS were immunostained live (unpermeabilised cells) for surface MMP14.

4.2.3 MMP14 incorporated into the plasma membrane is newly synthesised upon LPS stimulation

To verify that the increased surface MMP14 represents mostly de novo protein,

RAW264.7 macrophages were LPS stimulated for 12 h and treated in the presence or

absence of 1 μg/ml cycloheximide for the final 4 h to inhibit protein synthesis. Cell lysates

were then immunoblotted for MMP14 to determine the effect of inhibiting protein

synthesis on total MMP14 levels (Figure 4.3A). Overall MMP14 levels were strongly

reduced following cycloheximide treatment (Figure 4.3A). Next, cells treated with

cycloheximide were immunostained live for surface MMP14 and imaged by fluorescence

microscopy (Figure 4.3B). The cycloheximide‐treated cells appeared elongated in

comparison to control cells and had very low levels of surface MMP14 in comparison to

control cells, suggesting the pool at the cell surface after LPS activation is newly

synthesised MMP14 (Figure 4.3B). The change in MMP14 surface levels was quantified

by flow cytometry (Figure 4.3C). The population of cycloheximide‐treated macrophages

showed a strong shift to the left (indicating lower levels) on the histograms for surface

MMP14 signal (Figure 4.3C). While some cycloheximide‐treated cells had residual levels

of surface MMP14, an increasing portion of the population possessed no MMP14 at the

surface at all (Figure 4.3C). The number of macrophage positive for surface MMP14 was

reduced from 62% (control) to 38% after cycloheximide treatment (*P≤0.05) (Figure

4.3D). When the surface MMP14 median fluorescence intensity was analysed, the overall

levels were found to be reduced by more than half (**P≤0.01) (Figure 4.3E). This

confirms that the majority of MMP14 into the plasma membrane after LPS activation

requires the de novo synthesis of the protein.


Figure 4.3. MMP14 incorporated into the plasma membrane upon LPS stimulation is newly synthesised. (A) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 12 h and treated with cycloheximide (CHX) for the last 4 h to stop protein synthesis. Cell lysates were collected and were immunoblotted for MMP14 and actin. (B) RAW264.7 macrophages grown on coverslips, stimulated with 100 ng/ml LPS for 9 h and treated with cycloheximide (CHX) for the last 4 h to stop protein synthesis were immunostained live (unpermeabilised cells) for surface MMP14 and fixed. Scale bar is 20 μm. (C) Macrophages stimulated with 100 ng/ml LPS for 9 hours and treated with cycloheximide (CHX) for the last 4 h to stop protein synthesis were immunostained live (unpermeabilised cells) with anti-MMP14 antibody or with an isotype control antibody, fixed and analysed using a FACSAriaIII flow cytometer. The overlaid histograms show surface MMP14-immunostaining for cycloheximide treated and control cells. (D) Bar graph shows the percentage of cells positive for surface MMP14, n=3; mean± SEM, *P<0.05 (unpaired, one-tailed Student’s t test). (E) Bar graph shows the relative MMP14 surface levels (MFI), n=3; mean± SEM, **P≤0.01 (unpaired, one-tailed Student’s t test).


4.2.4 Live imaging of MMP14 transport in macrophages

To visualise live the dynamics of intracellular transport and MMP14 surface

delivery two different constructs encoding MMP14, one with a GFP tag and the other

with a mCherry tag, were sourced. RAW264.7 macrophages were transiently transfected

for 6 h with either MMP14‐GFP or MMP14‐mCherry and lysates from these cells were

immunoblotted for MMP14 (Figure 4.4A). A band can be seen around 55 kDa in the

transfected and control cells, most likely representing endogenous MMP14 (Figure

4.4A). In lysates of cells transfected with MMP14‐GFP a strong higher molecular weight

band was detected between the 75 and 100 kDa markers consistent with it being MMP14

(54 kDa) tagged with GFP (27 kDa) (lane 2, Figure 4.4A). An additional band of a similar

size was also observed in lysates from MMP14‐mCherry‐transfected cells signifying

MMP14 fused to mCherry (28 kDa) (lane 3, Figure 4.4A). However, its expression level

was much lower than MMP14‐GFP. It was therefore decided that MMP14‐GFP was more

suitable for monitoring trafficking events in macrophages.

To determine the ideal time point for live imaging experiments RAW264.7

macrophages were transfected with the MMP14‐GFP construct and fixed at 0, 3, 6 or 9 h

post transfection and MMP14‐GFP expression examined by fluorescence microscopy

(Figure 4.4B). Low levels of MMP14‐GFP fluorescence were noticeable from 3 h but

MMP14‐GFP fluorescence was most intense between 6 and 9 h post‐transfection. The

majority of tagged‐MMP14 localised to perinuclear structures but it could also be found

in punctate structures in the cell periphery (Figure 4.4.B). It was thus decided to use 6 h

post transfection as a suitable time point to examine intracellular transport of MMP14‐

GFP.

To record the dynamic movement of MMP14‐GFP containing vesicles by live cell

imaging, images of macrophages were acquired at rates of 1 frame every 1 s over 1 min

(Movie 1 and Figure 4.4C). Tracking of individual vesicles shows directed transport of

MMP14 from the perinuclear region to the periphery of the macrophage (Movie 1 and

Figure 4.4C). Directionality was measured by comparing the displacement (Euclidean

distance) and track length (accumulated distance), where a directionality of 1 equals a

straight‐line from start to endpoint and is usually indicative that direction of movement

is coupled to an external cue. Movement of MMP14 vesicles was not random but showed

high directionality (0.93±0.01). The average velocity of these moving vesicles was

0.65±0.16 μm/s. Tracking these long‐distance movements showed MMP14 appeared to

follow a direct path towards the cell surface. Together this suggests that the transport of

MMP14 cargo could potentially involve the cytoskeleton for a polarised delivery.



Figure 4.4. Fluorescently-labelled MMP14 can be transiently overexpressed to monitor MMP14 trafficking in live macrophages. (A) Cell lysates from macrophages transfected with the MMP14-GFP or MMP14-mCherry and control cells were immunoblotted for MMP14 and actin. (B) Cells grown on coverslips were transiently transfected with MMP14-GFP and fixed at 0, 3, 6 and 9 h post transfection. MMP14-GFP fluorescence intensity, as well as distribution, were assessed by microscopy (grey scale). Right hand panel shows the pseudo-colouring using an intensity-coded palette where black is low and white is high. Scale bar is 20 μm. (C) Cells grown on Matek dishes, which contain a 12-mm diameter round glass coverslip built into the bottom of the plate, were transiently transfected with MMP14-GFP. At 6 h post transfection cells were imaged to record the dynamic movement of MMP14-GFP containing vesicles. Images were acquired at rates of one frame every 1 s for 1 min. A selection of images are shown and localisation of a MMP14-GFP positive vesicle moving from the cell body to the periphery is indicated. Scale bar is 10 μm. The movie is available at the following web address: https://mediawarehouse.qut.edu.au/QMW/player/?dID=40717&dDocName=QMW_032392

4.2.5 MMP14 cell surface delivery is not dependent on intact microtubules or actin filaments

The high level of directionality observed for MMP14‐containing vesicles that

appeared to move along linear tracks from the cell body towards the cell periphery

suggested involvement of MMP14 positive vesicles along the cytoskeleton. Thus, the

effect of disrupting the cytoskeleton on MMP14 cell surface delivery was next tested.

First, RAW264.7 macrophages stimulated with 100 ng/ml LPS for 12 h were treated with

nocodazole for the last 3 h to disrupt microtubule networks (Figure 3.10B and Figure

4.5). Successful microtubule disruption was confirmed (as seen in Figure 3.10B). Cell

lysates were collected and immunoblotted for MMP14 and actin. It can be seen that the

total level of MMP14 is unaffected by microtubule disruption (Figure 4.5A). The surface

delivery of MMP14 was next tested by again treating cells in the presence or absence of

nocodazole. The delivery of MMP14 to the plasma membrane appears to be unaffected

by the disruption of the microtubule network as MMP14 surface levels are very similar

to those on control cells (Figure 4.5B). This was confirmed using flow cytometry, where

histograms of MMP14‐labelled macrophages following nocodazole‐treatment looked

similar to those in control cells (Figure 4.5C). The number of cells positive for surface

MMP14 was identical for treated and control macrophages (62%) (Figure 4.5D) and the

median fluorescence intensities did also not differ significantly between the two

populations (P>0.05) (Figure 4.5E). Together, this suggests that disruption of

microtubules under these conditions does not affect the trafficking of MMP14 to the cell

surface in macrophages.

In cancer cells, MMP14 has been reported to localise to actin‐rich structures at

the cell surface thought to assist migration [Poincloux et al., 2009]. To determine

whether this is the case in macrophages, RAW264.7 cells grown on coverslips and

https://mediawarehouse.qut.edu.au/QMW/player/?dID=40717&dDocName=QMW_032392


stimulated with 100 ng/ml LPS for 9 h were costained for MMP14 and F‐actin (Figure

4.6A). Some MMP14 was located along ruffles, sheet‐like protrusive actin structures but

there was also a large pool of MMP14 that did not appear to locate close to actin

structures. To test the requirement of actin filaments for MMP14 surface delivery,

RAW264.7 macrophages stimulated with 100 ng/ml LPS for 12 h were treated with

cytochalasin for the last 2 h to disrupt cortical actin (Figure 4.6B‐D). Immunoblotting of

cell lysates from cytochalasin‐treated macrophages and control cells for MMP14 and

actin confirmed that total MMP14 levels remained unaffected by the treatment (Figure

4.6B). LPS‐activated macrophages treated in the presence or absence of cytochalasin

were immunostained live and the surface levels of MMP14 were determined through

fluorescence microscopy and flow cytometry. By microscopy, cytochalasin treatment did

not appear to alter MMP14 surface levels (Figure 4.6C). This was confirmed by flow

cytometry showing surface MMP14 expression on cytochalasin‐treated macrophages

looked similar to those of control cells (Figure 4.6D). The number of MMP14 surface

positive cells was similar for both groups (62%) (Figure 5.6E). Similarly, there were no

significant differences in median fluorescence intensities of MMP14 between the

cytochalasin treated and control cells (Figure 5.6F). This suggests that despite the

localisation of some MMP14 to actin structures cell surface delivery of the protein is not

greatly influenced by the presence or absence of cortical actin.


Figure 4.5. MMP14 cell surface levels are not dependent on microtubules. (A) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 12 h and treated with nocodazole (Noc) for the last 3 h to disrupt microtubules. Cell lysates were collected and analysed by immunoblotting for MMP14 and actin. (B) RAW264.7 macrophages stimulated with 100 ng/ml LPS for 9 h and treated with Nocodazole (Noc) for the last 3 h to disrupt microtubules were immunostained for surface MMP14 (unpermeabilised cells), and fixed. Scale bar is 20 μm. (C) Macrophages stimulated with 100 ng/ml LPS for 9 h and treated with nocodazole (Noc) for the last 3 h to disrupt microtubules were immunostained live (unpermeabilised cells) with anti-MMP14 antibody or with an isotype control antibody, fixed and analysed using a FACSAriaIII flow cytometer. The overlaid histograms show surface MMP14-immunostained cell populations in comparison to the control cells. (D) Bar graph shows the percentage of cells positive for surface MMP14, n=3; mean± SEM, n.s. P>0.05 (unpaired, one-tailed Student’s t test). (E) Bar graph shows the relative MMP14 surface levels (MFI), n=3; mean± SEM, n.s. P>0.05 (unpaired, one-tailed Student’s t test).



Figure 4.6. Disruption of cortical actin does not affect MMP14 cell surface levels. (A) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 9 h, fixed and immunostained for MMP14 (red) and F-actin (green). Lower panels show pools of MMP14 that locate along ruffles, sheet-like protrusive actin structures at the cell surface. Scale bar is 20 μm. (B) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 12 h and treated with cytochalasin (Cyto) for the last 2 h to disrupt cortical actin. Cell lysates were collected and analysed by immunoblotting for MMP14 and actin. (C) RAW264.7 macrophages stimulated with 100 ng/ml LPS for 9 h and treated with cytochalasin (Cyto) for the last 2 h to disrupt cortical actin were immunostained for surface MMP14 (unpermeabilised cells), and fixed. Scale bar is 20 μm. (D) Macrophages stimulated with 100 ng/ml LPS for 9 h and treated with cytochalasin (Cyto) for the last 2 h to disrupt cortical actin were immunostained live (unpermeabilised cells) with antibodies specific for MMP14 or with an isotype control antibody, fixed and analysed using a FACSAriaIII flow cytometer. The overlaid histograms show surface MMP14-immunostained cell populations in comparison to the control cells. (E) Bar graph shows the percentage of cells positive for surface MMP14, n=3; mean± SEM, n.s. P>0.05 (unpaired, one-tailed Student’s t test). (F) Bar graph shows the relative MMP14 surface levels (MFI), n=3; mean± SEM, n.s. P>0.05 (unpaired, one-tailed Student’s t test).

4.2.6 Newly synthesised MMP14 is trafficked via the Golgi complex

MMP14 has an N‐terminal signal peptide targeting it for transient retention in

the ER membrane during its biosynthesis [Egeblad and Werb, 2002]. Unless a protein

possesses a retention signal in addition to the signal peptide localisation to the ER during

biosynthesis can be very transient, which can make the detection of proteins that are not

resident to the ER difficult [Alberts et al., 2002]. Accordingly, when simultaneously

labelling MMP14 and the ER marker PDI in RAW264.7 macrophages stimulated with 100

ng/ml LPS for 6 h no overlap can be observed (Figure 4.7A). MMP14 is very quickly

shuttled to the Golgi apparatus before being processed to the mature form and

transported to the cell surface to be incorporated into the plasma membrane. In

macrophages this route to the cell surface is unclear. To confirm MMP14 localisation to

the Golgi complex in RAW264.7 macrophages, cells were stimulated with 100 ng/ml LPS

for 6 h followed by immunostaining of MMP14 and the cis‐Golgi protein GM130 (Figure

4.7B). It can be seen from Figure 4.7B that a perinuclear pool of MMP14 co‐localises with

GM130 on the Golgi complex as indicated by strong overlap of their fluorescence signals

(Figure 4.7B). To test whether MMP14 trafficking via the Golgi complex is a prerequisite

for its incorporation into the plasma membrane, RAW264.7 macrophages were

stimulated with 100 ng/ml LPS for 12 h and treated with brefeldin A for the last 6 h,

which leads to collapse of the Golgi complex. Immunoblotting of lysates shows that

disrupting transport through the Golgi complex did not affect total MMP14 levels as

compared to control cells (Figure 4.7C). Next, surface MMP14 was immunostained live

in unpermeabilised macrophages to determine the effect of Golgi stack dispersal on

delivery of MMP14 to the plasma membrane. Fluorescence imaging demonstrated that

surface MMP14 was almost absent in macrophages treated with brefeldin A, while


control cells showed high levels of surface MMP14 (Figure 4.7D). This was confirmed by

flow cytometry where histograms for MMP14 surface levels reflecting the dramatic left

shift for populations that had been treated with the brefeldin A (Figure 4.7E). Only 8%

of macrophages remain positive for surface MMP14 following brefeldin A treatment,

which represents an almost 8‐fold decrease compared to the control with 62% of

MMP14‐positive macrophages (**P≤0.01) (Figure4.7F). Relative to the control, median

fluorescence intensities were also reduced by 88% after brefeldin A treatment

(***P≤0.001) (Figure 4.7G). Taken together, this provides strong evidence that MMP14

trafficking through the Golgi complex is crucial for MMP14 cell surface delivery in

macrophages.

Figure 4.7. Newly synthesised MMP14 is trafficked via the Golgi complex. (A) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 6 h, fixed and immunostained for MMP14 (red) and the ER marker PDI (green). The smaller panels on the right show an enlarged region where perinuclear MMP14 is in proximity to PDI but does not overlap. Scale bar is 20 μm. (B) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 6 h, fixed and immunostained for MMP14 (red) and the cis-Golgi marker GM130 (green). The smaller panels on the right show an enlarged region where perinuclear MMP14 co-localises with GM130. Scale bar is 20 μm. (C) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 12 h and treated with brefeldin A (BFA) for the last 6 h to disrupt the Golgi complex. Cell lysates were collected and analysed by immunoblotting for MMP14 and actin. (D) RAW264.7 macrophages stimulated with 100 ng/ml LPS for 9 h and treated with brefeldin A (BFA) for the last 6 h to disrupt the Golgi complex were immunostained live (unpermeabilised cells) for surface MMP14, and fixed. Scale bar is 20 μm. (E) Macrophages stimulated with 100 ng/ml LPS for 9 hours and treated with brefeldin A (BFA) for the last 6 h were immunostained with an anti-MMP14 antibody or with an isotype control antibody, fixed and analysed using a FACSAriaIII flow cytometer. The overlaid histograms show surface MMP14-immunostained cell populations in comparison to the control macrophages. (F) Bar graph shows the percentage of cells positive for surface MMP14, n=3; mean± SEM, **P≤0.01 (unpaired, one-tailed Student’s t test). (G) Bar graph shows the relative MMP14 surface levels (MFI), n=3; mean± SEM, ****P≤0.0001 (unpaired, one-tailed Student’s t test).



4.2.7 R-SNARE VAMP4 regulates delivery of MMP14 from the Golgi network en route to the cell surface

Having shown that MMP14 is transported through the Golgi complex and that it

is also located in late endosomes/lysosomes possibly en route to the cell surface it was

next determined which SNARE proteins might regulate this transport. As SNAREs have

precise subcellular localisations and mediate specific fusion events, identification of the

trafficking machinery proteins involved in the incorporation of MMP14 into the plasma

membrane would further pinpoint the intracellular trafficking pathways for this protein.

VAMP4 is located in the Golgi complex in RAW264.7 macrophages [Lei et al.,

2012] and is known to regulate fusion from and to the Golgi network in other cell types.

As trafficking of newly made MMP14 via the Golgi was identified as prerequisite for its

incorporation into the plasma membrane, VAMP4 knockdown was tested to determine

whether it could regulate transport of MMP14. RAW264.7 macrophages were

transfected with non‐targeted (scrambled) or VAMP4 siRNA on day 1, this was repeated

on day 2 and then the cells were stimulated with LPS for 15 h. As determined in previous

experiments, VAMP4 levels were reduced by 38% (**P≤0.01) without affecting any of

the other tested R‐SNAREs (Figure 3.13). To test whether SNARE knockdown affected

levels of total MMP14 cell lysates were immunoblotted for MMP14 (Figure 4.8B).

Comparison of the intensity of MMP14 bands of samples following targeted knockdown

of SNARE proteins to the scrambled control showed that MMP14 levels remained mostly

unchanged. MMP14 surface levels following knockdown of VAMP4 were assessed

through immunolabelling of unpermeabilised macrophages (Figure 4.8C). A reduction in

MMP14 surface levels was seen by fluorescence microscopy when compared to a

scrambled control. Quantification of flow cytometry data revealed a 40% reduction in

cell numbers (**P≤0.01) and 38% reduction of overall MMP14 surface levels (n.s.

P>0.05) for cells treated with VAMP4 siRNA compared to scrambled siRNA treated cells

(Figure 4.8D). This suggests that VAMP4‐mediated trafficking of MMP14 from the Golgi

complex is required for the delivery of MMP14 to the cell surface.


Figure 4.8. R-SNARE VAMP4 regulates delivery of MMP14 from the Golgi network en route to the cell surface. (A) Schematic of the subcellular distribution of SNARE proteins that were targeted using siRNA knockdown. (B) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 15 h following transfection with scrambled (scr), VAMP2, VAMP3, VAMP4, VAMP7 or VAMP8 siRNA. Whole cell lysates were assayed for total MMP14 levels by immunoblotting, with actin serving as a loading control. (C) RAW264.7 macrophages grown on coverslips were immunostained for surface MMP14 (unpermeabilised cells) and fixed. Scale bar is 20 μm. (D) Live cells were immunostained for surface MMP14 or with an isotype control antibody (unpermeabilised cells), fixed and analysed using a FACSAriaIII flow cytometer. Bar graph shows the relative percentage of cells positive for surface MMP14 or fold change of MMP14 surface levels (MFI) n=3; mean± SEM, **P≤0.01 (one-way ANOVA with Dunnett’s post-hoc test).


4.2.8 MMP14 is not trafficked via the recycling endosome

After passing through the Golgi apparatus MMP14 is then transported to the

surface anchored in the membrane of vesicles, which could occur either directly or via

other organelles, such as the recycling endosomes or lysosomes. It has been suggested

that trafficking of newly synthesised MMP14 to the cell surface could occur via the

recycling endosome, as its transport to the plasma membrane in primary macrophages

was shown to be mediated by Rab8a, which can act as a recycling endosome Rab

[Wiesner et al., 2013]. However, no co‐localisation with Rab11, another recycling

endosome Rab, was found [Wiesner et al., 2013]. It was therefore next determined

whether MMP14 localises to recycling endosomes. RAW264.7 macrophages stimulated

with 100 ng/ml LPS for 9 h were fixed, permeabilised and immunostained for MMP14

and the recycling endosome marker protein transferrin receptor (TfR). MMP14 was

observed in the perinuclear region and in punctate structures in the cytoplasm (Figure

4.9A). TfR also appeared in punctuate structures, however no overlap was seen with

MMP14 (Figure 4.9A). This suggests that newly made MMP14 is not trafficked via the

recycling endosome.

In fibrosarcoma cell lines MMP14 is trafficked via recycling endosomes and

delivery to the plasma membrane is dependent on the recycling endosome R‐SNARE

VAMP3 [Kean et al., 2009]. In cancer cells, MMP14 at the cell surface is continuously

internalised and then either degraded [Takino et al., 2003] or recycled back to the cell

surface [Remacle et al., 2003; Wang et al., 2004]. Thus, MMP14 that is delivered to the

cell surface from a recycling endosome could be newly made or might also be

endocytosed protein that is delivered back to the cell surface via a long‐loop recycling

pathway [Frittoli et al., 2011]. In macrophages, VAMP3 mediates transport of newly

synthesised proteins from the Golgi network to the recycling endosome and then from

the recycling endosome to the plasma membrane [Murray et al., 2005a; Murray et al.,

2005b]. It also mediates the transport of endocytosed material from the recycling

endosome to the cell surface [Veale et al., 2010] and so VAMP3 has the potential to

regulate both pathways. VAMP2 is an R‐SNARE that is located on early endosomes in

macrophages [Veale et al., 2010] and has been found to regulate early endocytic events

in neuronal and HeLa cells [Miller et al., 2011; Koo et al., 2011]. To test whether MMP14

could be trafficked through the recycling endosome en route to the cell surface or

recycled through the early endosome and recycling endosome back to the cell surface

VAMP2 or VAMP3 were targeted by siRNA knockdown. Following knockdown, VAMP2

levels were reduced by 38%, while VAMP3 levels were reduced by 70% (****P≤0.0001)


without affecting the levels of other tested R‐SNAREs (Figure 3.13). Total MMP14 levels

remained unaffected in these cells compared to scrambled treated cells (Figure 4.8B). No

changes in MMP14 surface levels were apparent by fluorescence microscopy when cells

were treated with VAMP2 or VAMP3 siRNA compared to scrambled treated cells (Figure

4.9B). Quantification of macrophages positive for surface MMP14 showed minimal

change in their populations, with a 7% reduction for VAMP2 siRNA treated cells and a

10% increase following knockdown of VAMP3 compared to scrambled treated cells,

although these changes were not significant (n.s. P>0.05)(Figure 4.9C). Similarly, overall

MMP14 surface levels were not significantly altered in these populations. These results

together suggest that VAMP3 does not mediate the transport of newly synthesised

MMP14 to the cell surface and the VAMP2 and VAMP3 data together suggest that unlike

some cancer cells the majority of surface MMP14 is not recycled through an early

endosomes and recycling endosomes in macrophages.



Figure 4.9. MMP14 is not trafficked via the recycling endosome. (A) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 9 h, fixed and immunostained for MMP14 (red) and the recycling endosome marker transferrin receptor (TfR) (green). Scale bar is 10 μm. (B) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 15 h following transfection with scrambled (scr), VAMP2 or VAMP3 siRNA. RAW264.7 macrophages grown on coverslips were immunostained for surface MMP14 (unpermeabilised cells) and fixed. Scale bar is 20 μm. (C) Live cells were immunostained for surface MMP14 or with an isotype control antibody (unpermeabilised cells), fixed and analysed using a FACSAriaIII flow cytometer. Bar graph shows the relative percentage of cells positive for surface MMP14 or fold change of MMP14 surface levels (MFI), n=3; mean± SEM, n.s. P>0.05 (one-way ANOVA with Dunnett’s post-hoc test).

4.2.9 The R-SNAREs VAMP7 and VAMP8 regulate delivery of MMP14 from late endosomes/lysosomes en route to the cell surface

In cancer cells, MMP14 is continuously internalised and either trafficked to late

endosomes/lysosomes for degradation or recycled back to the cell surface. However, in

macrophages MMP14 incorporated into the plasma membrane is mostly newly made

protein (Figure 4.3). Delivery of newly made MMP14 to the cell surface could

nonetheless occur via late endosomes/lysosomes. Hence, whether MMP14 localises to

late endosome/lysosomes was next tested using antibodies specific for lysosome‐

associated membrane protein 1 (LAMP1), a structural glycoprotein integrated into the

limiting membranes of both late endosomes and lysosomes and a commonly used

marker for these compartments [Chen et al., 1985; Eskelinen et al., 2003]. Thus,

macrophages stimulated with LPS for 9 h were, fixed permeabilised and immunostained

for MMP14 and LAMP1. A pool of MMP14 was found co‐localising with some LAMP1

compartments resulting in partial overlap of the two fluorescence signals (Figure 4.10A).

This suggests that some MMP14 could be trafficked through a late endosomal/lysosomal

compartment.

Two R‐SNAREs, VAMP7 and VAMP8, are known to regulate late

endosome/lysosome fusion [Pryor et al., 2004]. Both, VAMP7 and VAMP8, have also been

implicated in trafficking events from lysosome‐related organelles to the cell surface in a

range of immune cells [Braun et al., 2004; Dressel et al., 2010; Fader et al., 2009; Lippert

et al., 2007; Logan et al., 2006; Mollinedo et al., 2003; Paumet et al., 2000]. Hence, VAMP7

and VAMP8 protein levels were reduced using siRNA to test for their involvement in

MMP14 cell surface delivery via a late endosomal/lysosomal pathway (Figure 4.10B).

Following knockdown, VAMP7 levels were reduced by 58% (****P≤0.0001), while

VAMP8 levels were reduced by 71% (****P≤0.0001) without affecting any of the other

tested R‐SNAREs (Figure 3.13). Total MMP14 levels remained unaffected (Figure 4.8B).

Unpermeabilised macrophages were labelled live for surface MMP14 following siRNA

knockdown and imaged by microscopy. MMP14 levels at the cell surface appear reduced


in both VAMP7 and VAMP8 siRNA treated cells when compared to the control scrambled

siRNA treated cells (Figure 4.10B). From flow cytometry analysis it could be seen that

treatment with VAMP7 siRNA reduced the number of macrophage that are positive for

surface MMP14 by 31% (**P≤0.01) and VAMP8 siRNA treatment by 33% (**P≤0.01)

compared to siRNA treated cells (Figure 4.10C). Overall MMP14 levels were reduced by

37% and 43% following VAMP7 or VAMP8 siRNA treatment, respectively. This suggests

that MMP14 localising to late endosome/lysosome compartments is en route to the cell

surface and its transport either through these compartments is mediated by VAMP7 and

VAMP8.

Figure 4.10. R-SNAREs VAMP7 and VAMP8 regulate delivery of MMP14 from late endosomes/lysosomes en route to the cell surface. (A) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 9 h, fixed and immunostained for MMP14 (red) and the late endosome/lysosome marker LAMP1 (green). Scale bar is 20 μm. (B) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 15 h following transfection with scrambled (scr), VAMP7 or VAMP8 siRNA. RAW264.7 macrophages grown on coverslips were immunostained for surface MMP14 (unpermeabilised cells) and fixed. Scale bar is 20 μm. (C) Live cells were immunostained for surface MMP14 or with an isotype control antibody (unpermeabilised cells), fixed and analysed using a FACSAriaIII flow cytometer. Bar graph shows the relative percentage of cells positive for surface MMP14 or fold change of MMP14 surface levels (MFI) n=3; mean± SEM, *P<0.05 (one-way ANOVA with Dunnett’s post-hoc test).



