Study of molecular and immunological factors that determine the sensitivity to the antitumor effect of Tumor Necrosis Factor

Leander Huyghe

Thesis submitted in fulfillment of the requirements for the degree of DOCTOR IN SCIENCES: BIOTECHNOLOGY

Promoter: Prof. Dr. Peter Brouckaert

2008

Ghent University, Faculty of Sciences VIB Department for Molecular Biomedical Research

Cover pictures: C57BL/6 mouse, PECAM-1 stained tumor blood vessels, crystal structure of human Tumor Necrosis Factor (from Eck, M.J., Sprang, S.R. (1989) J. Biol. Chem. 264: 17595-605), May-Grünwald Giemsa stained blood smear showing lymphocyte and neutrophil, cultured B16Bl6 melanoma cells

Study of molecular and immunological factors that determine the sensitivity to the antitumor effect of Tumor Necrosis Factor

Leander Huyghe

Thesis submitted in fulfillment of the requirements for the degree of DOCTOR IN SCIENCES: BIOTECHNOLOGY

Promoter: Prof. Dr. Peter Brouckaert

2008

Ghent University, Faculty of Sciences VIB Department for Molecular Biomedical Research Examination committee:

Reading committee: Prof. Dr. Peter Brouckaert 1 Prof. Dr. Dirk Elewaut 2 Prof. Dr. Olivier Feron 3 Prof. Dr. Bart Vanhaesebroeck 4

Other members: Prof. Dr. Johan Grooten (chair) 5 Prof. Dr. Georges Leclercq 6 Prof. Dr. Claude Libert 1 Dr. Jeroen Hostens 7

1 VIB Department for Molecular Biomedical Research, Ghent University, Ghent, Belgium 2 Department of Rheumatology, Ghent University Hospital, Ghent, Belgium 3 Unit of Pharmacology and Therapeutics, University of Louvain (UCL) Medical School, Brussels, Belgium 4 Centre for Cell Signaling, Institute of Cancer, Queen Mary University of London, London, UK 5 Department of Molecular Biology, Ghent University, Ghent, Belgium 6 Department of Clinical Chemistry, Microbiology and Immunology, Ghent University Hospital, Ghent, Belgium 7 Philip Morris Research Laboratories, Leuven, Belgium i

TABLE OF CONTENTS

LIST OF ABBREVIATIONS ...... 1

OUTLINE OF THE THESIS ...... 7

INTRODUCTION ...... 9

Chapter 1: Tumor Necrosis Factor ...... 9 Chapter foreword ...... 11 1.1. A brief history ...... 11 1.2. TNF and its receptors ...... 12 1.3. Regulation of TNF production ...... 15 1.4. TNF-R1 signaling ...... 16 1.4.1. General aspects of TNF-R1 signaling ...... 16 1.4.2. The NF-B pathway ...... 17 1.4.3. The MAPK pathway ...... 20 1.4.4. TNF-induced cell death ...... 24 1.4.5. Other TNF-induced pathways ...... 31 1.5. TNF-R2 signaling ...... 35 1.6. TNF in inflammation and immunity ...... 36 1.7. TNF in cancer therapy ...... 43 1.8. References ...... 49

Chapter 2: Interferon- ...... 69 2.1. IFN- and its receptor ...... 71 2.2. IFN-R signaling ...... 72 2.3. IFN- in immunity ...... 75 2.4. Synergy with TNF ...... 78 2.5. References ...... 82

Chapter 3: The PI3K/Akt pathway ...... 91 3.1. PI3K ...... 93 3.1.1. The PI3K family ...... 93 3.1.2. Class I PI3Ks ...... 96 3.2. Akt ...... 102 3.3. mTOR ...... 107 3.4. References ...... 110

Chapter 4: Angiogenesis ...... 121 4.1. Background and definitions ...... 123 4.2. Regulation of angiogenesis ...... 126 4.2.1. Positive regulators of angiogenesis ...... 126 4.2.2. Negative regulators of angiogenesis ...... 134 4.3. Tumor vascularization ...... 136 4.3.1. Tumor angiogenesis ...... 136 4.3.2. EPC recruitment ...... 137 4.3.3. Vessel cooption ...... 138 4.3.4. Mosaic vessels ...... 139 4.3.5. Vasculogenic mimicry ...... 139 ii

4.3.6. Abnormalities in tumor blood vessels ...... 140 4.4. References ...... 141

Chapter 5: Vascular-targeting anticancer therapy ...... 151 5.1. Introduction ...... 153 5.2. Angiogenesis inhibitors ...... 154 5.3. Vascular-disrupting agents ...... 156 5.3.1. Tubulin-binding agents ...... 156 5.3.2. TNF and TNF-inducing flavonoids ...... 157 5.4. Ligand-directed vascular-targeting agents ...... 159 5.5. References ...... 161

Chapter 6: NKT cells ...... 165 6.1. Definition ...... 167 6.2. Development ...... 169 6.3. Function ...... 170 6.4. NKT cells in tumor immunity and cancer therapy ...... 172 6.5. References ...... 175

CHAPTER 7: RESEARCH OBJECTIVES ...... 179

EXPERIMENTAL WORK ...... 183

Chapter 8: Sensitization of the tumor vasculature to the vascular-disrupting effect of TNF by inhibition of the PI3K/Akt survival pathway ...... 183 Abstract ...... 185 8.1. Introduction ...... 186 8.2. Materials and methods ...... 187 8.2.1. Mice ...... 187 8.2.2. Cells ...... 188 8.2.3. Reagents ...... 188 8.2.4. Tumor experiments ...... 188 8.2.5. MTT cytotoxicity assay ...... 188 8.2.6. CD31 staining ...... 189 8.2.7. TUNEL staining ...... 189 8.2.8. In vivo Hoechst staining ...... 189 8.2.9. Dorsal skinfold window chamber model ...... 189 8.2.10. Statistics ...... 190 8.3. Results ...... 190 8.3.1. Inhibition of PI3K sensitizes to the antitumor effect of TNF ...... 190 8.3.2. The antitumor effect of TNF plus wortmannin is host-mediated ...... 194 8.3.3. Role of p110 isoforms in the sensitivity to the antitumor effect of TNF ...... 195 8.3.4. Inhibition of PI3K sensitizes to the vascular-disrupting effect of TNF ...... 196 8.3.5. Inhibition of Akt or mTOR sensitizes to the antitumor effect of TNF ...... 203 8.3.6. Synergy of TNF and anti-angiogenic therapy ...... 206 8.4. Discussion ...... 208 8.5. References ...... 212

Chapter 9: Invariant NKT cells contribute to tumor destruction by TNF ...... 215 Abstract ...... 217 iii

9.1. Introduction ...... 218 9.2. Materials and methods ...... 218 9.2.1. Mice ...... 218 9.2.2. Cells ...... 219 9.2.3. Reagents ...... 219 9.2.4. Tumor experiments ...... 220 9.2.5. Transfer of splenocytes and liver leukocytes ...... 220 9.2.6. In vivo depletion of NK and NK1.1+ cells ...... 220 9.2.7. Statistics ...... 221 9.3. Results ...... 221 9.3.1. Lymphocytes influence the response to TNF ...... 221 9.3.2. Invariant NKT cells contribute to the TNF-induced antitumor effect ...... 226 9.3.3. GalCer sensitizes to the antitumor effect of TNF ...... 228 9.4. Discussion ...... 231 9.5. References ...... 234

Chapter 10: The synergistic antitumor effect of TNF and IFN- is dependent on neutrophils ...... 237 Abstract ...... 239 10.1. Introduction ...... 240 10.2. Materials and methods ...... 241 10.2.1. Mice ...... 241 10.2.2. Cells ...... 241 10.2.3. Reagents ...... 241 10.2.4. Tumor experiments ...... 241 10.2.5. In vivo depletion of neutrophils, macrophages or NK1.1+ cells ...... 242 10.2.6. Statistics ...... 242 10.3. Results ...... 243 10.3.1. Lymphocytes are not required for the synergistic antitumor effect of TNF and IFN- ...... 243 10.3.2. Neutrophils are required for the synergistic antitumor effect of TNF and IFN- ...... 244 10.4. Discussion ...... 247 10.5. References ...... 249

CHAPTER 11: GENERAL SUMMARY & DISCUSSION ...... 251 11.1. Summary ...... 253 11.2. Discussion ...... 257 11.3. Conclusion & future perspectives ...... 262 11.4. References ...... 264

CHAPTER 12: ALGEMENE SAMENVATTING & DISCUSSIE ...... 267 12.1. Samenvatting ...... 269 12.2. Discussie ...... 274 12.3. Conclusie & perspectieven ...... 279 12.4. Referenties ...... 281

CHAPTER 13: ADDENDUM ...... 285

CURRICULUM VITAE ...... 289 iv

DANKWOORD ...... 292

List of abbreviations 1

LIST OF ABBREVIATIONS

2-5A synthetase 2’-5’ oligoadenylate synthetase 3’UTR 3’ untranslated region 4EBP1 eIF4E binding protein 1 GalCer -galactosylceramide ADAM a-disintigrin-and-metalloproteinase ADCC antibody-dependent cell-mediated cytotoxicity aFGF acidic fibroblast growth factor AI angiogenesis inhibitor AIP1 ASK-interacting protein 1 AMPK AMP-activated protein kinase Ang angiopoietin ARE AU-rich element ASK-1 apoptosis signal-regulating kinase-1 aSMase acidic sphingomyelinase Apaf-1 apoptotic protease activating factor-1 APC antigen-presenting cell ATF-2 activating transcription factor-2 ATM ataxia telangiectasia mutated ATR ATM-related -TRCP -transducin repeat-containing protein BCG Bacillus Calmette-Guérin bFGF basic fibroblast growth factor c-FLIP cellular FLICE-inhibitory protein CA4 combretastatin A4 CA4P combretastatin A4 phosphate CAD caspase-activated DNase CAM chorioallantoic membrane CARD caspase activation and recruitment domain CBP CREB binding protein CCL2 CC-chemokine ligand 2 CIITA class II transactivator CK2 casein kinase 2 COX cyclooxygenase cPLA2 cytosolic phospholipase A2 CRE cyclic AMP response element CREB cAMP response element-binding CTMP carboxyterminal modulator protein DAP death-associated protein DARC Duffy antigen/receptor for chemokines dATP 2’-deoxy ATP 2

DED death effector domain DIABLO direct IAP-binding protein with low pI DISC death-inducing signaling complex DD death domain DMXAA 5,6-dimethylxanthenone-4-acetic acid DNA-PK DNA-dependent protein kinase DP CD4+ CD8+ double positive dsRAD dsRNA specific adenosine deaminase dsRNA double-stranded RNA EC endothelial cell ECM extracellular matrix eEF2 eukaryotic elongation factor 2 EGF epidermal growth factor eIF eukaryotic initiation factor EMEA European Medicine Agency eNOS endothelial NOS EPC endothelial precursor cell ERK extracellular signal-related kinase FAA flavone acetic acid FADD Fas-associated death domain protein FAN factor associated with nSMase activation FAT focal adhesion targeting FDA Food and Drug Administration FGF fibroblast growth factor FLICE Fas-associated DD-like IL-1-converting enzyme FOXO Forkhead box O FRB FKBP12-rapamycin-binding GL G protein -like GADD153 growth arrest and DNA damage 153 GAG glycosaminoglycan GAP GTPase-activating protein GAS gamma-activated site GCK germinal center kinase GPCR G-protein-coupled receptor Grb10 growth factor receptor-interacting protein 10 GSK3 glycogen synthase kinase 3 HAT histone acetyltransferase HDAC histone deacetylase HGF hepatocyte growth factor signaling HIF-1 hypoxia-inducible factor-1 HRE HIF-responsive element Hsp90 heat shock protein-90 List of abbreviations 3

hTNF human TNF IAP inhibitor of apoptosis protein ICAD inhibitor of CAD ICAM intercellular adhesion molecule IFN- interferon- IFN-R  IFN- receptor IFP interstitial fluid pressure iGb3 isoglobotrihexosylceramide IHP isolated hepatic perfusion IB inhibitor of NF-B IKK IB kinase IL interleukin ILK integrin-linked kinase ILP isolated limb perfusion iNKT invariant NKT cell iNOS inducible NOS

InsP3 inositol-1,4,5-triphosphate IRF-1 interferon regulatory factor-1 JAK Janus kinase JIP1 JNK-interacting protein 1 JNK c-Jun NH2-terminal kinase KIR killer inhibitory receptor KitL Kit ligand L-NAME N-nitro-L-arginine-methyl-ester LLC Lewis lung carcinoma LOX lipoxygenase LPS lipopolysaccharide LT lymphotoxin LT leukotriene MAP2K MAPK kinase MAP3K MAPK kinase kinase MAPC multipotent adult progenitor cells MAPK mitogen-activated protein kinase MCP-1 monocyte chemoattractant protein-1 MDM2 murine double minute homolog 2 MEF mouse embryonic fibroblast MEKK MAPK/ERK kinase kinase methA methylcholanthrene A MHC major histocompatibility complex MK2 MAPK-activated protein kinase-2 MKK MAPK kinase MLC myosin light chain 4

MLCK MLC kinase MLK-3 mixed lineage kinase-3 MMP matrix metalloproteinase MnSOD manganous SOD mRNA messenger RNA MSK-1 mitogen- and stress-activated kinase-1 MT1-MMP membrane type 1-MMP MTD maximum tolerated dose mTNF murine TNF mTOR mammalian target of rapamycin mTORC mTOR complex NEMO NF-B essential modulator NF-B nuclear factor-B NK natural killer nNOS neuronal NOS NO nitric oxide NOS NO-synthase NRP neuropilin NSD neutral sphingomyelinase domain nSMase neutral sphingomyelinase PAF platelet-activating factor PAI-1 plasminogen activator inhibitor-1 PAK2 p21-activated caspase 2 PAMP pathogen-associated molecular pattern PARP poly(ADP-ribose) polymerase PCNA proliferating cell nuclear antigen PDGF platelet-derived growth factor

PDK-1 PI(3,4,5)P3-dependent protein kinase-1 PECAM platelet-endothelial cell adhesion molecule PG prostaglandin PGI prostacyclin PH pleckstrin homology PI phosphatidylinositol PI(3)P PI 3-phosphate

PI(3,4)P2 PI 3,4-biphosphate

PI(3,4,5)P3 PI 3,4,5-triphosphate PI3K PI 3-kinase PI4K PI 4-kinase PIAS protein inhibitor of activated STAT PIKK PI3K-related kinase PKA protein kinase A PKB protein kinase B List of abbreviations 5

PKC protein kinase C PKR dsRNA activated protein kinase PLA2 phospholipase A2 PLAD preligand binding assembly domain PLC phospholipase C PlGF placental growth factor PP2A protein phosphatase 2A PRR pattern recognition receptor PS phosphatidylserine PSMA prostate-specific membrane antigen PTEN phosphatase and tensin homolog PTK protein tyrosine kinase Pyk2 proline-rich tyrosine kinase 2 pVHL von Hippel-Lindau protein RIP receptor-interacting protein RACK1 receptor for activated C-kinase 1 RAIDD RIP-associated ICH-1/CED-3 homologues protein with DD Rap-1 Ras GTPase-activating protein-1 Raptor regulatory associated protein of mTOR RasGAP Ras GTPase-activating protein Rheb Ras homolog enriched in brain Rictor rapamycin-insensitive companion of mTOR RIT radioimmunotherapy ROCK-I Rho kinase-I ROS reactive oxygen species RTK receptor tyrosine kinase S6K p70 S6 kinase SAPK stress-activated protein kinase SEK-1 SAPK/ERK kinase-1 sGC soluble guanylate cyclase SH2 Src homology 2 SHIP SH2 domain-containing inositol 5-phosphatase SHP-2 SH2 domain-containing tyrosine phosphatase-2 SIN1 SAPK-interacting protein kinase 1 SIRS systemic inflammatory response syndrome SK-1 sphingosine kinase-1 Smac second mitochondria-derived activator of caspases SMase sphingomyelinase SMC smooth muscle cell SOCS1 suppressor of cytokine signaling 1 SOD superoxide dismutase SODD silencer of death domains 6

STAT1 signal transducer and activator of transcription 1 sTNF-R soluble TNF receptor TACE TNF--converting enzyme TAK1 TGF--activated kinase 1 TAP transporter associated with antigen processing tBid truncated Bid TCL1 T cell leukemia 1 TCR T cell receptor TGF- transforming growth factor- Th1 T helper type 1 TIA-1 T cell restricted intracellular antigen-1 TIAR TIA-1-related protein TIM TRAF-interacting motif TIMP tissue inhibitor of MMP TLR4 Toll-like receptor 4 TNF Tumor necrosis factor TNF-R TNF-receptor tPA tissue-type plasminogen activator TRADD TNF-receptor associated death domain protein TRAF TNF-receptor associated factor TRAIL TNF-related apoptosis inducing ligand TRB3 Tribbles homolog 3 TRID TNF-R1 internalization domain TSC tuberous sclerosis complex TSI tumor size index TTP tristetraprolin TX thromoboxane uPA urokinase-type plasminogen activator uPAR uPA receptor VCAM vascular cell adhesion molecule VDA vascular-disrupting agent VE-cadherin vascular endothelial-cadherin VEGF vascular endothelial growth factor VEGI vascular endothelial growth inhibitor VTA vascular-targeting agent vWF von Willebrand factor XAF-1 XIAP associated factor-1 XIAP X-chromosome-linked IAP Outline of the thesis 7

OUTLINE OF THE THESIS

Tumor necrosis factor (TNF) is an inflammatory cytokine with a potent antitumor activity. Both in experimental animal models and in human patients it can rapidly induce a haemorrhagic necrosis of tumors. The clinical use of TNF is, however, severely hampered by the potentially life-threatening side-effects which emerge when the high doses of TNF required to produce an antitumor effect are administered systemically. The research described in this dissertation is part of a larger research project that aims to unravel the molecular basis of the antitumor and shock-inducing effects of TNF, so that a safe, systemic TNF-based treatment could be developed. More in particular, the aim of the thesis was to study factors that influence the sensitivity to the antitumor activity of TNF, in order to interfere with them therapeutically, so that lower doses of TNF would be sufficient to obtain a similar antitumor effect.

In the first part of the thesis we introduce the subject by summarizing the current knowledge of the different molecular players and systems. The signaling cascade initiated by TNF and the inflammatory and antitumor properties of this cytokine are summarized in chapter 1. Already during the first experiments performed by Brouckaert and colleagues it became clear that co-treatment with another inflammatory cytokine, interferon- (IFN-), could significantly enhance the antitumor effect of TNF. These and similar findings by other groups has even led to the addition of IFN- to TNF-based isolated limb perfusion (ILP) settings. The mechanism behind this remarkable effect is however still largely unknown and is one of the subjects of this thesis. Therefore, IFN- is reviewed in the introductory chapter 2, with the emphasis on its role in immunity and its synergy with TNF. Another approach that is followed to develop a safer TNF-based anticancer therapy is to combine low doses of TNF with agents that sensitize for the antitumor effect of TNF. In a preliminary study we identified a phosphatidylinositol 3-kinase (PI3K) inhibitor as such an agent. A second aim of this research was to further elaborate this finding and identify new potential targets within the PI3K/Akt survival pathway. Therefore, the PI3K/Akt pathway and its most important components, PI3K, Akt and mammalian target of rapamycin (mTOR) are discussed in chapter 3 of the introduction. It has become clear that TNF-based anticancer therapies act principally by modifying and disrupting the tumor neovasculature. To better understand the complex nature of this tumor vasculature, the process of (tumor) angiogenesis will be discussed in chapter 4. A general overview of current vascular-targeted therapies is given in chapter 5. A third and final subject of this thesis is the role of the immune system in the antitumor effects of TNF and IFN-. Some reports suggest that certain immune cells may actively contribute to the antitumor response initiated by TNF. However, the nature of these immune cells remains unclear. As our research, described in this thesis, points to a potential role for natural killer T (NKT) cells, a relatively recently discovered type 8

of immune cell with some remarkable properties, these cells will be discussed in chapter 6 of the introduction.

In the second part of the thesis we report the results of our experimental work. In chapter 8 we report our observations regarding the effects interfering with the PI3K/Akt pathway has on the antitumor effect of TNF. In chapter 9 we report our observations on the role of NKT cells in this antitumor effect while in chapter 10 the mechanisms underlying the synergistic effect of IFN- are reported.

Finally, chapter 11 of the thesis contains a general summary and discussion of our results.

Chapter 1 Tumor Necrosis Factor

Introduction: Tumor Necrosis Factor 11

1. TUMOR NECROSIS FACTOR

Chapter foreword

Tumor necrosis factor (TNF) is a fascinating protein. Its name as well as its history are closely linked to cancer. However, despite the fact that everyone nowadays agrees that this is a protein with great anticancer potential, still no TNF-based therapy has been proven to prolong survival of human cancer patients. TNF is primarily a pro- inflammatory cytokine, and it has become clear that it plays a central role in inflammatory disorders such as sepsis, rheumatoid arthritis, inflammatory bowel disease and others. Hence, anti-TNF therapy has found its way to clinical practice, and often with success. Decades of research have now made clear that TNF has a remarkably broad range of often contradictory activities: it is involved in inflammation as well as anti-inflammation, angiogenesis as well as disruption of blood vessels, carcinogenesis as well as tumor necrosis, and cell survival as well as cell death. Therefore, TNF is often called a „Janus face‟, referring to the Roman god of gates, doors, beginnings and endings, who was usually depicted with two faces looking in opposite directions. Yet, despite its long history and decades of research, many aspects of its biology are still covered with mystery. Fortunately, there is great scientific interest in TNF: searching PubMed for “tumor necrosis factor” delivers more than 88000 articles, of which an overwhelming 6300 were published during the last year alone. In this chapter some general aspects of TNF biology will be discussed, with the emphasis on signaling to cell death and inflammation, and on its remarkable antitumor effect.

1.1. A brief history

The story of the discovery of TNF began with William B. Coley, a surgeon at Memorial Hospital in New York from 1892 to 1931. During the 19th century several physicians reported that cancer patients who suffered from severe bacterial infections sometimes had spontaneous regression of their tumors. Coley was the first to translate this observation into clinical practice. In the beginning he treated cancer patients with live bacteria, with some success, but due to difficulties to control the infection, Coley was forced to change his strategy. Instead of using live bacteria he now treated cancer patients with extracts of killed bacteria, which came to be known as Coley‟s toxins. These reproduced many of the symptoms of bacterial infection, such as fever and chills, but they could be administered without the risk of producing a real infection. Although Coley‟s approach was quite successful in the beginning of the 20th century, it was soon abandoned when chemotherapeutics and radiation therapy entered the field of cancer therapy 1. 12 Chapter 1

The interest in Coley‟s toxins however lived on in the laboratory. Investigators confirmed that some infectious agents and their products had anticancer effects in animals. In particular, they demonstrated that the injection of gram-negative bacteria could cause haemorrhagic necrosis of mouse tumors. In 1943 Murray J. Shear and his colleagues identified and purified the active component of gram-negative bacteria: a complex fat-and-sugar compound of the bacterial cell wall, now called lipopolysaccharide (LPS) 2. Although LPS could induce haemorrhagic necrosis of tumors in mice, in test-tube studies it could not inhibit or kill tumor cells directly, suggesting that its in vivo action was indirect and mediated by something in the host. More than 30 years later Lloyd J. Old and his colleagues identified in the serum of mice infected with Bacillus Calmette-Guérin (BCG) and 1 to 3 weeks later treated with LPS a factor that was capable of inducing haemorrhagic necrosis and regression of a methylcholanthrene A (methA) induced sarcoma in mice. Since the same factor had cytotoxic activity against certain tumor cell lines in vitro, among others the mouse fibrosarcoma cell line L929, they decided to name this factor “Tumor Necrosis Factor” 3. At the same time, another research group identified a serum factor that was a central mediator of cachexia, an infection- and cancer-associated chronic weight loss, and hence it was called “cachectine”. After purification and partial sequencing cachectine appeared to be identical to TNF 4. In 1984 and 1985 the genes for human and mouse TNF were cloned 5-7.

1.2. TNF and its receptors

TNF belongs to a superfamily of TNF protein homologues (with 15-20% identity to each other) consisting of 19 members that signal through 29 different receptors (Fig.1.1.). Unlike TNF, most members have specific activities and are specialized in one type of signal. For example, TNF-related apoptosis-inducing ligand (TRAIL) and FasL appear to be specialized in apoptosis. The human TNF gene is located on chromosome 6p21.3 and is mainly expressed by activated macrophages, natural killer (NK) cells and T lymphocytes, although many other cell types (e.g. fibroblasts, Kuppfer cells, smooth muscle cells, keratinocytes, astrocytes, and tumor cells) have been shown to contain TNF messenger RNA (mRNA) 8. TNF expression results in a 26kDa membrane-bound pro-peptide (pro- TNF) which associates into homotrimers. TNF is released from the cell surface after cleavage by the TNF--converting enzyme (TACE), a member of the a-disintigrin- and-metalloproteinase (ADAM) family 9, 10. The soluble active form of TNF is a 52kDa homotrimer 11.

TNF transduces its signals via two cell surface receptors: TNF-receptor 1 (TNF-R1) (or p55TNFR or CD120a), constitutively expressed on all cell types, and TNF-R2 (or Introduction: Tumor Necrosis Factor 13

p75TNFR or CD120b), of which the expression is inducible and appears to be restricted to immune and endothelial cells 12. TNF-R1 can be fully activated by both the membrane-bound and soluble forms of TNF, while TNF-R2 only responds fully to the membrane-bound form of the TNF homotrimer 13. Like TNF, these receptors can self-assemble into homomultimers. This ligand-independent assembly is mediated by an extracellular domain called preligand binding assembly domain (PLAD) 14. Both TNF-receptors can undergo ectodomain shedding to produce soluble TNF- receptors (sTNF-Rs). These sTNF-Rs can act as circulating decoy receptors and this is an important mechanism of negative regulation of the biological activity of TNF 15. Like TNF, TNF-R2 is cleaved by TACE 16. It is still unclear which proteolytic enzymes are responsible for the cleavage of TNF-R1, but recently it was shown that the shedding of TNF-R1 was greatly reduced in TACE-deficient neutrophils and macrophages 17. The cleavage of TNF-R1 is obviously an important step in the regulation of cellular TNF responsiveness as cleavage-resistant TNF-R1 mutations are linked with a dominantly inherited auto-inflammatory syndrome, called TNF-R1- associated periodic syndrome 18. Elevated levels of sTNF-Rs are also found in certain diseases, e.g. in rheumatoid arthritis high serum levels of sTNF-Rs are associated with a greater risk of mortality due to cardiovascular disease 19. While the extracellular region of both TNF receptors is very similar, their intracellular regions display no homology 20. The intracellular region of the TNF-R1 contains a death domain (DD), which is required for most TNF-R1-mediated signaling events, including gene induction, death-inducing signaling complex (DISC) formation and the activation of the acidic sphingomyelinase (aSMase) 21-23. The TNF-R1 also contains a second, more membrane-proximal, domain, called neutral sphingomyelinase domain (NSD), which is important for the recruitment and activation of the neutral sphingomyelinase (nSMase) 24, 25. The intracellular region of the TNF-R2 is much shorter and transduces signals through its TNF receptor associated factor (TRAF)- interacting motif (TIM) 26. This domain is required for the recruitment of TRAF family members, which in turn can associate with members of the mitogen-activated protein kinase (MAPK) and nuclear factor-B (NF-B) pathways, and anti-apoptotic factors such as the inhibitor of apoptosis protein (IAP) family members c-IAP1 and 2 22, 27.

14 Chapter 1

Fig. 1.1. The TNF superfamily. TNF superfamily numbers are indicated in green boxes, TNF receptor superfamily numbers in blue boxes. Most ligands are type-II transmembrane proteins (with carboxyterminal extracellular domain responsible for binding the receptor, aminoterminal intracellular domain, and a single transmembrane domain). LT and TL1A are secreted proteins. BAFF, EDA1, EDA2, APRIL and TWEAK are released from the cell surface by members of the Furin protease family; TNF, LT, RANKL and FasL are cleaved off by other proteases. Most receptors are type-I transmembrane proteins (extracellular N-terminus and intracellular C-terminus). BCMA, TACI, BAFF- R and XEDAR lack a signal peptide sequence and are considered type-III transmembrane proteins. OPG and DcR3 are secreted proteins. Encircled receptors can be found in soluble form and act as circulating decoy receptors. DD-containing receptors are represented in red, decoy receptors in green. The mouse homologues for the human TRAIL-R1 and TRAIL-R2, for the human TRAIL-R3, and for the human TRAIL-R4 are respectively TRAIL-R2, DcTRAIL-R1, and DcTRAIL-R2L and DcTRAIL- R2S. ** Do not belong to the TNF Superfamily. Adapted from Mocellin et al., 2005 22.

Introduction: Tumor Necrosis Factor 15

1.3. Regulation of TNF production

The production of TNF is regulated on the transcriptional, translational and post- translational level. The TNF gene promoter contains potential binding sites for several transcription factors, including AP-1, AP-2, NF-B, Ets, NFAT, cyclic AMP response element (CRE) and C/EBP, which allow the regulation of gene expression by many different stimuli 28-34. For example, in macrophages the induction of TNF by LPS requires the transcription factor NF-B 35. Besides LPS, also other microbial compounds and cytokines such as interleukin (IL)-1, IL-2, GM-CSF, and TNF itself can induce TNF production. Several anti-inflammatory agents, such as glucocorticoids and the cytokines IL-10 and transforming growth factor- (TGF-), function as inhibitors of TNF production 8. The 3‟untranslated region (3‟UTR) of the TNF mRNA contains an AU-rich element (ARE) that is important for the translational regulation of TNF 36. This ARE interacts with a variety of ARE-binding proteins that regulate the transport of mRNA, its decay in the cytoplasm and its translational efficiency 37. In nonstimulated macrophages, spliced TNF mRNA is retained in the nucleus by an ARE-dependent mechanism. Stimulation with LPS induces export of the mRNA into the cytoplasm in a Tpl2- and extracellular signal-related kinase (ERK)-dependent pathway 38. Once in the cytoplasm, the rate of mRNA decay is stringently controlled. Decay is mediated through binding of the zinc finger protein tristetraprolin (TTP) to the ARE and requires proteasomal activity 39, 40. Phosphorylation of TTP by MAPK-activated protein kinase-2 (MK2), a downstream target of p38 MAPK, leads to reduced affinity for the mRNA and increased mRNA stability 41. Recently it was shown that TNF mRNA stabilization also requires activation of the ERK pathway 40. Two other RNA binding proteins with high affinity for the TNF ARE have been identified: T cell restricted intracellular antigen-1 (TIA-1) and TIA-1-related protein (TIAR). TIA-1, and probably also TIAR, is involved in silencing at the level of translation 37, 42. Downstream of the ARE another mRNA-destabilizing element has been identified which is the target of a second mRNA degradation pathway that is independent of ARE-mediated decay 43. Finally, also post-translational regulation of TNF production may occur 44.

16 Chapter 1

1.4. TNF-R1 signaling

1.4.1. General aspects of TNF-R1 signaling

The initial step in TNF-R1 signaling is binding of the TNF homotrimer to the extracellular domain of pre-assembled receptor complexes. This induces a conformational change that results in the release of the inhibitory protein silencer of death domains (SODD) from the TNF-R1 intracellular DD. The function of SODD is to repress spontaneous TNF-R1 signaling by blocking the binding of adaptor proteins to the DD of TNF-R1 45. After SODD release, TNF-receptor associated DD protein (TRADD) binds to the DD of TNF-R1 and functions as a platform to recruit the additional adaptor proteins receptor-interacting protein (RIP)-1, via its DD, and TRAF, via an aminoterminal domain of TRADD 46, 47. Within minutes after TNF binding a membrane-bound complex is formed composed of TNF-R1, TRADD, RIP, TRAF-2 and c-IAP1 (Fig.1.2.). This complex is called complex I and induces activation of NF-B (see chapter 1.4.2.), but no apoptosis 48. Complex I is formed in cholesterol- and sphingolipid-enriched membrane microdomains, termed lipid rafts. It was shown that disruption of lipid rafts by cholesterol depletion results in a switch of TNF-induced NF-B activation to apoptosis 49. However, the role of lipid rafts in TNF-R1 signaling may be cell-type dependent. In primary mouse macrophages it was shown that lipid rafts are required for TNF-R1 signaling to ERK, but not to NF-B 50. In a human endothelial cell line disruption of lipid rafts resulted in inhibition of activation of phosphatidylinositol (PI) 3-kinase (PI3K), but not of NF-B signaling 51.

After a lag of several hours and internalization of complex I, a secondary complex is formed in the cytoplasm, which lacks TNF-R1 but includes Fas-associated DD protein (FADD) and the procaspases-8 and/or -10 (Fig.1.2.) 48. FADD is recruited to the complex by homophilic interaction of the carboxyterminal DD of FADD with the DD of TRADD, whereas the procaspases bind to FADD by homophilic interaction between the aminoterminal death effector domain (DED) of FADD and a DED in the prodomain of the procaspases 52. This secondary complex is called complex II or DISC and can initiate apoptosis (see chapter 1.4.4.). It was recently shown that DISC formation is critically dependent on endocytosis of the TNF-R1, and a domain, named TNF-R1 internalization domain (TRID), has been identified that is required for internalization of the receptor. The vesicles containing the internalized TNF-R1 are called TNF receptosomes 53.

Introduction: Tumor Necrosis Factor 17

Fig. 1.2. TNF-R1 signaling via two sequential complexes. After binding of TNF to TNF-R1, rapid recruitment of TRADD, RIP and TRAF2 occurs (Complex I). Subsequently TNF-R1, TRADD and RIP become modified () and the complex dissociates from TNF-R1. The liberated DD of TRADD and/or RIP now bind FADD, resulting in recruitment of procaspase-8/10 (Complex II). If NF-B activation triggered by complex I is successful, c-FLIP levels are sufficiently elevated to block the activation of caspase-8/10 by complex II. From Micheau et al., 2003 48.

1.4.2. The NF-B pathway

NF-B is a dimeric transcription factor that regulates the expression of many genes involved in inflammation and immunity. NF-B also upregulates the expression of several genes that participate in controlling cell survival, such as, Bcl-xL, cellular Fas- associated DD-like IL-1-converting enzyme (FLICE)-inhibitory protein (c-FLIP), proteins of the IAP family, and the Bcl-2 homologue A1, which are important to counteract the TNF-R1-induced apoptotic machinery 54-60. The term NF-B refers to a collection of homo- or heterodimers formed by members of the NF-B/Rel family, which consists of five members: NF-B1 (or p50) and its precursor p105, NF-B2 (or p52) and its precursor p100, RelA (or p65), RelB, and c- Rel. NF-B/Rel family members are involved in cellular responses to stimuli such as cytokines, for example TNF and IL-1, bacterial components, for example LPS, and some forms of physical stress, for example reactive oxygen species (ROS) or ultraviolet radiation 61. The NF-B/Rel proteins have a conserved Rel homology 18 Chapter 1

domain that mediates dimerization, nuclear translocation, DNA binding and interaction with members of the inhibitor of NF-B (IB) family 62. In unstimulated cells NF-B activity is inhibited by association with IB proteins, which mask its nuclear localization sequence and inhibit its DNA-binding activity 63. NF-B can be activated through two distinct pathways. Pro-inflammatory stimuli, such as LPS, TNF or IL-1, activate NF-B by the so-called canonical (or type 1) NF-B pathway, whereas non-inflammatory stimuli, associated with the development of lymphoid organs, have been shown to activate NF-B by an alternative (or type 2) pathway 64. The primary NF-B isoform induced by the canonical pathway is the NF-B1:RelA dimer, also called classical NF-B 65.

The activation of NF-B by TNF is mediated by the canonical IB kinase (IKK) complex (Fig.1.3.). This complex consists of a heteromer of two related IB kinases IKK1 and IKK2 (or IKK and IKK), the regulatory protein NF-B essential modulator (NEMO) (or IKK), a homodimer of the heat shock protein-90 (Hsp90), and two or three molecules of the Hsp90-associated protein cdc37 66-69. TRAF-2 plays an essential role in the recruitment of the IKK complex to the TNF-R1 signaling complex I, whereas RIP is essential for the activation of the IKK complex 70. RIP recruits the MAPK kinase kinase (MAP3K) TGF--activated kinase 1 (TAK1) to the complex and TAK1 cooperates with a second MAP3K, MAPK/ERK kinase kinase 3 (MEKK3), to activate the IKK complex 71-73. It has been shown that ubiquitination of RIP, probably by TRAF-2, is required for NF-B activation. The polyubiquitin chains (lysine 63-linked) on RIP may provide a docking site for the association with the TAK1-associated protein TAB2 74-76. Upon activation, the IKK complex phosphorylates the canonical IB proteins IB, IB and IB on two N-terminal serines, which then act as a docking platform for the ubiquitin ligase -transducin repeat-containing protein (-TRCP) 77, 78. Subsequent ubiquitination (lysine 48-linked chains) induces proteasome-mediated degradation of the IB proteins without affecting the bound NF-B dimer. The liberated dimer can then bind DNA at B sites and activate or repress gene transcription 65.

Introduction: Tumor Necrosis Factor 19

Fig. 1.3. Current model of NF-B activation by TNF. P: phosphorylation, Ub: ubiquitin(ation).

Next to signal-induced degradation of the inhibitory IB proteins, NF-B transcriptional activity is also regulated through crosstalk with other signaling pathways. For example, activation of the serine/threonine kinase Akt in response to IL-1 has been shown to increase the DNA-binding activity of NF-B by phosphorylation of the NF-B1 subunit 79. The PI3K/Akt pathway has also been implicated as a critical mediator for TNF-dependent activation of NF-B in certain cell types 80, 81. Furthermore, casein kinase II (CKII), IKK2, protein kinase A (PKA) and mitogen- and stress-activated kinase-1 (MSK-1) have been shown to directly phosphorylate RelA, thereby increasing its transcriptional activity 82-85. Also protein kinase C (PKC), glycogen synthase kinase 3 (GSK), IKK1, and the MAPKs p38 and ERK have been shown to regulate the transcriptional activity of NF-B, presumably by direct or indirect phosphorylation 81, 86-89. Finally, several studies have shown that acetylation of NF-B occurs and that NF-B can associate with histone 20 Chapter 1

acetyltransferases (HATs), e.g. cAMP response element-binding (CREB)-binding protein (CBP), and histone deacetylases (HDACs) 90-92.

Upon cell stimulation increases in nuclear NF-B activity can be detected within 10 minutes, and some NF-B-responsive promoters are induced almost immediately. One such early responding promoter is that of IB, which mediates a powerful negative feedback loop that is responsible for repression of NF-B activity upon removal of the stimulus 93. During chronic stimulation this may result in oscillatory NF-B activity 94. However, a second negative feedback loop mediated by IB, which is delayed and functions in anti-phase to IB, may counteract this oscillatory activity 95.

1.4.3. The MAPK pathway

MAPKs are serine/threonine kinases that respond to extracellular stimuli, mitogens, and regulate various cellular activities, including gene expression, cell division, differentiation, migration and survival/apoptosis. Extracellular stimuli lead to activation of a MAPK via a signaling cascade composed of MAPK, MAPK kinase (MAP2K or MKK) and MAPK3K (Fig.1.4.). A MAP3K that is activated by an extracellular stimulus phosphorylates a MAP2K on specific serine and threonine residues, and subsequently this MAP2K activates a MAPK through phosphorylation of specific serine and tyrosine residues. The activated MAPK can then activate its substrates through phosphorylation of specific serine and threonine residues. The MAPK family can be divided into three groups: p38 MAPK, which contains a Thr- Ala-Tyr motif, c-Jun NH2-terminal kinas/stress activated protein kinase (JNK/SAPK), which contains a Thr-Pro-Tyr motif, and ERK, which contains a Thr-Glu-Tyr motif 96, 97. TNF can activate all three MAPKs.

