The role of the metalloproteases ADAM8, 9, 10 and 17 in leukocyte migration in vitro

Von der Fakultät für Mathematik, Informatik und Naturwissenschaften der RWTH Aachen University zur Erlangung des akademischen Grades eines Doktors der Naturwissenschaften genehmigte Dissertation

vorgelegt von

Diplom-Humanbiologen

Franz Martin Heß

aus Darmstadt

Berichter: Universitätsprofessor Dr. rer. nat. Andreas Ludwig

Universitätsprofessor Dr. rer. nat. Martin Zenke

Tag der mündlichen Prüfung: 26.06.2012

Diese Dissertation ist auf den Internetseiten der Hochschulbibliothek online verfügbar.

Index

1. Introduction...... 1

1.1. Aim of this study...... 18

2. Materials and Methods...... 20

2.1. Materials, Chemicals and Equipment...... 20

2.2. culture methods...... 33

2.3. biochemistry methods ...... 35

2.4. Molecular biology methods...... 37

2.5. Functional assays...... 40

2.6. Statistics ...... 42

3. Results...... 43

3.1. Comparison of ADAM8, 9, 10 and 17 expression in different leukocytic, endothelial and epithelial cell culture lines...... 43

3.2. The effect of ADAM-inhibitors on leukocyte migration...... 44

3.3. Lentiviral shRNA-based knock-down of ADAM 8, 9, 10 and 17...... 48

3.4. Effect of ADAM10 and 17 silencing on shedding activity...... 55

3.5. Role of ADAM 10 and ADAM 8, 9, 17 for transmigration and chemotaxis of THP1 cells 58

3.6. Role of ADAMs for the sensitivity of LV-THP1 cells towards different MCP-1 concentrations ...... 70

3.7. The effect of ADAM8, 9, 10 and 17 knock-down on the Ca2+ response of THP1 cells to MCP-1 ...... 72

3.8. Influence of ADAM8, 9, 10 and 17 knock-down on adhesion of THP1 cells to VCAM-1 and fibronectin under static conditions ...... 76

3.9. Role of ADAM8, 9, 10 and 17 silencing on epithelial cell migration...... 79

3.10. Remarks...... 82

4. Discussion ...... 83

4.1. Conclusion...... 97

I

5. Summary...... 104

6. Literature ...... 106

7. List of abbreviations ...... 122

8. List of figures...... 128

9. List of tables...... 131

10. Appendix ...... 132

10.1. Plasmid maps ...... 132

10.2. Complementary figure: Comparison of ADAM8, 9, 10 and 17 expression in different leukocytic, endothelial and epithelial cell culture lines...... 136

10.3. Complementary figures: Relative number of migrated ADAM8, ADAM9 or ADAM17- KD LV-THP1 in transmigration and chemotaxis assay...... 137

11. Publications...... 139

12. Curriculum Vitae...... 140

13. Declaration ...... 141

14. Acknowledgements...... 142

II

1. Introduction

Inflammation is the organism’s defence response to invading pathogens such as bacteria, virus or parasites as well as to traumas caused by tissue damage, wound injury or tumours. Without inflammation the organism would be unable to fend off infections or induce tissue repair after a trauma. However, the coin has two faces: excessive inflammatory responses are responsible for a variety of disorders including septic shock, autoimmune diseases or chronic inflammatory diseases such as asthma, atherosclerosis, fibrosis, rheumatoid arthritis and even cancer. Given its importance and the consequences for the organism, inflammation is as tightly controlled as embryogenesis, metabolism, cell proliferation and other crucial complex biological processes.

Metaphorically speaking, the vasculature constitutes the front line in the war of inflammation. It provides a barrier for the circulatory system, which transports the troops – the leukocytes – to the battle being waged by the inflammatory response. One of the first responses in the inflammation process is the dilatation of the vasculature near the site of inflammation. This results in a reduction of the blood flow rate. The permeability of the endothelium in the microvessels increases and leukocytes are recruited to the site of inflammation. Later stages of inflammation include neovascularisation and the repair of the affected tissue. Different cell vascular cells take part in these processes: endothelial cells, smooth muscle cells, resident macrophages, platelets and the leukocytes of the blood flow.

The process of leukocyte recruitment itself is subdivided into several steps: the inflammatory activation of the tissue at the inflammation site, the rolling of the leukocytes on the activated endothelium, their tight adhesion and finally the transmigration or diapedesis of the leukocytes through the endothelium. Having overcome the endothelial barrier the leukocytes migrate, guided by the gradient of chemotactic stimuli, to the inflammation site, where they do their duty fighting in the cause of inflammation.

Since leukocyte recruitment to the site of inflammation is one of the critical hallmarks in inflammation, it is tightly controlled and finely tuned. that are involved in the modulation of these events belong to the families of , chemokines, growth factors and adhesion molecules. To coordinate this critical process most of the mediators of inflammation are tightly regulated on the transcriptional level. Several of the mentioned molecules undergo posttranslational modifications as well.

1

Many of the effector proteins of inflammation are initially expressed as transmembrane proteins on cells of the vasculature or leukocytes. During the process of limited or shedding, soluble variants of these molecules are released into the surrounding tissue. Several proteins of the A Disintegin And (ADAM) family are involved in the shedding of proteins that regulate inflammation responses and leukocyte recruitment1–4. The most prominent member of the ADAM family, ADAM17 actually received its common name TACE (tumour necrosis factor-α-converting ) because it activates one of the most important mediators of inflammation – the TNF-α – by proteolytic cleavage5.

Figure 1: The process of leukocyte recruitment. In the early inflammatory activation cytokines and chemokines are released by resident macrophages and smooth muscle cells of the affected tissue. Cytokines like TNFα and IL-1 lead to the activation of the adjacent endothelial cells which shed cell- molecules and express adhesion molecules on their surface. The endothelial permeability increases, chemokines diffuse easier into the vascular space and are presented on the activated endothelium. Leucocytes of the vasculture interact via their adhesion molecules with the adhesion molecules of the activated endothelial cells. This initiates the leukocytes’ process of slow rolling on the endothelium. The interaction of leukocytic with endothelial adhesion molecules VCAM-1, ICAM-1 and membrane bound chemokines leads to the firm adhesion of the leukocytes. The increased permeability of the endothelium facilitates the subsequent transendothelial migration or diapedesis of the leukocytes.

Cytokines

The first and most crucial step in almost all inflammatory responses is the release of cytokines by cells in the affected tissue. Cytokines are small signalling proteins which are 2

responsible for intercellular communication and the coordination of the inflammatory responses. The released cytokines bind specific receptors on neighbouring cells (paracrine signalling) or even on the cells which initially produced them (autocrine signalling). Often the released cytokines have systemic effects on the whole organism as well (endocrine signalling). The activated cytokine receptors initiate intracellular signalling cascades which result in changes in the function of the affected cells.

Cytokines are released by all nucleated cells of the body. Among the most prominent cytokines are TNFα, TGFβ (transforming growthfactor-β), the Interferons and numerous members of the Interleukin family. Resident macrophages are the main producers of cytokines in the early stage of inflammation. Endothelial and epithelial cells together with smooth muscle cells are other key contributors in the inflammatory cytokine release.

Early on in the inflammation process, Interleukin-1 (IL-1) and TNFα lead to the activation of the endothelium. In response, the activated endothelial cells express adhesion molecules such as P- and E-, VCAM-1 and ICAM-1. The released TNFα leads to increased endothelial permeability, which allows for the perfusion into the vasculature of chemokines produced by the inflamed tissue cells.

Chemokines

Chemokines are a subclass of cytokines that mediate the directed chemotaxis of leukocytes and other responsive cells, hence their name, which is an abbreviation for chemotactic cytokines. Chemokines have a distinct tertiary protein structure which is generally formed by two disulfide bonds between four cysteins in their protein sequence. At their N-terminus, the spacing between the two neighbouring cysteins which form the two disulfide bonds is used for their scientific nomenclature. In the case of CC-chemokines there is no residue between those cysteins, while in CXC-chemokines there is one additional amino acid.

In CX3CL1 – the only member of the CX3C-familiy – there are three amino acid residues between the two cysteins.

Chemokines are recognized by their corresponding chemokine receptors on cells which are recruited to the site of chemokine release6. The attracted cells migrate along the concentration gradient of the released chemokines7. Chemokine receptors possess seven transmembrane spanning regions and signal via coupled G-proteins. After binding their corresponding chemokines the chemokine receptors activate their G-proteins. This leads to the release of the second messengers diacylglycerol (DAG) and inositol triphosphate (IP3).

3

IP3 initiates the release of calcium from the intracellular stores of the endoplasmatic reticulum. The released calcium is part of the triggered signal cascade that finally leads to events such as chemotaxis and the activation of adhesion molecules such as the integrins expressed on leukocytes.

A well studied example of chemokine induced recruitment of leukocytes is the recruitment of neutrophil granulocytes induced by IL-8 (CXCL8)8. CCL2 or monocyte chemotactic protein-1 (MCP-1) is another chemokine whose function is well known9. MCP-1 binds the CCR2 on monocytes which are then guided to the inflammation site10. The monocytes which migrate out of bloodstream can differentiate into resident macrophages. In the inflamed tissue, it is their task to take on bacterial pathogens via phagocytosis and therefore they are crucial for the first line of defence. They are also capable of releasing important cytokines and to present foreign antigens. They thereby initiate and modulate later stages of immune response. One prominent example for the role of the MCP-1-CCR2-axis in disease is the recruitment of monocytes in the formation of plaques in arthrosclerosis11.

There are only two chemokines known which are expressed as transmembrane protein: CX3CL1 and CXCL16. Both chemokines are shedding substrates for members of the ADAM family. After the shedding process their chemokine domain is released from the endothelial cell surface where both are predominantly expressed. Soluble CX3CL1 is a potent chemoattractant for T-cells and monocytes whereas the membrane-bound form induces strong leukocyte adhesion12,13.

Adhesion molecules

Adhesion molecules are the other major class of molecules (besides the above-described cytokines and chemokines) that is needed for leukocyte recruitment.

Shortly after the activation by cytokines, the endothelial cells express P-selectin and, with a short delay, E-selectin. Leukocytes bind the endothelial with surface molecules, such as PSGL-1 (P-selectin ligand-1), carrying a Sialyl-LewisX sugar motif that is recognized by the lectin domain of selectins. This interaction results in the so-called “slow rolling” of the leukocytes on the endothelium which slows their movement in the bloodflow.

Endothelial proteins of the immunoglobulin superfamily such as ICAM-1 (intercellular adhesion molecule 1) and VCAM-1 (vascular 1) are bound by integrins expressed on the rolling leukocytes. This leads to the next step in leukocyte recruitment, the

4

tight adhesion to the activated endothelium. Integrins are heterodimers which are composed of specific pairs of alpha (α) and beta (β) subunits.

In addition to the adhesion molecules facilitating the adhesion of leukocytes to the endothelium, the adhesion molecules connecting the endothelial cells to each other are of great importance for leukocyte recruitment. Endothelial and epithelial cells are connected by tight junctions which are impermeable to fluid and inhibit the free diffusion of molecules. Proteins that form homodimers between neighbouring endothelial cells like VE-Cadherin and JAM-A are essential for the integrity of the endothelial adherens and tight junctions, respectively. Both mentioned molecules are shedding substrates for members of the ADAM- family and undergo proteolytic cleavage on the surface of inflamed endothelial cells14–16. This leads to increased endothelium permeability and the easier diffusion of small molecules through the endothelium layer.

Heperansulfat- (HPSG) are not considered to be classical adhesion molecules, as are the selectins, the immunoglobins or the integrins. However this class of proteins possesses a modulating function in inflammation and influences leukocyte recruitment17,18. The of the syndecans is the most prominent class of HPSGs. There are four known syndecans in humans; these are predominantly expressed on endothelial and epithelial cells. Syndecans bind and present secreted cytokines, chemokines and growth factors on the surface of endothelial cells. Thereby, they act as co-receptors for the receptors of the mentioned molecules19,20. The surface expression of syndecans is regulated by limited proteolysis. For instance, stimulation of lung epithelial cells by TNF-α and INF-γ leads to an increased shedding of Syndecan-1 and -4 by ADAM1721.

Leukocyte diapedesis

The last step in the process of leukocyte recruitment to the site of inflammation is the leukocyte transmigration or diapedesis through the activated endothelium. Once they have left the blood vessel the leukocytes are guided by the chemotactic gradient of chemokines such as IL8 (neutrophiles) or MCP-1 (monocytes) to the inflammation site. There they take on and destroy pathogens by phagocytosis, by producing reactive oxygen species and by releasing effector proteins. The step of endothelial transmigration constitutes the point of no return for the recruited cells since all previous steps in the recruitment process are reversible: the cells cannot leave the inflamed tissue without further differentiation into another celltype22.

5

Although most leukocytes take the paracellular route between two adjacent endothelial cells, transcelluar migration has been observed as well22,23. To migrate through the endothelium the leukocytes undergo drastic changes in their shape. The signalling of the chemokine receptors (see above) lead to the activation of several signalling kinases, such as members of the src-kinase family. This results in the activation of small GTPases of the Rho family and subsequently to actin polymerization and leukocyte polarization. The migrating cells form protrusions such as podosomes or lamellipodia in which the actin filaments accumulate. With these protrusions at their leading edge the cells squeeze themselves through the junctions between the endothelial cells24.

Leukocytic integrins serve as anchors for the filamentous actin on the plasma membrane. To perform the process of diapedesis, the leukocytic integrins interact with clustered ICAM-1 and VCAM-1 on the surface of the endothelial cells25,26. While ICAM-1 is bound by LFA-1

(lymphocyte function-associated antigen-1, αLβ2 ) VCAM-1 is recognized by VLA-4

(very late antigen-4, α4β1 integrin). Additionally, junctional adhesion molecules (JAM) are needed for the interaction of leukocytic adhesion molecules in diapedesis, as was demonstrated both in vitro and in vivo14,27. Leukocytic integrins, during further migration within the tissue to the inflammation site, interact with components of the ECM such as collagen and fibronectin.

Cell migration is a ubiquitous process necessary for many biological events beyond leukocyte recruitment. Other crucial events include embryonic development and endothelial or epithelial cell migration in the process of wound healing. Tumour cell migration and metastasis are examples of the pathologic consequences of cell migration.

The function of from the ADAM family has been extensively studied in the last two decades and their crucial role in inflammation is evident1–3. However there is little known about the direct role of leukocytic-expressed ADAMs in the process of transendothelial migration.

Classification of Metalloproteinases

Limited regulated proteolysis, also known as shedding, is an important posttranslational modification in higher organisms. of the metalloprotease superfamily are the most important . There are about 30 different families of metalloprotease which are defined by their structural characteristics, the sequence homologies of the proteins and similarities in their three-dimensional structure28. Metalloproteases which share the minimal

6

consensus sequence HEXXH in their catalytic domain are members of the zincins. In the consensus sequence, the two zinc-binding histidines (H) flank the catalytic glutamate residue (E). The group of zincins is divided into three superfamilies: the gluzincins, the aspzincins and the metzincins (Figure 2)29.

The metzincins are mostly multidomain proteins which possess globular catalytic domains with a size of 130-260 amino acid-residues30. These domains contain a HEXXHXXXXH…M zinc-binding consensus motif where three histidines bind the zinc ion and a conserved methionine-turn at the active site31. There are several families in the metzincins – these include the , pappalysins, , matrix metalloproteinases and /repro-lysins32.

Figure 2: Family tree of the metzincins. The drawing shows the three dimensional structure of the catalytic site of metzincins with the common HEXXHXXXXH…M zinc-binding consensus motif and the methionine-turn. Figure taken from32

The family of the adamalysins can be further divided into the sub-families of the A Disintegin And Metalloproteases (ADAMs) and the Snake Venom MetalloProteaes (SVMPs). While proteins of the ADAM-family are transmembrane proteins, the members of the SVMP family are soluble proteins. ADAMs are unique among other metalloproteases because they possess a potential adhesion domain (the domain) as well as a potential domain.

7

The members of the ADAM-TS-family (ADAM with thrombospondin domains) lack the transmembrane domain of ADAMs and function as secreted proteins33. The thrombospondin domain enables the ADAM-TS to bind to components of the ECM34.

The ADAM family of metalloproteinases and their structure

Twenty years ago ADAM1 and 2, the first members of the ADAM family, and their role in the fertilization process were identified35,36. Since then more than 40 ADAM have been discovered in various species37. The contains 23 known ADAM genes33. However, only seventeen of the 29 known mammalian ADAMs are active : most probably because many members lost critical catalytic residues over the course of evolution38.

ADAMs are composed of an N-terminal pro-domain, a catalytic domain, a disintegrin domain, a cystein-rich region, an EGF-like (epidermal growth factor) region followed by a single-span transmembrane region and a cytoplasmic tail (Figure 3). The other ADAM domains besides the catalytic domains can critically influence the proteolytic function.

Figure 3: Drawing showing the distinct domain of ADAM protease. The prodomain (PRO), metalloproteinase domain (MP), disintegin domain (DIS), cystein-rich region (CR), epidermal growth factor like repeat (EGF), transmembrane region (TM) and the cytoplasmic domain (CD). Modified from32

8

The prodomain of ADAM proteins facilitates correct protein folding during biosynthesis by holding the zinc ion and the protease domain in a latent state33. It also suppresses the activity of the proteases during maturation and intracellular transport. The prodomain of most ADMAs is cleaved from the catalytic-domain by furin proteases during the maturation of ADAMs in the secretory pathway39. It seems, however, that in cases of some ADAM proteases it remains bound within the catalytic cleft where it could fulfil regulatory functions40,41.

It has been shown that the cystein rich domain of some ADAMs can interact with heparin sulfate proteoglycans such as syndecan and with ECM proteins such as fibronetin42–44. It is thought hat this adhesive interaction is important for the substrate specificity of the protease45–47. However, the ADAM domain that mediates most of the adhesive functions of ADAMs is the disintegrin domain38.

The term disintregrin originates from the closely related SVMPs (snake venom metalloproteases, see above) where the disintegrin domain binds to the platelet integrin receptors. This inhibits platelet integrins from interacting with their natural ligands such as fibrinogen, which eventually causes the inhibition of platelet aggregation. This, together with the cleavage of basement membrane proteins by the proteoltic domain of the SVMPs, leads to the typical haemorrhaging caused by those snake venoms containing SVMPs48,49.

In ADAMs the disintegrin domain is about 90 amino acids long37. With the exception of ADAM15, the disintegrin-like domains of ADAM-proteases do not contain the classical RGD (Arg-Lys-Asp)-integrin binding motif of SVMP50. For several ADAMs it has been reported that they can bind to integrins through a conserved RX6DLPEF motif in their disintegrin domain51,52. However, ADAM10 and 17 do not posses this motif and the structural basis of their interaction with integrins remains to be further investigated. The interaction of ADAMs with integrins shows some redundancies since a single integrin can bind to several members of the ADAM family while one ADAM can interact with multiple integrins53. It has been shown that the inaction of ADAMs with integrins via their disintegrin domain can in some cases block54 and in other cases promote55,56 cell migration independent of their proteolytic activity.

The cytoplasmic domains of ADAM proteases vary in length and sequence. It is speculated that this domain is involved in the inside-out regulation of their proteolytic activity as well as in outside-in signalling. Many ADAMs contain a for proteins containing a Src homology 3 (SH3)-domain37. ADAM8, 9, 10 and 17, whose role in the migration of leukocytes were analysed in this study, contain such SH3-binding sites. Furthermore, the cytoplasmic tails of several ADAMs contain potential phosphorylation sites. It is speculated that

9

phosphorylation of residues from the cytoplasmic tail affects ADAM proteolytic activity and intracellular trafficking57. The phosphorylated cytoplasmic tail of ADAMs could also act as adaptor in the assembly of protein complexes at the critical site of functional activity48.

The most prominent function of ADAMs is their shedding activity which is performed by their catalytic domain. The proteolytic function is best studied for ADAM10 and ADAM17. As outlined below these proteases shed numerous substrates on the cell surface and seem to be vital for various events during development, immune regulation and regeneration. In contrast, significantly less substrates exist for ADAM8 and ADAM9 and only little is known about the role of these proteases under physiological and pathophysiological conditions.

A

B C

Figure 4: The implication of ADAM-shedding on various substrate molecules. A: Shedding of membrane-bound proforms of cytokines or growth factors which results in the release of their soluble form. The mature factors are then able to act either as agonist or antagonist of their respective receptors. B: Shedding of adhesion molecules. ADAMs are known to shed adhesion molecules from endothelial and leukocytic cells and thereby facilitate cell detachment. C: Regulated intramembrane proteolysis: ADAMs can cleave the extracellular domain of membrane proteins which are then further processed by the γ-secretase complex. In case of Notch, cleavage by ADAM10 enables the γ-secretase to further process the truncated molecule within the . This results in the release of the intracellular fragment that then acts as transcription factor.

10

Shedding refers to the ectodomain release of a transmembrane protein by limited proteolysis. Several cytokines and growth factors are initially expressed as membrane-bound proforms. Shedding of these proforms results in their release as soluble factors which enables the molecules to function as auto- or parakrine stimulus (Figure 4A). Prominent examples of such factors are TNFα, the chemokine CX3CL1 and all members of the EGF-like family. Several adhesion molecules are also subject to shedding (Figure 4B). Their proteolytic cleavage results in the loosening of intercellular adhesion. For example, the shedding of the endothelial adhesion molecules JAM-A and VE-Cadherin results in increased vascular permeability due to the loosening of the intercellular tight and adhesion junctions. Regulated intramembrane proteolysis by the γ-secretase-complex – which results in the release of an intracellular fragment – generally requires the preceding shedding of the extracellular domain of the cleaved protein (Figure 4C). Notch and the amyloid precursor protein (APP) are well-studied proteins that undergo a so-called α-cleavage by ADAM proteases before being further processed by the γ-secretase-complex.

ADAM17

ADAM17 was the first member of ADAM family of metalloproteases which was identified as a sheddase58. Due to its proteolytic processing of TNFα and its release as a soluble protein, it was initially named TACE (TNF-alpha Converting Enzyme)5. As mentioned above, TNFα is one of the key cytokines in inflammatory responses. In the following years it was shown that multiple proteins involved in inflammation, neurodegenerative diseases, such as Alzheimer’s disease and cancer, are shedding substrates of ADAM1759. Until now 76 shedding substrates of ADAM7 have been identified1.

The importance of ADAM17 (particularly in development) is underlined by the fact that gene- targeted mice with the functional disruption of the ADAM17 metalloproteinase domain die between embryonic day 17.5 or on the day after birth. Foetuses from these mice show abnormalities such as open eyelids and a lack of a conjunctival sac. These mice display the same phenotype as mice deficient for transforming growth factor alpha (TGFα) 60. TGFα is synthesized as a transmembrane precursor and has to be released from the cell surface by ADAM17 cleavage to be functional.

The critical role of ADAM17 in inflammation was recently shown by studies of transgenic mice with tissue-specific deletion of ADAM17. Conditional ADAM17-/- mice whose leukocytes lack ADAM17 are protected from endotoxin-induced septic shock61, have better survival rates

11

than wild-type control mice in e. coli-mediated periotinitis62 and show reduced inflammation symptoms in LPS-induced pulmonary inflammation63. In all three studies ADAM17-/- mice showed markedly decreased TNFα levels after pathogen stimulation. The reduced TNFα levels resulted in lower levels of the early responding cytokines IL1β and IL-662,63. Neutrophil accumulation in the inflamed peritoneal cavity and lung was significantly reduced even though it occurred significant earlier in the ADAM17-/-. This was attributed to the higher L- selectin expression on the ADAM17-/--leukocytes since ADAM17 is the major of the leukocytic adhesion molecule62–65.

A recent study by our group showed that the deletion of endothelial ADAM17 in mice or its pharmaceutical inhibition reduced vascular permeability, edema formation and reduced leukocyte recruitment in LPS-induced lung inflammation. Applying TNFα did not restore leukocyte recruitment and edema formation. This indicates that the shedding of further endothelial molecules by ADAM17 is necessary for the establishment of pulmonary inflammation66.

Other examples of inflammatory cyto- and chemokines that are shed by ADAM17 are CX3CL167,68, FLT3 ligand69 and HB-EGF70. However, cytokine receptors are also subject to shedding by ADAM17. Prominent examples include the TNFα receptor p55 (TNF-R-I)71 and p75 (TNF-R-II)60, the IL6-receptor72 and the IL1-receptor-II71. ADAM17 sheds as well adhesion molecules which play crucial roles in leukocyte recruitment and inflammation such as the above-mentioned L-selectin60, JAM-A14, ICAM-173, VCAM-174, CD4475 and syndecan-1 and -421.

The regulation of ADAM17 activity is of critical importance to the inflammation process1,3. Its shedding activity can be enhanced by stimulation with the phorbolester PMA67,68,74 and as well with TNFα and IFNγ14. It has been shown that isoforms of Phosphokinase C, as well as MAP/ERK kinases, Src kinase and p38, are all involved in the regulation of ADAM17 activity76– 79. TIMP-3 (tissue inhibitor of metalloproteases-3) is an endogenous inhibitor that regulates ADAM17 activity80.

ADAM10

ADAM10 is another, better characterized member of the ADAM-family of metalloproteases. It is known for its prominent roles in neurodevelopment81,82 and neuroprotection 2,83 and inflammation3. ADAM10 and ADAM17 share several shedding substrates. For instance, they both function as so-called α-secretases by cleaving the amyloid precoursor protein (APP).

12

The α-cleavage of APP by ADAMs protect it from being processed by β-secretases and from subsequent processing by the γ-secretase, which results in the release of an intracellular fragment and the formation of amyloid plaques. Therefore, the α-secretases activity of ADAMs prevents amyloid plaque formation in the development of Alzheimer’s disease82,84. Other common substrates of ADAM10 and 17 are the transmembrane chemokine CX3CL113,68, the receptor of IL-685 and TNFα86, among many other substrates.

Mice carrying a targeted disruption of the ADAM10 gene have a severe phenotype with development defects in the central nervous and cardiovascular system which causes their death at embryonic day 9.587. The ADAM10-deficient mice display a similar phenotype just as mice deficient for Notch87,88. The Notch-signalling pathway is highly evolutionary conserved as it important in the fruit-fly Drosophilia, the nematode c. elegans and mammals. It is essential for the establishment of intercellular contacts and regulates cell fate at all developmental stages of the nervous system. It is thought that the ADAM10 knockout in mice has similar effects like the deletion of the ADAM10-homolog Kuzbian in Drosophila89. In Kuzbian deficient fruit-flies the transmembrane protein Notch cannot be shed and signalling via its ligands Delta and Serrate is aborted, which leads to severe developmental abnormalities90.

Other molecules which have crucial roles in neurodevelopment and which are subject to ADAM10-shedding are members of the ephrin-ligand family91,92. Ephrin guides migrating which bind ephrins through their Eph receptors. The receptor-ligand interaction induces bidirectional signalling between the cells and a cell-cell repellent reaction. The study describing the shedding of Ephrin-A5 by ADAM10 also indicated that trans-shedding-events by ADAMs can occur92.

The important role of ADAM10 in cancer has been summarised in several reviews 2,37,83,93,94. ADAM10 releases the epidermal growth factor receptor (ErbB) HER2/ErbB2 from the cell membrane of tumour cells. This results in the ligand independent signalling of the receptor which leads to the proliferation of the cells95. Together with ADAM17, ADAM10 is the major sheddase for many ErbB ligands96 such as EGF (epidermal growth factor) and betacullin97–99. The shedding of adhesion molecules such -CAM100, CD44101, C4.4A102, N- and E- Cadherin103,104 by ADAM10 contributes to tumour cell migration and metastasis.

In inflammatory events, the ADAM10 mediated shedding of CD44, IL-6-receptor and TNFα contributes to inflammatory responses. Compared to ADAM17, ADAM10 seems to be less important for TNFα shedding and therefore is not made responsible for the generation of proinflammatory TNFα. ADAM10 also sheds the transmembrane chemokines CX3CL1105 and

13

CXCL16106. Both transmembrane molecules mediate leukocyte adhesion by tightly binding their chemokine domain to the chemokine receptors CX3CR1 and CXCR6, respectively. Shedding of the transmembrane chemokines resolves this interaction and leads to the release of bound leukocytes12,107. ADAM10 shedding of the endothelial adhesion molecule VE- Cadherin contributes to increased vascular permeability and leukocyte migration15.

Figure 5: Involvement of ADAMs in inflammatory activation and leukocyte recruitment. This figure complements Figure 1 by showing the role of different of shedding events of ADAM proteases in leukocyte recruitment. The most prominent shedding substrates of ADAM8, 10 and 17 which are involved in activation of the inflamed tissue and leukocyte rolling, adhesion, transmigration and chemotaxis are shown. The influence of the shedding of these substrates is discussed in detail in the text. Modified from3.

Fas ligand (FasL) is an important regulatory lymphoid surface protein which is shed by ADAM10108. Binding of FasL, which is expressed by activated T Cells and natural killer (NK) cells, to its receptor Fas on adjacent cells triggers the of the cells expressing the . It is thought that cleveage of FasL is an important regulator of apoptosis109. The low affinity IgE receptor CD23 – which is expressed on B-cells and is involved in allergic

14

responses by its regulation of IgE production110 – is shed by ADAM10 as well111. Transgenic mice with a B-cell specific ADAM10 knockout show defects in B-cell development112 and in their response to T-cell dependent antigen immunization113.

Like other ADAMs, ADAM10 is expressed as an inactive protease since its prodomain holds the catalytic domain in an inactive state. It is activated by proteolytic cleavage of the furin protease during its maturation in the secretory pathway. ADAM10 proteolytic activity can be suppressed by the addition of its own soluble prodomain114. ADAM10 activity is potently blocked by the endogenous inhibitory proteins TIMP1 and to a lesser degree by TIMP3115. In cell culture, ADAM10 activity can be induced by stimulation with the calcium-ionophore ionomycin12. It has also been shown that ADAM10 shedding activity is affected by changes in membrane fluidity116.

ADAM8

ADAM8 and its function is less studied and understood than the more prominent members of the ADAM family ADAM10 and ADAM1789. However, it has been shown that ADAM8 is capable of cleaving such ADAM10 and ADAM17 substrates as APP117, TNFα-RI118, L-seletin119 and VCAM-1120. A screen with containing cleavage sites of substrates from other ADAM proteases revealed that CD16, CD23, TGFα and TNFα are potential ADAM8 substrates117. Experiments with ADAM8-deficient mice revealed that ADAM8 is a non- essential protein in healthy since ADAM8-deficient mice show no pathological abnormalities121. One explanation of this finding may be the redundancy of substrates that ADAM8 shares with other ADAM proteases.

Similar to most metalloproteases of the metzincin family, ADAM8 has been associated with processes of cancer development and progression. Many tumours show high expression and protein level of ADAM8 which is a indicator of increased tumour invasiveness and a reduced survival rate for the tumour patients. This has been demonstrated for pancreatic duct carcinoma122, primary brain tumours123, primary lung cancer124, prostate cancer125 and renal cell carcinoma126.

It has been shown that ADAM8 is involved in the inflammatory process of rheumatoid arthritis. In human patients with this disease an increased number of ADAM8-expressing cells was found at the edge between the eroded cartilage and bone127. These findings were corroborated by a study showing that mice expressing an catalytic inactive variant of ADAM8 did not develop such severe arthritis as wild-type control mice128.

15

ADAM8 has been associated with the establishment of asthma in ovalbumin-induced murine asthma models. Its expression was upregulated in airway epithelia, airway smooth muscle cells and leukocytic cells that infiltrated the lung parenchyma after challenging the animals with the allergen129,130. ADAM8-/--mice as well as chimeras of wt mice which had been transplanted with bone marrow from ADAM8-/--mice were found to be less responsive to allergen exposure in experimental asthma. In mice with ADAM8-/- leukocytes, less infiltrating cells were found in the lung tissue and the bronchio-alveolar lavage. This indicates that ADAM8 influences leukocyte recruitment and migration in experimental asthma131.

A study of the same group found that ADAM8 influences T-cell maturation in the thymic medulla which coincided with changes in general morphology of the thymus132. A recent study showed that ADAM8 influences the rolling of leukocytes via the shedding of P-selectin glycoprotein ligand-1 (PSGL-1) and thereby affects leukocyte migration133.

ADAM8 does not need a second protease for the removal of its inhibitory prodomain like ADAM10 since this happens by autoproteolysis134. ADAM8 is not inhibited by any of the known TIMP members135.

ADAM9

ADAM9 or meltrin-γ is a member of the ADAM-family whose function still remains unclear. Knock-out mice lacking ADAM9 expression show no deficits in development and adult life and are fertile and viable136. It was found though that ADAM9-/- mice develop retinal degeneration later in life137.

ADAM9 may act as α-secretase of APP138 but does not seem to be as relevant as ADAM10 and ADAM17. ADAM9 has also been shown to be a sheddase of HB-EGF (heparin-binding EGF-like growth factor)139. The shedding of both APP and HB-EGF was not affected in ADAM9-/- animals, thus indicating that the shedding activity of other ADAMs may compensate for ADAM9 loss. Interestingly, ADAM10 is a shedding substrate of ADAM9, as was shown by two working groups140,141. Other shedding substrates of ADAM9 include cKit-ligand142, the Notch ligand Delta-like-1143, VCAM-1144 and VE-Cadherin144.