4.3 Discussion

MMP14 expression is highly elevated in chronic wounds and as an activator of

MMP9 it can contribute to excessive matrix degradation and tissue damage. Additionally,

MMP14 also promotes the movement of macrophages into wounds allowing them to

further perpetuate the inflammatory state of a non‐healing wound. Hence, preventing

delivery of MMP14 to the cell surface, where the enzyme is incorporated into the plasma

membrane and active, might not only lead to a reduction in proteolytic potential of MMP9

but might also reduce tissue infiltration by macrophages and therefore dampen

inflammation. Thus, this chapter aimed to identify key organelles and machinery

involved in the intracellular transport and cell surface delivery of MMP14 in

macrophages.

The results showed that MMP14 expression is upregulated in macrophages and

this newly synthesised MMP14 is incorporated into the plasma membrane upon LPS

activation. The cell surface delivery of MMP14 appeared to be highly directional and

along linear tracks from the cell body to the periphery but disruption of elements of the

cytoskeleton did not affect levels of surface MMP14. MMP14 localises to the Golgi

complex and is trafficked via this organelle prior to its incorporation into the plasma

membrane. MMP14 was also found in late endosomes/lysosomes and targeted siRNA

knockdown of the late endosome and lysosome R‐SNAREs VAMP7 and VAMP8 showed

that MMP14 is trafficked through these organelles. In cancer cells a VAMP3 mediated

recycling pathway has been found in some cells to regulate surface levels of MMP14 but

knockdown of VAMP2 and VAMP3 had little effect on surface MMP14 levels. This

suggests that MMP14 might be trafficked from the Golgi complex to late

endosomes/lysosomes prior to its surface delivery and data from VAMP4 siRNA treated

cells suggest VAMP4 could potentially regulate this step. Thus, the knockdown of specific

SNARE proteins known to mediate different pathways allowed identification of specific

machinery involved in the intracellular trafficking of MMP14 and its delivery to the cell

surface.


4.3.1 MMP14 expression and cell surface delivery in RAW264.7 macrophages

MMP14 expression

In contrast to cancer cells where MMP14 appears to be constitutively expressed

and levels often correlate with invasiveness, MMP14 expression in macrophages is

transient and induced upon exposure to external stimuli [Hald et al., 2012; Ming et al.,

2012; Reel et al., 2011]. In chronic wounds, macrophage numbers, as well as MMP14

expression, appear to be highly elevated compared to normal skin leading to enhanced

matrix turnover [Norgauer et al., 2002]. This underpins the importance of spatio‐

temporal control of MMP14 activity to avoid unrestrained infiltration of macrophages

into the wound and excessive tissue damage. MMP14 expression has been shown to be

upregulated at the RNA level in human and mouse monocytes/macrophages following 2,

6, 18 h and 24 h LPS stimulation [Hald et al., 2012; Ming et al., 2012; Murray et al., 2013;

Reel et al., 2011]. It has been shown that in RAW264.7 macrophages, upregulation of

MMP14 mRNA is 10‐, 13‐ and 6‐fold following 2, 6 and 18 h of LPS stimulation,

respectively, suggesting temporal upregulation of MMP14 [Hald et al., 2012]. At the

protein level, MMP14 was shown to be upregulated in human macrophages after 24 h of

IFN and LPS activation, with levels increasing about 3‐fold [Johnson et al., 2014].

Accordingly, herein unstimulated RAW264.7 macrophages showed low levels of MMP14.

Protein levels then increased by 2.7‐fold after stimulation with LPS and peaked around

15 h post‐stimulus. Thus, activation of macrophages with LPS increased MMP14

synthesis. Interestingly, other factors found in wounds also upregulate MMP14

expression, for example monocytes differentiated on fibronectin have increased levels of

mRNA, peaking around 18 h, suggesting the wound environment could upregulate

MMP14 [Reel et al., 2011]. An effect of the substratum on MMP14 levels has also been

observed in breast adenocarcinoma cells where endocytosis events of MMP14 only occur

in the absence of a matrix substratum and can be inhibited through plating of cells on

collagen [Bravo‐Cordero et al., 2007; Lafleur et al., 2006]. In accordance with this,

MMP14 surface levels were upregulated in RAW264.7 macrophages that were not LPS‐

stimulated but plated onto fibronectin or gelatin, a hydrolysed form of collagen, to a

similar extent as observed following stimulation with the endotoxin.

After synthesis, at least two forms of MMP14 are present in the cell; MMP14 is

initially synthesised as a zymogen (66 kDa) that requires cleavage of the pro‐peptide by

furin in the TGN to form the mature active form (54 kDa) [Mazzone et al., 2004]. Although

both forms of MMP14, cleaved and uncleaved, could potentially be expected to be


present in cell lysates only one band, representing the mature form, was detected in

immunoblots of LPS‐stimulated RAW264.7 macrophages. The MMP14 antibody that was

used for immunoblotting was raised against an immunogen sequence derived from a

region within amino acids 150‐180 of the protein and should therefore be able to

recognise both forms of MMP14. Localisation to the ER during biosynthesis can be very

transient [Alberts et al., 2002] and the absence of the pro‐MMP14 band herein suggest

that the transport of MMP14 through the ER and the Golgi apparatus is relatively fast

leading to quick turnover of pro‐MMP14 to the mature form in the TGN. Accordingly,

immunofluorescence images with MMP14 and the ER marker PDI showed no overlap in

LPS‐stimulated RAW264.7 macrophages (Figure 4.7A). The fact that MMP14 exhibits

multiple O‐glycosylation sites but in contrast to MMP9 no N‐glycosylation sites [Remacle

et al., 2006] also suggests a short retention time within the ER which could lead to faster

processing of pro‐MMP14 in the TGN.

In RAW264.7 macrophages that had been stimulated with LPS for 9 h, the

majority of MMP14 was found to localise to the perinuclear Golgi complex (Figure 4.1B).

After 18 h of LPS stimulation more MMP14 localised to punctate structures in the cell

periphery and near the cell surface. This suggests that upon LPS stimulation newly made

protein is delivered to the cell surface for incorporation into the plasma membrane.

MMP14 cell surface delivery

Transport of overexpressed MMP14 from perinuclear regions, most likely the

Golgi complex, into cell periphery and protrusions for incorporation into the plasma

membrane could be observed by live cell imaging (Figure 4.4.C). Small MMP14‐GFP‐

positive vesicles appeared to follow straight paths in the cytosol towards the cell surface

(Figure 4.4C). In some instances, what appeared to be two separate vesicles, moved along

the same track in the same direction with similar velocity (Figure 4.4C). The average

velocity of these moving vesicles was with 0.65±0.16 μm/s very similar to what has

previously been reported for fluorescently‐tagged MMP14 transport in primary

macrophages, which exhibited velocities of about 0.8 μm/s [Wiesner et al., 2010].

Together, this suggests that transport involves some kind of tracks. Occasionally, the

small fast‐moving vesicles travelled via slightly larger and mostly stationary MMP14‐

GFP‐positive vesicular endosomes, which did not change in fluorescence (Figure 4.4C,

16” frame). Similar trafficking behaviour has previously been observed for cell surface

delivery of LAMP1‐positive vesicles in RAW264.7 macrophages where these kind of

vesicular endosomes were identified as late endosomal/lysosomal compartments

[Offenhauser, 2011].


Immunolabelling of unpermeabilised cells allowed for specific detection of

surface MMP14 and showed that in RAW264.7 macrophages the delivery of MMP14 to

plasma membrane was induced upon LPS stimulation (Figure 4.2). Increased levels of

MMP14 at the cell surface could be observed from 6 h post‐stimulation and continued to

increase over an 18 h period. While total levels of MMP14 increased only by about three‐

fold levels of surface MMP14 increased by over 10 times following LPS stimulation

suggesting that the majority of newly made MMP14 is delivered to the cell surface. That

the incorporation of MMP14 into the plasma membrane is coupled to the de novo

synthesis of the protein was confirmed in RAW264.7 macrophages when protein

synthesis was disrupted following cycloheximide treatment, which led to significantly

reduced MMP14 surface presentation (Figure 4.3). This is in contrast to cancer cells,

where a significant fraction of the MMP14 that is delivered to the cell surface occurs from

endosome compartments that contain endocytosed rather than newly made protein,

similar to endocytic recycling of integrins [Poincloux et al., 2009].

4.3.2 The cytoskeleton and MMP14 cell surface delivery

Live imaging of MMP‐GFP showed the average velocity of the anterograde

vesicles was 0.65±0.16 μm/s and that the long‐distance movement of these MMP14

vesicles was not random but showed high directionality (0.93±0.01) and appeared to

follow straight paths in the cytosol towards the cell surface. Movement through the

cytoplasm by diffusion alone, which is in the order of 0.1 μm/s, cannot account for the

rate, directionality, and destinations of MMP‐GFP suggesting that transport involves

some kind of tracks [Sarfaraz and Brown, 2016]. The cytoskeleton can assist in the

anterograde transport of cargo‐containing vesicles to the cell surface [Anitei and Hoflack,

2012]. Anterograde transport of MMP14 vesicles in macrophages derived from human

monocytes has been reported to occur along microtubules with velocities of 0.78 μm/s

and involves the kinesins KIF5b, KIF3A and KIF3B [Wiesner et al., 2010]. Retrograde

transport and internalisation of MMP14 was observed suggesting that MMP14 surface

levels were at least partly influenced by dynein‐ and kinesin‐dependent recycling of

MMP14 and not solely by newly made protein. However, the majority of the data

obtained in this study relied on the transient overexpression of mCherry‐labelled

MMP14 and assessment of MMP14‐mCherry surface levels was conducted 18 h post‐

transfection, which means it might not represent early trafficking events of newly

synthesised MMP14. Importantly, these studies were performed in unactivated

macrophages. It is known that transport pathways in macrophages can be dramatically

altered after activation [Murray and Stow, 2014].


It was also reported that MMP14‐positive vesicles could spontaneously switch

direction. This is indicative of the presence of both, antero‐ and retrograde motor

proteins on the same vesicle. The bidirectional movement of late endosomes has been

described as a ‘tug of war’ between dynein and kinesin motors [Granger et al., 2014]. The

involvement of kinesins in the surface delivery of MMP14 was evaluated in macrophages

with reduced kinesin levels, while microtubules remained intact [Granger et al., 2014].

This might have favoured the retrograde dynein‐mediated transport of MMP14 rather

than just inhibiting anterograde transport. This study also showed that not all surface

MMP14 is regulated by kinesins indicating that MMP14 cell surface delivery can also

occur in the absence of kinesins or that anterograde transport is also influenced by other

factors [Granger et al., 2014].

Nonetheless, a role for microtubules in the polarised delivery to invasive

structures has been postulated [Linder, 2007] and transport of MMP14 vesicles along

microtubules towards podosomes has been observed in macrophages [Wiesner et al.,

2010]. Thus, investigating the role of the microtubule network on MMP14 surface levels,

as well as the enzyme’s polarised delivery to invasive structures to allow matrix

degradation and migration, could be informative. Disrupting microtubule networks with

5 μM/ml nocodazole did not affect the levels of MMP14 at the cell surface in RAW264.7

macrophages (Figure 4.5B and C). A lack of an effect could be due to the fact that upon

nocodazole‐induced microtubule disruption Golgi membrane components gradually

redistribute from the MTOC to peripheral sites. Once these mini‐Golgi stacks are re‐

established, efficient transport from dispersed Golgi sites to the cell surface is restored

even though the microtubule network is still disturbed but is no longer directed to

specific regions at the cell surface [Cole et al., 1996]. To add to this, microtubules only

increase the probability for a cargo vesicle to encounter the target membrane, which

means that the secretion in nocodazole‐treated samples is only slowed and not fully

blocked [Wacker et al., 1997]. In a study investigating the dynamic movements of

MMP14‐mCherry in breast cancer cells, these MMP14‐containing vesicles showed

reduced mobility following nocodazole treatment but vesicles were not absolutely

stationary and the effect on cell surface levels of the protein were not investigated

[Marchesin et al., 2015]. As treatment with nocodazole was only for the last 3 h of 12 h

LPS stimulation, there is a possibility that any effect could have been too small to be

detected given the large amounts of MMP14 already being delivered to the cell surface

during the absence of nocodazole. Differences in treatment times, as well as nocodazole

concentration, and the activation of the macrophages with LPS might have added to the

absence of an effect. However, this can only be speculated as dosage and timings for the


nocodazole treatment in the breast cancer cell study were not mentioned. It might,

however, be valuable to test whether microtubules disruption does influence the

distribution of MMP14 to invasive structures despite total surface levels of the protein

being unaffected.

In macrophages derived from human monocytes MMP14‐containing vesicles

were also found to move along actin filaments at the cell periphery with velocities of 0.2‐

0.5 μm/s [Wiesner et al., 2010]. In LPS‐stimulated RAW264.7 macrophages a pool of

endogenous MMP14 was found to locate along ruffles, sheet‐like protrusive actin

structures. However, disruption of cortical actin with cytochalasin did not affect MMP14

levels at the cell surface. Treatment of RAW264.7 macrophages with cytochalasin for 2 h

had previously been shown to disrupt cortical actin and an effect on lysozyme secretion

could be detected [Offenhauser, 2011]. Longer incubation times in the presence of

cytochalasin are not recommended as it can induce apoptosis [Rubtsova et al., 1998]. But

significant levels of surface MMP14 can only be observed from 9 h post‐stimulation. As a

consequence, cells were treated with cytochalasin for the last 2 h of 9 h LPS stimulation

and any effect could have been too small to be detected given the large amounts of

MMP14 already being delivered to the cell surface before start of the treatment. The actin

skeleton has also been reported to have both, supporting as well as hindering roles

during the secretion process [Porat‐Shliom et al., 2013]. Actin filaments can assist in the

guidance of exocytic vesicles to sites of fusion, but actin can also act as physical barrier

for any cargo trying to access the plasma membrane [Porat‐Shliom et al., 2013]. So

disruption of the cortical actin could allow for easier access of MMP14‐containing

vesicles to the plasma membrane annulling any loss in vesicle guidance. It may also be

the case that the actin filaments are necessary for proper localisation and function at the

cell surface, rather than being a general prerequisite for the cell surface delivery. In fact,

tethering of MMP14’s cytoplasmic tail to F‐actin, through the action of N‐WASP, was

found to be necessary to stabilise MMP14 to invasive structures (invadosomes) and

induce matrix degradation in breast cancer cells [Yu et al., 2012]. Macrophages do not

have invadosomes, but instead have related structures termed podosomes, to which

MMP14 also localises [Nusblat et al., 2011]. Thus, it could be interesting in future to test

whether cytochalasin‐treated macrophages show reduced degradative capacity despite

MMP14 surface levels remaining unchanged. The emerging picture for cancer cells is that

the formation of invasive structures for subsequent matrix degradation relies on the

local coordination of the cytoskeleton to assist delivery of MMP14 to these structures

[Poincloux et al., 2009]. MMP14 is also implicated in 2D migration of macrophages and

is delivered to podosomes at the front of migrating cells. But whether MMP14 is


specifically transported to podosomes or is also delivered to other places on the

macrophage cell surface and then to the podosomes in macrophages has not been fully

investigated yet [Verollet et al., 2011]. Trafficking of MMP14 to specific areas of the

plasma membrane has also been suggested to rely on interactions with podosome‐

associated matrix receptors, such as CD44 and integrins [Wiesner et al., 2010].

4.3.3 Trafficking pathways for MMP14 in RAW264.7 macrophages

Trafficking of newly made MMP14 via the Golgi complex

MMP14 has an N‐terminal signal peptide targeting it for translation into the ER

membrane during biosynthesis [Egeblad and Werb, 2002]. Following translation into the

ER, MMP14 is shuttled through the Golgi apparatus. While localisation of MMP14 to the

ER appeared to be only very transient, MMP14 was found to locate to perinuclear

structures in a pattern typical for the Golgi complex in LPS‐stimulated RAW264.7

macrophages. A recent study using primary human macrophages revealed that MMP14

co‐localised with Golgi components, such as GM130, a cis‐Golgi marker, and TGN46, a

trans‐Golgi protein [Wiesner et al., 2010]. Co‐localisation with GM130 in RAW264.7

macrophages established that newly made MMP14 also locates to the Golgi apparatus in

these LPS activated cells. More interestingly, disruption of the Golgi complex with

brefeldin A abolished incorporation of MMP14 into the plasma membrane, which

confirms that trafficking via this organelle is essential for the cell surface delivery in

activated macrophages.

As trafficking of nascent MMP14, via the Golgi network, was identified as

prerequisite for its incorporation into the plasma membrane SNARE machinery that

could mediate this pathway in macrophages was investigated. As VAMP4 is located in the

Golgi complex in RAW264.7 macrophages [Lei et al., 2012], and known to regulate fusion

from the Golgi network in other cell types, it was considered a possible candidate for

MMP14 trafficking. A significant reduction of MMP14 surface levels was observable

following knockdown of VAMP4 suggesting that this SNARE is required for trafficking of

MMP14 from the Golgi to the cell surface. VAMP4 could mediate either direct trafficking

of MMP14 to the cell surface or via other organelles. In activated NK cells, VAMP4

localises to lytic granules, often referred to as ‘secretory lysosomes’ or ‘lysosome‐related

organelles’ and is indispensable for LAMP1 cell surface delivery from these organelles

[Krzewski et al., 2011]. VAMP4 (R‐SNARE), as well as Stx7 (Qa‐SNARE), Vti1b (Qb‐

SNARE) and Stx8 (Qc‐SNARE), also localises to the lysosomal related lytic granules in


cytotoxic T lymphocytes [Pattu et al., 2012]. These SNAREs could theoretically function

as a SNARE complex. VAMP4 might, thus, mediate trafficking of MMP14 from the Golgi

network to late endosomal/lysosomal compartments in macrophages, from where the

protein could then be incorporated into the plasma membrane. However, to confirm this

hypothesis, the existence of a VAMP4/Stx7/Vti1b/Stx8 complex in macrophages and the

involvement of the Q‐SNARE Stx7, Vti1b and Stx8 in MMP14 transport would have to be

tested.

Trafficking of newly made MMP14 via Recycling Endosome

Evidence that MMP14 traffics through recycling endosomes has been shown for

fibrosarcoma and breast carcinoma cell lines [Kean et al., 2015 Remacle et al., 2005].

However, this could potentially be a result of MMP14 recycling events that continuously

occur in cancer cells. In primary human macrophages it has been proposed that newly

made MMP14 could be trafficked via the recycling endosome as the GTPase Rab8a, which

can mediate biosynthetic traffic through this compartment, was found to regulate

MMP14 surface levels [Henry and Sheff, 2007; Wiesner et al., 2013]. Yet, the exact route

is unclear and no co‐localisation with Rab11, a more established recycling endosome

Rab, was found [Wiesner et al., 2013]. Herein it was found that in RAW264.7

macrophages, MMP14 did not co‐localise with TfR, another marker for recycling

endosomes, suggesting newly made protein is not trafficked via this organelle. This

confirms a study utilising an adenocarcinoma cell line where trafficking of newly made

MMP14 was also found to depend on Rab8, but did not co‐localise with typical markers

of the recycling endosomes, such as Rab11, transferrin and TfR suggesting it was not

trafficked through this compartment [Bravo‐Cordero et al., 2007]. Interestingly, Rab8

has also been associated with ruffle formation and membrane recycling in fibrosarcoma

cells [Hattula et al., 2006]. As pools of MMP14 were found to locate to ruffle structures

in RAW264.7 macrophages trafficking of newly made protein to these structures could

be a possibility.

Recycling of MMP14

MMP14 is synthesised and delivered to the cell surface in macrophages upon LPS

activation but surface MMP14 could also originate from previously endocytosed protein

that is recycled back via a short‐loop (from the early endosome directly back to the cell

surface) or long‐loop (from the early endosome to the recycling endosome and then to

the surface) pathway. In cervical adenocarcinoma, fibrosarcoma and breast

adenocarcinoma cell lines, the MMP14 delivered to the cell surface predominantly stems

from pools of previously endocytosed protein that is recycled back to the cell surface


[Frittoli et al., 2014; Remacle et al., 2003; Remacle et al., 2005]. Rab 4 and Rab5 have

been identified as regulators for fast short‐loop recycling of MMP14 from the early

endosome to the plasma membrane, while Rab11 mediates the slower long‐loop delivery

of recycled MMP14 via the recycling endosome to the cell surface [Frittoli et al., 2014;

Remacle et al., 2003; Remacle et al., 2005]. The R‐SNARE VAMP3 has also been reported

to regulate MMP14 surface delivery from the recycling endosome in fibrosarcoma cells

[Kean et al., 2009]. In unactivated primary macrophages, Rab5a is responsible for

endocytosis of MMP14, short‐loop recycling of MMP14 is mediated by Rab14, while

Rab22a allows shuttling from the early endosome to the recycling endosome in a long‐

loop recycling pathway [Wiesner et al., 2013]. However, SNARE proteins that could be

involved in such a process and whether this was a major transport route in activated

macrophages had not been tested. VAMP2 is located on early endosomes in RAW264.7

macrophages [Veale et al., 2010] and has been found to regulate endocytic events in

neuronal and cervical adenocarcinoma cells [Miller et al., 2011; Koo et al., 2011]. VAMP3

is a recycling endosome SNARE and in macrophages mediates transport of both, newly

synthesised proteins from the Golgi network to the recycling endosome and then from

the recycling endosome to the plasma membrane [Murray et al., 2005a; Murray et al.,

2005b] as well as endocytosed material from the recycling endosome to the cell surface

[Veale et al., 2010]. To test whether VAMP2 or VAMP3 could be involved in MMP14

trafficking via an endocytosis/recycling pathway in RAW264.7 macrophages, levels of

these SNAREs were reduced by siRNA knockdown. No significant effects were seen with

MMP14 surface levels. From this it can be concluded that VAMP2 and VAMP3 may not

mediate trafficking MMP14 and that the majority of the surface MMP14 is not

transported through the recycling endosome as nascent MMP14 or recycled via the early

endosome or early then recycling endosome. To further confirm this, MMP14 co‐

localisation with markers for this pathway, e.g. EEA1, as well as involvement of other

trafficking machinery for such events, should be tested in future experiments.

Interestingly, it appears that in breast adenocarcinoma cells endocytosis of MMP14 only

occur in the absence of a matrix substratum and it can be inhibited through plating of

cells on collagen suggesting that endocytosis might only occurs under specific conditions

[Bravo‐Cordero et al., 2007; Lafleur et al., 2006].

Trafficking via Late Endosome/Lysosome

MMP14 locates to late endosome/lysosome compartment in melanoma,

fibrosarcoma, breast and cervical adenocarcinoma cells [Clancy et al., 2015; Loskutov et

al., 2014, Macpherson et al., 2014; Monteiro et al., 2013; Steffen et al., 2008; Williams and


Coppolino, 2011; Yu et al., 2012]. Whether MMP14 can also be found in late

endosomes/lysosomes of macrophages was determine herein. MMP14 co‐localises with

LAMP1, a late endosome/lysosome marker, in RAW264.7 macrophages. In fibrosarcoma,

breast and cervical adenocarcinoma cells, MMP14 has been reported to be transported

via the late endosome/lysosome en route to the cell surface [Loskutov et al., 2014;

Macpherson et al., 2014; Marchesin et al., 2015; Monteiro et al., 2013; Steffen et al., 2008;

Williams & Coppolino, 2011; Yu et al., 2012]. In these cells, MMP14 trafficking from such

organelles to the cell surface was found to be mediated by GTPase Rab7 and Rab27, as

well as the late endosomal/lysosomal R‐SNARE VAMP7 [Macpherson et al., 2014; Steffen

et al., 2008; Williams and Coppolini, 2011]. Whether incorporation of MMP14 into the

plasma membrane occurs from these organelles or MMP14 is targeted to these

organelles for degradation was further investigated. In macrophages this route has been

shown for secretion of the actin‐binding protein flightless‐1 or the glycoside hydrolase

lysozyme from LAMP1‐positive late endosomes/lysosomes in RAW264.7 macrophages

but has not been investigated for MMP14 surface delivery [Lei et al., 2012].

VAMP7 and VAMP8 can both regulate homotypic fusion of late endosomes, while

VAMP7 also regulates late endosome to lysosome fusion [Pryor et al., 2004]. VAMP7 has

further been shown to regulate fusion of late endosome/lysosome membrane with the

cell surface during phagocytosis in RAW264.7 macrophages [Braun et al., 2004]. VAMP8

has been reported to mediate stimulated fusion of granules and secretory lysosomes

with the plasma membrane in several other immune cells, such as cytotoxic T

lymphocytes, basophils, mast cells, neutrophils and eosinophils [Dressel et al., 2010;

Lippert et al., 2007; Logan et al., 2006; Mollinedo et al., 2003; Paumet et al., 2000. To test

whether MMP14 was delivered to the cell surface via late endosomes/lysosomes

RAW264.7 macrophages were treated with VAMP7 or VAMP8 siRNA. Cells with reduced

levels of VAMP7 or VAMP8 showed significantly reduced levels of MMP14 at the cell

surface suggesting that these SNAREs are indeed involved in the incorporation MMP14

into the plasma membrane from a late endosomal/lysosomal pathway after LPS

activation. Thus, MMP14 is synthesised in the ER, transported to the Golgi complex and

on to the late endosomes/lysosomes where it is delivered to cell surface (Figure 4.10).

Transport from the Golgi complex requires VAMP4 mediated fusion and transport to or

between the late endosomes/lysosomes to the cell surface requires VAMP7 and VAMP8

(Figure 4.10). In fibrosarcoma cells, the SNARE complex VAMP7/Stx4/SNAP23 has been

proposed to mediate trafficking of MMP14 to invadopodia [Williams et al., 2014].

Whether these same SNAREs also regulate delivery to podosomes in macrophages is

tested in the next chapter.


Figure 4.11. Schematic of MMP14 trafficking in macrophages. Newly made MMP14 is trafficked from the Golgi complex to late endosome/lysosomes prior to its surface delivery, which might be regulated by VAMP4. Incorporation of newly made MMP14 into the plasma membrane through a late endosomal/lysosomal pathway and is mediated by R-SNAREs VAMP7/VAMP8.


140 Chapter 5: MMP14 and macrophage migration

Chapter 5: MMP14 and macrophage migration 5.1 Introduction

MMP14 promotes tumour growth and local invasion with its expression levels

correlating with tumour progression and metastasis [Poincloux et al., 2009]. Similarly,

MMP14 is believed to play a role in macrophage migration and infiltration of tissue

[Verollet et al., 2011]. As high numbers of macrophages contribute to prolonged

inflammation and wound chronicity it would be advantageous to elucidate mechanisms

and influence of MMP14 incorporation into the plasma membrane for its activity and

whether disruption of MMP14 cell surface delivery affects macrophage invasion.

Although the zymogen form of MMP14 is activated in the trans‐Golgi network

prior to its incorporation into the plasma membrane it is clear that the active protease

needs to be at the cell surface in order to access and degrade substrates, such as matrix

material. Incorporation into the plasma membrane is achieved through a final fusion step

mediated through SNARE complex formation at the cell surface. Evidence suggests that

incorporation of MMP14 into the plasma membrane can specifically occur at ECM‐

degrading adhesion structures called podosomes to fulfil its role in matrix degradation

and cell migration [Azzouzi et al., 2016; Wiesner et al., 2010; Wiesner et al., 2013]. In

other cell types surface Q‐SNAREs Stx4 and SNAP23 have been proposed to mediate

trafficking of MMP14 to invadopodia [Williams et al., 2014]. In macrophages,

Stx4/SNAP23 has been shown to regulate the polarised delivery of recycling endosome

to the cell surface, such that membrane and cargo are focally delivered to where they are

needed on the cell surface [Murray and Stow, 2014]. Whether this is the case for late

endosome/lysosome delivery to podosomes/cell surface in macrophages had not been

investigated. Importantly, whether disruption of this pathway can alter matrix

degradation and macrophage migration and the effect on invasion has not been tested.

Therefore, this chapter begins with the identification of the surface SNARE

machinery responsible for the final step to incorporate MMP14 into the plasma

membrane in macrophages, as well as the characterisation of the localisation of cell

surface‐delivered MMP14 upon matrix contact and studies the involvement of this

particular MMP in matrix degradation. As intracellular trafficking pathways and SNARE

machinery proteins that are responsible for incorporation of MMP14 into the plasma

Chapter 5: MMP14 and macrophage migration 141

membrane were identified and disruption of these pathways restricted cell surface

delivery of MMP14, it was tested whether disruption of these pathways would affect

matrix degradation and macrophage migration in different models of 2D and 3D

migration. The results from this chapter could provide new targets to modulate

macrophage migration to reduce inflammation and improve wound healing.