Introduction: Tumor Necrosis Factor 21

Fig. 1.4. Major MAPK cascades in mammalian cells. Adapted from Zhang et al., 2002 97.

 JNK

TNF-R1-induced activation of JNK occurs through a TRAF-2-dependent pathway (Fig.1.5.) 98-100. JNK can be activated by two MKKs, MKK4 and MKK7. TNF activates only MMK7, not MKK4 101. However, MMK4 appears to be required for optimal TNF-induced JNK activation 102. MKK4 and MKK7 are target of a variety of MAP3Ks, among these are apoptosis signal-regulating kinase-1 (ASK-1), TAK1 and MEKK1. All of these kinases have been shown to interact with TRAF-2 103-105. The role of MEKK1 in TNF-mediated JNK activation is unclear. One group reported that embryonic stem cells with a functional deletion of MEKK1 showed a total block of TNF-induced JNK activation, while another group reported that MEKK1 knockout macrophages showed normal JNK activation in response to TNF 106, 107. However, since germinal center kinase (GCK), a known activator of MEKK1, also interacts with TRAF-2, it is likely that, at least in certain cell types or conditions, MEKK1 is involved in JNK activation downstream of TNF-R1 108. The role of ASK-1 in TNF-R1 signaling is much better described. Overexpression of a dominant negative form of ASK-1 or genetic deletion of ASK-1 results not only in defective JNK activation but also in inhibition of apoptosis in response to TNF treatment 104, 109, 110. The current model for ASK-1 activation by TNF involves several critical steps, including the release of the inhibitory proteins thioredoxin and 14-3-3, TRAF-dependent 22 Chapter 1

dimerization/polymerization, and ASK-1 autophosphorylation 111, 112. Thioredoxin is a protein with a redox-active center composed of two cysteine residues. Only the reduced form of thioredoxin is able to interact with and inhibit ASK-1 113. Stimulation of TNF-R1 can result in a TRAF-2-dependent increase of ROS of mitochondrial origin 114. This increase of ROS leads to oxidation of thioredoxin, which is then no longer able to interact with ASK-1. The released ASK-1 can then associate with TRAF-2 to activate JNK 115. Besides thioredoxin, also 14-3-3, a family of dimeric phosphoserine-binding proteins, has been identified as an inhibitor of ASK-1 116. The release of 14-3-3 involves the ASK-interacting protein 1 (AIP1), a member of the Ras GTPase-activating protein family. In endothelial cells, AIP1 is localized to the plasma membrane, where it associates with TNF-R1. In response to TNF, AIP1 becomes phosphorylated at the 14-3-3 binding site by RIP, dissociates from the TNF-R1 and forms in the cytoplasm a complex composed of TRADD, TRAF-2, RIP and AIP1, called AIP1 complex. AIP1 then associates with ASK-1 close to its 14-3-3 binding site, leading to dissociation of 14-3-3 from ASK-1 and subsequent activation of JNK 117, 118. The JNK pathway plays an important role in TNF-induced apoptosis (see next chapter).

Fig. 1.5. Current model of JNK activation by TNF. P: phosphorylation. Introduction: Tumor Necrosis Factor 23

 p38

The activation of the p38 MAPK pathway by TNF occurs through a similar mechanism as the activation of JNK and also involves TRAF-2, AIP1, ASK-1 and MEKK1 108, 110, 118, 119. Nevertheless, there is clear evidence that upstream and downstream of the MAP3K level differences exist between JNK and p38 activation. For example, GCKs are unable to stimulate the p38 MAPK cascade 120, 121. Moreover, it has been shown that a deletion mutant of RIP, lacking its intermediate domain, abolished the TRAF-2-induced p38 activation, but failed to inhibit the activation of JNK 108. Also the MKKs involved appear to be different. While MKK4/MKK7 double knockout mouse embryonic fibroblasts (MEFs) were defective in the TNF-induced activation of JNK, the activation of p38 in these MEFs was largely unaffected 102. On the other hand, MEFs of MKK3-deficient mice showed a strong reduction in TNF- induced activation of p38, but not of JNK 122. The p38 pathway is an important mediator of the TNF-induced inflammation. The production of the inflammatory cytokines IL-1 and IL-6 upon TNF stimulation was almost completely blocked in MKK3 knockout MEFs 122. The p38 MAPK pathway acts by activation of a variety of transcription factors, among others activating transcription factor-2 (ATF-2), Elk-1, CREBs, growth arrest and DNA damage 153 (GADD153), MEF2A and MEF2C, and Sap-1a 123-126. The p38 pathway can also increase the stability of ARE-containing mRNAs through activation of its downstream target MK2 127. TTP, an ARE-destabilizing protein is phosphorylated by MK2, which results in decreased affinity of TTP for the ARE-containing mRNA and increased mRNA stability 41. Many inflammatory cytokines, including IL-1 and IL-6, have an ARE-sequence in the 3‟ UTR of their mRNA 128. Another important downstream target of the p38 pathway is cytosolic phospholipase A2 (cPLA2) (see chapter 1.4.5) 129. The p38 MAPK has also been implicated as a mediator of NF-B transactivation 88.

 ERK

TNF can also activate the ERK pathway, but it is still unclear by which mechanism this occurs 130. More than ten years ago a protein termed MAPK-activating DD protein (MADD) was identified as a link between the TNF-R1 and ERK activation, but very few publications on this protein have appeared ever since 131. Recently, it was shown that FADD, caspase-8 and c-FLIP are required for ERK activation in response to TNF, suggesting that ERK activation occurs at the level of complex II 132.

24 Chapter 1

1.4.4. TNF-induced cell death

TNF can induce two main forms of programmed cell death: (1) classical apoptosis and apoptosis-like cell death, characterized by chromatin condensation and fragmentation, membrane blebbing, and generation of apoptotic bodies, and (2) necrosis-like cell death, characterized mainly by a disruption of membrane integrity and a lack of the nuclear features of apoptosis (Fig.1.6.) 133.

Fig. 1.6. Characteristics of apoptosis and necrosis.

 Classical apoptosis

Caspases are cysteine proteases that cleave their target proteins at specific aspartic acid residues. They play a central role in apoptosis through a so-called caspase cascade. There are two types of apoptotic caspases, called initiator caspases, e.g. the human caspase-2, -8, -9, -10 and -12, and effector caspases, e.g. the human caspase-3, -6, -7 and -14. The mouse genome lacks a counterpart of human caspase-10 134. In most cells, caspases are constitutively present as inactive procaspases. A procaspase consist of an aminoterminal prodomain, followed by a large p20 subunit and a carboxyterminal small p10 subunit. Activation of the caspase requires a first cleavage between the p10 and p20 subunits and a second cleavage to remove the prodomain. Two p10 and two p20 subunits associate then to form the active caspase. Since both cleavages required for activation occur at conserved aspartic acid residues, Introduction: Tumor Necrosis Factor 25

procaspases are a substrate of other caspases or can self-activate. Upon activation, initiator caspases can activate effector caspases, which can then cleave other protein substrates within the cell at specific aspartic acid residues, resulting in the apoptotic process. The best characterized mechanism of TNF-induced apoptosis is mediated by the cytoplasmic complex II (Fig.1.7.). Procaspase-8 is recruited to this complex by interaction of one of the two DEDs in the prodomain of procaspase-8 with the DED of FADD 52. This complex is also called DISC, in analogy with the FADD- and caspase- 8-containing complex that is formed upon stimulation of Fas and TRAIL receptors. The second DED in the prodomain of caspase-8 allows clustering of procaspase-8, followed by self-activation of the caspase. Caspase-8 can then activate the effector caspases-3 and -7, which in turn cleave specific substrates. Besides FADD, a second DD-containing adaptor protein has been identified that can recruit caspases downstream of TNF-R1. This adaptor protein, called RIP-associated ICH-1/CED-3 homologues protein with DD (RAIDD), can associate via its carboxyterminal DD with RIP, and via its aminoterminal caspase activation and recruitment domain (CARD) with the CARD in the prodomain of caspase-2, allowing autocatalytic activation of caspase-2 135. A characteristic substrate of effector caspases, in particular caspase-3, is inhibitor of caspase-activated DNase (ICAD). Cleavage of ICAD leads to activation of the caspase-activated DNase (CAD), which then degrades the chromosomal DNA into nucleosomal units (multimers of approx. 180 base pairs), a hallmark of apoptotic cell death 136, 137. Other important substrates of caspases are poly(ADP-ribose) polymerase (PARP), p21-activated kinase 2 (PAK2), Rho kinase-I (ROCK-I), actin, gelsolin, and nuclear lamins 138-145. PARP is a nuclear enzyme that mediates repair of DNA single strand breaks via the activation and recruitment of DNA repair enzymes. Lamins are structural components of the nuclear envelope, and cleavage of nuclear lamins during apoptosis causes nuclear fragmentation and chromatin condensation. Caspase- mediated cleavage of PAK2 produces an activated fragment, which may be involved in inducing certain morphological changes that are characteristic for apoptosis 144. Cleavage of ROCK-I by caspase-3 also generates a truncated active form. The activity of ROCK proteins is necessary and sufficient for the formation of membrane blebs during apoptosis and for re-localization of fragmented DNA into blebs and apoptotic bodies 146.

In certain cell types, called type I cells, the death receptor induced activation of caspase-8 alone is sufficient to robustly activate caspase-3 and the execution phase of apoptosis. However, in most cell types, called type II cells, e.g. hepatocytes, caspase- 8-mediated activation of caspase-3 is inefficient and the apoptotic process therefore depends on a mitochondria-dependent amplification loop (Fig.1.7.) 147, 148. In type II cells, active caspase-8 can cleave the Bcl-2 family member Bid 149, 150. The site of action of Bcl-2 family members is the outer mitochondrial membrane, where they play an important role in controlling the release of pro-apoptotic proteins, in particular 26 Chapter 1

cytochrome c, from the mitochondria. Bcl-2 family members can be either pro- or

anti- apoptotic. Anti-apoptotic Bcl-2 family members, such as Bcl-2 and Bcl-xL, can neutralize pro-apoptotic family members, such as Bax, Bak, Bad and Bid, by forming heterodimers. Therefore, the balance between the pro- and anti-apoptotic members will determine the sensitivity of a cell to apoptosis 151. Cleavage of Bid by caspase-8 produces a potent pro-apoptotic carboxyterminal fragment of Bid, called truncated Bid (tBid). tBid translocates to the outer mitochondrial membrane, where it triggers the homo-oligomerization of Bax and Bak, resulting in the release of cytochrome c and second mitochondria-derived activator of caspases (Smac)/direct IAP-binding protein with low pI (DIABLO) 152-155. tBid has also been shown to interact with components of the mitochondrial permeability transition pore, resulting in mitochondrial membrane depolarization 156, 157. It has been reported that this mitochondrial membrane transition is essential for TNF-induced apoptosis in primary rat hepatocytes 158. The release of cytochrome c from mitochondria into the cytosol allows the formation of a caspase-3-activating complex, called apoptosome. This apoptosome is a wheel-like protein structure, consisting of an oligomer of apoptotic protease activating factor-1 (Apaf-1), cytochrome c, caspase-9, and ATP or 2‟-deoxy ATP (dATP) 159. Apaf-1 contains a CARD that interacts with a CARD in the prodomain of caspase-9. The apoptosome can then cleave and activate caspase-3 160, 161. Complementary to the actions of the apoptosome, Smac/DIABLO binds and antagonizes IAPs 155, 162. Activated caspase-3 is able to activate caspase-8, thereby creating a positive feedback loop 163.

Introduction: Tumor Necrosis Factor 27

Fig. 1.7. Cell death pathways induced by TNF. Note that not all pathways may be induced simultaneously. Black arrows: apoptotic pathway including mitochondrial amplification loop. Green arrows: lysosomal death pathway. Blue arrows: necrotic death pathway.

 Lysosomal death pathway

Another apoptosis pathway activated by TNF is the lysosomal death pathway, which is mainly mediated by cathepsins (Fig.1.7.). Cathepsins are proteases of which the activity is usually confined to the lysosomes. Like caspases, they become activated after proteolytic cleavage. The hallmark of the lysosomal death pathway is the lysosomal membrane permeabilization that results in the release of active cathepsins into the cytosol. It has been shown that TNF can induce the release of the lysosomal cysteine protease cathepsin B in certain cell types (e.g. hepatocytes, MEFs, and a number of cancer cell lines) and that cathepsin B is an important mediator of TNF- induced apoptosis in these cells 148, 164, 165. Moreover, mice deficient for cathepsin B are resistant to TNF-induced hepatocyte apoptosis and liver injury 166. Active cathepsin B can cleave several substrates involved in apoptosis, such as caspases and Bid 167-169. However, it is currently unclear whether Bid and caspases act really downstream of the released cathepsin B. Werneburg and colleagues reported that, in cultured hepatocytes, TNF-induced lysosomal membrane permeabilization required caspase-8, Bid, and cathepsin B itself 170, 171. They suggested that a truncated form of 28 Chapter 1

Bid (different from tBid) is able to associate with the lysosomal membrane and mediate the release of cathepsin B in analogy with the release of cytochrome c from the mitochondria. More recently, Gyrd-Hansen and co-workers reported that TNF- induced lysosomal membrane permeabilization and apoptosis in MEFs critically depended on caspase-9 172. Other factors implicated in the TNF-induced lysosomal death pathway are the sphingomyelinases (SMases). TNF can activate both nSMase and aSMase (see next chapter). SMases are enzymes that catalyze the hydrolysis of sphingomyelin, an important component of cell membranes, to phosphocholine and ceramide. It was shown that lysosomal membrane permeabilization observed in cultured hepatocytes after stimulation with TNF is dependent on factor associated with nSMase activation (FAN), an adaptor protein that interacts with the TNF-R1 and is required for nSMase activation 171, 173. Interestingly, it was recently shown that an internalization-defective TNF-R1 which is unable to assemble a DISC still allowed for residual activation of caspase-3 and cell death. This activation was independent of caspase-9, but was critically dependent on interaction of the TNF-R1 with FAN 174.

Also the aspartyl protease cathepsin D has been implicated as a mediator of TNF- induced cell death 175. The activation of aSMase upon TNF-R1 stimulation results in the production of ceramide in the endolysosomal compartment. Heinrich and colleagues showed that ceramide can interact with cathepsin D, followed by proteolytic activation of the cathepsin 176. They further reported that TNF and ceramide induce the translocation of active cathepsin D from the endosomal compartment to the cytosol in HeLa cells, and that cathepsin D can cleave Bid, leading to activation of the mitochondrial apoptosis pathway. Moreover, they showed that the TNF-induced cleavage of Bid was absent in cathepsin D- or aSMase-deficient cells. Interestingly, translocation of cathepsin B was not observed after TNF or ceramide treatment of HeLa cells 177.

 Necrotic cell death

TNF can also induce a necrosis-lie cell death that depends on the production of ROS (Fig.1.7.). Evidence for a caspase-independent TNF-induced cell death comes from experiments with cells or mice treated with the pan-caspase inhibitor zVAD-fmk (which also inhibits cathepsin B 178): L929 mouse fibrosarcoma cells or human neutrophils treated with TNF plus zVAD-fmk display augmented, instead of inhibited, TNF-induced cell death; and treatment of mice with TNF plus zVAD-fmk exacerbated the TNF-induced toxicity by enhancing oxidative stress and mitochondrial damage 133, 179, 180. In L929 cells, the augmented cell death after TNF and zVAD-fmk treatment could be reversed with a mitochondrial permeability transition inhibitor or a PARP inhibitor 180. Furthermore, depletion of reduced glutathione can sensitize cultured mouse hepatocytes to TNF-induced cell death, suggesting that ROS play an important role in TNF-induced liver injury 181. ROS, Introduction: Tumor Necrosis Factor 29

including superoxide anion, hydrogen peroxide and hydroxyl radicals, are normally rapidly eliminated by antioxidant enzymes, such as superoxide dismutases (SODs), catalase, glutathione peroxidases and peroxiredoxins 182. However, in certain circumstances, such as in cells treated with TNF and zVAD-fmk, excessive ROS production by the mitochondria may occur 183. This leads to DNA damage, followed by the activation of the DNA repair enzyme PARP-1. Excessive action of PARP-1 consumes high amounts of energy, leading to ATP depletion, thereby creating conditions that favor necrosis instead of apoptosis 184, 185. It is currently unclear how TNF triggers the mitochondria-dependent production of ROS, but the Bcl-2 family may play an important role herein 186, 187. It was reported that Bid is required for the majority of TNF-induced mitochondrial ROS production in cultured mouse hepatocytes 188. TNF can also induce ROS production by NADPH oxidases. NADPH oxidases are membrane-bound enzyme complexes specialized in generating superoxide from

NADPH and O2. Recently, the NAPDH oxidase components Nox1 and the small GTPase Rac1 were shown to play an important role in TNF-induced necrotic cell death of mouse fibroblasts 189, 190. Interestingly, in endothelial cells an NADPH oxidase consisting of Nox1, Nox2, Nox4, Nox5, p22phox, p47phox and Rac1, localizes (together with eNOS and nSMase; see further) to caveolae, a type of membrane rafts. Perturbation of these caveolae by cholesterol-sequestering was shown to inhibit TNF- induced ROS production 191.

 Regulation of TNF-induced cell death

In contrast to other death receptors, like Fas or TRAIL receptors, TNF-R1 only signals to cell death in certain circumstances. Most often, TNF-R1 signaling results in the transcription of inflammatory genes, suggesting that the TNF-R1 also provides a mechanism to counteract the various TNF-induced cell death pathways. Consistent with this, mice deficient in RelA or other components involved in NF-B activation are embryonically lethal or die early after birth due to massive TNF-dependent liver failure 192-195. Thus, the NF-B pathway plays an essential role in blocking the cell death machinery induced by TNF. NF-B activation induces the expression of many 54-60 anti-apoptotic factors, including IAP family members, Bcl-xL, c-FLIP and A1 . c- FLIP is recruited to complex II, where it inhibits caspase-8 activation. So, when complex I or other exogenous factors trigger sufficient NF-B activation, gene expression of c-FLIP is sufficiently induced to block the activation of caspase-8 by complex II 48. On the other hand, once the apoptotic program is triggered, caspases interfere with the activation of NF-B by cleavage of several components utilized by this pathway, including RIP, IKK2, NF-B1, RelA and Akt 196-200. RIP is already cleaved by caspase-8 during the initiation phase of apoptosis, thereby blocking the anti-apoptotic NF-B response 201. 30 Chapter 1

It was reported that TNF induced the accumulation of ROS in RelA-deficient MEFs, but not in wild-type cells. This suggests that another survival mechanism induced by NF-B is inhibition ROS accumulation 202. Notably, manganous SOD (MnSOD), which acts as a scavenger of superoxide radicals, is an NF-B-dependent target gene of TNF 203, 204.

JNK is an important modulator of TNF-induced apoptosis. JNK activation can have both pro- and anti-apoptotic effects. TRAF-2-deficient MEFs, which have largely intact NF-B signaling but reduced JNK activation, and JNK1 or JNK2 knockout MEFs display increased sensitivity to TNF-induced apoptosis 100, 205. On the other hand, ASK1 knockout MEFs, which have normal rapid TNF-induced JNK activation, but reduced prolonged activation, are significantly protected against apoptosis 109. This suggests that transient JNK signaling mediates anti-apoptotic effects, while sustained JNK signaling is associated with apoptosis. Consistent with this, it was found that the transient activation of JNK is inhibited in MKK7-deficient MEFs, while the prolonged activation of JNK, which correlates with apoptosis induction, is blocked by caspase inhibitors 102, 206. The length of JNK activation is also controlled by NF-B; JNK activation by TNF is transient in wild-type fibroblasts, but prolonged in cells deficient for either IKK2 or RelA 207. The NF-B target genes A20, GADD45 and X-chromosome-linked IAP (XIAP) have been shown to attenuate TNF-induced JNK activation upon ectopic expression 207-209. GADD45 can directly bind with MKK7, thereby blocking its catalytic activity 210, 211. The role of A20 is less clear. A20-deficient MEFs displayed unaltered TNF-induced JNK activation, suggesting that this protein is not involved in the attenuation of JNK activation, or a yet to be identified compensation mechanism exists 212. TNF-induced ROS production may induce the sustained activation of JNK through oxidation of the ASK-1 inhibitor thioredoxin. Since NF-B can inhibit the TNF-induced accumulation of ROS, presumably by upregulation of MnSOD, this forms another mechanism by which NF- B could inhibit prolonged JNK activation 202, 213. The anti-apoptotic effect of JNK signaling is mainly mediated by activation and increased expression of the transcription factor JunD, a component of the dimeric transcription factor AP-1. JunD can collaborate with NF-B to increase the expression of anti-apoptotic genes such as cIAP2. Interestingly, it was observed that the TNF- induced activation of Akt was reduced in JNK1/2 double knockout MEFs 214. JNK can also induce the expression of pro-apoptotic genes by activation of the transcription factor c-Jun, another component of AP-1. c-Jun/AP-1 mediates the transcription of Bim, a pro-apoptotic member of the Bcl-2 family 215. JNK has been shown to phosphorylate Bcl-2 family members. This phosphorylation inhibits the anti-apoptotic family member Bcl-2, but activates the pro-apoptotic family members Bim and Bmf 216-218. Moreover, it has been reported that the TNF-induced activation of MKK7/JNK can lead to the caspase-8-independent cleavage of Bid. The cleavage Introduction: Tumor Necrosis Factor 31

product, termed jBid, can translocate to the mitochondria and trigger the release of Smac/DIABLO, but not of cytochrome c. It is proposed that the released Smac/DIABO then interferes with the cIAP1/TRAF-2 complex to allow activation of caspase-8 219. However, evidence for an inhibitory association of cIAPs and caspase-8 is still lacking 220. More recently, it was shown that in hepatocytes TNF induces JNK- mediated phosphorylation and activation of the E3 ubiquitin ligase Itch, which then specifically ubiquitinates the caspase-8 inhibitor c-FLIP, thereby inducing its degradation 221, 222.

1.4.5. Other TNF-induced pathways

 Phospholipase A2

Phospholipase A2 (PLA2) was one of the first identified downstream targets of TNF 223. cPLA2 is an important regulator of the inflammatory process. It is a Ca2+ dependent lipase that specifically recognizes the sn-2 acyl bond of phospholipids and hydrolyzes this bond, thereby releasing arachidonic acid and lysophospholipids. cPLA2 does not require Ca2+ for its catalytic activity, but elevated intracellular Ca2+ concentrations are required to induce its translocation to cellular membranes (including plasma membrane, nuclear envelope, and endoplasmatic reticulum), where its substrate is 224. The activity of cPLA2 is also regulated by phosphorylation of the serine residues Ser505 and Ser727. The TNF-R1-induced activation of cPLA2 critically depends on the p38 MAPK pathway, which mediates phosphorylation of cPLA2 on both Ser505 and Ser727 129, 225. The ERK pathway may also contribute to cPLA2 activation 226. As translocation of cPLA2 to cellular membranes is stimulated by an increase of intracellular Ca2+ concentration and TNF-R1 stimulation does not induce such an increase, it is to be expected that TNF alone will only induce a moderate activation of cPLA2 226, 227. However, very early during the inflammatory process, G-protein-coupled receptors (GPCRs), such as the histamine H1 receptor, become activated on endothelial cells, leading to activation of phospholipase C (PLC) followed by an increase of cytosolic free Ca2+ levels by releasing the Ca2+ from endoplasmatic reticulum stores and entry of extracellular Ca2+. This rise in cytosolic Ca2+ concentration supports then the TNF-induced activation of cPLA2 later on during the inflammatory process 228. In addition, it has been reported that TNF can induce gene expression of cPLA2 229. Free arachidonic acid produced by cPLA2 is rapidly processed by cyclooxygenases (COXs) and lipoxygenases (LOXs) to eicosanoids (Fig.1.8.). Eicosanoids are paracrine and autocrine signaling molecules that play an important role in inflammation, hemostasis and the regulation of vascular tone. COXs convert arachidonic acid to prostaglandin (PG) D2, E2 and F2, prostacyclin (PGI2) and thromoboxane (TX), while 5-LOX converts arachidonic acid to leukotriens (LT) 230. PGs are produced by all cells at a basal level, but their production is strongly induced 32 Chapter 1

231, 232 during inflammation by upregulation of COX-2 expression by TNF . PGD2 and

PGE2 both act as vasodilators. TXs are produced by activated platelets. TXs stimulate platelet aggregation and induce vasoconstriction. The effects of TXs are balanced by

PGI2 produced in endothelial cells, which inhibits platelet aggregation and induces vasodilation. LTs are produced in leukocytes, such as granulocytes, mast cells and

monocytes, and are important mediators of inflammation. LTB4, for example, has a chemotactic and anti-apoptotic effect on neutrophils 230, 233. An increased production

of eicosanoids, especially of PGE2 and LTB4, has been observed in many types of cancer. Moreover, inhibitors of COX, LOX, and eicosanoid signaling have been reported to induce apoptosis of certain cancer cell lines, suggesting that eicosanoids can confer survival signals to these cells 234-239. Importantly, it has also been shown that inhibitors of arachidonic acid metabolism can sensitize cancer cells to TNF- induced apoptosis 240, 241.

Fig. 1.8. Biosynthesis and important properties of eicosanoids. Major sites of eicosanoid synthesis are indicated in italic. The immune cells wherein leukotriens are synthesized include mast cells, neutrophils, eosinophils, basophils, monocytes and alveolar macrophages.

cPLA2 may also play an important role in the TNF-induced ROS production. Cauwels and colleagues demonstrated that the zVAD-fmk-induced sensitization of mice to TNF-induced lethality, which is associated with excessive ROS production and Introduction: Tumor Necrosis Factor 33

mitochondrial damage, could be completely reverted by either antioxidant treatment or inhibition of PLA2 179. How PLA2 activity leads to ROS production is not fully understood at this moment, but several reports have implied that the oxidative metabolism of AA into eicosanoids would generate ROS 242-244. Furthermore, it was shown that PLA2 may also trigger mitochondrial ROS production 243, 245.

 NO synthases

When TNF is administered to animals in a sufficiently high dose, it induces a lethal shock syndrome. This shock syndrome is characterized by a severe hypotension that is induced by the vasodilator nitric oxide (NO) 246, 247. NO is produced by NO synthases (NOSs) from the terminal nitrogen atom of L-arginine in the presence of NADPH and

O2. Three isoforms of NOS have been identified: neuronal NOS (nNOS) (or NOS-1), inducible NOS (iNOS) (or NOS-2), and endothelial NOS (eNOS) (or NOS-3). eNOS is constitutively expressed in endothelial cells (and other cells) and is activated by an increase in intracellular Ca2+ concentration and by phosphorylation. iNOS on the other hand does not require an activation step, but its activity is controlled at the transcriptional level. iNOS is predominantly induced in immune cells and cells of the cardiovascular system 248-250. NO is a short-lived signaling molecule that is freely diffusible across membranes. As it is a free radical, it rapidly reacts with proteins and other molecules. One of the main targets of NO is soluble guanylate cyclase (sGC). Nitrosylation of the haem-iron in sGC results in a strong activation of the enzyme. Activated sGC produces the second messenger cGMP, which mediates many of the cardiovascular effects induced by NO, including vasorelaxation, increase of vascular permeability, as well as inhibition of platelet aggregation and anti-oxidant effects 249. NO can also react with superoxide to produce peroxynitrite, a powerful oxidant which mediates the cytotoxic effects of NO, such as DNA damage, tyrosine nitration, LDL oxidation, and inhibition of mitochondrial respiration. The latter reaction is one of the effector mechanisms used by macrophages and neutrophils to kill pathogens 251. TNF induces the production of NO by activating eNOS and by inducing the expression of iNOS. The induction of iNOS gene expression by TNF, or other inflammatory stimuli, is critically dependent on NF-B 252-254. However, the JNK, ERK, and p38 MAPK pathways may also be involved, depending on the cell type and stimuli investigated 255-257. The activation of eNOS by TNF requires sequential action of nSMase, sphingosine kinase-1 (SK-1), PI3K and Akt 258, 259. Akt activates eNOS by phosphorylation of a specific serine residue 260, 261. Since it has been shown that TNF- R1, nSMase as well as eNOS are present in caveolae, and that disruption of caveolae can interfere with the TNF-induced activation of PI3K, it is most likely that the activation of eNOS by TNF also occurs in these membrane raftse 51, 262, 263. TNF also negatively regulates eNOS by decreasing its promoter activity and by destabilizing its mRNA 264-266. 34 Chapter 1

The NO production induced by TNF may have, as many aspects in TNF biology, both beneficial and detrimental effects. Cauwels and colleagues demonstrated that inhibition of sGC completely protected mice against the TNF-induced shock and lethality. However, when mice were treated with a non-selective NOS inhibitor, no protection against TNF-induced toxicity and death could be observed. This suggests that NO also mediates an sGC-independent protective effect in this lethal shock model 267. More recently, Cauwels and colleagues reported that this beneficial effect of NO consists of tempering the TNF-induced oxidative stress, in particular the lipid peroxidation in liver and kidney, which was shown to be largely dependent on PLA2 268. One of the mechanisms by which NO can protect against oxidative stress is by neutralizing superoxide. An excess of superoxide or equimolar concentrations of NO and superoxide promote the formation of toxic peroxynitrite and lipid peroxidation, whereas an excess of NO neutralizes superoxide and inhibits lipid peroxidation 269. NO can also protect against apoptosis, e.g. in endothelial cells and hepatocytes 270. It has been shown that NO can inhibit caspases, including caspase-3, -8, and -9, by S- nitrosylation of specific cysteine residues 271-274. Furthermore, low, physiological concentrations of NO have been shown to inhibit mitochondrial permeability transition and cytochrome c release, whereas higher concentrations had adverse effects 275. The pro-apoptotic effects of NO mainly comprise the induction of DNA and mitochondrial damage 270. In addition, NO may inhibit NF-B by S-nitrosylation of the NF-B1 subunit 276.

 SMases

As discussed above, engagement of TNF with TNF-R1 has been reported to activate both nSMase and aSMase, resulting in sphingomyelin degradation and formation of ceramide. aSMase has a low pH optimum and is mainly present in the endosomal- lysosomes compartment. However, it has been shown that stimulation of cells can result in the translocation of aSMase to the outer leaflet of the cell membrane 277, 278. The TNF-triggered activation of aSMase is dependent on both TRADD and FADD, placing the activation of aSMase somewhere at the level of complex II 23, 279. In addition, it was shown that activation of aSMase is dependent on TRID, and it was suggested that aSMase activation occurs after fusion of TNF receptosomes with aSMase-containing trans-Golgi vesicles 53. The TNF-induced activation of aSMase results in the production of ceramide in the endolysosomal compartment, which can lead to activation of cathepsin D and subsequent activation of the mitochondrial apoptosis pathway 176, 177. The importance of this pathway in vivo is demonstrated by the fact that aSMase knockout mice are resistant to TNF-induced hepatocyte apoptosis and liver injury 280. nSMase has a neutral pH optimum and is mainly associated with the plasma membrane. The activation of nSMase by TNF requires the adapter protein FAN, which directly binds to a short motif in the intracellular part of the TNF-R1 termed NSD 173. A few years ago, the scaffolding protein receptor for activated C- Introduction: Tumor Necrosis Factor 35

kinase 1 (RACK1) was identified as an interaction partner of FAN, but it is still unclear how this protein is involved in the regulation of nSMase 281, 282. Expression of a dominant negative form of FAN in human fibroblasts abrogated the TNF-induced ceramide generation and reduced caspase processing and apoptosis, suggesting that nSMase rather than aSMase is responsible for ceramide production and regulation of cell death in these cells 283. Moreover, expression of a dominant negative FAN in hepatocytes blocked the TNF-induced activation of caspase-8, lysosomal permeabilization, and subsequent release of cathepsin B 171. Another mechanisms by which ceramide may contribute to apoptosis is by inducing clustering of receptors. It has been shown that activation of Fas induced the translocation of aSMase to the cell membrane, followed by ceramide production in lipid rafts and clustering of Fas within these rafts. These events were shown to be essential for Fas-induced apoptosis of lymphocytes and hepatocytes 277, 278, 284-286. Increased amounts of ceramide or its derivate ganglioside GD3 are also found in the mitochondria of TNF-treated cells, where it may induce mitochondrial permeability transition, cytochrome c release and ROS production 287-289.

 The PI3K/Akt pathway

Several reports have shown that TNF is capable of activating the PI3K/Akt pathway and that this pathway may play an important role in the activation of NF-B 80, 290-292. In agreement with this it was shown that re-expression of phosphatase and tensin homolog (PTEN), a phosphatase that counteracts PI3K by dephosphorylating the 3‟- position of the inositol ring of inositol phospholipids, inhibits the transcription of NF- B-driven genes in certain cancer cells 293, 294. It is still unclear how TNF can trigger the activation of PI3K, but it most likely involves the sequential activation of nSMase and SK-1 259, 295. The PI3K/Akt pathway is essentially a survival pathway and will be discussed elsewhere in this thesis (see chapter 3).

1.5. TNF-R2 signaling

As mentioned above, TNF-R2 only fully responds to the membrane-bound form of TNF, and thus requires direct cell-to-cell contact to become activated. Upon activation, TRAF-2 is recruited to the intracellular TIM of TNF-R2. TRAF-2 then recruits TRAF-1 and cIAP1 and -2 to the receptor complex and mediates further signaling that may lead to the activation of NF-B and JNK 26, 27, 296. However, it is still unclear which signaling events occur downstream of TNF-R2.

Although TNF-R2 has no DD and cannot directly induce the activation of caspases or other cell death effectors, it may regulate the apoptotic process through crosstalk with TNF-R1. cIAP1, that is recruited to the TNF-R2 signaling complex, has ubiquitin 36 Chapter 1

protein ligase (E3) activity 297. It was shown that TNF-induced activation of TNF-R2 leads to cIAP1-dependent ubiquitination and proteasomal degradation of TRAF-2 and ASK-1 298. Degradation of ASK-1 may be an important mechanism to prevent prolonged TNF-induced MAPK signaling 299. Depletion of TRAF-2 on the other hand may significantly sensitize for TNF-R1-mediated apoptosis 300. Furthermore, TNF-R2 has been shown to induce cell death through expression of membrane-bound TNF and subsequent activation of TNF-R1 301. This mechanism may play an important role in the apoptosis of activated T cells 302, 303.

1.6. TNF in inflammation and immunity

Inflammation is the body‟s response to infection or other forms of tissue injury. It serves to keep the infection localized, to trigger and amplify the innate, and later adaptive, immune responses, and to promote the repair of the injured tissue. The five basic symptoms of acute inflammation are swelling, pain, heat, redness of the skin, and, possibly, loss of function of the organ or tissue involved. These symptoms are mostly due to vasodilation and increased permeability of the local blood vessels and infiltration of plasma proteins and leukocytes to the site of infection. Inflammation at a site of infection is in most cases initiated when tissue macrophages respond to pathogens. Small molecular motifs, termed pathogens-associated molecular patterns (PAMPs), found consistently on the surface of pathogens are recognized by pattern recognition receptors (PRRs) on tissue macrophages and other cells. For example, LPS, a major component of the outer membrane of Gram-negative bacteria, is specifically recognized by the PRR Toll-like receptor 4 (TLR4) 304. Tissue macrophages that are activated in this way very rapidly produce PGs, LTs, and platelet-activating factor (PAF), which together are called lipid mediators of inflammation because they are produced through enzymatic pathways that degrade membrane phospholipids. Subsequently, activated macrophages synthesize and secrete chemokines and cytokines, in particular TNF 305-307. Another way in which recognition of pathogens triggers an inflammatory response is through activation of the complement system. Several complement components can bind directly to PAMPs on the surface of pathogens, thereby inducing a complement cascade that eventually results in the production of the cleavage product C5a, which is a powerful small peptide mediator of inflammation. The actions of C5a include increasing the vascular permeability, inducing the expression of adhesion molecules on endothelial cells and neutrophils, attracting circulating granulocytes and monocytes, and activating neutrophils and local mast cells 308, 309. Activated mast cells will in turn release granules containing the inflammatory molecule histamine and TNF 310, 311.

Endothelial cells play a central role in the inflammatory process. Histamine and the lipid mediators that are released very early in the inflammatory process signal via Introduction: Tumor Necrosis Factor 37

GPCRs that are present on endothelial cells. These mediators are responsible for the so-called type I activation of endothelial cells. Triggering of the involved GPCRs results in activation of phospholipase C (PLC), which catalyzes the release of inositol-1,4,5-triphosphate (InsP3) from the membrane lipid PI-4,5-biphosphate. InsP3 induces the release of Ca2+ from endoplasmatic reticulum stores. The cytosolic free Ca2+ concentrations may be further increased by entry of extracellular Ca2+. The elevated cytosolic Ca2+ levels allow then the activation of cPLA2 and, through binding with calmodulin, of eNOS. The arachidonic acid produced by cPLA2 is rapidly converted to PGI2, which synergizes with NO to induce vasodilation. Triggering of the GPCRs also induces activation of myosin light chain (MLC) kinase (MLCK), which phosphorylates MLC. Phosphorylated MLC initiates contraction of actin filaments that are attached to tight junctions and adherens junctions, resulting in the formation of gaps between adjacent endothelial cells, which enhances leakage of plasma proteins and facilitates the transmigration of neutrophils. Type I activated endothelial cells also display P-selectin and a membrane-bound form of PAF on the luminal plasma membrane, which together initiate neutrophils extravasation (Fig.1.9.). The endothelial cell membranes at the cell-cell junctions are enriched in platelet-endothelial cell adhesion molecule (PECAM) and CD99 (or CD-Like 2 in mice 312), two proteins that are essential for the transmigration of neutrophils and monocytes through homophilic interactions (see Pober et al., 2007 and references therein 228).

38 Chapter 1

Fig. 1.9. Leukocyte extravasation and cell adhesion molecules involved in the adhesion of leukocytes to endothelial cells. Adapted from Ulbrich et al., 2003 313.