It has been shown for ADAM9 that it can function as an adhesion molecule by binding to

56 52 145 α6β1 , α9β1 , αvβ5 integrins via its disintegrin domain. Studies of ADAM9 expression in rat kidneys showed that is located on the basolateral surface of tubular cells. It mediates the adhesion to the basal membrane via the binding of the disintegrin domain to β1-integrin and is essential for the maintenance of cell morphology146,147. 16

ADAM9 overexpression is a poor marker for the prognosis in prostate cancer148 and renal cancer149. It is overexpressed in breast150, pancreas151, skin152, liver153 and lung154 carcinoma as well.

ADAM9-/- mice show accelerated wound healing compared to wild-type control mice. The faster closure of skin wounds occurred without changes in the numbers of infiltrating leukocytes in the wound. The authors of this study attributed the faster recovery to increased keratinocytes migration, since re-epithelialization was significantly faster in ADAM9-/- than in wild-type animals155. To date no endogenous inhibitory proteins for ADAM9 have been reported135.

Table 1: Shedding substrates of ADAM8, 9, 10 and 17 which are involved in inflammatory responses and leukocyte migration. The substrates and their shedding by the ADAM proteases are discussed in detail in the text.

cytokines cytokine or leukocytic endothelial cleavage of or growth growth adhesion adhesion regulators of factors factor molecules molecules leukocyte receptors development

ADAM8 cKit-ligand TNF-RI L-selectin VCAM-1 TGFα PSGL-1 TNFα

ADAM9 cKit-ligand VCAM-1 Delta-like-1 HB-EGF VE-Cadherin

ADAM10 beta-cellulin IL6-R CD44 VE-Cadherin CD23 CX3CL1 VEGF-RII FasL CXCL16 Notch EGF TNFα

ADAM17 CX3CL1 IL1-RII CD44 JAM-A FLT3 ligand IL6-R L-selectin ICAM-1 HB-EGF TNF-RI VCAM-1 TNFα TNF-RII Syndecan1 TGFα Syndecan4

17

1.1. Aim of this study

The transendothelial migration or diapedesis of leukocytes is a very critical step in inflammatory responses to infections or trauma. Several studies were able to show that endothelial-expressed ADAM17 and ADAM10 are important regulators of inflammation and leukocyte recruitment14–16,156. In vivo studies revealed that mice lacking functional ADAM17 in leukocytes show strongly reduced recruitment of neutrophils to inflammatory sites62,63,157. Relatively recently, ADAM8 emerged as another member of the ADAM family to influence leukocyte recruitment. ADAM8-KO mice are protected from developing experimental asthma131 and it was shown that ADAM8 cleaves the leukocytic adhesion molecule PSGL-1 and thereby influences the leukocyte rolling133. Two studies indicate the potential role of leukocytic ADAM10 in transendothelial migration15,158. However, both studies focused on the role of endothelial ADAM10 in this process and did not investigate the role of leukocytic- expressed ADAM10 in detail. To date there is no published study directly comparing the influence of leukocytic expressed ADAM8, 9, 10 and 17 on transendothelial leukocyte migration in the same experimental setup.

The aim of this study was to analyse the role of leukocytic-expressed ADAM8, 9, 10 and 17 in the process of transendothelial migration and chemotaxis. Therefore, the mRNA expression levels of these metalloproteases were assessed in cell culture lines which are commonly used in cell culture migration assays. The effect generated by specific pharmacological inhibitors against ADAM10 and ADAM17 in the migration of monocytic cells was tested in an in vitro migration assay.

To analyse the effect of ADAM8, 9, 10 or 17 on the migratory abilities of leukocytes, a lentiviral-mediated shRNA system for transcriptional gene silencing was established. The gene silencing of each target protease in stably transduced monocytic cells was analysed and confirmed. The migratory behaviour of the ADAM8, 9, 10 or 17 deficient monocytic cell lines was examined in transendothelial migration and chemotaxis experiments using the chemokine MCP-1 as stimulant for directed cell migration.

The effect of the gene silencing of the targeted proteases on the dose-dependent responsiveness of the monocytic ADAM knock-down cell lines to MCP-1 was analysed. Further investigations looked at whether the lack of ADAM8, 9, 10 or 17 influenced the intracellular calcium-transients of the stable monocytic cell lines in response to MCP-1 stimulation. The influence of the gene silencing of the four proteases on monocytic cell

18

adhesion was assayed. This was performed by testing the adhesion of the cell lines to immobilized integrin ligands (VCAM-1, ICAM-1 and fibronectin) under static conditions.

Wound healing is another process aside from leukocyte recruitment that involves cell migration. To asses the question whether ADAMs also contribute to the movement of adherent cells, in vitro wound healing assays were performed with epithelial cells treated with sHRNA for knock down of ADAM8, 9, 10 or 17.

19

2. Materials and Methods

2.1. Materials, Chemicals and Equipment

Equipment

Cell culture bench Hera Safe KS 12 Thermo Fischer, Waltham, MA USA

Cell culture incubator C150 Binder, Tuttlingen, Ger

Centrifuge 5415R Eppendorf, Hamburg, Ger

Centrifuge 5810R Eppendorf, Hamburg, Ger

Centrifuge Avanti J-25 Beckman Coulter, Krefeld, Ger

FL-20-M Fluo-Link (UV-lamp) Bachofer, Reutlingen, Ger

Flowcytometer Canto II BD Biosciences, Heidelberg, Ger

Flowcytometer LSRFortessa BD Biosciences, Heidelberg, Ger

Fluorescence mikroscope DMIRB Leica, Wetzlar, Ger

FLUOstar OPTIMA - plate reader BMG-Labtech, Ortenberg, Ger

Gelelektrophoresechamber BIO RAD, Hercules, USA

Haemocytometer Brand, Wertheim, Ger

LAS-3000 (Western Blot) Fujifilm, Düsseldorf, Ger

LightCycler 480 Roche, Penzberg, Ger

Microscop DMIL Leica, Wetzlar, Ger

Minitron (shaker for bacteria) INFORS HT Bottmingen, Ch

MS2-Minishaker (Vortex) IKA® Werke GmbH & Co. KG, Staufen, Ger

NanoDrop-ND-1000 PEQLAB Biotechnologie GmbH, Erlangen, Ger

PCR-Cycler Biometra Analytik GmbH, Göttingen, Ger

Pipets Eppendorf, Hamburg, Ger

Polyacrylamidgelelectrophoresechamber BIO-RAD, Hercules, USA

Thermomixer comfort Eppendorf, Hamburg, Ger

Transblot® SD (Semi-dry-Blot) BIO-RAD, München, Ger

20

Ultracentrifuge L7-65 Beckman Coulter, Krefeld, Ger

Consumables

All cell culture consumables (cell culture dishes, cell culture flaks, multi-well plates) were purchased from Sarstedt (Nürnberg, Ger).

Cryovials Starlab, Merenschwand, Schweiz

Falcon-Tubes Greiner Bio-One GmbH, Frickenhausen, Ger

Filter pipet tips Tip One Starlab, Merenschwand, CH

Glas-pipets Brand GmbH & Co KG, Wertheim, Ger

Hybond-P (PVDF-Membran) Amersham, Freiburg, Ger

Nunc-Immuno™ Plate (ELISA) Nunc A/S, Roskilde, Denmark

Pipet tips (1000µL) Heinz Herenz Medizinalbedarf, Hamburg, Ger

Pipet tips (100/10µL) Sarstedt, Nürnberg, Ger

Polycarbonate ultra-centrifugation tubes Nalgene, Rochester, NY USA

Polystyrene tube, 5mL (FACS-tubes) BD biosciences, Heidelberg, Ger qPCR-plates (96 wells) BIO Plastics, Landgraaf, NL

Transwellfilter polycarbonat membran 8 µm Corning Inc. Costar, Schiphol-Rijk, NL

Vivaspin 500 (Proteinfilter) Satorius AG, Göttingen, Ger

Whatman paper (gel-blotting-paper) Carl Roth GmbH & Co KG, Karlsruhe, Ger

Chemicals

All chemicals used (if not mentioned below) were purchased from Merck (Darmstadt, Ger), Sigma-Aldrich (St. Louis, USA) and Roth (Karlsruhe, Ger)

21

BM Blue POD Substrate soluble (ELISA) Roche, Prenzberg, Ger

BSA PAA Laboratories GmbH, Linz, Austria

Collagen-G Biochrome, Berlin, Ger

Complete Protease Inhibitor Roche, Prenzberg, Ger

Fibronectin Biochrome, Berlin, Ger

Fluo-3-AM Biotium, Hayward, CA USA

GelRed (gel-electrophorese of DNA) Biotium, Hayward, CA USA

GI254023X GlaxoSmithKline, Stevenage, UK

GW280264X GlaxoSmithKline, Stevenage, UK jetPEI Poly Plus Transfections, Illkirch, F

Lipofectamine Invitrogen, Karlsruhe, Ger

Loading Dye for gel-electrophorese (DNA)

6x Orange Fermentas GmbH, St. Leon-Rot, Ger

Pluronic-F127 Biotium, Hayward, CA USA

Kits

Unless specifically noted all kits were used according to the manuals provided by the manufactures.

Detection System

ECL Plus/Advanced Western Blotting Amersham, Little Chalfont, UK

LightCycler480 DNA SYBR Green I Master Roche, Penzberg, Ger

Lumigen TMA-6 (Western Blot) GE Healthcare, München, Ger

NucleoBond Finalizer (DNA-precipitation) Macherey-Nagel, Düren, Ger

NucleoBond PC 500 (maxiprep) Macherey-Nagel, Düren, Ger

NucleoSpin Extract II (DNA-gelextraktion) Macherey-Nagel, Düren, Ger

NucleoSpin Plasmid (miniprep) Macherey-Nagel, Düren, Ger

22

Protein-Assay (Bradford) BIO-RAD, München, Ger

RNeasy Mini Kit (RNA-Isolation) Qiagen, Hilden, Ger

RevertAid™ H Minus First

Strand (cDNA-synthesis) Fermentas GmbH, St. Leon-Rot, Ger

Z-CompetentTM E. Coli Transformation Buffer Set Zymo Research, Irvine, CA USA

Buffer and Solutions

Annealing-buffer for shRNA-Oligos (x2) 200nm potassium-acetate

60nM HEPES

4mM magnesium-acetate

pH 7.4

Ca2+-assay buffer HBSS (PAA)

10mM HEPES

720mg/l Probenicid

Ca2+-staining buffer Ca2+-assay buffer

2.65mM Fluo-3-AM

32nM Pluronic F-127

1% FCS

FACS-buffer Dulbecco’s PBS (PAA)

1% FCS

5mM EDTA

23

Glucoronidase substrate buffer 100mM sodium-acetate

10mM p-nitrophenyl-β-D-glucoronide

pH 4

Glucoronidase stopping buffer 400mM glycin

pH 10.3

LB-Medium 10g bakto-trypton

5g yeast extract

10g NaCl

ad 1l H2O bidest.

pH 7.0

LB-Agar 10g bakto-trypton

5g yeast-extract

10g NaCl

15g bakto-agar

ad 1l H2O bidest.

pH 7.0

Extractionbuffer for CX3CL1 PBS

1% Triton X-100

1x Complete Protease Inhibitor

24

Lysisbuffer for ADAM extraction 5 mM Tris

1 mM EGTA

250 mM Saccharose

1% Triton X-100

0.5% SDS

1x Complete Protease Inhibitor

PBS 150mM NaCl

120mM KCl

10nM Na2HPO4/KH2PO4

pH 7.4

PBST PBS

0.1% Tween 20

Running-buffer (SDS-PAGE) 25mM Tris/HCl

200mM Glycin

3.5mM SDS

pH 7.5

SDS-samplepuffer (SDS-PAGE) 20% Glycin

4.5% SDS

125mM Tris/HCl

a spatula tip of bromphenol blue

pH 6.8

for reducing conditions: 5% mercapto-

25

Separationgel-buffer (SDS-PAGE) 90.85g Tris/HCl

2g SDS

ad 500 mL H2O bidest.

pH 8.8

SOC-Medium 0.5g yeast extrakt

1g bacto-pepton

1g NaCl

20mM Glukose

ad 100 mL H2O bidest.

Stackinggel-buffer (SDS-PAGE) 30.28g Tris/HCl

2g SDS

ad 500 mL H2O bidest.

pH 6.8

Stripping-buffer (for PVDF-Membranen) 2% SDS

60.25mM Tris/HCl

0.83% β-Mercaptoethanol

in H2O bidest.

pH 6.8

TAE (x50) 242g Tris

57.1ml glacial acetic acid

1mM EDTA

ad 1l H2O bidest.

pH 7.4

26

Transferbuffer (western blot) 192mM Glycin

25mM Tris/HCl

10% Methanol

Cellculture-media

The following cell culture media, buffers and supplements were purchased from PAA Laboratories GmbH, Linz, Austria:

Dulbecco’s modified Eagle Medium (DMEM) High Glucose (4,5g/mL)

Medium 199 with Earl’s Salts (M199)

Roswell Park Memorial Institute (RPMI)

Dulbecco’s PBS without Ca and Mg (PBS)

Hanks Balanced Salt Solution containing Ca and Mg (HBSS)

Fetal Calf Serum (FCS)

Penicilin-Stryptomycin (PS)

Trypsin-EDTA (TE)

The basic endothelial cell medium and growth factor kit EGM-2MV single quots for the culture of the endothelial cell line HMVEC was purchased form Lonza, Walkersville, MD USA. Endothelial Cell Growth Medium including the supplements for the culture of the endothelial HUVEC cells was purchased from Promocell, Heidelberg, Ger.

Enzymes and recombinant proteins

All enzymes and reaction buffer used in the molecular cloning experiments were purchased from Fermentas GmbH, St. Leon-Rot, Ger. For the selection of the different buffers for the restriction enzymes the suggestion of the supplier were followed.

All recombinant cytokines, chemokines and the adhesion molecules VCAM-1 and I-CAM-1 were purchased from Peprotech, London, UK.

27

Streptavidin conjugated with peroxidase for the detection of biotin conjugated proteins in the ELISA was purchased from Roche, Prenzberg, Ger.

Antibodies

All primary used in this study and their working dilutions are listed in Table 2.

Table 2: List of primary antibodies and their working dilutions.

Name Company Isotype Specifity Dilution Dilution ELISA Blot FACS

ADAM8 R&D Mouse Human 1:500 1-5µg/ml - (MAB10311) IgG Systems 1 (AS 498 – 653)

ADAM9 R&D Mouse Human 1:500 - - (MAB939) IgG Systems 1 (Ectodomain)

ADAM10 R&D Mouse Human 1:500 1µg/ml - (MAB1427) IgG Systems 2B (N-terminal)

ADAM17 R&D Mouse Human - 4µg/ml - (MAB9301) IgG Systems 1 (N-terminal)

CX3CL1-Biotin R&D Mouse Human, mouse - - 300ng/ml (BAM365) IgG Systems 1

CX3CL1 R&D Mouse Human - - 4µg/ml (MAB3652) IgG Systems 1 (chemokine domain)

Isotypecontrol R&D Mouse Isotype control - 1µg/ml - A10 IgG Systems 2B IgG (MAB0042) 2B

Isotypecontrol R&D Mouse Isotype control - 4µg/lm - A17 (MAB002) IgG1 Systems IgG1

β-Actin Abcam Mouse Human, hamster, 1:500 - - IgG1 rat, dog, pig, rabbit, mouse

The secondary antibodies used in this study are listed in Table 3.

28

Table 3: List of secondary antibodies and their working dilutions.

Name Conjugate Company Dilution blot Dilution FACS

Goat anti-mouse PE Jackson Immuno Research - 1:100

Goat anti-mouse APC Jackson Immuno Research - 1:100

Goat anti-mouse HRP Jackson Immuno Research 1:25,000 -

shRNA sequences and oligonucleotides

All shRNA sequences that were cloned into the pLVTHM plasmid were designed after the scheme depicted in Figure 6. The 19 nucleotide long sense siRNA sequence and the reverse complementary anti-sense siRNA sequence were separated by the 7 nucleotide spanning loop. At the 5’-end of the sequence tandem was the transcription initiation site for the RNA PolIII, while the multiple thymidins at the 3’-end of the oligo encode the stop signal for the polymerase. After the transcription of the siRNA sequence tandem, both, the sense and the anti-sense, siRNA sequence anneal and form, together with the loop region, the small hairpin RNA (shRNA) which is then further processed to siRNA. The forward oligo (upper sequence Figure 6) and reverse oligo (lower sequence Figure 6) are complementary. Both oligos were designed to form overlapping “sticky” ends at the 5’ and 3’ end. This enabled the directed ligation of the annealed oligos into the opened pLVTHM plasmid which had been digested with the corresponding restriction enzymes MluI and ClaI.

Figure 6: General design of oligonucleotides containing the shRNA target sequences. The sequence above represents the forward strand and the sequence below the reverse stand of the oligonucleotide pair. The sense and the anti-sense siRNA are separated by the loop region. There are overhanging “sticky” ends at either end of the oligo-dimer. The annealed oligos were ligated into the opened pLVTHM plasmid which had been digested with the restriction enzymes MluI and ClaI. After http://lentiweb.com/cloning_strategies.php, design by Maciej Wiznerowicz.

The sequences of the oligonucleotides containing the shRNA sequences against the ADAM proteinases are listed in Table 4 (only the sequences of the forward oligos are shown). All oligonucleotides were ordered from Eurofins MWG Operon (Ebersberg, Ger). The oligos containing the shRNA sequences against ADAM8 and 9 have a slightly different design since

29

they were first inserted into the pSuper plasmid (opened with BglII and HindIII) and subsequently subcloned into the lentiviral pLVTHM vector using the restriction enzymes EcoRI and ClaI.

The pLVTHM-scramble plasmid was a generous gift of Dr Antonio Sechi of the working group of Prof Martin Zenke, Department for Cell Biology, University Hospital Aachen.

Table 4: List of oligonucleotides for the generation of pLVTHM plasmids for the down regulation of our target ADAM proteinases.

Name Oligonucleotide sequence (target sequence are shown in bold) scramble CGC GTC CCC CCG TCA CAT CAA TTG CCG TTT CAA GAG AAC GGC AAT TGA TGT GAC GGT TTT TGG AAA T shADAM8-903 GAT CCC CGC ATG ACA ACG TAC AGC TCT TCA AGA GAG AGC TGT ACG TTG TCA TGC TTT TTA shADAM8-1831 GAT CCC CAG AGA AGG TTT GCT GGA AAT TCA AGA GAT TTC CAG CAA ACC TTC TCT TTT TTA shADAM8-2645 GAT CCC CGC TGC TGT TCT AAC CTC AGT TCA AGA GAC TGA GGT TAG AAC AGC AGC TTT TTA shADAM9-1967 GAT CCC CGA AAC TTC CAG TGT GTA GAT TCA AGA GAT CTA CAC ACT GGA AGT TTC TTT TTA shADAM9-2260 GAT CCC CCC AGA GAA GTT CCT ATA TAT TCA AGA GAT ATA TAG GAA CTT CTC TGG TTT TTA shADAM10-650 CGC GTC CCC GAC ATT TCA ACC TAC GAA TTT CAA GAG AAT TCG TAG GTT GAA ATG TCT TTT TGG AAA T shADAM10-1947 CGC GTC CCC ACA GTG CAG TCC AAG TCA ATT CAA GAG ATT GAC TTG GAC TGC ACT GTT TTT TGG AAA T shADAM17-2061 CGC GTC CCC AGG AAA GCC CTG TAC AGT ATT CAA GAG ATA CTG TAC AGG GCT TTC CTT TTT TGG AAA T shADAM17-2642 CGC GTC CCC GAA ACA GAG TGC TAA TTT ATT CAA GAG ATA AAT TAG CAC TCT GTT TCT TTT TGG AAA T

The sequences of the oligonucleotides, which were used as primers in the quantitative PCR (qPCR), to determine the transcript level of the respective target mRNA, are listed in Table 5.

30

Table 5: Sequences of qPCR-primers for ADAM8, 9, 10, 17 and GAPDH used in this study.

Name Sequence Bp length Melting temperature hADAM8 qRT for GTGTGTGTACGTGTCTCCAGGT 22 57°C hADAM8 qRT rev CAGACAAGA TAGCTGACTCTCCC 23 57°C hADAM9 qRT for GGTGACAGATTTGGCAATTGTG 22 53°C hADAM9 qRT rev TTGTGCCTTCGTTAACCATCC 21 52°C hADAM10 qRT for GGATTGTGGCTCATTGGTGGGCA 23 59°C hADAM10 qRT rev ACTCTCTCGGGGCCGCTGAC 20 60°C hADAM17 qRT for GAAGTGCCAGGAGGCGATTA 20 54°C hADAM17 qRT rev CGGGCACTCACTGCTATTACC 21 56°C

GAPDH qRT for CGGGGCTCTCCAGAACATCATCC 23 61°C

GAPDH qRT rev CCAGCCCCAGCGTCAAAGGTG 21 60°C

Plasmids

The plasmids used in this study are listed in Table 6. Vector maps of the plasmids can be found in Supplements. The pSuper plasmid was originally generated and described by Brummelkamp et al.159. The lentiviral production system and the three lentiviral vectors listed in Table 6 were developed by the group of Prof Didier Trono, Ecole Polytechnique Fédérale de Lausanne160,161. The lentiviral plasmids were ordered from the non-profit plasmid-storage bank Addgene, Cambridge, MA, USA (www.addgene.org).

Table 6: List of plasmids used in this study.

Name Properties Addgene number pSuper shRNA expression plasmid pLVTHM Lentiviral shRNA expression plasmid 12247 encodes the reporter gene GFP pMD2 Lentiviral envelope plasmid encoding VSV-G 12259 psPAX Lentiviral packaging plasmid encoding the HIV gag-pol gene 12260 and further auxiliary HIV proteins

31

Bacteria strains

For the cloning and propagation of the plasmids the following E. coli-strain was used: DH5α - Genotype DH5α: supE44 ∆lacU 169(φ80lacZ ∆M15) hsd R17 recA1 end A1 gyr196 thi-1 relA1.

Cell lines

All cell lines which were used in this study are listed in Table 7 . In general the culture media contained 10% FCS and penicllin (100U/ml) and streptomycin (100µg/ml). The media for the endothelial cells HMVEC and HUVEC contained the supplements which were provided by the supplier. All cells were cultured at 37°C and 5% CO2 in a humid atmosphere.

Table 7: List of cell lines used in this study and their respective culture media.

Name Description Culture Medium

A549 human alveolar basal epithelial adenocarcinoma DMEM

ECV304 human bladder carcinoma M199

ECV-CX3CL1 ECV304 stably expressing CX3CL1 M199 with G418 (1:100)

HEK293 human embryonic kidney cell line 293 DMEM

HEK293T Variant of HEK293-cells, stably expressing the DMEM large T-antigen Collagen-coat

HMVEC human microvascular endothelial cells EGM-2MV single quots

HUVEC human umbilical vein endothelial cells Endothelial Cell Growth Medium Collagen-coat

Jurkat human T cell leukemia cells RPMI 1640

THP1 human acute monocytic leukemia cells RPMI 1640

32

2.2. Cell culture methods

Propagation of cells

All cells were kept in culture dishes (10cm diameter). When the adherent cell lines reached full confluency, they were washed with PBS and trypsin/EDTA was added to the cells. The cells were incubated at 37°C until they had completely detached from the dish. Then medium containing FCS was added to the cells and they were singled out by up pipetting. Subsequently, the cells were seeded out at the desired dilution into dishes containing fresh medium.

The suspension cells (THP1 and Jurkat) were kept at densities of 2-10 x105cells/ml. After generating a homogenous cell suspension by pipetting, the cells were split at ratios ranging from 2-5 into dishes containing fresh medium. Every second split, their cell density was determined by counting the cells with a haemocytometer (see below). The cells were then centrifuged and resuspended in 10ml fresh medium seeded out at a density of 2x105cells/ml.

Freezing and thawing of cells

To preserve the cell lines for a longer period, aliquots of each cell lines were stored in liquid nitrogen (-196°C). To prepare the cells for the cryo-conservation, the cells were harvested at an early passage as described above and centrifuged at 300g for 5min. They were then resuspended in 1ml of cryo-medium (70% medium with supplements, 20% FCS, 10% DMSO) and filled into screw-top cryo-vials. To freeze the cells, the vials were placed in a Styrofoam-box and frozen step-wise (2h at -20°C; 48h at -80°C) before storing them in liquid nitrogen.

To take a fresh batch of cells into culture the cryo-vials were taken out of the liquid nitrogen and thawed in the waterbath at 37°C. Subsequently the cells were washed with 10ml of fresh medium and centrifuged at 300g for 5min before being resuspended in 10ml of fresh cell culture medium and being seeded out on cell culture dishes. After 24h, the medium of the cells was changed to remove dead cells and debris.

33

Cell counting

Before cell-counting, the cells were harvested as described above and 10µl of the cell suspension were mixed with 10µl of a 0.4% Trypan blue solution. 10µl of this mixture were added to the haemocytometer and the cells were counted. Dead cells incorporate Trypanblue due to their damaged cell membrane and appear blue in the haemocytometer. Blue stained cells were excluded from the cell count.

Production and concentration of lentiviral particles

The 293T cells were used as producer cell line for production of lentiviral particles. To produce lentivirus, the 293T cells were transfected with the lentiviral plasmids pLVTHM, pMD2 and psPAX2. 48h before the transfection, 2x106 293T cells were seeded on a collagen- coated 10cm culture dish. 2-3h before the transfection, the culture medium was changed. Per dish 5µg pLVTHM, 3.25µg psPAX2 and 1.75µg pMD2 were added to 250µ NaCl-solution (150mM). In a separate tube, 20µl JetPEI was mixed with 250µl NaCl-solution. Subsequently, the JetPEI containing solution was added to the DNA and the solution was briefly vortex. After an incubation of 15-30min, the DNA/JetPEI solution was added dropwise to the cells. 24h after the transfection the medium was replaced with 10ml of fresh medium. After a further incubation of 48h, the supernatant containing the lentiviral particles was harvested. It was centrifuged at 500g to remove debris and sterile filtered, using a 0.2µm filter. To produce larger amounts of lentiviral containing supernatant, 15cm dishes were used and the amount of cells, DNA, NaCl, JetPEI and medium was scaled up by the factor 2.5.

Concentration of the lentiviral containing supernatant was performed via ultracentrifugation. 25-30ml of supernatant was centrifuged in screw-top ultracentrifuge tubes in the ultracentrifuge L7-65 at 50,000g for 2h. After the centrifugation the medium was discarded and the pellet containing the lentiviral particles was dissolved in 50µl PBS. The concentrated lentivirus was aliquoted and stored at -80°C.

Transduction of target cells

Adherent cells were generally transduced with unconcentrated lentivirus containing supernatant. Just before the transduction, the target cells were harvested as described above and seeded out on six-well plates. The wells of the six-well-plates contained 2ml of fresh medium and 1ml of lentiviral supernatant. Polybrene was added to the medium at a 34

final concentration of 4µg/ml to increase the transduction efficiency. To transduce suspension cells such as THP1 cells, the cells were counted and seeded out in 24-well plates in a volume of 500µl at a density of 4x105 cells/ml. Polybrene was added at a concentration of 4µg/ml and subsequently 2µl concentrated lentivirus was added to the cell suspension. After 24h of incubation the medium was changed. To determine the efficiency of the transfection the GFP-expression of the cells was analysed 72h after the transduction using either a fluorescent microscope (488nm excitation) or a flowcytometer (FITC channel).

2.3. Protein biochemistry methods

Westernblotting

Sample-preparation, preparation of SDS-PAGES, electrophoreses of the samples and semi- dry blotting of the proteins on the PVDF-Membran was performed according to the protocols of Molecular Cloning162.

After the prepared samples had been run of the SDS-PAGE and the proteins blotted onto the PVDF-membrane via semi-dryblotting, the blotted membranes were blocked overnight with PBS-T containing 5% milk powder under constant shaking at 4°C. The primary antibodies against the protein of interest were added to the membranes in blocking buffer at the concentrations listed in Table 2 and incubated for 1h at RT. Unbound primary antibody was removed by washing the membrane three times in PBS-T for 10min. The corresponding secondary antibodies were diluted in PBS-T at the concentration listed in Table 3, added to the membrane and incubated for 1h. After three washes, the signal of horse-radish- peroxidase (HRP) coupled to the secondary antibody was analysed in the LAS-3000-scanner. The ECL Plus (Amersham, Little Chalfont, UK) was used as substrate solution for the detection of the HRP.

CX3CL1-ELISA

A CX3CL1-ELISA (Enzyme-linked Immuno-sorbant Assay) was used to determine the amount of CX3CL1 released by stable cell line ECV-CX3CL1. All samples and standards (312 to 4,875ng recombinant CX3CL1) were diluted in PBS-T containing 2% BSA. 96-well plates (Nunc-Immuno™, Nunc A/S, Roskilde, Denmark) were coated with 4µg/ml CX3CL1 catching antibody (MAB3652, R&D Sytems) overnight at RT. Subsequently, the wells were blocked for

35

2h with PBS-T containing 2% BSA. After washing the wells three times with PBS-T, the samples and the CX3CL-1 standard were added and incubated for 1h at RT. After three washes, the anti-CX3CL1 biotin-conjugated antibody was added at a concentration of 300ng/ml diluted in PBS-T and incubated for 1h. After washing three times with PBS-T, Strepavidin-POD was added at a dilution of 1:5000 (in blocking buffer) and incubated for 1h. After 4 washes with PBS-T, the substrate solution was added and incubated for 10-20min until the reaction was stopped by the addition of 100µl 5% H2SO4. The absorption of the wells was measured at 450nm in a plate reader, subtracting the readout of 550nm to correct the value.

Glucoronidase-assay

In general, the glucoronidase assay was used to determine the number of cells which had migrated in a transmigration or chemotaxis assay or that had adhered to the different substrates in the adhesion assay. In the glucoronidase assay, the cell number of a given sample was determined by measuring the β-glucoronidase-activity of the sample via the synthetic substrate p-nitrophenyl-β-D-glucoronide.

For the glucoronidase assay, the cells were lysed using a final concentration of 0.1% Triton. Then 100µl of the lysed cells were transferred to a 96-well and 100µl of the substrate buffer was added. After an overnight incubation at 37°C, 100µl of the stopping buffer was added and the absorption at 405nm was determined in the plate reader. By comparing the readout values of the samples to the readout of a standard dilution of cells (starting number of cells used in particular assay stepwise diluted 1:2), it was possible to determine the cell number of the sample.

Surface staining for flowcytometry

To stain cellular surface proteins for the flowcytometry experiments, the cells were harvested as described above and added into FACS-tubes containing 1.5ml of ice-cold PBS. After centrifuging for 5min at 300g (this centrifugation speed and duration applied to all further steps), the cells were resuspended in 50µl of FACS-buffer containing the respective antibodies at the concentrations listed in Table 2. After an incubation of 30-45min on ice, 1ml of FACS-buffer was added, the cells were briefly vortexed and then centrifuged. The cells were then resuspended in FACS-buffer containing the secondary antibodies (diluted

36

1:100). After an incubation of 30-45min on ice, 1.5ml ice-cold PBS was added to the cells and the tube was briefly vortexed before being centrifuged. After the last wash the cells were resuspended in 100µl ice-cold PBS and 100µl of 4% PFA was added to fix the cells and to deactivate any remaining lentiviral particles. The stained cells were stored at 4°C. Before the flowcytometry experiments 300µl of PBS was added to the cells to dilute the PFA.

2.4. Molecular biology methods

Cloning of pLVTHM plasmids

To insert the oligonucleotides containing the target shRNA sequences into the pLVTHM plasmid, the lyophilised oligos were first dissolved in H20 at a concentration 300pmol/µl. 2µl of each oligo were added to 21µl of H20 and 25µl of the annealing buffer (x2) were added. The oligos were annealed in the PCR-cycler by heating them to 95°C for 4min and a stepwise cooling process 4min each at 85°C, 82°C, 80°C, 78°C, 75°C, 72°C, 70°C). After pausing at 70°C for 10min, the oligos were taken out of the PCR-cycler and cooled to RT on the bench. Subsequently the oligos were phosphorylated using PNK-kinase with the following protocol:

5µl annealed oligos

12µl H20

2µl T4 buffer (x10)

1µl PNK kinase

Incubation at 37°C for 30min

Heat-deaktivation of PNK at 70°C for 10min

In the meantime, 1µg of the empty pLVTHM vector had been digested with the restriction enzymes ClaI and MluI. The linearized plasmid-backbone was separated from the 6bp- residue by agarose gel electrophoreses on 0.8% agarose-gel. After cutting the plasmid out of the agarose-gel, a gel-purification (NucleoSpin Extract II-kit) was performed and the plasmid was eluted with 20µl H20 which was warmed to 70°C. The ligation of the phosphorylated oligos and the opened plasmid was performed using the following protocol:

37

5µl phosphorylated oligos

1µl pLVTHM (20-50ng DNA)

11µl H20

2µl T4 ligase buffer (x10)

1µl T4 ligase (1U)

After an incubation of 60-120min at RT, 4µl of the ligation-solution were used for the transformation (see below) of 100µl competent bacteria. The transformed bacteria were plated out on LB-agar plates containing the carbenicillin. After incubating the agar plates overnight at 37°C, the colonies that had grown on the agar plate were picked with sterile pipette-tips and inoculated 3ml of LB containing carbenicillin (50µg/ml). After growing the bacteria cultures overnight (37°C, constant shaking), miniprep (NucleoSpin Plasmid, eluting with 50µl TE) were performed to extract the plasmid DNA. The screening for positive clones was performed by analytic restriction diges,t using the restriction enzymes ClaI and EcoRI (2µg MiniPrep DNA, 0.2µl of each enzyme) and a subsequent agarose gel electrophoreses. The successful insertion of the oligos was confirmed when the digested plasmid showed a band of 286bp together with the 10.8kb band of backbone instead of a 236bp fragment which appears when digesting the empty pLVTHM plasmid.