5.2 Results

5.2.1 The Q-SNAREs Stx4 and SNAP23 regulate delivery of MMP14 to the cell surface and incorporation into the plasma membrane

In the previous chapter, three R‐SNAREs responsible for the intracellular

trafficking of newly made MMP14 from the Golgi complex and late

endosomes/lysosomes towards the cell surface have been identified. Ultimately, to allow

cell surface delivery of MMP14, an R‐SNARE must form a complex with Q‐SNAREs to

regulate fusion with the plasma membrane. To identify possible regulators of MMP14

incorporation into the plasma membrane, the four main Q‐SNAREs (Stx2, Stx3, Stx4 and

SNAP23) found on the macrophage cell surface and most likely to mediate the final step

of cell surface delivery [Veale et al., 2010] were targeted for knockdown with specific

siRNA (Figure 5.1). As with the R‐SNAREs, RAW264.7 macrophages were transfected

with siRNA to these SNARE proteins on day 1, this was repeated on day 2 and then cells

were stimulated with LPS for 15 h. Knockdown was successful for all four tested surface

SNARE proteins and did not alter the levels of other tested SNARE proteins. Knockdown

efficiencies for Stx2, Stx3, Stx4 and SNAP23 were 74±6%, 50±8%, 55±19% and 20±6%,

respectively (Figure 3.12). A reduction in these surface Q‐SNAREs did not affect the

levels of total MMP14, as apparent from immunoblotting cell lysates for MMP14

following siRNA knockdown (Figure 5.1A). To determine the effect of loss of function of

these SNAREs cells were immunolabelled for surface MMP14 and analysed by

fluorescence microscopy (Figure 5.1B). Surface MMP14 levels did not significantly alter

following knockdown of Stx2 or Stx3 (Figure 5.1B). However, in macrophages treated

with Stx4 or SNAP23 siRNA MMP14 surface levels were strongly reduced (Figure 5.1B).

By flow cytometry it can be seen that the relative number of macrophages positive for

surface MMP14 did not significantly change following knockdown of Stx2 or Stx3

compared to cells treated with scrambled siRNA (P>0.05) (Figure 5.1C). In contrast,

when Stx4 levels were reduced (55%), macrophage numbers positive for surface MMP14

decreased by 28% (Figure 5.1C). Reducing SNAP23 levels by only 20% led to a 33%

decrease of the percentage of macrophages that remain positive for surface MMP14

compared to the scrambled control (Figure 5.1C). Overall, MMP14 surface levels were

reduced by 27% and 43% following Stx4 or SNAP23 siRNA treatment, respectively.

These results suggest that Stx4 and SNAP23 are the Q‐SNAREs responsible for surface

delivery of MMP14.


Figure 5.1. The Q-SNAREs Stx4 and SNAP23 regulate delivery of MMP14 to the cell surface. (A) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 15 h following transfection with scrambled (scr), Stx2, Stx3, Stx4 or SNAP23 siRNA. Whole cell lysates were assayed for total MMP14 levels by immunoblotting, with actin serving as a loading control. (B) RAW264.7 macrophages grown on coverslips were stimulated with 100 ng/ml LPS for 15 h following transfection with scrambled (scr), Stx2, Stx3, Stx4 or SNAP23 siRNA, immunostained for surface MMP14 (unpermeabilised cells) and fixed. Scale bar is 20 μm. (C) Live cells stimulated with 100 ng/ml LPS for 15 h following transfection with scrambled (scr), Stx2, Stx3, Stx4 or SNAP23 siRNA were immunostained for surface MMP14 or with an isotype control antibody (unpermeabilised cells), fixed and analysed using a FACSAriaIII flow cytometer. Bar graph shows fold change of the percentage of cells positive for surface MMP14 or fold change of MMP14 surface levels (MFI), n=3; mean±SEM, *P<0.05, **P≤0.01.


To confirm a role for Stx4 and SNAP23 in MMP14 cell surface delivery,

macrophages stimulated with 100 ng/ml LPS for 9 h were simultaneously

immunolabelled for Stx4 or SNAP23 and MMP14 to determine whether they co‐localise

at the cell surface. A partial overlap of a pool of MMP14 near the cell surface with Stx4

(Figure 5.2A) and SNAP23 (Figure 5.2B) can be observed. This suggests that Qa‐SNARE

Stx4 and Qbc‐SNARE SNAP23, perhaps as Q‐SNARE complex, mediate the final step of

MMP14 cell surface delivery.


Figure 5.2. The Q-SNAREs Stx4 and SNAP23 co-localise with MMP14 on the plasma membrane. (A) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 9 h, fixed and immunostained for MMP14 (red) and the surface SNARE Stx4 (green) to show co-localisation. Enlarged regions (lower panels) show partial overlap of MMP14 and Stx4. Scale bar is 10 μm. (B) RAW264.7 macrophages were stimulated with 100 ng/ml LPS for 9 h, fixed and immunostained for MMP14 (red) and surface SNARE SNAP23 (green) to show co-localisation. Enlarged regions (lower panels) show partial overlap at the cell surface. Scale bar is 10 μm.


5.2.2 MMP14 localises to cell structures implicated in cell migration and matrix degradation in response to matrix contact

Once delivered to the cell surface and incorporated into the plasma membrane,

MMP14 has been reported to localise to actin‐rich structures thought to assist matrix

degradation and migration [Poincloux et al., 2009; Verollet et al., 2011]. To determine

where MMP14 is localised on the macrophage cells surface, RAW264.7 macrophages

grown on gelatin‐coated coverslips for 18 h were immunostained for MMP14 and F‐actin

(Figure 5.3). Cells migrating on flat two‐dimensional matrix surfaces, such as the one

depicted in Figure 5.3, exhibit sheet‐like protrusions (lamellipodia) at the leading edge

of the plasma membrane. Lamellipodia‐based migration is characterised by increased

actin polymerisation and adhesion complex turn‐over, and rear detachment at the

trailing edge of the cell. Thin actin rich protrusions, called filipodia, can also be found at

the leading edge advancing the lamellipodium (Figure 5.3, Inset 2). Clustered behind the

leading edge at the ventral cell surface are podosomes, dynamic actin‐rich adhesion

structures that can be found in migratory cells of monocytic myeloid lineage [Murphy

and Courtneidge, 2011]. Sometimes these podosomes can be found arranged to ring‐like

structures that are termed rosettes (Figure 5.3, Inset 1). Podosomes are able to degrade

the underlying matrix through the local secretion of proteases, such as MMP14 [Murphy

and Courtneidge, 2011]. Indeed, MMP14 at the cell surface was found to localise to

podosomes of RAW264.7 macrophages that had been grown on gelatin for 18 h (Figure

5.3, Inset 1). In these cells, surface MMP14 was also found at other actin‐rich structures,

such as the leading edge, as well as trailing edge and on filopodia (Figure 5.3, Inset 2).

Together, this supports the idea of the polarised delivery of MMP14 to actin rich

podosomes implicated in macrophage migration in response to matrix contact for

degradation.


Figure 5.3. MMP14 localises to cell structures implicated in cell migration in response to matrix contact. RAW264.7 macrophages grown on gelatin-coated coverslips for 18 h were immunostained live (unpermeabilised cells) for surface MMP14 (green), fixed and then stained for F-actin (red). Insets show regions where MMP14 localises to cell membrane structures that are implicated in cell migration. Inset 1 shows MMP14 at podosomes. Inset 2 shows MMP14 at filopodia. Scale bar is 10 μm.


In order to confirm that those regions facilitate degradation of the substratum an

assay was developed. This entailed plating the cells on glass coverslips that had been

coated with a thin layer of fluorescently‐labelled gelatin matrix. As degradation of this

gelatin is accompanied by a loss of fluorescence, it allows for the detection of areas of

degradation and whether this corresponds to surface MMP14 as well as testing the effect

of inhibiting particular proteins or machinery. To investigate degradative behaviour of

RAW264.7 macrophages, cells were seeded onto coverslips coated with fluorescently

labelled gelatin, incubated for 18 h and immunostained for surface MMP14 (Figure 5.4).

It was observed that all macrophages displayed proteolytic behaviour to varying degrees

and degradation patterns, i.e. areas devoid of gelatin fluorescence, appeared punctate or

blotchy and might be attributed the secretion of enzymes from podosomes in these areas.

Interestingly, surface MMP14 also appeared in punctate patterns and intense MMP14

fluorescence signal (Figure 5.4A lower panel, arrow 1) localised to areas of strongly

reduced gelatin fluorescence, while MMP14 was absent from areas where gelatin

fluorescence remained high (Figure 5.4A lower panel, arrow 2) indicating that MMP14

is responsible for the degradation of the gelatin matrix (Figure 5.4A, lower panel). The

intensity profile (Figure 5.4B) of the line scan (dashed line in lower panel of Figure 5.4A)

demonstrates spatial separation of the intensity peaks (Figure 5.4B, arrows 1 and 2) for

gelatin and MMP14 fluorescence signal. It is important to note that MMP14 was not

detected at all areas of gelatin degradation. This could either suggest that surface MMP14

was downregulated following successful degradation of the substratum or degradation

of these areas was MMP14‐independent.


Figure 5.4. MMP14 localises to cell structures implicated matrix degradation in response to matrix contact. (A) RAW264.7 macrophages were grown coverslips coated with fluorescent gelatin (green) for 18 h, immunostained live (unpermeabilised cells) for surface MMP14 (red) and then fixed. Enlarged inset (lower panel) shows region where MMP14 localises to areas of gelatin degradation. (B) The intensity profile of the line scan (dashed line in magnified inset, lower panel) demonstrates spatially separated intensity peaks (arrows 1 and 2) of gelatin (green) and MMP14 (red) fluorescence. Scale bar is 20 μm.


5.2.3 Gelatin degradation by MMP14 can be inhibited through targeting the key SNAREs implicated in MMP14 surface delivery

Localisation of MMP14 to cell structures implicated in cell migration and matrix

degradation suggested that proficient delivery of MMP14 to the cell surface is required

for gelatin degradation by RAW264.7 macrophages. Thus, it was next tested whether

targeted siRNA knockdown of those SNARE proteins that were identified to mediate

MMP14 cell surface delivery would affect gelatin degradation. Firstly, the R‐SNARE

implicated in transport of MMP14 from the Golgi complex, VAMP4, was tested for its

ability to regulate gelatin degradation. Macrophages were transfected with siRNA to

VAMP4, along with siRNA to the early endosome and recycling endosome SNAREs

VAMP2 and VAMP3 that were shown not to regulate MMP14 surface delivery and with

scrambled siRNA as a control. These cells were seeded onto fluorescently labelled gelatin

and incubated for a further 18 h. Macrophages were fixed, stained for F‐actin and gelatin

degradation was imaged by fluorescence microscopy (Figure 5.5A). With the control

scrambled siRNA treated cells it can be seen that some areas below the cells have

reduced or no fluorescence indicating that the gelatin has been degraded (Figure 5.5A).

Similar patterns of reduced fluorescence were seen when the cells treated with VAMP2

or VAMP3 siRNA were plated on gelatin (Figure 5.5A). However, the area below cells

treated with VAMP4 siRNA had much higher levels of fluorescence and it can be seen that

there is some degradation but it was much less than the controls (Figure 5.5A). Areas of

gelatin degradation were next quantified and normalised to the area of the cells (Figure

5.5B). Proficient gelatin degradation was observed in RAW264.7 macrophages that had

been transfected with non‐targeted (scrambled) siRNA. Macrophages with reduced

levels of VAMP2 and VAMP3, which did not have significant effects on MMP14 surface

levels, showed little change in the degradation of gelatin (Figure 5.5A). The degradative

capability was reduced by 2% and 11% for VAMP2 and VAMP3, respectively (Figure

5.5B, n.s. P>0.05). However, VAMP4 knockdown, which reduced MMP14 surface levels

by 48% (Figure 4.8), diminished gelatin degradation by 54% (Figure 5.5B,

****P≤0.0001). This confirms a role for VAMP4 in the surface delivery of MMP14 and

degradation of its substrate gelatin.


Figure 5.5. Gelatin degradation by MMP14 can be inhibited through targeting the Golgi complex R-SNAREs VAMP4 or the late endosome/lysosome R-SNAREs VAMP7 or VAMP8. (A) RAW264.7 macrophages were grown on coverslips coated with fluorescent gelatin (upper panel) for 18 h following transfection with scrambled (scr), VAMP2, VAMP3, VAMP4, VAMP7 or VAMP8 siRNA, fixed and stained for F-actin (lower panel). Scale bar is 50 μm. (B) Quantification of gelatin degradation by RAW264.7 macrophages following knockdown of VAMP2, VAMP3, VAMP4, VAMP7 or VAMP8. Areas of gelatin degradation were determined and normalised to the total surface area of the cells. Bar graphs show the degradation of gelatin by RAW264.7 macrophages following knockdown of VAMP2, VAMP3, VAMP4, VAMP7 or VAMP8 relative to cells that had been transfected with scrambled (scr) siRNA. The graph shows the analysis of 200 cells per siRNA treatment across two biological repeats; mean± SEM, ****P≤0.0001 (one-way ANOVA with Dunnett’s post-hoc test).


Next, the late endosome/lysosome R‐SNAREs, VAMP7 and VAMP8, identified in

chapter 4 to regulate the transport of MMP14 to the cell surface from the late

endosome/lysosome, were tested for a role in regulating gelatin degradation.

Macrophages were transfected with siRNA to VAMP7 and VAMP8 and plated on

fluorescently labelled gelatin. Figure 5.5B shows that the area below the cells has

considerably less degradation (more fluorescence) than the control scrambled siRNA

treated cells. This was quantified and gelatinolytic activity was seen to be decreased by

45% following VAMP7 knockdown (Figure 5.5B, ****P≤0.0001), when MMP14 surface

levels were reduced by 37% (Figure 4.10). Similarly, when MMP14 surface levels were

reduced by 43% due to VAMP8 knockdown there was a 51% drop of gelatin degradation

(Figure 5.5B, ****P≤0.0001). Together these results suggest that targeting machinery

proteins in the MMP14 intracellular pathway can lead to a reduction in surface MMP14,

which in turn leads to a reduction in gelatin degradation.

Next, it was determined whether knockdown of the surface Q‐SNAREs, Stx4 and

SNAP23, that have been found mediate the final step of MMP14 surface delivery (Figure

5.1), also regulates the gelatinase activity of macrophages. Macrophages were

transfected with siRNA to Stx4 and SNAP23, as well as siRNA to the surface SNAREs Stx2

and Stx3 that were found to have no significant effect on MMP14 surface delivery (Figure

5.1) as controls, along with scrambled siRNA. Next cells were seeded onto fluorescently

labelled gelatin and incubated for a further 18 h. Macrophages were fixed, stained for F‐

actin and gelatin degradation assessed by fluorescence microscopy (Figure 5.6A). Again,

the areas below the cells treated with scrambled siRNA were reduced in fluorescence,

suggesting degradation has occurred (Figure 5.6A). Cells treated with Stx2 or Stx3 siRNA

showed similar levels of degradation to the control scrambled siRNA treated cells, but

cells treated with Stx4 or SNAP23 siRNA showed reduced degradation, as indicated by

the higher levels of fluorescence below the cells (Figure 5.6A). The areas of gelatin

degradation were quantified and normalised to the area of the cells (Figure 5.6B). Gelatin

degradation analysis of macrophages with reduced levels of Stx2 revealed that the

degradative capability in these cells was reduced marginally (10%; Figure 5.6B, n.s.

P>0.05). Macrophages with reduced levels of Stx3, which also does not hinder MMP14

surface delivery (Figure 5.1), showed no change in the degradation of gelatin (Figure

5.6B, n.s. P>0.05). Macrophages with reduced levels of Stx4 or SNAP23, which does

significantly disrupt incorporation of MMP14 into the plasma membrane (Figure 5.1),

had reduced gelatinolytic behaviour (Figure 5.6A). Knockdown of Stx4, which reduced

MMP14 surface levels by 27%, diminished gelatin degradation by a striking 41% (Figure

5.6B, ***P≤0.001). SNAP23 knockdown, where MMP14 surface levels were found to be


reduced by 43% (Figure 5.1), led to gelatinolytic activity being reduced by 32% (Figure

5.6B, **P≤0.01). Thus, Stx4 and SNAP23 regulate the final stages of delivery of MMP14 to

the cell surface and the degradation of gelatin.

To confirm that MMPs are responsible for the gelatinase activity of RAW264.7

macrophages, cells were first seeded on fluorescently labelled gelatin and incubated for

18 h in the absence or presence of GM6001, a broad‐spectrum matrix metalloproteinase

inhibitor (also known as galardin or ilomastat) (Figure 5.6C‐D). Specifically, GM6001 is

known to inhibit MMP1, MMP2, MMP3, MMP7, MMP8, MMP9, MMP12, MMP13, MMP14,

MMP15, MMP16, MMP20 and MMP26. Cells were imaged by fluorescence microscopy

(Figure 5.6C) and it can be seen that upon inhibition of MMP activity macrophages

showed reduced degradative capability when compared to control cells. Areas of

degradation were quantified and normalised to cell area, which was determined from

staining of the cell perimeter for F‐actin (Figure 5.6C, lower panel). In the presence of

GM6001, degradation was reduced by 65% (*P<0.05) when compared to the control

treated cells (Figure 5.6D), indicating that MMP activity is required for full degradative

capacity.

To address the specific involvement of MMP14 in the degradation of the gelatin matrix

by RAW264.7 macrophages, cells seeded on fluorescent gelatin were incubated for 18 h

in the presence or absence of a commercially available functionally blocking MMP14

antibody that binds to the catalytic domain of the enzyme (Figure 5.6C‐D). This antibody

was originally used to investigate MMP14‐dependent gelatinolytic activity in endothelial

cells [Galvez et al., 2001]. Fluorescence images show that gelatin degradation by

macrophages was almost absent in the presence of the MMP14 antibody (Figure 5.6C).

Quantification of degraded areas revealed that the gelatinolytic capacity of anti‐MMP14‐

treated macrophages was reduced by 97% (**P≤0.01) when compared to control cells

that were treated with an isotype‐matched antibody (Figure 5.6D). This indicates that

gelatin degradation by macrophages is dependent on functional MMP14. These results

suggest that the reduced degradation seen when Stx4, SNAP23, VAMP4, VAMP7 or

VAMP8 are knocked down is due to the reduced delivery of MMP14 to the cell surface. In

summary, upon disruption of MMP14 cell surface delivery through targeting the SNAREs

that mediate this process, namely VAMP4, VAMP7, VAMP8, Stx4 or SNAP23, MMP14‐

dependant degradation of gelatin matrix can be reduced.

154

Chapter 5: MM

P14 and macrophage m

igration


Figure 5.6. Gelatin degradation by MMP14 can be inhibited through targeting the Q-SNARE complex Stx4/SNAP23. (A) RAW264.7 macrophages were grown coverslips coated with fluorescent gelatin (upper panel) for 18 h following transfection with scrambled (scr), Stx2, Stx3, Stx4 or SNAP23 siRNA, fixed and stained for F-actin (lower panel). Scale bar is 50 μm. (B) Quantification of gelatin degradation by RAW264.7 macrophages following knockdown of Stx2, Stx3, Stx4 or SNAP23. Areas of gelatin degradation were determined and normalised to the total surface area of the cells. Bar graphs show the degradation of RAW264.7 macrophages following knockdown of Stx2, Stx3, Stx4 or SNAP23 relative to cells that had been transfected with scrambled (scr) siRNA Analysis of 200 cells per siRNA treatment across two biological repeats; mean ± SEM, **P≤0.01, ***P≤0.001 (one-way ANOVA with Dunnett’s post-hoc test. (C) RAW264.7 macrophages were grown on coverslips coated with fluorescent gelatin (upper panel) for 18 h in the presence or absence of GM6001 or anti-MMP14 antibody, fixed and stained for F-actin (lower panel). Scale bar is 50 μm. (D) Quantification of gelatin degradation by RAW264.7 macrophages in the presence and absence of GM6001 or anti-MMP14 antibody. Areas of gelatin degradation were determined and normalised to the total surface area of the cells. Bar graphs show the degradation of RAW264.7 macrophages in the presence of anti-MMP14 antibody relative to control cells grown in the presence of an isotype control antibody. n=3; mean ± SEM, *P<0.05, **P≤0.01 (unpaired, two-tailed Student’s t test).

5.2.4 3D migration of macrophages through different matrices

During migration into a wound macrophages will have to perform 3D migration

through tissue. The fact that matrix degradation is MMP14 dependant and can be

disrupted by targeting SNARE machinery involved in the cell surface delivery of MMP14

suggests that this could also be true for macrophage migration through 3D matrices. To

be able to test the effect of disrupting intracellular MMP14 trafficking on mesenchymal

migration of macrophages, it was first necessary to develop an assay that allows the

analysis of macrophage migration through a matrix along a chemotactic gradient. For

this, RAW264.7 macrophages were seeded onto the bottom side of a Transwell®

membrane where the upper well contains a thick layer of Matrigel™ or collagen I matrix

with media containing the chemoattractant fMLP placed on the matrix. Cells were

allowed to migrate for 3 days through Matrigel™ or collagen I matrix. After the 3 days,

macrophages in the Transwell® were fixed, stained for F‐actin and visualised by

confocal microscopy to assess macrophage migration (Figure 5.7). At the bottom of the

gel matrix, macrophages were found passing through the randomly distributed pores of

the membrane (Figure 5.7A). Within the gel, invading cells were found in grape‐like

clusters across the gel (Figure 5.7B). These macrophages that have successfully invaded

the gel exhibited actin‐rich protrusions that resembled filopodia and lamellipodia

(Figure 5.7C and 5.7D) and actin‐rich podosomes near the leading edge (Figure 5.7D).

MMP14 was found to be enriched in these 3D podosomes (Figure 5.8E). As macrophages

successfully migrated into 3D matrices, both Matrigel™ and collagen I, and exhibit a

mesenchymal phenotype with MMP14 localising to cell structures implicated in

migration, the assay was considered adequate for studying the effects of manipulating

MMP14 trafficking pathways on macrophage migration in to 3D matrices.



Figure 5.7. 3D migration of macrophages through Matrigel™ and collagen I matrices. (A) An inverted invasion assay performed to assess 3D migration of RAW264.7 macrophages. Macrophages were allowed to invade into a collagen I plug for 72 h along a gradient of fMLP. Cells were stained with TRITC-phalloidin (red) and macrophages passing through pores of the membrane (grey) were imaged by bright field and confocal microscopy. Scale bar is 100 μm. (B) Invading cells within a Matrigel™ plug were stained with TRITC-phalloidin and visualised by confocal microscopy. Serial optical sections of the gel plugs were captured at 3 μm intervals and a typical example of groups of invaded cells shown as a Maximum Intensity Projection; scale bar is 100 μm. (C-D) An inverted invasion assay was performed to assess the morphology of RAW264.7 macrophages migrating through collagen I. Macrophages were allowed to invade into a collagen plug for 72 h along a gradient of fMLP. Invading cells (within the gel plug) were stained with TRITC-phalloidin and morphology visualised by confocal microscopy. Scale bar is 25 μm. (E) An inverted invasion assay was performed to assess MMP14 localisation in RAW264.7 macrophages during 3D migration. Macrophages were allowed to invade into a collagen plug for 72 h along a gradient of fMLP. Invading cells (within the gel plug) were stained with MMP14 (green) and TRITC-phalloidin (red) and the morphology visualised by confocal microscopy. Scale bar is 10 μm.

5.2.5 SNAREs regulate macrophage invasion of Matrigel™

In order for macrophages to migrate into a wound site, monocytes recruited from

the blood stream are first required to emigrate from the blood vessel into the

surrounding tissue in a process called diapedesis. Monocyte migration through the

vessel walls involves transmigration through the endothelium, as well as degradation of

the associated basement membrane. Transendothelial migration of monocytes is

MMP14‐dependent [Matias‐Roman et al., 2005; Gonzalo et al., 2010; Sithu et al., 2007].

However, there is conflicting data about whether degradation of the basement

membrane by monocytes/macrophages is MMP‐dependant [Agrawal et al., 2008;

Murray et al., 2013] or MMP‐independent [Van Goethem et al., 2010]. Matrigel™ is a gel‐

like mixture of ECM proteins, mainly laminin, collagen IV, heparin sulphate

proteoglycans and entactin, secreted by a mouse sarcoma cell line that has been widely

used to mimic the basement membrane. It is therefore a good model to study the effect

of targeting MMP14 cell surface delivery on the ability of macrophages to cross the

basement membrane.

Hence, it was tested whether reducing the levels of SNARE machinery identified

as being involved in the intracellular trafficking and surface delivery of MMP14 would

affect macrophage migration through Matrigel™. As knockdown of R‐SNAREs VAMP4,

VAMP7 and Q‐SNAREs Stx4 and SNAP23 had dramatic effects on both, MMP14 cell

surface delivery and degradative function, the effect of reducing the levels of these

SNAREs was investigated. RAW264.7 macrophages were transfected with non‐targeted

(scrambled), VAMP4, VAMP7, Stx4 or SNAP23 siRNA on day 1, this was repeated on day

2 and then cells were seeded onto the bottom side of a Transwell® membrane and


allowed to migrate for 3 days through a thick layer of undiluted Matrigel™ along a

chemotactic gradient of fMLP. After the 3 days, cells were fixed, stained for F‐Actin and

visualised by confocal microscopy (Figure 5.8). Images taken before the Transwell®

membrane show that cell numbers between different siRNA treatment groups

(scrambled, VAMP4, VAMP7, Stx4 or SNAP23) remained similar (Figure 5.8A, first

panels). Macrophages treated with scrambled siRNA were found to invade into the gel in

grape‐like clusters (Figure 5.8A) as described in section 5.2.4. These clusters of varying

sizes were found all across the gel and invaded on average about 140 µm into the gel.

Invasion of the overall macrophage population was quantified by adding up the area

covered by invading cells for each section as an indicator of cell numbers across a certain

distance. When disrupting MMP14 trafficking from the Golgi complex through siRNA

targeted knockdown of R‐SNARE VAMP4, macrophages were found to invade the gel only

to about 100 µm as apparent from comparing sequences of optical sections from invaded

cells (Figure 5.8A). The size of the macrophage clusters within the gel were also much

smaller as less cells passed through the Transwell® membrane (Figure 5.8A). Overall,

invasion of macrophages that had been treated with VAMP4 siRNA was reduced by 99%

in comparison to the control (Figure 5.8B). Macrophages treated with siRNA targeting

the late endosomal/lysosomal R‐SNARE VAMP7 invaded only about as far as 60 µm into

the gel and overall much less cells were found to have invasive potential. When

quantified invasion was found to be reduced by 95% following treatment with VAMP7

siRNA (Figure 5.8B). When disrupting incorporation of MMP14 into the plasma

membrane through targeting of surface Q‐SNAREs Stx4 or SNAP23 macrophages did not

invade as far (85 and 65 µm, respectively) into the Matrigel™ as control cells (Figure

5.8A). For the whole population, invasion was diminished by over 90% or 94% following

treatment with Stx4 or SNAP23 siRNA, respectively (Figure 5.8B). This suggests that

targeting these SNAREs, which are involved in the cell surface delivery of MMP14,

impedes 3D migration of RAW264.7 macrophages through Matrigel™.

To confirm that the observed effect of diminished macrophage invasion

following SNARE knockdown is a result of reduced MMP14 activity at the cell surface, 3D

migration of macrophages through Matrigel™ in the absence or presence of MMP

inhibitor GM6001 or anti‐MMP14 antibody was assessed. Comparing sequences of

optical sections from invaded cells it becomes clear that GM6001‐ or MMP14‐antibody‐

treated macrophages did not invade as far into the Matrigel™ as untreated cells (Figure

5.8C). While the depth of the clusters of invaded cells was highly variable, on average

control macrophages were found to migrate 145 μm into the Matrigel™ within the three

days while GM6001‐ or MMP14‐antibody‐treated cells invaded less than 65 μm within


the 3 days. Overall invasion was reduced by 77% (**P≤0.01) or 78 % (**P≤0.01)

following treatment with GM6001‐ or MMP14‐antibody, respectively (Figure 5.8D). This

means generally less macrophages invaded the gel upon inhibition of MMPs or MMP14

and those cells that did invade into the Matrigel™ plug were unable to travel as far.

Taken together, the data suggests that MMP14 is necessary for efficient

infiltration of 3D Matrigel™. Importantly, knockdown of key SNARE proteins identified

to regulate MMP14 transport to the cell surface lead to a loss in the ability to efficiently

migrate through Matrigel.