Signals by GPCRs last for only 10-20 minutes, after which the receptors become desensitized. A more persistent form of endothelial cell activation, called type II activation, is induced by cytokines. TNF is the prototypic mediator of type II activation (Fig.1.10). Since the pro-inflammatory effects of TNF involve new gene transcription mediated by NF-B and AP-1, type II activation of endothelial cells requires a longer time to be initiated 228. Type I agonist, such as histamine, have also been shown to induce TACE-dependent shedding of TNF-R1 from the surface of endothelial cells, thereby desensitizing these cells to TNF. TNF-R1 molecules stored in the Golgi apparatus are then brought to the cell surface to restore the sensitivity 314. This may be another reason for the temporal lag between type I and type II activation. Similar to type I activation, type II activation leads to increased blood flow, vascular leakage and leukocyte recruitment. TNF increases the blood flow by activating cPLA2 and eNOS through phosphorylation, and upregulation of COX-2 expression, 223, 231, 259 resulting in a strong production of the vasodilators PGI2 an NO . TNF has also been shown to induce the formation of gaps between adjacent endothelial cells, thereby increasing the endothelial permeability and leakage of plasma proteins to the site of inflammation. How TNF increases endothelial permeability is not yet completely understood but Petrache and colleagues reported that TNF induces opening of vascular endothelial (VE)-cadherin junctions by a mechanism that is independent of MLC phosphorylation but involves a p38-dependent rearrangement of Introduction: Tumor Necrosis Factor 39

the actin and tubulin cytoskeleton 315. More recently, Angelini and colleagues have shown that TNF induces tyrosine phosphorylation of VE-cadherin through activation of the protein tyrosine kinase (PTK) Fyn, a member of the Src family. Tyrosine phosphorylation of VE-cadherin decreases the cadherin ectodomain homophilic adhesion, thereby facilitating the formation of intercellular gaps 316. TNF furthermore induces the expression of chemokines and adhesion molecules which results in a more sustained recruitment of leukocytes in comparison to type I activation. TNF-activated endothelial cells have been shown to express E-selectin, vascular cell adhesion molecule (VCAM)-1 and intercellular adhesion molecule (ICAM)-1 317. In mice, but not in humans, TNF also increases the expression of P-selectin (Fig.1.9.) 318. E- selectin has a similar function as P-selectin and mainly mediates the recruitment of neutrophils 319, 320. ICAM-1 is the ligand for the integrins LFA-1 and Mac-1, VCAM- 1 is the ligand for the integrin VLA-4. The expression of VCAM-1 and ICAM-1, together with the TNF-induced expression of chemokines like CC-chemokine ligand 2 (CCL2), formerly called monocyte chemoattractant protein-1 (MCP-1), favors the recruitment of monocytes, resulting in a switch of neutrophil-rich infiltrates to mononuclear cell-rich infiltrates 228, 321.

Fig. 1.10. TNF in type II activation of endothelial cells. See text for details. Adapted from Pober et al., 2007 228.

40 Chapter 1

Another protective mechanism triggered by TNF is the coagulation cascade. TNF inhibits the synthesis of the anti-coagulant protein thrombomodulin, while it induces the expression of tissue factor, the principal initiator of coagulation 322-325. Intravascular thrombus formation leads to occlusion of local blood vessels and serves to prevent the spread of pathogens through the bloodstream. Thrombosis may, however, also induce vascular damage and haemorrhage 228. Furthermore, TNF, alone or in combination with other mediators that are released during inflammation, such as interferon- (IFN-), may induce endothelial cell death 326-329. These vascular- damaging properties of TNF will be further discussed elsewhere in this thesis (see chapter 5).

TNF is also an important regulator of the innate and adaptive immune responses that accompany the inflammatory process (Table 1.1.). In short, TNF attracts neutrophils and macrophages to the site of inflammation and stimulates their effector functions. At the same time, TNF stimulates the migration of dendritic cells and other antigen- presenting cells to the draining lymph nodes, thereby supporting the initiation of an adaptive immune response. Furthermore, TNF has been shown to play an important role in the regulation of proliferation, apoptosis and some effector functions of lymphocytes. The important role of TNF in immunity is delineated by the fact that TNF knockout mice or TNF-R1 knockout mice are highly susceptible to infection by many pathogens, including Candida albicans, Mycobacterium tuberculosis, Listeria monocytogenes, Salmonella typhimurium, Trypanosoma cruzi, and Leishmania major 330-335. Surprisingly however, TNF knockout mice appear to have relatively normal macrophage phagocytic activity, T cell proliferation, cytokine release and cytotoxicity, and B cell response to thymus-independent antigens. Moreover, TNF knockout mice developed a delayed, but vigorous inflammatory response to heat- killed Corynebacterium parvum, leading to death, suggesting that TNF may also have an important anti-inflammatory function.

Introduction: Tumor Necrosis Factor 41

Table 1.1. Effects of TNF on leukocytes. Leukocytes Effects References

Neutrophils Activation, stimulation of phagocytosis, oxidative burst, release [133, 336-341]

of elastase and production of LTB4 Increased expression of Mac-1, increased adhesion to endothelial cells and extracellular matrix proteins Apoptosis

Mast cells Activation, release of histamine and TNF [342, 343]

Monocytes/ Activation, oxidative burst, phagocytosis, production of TNF [8, 311, 344-350] macrophages Chemotaxis Increased IFN--induced NO production and tumoricidal activity Inhibition of proliferation and differentiation Apoptosis

T lymphocytes Stimulation of colony formation, proliferation [302, 351-353] Stimulation of cytotoxic T cell invasiveness Apoptosis of mature T cells

B lymphocytes Superoxide production in EBV-transformed B cell line [354]

NK cells Increased IFN- production and cytotoxic activity [355, 356]

Dendritic cells Development, differentiation [357-359] Migration to lymph nodes

If the acute inflammatory reactions fail to resolve the infection, or when a cellular immune response is triggered (e.g. in the case of viral infections, certain malignancies, or autoimmune responses), the inflammatory process will evolve to a more chronic form. Chronic inflammation may last anywhere from a week to an indefinite time frame. It is characterized by infiltration of T lymphocytes, monocytes/macrophages and plasma cells, the so-called mononuclear infiltrate. A typical pathogen that will induce chronic inflammation is Mycobacterium tuberculosis. Tuberculosis is associated with the formation of lesions in the lungs, called granulomas, consisting of a central area of infected macrophages, surrounded by a mantle of activated lymphocytes. The formation of granulomas serves to prevent the spread of the pathogen, but it often fails to completely eradicate the pathogen. Surviving mycobacteria may become dormant, causing no disease. However, when the immune system is weakened, for example in HIV-infected patients, reactivation of these dormant mycobacteria may occur, causing disease and even death 360. It was shown that TNF is expressed in granulomas and is important for the organization of granulomas and the control of infection. Neutralization of TNF in mice with latent Mycobacterium tuberculosis infection resulted in outbreak of the disease and 100% mortality 361. Reactivation of latent tuberculosis infections was also observed in human patients that were treated with TNF-blocking agents 362. Thus, screening for latent tuberculosis is now indicated before the start of anti-TNF therapy. TNF- 42 Chapter 1

blocking agents that are currently used in the clinic include infliximab, etanercept, lenercept, adalimumab and certolizumab. They are used to treat patients with chronic inflammatory disorders, such as rheumatoid arthritis, inflammatory bowel disease, and psoriasis 363-366. The good clinical responses obtained with these agents demonstrates the important role of TNF in the maintenance of chronic inflammation 367.

If an infection spreads to the bloodstream, the same mechanisms by which TNF so effectively contains local infection may become catastrophic. The presence of infection in the bloodstream may induce the systemic release of pro-inflammatory cytokines, particularly TNF, which may cause a pathological condition called sepsis or, generally, systemic inflammatory response syndrome (SIRS). It is one of the major causes of death in intensive care units. The systemic release of TNF induces vasodilation and loss of plasma volume due to increased vascular permeability, which may result in a life-threatening hypovolemic shock, or septic shock. TNF may furthermore trigger disseminated intravascular coagulation, leading to occlusion of small vessels and massive consumption of clotting proteins. This condition may lead to failure of vital organs such as the kidneys, liver, heart and lungs, and eventually death 311. TNF was identified as one of the main mediators of septic shock. The injection of high doses of TNF or LPS causes a septic shock-like syndrome in mice, while TNF- R1-deficient mice were shown to be resistant to lethal doses of LPS 332, 368. Infusion of anti-TNF antibodies in baboons or mice could protect against septic shock induced by LPS or Escherichia coli 369-371. However, in these animal studies the TNF-neutralizing antibodies were only effective when administered before or just after the challenge. Clinical trials with anti-TNF therapy in sepsis patients did not show improved survival, and in one case even an increase of mortality was reported 372-375. This suggests that, although TNF is clearly involved in the development of septic shock, it has no essential role during the later stages of the disease. Therefore, the focus of anti- cytokine therapy in sepsis has now shifted to cytokines downstream of TNF 376.

Introduction: Tumor Necrosis Factor 43

1.7. TNF in cancer therapy

Carswell and colleagues demonstrated already in 1975 that serum containing TNF could induce haemorrhagic necrosis of methA sarcomas and other transplanted tumors in mice 3. In vitro these methA tumor cells were rather insensitive to TNF, which suggested that the in vivo antitumor effect of TNF was not due to direct cytotoxic effects on tumor cells. The same was observed by Brouckaert and colleagues in 1986, who used the first completely syngeneic tumor model, namely C57BL/6 mice injected subcutaneously with B16Bl6 melanoma cells. These B16Bl6 cells were resistant to the cytotoxic effects of TNF in vitro, unless TNF was given together with IFN-. In vivo on the other hand, TNF induced complete eradication of B16Bl6 tumors. In the same report they demonstrated the species-specificity of TNF. While a complete regression of the tumor could be induced with murine TNF (mTNF), human TNF (hTNF) could only do so in combination with IFN-. Moreover, it was observed that mTNF caused considerably higher systemic toxicity in mice than hTNF 377, 378. It was found that hTNF can only trigger the murine TNF-R1, not the TNF-R2 379. However, Ameloot and colleagues demonstrated that not the difference in receptor binding, but different pharmacokinetics are responsible for the different toxicity of hTNF and mTNF. They showed that hTNF is cleared much faster than mTNF, and that repetitive administration of low doses of hTNF induced toxicity similar to mTNF, suggesting that prolonged exposure rather than peak levels determine the toxic effects of TNF in mice 380. It has now become clear that the antitumor effect of TNF is completely host-mediated. Stoelcker and co-workers showed that TNF administration could still result in tumor necrosis in TNF-R2 deficient mice, but not in TNF-R1 deficient mice. After implantation of TNF-R1 knockout tumors to wild-type mice, it was shown that TNF- induced tumor necrosis was completely independent of TNF-R1 expression on tumor cells. Moreover, bone marrow transplantation of wild-type hematopoietic cells to TNF-R1 knockout mice could not restore the antitumor effect of TNF in these mice, suggesting that leukocytes are no direct targets cells for the antitumor effect of TNF 381. Several reports have suggested that the endothelial cells of the tumor vasculature are the direct target cells of TNF, mostly based on the observation that TNF appears to induce a selective destruction of the tumor vasculature 382-384. However, there was no solid proof for this assumption until recently Goethals and colleagues demonstrated that mice expressing the TNF-R1 exclusively on endothelial cells (controlled by a Flk-1-promoter) were still sensitive to the antitumor effect of TNF, while the toxic side-effects of TNF were markedly reduced 385. These date together with observations made by other groups clearly demonstrate that the endothelial cells of the tumor vasculature are the main target cells for the antitumor effect of TNF.

Despite the success of TNF as an antitumor agent in experimental animal models, the first clinical trials performed with TNF were disappointing. Doses of 10 to 50 times 44 Chapter 1

lower than the estimated effective dose already caused severe toxic side-effects, such as hypotension, thrombocytopenia and liver toxicity, characteristics of septic shock- like syndrome 386-391. Unfortunately, doses close to the maximum tolerated dose (about 150-400µg/m2) produced only rare and minimal tumor responses 392-395. These disappointing results prompted Lienard, Lejeune and colleagues in 1988 to administer TNF via the technique of isolated limb perfusion (ILP) 396, 397. In ILP, isolation of the blood circulation of the affected limb is achieved by clamping the major vein and artery, ligation of the collateral vessels, and application of a tourniquet around the basis of the limb to compress the remaining small vessels. The main artery and vein is cannulated and connected to an oxygenated extracorporeal circuit wherein the drugs are injected (Fig.1.11.). In this way, high local concentrations of a drug can be achieved with minimal risk of systemic leakage and side-effects. After the procedure (60-90 minutes of perfusion) a washout is performed to ensure minimal systemic exposure to the drug 398. ILP can be performed with mild hyperthermia (38,5-40°C) to improve local drug uptake 399.

Fig. 1.11. Setup of ILP for a large soft tissue sarcoma in the upper leg. Adapted from Lejeune et al., 2006 400.

The standard drug used in ILP is the alkylating agent melphalan. ILP with melphalan alone was found to produce a complete response rate of around 50% in in-transit melanoma metastasis, but had only minimal effect on unresectable soft tissue sarcomas of the limbs. Some pilot studies of ILP with TNF alone were performed, but Introduction: Tumor Necrosis Factor 45

only minimal responses were observed 401. However, when Lienard and colleagues added TNF and IFN- to ILP with melphalan, tumor response was significantly increased 396. Nowadays, the complete response rate for ILP with TNF plus melphalan (with or without IFN-) is around 80% for melanoma and around 20% for sarcoma, and an objective response rate of respectively >90% and around 80%. For soft tissue sarcoma this results in limb salvage in around 80% of the cases, where amputation used to be the only adequate treatment 400. There is evidence for two important effects of TNF in ILP: (1) early after infusion (within 30 minutes) it increases the permeability of the tumor vasculature, resulting in a higher drug uptake in the tumor, and (2) later (24 hours to a few days) it induces a selective destruction of the tumor vasculature 400. Increase of vascular permeability is a well-known effect of TNF, and it appears that tumor-associated blood vessels are more sensitive for this effect than normal vessels. In rat limb sarcoma models ILP with TNF was shown to increase the penetration of melphalan or doxorubicin in the tumor, while there was no change in uptake in normal skin and muscle 402, 403. It was furthermore demonstrated that systemic administration of TNF selectively increased the uptake of monoclonal antibodies or stealth liposomal doxorubicin (Doxil) in a human colorectal carcinoma xenograft in nude mice or in a B16Bl6 melanoma in C57BL/6 mice, respectively 404, 405. An increased drug uptake resulting from an increase in vascular permeability may however only be a short-lived effect as an increase in vascular permeability may also cause an elevation of interstitial fluid pressure (IFP) which is known to counteract drug uptake, in particular of macromolecules 406. Angiograms taken from patients treated with TNF-based ILP confirmed that TNF induces a selective destruction of the tumor vasculature, while leaving the normal vasculature unharmed (Fig.1.12.) 398, 400. TNF also triggers an inflammatory response during ILP, resulting in the rapid recruitment of polymorphonuclear cells to the tumor site, followed by the infiltration of macrophages and lymphocytes after several days 407. It is still unclear whether these immune cells play a role in the vascular- or tumor-destructing effects of the treatment. However, at least one study showed that total body irradiation of tumor-bearing rats, thereby inducing general leukopenia, markedly decreased the tumor response and skin necrosis at the tumor site following ILP with TNF and melphalan 408.

46 Chapter 1

Fig. 1.12. Selective destruction of the tumor-associated vasculature in TNF-based ILP. Angiograms of unresectable sarcoma of the leg were taken before (left) or 7 days after ILP with TNF, IFN- and melphalan (right). From Lejeune et al., 2006 400.

The success of ILP for the treatment of unresectable tumors of the limbs has led to the adjustment of this technique for the treatment of other organs, particularly the liver, which is then called isolated hepatic perfusion (IHP). Clinical trials of IHP with melphalan were successful, however, only minimal improvement of tumor response was noticed when TNF was added 409-411.

Despite the antitumor activities just mentioned, no TNF-based treatment thus far has proved to affect patient survival. The reason for this discrepancy lies in the fact that patient survival is mainly determined by the presence of distant metastasis, which are not affected by ILP. Thus, in order to really know the anticancer potential of TNF, systemic administration is required. Several strategies may be followed to allow systemic treatment with TNF: (1) combining low-dose TNF with other drugs to increase drug uptake, (2) the development of TNF muteins with lower toxicity, (3) combining low-dose TNF with sensitizers for the antitumor effect of TNF, (4) specific delivery systems to target TNF to the tumor, and (5) using high-dose TNF in combination with inhibitors of the toxic side-effects of TNF. The latter may be the easiest solution. Cauwels and colleagues demonstrated that inhibition of sGC protects mice against the lethality induced by treatment with mTNF plus IFN-, without interfering with the antitumor activities of this treatment 267. However, using potentially lethal doses of TNF together with an inhibitor of the toxic side-effects is not an attractive solution for human cancer therapy as failure of the inhibitor could lead to death of the patient. It was shown by Ameloot and colleagues that prolonged exposure to TNF rather than high peak concentrations causes toxicity, suggesting that TNF muteins with faster clearance will most likely be less toxic 380. However, as Introduction: Tumor Necrosis Factor 47

demonstrated by the lower antitumor efficacy of hTNF in murine tumor models, faster cleared TNF muteins will probably require combination with sensitizers for the antitumor activity of TNF, such as IFN- 377. Most chemical sensitizers that have been identified interfere with the anti-apoptotic machinery induced by TNF, in essence the NF-B and PI3K/Akt pathways. For instance, gambogic acid, a drug that has recently been shown to suppress NF-B activation through a mechanism that involves the transferrin receptor, can sensitize a human leukemia cell line to apoptosis induced by TNF 412. Another promising NF-B pathway inhibitor is the proteasome inhibitor bortezomib. This drug has already entered clinical trials in combination with conventional chemotherapeutics 413. The first combined treatments of TNF and bortezomib in tumor-bearing mice showed a significantly increased response compared to either drug alone 414. Matschurat and colleagues demonstrated that co- injection of the PI3K inhibitor wortmannin and TNF markedly reduced the effective dose of TNF to induce haemorrhage and necrosis in a murine methA tumor model. Interestingly, in the same report it was shown that inhibition of PI3K increased the TNF-induced expression of tissue factor on HUVECs 415. Another interesting target is eNOS. The coadministration of TNF and the NOS inhibitor N-nitro-L-arginine- methyl-ester (L-NAME) significantly increased the tumor response in ILP treatment of soft tissue sarcoma in rats, and in B16Bl6 melanoma-bearing mice 416, 417. It was also shown that eNOS deficient mice are more sensitive to the antitumor effect of TNF 417. Goethals and colleagues also found that inhibition of PLA2 with the inhibitors aristolochic acid or ATK significantly lowered the effective dose of TNF in the murine B16Bl6 melanoma model 417. However, the mechanism involved in this sensitization remains to be elucidated. New delivery systems are being developed to selectively target TNF to the tumor microenvironment, with minimal exposure to the rest of the body (summarized in Table 1.2.). Some of these TNF delivery systems, including NGR-TNF and TNFerade, are currently tested in clinical studies, with encouraging results.

48 Chapter 1

Table 1.2. Specific TNF delivery systems. Name Description / Results References

TNF-PEGL TNF encapsulated in pegylated (Stealth) liposomes, which increases its [418, 419] circulation time and accumulation in tumor tissue. TNF-PEGL has been shown to enhance the antitumor activity of Doxil in rats without causing the toxic side-effects of free TNF.

NGR-TNF Fusion protein of TNF with the tumor-homing peptide Cys-Asn-Gly-Arg- [420-423] Cys (CNGRC), a ligand of aminopeptidase-N, which is expressed on tumor vasculature. The therapeutic index of NGR-TNF was shown to be 12-15 higher than free TNF. NGR-TNF has potent antitumor activity in combination with chemotherapeutics in murine tumor models. NGR-TNF is currently being tested in clinical studies.

TNF-L19 Fusion protein of TNF and an antibody fragment (L19) specific to the [424-426] extradomain B of fibronectin, a marker of angiogensis. L19 has been shown to specifically target tumors in animal models and in cancer patients. TNF-L19 was shown to have synergistic antitumor activity with melphalan or IL-12-L19 in murine tumor models.

TNFerade A replication-deficient adenovector carrying the TNF cDNA regulated by [427-430] the radiation-sensitive promoter Egr-1. TNFerade allows irradiation-induced expression of TNF in tumor tissue, with minimal systemic exposure and toxic side-effects. Phase I and II clinical trials of TNFerade with radiation therapy and other chemotherapeutics showed impressive tumor responses in patients with prior treatment-refractory solid tumors.

PT-cAu-TNF TNF encapsulated in PEG-coated colloidal gold nanoparticles. [431-433] PT-cAu-TNF has been shown to rapidly accumulate in solid tumors, with little to no accumulation in other organs following intravenous injection in mice. PT-cAu-TNF alone or in combination with thermal therapy has potent antitumor activity in murine tumor models.

Introduction: Tumor Necrosis Factor 49

1.8. References

1. Old, L.J. Tumor necrosis factor. Scientific American 258, 59-60, 69-75 (1988). 2. Shear, M.J., Turner, F.C., Perrault, A., Shovelton, T. Chemical treatment of tumors. V. Isolation of the hemorrhage-producing fraction from Serratia marcescens (Bacillus prodigiosus) culture filtrate. J Natl Cancer Inst 4, 81-97 (1943). 3. Carswell, E.A. et al. An endotoxin-induced serum factor that causes necrosis of tumors. Proceedings of the National Academy of Sciences of the United States of America 72, 3666- 3670 (1975). 4. Cerami, A., Ikeda, Y., Le Trang, N., Hotez, P.J. & Beutler, B. Weight loss associated with an endotoxin-induced mediator from peritoneal macrophages: the role of cachectin (tumor necrosis factor). Immunology letters 11, 173-177 (1985). 5. Fransen, L. et al. Molecular cloning of mouse tumour necrosis factor cDNA and its eukaryotic expression. Nucleic acids research 13, 4417-4429 (1985). 6. Marmenout, A. et al. Molecular cloning and expression of human tumor necrosis factor and comparison with mouse tumor necrosis factor. European journal of biochemistry / FEBS 152, 515-522 (1985). 7. Pennica, D. et al. Human tumour necrosis factor: precursor structure, expression and homology to lymphotoxin. Nature 312, 724-729 (1984). 8. Oppenheim, J., Feldmann, M. Cytokine Reference. Volume 1: Ligands. (Academic Press, San Diego; 2000). 9. Moss, M.L. et al. Cloning of a disintegrin metalloproteinase that processes precursor tumour- necrosis factor-alpha. Nature 385, 733-736 (1997). 10. Black, R.A. et al. A metalloproteinase disintegrin that releases tumour-necrosis factor-alpha from cells. Nature 385, 729-733 (1997). 11. Wingfield, P., Pain, R.H. & Craig, S. Tumour necrosis factor is a compact trimer. FEBS letters 211, 179-184 (1987). 12. Loetscher, H. et al. Two distinct tumour necrosis factor receptors--members of a new cytokine receptor gene family. Oxford surveys on eukaryotic genes 7, 119-142 (1991). 13. Grell, M. et al. The transmembrane form of tumor necrosis factor is the prime activating ligand of the 80 kDa tumor necrosis factor receptor. Cell 83, 793-802 (1995). 14. Chan, F.K. et al. A domain in TNF receptors that mediates ligand-independent receptor assembly and signaling. Science (New York, N.Y 288, 2351-2354 (2000). 15. Bemelmans, M.H., van Tits, L.J. & Buurman, W.A. Tumor necrosis factor: function, release and clearance. Critical reviews in immunology 16, 1-11 (1996). 16. Solomon, K.A., Pesti, N., Wu, G. & Newton, R.C. Cutting edge: a dominant negative form of TNF-alpha converting enzyme inhibits proTNF and TNFRII secretion. J Immunol 163, 4105- 4108 (1999). 17. 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, 173-176 (2007). 18. McDermott, M.F. et al. Germline mutations in the extracellular domains of the 55 kDa TNF receptor, TNFR1, define a family of dominantly inherited autoinflammatory syndromes. Cell 97, 133-144 (1999). 19. Mattey, D.L., Glossop, J.R., Nixon, N.B. & Dawes, P.T. Circulating levels of tumor necrosis factor receptors are highly predictive of mortality in patients with rheumatoid arthritis. Arthritis and rheumatism 56, 3940-3948 (2007). 20. Dembic, Z. et al. Two human TNF receptors have similar extracellular, but distinct intracellular, domain sequences. Cytokine 2, 231-237 (1990). 21. Tartaglia, L.A., Ayres, T.M., Wong, G.H. & Goeddel, D.V. A novel domain within the 55 kd TNF receptor signals cell death. Cell 74, 845-853 (1993). 22. Mocellin, S., Rossi, C.R., Pilati, P. & Nitti, D. Tumor necrosis factor, cancer and anticancer therapy. Cytokine & growth factor reviews 16, 35-53 (2005). 23. Schwandner, R., Wiegmann, K., Bernardo, K., Kreder, D. & Kronke, M. TNF receptor death domain-associated proteins TRADD and FADD signal activation of acid sphingomyelinase. The Journal of biological chemistry 273, 5916-5922 (1998). 24. Adam, D., Adam-Klages, S. & Kronke, M. Identification of p55 tumor necrosis factor receptor-associated proteins that couple to signaling pathways not initiated by the death domain. Journal of inflammation 47, 61-66 (1995). 50 Chapter 1

25. Adam-Klages, S. et al. Distinct adapter proteins mediate acid versus neutral sphingomyelinase activation through the p55 receptor for tumor necrosis factor. Journal of leukocyte biology 63, 678-682 (1998). 26. Rothe, M., Wong, S.C., Henzel, W.J. & Goeddel, D.V. A novel family of putative signal transducers associated with the cytoplasmic domain of the 75 kDa tumor necrosis factor receptor. Cell 78, 681-692 (1994). 27. Rothe, M., Pan, M.G., Henzel, W.J., Ayres, T.M. & Goeddel, D.V. The TNFR2-TRAF signaling complex contains two novel proteins related to baculoviral inhibitor of apoptosis proteins. Cell 83, 1243-1252 (1995). 28. Collart, M.A., Baeuerle, P. & Vassalli, P. Regulation of tumor necrosis factor alpha transcription in macrophages: involvement of four kappa B-like motifs and of constitutive and inducible forms of NF-kappa B. Molecular and cellular biology 10, 1498-1506 (1990). 29. Kramer, B., Wiegmann, K. & Kronke, M. Regulation of the human TNF promoter by the transcription factor Ets. The Journal of biological chemistry 270, 6577-6583 (1995). 30. Pope, R.M., Leutz, A. & Ness, S.A. C/EBP beta regulation of the tumor necrosis factor alpha gene. The Journal of clinical investigation 94, 1449-1455 (1994). 31. Rhoades, K.L., Golub, S.H. & Economou, J.S. The regulation of the human tumor necrosis factor alpha promoter region in macrophage, T cell, and B cell lines. The Journal of biological chemistry 267, 22102-22107 (1992). 32. Tsai, E.Y., Jain, J., Pesavento, P.A., Rao, A. & Goldfeld, A.E. Tumor necrosis factor alpha gene regulation in activated T cells involves ATF-2/Jun and NFATp. Molecular and cellular biology 16, 459-467 (1996). 33. Newell, C.L., Deisseroth, A.B. & Lopez-Berestein, G. Interaction of nuclear proteins with an AP-1/CRE-like promoter sequence in the human TNF-alpha gene. Journal of leukocyte biology 56, 27-35 (1994). 34. Zagariya, A. et al. Tumor necrosis factor alpha gene regulation: enhancement of C/EBPbeta- induced activation by c-Jun. Molecular and cellular biology 18, 2815-2824 (1998). 35. Shakhov, A.N., Collart, M.A., Vassalli, P., Nedospasov, S.A. & Jongeneel, C.V. Kappa B- type enhancers are involved in lipopolysaccharide-mediated transcriptional activation of the tumor necrosis factor alpha gene in primary macrophages. The Journal of experimental medicine 171, 35-47 (1990). 36. Caput, D. et al. Identification of a common nucleotide sequence in the 3'-untranslated region of mRNA molecules specifying inflammatory mediators. Proceedings of the National Academy of Sciences of the United States of America 83, 1670-1674 (1986). 37. Zhang, T., Kruys, V., Huez, G. & Gueydan, C. AU-rich element-mediated translational control: complexity and multiple activities of trans-activating factors. Biochemical Society transactions 30, 952-958 (2002). 38. Dumitru, C.D. et al. TNF-alpha induction by LPS is regulated posttranscriptionally via a Tpl2/ERK-dependent pathway. Cell 103, 1071-1083 (2000). 39. Carballo, E., Lai, W.S. & Blackshear, P.J. Feedback inhibition of macrophage tumor necrosis factor-alpha production by tristetraprolin. Science (New York, N.Y 281, 1001-1005 (1998). 40. Deleault, K.M., Skinner, S.J. & Brooks, S.A. Tristetraprolin regulates TNF TNF-alpha mRNA stability via a proteasome dependent mechanism involving the combined action of the ERK and p38 pathways. Molecular immunology 45, 13-24 (2008). 41. Hitti, E. et al. Mitogen-activated protein kinase-activated protein kinase 2 regulates tumor necrosis factor mRNA stability and translation mainly by altering tristetraprolin expression, stability, and binding to adenine/uridine-rich element. Molecular and cellular biology 26, 2399-2407 (2006). 42. Piecyk, M. et al. TIA-1 is a translational silencer that selectively regulates the expression of TNF-alpha. The EMBO journal 19, 4154-4163 (2000). 43. Stoecklin, G., Lu, M., Rattenbacher, B. & Moroni, C. A constitutive decay element promotes tumor necrosis factor alpha mRNA degradation via an AU-rich element-independent pathway. Molecular and cellular biology 23, 3506-3515 (2003). 44. Zuckerman, S.H., Evans, G.F., Snyder, Y.M. & Roeder, W.D. Endotoxin-macrophage interaction: post-translational regulation of tumor necrosis factor expression. J Immunol 143, 1223-1227 (1989). 45. Jiang, Y., Woronicz, J.D., Liu, W. & Goeddel, D.V. Prevention of constitutive TNF receptor 1 signaling by silencer of death domains. Science (New York, N.Y 283, 543-546 (1999). Introduction: Tumor Necrosis Factor 51

46. Hsu, H., Shu, H.B., Pan, M.G. & Goeddel, D.V. TRADD-TRAF2 and TRADD-FADD interactions define two distinct TNF receptor 1 signal transduction pathways. Cell 84, 299-308 (1996). 47. Hsu, H., Huang, J., Shu, H.B., Baichwal, V. & Goeddel, D.V. TNF-dependent recruitment of the protein kinase RIP to the TNF receptor-1 signaling complex. Immunity 4, 387-396 (1996). 48. Micheau, O. & Tschopp, J. Induction of TNF receptor I-mediated apoptosis via two sequential signaling complexes. Cell 114, 181-190 (2003). 49. Legler, D.F., Micheau, O., Doucey, M.A., Tschopp, J. & Bron, C. Recruitment of TNF receptor 1 to lipid rafts is essential for TNFalpha-mediated NF-kappaB activation. Immunity 18, 655-664 (2003). 50. Doan, J.E., Windmiller, D.A. & Riches, D.W. Differential regulation of TNF-R1 signaling: lipid raft dependency of p42mapk/erk2 activation, but not NF-kappaB activation. J Immunol 172, 7654-7660 (2004). 51. D'Alessio, A., Al-Lamki, R.S., Bradley, J.R. & Pober, J.S. Caveolae participate in tumor necrosis factor receptor 1 signaling and internalization in a human endothelial cell line. The American journal of pathology 166, 1273-1282 (2005). 52. Carrington, P.E. et al. The structure of FADD and its mode of interaction with procaspase-8. Molecular cell 22, 599-610 (2006). 53. Schneider-Brachert, W. et al. Compartmentalization of TNF receptor 1 signaling: internalized TNF receptosomes as death signaling vesicles. Immunity 21, 415-428 (2004). 54. Pahl, H.L. Activators and target genes of Rel/NF-kappaB transcription factors. Oncogene 18, 6853-6866 (1999). 55. Stehlik, C. et al. Nuclear factor (NF)-kappaB-regulated X-chromosome-linked iap gene expression protects endothelial cells from tumor necrosis factor alpha-induced apoptosis. The Journal of experimental medicine 188, 211-216 (1998). 56. Chen, C., Edelstein, L.C. & Gelinas, C. The Rel/NF-kappaB family directly activates expression of the apoptosis inhibitor Bcl-x(L). Molecular and cellular biology 20, 2687-2695 (2000). 57. Micheau, O., Lens, S., Gaide, O., Alevizopoulos, K. & Tschopp, J. NF-kappaB signals induce the expression of c-FLIP. Molecular and cellular biology 21, 5299-5305 (2001). 58. Chu, Z.L. et al. Suppression of tumor necrosis factor-induced cell death by inhibitor of apoptosis c-IAP2 is under NF-kappaB control. Proceedings of the National Academy of Sciences of the United States of America 94, 10057-10062 (1997). 59. Wang, C.Y., Mayo, M.W., Korneluk, R.G., Goeddel, D.V. & Baldwin, A.S., Jr. NF-kappaB antiapoptosis: induction of TRAF1 and TRAF2 and c-IAP1 and c-IAP2 to suppress caspase-8 activation. Science (New York, N.Y 281, 1680-1683 (1998). 60. Zong, W.X., Edelstein, L.C., Chen, C., Bash, J. & Gelinas, C. The prosurvival Bcl-2 homolog Bfl-1/A1 is a direct transcriptional target of NF-kappaB that blocks TNFalpha-induced apoptosis. Genes & development 13, 382-387 (1999). 61. Baud, V. & Karin, M. Signal transduction by tumor necrosis factor and its relatives. Trends in cell biology 11, 372-377 (2001). 62. Perkins, N.D. The Rel/NF-kappa B family: friend and foe. Trends in biochemical sciences 25, 434-440 (2000). 63. Tergaonkar, V., Correa, R.G., Ikawa, M. & Verma, I.M. Distinct roles of IkappaB proteins in regulating constitutive NF-kappaB activity. Nature cell biology 7, 921-923 (2005). 64. Senftleben, U. et al. Activation by IKKalpha of a second, evolutionary conserved, NF-kappa B signaling pathway. Science (New York, N.Y 293, 1495-1499 (2001). 65. Hoffmann, A. & Baltimore, D. Circuitry of nuclear factor kappaB signaling. Immunological reviews 210, 171-186 (2006). 66. Zandi, E., Rothwarf, D.M., Delhase, M., Hayakawa, M. & Karin, M. The IkappaB kinase complex (IKK) contains two kinase subunits, IKKalpha and IKKbeta, necessary for IkappaB phosphorylation and NF-kappaB activation. Cell 91, 243-252 (1997). 67. Rothwarf, D.M., Zandi, E., Natoli, G. & Karin, M. IKK-gamma is an essential regulatory subunit of the IkappaB kinase complex. Nature 395, 297-300 (1998). 68. Yamaoka, S. et al. Complementation cloning of NEMO, a component of the IkappaB kinase complex essential for NF-kappaB activation. Cell 93, 1231-1240 (1998). 69. Chen, G., Cao, P. & Goeddel, D.V. TNF-induced recruitment and activation of the IKK complex require Cdc37 and Hsp90. Molecular cell 9, 401-410 (2002). 70. Devin, A. et al. The distinct roles of TRAF2 and RIP in IKK activation by TNF-R1: TRAF2 recruits IKK to TNF-R1 while RIP mediates IKK activation. Immunity 12, 419-429 (2000). 52 Chapter 1

71. Takaesu, G. et al. TAK1 is critical for IkappaB kinase-mediated activation of the NF-kappaB pathway. Journal of molecular biology 326, 105-115 (2003). 72. Blonska, M., You, Y., Geleziunas, R. & Lin, X. Restoration of NF-kappaB activation by tumor necrosis factor alpha receptor complex-targeted MEKK3 in receptor-interacting protein- deficient cells. Molecular and cellular biology 24, 10757-10765 (2004). 73. Blonska, M. et al. TAK1 is recruited to the tumor necrosis factor-alpha (TNF-alpha) receptor 1 complex in a receptor-interacting protein (RIP)-dependent manner and cooperates with MEKK3 leading to NF-kappaB activation. The Journal of biological chemistry 280, 43056- 43063 (2005). 74. Li, H., Kobayashi, M., Blonska, M., You, Y. & Lin, X. Ubiquitination of RIP is required for tumor necrosis factor alpha-induced NF-kappaB activation. The Journal of biological chemistry 281, 13636-13643 (2006). 75. Kanayama, A. et al. TAB2 and TAB3 activate the NF-kappaB pathway through binding to polyubiquitin chains. Molecular cell 15, 535-548 (2004). 76. Wertz, I.E. et al. De-ubiquitination and ubiquitin ligase domains of A20 downregulate NF- kappaB signalling. Nature 430, 694-699 (2004). 77. Ben-Neriah, Y. Regulatory functions of ubiquitination in the immune system. Nature immunology 3, 20-26 (2002). 78. Nakayama, K. et al. Impaired degradation of inhibitory subunit of NF-kappa B (I kappa B) and beta-catenin as a result of targeted disruption of the beta-TrCP1 gene. Proceedings of the National Academy of Sciences of the United States of America 100, 8752-8757 (2003). 79. Koul, D., Yao, Y., Abbruzzese, J.L., Yung, W.K. & Reddy, S.A. Tumor suppressor MMAC/PTEN inhibits cytokine-induced NFkappaB activation without interfering with the IkappaB degradation pathway. The Journal of biological chemistry 276, 11402-11408 (2001). 80. Reddy, S.A., Huang, J.H. & Liao, W.S. Phosphatidylinositol 3-kinase as a mediator of TNF- induced NF-kappa B activation. J Immunol 164, 1355-1363 (2000). 81. Sizemore, N., Lerner, N., Dombrowski, N., Sakurai, H. & Stark, G.R. Distinct roles of the Ikappa B kinase alpha and beta subunits in liberating nuclear factor kappa B (NF-kappa B) from Ikappa B and in phosphorylating the p65 subunit of NF-kappa B. The Journal of biological chemistry 277, 3863-3869 (2002). 82. Zhong, H., Voll, R.E. & Ghosh, S. Phosphorylation of NF-kappa B p65 by PKA stimulates transcriptional activity by promoting a novel bivalent interaction with the coactivator CBP/p300. Molecular cell 1, 661-671 (1998). 83. Wang, D., Westerheide, S.D., Hanson, J.L. & Baldwin, A.S., Jr. Tumor necrosis factor alpha- induced phosphorylation of RelA/p65 on Ser529 is controlled by casein kinase II. The Journal of biological chemistry 275, 32592-32597 (2000). 84. Sakurai, H., Chiba, H., Miyoshi, H., Sugita, T. & Toriumi, W. IkappaB kinases phosphorylate NF-kappaB p65 subunit on serine 536 in the transactivation domain. The Journal of biological chemistry 274, 30353-30356 (1999). 85. Vermeulen, L., De Wilde, G., Van Damme, P., Vanden Berghe, W. & Haegeman, G. Transcriptional activation of the NF-kappaB p65 subunit by mitogen- and stress-activated protein kinase-1 (MSK1). The EMBO journal 22, 1313-1324 (2003). 86. Leitges, M. et al. Targeted disruption of the zetaPKC gene results in the impairment of the NF-kappaB pathway. Molecular cell 8, 771-780 (2001). 87. Hoeflich, K.P. et al. Requirement for glycogen synthase kinase-3beta in cell survival and NF- kappaB activation. Nature 406, 86-90 (2000). 88. Vanden Berghe, W. et al. p38 and extracellular signal-regulated kinase mitogen-activated protein kinase pathways are required for nuclear factor-kappaB p65 transactivation mediated by tumor necrosis factor. The Journal of biological chemistry 273, 3285-3290 (1998). 89. Luschen, S., Scherer, G., Ussat, S., Ungefroren, H. & Adam-Klages, S. Inhibition of p38 mitogen-activated protein kinase reduces TNF-induced activation of NF-kappaB, elicits caspase activity, and enhances cytotoxicity. Experimental cell research 293, 196-206 (2004). 90. Chen, L., Fischle, W., Verdin, E. & Greene, W.C. Duration of nuclear NF-kappaB action regulated by reversible acetylation. Science (New York, N.Y 293, 1653-1657 (2001). 91. Ashburner, B.P., Westerheide, S.D. & Baldwin, A.S., Jr. The p65 (RelA) subunit of NF- kappaB interacts with the histone deacetylase (HDAC) corepressors HDAC1 and HDAC2 to negatively regulate gene expression. Molecular and cellular biology 21, 7065-7077 (2001). 92. Zhong, H., May, M.J., Jimi, E. & Ghosh, S. The phosphorylation status of nuclear NF-kappa B determines its association with CBP/p300 or HDAC-1. Molecular cell 9, 625-636 (2002). Introduction: Tumor Necrosis Factor 53