Transformation of bacteria

Chemical competent bacteria (Z-CompetentTM E. Coli Transformation Buffer Set) were transformed by heat shock. After thawing the frozen bacteria on ice, 4µl of ligation reaction or 0.1µg of purified plasmid were added to the bacteria suspension. After an incubation of 20-30min on ice, the bacteria were plated out on prewarmed (37°C) LB-agar plates and incubated for 18h at 37°.

Propagation of plasmids (Maxiprep)

Plasmids were propagated by inoculated 200ml LB containing carbenicillin (50µg/ml) with either 100µl of a transformed bacteria culture or a bacteria colony picked from an agar plate. After an overnight incubation at 37°C under constant shaking maxiprep (NucleoBond PC 500) was performed according to the manual of the kit. 38

Measuring DNA and RNA concentration

The NanoDrop-ND-1000 (PEQLAB Biotechnologie GmbH, Erlangen, Ger) was used to determine the the concentration of DNA preparations and RNA samples.

RNA isolation and cDNA synthesis

For the RNA-isolation of our cells of interest, cells were harvested as described above. The cells were counted and 1x106 cells were then washed with PBS. If the RNA-isolation was not performed directly after the cell harvesting, the pellets of the cells were temporarily stored at -80°C. The RNA was isolated with the RNeasy Mini Kit (Qiagen, Hilden, Ger), according to the manual of the kit. To avoid contamination by genomic DNA, DNase-treatment was performed during the isolation. The RNA was eluted with 20µl RNase free water. The RNA content of the samples was determined using the NanoDrop-ND-100 and the samples were stored at -80°C. cDNA sythesis was performed with the RevertAid™ H Minus First Strand-Kit using 1µg of the isolated RNA per reaction.

Quantitative PCR

The quantitative PCR (qPCR) was performed in the LightCycler 480 (Roche, Penzberg, Ger) using 96-well plates. The mastermix for the qPCR-reaction was made using the following recipe (volumes were multiplied by number of total samples):

5,5µL Sybr Green I Master-Mix (Roche, Prenzberg, D)

3,3µL H20

0,55µL forward primer (6,25pmol/µL)

0,55µL reverse primer (6,25pmol/µL)

9µl of the qPCR master and 1µl of sample-cDNA were added to each wells. Generally, measured triplets of 1:2 and 1:16 dilutions of the sample cDNA were measured. The annealing temperatures and number of cycles which were used for each target gene are listed in Table 8. The housekeeping gene GAPDH was used as reference genes for the transcription levels of our target genes.

39

Table 8: Annealing temperatures and number of cycles for the qPCR for each target gene.

Template Annealing- Number of cycles Temperatur

hADAM8 62 °C 42

hADAM9 58 °C 40

hADAM10 61 °C 45

hADAM17 55 °C 45

GAPDH 66 °C 33

2.5. Functional assays

Transmigration and chemotaxis assays

For the transmigration assay, 5,000 HMVEC were seeded out on transwell inserts with a pore size of 8µm. The transwell insert contained 100µl of medium in the insert and 600µl of medium in lower chamber (a well of a 24-well plate). The cells were grown to full confluency for 4d while the medium was changed daily. To not disturb the cell layer during the medium change, only half of the medium of the insert and the lower compartment was changed.

On the day of the assay, the migrating THP1 cells were harvested and counted. After a centrifugation at 300g for 5min, the cells were resuspended at 2x106 cells per ml in RPMI without FCS and PS containing 0.05% BSA (transmigration medium). In case of a chemotaxis experiment, the transwell filter without the endothelial layer were preincubated with 600+100µl of transmigration medium in the incubator in the meantime. Just before the assay, the medium of the transwell insert with (transmigration) or without endothelial layer (chemotaxis) were put in the free spot beside the well containing the prewarmed 600µl transmigration medium with the desired MCP-1 concentration. The medium of the transwell insert was taken off and 100µl of the cell suspension (containing 2x105 cells) were added to the insert. The insert was then placed in the well, containing the transmigration medium with the desired MCP-1 concentration. After an incubation of 2h in the cell culture incubator the inserts were taken out of their wells and the transmigrated cells were analysed with either flowcytometry and/or by the glucoronidase assay (see Results).

40

Ca2+-influx assay

For the Ca2+-influx assay, the THP1 cells were harvested as described above, counted and centrifuged at 300g for 5 min. The cells were resuspended in Ca2+-staining buffer at a density of 2x106cells/ml. After an incubation of 30min at 37°C, the cells were washed twice (centrifugation at 300g for 5 min) with 5ml of prewarmed Ca2+-assay buffer. During the second wash, an aliquot of the cell suspension was counted and the cells were resuspended at a density of 4x106cells/ml. 100µl of the cell suspension (4x105 cells) were added to a black 96-well-plate with a clear bottom. The Ca2+-influx upon the stimulation with MCP-1 was then measured at a constant temperature of 37°C in the FLUOstar OPTIMA plate-reader. After establishing the fluorescent baseline of the cells for 10s, either 5µl (1nM MCP-1 final) or 17µl (3nM MCP-1 final) of a 21nM MCP-1 solution (diluted in Ca2+-assay buffer containing 0.5% BSA) was injected to the labelled cells. As buffer control, the same volume of Ca2+-assay buffer containing 0.5% BSA but no MCP-1 was injected. The fluorescence was constantly measured at an excitation wavelength of 488nm and an emission wavelength of 526nm. After the stimulation of the cells, the labelling of the THP1 with Fluo-3-AM was verified by injecting the detergent digitonin (125µg/ml final). This lead to a maximum increase of the signal since because of the permeabillisation of the cell membranes the Fluo-3-AM stored inside the cells could bind to the free Ca2+ of the assay buffer. To determine the background fluorescence, EGTA (15mM final) was injected to disrupt the complex and capture all free Ca2+.

Adhesion assay

For the adhesion assay 96-well-plates (Nunc-Immuno™ Plate) were coated with 40µl soluble VCAM-1 (1.5µg/ml), ICAM-1 (1.5µg/ml) and fibronectin (6µg/ml) for 1h at RT. After the incubation, the wells with the adhesion molecules and the control wells were blocked with 200µl of 1% BSA (all diluted PBS (PAA)). The wells were washed twice with PBS and once with RPMI with 0.05% BSA. In the meantime, the THP1 were harvested, counted and centrifuged (300g for 5min). The THP1 were resuspended in RPMI with 0.05% BSA at a density of 1x106cells/ml. 100 µl of the cell suspension was added to the coated wells. The THP1 were incubated for 10min in the presence or absence of 3nM MCP-1 at 37°C. After the incubation, the wells were twice washed with 200µl prewarmed RPMI containing 0.05% BSA. Then 100µl of RPMI with 0.05% BSA containing 0.1% Triton X-100 were added to the wells

41

to lyse the adherent cells. The cell number of adherent cells was determined with glucoronidase assay (see above).

Scratch assay

The wound-healing properties of the ADAM-KD LV-ECV-304 were analysed with the scratch assay. The cells were cultured in six-wells to full confluency. On the day of the assay, a scratch was made on the cell layer using a 100µl pipette tip. Microscopic pictures of the scratch were taken 0h, 2h, 4h, 6h, 8h and 24h after the injury. The migration behaviour of LV-ECV 304 was evaluated by counting the cells that had invaded a defined area of the scratch.

2.6. Statistics

All data sets were analysed with the statistic software GraphPad Prism 5 (GraphPad Inc., San Diego, CA USA) using one-way analysis of variance (ANOVA) and the Dunnet-test as post- hoc test. If not indicated otherwise, it was always tested against either the solvent control or the scramble control cell lines. The data were considered significantly different at probability values smaller 0.05 (0.01, 0.001), marked with * (**, ***). All figures of this work show the mean ±SD (standard devariation) if not indicated otherwise.

42

3. Results

3.1. Comparison of ADAM8, 9, 10 and 17 expression in different leukocytic, endothelial and epithelial cell culture lines

To get an overview about the relative expression of ADAM proteases in different cell types, mRNA expressions profiles of ADAM 8, 9, 10 and 17 were determined by quantative real- time PCR (qPCR) in cultured cell lines and primary cells of different origin: ECV304 (human bladder carcinoma cell line), A549 (human alveolar basal epithelial adenocarcinoma cell line), HEK293 (human embryonic kidney cell line), HMVEC (human microvascular endothelial cell line), HUVEC (primary human umbilical vein endothelial cells), THP1 (primary human acute monocytic leukemia cells) and Jurkat (primary human T cell leukemia cells). While ECV304 and A549 are epithelial cells, HEK293 cells may represent fibroblasts or epithelial cells, HMVEC and HUVEC are both primary endothelial cells and THP1 and Jurkat are both leukocytic cell lines.

Figure 7 shows a heat map which summarises the results of several qPCRs (n=4) of the mRNA level of the four ADAM proteinases in the mentioned cell lines. The heat map shows that each cell line displays a distinct ADAM profile.

Among the tested cell lines, ECV304 cells had the highest level of ADAM8. Their ADAM9 and 10 expression levels were about average while their ADAM17 expression level was among the lowest. By contrast, A549 cells - the other examined epithelial cell line - had a low level of ADAM8, a comparatively high level of ADAM9 and relative low levels of ADAM10 and 17. In comparison to the former cell lines, HEK293 cells showed high levels of both ADAM9 and 17 and average levels of ADAM8 and 9.

The two endothelial cell lines HMVEC and HUVEC had distinct expression levels of ADAM10 and 17. HMVEC, which were derived from the microvasculature, had high levels of ADAM10 and 17. HUVEC, which were isolated from the umbilical vein, expressed both proteinases only at a basal level. Both cell lines had a low level of ADAM8 and a rather high level of ADAM9.

The two leukocytic cell lines THP1 and Jurkat both expressed ADAM8 slightly above average and ADAM9 at a very low level. Their expression profiles of ADAM10 and 17 differed though. While Jurkat cells expressed both ADAM10 and 17 at a basal level, THP1 cells showed high ADAM10 mRNA expression and an above average expression level of ADAM17.

43

These findings demonstrate that all the tested proteases are abundantly expressed in different cell types but with varying expression levels.

The high expression of ADAM8, 10 and 17 in THP1 cells make them a good model cell line for the study of these proteases in leukocyte migration.

Figure 7: Heat map of mRNA levels of ADAM8, 9, 10 and 17 in several cell culture lines. The amount of each ADAM-mRNA was determined using qPCR (n=4). The expression rate of the ADAM genes was normalised to the housekeeping gene GAPDH. The colour of each square represents at which level a gene is expressed compared to the average of the examined cell lines. Light green indicates that a gene is expressed at a very high rate while red shows that a gene is very weakly expressed in the corresponding cell line. Average expression levels are show in brown. (For exact quantitative data see Appendix)

3.2. The effect of ADAM-inhibitors on leukocyte migration

To study the importance of ADAM proteases for leukocyte migration an in vitro migration assay was used. In vitro migration experiments can be performed with primary leukocytes isolated from different blood donors or leukocytic cell lines. Since primary leukocytes often show donor specific behaviour in migration assays and are not easily accessible for genetic manipulation it was decided to use THP1163 cells in the migration assays. THP1 cells have been widely used to study a variety of functions and mechanisms of monocytic cells in vitro assays over the last decades. Like native monocytes THP1 cells are chemotactically attracted by the chemokine CCL2, commonly known as MCP-1 (monocyte chemotactic protein-1)164–166.

44

MCP-1-induced transendothelial migration of THP1 cells

THP1 cell migration was investigated in an endothelial transmigration assay using transwell inserts. To mimic the physiological endothelium of a microvessel, the site where recruitment and transmigration of leukocytes occurs in vivo, human microvascular endothelial cells (HMVEC) were cultured on transwell filter inserts. The confluent HMVEC monolayer, that had grown on the filter inserts, was then used as migration barrier in the transmigration assay. HMVEC cells were used since their microvascular endothelial phenotype resembles the in vivo situation much better than that of human umbilical vein endothelial cells (HUVEC) which are commonly used in these assays. MCP-1 served as chemoattractant in the migration studies. MCP-1 was added to the cell culture wells before the transwells were inserted and THP1 cells were added to the inserts. To establish and optimize the transmigration assay the optimal MCP-1 concentration at which THP1 cells are attracted by MCP-1 was determined. As shown in Figure 8 the highest THP1 cell migration rate through the layer of HMVEC was observed at a concentration of 3nM MCP-1.

100

80

60

40

% migrated cells migrated % 20

0

M 0nM 1n 3nM 10nM 30nM MCP-1 conc.

Figure 8: Concentration kinetic of MCP-1-induced THP1 cell transmigration through HMVEC. The transmigration assay was performed using a transwell set up with detachable filter insert with defined pores of 8µm. On the filter bottom of the insert, a HMVEC monolayer had been grown to complete confluence. MCP-1 was added at the indicated concentration to the lower chamber of the transwell setup. 200 000 THP1 cells were then added to the upper chamber of the transwell and incubated for 2h. After the incubation period, the number of migrated cells in the lower chamber was determined using the glucoronidase assay. Data is shown as mean in % of total applied cells ± SD, n=2.

45

Influence of endothelial cell in the in vitro transmigration assay

The influence of the HMVEC layer in the in vitro transmigration assay of THP1 was analysed by comparing the migration rates of THP1 in the transmigration assay in presence of HMVEC to the number of migrated cells in the chemotaxis assay. For the chemotaxis assays, the same transwell inserts with 8µm pores were used as in the transmigration assay but without the confluent endothelial cell layer.

Comparing the number of cells that migrated in the transendothelial migration to the percentage of the initial cell number that migrated in the chemotaxis assays (Figure 9) it is apparent that the cells migrate towards the chemotactic stimulus of MCP-1 at a threefold higher rate through the transendothelial barrier than through the filter membrane alone. In the absence of chemotactic stimulus, the random migration in presence of endothelial cells is even 10-fold higher. Instead of being just another additional barrier which the migrating cells have to overcome, the migration of the THP1 is supported by the presence of the HMVEC layer.

60

50

40

30

10

% migrated cells migrated % 5

0

Filter no stim. Filter + MCP-1 HMVEC no stimHMVEC + MCP-1

migration conditions

Figure 9: Comparison of THP1 cells migration in in vitro chemotaxis and transdendothelial migration assays. THP1 cells migrated in absence (filter) and presence (HMVEC) of a confluent HMVEC layer. Random migration in absence of a chemotactic stimulus (open columns, no stim) was compared to directed migration towards 3nM MCP-1 (striated columns, + MCP-1). 200 000 THP1 cells were added to the upper chamber of the transwell and incubated for 2h. After the incubation period, the number of migrated cells in the lower chamber was determined by counting, using flow cytometry. Data is shown as mean in % of total applied cells ± SD, n=4.

46

Effect of the ADAM inhibitors GI254023 and GW282406 on the chemotaxis and transmigration of THP1 cells

Having found the optimal MCP-1 stimulus concentration for the induction of THP1 cell transmigration, inhibition studies were performed to obtain first insight in the potential involvement of ADAM proteases in THP1 cell migration.

For these experiments the inhibitors GI254023 and GW282406 were used that had been characterized previously for inhibition of ADAM10 and 17 activity 15,21,108,167. While the inhibitor GI254023 preferentially blocks ADAM10 but not ADAM17, GW282406 inhibits both ADAM10 and 17 with high potency. The effect of GI254023 and GW282406 on the migration response of THP1 was analysed both in chemotaxis and transmigration assays.

In first experiments the effect of both inhibitors on MCP-1 induced transmigration of THP1 cells through cultured endothelial cells was analysed. For the transmigration assay THP1 cells were pretreated with the inhibitors GI254023 or GW282406 at a concentration of 10µM inhibitor for 1h. DMSO which was used as solvent for both inhibitors served as negative control and was used in parallel. The migration assay itself was performed in absence of the inhibitors or DMSO to exclude inhibition of ADAMs on endothelial cells and to specifically examine the role of leukocytic ADAMs for migration. For this purpose, THP1 cells were washed prior to the assay to remove free inhibitors. Pretreatment with GI254023 resulted in a slight but significant THP1 transmigration reduction through the HMVEC monolayer (Figure 10A) whereas the slight reduction by GW282406 did not reach significance. This can be explained by the fact that both GI254023 and GW282406 are reversible antagonists of ADAM10 and ADAM10 and ADAM17 respectively. It is likely that the inhibitors diffused of the ADAM proteases within the incubation period of 2h and that the proteases thereby regained their full proteolytic activity during the assay.

In the next experiments cell migration was performed without endothelial cells and it was analysed whether the ADAM inhibitors are more effective when present throughout the assay. For the chemotaxis assays, the same transwell inserts were used as in the transmigration assay but without the endothelial cell layer. THP1 cells were preincubated for 1h with the inhibitors GI254023 or GW282406 at a concentration of 10µM. Following pretreatment, the inhibitors were not removed and the chemotaxis assay was performed in the presence of the inhibitors in both compartments of the transwell setup. The results depicted in Figure 10A showed that both inhibitors GI254023 and GW282406 had a profound impact on the THP1 cell migration when present during the chemotaxis assay. Since both inhibitors inhibited ADAM10, this result indicated that ADAM10 could play a crucial role 47

during the migration process of the THP1 cells. The additional ADAM17 inhibition by GW282406 which inhibits both ADAM10 and 17 did not result in a further decrease of THP1 cell migration. Since there was no additive effect of the additional ADAM17 inhibition this finding suggests that ADAM17 could be of minor importance for the migration process.

A 1.5 B 1.5

1.0 1.0 *

0.5 0.5 *** ** migration index migration migration index migration

0.0 0.0 I O GI G W GW G MS MSO D D Inhibitors Inhibitors

Figure 10: Influence of ADAM inhibitors on THP1 migration. A: Transmigration of THP1 cells which had been incubated with the ADAM specific inhibitors GI254023, GW282406 or solvent control for 1h prior to the assay. Of each pretreated group, 200 000 THP1 cells were added to the filter inlay of a transwell set up on which a monolayer of HMVEC had been grown to full confluency. Just before the assay, the cells had been washed to exclude an effect of the inhibitors on the ADAMs of the endothelial cells. B: Chemotaxis of THP1 cells in the presence of ADAM inhibitors GI254023 (GI, inhibits ADAM10) and GW282406 (GW, inhibits ADAM10 and 17) at a concentration of 10µM. DMSO at a concentration of 0.1% was used as a negative control. 200 000 THP1 cells, which had been preincubated for 1h with the respective inhibitor, were added to the filter inlay of a transwell setup with 8µm pores. In both sets of experiments MCP-1 was added at a concentration of 3nM to the lower compartment of the transwell setup as chemotactic stimulus. The number of migrated cells in the lower compartment of the transwell was determined subsequent, to the incubation period of 2h using the glucoronidase assay. The experiments are expressed as migration index relative to the DMSO solvent control which was set to 1. Data are shown as mean ±SD, n=3, *p<0.05, **p<0.01, ***p<0.001 compared to solvent control.

3.3. Lentiviral shRNA-based knock-down of ADAM 8, 9, 10 and 17

The use of pharmaceutical inhibitors like GI254023 and GW282406 has the disadvantage that inhibition of non-target molecules such as other ADAMs and proteases can never be fully excluded. To further support the pharmacological inhibition experiments and to elucidate the involvement of other ADAM proteases in the migration process of leukocytes the specific transcriptional inhibition technique by RNA interference (RNAi) was employed.

Since the discovery of gene silencing by small interfering RNA (siRNA) in 2001 in mammalian cells168 the RNAi technology has become a invaluable tool for the study of gene function.

48

There are two general methods to apply siRNAs to the target cells. One is the direct transfection of synthetic short RNA duplexes. In the second method, vectors containing a RNA polymerase III promoter and the sequence for a short hairpin RNA (shRNA) are transfected into the target cells. The transcription machinery of the transfected cells subsequently produces the shRNA which then gets further processed to siRNA.

However, one of the major obstacles of gene knock down (KD) via siRNA remains the gene delivery of recombinant siRNA or shRNA encoding plasmids to the cells of interest. Very few cultured cell lines are easily transfected at high rates by conventional transfection methods. Leukocytic cell lines like THP1 and primary cell lines such as HUVEC are even more difficult to transfect than most standard cell lines. Since those were the cell lines of interest of this study, it was decided to make use of the advanced and powerful gene delivery by lentiviral vectors.

Lentiviral transduction efficiency test

Over the last few years, lentiviral transduction has become a widely established gene delivery method for in vitro experiments as well as for several in vivo setups. Lentiviral transduction has two key advantages over the adenoviral and retroviral transduction system. The lentivirus can transduce dividing, as well as non-dividing, cells and it integrates its genetic message stably into the hostgenome. It can also be equipped with a great variance of surface proteins a process termed pseudotyping. The pseudotyping of lentivirus alters its target cell specifity and its biochemical and mechanical properties. The most common, used surface protein is the vesicular stomatitis virus glykoprotein (VSV-G) which provides the lentivirus with a broad tropism and great biochemical and mechanical robustness.

In the initial lentiviral experiment it was tested whether the difficult to transfect cell lines were transducable by lentivirus. Cell culture supernatant from 293T producer cells containing lentiviral particles was added to cultured HUVEC. Those lentiviral particles were generated by the 293T cells after they had been transfected the with the lentiviral shRNA plasmid pLVTHM160, the gag-pol plasmid psPax2 and the envelope plasmid pMD.2. The pLVTHM plasmid encodes a shRNA cassette and the gene for the green fluorescent protein (GFP) as a reporter gene. Therefore, successful transduction of target cells can be easily monitored via fluorescent microscopy or flow cytometry. Three days after the lentivirus application the transduced HUVEC cells were analysed in a fluorescent microscope. In the initial experiment transduction rates of virtually 100% was achieved (Figure 11). Encouraged by this positive

49

outcome, cloning and screening of potential shRNA sequences against the targets ADAM10 and 17 was started.

A B

Figure 11: Analysis of lentiviral transduced HUVEC. HUVEC were incubated with cell culture supernatant containing GFP-encoding lentivus for 24h. After medium change, cells were further incubated for 48h and then analysed by bright field microscopy (A) and fluorescence microscopy (B).

Non-viral shRNA-KD of ADAM10

In a previous study, gene knock down of ADAM10 via shRNA was achieved with a shRNA sequence expressed by the non-viral pSuper plasmid103,159. Since the pSuper plasmid utilizes the same shRNA cassette as the lentiviral pLVTHM plasmid, it was the aim to insert the pSUPER shRNA cassette into the lentiviral setup. First, it was necessary to confirm that the shRNA sequences mediate effective downregulation of ADAM10. HEK293 cells were transfected with the ADAM10-pSuper vector using a liposomal transfection reagent and after 3 days ADAM10 surface expression was analysed to determine the gene KD efficiency. As expected when using liposomal transfection, only a part of the cells were successfully transfected. Those cells that were apparently successfully transfected with the pSuper- ADAM10 plasmid showed a markedly reduced ADAM10 surface staining (Figure 12). On the one hand this result confirms that the shRNA sequence works effectively and on the other hand it demonstrates that liposomal transfection is not sufficient to target the majority of cells. This underlines that the importance to combine effective shRNA sequences with effective transduction methods to yield a pronounced gene knock-down.

50

Figure 12: Surface expression of ADAM10 on HEK293 cells transiently transfected with pSuper-ADAM10. HEK293 cells were transfected with pSuper-ADAM10 (dotted line) and scramble control-pSuper (solid black line). 72h after the transfection, ADAM10 surface expression was analysed in a flow cytometer. While no reduction of surface expression was detectable for the scramble transfected cells, about 50% of the pSuper-ADAM10 transfected cells show a clearly reduced ADAM10 surface staining. Those cells represent the successfully transfected cells where the gene KD of ADAM10 was in progress. The part of the population which expresses ADAM10 as high as the scramble control was apparently not transfected. The isotype control is shown as grey continuous line. Data are shown as representative histogram of at least three independent experiments.

Lentivirus-mediated shRNA-KD of ADAM10

Since the previous experiment had demonstrated that the shRNA sequence of the pSuper- ADAM10 plasmid works for ADAM10-KD, this sequence (ADAM10-650) was cloned into the pLVTHM plasmid. To confirm the results by another shRNA, a second plasmid with a different shRNA-sequence against ADAM10 (ADAM10-1947) was generated. These plasmids were transfected into 293T producer cells to generate lentivirus containing supernatants. Primary HUVEC cells were transduced using the lentiviral supernatants to confirm the ability of the pLVTHM-ADAM10 plasmids to silence ADAM10.

Similar to the initial lentiviral transduction experiment, high transduction rates of about 95% were achieved. The transduction efficiency was determined by analysing GFP expression of the transduced cells in the flow cytometer (Figure 13A). KD of ADAM10 by the two shRNA sequences ADAM10-650 and ADAM10-1947 was demonstrated by flow cytometry analysis of ADAM10 surface expression (Figure 13B and C). Quantitative real-time PCR (qPCR) analysis of ADAM10 mRNA of ECV cells transduced with both ADAM10-pLVTHM lentivirus and scramble control (scr) lentivirus proved that the ADAM10-shRNA sequences induced the silencing of ADAM10 mRNA in the transduced cells (Figure 13D).

51

A B

C D 1.5

1.0

0.5 (rel. to(rel.GAPDH) ADAM10 transkript. level transkript. ADAM10 0.0

scr 650 47 0- 0-19 A1 A1 LV-ECV-KD-lines

Figure 13: Analysis of ADAM10 gene KD by pLVTHM-vectors ADAM10-650 and ADAM10- 1947. A: Histogram showing distribution of GFP-expression by lentiviral transduced HUVEC. All lentiviral transduced HUVEC show the same level of GFP fluorescence which indicates similar transduction rates. B: Histogram showing distribution of ADAM10 surface expression by GFP positive transduced cells. Cells were stained with a mouse monoclonal primary antibody against ADAM10 and a secondary PE-conjugated anti-mouse antibody. ADAM10-650 transduced HUVEC (dashed line) showed a drastically reduced ADAM10 expression compared to the scramble control (solid black line). The solid grey line represents the isotype control. C: Histogram of pLVTHM-ADAM10-1947 (dashed line) transduced HUVEC. The sequence ADAM10-1947 is also able to silence ADAM10 expression. D: Analysis of ADAM10 mRNA in pLVTHM transduced ECV level by quantitative PCR. Cells transduced with lentivirus carrying shRNA against ADAM10 have a lower copy number of ADAM10 mRNA compared to the scramble control. The mRNA of the housekeeping gene GAPDH was used to normalise for variations in the amount of RNA between the different samples. All diagrams of this figure are representative experiments of several experiments performed.

Lentiviral KD of ADAM17

Based on the success of the ADAM10 KD by lentiviral transduction, pLVTHM plasmids with shRNA sequences against ADAM17 were generated. Transcriptional silencing via synthetic siRNA or vector based shRNA rely on the same cellular mechanisms. Therefore, one successful siRNA, sequence which was used in a previous study of our group, was cloned into the pLVTHM vector generating the pLVTHM-ADAM17-2061 construct14. To confirm the

52

results by another shRNA against ADAM17 several other pLVTHM-ADAM17 vectors were designed and constructed. Theses sequences were designed with the help of algorisms that predict potential shRNA sequences169 (pLVTHM-ADAM17-1631 and others).

A B

C D 1.5

1.0

0.5 (rel. to to GAPDH) (rel. ADAM17 transkript. level transkript. ADAM17 0.0

scr -2061 A17 LV-ECV-KD-lines

Figure 14: Gene silencing of ADAM17 with pLVTHM-lentiviral vectors. A: Histogram for GFP-expression of lentiviral transduced HUVEC. The lentiviral transduced HUVEC exhibit similar levels of GFP fluorescence. B: ADAM17-surface staining of HUVEC transduced with pLVTHM-ADAM17-2061 (dashed line). The shRNA sequence ADAM17-2061 induces strong reduction of ADAM17 surface expression compared to the HUVEC transduced with the scramble-control (solid black line). C: ADAM17 surface expression on HUVEC transduced with pLVTHM-ADAM17-1631 (dashed line). This shRNA sequence was non-functional as the ADAM17 expression of the HUVEC transduced with ADAM17-1631 is as high as the cells transduced with the scramble-control (black solid line). D: ADAM17-mRNA level of ECV304 transduced with ADAM17-2061 (filled bar) compared to scramble control (open chart). The analysis of ADAM17 mRNA level in qPCR proved that the KD of ADAM17 surface expression is caused by the transcriptional silencing. All diagrams of this figure are representative experiments of more than three experiments performed.

After production of lentiviral supernatants with the pLVTHM-ADAM17 plasmids, transduced HUVEC were analysed to determine the efficiency of the ADAM17-shRNA sequences. Cells were analysed by flow cytometry and qPCR. As shown in Figure 14A all transduced HUVEC had comparable levels of GFP expression which indicates a high and uniform transduction

53

rate. The shRNA-sequence ADAM17-2061 clearly reduced ADAM17 surface expression on HUVEC (Figure 14B) and the number of ADAM17 mRNA transcripts of transduced ECV304 cells (Figure 14D).

By contrast, the sequence ADAM17-1631 did not effect the ADAM17 surface expression (Figure 14C) even though the cells were successfully transduced (Figure 14A long dashed dark line). The same results were obtained with two other shRNA sequences against ADAM17 (data not shown). This outcome underlines that not every shRNA sequence that aligns to its target mRNA can be used to silence gene transcription. The ability of a shRNA- sequence to silence its target gene can be predicted to a certain degree via complex algorisms169 but its efficiency to silence a gene in vitro still has to be determined experimentally in each target cell type.

Specificity of ADAM10 or ADAM17 gene silencing by shRNA

The overexpression of short RNAs in the cell and the lentiviral transduction can have general effects on the transcription machinery of cells transduced with lentiviral shRNA vectors. To address this problem, the effects of shRNA treatment against a target molecule must always be compared to a non specific shRNA sequences (scramble control) that does not interact with any known mRNA.

One further major obstacle of the RNAi technique are unwanted off target effects caused by unspecific binding of the siRNA sequences to none target mRNAs. Therefore, it is mandatory to study at least two different shRNA sequences against a given target molecule. Possible off target effects of shRNA sequences against closely related molecules should as well always be excluded. The proteases ADAM10 and ADAM17 are a classical example of two closely related proteins. They share not only sequence homologies in their mRNA sequence and their protein motifs, but as well exhibit similar biological functions. Furthermore, deficiency of either ADAM10 or 17 may lead to compensation by increasing the expression of the related protease as it has been reported for fibroblast generated from ADAM10 or ADAM17 deficient mice97. Therefore, the influence of shRNA sequences against ADAM10 on the ADAM17 expression and vice versa was analysed on all transduced cell lines.

A typical analysis of the side effects of shRNA sequences against ADAM10 or ADAM17 on the protein expression of respective other ADAM protease is shown in Figure 15. The ADAM10 surface expression of HUVEC transduced with lentivirus containing shRNA against ADAM17 (A17-2061) was as high as the ADAM10 surface expression of cells transduced with the

54

scramble control virus (scr) (Figure 15A). Cells with the gene silencing of ADAM10 (A10-650) showed a strongly reduced ADAM10 surface expression as seen before. In comparison, the ADAM17 surface expression of HUVEC transduced with lentivirus containing shRNA sequences against ADAM10 was comparable to that of scramble transduced HUVEC while the effect of the ADAM17-shRNA was clearly noticeable (Figure 15B).

D B

Figure 15: Effects of shRNA against ADAM10 or ADAM17 on the surface expression of the respective other ADAM protease. Surface expression of ADAM10 (A) and ADAM17 (B) on HUVEC transduced with lentivirus expressing the scramble control, ADAM10-650 (dashed line) or ADAM10-2061 shRNA (dotted line). Cells transduced with other successful shRNA sequences against ADAM10 or ADAM17 (ADAM10-1947, ADAM17-2642) showed also no influence on surface expression of the respective other protease. The diagrams of this figure are representative experiments of several experiments performed.