160

Chapter 5: MM


igration

Chapter 5: M

MP14 and m

acrophage migration

161

Figure 5.8. 3D migration through Matrigel™ can be inhibited through targeting SNAREs that are involved in the cell surface delivery of MMP14. (A) An inverted invasion assay was performed to compare 3D migration of RAW264.7 macrophages through Matrigel™ following their transfection with scrambled (scr), VAMP4, VAMP7, Stx4 or SNAP23 siRNA. Macrophages were allowed to invade into a Matrigel™ plug for 72 h along a gradient of fMLP. Invading (within the plug) and non-invading cells (on the membrane) were stained with TRITC-Phalloidin and visualised by confocal microscopy. Serial optical sections of the Matrigel™ plug were captured at 30 μm intervals. Typical examples are presented as a sequence in which the individual optical sections are placed alongside one another with increasing depth from left to right, as indicated. (B) Bar graph represents invasion of macrophages following transfection with scrambled (scr), VAMP4 or SNAP23 siRNA quantified by adding up the area covered by invading cells (as an indicator of cell numbers) for each section and is expressed relative to the control; VAMP4 and SNAP23 n=2, VAMP7 and Stx4 n=1. (C) An inverted invasion assay was performed to compare 3D migration through Matrigel™ of RAW264.7 macrophages in the absence or presence of MMP inhibitor GM6001 or anti-MMP14 antibody. Macrophages were allowed to invade into a Matrigel™ plug for 72 h along a gradient of fMLP. Invading (within the plug) and non-invading cells (on the membrane) were stained with TRITC-Phalloidin and visualised by confocal microscopy. Serial optical sections of the Matrigel™ plug were captured at 15 μm intervals and typical examples are presented as a sequence in which the individual optical sections are placed alongside one another with increasing depth from left to right, as indicated. (D) Bar graph represents invasion of macrophages in the absence or presence of MMP inhibitor GM6001 or anti-MMP14 antibody quantified by adding up the area covered by invading cells for each section and is expressed relative to the control; n=3, mean± SEM, **P≤0.01 (one-way ANOVA with Dunnett’s post-hoc test.


5.2.6 3D migration through collagen I can be inhibited through targeting SNAREs that are involved in the cell surface delivery of MMP14

After having emigrated from blood vessels, macrophages have to migrate

through a 3D network of interstitial type I collagen on their way to the wound site. In

order to move through dense matrix material macrophages create a path by

proteolytically degrading the matrix. In macrophages, the proteolytic processing during

this mesenchymal mode of migration has been shown to occur at cell protrusions where

3D podosomes form [Van Goethem et al., 2010; Van Goethem et al., 2011]. To study the

role of SNARE‐mediated trafficking of MMP14 in migration of macrophages through

collagen, RAW264.7 cells were transfected with non‐targeted (scrambled), VAMP4,

VAMP7, Stx4 or SNAP23 siRNA on day 1, this was repeated on day 2 and then cells were

seeded onto the bottom side of a Transwell® membrane and allowed to migrate for 3

days through a thick layer of gelled collagen I along a chemotactic gradient of fMLP. After

the 3 days, cells were fixed, stained for F‐Actin and visualised by confocal microscopy

(Figure 5.9).

Images taken before the Transwell® membrane show that cell numbers between

different siRNA treatment groups (scrambled, VAMP4, VAMP7, Stx4 or SNAP23)

remained similar (Figure 5.9A, first panels). As before, macrophages treated with

scrambled siRNA were found to invade into the collagen gel in grape‐like clusters (Figure

5.9A). These clusters of varying sizes were found all across the gel and invaded on

average about 94 µm into the gel. Invasion of the overall macrophage population was

quantified by adding up the area covered by invading cells for each section as an

indicator of cell numbers across a certain distance. First, the effect of reducing levels of

R‐SNARE VAMP4 was investigated as this had a dramatic effect on MMP14 cell surface

delivery as well as degradative function. When disrupting MMP14 trafficking from the

Golgi complex through siRNA targeted knockdown of R‐SNARE VAMP4, macrophages

were found to invade the gel only to about 60 µm as determined from comparing

sequences of optical sections from invaded cells (Figure 5.8A). The size of the

macrophage clusters within the gel were also much smaller as less cells passed through

the Transwell® membrane (Figure 5.8A). Overall, invasion of macrophages that had

been treated with VAMP4 siRNA was reduced by 72% (P<0.05) in comparison to the

control (Figure 5.8B). Similarly, macrophages treated with siRNA targeting the late

endosomal/lysosomal R‐SNARE VAMP7 invaded only about as far as 74 µm into the gel

and overall much less cells were found to have invasive potential. When quantified

invasion was found to be reduced by 59% (n.s.) following treatment with VAMP7 siRNA


(Figure 5.8B). The effect of reducing levels of the Q‐SNAREs Stx4 and SNAP23 was also

investigated, as these SNAREs were found to be responsible for incorporation of MMP14

into the plasma membrane, and, as such, regulate the degradative function of

macrophages. It was observed that depth of invasion into collagen I of invading

macrophages was reduced to 60 and 58 µm following transfection with Stx4 or SNAP23

siRNA, respectively, compared to the scrambled control (Figure 5.9A). Invasiveness of

macrophages overall was diminished by 57% following treatment with Stx4 siRNA (n.s.

P>0.05) while it was impacted by 78% upon SNAP23 knockdown (**P≤0.01; Figure

5.9B).

Using the same inverted migration assay, RAW264.7 cells were allowed to

migrate through a thick layer of gelled collagen I in the absence or presence of either

MMP inhibitor GM6001 or MMP14 antibody to confirm that the observed effect on

macrophage chemotactic migration is due to altered MMP14 levels. Treatment with

inhibitor or antibody did not negatively affect cell viability as cell numbers of non‐

invading cells, i.e. cells that did not cross the membrane, between control and treated

groups were similar (Figure 5.9C, first panel). Comparing sequences of optical sections

of the gels it becomes clear that GM6001‐ or MMP14‐antibody‐treated macrophages only

invade about half as far into the collagen I plug compared to untreated cells (Figure 5.9C).

Relative invasion of the macrophage population was quantified by adding up the area

covered by invading cells for each section and revealed that invasion was reduced by

87% (**P≤0.01) or 91% (**P≤0.01) following treatment with GM6001‐ or MMP14‐

antibody, respectively (Figure 5.8E). This data suggests that MMP14 is necessary for

efficient migration through 3D collagen I.

In summary, some macrophages continued to migrate into the collagen I matrix

following SNARE knockdown. However, significantly less macrophages had invasive

properties when pathways responsible for trafficking of MMP14 to the plasma

membrane were disrupted. Therefore, targeting SNAREs that are involved in the cell

surface delivery of MMP14, namely VAMP4, VAMP7, Stx4 and SNAP23, hampers MMP14‐

dependent 3D migration of RAW264.7 macrophages through collagen I.

164

Chapter 5: MM


igration

Chapter 5: M

MP14 and m

acrophage migration

165

Figure 5.9. 3D macrophage migration through collagen can be inhibited through targeting SNAREs that are involved in cell surface delivery of MMP14. (A) An inverted invasion assay was performed to compare 3D migration of RAW264.7 macrophages through collagen I following transfection with scrambled (scr), VAMP4, VAMP7, Stx4 or SNAP23 siRNA. Macrophages were allowed to invade into a collagen I plug for 72 h along a gradient of fMLP. Invading (within the plug) and non-invading cells (bottom of the membrane) were stained with TRITC-Phalloidin and visualised by confocal microscopy. Serial optical sections of the collagen I plug were captured at 30 μm intervals and typical examples are presented as a sequence in which the individual optical sections are placed alongside one another with increasing depth from left to right, as indicated. (B) Bar graph represents invasion of macrophages following transfection with scrambled (scr), VAMP4, VAMP7, Stx4 or SNAP23 siRNA quantified by adding up the area covered by invading cells for each section and is expressed relative to the control; n=3 for VAMP4 and VAMP7, n=4 for Stx4 and SNAP23, mean± SEM, *P<0.05, **P≤0.01 (one-way ANOVA with Dunnett’s post-hoc test). (C) An inverted invasion assay was performed to compare 3D migration through collagen I of RAW264.7 macrophages in the absence or presence of MMP inhibitor GM6001 or anti-MMP14 antibody. Macrophages were allowed to invade into a collagen I plug for 72 h along a gradient of fMLP. Invading (within the plug) and non-invading cells (before the membrane) were stained with TRITC-Phalloidin and visualised by confocal microscopy. Serial optical sections of the collagen I plug were captured at 15 μm intervals and typical examples are presented as a sequence in which the individual optical sections are placed alongside one another with increasing depth from left to right, as indicated. (D) Bar graph represents invasion of macrophages in the absence or presence of MMP inhibitor GM6001 or anti-MMP14 antibody quantified by adding up the area covered by invading cells for each section and is expressed relative to the control; n=3, mean± SEM, **P≤0.01 (one-way ANOVA with Dunnett’s post-hoc test).


5.3 Discussion

Macrophage infiltration of wound tissue during the inflammatory phase is part

of normal wound healing. However, increased numbers of macrophages leading to

excessive inflammation are associated with poor wound healing outcomes. The influx of

macrophages into the tissue is believed to be dependent on the activity of MMP14. In this

chapter, SNARE machinery responsible for the incorporation of MMP14 into the plasma

membrane was identified. Furthermore, it was determined whether the SNARE proteins

that were identified to regulate the transport of MMP14 to the cell surface also regulate

degradation and invasion.

Initially, two Q‐SNAREs, namely Stx4 and SNAP23, were identified to mediate the

incorporation of MMP14 into the plasma membrane of macrophages. It was observed

that the incorporation of MMP14 at the cell surface does not occur at random but that

MMP14 localises to filopodia and podosomes, F‐actin‐rich cell membrane structures

implicated in cell migration, in response to matrix contact. Once MMP14 is specifically

delivered to these regions it facilitates matrix degradation and macrophage movement.

Delivery of MMP14 to areas of matrix degradation was found to be reliant on the action

of specific SNARE proteins. MMP14‐dependant degradation of gelatin could be repressed

through targeting SNARE proteins that were previously identified as mediators of

MMP14 cell surface delivery, namely VAMP4, VAMP7, VAMP8, Stx4 and SNAP23. 3D

migration of macrophages through different ECM materials, thick layers of Matrigel™ or

gelled collagen I, was studied using an inverted invasion assay. Infiltration of these gel

plugs by macrophages occurred in grape‐like cell clusters. Individual cells that

successfully invaded the gel matrix exhibited actin‐rich protrusions that resembled

filopodia and 3D podosomes. The numbers of macrophages that were able to migrate

into gels of Matrigel™ or collagen I as well as the distance covered by invading cells was

dependent on MMP14 activity at the cell surface. As a consequence, disrupting SNARE

machinery that is required for the incorporation of MMP14 into the plasma membrane

significantly reduced the ability of macrophages to effectively invade into both,

Matrigel™ and collagen I gels. Thus, targeting these SNAREs might allow attenuation of

macrophage tissue infiltration and inflammation.


5.3.1 Trafficking to the cell surface

Ultimately, to allow cell surface delivery of MMP14 an R‐SNARE must form a

complex with Q‐SNAREs to regulate fusion with the plasma membrane. In various cell

types, the Q‐SNAREs Stx4 and SNAP23 mediate this final fusion event to deliver MMP14

to the cell surface [Kean et al., 2009; Miyata et al., 2004; Williams et al., 2014].

Knockdown of the four main surface Q‐SNAREs (Stx2, Stx3, Stx4 and SNAP23) found on

the macrophage cell surface confirmed that Stx4 and SNAP23 mediate MMP14 cell

surface delivery in macrophages (Figure 5.1). Co‐localisation of MMP14 with Stx4 and

SNAP23 at the cell perimeter in RAW264.7 cells provided final clues that these SNAREs

regulate MMP14 incorporation into the plasma membrane in macrophages.

Protein cargo can be delivered from the Golgi complex to the cell surface directly

and R‐SNARE VAMP4 is known to interact with the Qa‐SNARE Stx4 in neuroendocrine

pheochromocytoma (PC12) cells [Thomas et al., 1999], two‐hybrid systems [Rolland et

al., 2014; Rual et al., 2005] and affinity capture mass spectrometry of human embryonic

kidney cells [Huttlin et al., 2015]. Furthermore, VAMP4 has been found to interact with

Qbc‐SNARE SNAP23. This opens up the possibility of a VAMP4/Stx4/SNAP23 complex

and therefore transport of MMP14 from the Golgi complex to the cell surface directly.

However, VAMP4 has also been shown to regulate delivery from lysosome‐related

organelles to the cell surface in activated NK cells [Krzewski et al., 2011]. As newly made

MMP14 has also been found in late endosomes/lysosomes, it could be more likely that

VAMP4 is regulating the transport of MMP14 from the Golgi complex to late

endosomes/lysosomes. Subsequently, incorporation of newly made MMP14 into the

plasma membrane from a late endosomal/lysosomal pathway could be regulated by a

possible complex of VAMP7 (R‐SNARE), Stx4 (Qa‐SNARE) and SNAP23 (Qbc‐SNARE)

(Figure 5.10). A VAMP7/Stx4/SNAP23 complex has previously been found to allow

MMP14 cell surface delivery, as well as targeting it to invasive structures and is required

for cell invasion of MDA‐MB‐231 cells [Williams et al., 2014].


Figure 5.10. Schematic of MMP14 trafficking to podosomes in macrophages. In response to external stimuli, MMP14 is synthesised and then transported through a biosynthetic pathway via the Golgi complex as well as late endosomes and lysosomes to the plasma membrane. Cell-matrix contact, facilitated through integrin-mediated adhesion (orange), initiates the assembly of podosomes, which are specialised actin-rich membrane protrusions of invading cells that facilitate matrix degradation. The microtubule (green) and actin (red) cytoskeleton and associated scaffolding proteins are believed to be involved in the polarised delivery of MMP14 to podosome sites but were not further investigated in this study. The R-SNARE protein VAMP7 (pink) was found to regulate delivery of MMP14 (olive) to podosomes, indicating that the SNARE machinery regulates the fusion of MMP14-containing vesicles with the plasma membrane. The plasma-membrane Q-SNAREs Stx4 and SNAP23 (aqua) were also identified to mediate MMP14-dependant matrix degradation and migration.


5.3.2 Localisation of MMP14 to structures for migration and degradation

Following cell‐matrix contact MMP14 was found to localise to F‐actin rich

membrane structures that resembled podosomes and filopodia, structures that are

implicated in matrix degradation and cell migration (Figure 5.10). Specific localisation of

MMP14 to lamellipodia and filopodia has previously been shown in human peripheral

blood mononuclear cells following attachment to fibronectin [Matias‐Roman et al.,

2005]. The polarised delivery of MMP14 to the leading edge was also previously shown

for monocytes migrating over activated endothelial cells through co‐localisation staining

with leading edge marker profilin [Matias‐Roman et al., 2005]. Furthermore, MMP14

localisation to podosomes has previously been reported for primary macrophages, as

well as RAW264.7 cells, and is crucial for matrix degradation at podosomes [Nusblat et

al., 2011; Wiesner et al., 2010]. Previously, degradation of fibronectin, collagen I and

gelatin has been shown to correlate with the presence of overexpressed MMP14 at the

cell surface of primary macrophages [Wiesner et al., 2010]. The study herein showed

endogenous MMP14 localised to areas of gelatin degradation, indicating that MMP14

may be responsible for the degradation of the gelatin matrix. The pattern of gelatin

degradation suggested podosome formation in these areas (Figure 5.10). However, to be

fully confident, staining for podosome‐specific markersan such as cortactin, talin, TKS5,

vinculin or N‐WASP would have to be conducted.

The recruitment of MMP14‐containing vesicles towards podosome sites occurs

via podosome‐contacting microtubules and motorproteins KIF5B (kinesin‐1 subunit)

and KIF3A/KIF3B (kinesin‐2 subunits) in primary macrophages [Wiesner et al., 2010].

In breast adenocarcinoma cells it has been show that the GTP‐binding protein ARF6 at

the plasma membrane interacts through the scaffold effector proteins JIP3/JIP4 with

kinesin‐1 to control the positioning and tubulation of MMP14‐positive endosomes.

Activation of the actin cytoskeleton regulator Arp2/3 complex, which is dependent on

actin‐regulator WASH, further coordinates the delivery of MMP14 to invadopodia

[Marchesin et al., 2015; Monteiro et al., 2013]. Furthermore, WASH interacts with the

exocyst complex to tether MMP14‐positive late endosomes to the target membrane in

the same cells [Monteiro et al., 2013]. Finally, delivery of MMP14 to podosomes was

found to be dependent on N‐WASP, inducer of actin polymerisation, in RAW264.7

macrophages [Nusblat et al., 2011]. N‐WASP‐mediated tethering of MMP14’s

cytoplasmic tail to F‐actin was found to be necessary to stabilise MMP14 to invasive

structures and induce matrix degradation in breast cancer cells [Yu et al., 2012].


Others have found different SNARE trafficking machinery proteins that also

regulate MMP14 transport and membrane docking, such as the late

endosomal/lysosomal SNARE protein VAMP7 and plasma membrane SNAREs Stx4 and

SNAP23, which were found to mediate the delivery of MMP14 to invadopodia in breast

adenocarcinoma cells [Steffen et al., 2008; Williams et al., 2014]. However, the role of

these or other SNAREs in the delivery of MMP14 to podosomes and the effect of blocking

their function on matrix degradation by macrophages has not been investigated.

5.3.3 Matrix degradation by macrophages can be inhibited through targeting SNAREs that are involved in the cell surface delivery of MMP14

As MMP14 is responsible for gelatin degradation by RAW264.7 macrophages, it

was next tested whether disruption of MMP14 trafficking through targeted siRNA

knockdown of those SNARE proteins that were identified to mediate incorporation of

MMP14 into the plasma membrane would affect gelatin degradation. And indeed, upon

disruption of MMP14 cell surface delivery through targeting VAMP4, VAMP7, VAMP8,

Stx4 or SNAP23, MMP14‐dependant degradation of gelatin matrix was significantly

reduced. MMP14 was found to co‐localise with Stx4 and SNAP23 at the plasma

membrane of macrophages in this study herein. MMP14 has also previously been shown

to co‐localise with VAMP7, Stx4 and SNAP23 at focal sites of matrix degradation and

targeting these SNAREs was found to impede delivery of MMP14 to podosomes and

MMP14‐mediated gelatin degradation by breast adenocarcinoma cancer [Steffen et al.,

2008; Williams et al., 2014] endorsing the results seen for RAW264.7 macrophages

(Figure 5.10). In fibrosarcoma cells, disrupting trafficking facilitated by VAMP3 also

negatively affected gelatin proteolysis [Kean et al., 2009]. However, proteolytic activity

was only reduced by about 60% and could therefore be attributed to recycled pools of

MMP14 rather than newly made protein. In RAW264.7 macrophages, reducing VAMP3

levels did not significantly affect gelatinolytic activity. This is not surprising as MMP14

surface levels were not found to be decreased in RAW264.7 macrophages with reduced

VAMP3. A role for VAMP4 and VAMP8 in the regulation of MMP14‐dependent matrix

degradation has not previously been reported. In macrophages, studies that investigate

the effect of disrupting trafficking machinery on matrix degradation have so far been

limited to Rab proteins only [Wiesner et al., 2013]. Thus, the here identified SNAREs

proteins represent novel targets to modify the degradative capacity of macrophages. It

was next tested whether targeting the SNAREs would also influence macrophage

migration.


5.3.4 3D migration of macrophages can be inhibited through targeting SNAREs that are involved in the cell surface delivery of MMP14

In vivo monocytes are recruited through adhesion to the activated endothelium

of the blood vessel, rolling then crawling over it, before subsequent diapedesis and

emigration through the basement membrane to then migrate towards the wound

[Verollet et al., 2011]. Transendothelial migration of murine bone marrow‐derived

macrophages and human peripheral blood monocytes has been shown to be MMP14‐

dependent based on knockout, knockdown or functional blocking of MMP14 [Gonzalo et

al., 2010; Matias‐Roman et al., 2005; Sithu et al., 2007]. In an ex vivo model using

mesothelial and epithelial basement membrane, it was shown that the activity of MMP14

and other transmembrane MMPs are responsible for basement membrane breaching by

human breast adenocarcinoma cells [Hotary et al., 2006]. It was therefore interesting to

characterise the role of SNARE machinery proteins in the migration of RAW264.7

macrophages through a basement membrane‐like gel and whether basement membrane

degradation could be manipulated by targeting SNAREs that are involved in MMP14

trafficking. For this, macrophage migration through Matrigel™ was assessed.

In dense matrices such as Matrigel™ and gelled collagen I macrophages use the

mesenchymal migration mode, which is characterised by the requirement of proteases

to create their path. Through the use of proteases macrophages create tunnels when

migrating through an extracellular matrix [Van Goethem et al., 2011]. This might explain

why macrophages were often found as cell clusters within the gel: Once a path has

already been established within the gel, other macrophages might follow as a result of a

principle of least effort. RAW264.7 macrophages were found to migrate on average 145

μm into the Matrigel™ within the three days, which corresponds to an average velocity

of about 30 nm per minute. As the basement membrane is only about 100 nm in

thickness [Kalluri, 2003], breaching of the basement membrane could be expected to

occur in only a few minutes. Considering that Matrigel™ is less crosslinked and more

vulnerable to degradation compared to native basement membrane [Poincloux et al.,

2009], this time frame is in agreement with in vivo imaging studies showing that the

majority of leukocytes extravasation takes between 10 and 20 min [Hyun et al., 2012].

When reducing the levels of SNARE proteins that are involved in the cell surface delivery

of MMP14, namely VAMP4, VAMP7, Stx4 and SNAP23, MMP14‐dependent 3D migration

of RAW264.7 macrophages through Matrigel™ was dramatically impaired. The observed

effect was similar to treatment of RAW264.7 macrophages with GM6001 or function

blocking MMP14‐antibody. This suggests a direct involvement of MMP14 in the


proteolytic degradation process rather than it being only required for the activation of

secreted MMPs, as previously been reported for basement membrane breaching of

breast adenocarcinoma cells [Hotary et al., 2006]. It also shows the influence of

disrupting the SNARE machinery is specific to disturbed MMP14 trafficking. A negative

effect on invasion through Matrigel™ following the inhibition of VAMP7, Stx4 and

SNAP23 has previously been shown for breast adenocarcinoma and fibrosarcoma cells

[Steffen et al., 2008; Williams and Coppolino, 2011; Williams et al., 2014]. VAMP4 has not

yet been reported to regulate migration through Matrigel™ in any cell type. Studies

examining the effect of disrupting MMP14‐related trafficking machinery on macrophage

penetration of Matrigel™ have not been described either.

It is important to note that following SNARE‐targeted siRNA treatment some

macrophages were still able to penetrate a basement membrane‐resembling gel. This

could be owed to the fact that siRNA knockdown was not complete and some

macrophages were known to still be positive for MMP14 at the cell surface even when

overall levels were reduced. However, macrophage movement appeared to be much

slower. Thus, it could be hypothesised that the extravasation process in vivo would

possibly be delayed. Overall there were also a smaller number of macrophages that

retained their invasive properties upon inhibition of MMP14 meaning that potentially

less macrophages would be able to infiltrate the tissue in vivo. Taken together, this

demonstrates that MMP14 is necessary for efficient migration through a basement

membrane‐resembling matrix and targeting of SNARE machinery that was identified

here to be involved in the cell surface delivery of MMP14 could potentially reduce

macrophage extravasation from the bloodstream into the surrounding tissue.

Following the successful emigration from the blood vessel by diapedesis,

macrophages will then have to migrate through the interstitial tissue, which is a complex

3D environment made from different extracellular matrix components, primarily type I

collagen, and can vary in architecture and rigidity. Depending on the architecture of the

matrix, macrophages can either use the amoeboid mode, which does not use proteases,

to migrate through materials with large pores such as pepsin‐extracted fibrillar collagen

I or the use the mesenchymal mode, which is dependent on protease activity, to migrate

through dense matrices, such as gelled collagen I [Cougoule et al., 2010; Van Goethem et

al., 2010]. When using the mesenchymal mode, macrophages form 3D podosomes at the

tip of cell [Van Goethem et al., 2010; Van Goethem et al., 2011]. Accordingly, RAW264.7

macrophages that successfully invaded into gels of collagen I also exhibited actin‐rich

protrusions that resembled filopodia and 3D podosomes as would be found in cells using


the mesenchymal mode of migration. It is known that proteolytic degradation of the

matrix takes place at these protrusions and the secretion of MMPs and other proteases

is believed to occur at podosomes [Van Goethem et al., 2011; Wiesner et al., 2013;

Wiesner et al., 2014]. Whether the actin‐rich structures are indeed podosomes would

have to be confirmed through the staining with known podosome markers, such as

vinculin or paxillin. It was possible to confirm the localisation of MMP14 to these 3D

podosomes in this model. In human peripheral blood monocytes embedded in collagen I

MMP14 was shown to be enriched at F‐actin‐rich cell protrusions resembling 3D

podosomes, which locally degraded the collagen matrix [Wiesner et al., 2013].

It was therefore interesting to characterise the role of SNARE machinery proteins

in the migration of RAW264.7 macrophages through 3D collagen I and whether it could

be manipulated by targeting SNAREs that are involved in MMP14 trafficking. Indeed,

targeting SNARE machinery involved in the cell surface delivery of MMP14, namely

VAMP4, VAMP7, Stx4 and SNAP23, hampered MMP14‐dependent 3D migration of

RAW264.7 macrophages in an inverted 3D invasion assay using a gel of native collagen

I, as significantly less RAW264.7 macrophages had invasive properties. Similarly,

reduced invasion of RAW264.7 macrophages in the presence of function blocking

MMP14 antibody supports the idea that this particular MMP is essential for the invasion

of macrophages into collagen I‐rich matrices but, more importantly, shows that MMP14‐

mediated migration can be specifically targeted by disrupting the mentioned trafficking

machinery. Although the role of SNARE machinery in 3D migration of macrophages has

not previously been elucidated, primary macrophage invasion into 3D collagen gels has

been shown to be inhibited when Rab trafficking machinery responsible for the cell

surface delivery of MMP14 was disrupted [Wiesner et al., 2013]. Rab GTPases initiate

vesicle docking and facilitate SNARE complex formation, while formation of the SNARE

complex mediates membrane fusion. Together, Rabs and SNAREs regulate selective

delivery of vesicle cargo to the proper organelle or target membrane.

The reported findings on the effects of disrupting MMP14 trafficking are not

surprising considering that intact collagen fibres, due to their triple‐helical structure, can

only be cleaved by a limited number of proteases, which include some of the MMPs,

cathepsins and a few serine proteinases [Song et al., 2006]. In terms of the MMPs, fibrillar

collagen cleavage can only be initiated by MMP1, MMP8, MMP13 and MMP14, while

fragments of the cleaved fibrillar collagen may subsequently be further degraded by

MMP2 and MMP9 [McKleroy et al., 2013]. In a study with human monocyte‐derived

classically activated macrophages, MMP14 was expressed at much higher levels than


MMP1 and MMP8, while MMP13 was not expressed at all [Huang et al., 2012].

Furthermore, MMP14 proteolytically activates MMP‐8 and MMP‐13 [Barbolina and

Stack, 2008]. Together this would support an important role for the proper trafficking of

MMP14 in the migration of macrophages through collagen matrices.

It has been suggested that macrophages could use different sets of proteases

subject to the composition of the 3D matrix they encounter [Wiesner et al., 2014]. Thus,

differences in the architecture and connectivity of the matrix used in individual studies

could have a strong influence on the degree of protease‐dependency of mesenchymal

macrophage migration. collagen I can be of gel‐like or fibrillar architecture, which differs

in fibre density and fibre thickness, dependent on factors such as monomer

concentration, temperature or pH during the polymerisation process [Wiesner et al.,

2014]. Therefore, the polymerisation process strongly influences the effective pore size

of the resulting matrix. But it is not only the polymerisation conditions that have an effect

on the rigidity of the matrix but also the origin, tissue source and collagen isolation

procedure used [Wiesner et al., 2014]. The degree of intermolecular connectivity of

pepsin‐extracted collagens compared to collagens isolated without the use of enzymes

differs considerably. Although pepsin‐treated collagen appears to have a similar

architecture to enzyme‐free collagen (i.e. pore size and density of fibrils), it possesses

less intermolecular cross‐links and is therefore less rigid. This might be an explanation

for why mesenchymal migration of macrophages through pepsin‐treated collagen is

MMP‐independent [Van Goethem et al., 2010] while MMP activity was required for

migration in enzyme‐free collagen, as observed herein and in previous studies [Wiesner

et al., 2014]. The extracellular matrix that comprises the connective tissue in vivo is

expected to be even more heterogeneous and will encompass both, dense and loose areas

[Wiesner et al., 2014]. Nonetheless, the data presented here suggests that targeting of

SNARE machinery identified to be involved in the cell surface delivery of MMP14 could

potentially reduce macrophage tissue infiltration.