93. Hoffmann, A., Levchenko, A., Scott, M.L. & Baltimore, D. The IkappaB-NF-kappaB signaling module: temporal control and selective gene activation. Science (New York, N.Y 298, 1241-1245 (2002). 94. Nelson, D.E. et al. Oscillations in NF-kappaB signaling control the dynamics of gene expression. Science (New York, N.Y 306, 704-708 (2004). 95. Kearns, J.D., Basak, S., Werner, S.L., Huang, C.S. & Hoffmann, A. IkappaBepsilon provides negative feedback to control NF-kappaB oscillations, signaling dynamics, and inflammatory gene expression. The Journal of cell biology 173, 659-664 (2006). 96. Johnson, G.L. & Lapadat, R. Mitogen-activated protein kinase pathways mediated by ERK, JNK, and p38 protein kinases. Science (New York, N.Y 298, 1911-1912 (2002). 97. Zhang, W. & Liu, H.T. MAPK signal pathways in the regulation of cell proliferation in mammalian cells. Cell research 12, 9-18 (2002). 98. Reinhard, C., Shamoon, B., Shyamala, V. & Williams, L.T. Tumor necrosis factor alpha- induced activation of c-jun N-terminal kinase is mediated by TRAF2. The EMBO journal 16, 1080-1092 (1997). 99. Natoli, G. et al. Activation of SAPK/JNK by TNF receptor 1 through a noncytotoxic TRAF2- dependent pathway. Science (New York, N.Y 275, 200-203 (1997). 100. Yeh, W.C. et al. Early lethality, functional NF-kappaB activation, and increased sensitivity to TNF-induced cell death in TRAF2-deficient mice. Immunity 7, 715-725 (1997). 101. Moriguchi, T. et al. A novel SAPK/JNK kinase, MKK7, stimulated by TNFalpha and cellular stresses. The EMBO journal 16, 7045-7053 (1997). 102. Tournier, C. et al. MKK7 is an essential component of the JNK signal transduction pathway activated by proinflammatory cytokines. Genes & development 15, 1419-1426 (2001). 103. Baud, V. et al. Signaling by proinflammatory cytokines: oligomerization of TRAF2 and TRAF6 is sufficient for JNK and IKK activation and target gene induction via an amino- terminal effector domain. Genes & development 13, 1297-1308 (1999). 104. Nishitoh, H. et al. ASK1 is essential for JNK/SAPK activation by TRAF2. Molecular cell 2, 389-395 (1998). 105. Ninomiya-Tsuji, J. et al. The kinase TAK1 can activate the NIK-I kappaB as well as the MAP kinase cascade in the IL-1 signalling pathway. Nature 398, 252-256 (1999). 106. Yujiri, T. et al. MEK kinase 1 (MEKK1) transduces c-Jun NH2-terminal kinase activation in response to changes in the microtubule cytoskeleton. The Journal of biological chemistry 274, 12605-12610 (1999). 107. Xia, Y. et al. MEK kinase 1 is critically required for c-Jun N-terminal kinase activation by proinflammatory stimuli and growth factor-induced cell migration. Proceedings of the National Academy of Sciences of the United States of America 97, 5243-5248 (2000). 108. Yuasa, T., Ohno, S., Kehrl, J.H. & Kyriakis, J.M. Tumor necrosis factor signaling to stress- activated protein kinase (SAPK)/Jun NH2-terminal kinase (JNK) and p38. Germinal center kinase couples TRAF2 to mitogen-activated protein kinase/ERK kinase kinase 1 and SAPK while receptor interacting protein associates with a mitogen-activated protein kinase kinase kinase upstream of MKK6 and p38. The Journal of biological chemistry 273, 22681-22692 (1998). 109. Tobiume, K. et al. ASK1 is required for sustained activations of JNK/p38 MAP kinases and apoptosis. EMBO reports 2, 222-228 (2001). 110. Ichijo, H. et al. Induction of apoptosis by ASK1, a mammalian MAPKKK that activates SAPK/JNK and p38 signaling pathways. Science (New York, N.Y 275, 90-94 (1997). 111. Gotoh, Y. & Cooper, J.A. Reactive oxygen species- and dimerization-induced activation of apoptosis signal-regulating kinase 1 in tumor necrosis factor-alpha signal transduction. The Journal of biological chemistry 273, 17477-17482 (1998). 112. Tobiume, K., Saitoh, M. & Ichijo, H. Activation of apoptosis signal-regulating kinase 1 by the stress-induced activating phosphorylation of pre-formed oligomer. Journal of cellular physiology 191, 95-104 (2002). 113. Saitoh, M. et al. Mammalian thioredoxin is a direct inhibitor of apoptosis signal-regulating kinase (ASK) 1. The EMBO journal 17, 2596-2606 (1998). 114. Chandel, N.S., Schumacker, P.T. & Arch, R.H. Reactive oxygen species are downstream products of TRAF-mediated signal transduction. The Journal of biological chemistry 276, 42728-42736 (2001). 115. Liu, H., Nishitoh, H., Ichijo, H. & Kyriakis, J.M. Activation of apoptosis signal-regulating kinase 1 (ASK1) by tumor necrosis factor receptor-associated factor 2 requires prior 54 Chapter 1

dissociation of the ASK1 inhibitor thioredoxin. Molecular and cellular biology 20, 2198-2208 (2000). 116. Zhang, L., Chen, J. & Fu, H. Suppression of apoptosis signal-regulating kinase 1-induced cell death by 14-3-3 proteins. Proceedings of the National Academy of Sciences of the United States of America 96, 8511-8515 (1999). 117. Zhang, R. et al. AIP1 mediates TNF-alpha-induced ASK1 activation by facilitating dissociation of ASK1 from its inhibitor 14-3-3. The Journal of clinical investigation 111, 1933-1943 (2003). 118. Zhang, H. et al. RIP1-mediated AIP1 phosphorylation at a 14-3-3-binding site is critical for tumor necrosis factor-induced ASK1-JNK/p38 activation. The Journal of biological chemistry 282, 14788-14796 (2007). 119. Carpentier, I. et al. TRAF2 plays a dual role in NF-kappaB-dependent gene activation by mediating the TNF-induced activation of p38 MAPK and IkappaB kinase pathways. FEBS letters 425, 195-198 (1998). 120. Diener, K. et al. Activation of the c-Jun N-terminal kinase pathway by a novel protein kinase related to human germinal center kinase. Proceedings of the National Academy of Sciences of the United States of America 94, 9687-9692 (1997). 121. Nakano, K., Yamauchi, J., Nakagawa, K., Itoh, H. & Kitamura, N. NESK, a member of the germinal center kinase family that activates the c-Jun N-terminal kinase pathway and is expressed during the late stages of embryogenesis. The Journal of biological chemistry 275, 20533-20539 (2000). 122. Wysk, M., Yang, D.D., Lu, H.T., Flavell, R.A. & Davis, R.J. Requirement of mitogen- activated protein kinase kinase 3 (MKK3) for tumor necrosis factor-induced cytokine expression. Proceedings of the National Academy of Sciences of the United States of America 96, 3763-3768 (1999). 123. Zhao, M. et al. Regulation of the MEF2 family of transcription factors by p38. Molecular and cellular biology 19, 21-30 (1999). 124. van Dam, H. et al. ATF-2 is preferentially activated by stress-activated protein kinases to mediate c-jun induction in response to genotoxic agents. The EMBO journal 14, 1798-1811 (1995). 125. Wang, X.Z. & Ron, D. Stress-induced phosphorylation and activation of the transcription factor CHOP (GADD153) by p38 MAP Kinase. Science (New York, N.Y 272, 1347-1349 (1996). 126. Janknecht, R. & Hunter, T. Convergence of MAP kinase pathways on the ternary complex factor Sap-1a. The EMBO journal 16, 1620-1627 (1997). 127. Winzen, R. et al. The p38 MAP kinase pathway signals for cytokine-induced mRNA stabilization via MAP kinase-activated protein kinase 2 and an AU-rich region-targeted mechanism. The EMBO journal 18, 4969-4980 (1999). 128. Miyazawa, K. et al. Regulation of interleukin-1beta-induced interleukin-6 gene expression in human fibroblast-like synoviocytes by p38 mitogen-activated protein kinase. The Journal of biological chemistry 273, 24832-24838 (1998). 129. Waterman, W.H., Molski, T.F., Huang, C.K., Adams, J.L. & Sha'afi, R.I. Tumour necrosis factor-alpha-induced phosphorylation and activation of cytosolic phospholipase A2 are abrogated by an inhibitor of the p38 mitogen-activated protein kinase cascade in human neutrophils. The Biochemical journal 319 ( Pt 1), 17-20 (1996). 130. Winston, B.W. & Riches, D.W. Activation of p42mapk/erk2 following engagement of tumor necrosis factor receptor CD120a (p55) in mouse macrophages. J Immunol 155, 1525-1533 (1995). 131. Schievella, A.R., Chen, J.H., Graham, J.R. & Lin, L.L. MADD, a novel death domain protein that interacts with the type 1 tumor necrosis factor receptor and activates mitogen-activated protein kinase. The Journal of biological chemistry 272, 12069-12075 (1997). 132. Luschen, S. et al. The Fas-associated death domain protein/caspase-8/c-FLIP signaling pathway is involved in TNF-induced activation of ERK. Experimental cell research 310, 33- 42 (2005). 133. Maianski, N.A., Roos, D. & Kuijpers, T.W. Tumor necrosis factor alpha induces a caspase- independent death pathway in human neutrophils. Blood 101, 1987-1995 (2003). 134. Reed, J.C. et al. Comparative analysis of apoptosis and inflammation genes of mice and humans. Genome research 13, 1376-1388 (2003). 135. Duan, H. & Dixit, V.M. RAIDD is a new 'death' adaptor molecule. Nature 385, 86-89 (1997). Introduction: Tumor Necrosis Factor 55

136. Sakahira, H., Enari, M. & Nagata, S. Cleavage of CAD inhibitor in CAD activation and DNA degradation during apoptosis. Nature 391, 96-99 (1998). 137. Enari, M. et al. A caspase-activated DNase that degrades DNA during apoptosis, and its inhibitor ICAD. Nature 391, 43-50 (1998). 138. Lazebnik, Y.A., Kaufmann, S.H., Desnoyers, S., Poirier, G.G. & Earnshaw, W.C. Cleavage of poly(ADP-ribose) polymerase by a proteinase with properties like ICE. Nature 371, 346-347 (1994). 139. Lazebnik, Y.A. et al. Studies of the lamin proteinase reveal multiple parallel biochemical pathways during apoptotic execution. Proceedings of the National Academy of Sciences of the United States of America 92, 9042-9046 (1995). 140. Takahashi, A. et al. Cleavage of lamin A by Mch2 alpha but not CPP32: multiple interleukin 1 beta-converting enzyme-related proteases with distinct substrate recognition properties are active in apoptosis. Proceedings of the National Academy of Sciences of the United States of America 93, 8395-8400 (1996). 141. Kothakota, S. et al. Caspase-3-generated fragment of gelsolin: effector of morphological change in apoptosis. Science (New York, N.Y 278, 294-298 (1997). 142. Kayalar, C., Ord, T., Testa, M.P., Zhong, L.T. & Bredesen, D.E. Cleavage of actin by interleukin 1 beta-converting enzyme to reverse DNase I inhibition. Proceedings of the National Academy of Sciences of the United States of America 93, 2234-2238 (1996). 143. Mashima, T., Naito, M., Fujita, N., Noguchi, K. & Tsuruo, T. Identification of actin as a substrate of ICE and an ICE-like protease and involvement of an ICE-like protease but not ICE in VP-16-induced U937 apoptosis. Biochemical and biophysical research communications 217, 1185-1192 (1995). 144. Rudel, T. & Bokoch, G.M. Membrane and morphological changes in apoptotic cells regulated by caspase-mediated activation of PAK2. Science (New York, N.Y 276, 1571-1574 (1997). 145. Sebbagh, M. et al. Caspase-3-mediated cleavage of ROCK I induces MLC phosphorylation and apoptotic membrane blebbing. Nature cell biology 3, 346-352 (2001). 146. Coleman, M.L. et al. Membrane blebbing during apoptosis results from caspase-mediated activation of ROCK I. Nature cell biology 3, 339-345 (2001). 147. Scaffidi, C. et al. Two CD95 (APO-1/Fas) signaling pathways. The EMBO journal 17, 1675- 1687 (1998). 148. Guicciardi, M.E. et al. Cathepsin B contributes to TNF-alpha-mediated hepatocyte apoptosis by promoting mitochondrial release of cytochrome c. The Journal of clinical investigation 106, 1127-1137 (2000). 149. Luo, X., Budihardjo, I., Zou, H., Slaughter, C. & Wang, X. Bid, a Bcl2 interacting protein, mediates cytochrome c release from mitochondria in response to activation of cell surface death receptors. Cell 94, 481-490 (1998). 150. Li, H., Zhu, H., Xu, C.J. & Yuan, J. Cleavage of BID by caspase 8 mediates the mitochondrial damage in the Fas pathway of apoptosis. Cell 94, 491-501 (1998). 151. Antonsson, B. Mitochondria and the Bcl-2 family proteins in apoptosis signaling pathways. Molecular and cellular biochemistry 256-257, 141-155 (2004). 152. Wei, M.C. et al. Proapoptotic BAX and BAK: a requisite gateway to mitochondrial dysfunction and death. Science (New York, N.Y 292, 727-730 (2001). 153. Korsmeyer, S.J. et al. Pro-apoptotic cascade activates BID, which oligomerizes BAK or BAX into pores that result in the release of cytochrome c. Cell death and differentiation 7, 1166- 1173 (2000). 154. Kandasamy, K. et al. Involvement of proapoptotic molecules Bax and Bak in tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-induced mitochondrial disruption and apoptosis: differential regulation of cytochrome c and Smac/DIABLO release. Cancer research 63, 1712-1721 (2003). 155. Verhagen, A.M. et al. Identification of DIABLO, a mammalian protein that promotes apoptosis by binding to and antagonizing IAP proteins. Cell 102, 43-53 (2000). 156. Zamzami, N. et al. Bid acts on the permeability transition pore complex to induce apoptosis. Oncogene 19, 6342-6350 (2000). 157. Madesh, M., Antonsson, B., Srinivasula, S.M., Alnemri, E.S. & Hajnoczky, G. Rapid kinetics of tBid-induced cytochrome c and Smac/DIABLO release and mitochondrial depolarization. The Journal of biological chemistry 277, 5651-5659 (2002). 158. Bradham, C.A. et al. The mitochondrial permeability transition is required for tumor necrosis factor alpha-mediated apoptosis and cytochrome c release. Molecular and cellular biology 18, 6353-6364 (1998). 56 Chapter 1

159. Riedl, S.J. & Salvesen, G.S. The apoptosome: signalling platform of cell death. Nature reviews 8, 405-413 (2007). 160. Yin, Q. et al. Caspase-9 holoenzyme is a specific and optimal procaspase-3 processing machine. Molecular cell 22, 259-268 (2006). 161. Acehan, D. et al. Three-dimensional structure of the apoptosome: implications for assembly, procaspase-9 binding, and activation. Molecular cell 9, 423-432 (2002). 162. Chai, J. et al. Structural and biochemical basis of apoptotic activation by Smac/DIABLO. Nature 406, 855-862 (2000). 163. Engels, I.H. et al. Caspase-8/FLICE functions as an executioner caspase in anticancer drug- induced apoptosis. Oncogene 19, 4563-4573 (2000). 164. Foghsgaard, L. et al. Cathepsin B acts as a dominant execution protease in tumor cell apoptosis induced by tumor necrosis factor. The Journal of cell biology 153, 999-1010 (2001). 165. Liu, J. et al. Cathepsin B and its interacting proteins, bikunin and TSRC1, correlate with TNF- induced apoptosis of ovarian cancer cells OV-90. FEBS letters 580, 245-250 (2006). 166. Guicciardi, M.E., Miyoshi, H., Bronk, S.F. & Gores, G.J. Cathepsin B knockout mice are resistant to tumor necrosis factor-alpha-mediated hepatocyte apoptosis and liver injury: implications for therapeutic applications. The American journal of pathology 159, 2045-2054 (2001). 167. Stoka, V. et al. Lysosomal protease pathways to apoptosis. Cleavage of bid, not pro-caspases, is the most likely route. The Journal of biological chemistry 276, 3149-3157 (2001). 168. Turk, B., Turk, D. & Turk, V. Lysosomal cysteine proteases: more than scavengers. Biochimica et biophysica acta 1477, 98-111 (2000). 169. Vancompernolle, K. et al. Atractyloside-induced release of cathepsin B, a protease with caspase-processing activity. FEBS letters 438, 150-158 (1998). 170. Werneburg, N.W., Guicciardi, M.E., Bronk, S.F. & Gores, G.J. Tumor necrosis factor-alpha- associated lysosomal permeabilization is cathepsin B dependent. American journal of physiology 283, G947-956 (2002). 171. Werneburg, N., Guicciardi, M.E., Yin, X.M. & Gores, G.J. TNF-alpha-mediated lysosomal permeabilization is FAN and caspase 8/Bid dependent. American journal of physiology 287, G436-443 (2004). 172. Gyrd-Hansen, M. et al. Apoptosome-independent activation of the lysosomal cell death pathway by caspase-9. Molecular and cellular biology 26, 7880-7891 (2006). 173. Adam-Klages, S. et al. FAN, a novel WD-repeat protein, couples the p55 TNF-receptor to neutral sphingomyelinase. Cell 86, 937-947 (1996). 174. Neumeyer, J. et al. TNF-receptor I defective in internalization allows for cell death through activation of neutral sphingomyelinase. Experimental cell research 312, 2142-2153 (2006). 175. Deiss, L.P., Galinka, H., Berissi, H., Cohen, O. & Kimchi, A. Cathepsin D protease mediates programmed cell death induced by interferon-gamma, Fas/APO-1 and TNF-alpha. The EMBO journal 15, 3861-3870 (1996). 176. Heinrich, M. et al. Cathepsin D targeted by acid sphingomyelinase-derived ceramide. The EMBO journal 18, 5252-5263 (1999). 177. Heinrich, M. et al. Cathepsin D links TNF-induced acid sphingomyelinase to Bid-mediated caspase-9 and -3 activation. Cell death and differentiation 11, 550-563 (2004). 178. Schotte, P., Declercq, W., Van Huffel, S., Vandenabeele, P. & Beyaert, R. Non-specific effects of methyl ketone peptide inhibitors of caspases. FEBS letters 442, 117-121 (1999). 179. Cauwels, A., Janssen, B., Waeytens, A., Cuvelier, C. & Brouckaert, P. Caspase inhibition causes hyperacute tumor necrosis factor-induced shock via oxidative stress and phospholipase A2. Nature immunology 4, 387-393 (2003). 180. Huang, J., Wu, L., Tashiro, S., Onodera, S. & Ikejima, T. The augmentation of TNFalpha- induced cell death in murine L929 fibrosarcoma by the pan-caspase inhibitor Z-VAD-fmk through pre-mitochondrial and MAPK-dependent pathways. Acta medica Okayama 59, 253- 260 (2005). 181. Nagai, H., Matsumaru, K., Feng, G. & Kaplowitz, N. Reduced glutathione depletion causes necrosis and sensitization to tumor necrosis factor-alpha-induced apoptosis in cultured mouse hepatocytes. Hepatology (Baltimore, Md 36, 55-64 (2002). 182. Thannickal, V.J. & Fanburg, B.L. Reactive oxygen species in cell signaling. Am J Physiol Lung Cell Mol Physiol 279, L1005-1028 (2000). 183. Los, M. et al. Activation and caspase-mediated inhibition of PARP: a molecular switch between fibroblast necrosis and apoptosis in death receptor signaling. Molecular biology of the cell 13, 978-988 (2002). Introduction: Tumor Necrosis Factor 57

184. Leist, M., Single, B., Castoldi, A.F., Kuhnle, S. & Nicotera, P. Intracellular adenosine triphosphate (ATP) concentration: a switch in the decision between apoptosis and necrosis. The Journal of experimental medicine 185, 1481-1486 (1997). 185. Ha, H.C. & Snyder, S.H. Poly(ADP-ribose) polymerase is a mediator of necrotic cell death by ATP depletion. Proceedings of the National Academy of Sciences of the United States of America 96, 13978-13982 (1999). 186. Hockenbery, D.M., Oltvai, Z.N., Yin, X.M., Milliman, C.L. & Korsmeyer, S.J. Bcl-2 functions in an antioxidant pathway to prevent apoptosis. Cell 75, 241-251 (1993). 187. Xiang, J., Chao, D.T. & Korsmeyer, S.J. BAX-induced cell death may not require interleukin 1 beta-converting enzyme-like proteases. Proceedings of the National Academy of Sciences of the United States of America 93, 14559-14563 (1996). 188. Ding, W.X., Ni, H.M., DiFrancesca, D., Stolz, D.B. & Yin, X.M. Bid-dependent generation of oxygen radicals promotes death receptor activation-induced apoptosis in murine hepatocytes. Hepatology (Baltimore, Md 40, 403-413 (2004). 189. Kim, Y.S., Morgan, M.J., Choksi, S. & Liu, Z.G. TNF-induced activation of the Nox1 NADPH oxidase and its role in the induction of necrotic cell death. Molecular cell 26, 675- 687 (2007). 190. Vanden Berghe, T., Declercq, W. & Vandenabeele, P. NADPH oxidases: new players in TNF- induced necrotic cell death. Molecular cell 26, 769-771 (2007). 191. Yang, B. & Rizzo, V. TNF-alpha potentiates protein-tyrosine nitration through activation of NADPH oxidase and eNOS localized in membrane rafts and caveolae of bovine aortic endothelial cells. Am J Physiol Heart Circ Physiol 292, H954-962 (2007). 192. Tanaka, M. et al. Embryonic lethality, liver degeneration, and impaired NF-kappa B activation in IKK-beta-deficient mice. Immunity 10, 421-429 (1999). 193. Rudolph, D. et al. Severe liver degeneration and lack of NF-kappaB activation in NEMO/IKKgamma-deficient mice. Genes & development 14, 854-862 (2000). 194. Beg, A.A., Sha, W.C., Bronson, R.T., Ghosh, S. & Baltimore, D. Embryonic lethality and liver degeneration in mice lacking the RelA component of NF-kappa B. Nature 376, 167-170 (1995). 195. Kelliher, M.A. et al. The death domain kinase RIP mediates the TNF-induced NF-kappaB signal. Immunity 8, 297-303 (1998). 196. Lin, Y., Devin, A., Rodriguez, Y. & Liu, Z.G. Cleavage of the death domain kinase RIP by caspase-8 prompts TNF-induced apoptosis. Genes & development 13, 2514-2526 (1999). 197. Tang, G., Yang, J., Minemoto, Y. & Lin, A. Blocking caspase-3-mediated proteolysis of IKKbeta suppresses TNF-alpha-induced apoptosis. Molecular cell 8, 1005-1016 (2001). 198. Ravi, R., Bedi, A., Fuchs, E.J. & Bedi, A. CD95 (Fas)-induced caspase-mediated proteolysis of NF-kappaB. Cancer research 58, 882-886 (1998). 199. Levkau, B., Scatena, M., Giachelli, C.M., Ross, R. & Raines, E.W. Apoptosis overrides survival signals through a caspase-mediated dominant-negative NF-kappa B loop. Nature cell biology 1, 227-233 (1999). 200. Bachelder, R.E. et al. p53 inhibits alpha 6 beta 4 integrin survival signaling by promoting the caspase 3-dependent cleavage of AKT/PKB. The Journal of cell biology 147, 1063-1072 (1999). 201. Martinon, F., Holler, N., Richard, C. & Tschopp, J. Activation of a pro-apoptotic amplification loop through inhibition of NF-kappaB-dependent survival signals by caspase- mediated inactivation of RIP. FEBS letters 468, 134-136 (2000). 202. Sakon, S. et al. NF-kappaB inhibits TNF-induced accumulation of ROS that mediate prolonged MAPK activation and necrotic cell death. The EMBO journal 22, 3898-3909 (2003). 203. Wong, G.H. & Goeddel, D.V. Induction of manganous superoxide dismutase by tumor necrosis factor: possible protective mechanism. Science (New York, N.Y 242, 941-944 (1988). 204. Wong, G.H., Elwell, J.H., Oberley, L.W. & Goeddel, D.V. Manganous superoxide dismutase is essential for cellular resistance to cytotoxicity of tumor necrosis factor. Cell 58, 923-931 (1989). 205. Hochedlinger, K., Wagner, E.F. & Sabapathy, K. Differential effects of JNK1 and JNK2 on signal specific induction of apoptosis. Oncogene 21, 2441-2445 (2002). 206. Muhlenbeck, F. et al. TRAIL/Apo2L activates c-Jun NH2-terminal kinase (JNK) via caspase- dependent and caspase-independent pathways. The Journal of biological chemistry 273, 33091-33098 (1998). 58 Chapter 1

207. Tang, G. et al. Inhibition of JNK activation through NF-kappaB target genes. Nature 414, 313-317 (2001). 208. Lin, A. Activation of the JNK signaling pathway: breaking the brake on apoptosis. Bioessays 25, 17-24 (2003). 209. De Smaele, E. et al. Induction of gadd45beta by NF-kappaB downregulates pro-apoptotic JNK signalling. Nature 414, 308-313 (2001). 210. Papa, S. et al. Gadd45 beta mediates the NF-kappa B suppression of JNK signalling by targeting MKK7/JNKK2. Nature cell biology 6, 146-153 (2004). 211. Papa, S. et al. Insights into the structural basis of the GADD45beta-mediated inactivation of the JNK kinase, MKK7/JNKK2. The Journal of biological chemistry 282, 19029-19041 (2007). 212. Lee, E.G. et al. Failure to regulate TNF-induced NF-kappaB and cell death responses in A20- deficient mice. Science (New York, N.Y 289, 2350-2354 (2000). 213. Wajant, H., Pfizenmaier, K. & Scheurich, P. Tumor necrosis factor signaling. Cell death and differentiation 10, 45-65 (2003). 214. Lamb, J.A., Ventura, J.J., Hess, P., Flavell, R.A. & Davis, R.J. JunD mediates survival signaling by the JNK signal transduction pathway. Molecular cell 11, 1479-1489 (2003). 215. Shaulian, E. & Karin, M. AP-1 as a regulator of cell life and death. Nature cell biology 4, E131-136 (2002). 216. Maundrell, K. et al. Bcl-2 undergoes phosphorylation by c-Jun N-terminal kinase/stress- activated protein kinases in the presence of the constitutively active GTP-binding protein Rac1. The Journal of biological chemistry 272, 25238-25242 (1997). 217. Yamamoto, K., Ichijo, H. & Korsmeyer, S.J. BCL-2 is phosphorylated and inactivated by an ASK1/Jun N-terminal protein kinase pathway normally activated at G(2)/M. Molecular and cellular biology 19, 8469-8478 (1999). 218. Lei, K. & Davis, R.J. JNK phosphorylation of Bim-related members of the Bcl2 family induces Bax-dependent apoptosis. Proceedings of the National Academy of Sciences of the United States of America 100, 2432-2437 (2003). 219. Deng, Y., Ren, X., Yang, L., Lin, Y. & Wu, X. A JNK-dependent pathway is required for TNFalpha-induced apoptosis. Cell 115, 61-70 (2003). 220. Deveraux, Q.L. et al. IAPs block apoptotic events induced by caspase-8 and cytochrome c by direct inhibition of distinct caspases. The EMBO journal 17, 2215-2223 (1998). 221. Chang, L. et al. The E3 ubiquitin ligase itch couples JNK activation to TNFalpha-induced cell death by inducing c-FLIP(L) turnover. Cell 124, 601-613 (2006). 222. Geisler, F., Algul, H., Paxian, S. & Schmid, R.M. Genetic inactivation of RelA/p65 sensitizes adult mouse hepatocytes to TNF-induced apoptosis in vivo and in vitro. Gastroenterology 132, 2489-2503 (2007). 223. Clark, M.A., Chen, M.J., Crooke, S.T. & Bomalaski, J.S. Tumour necrosis factor (cachectin) induces phospholipase A2 activity and synthesis of a phospholipase A2-activating protein in endothelial cells. The Biochemical journal 250, 125-132 (1988). 224. Schievella, A.R., Regier, M.K., Smith, W.L. & Lin, L.L. Calcium-mediated translocation of cytosolic phospholipase A2 to the nuclear envelope and endoplasmic reticulum. The Journal of biological chemistry 270, 30749-30754 (1995). 225. Hefner, Y. et al. Serine 727 phosphorylation and activation of cytosolic phospholipase A2 by MNK1-related protein kinases. The Journal of biological chemistry 275, 37542-37551 (2000). 226. Jupp, O.J., Vandenabeele, P. & MacEwan, D.J. Distinct regulation of cytosolic phospholipase A2 phosphorylation, translocation, proteolysis and activation by tumour necrosis factor- receptor subtypes. The Biochemical journal 374, 453-461 (2003). 227. McFarlane, S.M., Anderson, H.M., Tucker, S.J., Jupp, O.J. & MacEwan, D.J. Unmodified calcium concentrations in tumour necrosis factor receptor subtype-mediated apoptotic cell death. Molecular and cellular biochemistry 211, 19-26 (2000). 228. Pober, J.S. & Sessa, W.C. Evolving functions of endothelial cells in inflammation. Nat Rev Immunol 7, 803-815 (2007). 229. Wu, T., Ikezono, T., Angus, C.W. & Shelhamer, J.H. Tumor necrosis factor-alpha induces the 85-kDa cytosolic phospholipase A2 gene expression in human bronchial epithelial cells. Biochimica et biophysica acta 1310, 175-184 (1996). 230. Leval, X., Julemont, F., Delarge, J., Pirotte, B. & Dogne, J.M. New trends in dual 5- LOX/COX inhibition. Current medicinal chemistry 9, 941-962 (2002). Introduction: Tumor Necrosis Factor 59

231. Surh, Y.J. et al. Molecular mechanisms underlying chemopreventive activities of anti- inflammatory phytochemicals: down-regulation of COX-2 and iNOS through suppression of NF-kappa B activation. Mutation research 480-481, 243-268 (2001). 232. Yan, Y. et al. NFAT3 is specifically required for TNF-alpha-induced cyclooxygenase-2 (COX-2) expression and transformation of Cl41 cells. Journal of cell science 119, 2985-2994 (2006). 233. Petrin, D., Turcotte, S., Gilbert, A.K., Rola-Pleszczynski, M. & Stankova, J. The anti- apoptotic effect of leukotriene B4 in neutrophils: a role for phosphatidylinositol 3-kinase, extracellular signal-regulated kinase and Mcl-1. Cellular signalling 18, 479-487 (2006). 234. Avis, I. et al. Five-lipoxygenase inhibitors can mediate apoptosis in human breast cancer cell lines through complex eicosanoid interactions. Faseb J 15, 2007-2009 (2001). 235. Yamazaki, R., Kusunoki, N., Matsuzaki, T., Hashimoto, S. & Kawai, S. Selective cyclooxygenase-2 inhibitors show a differential ability to inhibit proliferation and induce apoptosis of colon adenocarcinoma cells. FEBS letters 531, 278-284 (2002). 236. Sun, Y., Tang, X.M., Half, E., Kuo, M.T. & Sinicrope, F.A. Cyclooxygenase-2 overexpression reduces apoptotic susceptibility by inhibiting the cytochrome c-dependent apoptotic pathway in human colon cancer cells. Cancer research 62, 6323-6328 (2002). 237. Wu, T., Leng, J., Han, C. & Demetris, A.J. The cyclooxygenase-2 inhibitor celecoxib blocks phosphorylation of Akt and induces apoptosis in human cholangiocarcinoma cells. Molecular cancer therapeutics 3, 299-307 (2004). 238. Tong, W.G., Ding, X.Z., Talamonti, M.S., Bell, R.H. & Adrian, T.E. LTB4 stimulates growth of human pancreatic cancer cells via MAPK and PI-3 kinase pathways. Biochemical and biophysical research communications 335, 949-956 (2005). 239. Tong, W.G., Ding, X.Z. & Adrian, T.E. The mechanisms of lipoxygenase inhibitor-induced apoptosis in human breast cancer cells. Biochemical and biophysical research communications 296, 942-948 (2002). 240. Vondracek, J. et al. Inhibitors of arachidonic acid metabolism potentiate tumour necrosis factor-alpha-induced apoptosis in HL-60 cells. European journal of pharmacology 424, 1-11 (2001). 241. Kovarikova, M., Hofmanova, J., Soucek, K. & Kozubik, A. The effects of TNF-alpha and inhibitors of arachidonic acid metabolism on human colon HT-29 cells depend on differentiation status. Differentiation; research in biological diversity 72, 23-31 (2004). 242. Caccese, D. et al. Superoxide anion and hydroxyl radical release by collagen-induced platelet aggregation--role of arachidonic acid metabolism. Thrombosis and haemostasis 83, 485-490 (2000). 243. Muralikrishna Adibhatla, R. & Hatcher, J.F. Phospholipase A2, reactive oxygen species, and lipid peroxidation in cerebral ischemia. Free radical biology & medicine 40, 376-387 (2006). 244. Nanda, B.L. et al. PLA2 mediated arachidonate free radicals: PLA2 inhibition and neutralization of free radicals by anti-oxidants--a new role as anti-inflammatory molecule. Current topics in medicinal chemistry 7, 765-777 (2007). 245. Nethery, D. et al. PLA(2) dependence of diaphragm mitochondrial formation of reactive oxygen species. J Appl Physiol 89, 72-80 (2000). 246. Kilbourn, R.G. et al. NG-methyl-L-arginine inhibits tumor necrosis factor-induced hypotension: implications for the involvement of nitric oxide. Proceedings of the National Academy of Sciences of the United States of America 87, 3629-3632 (1990). 247. Brouckaert, P. et al. Inhibition of and sensitization to the lethal effects of tumor necrosis factor. Journal of inflammation 47, 18-26 (1995). 248. Alderton, W.K., Cooper, C.E. & Knowles, R.G. Nitric oxide synthases: structure, function and inhibition. The Biochemical journal 357, 593-615 (2001). 249. Guzik, T.J., Korbut, R. & Adamek-Guzik, T. Nitric oxide and superoxide in inflammation and immune regulation. J Physiol Pharmacol 54, 469-487 (2003). 250. Michel, T. & Feron, O. Nitric oxide synthases: which, where, how, and why? The Journal of clinical investigation 100, 2146-2152 (1997). 251. Pacher, P., Beckman, J.S. & Liaudet, L. Nitric oxide and peroxynitrite in health and disease. Physiological reviews 87, 315-424 (2007). 252. Kleinert, H., Euchenhofer, C., Ihrig-Biedert, I. & Forstermann, U. In murine 3T3 fibroblasts, different second messenger pathways resulting in the induction of NO synthase II (iNOS) converge in the activation of transcription factor NF-kappaB. The Journal of biological chemistry 271, 6039-6044 (1996). 60 Chapter 1

253. Jeon, Y.J., Han, S.H., Lee, Y.W., Yea, S.S. & Yang, K.H. Inhibition of NF-kappa B/Rel nuclear translocation by dexamethasone: mechanism for the inhibition of iNOS gene expression. Biochemistry and molecular biology international 45, 435-441 (1998). 254. Kleinert, H., Schwarz, P.M. & Forstermann, U. Regulation of the expression of inducible nitric oxide synthase. Biological chemistry 384, 1343-1364 (2003). 255. Da Silva, J., Pierrat, B., Mary, J.L. & Lesslauer, W. Blockade of p38 mitogen-activated protein kinase pathway inhibits inducible nitric-oxide synthase expression in mouse astrocytes. The Journal of biological chemistry 272, 28373-28380 (1997). 256. Cruz, M.T., Duarte, C.B., Goncalo, M., Carvalho, A.P. & Lopes, M.C. Involvement of JAK2 and MAPK on type II nitric oxide synthase expression in skin-derived dendritic cells. The American journal of physiology 277, C1050-1057 (1999). 257. Chan, E.D., Winston, B.W., Uh, S.T., Remigio, L.K. & Riches, D.W. Systematic evaluation of the mitogen-activated protein kinases in the induction of iNOS by tumor necrosis factor-alpha and interferon-gamma. Chest 116, 91S-92S (1999). 258. Kawanaka, H. et al. Activation of eNOS in rat portal hypertensive gastric mucosa is mediated by TNF-alpha via the PI 3-kinase-Akt signaling pathway. Hepatology (Baltimore, Md 35, 393- 402 (2002). 259. De Palma, C., Meacci, E., Perrotta, C., Bruni, P. & Clementi, E. Endothelial nitric oxide synthase activation by tumor necrosis factor alpha through neutral sphingomyelinase 2, sphingosine kinase 1, and sphingosine 1 phosphate receptors: a novel pathway relevant to the pathophysiology of endothelium. Arteriosclerosis, thrombosis, and vascular biology 26, 99- 105 (2006). 260. Dimmeler, S. et al. Activation of nitric oxide synthase in endothelial cells by Akt-dependent phosphorylation. Nature 399, 601-605 (1999). 261. Fulton, D. et al. Regulation of endothelium-derived nitric oxide production by the protein kinase Akt. Nature 399, 597-601 (1999). 262. Shaul, P.W. et al. Acylation targets endothelial nitric-oxide synthase to plasmalemmal caveolae. The Journal of biological chemistry 271, 6518-6522 (1996). 263. Veldman, R.J. et al. A neutral sphingomyelinase resides in sphingolipid-enriched microdomains and is inhibited by the caveolin-scaffolding domain: potential implications in tumour necrosis factor signalling. The Biochemical journal 355, 859-868 (2001). 264. Anderson, H.D., Rahmutula, D. & Gardner, D.G. Tumor necrosis factor-alpha inhibits endothelial nitric-oxide synthase gene promoter activity in bovine aortic endothelial cells. The Journal of biological chemistry 279, 963-969 (2004). 265. Yoshizumi, M., Perrella, M.A., Burnett, J.C., Jr. & Lee, M.E. Tumor necrosis factor downregulates an endothelial nitric oxide synthase mRNA by shortening its half-life. Circulation research 73, 205-209 (1993). 266. Mohamed, F. et al. Lack of role for nitric oxide (NO) in the selective destabilization of endothelial NO synthase mRNA by tumor necrosis factor-alpha. Arteriosclerosis, thrombosis, and vascular biology 15, 52-57 (1995). 267. Cauwels, A. et al. Protection against TNF-induced lethal shock by soluble guanylate cyclase inhibition requires functional inducible nitric oxide synthase. Immunity 13, 223-231 (2000). 268. Cauwels, A., Bultinck, J. & Brouckaert, P. Dual role of endogenous nitric oxide in tumor necrosis factor shock: induced NO tempers oxidative stress. Cell Mol Life Sci 62, 1632-1640 (2005). 269. Rubbo, H. et al. Nitric oxide regulation of superoxide and peroxynitrite-dependent lipid peroxidation. Formation of novel nitrogen-containing oxidized lipid derivatives. The Journal of biological chemistry 269, 26066-26075 (1994). 270. Choi, B.M., Pae, H.O., Jang, S.I., Kim, Y.M. & Chung, H.T. Nitric oxide as a pro-apoptotic as well as anti-apoptotic modulator. Journal of biochemistry and molecular biology 35, 116-126 (2002). 271. Rossig, L. et al. Nitric oxide inhibits caspase-3 by S-nitrosation in vivo. The Journal of biological chemistry 274, 6823-6826 (1999). 272. Kim, Y.M. et al. Nitric oxide prevents tumor necrosis factor alpha-induced rat hepatocyte apoptosis by the interruption of mitochondrial apoptotic signaling through S-nitrosylation of caspase-8. Hepatology (Baltimore, Md 32, 770-778 (2000). 273. Kim, J.E. & Tannenbaum, S.R. S-Nitrosation regulates the activation of endogenous procaspase-9 in HT-29 human colon carcinoma cells. The Journal of biological chemistry 279, 9758-9764 (2004). Introduction: Tumor Necrosis Factor 61