3.4. Effect of ADAM10 and 17 silencing on shedding activity

To confirm that gene silencing of ADAM10 and 17 leads to reduced cleavage of their substrates on the cell surface, the shedding activity of cells transduced with pLVTHM vectors against ADAM10 or ADAM17 was analysed. One of the common shedding substrates of both ADAM10 and 17, is the chemokine CX3CL1 (Fraktalcine). CX3CL1 is one of only two chemokines (the other being CXCL16) which are expressed as transmembrane protein. Upon cleavage by ADAM10 and 17, its chemokine domain gets released from the cell surface of endothelial cells, where it is predominantly expressed.

Soluble CX3CL1 is a potent chemoattractant for T-Cells and monocytes whereas the membrane-bound form induces strong adhesion of leukocytes. It has been shown that ADAM10 contributes mainly to the constitutive CX3CL1 shedding. In contrast, increased shedding of CX3CL1 by ADAM17 is observed after stimulation with the phorbolester PMA (phorbol 12-myristate 13-acetate) or inflammatory stimuli such as Tumour Necrosis Factor- alpha (TNF-α) and Interferon-gamma (INF-γ).

55

ECV304 cells that stably expressed CX3CL1 (ECV-CX3CL1) were transduced with lentivirus containing the shRNA sequences against ADAM10 and 17. The influence of the gene silencing of ADAM10 or ADAM17 on the CX3CL1 shedding activity was analysed by measuring the amount of released CX3CL1 with ELISA.

As depicted in the left panel of Figure 16, unstimulated scramble transduced CX3CL1-ECV304 released about 60% of their total CX3CL1 into their supernatant. In comparison, unstimulated cells with lentiviral induced ADAM10-KD released significantly less (only 30%) of their CX3CL1. Unstimulated ADAM17-deficient ECV-CX3CL1 cells shed about 45% of their CX3CL1. This was less than the scramble control transduced cells, but still more than the CX3Cl1 shedding observed for ADAM10-deficient ECV-CX3CL1 cells.

100 constitutive 80 PMA induced *** 60 * * 40

20 % released CX3CL1 % released

0 r c 1 0 1 s 50 6 5 6 0 scr -6 0- -2 0 -206 1 7 A 1 A1 10-1947 A10-1947A A A17

LV-ECV-CX3CL1-KD-lines

Figure 16: Effect of ADAM10 and 17 KD on the shedding of CX3CL1. ECV-304, which stably expressed CX3CL1, were transduced with lentiviral particles containing the control shRNA sequence (scr) and sequences against ADAM10 (A10-650 and 1947) and ADAM17 (A17-2061). The cells were cultured on 6-well dishes until they reached complete confluency. Before the CX3CL1 shedding assay, the cells were washed with PBS and received 1ml fresh serum-free medium. Cell were left unstimulated (open columns) or treated with 200ng/ml PMA (hatched columns on the right panel of the diagram). After incubating the cells for 2h the supernatant was taken off and the remaining cells were lysed with PBS containing 1% Triton. The amount of CX3CL1 of lysate and supernatant was determined by ELISA. The columns represent the percentage of CX3CL1 in the supernatant of the total amount of CX3CL1 (amount of cell-bound CX3CL1 in lysate plus amount CX3CL1 of the supernatant). Data are shown as mean ±SD, n=4, *p<0.05, ***p<0.001 compared to scramble control.

56

The stimulation of the ECV-CX3CL1 cells with PMA led to an increased release of CX3CL1 as expected (left panel Figure 16). In these PMA-stimulated cells, an effect of the ADAM17-KD was clearly visible whereas ADAM10 KD was ineffective. After stimulation with PMA, ADAM17 shRNA transduced ECV-CX3CL1 cells released only 55% of their total CX3CL1 into their cell culture medium. ADAM10-shRNA transduced ECV304-CX3CL1 as well as cells transduced with the scramble control virus released 75% of their total CX3CL1.

This experiment was the prove of principle for the lentiviral shRNA approach to downregulate ADAM10 and 17 since it is coherent to previous studies in which ADAM10 and 17 activity was inhibited with the inhibitors GI254023 and GW28240613,67.

Lentiviral silencing of ADAM8 and 9

While the function of ADAM10 and 17 has been studied extensively in the last decade, other proteins of the ADAM family received less attention. Proteolytic activity and shedding of surface molecules has as well been demonstrated for ADAM8 and ADAM9. Therefore, both proteases were included in the lentiviral knock down experiments with the aim to study their role in leukocyte recruitment. For this purpose, pLVTHM-plasmids with shRNA sequences against both ADAM8 and ADAM9 were generated.

Figure 17A shows the ADAM8 mRNA level of ECV-304 cells that had been transduced with pLVTHM-lentivirus containing shRNA sequences against ADAM8. All three sequences against ADAM8 induce the silencing of its mRNA significantly as shown by qPCR. Unfortunately, the KD of ADAM8 could not be shown on protein basis, since none of the tested antibodies against ADAM8 could detect the protease by either western blotting or surface staining for flow cytometry.

Successful gene silencing of ADAM9 was achieved as well, as can be seen in the results of the qPCR against ADAM9 (Figure 17B). The western blot for ADAM9 in the lysates of EVC transduced with shRNA sequences against ADAM8 and 9 (Figure 17C) showed that the cells transduced with ADAM9-pLVTHM vectors have almost no detectable ADAM9. In contrast, the cells transduced with ADAM8-pLVTHM vectors expressed ADAM9 at the same level as scramble transduced cells confirming specificity of the shRNA sequences.

57

A 1.5 B 2.0

* 1.5 1.0 * 1.0 ** 0.5 (rel. to GAPDH) (rel.to (rel. GAPDH) (rel. to 0.5 *** *** ADAM8 mRNA level mRNA ADAM8 ADAM9 mRNA level mRNA ADAM9

0.0 0.0 5 r 31 0 scr 903 8 sc 8- -264 -226 A 8 9 A8-1 A A9-1967 A

LV-ECV-KD-lines LV-ECV-KD-lines

C

Figure 17: Gene silencing of ADAM8 and 9. A: Quantitative real-time PCR analysis of ADAM8 transcripts of ECV cells transduced with pLVTHM lentivirus containing shRNA sequences against ADAM8 or scramble control (n=5). B: Expression of ADAM9 mRNA by ECV transduced with shRNA sequences against ADAM9 or scramble control. Levels of transcripts were analysed with qPCR and normalised to the housekeeping gene GAPDH (n=4). C: Western blot of ADAM9 in lysates of ECV cells transduced with the indicated pLVTHM lentivirus (representative of several experiments). The data presented in the bar charts are shown as mean ±SD, *p<0.05, **p<0.01, ***p<0.001 compared to scramble control.

3.5. Role of ADAM 10 and ADAM 8, 9, 17 for transmigration and chemotaxis of THP1 cells

To further investigate the involvement of ADAM10 and that of ADAM8, 9 and 17 in the process of leukocytic transmigration, stably transduced LV-THP1 cell lines with a gene KD for either of these ADAM-proteinases were generated. The analysis of the mRNA expression

58

profiles of the four ADAMs (Figure 7) had shown that THP1 cells fitted this purpose since they expressed ADAM10 and 17 at higher levels than Jurkat cells, the other leukocytic cell lines which had been analysed. To generate the gene KD of the target molecules, those of the previous described lentiviral shRNA vectors which had been shown to successfully silence these genes in HUVEC and ECV304 cells, were used.

A B

C D

Figure 18: Flow cytometry analysis of ADAM10 and 17 surface expression on lentiviral transduced THP1. A: ADAM10 surface staining of the stable cell line LV-THP1-ADAM10-650 (dashed line) compared to that of LV-THP1-scr (solid black line) which served as the nonspecific shRNA control. B: ADAM10 surface expression of the cell line ADAM10-1947. C: ADAM17 surface expression of LV-THP1-ADAM17-2061. D: ADAM17 surface expression of LV-THP1-2642. The cells were stained with a primary mouse monoclonal antibody against ADAM10 or ADAM17 and a secondary anti-mouse antibody which was conjugated with APC. The data are shown as representative histograms of at least three experiments.

Stable gene silencing of ADAM8, 9, 10 and 17 in lentiviral transduced THP1 cells

THP1 are suspension cells which are generally more difficult to transduce with lentiviral particles than adherent cells. To transduce the THP1 cells lentiviral particles, which had been concentrated via ultracentrifugation, were used. The unconcentrated lentiviral supernatant,

59

which had been used in the previous experiment to transduce adherent cells like HUVEC and ECV304, did not contain enough viral particles per ml to achieve a sufficient transduction rate.

The GFP, encoded by the pLVTHM-plasmids, proved to be very helpful to determine the successful transduction of the THP1 cells since it can be easily detected in the flow cytometer. In general, the THP1 cells were transduced once or twice with the concentrated lentivirus until a transduction rate of almost 100% was achieved. After the transduction, the newly stable LV-THP1 cell lines were kept in culture to determine the KD efficiency and to use them in the transmigration experiments.

A 1.5 B 1.5

1.0 * 1.0 * ** * 0.5 0.5 * (rel. GAPDH) (rel. to (rel. GAPDH) to ADAM8 transkript. level transkript. ADAM8 ADAM9 transkript. level transkript. ADAM9 0.0 0.0 P 3 1 5 F 3 4 P G -90 F r -18 -26 1967 c A8 9- s A8 A8 scr G A A9-2260

LV-THP1-KD-lines LV-THP1-KD-lines

C 1.5 D 1.5

1.0 1.0

0.5 0.5 * (rel. to GAPDH) (rel. to (rel. to (rel.to GAPDH) ADAM10 transkript. level transkript. ADAM10 ADAM17 transkript. level transkript. ADAM17 * 0.0 0.0

cr 0 s 47 -65 9 scr -1 A10 10 17-2642 A A17-2061 A LV-THP1-KD-lines LV-THP1-KD-lines

Figure 19: qPCR analysis of ADAM8, 9, 10 and 17 mRNA transcript levels in stable LV- THP1 cell lines. A: Downregulation of ADAM8 mRNA in the stable cell lines LV-THP1 ADAM8-903, 1831 and 2645 (n=3). B: ADAM9 mRNA levels of the cell lines LV-THP1-ADAM9-1967 and 2210 (n=3). C: Example of the ADAM10-mRNA silencing in the cell lines LV-THP1 ADAM10-650 and 1947. D: Expression level of ADAM17 mRNA in the cell lines LV-THP1-ADAM17-2061 and 2642 (n=3). The data presented are shown as mean ±SD, *p<0.05, **p<0.01 compared to scramble control.

60

The successful gene silencing of ADAM10 and 17 on protein level is depicted in Figure 18. The gene KD was confirmed in the flow cytometer analysing the surface expression of either protein. Figure 18A and B show the reduction of surface ADAM10 induced by the lentiviral vectors pLVTHM-ADAM10-650 and 1947. The reduction of surface ADAM17 by the vectors pLVTHM-ADAM17-2061 and 2642 is shown by Figure 18C and D.

Successful gene silencing of ADAM8, ADAM9, ADAM10 or ADAM17 in THP1 cells, stably transduced with the respective lentivirus, was demonstrated by qPCR analysis of the proteases mRNA levels (Figure 19A to D). Together with the protein expression data, these results indicate that all shown shRNA sequences can be used to silence expression of the respective ADAM protease in THP1 cells.

Refinement of the transmigration assay read out

In preliminary transmigration experiments, THP1 cell lines with gene silencing of ADAM10 or ADAM17 were used in parallel with the scramble control cells. The glucoronidase assay was used as read out to determine the number of transmigrated cells. Unfortunately, the variations between each set of experiments had been high and the data generated in those experiments did not meet the initial expectations. Therefore, the read out of the original transmigration assay was refined. The aim was to establish a transmigration set up with an internal control. This set up should include the scramble control cells in the same sample with the cells with the gene KD for the ADAM proteinase.

To distinguish scramble control cells and ADAM-KD LV-THP1, the sequence of the GFP in the pLVTHM scramble plasmid was exchanged for the sequence of cyan fluorescent protein (CFP). CFP can be excitated with the 405nm laser of the flow cytometer while GFP is excitated by the 488nm laser. This made it possible to detect cells expressing either fluorescent protein in different channels of the flow cytometer. Using the newly generated plasmid, the scramble LV-THP1 control cells (expressing CFP) could be mixed with ADAM-KD LV-THP1 (expressing GFP) in the same sample.

Figure 20A shows a schematic drawing of the modified transmigration assay. Both cell lines were mixed before the assay and added to the filter insert of the transwell setup. After the incubation period, the cell population in the lower compartment was analysed for their ratio of CFP- versus GFP-positive cells by flow cytometry. The CFP/GFP-ratio of the initial cell mixture, that was added to the transwell set up at the beginning of the transmigration, was

61

used as reference to determine the influence of the gene silencing on the migratory abilities of the THP1 cells.

A typical flow cytometry analysis of CFP-labelled scramble control cells and GFP-labelled ADAM10-KD LV-THP1 cells is shown in Figure 20B. The panels at the left show the analysis of the loading control of the cell mixtures of CFP- and GFP-labelled cells that were added into the transwell insert. The panels at the right are derived from the analysis of transmigrated cells. Quadrant Q1 of each panel is the gate for CFP-positive cells which represent the scramble-CFP-LV-THP1 cells while quadrant Q3 is the gate for the GFP-positive LV-THP1 cells. Both upper panels show the control sample with scramble-CFP and scramble-GFP cells. This scramble control sample was included in each set of experiments. The lower panels show the result of a transmigration experiment in which CFP-positive scramble control migrated together with GFP-labelled ADAM10-KD LV-THP1 cells.

Both upper panels of Figure 20 show that the ratio of GFP- versus CFP-labelled scramble cells stayed the same before and after the transmigration. In the loading control of the scramble sample (upper left panel) 38% of the cells expressed CFP and 55% expressed GFP. After the transmigration (upper right panel), the transmigrated cells contained 37% CFP positive and 54% GFP positive cells. Percentage of GFP-positive cells in the loading control before transmigration (55%) did not differ from percentage of GFP positive cells in the transmigrated cell population (54%). Vice versa, the percentage of CFP-positive cells remained the same before and after transmigration. This indicates that both, the CFP- and the GFP labelled LV-THP1-scramble cells, migrated at about the same rate. For subsequent experiments changes in the distribution were compared by dividing the percentage of GFP positive cells before transmigration by that after transmigration, respectively.

As an example, the flow cytometric analysis of a transmigration experiment with ADAM10-KD LV-THP1 is shown in the lower panels of Figure 20B. Both panels show dot blot diagrams of a cell mixture of CFP-labelled scramble LV-THP1 control cells and GFP-labelled ADAM10-KD cells before and after transmigration. The lower left panel shows that the loading control before transmigration contained 36% CFP positive scramble cells and 53% GFP positive ADAM10-KD cells. Of the transmigrated cells (lower right panel), 62% are CFP positive and only 27% GFP positive. In this experiment the ratio of GFP positive cells before (53%) versus after (27%) the transmigration is at 0.51 (0.27/0.53). This indicates that in this experiment THP1 cells lacking ADAM10 migrated at a clearly lower rate than the control cells.

62

A

B

Figure 20: The refined transmigration assay using CFP labelled control cells and GFP labelled KD cells. A: Schematic drawing of the refined transmigration assay. CFP-labelled LV-THP1 scramble control cells (blue-labelled cells) and GFP-labelled ADAM-KD LV-THP1 cells were added into the transwell insert. At the end of the incubation period the cells that had transmigrated into the lower compartment were collected and analysed by flow cytometry. B: Dot blot diagrams of transmigrated CFP and GFP-labelled LV-THP1 which had been analysed by flow cytometry. Cells in quadrants Q1 are CFP positive while quadrant Q3 includes GFP positive cells. The upper panels show the ratio of GFP- scamble and CFP-scramble LV-THP1 before (loading control on the left) and after the transmigration (right panel). Both lower panels show the ratio of scramble-CFP and ADAM10-KD LV-THP1 before (left panel) and after the transmigration (right panel).

63

With the refined transmigration setup it is possible to look exclusively at cells carrying the knock-down of the target genes, excluding non-transduced cells. Another advantage of this system is that it gives highly reproducible results and that it allows the use of flow cytometry as a read out method. Flow cytometry is a fast and straight forward read-out system which is capable of analysing high cell numbers and many different populations in one experiment.

Effect of ADAM10 knock down on THP1 cell chemotaxis and transmigration

The inhibition of THP1 cell chemotaxis and transmigration by the ADAM inhibitors GI254023 and GW282406 (Figure 10) had indicated that ADAM10 seems to play a more important role in the migration of THP1 cells than ADAM17. Therefore, the LV-THP1 with the gene silencing of ADAM10 were tested first in the refined transmigration setup. Figure 21A shows that less ADAM10-KD cells migrated in response to MCP-1 trough the endothelial monolayer compared to the scramble control cells. The ADAM10-650 LV-THP1 cells migrated at a rate of 80% compared to the included scramble control, while the ADAM10-1947 cells migrated at a rate of only 61%.

In addition, both ADAM10 deficient cell lines showed a migration deficit in MCP-1 induced chemotaxis when endothelial cells are not present. The cell line ADAM10-650 migrated at a rate of 80% and the ADAM10-1947 at only 65% (Figure 21B).

This findings are in accordance with the inhibitor studies where the inhibition of ADAM10 led to a strongly decreased migration, especially in the chemotaxis assays that were performed in the presence of the inhibitors (Figure 10A). It is noteworthy though, that the inhibition of the ADAMs with the inhibitors GI254023 (inhibits ADAM10) and GW282406 (inhibits ADAM10 and 17) decreased the transmigration rate of the wt THP1 cells to 21% and 32%, respectively. This may be caused by the fact, that at the concentration the inhibitors were used, all ADAM10-molecules present are inhibited. The shRNA treatment leads to a down regulation of about 80-90% (see Figure 18A and B and Figure 19C), which means that the cells still possessed 10-20% of their initial amount of ADAM10 molecules.

The undirected transendothelial migration in absence of the chemotactic stimulus MCP-1 through the endothelial layer (Figure 21C) is as well impaired by the knock-down of ADAM10 (ADAM10-650 95% and ADAM10-1947 75%), although only the ADAM10-1947 cell line showed a significant difference to the scramble LV-THP1. The same observation was made for the undirected chemotaxis (Figure 21D) where the ADAM10-650 cells migrated at an

64

average rate of 77% and the ADAM10-1947 at a rate of 94% compared to the included scramble LV-THP1.

A B

1.0 1.0 * * ** **

0.5 0.5 change CFP/GFPof ratio change CFP/GFPof ratio 0.0 0.0 0 7 0 5 scr 5 47 scr -6 -6 9 0 -1 10 10 A1 10-194 A A A

LV-THP1-KD-lines LV-THP1-KD-lines

C D

1.5 1.0 *** 1.0 * 0.5 0.5 change of CFP/GFP ratio change of CFP/GFP ratio 0.0 0.0

47 scr 650 9 scr 47 - 19 0-1 - A10 10 A1 A10-650 A

LV-THP1-KD-lines LV-THP1-KD-lines

Figure 21: Transmigration and chemotaxis of ADAM10-LV-THP1 in response to MCP-1. A: Transmigration of ADAM10-KD LV-THP1 cells through an endothelial layer of HMVEC with MCP-1 as chemotactic stimulus. B: Chemotaxis of LV-THP1 with shRNA KD of ADAM10 through 8nm filters in presence of MCP-1 as chemotactic stimulus. C: Random transmigration rate of ADAM10 deficient LV- THP1 through an endothelial layer of HMVEC in the absence of a chemotactic stimulus. D: Random migration of ADAM10-KD LV-THP1 in the chemotaxis assay with no chemotactic stimulus present in the lower chamber of the transwell. All data are shown as the average change in the CFP/GFP ratio of the transmigrated cells compared to the CFP/GFP ratio of the input cell mixture that was added to the insert of the transwell. Data are shown as mean ±SD, n=6, *p<0.05, **p<0.01, ***p<0.001 compared to scramble control.

In parallel to the relative quantification, also absolute numbers of transmigrated cells were determined. The number of transmigrated cells of each sample was determined to have an alternative readout value for the transmigration. To count the number of transmigrated cells by flow cytometry, a defined amount of counting beads was added to each sample before

65

the flow cytometry analysis. The flow cytometry analysis was stopped, when a defined number of the counting beads had been detected. Since the ADAM10-KD LV-THP1 cells could be differentiated from the scramble control cells because of their different fluorescent proteins, it could be determined how many cells of each population had transmigrated.

A GFP-positive LV-THP1 B CFP-positve scramble LV-THP1 50 50

40 40

30 30

20 ** * 20 *** 10 10 ** % migrated GFP-pos. cells %migrated GFP-pos. 0 cells CFP-pos. migrated % 0 0 cr 5 m scr s 6 ti - s 0 o A10-650 n A10-1947 A1 r A10-1947 c scr no stim s LV-THP1-KD-lines GFP positive-LV-THP1-KD-lines with which the CFP-scramble LV-THP1 migrated

C GFP-positive LV-THP1 D CFP-positve scramble LV-THP1 15 15

10 10

*** 5 5 *** ** *** % migrated CFP-pos. cells migrated %CFP-pos. % migrated GFP-pos. cells GFP-pos. migrated % 0 0

m 50 m scr i scr 6 ti st - 1947 s 1947 - o A10-650 no A10 n A10- A10 scr scr LV-THP1-KD-lines GFP positve-LV-THP1-KD-lines with which the CFP-scramble LV-THP1 migrated

Figure 22: Percentage of migrated cells in the transmigration and chemotaxis assays of ADAM10-LV-THP1. A: Percentage of GFP-positive scramble (scr) and ADAM10-KD (A10-650 and A10-1947) LV-THP1 cells that migrated through the transendothelial barrier. B: Number of CFP-labelled control cells which migrated together with scramble control cells (scr) or ADAM10-KD (A10-650 and A10-1947) LV-THP1 cells. C: Migration rates of GFP expressing ADAM10-KD LV-THP1 cells in the chemotaxis assays expressed in percentage of inserted cells. D: Percentage of migrated CFP-labelled control cells that competed with scramble or ADAM10-KD cells in the chemotaxis assay. All data is shown as the average percentage of transmigrated cells compared to the number of cells that were added to the insert of the transwell. To show the effect of the MCP-1 stimulus the percentage of scramble cells that migrated in the absence of MCP-1 (scr no stim) is shown in all diagrams. Data are shown as mean ±SD, n=5, *p<0.05, **p<0.01, compared to scramble control.

66

As expected, the GFP-labelled ADAM10-KD LV-THP1 cells transmigrated through the endothelial layer at a significant lower rate than the CFP-labelled scramble-control cells (Figure 22A). The transmigration of the GFP-positive scramble cells which migrated together with ADAM10-KD LV-THP1 cells, was not significantly affected (Figure 22B A10-650 and A10- 1947). However, there was a notable decrease in their number as well. When CFP-positive scramble cells migrated with GFP-labelled control cells 30% of the inserted CFP-positive cells migrated. When the CFP-positive scramble cells competed with ADAM10-KD LV-THP1, an average of only 21% (ADAM10-650) respectively 22% (ADAM10-1947) migrated.

In the chemotaxis assays, the GFP-positive ADAM10-KD LV-THP1 migrated at a significant lower percentage compared to the GFP-scramble control cells (Figure 22C). Similar to the transendothelial migration, the total number of CFP-labelled scramble control cells was reduced when they migrated together with ADAM10-KD LV-THP1 cells instead of scramble GFP-positive control cells. (Figure 22D). However, this effect was not significant.

In summary, gene knock-down of ADAM10 results in reduced THP1 cell migration towards MCP-1 independent of the presence or absence of endothelial cells. It also seems possible that ADAM10-downregulation could affect migration of neighbouring cells with normal levels of ADAM10. This is indicated by the attenuated migration of CFP-labelled scramble controls in the above described experiment. Although this effect did not reach significance, it could well be that non migrating cells with knockdown of ADAM10 compete with the control cells for limited binding contacts on endothelial cells or chemotaxis membranes and thereby hinder their migration.

Role of other ADAM proteases for transmigration and chemotaxis of THP1 cells

The results of the transmigration and chemotaxis assays performed with ADAM10-KD LV- THP1 cells (Figure 21 and Figure 22) confirmed the findings obtained with the ADAM inhibitors GI254023 and GW282406. Both, pharmaceutical and transcriptional inhibition, indicated a crucial role of ADAM10 in the transmigration of THP1 cells. Due to the lack of specific inhibitors, the influence of other ADAMs such as ADAM8, ADAM9 or ADAM17 on leukocytes could not be directly investigated by inhibition experiments. To study the involvement of other proteases in the migration process, THP1 cell lines with the gene silencing of ADAM8, 9 or 17 (see Figure 18 and Figure 19) were investigated.

67

Effect of ADAM8 knock down on monocytic migration

Figure 23 shows the results of the transmigration and the chemotaxis assays performed with the ADAM8-KD LV-THP1 cell lines. Surprisingly, the gene silencing of ADAM8 had an even higher impact on the leukocyte transmigration (Figure 23A) than the KD of ADAM10 (Figure 21A and Figure 22A). All three ADAM8 shRNA sequences tested reduced the transendothelial migration of ADAM8-KD LV-THP1 significantly. Their inhibitory effect was stronger than that of the ADAM10-KD shRNA sequences when comparing the change of CFP/GFP ratio (Figure 23A). In line with the transmigration experiments, the migration rates of all there ADAM8 silenced LV-THP1 cell lines were also strongly impaired in the chemotaxis assays (Figure 23B).

A 1.5 B 1.5

1.0 1.0 * ** ** ** 0.5 *** 0.5 *** change CFP/GFP of ratiochange change of ratio CFP/GFP change 0.0 0.0 5 1 r 03 scr 83 645 sc 64 -1 2 -2 A8-903 8 A8-9 A A8- A8-1831 A8

LV-THP1-KD-lines LV-THP1-KD-lines

Figure 23: Effect of ADAM8-KD on transmigration and chemotaxis of THP1 cells. A: Transmigration of ADAM8-KD LV-THP1 cells through an endothelial layer of HMVEC with MCP-1 as chemotactic stimulus. B: Chemotactic migration of LV-THP1 with shRNA silencing of ADAM8 through 8nm filter in presence of MCP-1 as chemotactic stimulus. Data is shown as change of CFP/GFP ratio before and after the transmigration. Data are shown as number of migrated cells relative to the scramble control. The data of all panels are shown as mean ±SD, n=4, *p<0.05, **p<0.01, ***p<0.001 compared to scramble control. Diagrams expressing the effect of the gene silencing of ADAM8 as relative number of migrated LV-THP1 cells are found in the Appendix.

Effect of ADAM9 knock down

In contrast to the gene silencing of ADAM8 (Figure 23) and ADM10 (Figure 21 and Figure 22), gene silencing of ADAM9 in LV-THP1 cells did not reduce the migratory response of the LV-THP1 cells (Figure 24). LV-THP1 with silenced ADAM9 expression migrated almost at the same rate as the scramble control cell (ADAM9-1967) or even at a significantly higher rate

68

than the control cells (ADAM9-1967). The ADAM9-KD LV-THP1 cells showed this migration behaviour in presence (Figure 24A) or absence of the endothelial layer (Figure 24B).

A 2.0 B 2.0

1.5 * 1.5

1.0 1.0

0.5 0.5 change of CFP/GFP of ratio CFP/GFP change change of CFP/GFP of ratio CFP/GFP change 0.0 0.0 7 6 r scr 9 260 sc -2 -2260 9 A9-1 A9 A9-1967 A

LV-THP1-KD-lines LV-THP1-KD-lines

Figure 24: Migration rates of ADAM9-LV-THP1 cells in transmigration and chemotaxis assays. A: Transmigration assay of LV-THP1 with a gene silencing of ADAM9 through an endothelial layer of HMVEC with MCP-1 as chemotactic stimulus. B: Chemotaxis assay of ADAM9-KD LV-THP1 through 8nm filter in presence of MCP-1 as chemotactic stimulus. Data is shown as change of CFP/GFP ratio before and after the transmigration. The data of all panels are shown as mean ±SD, n=4, *p<0.05, compared to scramble control. Diagrams expressing the effect of the gene silencing of ADAM9 as relative number of migrated LV-THP1 cells are found in the Appendix.

Effect of ADAM17 Knock-down

The analysis of the migratory behaviour of the ADAM17-KD LV-THP1, (Figure 25) largely confirmed the inhibition studies (Figure 10) which had shown that combined inhibition of ADAM10 and ADAM17 by GW282406 does not result in an additional migratory deficit compared to preferential inhibition of ADAM10 by GI254023. While the shRNA sequence ADAM17-2642 did not affect the migration rate of the LV-THP1 the second sequence ADAM17-2061 decreased the CFP/GFP-ratio of scramble versus ADAM17-KD cells slightly but significantly (Figure 25A). This indicates that ADAM17-KD cells may have only slightly reduced migratory abilities in direct comparison to the scramble cells. Both ADAM17-KD LV- THP1 cell lines did not migrate at a significant lower rate in the chemotaxis assay (Figure 25).

69

A 1.5 B 1.5

* 1.0 1.0

0.5 0.5 change ratioCFP/GFP of change change of CFP/GFP ratio of CFP/GFP change 0.0 0.0 r c 2 cr s 64 s -2 7 A17-2061 A1 A17-2061 A17-2642 LV-THP1-KD-lines LV-THP1-KD-lines

Figure 25: Transmigration and chemotaxis assays with ADAM17-LV-THP1. A: Transmigration of ADAM17 gene silenced LV-THP1 cells through an endothelial layer of HMVEC with MCP-1 as chemotactic stimulus. B: Chemotaxis assay of LV-THP1 with ADAM17 gene silencing through 8nm filter in presence of MCP-1 as chemotactic stimulus. Data is shown as change of CFP/GFP ratio before and after the transmigration. The data of all panels are shown as mean ±SD, n=4, *p<0.05, compared to scramble control. Diagrams expressing the effect of the gene silencing of ADAM17 as relative number of migrated LV-THP1 cells are found in the Appendix.

3.6. Role of ADAMs for the sensitivity of LV-THP1 cells towards different MCP-1 concentrations

The transmigration and chemotaxis experiments had shown that LV-THP1 cells with a gene silencing of ADAM8 and 10 (Figure 21 and Figure 23) have migratory deficiencies compared to scramble LV-THP1. THP1 cells with a gene knock down of ADAM9 and 17 in contrast migrate at more or less the same rate as scramble cells (Figure 24 and Figure 25). These findings gave rise to the question how ADAM8 and 10 contribute to the migration of the leukocytes.

Gene knock down of ADAM8 and ADAM10 showed the same effect on transendothelial migration and chemotaxis of THP1 cells. The latter findings demonstrate that the migration deficit was independent of the endothelial cell layer. One possible explanation may be that the ADAM8-KD and ADAM10-KD affects the sensitivity of THP1 towards MCP-1.

To test this hypothesis, transmigration experiments at MCP-1 concentrations, ranging from 0nM to 30nM MCP-1, were performed with the different ADAM-KD LV-THP1 cell lines. Figure 26A shows the dose response curve of the five LV-THP1 cell lines tested. The curve has the typical bell shaped outline which is common for dose response curves of chemokines. The ADAM9-1967 LV-THP1 (up-pointing triangles) migrated at all MCP-1 concentration at higher

70

rates than the control scramble cell line (circles) which was in accordance with the results of the initial transmigration experiments (Figure 24). However, a lateral shift of the peak of the

A 40 Scr A8-903 30 A9-2260 A10-1947 20 A17-2642

10 % of migrated cells migrated of %

0

nM nM 1nM 3nM 0 0 0nM 1 3 MCP-1 concentration

B 2.0 scr A8-903 1.5 A9-2260 A10-1947 1.0 A17-2642

0.5 change of change CFP/GFPratio 0.0

M M M M M M M M n n nM nM 1n 3 1n 3nM 1nM 3n 0n 1 3nM 1nM 3nM 10nM30nM 10n 30 10nM3 10nM30 10n 30nM

MCP-1 concentration

Figure 26: Concentration dependent transmigration of ADAM-KD LV-THP1 towards MCP-1. A: Dose response curve of LV-THP1 with gene silencing of ADAM8, 9, 10 and 17 and scramble control cells towards MCP. The results of one representative experiment are shown. B: Change of the CFP/GFP ratios of scramble-CFP cells versus ADAM-KD-GFP LV-THP1 over the course of the different MCP-1 concentrations. The data are shown as mean ±SD, n=3.

concentration dependent migration curve could not be observed. This would have indicated a higher or lower sensitivity of ADAM9-KD LVTHP1 to MCP-1. Accordingly to the transmigration

71

experiments of ADAM8, 10 and 17-KD-LV-THP1 at 3nM (Figure 22, Figure 23, Figure 25) the percentage of total migrated cells of the ADAM8 (squares), ADAM10 (down pointing triangles) and ADAM17 (diamonds) deficient LV-THP1 was lower than that of the scramble LV-THP1. However, a lateral shift of the curve was not observed.

The change of the CFP/GFP ratios of scramble-CFP cells versus ADAM-KD-GFP LV-THP1 over the course of the different MCP-1 concentration is shown in Figure 26B. While ADAM8 or 10 deficient cells migrated constantly at lower rates than the scramble cells, there is no concentration at which these cells would migrate at a significant higher rate compared to 3nM MCP-1 which was the concentration at which all previous transmigration and chemotaxis experiments had been performed. Neither ADAM9 deficient cells nor ADAM17 deficient LV- THP1 cells showed a significant change of their CFP/GFP ratio of scramble-CFP versus ADAM- KD-GFP LV-THP1 at any MCP-1 concentration. While ADAM9-KD LV-THP1 migrated at a constant higher rate than the scramble control cells, the ADAM17-KD LV-THP1 migrate at about the same rate as the scramble control cells.