Chapter 6: General discussion 175

Chapter 6: General discussion

The overall aim of this thesis was to identify the intracellular trafficking

pathways and trafficking machinery proteins responsible for the cell surface delivery of

important MMPs and investigate the effect of disrupting these pathways on macrophage

migration. Part of the biosynthetic route of MMP9 trafficking in macrophages was

identified. It would appear that there is crosstalk with endocytic pathways in terms of

regulating extracellular levels of MMP9 and SNAREs that might regulate these pathways

were identified. Intracellular trafficking pathways for the cell surface delivery of MMP14

were also investigated and trafficking of newly made MMP14 via the Golgi complex could

be confirmed in macrophages. The R‐SNARE VAMP4 was identified to mediate the post‐

Golgi trafficking of MMP14. This study shows for the first time that newly made MMP14

is trafficked via late endosomes/lysosomes en route to the cell surface and that the R‐

SNAREs VAMP7 and VAMP8 mediate this pathway in macrophages. Incorporation of

MMP14 into the plasma membrane was found to be regulated by the Q‐SNAREs Stx4 and

SNAP23. The SNARE‐mediated delivery of MMP14 to cell surface areas of matrix

degradation could successfully be disrupted. The disruption also attenuated macrophage

migration in vitro. These findings might lead to development of new therapeutics to

reduce macrophage tissue infiltration and inflammation.

The role of MMP9 and MMP14 secreted from macrophages during wound healing

In response to wounding, immune cells are recruited to the site of injury

following haemostasis. This inflammatory phase of the wound healing process is

important for wound cleansing, pathogen defence and promotion of subsequent healing

processes. Macrophages play an influential role on the level of inflammation within the

tissue, and as such their recruitment from the blood stream towards the wound has to

be highly efficient, but also tightly regulated. For the efficient degradation of the blood

vessel basement membrane and the interstitial tissue macrophages are equipped with

potent matrix‐degrading enzymes, such as MMP9 and MMP14. These are specifically

delivered to cell membrane structures implicated in matrix degradation, invasion and

migration in response to external stimuli, such as matrix components or pathogens. To

avoid excessive tissue damage and inflammation, expression and secretion of these

enzymes is temporally and spatially regulated. Accordingly, MMP9 and MMP14 levels

were very low in unactivated macrophages and their cell surface delivery absent. But

expression and secretion of newly made protein was found highly upregulated from 3 h

176 Chapter 6: General discussion

to 18 h post‐stimulation with the bacterial cell wall component LPS. MMP14 cell surface

delivery was also found upregulated upon cell‐matrix contact. Activation of macrophages

with bacterial and matrix components simulates macrophages encountering pathogen‐

and damage‐associated molecular patterns within the tissue following wounding.

Similar time‐ and stimulus‐dependant regulation has also previously been shown for

other MMPs, such as MMP10, MMP12 and MMP13 at the mRNA level in macrophages

[Hald et al., 2012; Murray et al., 2013].

The observed time‐dependent upregulation of proteases, such as MMP9 and

MMP14, is in alignment with the concept that macrophage tissue infiltration as well as

proteolytic activity is transient and regulated in order to avoid excessive tissue damage.

In acute wounds, macrophages can be found to invade the wound site from about one

day following injury and numbers peak around day 2 post‐wounding [Sindrilaru et al.,

2011]. In a normal healing wound, macrophage numbers will then start to decline over

the next few days. However, chronically inflamed tissue will continue to be infiltrated by

macrophages that secrete more MMPs contributing to the continuous cycle of

inflammation, as seen in chronic wounds. When dysregulated, elevated levels of certain

MMPs do not only lead to uncontrolled matrix degradation in these wounds, but also

causes the depletion of important growth factors and activation of pro‐inflammatory

cytokines. For example, MMP9 can cleave the precursors of cytokines TNFα and IL‐1β

and therefore increase pro‐inflammatory signalling [Gearing et al., 1994; Schonbeck et

al., 1998]. Through cleavage of TGFβ precursor, MMP9 also increases the bioavailability

of TGFβ, which is believed to play a role in hypertrophic scarring [Yu and Stamenkovic,

2000]. MMP14 can also activate TNFα and therefore has pro‐inflammatory potential

[D’Ortho et al., 1997]. Furthermore, MMP14 has an indirect effect on cell migration

through shedding of the cell surface glycoprotein CD44 [Kajita et al., 2001].

Unrestrained MMP expression and activity does not only play a role in

inflammation in chronic wounds, it also contributes to tumour growth, invasion and

metastasis through proteolytic degradation of the ECM, alteration of cell–cell and cell–

ECM interactions, migration and angiogenesis [Gialeli et al., 2010]. Particularly MMP14

has been considered a key player in these processes as levels correlate with the

invasiveness and metastasis of the cancer. Hence, a lot of studies have investigated the

trafficking of MMP14 in cancer cell lines and its role in cell migration. In these cell lines,

MMP14 appears to be constitutively expressed and surface pools of the enzyme are

endocytosed and recycled in the absence of cell‐matrix contact. This is in strong contrast

to the observations made for MMP14 expression in macrophages.


Whether the trafficking routes for MMP14 in macrophages are different to those

identified in cancer cell lines is largely unknown. There is currently only one other study

that has begun to investigate the trafficking pathways of MMP14 in primary

macrophages and as such these pathways in macrophages remained still poorly

understood and required further investigation in order to manipulate macrophage

infiltration and tissue inflammation.

Intracellular trafficking and secretion of MMP9 and MMP14 in macrophages

To be functional MMPs need to be delivered to the cell surface where they can

access their substrates. The newly made enzymes are trafficked via the Golgi

compartment en route to the cell surface. Apart from the Golgi complex, newly made

MMP9 and MMP14 were also found in structures in the periphery of the cell, which were

identified as late endosomal/lysosomal compartments. These organelles are well known

for their involvement in protein degradation but can also have secretory functions.

Whether trafficking via these organelles to the cell surface is a requisite for MMP9 and

MMP14 secretion could be further tested by blocking this pathway through a HRP

inactivation assay [Laulagnier et al., 2011]. Targeting the SNARE proteins Stx2, SNAP23,

VAMP2, VAMP3, VAMP4, VAMP7 and VAMP8 led to an increased release of MMP9 levels

in the extracellular environment rather than reducing its secretion. This suggests that, in

addition to a biosynthetic pathway, endocytic mechanisms are involved in regulating

MMP9 levels.

Co‐localisation with organelle markers, such as the transferrin receptor (TfR), a

recycling endosome marker, or early endosome antigen 1 (EEA1), an early endosome

marker, could give further evidence for MMP9 endocytic trafficking. Trafficking of newly

made MMP9 could also be imaged live by transfecting the cells to express MMP9 fused

to a photoactivatable/photoconvertible tag in conjunction with live‐staining organelle

markers. This would allow focussing on MMP9‐positive vesicles leaving the Golgi

complex and would provide evidence to whether these vesicles traffic through late

endosomal/lysosomal compartments before being delivered to the cell surface.

Comparing co‐localisation of MMP9 and late endosome/lysosome markers following

treatment with cycloheximide to disrupt protein synthesis might also give further insight

into whether newly made MMP9 is trafficked via the late endosome/lysosome as one

would expect less MMP9 fluorescence in these compartments if this was the case. More

understanding of the endocytic pathways of MMP9 could also be achieved by performing

a cell surface biotinylation assay [Schmidt et al., 1997]. This assay would involve


precipitation of surface molecules that are endocytosed following biotin‐labelling and

samples could be probed for the presence of MMP9. Detection of MMP9 would confirm

endocytic pathways in macrophages. Binding of MMP9 to LRP‐1 is a known clearance

mechanism for previously secreted MMP9 via endocytic pathways [Hahn‐Dantona et al.,

2001]. Thus, MMP9‐LRP‐1‐coimmunoprecipitation experiments as well as

immunofluorescence co‐localisation studies of these two proteins should be considered

to further investigate LRP‐1‐mediated endocytosis of MMP9 in macrophages. A LRP‐1‐

mediated clearing mechanism of MMP9 might explain why knockdown of the SNAREs

VAMP3, VAMP4, VAMP7, VAMP8, Stx2 and SNAP23 increased the level of extracellular

MMP9 as these SNAREs might regulate LRP‐1 trafficking.

Nonetheless, other SNARE machinery could also be tested to possibly identify

SNARE proteins more specific for the biosynthetic pathway for newly made MMP9

without affecting LRP‐1 trafficking. If newly made MMP9 was found in the recycling

endosome of macrophages, the Q‐SNAREs Stx7 (Qa), Vti1b (Qb) and Stx6 (Qc) could be

tested as a STX7/Vti1b/Stx6 complex has been implicated in trafficking of TNFα and IL‐

6 from the Golgi complex to the recycling endosome prior to their secretion from

macrophages [Manderson et al., 2007; Murray et al., 2005]. If MMP9 is trafficked from

the Golgi complex to the early endosome the Q‐SNAREs Stx16 (Qa), Vti1a (Qb) and Stx6

(Qc) could be tested for their role in MMP9 trafficking as they are known to mediate this

pathway [Scheller, 2013]. Trafficking of MMP9 from late endosomes to lysosomes could

potentially be mediated by Stx7 (Qa), Vti1b (Qb) and Stx8 (Qc) and could therefore also

be tested [Scheller, 2013]. Involvement of the SNAREs Stx13 (Qa) and SNAP25 or

SNAP29 (Qbc) in MMP9 secretion could also be examined as these SNAREs mediate

fusion events at the plasma membrane [Hong and Lev, 2014].

If MMP9 secretion cannot be targeted directly other options to specifically block

MMP9 activity could be pursued, for example on the level of MMP9 activation. An

MMP14/MMP2/TIMP complex at the cell surface has been suggested to activate MMP9

from its zymogen form [Itoh et al., 2001; Itoh and Seiki, 2004; Toth, 2003]. Thus,

targeting this complex might allow influencing MMP9 activity on that level. If any of the

components of this complex became unavailable, then activation of MMP9 is expected to

be disturbed and ECM damage might be reduced. Trafficking pathways necessary for the

cell surface delivery of MMP14 have been identified here and were successfully

disrupted by targeting responsible SNARE machinery leading to significantly reduced

MMP14 surface levels. Whether this would affect zymogen activation of MMP9 remains

to be elucidated. Interestingly, MMP14 has also been identified to cleave LRP‐1 [Rozanov


et al., 2003]. Therefore, targeting MMP14 could not only have an inhibiting effect on

MMP9 activation but might also improve MMP9 clearance as more LRP‐1 becomes

available.

Trafficking of MMP14 from the trans‐Golgi network was found to be regulated by

the R‐SNARE VAMP4. To allow membrane fusion, an R‐SNARE must form a complex with

two (Qa and Qbc) or three Q‐SNAREs (Qa, Qb and Qc). Since VAMP4 could mediate either

trafficking of MMP14 to the cell surface directly or via other organelles, partnering Q‐

SNAREs would have to be identified but are unknown in macrophages. One option is that

VAMP4 partners with Stx4/SNAP23 to allow direct trafficking of MMP14 from the Golgi

network to the plasma membrane There are, however, some indications that VAMP4

could mediate trafficking of MMP14 from the Golgi complex to the late

endosome/lysosome, as VAMP4 localises to lysosome‐related organelles in activated NK

cells and cytotoxic lymphocytes [Krzewski et al., 2011; Pattu et al., 2012]. Stx7 (Qa‐

SNARE), Vti1b (Qb‐SNARE) and Stx8 (Qc‐SNARE) could possibly form a SNARE complex

with VAMP4 (R‐SNARE), as they all localise to the lysosomal related lytic granules in

cytotoxic T lymphocytes [Pattu et al., 2012]. To confirm this hypothesis, the existence of

a VAMP4/Stx7/Vti1b/Stx8 complex in macrophages could be tested through

immunoprecipitation experiments while the involvement of the Q‐SNARE Stx7, Vti1b

and Stx8 in MMP14 transport could be tested through siRNA targeted knockdown of

these SNAREs. Trafficking of MMP14 from late endosomal/lysosomal structures was

found to be mediated by R‐SNAREs VAMP7 and VAMP8 and the Q‐SNAREs Stx4 and

SNAP23 were found to mediate the final step of incorporation of MMP14 into the plasma

membrane. Immunoprecipitation experiments could establish if VAMP7 (R‐SNARE) is

forming a complex with Stx4 (Qa‐SNARE) and SNAP23 (Qbc‐SNARE) or whether VAMP4

could also interact with Stx4/SNAP23 and allow direct trafficking of MMP14 from the

Golgi network to the plasma membrane. A VAMP7/Stx4/SNAP23 SNARE complex has

been proposed to mediate trafficking of MMP14 from the lysosome to the plasma

membrane in breast adenocarcinoma cells [Williams et al., 2014].

Long‐range transport of vesicles from the trans‐Golgi network towards the cell

periphery can be assisted by elements of the cytoskeleton, such as microtubules and

associated motor proteins, and has previously been reported for MMP9 and MMP14.

Surprisingly, disruption of the cytoskeleton did not affect levels of secreted MMP9 or

MMP14 in this study. This could be owed to the fact that any changes could have been

too small to detect as treatment with cytoskeleton‐disrupting agents did not occur for

the whole period that cells were stimulated with LPS due to the cytotoxic effects of these


drugs. It has also been reported that cytoskeleton disruption has more influence on the

rate of long‐range transport rather than fully abolishing cell surface delivery of cargo.

Live imaging of fluorescently labelled MMP9 or MMP14 in the presence of cytoskeleton‐

disrupting agents could allow determining if there is a difference in the transport speed

of vesicles in comparison to control cells. As the cytoskeleton is also important for the

polarised delivery of the protein cargo it would be interesting to investigate whether the

specific localisation of MMP14 at the cell surface of polarised macrophages remains

equally unaffected or whether delivery of MMP14 to specialised structures important for

cell migration, such as podosomes, and degradative function is disturbed upon

cytoskeleton disruption. Previous studies have shown that many cytoskeleton‐

associated molecules, such as JIP3/4, WASH and N‐WASP, are involved in the proper

trafficking of MMP14 to podosomes and indispensable for MMP14 localisation to

podosomes and matrix degradation. As such, these proteins have been suggested as

potential targets in aggressive cancers. However, agents that could disrupt the action of

these machinery proteins would not be specific enough for an application on non‐healing

wounds and negatively affect the cellular behaviour of various other cell types in the

wound environment.

The polarised cell surface delivery of MMP14 to membrane structures important

for migration and invasion is therefore a highly efficient but also regulated process and

allows to limit matrix degradation to necessary areas of the cell such as the leading edge.

As a transmembrane protein MMP14 activity is also restricted to the pericellular space.

MMP9 on the other hand is secreted as a soluble enzyme. To concentrate the activity of

soluble MMPs, such as MMP9, to the pericellular space they can bind back to the cell

surface at plasma membrane structures implicated in cell migration. MMP9 is known to

bind to CD44, as well as integrins β1 and β5 [Deryugina and Quigley, 2011]. Interestingly,

these molecules are also found at podosomes allowing a highly efficient migration

process. Co‐localisation studies would confirm if these interactions are also present in

macrophages. Whether disruption of the interaction between MMP9 and its surface

receptors has an impact on the efficiency of macrophage invasion could also be

investigated.

Disrupting MMP14 trafficking to attenuate matrix degradation and cell migration

Upon successful delivery of MMP14 to podosomes, MMP14 is responsible for

matrix degradation and macrophage migration. Thus, when targeting SNARE machinery

proteins that were identified to mediate MMP14 cell surface delivery, namely VAMP4.


VAMP7, VAMP8, Stx4 and SNAP23, macrophage degradative capacity and migration in

vitro were significantly impaired. This effect appeared to be specific for those SNAREs

that are involved in MMP14 trafficking, as for example treatment with VAMP3 siRNA,

which has previously been shown to affect integrin recycling and persistent migration

on fibronectin in macrophages [Veale et al., 2011], did not significantly reduce

degradation of gelatin. Similarly, treatment with VAMP2, Stx2 or Stx3 siRNA did also not

change matrix degradation.

The effect of disrupting trafficking machinery responsible for MMP14‐mediated

matrix degradation and protease‐dependent cell migration could lead to development of

therapeutics that attenuate macrophage tissue infiltration and inflammation. However,

it would be important to confirm these effects on protease‐dependent migration in vivo

as the definition of amoeboid and mesenchymal migration modes by macrophages have

been identified in vitro [Verrolet et al., 2011]. The cell morphology of can be an indication

whether macrophages are using the amoeboid or mesenchymal migration mode and as

such be a predictor of the involvement of proteases. Studies using techniques that allow

intravital imaging have shown that migrating macrophages exhibit mesenchymal

characteristics with a polarised cell shape and cytoplasmic protrusions, rather than the

rounded cell morphology that is characteristic for the amoeboid mode of migration and

observed with neutrophils and T lymphocytes [Egen et al., 2008; Leimgruber et al.,

2009]. Another indicator for mesenchymal migration is that upon deletion or mutation

of mesenchymal migration effectors macrophages show altered migration in vivo.

Macrophages from patients expressing truncated forms of actin polymerisation

regulator WASP, which consequently lack podosomes, show defective cell migration and

impaired immunostimulatory activation [Linder et al., 1999]. Thus, it seems to be very

likely that macrophages do use the mesenchymal migration mode in vivo. But more work

will be required to confirm this as well as endorse the role of MMP14 in macrophage

migration in vivo. In cancer, MMP14 expression is correlated with invasiveness

illustrating the significance of MMP14 for tissue infiltration in these cells.

Interestingly, conditional knockout of MMP14 in the monocyte/macrophage

lineage leads to a reduction in the number of macrophages in inflamed tissue when

compared with wildtype mice in a contact dermatitis model. However, no differences

were found for wound closure, granulation tissue formation, and angiogenesis between

wildtype and knockout groups [Klose et al., 2013]. Potentially, it could be more relevant

to investigate the role of disrupting MMP14‐mediated macrophage migration in a mouse

model with impaired wound healing. Widely used models of impaired healing involve


mice with type 2 diabetes as a result of leptin (ob/ob) ‐ or leptin receptor (db/db)

knockout [Baetens et al., 1978; Boquist et al., 1974].

Targeting SNAREs with Clostridial neurotoxins

The family of clostridial neurotoxins (CNTs) is comprised of one tetanus

neurotoxin (TeNT) and seven (A to G) distinct botulinum neurotoxins (BoN). They

specifically bind to molecules on the cell surface of neuronal cells, which allows

endocytosis of the toxin followed by membrane translocation of the catalytic domain and

finally proteolytic cleavage. To fulfil this multistep process, the two‐chain toxin is

encompasses three domains. Binding of the C‐terminal domain of the heavy chain (HC)

to gangliosides and synaptic vesicle protein SV2, which are found particularly enriched

at the presynaptic membrane, allows receptor‐mediated endocytosis of the toxin. As the

pH decreases the toxin undergoes structural rearrangement so that the N‐terminal

domain of the heavy chain (HN) initiates membrane translocation of the light chain which

is the domain of the toxin that contains the protease activity. Depending on the substrate

specificity of the individual toxin they can proteolytically cleave the SNAREs VAMP2,

VAMP3, Stx1, or SNAP25 leading to a defect in the release of the neurotransmitter

acetylcholine causing the diseases tetanus and botulism.

Synthesised, unmodified BoNT‐A and B are used as therapeutics in the clinic to

treat diseases such as upper motor neuron syndrome, focal hyperhidrosis,

blepharospasm, strabismus, chronic migraine and bruxism but have also been utilised

for cosmetic treatments. Although these toxins are known to specifically bind to motor

neurons and Stx1 is only expressed in neuronal cells, it could be shown that unmodified

BoNT‐A can inhibit insulin secretion from beta cells [Boyd et al., 1995]. This served as a

proof of concept that CNTs could be modified to target non‐neuronal secretory

processes. Through engineered replacement of the motor neuron‐specific cell binding

domain (HC) it was possible to inhibit secretion in epithelial cell line [Foster et al., 2006].

Furthermore, it has been suggested that the L chain of CNTs can also be catalytically

modified to specifically target other SNARE family protein members [Pickett and Perrow,

2011]. For example, BoNT‐E was successfully modified (BoNT‐E K224D) to cleave

SNAP23 in addition to SNAP25 and when added to TNFα‐stimulated HeLA cells reduced

IL‐8 secretion, which is usually mediated by SNAP23 [Chen et al., 2009]. It has

furthermore been suggested that engineering of these BoNT to target the R‐SNARE

VAMP7 will also be possible [Binz et al., 2010].


This opens up the opportunity to use modified CNTs to explicitly disrupt MMP14

trafficking in macrophages.

Being able to selectively inhibit macrophage‐derived MMP14 activity is

important as not only macrophages but also fibroblasts, endothelial cells as well as

keratinocytes express MMP14. Therefore, non‐specific inhibition of MMP14 could be

expected to impede not only macrophage but also fibroblast, endothelial and

keratinocyte migration and as such negatively affect important healing processes like re‐

epithelialisation and angiogenesis. The binding domain (HC) of the toxin could be

modified to interact with macrophage‐specific surface molecules, for example CD68 (low

density lipoprotein/Macrosialin, pan marker for macrophages). However, as

macrophages of the anti‐inflammatory M2 type are considered to promote healing,

markers that are known to be highly expressed on pro‐inflammatory M1 macrophages

within the wound tissue, such as Ly6C (mouse marker) or CD192/CCR2 (human and

mouse marker), could be targeted. In combination with re‐engineering of the catalytic

domain (L chain) to specifically target the SNAREs VAMP7 or SNAP23, this might allow

disruption of inflammatory macrophage‐specific MMP14 trafficking and attenuation of

macrophage invasion. Therefore, the findings from this work together with future drug

development might lead to the advancement of chronic wound treatments.

184 References

References Agrawal, S., Anderson, P., Durbeej, M., van Rooijen, N., Ivars, F., Opdenakker, G., & Sorokin,

L. M. (2006). Dystroglycan is selectively cleaved at the parenchymal basement membrane at sites of leukocyte extravasation in experimental autoimmune encephalomyelitis. J Exp Med, 203(4), 1007‐1019. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16585265. doi:10.1084/jem.20051342

Andrei, C., Dazzi, C., Lotti, L., Torrisi, M. R., Chimini, G., & Rubartelli, A. (1999). The

secretory route of the leaderless protein interleukin 1beta involves exocytosis of endolysosome‐related vesicles. Mol Biol Cell, 10(5), 1463‐1475. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10233156.

Anitei, M., & Hoflack, B. (2012). Bridging membrane and cytoskeleton dynamics in the

secretory and endocytic pathways. Nat Cell Biol, 14(1), 11‐19. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22193159. doi:10.1038/ncb2409

Appenzeller‐Herzog, C., & Hauri, H. P. (2006). The ER‐Golgi intermediate compartment

(ERGIC): in search of its identity and function. J Cell Sci, 119(Pt 11), 2173‐2183. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16723730. doi:10.1242/jcs.03019

Baetens, D., Stefan, Y., Ravazzola, M., Malaisse‐Lagae, F., Coleman, D. L., & Orci, L. (1978).

Alteration of islet cell populations in spontaneously diabetic mice. Diabetes, 27(1), 1‐7. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/340309.

Bannon, P., Wood, S., Restivo, T., Campbell, L., Hardman, M. J., & Mace, K. A. (2013).

Diabetes induces stable intrinsic changes to myeloid cells that contribute to chronic inflammation during wound healing in mice. Dis Model Mech, 6(6), 1434‐1447. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24057002. doi:10.1242/dmm.012237

Barbolina, M. V., & Stack, M. S. (2008). Membrane type 1‐matrix metalloproteinase:

substrate diversity in pericellular proteolysis. Semin Cell Dev Biol, 19(1), 24‐33. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17702616. doi:10.1016/j.semcdb.2007.06.008

Barletta, J. P., Angella, G., Balch, K. C., Dimova, H. G., Stern, G. A., Moser, M. T., . . . Schultz,

G. S. (1996). Inhibition of pseudomonal ulceration in rabbit corneas by a synthetic matrix metalloproteinase inhibitor. Invest Ophthalmol Vis Sci, 37(1), 20‐28. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/8550324.

Beidler, S. K., Douillet, C. D., Berndt, D. F., Keagy, B. A., Rich, P. B., & Marston, W. A. (2008).

Multiplexed analysis of matrix metalloproteinases in leg ulcer tissue of patients with chronic venous insufficiency before and after compression therapy. Wound Repair Regen, 16(5), 642‐648. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19128259. doi:10.1111/j.1524‐475X.2008.00415.x

Beraud‐Dufour, S., & Balch, W. (2002). A journey through the exocytic pathway. J Cell Sci,

115(Pt 9), 1779‐1780. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11956309.

http://www.ncbi.nlm.nih.gov/pubmed/16585265










References 185

Bertho, N., Cerny, J., Kim, Y. M., Fiebiger, E., Ploegh, H., & Boes, M. (2003). Requirements

for T cell‐polarized tubulation of class II+ compartments in dendritic cells. J Immunol, 171(11), 5689‐5696. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/14634076.

Bhuin, T., & Roy, J. K. (2014). Rab proteins: the key regulators of intracellular vesicle

transport. Exp Cell Res, 328(1), 1‐19. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25088255. doi:10.1016/j.yexcr.2014.07.027

Bianchi, M. E. (2007). DAMPs, PAMPs and alarmins: all we need to know about danger. J

Leukoc Biol, 81(1), 1‐5. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17032697. doi:10.1189/jlb.0306164

Bjorklund, M., & Koivunen, E. (2005). Gelatinase‐mediated migration and invasion of

cancer cells. Biochim Biophys Acta, 1755(1), 37‐69. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15907591. doi:10.1016/j.bbcan.2005.03.001

Blott, E. J., & Griffiths, G. M. (2002). Secretory lysosomes. Nat Rev Mol Cell Biol, 3(2), 122‐

131. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11836514. doi:10.1038/nrm732

Bobrie, A., Krumeich, S., Reyal, F., Recchi, C., Moita, L. F., Seabra, M. C., . . . Thery, C. (2012).

Rab27a supports exosome‐dependent and ‐independent mechanisms that modify the tumor microenvironment and can promote tumor progression. Cancer Res, 72(19), 4920‐4930. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22865453. doi:10.1158/0008‐5472.CAN‐12‐0925

Boes, M., Cerny, J., Massol, R., Op den Brouw, M., Kirchhausen, T., Chen, J., & Ploegh, H. L.

(2002). T‐cell engagement of dendritic cells rapidly rearranges MHC class II transport. Nature, 418(6901), 983‐988. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12198548. doi:10.1038/nature01004

Boquist, L., Hellman, B., Lernmark, A., & Taljedal, I. B. (1974). Influence of the mutation

"diabetes" on insulin release and islet morphology in mice of different genetic backgrounds. J Cell Biol, 62(1), 77‐89. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/4135113.

Brancato, S. K., & Albina, J. E. (2011). Wound macrophages as key regulators of repair:

origin, phenotype, and function. Am J Pathol, 178(1), 19‐25. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21224038. doi:10.1016/j.ajpath.2010.08.003

Braun, V., Fraisier, V., Raposo, G., Hurbain, I., Sibarita, J. B., Chavrier, P., . . . Niedergang, F.

(2004). TI‐VAMP/VAMP7 is required for optimal phagocytosis of opsonised particles in macrophages. EMBO J, 23(21), 4166‐4176. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15470500. doi:10.1038/sj.emboj.7600427

Bravo‐Cordero, J. J., Marrero‐Diaz, R., Megias, D., Genis, L., Garcia‐Grande, A., Garcia, M.

A., . . . Montoya, M. C. (2007). MT1‐MMP proinvasive activity is regulated by a











186 References

novel Rab8‐dependent exocytic pathway. EMBO J, 26(6), 1499‐1510. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17332756. doi:10.1038/sj.emboj.7601606

Brown, B. N., Sicari, B. M., & Badylak, S. F. (2014). Rethinking regenerative medicine: a

macrophage‐centered approach. Front Immunol, 5, 510. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25408693. doi:10.3389/fimmu.2014.00510

Brzezinska, A. A., Johnson, J. L., Munafo, D. B., Crozat, K., Beutler, B., Kiosses, W. B., . . . Catz,

S. D. (2008). The Rab27a effectors JFC1/Slp1 and Munc13‐4 regulate exocytosis of neutrophil granules. Traffic, 9(12), 2151‐2164. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18939952. doi:10.1111/j.1600‐0854.2008.00838.x

Buccione, R., Orth, J. D., & McNiven, M. A. (2004). Foot and mouth: podosomes,

invadopodia and circular dorsal ruffles. Nat Rev Mol Cell Biol, 5(8), 647‐657. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15366708. doi:10.1038/nrm1436

Bucky, L. P., Vedder, N. B., Hong, H. Z., Ehrlich, H. P., Winn, R. K., Harlan, J. M., & May, J. W.,

Jr. (1994). Reduction of burn injury by inhibiting CD18‐mediated leukocyte adherence in rabbits. Plast Reconstr Surg, 93(7), 1473‐1480. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7911582.