274. Mitchell, D.A., Morton, S.U., Fernhoff, N.B. & Marletta, M.A. Thioredoxin is required for S- nitrosation of procaspase-3 and the inhibition of apoptosis in Jurkat cells. Proceedings of the National Academy of Sciences of the United States of America 104, 11609-11614 (2007). 275. Brookes, P.S. et al. Concentration-dependent effects of nitric oxide on mitochondrial permeability transition and cytochrome c release. The Journal of biological chemistry 275, 20474-20479 (2000). 276. Huerta-Yepez, S. et al. Nitric oxide sensitizes prostate carcinoma cell lines to TRAIL- mediated apoptosis via inactivation of NF-kappa B and inhibition of Bcl-xl expression. Oncogene 23, 4993-5003 (2004). 277. Grassme, H. et al. CD95 signaling via ceramide-rich membrane rafts. The Journal of biological chemistry 276, 20589-20596 (2001). 278. Grassme, H., Schwarz, H. & Gulbins, E. Molecular mechanisms of ceramide-mediated CD95 clustering. Biochemical and biophysical research communications 284, 1016-1030 (2001). 279. Wiegmann, K. et al. Requirement of FADD for tumor necrosis factor-induced activation of acid sphingomyelinase. The Journal of biological chemistry 274, 5267-5270 (1999). 280. Garcia-Ruiz, C. et al. Defective TNF-alpha-mediated hepatocellular apoptosis and liver damage in acidic sphingomyelinase knockout mice. The Journal of clinical investigation 111, 197-208 (2003). 281. Tcherkasowa, A.E. et al. Interaction with factor associated with neutral sphingomyelinase activation, a WD motif-containing protein, identifies receptor for activated C-kinase 1 as a novel component of the signaling pathways of the p55 TNF receptor. J Immunol 169, 5161- 5170 (2002). 282. Mohlig, H. et al. The WD repeat protein FAN regulates lysosome size independent from abnormal downregulation/membrane recruitment of protein kinase C. Experimental cell research 313, 2703-2718 (2007). 283. Segui, B. et al. Involvement of FAN in TNF-induced apoptosis. The Journal of clinical investigation 108, 143-151 (2001). 284. Paris, F. et al. Natural ceramide reverses Fas resistance of acid sphingomyelinase(-/-) hepatocytes. The Journal of biological chemistry 276, 8297-8305 (2001). 285. Cremesti, A. et al. Ceramide enables fas to cap and kill. The Journal of biological chemistry 276, 23954-23961 (2001). 286. Grassme, H., Cremesti, A., Kolesnick, R. & Gulbins, E. Ceramide-mediated clustering is required for CD95-DISC formation. Oncogene 22, 5457-5470 (2003). 287. Garcia-Ruiz, C., Colell, A., Mari, M., Morales, A. & Fernandez-Checa, J.C. Direct effect of ceramide on the mitochondrial electron transport chain leads to generation of reactive oxygen species. Role of mitochondrial glutathione. The Journal of biological chemistry 272, 11369- 11377 (1997). 288. Garcia-Ruiz, C., Colell, A., Paris, R. & Fernandez-Checa, J.C. Direct interaction of GD3 ganglioside with mitochondria generates reactive oxygen species followed by mitochondrial permeability transition, cytochrome c release, and caspase activation. Faseb J 14, 847-858 (2000). 289. Garcia-Ruiz, C. et al. Trafficking of ganglioside GD3 to mitochondria by tumor necrosis factor-alpha. The Journal of biological chemistry 277, 36443-36448 (2002). 290. Ozes, O.N. et al. NF-kappaB activation by tumour necrosis factor requires the Akt serine- threonine kinase. Nature 401, 82-85 (1999). 291. Pastorino, J.G., Tafani, M. & Farber, J.L. Tumor necrosis factor induces phosphorylation and translocation of BAD through a phosphatidylinositide-3-OH kinase-dependent pathway. The Journal of biological chemistry 274, 19411-19416 (1999). 292. Burow, M.E. et al. PI3-K/AKT regulation of NF-kappaB signaling events in suppression of TNF-induced apoptosis. Biochemical and biophysical research communications 271, 342-345 (2000). 293. Mayo, M.W. et al. PTEN blocks tumor necrosis factor-induced NF-kappa B-dependent transcription by inhibiting the transactivation potential of the p65 subunit. The Journal of biological chemistry 277, 11116-11125 (2002). 294. Gustin, J.A., Maehama, T., Dixon, J.E. & Donner, D.B. The PTEN tumor suppressor protein inhibits tumor necrosis factor-induced nuclear factor kappa B activity. The Journal of biological chemistry 276, 27740-27744 (2001). 295. Osawa, Y. et al. TNF-alpha-induced sphingosine 1-phosphate inhibits apoptosis through a phosphatidylinositol 3-kinase/Akt pathway in human hepatocytes. J Immunol 167, 173-180 (2001). 62 Chapter 1

296. Dempsey, P.W., Doyle, S.E., He, J.Q. & Cheng, G. The signaling adaptors and pathways activated by TNF superfamily. Cytokine & growth factor reviews 14, 193-209 (2003). 297. Yang, Y.L. & Li, X.M. The IAP family: endogenous caspase inhibitors with multiple biological activities. Cell research 10, 169-177 (2000). 298. Li, X., Yang, Y. & Ashwell, J.D. TNF-RII and c-IAP1 mediate ubiquitination and degradation of TRAF2. Nature 416, 345-347 (2002). 299. Zhao, Y., Conze, D.B., Hanover, J.A. & Ashwell, J.D. Tumor necrosis factor receptor 2 signaling induces selective c-IAP1-dependent ASK1 ubiquitination and terminates mitogen- activated protein kinase signaling. The Journal of biological chemistry 282, 7777-7782 (2007). 300. Weiss, T. et al. Enhancement of TNF receptor p60-mediated cytotoxicity by TNF receptor p80: requirement of the TNF receptor-associated factor-2 binding site. J Immunol 158, 2398- 2404 (1997). 301. Grell, M. et al. Induction of cell death by tumour necrosis factor (TNF) receptor 2, CD40 and CD30: a role for TNF-R1 activation by endogenous membrane-anchored TNF. The EMBO journal 18, 3034-3043 (1999). 302. Zheng, L. et al. Induction of apoptosis in mature T cells by tumour necrosis factor. Nature 377, 348-351 (1995). 303. Teh, H.S., Seebaran, A. & Teh, S.J. TNF receptor 2-deficient CD8 T cells are resistant to Fas/Fas ligand-induced cell death. J Immunol 165, 4814-4821 (2000). 304. Hoshino, K. et al. Cutting edge: Toll-like receptor 4 (TLR4)-deficient mice are hyporesponsive to lipopolysaccharide: evidence for TLR4 as the Lps gene product. J Immunol 162, 3749-3752 (1999). 305. Nathan, C.F. Secretory products of macrophages. The Journal of clinical investigation 79, 319-326 (1987). 306. Billack, B. Macrophage activation: role of toll-like receptors, nitric oxide, and nuclear factor kappa B. American journal of pharmaceutical education 70, 102 (2006). 307. Janeway, C.A., Jr. & Medzhitov, R. Innate immune recognition. Annual review of immunology 20, 197-216 (2002). 308. Guo, R.F., Riedemann, N.C. & Ward, P.A. Role of C5a-C5aR interaction in sepsis. Shock (Augusta, Ga 21, 1-7 (2004). 309. Riedemann, N.C., Guo, R.F. & Ward, P.A. A key role of C5a/C5aR activation for the development of sepsis. Journal of leukocyte biology 74, 966-970 (2003). 310. Gordon, J.R. & Galli, S.J. Mast cells as a source of both preformed and immunologically inducible TNF-alpha/cachectin. Nature 346, 274-276 (1990). 311. Janeway, C.A., Travers, P., Walport, M., Shlomchik, M. Immunobiology: The immune system in health and disease, Edn. 5. (Garland Publishing, New York; 2001). 312. Schenkel, A.R., Dufour, E.M., Chew, T.W., Sorg, E. & Muller, W.A. The Murine CD99- Related Molecule CD99-Like 2 (CD99L2) Is an Adhesion Molecule Involved in the Inflammatory Response. Cell communication & adhesion 14, 227-237 (2007). 313. Ulbrich, H., Eriksson, E.E. & Lindbom, L. Leukocyte and endothelial cell adhesion molecules as targets for therapeutic interventions in inflammatory disease. Trends in pharmacological sciences 24, 640-647 (2003). 314. Wang, J. et al. Histamine antagonizes tumor necrosis factor (TNF) signaling by stimulating TNF receptor shedding from the cell surface and Golgi storage pool. The Journal of biological chemistry 278, 21751-21760 (2003). 315. Petrache, I., Birukova, A., Ramirez, S.I., Garcia, J.G. & Verin, A.D. The role of the microtubules in tumor necrosis factor-alpha-induced endothelial cell permeability. American journal of respiratory cell and molecular biology 28, 574-581 (2003). 316. Angelini, D.J. et al. TNF-alpha increases tyrosine phosphorylation of vascular endothelial cadherin and opens the paracellular pathway through fyn activation in human lung endothelia. Am J Physiol Lung Cell Mol Physiol 291, L1232-1245 (2006). 317. Cotran, R.S. & Pober, J.S. Cytokine-endothelial interactions in inflammation, immunity, and vascular injury. J Am Soc Nephrol 1, 225-235 (1990). 318. Pan, J., Xia, L. & McEver, R.P. Comparison of promoters for the murine and human P- selectin genes suggests species-specific and conserved mechanisms for transcriptional regulation in endothelial cells. The Journal of biological chemistry 273, 10058-10067 (1998). 319. Kunkel, E.J. & Ley, K. Distinct phenotype of E-selectin-deficient mice. E-selectin is required for slow leukocyte rolling in vivo. Circulation research 79, 1196-1204 (1996). Introduction: Tumor Necrosis Factor 63

320. Munro, J.M., Pober, J.S. & Cotran, R.S. Recruitment of neutrophils in the local endotoxin response: association with de novo endothelial expression of endothelial leukocyte adhesion molecule-1. Laboratory investigation; a journal of technical methods and pathology 64, 295- 299 (1991). 321. Denk, A. et al. Activation of NF-kappa B via the Ikappa B kinase complex is both essential and sufficient for proinflammatory gene expression in primary endothelial cells. The Journal of biological chemistry 276, 28451-28458 (2001). 322. Sohn, R.H. et al. Regulation of endothelial thrombomodulin expression by inflammatory cytokines is mediated by activation of nuclear factor-kappa B. Blood 105, 3910-3917 (2005). 323. Conway, E.M., Bach, R., Rosenberg, R.D. & Konigsberg, W.H. Tumor necrosis factor enhances expression of tissue factor mRNA in endothelial cells. Thrombosis research 53, 231- 241 (1989). 324. Bevilacqua, M.P. et al. Recombinant tumor necrosis factor induces procoagulant activity in cultured human vascular endothelium: characterization and comparison with the actions of interleukin 1. Proceedings of the National Academy of Sciences of the United States of America 83, 4533-4537 (1986). 325. Parry, G.C. & Mackman, N. Transcriptional regulation of tissue factor expression in human endothelial cells. Arteriosclerosis, thrombosis, and vascular biology 15, 612-621 (1995). 326. Li, J.H. & Pober, J.S. The cathepsin B death pathway contributes to TNF plus IFN-gamma- mediated human endothelial injury. J Immunol 175, 1858-1866 (2005). 327. Petrache, I. et al. Caspase-dependent cleavage of myosin light chain kinase (MLCK) is involved in TNF-alpha-mediated bovine pulmonary endothelial cell apoptosis. Faseb J 17, 407-416 (2003). 328. Petrache, I., Crow, M.T., Neuss, M. & Garcia, J.G. Central involvement of Rho family GTPases in TNF-alpha-mediated bovine pulmonary endothelial cell apoptosis. Biochemical and biophysical research communications 306, 244-249 (2003). 329. Deshpande, S.S., Angkeow, P., Huang, J., Ozaki, M. & Irani, K. Rac1 inhibits TNF-alpha- induced endothelial cell apoptosis: dual regulation by reactive oxygen species. Faseb J 14, 1705-1714 (2000). 330. Marino, M.W. et al. Characterization of tumor necrosis factor-deficient mice. Proceedings of the National Academy of Sciences of the United States of America 94, 8093-8098 (1997). 331. Flynn, J.L. et al. Tumor necrosis factor-alpha is required in the protective immune response against Mycobacterium tuberculosis in mice. Immunity 2, 561-572 (1995). 332. Pfeffer, K. et al. Mice deficient for the 55 kd tumor necrosis factor receptor are resistant to endotoxic shock, yet succumb to L. monocytogenes infection. Cell 73, 457-467 (1993). 333. Everest, P., Roberts, M. & Dougan, G. Susceptibility to Salmonella typhimurium infection and effectiveness of vaccination in mice deficient in the tumor necrosis factor alpha p55 receptor. Infection and immunity 66, 3355-3364 (1998). 334. Castanos-Velez, E. et al. Trypanosoma cruzi infection in tumor necrosis factor receptor p55- deficient mice. Infection and immunity 66, 2960-2968 (1998). 335. Vieira, L.Q. et al. Mice lacking the TNF receptor p55 fail to resolve lesions caused by infection with Leishmania major, but control parasite replication. J Immunol 157, 827-835 (1996). 336. Zarkesh-Esfahani, H. et al. Leptin indirectly activates human neutrophils via induction of TNF-alpha. J Immunol 172, 1809-1814 (2004). 337. Abe, Y., Osuka, Y., Nakata, T., Kashu, Y. & Kimura, S. The functional role of 55- and 75- kDa tumour necrosis factor receptors in human polymorphonuclear cells in vitro. Cytokine 7, 39-49 (1995). 338. Schleiffenbaum, B. & Fehr, J. The tumor necrosis factor receptor and human neutrophil function. Deactivation and cross-deactivation of tumor necrosis factor-induced neutrophil responses by receptor down-regulation. The Journal of clinical investigation 86, 184-195 (1990). 339. Montecucco, F. et al. Tumor necrosis factor-alpha (TNF-alpha) induces integrin CD11b/CD18 (Mac-1) up-regulation and migration to the CC chemokine CCL3 (MIP-1alpha) on human neutrophils through defined signalling pathways. Cellular signalling 20, 557-568 (2008). 340. Meyer, J.D. et al. Tumor necrosis factor-enhanced leukotriene B4 generation and chemotaxis in human neutrophils. Arch Surg 123, 1454-1458 (1988). 341. Kapp, A. & Zeck-Kapp, G. Activation of the oxidative metabolism in human polymorphonuclear neutrophilic granulocytes: the role of immuno-modulating cytokines. The Journal of investigative dermatology 95, 94S-99S (1990). 64 Chapter 1

342. van Overveld, F.J., Jorens, P.G., Rampart, M., de Backer, W. & Vermeire, P.A. Tumour necrosis factor stimulates human skin mast cells to release histamine and tryptase. Clin Exp Allergy 21, 711-714 (1991). 343. Coward, W.R. et al. NF-kappa B and TNF-alpha: a positive autocrine loop in human lung mast cells? J Immunol 169, 5287-5293 (2002). 344. Hoffman, M. & Weinberg, J.B. Tumor necrosis factor-alpha induces increased hydrogen peroxide production and Fc receptor expression, but not increased Ia antigen expression by peritoneal macrophages. Journal of leukocyte biology 42, 704-707 (1987). 345. Ding, A.H., Nathan, C.F. & Stuehr, D.J. Release of reactive nitrogen intermediates and reactive oxygen intermediates from mouse peritoneal macrophages. Comparison of activating cytokines and evidence for independent production. J Immunol 141, 2407-2412 (1988). 346. Wang, J.M., Walter, S. & Mantovani, A. Re-evaluation of the chemotactic activity of tumour necrosis factor for monocytes. Immunology 71, 364-367 (1990). 347. Hori, K., Ehrke, M.J., Mace, K. & Mihich, E. Effect of recombinant tumor necrosis factor on tumoricidal activation of murine macrophages: synergism between tumor necrosis factor and gamma-interferon. Cancer research 47, 5868-5874 (1987). 348. Talmadge, J.E. et al. Immunomodulatory properties of recombinant murine and human tumor necrosis factor. Cancer research 48, 544-550 (1988). 349. Ellerin, T., Rubin, R.H. & Weinblatt, M.E. Infections and anti-tumor necrosis factor alpha therapy. Arthritis and rheumatism 48, 3013-3022 (2003). 350. Acharya, A. & Singh, S.M. Effect of TNF-alpha on the induction of apoptosis in murine macrophages: role of interleukin-1 beta converting enzyme. Int J Immunopathol Pharmacol 14, 5-10 (2001). 351. Zucali, J.R., Elfenbein, G.J., Barth, K.C. & Dinarello, C.A. Effects of human interleukin 1 and human tumor necrosis factor on human T lymphocyte colony formation. The Journal of clinical investigation 80, 772-777 (1987). 352. Tartaglia, L.A. et al. Stimulation of human T-cell proliferation by specific activation of the 75-kDa tumor necrosis factor receptor. J Immunol 151, 4637-4641 (1993). 353. La Riviere, G., Klein Gebbinck, J.W., Schipper, C.A. & Roos, E. Tumour necrosis factor- alpha stimulates invasiveness of T-cell hybridomas and cytotoxic T-cell clones by a pertussis toxin-insensitive mechanism. Immunology 75, 269-274 (1992). 354. Hancock, J.T., Henderson, L.M. & Jones, O.T. Superoxide generation by EBV-transformed B lymphocytes. Activation by IL-1 beta, TNF-alpha and receptor independent stimuli. Immunology 71, 213-217 (1990). 355. Wherry, J.C., Schreiber, R.D. & Unanue, E.R. Regulation of gamma interferon production by natural killer cells in scid mice: roles of tumor necrosis factor and bacterial stimuli. Infection and immunity 59, 1709-1715 (1991). 356. Ostensen, M.E., Thiele, D.L. & Lipsky, P.E. Tumor necrosis factor-alpha enhances cytolytic activity of human natural killer cells. J Immunol 138, 4185-4191 (1987). 357. Cumberbatch, M. & Kimber, I. Dermal tumour necrosis factor-alpha induces dendritic cell migration to draining lymph nodes, and possibly provides one stimulus for Langerhans' cell migration. Immunology 75, 257-263 (1992). 358. Santiago-Schwarz, F., Divaris, N., Kay, C. & Carsons, S.E. Mechanisms of tumor necrosis factor-granulocyte-macrophage colony-stimulating factor-induced dendritic cell development. Blood 82, 3019-3028 (1993). 359. Chomarat, P., Dantin, C., Bennett, L., Banchereau, J. & Palucka, A.K. TNF skews monocyte differentiation from macrophages to dendritic cells. J Immunol 171, 2262-2269 (2003). 360. Kaufmann, S.H. Protection against tuberculosis: cytokines, T cells, and macrophages. Annals of the rheumatic diseases 61 Suppl 2, ii54-58 (2002). 361. Mohan, V.P. et al. Effects of tumor necrosis factor alpha on host immune response in chronic persistent tuberculosis: possible role for limiting pathology. Infection and immunity 69, 1847- 1855 (2001). 362. Keane, J. TNF-blocking agents and tuberculosis: new drugs illuminate an old topic. Rheumatology (Oxford, England) 44, 714-720 (2005). 363. Fernandez-Nebro, A. et al. Effectiveness, predictive response factors, and safety of anti-tumor necrosis factor (TNF) therapies in anti-TNF-naive rheumatoid arthritis. The Journal of rheumatology 34, 2334-2342 (2007). 364. Kozuch, P.L. & Hanauer, S.B. Treatment of inflammatory bowel disease: A review of medical therapy. World J Gastroenterol 14, 354-377 (2008). Introduction: Tumor Necrosis Factor 65

365. Menter, A. et al. Adalimumab therapy for moderate to severe psoriasis: A randomized, controlled phase III trial. Journal of the American Academy of Dermatology 58, 106-115 (2008). 366. Tyring, S. et al. Long-term safety and efficacy of 50 mg of etanercept twice weekly in patients with psoriasis. Archives of dermatology 143, 719-726 (2007). 367. Bradley, J. TNF-mediated inflammatory disease. J Pathol 214, 149-160 (2008). 368. Tracey, K.J. et al. Shock and tissue injury induced by recombinant human cachectin. Science (New York, N.Y 234, 470-474 (1986). 369. Beutler, B., Milsark, I.W. & Cerami, A.C. Passive immunization against cachectin/tumor necrosis factor protects mice from lethal effect of endotoxin. Science (New York, N.Y 229, 869-871 (1985). 370. Tracey, K.J. et al. Anti-cachectin/TNF monoclonal antibodies prevent septic shock during lethal bacteraemia. Nature 330, 662-664 (1987). 371. Silva, A.T., Bayston, K.F. & Cohen, J. Prophylactic and therapeutic effects of a monoclonal antibody to tumor necrosis factor-alpha in experimental gram-negative shock. The Journal of infectious diseases 162, 421-427 (1990). 372. Abraham, E. et al. Efficacy and safety of monoclonal antibody to human tumor necrosis factor alpha in patients with sepsis syndrome. A randomized, controlled, double-blind, multicenter clinical trial. TNF-alpha MAb Sepsis Study Group. Jama 273, 934-941 (1995). 373. Abraham, E. et al. Double-blind randomised controlled trial of monoclonal antibody to human tumour necrosis factor in treatment of septic shock. NORASEPT II Study Group. Lancet 351, 929-933 (1998). 374. Fisher, C.J., Jr. et al. Treatment of septic shock with the tumor necrosis factor receptor:Fc fusion protein. The Soluble TNF Receptor Sepsis Study Group. The New England journal of medicine 334, 1697-1702 (1996). 375. Abraham, E. et al. Lenercept (p55 tumor necrosis factor receptor fusion protein) in severe sepsis and early septic shock: a randomized, double-blind, placebo-controlled, multicenter phase III trial with 1,342 patients. Critical care medicine 29, 503-510 (2001). 376. Clark, I.A. How TNF was recognized as a key mechanism of disease. Cytokine & growth factor reviews 18, 335-343 (2007). 377. Brouckaert, P.G., Leroux-Roels, G.G., Guisez, Y., Tavernier, J. & Fiers, W. In vivo anti- tumour activity of recombinant human and murine TNF, alone and in combination with murine IFN-gamma, on a syngeneic murine melanoma. International journal of cancer 38, 763-769 (1986). 378. Brouckaert, P., Libert, C., Everaerdt, B. & Fiers, W. Selective species specificity of tumor necrosis factor for toxicity in the mouse. Lymphokine and cytokine research 11, 193-196 (1992). 379. Lewis, M. et al. Cloning and expression of cDNAs for two distinct murine tumor necrosis factor receptors demonstrate one receptor is species specific. Proceedings of the National Academy of Sciences of the United States of America 88, 2830-2834 (1991). 380. Ameloot, P. et al. Bioavailability of recombinant tumor necrosis factor determines its lethality in mice. European journal of immunology 32, 2759-2765 (2002). 381. Stoelcker, B. et al. Tumor necrosis factor induces tumor necrosis via tumor necrosis factor receptor type 1-expressing endothelial cells of the tumor vasculature. The American journal of pathology 156, 1171-1176 (2000). 382. Shimomura, K. et al. Recombinant human tumor necrosis factor-alpha: thrombus formation is a cause of anti-tumor activity. International journal of cancer 41, 243-247 (1988). 383. Van de Wiel, P.A., Bloksma, N., Kuper, C.F., Hofhuis, F.M. & Willers, J.M. Macroscopic and microscopic early effects of tumour necrosis factor on murine Meth A sarcoma, and relation to curative activity. J Pathol 157, 65-73 (1989). 384. Eggermont, A.M. et al. Isolated limb perfusion with high-dose tumor necrosis factor-alpha in combination with interferon-gamma and melphalan for nonresectable extremity soft tissue sarcomas: a multicenter trial. J Clin Oncol 14, 2653-2665 (1996). 385. Goethals, A., Pasparakis, M., Hostens, J., Breier, G., Kollias, G., Brouckaert, P. Expression of TNF-R1 exclusively on endothelial cells is sufficient to obtain tumor destruction by TNF. (Article submitted) (2008). 386. Kimura, K. et al. Phase I study of recombinant human tumor necrosis factor. Cancer chemotherapy and pharmacology 20, 223-229 (1987). 387. Creaven, P.J. et al. Phase I clinical trial of recombinant human tumor necrosis factor. Cancer chemotherapy and pharmacology 20, 137-144 (1987). 66 Chapter 1

388. Spriggs, D.R. et al. Recombinant human tumor necrosis factor administered as a 24-hour intravenous infusion. A phase I and pharmacologic study. J Natl Cancer Inst 80, 1039-1044 (1988). 389. Sherman, M.L. et al. Recombinant human tumor necrosis factor administered as a five-day continuous infusion in cancer patients: phase I toxicity and effects on lipid metabolism. J Clin Oncol 6, 344-350 (1988). 390. Creaven, P.J. et al. A phase I clinical trial of recombinant human tumor necrosis factor given daily for five days. Cancer chemotherapy and pharmacology 23, 186-191 (1989). 391. Gamm, H., Lindemann, A., Mertelsmann, R. & Herrmann, F. Phase I trial of recombinant human tumour necrosis factor alpha in patients with advanced malignancy. Eur J Cancer 27, 856-863 (1991). 392. Lenk, H., Tanneberger, S., Muller, U., Ebert, J. & Shiga, T. Phase II clinical trial of high-dose recombinant human tumor necrosis factor. Cancer chemotherapy and pharmacology 24, 391- 392 (1989). 393. Kemeny, N., Childs, B., Larchian, W., Rosado, K. & Kelsen, D. A phase II trial of recombinant tumor necrosis factor in patients with advanced colorectal carcinoma. Cancer 66, 659-663 (1990). 394. Heim, M.E. et al. Tumor necrosis factor in advanced colorectal cancer: a phase II study. A trial of the phase I/II study group of the Association for Medical Oncology of the German Cancer Society. Onkologie 13, 444-447 (1990). 395. Skillings, J. et al. A phase II study of recombinant tumor necrosis factor in renal cell carcinoma: a study of the National Cancer Institute of Canada Clinical Trials Group. J Immunother 11, 67-70 (1992). 396. Lienard, D., Ewalenko, P., Delmotte, J.J., Renard, N. & Lejeune, F.J. High-dose recombinant tumor necrosis factor alpha in combination with interferon gamma and melphalan in isolation perfusion of the limbs for melanoma and sarcoma. J Clin Oncol 10, 52-60 (1992). 397. Lienard, D., Lejeune, F.J. & Ewalenko, P. In transit metastases of malignant melanoma treated by high dose rTNF alpha in combination with interferon-gamma and melphalan in isolation perfusion. World journal of surgery 16, 234-240 (1992). 398. Eggermont, A.M., de Wilt, J.H. & ten Hagen, T.L. Current uses of isolated limb perfusion in the clinic and a model system for new strategies. The lancet oncology 4, 429-437 (2003). 399. Omlor, G. Optimization of isolated hyperthermic limb perfusion. World journal of surgery 17, 134 (1993). 400. Lejeune, F.J., Lienard, D., Matter, M. & Ruegg, C. Efficiency of recombinant human TNF in human cancer therapy. Cancer Immun 6, 6 (2006). 401. Posner, M.C., Lienard, D., Lejeune, F.J., Rosenfelder, D. & Kirkwood, J. Hyperthermic isolated limb perfusion with tumor necrosis factor alone for melanoma. The cancer journal from Scientific American 1, 274-280 (1995). 402. de Wilt, J.H. et al. Tumour necrosis factor alpha increases melphalan concentration in tumour tissue after isolated limb perfusion. British journal of cancer 82, 1000-1003 (2000). 403. van der Veen, A.H. et al. TNF-alpha augments intratumoural concentrations of doxorubicin in TNF-alpha-based isolated limb perfusion in rat sarcoma models and enhances anti-tumour effects. British journal of cancer 82, 973-980 (2000). 404. Folli, S. et al. Tumor-necrosis factor can enhance radio-antibody uptake in human colon carcinoma xenografts by increasing vascular permeability. International journal of cancer 53, 829-836 (1993). 405. Brouckaert, P. et al. Tumor necrosis factor-alpha augmented tumor response in B16BL6 melanoma-bearing mice treated with stealth liposomal doxorubicin (Doxil) correlates with altered Doxil pharmacokinetics. International journal of cancer 109, 442-448 (2004). 406. Jain, R.K. Barriers to drug delivery in solid tumors. Scientific American 271, 58-65 (1994). 407. Renard, N. et al. Early endothelium activation and polymorphonuclear cell invasion precede specific necrosis of human melanoma and sarcoma treated by intravascular high-dose tumour necrosis factor alpha (rTNF alpha). International journal of cancer 57, 656-663 (1994). 408. Manusama, E.R. et al. Assessment of the role of neutrophils on the antitumor effect of TNFalpha in an in vivo isolated limb perfusion model in sarcoma-bearing brown Norway rats. The Journal of surgical research 78, 169-175 (1998). 409. Alexander, H.R., Jr., Bartlett, D.L. & Libutti, S.K. Current status of isolated hepatic perfusion with or without tumor necrosis factor for the treatment of unresectable cancers confined to liver. The oncologist 5, 416-424 (2000). Introduction: Tumor Necrosis Factor 67

410. Alexander, H.R. et al. A phase I-II study of isolated hepatic perfusion using melphalan with or without tumor necrosis factor for patients with ocular melanoma metastatic to liver. Clin Cancer Res 6, 3062-3070 (2000). 411. Alexander, H.R., Jr. et al. Hepatic vascular isolation and perfusion for patients with progressive unresectable liver metastases from colorectal carcinoma refractory to previous systemic and regional chemotherapy. Cancer 95, 730-736 (2002). 412. Pandey, M.K. et al. Gambogic acid, a novel ligand for transferrin receptor, potentiates TNF- induced apoptosis through modulation of the nuclear factor-kappaB signaling pathway. Blood 110, 3517-3525 (2007). 413. Zavrski, I. et al. Molecular and clinical aspects of proteasome inhibition in the treatment of cancer. Recent results in cancer research. Fortschritte der Krebsforschung 176, 165-176 (2007). 414. Nowis, D. et al. TNF potentiates anticancer activity of bortezomib (Velcade) through reduced expression of proteasome subunits and dysregulation of unfolded protein response. International journal of cancer 121, 431-441 (2007). 415. Matschurat, S. et al. Negative regulatory role of PI3-kinase in TNF-induced tumor necrosis. International journal of cancer 107, 30-37 (2003). 416. de Wilt, J.H. et al. Nitric oxide synthase inhibition results in synergistic anti-tumour activity with melphalan and tumour necrosis factor alpha-based isolated limb perfusions. British journal of cancer 83, 1176-1182 (2000). 417. Goethals, A., Brouckaert, P. (unpublished data). (2007). 418. van der Veen, A.H., Eggermont, A.M., Seynhaeve, A.L., van, T. & ten Hagen, T.L. Biodistribution and tumor localization of stealth liposomal tumor necrosis factor-alpha in soft tissue sarcoma bearing rats. International journal of cancer 77, 901-906 (1998). 419. ten Hagen, T.L., Seynhaeve, A.L., van Tiel, S.T., Ruiter, D.J. & Eggermont, A.M. Pegylated liposomal tumor necrosis factor-alpha results in reduced toxicity and synergistic antitumor activity after systemic administration in combination with liposomal doxorubicin (Doxil) in soft tissue sarcoma-bearing rats. International journal of cancer 97, 115-120 (2002). 420. Curnis, F. et al. Enhancement of tumor necrosis factor alpha antitumor immunotherapeutic properties by targeted delivery to aminopeptidase N (CD13). Nature biotechnology 18, 1185- 1190 (2000). 421. Curnis, F., Sacchi, A. & Corti, A. Improving chemotherapeutic drug penetration in tumors by vascular targeting and barrier alteration. The Journal of clinical investigation 110, 475-482 (2002). 422. Sacchi, A. et al. Synergistic antitumor activity of cisplatin, paclitaxel, and gemcitabine with tumor vasculature-targeted tumor necrosis factor-alpha. Clin Cancer Res 12, 175-182 (2006). 423. van Laarhoven, H.W. et al. Effects of the tumor vasculature targeting agent NGR-TNF on the tumor microenvironment in murine lymphomas. Investigational new drugs 24, 27-36 (2006). 424. Santimaria, M. et al. Immunoscintigraphic detection of the ED-B domain of fibronectin, a marker of angiogenesis, in patients with cancer. Clin Cancer Res 9, 571-579 (2003). 425. Borsi, L. et al. Selective targeted delivery of TNFalpha to tumor blood vessels. Blood 102, 4384-4392 (2003). 426. Halin, C. et al. Synergistic therapeutic effects of a tumor targeting antibody fragment, fused to interleukin 12 and to tumor necrosis factor alpha. Cancer research 63, 3202-3210 (2003). 427. Weichselbaum, R.R. et al. Radiation-induced tumour necrosis factor-alpha expression: clinical application of transcriptional and physical targeting of gene therapy. The lancet oncology 3, 665-671 (2002). 428. Mundt, A.J. et al. A Phase I trial of TNFerade biologic in patients with soft tissue sarcoma in the extremities. Clin Cancer Res 10, 5747-5753 (2004). 429. Senzer, N. et al. TNFerade biologic, an adenovector with a radiation-inducible promoter, carrying the human tumor necrosis factor alpha gene: a phase I study in patients with solid tumors. J Clin Oncol 22, 592-601 (2004). 430. McLoughlin, J.M. et al. TNFerade, an adenovector carrying the transgene for human tumor necrosis factor alpha, for patients with advanced solid tumors: surgical experience and long- term follow-up. Annals of surgical oncology 12, 825-830 (2005). 431. Paciotti, G.F. et al. Colloidal gold: a novel nanoparticle vector for tumor directed drug delivery. Drug delivery 11, 169-183 (2004). 432. Visaria, R.K. et al. Enhancement of tumor thermal therapy using gold nanoparticle-assisted tumor necrosis factor-alpha delivery. Molecular cancer therapeutics 5, 1014-1020 (2006). 68 Chapter 1

433. Visaria, R. et al. Nanotherapeutics for enhancing thermal therapy of cancer. Int J Hyperthermia 23, 501-511 (2007).

Chapter 2 Interferon-

Introduction: Interferon- 71

2. INTERFERON-

2.1. IFN- and its receptor

Interferons are a family of proteins that were originally discovered as agents that interfere with viral replication 1. They are classified into type I and type II according to receptor specificity and acid stability. Type I IFNs include IFN-, IFN-, IFN-, and IFN-, all of which are acid-stable and bind to a common receptor known as the IFN- receptor. IFN- is the only type II IFN. It is acid-labile and signals through the IFN- receptor (IFN-R). The major function of type I IFNs is indeed the defense against viral infection, but IFN-, also called immune IFN, has besides its anti-viral activity a more general function in the immune system 2. Initially it was believed that IFN- is produced exclusively by activated NK cells, CD4+ T helper type 1 (Th1) cells, and CD8+ cytotoxic T cells, but now it is clear that many other immune cells can produce IFN-, including NKT cells,  T cells, and professional antigen-presenting cells (APCs) (dendritic cells, macrophages and B cells) 2-6. IFN- production early during infection is mainly controlled by cytokines secreted by APCs. Especially IL-12 and IL-18 appear to play a major role in the stimulation of IFN- production. IL-12 and IL-18 have been shown to synergistically increases IFN- production in dendritic cells, macrophages, NK, Th1 and B cells 7-9. T lymphocytes are the major source of IFN- in the adaptive immune response and secrete IFN- upon triggering of their T cell receptor (TCR). Negative regulators of IFN- production include IL-4, IL-10, TGF-, and glucocorticoids 10. The promoter region of the IFN- gene contains binding sites for many transcription factors, such as NFAT, NF-B, STAT4 and T-bet, suggesting that the regulation of its expression is very complex 2. IFN- gene expression results in a polypeptide of 17kDa (in humans), followed by N-glycosylation, association of two polypeptides into a non- covalent homodimer, which is the biological active form of IFN-, and secretion 11-13. N-glycosylation of the protein is not required for its biological activity but was shown to influence its clearance from circulation 14-16.

The IFN-R is comprised of two subunits, called IFN-R1, which is responsible for ligand binding, and IFN-R2 17, 18. Both receptor chains contain sequences in their carboxyterminal intracellular regions for constitutive association with members of the Janus kinase (JAK) family of tyrosine kinases. IFN-R1 specifically associates with JAK-1, while IFN-R2 is constitutively associated with JAK-2 19-21. The IFN-R is constitutively expressed on most cells 22. IFN-R1 is usually in surplus, but the expression level of IFN-R2 appears to be tightly regulated according to the state of cellular differentiation and activation. For example, CD4+ Th1 cells express only low levels of IFN-R2n and are insensitive to the growth-inhibitory effects of 72 Chapter 2

IFN-, while CD4+ Th2 cells express high levels of IFN-R2 and stimulation of these cells with IFN- inhibits their proliferation and can induce apoptosis 10, 23-25.

2.2. IFN-R signaling

Signal transduction starts with the interaction of the IFN- homodimer with two IFN- R1 chains, thereby inducing dimerization of the receptor. This brings inactive JAKs in close proximity with one another, resulting in activation of JAK-2 through autophosphorylation (Fig.2.1.). Activated JAK-2 phosphorylates and activates JAK-1, which then phosphorylates a specific tyrosine residue near the carboxyterminus of the IFN-R1 chain 26, 27. This phosphorylated tyrosine residue serves as a docking site for the Src homology 2 (SH2) domain of the transcription factor signal transducer and activator of transcription 1 (STAT1) 28. After association with the receptor, STAT1 is tyrosine phosphorylated by JAKs, which leads to rapid dissociation of the receptor and SH2 domain-dependent dimerization of STAT1 29, 30. At some point during its activation, STAT1 becomes phosphorylated on a specific serine residue, which is required for maximal transcriptional activity 31. Recently it was shown that IFN-- induced serine phosphorylation of STAT1 is regulated by PI3K, mammalian target of rapamycin (mTOR) and PKC activity, and occurs in a macromolecular complex that contains STAT1, mTOR and PKC 32-34. PKC has also been shown to mediate serine phosphorylation of STAT1 in response to type I IFNs 35. Upon activation, the STAT1 homodimer is transported to the nucleus where it can bind to specific DNA sequences, known as gamma-activated sites (GASs), to initiate or suppress gene transcription 36. This results within 15-30 minutes after stimulation with IFN- in a first wave of transcription, called primary response. Many of the induced genes are in fact transcription factors, most importantly IFN regulatory factor-1 (IRF-1). These transcription factors further drive the regulation of the next wave of transcription, called secondary response 10. The total number of IFN--regulated genes is estimated to be around 500 37.