3.7. The effect of ADAM8, 9, 10 and 17 knock-down on the Ca2+ response of THP1 cells to MCP-1

The results of the concentration dependent transmigration experiments (Figure 26) had shown that the differences in the migration rates of LV-THP1 cells with a gene silencing of ADAM8, 9, 10 or 17 were not caused by a decreased or increased sensitivity of the cells towards MCP-1. Another possible reason for the migratory deficits of ADAM8 and ADAM10- KD LV-THP1 could be an impaired MCP-1 signalling. One important step in the signalling event of chemokines is the release of the second messenger Ca2+ from the endoplasmatic reticulum which occurs after the binding of chemokines to their respective G-protein coupled receptor. In case of MCP-1, the signalling and the Ca2+-release is mediated by the seven- membrane spanning G-protein coupled receptor CCR2170.

MCP-1-induced Ca2+response of THP1 cells

To test the hypothesis of an impaired MCP-1-induced signalling of THP1 cells caused by the inhibition or gene silencing of ADAM-proteinases, THP1 cells were assayed for MCP-1 induced changes in cytoplasmic free Ca2+ using the FLUOstar OPTIMA plate-reader This fluorescence plate-reader is equipped with two injectors which make it possible to

72

automatically inject stimuli such as MCP-1, while messuring fluorescence of cells labelled with fluorescent dyes like Flou-3 or Fura-2. These fluorescent dyes increase or shift their fluorescence signal after binding the Ca2+, released by the intracellular stores of the cells. In this study, Fluo-3 was used as Ca2+ -indicator which is a well established fluorochrome and has been widely used in Ca2+-flux assays in plate-reader setups.

A

16000 3nM MCP 14000 buffer control

12000

10000 rel. light units light rel. 8000

6000 0 50 100 150 time (s)

B 8000

6000

4000

in rel. light units in light rel. 2000 increase of increase flourescence 0 P C ffer uffer B M MCP Bu n l 1nmM 3 µ 5µl 17

Figure 27: Ca2+-flux of wt THP1 after MCP-1 stimulation. A: Diagram showing the time course of fluorescence increase after stimulation of the Flou3-labelled THP1 cells with 3nM MCP-1. A representative experiment showing the average fluorescence of a triplet of wells containing Fluo3-labelled THP1 cells is shown. B: Comparison of the average fluorescence increase after stimulation with 1 and 3nM MCP-1 and their corresponding buffer controls. The data are shown as mean ±SD, n=3.

73

A typical Ca2+-flux measurement is shown in Figure 27A. During the first 10s of the experiment, the base line fluorescence of Flou3-labelled cells was measured. Then MCP-1 (3nM solid line) was injected which led to a transient increase in fluorescence indicating the increase of cytoplasmic free Ca2+ in the cells. The response reached its maximum within 5s and returned to the baseline level within 90s. As control, the same volume of buffer was injected to a different well of the labelled cells (dotted line) to exclude a false positive effect caused by the mechanical stimulation of the cells due to the injection of the MCP-1 solution.

The Ca2+-response of the THP1 cells to different concentrations of MCP-1 was tested. Besides, the concentration of 3nM MCP-1 at which the transmigration and chemotaxis experiments had generally been performed 1nM MCP-1 was used as stimulation concentration in all Ca2+-influx experiments. Figure 27B summarises the results of the stimulation with both MCP-1 concentrations and their corresponding buffer control for three experiments with wt THP1 cells.

Influence of the ADAM inhibitors GI254023 and GW282406 on the Ca2+-response of wt THP1

The Ca2+-flux assay was first used to study the influence of the ADAM inhibitors GI254023 and GW282406 on the stimulation of wt THP1 cells by MCP-1. The cells were pretreated with the inhibitors at a concentration of 10µM for 1h and subsequently stimulated with MCP-1 (1nM and 3nM). The inhibition of the ADAM proteases by the inhibitors slightly decreased Ca2+-transients in response to 3nM MCP-1, but this effect was not significant (Figure 28). Similar results were obtained when cells were stimulated with 1nM MCP-1 (not shown) (compare to Figure 27B). These experiments indicate that the metalloproteinases ADAM10 and ADAM17 do not play a crucial role in the induction of intracellular calcium-transients by MCP-1.

74

8000

6000

4000

in rel. light in units rel. light 2000 increase flourescenceincrease of 0 I G GW DMSO Inhibitors

Figure 28: MCP-1 induced Ca2+-response of THP1 cells treated with the ADAM inhibitors GI254023 and GW282406. ADAM inhibition of wt THP1 with the specific inhibitor a GI254023 (10µM) and GW282406 (10µM) did not result in changed Ca2+-influx after stimulation with 3nM MCP-1 compared to the DMSO solvent control. The data are shown as mean ±SD, n=3.

Influence of the gene silencing of ADAM8, 9, 10 and 17 on the Ca2+- response of LV-THP1

The influence of the mRNA knockdown of ADAM proteases on calcium-transients in MCP-1 stimulated THP1 cells was analysed next. The stable ADAM8, 9, 10 or 17-KD THP1 cell lines (see Figure 18 and Figure 19) were assayed for calcium-transients in response to 3 nM MCP- 1. Cells treated with scramble control shRNA were investigated in parallel (Figure 29).

In comparison to the scramble control, neither ADAM8, ADAM9, ADAM10 nor ADAM17-KD LV-THP1 cells showed a significantly altered calcium response to MCP-1 stimulation.

Altogether, the inhibition as well as the mRNA knockdown experiments, indicated that ADAMs do not influence the calcium response in THP1 cells. Therefore, the observed migratory deficits of THP1 cells upon inhibition or silencing of ADAM10 and 8 cannot be explained by an alteration of the Ca2+-response towards MCP-1.

75

A 5000 B 5000

4000 4000

3000 3000

2000 2000 in rel. light in units rel. light in rel. light in units light rel. 1000 1000 increase increase flourescence of increase of increase flourescence 0 0 r 7 cr 03 31 45 sc s 9 8 6 8- -1 -2 A A9-196 A9-2260 A8 A8

LV-THP1-KD-lines LV-THP1-KD-lines

C 8000 D 5000

4000 6000 3000 4000 2000

in rel. light units rel. light in 2000 in rel. light units rel. light in 1000 increase of increase flourescence increase of increase flourescence 0 0 0 scr 947 -65 1 scr 0 -2061 -2642 A1 A10- 17 A17 A LV-THP1-KD-lines LV-THP1-KD-lines

Figure 29: Ca2+-response of LV-THP1 with gene silencing of ADAM8, 9, 10 or 17. Ca2+-influx of ADAM8-KD LV-THP1 (A), ADAM9-KD LV-THP1 (B), ADAM10-KD LV-THP1 (C) or ADAM17- KD LV-THP1 (D) after MCP-1 stimulation (3nM). The data are shown as mean ±SD, n=3.

3.8. Influence of ADAM8, 9, 10 and 17 knock-down on adhesion of THP1 cells to VCAM-1 and fibronectin under static conditions

The Ca2+-influx experiments showed that the investigated ADAMs do not influence MCP-1 induced cell migration on the level of intracellular Ca2+-transients. To gain more insight in the involvement of the ADAM proteinases in the process of leukocyte migration, adhesion experiments under static conditions were performed with the THP1 cells.

As substrates for adhesion VCAM-1, ICAM-1 and fibronectin were chosen. These substrates play a crucial role in cell attachment during leukocyte migration and transmigration. Fibronectin is a component of the and bound by integrins that are

76

expressed on THP1 cells. ICAM-1 and VCAM-1 are expressed on endothelial cells and are bound by the integrins LFA1 and VLA4 of leukocytes, respectively. THP1 cells have been found to adhere under static conditions to VCAM-1 and fibronectin but not to ICAM-1166. In the mentioned study, adherence to both, VCAM-1 and fibronectin, was enhanced by stimulation with MCP-1. BSA coated wells served as control.

Adhesion of THP1 cells to V-CAM and fibronectin under static conditions

The experimental setup of the adhesion assay of Ashida et al.166 was adapted to examine whether the inhibition or the gene silencing of ADAM8, 9, 10 or 17 affected the adhesion of THP1. Unstimulated and MCP-1 stimulated wt THP1 were investigated for adhesion to cell culture plates coated with VCAM-1, ICAM-1, fibronectin and BSA (Figure 30). In accordance to Ashida and co-workers it was observed that THP1 adhered tightly to VCAM-1 and

80

60

40

20 % adherent cells %adherent

1 0

1 1 1 -1 im P- P CP- CP- stim M MC + + + MC n in no M-1 no st M-1 no stim SA no stim M-1 + M M-1 cti ect B A A A A e BSA C IC on V VC IC on fibr fibr

adhesion substrates and stimulations

Figure 30: Adhesion of THP1 to VCAM1, ICAM-1 and fibronectin under static conditions with and without the stimulation by 3nM MCP-1. THP1 cells in absence or presence of 3nM MCP-1 were added wells of a 96-well plates that were coated with VCAM-1, ICAM-1 and fibronectin. BSA coated wells served as negative control. After washing the well two times and lysis of cells, the cell count was determined using the glucoronidase assay. The data are shown as mean ±SD, n=3.

77

fibronectin but not to VCAM-1 and the BSA control. In contrast to the mentioned study, the stimulation with MCP-1 did not result in adhesion changes of the THP1 cells.

The influence of the ADAM inhibitors GI254023 and GW282406 on the adhesion of wt THP1 was examined next. Compared to the solvent control, neither GI254023 nor GW282406 affected adhesion of wt THP1 to VCAM-1 (Figure 31A) or fibronectin (Figure 31B). As seen before, almost no adhesion to ICAM-1 and the BSA control was observed (data not shown). The same observation was made when cells were not stimulated with MCP-1 (compare to Figure 30).

A 80 B 80

60 60

40 40

20 20 % adherent cells % adherent % adherent cells %adherent

0 0 I GI O G GW S GW DMSO DM

Inhibitors Inhibitors

Figure 31: Static adhesion of wt THP1 treated with the ADAM inhibitors GI254023 and GW282406. A: Static adhesion of wt THP1 to VCAM-1 under the pharmaceutical inhibition of ADAM10 (GI254023) and ADAM10 and 17 (GW282406). Each inhibitor was used at a concentration of 10µM. DMSO (0.1%) served as solvent control. B: Adhesion of inhibitor treated THP1 to fibronectin. The data depicted in figures A and B show the level of adhesion with a stimulation of MCP-1 at 3nM. Unstimulated THP1 adhered at equally rates to both substrates. The percentages indicate the number of adherent cells compared to the number of cells that were initially added to the wells containing the immobilized adhesion substrates. The data are shown as mean ±SD, n=3.

To confirm these results and to include ADAM8 and ADAM9, the effect of ADAM8, 9, 10 or 17 gene silencing on adhesion of THP1 cells was analysed. Similar to the results with GI254023 and GW282406 treated wt THP1 cells (Figure 31), no significant effect of the gene silencing of either ADAM-proteinase on the adhesion to VCAM-1 (Figure 32A) or fibronectin (Figure 32B) was observed.

Combining the results of the inhibition experiments with GI254023 and GW282406 and the mRNA silencing experiments, an influence of ADAM8, 9, 10 or 17 on the static adhesion of THP1 to VCAM-1 and fibronectin could be excluded.

78

A 100 B 100

80 80

60 60

40 40

20 % of adherent cells adherent % of 20 % of adherent of %cells adherent

0 0 0 03 31 67 60 5 47 61 r 3 1 0 1 scr 9 9 2 6 9 0 c 0 3 18 645 - s 9 8 967 947 -2 0 -2642 - 1 65 1 06 A8- 8- 8 9-2 0-1 7-2 7 8 -1 -2 A A A9-1 A A1 1 -1 A 8 7 A A A1 A A8-2645A9- A9-2260A10- A10- A-1 A17-2642 LV-THP1-KD-lines LV-THP1-KD-lines

Figure 32: Static adhesion of LV-THP1 with gene silencing of ADMA8, 9 10 or 17. Static adhesion of ADAM-KD LV-THP1 to immobilized VCAM-1 (A) or fibronectin (B). The data depicted in figures A and B show the level of adhesion after stimulation of MCP-1 at 3nM. Unstimulated THP1 adhered at equal rates to both substrates irrespective of either ADAM8, 9, 10 or 17 gene silencing. Adherence is expressed as percentage of total number of cells that were initially added to each wells containing the immobilized adhesion substrates. The data are shown as mean ±SD, n=3.

3.9. Role of ADAM8, 9, 10 and 17 gene silencing on epithelial cell migration

The transmigration and chemotaxis assays had demonstrated that gene silencing of ADAM8 and 10 profoundly reduces chemotaxis and transmigration of THP1 cells, whereas silencing of ADAM9 and 17 showed no comparable effects on the migratory abilities. To this end, only the migration of leukocytic cells had been investigated. To answer the question whether the observed effects also account for non leukocytic cells, migration experiments with an adherent epithelial cell line were performed.

A common in vitro assay, to study the migration of epithelial or endothelial cells, is the so called “scratch assay”. A confluent cell layer is injured by a defined scratch and the closure of the gap in the cell layer by the immigrating cells is monitored over time by microscopy. This assay is used as a model for the wound closure process of endothelial or epithelial tissues.

Scratch assays were performed with the epithelial cell line ECV304 to study the role of ADAM8, 9, 10 and 17. The scratch assays were conducted with cells which had been stably transduced with the ADAM8, 9, 10 or 17 pLVTHM lentivirus, respectively (see Figure 13 to

79

Figure 17). Figure 33 shows the microscopic picture of the migrated ADAM8-KD LV-ECV 304 in comparison to the scramble control cells just after the scratch and 24h after the migratory process had started. In all experiments additional photos of the scratch were taken 2h, 4h and 8h after the scratch (data not shown).

Figure 33: Microscopic images of the scratch assay performed with ADAM8-KD LV-ECV304 and scramble control cells. Directly after the scratch was made, the defined gap in the confluent cell layer had sharp edges (0h). 24h after the scratch, the process of gap closure is clearly detectable when compared to the original size of the scratch indicated by the red square. While the gap of the scramble control ECV304 cell layer is almost closed, fewer cells have immigrated into the gap of the two different ECV304 cell lines with the gene silencing of ADAM8.

The migration behaviour of the different ECV304 cell lines with gene silencing of ADAM8, 9, 10 or 17 was evaluated by counting the cells that had invaded a defined area of the gap within the cell layer (red squares in Figure 33). The results of all scratch assays with different cell lines are summarised in Figure 34.

Compared to the scramble control, ADAM8-KD LV-ECV 304 showed a significant migration deficit in the scratch assay (Figure 34A). By contrast, ADAM9-KD and ADAM-17 LV-ECV 304 showed no significant difference in their wound healing properties compared to the scramble control cells (Figure 34B and D). These findings are in accordance to the chemotaxis and transmigration experiments showing that the migratory response of THP1 cells depends on ADAM8 but not on ADAM9 and only to minor extend on ADAM17.

80

A significant migratory impairment of ADAM10 silencing on wound closure of ECV304 cells in the scratch assay could not be detected (Figure 34C). While there was still a trend of a reduced migration rate this did not result in a significant difference. This is in contrast to the results of the chemotaxis or transmigration experiments performed with THP1 cells that had demonstrated a crucial role of ADAM10 for leukocytic cell migration. Thus, the contribution of ADAM10 to cell migration can be different, depending on the type of cell and its migratory response.

A 1.5 B 1.5

1.0 1.0 * ** (rel. scr) to 0.5 (rel.to scr) 0.5 migration index migration migration index migration

0.0 0.0

45 scr scr -1831 -26 -1967 A8 A8 A9 A9-2260 LV-ECV-KD-lines LV-ECV-KD-lines

C 1.5 D 1.5

1.0 1.0

(rel. to scr) to (rel. 0.5 (rel. to scr) 0.5 migration index migration migration index migration

0.0 0.0 r 0 7 c 1 2 5 4 s 06 64 scr -6 0 -19 -2 -2 1 17 A A10 A A17 LV-ECV-KD-lines LV-ECV-KD-lines

Figure 34: Summary of the scratch assays performed with LV-ECV 304 with a gene silencing of ADAM8, 9, 10 and 17. A: Number of migrated ADAM8-KD LV-ECV304 compared to scramble control cells. B: Scratch assay results of ADAM9-KD LV-ECV 304. C: Results of wound closure of LV-ECV 304 with a gene silencing of ADAM10. D: Performance of ADAM17-KD LV-ECV-304 in the scratch assay. To be able to compare the performance of the different LV-ECV 304 cell lines in the scratch assay, the number of immigrated cells in the defined area of the gap after 8h were subtracted from the number of cells immigrated into the same area after 24h. The data is shown as migration index relative to number of immigrated scramble control cells. The data of all panels are shown as mean ±SD, n=6, *p<0.05, **p<0.01, compared to scramble control.

81

3.10. Remarks

The original data analysed and described by Figure 7, Figure 17, Figure 33 and Figure 34 were generated during the diploma thesis of Henriette Alert171 which was directly supervised by the author of the present PhD thesis, F. Martin Hess. All mentioned experiments were designed by F. Hess and separately evaluated for this thesis. H. Alert expressed her explicit consent to use the mentioned data for this PhD thesis.

82

4. Discussion

Leukocyte recruitment and subsequent transendothelial migration is one of the most critical hallmarks in inflammation. It has been shown that endothelial ADAM1015,16 and ADAM1714,66 influence endothelial permeability and thereby leukocyte diapedesis. In vivo studies with conditional transgenic mice showed that neutrophil influx was strongly reduced when leukocytes lacked ADAM1761–63. Recently it was reported that the absence of ADAM8 in leukocytes reduces the number of infiltrating leukocytic cells in an experimental asthma model131 and that PSGL-1 shedding by ADAM8 regulates the rolling of neutrophils on endothelium133. While it has been shown that ADAM10 is essential for lymphoid development109,113, there is no published study to date on the direct influence of leukocytic- expressed ADAM10 in immune cell recruitment and migration .

In this study, the influence of leukocytic-expressed ADAM8, 9, 10 and 17 on leukocyte migration was investigated by using ADAM-specific inhibitors and gene silencing of the four ADAM proteases. As a model, the monocytic cell line THP1 was used and tested for in vitro transendothelial migration and chemotaxis in response to the chemokine MCP-1. To our knowledge this is the first study directly comparing the influence of all four mentioned ADAM proteases in the same set up of migration assays.

The migration experiments revealed that directed transendothelial migration and THP1 cell chemotaxis in response to MCP-1 was significantly reduced by the pharmaceutical inhibition of ADAM10 and gene silencing of ADAM8 or 10, suggesting that these proteases are critically involved in cell migration. By contrast, gene silencing of ADAM17 resulted in only a small reduction of THP1 cell migration and the combined inhibition of ADAM10 and ADAM17 did not decrease migration further than inhibition of ADAM10 alone. This indicated ADAM17’s minor role in cell migration, while ADAM9 does not seem to be required at all as shown by the silencing of the latter protease. As demonstrated by calcium flux experiments and adhesion assays, the role of ADAM8 and ADAM10 in THP1 cell migration cannot be explained by the changed sensitivity of THP1 cells towards MCP-1 or by decreased static adhesion of THP1 cells to VCAM-1, ICAM-1 and fibronectin. Notably, the importance of ADAM8 and probably, of ADAM10, is not limited to leukocytic cells but also accounts for adherent cells such as ECV304 cells that were investigated for adhesion in scratch wound healing assays.

83

Table 9: Summary table of the outcome of the performed functional assays of this study. For experiments performed with cell lines which had been lentiviral transduced with shRNA sequences against a ADAM protease, one symbol indicates the results of one of the two or three different cell lines with the gene silencing of the mentioned ADAM.

Differential expression of ADAMs in different cell culture lines

The expression of mRNA transcripts of the four metalloproteases ADAM8, 9, 10 and 17 was analysed in seven cultured or primary cell lines. These cell lines have been used in previous studies that looked into the influence of ADAM10 and/or ADAM17 at various stages of inflammatory processes and leukocyte recruitment12,14,15,21,66,107,172. However, an analysis by quantitative real-time PCR showed that all examined cell lines expressed the four proteases with remarkable and unexpected differences in the expression profiles.

The two epithelial cancer cell lines ECV304 (human bladder carcinoma cell line) and A549 (human alveolar basal epithelial adenocarcinoma cell line) differed mostly in their ADAM8 expression. ADAM8 overexpression has been shown in many solid tumours (see Introduction, for a review of expression of ADAMs in different tumours see173) and is a marker of poor prognosis and increased invasiveness. Since both cell lines stem from different epithelia it is difficult to say whether their differences in ADAM expression stem from this cause or whether the tumour lines acquired the difference in during their

84

tumourigenesis. Nor can it be excluded that their ADAM expression changed during their time as cell culture lines: A549 have been used in research since 1972174 and ECV304 since 1970175,176.

The difference in expression of ADAM10 and 17 in the primary endothelial cell HMVEC and HUVEC was very striking. HMVEC, which are derived from microvessels of the adult lung, had comparatively high expression levels of ADAM10 and 17. By contrast, HUVEC, which were isolated from the human umbilical vein, showed low mRNA levels of these proteases. It is tempting to attribute theses differences in expression level to the micro- and macrovascular origin of both primary endothelial cell types. Since both vascular environments fulfil different purposes it is conceivable that vascular endothelial cells have different expression patterns of ADAM proteases. But one should also take into account that both cell lines were isolated from their respective donors at extremely different ages and stages of development, which could also affect the expression patterns of ADAM10 and 17.

Differential expression of ADAM10 and 17 was also observed for the two leukocytic cell lines THP1 (primary human acute monocytic leukaemia cells) and Jurkat (primary human T cell leukaemia cells). While both cell lines show comparatively high levels of ADAM8 and low levels of ADAM9, THP1 cells express both ADAM10 and 17 at high levels while Jurkat cells are amongst the cell lines with the lowest mRNA levels of both proteases. That is line with the proposition that ADAM8, 10 and 17 may fulfil different essential functions in the recruitment of cells of different leukocytic lineage4. A recent study showed that ADAM17 knockout leads to an increased adhesion of neutrophils in the in vivo cremaster model. However, the adhesion and recruitment of ADAM17-null monocytes did not exhibit any alterations in this study177.

Both cell lines, THP1 and Jurkat cells, are derived from leukemic patients. Similar to the epithelial cell lines ECV304 and A549, the question remains whether the changes in ADAM expression resemble their monocytic and lymphoid lineage or whether they are caused by their oncogenesis. Both cell lines have been widely used in research since the late 1970 and it is equally possible that their ADAM expression profile changed over the course of their cultivation163,178.

However, the mRNA expression analysis showed that THP1 cells were the ideal cell line to study the role of leukocytic-expressed ADAM proteases in the migration of leukocytes. They express comparative high level of ADAM8, ADAM10 and ADAM17 and the cells have been used in various studies as a model cell line for in vitro assays examining monocyte recruitment and transendothelial migration164–166,179.

85

Endothelial cells support THP1 cell migration

After the optimal MCP-1 concentration for directed THP1 migration in the in vitro transendothelial transmigration had been found, the influence of the endothelial layer in the migration process was analysed. Surprisingly, the migration rates of THP1 cells through the endothelial layer were more than threefold higher than in the chemotaxis assays in the absence of the endothelial layer when the cells migrated towards MCP-1. Apparently the endothelial layer does not constitute an additional barrier migrating THP1 cells have to overcome. Instead, the endothelial cells obviously assist the leukocytes in their transmigration.

One possible explanation for this finding could be that the endothelial cells bind MCP-1 from the lower compartment and present the chemokine on their surface to the THP1 cells in the upper compartment of the transwell insert. The local high concentration could lead to enhanced transmigration. There are several endothelial proteins which are known to take part in this process. For example, it has been shown that the Duffy antigen receptor for chemokines (DARC) transported MCP-1 through a endothelial barrier in a very similar experimental setup180. Proteoglycans like the proteins of the syndecan-family are known to present chemokines and other inflammatory modulators on the surface of endothelial cells and thereby act as coreceptors for leukocytic chemokine receptors181.

Another possible explanation could be that endothelial adhesion molecules and complex changes in the endothelial cell morphology actively support leukocyte transmigration22,182. It has been shown that the binding of leukocytic integrins, such as LFA-1 and VLA-1, to endothelial ICAM-1 and VCAM-1 leads to their clustering on the apical endothelial plasmamembrane25,26. This initiates a release of intracellular calcium and the activation of intracellular signalling molecules such as src-kinases and the GTPases RhoA and Rac-1. The signal cascade results in the polymerisation of actin near the plasmamembrane and a contraction of the endothelial cells183–186. Due to the proteolysis of adhesion molecules such as VE-Cadherin and JAM-A by ADAM proteases14,15 and phosphorylation of VE-Cadherin187, the inter-cellular endothelial junctions become permissive188. Signalling events also trigger the transport of vesicles from the lateral border recycling compartment (LBRC) to the lateral plasmamembrane at the cellular junction of the endothelial cells. The vesicles fuse to the plasmamembrane189,190 and phosphorylated PECAM – which is stored in the vesicles of the LBRC – is inserted into the junctional plasmamembrane191. The phosphorylated PECAM undergoes a homophilic interaction with leukocytic PECAM which facilitates the movement of the leucocytes through the endothelial junctions192.

86

The latter explanation that the endothelial cells are actively engaged in the diapedesis of the leukocytes fits the finding that the undirected migration was 10 times higher in the presence of an endothelial layer than in its absence. It is conceivable that consequences of both processes – the transport and presentation of chemokines by the endothelial cells and the active assistance of the endothelial cells in the migration – contribute to the increased migration in presence of endothelial cells.

Pharmaceutical inhibition of ADAM10 inhibits transendothelial migration and THP1 cell chemotaxis

The in vitro transendothelial migration of THP1 cells towards MCP-1 was significantly inhibited when THP1 cells were pretreated before the migration experiment with GI254023, which preferentially inhibits ADAM10. Pretreatment of the leukocytic cells with GW282406, which inhibits both ADAM10 and 17, did not result in a significantly reduced migration of the THP1 cells. Since this experiment sought only to examine the role of leukocytic-expressed ADAMs, the transmigration experiment itself was performed in absence of the inhibitors.

In their study showing the importance of ADAM10 shedding of endothelial-expressed VE- Cadherin for endothelial permeability and leukocyte transendothelial migration, Schulz et al. performed similar experiments15. The authors investigated migration of activated primary T- cells through HUVEC in the presence of GI254023 or GW282406. In their experiments, the use of both inhibitors resulted in a significant and more pronounced reduction in transmigration. Schulz and co-workers attributed the reduced migration to the inhibited shedding of VE-Cadherin.

The reduced effect of the inhibitors on the transendothelial migration of THP1 cells which was observed in this study can be attributed to the fact that the inhibitors were used for the preincubation of THP1 cells only and removed prior to the migration assay. Since GI254023 and GW282406 are reversible, it is conceivable that during the time course of the incubation of the transmigration experiment the inhibitors diffused off the ADAM proteases. The proteases thereby regained there full proteolytic activity during the experiment which led to the inhibitors’ reduced effect on migration.

Clear inhibition was observed in THP1 cell chemotaxis experiments without endothelial cells. In these experiments the inhibitors were present throughout the assay. ADAM10 inhibition by GI254023X profoundly decreased migration. Combined inhibition of ADAM17 and ADAM10 by GW2280264 did not result in an additional reduction of the migration rate compared to

87

GI254023. This indicates that ADAM10 inhibition alone is responsible for the observed effect and that ADAM17 is not essentially involved in the directed chemotaxis of THP1 cells.

The chemotaxis assays were performed in the absence of potential protease substrates such as endothelial adhesion proteins or components of the ECM, such as matrigel, collagen or fibronectin. These are commonly used in invasion assays of tumour cells. Donners et al. made similar observations in their study regarding the influence of the VEGF-receptor (VEGF- R) on the migration of endothelial cells158. In one experiment of their study, Donners et al. looked at GI254023’s effect on the migration of primary monocytes towards fMLP (N-formyl- methionine-leucine-phenylalanine) in a similar chemotaxis assay. Monocyte chemotaxis was inhibited in the presence of GI254023 although the observed effect was not as strong as in the experiments of the present study. The authors of the mentioned study hypothesized that ADAM10 may shed adhesion molecules which would facilitate the detachment of the monocytes. However, since the VEGF-R shedding of ADAM10 was the main aspect of their study the authors did not look into the role of monocytic ADAM10 in detail.

Taken together, there is a strong indication that THP1 cell expressed ADAM10 plays a crucial role in migration. ADAM10 seems be to either involved in the interaction with leukocytic adhesion molecules which are needed for the chemotaxis or in signalling processes in THP1 cells which effect their migration. To gain more insight into this problem and to complement the inhibitor experiments, further experiments had to be performed with cells with ADAM10 and 17 transcriptional inhibition.

ShRNA knockdown of ADAMs

To further analyse the function of leukocytic-expressed ADAMs in leukocyte diapedesis and chemotaxis, a lentiviral-based shRNA approach for the gene silencing of ADAM8, 9 10 and 17 was successfully established. The lentiviral transduction allowed targeting of cell lines and primary cells for the gene transfer of shRNA sequences which are generally inaccessible for standard transfection methods. The shRNA sequences designed against each target protease silenced their mRNA and protein expression with high efficiency. In this study the successful gene silencing of ADAM8, 9, 10 and 17 could be shown for HUVEC, ECV304 cells and THP1 cells. Unspecific silencing effects of the ADAM-specific shRNAs against non-target proteases were excluded; this underlines the experimental value of this method. That the gene knock- down of ADAM10 and 17 effect the shedding functions of both proteases was proven in

88

shedding experiments with CX3CL1, which is a well-described substrate for both proteases67,107.

Since it enables targeting only one of the cell types, the established gene silencing system is a highly useful tool in experimental systems such as in vitro transendothelial migration, where several different cell types interact. Pharmaceutical inhibitors are of limited use in these experiments since it is nearly impossible to determine which cell type is responsible for the observed effect of the inhibitors. Another problem with pharmaceutical inhibitors is their specificity. For example, has it been shown that GI254023 inhibits ADAM10 preferential over ADAM17 but it remains to be investigated if any other of the remaining 15 mammalian proteolytic active ADAMs is inhibited by this compound. However, combining both approaches (transcriptional silencing and pharmaceutical inhibition) is advisable for studies such as this one, since both approaches complement each other.

Lentiviral-delivered shRNA sequences which suppress gene expression of ADAM proteases have become important for studying the function of ADAM10 and ADAM17. Transgenic mice lacking either proteases die during embryogenesis or shortly after birth60,87 and conditional transgenic mice with tissue-specific deletions of ADAM10 or 17 have only recently become available16,61,66,113. Before the conditional transgenic mice were established, mouse embryonic fibroblasts (MEF cells) isolated from ADAM10 or 17-deficient transgenic embryos have been valuable tools to analyse the shedding of substrates in both proteases60,67,87,106. However, MEF cells are only of limited use when the functions of ADAMs in a specific cell type, such as endothelial cells or cells of different leukocytic lineages, are the topic of investigation. These cell types can now be transduced with the lentiviral shRNA vectors and the efficiency of the transduction can be easily monitored thanks to the reporter gene GFP which is used in the lentiviral system established in this study. The lentiviral shRNA system has been used in three additional studies of our group, where it was used alongside ADAM specific inhibitors and tissue specific ADAM17-null mice as well21,66,107.

shRNA KD of ADAM10 confirms the results of the inhibitor studies

The migration experiments performed with lentiviral-transduced THP1 cells (LV-THP1) with a ADAM10 gene silencing confirmed the findings of the migration experiments performed with the ADAM specific inhibitors. Both ADAM10-KD LV-THP1 cell lines migrated at a significantly lower rate than the scramble control cell in the directed transendothelial migration. The

89

random transendothelial migration of ADAM10-KD LV-THP1 cells in absence of MCP-1 was reduced as well.

The same effect was observed in the chemotaxis assay where the ADAM10-KD cells only migrated through polycarbonate filter membranes. The relative migration rates of the ADAM10-KD LV-THP1 compared to the scramble control cells was very similar in the transmigration and chemotaxis assay. However, when looking at the total number of transmigrated cells, it was again apparent that three times as many cells migrated in the transendothelial migration than in the chemotaxis assays. This observation was independent of the ADAM10 silencing.

In their study regarding ADAM10 shedding of VE-Cadherin and its influence on leukocyte recruitment, Schulz and co-workers performed transmigration experiments in the presence of GI254023 (see above) and with T-cells transfected with siRNA against ADAM1015. In both experimental set-ups, ADAM10 inhibition resulted in a reduced migration of the T-Cells. The authors speculated that a possible trans-cleavage of VE-Cadherin by ADAM10 could be responsible for this effect, as other studies had shown that ADAM10 is capable of cleaving substrate in trans92.