Cai, H., Yu, S., Menon, S., Cai, Y., Lazarova, D., Fu, C., . . . Ferro‐Novick, S. (2007). TRAPPI

tethers COPII vesicles by binding the coat subunit Sec23. Nature, 445(7130), 941‐944. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17287728. doi:10.1038/nature05527

Carroll, K. S., Hanna, J., Simon, I., Krise, J., Barbero, P., & Pfeffer, S. R. (2001). Role of Rab9

GTPase in facilitating receptor recruitment by TIP47. Science, 292(5520), 1373‐1376. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11359012. doi:10.1126/science.1056791

Caswell, P. T., Chan, M., Lindsay, A. J., McCaffrey, M. W., Boettiger, D., & Norman, J. C.

(2008). Rab‐coupling protein coordinates recycling of alpha5beta1 integrin and EGFR1 to promote cell migration in 3D microenvironments. J Cell Biol, 183(1), 143‐155. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18838556. doi:10.1083/jcb.200804140

Caviston, J. P., & Holzbaur, E. L. (2006). Microtubule motors at the intersection of

trafficking and transport. Trends Cell Biol, 16(10), 530‐537. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16938456. doi:10.1016/j.tcb.2006.08.002

Chakrabarti, S., Zee, J. M., & Patel, K. D. (2006). Regulation of matrix metalloproteinase‐9

(MMP‐9) in TNF‐stimulated neutrophils: novel pathways for tertiary granule release. J Leukoc Biol, 79(1), 214‐222. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16275891. doi:10.1189/jlb.0605353

Chen, D., Lemons, P. P., Schraw, T., & Whiteheart, S. W. (2000). Molecular mechanisms of

platelet exocytosis: role of SNAP‐23 and syntaxin 2 and 4 in lysosome release.











References 187

Blood, 96(5), 1782‐1788. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10961877.

Chen, J. W., Murphy, T. L., Willingham, M. C., Pastan, I., & August, J. T. (1985). Identification

of two lysosomal membrane glycoproteins. J Cell Biol, 101(1), 85‐95. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/2409098.

Chin, T., & Hyakusoku, H. (2003). A new device: A portable and effective irrigation system

for pressure ulcers. Plast Reconstr Surg, 112(7), 1973‐1974. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/14663270.

Chow, A., Toomre, D., Garrett, W., & Mellman, I. (2002). Dendritic cell maturation triggers

retrograde MHC class II transport from lysosomes to the plasma membrane. Nature, 418(6901), 988‐994. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12198549. doi:10.1038/nature01006

Chow, A. Y., & Mellman, I. (2005). Old lysosomes, new tricks: MHC II dynamics in DCs.

Trends Immunol, 26(2), 72‐78. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15668121. doi:10.1016/j.it.2004.11.008

Clancy, J. W., Sedgwick, A., Rosse, C., Muralidharan‐Chari, V., Raposo, G., Method, M., . . .

D'Souza‐Schorey, C. (2015). Regulated delivery of molecular cargo to invasive tumour‐derived microvesicles. Nat Commun, 6, 6919. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25897521. doi:10.1038/ncomms7919

Clark, R., & Griffiths, G. M. (2003). Lytic granules, secretory lysosomes and disease. Curr

Opin Immunol, 15(5), 516‐521. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/14499259.

Cole, N. B., Sciaky, N., Marotta, A., Song, J., & Lippincott‐Schwartz, J. (1996). Golgi dispersal

during microtubule disruption: regeneration of Golgi stacks at peripheral endoplasmic reticulum exit sites. Mol Biol Cell, 7(4), 631‐650. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/8730104.

Cornfine, S., Himmel, M., Kopp, P., El Azzouzi, K., Wiesner, C., Kruger, M., . . . Linder, S.

(2011). The kinesin KIF9 and reggie/flotillin proteins regulate matrix degradation by macrophage podosomes. Mol Biol Cell, 22(2), 202‐215. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21119006. doi:10.1091/mbc.E10‐05‐0394

Cougoule, C., Le Cabec, V., Poincloux, R., Al Saati, T., Mege, J. L., Tabouret, G., . . .

Maridonneau‐Parini, I. (2010). Three‐dimensional migration of macrophages requires Hck for podosome organization and extracellular matrix proteolysis. Blood, 115(7), 1444‐1452. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19897576. doi:10.1182/blood‐2009‐04‐218735

Cowin, A. J., Brosnan, M. P., Holmes, T. M., & Ferguson, M. W. (1998). Endogenous

inflammatory response to dermal wound healing in the fetal and adult mouse. Dev Dyn, 212(3), 385‐393. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9671942. doi:10.1002/(SICI)1097‐0177(199807)212:3<385::AID‐AJA6>3.0.CO;2‐D












188 References

Cullen, B., Smith, R., McCulloch, E., Silcock, D., & Morrison, L. (2002). Mechanism of action of PROMOGRAN, a protease modulating matrix, for the treatment of diabetic foot ulcers. Wound Repair Regen, 10(1), 16‐25. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11983003.

d'Ortho, M. P., Will, H., Atkinson, S., Butler, G., Messent, A., Gavrilovic, J., . . . Murphy, G.

(1997). Membrane‐type matrix metalloproteinases 1 and 2 exhibit broad‐spectrum proteolytic capacities comparable to many matrix metalloproteinases. Eur J Biochem, 250(3), 751‐757. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9461298.

Dahl, R., Titlestad, I., Lindqvist, A., Wielders, P., Wray, H., Wang, M., . . . Holt, A. (2012).

Effects of an oral MMP‐9 and ‐12 inhibitor, AZD1236, on biomarkers in moderate/severe COPD: a randomised controlled trial. Pulm Pharmacol Ther, 25(2), 169‐177. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22306193. doi:10.1016/j.pupt.2011.12.011

Daley, J. M., Brancato, S. K., Thomay, A. A., Reichner, J. S., & Albina, J. E. (2010). The

phenotype of murine wound macrophages. J Leukoc Biol, 87(1), 59‐67. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20052800.

Daley, J. M., Reichner, J. S., Mahoney, E. J., Manfield, L., Henry, W. L., Jr., Mastrofrancesco,

B., & Albina, J. E. (2005). Modulation of macrophage phenotype by soluble product(s) released from neutrophils. J Immunol, 174(4), 2265‐2272. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15699161.

Darby, I. A., Laverdet, B., Bonte, F., & Desmouliere, A. (2014). Fibroblasts and

myofibroblasts in wound healing. Clin Cosmet Investig Dermatol, 7, 301‐311. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25395868. doi:10.2147/CCID.S50046

De Matteis, M. A., & Luini, A. (2008). Exiting the Golgi complex. Nat Rev Mol Cell Biol, 9(4),

273‐284. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18354421. doi:10.1038/nrm2378

del Zoppo, G. J., Frankowski, H., Gu, Y. H., Osada, T., Kanazawa, M., Milner, R., . . . Koziol, J.

A. (2012). Microglial cell activation is a source of metalloproteinase generation during hemorrhagic transformation. J Cereb Blood Flow Metab, 32(5), 919‐932. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22354151. doi:10.1038/jcbfm.2012.11

Dell'Angelica, E. C., Mullins, C., Caplan, S., & Bonifacino, J. S. (2000). Lysosome‐related

organelles. FASEB J, 14(10), 1265‐1278. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10877819.

Devalaraja, R. M., Nanney, L. B., Du, J., Qian, Q., Yu, Y., Devalaraja, M. N., & Richmond, A.

(2000). Delayed wound healing in CXCR2 knockout mice. J Invest Dermatol, 115(2), 234‐244. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10951241. doi:10.1046/j.1523‐1747.2000.00034.x

Devy, L., & Dransfield, D. T. (2011). New Strategies for the Next Generation of Matrix‐

Metalloproteinase Inhibitors: Selectively Targeting Membrane‐Anchored MMPs











References 189

with Therapeutic Antibodies. Biochem Res Int, 2011, 191670. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21152183. doi:10.1155/2011/191670

Devy, L., Huang, L., Naa, L., Yanamandra, N., Pieters, H., Frans, N., . . . Dransfield, D. T.

(2009). Selective inhibition of matrix metalloproteinase‐14 blocks tumor growth, invasion, and angiogenesis. Cancer Res, 69(4), 1517‐1526. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19208838. doi:10.1158/0008‐5472.CAN‐08‐3255

DiPietro, L. A., Burdick, M., Low, Q. E., Kunkel, S. L., & Strieter, R. M. (1998). MIP‐1alpha

as a critical macrophage chemoattractant in murine wound repair. J Clin Invest, 101(8), 1693‐1698. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9541500. doi:10.1172/JCI1020

Doherty, G. J., & McMahon, H. T. (2009). Mechanisms of endocytosis. Annu Rev Biochem,

78, 857‐902. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19317650. doi:10.1146/annurev.biochem.78.081307.110540

Dovi, J. V., He, L. K., & DiPietro, L. A. (2003). Accelerated wound closure in neutrophil‐

depleted mice. J Leukoc Biol, 73(4), 448‐455. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12660219.

Dressel, R., Elsner, L., Novota, P., Kanwar, N., & Fischer von Mollard, G. (2010). The

exocytosis of lytic granules is impaired in Vti1b‐ or Vamp8‐deficient CTL leading to a reduced cytotoxic activity following antigen‐specific activation. J Immunol, 185(2), 1005‐1014. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20543108. doi:10.4049/jimmunol.1000770

Duellman, T., Burnett, J., Shin, A., & Yang, J. (2015). LMAN1 (ERGIC‐53) is a potential

carrier protein for matrix metalloproteinase‐9 glycoprotein secretion. Biochem Biophys Res Commun, 464(3), 685‐691. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/26150355. doi:10.1016/j.bbrc.2015.06.164

Duellman, T., Burnett, J., & Yang, J. (2015). Functional Roles of N‐Linked Glycosylation of

Human Matrix Metalloproteinase 9. Traffic, 16(10), 1108‐1126. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/26207422. doi:10.1111/tra.12312

Duffield, J. S., Forbes, S. J., Constandinou, C. M., Clay, S., Partolina, M., Vuthoori, S., . . .

Iredale, J. P. (2005). Selective depletion of macrophages reveals distinct, opposing roles during liver injury and repair. J Clin Invest, 115(1), 56‐65. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15630444. doi:10.1172/JCI22675

Egeblad, M., & Werb, Z. (2002). New functions for the matrix metalloproteinases in

cancer progression. Nat Rev Cancer, 2(3), 161‐174. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11990853. doi:10.1038/nrc745

Egen, J. G., Rothfuchs, A. G., Feng, C. G., Winter, N., Sher, A., & Germain, R. N. (2008).

Macrophage and T cell dynamics during the development and disintegration of mycobacterial granulomas. Immunity, 28(2), 271‐284. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18261937. doi:10.1016/j.immuni.2007.12.010












190 References

Ellett, F., Pase, L., Hayman, J. W., Andrianopoulos, A., & Lieschke, G. J. (2011). mpeg1

promoter transgenes direct macrophage‐lineage expression in zebrafish. Blood, 117(4), e49‐56. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21084707. doi:10.1182/blood‐2010‐10‐314120

Eming, S., Smola, H., Hartmann, B., Malchau, G., Wegner, R., Krieg, T., & Smola‐Hess, S.

(2008). The inhibition of matrix metalloproteinase activity in chronic wounds by a polyacrylate superabsorber. Biomaterials, 29(19), 2932‐2940. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18400293. doi:10.1016/j.biomaterials.2008.03.029

Eming, S. A., Krieg, T., & Davidson, J. M. (2007). Inflammation in wound repair: molecular

and cellular mechanisms. J Invest Dermatol, 127(3), 514‐525. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17299434. doi:10.1038/sj.jid.5700701

Eskelinen, E. L., Tanaka, Y., & Saftig, P. (2003). At the acidic edge: emerging functions for

lysosomal membrane proteins. Trends Cell Biol, 13(3), 137‐145. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12628346.

Fader, C. M., Sanchez, D. G., Mestre, M. B., & Colombo, M. I. (2009). TI‐VAMP/VAMP7 and

VAMP3/cellubrevin: two v‐SNARE proteins involved in specific steps of the autophagy/multivesicular body pathways. Biochim Biophys Acta, 1793(12), 1901‐1916. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19781582. doi:10.1016/j.bbamcr.2009.09.011

Fasshauer, D. (2003). Structural insights into the SNARE mechanism. Biochim Biophys

Acta, 1641(2‐3), 87‐97. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12914950.

Ferguson, M. W., & O'Kane, S. (2004). Scar‐free healing: from embryonic mechanisms to

adult therapeutic intervention. Philos Trans R Soc Lond B Biol Sci, 359(1445), 839‐850. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15293811. doi:10.1098/rstb.2004.1475

Fridman, R., Toth, M., Chvyrkova, I., Meroueh, S. O., & Mobashery, S. (2003). Cell surface

association of matrix metalloproteinase‐9 (gelatinase B). Cancer Metastasis Rev, 22(2‐3), 153‐166. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12784994.

Friedl, P., & Weigelin, B. (2008). Interstitial leukocyte migration and immune function.

Nat Immunol, 9(9), 960‐969. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18711433. doi:10.1038/ni.f.212

Frittoli, E., Palamidessi, A., Disanza, A., & Scita, G. (2011). Secretory and endo/exocytic

trafficking in invadopodia formation: the MT1‐MMP paradigm. Eur J Cell Biol, 90(2‐3), 108‐114. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20605060. doi:10.1016/j.ejcb.2010.04.007

Frykberg, R. G., & Banks, J. (2015). Challenges in the Treatment of Chronic Wounds. Adv

Wound Care (New Rochelle), 4(9), 560‐582. Retrieved from











References 191

http://www.ncbi.nlm.nih.gov/pubmed/26339534. doi:10.1089/wound.2015.0635

Fuchs, P., Hartmann, T. L., Schrimpf, C., Haunschild, J., Litzenburger, T., & Pallua, N.

(2006). A recombinant anti‐ICAM‐1 Fab fragment is as effective as the complete IgG antibody in treatment of burns in rabbits. Burns, 32(4), 430‐435. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16621320. doi:10.1016/j.burns.2005.11.019

Futter, C. E., Connolly, C. N., Cutler, D. F., & Hopkins, C. R. (1995). Newly synthesized

transferrin receptors can be detected in the endosome before they appear on the cell surface. J Biol Chem, 270(18), 10999‐11003. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7738042.

Galvez, B. G., Matias‐Roman, S., Albar, J. P., Sanchez‐Madrid, F., & Arroyo, A. G. (2001).

Membrane type 1‐matrix metalloproteinase is activated during migration of human endothelial cells and modulates endothelial motility and matrix remodeling. J Biol Chem, 276(40), 37491‐37500. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11448964. doi:10.1074/jbc.M104094200

Gawden‐Bone, C., Zhou, Z., King, E., Prescott, A., Watts, C., & Lucocq, J. (2010). Dendritic

cell podosomes are protrusive and invade the extracellular matrix using metalloproteinase MMP‐14. J Cell Sci, 123(Pt 9), 1427‐1437. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20356925. doi:10.1242/jcs.056515

Gearing, A. J., Beckett, P., Christodoulou, M., Churchill, M., Clements, J., Davidson, A. H., . . .

et al. (1994). Processing of tumour necrosis factor‐alpha precursor by metalloproteinases. Nature, 370(6490), 555‐557. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/8052310. doi:10.1038/370555a0

Gethin, G. (2012). Understanding the inflammatory process in wound healing. Br J

Community Nurs, Suppl, S17‐18, S20, S22. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22584180.

Geuze, H. J. (1998). The role of endosomes and lysosomes in MHC class II functioning.

Immunol Today, 19(6), 282‐287. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9639994.

Ghaemmaghami, S., Huh, W. K., Bower, K., Howson, R. W., Belle, A., Dephoure, N., . . .

Weissman, J. S. (2003). Global analysis of protein expression in yeast. Nature, 425(6959), 737‐741. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/14562106. doi:10.1038/nature02046

Gialeli, C., Theocharis, A. D., & Karamanos, N. K. (2011). Roles of matrix

metalloproteinases in cancer progression and their pharmacological targeting. FEBS J, 278(1), 16‐27. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21087457. doi:10.1111/j.1742‐4658.2010.07919.x

Golub, L. M., Lee, H. M., Lehrer, G., Nemiroff, A., McNamara, T. F., Kaplan, R., &

Ramamurthy, N. S. (1983). Minocycline reduces gingival collagenolytic activity during diabetes. Preliminary observations and a proposed new mechanism of











192 References

action. J Periodontal Res, 18(5), 516‐526. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/6315909.

Golub, L. M., McNamara, T. F., D'Angelo, G., Greenwald, R. A., & Ramamurthy, N. S. (1987).

A non‐antibacterial chemically‐modified tetracycline inhibits mammalian collagenase activity. J Dent Res, 66(8), 1310‐1314. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/3040832.

Gonzalo, P., Guadamillas, M. C., Hernandez‐Riquer, M. V., Pollan, A., Grande‐Garcia, A.,

Bartolome, R. A., . . . Arroyo, A. G. (2010). MT1‐MMP is required for myeloid cell fusion via regulation of Rac1 signaling. Dev Cell, 18(1), 77‐89. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20152179. doi:10.1016/j.devcel.2009.11.012

Gooyit, M., Peng, Z., Wolter, W. R., Pi, H., Ding, D., Hesek, D., . . . Chang, M. (2014). A

chemical biological strategy to facilitate diabetic wound healing. ACS Chem Biol, 9(1), 105‐110. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24053680. doi:10.1021/cb4005468

Goren, I., Muller, E., Schiefelbein, D., Christen, U., Pfeilschifter, J., Muhl, H., & Frank, S.

(2007). Systemic anti‐TNFalpha treatment restores diabetes‐impaired skin repair in ob/ob mice by inactivation of macrophages. J Invest Dermatol, 127(9), 2259‐2267. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17460730. doi:10.1038/sj.jid.5700842

Granger, E., McNee, G., Allan, V., & Woodman, P. (2014). The role of the cytoskeleton and

molecular motors in endosomal dynamics. Semin Cell Dev Biol, 31, 20‐29. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24727350. doi:10.1016/j.semcdb.2014.04.011

Gray, C., Loynes, C. A., Whyte, M. K., Crossman, D. C., Renshaw, S. A., & Chico, T. J. (2011).

Simultaneous intravital imaging of macrophage and neutrophil behaviour during inflammation using a novel transgenic zebrafish. Thromb Haemost, 105(5), 811‐819. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21225092. doi:10.1160/TH10‐08‐0525

Greenberg, S., & Grinstein, S. (2002). Phagocytosis and innate immunity. Curr Opin

Immunol, 14(1), 136‐145. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11790544.

Greenhalgh, D. G. (1998). The role of apoptosis in wound healing. Int J Biochem Cell Biol,

30(9), 1019‐1030. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9785465.

Greenwald, R. A., Golub, L. M., Lavietes, B., Ramamurthy, N. S., Gruber, B., Laskin, R. S., &

McNamara, T. F. (1987). Tetracyclines inhibit human synovial collagenase in vivo and in vitro. J Rheumatol, 14(1), 28‐32. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/3033237.

Griffiths, G. M. (2003). Endocytosing the death sentence. J Cell Biol, 160(2), 155‐156.

Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12538637. doi:10.1083/jcb.200212143












References 193

Gu, Z., Noss, E. H., Hsu, V. W., & Brenner, M. B. (2011). Integrins traffic rapidly via circular dorsal ruffles and macropinocytosis during stimulated cell migration. J Cell Biol, 193(1), 61‐70. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21464228. doi:10.1083/jcb.201007003

Hahn‐Dantona, E., Ruiz, J. F., Bornstein, P., & Strickland, D. K. (2001). The low density

lipoprotein receptor‐related protein modulates levels of matrix metalloproteinase 9 (MMP‐9) by mediating its cellular catabolism. J Biol Chem, 276(18), 15498‐15503. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11279011. doi:10.1074/jbc.M100121200

Hald, A., Rono, B., Lund, L. R., & Egerod, K. L. (2012). LPS counter regulates RNA

expression of extracellular proteases and their inhibitors in murine macrophages. Mediators Inflamm, 2012, 157894. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22529519. doi:10.1155/2012/157894

Hanania, R., Sun, H. S., Xu, K., Pustylnik, S., Jeganathan, S., & Harrison, R. E. (2012).

Classically activated macrophages use stable microtubules for matrix metalloproteinase‐9 (MMP‐9) secretion. J Biol Chem, 287(11), 8468‐8483. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22270361. doi:10.1074/jbc.M111.290676

Hattula, K., Furuhjelm, J., Tikkanen, J., Tanhuanpaa, K., Laakkonen, P., & Peranen, J.

(2006). Characterization of the Rab8‐specific membrane traffic route linked to protrusion formation. J Cell Sci, 119(Pt 23), 4866‐4877. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17105768. doi:10.1242/jcs.03275

Hennigan, R. F., Hawker, K. L., & Ozanne, B. W. (1994). Fos‐transformation activates

genes associated with invasion. Oncogene, 9(12), 3591‐3600. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7970719.

Henry, L., & Sheff, D. R. (2008). Rab8 regulates basolateral secretory, but not recycling,

traffic at the recycling endosome. Mol Biol Cell, 19(5), 2059‐2068. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18287531. doi:10.1091/mbc.E07‐09‐0902

Hong, W. (2005). SNAREs and traffic. Biochim Biophys Acta, 1744(2), 120‐144. Retrieved

from http://www.ncbi.nlm.nih.gov/pubmed/15893389. doi:10.1016/j.bbamcr.2005.03.014

Hoogenboom, H. R. (2005). Selecting and screening recombinant antibody libraries. Nat

Biotechnol, 23(9), 1105‐1116. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16151404. doi:10.1038/nbt1126

Hopkinson‐Woolley, J., Hughes, D., Gordon, S., & Martin, P. (1994). Macrophage

recruitment during limb development and wound healing in the embryonic and foetal mouse. J Cell Sci, 107 ( Pt 5), 1159‐1167. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7929625.

Hotary, K., Li, X. Y., Allen, E., Stevens, S. L., & Weiss, S. J. (2006). A cancer cell

metalloprotease triad regulates the basement membrane transmigration program. Genes Dev, 20(19), 2673‐2686. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16983145. doi:10.1101/gad.1451806












194 References

Huang, W. C., Sala‐Newby, G. B., Susana, A., Johnson, J. L., & Newby, A. C. (2012). Classical

macrophage activation up‐regulates several matrix metalloproteinases through mitogen activated protein kinases and nuclear factor‐kappaB. PLoS One, 7(8), e42507. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22880008. doi:10.1371/journal.pone.0042507

Hyun, Y. M., Sumagin, R., Sarangi, P. P., Lomakina, E., Overstreet, M. G., Baker, C. M., . . .

Kim, M. (2012). Uropod elongation is a common final step in leukocyte extravasation through inflamed vessels. J Exp Med, 209(7), 1349‐1362. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22711877. doi:10.1084/jem.20111426

Imai, K., & Takaoka, A. (2006). Comparing antibody and small‐molecule therapies for

cancer. Nat Rev Cancer, 6(9), 714‐727. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16929325. doi:10.1038/nrc1913

Itoh, Y., & Seiki, M. (2004). MT1‐MMP: an enzyme with multidimensional regulation.

Trends Biochem Sci, 29(6), 285‐289. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15276180. doi:10.1016/j.tibs.2004.04.001

Itoh, Y., Takamura, A., Ito, N., Maru, Y., Sato, H., Suenaga, N., . . . Seiki, M. (2001).

Homophilic complex formation of MT1‐MMP facilitates proMMP‐2 activation on the cell surface and promotes tumor cell invasion. EMBO J, 20(17), 4782‐4793. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11532942. doi:10.1093/emboj/20.17.4782

Jacob, A., Jing, J., Lee, J., Schedin, P., Gilbert, S. M., Peden, A. A., . . . Prekeris, R. (2013).

Rab40b regulates trafficking of MMP2 and MMP9 during invadopodia formation and invasion of breast cancer cells. J Cell Sci, 126(Pt 20), 4647‐4658. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23902685. doi:10.1242/jcs.126573

Jacob, S. S., Shastry, P., & Sudhakaran, P. R. (2002). Monocyte‐macrophage differentiation

in vitro: modulation by extracellular matrix protein substratum. Mol Cell Biochem, 233(1‐2), 9‐17. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12083384.

Jahn, R., & Scheller, R. H. (2006). SNAREs‐‐engines for membrane fusion. Nat Rev Mol Cell

Biol, 7(9), 631‐643. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16912714. doi:10.1038/nrm2002

Johnson, J. L., Jenkins, N. P., Huang, W. C., Di Gregoli, K., Sala‐Newby, G. B., Scholtes, V. P.,

. . . Newby, A. C. (2014). Relationship of MMP‐14 and TIMP‐3 expression with macrophage activation and human atherosclerotic plaque vulnerability. Mediators Inflamm, 2014, 276457. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25301980. doi:10.1155/2014/276457

Johnson, K. E., & Wilgus, T. A. (2014). Vascular Endothelial Growth Factor and

Angiogenesis in the Regulation of Cutaneous Wound Repair. Adv Wound Care (New Rochelle), 3(10), 647‐661. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25302139. doi:10.1089/wound.2013.0517











References 195

Jones, M. C., Caswell, P. T., & Norman, J. C. (2006). Endocytic recycling pathways:

emerging regulators of cell migration. Curr Opin Cell Biol, 18(5), 549‐557. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16904305. doi:10.1016/j.ceb.2006.08.003

Kahraman, A., Bronk, S. F., Cazanave, S., Werneburg, N. W., Mott, J. L., Contreras, P. C., &

Gores, G. J. (2009). Matrix metalloproteinase inhibitor, CTS‐1027, attenuates liver injury and fibrosis in the bile duct‐ligated mouse. Hepatol Res, 39(8), 805‐813. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19624765. doi:10.1111/j.1872‐034X.2009.00541.x

Kajita, M., Itoh, Y., Chiba, T., Mori, H., Okada, A., Kinoh, H., & Seiki, M. (2001). Membrane‐

type 1 matrix metalloproteinase cleaves CD44 and promotes cell migration. J Cell Biol, 153(5), 893‐904. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11381077.

Kalluri, R. (2003). Basement membranes: structure, assembly and role in tumour

angiogenesis. Nat Rev Cancer, 3(6), 422‐433. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12778132. doi:10.1038/nrc1094

Kean, M. J., Williams, K. C., Skalski, M., Myers, D., Burtnik, A., Foster, D., & Coppolino, M. G.

(2009). VAMP3, syntaxin‐13 and SNAP23 are involved in secretion of matrix metalloproteinases, degradation of the extracellular matrix and cell invasion. J Cell Sci, 122(Pt 22), 4089‐4098. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19910495. doi:10.1242/jcs.052761

Khanna, S., Biswas, S., Shang, Y., Collard, E., Azad, A., Kauh, C., . . . Roy, S. (2010).

Macrophage dysfunction impairs resolution of inflammation in the wounds of diabetic mice. PLoS One, 5(3), e9539. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20209061. doi:10.1371/journal.pone.0009539

Kim, M. H., Liu, W., Borjesson, D. L., Curry, F. R., Miller, L. S., Cheung, A. L., . . . Simon, S. I.

(2008). Dynamics of neutrophil infiltration during cutaneous wound healing and infection using fluorescence imaging. J Invest Dermatol, 128(7), 1812‐1820. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18185533. doi:10.1038/sj.jid.5701223

Kindzelskii, A. L., Amhad, I., Keller, D., Zhou, M. J., Haugland, R. P., Garni‐Wagner, B. A., . . .

Petty, H. R. (2004). Pericellular proteolysis by leukocytes and tumor cells on substrates: focal activation and the role of urokinase‐type plasminogen activator. Histochem Cell Biol, 121(4), 299‐310. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15042374. doi:10.1007/s00418‐004‐0639‐3

Kleijmeer, M., Ramm, G., Schuurhuis, D., Griffith, J., Rescigno, M., Ricciardi‐Castagnoli, P.,

. . . Geuze, H. J. (2001). Reorganization of multivesicular bodies regulates MHC class II antigen presentation by dendritic cells. J Cell Biol, 155(1), 53‐63. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11581285. doi:10.1083/jcb.200103071

Kleizen, B., & Braakman, I. (2004). Protein folding and quality control in the endoplasmic

reticulum. Curr Opin Cell Biol, 16(4), 343‐349. Retrieved from










196 References

http://www.ncbi.nlm.nih.gov/pubmed/15261665. doi:10.1016/j.ceb.2004.06.012

Klose, A., Zigrino, P., & Mauch, C. (2013). Monocyte/macrophage MMP‐14 modulates cell

infiltration and T‐cell attraction in contact dermatitis but not in murine wound healing. Am J Pathol, 182(3), 755‐764. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23438475. doi:10.1016/j.ajpath.2012.11.028

Kobayashi, T., Beuchat, M. H., Lindsay, M., Frias, S., Palmiter, R. D., Sakuraba, H., . . .