Important genes that are upregulated during the primary response include genes related to antigen presentation (CIITA, TAP-1/2, LMP-2/7/10), cell adhesion (VLA-4, ICAM-1, VCAM-1), and negative regulation of IFN signaling (suppressor of cytokine signaling 1 (SOCS1) and IRF-8). IRF-1 binds to a DNA sequence called IFN- stimulated response element (ISRE) and upregulates the expression of iNOS, IFN-, caspase-1, STAT1 and IRF-2 during the secondary response 37. IRF-2 and IRF-8 also bind to ISREs, but act as transcriptional repressors, thereby antagonizing IRF-1 38-40. IRF-1 plays an important role in immune development, which is highlighted by the phenotype of IRF-1 deficient mice. These mice have a reduced number of mature CD8+ T cells due to impaired peptide loading of MHC class I, which is required for Introduction: Interferon- 73

the maturation of these cells in the thymus 41, 42. IRF-1 knockout mice appear to have normal thymic maturation of CD4+ T cells but display increased numbers of memory or effector CD4+ T cells 43. Moreover, these mice fail to develop a Th1-type immune response and instead produce a Th2-type immune response 44, 45. Since IRF-1 controls the expression of IL-15, a cytokine that is crucial for the maturation of NK cells in the bone marrow, IRF-1 knockout mice also lack mature NK and NK1.1+ T cells 46-48.

Fig. 2.1. IFN- signaling pathways. Negative regulation of IFN- signaling is indicated with red arrows. P: phosphorylation, encircled P: phosphotyrosine or phosphoserine, Y: tyrosine residue, S: serine residue. Adapted from Ramana et al., 2002 and Schroder et al., 2004 10, 49.

SOCS proteins negatively regulate cytokine signaling by binding via an SH2 domain to phosphorylated tyrosine residues in cytokine receptors or associated kinases, such as JAKs. Besides an SH2 domain, SOCS proteins also contain a second domain, named SOCS-box, by which they can interact with the elongins B and C and cullin-2 to form an E3 ubiquitin ligase complex. So, binding of SOCS proteins may lead to ubiquitination of associated signaling proteins, which marks these proteins for proteasomal degradation 50-52. IFN- has been shown to upregulate the expression of SOCS1 and SOCS3 53. SOCS1 inhibits IFN- signaling by binding to a 74 Chapter 2

phosphorylated tyrosine residue in the catalytic site of JAK-2, which inhibits its kinase activity 54. Expression of SOCS3 has also been shown to block activation of STAT1 in response to IFN-, although it is less clear how SOCS3 mediates this inhibitory effect 55. It was suggested that SOCS3 inhibits activation of JAKs by direct interaction with cytokine receptors 56-58. Other negative regulators of IFN- signaling are SH2 domain-containing tyrosine phosphatase-2 (SHP-2) and protein inhibitor of activated STAT (PIAS). SHP-2 has been shown to interact with JAK-1 and JAK-2, which results in inactivation of these JAKs 59, 60. PIAS proteins on the other hand can associate with and inhibit activated STAT1 homodimers. PIAS1 mediates this inhibitory effect by blocking the DNA binding activity of STAT1, while PIASy functions as a transcriptional corepressor of STAT1 without blocking its DNA binding activity 61, 62.

Besides STAT1 homodimers, IFN- signaling has also been shown to produce STAT1:STAT2 heterodimers and STAT1:STAT2:IRF-9 and STAT1:STAT1:IRF-9 heterotrimers, which allows further modulation of gene expression downstream of the IFN-R 63, 64. It has now become clear that IFN- can, in addition to the well-known JAK-STAT pathway, also activate additional signaling pathways, which can lead to STAT1-independent gene transcription (Fig.2.1.) 49. It has been shown that JAK-2 associates with and activates the MAPK proline-rich tyrosine kinase 2 (Pyk2) and that Pyk2 then mediates further signaling to ERK and serine phosphorylation of STAT1 65. JAK-2 and Pyk2 have also been shown to associate with the Src family member Fyn 66. IFN- can furthermore activate the adaptor proteins c-Cbl, CrkL, CrkII and Vav, and the G-protein-linked signaling molecules C3G and Ras GTPase-activating protein-1 (Rap-1) 49, 67-69.

Introduction: Interferon- 75

2.3. IFN- in immunity

The IFN system regulates the host defense against viral infection. While both types of IFNs trigger direct cellular defense mechanisms against viral infection, IFN- also coordinates the innate and adaptive immune responses that will establish a longer- lasting protection against viral infection. The immunoregulatory functions of IFN- include the regulation of antigen presentation, Th1- versus Th2-type immune response, leukocyte trafficking, and phagocyte activity.

The direct antiviral activities of IFN- are mainly mediated by transcriptional induction of three genes, namely double-stranded RNA (dsRNA) activated protein kinase (PKR), 2’-5’ oligoadenylate synthetase (2-5A synthetase), and dsRNA specific adenosine deaminase (dsRAD) 37. PKR is activated by binding to dsRNA structures. Both RNA and DNA viruses produce RNA intermediates that can activate PKR. Once activated, PKR phosphorylates the -subunit of eukaryotic initiation factor 2 (eIF2), which inhibits viral and cellular protein synthesis 70, 71. PKR has also been implicated in the activation of NF-B and the splicing of TNF mRNA 72, 73. 2-5A synthetase is also stimulated by dsRNA but converts ATP into 2’-5’-linked adenylates. These adenylates in turn activate RNaseL, which degrades single-stranded RNAs, thereby inhibiting protein synthesis and viral growth 74-76. Finally, dsRAD catalyzes the deamination of adenosine to inosine, which produces edited mRNAs whose translation does not lead to a functional protein 37, 77-79.

IFN- increases antigen presentation by both MHC class I and class II. Upregulation of antigen presentation by MHC class I is important for host responses to intracellular pathogens as it increases the potential for cytotoxic CD8+ T cell recognition and thus promotes cell-mediated immunity. IFN- induces the expression of the proteasome subunits LMP-2, LMP-7 and MECL-1, which replace the constitutive subunits 1, 2, and 5 of the proteasome to form the immunoproteasome 80. Peptide fragments generated by this immunoproteasome are transported by the transporter associated with antigen processing (TAP) to the ER lumen, where they are loaded on the MHC class I complex which is then brought to the cell surface. IFN- upregulates the expression of both TAP-1 and TAP-2 subunits of the TAP transporter as well as the MHC class I heavy chain and 2-microglobulin subunits of the MHC class I complex. Other MHC class I pathway components upregulated by IFN- include tapasin and the proteasome activator PA28 (see Schroder et al., 2004 and references therein 10). Unlike type I IFNs, IFN- can efficiently upregulate the MHC class II antigen presentation pathway and thus promote the activation of specific CD4+ T cells. MHC class II molecules present peptides derived from extracellular pathogens or pathogens that are living in macrophage vesicles. Internalized proteins and pathogens are degraded to peptides by lysosomal cathepsins. IFN- has been shown to upregulate 76 Chapter 2

the expression of the cathepsins B, D, H and L 81-83. Vesicles derived from the ER that contain MHC class II molecules stabilized with the invariant chain protein Ii fuse with endolysosomal vesicles that contain the degraded proteins, whereafter the Ii chain is degraded, peptides are loaded on the MHC class II complex aided by the protein DM, and the complex is brought to the cell surface. The expression of all subunits of the MHC class II complex, the Ii chain and DM are upregulated by the class II transactivator (CIITA), which is induced by IFN- 37, 84. CIITA has no DNA binding activity, but regulates the expression of class II genes by binding with the transcription factor RFX 85. CIITA was shown to be the limiting factor for the class II gene expression 86.

As mentioned above, IFN- is mainly produced by NK, NKT, CD8+ T and Th1 cells. These are effector cells of the cell-mediated or Th1-type immunity, and the IFN- produced by these cells will further skew the immune response towards a Th1 phenotype. The phenotype adopted by naïve CD4+ T cells during T cell activation is strongly influenced by the cytokines present at the time of TCR triggering. IFN- and IL-12 direct differentiation to Th1, while IL-4 directs differentiation to Th2. IL-12 is secreted by activated dendritic cells, neutrophils and macrophages, and stimulates IFN- production by NK cells and naïve CD4+ T cells 87-89. IFN- on the other hand stimulates IL-12 production by macrophages and neutrophils, thus creating a positive feedback loop 90, 91. Moreover, IFN- inhibits the proliferation of Th2 cells, and inhibits the secretion of IL-4 by these cells, which further polarizes the immune response towards a Th1 phenotype 23, 92. Consistent with its role in promoting cell-mediated immunity, IFN- stimulates immunoglobulin isotype switching in B cells to IgG2a and IgG3, which together with the upregulation of FcRI on macrophages and other cells increases the phagocytosis of antibody-coated pathogens and antibody-dependent cell-mediated cytotoxicity (ADCC) 93-95. IFN- also antagonizes the IL-4-induced isotype switching to IgG1 and IgE and upregulation of FcRII on B cells, basophils and mast cells 6, 37, 96, 97. The importance of IFN- in regulating the immune response is illustrated by the fact that some chronic allergic diseases such as asthma have been correlated with reduced IFN- production 98. Allergic diseases are closely related to Th2-type immune responses, which are characterized by high levels of IL-4, IL-5, IL-9 and IL-13, IgE isotype switching, and the recruitment and activation of effector cells such as eosinophils and mast cells. It has been suggested that the induction of Th1 responses, for example with IFN- or IL-12, could protect against Th2-related allergic diseases (Fig.2.2.) 2.

Introduction: Interferon- 77

Fig. 2.2. IFN- in the control of Th immune and allergic responses. IFN- may counteract several steps of allergic responses, by enhancing IL-12 production by APCs, inducing Th1 differentiation, suppressing the proliferation and function of Th2 cells, blocking IgE isotype switching in B cells, and inducing eosinophils apoptosis. From Teixeira et al., 2005 2.

IFN- orchestrates the recruitment of monocytes and lymphocytes to sites of inflammation by inducing the expression of specific chemokines and adhesion molecules. IFN- upregulates the expression of the adhesion molecules ICAM-1 and VCAM-1 on endothelial cells, and induces the expression of chemokines (e.g. CXCL9, CXCL10, CCL2, CCL3, CCL4 and CCL5) in macrophages, lymphocytes, endothelial cells and fibroblasts 37, 99-104. Many of these IFN--induced chemokines, which are secreted by various cell types at the site of inflammation, are captured and presented on the luminal surface of endothelial cells by association with glycosaminoglycans (GAGs), such as heparan sulphate proteoglycans, and Duffy antigen/receptor for chemokines (DARC) 105, 106.

Another important function of IFN- is the stimulation of microbicidal activities of macrophages and neutrophils. Stimulation of these cells with IFN- increases their phagocytic activity 10, 107. Phagocytes primarily kill ingested pathogens by producing ROS and NO in a process known as oxidative (or respiratory) burst. IFN- upregulates the expression of several subunits of the NADPH oxidase, which is the enzyme responsible for the production of superoxide during oxidative burst 108-110. The NO production in phagocytes is mediated by iNOS. IFN- enhances NO production by upregulating the expression of argininosuccinate synthetase, which produces the L- arginine substrate of iNOS, GTP-cyclohydrolase I, which supplies the tetrahydrobiopterin cofactor required for NO production, and the iNOS enzyme itself 111-114. IFN- furthermore upregulates the expression of FcRI and the complement 78 Chapter 2

receptor CR3 on macrophages and neutrophils, thereby promoting the phagocytosis and killing of antibody- or complement-coated pathogens 93, 115-117.

2.4. Synergy with TNF

Since both TNF and IFN- are produced simultaneously at sites of inflammation, potentially by the same cells, it is not really surprising that these two cytokines can influence each other actions. In fact, same actions only become apparent when both cytokines are present. For example, Williamson and colleagues observed already in 1983 that some tumor cell lines were insensitive to the cytotoxic effects of TNF in vitro, but when they treated these cells with a combination of TNF and IFN-, a pronounced cytotoxic effect was observed 118. During the next years, a synergistic cytotoxic or anti-proliferative effect of TNF and IFN- was found for many cell lines, including many tumor cell lines such as the murine melanoma line B16Bl6, and human breast, cervix, colon, and pancreas carcinoma cell lines, but also non- transformed cells such as MEFs and HUVECs 119-123. As IFN- was found to upregulate the expression of TNF-R2 on several tumor cell lines, it was originally thought that this would at least in part explain the synergistic effect of IFN- and TNF 124-126. However, this hypothesis was soon abandoned since IFN- also upregulated TNF-R2 expression in other cell lines without sensitizing for TNF cytotoxicity, and the synergistic cytotoxic effect of TNF and IFN-was not abrogated by IFN-, even though the latter cytokine inhibited the IFN--induced upregulation of TNF-R2 127-129. Other candidate proteins upregulated by IFN- that could explain the synergistic cytotoxic effect of TNF and IFN- are the cathepsins B and D and several caspases 81, 83, 130-132. IFN- upregulates the expression of caspase-1 in a IRF-1-dependent manner and caspase-1 was shown to be required for IFN--induced apoptosis of HeLa and some other cell lines 130, 133. The cathepsins B and D have been implicated in TNF- induced cell death of hepatocytes and some other cells 134-136. It was shown that cathepsin D is involved in IFN- or TNF-induced cell death in HeLa cells 83. Recently, Li and colleagues demonstrated that IFN- can trigger the release of cathepsin B from lysosomes in cultured endothelial cells and that cathepsin B contributes to TNF plus IFN--induced cell death in these cells 137. IFN- was also shown to increase the expression of procaspase-8 in cultured endothelial cells and sensitize these cells to Fas-mediated apoptosis 131. IFN- also upregulates the expression of the pro-apoptotic proteins Bak, XIAP associated factor-1 (XAF-1), death-associated proteins (DAPs), and DAP-kinase, while it downregulates the expression of the anti-apoptotic proteins Bcl-2 and Hsp27 138-143. XAF-1 can associate with XIAP to block its caspase-inhibiting and survivin- stabilizing activity 144. The chaperone protein Hsp27 can protect cells from apoptosis by interacting with the cytochrome c released from the mitochondria, thereby Introduction: Interferon- 79

preventing the apoptosome-mediated activation of procaspase-9 145, 146. The pro- apoptotic activities of DAPs and DAP-kinase are still largely unknown. Inhibition of DAP3 expression by antisense RNA was shown to protect cells from IFN- or TNF- induced apoptosis 139, 147. DAP3 has also been linked to anoikis. Cells attached to the extracellular matrix receive survival signals via integrins, in essence by activation of the PI3K/Akt pathway. Akt can phosphorylate and inactivate DAP3. Detachment of cells would lead to activation of DAP3, followed by association of DAP3 with FADD and activation of caspase-8 148, 149. DAP-kinase is Ca2+/calmodulin-dependent, DD- containing serine/threonine kinase that was demonstrated to participate in apoptosis induced by various stimuli 150-153. Recently it was shown that DAP-kinase, which is localized to the actin cytoskeleton, phosphorylates MLC, induces cytoskeletal reorganization, and inhibits integrin-mediated cell adhesion and signal transduction 154-156.

Many genes that are inducible by IFN are also inducible by TNF, and often the induction is synergistic. Among many others this is the case for the monocyte/lymphocyte chemokines CCL2, CCL5 and CXCL10, the adhesion molecules ICAM-1 and VCAM-1, and iNOS 103, 157-161. Most of these genes contain an ISRE or GAS sequence as well as a B-site in their promoter to allow the binding of both NF-B and an IFN--induced transcription factor 37. Moreover, it was shown that NF-B can physically interact with STAT1, IRF-1 and IRF-2 162-164. Suk and colleagues demonstrated that transfection of a phosphorylation-defective STAT1 inhibited TNF plus IFN--induced apoptosis in a cervical cancer cell line, whereas transfection of IRF-1 sensitized these cells to TNF-induced apoptosis, suggesting that IRF-1 is required for synergistic cytotoxic effect of TNF and IFN- 165. IFN- can also enhance the TNF-induced activation of NF-B by stimulating degradation of IB, possibly in a PKR-dependent manner 72, 166, 167. TNF on the other hand may enhance the activation of STAT1 through phosphorylation of JAK-2 168. More recently, it was shown that inactive STAT1 can interact with the TNF- R1/TRADD complex and inhibit the TNF-induced activation of NF-B. Stimulation of cells with IFN- would then lead to activation and nuclear translocation of STAT1 and thus depletion of STAT1 from the cytoplasm, allowing for optimal NF-B activation by TNF 169, 170. Furthermore, TRADD and STAT1 were shown to form a nuclear complex that may regulate the transcriptional activity of STAT1 171. So, TNF and IFN- appear to reciprocally control each others’ signaling pathways at many levels.

The in vitro synergistic effects of TNF and IFN- on tumor cell lines prompted researchers to use this combination in experimental tumor models. In 1986, Balkwill and colleagues were the first to report a synergistic antitumor effect of hTNF and IFN- on human tumor xenografts in nude mice 172. However, they treated these nude 80 Chapter 2

mice with human IFN-, which has no effect on murine cells, suggesting that this synergistic antitumor effect was due to direct cytotoxic effects on the tumor cells. In the same year, Brouckaert and colleagues reported the synergistic antitumor activity of hTNF and murine IFN- in a syngeneic tumor model, namely murine B16Bl6 tumors grown subcutaneously in C57BL/6 mice 120. While hTNF alone had a static effect on tumor growth in these mice, the combination of hTNF and IFN- induced complete necrosis and regression of these tumors. In the years after, many reports have been published demonstrating the synergistic antitumor activity of TNF and IFN- in various experimental tumor models 173-178. These observations in mice has led to the addition of IFN- to ILP with TNF and melphalan for the treatment of human patients with unresectable sarcomas or melanomas of the limbs. Although the differences in response rates are small and statistically non-significant, clinical studies using ILP with the triple combination of TNF, IFN- and melphalan appear to result in consistently higher response rates than those without IFN-, suggesting that the synergistic effect of TNF and IFN- is also present in this system (Fig.2.3.) 179.

Fig. 2.3. Complete response rates obtained after ILP in melanoma. Complete response rates obtained of after ILP with TNF, IFN- and melphalan are indicated in red, with TNF and melphalan in orange, and with melphalan alone in blue. The data are from 11 clinical studies (phase II and III) and include more than 200 patients. From Lejeune et al., 2006 179.

The mechanism underlying the in vivo synergistic antitumor effect of TNF and IFN- is still largely unknown. However, as was the case for TNF monotherapy, histological analysis of tumors treated with TNF plus IFN- revealed a selective destruction of the tumor vasculature, suggesting that endothelial cells are also the target cells of IFN- 178, 180. This was confirmed by the work of Hostens and colleagues, who created a Introduction: Interferon- 81

transgenic mouse strain that expresses a dominant negative IFN-R1 chain only in endothelial cells (controlled by a Flk-1-promoter), thereby rendering these cells insensitive to IFN-. The synergistic antitumor effect of TNF and IFN- was complexly abrogated in these mice, although all other cells, including leukocytes and tumor cells, still expressed a functional IFN-R, demonstrating that a functional IFN- R on endothelial cells is essential for the synergistic antitumor activity of TNF and IFN- 181. They furthermore showed, using IRF-1 knockout mice, that the transcription factor IRF-1 is to some extent important for the antitumor effects of TNF and IFN-. In fact, the synergism of TNF and IFN- is still present in these mice, but neither treatment with mTNF alone or with hTNF plus IFN- resulted in complete regression of the B16Bl6 tumor. This defect could be partly restored by reconstitution with wild-type splenocytes, suggesting that certain immune cells could play a role in the antitumor effect of TNF and IFN- 182. Rüegg and co-workers implied the integrin 183 V3 in the vascular-disrupting effect of TNF and IFN- . They found that treatment of HUVECs with TNF and IFN- suppressed the activation of this integrin, leading to decreased endothelial cell adhesion and survival. They furthermore observed detachment and apoptosis of v3-positive endothelial cells of the tumor vasculature in melanoma patients treated with TNF and IFN-.

82 Chapter 2

2.5. References

1. Isaacs, A. & Lindenmann, J. Virus interference. I. The interferon. Proceedings of the Royal Society of London. Series B, Containing papers of a Biological character 147, 258-267 (1957). 2. Teixeira, L.K., Fonseca, B.P., Barboza, B.A. & Viola, J.P. The role of interferon-gamma on immune and allergic responses. Memorias do Instituto Oswaldo Cruz 100 Suppl 1, 137-144 (2005). 3. Farrar, M.A. & Schreiber, R.D. The molecular cell biology of interferon-gamma and its receptor. Annual review of immunology 11, 571-611 (1993). 4. Hammond, K.J. et al. NKT cells are phenotypically and functionally diverse. European journal of immunology 29, 3768-3781 (1999). 5. Frucht, D.M. et al. IFN-gamma production by antigen-presenting cells: mechanisms emerge. Trends in immunology 22, 556-560 (2001). 6. Janeway, C.A., Travers, P., Walport, M., Shlomchik, M. Immunobiology: The immune system in health and disease, Edn. 5. (Garland Publishing, New York; 2001). 7. Yoshimoto, T. et al. IL-12 up-regulates IL-18 receptor expression on T cells, Th1 cells, and B cells: synergism with IL-18 for IFN-gamma production. J Immunol 161, 3400-3407 (1998). 8. Munder, M., Mallo, M., Eichmann, K. & Modolell, M. Murine macrophages secrete interferon gamma upon combined stimulation with interleukin (IL)-12 and IL-18: A novel pathway of autocrine macrophage activation. The Journal of experimental medicine 187, 2103-2108 (1998). 9. Lauwerys, B.R., Renauld, J.C. & Houssiau, F.A. Synergistic proliferation and activation of natural killer cells by interleukin 12 and interleukin 18. Cytokine 11, 822-830 (1999). 10. Schroder, K., Hertzog, P.J., Ravasi, T. & Hume, D.A. Interferon-gamma: an overview of signals, mechanisms and functions. Journal of leukocyte biology 75, 163-189 (2004). 11. Scahill, S.J., Devos, R., Van der Heyden, J. & Fiers, W. Expression and characterization of the product of a human immune interferon cDNA gene in Chinese hamster ovary cells. Proceedings of the National Academy of Sciences of the United States of America 80, 4654- 4658 (1983). 12. Rinderknecht, E., O'Connor, B.H. & Rodriguez, H. Natural human interferon-gamma. Complete amino acid sequence and determination of sites of glycosylation. The Journal of biological chemistry 259, 6790-6797 (1984). 13. Nagata, K., Izumi, T., Kitagawa, T. & Yoshida, N. Oligomeric structure of recombinant human and murine immune interferons by means of sedimentation equilibrium. Journal of interferon research 7, 313-320 (1987). 14. Kelker, H.C., Yip, Y.K., Anderson, P. & Vilcek, J. Effects of glycosidase treatment on the physicochemical properties and biological activity of human interferon-gamma. The Journal of biological chemistry 258, 8010-8013 (1983). 15. Arakawa, T., Hsu, Y.R., Chang, D., Stebbing, N. & Altrock, B. Structure and activity of glycosylated human interferon-gamma. Journal of interferon research 6, 687-695 (1986). 16. Sareneva, T., Cantell, K., Pyhala, L., Pirhonen, J. & Julkunen, I. Effect of carbohydrates on the pharmacokinetics of human interferon-gamma. Journal of interferon research 13, 267-269 (1993). 17. Soh, J. et al. Identification and sequence of an accessory factor required for activation of the human interferon gamma receptor. Cell 76, 793-802 (1994). 18. Marsters, S.A., Pennica, D., Bach, E., Schreiber, R.D. & Ashkenazi, A. Interferon gamma signals via a high-affinity multisubunit receptor complex that contains two types of polypeptide chain. Proceedings of the National Academy of Sciences of the United States of America 92, 5401-5405 (1995). 19. Kaplan, D.H., Greenlund, A.C., Tanner, J.W., Shaw, A.S. & Schreiber, R.D. Identification of an interferon-gamma receptor alpha chain sequence required for JAK-1 binding. The Journal of biological chemistry 271, 9-12 (1996). 20. Kotenko, S.V. et al. Interaction between the components of the interferon gamma receptor complex. The Journal of biological chemistry 270, 20915-20921 (1995). 21. Sakatsume, M. et al. The Jak kinases differentially associate with the alpha and beta (accessory factor) chains of the interferon gamma receptor to form a functional receptor unit capable of activating STAT transcription factors. The Journal of biological chemistry 270, 17528-17534 (1995). Introduction: Interferon- 83

22. Valente, G. et al. Distribution of interferon-gamma receptor in human tissues. European journal of immunology 22, 2403-2412 (1992). 23. Gajewski, T.F. & Fitch, F.W. Anti-proliferative effect of IFN-gamma in immune regulation. I. IFN-gamma inhibits the proliferation of Th2 but not Th1 murine helper T lymphocyte clones. J Immunol 140, 4245-4252 (1988). 24. Pernis, A. et al. Lack of interferon gamma receptor beta chain and the prevention of interferon gamma signaling in TH1 cells. Science (New York, N.Y 269, 245-247 (1995). 25. Bernabei, P. et al. Interferon-gamma receptor 2 expression as the deciding factor in human T, B, and myeloid cell proliferation or death. Journal of leukocyte biology 70, 950-960 (2001). 26. Briscoe, J. et al. Kinase-negative mutants of JAK1 can sustain interferon-gamma-inducible gene expression but not an antiviral state. The EMBO journal 15, 799-809 (1996). 27. Greenlund, A.C., Farrar, M.A., Viviano, B.L. & Schreiber, R.D. Ligand-induced IFN gamma receptor tyrosine phosphorylation couples the receptor to its signal transduction system (p91). The EMBO journal 13, 1591-1600 (1994). 28. Heim, M.H., Kerr, I.M., Stark, G.R. & Darnell, J.E., Jr. Contribution of STAT SH2 groups to specific interferon signaling by the Jak-STAT pathway. Science (New York, N.Y 267, 1347- 1349 (1995). 29. Greenlund, A.C. et al. Stat recruitment by tyrosine-phosphorylated cytokine receptors: an ordered reversible affinity-driven process. Immunity 2, 677-687 (1995). 30. Bach, E.A., Aguet, M. & Schreiber, R.D. The IFN gamma receptor: a paradigm for cytokine receptor signaling. Annual review of immunology 15, 563-591 (1997). 31. Wen, Z., Zhong, Z. & Darnell, J.E., Jr. Maximal activation of transcription by Stat1 and Stat3 requires both tyrosine and serine phosphorylation. Cell 82, 241-250 (1995). 32. Nguyen, H., Ramana, C.V., Bayes, J. & Stark, G.R. Roles of phosphatidylinositol 3-kinase in interferon-gamma-dependent phosphorylation of STAT1 on serine 727 and activation of gene expression. The Journal of biological chemistry 276, 33361-33368 (2001). 33. Kristof, A.S., Marks-Konczalik, J., Billings, E. & Moss, J. Stimulation of signal transducer and activator of transcription-1 (STAT1)-dependent gene transcription by lipopolysaccharide and interferon-gamma is regulated by mammalian target of rapamycin. The Journal of biological chemistry 278, 33637-33644 (2003). 34. Deb, D.K. et al. Activation of protein kinase C delta by IFN-gamma. J Immunol 171, 267-273 (2003). 35. Uddin, S. et al. Protein kinase C-delta (PKC-delta ) is activated by type I interferons and mediates phosphorylation of Stat1 on serine 727. The Journal of biological chemistry 277, 14408-14416 (2002). 36. Decker, T., Kovarik, P. & Meinke, A. GAS elements: a few nucleotides with a major impact on cytokine-induced gene expression. J Interferon Cytokine Res 17, 121-134 (1997). 37. Boehm, U., Klamp, T., Groot, M. & Howard, J.C. Cellular responses to interferon-gamma. Annual review of immunology 15, 749-795 (1997). 38. Harada, H. et al. Structurally similar but functionally distinct factors, IRF-1 and IRF-2, bind to the same regulatory elements of IFN and IFN-inducible genes. Cell 58, 729-739 (1989). 39. Nelson, N., Marks, M.S., Driggers, P.H. & Ozato, K. Interferon consensus sequence-binding protein, a member of the interferon regulatory factor family, suppresses interferon-induced gene transcription. Molecular and cellular biology 13, 588-599 (1993). 40. Mamane, Y. et al. Interferon regulatory factors: the next generation. Gene 237, 1-14 (1999). 41. Matsuyama, T. et al. Targeted disruption of IRF-1 or IRF-2 results in abnormal type I IFN gene induction and aberrant lymphocyte development. Cell 75, 83-97 (1993). 42. White, L.C. et al. Regulation of LMP2 and TAP1 genes by IRF-1 explains the paucity of CD8+ T cells in IRF-1-/- mice. Immunity 5, 365-376 (1996). 43. McElligott, D.L. et al. CD4+ T cells from IRF-1-deficient mice exhibit altered patterns of cytokine expression and cell subset homeostasis. J Immunol 159, 4180-4186 (1997). 44. Taki, S. et al. Multistage regulation of Th1-type immune responses by the transcription factor IRF-1. Immunity 6, 673-679 (1997). 45. Lohoff, M. et al. Interferon regulatory factor-1 is required for a T helper 1 immune response in vivo. Immunity 6, 681-689 (1997). 46. Ogasawara, K. et al. Requirement for IRF-1 in the microenvironment supporting development of natural killer cells. Nature 391, 700-703 (1998). 47. Ohteki, T. et al. The transcription factor interferon regulatory factor 1 (IRF-1) is important during the maturation of natural killer 1.1+ T cell receptor-alpha/beta+ (NK1+ T) cells, natural 84 Chapter 2

killer cells, and intestinal intraepithelial T cells. The Journal of experimental medicine 187, 967-972 (1998). 48. Taniguchi, T., Ogasawara, K., Takaoka, A. & Tanaka, N. IRF family of transcription factors as regulators of host defense. Annual review of immunology 19, 623-655 (2001). 49. Ramana, C.V., Gil, M.P., Schreiber, R.D. & Stark, G.R. Stat1-dependent and -independent pathways in IFN-gamma-dependent signaling. Trends in immunology 23, 96-101 (2002). 50. Krebs, D.L. & Hilton, D.J. SOCS proteins: negative regulators of cytokine signaling. Stem cells (Dayton, Ohio) 19, 378-387 (2001). 51. Zhang, J.G. et al. The conserved SOCS box motif in suppressors of cytokine signaling binds to elongins B and C and may couple bound proteins to proteasomal degradation. Proceedings of the National Academy of Sciences of the United States of America 96, 2071-2076 (1999). 52. Kamura, T. et al. The Elongin BC complex interacts with the conserved SOCS-box motif present in members of the SOCS, ras, WD-40 repeat, and ankyrin repeat families. Genes & development 12, 3872-3881 (1998). 53. Sakamoto, H. et al. A Janus kinase inhibitor, JAB, is an interferon-gamma-inducible gene and confers resistance to interferons. Blood 92, 1668-1676 (1998). 54. Yasukawa, H. et al. The JAK-binding protein JAB inhibits Janus tyrosine kinase activity through binding in the activation loop. The EMBO journal 18, 1309-1320 (1999). 55. Song, M.M. & Shuai, K. The suppressor of cytokine signaling (SOCS) 1 and SOCS3 but not SOCS2 proteins inhibit interferon-mediated antiviral and antiproliferative activities. The Journal of biological chemistry 273, 35056-35062 (1998). 56. Hansen, J.A., Lindberg, K., Hilton, D.J., Nielsen, J.H. & Billestrup, N. Mechanism of inhibition of growth hormone receptor signaling by suppressor of cytokine signaling proteins. Molecular endocrinology (Baltimore, Md 13, 1832-1843 (1999). 57. Nicholson, S.E. et al. Suppressor of cytokine signaling-3 preferentially binds to the SHP-2- binding site on the shared cytokine receptor subunit gp130. Proceedings of the National Academy of Sciences of the United States of America 97, 6493-6498 (2000). 58. Dey, B.R., Furlanetto, R.W. & Nissley, P. Suppressor of cytokine signaling (SOCS)-3 protein interacts with the insulin-like growth factor-I receptor. Biochemical and biophysical research communications 278, 38-43 (2000). 59. Yin, T., Shen, R., Feng, G.S. & Yang, Y.C. Molecular characterization of specific interactions between SHP-2 phosphatase and JAK tyrosine kinases. The Journal of biological chemistry 272, 1032-1037 (1997). 60. You, M., Yu, D.H. & Feng, G.S. Shp-2 tyrosine phosphatase functions as a negative regulator of the interferon-stimulated Jak/STAT pathway. Molecular and cellular biology 19, 2416- 2424 (1999). 61. Liu, B. et al. Inhibition of Stat1-mediated gene activation by PIAS1. Proceedings of the National Academy of Sciences of the United States of America 95, 10626-10631 (1998). 62. Liu, B., Gross, M., ten Hoeve, J. & Shuai, K. A transcriptional corepressor of Stat1 with an essential LXXLL signature motif. Proceedings of the National Academy of Sciences of the United States of America 98, 3203-3207 (2001). 63. Bluyssen, A.R., Durbin, J.E. & Levy, D.E. ISGF3 gamma p48, a specificity switch for interferon activated transcription factors. Cytokine & growth factor reviews 7, 11-17 (1996). 64. Matsumoto, M. et al. Activation of the transcription factor ISGF3 by interferon-gamma. Biological chemistry 380, 699-703 (1999). 65. Takaoka, A. et al. Protein tyrosine kinase Pyk2 mediates the Jak-dependent activation of MAPK and Stat1 in IFN-gamma, but not IFN-alpha, signaling. The EMBO journal 18, 2480- 2488 (1999). 66. Uddin, S. et al. Interaction of p59fyn with interferon-activated Jak kinases. Biochemical and biophysical research communications 235, 83-88 (1997). 67. Platanias, L.C. & Fish, E.N. Signaling pathways activated by interferons. Experimental hematology 27, 1583-1592 (1999). 68. English, B.K., Orlicek, S.L., Mei, Z. & Meals, E.A. Bacterial LPS and IFN-gamma trigger the tyrosine phosphorylation of vav in macrophages: evidence for involvement of the hck tyrosine kinase. Journal of leukocyte biology 62, 859-864 (1997). 69. Alsayed, Y. et al. IFN-gamma activates the C3G/Rap1 signaling pathway. J Immunol 164, 1800-1806 (2000). 70. Meurs, E. et al. Molecular cloning and characterization of the human double-stranded RNA- activated protein kinase induced by interferon. Cell 62, 379-390 (1990). Introduction: Interferon- 85

71. Lee, S.B. et al. The interferon-induced double-stranded RNA-activated human p68 protein kinase potently inhibits protein synthesis in cultured cells. Virology 192, 380-385 (1993). 72. Zamanian-Daryoush, M., Mogensen, T.H., DiDonato, J.A. & Williams, B.R. NF-kappaB activation by double-stranded-RNA-activated protein kinase (PKR) is mediated through NF- kappaB-inducing kinase and IkappaB kinase. Molecular and cellular biology 20, 1278-1290 (2000). 73. Osman, F., Jarrous, N., Ben-Asouli, Y. & Kaempfer, R. A cis-acting element in the 3'- untranslated region of human TNF-alpha mRNA renders splicing dependent on the activation of protein kinase PKR. Genes & development 13, 3280-3293 (1999). 74. Williams, B.R. & Kerr, I.M. Inhibition of protein synthesis by 2'-5' linked adenine oligonucleotides in intact cells. Nature 276, 88-90 (1978). 75. Floyd-Smith, G., Slattery, E. & Lengyel, P. Interferon action: RNA cleavage pattern of a (2'- 5')oligoadenylate--dependent endonuclease. Science (New York, N.Y 212, 1030-1032 (1981). 76. Cole, J.L., Carroll, S.S., Blue, E.S., Viscount, T. & Kuo, L.C. Activation of RNase L by 2',5'- oligoadenylates. Biophysical characterization. The Journal of biological chemistry 272, 19187-19192 (1997). 77. Bass, B.L., Weintraub, H., Cattaneo, R. & Billeter, M.A. Biased hypermutation of viral RNA genomes could be due to unwinding/modification of double-stranded RNA. Cell 56, 331 (1989). 78. Liu, Y. & Samuel, C.E. Mechanism of interferon action: functionally distinct RNA-binding and catalytic domains in the interferon-inducible, double-stranded RNA-specific adenosine deaminase. Journal of virology 70, 1961-1968 (1996). 79. Patterson, J.B., Thomis, D.C., Hans, S.L. & Samuel, C.E. Mechanism of interferon action: double-stranded RNA-specific adenosine deaminase from human cells is inducible by alpha and gamma interferons. Virology 210, 508-511 (1995). 80. Groettrup, M., Khan, S., Schwarz, K. & Schmidtke, G. Interferon-gamma inducible exchanges of 20S proteasome active site subunits: why? Biochimie 83, 367-372 (2001). 81. Lah, T.T., Hawley, M., Rock, K.L. & Goldberg, A.L. Gamma-interferon causes a selective induction of the lysosomal proteases, cathepsins B and L, in macrophages. FEBS letters 363, 85-89 (1995). 82. Lafuse, W.P., Brown, D., Castle, L. & Zwilling, B.S. IFN-gamma increases cathepsin H mRNA levels in mouse macrophages. Journal of leukocyte biology 57, 663-669 (1995). 83. Deiss, L.P., Galinka, H., Berissi, H., Cohen, O. & Kimchi, A. Cathepsin D protease mediates programmed cell death induced by interferon-gamma, Fas/APO-1 and TNF-alpha. The EMBO journal 15, 3861-3870 (1996). 84. Chang, C.H. & Flavell, R.A. Class II transactivator regulates the expression of multiple genes involved in antigen presentation. The Journal of experimental medicine 181, 765-767 (1995). 85. Mach, B., Steimle, V., Martinez-Soria, E. & Reith, W. Regulation of MHC class II genes: lessons from a disease. Annual review of immunology 14, 301-331 (1996). 86. Boss, J.M. Regulation of transcription of MHC class II genes. Current opinion in immunology 9, 107-113 (1997). 87. Macatonia, S.E. et al. Dendritic cells produce IL-12 and direct the development of Th1 cells from naive CD4+ T cells. J Immunol 154, 5071-5079 (1995). 88. Trinchieri, G. Interleukin-12: a proinflammatory cytokine with immunoregulatory functions that bridge innate resistance and antigen-specific adaptive immunity. Annual review of immunology 13, 251-276 (1995). 89. Heufler, C. et al. Interleukin-12 is produced by dendritic cells and mediates T helper 1 development as well as interferon-gamma production by T helper 1 cells. European journal of immunology 26, 659-668 (1996). 90. Yoshida, A., Koide, Y., Uchijima, M. & Yoshida, T.O. IFN-gamma induces IL-12 mRNA expression by a murine macrophage cell line, J774. Biochemical and biophysical research communications 198, 857-861 (1994). 91. Cassatella, M.A. et al. Interleukin-12 production by human polymorphonuclear leukocytes. European journal of immunology 25, 1-5 (1995). 92. Maggi, E. et al. Reciprocal regulatory effects of IFN-gamma and IL-4 on the in vitro development of human Th1 and Th2 clones. J Immunol 148, 2142-2147 (1992). 93. Erbe, D.V., Collins, J.E., Shen, L., Graziano, R.F. & Fanger, M.W. The effect of cytokines on the expression and function of Fc receptors for IgG on human myeloid cells. Molecular immunology 27, 57-67 (1990). 86 Chapter 2