However, the findings of the present study indicate ADAM10’s more general involvement in leukocyte migration, in addition to its VE-Cadherin shedding. The migration rates of cells with ADAM10 gene silencing were reduced in transendothelial migration and in chemotaxis experiments where no endothelial cells and thereby no VE-Cadherin were present.

ADAM8 is needed for THP1 cell chemotaxis and transmigration

Migration experiments performed with ADAM8-KD LV-THP1 cell lines showed that in both transmigration and chemotaxis, ADAM8-deficient cells migrated at lower rates than the scramble control cells. The inhibitory effect of the ADAM8 silencing was even more pronounced than that of the ADAM10 gene silencing.

ADAM8 is not essential for the organism since ADAM8-deficient mice are fully viable121. However, Naus and co-workers have shown that these mice are protected from developing experimental asthma131. The role of leukocytic ADAM8 was highlighted by the findings that wt mice which had been transplanted with ADAM8-/- bone marrow showed reduced asthma symptoms and a reduced number of infiltrating leukocytes in the lung131. Naus et al. performed several separate in vitro migration assays of ADAM8-/- macrophages, dendritic cells and T-cells in which the cells migrated through matrigel. The authors managed to show 90

that the migration capabilities of all the analysed leukocytic subsets were negatively affected compared to wt control cells. The authors hypothesized that ADAM8 either is important for the digestion of components of the ECM or that ADAM8 is involved in activation cascades or signalling events. As a third option, Naus et al. pointed to a possible role of the disintegrin/cystin-rich domain of ADAM8 which may be important for cell-ECM attachment.

The group of Frederico Díaz-González could show that ADAM8 is responsible for the cleavage of L-selectin in neutrophils119. Leukocytic L-selectin is important in the first step of the rolling process of leukocytes, where it interacts with sulphated carbohydrates of GlyCam-1 and CD34 expressed on endothelial cells. Inhibition of L-selectin shedding leads to a reduction of rolling velocity and increases neutrophil adhesion to the endothelium193. In a very recent study the same group showed via immunoprecipitation that leukocytic ADAM8 interacts with the P-selectin glycoprotein ligand-1 (PSGL-1) through activated ezrin-radixin-moesin actin- binding proteins (ERM)133. PSGL-1 binds to endothelial P- and E-selectins and is involved in tethering and rolling leukocytes on the endothelial cells. The interaction of PSGL-1 with ADAM8 results in the shedding of the adhesion molecule. The authors showed with in vitro flow adhesion experiments that ADAM8 gene silencing leads to an increased rolling of the neutrophil-like cell line HL-60.

In the present PhD thesis the migration experiments ADAM8-KD LV-THP1 were performed under static conditions and the migration deficit of cells lacking ADAM8 was also observed in the chemotaxis assays in the absence of L-selectin and PSGL-1 binding partners. However, the finding that ADAM8 might interact with adhesion molecules via adaptor molecules such as ERM is intriguing. It is possible that ADAM8 is needed for the assembly of multiprotein complexes which serve as anchorage for the actin cytoskeleton of the leukocytes. It is conceivable that the interaction between ADAM8’s cytoplasmic tail and adaptor proteins such as ERM are responsible for this. Equally, the disintegrin domain of ADAM8 and ADAM10 could play a crucial role in a similar clustering process on the extracellular side of the plasma membrane.

ADAM9 gene silencing in THP1 cells does not result in migratory inhibition

The results of the transmigration and chemotaxis experiments with ADAM9-silenced LV-THP1 clearly showed that ADAM9 knockdown did not result in an inhibition of THP1 cell migration. One of the two ADAM9-KD LV-THP1 cell lines (ADAM9-2260) actually migrated at

91

significantly higher rates in the transendothelial migration assay than the scramble control cells. Even in the chemotaxis assay the migration rates of this cell line was elevated, though not statistically significant. Since the increase in migration was only observed for one of two shRNA sequences it cannot be concluded from these experiments that ADAM9 gene silencing increases the migratory abilities of THP1 cells. However, the findings show no inhibitory influence of the ADAM9-KD.

The results of the migration experiments accord with a study which analysed wound healing in ADAM9-/- mice155. Although wounds healed faster in ADAM9-/- mice, the authors of the study did not observe increased numbers of infiltrating leukocytes in the wound area. One explanation for the increased migration of one of the ADAM9-KD LV-THP1 could be the ADAM10 shedding by ADAM9, which was reported by two different groups141,194. The findings presented by both studies indicate more of a modulating role of ADAM10 shedding by ADAM9. Henriette Alert’s diploma thesis, which was supervised by the author of the present study, showed that ADAM9 gene silencing in ECV304 cells did not result in changes of ADAM10 surface expression171.

shRNA KD of ADAM17 effects chemotaxis and transmigration of THP1 cells only slightly

The transmigration and chemotaxis experiments performed with THP1 cells with a gene silencing of ADAM17 showed that ADAM17 is not as important as ADAM8 or ADAM10 for the THP1 cell migration. Only one of the two THP1 cell lines with ADAM17 gene silencing migrated at a slightly but significantly lower rate in the transendothelial migration experiments than the scramble control cells. In the chemotaxis assay both ADAM17-KD cell lines showed no significantly reduced migration. The migration experiments performed with the inhibitor GW282406 had already shown that ADAM17 may be of minor importance in the migration of THP1 compared to ADAM10. The combined inhibition of ADAM17 and ADAM10 did not result in an additional inhibition of the migration compared to the inhibition of ADAM10 alone by GI254023. Thus, ADAM17 does not seem to play a crucial role in the migration set-up of the present study.

Several studies have shown that mice with a conditional deletion of ADAM17 in leukocytes had a significantly reduced neutrophil infiltration and better survival rates compared to wt mice in LPS-induced sepsis models or e. coli-mediated peritonitis61–63. This was mostly attributed to the reduced TNFα shedding by the ADAM17-/- leukocytes, which resulted in

92

reduced levels of inflammatory cytokines IL-1β and IL-6 as well. The inhibition of L-selectin shedding in the ADAM17 leukocytes was made responsible for the observed, significantly earlier neutrophil recruitment. According to the authors of the study, the increased L-selectin expression led to tighter adhesion and mediated faster recruitment of the neutrophils.

A very recent study by Tang et al. confirmed the relevance of the increased L-selectin expression for the significantly earlier recruitment of ADAM17-/- neutrophils in a sterile peritonitis model and showed that these neutrophils exhibited slower rolling velocities in the in vivo cremaster model177. Tang and co-workers were able to reduce the accelerated recruitment of ADAM17-/- neutrophils by applying a blocking antibody against L-selectin; this reduced the recruitment of wt and ADAM17-/- neutrophils to the same level. However, the authors also found that the recruitment of ADAM17-/- monocytes was not altered by the gene knockout even though they exhibited higher L-selectin levels as well. The authors concluded that L-selectin plays a crucial role in neutrophil recruitment but it is of minor relevance for the recruitment of monocytes.

The results obtained in this thesis from the ADAM17-KD LV-THP1 cells are in line with the finding of Tang et al.. THP1 cells resemble a cell culture model for monocytes and it is possible that the same experiments with primary ADAM17 deficient neutrophils from transgenic mice or with different cell lines such as Jurkat cells (cell culture model for T-cells) or HL-60 (cell culture model for neutrophils) would have resulted in a different outcome.

Role of ADAMs in the sensitivity of LV-THP1 cells towards different MCP-1 concentrations

One possible explanation for the migratory deficits of THP1 cells with a gene silencing of ADAM8 or ADAM10 in the chemotaxis assay could have been that the THP1 cells undergo changes in their sensitivity towards MCP-1 when ADAM8 or ADAM10 is silenced. To test this hypothesis, the transmigration of LV-THP1 with a gene silencing of ADAM8, 9, 10 or 17 was tested using different MCP-1 concentrations ranging from 0nM to 30nM.

As expected, ADAM8 and ADAM10 silencing led to decreased migration, but no significant shift of the dose response curve towards MCP-1 was observed when ADAM8 or ADAM10 gene-silenced THP1 cells were compared with the control. While ADAM9 deficient THP1 cells showed higher migration rates than scramble control cells at all tested concentrations, ADAM17-KD LV-THP1 cells migrated at about the same rate as the control cells.

93

Thus, the inhibitory effect of the ADAM8 and ADAM10 gene silencing is independent of the concentration of the chemotactic stimulus MCP-1. These results show that both proteases do not influence the THP1 migration through a modulation of the concentration-dependent sensitivity of THP1 cells for MCP-1.

No influence of the gene silencing of ADAM8, 9, 10 or 17 on the MCP- 1 induced Ca2+-transients in THP1 cells

It was shown that the inhibitory effect of gene silencing of ADAM8 and 10 on the transendothelial migration of THP1 cells was independent of the concentration of the chemotactic stimulus MCP-1. However, the question remained whether the knockdown of both proteases effects the signal transduction of THP1 cells in response to MCP-1 in general. To answer this question, the Ca2+-transients of THP1 cells were analysed in response to MCP-1.

The Ca2+-transient experiments showed that neither the inhibition of THP1 cells with GI254023 or GW282406 nor the gene silencing of ADAM8, 9, 10 or 17 resulted in a significant change in the Ca2+-release in the THP1 cells after the MCP-1 stimulation. These findings ruled out that the observed migratory inhibitions of inhibitor-treated THP1 cells and THP1 cells with a gene silencing of ADAM8 and 10 were caused by an alteration in the direct signal transduction of the MCP-1 receptor CCR2. However, the possibility remains that the gene silencing of ADAM8 or 10 effects the arrestin-mediated internalization of CCR2 or interferes with the complex signalling downstream of the Ca2+-transients in which kinases such as ERK or MAPK are involved195

The adhesion of THP1 cells to VCAM-1 and fibronectin is not altered by pharmaceutical inhibition of ADAM10 and 17 or by gene silencing of ADAM8, 9, 10 or 17

The tight adhesion of leukocytes to the endothelial cells in the vasculature is the step in leukocyte recruitment that directly precedes their diapedesis. Tight adhesion of leukocytes is chiefly mediated by the interaction of leukocytic integrins LFA-1 (αLβ2 integrin) and VLA-4

(α4β1 integrin) to the endothelial adhesion molecules ICAM-1 and VCAM-1, respectively.

Fibronectin, an important component of the ECM, is bound by the integrin heterodimers α5β1 and VLA-4166. To test whether the pharmaceutical inhibition of ADAM 10 and 17 or the gene 94

silencing of ADAM8, 9, 10 or 17 changes THP1 cell adhesion, adhesion assays under static conditions were performed using the recombinant substrates VCAM-1, ICAM-1 and fibronectin.

The adhesion assays showed that neither the pharmaceutical inhibition nor the gene silencing of the target proteases resulted in changes to the adhesion of THP1 to VCAM-1, ICAM-1 or fibronectin. Accordingly to a study by Ashida et al.166 THP1 adhered strongly to VCAM-1 and fibronectin while only a very weak binding was observed for ICAM-1. In the present thesis, however, it could not be confirmed that the stimulation of THP1 cells with MCP-1 results in a stronger adhesion of the cells to any of the substrates. The unchanged adhesion of THP1 cells after MCP-1 stimulation was independent of ADAM inhibition or gene silencing. Therefore, the results of the adhesion assays indicate that the investigated ADAM proteases are not involved in modulating the interaction of either VLA-4 to VCAM-1 or that of

α5β1-integrin to fibronectin in THP1 cells.

No differences were observed in the adhesion of the THP1 cells to ICAM-1 in the experimental setup used in this study and therefore no influence of the investigated proteases on the activity of LFA-1 (αLβ2 integrin) was found. However, there are accumulating indications for the shedding of leukocytic β2 integrin in leukocytic migration. It is quite possible that the chosen experimental setup did not fully reflect the role of LFA-1 in chemotaxis and transmigration, since it is reported that LFA-1 is only transiently activated during migration.

Weber et al. showed in their study that β2-integrin is needed for the adhesion of PMA- stimulated primary monocytes to ICAM-1. In their study, the use of a blocking antibody against β2-integrin inhibited the in vitro transendothelial migration of the stimulated monocytes196. The authors of this study found that differential activation kinetics of different integrin-β2 heterodimers are responsible for the enhanced adhesion of monocytes after SDF- 1 (CXCL12) stimulation. While the SDF-1 stimulation resulted in a sustained activation of

Mac-1 (αMβ2 integrin), Weber et al. reported a transient activation of monocytic LFA-1 (αLβ2 integrin) after SDF-1 stimulation.

There are several reports that indicate a critical role of integrin shedding during the process of leukocyte recruitment. It was shown that cleavage of the αM integrin of Mac-1 (αMβ2 integrin) by the serin proteases elastase, proteinase-3, and is needed for the

197 successful detachment of neutrophils during chemotaxis . β2 integrin shedding has been observed on infiltrating neutrophils in patients exhibiting a cutaneous inflammatory response to the blistering agent cantharidin198. A proteomics study which compared the shed proteins

95

of wt macrophages to macrophages expressing constitutively active MMP-9 ( 9) identified β2 integrin as one of the major molecules shed by the MMP-9 overexpressing macrophages199.

A very recent publication by Elaine Raines’s group reported that β2 integrin shedding was needed for macrophage efflux after intraperitoneal stimulation with thioglycollate200. The authors showed that β2 integrin shedding could be inhibited with the broad spectrum zinc- metalloproteinase inhibitor GM6001. This resulted in a reduced efflux of the peritoneal macrophages. In comparison, macrophages lacking β2 integirn migrated faster out of the peritoneal cavity than control macrophages. The same effect was observed when the authors used an inhibiting antibody against β2 integrin in experiments with wt mice. However, the authors detected the same levels of soluble β2 integrin in transgenic animals with a complete genetic deletion of MMP-9 and in transgenic mice with myeloid ADAM10 or hematopoietic ADAM17 deficiency as in control animals. Thereby they ruled out that any of the three proteases is responsible for the in vivo β2 integrin shedding of macrophages. They attributed differences with their previous study to the different in vivo and in vitro situation. While several enzymes appear to shed the same substrates in vitro there is accumulating evidence that in vivo shedding is cell- and context specific.

ADAM8 gene silencing affects ECV304 cell migration

To analyse whether the gene silencing of ADAM8 or 10 also affects the migration of other cells such as epithelial cells, in vitro wound healing assays were performed with lentiviral- transduced ECV304 cells with gene silencing of ADAM8, 9, 10 or 17. The results of the wound healing assays showed that only the ADAM8 gene silencing resulted in a significantly reduced migration of the ECV304 cells. While there was also a reduced migration in one of two cell lines with ADAM10 gene silencing, the inhibitory effect was not significant.

The findings of the in vitro wound healing assay indicate that ADAM8 could play a general role in the migration processes of various cells. Reports showing that ADAM8 overexpression is related with invasiveness and metastasis in cancer back up the finding of the inhibition on migration after gene silencing of ADAM8122–126. While it seems that ADAM10 is not as essential as ADAM8 for the motility of the cancer cell line ECV304, it was reported that ADAM10 contributes to the migration of endothelial cell lines. In the already-mentioned study by Donners and co-workers, it was demonstrated that ADAM10 inhibition not only

96

resulted in strongly reduced VEGF-R shedding but in a reduced migration of endothelial and monocytic cells as well158.

It would be interesting to analyse and compare keratinocytic cell lines with gene silencing of ADAM8, 9, 10 or 17 in the in vitro wound healing assay. A faster wound healing – attributed to increased keratinocyte motility – was observed in transgeneic mice lacking ADAM9 expression155. It is conceivable that the shRNA knockdown of ADAM9 in keratinocytic cell lines could lead to a similar effect as observed in the ADAM9-null mice.

4.1. Conclusion

This study shows that ADAM8 and 10 have significant influence on the transendothelial migration and chemotaxis of the monocytic cell line THP1. Notwithstanding, the precise molecular mechanism by which the two proteases influence the migration of the monocytic cell lines could not be identified. It could be excluded that ADAM8 or 10 interfere with the direct MCP-1 signalling of THP1 and with their adhesion to the adhesion molecules VCAM-1 and fibronectin. The observed migration deficit was independent of the presence or absence of an endothelial cell layer. Therefore, it seems likely that both proteases interfere with a process in the THP1 cells themselves and that shedding events such as cleavage of endothelial adhesion molecules in trans are not responsible for the observed effect. Two possible mechanisms by which both ADAM proteases could influence the migration of the THP1 cells could either be the shedding or clustering of adhesion molecules or the shedding of growth factors which lead to an autokrine stimulation of the cells.

Possible shedding or clustering of adhesion molecules by ADAM8 or ADAM10

A very recent report showed that β2 integrin is cleaved by a protease inhibited by the broad 200 spectrum inhibitor GM6001 . The role of integrin heterodimers containing β2 integrin, such as LFA-1 or Mac-1, in leukocyte chemotaxis and transmigration is widely accepted201,202.

ADAM8 or ADAM10 expressed by THP1 cells could play a crucial role in shedding β2 integrin that is bound to one of its endothelial substrates, thereby releasing the cell from the endothelial cells it is attached to (Figure 35). It is conceivable that one of the two proteases or even both contribute to the disassembly of focal adhesion of migrating cells by integrin shedding. This would also explain why THP1 cells with a gene silencing of ADAM8 or 10

97

showed migration deficits in the chemotaxis assay in the absence of endothelial cells and components of the ECM. The reduced shedding of other adhesion molecules such as CD44 and PSGL-1, which are cleaved by ADAM10109,203 and ADAM8133, respectively, could be the reason for the migratory deficits of THP1 cells with a gene silencing of ADAM10 or 8.

Figure 35: Hypothesis 1: Shedding of integrins as possible contribution of ADAM8 or 10 to the migration of THP1 cells. Proteolytic cleavage of integrins could facilitate the release of migrating THP1 cells from adhesion molecules on endothelial cells. It is equally conceivable that ADAM8 or 10 could be essential for dissolving focal adhesions of migrating cells. That would explain why the gene silencing of ADAM8 or 10 inhibited the migration of the THP1 cells in the chemotaxis assay in the absence of endothelial cells.

Although Gomez et al. did not find elevated levels of β2 integrin in mice with a myeloid 200 ADAM10 deletion , it is possible that ADAM10 is able to cleave the β2 integrin of the THP1 cells in vitro or that ADAM8 is responsible for cleaving β2 integrin. To investigate whether β2 integrin shedding is effected by ADAM8 or ADAM10, the supernatants of ADAM-KD LV-THP1 could be analysed by ELISA or western blotting against β2 integrin. Moreover, the surface staining of β2 integrin might be analysed on ADAM-KD LV-THP1. Another option would be to analyse the migration of ADAM8-KD or ADAM10-KD LV-THP1 in the presence of an inhibiting antibody against β2 integrin.

Due to the prominent role of ADAM proteases in the shedding of various surface molecules, it is often overlooked that they can interact with integrins via their extracellular disintegrin-

98

domain. It is possible that either ADAM8 or 10 or even both proteases are involved in forming clusters of integrins on the plasma membrane of adhering THP1 cells. Such clusters of integrins are essential for the outside-in signalling of integrins and contribute to the formation of signalosomes at the plasma membrane which are required for the recruitment of protein tyrosine kinases201,204.

Another possible interaction interface of ADAM8 or 10 with integrins or other adhesion molecules needed for migration (besides their extracellular disintegrin domain) could be their cytoplasmic tail domains. Domínguez-Luis and co-workers nicely managed to show that proteins of the ezrin-radixin-moesin family (ERM proteins) function as adaptor molecules between ADAM8 and its shedding substrate PSGL-1133. ERM proteins link integral membrane proteins to the cortical actin cytoskeleton and have a regulatory effect on Rho GTPases205,206. It can be envisaged that the absence of either protease could result in a reduced recruitment of adaptor proteins, such as ERM proteins, to the leading edge of the migrating cells with the consequence of an affected Rho GTPase signalling and changes in the actin cytoskeleton.

Lastly, the interaction of ADAM10 with integrins in membrane microdomains organized by proteins of the tetraspanin family could provide an explanation for the observed migratory deficits of THP1 cells with a gene silencing of ADAM10 or 8. These microdomains are thought to play crucial roles in leukocyte functions such as activation, motility and antigen presentation207. It has been shown that ADAM10, together with many integrins and other proteases, is recruited to such membrane microdomains, which themselves are stabilized by tetraspanins208–210. Interestingly, it was reported that ADAM17, in contrast to ADAM10, is recruited to membrane microdomains composed of different tetraspanins211,212.

However, the observed inhibitory effect of the ADAM specific inhibitors on the migration of THP1 cells raises the question of how small molecules can influence this potential cluster- forming function of ADAMs and whether this is the exclusive mechanism by which ADAM10 contributes to the migration of THP1 cells. Nevertheless, it could be possible that the inhibitor which binds to ADAM10’s catalytic cleft holds the protease in a conformation state which inhibits its interaction via its disintegrin domain. In addition, the inactive conformation of the catalytic domain with the bound inhibitor could possibly interfere with protein interactions of the cytoplasmic tail. Such a mechanism has been reported for the extracellular domains of integrins which are involved in both outside-in and inside-out signalling202,213.

99

A B

C

Figure 36: Hypothesis 2: Interaction of ADAM8 or 10 with integrins and possible other adhesion molecules is needed for the formation of high molecular clusters at the plasmamembrane of migrating THP1 cells. There are several possibilities as to how ADAM8 or 10 could contribute to the formation of such high molecular clusters. A: The ADAM proteases could bind to integrins with their disintegrin-like domain and thereby facilitate the formation of multi-molecular clusters. B: For ADAM8, it has been shown that it interacts with adhesion molecules via adaptor proteins of the ERM-family. It is possible that such interactions are needed for the efficient migration of THP1 cells. C: Lastly, there is the possibility that ADAM10 or 8 are needed for the stabilization of membrane-microdomains which use tetraspanins as scaffold proteins.

100

To analyse whether ADAM8 or 10 participate in the formation of multi-molecular clusters or complexes on the plasma membrane, it would be feasible to perform immunoprecipitation (Co-IP) with antibodies against ADAM8 and 10. Analysing the protein content of such complexes with mass spectroscopy could be very helpful to identify potential binding partners of the ADAM proteins.

Another possibility to analyse further the role of ADAM8 and 10 in the migration of leukocytic cells could be to monitor migrating THP1 cells with high-resolution live cell microcopy, using actin tagged to fluorescent proteins to monitor the actin dynamics during migration. This would enable deciding whether the gene silencing of ADAM8 or 10 affects the actin assembly at the leading edge, or the formation of podosomes of (trans)-migrating THP1 cells or whether the cells are unable to resolve their adhesion to the substrate they are migrating on.

It could well be that the proposed clustering function of ADAM8 or 10 is not the only mechanism by which they contribute to the migration of leukocytic cells but that the insertion of ADAMs into such multi-protein complexes is essential for their substrate recognition or to be in proximity to their substrates, respectively. That the interaction between ADAMs and integrins is a topic worth investigating was shown by a very recent publication. In this study the authors were able to show that in kidney carcinoma cells, the binding of β1-integrin to ADAM17 suppresses its G-coupled receptor regulated activity214.

ADAM8 or 10 could be needed for the autokrine stimulation of THP1 cells

ADAM8 or 10 may contribute to the migration of THP1 cells by shedding stimulatory growth factors which induce a migratory phenotype of the cells by auto- or parakrine stimulation. Members of the EGF-like growth factor family could be possible shedding substrates of ADAM10 or 8 which are involved in such a stimulation of THP1 cells. EGF-like growth factors are expressed as transmembrane proforms and undergo proteolytic cleavage to become mature soluble factors215. EGF-like growth factors bind to receptors of the ErbB family, which induces their dimerization and triggers intracellular signalling cascades. ADAM10 is the major sheddase of EGF and betacellulin and has been implicated in the shedding of HB-EGF96,216. For ADAM8, it has been shown that it is capable of cleaving TGFα117 and that its overexpression leads to an increased shedding of neuregulin-1β2217.

101

Figure 37: Hypothesis 3: ADAM8 or 10 contribute to the migration of THP1 cells by shedding ligands of receptors of the ErbB family, leading to an auto- or parakrine stimulation which promotes a migratory phenotype of the monocytic cells.

To investigate further as to whether the inhibited shedding of ErbB ligands affects THP1 cell migration, supernatants of THP1 cells with a gene silencing of ADAM8 or 10 could be analysed for their levels of soluble EGF-like growth factors. The relevance of shed growth factors could be determined by performing transmigration experiments with ADAM8 or ADAM10 THP1 cells in the presence of supernatants from wtTHP1 cells or by supplementation with recombinant growth factors. Furthermore, THP1 cell transmigration could be performed in the presence of an EGF-receptor inhibitor and compared to cell migration in the presence of GI254023. Since EGF-R (ErbB1) is the most prominent member of the ErbB family and binds to many EGF-like factors (EGF, TGFα, amphiregulin, HB-EGF, epiregulin and betacellulin218), it constitutes the first logical target for such experiments.

Lastly, there is the possibility that Notch cleavage by ADAM10 and its signalling could be responsible for promoting a migratory phenotype in THP1 cells. It has been shown that Notch shedding by ADAM10 is essential for lymphoid development. However, there is still controversy regarding the influence of Notch signalling in the development and differentiation of myeloid cells109,113. However, the outcome of the migration experiments in presence of the ADAM inhibitors argues against this hypothesis. Since Notch signalling mainly effects the transcription of its target genes, it would take several hours until the inhibition of ADAM10 would show its effect on Notch signalling. In contrast, the inhibitors induced a strong inhibition of migration of the THP1 cells within three hours.

102

It is possible that ADAM8 and ADAM10 influence the migration of the THP1 cells via the same mechanism but it is also possible that both proteases act on the migration of THP1 via completely different mechanisms. Moreover, ADAM8 or ADAM10 could affect migration of the monocytic cells via more than one of the above-discussed molecular mechanisms and the combined effect of several mechanisms could then lead to the observed migratory deficits.

In addition to the above-suggested in vitro experiments, the analysis and comparison of leukocyte migration in transgenic mice with specific deletion of ADAM10 or 17 could be very helpful to elucidate ADAM10’s role in leukocyte migration. While several studies have analysed the in vivo recruitment of ADAM17-/- leukocytes61–63,177, there are still no published studies which look into the recruitment of leukocytes in transgenic mice with a leukocyte- specific deletion of ADAM10. Fortunately, conditional transgenic mice with a deletion of ADAM10 in all myeloid cells are about to become available in our research group. Additionally, mice which have a leukocyte specific knockout of ADAM17 are currently being bred in our laboratory. Both transgenic mice strains are viable and it will be interesting to compare the in vivo migration of ADAM10-/- to ADAM17-/- leukocytes. Additional insight could be gained by including ADAM8-/- mice in this study.

103

5. Summary

Proteases of the ADAM family are involved in the proteolytic shedding of various cell surface molecules. Thereby, they fulfil essential functions in embryogenesis, development, neuro- protection and cancer development. ADAM proteases also participate in all stages of inflammation, starting from early-response cytokine release and leukocyte recruitment to the process of tissue repair and damaged tissue revascularisation.

The transendothelial migration of leukocytes from the blood into the inflamed tissue is an important hallmark of inflammatory responses to infections or trauma. This migratory process is dependent on well-orchestrated responses of both endothelial and leukocytic cells. Endothelial ADAM10 and 17 have been shown to regulate endothelial permeability and thereby leukocyte transendothelial migration in inflammatory settings. In vivo studies demonstrated the importance of leukocytic-expressed ADAM8 and ADAM17 in the recruitment of leukocytes. While ADAM10 is important in the development of lymphoid cells, its role on leukocytes during transendothelial migration has yet to be fully elucidated.

Gene expression analysis of ADAM8, 9, 10 and 17 in several cell lines revealed distinct expression profiles of the four proteases in each of the examined cell lines. Due its high expression levels of ADAM8, 10 and 17 in the monocytic cell line THP1 it was decided to use this cell line to investigate the influence of the four ADAMs in the in vitro migration of leukocytes. In the present study the directed migration of THP1 cells to the chemotactic stimulus of the chemokine MCP-1 was analysed in the presence of endothelial cells (in vitro transendothelial migration) and in the absence of endothelial cells (in vitro chemotaxis).

The use of pharmaceutical ADAM inhibitors in in vitro transendothelial migration experiments revealed that transendothelial migration of THP1 cells was inhibited in the presence of an ADAM10 specific inhibitor. The same effect of the ADAM10 inhibitor was also observed the for the directed chemotaxis in the absence of potential endothelial substrates of ADAM10. The combined pharmaceutical inhibition of both ADAM10 and ADAM17 did not further suppress leukocyte migration in either transendothelial migration or chemotaxis. This indicates that ADAM17 is not as important as ADAM10 in the migration of THP1 cells.

To confirm these findings and further analyse the responsible mechanism, a lentiviral shRNA system for the transcriptional inhibition of ADAM8, 9, 10 and 17 was established. The knockdown of each protease was confirmed by flow cytometry and/or RT-PCR, respectively. The comparison of THP1 cells with a gene silencing of ADAM8, 9, 10 or 17 in transendothelial migration and chemotaxis experiments revealed that ADAM8 and ADAM10

104

are critically involved in THP1 cell migration. By contrast, ADAM9 expression did not seem to be required and ADAM17 was of minor importance for the migration of THP1 cells.

Subsequent experiments, which were performed to investigate a potential effect of ADAM8 and ADAM10 on signal transduction, showed that the sensitivity of THP1 cells towards the chemotactic stimulus MCP-1 was unaffected by the gene silencing of the investigated ADAM proteases. The analysis of the THP1 calcium response revealed that neither the pharmaceutical inhibition of ADAM10 and 17 nor the gene silencing of ADAM8, 9, 10 or 17 affected THP1 calcium transients after stimulation with MCP-1. Adhesion experiments were performed to analyse the adhesion of the monocytic cells under pharmaceutical inhibition or the gene silencing of the investigated proteases. In leukocyte recruitment, the adhesion of leukocytes directly precedes the transendothelial migration, and so an affect of ADAM8 or ADAM10 in this process could have explained their role during migration. However, the adhesion of THP1 cells to the recombinant integrin ligands VCAM-1, ICAM-1 and fibronection was not influenced by the pharmaceutical inhibition or the gene silencing of the investigated proteases.

Since neither the direct calcium signal in response to MCP-1 nor the adhesion of THP1 cells were affected by ADAM8 or ADAM10 gene silencing or by ADAM pharmaceutical inhibition, it appeared that both proteases are involved in a more general mechanism affecting cell migration. Indeed, ADAM8 gene silencing in the epithelial cancer cell line ECV304 resulted in a reduced migration of the cells in an in vitro wound-healing assay. Although ADAM10 knock- down resulted in a reduction of cell migration, the observed effect was not significant.

This is the first study to analyse and directly compare the role of the four proteases ADAM8, 9, 10 and 17 in the migration of leukocytic cells. It shows that ADAM8 and ADAM10 are crucial for in vitro transendothelial migration and chemotaxis of the monocytic cell line THP1. Furthermore, ADAM8 is critically involved in the migration of adherent epithelial cells in vitro. This study has identified the previously unknown importance of ADAM8 and ADAM10 in cell migration and therefore provides the basis for further analysis of the molecular mechanism by which ADAM8 and ADAM10 contribute to leukocyte migration. In-depth investigations are proposed, addressing the possibility that ADAM10 and ADAM8 directly interact and may also shed adhesion molecules required for the organization of the cytoskeleton during cell migration. These studies should also investigate the shedding of EGF-like growth factors by ADAM8 or ADAM10 which could promote a migratory phenotype of the cells via autokrine stimulation.

105

6. Literature

1. Scheller, J., Chalaris, A., Garbers, C. & Rose-John, S. ADAM17: a molecular switch to control inflammation and tissue regeneration. Trends in immunology 32, 380-7 (2011).

2. Pruessmeyer, J. & Ludwig, A. The good, the bad and the ugly substrates for ADAM10 and ADAM17 in brain pathology, inflammation and cancer. Seminars in cell & developmental biology 20, 164-74 (2009).

3. Dreymueller, D., Pruessmeyer, J., Groth, E. & Ludwig, A. The role of ADAM-mediated shedding in vascular biology. European journal of cell biology (2011).doi:10.1016/j.ejcb.2011.09.003

4. Zarbock, A. & Rossaint, J. Regulating inflammation: ADAM8--a new player in the game. European journal of immunology 41, 3419-22 (2011).

5. Black, R.A. et al. A metalloproteinase disintegrin that releases tumour-necrosis factor- alpha from cells. Nature 385, 729-33 (1997).

6. Murdoch, C. & Finn, A. Chemokine receptors and their role in inflammation and infectious diseases. Blood 95, 3032-43 (2000).

7. Springer, T.A. Traffic signals for lymphocyte recirculation and leukocyte emigration: The multistep paradigm. Cell 76, 301-314 (1994).

8. Baggiolini, M., Walz, A. & Kunkel, S.L. Neutrophil-activating -1/interleukin 8, a novel cytokine that activates neutrophils. The Journal of clinical investigation 84, 1045-9 (1989).

9. Leonard, E.J. & Yoshimura, T. Human monocyte chemoattractant protein-1 (MCP-1). Immunology today 11, 97-101 (1990).

10. Charo, I.F. & Peters, W. Chemokine receptor 2 (CCR2) in atherosclerosis, infectious diseases, and regulation of T-cell polarization. Microcirculation (New York, N.Y. : 1994) 10, 259-64 (2003).