Gruenberg, J. (1999). Late endosomal membranes rich in lysobisphosphatidic acid regulate cholesterol transport. Nat Cell Biol, 1(2), 113‐118. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10559883. doi:10.1038/10084

Koh, T. J., & DiPietro, L. A. (2011). Inflammation and wound healing: the role of the

macrophage. Expert Rev Mol Med, 13, e23. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21740602. doi:10.1017/S1462399411001943

Koo, S. J., Markovic, S., Puchkov, D., Mahrenholz, C. C., Beceren‐Braun, F., Maritzen, T., . . .

Haucke, V. (2011). SNARE motif‐mediated sorting of synaptobrevin by the endocytic adaptors clathrin assembly lymphoid myeloid leukemia (CALM) and AP180 at synapses. Proc Natl Acad Sci U S A, 108(33), 13540‐13545. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21808019. doi:10.1073/pnas.1107067108

Krzewski, K., Gil‐Krzewska, A., Watts, J., Stern, J. N., & Strominger, J. L. (2011). VAMP4‐

and VAMP7‐expressing vesicles are both required for cytotoxic granule exocytosis in NK cells. Eur J Immunol, 41(11), 3323‐3329. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21805468. doi:10.1002/eji.201141582

Laatsch, A., Panteli, M., Sornsakrin, M., Hoffzimmer, B., Grewal, T., & Heeren, J. (2012).

Low density lipoprotein receptor‐related protein 1 dependent endosomal trapping and recycling of apolipoprotein E. PLoS One, 7(1), e29385. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22238606. doi:10.1371/journal.pone.0029385

Lacy, P., & Stow, J. L. (2011). Cytokine release from innate immune cells: association with

diverse membrane trafficking pathways. Blood, 118(1), 9‐18. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21562044. doi:10.1182/blood‐2010‐08‐265892

Ladwig, G. P., Robson, M. C., Liu, R., Kuhn, M. A., Muir, D. F., & Schultz, G. S. (2002). Ratios

of activated matrix metalloproteinase‐9 to tissue inhibitor of matrix metalloproteinase‐1 in wound fluids are inversely correlated with healing of pressure ulcers. Wound Repair Regen, 10(1), 26‐37. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11983004.

Lafleur, M. A., Mercuri, F. A., Ruangpanit, N., Seiki, M., Sato, H., & Thompson, E. W. (2006).

Type I collagen abrogates the clathrin‐mediated internalization of membrane type 1 matrix metalloproteinase (MT1‐MMP) via the MT1‐MMP hemopexin domain. J Biol Chem, 281(10), 6826‐6840. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16407177. doi:10.1074/jbc.M513084200











References 197

Lau, K., Paus, R., Tiede, S., Day, P., & Bayat, A. (2009). Exploring the role of stem cells in

cutaneous wound healing. Exp Dermatol, 18(11), 921‐933. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19719838. doi:10.1111/j.1600‐0625.2009.00942.x

Le, N. T., Xue, M., Castelnoble, L. A., & Jackson, C. J. (2007). The dual personalities of matrix

metalloproteinases in inflammation. Front Biosci, 12, 1475‐1487. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17127395.

Lei, N., Franken, L., Ruzehaji, N., Offenhauser, C., Cowin, A. J., & Murray, R. Z. (2012).

Flightless, secreted through a late endosome/lysosome pathway, binds LPS and dampens cytokine secretion. J Cell Sci, 125(Pt 18), 4288‐4296. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22718342. doi:10.1242/jcs.099507

Leibovich, S. J., & Ross, R. (1975). The role of the macrophage in wound repair. A study

with hydrocortisone and antimacrophage serum. Am J Pathol, 78(1), 71‐100. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1109560.

Leimgruber, A., Berger, C., Cortez‐Retamozo, V., Etzrodt, M., Newton, A. P., Waterman, P.,

. . . Pittet, M. J. (2009). Behavior of endogenous tumor‐associated macrophages assessed in vivo using a functionalized nanoparticle. Neoplasia, 11(5), 459‐468, 452 p following 468. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19412430.

Ley, K., Laudanna, C., Cybulsky, M. I., & Nourshargh, S. (2007). Getting to the site of

inflammation: the leukocyte adhesion cascade updated. Nat Rev Immunol, 7(9), 678‐689. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17717539. doi:10.1038/nri2156

Lim, J. P., & Gleeson, P. A. (2011). Macropinocytosis: an endocytic pathway for

internalising large gulps. Immunol Cell Biol, 89(8), 836‐843. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21423264. doi:10.1038/icb.2011.20

Linder, S. (2007). The matrix corroded: podosomes and invadopodia in extracellular

matrix degradation. Trends Cell Biol, 17(3), 107‐117. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17275303. doi:10.1016/j.tcb.2007.01.002

Linder, S., Nelson, D., Weiss, M., & Aepfelbacher, M. (1999). Wiskott‐Aldrich syndrome

protein regulates podosomes in primary human macrophages. Proc Natl Acad Sci U S A, 96(17), 9648‐9653. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10449748.

Lippert, U., Ferrari, D. M., & Jahn, R. (2007). Endobrevin/VAMP8 mediates exocytotic

release of hexosaminidase from rat basophilic leukaemia cells. FEBS Lett, 581(18), 3479‐3484. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17618625. doi:10.1016/j.febslet.2007.06.057

Lippincott‐Schwartz, J., Cole, N., & Presley, J. (1998). Unravelling Golgi membrane traffic

with green fluorescent protein chimeras. Trends Cell Biol, 8(1), 16‐20. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9695802.












198 References

Lippincott‐Schwartz, J., Yuan, L. C., Bonifacino, J. S., & Klausner, R. D. (1989). Rapid redistribution of Golgi proteins into the ER in cells treated with brefeldin A: evidence for membrane cycling from Golgi to ER. Cell, 56(5), 801‐813. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/2647301.

Liu, S., & Storrie, B. (2012). Are Rab proteins the link between Golgi organization and

membrane trafficking? Cell Mol Life Sci, 69(24), 4093‐4106. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22581368. doi:10.1007/s00018‐012‐1021‐6

Liu, X., Yang, J. M., Zhang, S. S., Liu, X. Y., & Liu, D. X. (2010). Induction of cell cycle arrest

at G1 and S phases and cAMP‐dependent differentiation in C6 glioma by low concentration of cycloheximide. BMC Cancer, 10, 684. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21159181. doi:10.1186/1471‐2407‐10‐684

Liu, Y., Min, D., Bolton, T., Nube, V., Twigg, S. M., Yue, D. K., & McLennan, S. V. (2009).

Increased matrix metalloproteinase‐9 predicts poor wound healing in diabetic foot ulcers. Diabetes Care, 32(1), 117‐119. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18835949. doi:10.2337/dc08‐0763

Lobmann, R., Ambrosch, A., Schultz, G., Waldmann, K., Schiweck, S., & Lehnert, H. (2002).

Expression of matrix‐metalloproteinases and their inhibitors in the wounds of diabetic and non‐diabetic patients. Diabetologia, 45(7), 1011‐1016. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12136400. doi:10.1007/s00125‐002‐0868‐8

Lobmann, R., Zemlin, C., Motzkau, M., Reschke, K., & Lehnert, H. (2006). Expression of

matrix metalloproteinases and growth factors in diabetic foot wounds treated with a protease absorbent dressing. J Diabetes Complications, 20(5), 329‐335. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16949521. doi:10.1016/j.jdiacomp.2005.08.007

Loffek, S., Schilling, O., & Franzke, C. W. (2011). Series "matrix metalloproteinases in lung

health and disease": Biological role of matrix metalloproteinases: a critical balance. Eur Respir J, 38(1), 191‐208. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21177845. doi:10.1183/09031936.00146510

Logan, M. R., Lacy, P., Odemuyiwa, S. O., Steward, M., Davoine, F., Kita, H., & Moqbel, R.

(2006). A critical role for vesicle‐associated membrane protein‐7 in exocytosis from human eosinophils and neutrophils. Allergy, 61(6), 777‐784. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16677249. doi:10.1111/j.1398‐9995.2006.01089.x

Loots, M. A., Lamme, E. N., Zeegelaar, J., Mekkes, J. R., Bos, J. D., & Middelkoop, E. (1998).

Differences in cellular infiltrate and extracellular matrix of chronic diabetic and venous ulcers versus acute wounds. J Invest Dermatol, 111(5), 850‐857. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9804349. doi:10.1046/j.1523‐1747.1998.00381.x

Loskutov, Y. V., Kozyulina, P. Y., Kozyreva, V. K., Ice, R. J., Jones, B. C., Roston, T. J., . . .

Pugacheva, E. N. (2015). NEDD9/Arf6‐dependent endocytic trafficking of matrix metalloproteinase 14: a novel mechanism for blocking mesenchymal cell










References 199

invasion and metastasis of breast cancer. Oncogene, 34(28), 3662‐3675. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25241893. doi:10.1038/onc.2014.297

Low, Q. E., Drugea, I. A., Duffner, L. A., Quinn, D. G., Cook, D. N., Rollins, B. J., . . . DiPietro, L.

A. (2001). Wound healing in MIP‐1alpha(‐/‐) and MCP‐1(‐/‐) mice. Am J Pathol, 159(2), 457‐463. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11485904.

Lucas, T., Waisman, A., Ranjan, R., Roes, J., Krieg, T., Muller, W., . . . Eming, S. A. (2010).

Differential roles of macrophages in diverse phases of skin repair. J Immunol, 184(7), 3964‐3977. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20176743. doi:10.4049/jimmunol.0903356

Luini, A., Mironov, A. A., Polishchuk, E. V., & Polishchuk, R. S. (2008). Morphogenesis of

post‐Golgi transport carriers. Histochem Cell Biol, 129(2), 153‐161. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18214517. doi:10.1007/s00418‐007‐0365‐8

Luzio, J. P., Pryor, P. R., & Bright, N. A. (2007). Lysosomes: fusion and function. Nat Rev

Mol Cell Biol, 8(8), 622‐632. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17637737. doi:10.1038/nrm2217

Luzio, J. P., Pryor, P. R., Gray, S. R., Gratian, M. J., Piper, R. C., & Bright, N. A. (2005).

Membrane traffic to and from lysosomes. Biochem Soc Symp(72), 77‐86. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15649132.

MacDonald, K. P., Palmer, J. S., Cronau, S., Seppanen, E., Olver, S., Raffelt, N. C., . . . Hume,

D. A. (2010). An antibody against the colony‐stimulating factor 1 receptor depletes the resident subset of monocytes and tissue‐ and tumor‐associated macrophages but does not inhibit inflammation. Blood, 116(19), 3955‐3963. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20682855. doi:10.1182/blood‐2010‐02‐266296

Macedo, L., Pinhal‐Enfield, G., Alshits, V., Elson, G., Cronstein, B. N., & Leibovich, S. J.

(2007). Wound healing is impaired in MyD88‐deficient mice: a role for MyD88 in the regulation of wound healing by adenosine A2A receptors. Am J Pathol, 171(6), 1774‐1788. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17974599. doi:10.2353/ajpath.2007.061048

Mackay, A. R., Hartzler, J. L., Pelina, M. D., & Thorgeirsson, U. P. (1990). Studies on the

ability of 65‐kDa and 92‐kDa tumor cell gelatinases to degrade type IV collagen. J Biol Chem, 265(35), 21929‐21934. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/2174891.

Macpherson, I. R., Rainero, E., Mitchell, L. E., van den Berghe, P. V., Speirs, C.,

Dozynkiewicz, M. A., . . . Norman, J. C. (2014). CLIC3 controls recycling of late endosomal MT1‐MMP and dictates invasion and metastasis in breast cancer. J Cell Sci, 127(Pt 18), 3893‐3901. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25015290. doi:10.1242/jcs.135947

Mallard, F., Antony, C., Tenza, D., Salamero, J., Goud, B., & Johannes, L. (1998). Direct

pathway from early/recycling endosomes to the Golgi apparatus revealed











200 References

through the study of shiga toxin B‐fragment transport. J Cell Biol, 143(4), 973‐990. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9817755.

Mallard, F., Tang, B. L., Galli, T., Tenza, D., Saint‐Pol, A., Yue, X., . . . Johannes, L. (2002).

Early/recycling endosomes‐to‐TGN transport involves two SNARE complexes and a Rab6 isoform. J Cell Biol, 156(4), 653‐664. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11839770. doi:10.1083/jcb.200110081

Malsam, J., Kreye, S., & Sollner, T. H. (2008). Membrane fusion: SNAREs and regulation.

Cell Mol Life Sci, 65(18), 2814‐2832. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18726177. doi:10.1007/s00018‐008‐8352‐3

Manderson, A. P., Kay, J. G., Hammond, L. A., Brown, D. L., & Stow, J. L. (2007).

Subcompartments of the macrophage recycling endosome direct the differential secretion of IL‐6 and TNFalpha. J Cell Biol, 178(1), 57‐69. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17606866. doi:10.1083/jcb.200612131

Marchesin, V., Castro‐Castro, A., Lodillinsky, C., Castagnino, A., Cyrta, J., Bonsang‐Kitzis,

H., . . . Chavrier, P. (2015). ARF6‐JIP3/4 regulate endosomal tubules for MT1‐MMP exocytosis in cancer invasion. J Cell Biol, 211(2), 339‐358. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/26504170. doi:10.1083/jcb.201506002

Marom, B., Rahat, M. A., Lahat, N., Weiss‐Cerem, L., Kinarty, A., & Bitterman, H. (2007).

Native and fragmented fibronectin oppositely modulate monocyte secretion of MMP‐9. J Leukoc Biol, 81(6), 1466‐1476. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17327485. doi:10.1189/jlb.0506328

Martens, S., & McMahon, H. T. (2008). Mechanisms of membrane fusion: disparate

players and common principles. Nat Rev Mol Cell Biol, 9(7), 543‐556. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18496517. doi:10.1038/nrm2417

Martin, P., D'Souza, D., Martin, J., Grose, R., Cooper, L., Maki, R., & McKercher, S. R. (2003).

Wound healing in the PU.1 null mouse‐‐tissue repair is not dependent on inflammatory cells. Curr Biol, 13(13), 1122‐1128. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12842011.

Martin, P., & Leibovich, S. J. (2005). Inflammatory cells during wound repair: the good,

the bad and the ugly. Trends Cell Biol, 15(11), 599‐607. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16202600. doi:10.1016/j.tcb.2005.09.002

Martins, V. L., Caley, M., & O'Toole, E. A. (2013). Matrix metalloproteinases and epidermal

wound repair. Cell Tissue Res, 351(2), 255‐268. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22526628. doi:10.1007/s00441‐012‐1410‐z

Matias‐Roman, S., Galvez, B. G., Genis, L., Yanez‐Mo, M., de la Rosa, G., Sanchez‐Mateos, P.,

. . . Arroyo, A. G. (2005). Membrane type 1‐matrix metalloproteinase is involved in migration of human monocytes and is regulated through their interaction with fibronectin or endothelium. Blood, 105(10), 3956‐3964. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15665118. doi:10.1182/blood‐2004‐06‐2382












References 201

Matsuo, H., Chevallier, J., Mayran, N., Le Blanc, I., Ferguson, C., Faure, J., . . . Gruenberg, J. (2004). Role of LBPA and Alix in multivesicular liposome formation and endosome organization. Science, 303(5657), 531‐534. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/14739459. doi:10.1126/science.1092425

Maxfield, F. R., & McGraw, T. E. (2004). Endocytic recycling. Nat Rev Mol Cell Biol, 5(2),

121‐132. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15040445. doi:10.1038/nrm1315

Mazzone, M., Baldassarre, M., Beznoussenko, G., Giacchetti, G., Cao, J., Zucker, S., . . .

Buccione, R. (2004). Intracellular processing and activation of membrane type 1 matrix metalloprotease depends on its partitioning into lipid domains. J Cell Sci, 117(Pt 26), 6275‐6287. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15561768. doi:10.1242/jcs.01563

McCarty, S. M., Percival, S. L., Clegg, P. D., & Cochrane, C. A. (2013). The role of

polyphosphates in the sequestration of matrix metalloproteinases. Int Wound J. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23590276. doi:10.1111/iwj.12058

McKleroy, W., Lee, T. H., & Atabai, K. (2013). Always cleave up your mess: targeting

collagen degradation to treat tissue fibrosis. Am J Physiol Lung Cell Mol Physiol, 304(11), L709‐721. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23564511. doi:10.1152/ajplung.00418.2012

McLauchlan, H., Newell, J., Morrice, N., Osborne, A., West, M., & Smythe, E. (1998). A novel

role for Rab5‐GDI in ligand sequestration into clathrin‐coated pits. Curr Biol, 8(1), 34‐45. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9427626.

Mileski, W. J., Burkhart, D., Hunt, J. L., Kagan, R. J., Saffle, J. R., Herndon, D. N., . . .

Hansborough, J. (2003). Clinical effects of inhibiting leukocyte adhesion with monoclonal antibody to intercellular adhesion molecule‐1 (enlimomab) in the treatment of partial‐thickness burn injury. J Trauma, 54(5), 950‐958. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12777909. doi:10.1097/01.TA.0000030626.84680.11

Millar, A. L., Pavios, N. J., Xu, J., & Zheng, M. H. (2002). Rab3D: a regulator of exocytosis in

non‐neuronal cells. Histol Histopathol, 17(3), 929‐936. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12168804.

Miller, S. E., Sahlender, D. A., Graham, S. C., Honing, S., Robinson, M. S., Peden, A. A., &

Owen, D. J. (2011). The molecular basis for the endocytosis of small R‐SNAREs by the clathrin adaptor CALM. Cell, 147(5), 1118‐1131. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22118466. doi:10.1016/j.cell.2011.10.038

Ming, X. F., Rajapakse, A. G., Yepuri, G., Xiong, Y., Carvas, J. M., Ruffieux, J., . . . Yang, Z.

(2012). Arginase II Promotes Macrophage Inflammatory Responses Through Mitochondrial Reactive Oxygen Species, Contributing to Insulin Resistance and Atherogenesis. J Am Heart Assoc, 1(4), e000992. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23130157. doi:10.1161/JAHA.112.000992











202 References

Mirastschijski, U., Haaksma, C. J., Tomasek, J. J., & Agren, M. S. (2004). Matrix

metalloproteinase inhibitor GM 6001 attenuates keratinocyte migration, contraction and myofibroblast formation in skin wounds. Exp Cell Res, 299(2), 465‐475. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15350544. doi:10.1016/j.yexcr.2004.06.007

Mirza, R., DiPietro, L. A., & Koh, T. J. (2009). Selective and specific macrophage ablation

is detrimental to wound healing in mice. Am J Pathol, 175(6), 2454‐2462. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19850888. doi:10.2353/ajpath.2009.090248

Mirza, R., & Koh, T. J. (2011). Dysregulation of monocyte/macrophage phenotype in

wounds of diabetic mice. Cytokine, 56(2), 256‐264. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21803601. doi:10.1016/j.cyto.2011.06.016

Mirza, R. E., Fang, M. M., Ennis, W. J., & Koh, T. J. (2013). Blocking interleukin‐1beta

induces a healing‐associated wound macrophage phenotype and improves healing in type 2 diabetes. Diabetes, 62(7), 2579‐2587. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23493576. doi:10.2337/db12‐1450

Mirza, R. E., Fang, M. M., Weinheimer‐Haus, E. M., Ennis, W. J., & Koh, T. J. (2014).

Sustained inflammasome activity in macrophages impairs wound healing in type 2 diabetic humans and mice. Diabetes, 63(3), 1103‐1114. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24194505. doi:10.2337/db13‐0927

Miyoshi, H., Kanekura, T., Aoki, T., & Kanzaki, T. (2005). Beneficial effects of tissue

inhibitor of metalloproteinases‐2 (TIMP‐2) on chronic dermatitis. J Dermatol, 32(5), 346‐353. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16043896.

Mizushima, N. (2007). Autophagy: process and function. Genes Dev, 21(22), 2861‐2873.

Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18006683. doi:10.1101/gad.1599207

Mollinedo, F., Martin‐Martin, B., Calafat, J., Nabokina, S. M., & Lazo, P. A. (2003). Role of

vesicle‐associated membrane protein‐2, through Q‐soluble N‐ethylmaleimide‐sensitive factor attachment protein receptor/R‐soluble N‐ethylmaleimide‐sensitive factor attachment protein receptor interaction, in the exocytosis of specific and tertiary granules of human neutrophils. J Immunol, 170(2), 1034‐1042. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12517971.

Monteiro, P., Rosse, C., Castro‐Castro, A., Irondelle, M., Lagoutte, E., Paul‐Gilloteaux, P., . .

. Chavrier, P. (2013). Endosomal WASH and exocyst complexes control exocytosis of MT1‐MMP at invadopodia. J Cell Biol, 203(6), 1063‐1079. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24344185. doi:10.1083/jcb.201306162

Mori, R., Kondo, T., Ohshima, T., Ishida, Y., & Mukaida, N. (2002). Accelerated wound

healing in tumor necrosis factor receptor p55‐deficient mice with reduced leukocyte infiltration. FASEB J, 16(9), 963‐974. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12087057. doi:10.1096/fj.01‐0776com











References 203

Mosser, D. M., & Edwards, J. P. (2008). Exploring the full spectrum of macrophage activation. Nat Rev Immunol, 8(12), 958‐969. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19029990. doi:10.1038/nri2448

Moues, C. M., van Toorenenbergen, A. W., Heule, F., Hop, W. C., & Hovius, S. E. (2008). The

role of topical negative pressure in wound repair: expression of biochemical markers in wound fluid during wound healing. Wound Repair Regen, 16(4), 488‐494. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18638266. doi:10.1111/j.1524‐475X.2008.00395.x

Muller, W. A. (2003). Leukocyte‐endothelial‐cell interactions in leukocyte

transmigration and the inflammatory response. Trends Immunol, 24(6), 327‐334. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12810109.

Murphy, D. A., & Courtneidge, S. A. (2011). The 'ins' and 'outs' of podosomes and

invadopodia: characteristics, formation and function. Nat Rev Mol Cell Biol, 12(7), 413‐426. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21697900. doi:10.1038/nrm3141

Murray, M. Y., Birkland, T. P., Howe, J. D., Rowan, A. D., Fidock, M., Parks, W. C., &

Gavrilovic, J. (2013). Macrophage migration and invasion is regulated by MMP10 expression. PLoS One, 8(5), e63555. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23691065. doi:10.1371/journal.pone.0063555

Murray, R. Z. (2013). Quantifying alterations in cell migration: tracking fluorescently‐

tagged migrating cells by FACs and live‐imaging. Methods Mol Biol, 1037, 477‐494. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24029954. doi:10.1007/978‐1‐62703‐505‐7_28

Murray, R. Z., Kay, J. G., Sangermani, D. G., & Stow, J. L. (2005). A role for the phagosome

in cytokine secretion. Science, 310(5753), 1492‐1495. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16282525. doi:10.1126/science.1120225

Murray, R. Z., & Stow, J. L. (2014). Cytokine Secretion in Macrophages: SNAREs, Rabs, and

Membrane Trafficking. Front Immunol, 5, 538. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25386181. doi:10.3389/fimmu.2014.00538

Murray, R. Z., Wylie, F. G., Khromykh, T., Hume, D. A., & Stow, J. L. (2005). Syntaxin 6 and

Vti1b form a novel SNARE complex, which is up‐regulated in activated macrophages to facilitate exocytosis of tumor necrosis Factor‐alpha. J Biol Chem, 280(11), 10478‐10483. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15640147. doi:10.1074/jbc.M414420200

Mwaura, B., Mahendran, B., Hynes, N., Defreitas, D., Avalos, G., Adegbola, T., . . . Sultan, S.

(2006). The impact of differential expression of extracellular matrix metalloproteinase inducer, matrix metalloproteinase‐2, tissue inhibitor of matrix metalloproteinase‐2 and PDGF‐AA on the chronicity of venous leg ulcers. Eur J Vasc Endovasc Surg, 31(3), 306‐310. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16169261. doi:10.1016/j.ejvs.2005.08.007











204 References

Nagaoka, T., Kaburagi, Y., Hamaguchi, Y., Hasegawa, M., Takehara, K., Steeber, D. A., . . .

Sato, S. (2000). Delayed wound healing in the absence of intercellular adhesion molecule‐1 or L‐selectin expression. Am J Pathol, 157(1), 237‐247. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10880393. doi:10.1016/S0002‐9440(10)64534‐8

Newby, A. C. (2008). Metalloproteinase expression in monocytes and macrophages and

its relationship to atherosclerotic plaque instability. Arterioscler Thromb Vasc Biol, 28(12), 2108‐2114. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18772495. doi:10.1161/ATVBAHA.108.173898

Nielsen, E., Christoforidis, S., Uttenweiler‐Joseph, S., Miaczynska, M., Dewitte, F., Wilm, M.,

. . . Zerial, M. (2000). Rabenosyn‐5, a novel Rab5 effector, is complexed with hVPS45 and recruited to endosomes through a FYVE finger domain. J Cell Biol, 151(3), 601‐612. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11062261.

Nishimura, K., Watanabe, S., Hayashida, R., Sugishima, S., Iwasaka, T., & Kaku, T. (2015).

Binucleated HeLa cells are formed by cytokinesis failure in starvation and keep the potential of proliferation. Cytotechnology. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25894790. doi:10.1007/s10616‐015‐9869‐6

Norgauer, J., Hildenbrand, T., Idzko, M., Panther, E., Bandemir, E., Hartmann, M., . . .

Herouy, Y. (2002). Elevated expression of extracellular matrix metalloproteinase inducer (CD147) and membrane‐type matrix metalloproteinases in venous leg ulcers. Br J Dermatol, 147(6), 1180‐1186. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12452868.

Nusblat, L. M., Dovas, A., & Cox, D. (2011). The non‐redundant role of N‐WASP in

podosome‐mediated matrix degradation in macrophages. Eur J Cell Biol, 90(2‐3), 205‐212. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20732727. doi:10.1016/j.ejcb.2010.07.012

Okamura, Y., Watari, M., Jerud, E. S., Young, D. W., Ishizaka, S. T., Rose, J., . . . Strauss, J. F.,

3rd. (2001). The extra domain A of fibronectin activates Toll‐like receptor 4. J Biol Chem, 276(13), 10229‐10233. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11150311. doi:10.1074/jbc.M100099200

Olczyk, P., Mencner, L., & Komosinska‐Vassev, K. (2014). The role of the extracellular

matrix components in cutaneous wound healing. Biomed Res Int, 2014, 747584. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24772435. doi:10.1155/2014/747584

Olson, M. W., Bernardo, M. M., Pietila, M., Gervasi, D. C., Toth, M., Kotra, L. P., . . . Fridman,

R. (2000). Characterization of the monomeric and dimeric forms of latent and active matrix metalloproteinase‐9. Differential rates for activation by stromelysin 1. J Biol Chem, 275(4), 2661‐2668. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10644727.










References 205

Opdenakker, G., Van den Steen, P. E., Dubois, B., Nelissen, I., Van Coillie, E., Masure, S., . . . Van Damme, J. (2001). Gelatinase B functions as regulator and effector in leukocyte biology. J Leukoc Biol, 69(6), 851‐859. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11404367.

Otto, O., Rosendahl, P., Mietke, A., Golfier, S., Herold, C., Klaue, D., . . . Guck, J. (2015). Real‐

time deformability cytometry: on‐the‐fly cell mechanical phenotyping. Nat Methods, 12(3), 199‐202, 194 p following 202. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25643151. doi:10.1038/nmeth.3281

Overington, J. P., Al‐Lazikani, B., & Hopkins, A. L. (2006). How many drug targets are

there? Nat Rev Drug Discov, 5(12), 993‐996. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17139284. doi:10.1038/nrd2199

Pagan, J. K., Wylie, F. G., Joseph, S., Widberg, C., Bryant, N. J., James, D. E., & Stow, J. L.

(2003). The t‐SNARE syntaxin 4 is regulated during macrophage activation to function in membrane traffic and cytokine secretion. Curr Biol, 13(2), 156‐160. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12546791.

Paumet, F., Le Mao, J., Martin, S., Galli, T., David, B., Blank, U., & Roa, M. (2000). Soluble

NSF attachment protein receptors (SNAREs) in RBL‐2H3 mast cells: functional role of syntaxin 4 in exocytosis and identification of a vesicle‐associated membrane protein 8‐containing secretory compartment. J Immunol, 164(11), 5850‐5857. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10820264.

Pavlos, N. J., Xu, J., Riedel, D., Yeoh, J. S., Teitelbaum, S. L., Papadimitriou, J. M., . . . Zheng,

M. H. (2005). Rab3D regulates a novel vesicular trafficking pathway that is required for osteoclastic bone resorption. Mol Cell Biol, 25(12), 5253‐5269. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15923639. doi:10.1128/MCB.25.12.5253‐5269.2005

Perry, H. D., & Golub, L. M. (1985). Systemic tetracyclines in the treatment of noninfected

corneal ulcers: a case report and proposed new mechanism of action. Ann Ophthalmol, 17(12), 742‐744. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/4091374.