94. Snapper, C.M. et al. Induction of IgG3 secretion by interferon gamma: a model for T cell- independent class switching in response to T cell-independent type 2 antigens. The Journal of experimental medicine 175, 1367-1371 (1992). 95. Collins, J.T. & Dunnick, W.A. Germline transcripts of the murine immunoglobulin gamma 2a gene: structure and induction by IFN-gamma. International immunology 5, 885-891 (1993). 96. Finkelman, F.D. et al. Lymphokine control of in vivo immunoglobulin isotype selection. Annual review of immunology 8, 303-333 (1990). 97. Lee, C.E., Yoon, S.R. & Pyun, K.H. Mechanism of interferon-gamma down-regulation of the interleukin 4-induced CD23/Fc epsilon RII expression in human B cells: post-transcriptional modulation by interferon-gamma. Molecular immunology 30, 301-307 (1993). 98. Renzi, P.M. et al. Reduced interferon-gamma production in infants with bronchiolitis and asthma. American journal of respiratory and critical care medicine 159, 1417-1422 (1999). 99. Liao, F. et al. Human Mig chemokine: biochemical and functional characterization. The Journal of experimental medicine 182, 1301-1314 (1995). 100. Rollins, B.J., Yoshimura, T., Leonard, E.J. & Pober, J.S. Cytokine-activated human endothelial cells synthesize and secrete a monocyte chemoattractant, MCP-1/JE. The American journal of pathology 136, 1229-1233 (1990). 101. Taub, D.D. et al. Recombinant human interferon-inducible protein 10 is a chemoattractant for human monocytes and T lymphocytes and promotes T cell adhesion to endothelial cells. The Journal of experimental medicine 177, 1809-1814 (1993). 102. Taub, D.D., Conlon, K., Lloyd, A.R., Oppenheim, J.J. & Kelvin, D.J. Preferential migration of activated CD4+ and CD8+ T cells in response to MIP-1 alpha and MIP-1 beta. Science (New York, N.Y 260, 355-358 (1993). 103. Marfaing-Koka, A. et al. Regulation of the production of the RANTES chemokine by endothelial cells. Synergistic induction by IFN-gamma plus TNF-alpha and inhibition by IL-4 and IL-13. J Immunol 154, 1870-1878 (1995). 104. Hou, J., Baichwal, V. & Cao, Z. Regulatory elements and transcription factors controlling basal and cytokine-induced expression of the gene encoding intercellular adhesion molecule 1. Proceedings of the National Academy of Sciences of the United States of America 91, 11641- 11645 (1994). 105. Tanaka, Y. et al. T-cell adhesion induced by proteoglycan-immobilized cytokine MIP-1 beta. Nature 361, 79-82 (1993). 106. Middleton, J., Patterson, A.M., Gardner, L., Schmutz, C. & Ashton, B.A. Leukocyte extravasation: chemokine transport and presentation by the endothelium. Blood 100, 3853- 3860 (2002). 107. Ellis, T.N. & Beaman, B.L. Interferon-gamma activation of polymorphonuclear neutrophil function. Immunology 112, 2-12 (2004). 108. Newburger, P.E., Ezekowitz, R.A., Whitney, C., Wright, J. & Orkin, S.H. Induction of phagocyte cytochrome b heavy chain gene expression by interferon gamma. Proceedings of the National Academy of Sciences of the United States of America 85, 5215-5219 (1988). 109. Cassatella, M.A. et al. Molecular basis of interferon-gamma and lipopolysaccharide enhancement of phagocyte respiratory burst capability. Studies on the gene expression of several NADPH oxidase components. The Journal of biological chemistry 265, 20241-20246 (1990). 110. Gupta, J.W., Kubin, M., Hartman, L., Cassatella, M. & Trinchieri, G. Induction of expression of genes encoding components of the respiratory burst oxidase during differentiation of human myeloid cell lines induced by tumor necrosis factor and gamma-interferon. Cancer research 52, 2530-2537 (1992). 111. Schoedon, G., Troppmair, J., Adolf, G., Huber, C. & Niederwieser, A. Interferon-gamma enhances biosynthesis of pterins in peripheral blood mononuclear cells by induction of GTP- cyclohydrolase I activity. Journal of interferon research 6, 697-703 (1986). 112. McCall, T.B., Palmer, R.M. & Moncada, S. Induction of nitric oxide synthase in rat peritoneal neutrophils and its inhibition by dexamethasone. European journal of immunology 21, 2523- 2527 (1991). 113. Di Silvio, M. et al. Inducible nitric oxide synthase activity in hepatocytes is dependent on the coinduction of tetrahydrobiopterin synthesis. Advances in experimental medicine and biology 338, 305-308 (1993). 114. Nussler, A.K., Billiar, T.R., Liu, Z.Z. & Morris, S.M., Jr. Coinduction of nitric oxide synthase and argininosuccinate synthetase in a murine macrophage cell line. Implications for regulation of nitric oxide production. The Journal of biological chemistry 269, 1257-1261 (1994). Introduction: Interferon- 87

115. Petroni, K.C., Shen, L. & Guyre, P.M. Modulation of human polymorphonuclear leukocyte IgG Fc receptors and Fc receptor-mediated functions by IFN-gamma and glucocorticoids. J Immunol 140, 3467-3472 (1988). 116. Klebanoff, S.J. et al. Effects of gamma-interferon on human neutrophils: protection from deterioration on storage. Blood 80, 225-234 (1992). 117. Drevets, D.A., Leenen, P.J. & Campbell, P.A. Complement receptor type 3 mediates phagocytosis and killing of Listeria monocytogenes by a TNF-alpha- and IFN-gamma- stimulated macrophage precursor hybrid. Cellular immunology 169, 1-6 (1996). 118. Williamson, B.D., Carswell, E.A., Rubin, B.Y., Prendergast, J.S. & Old, L.J. Human tumor necrosis factor produced by human B-cell lines: synergistic cytotoxic interaction with human interferon. Proceedings of the National Academy of Sciences of the United States of America 80, 5397-5401 (1983). 119. Fransen, L., Van der Heyden, J., Ruysschaert, R. & Fiers, W. Recombinant tumor necrosis factor: its effect and its synergism with interferon-gamma on a variety of normal and transformed human cell lines. European journal of cancer & clinical oncology 22, 419-426 (1986). 120. Brouckaert, P.G., Leroux-Roels, G.G., Guisez, Y., Tavernier, J. & Fiers, W. In vivo anti- tumour activity of recombinant human and murine TNF, alone and in combination with murine IFN-gamma, on a syngeneic murine melanoma. International journal of cancer 38, 763-769 (1986). 121. Schmiegel, W.H. et al. Antiproliferative effects exerted by recombinant human tumor necrosis factor-alpha (TNF-alpha) and interferon-gamma (IFN-gamma) on human pancreatic tumor cell lines. Pancreas 3, 180-188 (1988). 122. Suffys, P., Beyaert, R., Van Roy, F. & Fiers, W. TNF in combination with interferon-gamma is cytotoxic to normal, untransformed mouse and rat embryo fibroblast-like cells. Anticancer research 9, 167-171 (1989). 123. Stolpen, A.H., Guinan, E.C., Fiers, W. & Pober, J.S. Recombinant tumor necrosis factor and immune interferon act singly and in combination to reorganize human vascular endothelial cell monolayers. The American journal of pathology 123, 16-24 (1986). 124. Aggarwal, B.B., Eessalu, T.E. & Hass, P.E. Characterization of receptors for human tumour necrosis factor and their regulation by gamma-interferon. Nature 318, 665-667 (1985). 125. Ruggiero, V., Tavernier, J., Fiers, W. & Baglioni, C. Induction of the synthesis of tumor necrosis factor receptors by interferon-gamma. J Immunol 136, 2445-2450 (1986). 126. Tsujimoto, M., Yip, Y.K. & Vilcek, J. Interferon-gamma enhances expression of cellular receptors for tumor necrosis factor. J Immunol 136, 2441-2444 (1986). 127. Tsujimoto, M., Feinman, R. & Vilcek, J. Differential effects of type I IFN and IFN-gamma on the binding of tumor necrosis factor to receptors in two human cell lines. J Immunol 137, 2272-2276 (1986). 128. Mori, H., Kondoh, H. & Tamaya, T. Synergistic effects of interferon-gamma and tumor necrosis factor against proliferation of gynecologic tumor cell lines. Gynecologic oncology 35, 209-214 (1989). 129. Aggarwal, B.B. & Eessalu, T.E. Induction of receptors for tumor necrosis factor-alpha by interferons is not a major mechanism for their synergistic cytotoxic response. The Journal of biological chemistry 262, 10000-10007 (1987). 130. Chin, Y.E., Kitagawa, M., Kuida, K., Flavell, R.A. & Fu, X.Y. Activation of the STAT signaling pathway can cause expression of caspase 1 and apoptosis. Molecular and cellular biology 17, 5328-5337 (1997). 131. Li, J.H. et al. Interferon-gamma augments CD95(APO-1/Fas) and pro-caspase-8 expression and sensitizes human vascular endothelial cells to CD95-mediated apoptosis. The American journal of pathology 161, 1485-1495 (2002). 132. Dai, C. & Krantz, S.B. Interferon gamma induces upregulation and activation of caspases 1, 3, and 8 to produce apoptosis in human erythroid progenitor cells. Blood 93, 3309-3316 (1999). 133. Tamura, T. et al. DNA damage-induced apoptosis and Ice gene induction in mitogenically activated T lymphocytes require IRF-1. Leukemia 11 Suppl 3, 439-440 (1997). 134. Guicciardi, M.E. et al. Cathepsin B contributes to TNF-alpha-mediated hepatocyte apoptosis by promoting mitochondrial release of cytochrome c. The Journal of clinical investigation 106, 1127-1137 (2000). 135. Foghsgaard, L. et al. Cathepsin B acts as a dominant execution protease in tumor cell apoptosis induced by tumor necrosis factor. The Journal of cell biology 153, 999-1010 (2001). 88 Chapter 2

136. Heinrich, M. et al. Cathepsin D targeted by acid sphingomyelinase-derived ceramide. The EMBO journal 18, 5252-5263 (1999). 137. Li, J.H. & Pober, J.S. The cathepsin B death pathway contributes to TNF plus IFN-gamma- mediated human endothelial injury. J Immunol 175, 1858-1866 (2005). 138. Deiss, L.P., Feinstein, E., Berissi, H., Cohen, O. & Kimchi, A. Identification of a novel serine/threonine kinase and a novel 15-kD protein as potential mediators of the gamma interferon-induced cell death. Genes & development 9, 15-30 (1995). 139. Kissil, J.L. et al. Isolation of DAP3, a novel mediator of interferon-gamma-induced cell death. The Journal of biological chemistry 270, 27932-27936 (1995). 140. Ahn, E.Y., Pan, G., Vickers, S.M. & McDonald, J.M. IFN-gamma upregulates apoptosis- related molecules and enhances Fas-mediated apoptosis in human cholangiocarcinoma. International journal of cancer 100, 445-451 (2002). 141. Chawla-Sarkar, M. et al. Apoptosis and interferons: role of interferon-stimulated genes as mediators of apoptosis. Apoptosis 8, 237-249 (2003). 142. Kamachi, M. et al. Regulation of apoptotic cell death by cytokines in a human salivary gland cell line: distinct and synergistic mechanisms in apoptosis induced by tumor necrosis factor alpha and interferon gamma. The Journal of laboratory and clinical medicine 139, 13-19 (2002). 143. Yonekura, N. et al. Interferon-gamma downregulates Hsp27 expression and suppresses the negative regulation of cell death in oral squamous cell carcinoma lines. Cell death and differentiation 10, 313-322 (2003). 144. Arora, V. et al. Degradation of survivin by the X-linked inhibitor of apoptosis (XIAP)-XAF1 complex. The Journal of biological chemistry 282, 26202-26209 (2007). 145. Garrido, C. et al. HSP27 inhibits cytochrome c-dependent activation of procaspase-9. Faseb J 13, 2061-2070 (1999). 146. Bruey, J.M. et al. Hsp27 negatively regulates cell death by interacting with cytochrome c. Nature cell biology 2, 645-652 (2000). 147. Kissil, J.L., Cohen, O., Raveh, T. & Kimchi, A. Structure-function analysis of an evolutionary conserved protein, DAP3, which mediates TNF-alpha- and Fas-induced cell death. The EMBO journal 18, 353-362 (1999). 148. Miyazaki, T. et al. Functional role of death-associated protein 3 (DAP3) in anoikis. The Journal of biological chemistry 279, 44667-44672 (2004). 149. Kim, H.R. et al. Mammalian dap3 is an essential gene required for mitochondrial homeostasis in vivo and contributing to the extrinsic pathway for apoptosis. Faseb J 21, 188-196 (2007). 150. Inbal, B. et al. DAP kinase links the control of apoptosis to metastasis. Nature 390, 180-184 (1997). 151. Cohen, O. et al. DAP-kinase participates in TNF-alpha- and Fas-induced apoptosis and its function requires the death domain. The Journal of cell biology 146, 141-148 (1999). 152. Pelled, D. et al. Death-associated protein (DAP) kinase plays a central role in ceramide- induced apoptosis in cultured hippocampal neurons. The Journal of biological chemistry 277, 1957-1961 (2002). 153. Jang, C.W. et al. TGF-beta induces apoptosis through Smad-mediated expression of DAP- kinase. Nature cell biology 4, 51-58 (2002). 154. Cohen, O., Feinstein, E. & Kimchi, A. DAP-kinase is a Ca2+/calmodulin-dependent, cytoskeletal-associated protein kinase, with cell death-inducing functions that depend on its catalytic activity. The EMBO journal 16, 998-1008 (1997). 155. Wang, W.J., Kuo, J.C., Yao, C.C. & Chen, R.H. DAP-kinase induces apoptosis by suppressing integrin activity and disrupting matrix survival signals. The Journal of cell biology 159, 169-179 (2002). 156. Kuo, J.C., Lin, J.R., Staddon, J.M., Hosoya, H. & Chen, R.H. Uncoordinated regulation of stress fibers and focal adhesions by DAP kinase. Journal of cell science 116, 4777-4790 (2003). 157. Paleolog, E.M., Aluri, G.R. & Feldmann, M. Contrasting effects of interferon gamma and interleukin 4 on responses of human vascular endothelial cells to tumour necrosis factor alpha. Cytokine 4, 470-478 (1992). 158. Hachicha, M., Rathanaswami, P., Schall, T.J. & McColl, S.R. Production of monocyte chemotactic protein-1 in human type B synoviocytes. Synergistic effect of tumor necrosis factor alpha and interferon-gamma. Arthritis and rheumatism 36, 26-34 (1993). Introduction: Interferon- 89

159. Ohmori, Y. & Hamilton, T.A. The interferon-stimulated response element and a kappa B site mediate synergistic induction of murine IP-10 gene transcription by IFN-gamma and TNF- alpha. J Immunol 154, 5235-5244 (1995). 160. Jahnke, A. & Johnson, J.P. Intercellular adhesion molecule 1 (ICAM-1) is synergistically activated by TNF-alpha and IFN-gamma responsive sites. Immunobiology 193, 305-314 (1995). 161. Kamijo, R. et al. Generation of nitric oxide and induction of major histocompatibility complex class II antigen in macrophages from mice lacking the interferon gamma receptor. Proceedings of the National Academy of Sciences of the United States of America 90, 6626- 6630 (1993). 162. Drew, P.D. et al. Interferon regulatory factor-2 physically interacts with NF-kappa B in vitro and inhibits NF-kappa B induction of major histocompatibility class I and beta 2- microglobulin gene expression in transfected human neuroblastoma cells. Journal of neuroimmunology 63, 157-162 (1995). 163. Drew, P.D. et al. NF kappa B and interferon regulatory factor 1 physically interact and synergistically induce major histocompatibility class I gene expression. J Interferon Cytokine Res 15, 1037-1045 (1995). 164. Ganster, R.W., Guo, Z., Shao, L. & Geller, D.A. Differential effects of TNF-alpha and IFN- gamma on gene transcription mediated by NF-kappaB-Stat1 interactions. J Interferon Cytokine Res 25, 707-719 (2005). 165. Suk, K. et al. Interferon gamma (IFNgamma ) and tumor necrosis factor alpha synergism in ME-180 cervical cancer cell apoptosis and necrosis. IFNgamma inhibits cytoprotective NF- kappa B through STAT1/IRF-1 pathways. The Journal of biological chemistry 276, 13153- 13159 (2001). 166. Kumar, A., Haque, J., Lacoste, J., Hiscott, J. & Williams, B.R. Double-stranded RNA- dependent protein kinase activates transcription factor NF-kappa B by phosphorylating I kappa B. Proceedings of the National Academy of Sciences of the United States of America 91, 6288-6292 (1994). 167. Cheshire, J.L. & Baldwin, A.S., Jr. Synergistic activation of NF-kappaB by tumor necrosis factor alpha and gamma interferon via enhanced I kappaB alpha degradation and de novo I kappaBbeta degradation. Molecular and cellular biology 17, 6746-6754 (1997). 168. Han, Y., Rogers, N. & Ransohoff, R.M. Tumor necrosis factor-alpha signals to the IFN- gamma receptor complex to increase Stat1alpha activation. J Interferon Cytokine Res 19, 731- 740 (1999). 169. Wang, Y., Wu, T.R., Cai, S., Welte, T. & Chin, Y.E. Stat1 as a component of tumor necrosis factor alpha receptor 1-TRADD signaling complex to inhibit NF-kappaB activation. Molecular and cellular biology 20, 4505-4512 (2000). 170. Wesemann, D.R. & Benveniste, E.N. STAT-1 alpha and IFN-gamma as modulators of TNF- alpha signaling in macrophages: regulation and functional implications of the TNF receptor 1:STAT-1 alpha complex. J Immunol 171, 5313-5319 (2003). 171. Wesemann, D.R., Qin, H., Kokorina, N. & Benveniste, E.N. TRADD interacts with STAT1- alpha and influences interferon-gamma signaling. Nature immunology 5, 199-207 (2004). 172. Balkwill, F.R. et al. Human tumor xenografts treated with recombinant human tumor necrosis factor alone or in combination with interferons. Cancer research 46, 3990-3993 (1986). 173. Balkwill, F.R., Ward, B.G. & Fiers, W. Effects of tumour necrosis factor on human tumour xenografts in nude mice. Ciba Foundation symposium 131, 154-169 (1987). 174. Rubin, S.C. et al. Synergistic activity of tumor necrosis factor and interferon in a nude mouse model of human ovarian cancer. Gynecologic oncology 34, 353-356 (1989). 175. Nio, Y. et al. In vivo effects of human recombinant tumor necrosis factor alone and in combination with other biological response modifiers on human digestive organ cancer xenografts transplanted in nude mice. Biotherapy (Dordrecht, Netherlands) 3, 337-344 (1991). 176. Sun, W.H. et al. In vivo cytokine gene transfer by gene gun reduces tumor growth in mice. Proceedings of the National Academy of Sciences of the United States of America 92, 2889- 2893 (1995). 177. Lasek, W. et al. Potentiation of antitumor effects of tumor necrosis factor alpha and interferon gamma by macrophage-colony-stimulating factor in a MmB16 melanoma model in mice. Cancer Immunol Immunother 40, 315-321 (1995). 178. de Kossodo, S. et al. Changes in endogenous cytokines, adhesion molecules and platelets during cytokine-induced tumour necrosis. British journal of cancer 72, 1165-1172 (1995). 90 Chapter 2

179. Lejeune, F.J., Lienard, D., Matter, M. & Ruegg, C. Efficiency of recombinant human TNF in human cancer therapy. Cancer Immun 6, 6 (2006). 180. Onishi, T. et al. [In vivo anti-tumour efficacy of tumour necrosis factor and interferon- alpha, - gamma on human renal cell carcinoma heterotransplanted in nude mice]. Nippon Gan Chiryo Gakkai shi 25, 1571-1578 (1990). 181. Hostens, J., Goethals, A., Breier, G., Schreiber, R.D., Brouckaert, P. IFN- synergistically enhances the anti-tumour effect of Tumour Necrosis Factor by activating the endothelial cells of the neovasculature. (article in preparation). 182. Hostens, J., Brouckaert, P. (unpublished data). (2003). 183. Ruegg, C. et al. Evidence for the involvement of endothelial cell integrin alphaVbeta3 in the disruption of the tumor vasculature induced by TNF and IFN-gamma. Nature medicine 4, 408- 414 (1998).

Chapter 3 The PI3K/Akt pathway

Introduction: The PI3K/Akt pathway 93

3. THE PI3K/AKT PATHWAY

3.1. PI3K

3.1.1. The PI3K family

PI3Ks are a family of related enzymes that are capable of phosphorylating the 3’-OH position of the inositol ring of inositol phospholipids, thereby producing three lipid products, namely PI 3-phosphate (PI(3)P), PI 3,4-biphosphate (PI(3,4)P2) and PI 2 3,4,5-triphosphate (PI(3,4,5)P3) . These phosphorylated PIs function as docking sites for proteins containing a pleckstrin homology (PH) domain, which then mediate further signaling. The serine/threonine kinase Akt, also called protein kinase B (PKB), was one of the first proteins identified with such a PH domain 3-5. The PH domain of Akt appears to specifically bind to PI3K lipid products, and a firm link between PI3K and Akt signaling has now been established, hence the name ‘PI3K/Akt pathway’.

To date, eight different mammalian PI3Ks have been identified, which have been divided into three classes, based on sequence homology, in vitro substrate preference, and method of activation and regulation (Fig.3.1. and 3.2.) 6. Only the class I PI3Ks are known to activate Akt and will be discussed in more detail below. The three class II PI3Ks, PI3K-C2, -C2 and –C2, are large proteins that are most likely constitutively associated with membrane structures, and in contrast to the class I and III PI3Ks do not appear to associate with regulatory adaptor proteins. They have several domains that are homologous to protein-protein interaction and lipid-binding domains found in other proteins. Particularly the characteristic carboxyterminal C2 domain, homologous to the Ca2+/lipid binding C2 domains found in classical PKC isoforms, could mediate the membrane association of class II PI3Ks. Their main in vivo product is PI(3)P, although in vitro they can also phosphorylate PI(4)P to 1 PI(4,5)P2 . Class II PI3Ks can be activated by many extracellular signals, including growth factors such as insulin and epidermal growth factor (EGF), chemokines such as MCP-1, and integrin engagement 7-10. However, the mechanism involved in their activation as well as their cellular functions are still largely unknown. PI3K-C2 and –C2 are ubiquitously expressed, while the expression of PI3K-C2 appears to be limited to hepatocytes 11, 12. The sole class III PI3K is Vps34 (sometimes referred to as hVps34 when the mammalian homologue of Vps34 is intended). Vps34 was originally identified as the only PI3K present in Saccharomyces cerevisiae 13, 14. Vps34 is associated with a regulatory subunit, Vps15, with serine/threonine kinase activity that is required for Vps34 function. Vps34 is ubiquitously expressed, is mainly associated with intracellular membranes, and uses only PI as a substrate, thereby producing PI(3)P. In yeast, Vps34 has an essential role in protein trafficking to the vacuole. This pathway and function are conserved in mammalian cells, where Vps34 is involved in vesicular trafficking in the endosomal/lysosomal system 1, 15, 16. 94 Chapter 3

Recently Vps34 has also been implicated in other important signaling processes, such as nutrient sensing in the mTOR/p70 S6 kinase (S6K) pathway 17, 18.

A number of other enzymes have been identified with catalytic domains that closely resemble those of PI3Ks (Fig.3.1.). These include the PI 4-kinases (PI4Ks), which phosphorylate the inositol ring of PI at the 4’-position to produce PI(4)P. There are also some PI3K-related protein serine/threonine kinases which are sometimes called class IV PI3Ks. In mammals, four such protein kinases have been identified: mTOR, ataxia telangiectasia mutated (ATM), ATM-related (ATR), and DNA-dependent protein kinase (DNA-PK). ATM, ATR and DNA-PK are all involved in the activation of DNA repair machinery and the linking of this to cell cycle control. The functions of mTOR will be discussed in more detail below (see chapter 3.3) 1, 6.

Fig. 3.1. The PI3K superfamily. Dendrogram showing the degree of sequence homology between the catalytic domains of PI3Ks and related kinases. From Stein et al., 2001 6.

Introduction: The PI3K/Akt pathway 95

.

1

, 2003 2003 ,

et et al.

From Foster

Fig. 3.2. PI3K The family members, their structure, subcellular localizationand function.

96 Chapter 3

3.1.2. Class I PI3Ks

Class I PI3Ks are heterodimeric proteins consisting of a catalytic subunit of approximately 110 kDa that is constitutively associated with a regulatory/adaptor subunit. Class I PI3Ks are largely cytosolic in resting cells, but upon stimulation are recruited to membranes via interactions with receptors or adaptor proteins. They are thought to function primarily at the plasma membrane, but there are reports of class I PI3Ks associated with vesicular and nuclear membranes. They become activated by a plethora of extracellular stimuli and have been implicated in a wide range of cellular processes, including cell growth, cell cycle progression, proliferation, motility,

adhesion and survival. The preferred in vitro substrate of class I PI3Ks is PI(4,5)P2, hence the principal lipid generated by their in vivo actions is believed to be

PI(3,4,5)P3. PI(3,4,5)P3 mediates further downstream signaling of class I PI3Ks by recruiting PH domain-containing proteins, including (1) members of the Tec family of tyrosine kinases, which link PI3K signaling to PLC and Ca2+/PKC activity; (2) guanine nucleotide exchange factors (GEFs) for Rho and Rac GTPases, which are involved in reorganization of the actin cytoskeleton; and (3) the serine/threonine kinases Akt and PI(3,4,5)P3-dependent protein kinase-1 (PDK-1) (see below) 19. Besides their lipid kinase activity, class I PI3Ks also have a protein serine kinase activity, which has been implicated in the regulation of the lipid kinase activity through either autophosphorylation of the catalytic subunit or transphosphorylation of the regulatory subunit 20-23. There are four distinct p110 isoforms identified in vertebrates, which have been divided in a class Ia and class Ib on the basis of their mechanism of activation and the regulatory subunit they associate with 1, 19.

The class Ia PI3K catalytic subunits are p110, p110 and p110. These p110 subunits are constitutively associated with a p85 regulatory/adaptor subunit via binding of the aminoterminal p85 binding domain of p110 with the inter-SH2 domain of p85. At least eight distinct p85 isoforms have been identified (encoded by three different genes): p85, p55, p50, p85 p55, and some other splice variants. There does not appear to be preferential association of any of the p110 isoforms with particular p85 isoforms, but there is variation in tissue distribution of catalytic and adapter subunit isoforms. p110, p110, p85 and p85 are expressed in almost all tissues. p110 and p55 on the other hand display a more restricted tissue distribution: p110 is predominantly expressed in leukocytes, while p55 is highly expressed in brain and testis 24-26. All of the p85 family members contain two SH2 domains which allow binding of the PI3K to specific phosphorylated tyrosine residues in receptors and other proteins. Some of the p85and p85 splice forms also possess an SH3 domain, which allows binging to proline-rich regions in proteins, and a Bcr/Rac GAP homology domain, which allows binding to some additional binding partners 27-29. The p110 subunit also contains a C2 domain for membrane association and a Ras- Introduction: The PI3K/Akt pathway 97

binding domain 30. Thus, the number of potential binding partners of the p85/p110 dimer is tremendous (Fig.3.3.).

Fig. 3.3. Domain structure, interaction partners and regulatory activities of the p85/p110 PI3K. Broken arrows represent the protein serine kinase activity of p110, which can lead to downregulation of the lipid kinase activity of the complex. This occurs by transphosphorylation of the inter-SH2 (iSH2) domain of p85 in the case of p110, or by autophosphorylation in the case of p110 and p110. Regulatory subunits stabilize the p110 proteins, but at the same time inhibit their catalytic activity. This inhibition can be alleviated by engagement of the SH2 domains of p85 with phosphotyrosine sequences (pY) in other proteins. The Bcl/Rac GAP homology (BH) domain of p85 and p85 is flanked by two proline-rich regions. In p55, p50 and p55 the aminoterminal (N) SH3 domain, proline-rich region and BH domain are absent. Adapted from Vanhaesebroeck et al., 2005 30.

The model for PI3K activation was first worked out in the context of the platelet- derived growth factor (PDGF) receptor in fibroblasts. Stimulation of this receptor leads to autophosphorylation of a specific tyrosine residue within the cytoplasmic tail of the receptor, which then recruits the p85/p110 PI3K by a high-affinity interaction of the SH2 domains in the p85 adaptor subunit with the tyrosine-phosphorylated sequence in the PDGF receptor 31-33. The p85-mediated translocation of the p110 catalytic subunit from the cytosol to the plasma membrane is sufficient for the activation of the PI3K, as constitutive membrane targeting of the p110 catalytic subunit creates a constitutively active PI3K that generates PI(3,4,5)P3 when expressed in cells 34. However, the p85 subunit is more than just a vehicle to bring the catalytic subunit to the plasma membrane, it may also regulate the activity of the PI3K. The catalytic activity of free p110 was shown to be higher than that of unbound p85/p110 heterodimers, suggesting that p85 has an inhibitory effect on the catalytic activity of the p110 subunit. Binding of p85 to specific phosphotyrosine sequences would alleviate this inhibition 35, 36. It is now clear that class Ia PI3Ks are activated by many receptors with intrinsic tyrosine kinase activity (RTKs) such as growth factor receptors, but also by non-receptor tyrosine kinases, such as JAKs and Src family 98 Chapter 3

members, which have been implicated in PI3K activity downstream of many cytokine receptors, B and T cell antigen receptors, co-stimulatory molecules such as CD28, integrins and other cell adhesion molecules. In addition, class I PI3Ks can be activated by the small GTPase Ras 2. Moreover, a recent study indicated that p110 also signals downstream of GPCRs 37.

The only class Ib catalytic subunit is p110. It associates with a p101 or p84 regulatory/adaptor subunit which appears to lack all the protein binding domains of p85, but contains a domain that is capable of binding to G subunits of heterotrimeric G-proteins 38, 39. Hence, the p110 functions downstream of GPCRs, such as chemokine receptors and the receptors for PAF and histamine. p110 is predominantly expressed in leukocytes, although its presence has also been demonstrated in endothelial cells and cardiomyocytes 40, 41. Like the class Ia p110 isoforms, p110 contains a C2 domain and a Ras-binding domain, which allow lipid membrane and Ras binding 1.

Two unrelated, broad-spectrum PI3K inhibitors are often used to study the involvement of PI3K signaling in cellular processes: wortmannin and LY294002. Wortmannin is an irreversible PI3K inhibitor, probably by covalent modification of a lysine residue in the ATP binding site of PI3K 42. It is reasonably specific for PI3K,

with an IC50 of 1-10nM for all class I PI3Ks, PI3K-C2, PI3K-C2, and class III PI3K, whereas PI3K-C2 and the class IV PI3K-related enzymes are markedly less sensitive to wortmannin. However, as wortmannin is an irreversible inhibitor, it may have cumulative effects. LY294002 on the other hand is a pure competitive inhibitor of ATP and far less potent than wortmannin. It inhibits all class I, II and III PI3Ks

with an IC50 of 1-50M. At these concentrations it also potently inhibits class IV PI3K-related enzymes such as mTOR and DNA-PK, and unrelated kinases such as casein kinase 2 (CK2) 6. Thus, some caution is needed when interpreting results obtained with these broad-spectrum inhibitors. Many pharmaceutical companies are now developing p110 isoform-selective inhibitors, which are suitable for studying the role of specific PI3K isoforms in cellular processes. Particularly p110- and p110- selective inhibitors are being developed, as these isoforms are promising targets for the treatment of allergic and chronic inflammatory diseases 43. Another approach to study the role of specific PI3K isoforms is the generation of p110 isoform-deficient mice. This has however proven to be difficult as both p110 and p110 knockout mice are not viable, and genetic deletion of p110 isoforms, particularly p110, was shown to alter the expression levels of p85 adaptor subunits, which may influence the phenotype of the knockout mice. The latter problem has been overcome by the generation of p110 isoform kinase-death knockin mice, which display normal expression levels of p110 and p85 subunits 30. The phenotypes of p110 isoform-deficient mice are summarized in Table 3.1. Notably, recent studies with Introduction: The PI3K/Akt pathway 99

heterozygous p110 knockin/wild-type mice and endothelial-specific p110 or p110 knockout mice pointed to a critical role of p110 in angiogenesis to control endothelial cell migration 44.

The PI(3,4,5)P3 produced by class I PI3Ks is degraded by two distinct lipid phosphatases: SH2 domain-containing inositol 5-phosphatase (SHIP) and PTEN.

SHIP dephosphorylates the 5'-position of the inositol ring of PI(3,4,5)P3, yielding 45 PI(3,4)P2 . Most PH domains of proteins that bind to PI(3,4,5)P3 can also bind to

PI(3,4)P2, although often with lower affinity. SHIP mediates an important negative 46, 47 feedback to PI3K signaling in leukocytes . PTEN dephosphorylates PI(3,4,5)P3 to

PI(4,5)P2, thereby directly counteracting the activity of class I PI3Ks. Cells that lack PTEN display constitutive activation of Akt and other PI3K-induced signaling pathways 48-50. As PTEN is capable of deactivating Akt signaling, it is not surprising that PTEN was identified as an important tumor suppressor gene. Mutations of PTEN are associated with Cowden syndrome, an inherited disorder characterized by the formation of benign tumors, called hamartomas, and an increased risk of developing several types of cancer, including cancers of the breast, thyroid, and uterus 51. PTEN mutations are frequently found in many types of cancer, especially glioblastoma, advanced prostate cancer, endometrial cancer, melanoma, and to a lesser extent breast, head and neck, and thyroid cancer 52-65. Moreover, p53, another major tumor suppressor, was shown to activate the transcription of PTEN and repress the transcription of p110, suggesting that p53 may also suppress tumorigenesis by acting as a negative regulator of PI3K signaling 66, 67. Another class of Akt activating mutations that was recently identified in human cancers are p110 gain-of-function mutations, which are particularly frequent in breast, ovarium, colorectal and other epithelial cancers 68-70. Two hotspots for gain-of-function mutations are identified in p110: one in the helical domain and a second in the kinase domain (Fig.3.4.) 71. Helical and kinase domain mutations were shown to operate by two different mechanisms, and both mutations can act synergistically. While kinase domain mutations require interaction with p85, helical domain mutations require interaction with Ras-GTP and are independent of p85 72.

100 Chapter 3

]

83

] ] -

] ]

87 90

] ] ] ] - -

73 74 44, 74 75 75, 76 40, 77 84 85 88

References [ [ [ [ [ [ [ [ [



and



antibody antibody

3, 3,

-



Mast cells: cells: Mast

; ;

apoptosisof DP

CD40, CD40,

-



CD3

-

T cells:

;

IgMor anti

-

number of neutrophils blood numberin of neutrophils

.



30

Dendritic cells, neutrophils cells, Dendritic

anti

;

Ag; Ag;

-

, 2005,

intolerant

-

et et al.

B cells: B affected not cells:

; ;

stimulated neutrophils stimulated

in response to response in

-

chemotaxis

adiposity

CD3, but normal proliferation to costimulation with with normal to costimulation proliferation but CD3,

glucose

-





Ag allergic challenge allergic Ag

-

migration of cultured endothelial of cells cultured migration

proliferation in response to stem cell to response factor stem in or proliferation IL

normal expression of p85 normalexpression

allergic response response toIgE allergic





, ,

proliferation proliferation





intolerant;

-

RI,



proliferation of mature T cells in response to in response T anti of cells proliferation mature

cell mass, slightly mass, cell

explants

-



ofFc



remodeling; remodeling;

oxidative burst in chemokine in burst oxidative

induced chemotaxis induced

-

activated endothelial endothelial cells, activated



-

.from Adapted Vanhaesebroeck

pancreatic pancreatic

in aortic ring aortic ring in

proliferation in response in response proliferation toanti



response to or adenosine IgE response

1 and leptin; glucose and 1

retina and hindbrain and retina embryonic



number in some tissues, tissues, some number in

-





RI,

chemokine



p85expression

number of myeloid cells in spleen blood in cells and of spleen myeloid number



stnatal stnatal





or impaired

sprouting and vascular sprouting

;

adhesionto TNF

lsof arthritis rheumatoid

mice.



expression in B cells expressionin

KO/KO; T cells: cells: KO/KO;T



e to insulin, e IGF



defects defects

CD28; Mast CD28; cells:

angiogenic

-

microvessel outgrowth microvessel

migration and migration chemotaxis,

intolerant

: decreased :

-

deficient deficient



andp50



-



Neutrophils:

;

stimulated insulin secretion, insulin stimulated

-

number of mature B cells, B1 and M1 cells absent, of absent, cells M1 mature B cells, number B1 and

glucose

induced induced



,p55

-

: :



CD3 and anti and CD3

-

body weight and skeletal muscle mass; 50% reduction in class Ia PI3K activity in liver, muscle and fat class and muscle Ia in tissue liver, 50% and skeletal reduction mass; muscle in PI3Kactivity weight body VEGF respons in signaling glucose and of thymocytes number aortic constriction after transverse dramaticdevelop myocardial damage contractility, cardiac p85

10 10 isoform

: increased, :

Phenotype angiogenesis severe in defects severe in the po angiogenesis delayed development normal slightly receptor, of adenosine stimulation after thymocytes of Fc downstream degranulation macrophages: mode in protected Largely macrophages: and Neutrophils phenotype cardiac No B cells production as same p110 B cells: anti and cytokine downstream production degranulation       



, ,

10.5)

11)

-

-

knockin

death

-

kinase

Viabiliy (E9.5 lethal Embryonic (E10 lethal Embryonic Viable (E3.5) lethal Embryonic Viable Viable Viable Viable Viable

KO/KO KI/KI KI/+ KO/+,

KO/KO

KO/+

KO/KO KI/KI

KO/KO KI/KI

         

able3.1. Viability and phenotype of p1

T

Genotype p110 p110 p110 p110 p110 p110 p110 p110 p110 p110 KO: KO: knockout, KI: Introduction: The PI3K/Akt pathway 101

Fig. 3.4. Mutations of the p110 gene PIK3CA found in human cancers. Arrowheads indicate the location of missense mutations. The percentages of mutations detected within the p85 binding domain, helical domain, and kinase domain are indicated below. From Samuels et al., 2004 71.

Inhibition of PI3K has been shown to sensitize many cancer cell lines and other cells to the cytotoxic effects of cytokines, chemotherapeutics or irradiation 91-95. Particularly interesting in light of this thesis is the finding that inhibition of PI3K can sensitize cultured endothelial cells to TNF-induced apoptosis 96-98. Madge and co- workers showed that TNF triggers a PI3K-mediated anti-apoptotic response in HUVECs that is independent of Akt and NF-B, and that inhibition of PI3K results in cathepsin B-dependent, but caspase-independent cell death. This suggests that PI3K can prevent the TNF-induced lysosomal permeabilization or activation of cathepsin B, and that this is sufficient to protect endothelial cells from apoptosis induced by TNF 98, 99. Matschurat and colleagues on the other hand reported that inhibition of PI3K strongly enhanced the expression of tissue factor on HUVECs costimulated with vascular endothelial growth factor (VEGF) and TNF. They furthermore showed that co-injection of wortmannin promoted the TNF-induced necrosis of MethA tumors in mice, suggesting that inhibition of PI3K can also sensitize tumor endothelial cells to TNF in vivo 100.