11. Zernecke, A. & Weber, C. Inflammatory mediators in atherosclerotic vascular disease. Basic research in cardiology 100, 93-101 (2005).

12. Hundhausen, C. et al. Regulated shedding of transmembrane chemokines by the disintegrin and metalloproteinase 10 facilitates detachment of adherent leukocytes. Journal of immunology (Baltimore, Md. : 1950) 178, 8064-72 (2007).

13. Hundhausen, C. et al. The disintegrin-like metalloproteinase ADAM10 is involved in constitutive cleavage of CX3CL1 (fractalkine) and regulates CX3CL1-mediated cell-cell adhesion. Blood 102, 1186-95 (2003).

14. Koenen, R.R. et al. Regulated release and functional modulation of junctional adhesion molecule A by disintegrin metalloproteinases. Blood 113, 4799-809 (2009).

106

15. Schulz, B. et al. ADAM10 regulates endothelial permeability and T-Cell transmigration by proteolysis of vascular endothelial cadherin. Circulation research 102, 1192-201 (2008).

16. Inoshima, I. et al. A Staphylococcus aureus pore-forming toxin subverts the activity of ADAM10 to cause lethal infection in mice. Nature medicine 17, 1310-4 (2011).

17. Li, Q., Park, P.W., Wilson, C.L. & Parks, W.C. Matrilysin shedding of syndecan-1 regulates chemokine mobilization and transepithelial efflux of neutrophils in acute lung injury. Cell 111, 635-46 (2002).

18. Manon-Jensen, T., Itoh, Y. & Couchman, J.R. Proteoglycans in health and disease: the multiple roles of syndecan shedding. The FEBS journal 277, 3876-89 (2010).

19. Charnaux, N. et al. RANTES (CCL5) induces a CCR5-dependent accelerated shedding of syndecan-1 (CD138) and syndecan-4 from HeLa cells and forms complexes with the shed ectodomains of these proteoglycans as well as with those of CD44. Glycobiology 15, 119-30 (2005).

20. Charnaux, N. et al. Syndecan-4 is a signaling molecule for stromal cell-derived factor-1 (SDF-1)/ CXCL12. The FEBS journal 272, 1937-51 (2005).

21. Pruessmeyer, J. et al. A disintegrin and metalloproteinase 17 (ADAM17) mediates inflammation-induced shedding of syndecan-1 and -4 by lung epithelial cells. The Journal of biological chemistry 285, 555-64 (2010).

22. Muller, W.A. Mechanisms of leukocyte transendothelial migration. Annual review of pathology 6, 323-44 (2011).

23. Engelhardt, B. & Wolburg, H. Mini-review: Transendothelial migration of leukocytes: through the front door or around the side of the house? European journal of immunology 34, 2955-63 (2004).

24. Carman, C.V. et al. Transcellular diapedesis is initiated by invasive podosomes. Immunity 26, 784-97 (2007).

25. Barreiro, O. et al. Dynamic interaction of VCAM-1 and ICAM-1 with moesin and ezrin in a novel endothelial docking structure for adherent leukocytes. The Journal of cell biology 157, 1233-45 (2002).

26. Carman, C.V. & Springer, T.A. A transmigratory cup in leukocyte diapedesis both through individual vascular endothelial cells and between them. The Journal of cell biology 167, 377-88 (2004).

27. Del Maschio, A. et al. Leukocyte recruitment in the cerebrospinal fluid of mice with experimental meningitis is inhibited by an antibody to junctional adhesion molecule (JAM). The Journal of experimental medicine 190, 1351-6 (1999).

28. Rawlings, N.D. & Barrett, A.J. Evolutionary families of metallopeptidases. Methods in enzymology 248, 183-228 (1995).

107

29. Stöcker, W. & Bode, W. Structural features of a superfamily of zinc-: the metzincins. Current opinion in structural biology 5, 383-90 (1995).

30. Gomis-Rüth, F.X. Catalytic domain architecture of metzincin metalloproteases. The Journal of biological chemistry 284, 15353-7 (2009).

31. Bode, W., Gomis-Rüth, F.X. & Stöckler, W. Astacins, serralysins, snake venom and matrix metalloproteinases exhibit identical zinc-binding environments (HEXXHXXGXXH and Met-turn) and topologies and should be grouped into a common family, the “metzincins”. FEBS letters 331, 134-40 (1993).

32. Murphy, G. The ADAMs: signalling scissors in the tumour microenvironment. Nature reviews. Cancer 8, 929-41 (2008).

33. van Goor, H., Melenhorst, W.B.W.H., Turner, A.J. & Holgate, S.T. Adamalysins in biology and disease. The Journal of pathology 219, 277-86 (2009).

34. Kuno, K. et al. Molecular cloning of a gene encoding a new type of metalloproteinase- disintegrin family protein with thrombospondin motifs as an inflammation associated gene. The Journal of biological chemistry 272, 556-62 (1997).

35. Blobel, C.P. et al. A potential fusion peptide and an integrin ligand domain in a protein active in -egg fusion. Nature 356, 248-52 (1992).

36. Wolfsberg, T.G. et al. The precursor region of a protein active in sperm-egg fusion contains a metalloprotease and a disintegrin domain: structural, functional, and evolutionary implications. Proceedings of the National Academy of Sciences of the United States of America 90, 10783-7 (1993).

37. Duffy, M.J., McKiernan, E., O’Donovan, N. & McGowan, P.M. Role of ADAMs in cancer formation and progression. Clinical cancer research : an official journal of the American Association for Cancer Research 15, 1140-4 (2009).

38. White, J.M. ADAMs: modulators of cell-cell and cell-matrix interactions. Current opinion in cell biology 15, 598-606 (2003).

39. Anders, A., Gilbert, S., Garten, W., Postina, R. & Fahrenholz, F. Regulation of the alpha-secretase ADAM10 by its prodomain and proprotein convertases. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 15, 1837-9 (2001).

40. Buckley, C.A. et al. Amino-terminal TACE prodomain attenuates TNFR2 cleavage independently of the cysteine switch. American journal of physiology. Lung cellular and molecular physiology 288, L1132-8 (2005).

41. Wewer, U.M. et al. ADAM12 is a four-leafed clover: the excised prodomain remains bound to the mature enzyme. The Journal of biological chemistry 281, 9418-22 (2006).

42. Iba, K. et al. The cysteine-rich domain of human ADAM 12 supports cell adhesion through syndecans and triggers signaling events that lead to beta1 integrin-dependent cell spreading. The Journal of cell biology 149, 1143-56 (2000).

108

43. Flannery, C.R. et al. Autocatalytic cleavage of ADAMTS-4 (-1) reveals multiple glycosaminoglycan-binding sites. The Journal of biological chemistry 277, 42775-80 (2002).

44. Gaultier, A., Cousin, H., Darribère, T. & Alfandari, D. ADAM13 disintegrin and cysteine- rich domains bind to the second heparin-binding domain of fibronectin. The Journal of biological chemistry 277, 23336-44 (2002).

45. Smith, K.M. et al. The cysteine-rich domain regulates ADAM protease function in vivo. The Journal of cell biology 159, 893-902 (2002).

46. Cousin, H., Gaultier, A., Bleux, C., Darribère, T. & Alfandari, D. PACSIN2 is a regulator of the metalloprotease/disintegrin ADAM13. Developmental biology 227, 197-210 (2000).

47. Serrano, S.M.T., Jia, L.-G., Wang, D., Shannon, J.D. & Fox, J.W. Function of the cysteine-rich domain of the haemorrhagic metalloproteinase : targeting adhesion proteins collagen I and von Willebrand factor. The Biochemical journal 391, 69-76 (2005).

48. Seals, D.F. & Courtneidge, S.A. The ADAMs family of metalloproteases: multidomain proteins with multiple functions. Genes & development 17, 7-30 (2003).

49. Lu, Q., Clemetson, J.M. & Clemetson, K.J. Snake venoms and hemostasis. Journal of thrombosis and haemostasis : JTH 3, 1791-9 (2005).

50. Nath, D. et al. Interaction of metargidin (ADAM-15) with alphavbeta3 and alpha5beta1 integrins on different haemopoietic cells. Journal of cell science 112 ( Pt 4, 579-87 (1999).

51. Reiss, K., Ludwig, A. & Saftig, P. Breaking up the tie: disintegrin-like metalloproteinases as regulators of cell migration in inflammation and invasion. Pharmacology & therapeutics 111, 985-1006 (2006).

52. Eto, K. et al. Functional classification of ADAMs based on a conserved motif for binding to integrin alpha 9beta 1: implications for sperm-egg binding and other cell interactions. The Journal of biological chemistry 277, 17804-10 (2002).

53. Arribas, J., Bech-Serra, J.J. & Santiago-Josefat, B. ADAMs, cell migration and cancer. Cancer metastasis reviews 25, 57-68 (2006).

54. Huang, J., Bridges, L.C. & White, J.M. Selective modulation of integrin-mediated cell migration by distinct ADAM family members. Molecular biology of the cell 16, 4982-91 (2005).

55. Mazzocca, A. et al. A secreted form of ADAM9 promotes carcinoma invasion through tumour-stromal interactions. Cancer research 65, 4728-38 (2005).

56. Nath, D. et al. Meltrin gamma(ADAM-9) mediates cellular adhesion through alpha(6)beta(1 )integrin, leading to a marked induction of fibroblast cell motility. Journal of cell science 113 ( Pt 1, 2319-28 (2000).

109

57. Soond, S.M., Everson, B., Riches, D.W.H. & Murphy, G. ERK-mediated phosphorylation of Thr735 in TNFalpha-converting enzyme and its potential role in TACE protein trafficking. Journal of cell science 118, 2371-80 (2005).

58. Werb, Z. CELL BIOLOGY:Enhanced: A Cellular Striptease Act. Science 282, 1279-1280 (1998).

59. Pruessmeyer, J. & Ludwig, A. The good, the bad and the ugly substrates for ADAM10 and ADAM17 in brain pathology, inflammation and cancer. Seminars in cell & developmental biology 20, 164-74 (2009).

60. Peschon, J.J. et al. An essential role for ectodomain shedding in mammalian development. Science (New York, N.Y.) 282, 1281-4 (1998).

61. Horiuchi, K. et al. Cutting edge: TNF-alpha-converting enzyme (TACE/ADAM17) inactivation in mouse myeloid cells prevents lethality from endotoxin shock. Journal of immunology (Baltimore, Md. : 1950) 179, 2686-9 (2007).

62. Long, C., Wang, Y., Herrera, A.H., Horiuchi, K. & Walcheck, B. In vivo role of leukocyte ADAM17 in the inflammatory and host responses during E. coli-mediated peritonitis. Journal of leukocyte biology 87, 1097-101 (2010).

63. Arndt, P.G. et al. Leukocyte ADAM17 regulates acute pulmonary inflammation. PloS one 6, e19938 (2011).

64. Bell, J.H., Herrera, A.H., Li, Y. & Walcheck, B. Role of ADAM17 in the ectodomain shedding of TNF-alpha and its receptors by neutrophils and macrophages. Journal of leukocyte biology 82, 173-6 (2007).

65. Li, Y., Brazzell, J., Herrera, A. & Walcheck, B. ADAM17 deficiency by mature neutrophils has differential effects on L-selectin shedding. Blood 108, 2275-9 (2006).

66. Dreymueller, D. et al. Lung endothelial ADAM17 regulates the acute inflammatory response to lipopolysaccharide. EMBO Molecular Medicine Epub ahead, (2012).

67. Ludwig, A. et al. Metalloproteinase inhibitors for the disintegrin-like metalloproteinases ADAM10 and ADAM17 that differentially block constitutive and phorbol ester-inducible shedding of cell surface molecules. Combinatorial chemistry & high throughput screening 8, 161-71 (2005).

68. Garton, K.J. et al. Tumour necrosis factor-alpha-converting enzyme (ADAM17) mediates the cleavage and shedding of fractalkine (CX3CL1). The Journal of biological chemistry 276, 37993-8001 (2001).

69. Horiuchi, K. et al. Ectodomain shedding of FLT3 ligand is mediated by TNF-alpha converting enzyme. Journal of immunology (Baltimore, Md. : 1950) 182, 7408-14 (2009).

70. Weskamp, G. et al. Pathological neovascularization is reduced by inactivation of ADAM17 in endothelial cells but not in pericytes. Circulation research 106, 932-40 (2010).

110

71. Reddy, P. et al. Functional analysis of the domain structure of tumour necrosis factor- alpha converting enzyme. The Journal of biological chemistry 275, 14608-14 (2000).

72. Althoff, K., Reddy, P., Voltz, N., Rose-John, S. & Müllberg, J. Shedding of interleukin-6 receptor and tumour necrosis factor alpha. Contribution of the stalk sequence to the cleavage pattern of transmembrane proteins. European journal of biochemistry / FEBS 267, 2624-31 (2000).

73. Tsakadze, N.L. et al. Tumour necrosis factor-alpha-converting enzyme (TACE/ADAM- 17) mediates the ectodomain cleavage of intercellular adhesion molecule-1 (ICAM-1). The Journal of biological chemistry 281, 3157-64 (2006).

74. Garton, K.J. et al. Stimulated shedding of vascular cell adhesion molecule 1 (VCAM-1) is mediated by tumour necrosis factor-alpha-converting enzyme (ADAM 17). The Journal of biological chemistry 278, 37459-64 (2003).

75. Stoeck, A. et al. A role for exosomes in the constitutive and stimulus-induced ectodomain cleavage of L1 and CD44. The Biochemical journal 393, 609-18 (2006).

76. Xu, P. & Derynck, R. Direct activation of TACE-mediated ectodomain shedding by p38 MAP kinase regulates EGF receptor-dependent cell proliferation. Molecular cell 37, 551-66 (2010).

77. Swendeman, S. et al. VEGF-A stimulates ADAM17-dependent shedding of VEGFR2 and crosstalk between VEGFR2 and ERK signaling. Circulation research 103, 916-8 (2008).

78. Lemjabbar-Alaoui, H., Sidhu, S.S., Mengistab, A., Gallup, M. & Basbaum, C. TACE/ADAM-17 phosphorylation by PKC-epsilon mediates premalignant changes in tobacco smoke-exposed lung cells. PloS one 6, e17489 (2011).

79. Göoz, M., Göoz, P., Luttrell, L.M. & Raymond, J.R. 5-HT2A receptor induces ERK phosphorylation and proliferation through ADAM-17 tumour necrosis factor-alpha- converting enzyme (TACE) activation and heparin-bound epidermal growth factor-like growth factor (HB-EGF) shedding in mesangial cells. The Journal of biological chemistry 281, 21004-12 (2006).

80. Amour, A. et al. TNF-alpha converting enzyme (TACE) is inhibited by TIMP-3. FEBS letters 435, 39-44 (1998).

81. Yang, P., Baker, K.A. & Hagg, T. The ADAMs family: coordinators of nervous system development, plasticity and repair. Progress in neurobiology 79, 73-94 (2006).

82. Lammich, S. et al. Constitutive and regulated alpha-secretase cleavage of Alzheimer’s amyloid precursor protein by a disintegrin metalloprotease. Proceedings of the National Academy of Sciences of the United States of America 96, 3922-7 (1999).

83. Duffy, M.J. et al. The ADAMs family of proteases: new biomarkers and therapeutic targets for cancer? Clinical proteomics 8, 9 (2011).

84. Buxbaum, J.D. et al. Evidence that tumour necrosis factor alpha converting enzyme is involved in regulated alpha-secretase cleavage of the Alzheimer amyloid protein precursor. The Journal of biological chemistry 273, 27765-7 (1998).

111

85. Matthews, V. et al. Cellular cholesterol depletion triggers shedding of the human interleukin-6 receptor by ADAM10 and ADAM17 (TACE). The Journal of biological chemistry 278, 38829-39 (2003).

86. Hikita, A. et al. Involvement of a disintegrin and metalloproteinase 10 and 17 in shedding of tumour necrosis factor-alpha. Biochemistry and cell biology = Biochimie et biologie cellulaire 87, 581-93 (2009).

87. Hartmann, D. et al. The disintegrin/metalloprotease ADAM 10 is essential for Notch signalling but not for alpha-secretase activity in fibroblasts. Human molecular genetics 11, 2615-24 (2002).

88. Krebs, L.T. et al. Notch signaling is essential for vascular morphogenesis in mice. Genes & development 14, 1343-52 (2000).

89. Klein, T. & Bischoff, R. Active metalloproteases of the A Disintegrin and Metalloprotease (ADAM) family: biological function and structure. Journal of proteome research 10, 17-33 (2011).

90. Sotillos, S., Roch, F. & Campuzano, S. The metalloprotease-disintegrin Kuzbanian participates in Notch activation during growth and patterning of Drosophila imaginal discs. Development (Cambridge, England) 124, 4769-79 (1997).

91. Hattori, M., Osterfield, M. & Flanagan, J.G. Regulated cleavage of a contact-mediated axon repellent. Science (New York, N.Y.) 289, 1360-5 (2000).

92. Janes, P.W. et al. Adam meets Eph: an ADAM substrate recognition module acts as a molecular switch for ephrin cleavage in trans. Cell 123, 291-304 (2005).

93. Moss, M.L., Stoeck, A., Yan, W. & Dempsey, P.J. ADAM10 as a target for anti-cancer therapy. Current pharmaceutical biotechnology 9, 2-8 (2008).

94. Murphy, G. The ADAMs: signalling scissors in the tumour microenvironment. Nature reviews. Cancer 8, 929-41 (2008).

95. Liu, P.C.C. et al. Identification of ADAM10 as a major source of HER2 ectodomain sheddase activity in HER2 overexpressing breast cancer cells. Cancer biology & therapy 5, 657-64 (2006).

96. Blobel, C.P., Carpenter, G. & Freeman, M. The role of protease activity in ErbB biology. Experimental cell research 315, 671-82 (2009).

97. Sahin, U. et al. Distinct roles for ADAM10 and ADAM17 in ectodomain shedding of six EGFR ligands. The Journal of cell biology 164, 769-79 (2004).

98. Zhou, B.-B.S. et al. Targeting ADAM-mediated ligand cleavage to inhibit HER3 and EGFR pathways in non-small cell lung cancer. Cancer cell 10, 39-50 (2006).

99. Sanderson, M.P. et al. ADAM10 mediates ectodomain shedding of the betacellulin precursor activated by p-aminophenylmercuric acetate and extracellular calcium influx. The Journal of biological chemistry 280, 1826-37 (2005).

112

100. Gavert, N. et al. Expression of L1-CAM and ADAM10 in human colon cancer cells induces metastasis. Cancer research 67, 7703-12 (2007).

101. Nagano, O. et al. Cell-matrix interaction via CD44 is independently regulated by different metalloproteinases activated in response to extracellular Ca(2+) influx and PKC activation. The Journal of cell biology 165, 893-902 (2004).

102. Esselens, C.W. et al. Metastasis-associated C4.4A, a GPI-anchored protein cleaved by ADAM10 and ADAM17. Biological Chemistry 389, 1075-1084 (2008).

103. Maretzky, T. et al. ADAM10 mediates E-cadherin shedding and regulates epithelial cell- cell adhesion, migration, and beta-catenin translocation. Proceedings of the National Academy of Sciences of the United States of America 102, 9182-7 (2005).

104. Reiss, K. et al. ADAM10 cleavage of N-cadherin and regulation of cell-cell adhesion and beta-catenin nuclear signalling. The EMBO journal 24, 742-52 (2005).

105. Hundhausen, C. et al. The disintegrin-like metalloproteinase ADAM10 is involved in constitutive cleavage of CX3CL1 (fractalkine) and regulates CX3CL1-mediated cell-cell adhesion. Blood 102, 1186-95 (2003).

106. Abel, S. et al. The transmembrane CXC-chemokine ligand 16 is induced by IFN- gamma and TNF-alpha and shed by the activity of the disintegrin-like metalloproteinase ADAM10. Journal of immunology (Baltimore, Md. : 1950) 172, 6362-72 (2004).

107. Schwarz, N. et al. Requirements for leukocyte transmigration via the transmembrane chemokine CX3CL1. Cellular and molecular life sciences : CMLS 67, 4233-48 (2010).

108. Schulte, M. et al. ADAM10 regulates FasL cell surface expression and modulates FasL- induced cytotoxicity and activation-induced cell death. Cell death and differentiation 14, 1040-9 (2007).

109. Gibb, D.R., Saleem, S.J., Chaimowitz, N.S., Mathews, J. & Conrad, D.H. The emergence of ADAM10 as a regulator of lymphocyte development and autoimmunity. Molecular immunology 48, 1319-27 (2011).

110. Conrad, D.H., Ford, J.W., Sturgill, J.L. & Gibb, D.R. CD23: an overlooked regulator of allergic disease. Current allergy and asthma reports 7, 331-7 (2007).

111. Lemieux, G.A. et al. The low affinity IgE receptor (CD23) is cleaved by the metalloproteinase ADAM10. The Journal of biological chemistry 282, 14836-44 (2007).

112. Gibb, D.R. et al. ADAM10 is essential for Notch2-dependent marginal zone B cell development and CD23 cleavage in vivo. The Journal of experimental medicine 207, 623-35 (2010).

113. Chaimowitz, N.S. et al. A Disintegrin and Metalloproteinase 10 Regulates Antibody Production and Maintenance of Lymphoid Architecture. Journal of immunology (Baltimore, Md. : 1950) 187, 5114-22 (2011).

113

114. Moss, M.L. et al. The ADAM10 prodomain is a specific inhibitor of ADAM10 proteolytic activity and inhibits cellular shedding events. The Journal of biological chemistry 282, 35712-21 (2007).

115. Amour, A. et al. The in vitro activity of ADAM-10 is inhibited by TIMP-1 and TIMP-3. FEBS letters 473, 275-9 (2000).

116. Reiss, K., Cornelsen, I., Husmann, M., Gimpl, G. & Bhakdi, S. Unsaturated Fatty Acids Drive Disintegrin and Metalloproteinase (ADAM)-dependent Cell Adhesion, Proliferation, and Migration by Modulating Membrane Fluidity. The Journal of biological chemistry 286, 26931-42 (2011).

117. Naus, S. et al. Identification of candidate substrates for ectodomain shedding by the metalloprotease-disintegrin ADAM8. Biological chemistry 387, 337-46 (2006).

118. Bartsch, J.W. et al. Tumour necrosis factor-alpha (TNF-alpha) regulates shedding of TNF-alpha receptor 1 by the metalloprotease-disintegrin ADAM8: evidence for a protease-regulated feedback loop in neuroprotection. The Journal of neuroscience : the official journal of the Society for Neuroscience 30, 12210-8 (2010).

119. Gómez-Gaviro, M. et al. Expression and regulation of the metalloproteinase ADAM-8 during human neutrophil pathophysiological activation and its catalytic activity on L- selectin shedding. Journal of immunology (Baltimore, Md. : 1950) 178, 8053-63 (2007).

120. Matsuno, O. et al. Role of ADAM8 in experimental asthma. Immunology letters 102, 67-73 (2006).

121. Kelly, K. et al. Metalloprotease-disintegrin ADAM8: expression analysis and targeted deletion in mice. Developmental dynamics : an official publication of the American Association of Anatomists 232, 221-31 (2005).

122. Valkovskaya, N. et al. ADAM8 expression is associated with increased invasiveness and reduced patient survival in pancreatic cancer. Journal of cellular and molecular medicine 11, 1162-74 (2007).

123. Wildeboer, D., Naus, S., Amy Sang, Q.-X., Bartsch, J.W. & Pagenstecher, A. Metalloproteinase ADAM8 and ADAM19 are highly regulated in human primary brain tumours and their expression levels and activities are associated with invasiveness. Journal of neuropathology and experimental neurology 65, 516-27 (2006).

124. Hernández, I., Moreno, J.L., Zandueta, C., Montuenga, L. & Lecanda, F. Novel alternatively spliced ADAM8 isoforms contribute to the aggressive bone metastatic phenotype of lung cancer. Oncogene 29, 3758-69 (2010).

125. Fritzsche, F.R. et al. ADAM8 expression in prostate cancer is associated with parameters of unfavorable prognosis. Virchows Archiv : an international journal of pathology 449, 628-36 (2006).

126. Roemer, A. et al. Increased mRNA expression of ADAMs in renal cell carcinoma and their association with clinical outcome. Oncology reports 11, 529-36 (2004).

114

127. Ainola, M. et al. Involvement of a disintegrin and a metalloproteinase 8 (ADAM8) in osteoclastogenesis and pathological bone destruction. Annals of the rheumatic diseases 68, 427-34 (2009).

128. Zack, M.D. et al. Reduced incidence and severity of experimental autoimmune arthritis in mice expressing catalytically inactive A disintegrin and metalloproteinase 8 (ADAM8). Clinical and experimental immunology 158, 246-56 (2009).

129. Chiba, Y. et al. Upregulation of ADAM8 in the airways of mice with allergic bronchial asthma. Lung 187, 179-85

130. King, N.E. et al. Expression and regulation of a disintegrin and metalloproteinase (ADAM) 8 in experimental asthma. American journal of respiratory cell and molecular biology 31, 257-65 (2004).

131. Naus, S. et al. The metalloprotease-disintegrin ADAM8 is essential for the development of experimental asthma. American journal of respiratory and critical care medicine 181, 1318-28 (2010).

132. Gossens, K., Naus, S., Holländer, G.A. & Ziltener, H.J. Deficiency of the metalloproteinase-disintegrin ADAM8 is associated with thymic hyper-cellularity. PloS one 5, e12766 (2010).

133. Domínguez-Luis, M. et al. The metalloprotease ADAM8 is associated with and regulates the function of the adhesion receptor PSGL-1 through ERM proteins. European journal of immunology 41, 3436-42 (2011).

134. Hall, T. et al. Autoactivation of human ADAM8: a novel pre-processing step is required for catalytic activity. Bioscience reports 29, 217-28 (2009).

135. Amour, A. et al. The enzymatic activity of ADAM8 and ADAM9 is not regulated by TIMPs. FEBS letters 524, 154-8 (2002).

136. Weskamp, G. et al. Mice lacking the metalloprotease-disintegrin MDC9 (ADAM9) have no evident major abnormalities during development or adult life. Molecular and cellular biology 22, 1537-44 (2002).

137. Parry, D.A. et al. Loss of the metalloprotease ADAM9 leads to cone-rod dystrophy in humans and retinal degeneration in mice. American journal of human genetics 84, 683-91 (2009).

138. Asai, M. et al. Putative function of ADAM9, ADAM10, and ADAM17 as APP alpha- secretase. Biochemical and biophysical research communications 301, 231-5 (2003).

139. Izumi, Y. et al. A metalloprotease-disintegrin, MDC9/meltrin-gamma/ADAM9 and PKCdelta are involved in TPA-induced ectodomain shedding of membrane-anchored heparin-binding EGF-like growth factor. The EMBO journal 17, 7260-72 (1998).

140. Tousseyn, T. et al. ADAM10, the rate-limiting protease of regulated intramembrane proteolysis of Notch and other proteins, is processed by ADAMS-9, ADAMS-15, and the gamma-secretase. The Journal of biological chemistry 284, 11738-47 (2009).

115

141. Parkin, E. & Harris, B. A disintegrin and metalloproteinase (ADAM)-mediated ectodomain shedding of ADAM10. Journal of neurochemistry 108, 1464-79 (2009).

142. Roghani, M. et al. Metalloprotease-disintegrin MDC9: intracellular maturation and catalytic activity. The Journal of biological chemistry 274, 3531-40 (1999).

143. Dyczynska, E. et al. Proteolytic processing of delta-like 1 by ADAM proteases. The Journal of biological chemistry 282, 436-44 (2007).

144. Guaiquil, V. et al. ADAM9 is involved in pathological retinal neovascularization. Molecular and cellular biology 29, 2694-703 (2009).

145. Zhou, M., Graham, R., Russell, G. & Croucher, P.I. MDC-9 (ADAM-9/Meltrin gamma) functions as an adhesion molecule by binding the alpha(v)beta(5) integrin. Biochemical and biophysical research communications 280, 574-80 (2001).

146. Mahimkar, R.M., Baricos, W.H., Visaya, O., Pollock, A.S. & Lovett, D.H. Identification, cellular distribution and potential function of the metalloprotease-disintegrin MDC9 in the kidney. Journal of the American Society of Nephrology : JASN 11, 595-603 (2000).

147. Mahimkar, R.M., Visaya, O., Pollock, A.S. & Lovett, D.H. The disintegrin domain of ADAM9: a ligand for multiple beta1 renal integrins. The Biochemical journal 385, 461- 8 (2005).

148. Fritzsche, F.R. et al. ADAM9 expression is a significant and independent prognostic marker of PSA relapse in prostate cancer. European urology 54, 1097-106 (2008).

149. Fritzsche, F.R. et al. ADAM9 is highly expressed in renal cell cancer and is associated with tumour progression. BMC cancer 8, 179 (2008).

150. O’Shea, C. et al. Expression of ADAM-9 mRNA and protein in human breast cancer. International journal of cancer. Journal international du cancer 105, 754-61 (2003).

151. Grützmann, R. et al. ADAM9 expression in pancreatic cancer is associated with tumour type and is a prognostic factor in ductal adenocarcinoma. British journal of cancer 90, 1053-8 (2004).

152. Zigrino, P., Mauch, C., Fox, J.W. & Nischt, R. Adam-9 expression and regulation in human skin melanoma and melanoma cell lines. International journal of cancer. Journal international du cancer 116, 853-9 (2005).

153. Tannapfel, A. et al. Identification of novel proteins associated with hepatocellular carcinomas using protein microarrays. The Journal of pathology 201, 238-49 (2003).

154. Shintani, Y. et al. Overexpression of ADAM9 in non-small cell lung cancer correlates with brain metastasis. Cancer research 64, 4190-6 (2004).

155. Mauch, C. et al. Accelerated wound repair in ADAM-9 knockout animals. The Journal of investigative dermatology 130, 2120-30 (2010).

116

156. Dreymueller, D. et al. Lung endothelial ADAM17 regulates the acute inflammatory response to lipopolysaccharide. EMBO molecular medicine (2012).doi:10.1002/emmm.201200217

157. Horiuchi, K. et al. Cutting Edge: TNF-{alpha}-Converting Enzyme (TACE/ADAM17) Inactivation in Mouse Myeloid Cells Prevents Lethality from Endotoxin Shock. J. Immunol. 179, 2686-2689 (2007).

158. Donners, M.M.P.C. et al. A disintegrin and metalloprotease 10 is a novel mediator of vascular endothelial growth factor-induced endothelial cell function in angiogenesis and is associated with atherosclerosis. Arteriosclerosis, thrombosis, and vascular biology 30, 2188-95 (2010).

159. Brummelkamp, T.R., Bernards, R. & Agami, R. Stable suppression of tumourigenicity by virus-mediated RNA interference. Cancer cell 2, 243-7 (2002).

160. Wiznerowicz, M. & Trono, D. Conditional Suppression of Cellular Genes : Lentivirus Vector-Mediated Drug-Inducible RNA Interference. Society 77, 8957-8961 (2003).

161. Barde, I., Salmon, P. & Trono, D. Production and titration of lentiviral vectors. Current protocols in neuroscience / editorial board, Jacqueline N. Crawley ... [et al.] Chapter 4, Unit 4.21 (2010).

162. Sambrook, J. & Russell, D.W. Molecular Cloning: A laboratory manual (third edition). 2344 (Cold Spring Harbor Laboratory Press,U.S.; 3rd Revised edition edition: 2000).

163. Tsuchiya, S. et al. Establishment and characterization of a human acute monocytic leukemia cell line (THP-1). International journal of cancer. Journal international du cancer 26, 171-6 (1980).

164. Ryu, J.-W. et al. Overexpression of uncoupling protein 2 in THP1 monocytes inhibits beta2 integrin-mediated firm adhesion and transendothelial migration. Arteriosclerosis, thrombosis, and vascular biology 24, 864-70 (2004).

165. Tikellis, C. et al. Reduced plaque formation induced by rosiglitazone in an STZ- diabetes mouse model of atherosclerosis is associated with downregulation of adhesion molecules. Atherosclerosis 199, 55-64 (2008).

166. Ashida, N., Arai, H., Yamasaki, M. & Kita, T. Distinct signaling pathways for MCP-1- dependent integrin activation and chemotaxis. The Journal of biological chemistry 276, 16555-60 (2001).

167. Koenen, R.R. et al. Regulated release and functional modulation of junctional adhesion molecule A by disintegrin metalloproteinases. Blood 113, 4799-809 (2009).

168. Elbashir, S.M. et al. Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells. Nature 411, 494-8 (2001).

169. Pei, Y. & Tuschl, T. On the art of identifying effective and specific siRNAs PERSPECTIVE. Nature Methods 3, 1-7 (2006).

117

170. Sozzani, S. et al. MCP-1 and CCR2 in HIV infection: regulation of agonist and receptor expression. Journal of leukocyte biology 62, 30-3 (1997).

171. Alert, H. Die Rolle der Metalloproteinasen ADAM8 und ADAM9 bei der Prozessierung von Syndekanen und der Migration epithelialer Zellen. (2011).