Peters, T., Sindrilaru, A., Hinz, B., Hinrichs, R., Menke, A., Al‐Azzeh, E. A., . . . Scharffetter‐

Kochanek, K. (2005). Wound‐healing defect of CD18(‐/‐) mice due to a decrease in TGF‐beta1 and myofibroblast differentiation. EMBO J, 24(19), 3400‐3410. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16148944. doi:10.1038/sj.emboj.7600809

Piccinini, A. M., & Midwood, K. S. (2010). DAMPening inflammation by modulating TLR

signalling. Mediators Inflamm, 2010. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20706656. doi:10.1155/2010/672395

Pickett, A., & Perrow, K. (2011). Towards new uses of botulinum toxin as a novel

therapeutic tool. Toxins (Basel), 3(1), 63‐81. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22069690. doi:10.3390/toxins3010063

Pierce, G. F. (2001). Inflammation in nonhealing diabetic wounds: the space‐time

continuum does matter. Am J Pathol, 159(2), 399‐403. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11485896. doi:10.1016/S0002‐9440(10)61709‐9












206 References

Pirila, E., Korpi, J. T., Korkiamaki, T., Jahkola, T., Gutierrez‐Fernandez, A., Lopez‐Otin, C., .

. . Sorsa, T. (2007). Collagenase‐2 (MMP‐8) and matrilysin‐2 (MMP‐26) expression in human wounds of different etiologies. Wound Repair Regen, 15(1), 47‐57. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17244319. doi:10.1111/j.1524‐475X.2006.00184.x

Poincloux, R., Lizarraga, F., & Chavrier, P. (2009). Matrix invasion by tumour cells: a focus

on MT1‐MMP trafficking to invadopodia. J Cell Sci, 122(Pt 17), 3015‐3024. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19692588. doi:10.1242/jcs.034561

Porat‐Shliom, N., Milberg, O., Masedunskas, A., & Weigert, R. (2013). Multiple roles for

the actin cytoskeleton during regulated exocytosis. Cell Mol Life Sci, 70(12), 2099‐2121. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22986507. doi:10.1007/s00018‐012‐1156‐5

Powelka, A. M., Sun, J., Li, J., Gao, M., Shaw, L. M., Sonnenberg, A., & Hsu, V. W. (2004).

Stimulation‐dependent recycling of integrin beta1 regulated by ARF6 and Rab11. Traffic, 5(1), 20‐36. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/14675422.

Pryor, P. R., Mullock, B. M., Bright, N. A., Lindsay, M. R., Gray, S. R., Richardson, S. C., . . .

Luzio, J. P. (2004). Combinatorial SNARE complexes with VAMP7 or VAMP8 define different late endocytic fusion events. EMBO Rep, 5(6), 590‐595. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15133481. doi:10.1038/sj.embor.7400150

Puri, C., Renna, M., Bento, C. F., Moreau, K., & Rubinsztein, D. C. (2014). ATG16L1 meets

ATG9 in recycling endosomes: additional roles for the plasma membrane and endocytosis in autophagosome biogenesis. Autophagy, 10(1), 182‐184. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24257061. doi:10.4161/auto.27174

Quaranta, V. (2000). Cell migration through extracellular matrix: membrane‐type

metalloproteinases make the way. J Cell Biol, 149(6), 1167‐1170. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10851014.

Raffetto, J. D. (2013). Inflammation in chronic venous ulcers. Phlebology, 28 Suppl 1, 61‐

67. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23482537. doi:10.1177/0268355513476844

Rapoport, T. A. (1992). Transport of proteins across the endoplasmic reticulum

membrane. Science, 258(5084), 931‐936. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1332192.

Rayment, E. A., Upton, Z., & Shooter, G. K. (2008). Increased matrix metalloproteinase‐9

(MMP‐9) activity observed in chronic wound fluid is related to the clinical severity of the ulcer. Br J Dermatol, 158(5), 951‐961. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18284390. doi:10.1111/j.1365‐2133.2008.08462.x

Reel, B., Sala‐Newby, G. B., Huang, W. C., & Newby, A. C. (2011). Diverse patterns of

cyclooxygenase‐independent metalloproteinase gene regulation in human











References 207

monocytes. Br J Pharmacol, 163(8), 1679‐1690. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21371008. doi:10.1111/j.1476‐5381.2011.01298.x

Reichel, C. A., Rehberg, M., Lerchenberger, M., Berberich, N., Bihari, P., Khandoga, A. G., . .

. Krombach, F. (2009). Ccl2 and Ccl3 mediate neutrophil recruitment via induction of protein synthesis and generation of lipid mediators. Arterioscler Thromb Vasc Biol, 29(11), 1787‐1793. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19608967. doi:10.1161/ATVBAHA.109.193268

Reiss, M. J., Han, Y. P., Garcia, E., Goldberg, M., Yu, H., & Garner, W. L. (2010). Matrix

metalloproteinase‐9 delays wound healing in a murine wound model. Surgery, 147(2), 295‐302. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20004432. doi:10.1016/j.surg.2009.10.016

Remacle, A., Murphy, G., & Roghi, C. (2003). Membrane type I‐matrix metalloproteinase

(MT1‐MMP) is internalised by two different pathways and is recycled to the cell surface. J Cell Sci, 116(Pt 19), 3905‐3916. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12915589. doi:10.1242/jcs.00710

Remacle, A. G., Chekanov, A. V., Golubkov, V. S., Savinov, A. Y., Rozanov, D. V., & Strongin,

A. Y. (2006). O‐glycosylation regulates autolysis of cellular membrane type‐1 matrix metalloproteinase (MT1‐MMP). J Biol Chem, 281(25), 16897‐16905. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16627478. doi:10.1074/jbc.M600295200

Remacle, A. G., Rozanov, D. V., Baciu, P. C., Chekanov, A. V., Golubkov, V. S., & Strongin, A.

Y. (2005). The transmembrane domain is essential for the microtubular trafficking of membrane type‐1 matrix metalloproteinase (MT1‐MMP). J Cell Sci, 118(Pt 21), 4975‐4984. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16219679. doi:10.1242/jcs.02610

Rhee, J. W., Lee, K. W., Kim, D., Lee, Y., Jeon, O. H., Kwon, H. J., & Kim, D. S. (2007). NF‐

kappaB‐dependent regulation of matrix metalloproteinase‐9 gene expression by lipopolysaccharide in a macrophage cell line RAW 264.7. J Biochem Mol Biol, 40(1), 88‐94. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17244487.

Rodero, M. P., Hodgson, S. S., Hollier, B., Combadiere, C., & Khosrotehrani, K. (2013).

Reduced Il17a expression distinguishes a Ly6c(lo)MHCII(hi) macrophage population promoting wound healing. J Invest Dermatol, 133(3), 783‐792. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23235530. doi:10.1038/jid.2012.368

Rodero, M. P., Licata, F., Poupel, L., Hamon, P., Khosrotehrani, K., Combadiere, C., &

Boissonnas, A. (2014). In vivo imaging reveals a pioneer wave of monocyte recruitment into mouse skin wounds. PLoS One, 9(10), e108212. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/25272047. doi:10.1371/journal.pone.0108212

Rosner, K., Ross, C., Karlsmark, T., Petersen, A. A., Gottrup, F., & Vejlsgaard, G. L. (1995).

Immunohistochemical characterization of the cutaneous cellular infiltrate in










208 References

different areas of chronic leg ulcers. APMIS, 103(4), 293‐299. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7612260.

Rozario, T., & DeSimone, D. W. (2010). The extracellular matrix in development and

morphogenesis: a dynamic view. Dev Biol, 341(1), 126‐140. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19854168. doi:10.1016/j.ydbio.2009.10.026

Rubtsova, S. N., Kondratov, R. V., Kopnin, P. B., Chumakov, P. M., Kopnin, B. P., & Vasiliev,

J. M. (1998). Disruption of actin microfilaments by cytochalasin D leads to activation of p53. FEBS Lett, 430(3), 353‐357. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9688570.

Rundhaug, J. E. (2005). Matrix metalloproteinases and angiogenesis. J Cell Mol Med, 9(2),

267‐285. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15963249. Sakagami, H., Kishino, K., Amano, O., Kanda, Y., Kunii, S., Yokote, Y., . . . Oizumi, T. (2009).

Cell death induced by nutritional starvation in mouse macrophage‐like RAW264.7 cells. Anticancer Res, 29(1), 343‐347. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19331171.

Sakamoto, T., & Seiki, M. (2009). Cytoplasmic tail of MT1‐MMP regulates macrophage

motility independently from its protease activity. Genes Cells, 14(5), 617‐626. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19371380. doi:10.1111/j.1365‐2443.2009.01293.x

Sbai, O., Ould‐Yahoui, A., Ferhat, L., Gueye, Y., Bernard, A., Charrat, E., . . . Khrestchatisky,

M. (2010). Differential vesicular distribution and trafficking of MMP‐2, MMP‐9, and their inhibitors in astrocytes. Glia, 58(3), 344‐366. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19780201. doi:10.1002/glia.20927

Schonbeck, U., Mach, F., & Libby, P. (1998). Generation of biologically active IL‐1 beta by

matrix metalloproteinases: a novel caspase‐1‐independent pathway of IL‐1 beta processing. J Immunol, 161(7), 3340‐3346. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9759850.

Schultz, G. S., Davidson, J. M., Kirsner, R. S., Bornstein, P., & Herman, I. M. (2011). Dynamic

reciprocity in the wound microenvironment. Wound Repair Regen, 19(2), 134‐148. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21362080. doi:10.1111/j.1524‐475X.2011.00673.x

Schultz, G. S., Sibbald, R. G., Falanga, V., Ayello, E. A., Dowsett, C., Harding, K., . . .

Vanscheidt, W. (2003). Wound bed preparation: a systematic approach to wound management. Wound Repair Regen, 11 Suppl 1, S1‐28. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12654015.

Schultz, G. S., Strelow, S., Stern, G. A., Chegini, N., Grant, M. B., Galardy, R. E., . . . et al.

(1992). Treatment of alkali‐injured rabbit corneas with a synthetic inhibitor of matrix metalloproteinases. Invest Ophthalmol Vis Sci, 33(12), 3325‐3331. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1385350.

Schultz, G. S., & Wysocki, A. (2009). Interactions between extracellular matrix and growth

factors in wound healing. Wound Repair Regen, 17(2), 153‐162. Retrieved from












References 209

http://www.ncbi.nlm.nih.gov/pubmed/19320882. doi:10.1111/j.1524‐475X.2009.00466.x

Serra, R., Buffone, G., Falcone, D., Molinari, V., Scaramuzzino, M., Gallelli, L., & de

Franciscis, S. (2013). Chronic venous leg ulcers are associated with high levels of metalloproteinases‐9 and neutrophil gelatinase‐associated lipocalin. Wound Repair Regen, 21(3), 395‐401. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23531093. doi:10.1111/wrr.12035

Simonsen, A., Lippe, R., Christoforidis, S., Gaullier, J. M., Brech, A., Callaghan, J., . . .

Stenmark, H. (1998). EEA1 links PI(3)K function to Rab5 regulation of endosome fusion. Nature, 394(6692), 494‐498. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9697774. doi:10.1038/28879

Sina, A., Lord‐Dufour, S., & Annabi, B. (2009). Cell‐based evidence for aminopeptidase

N/CD13 inhibitor actinonin targeting of MT1‐MMP‐mediated proMMP‐2 activation. Cancer Lett, 279(2), 171‐176. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19264392. doi:10.1016/j.canlet.2009.01.032

Sindrilaru, A., Peters, T., Wieschalka, S., Baican, C., Baican, A., Peter, H., . . . Scharffetter‐

Kochanek, K. (2011). An unrestrained proinflammatory M1 macrophage population induced by iron impairs wound healing in humans and mice. J Clin Invest, 121(3), 985‐997. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21317534. doi:10.1172/JCI44490

Sithu, S. D., English, W. R., Olson, P., Krubasik, D., Baker, A. H., Murphy, G., & D'Souza, S. E.

(2007). Membrane‐type 1‐matrix metalloproteinase regulates intracellular adhesion molecule‐1 (ICAM‐1)‐mediated monocyte transmigration. J Biol Chem, 282(34), 25010‐25019. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17591781. doi:10.1074/jbc.M611273200

Smeets, R., Ulrich, D., Unglaub, F., Woltje, M., & Pallua, N. (2008). Effect of oxidised

regenerated cellulose/collagen matrix on proteases in wound exudate of patients with chronic venous ulceration. Int Wound J, 5(2), 195‐203. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18494625. doi:10.1111/j.1742‐481X.2007.00367.x

Smith, G. N., Jr., Mickler, E. A., Hasty, K. A., & Brandt, K. D. (1999). Specificity of inhibition

of matrix metalloproteinase activity by doxycycline: relationship to structure of the enzyme. Arthritis Rheum, 42(6), 1140‐1146. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10366106. doi:10.1002/1529‐0131(199906)42:6<1140::AID‐ANR10>3.0.CO;2‐7

Soldati, T., & Schliwa, M. (2006). Powering membrane traffic in endocytosis and

recycling. Nat Rev Mol Cell Biol, 7(12), 897‐908. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17139330. doi:10.1038/nrm2060

Song, F., Wisithphrom, K., Zhou, J., & Windsor, L. J. (2006). Matrix metalloproteinase

dependent and independent collagen degradation. Front Biosci, 11, 3100‐3120. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16720379.











210 References

Stamenkovic, I. (2003). Extracellular matrix remodelling: the role of matrix metalloproteinases. J Pathol, 200(4), 448‐464. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12845612. doi:10.1002/path.1400

Stanley, A. C., & Lacy, P. (2010). Pathways for cytokine secretion. Physiology (Bethesda),

25(4), 218‐229. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20699468. doi:10.1152/physiol.00017.2010

Stanley, A. C., Lieu, Z. Z., Wall, A. A., Venturato, J., Khromykh, T., Hamilton, N. A., . . . Stow,

J. L. (2012). Recycling endosome‐dependent and ‐independent mechanisms for IL‐10 secretion in LPS‐activated macrophages. J Leukoc Biol, 92(6), 1227‐1239. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23012430. doi:10.1189/jlb.0412191

Stechmiller, J., Cowan, L., & Schultz, G. (2010). The role of doxycycline as a matrix

metalloproteinase inhibitor for the treatment of chronic wounds. Biol Res Nurs, 11(4), 336‐344. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20031955. doi:10.1177/1099800409346333

Steffen, A., Le Dez, G., Poincloux, R., Recchi, C., Nassoy, P., Rottner, K., . . . Chavrier, P.

(2008). MT1‐MMP‐dependent invasion is regulated by TI‐VAMP/VAMP7. Curr Biol, 18(12), 926‐931. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18571410. doi:10.1016/j.cub.2008.05.044

Stenmark, H. (2009). Rab GTPases as coordinators of vesicle traffic. Nat Rev Mol Cell Biol,

10(8), 513‐525. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19603039. doi:10.1038/nrm2728

Stow, J. L., Low, P. C., Offenhauser, C., & Sangermani, D. (2009). Cytokine secretion in

macrophages and other cells: pathways and mediators. Immunobiology, 214(7), 601‐612. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19268389. doi:10.1016/j.imbio.2008.11.005

Stow, J. L., Manderson, A. P., & Murray, R. Z. (2006). SNAREing immunity: the role of

SNAREs in the immune system. Nat Rev Immunol, 6(12), 919‐929. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17124513. doi:10.1038/nri1980

Stow, J. L., & Murray, R. Z. (2013). Intracellular trafficking and secretion of inflammatory

cytokines. Cytokine Growth Factor Rev, 24(3), 227‐239. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23647915. doi:10.1016/j.cytogfr.2013.04.001

Stramer, B. M., Mori, R., & Martin, P. (2007). The inflammation‐fibrosis link? A Jekyll and

Hyde role for blood cells during wound repair. J Invest Dermatol, 127(5), 1009‐1017. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17435786. doi:10.1038/sj.jid.5700811

Streit, M., Beleznay, Z., & Braathen, L. R. (2006). Topical application of the tumour

necrosis factor‐alpha antibody infliximab improves healing of chronic wounds. Int Wound J, 3(3), 171‐179. Retrieved from











References 211

http://www.ncbi.nlm.nih.gov/pubmed/16984574. doi:10.1111/j.1742‐481X.2006.00233.x

Subramaniam, M., Saffaripour, S., Van De Water, L., Frenette, P. S., Mayadas, T. N., Hynes,

R. O., & Wagner, D. D. (1997). Role of endothelial selectins in wound repair. Am J Pathol, 150(5), 1701‐1709. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9137094.

Szpaderska, A. M., Zuckerman, J. D., & DiPietro, L. A. (2003). Differential injury responses

in oral mucosal and cutaneous wounds. J Dent Res, 82(8), 621‐626. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12885847.

Takino, T., Miyamori, H., Kawaguchi, N., Uekita, T., Seiki, M., & Sato, H. (2003).

Tetraspanin CD63 promotes targeting and lysosomal proteolysis of membrane‐type 1 matrix metalloproteinase. Biochem Biophys Res Commun, 304(1), 160‐166. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12705901.

Tarlton, J. F., Bailey, A. J., Crawford, E., Jones, D., Moore, K., & Harding, K. D. (1999).

Prognostic value of markers of collagen remodeling in venous ulcers. Wound Repair Regen, 7(5), 347‐355. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10564563.

Toth, M., Chvyrkova, I., Bernardo, M. M., Hernandez‐Barrantes, S., & Fridman, R. (2003).

Pro‐MMP‐9 activation by the MT1‐MMP/MMP‐2 axis and MMP‐3: role of TIMP‐2 and plasma membranes. Biochem Biophys Res Commun, 308(2), 386‐395. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12901881.

Tran, T. H., Zeng, Q., & Hong, W. (2007). VAMP4 cycles from the cell surface to the trans‐

Golgi network via sorting and recycling endosomes. J Cell Sci, 120(Pt 6), 1028‐1041. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17327277. doi:10.1242/jcs.03387

van den Hoorn, T., Paul, P., Jongsma, M. L., & Neefjes, J. (2011). Routes to manipulate MHC

class II antigen presentation. Curr Opin Immunol, 23(1), 88‐95. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21112200. doi:10.1016/j.coi.2010.11.002

Van den Steen, P. E., Proost, P., Grillet, B., Brand, D. D., Kang, A. H., Van Damme, J., &

Opdenakker, G. (2002). Cleavage of denatured natural collagen type II by neutrophil gelatinase B reveals enzyme specificity, post‐translational modifications in the substrate, and the formation of remnant epitopes in rheumatoid arthritis. FASEB J, 16(3), 379‐389. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11874987. doi:10.1096/fj.01‐0688com

Van den Steen, P. E., Van Aelst, I., Hvidberg, V., Piccard, H., Fiten, P., Jacobsen, C., . . .

Opdenakker, G. (2006). The hemopexin and O‐glycosylated domains tune gelatinase B/MMP‐9 bioavailability via inhibition and binding to cargo receptors. J Biol Chem, 281(27), 18626‐18637. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16672230. doi:10.1074/jbc.M512308200

van der Goot, F. G., & Gruenberg, J. (2006). Intra‐endosomal membrane traffic. Trends Cell

Biol, 16(10), 514‐521. Retrieved from











212 References

http://www.ncbi.nlm.nih.gov/pubmed/16949287. doi:10.1016/j.tcb.2006.08.003

Van Goethem, E., Guiet, R., Balor, S., Charriere, G. M., Poincloux, R., Labrousse, A., . . . Le

Cabec, V. (2011). Macrophage podosomes go 3D. Eur J Cell Biol, 90(2‐3), 224‐236. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20801545. doi:10.1016/j.ejcb.2010.07.011

Van Goethem, E., Poincloux, R., Gauffre, F., Maridonneau‐Parini, I., & Le Cabec, V. (2010).

Matrix architecture dictates three‐dimensional migration modes of human macrophages: differential involvement of proteases and podosome‐like structures. J Immunol, 184(2), 1049‐1061. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20018633. doi:10.4049/jimmunol.0902223

Vandooren, J., Geurts, N., Martens, E., Van den Steen, P. E., & Opdenakker, G. (2013).

Zymography methods for visualizing hydrolytic enzymes. Nat Methods, 10(3), 211‐220. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23443633. doi:10.1038/nmeth.2371

Veale, K. J., Offenhauser, C., & Murray, R. Z. (2011). The role of the recycling endosome in

regulating lamellipodia formation and macrophage migration. Commun Integr Biol, 4(1), 44‐47. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21509176. doi:10.4161/cib.4.1.13569

Veale, K. J., Offenhauser, C., Whittaker, S. P., Estrella, R. P., & Murray, R. Z. (2010).

Recycling endosome membrane incorporation into the leading edge regulates lamellipodia formation and macrophage migration. Traffic, 11(10), 1370‐1379. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20604897. doi:10.1111/j.1600‐0854.2010.01094.x

Verollet, C., Charriere, G. M., Labrousse, A., Cougoule, C., Le Cabec, V., & Maridonneau‐

Parini, I. (2011). Extracellular proteolysis in macrophage migration: losing grip for a breakthrough. Eur J Immunol, 41(10), 2805‐2813. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21953638. doi:10.1002/eji.201141538

Vu, T. H., & Werb, Z. (2000). Matrix metalloproteinases: effectors of development and

normal physiology. Genes Dev, 14(17), 2123‐2133. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10970876.

Vyas, J. M., Kim, Y. M., Artavanis‐Tsakonas, K., Love, J. C., Van der Veen, A. G., & Ploegh, H.

L. (2007). Tubulation of class II MHC compartments is microtubule dependent and involves multiple endolysosomal membrane proteins in primary dendritic cells. J Immunol, 178(11), 7199‐7210. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17513769.

Wacker, I., Kaether, C., Kromer, A., Migala, A., Almers, W., & Gerdes, H. H. (1997).

Microtubule‐dependent transport of secretory vesicles visualized in real time with a GFP‐tagged secretory protein. J Cell Sci, 110 ( Pt 13), 1453‐1463. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9224763.

Wall, S. J., Bevan, D., Thomas, D. W., Harding, K. G., Edwards, D. R., & Murphy, G. (2002).

Differential expression of matrix metalloproteinases during impaired wound healing of the diabetes mouse. J Invest Dermatol, 119(1), 91‐98. Retrieved from











References 213

http://www.ncbi.nlm.nih.gov/pubmed/12164930. doi:10.1046/j.1523‐1747.2002.01779.x

Wang, X., Ma, D., Keski‐Oja, J., & Pei, D. (2004). Co‐recycling of MT1‐MMP and MT3‐MMP

through the trans‐Golgi network. Identification of DKV582 as a recycling signal. J Biol Chem, 279(10), 9331‐9336. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/14665622. doi:10.1074/jbc.M312369200

Wetzler, C., Kampfer, H., Stallmeyer, B., Pfeilschifter, J., & Frank, S. (2000). Large and

sustained induction of chemokines during impaired wound healing in the genetically diabetic mouse: prolonged persistence of neutrophils and macrophages during the late phase of repair. J Invest Dermatol, 115(2), 245‐253. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10951242. doi:10.1046/j.1523‐1747.2000.00029.x

Wiesner, C., El Azzouzi, K., & Linder, S. (2013). A specific subset of RabGTPases controls

cell surface exposure of MT1‐MMP, extracellular matrix degradation and three‐dimensional invasion of macrophages. J Cell Sci, 126(Pt 13), 2820‐2833. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23606746. doi:10.1242/jcs.122358

Wiesner, C., Faix, J., Himmel, M., Bentzien, F., & Linder, S. (2010). KIF5B and KIF3A/KIF3B

kinesins drive MT1‐MMP surface exposure, CD44 shedding, and extracellular matrix degradation in primary macrophages. Blood, 116(9), 1559‐1569. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20505159. doi:10.1182/blood‐2009‐12‐257089

Wiesner, C., Le‐Cabec, V., El Azzouzi, K., Maridonneau‐Parini, I., & Linder, S. (2014).

Podosomes in space: macrophage migration and matrix degradation in 2D and 3D settings. Cell Adh Migr, 8(3), 179‐191. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24713854.

Willenborg, S., Eckes, B., Brinckmann, J., Krieg, T., Waisman, A., Hartmann, K., . . . Eming,

S. A. (2014). Genetic ablation of mast cells redefines the role of mast cells in skin wound healing and bleomycin‐induced fibrosis. J Invest Dermatol, 134(7), 2005‐2015. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24406680. doi:10.1038/jid.2014.12

Willenborg, S., Lucas, T., van Loo, G., Knipper, J. A., Krieg, T., Haase, I., . . . Eming, S. A.

(2012). CCR2 recruits an inflammatory macrophage subpopulation critical for angiogenesis in tissue repair. Blood, 120(3), 613‐625. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22577176. doi:10.1182/blood‐2012‐01‐403386

Williams, K. C., & Coppolino, M. G. (2011). Phosphorylation of membrane type 1‐matrix

metalloproteinase (MT1‐MMP) and its vesicle‐associated membrane protein 7 (VAMP7)‐dependent trafficking facilitate cell invasion and migration. J Biol Chem, 286(50), 43405‐43416. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22002060. doi:10.1074/jbc.M111.297069

Williams, K. C., McNeilly, R. E., & Coppolino, M. G. (2014). SNAP23, Syntaxin4, and vesicle‐

associated membrane protein 7 (VAMP7) mediate trafficking of membrane type










214 References

1‐matrix metalloproteinase (MT1‐MMP) during invadopodium formation and tumor cell invasion. Mol Biol Cell, 25(13), 2061‐2070. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24807903. doi:10.1091/mbc.E13‐10‐0582

Woo, C. H., Lim, J. H., & Kim, J. H. (2004). Lipopolysaccharide induces matrix

metalloproteinase‐9 expression via a mitochondrial reactive oxygen species‐p38 kinase‐activator protein‐1 pathway in Raw 264.7 cells. J Immunol, 173(11), 6973‐6980. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15557194.

Xue, M., Le, N. T., & Jackson, C. J. (2006). Targeting matrix metalloproteases to improve

cutaneous wound healing. Expert Opin Ther Targets, 10(1), 143‐155. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16441234. doi:10.1517/14728222.10.1.143

Yang, Y. H., Li, D. L., Bi, X. Y., Sun, L., Yu, X. J., Fang, H. L., . . . Zang, W. J. (2015). Acetylcholine

Inhibits LPS‐Induced MMP‐9 Production and Cell Migration via the alpha7 nAChR‐JAK2/STAT3 Pathway in RAW264.7 Cells. Cell Physiol Biochem, 36(5), 2025‐2038. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/26202362. doi:10.1159/000430170

Yoo, H. G., Shin, B. A., Park, J. S., Lee, K. H., Chay, K. O., Yang, S. Y., . . . Jung, Y. D. (2002). IL‐

1beta induces MMP‐9 via reactive oxygen species and NF‐kappaB in murine macrophage RAW 264.7 cells. Biochem Biophys Res Commun, 298(2), 251‐256. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12387824.

Young, P. K., & Grinnell, F. (1994). Metalloproteinase activation cascade after burn injury:

a longitudinal analysis of the human wound environment. J Invest Dermatol, 103(5), 660‐664. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7963652.

Yu, Q., & Stamenkovic, I. (2000). Cell surface‐localized matrix metalloproteinase‐9

proteolytically activates TGF‐beta and promotes tumor invasion and angiogenesis. Genes Dev, 14(2), 163‐176. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10652271.

Yu, X., Zech, T., McDonald, L., Gonzalez, E. G., Li, A., Macpherson, I., . . . Machesky, L. M.

(2012). N‐WASP coordinates the delivery and F‐actin‐mediated capture of MT1‐MMP at invasive pseudopods. J Cell Biol, 199(3), 527‐544. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23091069. doi:10.1083/jcb.201203025

Zerial, M., & McBride, H. (2001). Rab proteins as membrane organizers. Nat Rev Mol Cell

Biol, 2(2), 107‐117. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11252952. doi:10.1038/35052055

Zhao, P., Yang, L., Lopez, J. A., Fan, J., Burchfield, J. G., Bai, L., . . . James, D. E. (2009).

Variations in the requirement for v‐SNAREs in GLUT4 trafficking in adipocytes. J Cell Sci, 122(Pt 19), 3472‐3480. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19759285. doi:10.1242/jcs.047449











Documents

INTRACELLULAR TRAFFICKING AND SECRETION OF MATRIX ... · INTRACELLULAR TRAFFICKING AND SECRETION OF MATRIX METALLOPROTEINASES DURING MACROPHAGE MIGRATION. Joan Rohl . Bachelor of