172. Andrzejewski, M.G. et al. Distinct role of the intracellular C-terminus for subcellular expression, shedding and function of the murine transmembrane chemokine CX3CL1. Biochemical and biophysical research communications 395, 178-84 (2010).

173. Mochizuki, S. & Okada, Y. ADAMs in cancer cell proliferation and progression. Cancer science 98, 621-8 (2007).

174. Giard, D.J. et al. In vitro cultivation of human tumours: establishment of cell lines derived from a series of solid tumours. Journal of the National Cancer Institute 51, 1417-23 (1973).

175. Brown, J. et al. Critical evaluation of ECV304 as a human endothelial cell model defined by genetic analysis and functional responses: a comparison with the human bladder cancer derived epithelial cell line T24/83. Laboratory investigation; a journal of technical methods and pathology 80, 37-45 (2000).

176. Bubeník, J., Perlmann, P., Helmstein, K. & Moberger, G. Cellular and humoral immune responses to human urinary bladder carcinomas. International journal of cancer. Journal international du cancer 5, 310-9 (1970).

177. Tang, J. et al. Adam17-dependent shedding limits early neutrophil influx but does not alter early monocyte recruitment to inflammatory sites. Blood 118, 786-94 (2011).

178. Schneider, U., Schwenk, H.U. & Bornkamm, G. Characterization of EBV-genome negative “null” and “T” cell lines derived from children with acute lymphoblastic leukemia and leukemic transformed non-Hodgkin lymphoma. International journal of cancer. Journal international du cancer 19, 621-6 (1977).

179. Dellacasagrande, J. et al. Reduced transendothelial migration of monocytes infected by Coxiella burnetii. Infection and immunity 68, 3784-6 (2000).

180. Pruenster, M. et al. The Duffy antigen receptor for chemokines transports chemokines and supports their promigratory activity. Nature immunology 10, 101-8 (2009).

181. Bartlett, A.H., Hayashida, K. & Park, P.W. Molecular and cellular mechanisms of syndecans in tissue injury and inflammation. Molecules and cells 24, 153-66 (2007).

182. Muller, W.A. Mechanisms of transendothelial migration of leukocytes. Circulation research 105, 223-30 (2009).

183. Durieu-Trautmann, O., Chaverot, N., Cazaubon, S., Strosberg, A.D. & Couraud, P.O. Intercellular adhesion molecule 1 activation induces tyrosine phosphorylation of the cytoskeleton-associated protein cortactin in brain microvessel endothelial cells. The Journal of biological chemistry 269, 12536-40 (1994).

118

184. van Buul, J.D., Kanters, E. & Hordijk, P.L. Endothelial signaling by Ig-like cell adhesion molecules. Arteriosclerosis, thrombosis, and vascular biology 27, 1870-6 (2007).

185. Cernuda-Morollón, E. & Ridley, A.J. Rho GTPases and leukocyte adhesion receptor expression and function in endothelial cells. Circulation research 98, 757-67 (2006).

186. Huang, A.J. et al. Endothelial cell cytosolic free calcium regulates neutrophil migration across monolayers of endothelial cells. The Journal of cell biology 120, 1371-80 (1993).

187. Turowski, P. et al. Phosphorylation of vascular endothelial cadherin controls lymphocyte emigration. Journal of cell science 121, 29-37 (2008).

188. Schulte, D. et al. Stabilizing the VE-cadherin-catenin complex blocks leukocyte extravasation and vascular permeability. The EMBO journal 30, 4157-70 (2011).

189. Mamdouh, Z., Kreitzer, G.E. & Muller, W.A. Leukocyte transmigration requires kinesin- mediated microtubule-dependent membrane trafficking from the lateral border recycling compartment. The Journal of experimental medicine 205, 951-66 (2008).

190. Mamdouh, Z., Chen, X., Pierini, L.M., Maxfield, F.R. & Muller, W.A. Targeted recycling of PECAM from endothelial surface-connected compartments during diapedesis. Nature 421, 748-53 (2003).

191. Dasgupta, B. & Muller, W.A. Endothelial Src kinase regulates membrane recycling from the lateral border recycling compartment during leukocyte transendothelial migration. European journal of immunology 38, 3499-507 (2008).

192. Muller, W.A., Weigl, S.A., Deng, X. & Phillips, D.M. PECAM-1 is required for transendothelial migration of leukocytes. The Journal of experimental medicine 178, 449-60 (1993).

193. Walcheck, B. et al. Neutrophil rolling altered by inhibition of L-selectin shedding in vitro. Nature 380, 720-3 (1996).

194. Cissé, M.A. et al. The disintegrin ADAM9 indirectly contributes to the physiological processing of cellular prion by modulating ADAM10 activity. The Journal of biological chemistry 280, 40624-31 (2005).

195. Ferguson, S.S. Evolving concepts in G protein-coupled receptor endocytosis: the role in receptor desensitization and signaling. Pharmacological reviews 53, 1-24 (2001).

196. Weber, K.S., Klickstein, L.B. & Weber, C. Specific activation of leukocyte beta2 integrins lymphocyte function-associated antigen-1 and Mac-1 by chemokines mediated by distinct pathways via the alpha subunit cytoplasmic domains. Molecular biology of the cell 10, 861-73 (1999).

197. Zen, K. et al. Cleavage of the CD11b extracellular domain by the leukocyte serprocidins is critical for neutrophil detachment during chemotaxis. Blood 117, 4885- 94 (2011).

119

198. Evans, B.J. et al. Shedding of lymphocyte function-associated antigen-1 (LFA-1) in a human inflammatory response. Blood 107, 3593-9 (2006).

199. Vaisar, T. et al. MMP-9 sheds the beta2 integrin subunit (CD18) from macrophages. Molecular & cellular proteomics : MCP 8, 1044-60 (2009).

200. Gomez, I.G. et al. Metalloproteinase-mediated shedding of integrin 2 promotes macrophage efflux from inflammatory sites. Journal of Biological Chemistry 287, 4581-9 (2011).

201. Ley, K., Laudanna, C., Cybulsky, M.I. & Nourshargh, S. Getting to the site of inflammation: the leukocyte adhesion cascade updated. Nature reviews. Immunology 7, 678-89 (2007).

202. Zarbock, A., Kempf, T., Wollert, K.C. & Vestweber, D. Leukocyte integrin activation and deactivation: novel mechanisms of balancing inflammation. Journal of molecular medicine (Berlin, Germany) (2011).doi:10.1007/s00109-011-0835-2

203. Stamenkovic, I. & Yu, Q. Shedding light on proteolytic cleavage of CD44: the responsible sheddase and functional significance of shedding. The Journal of investigative dermatology 129, 1321-4 (2009).

204. Liu, S. et al. A fragment of paxillin binds the alpha 4 integrin cytoplasmic domain (tail) and selectively inhibits alpha 4-mediated cell migration. The Journal of biological chemistry 277, 20887-94 (2002).

205. Hughes, S.C. & Fehon, R.G. Understanding ERM proteins--the awesome power of genetics finally brought to bear. Current opinion in cell biology 19, 51-6 (2007).

206. Ivetic, A. & Ridley, A.J. Ezrin/radixin/moesin proteins and Rho GTPase signalling in leucocytes. Immunology 112, 165-76 (2004).

207. Tarrant, J.M., Robb, L., van Spriel, A.B. & Wright, M.D. Tetraspanins: molecular organisers of the leukocyte surface. Trends in immunology 24, 610-7 (2003).

208. Wang, H.-X., Li, Q., Sharma, C., Knoblich, K. & Hemler, M.E. Tetraspanin protein contributions to cancer. Biochemical Society transactions 39, 547-52 (2011).

209. Le Naour, F. et al. Profiling of the tetraspanin web of human colon cancer cells. Molecular & cellular proteomics : MCP 5, 845-57 (2006).

210. Arduise, C. et al. Tetraspanins regulate ADAM10-mediated cleavage of TNF-alpha and epidermal growth factor. Journal of immunology (Baltimore, Md. : 1950) 181, 7002-13 (2008).

211. Xu, D., Sharma, C. & Hemler, M.E. Tetraspanin12 regulates ADAM10-dependent cleavage of amyloid precursor protein. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 23, 3674-81 (2009).

212. Gutiérrez-López, M.D. et al. The sheddase activity of ADAM17/TACE is regulated by the tetraspanin CD9. Cellular and molecular life sciences : CMLS 68, 3275-92 (2011).

120

213. Hogg, N., Patzak, I. & Willenbrock, F. The insider’s guide to leukocyte integrin signalling and function. Nature reviews. Immunology 11, 416-26 (2011).

214. Gooz, P. et al. A Disintegrin and Metalloenzyme (ADAM) 17 Activation Is Regulated by α5β1 Integrin in Kidney Mesangial Cells. PLoS ONE 7, e33350 (2012).

215. Sanderson, M.P., Dempsey, P.J. & Dunbar, A.J. Control of ErbB signaling through metalloprotease mediated ectodomain shedding of EGF-like factors. Growth factors (Chur, Switzerland) 24, 121-36 (2006).

216. Blobel, C.P. ADAMs: key components in EGFR signalling and development. Nature reviews. Molecular cell biology 6, 32-43 (2005).

217. Guaiquil, V.H. et al. ADAM8 is a negative regulator of retinal neovascularization and of the growth of heterotopically injected tumour cells in mice. Journal of molecular medicine (Berlin, Germany) 88, 497-505 (2010).

218. Zhang, H. et al. ErbB receptors: from oncogenes to targeted cancer therapies. The Journal of clinical investigation 117, 2051-8 (2007).

121

7. List of abbreviations

A8 ADAM8

A9 ADAM9

A10 ADAM10

A17 ADAM17

ADAM a disintegrin and metalloproteinase

ADAM-TS a disintegrin and metalloproteinase with thrombospondin 1

ADP adenosine diphosphate

Amp ampicillin

ANOVA analysis of variances

APC allophycocyanin

APP amyloid precursor protein

ATP adenosine triphosphate bp(s) base pairs

BSA bovine serum albumine

°C degrees Celsius

Ca2+ calcium2+

CD cluster of differentiation cDNA complementary DNA

CFP cyan fluorescent protein

CMV cytomegalovirus

Da dalton

DMEM Dulbecco’s modified Eagle’s medium

DMSO dimethyl sulfoxide

DNA deoxyribonucleic acid

E. coli Escherichia coli 122

E- Cadherin epithelial cadherin

ECL enhanced chemiluminescence

ECM extracellular matrix

EDTA ethylenediamine tetra-acetic acid

EGF epidermal growth factor

EGTA ethylene glycol tetraacetic acid

ELISA enzyme-linked immunosorbent assay

ERK extracellular-signal regulated kinase

E-selectin endothelial selectin et al. et alii

EtOH ethanol

FACS fluorescence activated cell sorter

FCS fetal calf serum

FITC fluorescein isothiocyanate fMLP N-formyl-methionine-leucine-phenylalanine

FSC forward scatter g g-force

G418 geneticin sulfate

GAPDH glyseraldehyde-3-phosphate dehydrogenase

G-CSF granulocyte-colony stimulating factor

GFP green fluorescent protein

GPCR G-protein coupled receptor

GTP guanosine triphosphate h hour

H2O water

HB-EGF heparin binding epidermal growth factor

HBSS Hank's buffered salt solution

123

HCl hydrochloric acid

HEPES 4-2-hydroxyethyl-1-piperazineethanesulfonic acid

HMVEC human micro-vascular endothelial cells

HRP horseradish peroxidase

HSPG heparan-sulphate

HUVEC human umbilical vein endothelial cells

ICAM intercellular adhesion molecule

IFN interferon

Ig immunoglobulin

IgG immunglobulin G

IL interleukin

IL6-R interleukin-6-receptor

IP3 inositol triphosphat

JAM junctional adhesion molecule kb kilo base pairs

KCl potassium chloride

KD knock-down kDa kilo Dalton l litre

LB lysogeny broth

LFA-1 lymphocyte-function associated antigen-1

LPS lipopolysaccharide

L-selectin leukocyte selectin

LV lentiviral transduced m milli

M Molar mAb monoclonal antibody

124

Mac-1 macrophage antigen-1

MAPK Mitogen aktivierte Protein Kinase

MCP-1 monocyte chemoattractant protein-1

MCS multiple cloning site mg Milligramm

Mg2+ magnesium2+ min minute ml millilitre mM millimolar

MMP matrix metalloproteinase mRNA messenger RNA n nano

N normality

N- Cadherin neuronal cadherin

NaCl sodium chloride

OD optical density

ORF open reading frame p pico p probability

P/S penicillin/streptomycin

PAGE polyacrylamide gel electrophoresis

PBS phosphate buffered saline

PCR polymerase chain reaction

PDVF polyvinylidene fluoride

PE phycoerythrin

PECAM platelet/endothelial cell adhesion molecule

PFA paraformaldehyd

125

pH potentia hydrogenii

PIP2 phosphatidylinositol 4,5-bisphosphate

PKC protein kinase C

PMA phorbol 12-myristate 13-acetate

POD peroxidase

P-selectin platelet selectin

PSGL-1 p-selection glycoprotein ligand-1 qPCR quantitative real-time polymerase chain reaction

RNA ribonucleic acid

RNAi RNA interference rpm rounds per minute

RPMI Roswell Park Memorial Institute

RT room temperature s second scr scramble shRNA

SD standard deviation

SDS sodium docecyl sulphate shRNA short hairpin RNA siRNA short interfering RNA

SOB super optimal broth

SSC sideward scatter stim. stimulation

SV-40 Simian Virus 40

SVMP snake venom metalloproteinase

TACE tumour necrosis factor alpha converting enzyme

TGFα transforming growth factor-alpha

TIMP tissue inhibitor of metalloproteinase

126

TNFα tumour necrosis factor-alpha

Tris tris (hydroxymethyl) aminomethane

Tween20 polyoxyethylenesorbitan monolaurate

U unit

V volts

VCAM-1 vascular cell adhesion molecule-1

VE-Cadherin vascular endothelial-Cadherin

VEGF-R vascular endothelial growth factor receptor

VLA-4 very late antigen 4

VSV-G vesicular stomatitis virus glykoprotein wt wildtype

Zn zinc

α alpha

β beta

γ gamma

µ mikro

µg microgram

µl microlitre

µm micrometer

Amino acids

A alanine (Ala) G glycine (Gly) M methionine (Met) S serine (Ser)

C cysteine (Cys) H histidine (His) N asparagine (Asn) T threonine (Thr)

D aspartic acid (Asp) I isoleucine (Ile) P proline (Pro) V valine (Val)

E glutamic acid (Glu) K lysine (Lys) Q glutamine (Gln) W tryptophane (Trp)

F phenylalanine (Phe) L leucine (Leu) R arginine (Arg) Y tyrosine (Tyr)

127

8. List of figures

Figure 1: The process of leukocyte recruitment...... 2

Figure 2: Family tree of the metzincins...... 7

Figure 3: Drawing showing the distinct domain of ADAM protease...... 8

Figure 4: The implication of ADAM-shedding on various substrate molecules...... 10

Figure 5: Involvement of ADAMs in inflammatory activation and leukocyte recruitment...... 14

Figure 6: General design of oligonucleotides containing the shRNA target sequences...... 29

Figure 7: Heat map of mRNA levels of ADAM8, 9, 10 and 17 in several cell culture lines.....44

Figure 8: Concentration kinetic of MCP-1-induced THP1 cell transmigration through HMVEC...... 45

Figure 9: Comparison of THP1 cells migration in in vitro chemotaxis and transdendothelial migration assays...... 46

Figure 10: Influence of ADAM inhibitors on THP1 migration...... 48

Figure 11: Analysis of lentiviral transduced HUVEC...... 50

Figure 12: Surface expression of ADAM10 on HEK293 cells transiently transfected with pSuper-ADAM10...... 51

Figure 13: Analysis of ADAM10 gene KD by pLVTHM-vectors ADAM10-650 and ADAM10- 1947...... 52

Figure 14: Gene silencing of ADAM17 with pLVTHM-lentiviral vectors...... 53

Figure 15: Effects of shRNA against ADAM10 or ADAM17 on the surface expression of the respective other ADAM protease...... 55

Figure 16: Effect of ADAM10 and 17 KD on the shedding of CX3CL1...... 56

Figure 17: Gene silencing of ADAM8 and 9...... 58

Figure 18: Flow cytometry analysis of ADAM10 and 17 surface expression on lentiviral transduced THP1...... 59

Figure 19: qPCR analysis of ADAM8, 9, 10 and 17 mRNA transcript levels in stable LV-THP1 cell lines...... 60

128

Figure 20: The refined transmigration assay using CFP labelled control cells and GFP labelled KD cells...... 63

Figure 21: Transmigration and chemotaxis of ADAM10-LV-THP1 in response to MCP-1...... 65

Figure 22: Percentage of migrated cells in the transmigration and chemotaxis assays of ADAM10-LV-THP1...... 66

Figure 23: Effect of ADAM8-KD on transmigration and chemotaxis of THP1 cells...... 68

Figure 24: Migration rates of ADAM9-LV-THP1 cells in transmigration and chemotaxis assays...... 69

Figure 25: Transmigration and chemotaxis assays with ADAM17-LV-THP1...... 70

Figure 26: Concentration dependent transmigration of ADAM-KD LV-THP1 towards MCP-1. 71

Figure 27: Ca2+-flux of wt THP1 after MCP-1 stimulation...... 73

Figure 28: MCP-1 induced Ca2+-response of THP1 cells treated with the ADAM inhibitors GI254023 and GW282406...... 75

Figure 29: Ca2+-response of LV-THP1 with gene silencing of ADAM8, 9, 10 or 17...... 76

Figure 30: Adhesion of THP1 to VCAM1, ICAM-1 and fibronectin under static conditions with and without the stimulation by 3nM MCP-1...... 77

Figure 31: Static adhesion of wt THP1 treated with the ADAM inhibitors GI254023 and GW282406...... 78

Figure 32: Static adhesion of LV-THP1 with gene silencing of ADMA8, 9 10 or 17...... 79

Figure 33: Microscopic images of the scratch assay performed with ADAM8-KD LV-ECV304 and scramble control cells...... 80

Figure 34: Summary of the scratch assays performed with LV-ECV 304 with a gene silencing of ADAM8, 9, 10 and 17...... 81

Figure 35: Hypothesis 1: Shedding of integrins as possible contribution of ADAM8 or 10 to the migration of THP1 cells...... 98

Figure 36: Hypothesis 2: Interaction of ADAM8 or 10 with integrins and possible other adhesion molecules is needed for the formation of high molecular clusters at the plasmamembrane of migrating THP1 cells...... 100

129

Figure 37: Hypothesis 3: ADAM8 or 10 contribute to the migration of THP1 cells by shedding ligands of receptors of the ErbB family, leading to an auto- or parakrine stimulation which promotes a migratory phenotype of the monocytic cells...... 102

Figure 38: Plasmid map of pSuper...... 132

Figure 39: Plasmid map of pLVTHM (Addgene number: 12247)...... 133

Figure 40: Plasmid map of psPAX2 (Addgene number: 12260)...... 134

Figure 41: Plasmid map of pMD2.G (Addgene number: 12259)...... 135

Figure 42: Expression levels of ADAM8, 9, 10 and 17 in several cell culture lines...... 136

Figure 43: Effect of ADAM8-KD on transmigration and chemotaxis of THP1 cells...... 137

Figure 44: Relative migration rates of ADAM9-LV-THP1 cell in transmigration and chemotaxis assays ...... 138

Figure 45: Transmigration and chemotaxis assays with ADAM17-LV-THP1...... 138

130

9. List of tables

Table 1: Shedding substrates of ADAM8, 9, 10 and 17 which are involved in inflammatory responses and leukocyte migration...... 17

Table 2: List of primary antibodies and their working dilutions...... 28

Table 3: List of secondary antibodies and their working dilutions...... 29

Table 4: List of oligonucleotides for the generation of pLVTHM plasmids for the down regulation of our target ADAM proteinases...... 30

Table 5: Sequences of qPCR-primers for ADAM8, 9, 10, 17 and GAPDH used in this study..31

Table 6: List of plasmids used in this study...... 31

Table 7: List of cell lines used in this study and their respective culture media...... 32

Table 8: Annealing temperatures and number of cycles for the qPCR for each target gene. 40

Table 9: Summary table of the outcome of the performed functional assays of this study...84

131

10. Appendix

10.1. Plasmid maps pSuper

DraIII BsaAI NaeI XmnI NgoMIV

TatI ScaI f1(+) orign Ecl136II SacI 3000 BstXI BstDSI SacII Amp NotI 500 Eco52I XbaI SpeI StyIBamHI 2500 pSuper-basic SmaI XmaI H1 prom. PstI 3176 bps EcoRI BpmI BbeIBstBI EheIAatII 1000 KasI NarI AhdI BstAPI 2000 BglII HindIII ClaI 1500 AccI HincII SalI XhoI EcoO109I pUC orgn Acc65I KpnI SapI

AlwNI AflIII PciI

Figure 38: Plasmid map of pSuper.

132

pLVTHM

XcmI PmlI

NotI Eco52I PshAI LTR

PpuMI SanDI BlnI 10000 SalI SfiI

SV40 ori 2000 pLVTHM BspMI EF1a 8000 11085 bps FseI

4000 SwaI Amp PacI eGFP PmeI 6000 FspI PvuI WPRE seq. H1 SpeI NdeI

SgrAI SnaBI++ XbaI Acc65I ClaI KpnI MluI BstXI BamHI EcoRI

Figure 39: Plasmid map of pLVTHM (Addgene number: 12247). MluI and ClaI were used to insert annealed oligonucleotides containing the shRNA sequences. EcoRI and ClaI were used to transfer H1-cassette together with the shRNA sequences from pSUPER. PmeI and SpeI were used to replace eGFP with eCFP. The site where the sequencing primer annealed is indicated by seq.

133

psPAX2

AatII DraIII SnaBI PvuI PmlI FspI PshAI SacII Bpu1102I CMVenh SgrAI

Amp CAGpro FseI

10000 StuI SfiI

2000 SphI SV40 ori psPAX2 gag

8000 11009 bps

NheI

4000

Tat/ref 6000 BclI

pol SbfI

Tth111I

Bsu36I AgeI SwaI AflII

Figure 40: Plasmid map of psPAX2 (Addgene number: 12260).

134

pMD2

Bsu36I ApaI PspOMI BtrI BsrGI XmnI ScaI PvuI SnaBI FspI CMV BsmBI HindIII

Ampicilin AhdI

5000 1000 beta-Globin intr. PmlI pMD2.G ClaI

pBR322 ori 5990 bps HaeII SalI 4000 2000

3000 VSV-G PshAI Eco52I Van91I NotI MscI EcoNI BstAPI PstI

AgeI BsaI BclI Bpu10I BstBI

Figure 41: Plasmid map of pMD2.G (Addgene number: 12259).

135

10.2. Complementary figure: Comparison of ADAM8, 9, 10 and 17 expression in different leukocytic, endothelial and epithelial cell culture lines

Figure 42 shows the data which were used to generate the heat map of mRNA levels of ADAM8, 9, 10 and 17 in several cell culture lines (see Results).

A 1.5 B 5

4 1.0 3

2 0.5 (rel. to GAPDH) to (rel. (rel. to (rel.to GAPDH) 1 level of ADAM8 mRNA of ADAM8 level ADAM9 transkript. level transkript. ADAM9 0.0 0 9 t V C 4 93 a C E CV HP1 E A549 VEC V E A5 K2 VEC T urk M THP1 Jurkat E J HEK293 H HU H HMVEC HU cell lines cell lines

C 3 D 5

4 2 3

2 1 (rel. to (rel. GAPDH) to (rel.to GAPDH) 1 ADAM17 transkript. level transkript. ADAM17 ADAM10 transkript. level transkript. ADAM10 0 0 t V 9 C V 9 a 4 E C 4 EC rkat E 5 293 HP1 rk EC A5 u A K T u THP1 J MV J HEK293 HMV HUVEC HE H HUVEC cell lines cell lines

Figure 42: Expression levels of ADAM8, 9, 10 and 17 in several cell culture lines. The figure shows the mRNA levels of ADAM8 (A), ADAM9 (B), ADAM10 (C) and ADAM17 (D) in the indicated cell lines. The amount of each ADAM-mRNA was determined using qPCR (n=4) and is shown as mean ±SD. The expression rate of the ADAM genes was normalised to the housekeeping gene GAPDH. mRNA expression levels are expressed in relation to those of ECV304 cells.

136

10.3. Complementary figures: Relative number of migrated ADAM8, ADAM9 or ADAM17-KD LV-THP1 in transmigration and chemotaxis assay.

A 1.5 B 1.5

1.0 1.0

** ** *** *

(rel.to scr) 0.5 (rel. scr) to 0.5

ratio of migrated cells ratio migrated of *** *** cells migrated ratioof 0.0 0.0 r 3 0 31 45 im sc 8 t cr 26 s 31 tim A8 9 18 A8 1 A8 A8 903 A8 A8 2645 scr no s scr no s LV-THP1-KD-lines LV-THP1-KD-lines

Figure 43: Effect of ADAM8-KD on transmigration and chemotaxis of THP1 cells (data expressed as number of migrated cells relative to the scr control cells). A: Relative number of migrated ADAM8 deficient LV-THP1 through an endothelial layer of HMVEC in the presence of MCP-1 as chemotactic stimulus. B: Relative number of migrated ADAM8-KD LV-THP1 in the chemotaxis assay. Data are shown as number of migrated cells relative to the scramble control. The number of all migrated cells (LV-THP1 scramble-CFP together with ADAM8-KD GFP expressing LV- THP1) was determined with the glucoronidase assay. To show the effect of the MCP-1 stimulus the percentage of scramble cells that migrated in the absence of MCP-1 (scr no stim) shown in all diagrams. The data of all panels are shown as mean ±SD, n=4, *p<0.05, **p<0.01, ***p<0.001 compared to scramble control.

137

A 2.0 B 2.5 *** 2.0 1.5 1.5 1.0 1.0 (rel. scr) to (rel. to scr) 0.5 0.5 ratio of migrated cells of ratio migrated ratio of migrated cells ratio migrated of *** *** 0.0 0.0 0 67 60 m cr 6 scr 9 2 s im st 9 1 9 2 A A A9 1967 A9 22 no cr scr no sti s LV-THP1-KD-lines LV-THP1-KD-lines

Figure 44: Relative migration rates of ADAM9-LV-THP1 cell in transmigration and chemotaxis assays (data expressed as number of migrated cells relative to the scr control cells). A: Relative number of migrated ADAM9 deficient LV-THP1 cells through an endothelial layer of HMVEC in the presence of MCP-1. B: Relative number of migrated ADAM9-KD LV-THP1 in the chemotaxis assay. Data are shown as number of migrated cells relative to the scramble control. The number of all migrated cells (LV-THP1 scramble-CFP together with ADAM9-KD GFP expressing LV-THP1) was determined with the glucoronidase assay. To show the effect of the MCP-1 stimulus the percentage of scramble cells that migrated in the absence of MCP-1 (scr no stim) shown in all diagrams. The data of all panels are shown as mean ±SD, n=4, ***p<0.001 compared to scramble control.

A 1.5 B 1.5

1.0 1.0 *** ***

(rel. scr) to 0.5 (rel.to scr) 0.5 *** ratio of migrated ratiocells of migrated

*** cells ratio migrated of 0.0 0.0 1 2 4 m scr 06 scr stim 7 2 7 26 2061 2642 1 1 o A 17 n A r scr no sti A17 A sc LV-THP1-KD-lines LV-THP1-KD-lines

Figure 45: Transmigration and chemotaxis assays with ADAM17-LV-THP1 (data expressed as number of migrated cells relative to the scr control cells). A: Relative number of migrated ADAM17 deficient LV-THP1 through an endothelial layer of HMVEC in the presence of MCP-1 as chemotactic stimulus. B: Relative number of migrated ADAM17-KD LV-THP1 in the chemotaxis assay. Data are shown as number as migrated cells relative to the scramble control. The number of all migrated cells (LV-THP1 scramble-CFP together with ADAM17-KD GFP expressing LV-THP1) was determined with the glucoronidase assay. To show the effect of the MCP-1 stimulus the percentage of scramble cells that migrated in the absence of MCP-1 (scr no stim) shown in all diagrams. The data of all panels are shown as mean ±SD, n=4, ***p<0.001 compared to scramble control.

138

11. Publications

Dreymueller, Daniela, Christian Martin, Tanja Kogel, Jessica Pruessmeyer, Franz Martin Hess, Keisuke Horiuchi, Stefan Uhlig, and Andreas Ludwig. “Lung endothelial ADAM17 regulates the acute inflammatory response to lipopolysaccharide.” EMBO Molecular Medicine; (2012 Feb 24). [Epub ahead of print]

Pruessmeyer, Jessica, Christian Martin, Franz Martin Hess, Nicole Schwarz, Sven Schmidt, Tanja Kogel, Nicole Hoettecke, Boris Schmidt, Antonio Sechi, Stefan Uhlig and Andreas Uhlig. “A disintegrin and metalloproteinase 17 (ADAM17) mediates inflammation-induced shedding of syndecan-1 and -4 by lung epithelial cells.” The Journal of biological chemistry 285, no. 1 (January 1, 2010): 555-64.

Schwarz, Nicole, Jessica Pruessmeyer, Franz Martin Hess, Daniela Dreymueller, Elena Pantaler, Anne Koelsch, Reinhard Windoffer, Mathias Voss, Alisina Sarabi, Christian Weber, Antonio Sechi, Stefan Uhlig and Andreas Ludwig. “Requirements for leukocyte transmigration via the transmembrane chemokine CX3CL1.” Cellular and molecular life sciences : CMLS 67, no. 24 (December 2010): 4233-48.

139

12. Curriculum Vitae

Franz Martin Heß

Born: May 26th, 1980 Place of birth: Darmstadt, Germany Nationality: German

Since December 2007 - PhD student at the Institute of Pharmacology and Toxicology, to present University Hospital, RWTH University Aachen, Germany Supervisor: Prof. Dr. rer. nat. Andreas Ludwig Topic of PhD thesis: „ The role of the metalloproteases ADAM8, 9, 10 and 17 in leukocyte migration in vitro”

June - August 2007 Research assistant at the Max-Planck-working groups for structural Molecularbiology, Hamburg

April 2006 - March 2007 Diploma thesis at the Institute for Zytobiology Philipps University Marburg, Germany „In Vivo Analysis of the apical and basolateral protein transport in polarized epithelial cells“

July - December 2004 Lab internship at the Neuropharmacology and Neuroinflammatory Research Group, RMIT University, Melbourne, Australia

March - June 2004 Study abroad at University of Melbourne, Melbourne, Australia

October 2000 Enrolment as student of Humanbiology at the Philipps-University Marburg, Germany

August 1990 - Secondary school at Gymnasium Michelstadt, Germany June 2000

July 1996 - June 1997 Exchange student at the Escuela Pablo Colon Berdecia in Barranquitas, Puerto Rico

140

13. Declaration

I hereby declare, that this thesis was carried out at RWTH Aachen University, within the Institute for Pharmacology and Toxicology and the Institute for Molecular and Cardiovascular Research. It was exclusively performed by myself, unless otherwise stated in the text. To my knowledge, it contains no material used in other publications or thesis, except where reference is made in the text.

F. Martin Heß , Aachen 03.7.2012

141

14. Acknowledgements

First of all I would like to thank my doctoral adviser Prof Dr Andreas Ludwig for his guidance and his helpful advice which he provided over the last years. He did not only help me with experiments and scientific matters but as well always had an open ear and understanding for matters exceeding science.

I want to thank Prof Dr Zenke for reviewing this thesis and for letting me work in his laboratory on the establishment of the lentiviral transduction system.

Prof Dr Baumgartner, thank you for agreeing to be my examiner in the upcoming PhD defence.

Without the help of Dr Antonio Sechi, who introduced me to lentiviral work and helped with the establishment of our lentiviral system, and Dr Kristin Seré this thesis would not be what it is today. Thank you for your advice and precious time.

Our technicians, Tanja Kogel and Melanie Esser, deserve my gratitude for keeping the laboratory running and for their help with experiments. The same goes for my congenial colleagues of the Ludwig group, the Institute of Pharmacology and the IMCAR. Especially, I would like to thank Michael, Mike, Svenja, Regina, Ali, Martin, Julian, Ralf and Ulla. The anatomists shall not be forgotten: thanks for the good times Christian, Jon and Tim!

Jette Alert deserves special mentioning since it was always a great pleasure to work with her.

Dr Phillip, you are the bon vivant who showed me how to survive writing a PhD thesis.

Brendan, my Californian Australian friend: I owe you big time for reading and correcting this PhD thesis.

My family in law, who took me in with open arms and hearts: Thank you for accepting me as son and brother.

Without my parents, Johanna and Franz Jakob Heß, I would not be where I am today. I want to thank you for being loving and understanding parents who raised me, supported me and accepted me as I am.

Finally, I thank my dear and beloved family, whose patience and benevolence gave me the time and strength to work on this thesis. My little man Matti: without your thousand hugs and good wishes your old man would not have made it! Kristine: thousand thanks for everything. We made it.

142