102 Chapter 3

3.2. Akt

The serine/threonine kinase Akt is a key target of class I PI3Ks and regulates cell survival, cell cycle progression as well as protein synthesis. In mammals, three closely related isoforms of Akt have been identified, Akt-1, -2, and -3 . All Akt isoforms have a broad tissue distribution and consist of an aminoterminal PH domain, a central kinase domain with an activation loop, and a carboxyterminal regulatory tail that contains a hydrophobic motif 101. The PH domain of Akt preferentially binds to 102 PI(3,4,5)P3 and with somewhat lower affinity to PI(3,4)P2 . Akt is cytosolic in unstimulated cells but translocates to the plasma membrane upon activation of PI3K 103 . Binding of Akt to PI(3,4,5)P3 does not activate its protein kinase activity, but it induces a conformational change which enables PDK-1 to phosphorylate a specific threonine residue (Thr308 in Akt-1) in the activation loop of Akt 102, 104, 105. PDK-1 is a serine/threonine kinase that contains a PH domain that binds with high affinity to

PI(3,4,5)P3 and PI(3,4)P2 . Like Akt, it is mainly cytosolic in unstimulated cells, and is recruited to the plasma membrane upon activation of PI3K 106, 107. PDK-1 was first identified by its ability to phosphorylate Thr308 of Akt-1, but meanwhile it has become clear that PDK-1 is also important for the regulation of other kinases that act downstream of PI3K, such as PKC isoforms and S6K, by phosphorylating key residues in their respective activation loops 108, 109. Full activation of Akt also requires phosphorylation of a serine residue (Ser473 in Akt-1) in the carboxyterminal hydrophobic motif 110. During the last ten years, a number of enzymes have been shown to be capable of phosphorylating this specific serine residue, including PDK-1 (in a complex with PKC-related kinase 2), integrin-linked kinase (ILK), PKC, PKCII, DNA-PK, ATM, and Akt itself through autophosphorylation 111-117. Most recently, mTOR, in a complex with Rictor, G protein -like (GL) and SAPK- interacting protein 1 (SIN1) (see below) was shown to mediate the phosphorylation of Akt-1 on Ser473 in many cell types 118-120. Thus, the enzyme responsible for serine phosphorylation of Akt will most likely differ depending on the cell type and/or the PI3K-activating stimulus. Once fully activated, Akt translocates from the plasma membrane to the cytosol and the nucleus, where it phosphorylates specific substrates and ultimately undergoes dephosphorylation by phosphatases, such as protein phosphatase 2A (PP2A), to return to its inactive state 121, 122.

Akt plays a key role in survival signaling (Fig.3.5.). Akt directly phosphorylates and inactivates proteins that promote apoptosis, including Bad, human caspase-9, and Forkhead box O (FOXO) transcription factors 123-128. The residue that Akt phosphorylates in human caspase-9 is not conserved in other species, and Akt is unable to phosphorylate other important caspases such as caspase-8 and -3, suggesting that caspase inhibition is not a major pathway by which Akt promotes cell survival 128, 129 . Akt-mediated phosphorylation of Bad causes Bad to dissociate from Bcl-xL at the mitochondrial membrane and to form a complex with 14-3-3 proteins in the Introduction: The PI3K/Akt pathway 103

cytoplasm 130. Similarly, phosphorylation of FOXO transcription factors promotes their export from the nucleus to the cytosol, where they interact with 14-3-3 proteins 124, 131. In addition, phosphorylation of FOXO transcription factors, and other Akt substrates such as tuberin, has been shown to promote the proteasomal degradation of these proteins 127, 132. FOXO transcription factors promote apoptosis by regulating the transcription of pro-apoptotic proteins, including FasL, TRAIL, TRADD, the pro- apoptotic Bcl-2 family member Bim, and Bcl-6, a strong transcriptional repressor of 124, 133-136 Bcl-xL . FOXO transcription factors also inhibit cell cycle progression by upregulating the expression of the cyclin-dependent kinase inhibitor p27Kip1 and downregulating the expression of cyclin D 137-139. Akt also directly promotes cell cycle progression by phosphorylating p27Kip1 and p21Cip1/WAF1. Akt-mediated phosphorylation of p27Kip1 induces nuclear exclusion of this proteins, thereby inhibiting its nuclear function, while in the case of p21Cip1/WAF1 this phosphorylation increases its stability and inhibits its interaction with proliferating cell nuclear antigen (PCNA), a protein involved in DNA replication 140, 141. Akt has also been shown to phosphorylate murine double minute homolog 2 (MDM2), which induces its translocation from the cytoplasm to the nucleus and enhances its E3 ubiquitin ligase activity for p53 142, 143. Ubiquitination of p53 targets this protein for proteasomal degradation and thus inhibits p53-mediated transcription, cell cycle arrest, and apoptosis. Another substrate of Akt is ASK-1, a MAP3K involved in the activation of JNK and p38 MAPKs. Phosphorylation of ASK-1 by Akt was shown to inhibit the activation of JNK and protect cells from ASK-1-induced apoptosis 144. Likewise, Akt has been shown to phosphorylate the MAP3Ks mixed lineage kinase-3 (MLK-3) and SAPK/ERK kinase-1 (SEK-1), thereby promoting cell survival 145, 146. Akt also stimulates the transcription of anti-apoptotic genes by activating the transcription factors NF-B and CREB. Akt can directly phosphorylate CREB, thereby increasing its transcriptional activity which may result in the expression of anti-apoptotic proteins such as Bcl-2 147, 148. Akt may affect NF-B through multiple mechanisms. IL-1 has been shown to stimulate the transactivation potential of NF-B by phosphorylation of RelA, and the DNA-binding activity of NF-B by phosphorylation of NF-B1, both in a PI3K/Akt-dependent manner 149-151. Furthermore, Akt can associate with and activate Cot/Tpl-2, a MAP3K that has been implicated in the activation of NF-B by activating the IKK complex 152. TNF has also been shown to induce Akt-dependent phosphorylation of IKK, and this phosphorylation event was shown to be essential to allow the activation of NF-B 153. Moreover, Akt was shown to transiently interact with the IKK complex and induce IKK activation in cells stimulated with PDGF 154. A recent report by Gustin and colleagues delineated that the involvement of Akt in NF-B activation may differ depending on the cell type investigated. They demonstrated that the ratio of IKK to IKK varies among cell types, and that cells that express high levels of IKK, such as 104 Chapter 3

HUVECs, do not depend on the PI3K/Akt pathway for TNF-induced activation of NF-B, while cells that express high levels of IKK, the target of Akt phosphorylation, such as the 293 cell line, do require PI3K signaling to allow NF-B activation 155. Finally, besides direct and transcriptional control of cell survival, Akt can also mediate metabolic control over cell survival through phosphorylation of GSK3 156. GSK3 was initially identified as the enzyme that regulates glycogen synthesis in response to insulin. When high levels of insulin are present, GSK3 is inhibited through phosphorylation by Akt, with the result that the storage of glucose as glycogen is promoted. It is now clear, however, that GSK3 phosphorylates, besides glycogen synthase, a broad range of substrates which are involved in cell metabolism, cell cycle progression, transcription, translation initiation, and even mitochondrial permeability transition; and GSK3 has emerged as a key target of Akt survival signaling under certain conditions 157-160.

Fig. 3.5. The PI3K/Akt survival pathway. Activated class I PI3Ks produce PI(3,4,5)P3, which results in the recruitment of PDK-1 and Akt to the plasma membrane, where Akt becomes activated by phosphorylation of a specific threonine (T) and serine (S) residue. Once activated, Akt translocates to the cytosol and nucleus where it promotes cell survival by direct inactivation of proteins involved in the apoptotic cascade, by regulating the activity of transcription factors that control the expression of pro- and anti-apoptotic genes, by promoting cell cycle progression, and by regulating cell metabolism, translation and cell growth. Encircled P: phosphoserine or -threonine.

As activation of Akt can promote cell survival under many conditions, including triggering of cell death receptors, oxidative stress, loss of cell adhesion, growth factor withdrawal, or starvation, it is not surprisingly that overactivation of Akt is found in cancer. Besides inactivation of the tumor suppressors PTEN or p53, or p110 gain-of- function mutations, this can also be achieved by overexpression of Akt itself, which is Introduction: The PI3K/Akt pathway 105

in most cases a consequence of Akt gene amplification. Overexpression of Akt is frequently found in breast, ovarian, prostate and pancreatic cancers 161-163. Perifosine is an Akt inhibitor that is currently evaluated in clinical trials. Perifosine is an orally bioavailable alkylphospholipid that structurally resembles naturally occurring phospholipids, such as PIs. The putative mechanism of its inhibitory effect on Akt is that it would interfere with the PH domain-mediated membrane recruitment of Akt 164.

A large number of proteins have been identified that can interact with Akt. Most of these proteins are substrates of Akt. However, several of these interacting proteins have been shown to regulate the activity of Akt (Fig.3.6.) 165. T cell leukemia 1 (TCL1) family members, including TCL1, TCL1b and MTCP1, are proto-oncogenes that specifically interact with the PH domain of Akt 166. Binding of TCL1 induces oligomerization of Akt, and this facilitates the phosphorylation of Akt. Thus, TCL1 functions as a coactivator of Akt 167. In addition, TCL1 may also promote the translocation of activated Akt to the nucleus 168. Other proteins that interact with the PH domain of Akt include JNK-interacting protein 1 (JIP1), growth factor receptor- binding protein 10 (Grb10), and Ras GTPase-activating protein (RasGAP), all of which appear to increase the activity of Akt 169-171. Hsp90 is a molecular chaperone that facilitates protein folding and protects proteins from degradation during stress. Hsp90 forms together with the co-chaperone Cdc37 a complex that binds a variety of cellular proteins, including tyrosine and serine/threonine kinases. One of these kinases is Akt, which interacts with the Hsp90/Cdc37 complex via its kinase domain 172. Since the Hsp90/Cdc37 complex binds to many proteins, this may facilitate the interaction of Akt with some of its substrates, such as eNOS 173. Homolog 3 of Drosophila Tribbles (TRB3) is a protein that interacts with the activation loop in the kinase domain of Akt and thereby inhibits both the serine and threonine phosphorylations of Akt that are required for its activation 174. TRB3 has also been shown to interact with the RelA subunit of NF-B, and downregulation of TRB3 by RNAi was shown to inhibit TNF-induced apoptosis of 293 cells 175. Carboxyterminal modulator protein (CTMP) inhibits the phosphorylation of Akt by interacting with the carboxyterminal regulatory tail of Akt. Unphosphorylated CTMP is membrane-associated and inactivates bound Akt at the plasma membrane. Stimulation of cells promotes the phosphorylation of CTMP and the release and subsequent activation of Akt 176. Another negative regulator of Akt is Keratin-10, a major component of intermediate filaments of the cytoskeleton. Binding of Keratin-10 sequestrates Akt to the cytoskeleton and inhibits its intracellular translocation 177.

106 Chapter 3

Fig. 3.6. Interacting proteins that regulate the activity of Akt. The interacting proteins are listed under the domain of Akt they interact with. * The domain of Akt that Keratin-10 interacts with is currently unknown. Adapted from Du and Tsichlis, 2005 165.

To determine the role of the individual Akt isoforms, Akt isoform-specific knockout and double knockout mice have been generated. Akt-1 knockout mice display increased perinatal mortality and a 20-30% reduction in body weight. MEFs isolated from Akt-1-deficient mice are more susceptible to apoptosis induced by TNF or other stimuli 178. Akt-1 knockout mice also show impaired platelet aggregation and responses, increased pathological angiogenesis and impaired maturation of new blood vessels 179, 180. Consistent with this vascular phenotype, Akt-1 was shown to be the main isoform expressed in endothelial cells 180. Akt-2-deficient mice have normal growth, but develop a diabetes-like syndrome with insulin resistance, hyperinsulinemia, hyperglycemia and glucose intolerance, suggesting an important role of Akt-2 in glucose metabolism 181, 182. A point mutation of Akt-2 has also been found in one case of familial diabetes 183. Like Akt-1-deficient mice, Akt-2-deficient mice show defects in platelet aggregation and thrombus formation 184. Akt-3 knockout mice are viable but exhibit a reduction in brain size due to decreased neuronal cell size and numbers 185, 186. Mice lacking both Akt-1 and Akt-2 die shortly after birth, display severe growth deficiencies and fail to develop normal skin, skeletal muscle, bones, and fat tissue 187. Mice lacking both Akt-1 and Akt-3 die around embryonic day 12, with severe impairments in growth, cardiovascular development, and organization of the nervous system 188. In contrast, mice deficient for both Akt-2 and Akt-3 are viable, but display a substantial reduction in body weight and organ size, and severe defects in glucose metabolism. Even mice lacking both Akt-2 and Akt-3 that have only one functional allele for Akt-1 were shown to be viable, demonstrating that minimal amounts of Akt-1 are sufficient for embryonic development and postnatal survival 189.

Introduction: The PI3K/Akt pathway 107

3.3. mTOR

Besides cell survival, the PI3K/Akt pathway is also a major regulator of mRNA translation, through activation of mTOR. Akt activates mTOR by inactivation of the tuberous sclerosis complex (TSC), which negatively regulates mTOR kinase activity. The TSC is formed by two proteins, called TSC1 (or hamartin) and TSC2 (or tuberin). Akt directly phosphorylates TSC2, which disrupts the complex with TSC1 and leads to proteasome-mediated degradation of this protein 127, 190, 191. Interestingly, also MK2 has been shown to phosphorylate TSC2, which creates a 14-3-3 binding site 192. The heterodimeric TSC complex negatively regulates mTOR by inhibiting Ras homolog enriched in brain (Rheb), a small GTPase required for the activation of mTOR 193-195. TSC2 has a GTPase-activating protein (GAP) activity for Rheb and stimulates the conversion of the active GTP-bound form of Rheb to the inactive GDP-bound form 196. Rheb interacts with mTOR and GL, and activates the kinase activity of mTOR via a still unknown mechanism (Fig.3.7.) 197. mTOR is a large serine/threonine kinase belonging to the PI3K-related kinase (PIKK) family, also called class IV PI3Ks. It contains from amino- to carboxyterminus: a protein-protein interaction domain containing 20 HEAT repeats, a focal adhesion targeting (FAT) domain, a FKB12-rapamycin-binding (FRB) domain, and a second carboxyterminal FAT domain. In the cell, mTOR is present in two distinct mTOR complexes (mTORCs), called mTORC1 and mTORC2. mTORC1 consists of three proteins: mTOR, GL (or mLST8), and regulatory associated protein of mTOR (Raptor). This complex largely controls translation and cell growth in response to nutrients. mTORC2 consists of mTOR, GL, rapamycin-insensitive companion of mTOR (Rictor), and SIN1. mTORC2 controls actin cytoskeleton dynamics, but has also been shown to be responsible for the activating serine phosphorylation of Akt (Fig.3.7.) 118, 198. Only mTORC1 is sensitive to direct inhibition by the drug rapamycin. Rapamycin binds within the cell with the protein FKBP12, and this complex binds to the FRB domain of mTORC1, but not of mTORC2. Consequently, acute rapamycin treatment does not lead to a decrease in Akt phosphorylation and activity 118. However, Sarbassov and co-workers demonstrated that prolonged exposure to rapamycin inhibits mTORC2 function by blocking the association of newly formed mTOR with Rictor and thus the formation of mTORC2 199. Therefore, prolonged treatment with rapamycin may inhibit Akt activity in certain cell types. This was recently shown to be the case in HUVECs stimulated with TNF, VEGF or insulin 200.

108 Chapter 3

Fig. 3.7. The mTOR pathway. See text for details. It is currently unclear whether Rheb also regulates the mTOR-Rictor complex (mTORC2). Adapted from Corradetti and Guan, 2006 198.

The best described function of mTORC1 is its regulation of translation by phosphorylating the key translation regulators S6K and eukaryotic initiation factor 4E (eIF4E) binding protein 1 (4EBP1) (Fig.3.7.). Under basal conditions, 4EBP1 and S6K are bound to, respectively, eIF4E on the mRNA 5’cap, and eIF3 in the translation pre-initiation complex; but upon stimulation, mTOR associates with eIF3 and phosphorylates S6K and 4EBP1, causing their dissociation from eIF3 and eIF4E 201. S6K then associates with PDK-1, which phosphorylates the activation loop of S6K, resulting in full activation 109. S6K subsequently phosphorylates the 40S ribosomal protein S6, the RNA-binding protein eIF4B and other targets. eIF4B can then associate with eIF3 and eIF4A, allowing the binding of eIF4G and the assembly of a functional translation initiation complex 201. eIF4G is a large scaffolding protein that interacts with many proteins. eIF4G is phosphorylated in a mTOR-dependent manner, although the function of this phosphorylation is still unclear 202. eIF4A is an ATP- dependent RNA helicase required to unwind regions of secondary structure in the mRNA 203. The translation initiation complex thus formed is particularly efficient in the translation of (1) mRNAs encoding proteins involved in ribosome biogenesis due to S6 phosphorylation, (2) mRNAs with 7-methyl guanine caps due to mTOR- Introduction: The PI3K/Akt pathway 109

dependent phosphorylation of 4E-BP1, and (3) mRNAs with long, complex 5′UTRs due to the recruitment of eIF4A 201, 204-207. In this way, mTOR may control up to 15% of the bulk translation of the cell 198. Translation regulation, especially through eIF4E, is directly linked to cell growth, via key growth promoting proteins such as c-Myc and cyclin D1 208. In addition to S6K and 4EBP1, mTOR also induces the activation of other proteins involved in translation and ribosome biogenesis. It was shown that TIF-1A, a transcription factor for RNA polymerase I, which is essential for the transcription of ribosomal RNAs, is activated in a mTOR-dependent manner 209. Eukaryotic elongation factor 2 (eEF2), a protein that mediates the translocation step of translation elongation in which the ribosome moves relative to the mRNA by one codon, is also regulated by mTOR. eEF2 is negatively regulated by eEF2 kinase. mTOR indirectly (partly through S6K) phosphorylates eEF2 kinase at several residues, thereby inactivating its kinase activity for eEF2 210-212. Recently it was shown that eEF2 kinase may also be involved in the activation of autophagy (see below) 213, 214.

Translation and ribosome biogenesis are among the most energetically intensive processes that a cell can perform 215. Therefore, these activities must be closely regulated during times of low energy (ATP) or nutrient (amino acid) availability. One mechanism by which low energy levels couple to the regulation of translation is via inhibition of mTOR. High amounts of AMP in the cell, which reflect a low energy status, activate AMP-activated protein kinase (AMPK). AMPK subsequently phosphorylates several targets to enhance catabolism and suppress anabolism, for example TSC2. AMPK-dependent phosphorylation of TSC2 enhances its GAP activity for Rheb and thus inhibits mTOR activity 216. Full activation of AMPK also requires phosphorylation of its activation loop by the serine/threonine kinase LKB1, which is a tumor suppressor that is mutated in the human Peutz-Jeghers syndrome 217, 218. The availability of amino acids also regulates mTOR activity. Particularly leucine appears to be effective in activating mTOR 219. Some recent reports have identified the class III PI3K hVps34 as a critical regulator of mTOR activity 17, 18, 220. hVps34 was shown to be activated by the presence of amino acids and subsequently activates mTOR and S6K downstream of TSC and Rheb, although the mechanism involved is still unclear. hVps34 also functions in the regulation of autophagy, and this process is negatively regulated by mTOR 221. Autophagy is a catabolic process involving the degradation of cellular components through the lysosomal machinery. Since this process likely supplies intracellular amino acids to maintain mTOR signaling, there is some logic in connecting these processes 222. Finally, many other forms of cellular stress, including hypoxia, ROS and osmotic stress, have been shown to regulate the activity of mTOR 198. 110 Chapter 3

3.4. References

1. Foster, F.M., Traer, C.J., Abraham, S.M. & Fry, M.J. The phosphoinositide (PI) 3-kinase family. Journal of cell science 116, 3037-3040 (2003). 2. Vanhaesebroeck, B. & Alessi, D.R. The PI3K-PDK1 connection: more than just a road to PKB. The Biochemical journal 346 Pt 3, 561-576 (2000). 3. Bellacosa, A., Testa, J.R., Staal, S.P. & Tsichlis, P.N. A retroviral oncogene, akt, encoding a serine-threonine kinase containing an SH2-like region. Science (New York, N.Y 254, 274-277 (1991). 4. Jones, P.F., Jakubowicz, T., Pitossi, F.J., Maurer, F. & Hemmings, B.A. Molecular cloning and identification of a serine/threonine protein kinase of the second-messenger subfamily. Proceedings of the National Academy of Sciences of the United States of America 88, 4171- 4175 (1991). 5. Coffer, P.J. & Woodgett, J.R. Molecular cloning and characterisation of a novel putative protein-serine kinase related to the cAMP-dependent and protein kinase C families. European journal of biochemistry / FEBS 201, 475-481 (1991). 6. Stein, R.C. Prospects for phosphoinositide 3-kinase inhibition as a cancer treatment. Endocrine-related cancer 8, 237-248 (2001). 7. Brown, R.A., Domin, J., Arcaro, A., Waterfield, M.D. & Shepherd, P.R. Insulin activates the alpha isoform of class II phosphoinositide 3-kinase. The Journal of biological chemistry 274, 14529-14532 (1999). 8. Turner, S.J., Domin, J., Waterfield, M.D., Ward, S.G. & Westwick, J. The CC chemokine monocyte chemotactic peptide-1 activates both the class I p85/p110 phosphatidylinositol 3- kinase and the class II PI3K-C2alpha. The Journal of biological chemistry 273, 25987-25995 (1998). 9. Arcaro, A. et al. Class II phosphoinositide 3-kinases are downstream targets of activated polypeptide growth factor receptors. Molecular and cellular biology 20, 3817-3830 (2000). 10. Zhang, J. et al. A type II phosphoinositide 3-kinase is stimulated via activated integrin in platelets. A source of phosphatidylinositol 3-phosphate. The Journal of biological chemistry 273, 14081-14084 (1998). 11. Ono, F. et al. A novel class II phosphoinositide 3-kinase predominantly expressed in the liver and its enhanced expression during liver regeneration. The Journal of biological chemistry 273, 7731-7736 (1998). 12. Misawa, H. et al. Cloning and characterization of a novel class II phosphoinositide 3-kinase containing C2 domain. Biochemical and biophysical research communications 244, 531-539 (1998). 13. Herman, P.K. & Emr, S.D. Characterization of VPS34, a gene required for vacuolar protein sorting and vacuole segregation in Saccharomyces cerevisiae. Molecular and cellular biology 10, 6742-6754 (1990). 14. Schu, P.V. et al. Phosphatidylinositol 3-kinase encoded by yeast VPS34 gene essential for protein sorting. Science (New York, N.Y 260, 88-91 (1993). 15. Wurmser, A.E., Gary, J.D. & Emr, S.D. Phosphoinositide 3-kinases and their FYVE domain- containing effectors as regulators of vacuolar/lysosomal membrane trafficking pathways. The Journal of biological chemistry 274, 9129-9132 (1999). 16. Backer, J.M. The regulation and function of Class III PI3Ks: novel roles for Vps34. The Biochemical journal 410, 1-17 (2008). 17. Byfield, M.P., Murray, J.T. & Backer, J.M. hVps34 is a nutrient-regulated lipid kinase required for activation of p70 S6 kinase. The Journal of biological chemistry 280, 33076- 33082 (2005). 18. Nobukuni, T. et al. Amino acids mediate mTOR/raptor signaling through activation of class 3 phosphatidylinositol 3OH-kinase. Proceedings of the National Academy of Sciences of the United States of America 102, 14238-14243 (2005). 19. Cantrell, D.A. Phosphoinositide 3-kinase signalling pathways. Journal of cell science 114, 1439-1445 (2001). 20. Dhand, R. et al. PI 3-kinase is a dual specificity enzyme: autoregulation by an intrinsic protein-serine kinase activity. The EMBO journal 13, 522-533 (1994). 21. Czupalla, C. et al. Identification and characterization of the autophosphorylation sites of phosphoinositide 3-kinase isoforms beta and gamma. The Journal of biological chemistry 278, 11536-11545 (2003). Introduction: The PI3K/Akt pathway 111

22. Vanhaesebroeck, B. et al. Autophosphorylation of p110delta phosphoinositide 3-kinase: a new paradigm for the regulation of lipid kinases in vitro and in vivo. The EMBO journal 18, 1292- 1302 (1999). 23. Foukas, L.C., Beeton, C.A., Jensen, J., Phillips, W.A. & Shepherd, P.R. Regulation of phosphoinositide 3-kinase by its intrinsic serine kinase activity in vivo. Molecular and cellular biology 24, 966-975 (2004). 24. Pons, S. et al. The structure and function of p55PIK reveal a new regulatory subunit for phosphatidylinositol 3-kinase. Molecular and cellular biology 15, 4453-4465 (1995). 25. Chantry, D. et al. p110delta, a novel phosphatidylinositol 3-kinase catalytic subunit that associates with p85 and is expressed predominantly in leukocytes. The Journal of biological chemistry 272, 19236-19241 (1997). 26. Vanhaesebroeck, B. et al. P110delta, a novel phosphoinositide 3-kinase in leukocytes. Proceedings of the National Academy of Sciences of the United States of America 94, 4330- 4335 (1997). 27. Otsu, M. et al. Characterization of two 85 kd proteins that associate with receptor tyrosine kinases, middle-T/pp60c-src complexes, and PI3-kinase. Cell 65, 91-104 (1991). 28. Antonetti, D.A., Algenstaedt, P. & Kahn, C.R. Insulin receptor substrate 1 binds two novel splice variants of the regulatory subunit of phosphatidylinositol 3-kinase in muscle and brain. Molecular and cellular biology 16, 2195-2203 (1996). 29. Dey, B.R., Furlanetto, R.W. & Nissley, S.P. Cloning of human p55 gamma, a regulatory subunit of phosphatidylinositol 3-kinase, by a yeast two-hybrid library screen with the insulin- like growth factor-I receptor. Gene 209, 175-183 (1998). 30. Vanhaesebroeck, B., Ali, K., Bilancio, A., Geering, B. & Foukas, L.C. Signalling by PI3K isoforms: insights from gene-targeted mice. Trends in biochemical sciences 30, 194-204 (2005). 31. Escobedo, J.A. et al. cDNA cloning of a novel 85 kd protein that has SH2 domains and regulates binding of PI3-kinase to the PDGF beta-receptor. Cell 65, 75-82 (1991). 32. Escobedo, J.A., Kaplan, D.R., Kavanaugh, W.M., Turck, C.W. & Williams, L.T. A phosphatidylinositol-3 kinase binds to platelet-derived growth factor receptors through a specific receptor sequence containing phosphotyrosine. Molecular and cellular biology 11, 1125-1132 (1991). 33. Parker, P.J. & Waterfield, M.D. Phosphatidylinositol 3-kinase: a novel effector. Cell Growth Differ 3, 747-752 (1992). 34. Klippel, A. et al. Membrane localization of phosphatidylinositol 3-kinase is sufficient to activate multiple signal-transducing kinase pathways. Molecular and cellular biology 16, 4117-4127 (1996). 35. Kodaki, T. et al. The activation of phosphatidylinositol 3-kinase by Ras. Curr Biol 4, 798-806 (1994). 36. Jimenez, C., Hernandez, C., Pimentel, B. & Carrera, A.C. The p85 regulatory subunit controls sequential activation of phosphoinositide 3-kinase by Tyr kinases and Ras. The Journal of biological chemistry 277, 41556-41562 (2002). 37. Guillermet-Guibert, J. et al. The p110beta isoform of phosphoinositide 3-kinase signals downstream of G protein-coupled receptors and is functionally redundant with p110gamma. Proceedings of the National Academy of Sciences of the United States of America 105, 8292- 8297 (2008). 38. Suire, S. et al. p84, a new Gbetagamma-activated regulatory subunit of the type IB phosphoinositide 3-kinase p110gamma. Curr Biol 15, 566-570 (2005). 39. Stephens, L.R. et al. The G beta gamma sensitivity of a PI3K is dependent upon a tightly associated adaptor, p101. Cell 89, 105-114 (1997). 40. Crackower, M.A. et al. Regulation of myocardial contractility and cell size by distinct PI3K- PTEN signaling pathways. Cell 110, 737-749 (2002). 41. Puri, K.D. et al. The role of endothelial PI3Kgamma activity in neutrophil trafficking. Blood 106, 150-157 (2005). 42. Wymann, M.P. et al. Wortmannin inactivates phosphoinositide 3-kinase by covalent modification of Lys-802, a residue involved in the phosphate transfer reaction. Molecular and cellular biology 16, 1722-1733 (1996). 43. Marone, R., Cmiljanovic, V., Giese, B. & Wymann, M.P. Targeting phosphoinositide 3- kinase-Moving towards therapy. Biochimica et biophysica acta 1784, 159-185 (2008). 44. Graupera, M. et al. Angiogenesis selectively requires the p110alpha isoform of PI3K to control endothelial cell migration. Nature 453, 662-666 (2008). 112 Chapter 3

45. Damen, J.E. et al. The 145-kDa protein induced to associate with Shc by multiple cytokines is an inositol tetraphosphate and phosphatidylinositol 3,4,5-triphosphate 5-phosphatase. Proceedings of the National Academy of Sciences of the United States of America 93, 1689- 1693 (1996). 46. Rauh, M.J. et al. Role of Src homology 2-containing-inositol 5'-phosphatase (SHIP) in mast cells and macrophages. Biochemical Society transactions 31, 286-291 (2003). 47. Helgason, C.D. et al. Targeted disruption of SHIP leads to hemopoietic perturbations, lung pathology, and a shortened life span. Genes & development 12, 1610-1620 (1998). 48. Cantley, L.C. & Neel, B.G. New insights into tumor suppression: PTEN suppresses tumor formation by restraining the phosphoinositide 3-kinase/AKT pathway. Proceedings of the National Academy of Sciences of the United States of America 96, 4240-4245 (1999). 49. Liliental, J. et al. Genetic deletion of the Pten tumor suppressor gene promotes cell motility by activation of Rac1 and Cdc42 GTPases. Curr Biol 10, 401-404 (2000). 50. Shan, X. et al. Deficiency of PTEN in Jurkat T cells causes constitutive localization of Itk to the plasma membrane and hyperresponsiveness to CD3 stimulation. Molecular and cellular biology 20, 6945-6957 (2000). 51. Gustafson, S., Zbuk, K.M., Scacheri, C. & Eng, C. Cowden syndrome. Seminars in oncology 34, 428-434 (2007). 52. Liu, W., James, C.D., Frederick, L., Alderete, B.E. & Jenkins, R.B. PTEN/MMAC1 mutations and EGFR amplification in glioblastomas. Cancer research 57, 5254-5257 (1997). 53. Rasheed, B.K. et al. PTEN gene mutations are seen in high-grade but not in low-grade gliomas. Cancer research 57, 4187-4190 (1997). 54. Wang, S.I. et al. Somatic mutations of PTEN in glioblastoma multiforme. Cancer research 57, 4183-4186 (1997). 55. Bostrom, J. et al. Mutation of the PTEN (MMAC1) tumor suppressor gene in a subset of glioblastomas but not in meningiomas with loss of chromosome arm 10q. Cancer research 58, 29-33 (1998). 56. Cairns, P. et al. Frequent inactivation of PTEN/MMAC1 in primary prostate cancer. Cancer research 57, 4997-5000 (1997). 57. Risinger, J.I., Hayes, A.K., Berchuck, A. & Barrett, J.C. PTEN/MMAC1 mutations in endometrial cancers. Cancer research 57, 4736-4738 (1997). 58. Tashiro, H. et al. Mutations in PTEN are frequent in endometrial carcinoma but rare in other common gynecological malignancies. Cancer research 57, 3935-3940 (1997). 59. Shi, W. et al. Dysregulated PTEN-PKB and negative receptor status in human breast cancer. International journal of cancer 104, 195-203 (2003). 60. Liaw, D. et al. Germline mutations of the PTEN gene in Cowden disease, an inherited breast and thyroid cancer syndrome. Nature genetics 16, 64-67 (1997). 61. Ueda, K. et al. Infrequent mutations in the PTEN/MMAC1 gene among primary breast cancers. Jpn J Cancer Res 89, 17-21 (1998). 62. Okami, K. et al. Analysis of PTEN/MMAC1 alterations in aerodigestive tract tumors. Cancer research 58, 509-511 (1998). 63. Dahia, P.L. et al. Somatic deletions and mutations in the Cowden disease gene, PTEN, in sporadic thyroid tumors. Cancer research 57, 4710-4713 (1997). 64. Tsao, H., Zhang, X., Benoit, E. & Haluska, F.G. Identification of PTEN/MMAC1 alterations in uncultured melanomas and melanoma cell lines. Oncogene 16, 3397-3402 (1998). 65. Guldberg, P. et al. Disruption of the MMAC1/PTEN gene by deletion or mutation is a frequent event in malignant melanoma. Cancer research 57, 3660-3663 (1997). 66. Stambolic, V. et al. Regulation of PTEN transcription by p53. Molecular cell 8, 317-325 (2001). 67. Singh, B. et al. p53 regulates cell survival by inhibiting PIK3CA in squamous cell carcinomas. Genes & development 16, 984-993 (2002). 68. Bachman, K.E. et al. The PIK3CA gene is mutated with high frequency in human breast cancers. Cancer biology & therapy 3, 772-775 (2004). 69. Campbell, I.G. et al. Mutation of the PIK3CA gene in ovarian and breast cancer. Cancer research 64, 7678-7681 (2004). 70. Levine, D.A. et al. Frequent mutation of the PIK3CA gene in ovarian and breast cancers. Clin Cancer Res 11, 2875-2878 (2005). 71. Samuels, Y. et al. High frequency of mutations of the PIK3CA gene in human cancers. Science (New York, N.Y 304, 554 (2004). Introduction: The PI3K/Akt pathway 113

72. Zhao, L. & Vogt, P.K. Helical domain and kinase domain mutations in p110{alpha} of phosphatidylinositol 3-kinase induce gain of function by different mechanisms. Proceedings of the National Academy of Sciences of the United States of America (2008). 73. Bi, L., Okabe, I., Bernard, D.J., Wynshaw-Boris, A. & Nussbaum, R.L. Proliferative defect and embryonic lethality in mice homozygous for a deletion in the p110alpha subunit of phosphoinositide 3-kinase. The Journal of biological chemistry 274, 10963-10968 (1999). 74. Foukas, L.C. et al. Critical role for the p110alpha phosphoinositide-3-OH kinase in growth and metabolic regulation. Nature 441, 366-370 (2006). 75. Bi, L., Okabe, I., Bernard, D.J. & Nussbaum, R.L. Early embryonic lethality in mice deficient in the p110beta catalytic subunit of PI 3-kinase. Mamm Genome 13, 169-172 (2002). 76. Brachmann, S.M., Ueki, K., Engelman, J.A., Kahn, R.C. & Cantley, L.C. Phosphoinositide 3- kinase catalytic subunit deletion and regulatory subunit deletion have opposite effects on insulin sensitivity in mice. Molecular and cellular biology 25, 1596-1607 (2005). 77. MacDonald, P.E. et al. Impaired glucose-stimulated insulin secretion, enhanced intraperitoneal insulin tolerance, and increased beta-cell mass in mice lacking the p110gamma isoform of phosphoinositide 3-kinase. Endocrinology 145, 4078-4083 (2004). 78. Sasaki, T. et al. Function of PI3Kgamma in thymocyte development, T cell activation, and neutrophil migration. Science (New York, N.Y 287, 1040-1046 (2000). 79. Laffargue, M. et al. Phosphoinositide 3-kinase gamma is an essential amplifier of mast cell function. Immunity 16, 441-451 (2002). 80. Del Prete, A. et al. Defective dendritic cell migration and activation of adaptive immunity in PI3Kgamma-deficient mice. The EMBO journal 23, 3505-3515 (2004). 81. Hirsch, E. et al. Central role for G protein-coupled phosphoinositide 3-kinase gamma in inflammation. Science (New York, N.Y 287, 1049-1053 (2000). 82. Camps, M. et al. Blockade of PI3Kgamma suppresses joint inflammation and damage in mouse models of rheumatoid arthritis. Nature medicine 11, 936-943 (2005). 83. Alloatti, G. et al. Phosphoinositide 3-kinase gamma-deficient hearts are protected from the PAF-dependent depression of cardiac contractility. Cardiovascular research 60, 242-249 (2003). 84. Patrucco, E. et al. PI3Kgamma modulates the cardiac response to chronic pressure overload by distinct kinase-dependent and -independent effects. Cell 118, 375-387 (2004). 85. Clayton, E. et al. A crucial role for the p110delta subunit of phosphatidylinositol 3-kinase in B cell development and activation. The Journal of experimental medicine 196, 753-763 (2002). 86. Jou, S.T. et al. Essential, nonredundant role for the phosphoinositide 3-kinase p110delta in signaling by the B-cell receptor complex. Molecular and cellular biology 22, 8580-8591 (2002). 87. Puri, K.D. et al. Mechanisms and implications of phosphoinositide 3-kinase delta in promoting neutrophil trafficking into inflamed tissue. Blood 103, 3448-3456 (2004). 88. Okkenhaug, K. et al. Impaired B and T cell antigen receptor signaling in p110delta PI 3- kinase mutant mice. Science (New York, N.Y 297, 1031-1034 (2002). 89. Ali, K. et al. Essential role for the p110delta phosphoinositide 3-kinase in the allergic response. Nature 431, 1007-1011 (2004). 90. Bilancio, A. et al. Key role of the p110delta isoform of PI3K in B-cell antigen and IL-4 receptor signaling: comparative analysis of genetic and pharmacologic interference with p110delta function in B cells. Blood 107, 642-650 (2006). 91. Shi, Y.Q., Blattmann, H. & Crompton, N.E. Wortmannin selectively enhances radiation- induced apoptosis in proliferative but not quiescent cells. International journal of radiation oncology, biology, physics 49, 421-425 (2001). 92. Tan, J. & Hallahan, D.E. Growth factor-independent activation of protein kinase B contributes to the inherent resistance of vascular endothelium to radiation-induced apoptotic response. Cancer research 63, 7663-7667 (2003). 93. Westfall, S.D. & Skinner, M.K. Inhibition of phosphatidylinositol 3-kinase sensitizes ovarian cancer cells to carboplatin and allows adjunct chemotherapy treatment. Molecular cancer therapeutics 4, 1764-1771 (2005). 94. Rychahou, P.G., Murillo, C.A. & Evers, B.M. Targeted RNA interference of PI3K pathway components sensitizes colon cancer cells to TNF-related apoptosis-inducing ligand (TRAIL). Surgery 138, 391-397 (2005). 95. Hatano, E. & Brenner, D.A. Akt protects mouse hepatocytes from TNF-alpha- and Fas- mediated apoptosis through NK-kappa B activation. American journal of physiology 281, G1357-1368 (2001). 114 Chapter 3

96. Zhang, L., Himi, T., Morita, I. & Murota, S. Inhibition of phosphatidylinositol-3 kinase/Akt or mitogen-activated protein kinase signaling sensitizes endothelial cells to TNF-alpha cytotoxicity. Cell death and differentiation 8, 528-536 (2001). 97. Clermont, F., Adam, E., Dumont, J.E. & Robaye, B. Survival pathways regulating the apoptosis induced by tumour necrosis factor-alpha in primary cultured bovine endothelial cells. Cellular signalling 15, 539-546 (2003). 98. Madge, L.A., Li, J.H., Choi, J. & Pober, J.S. Inhibition of phosphatidylinositol 3-kinase sensitizes vascular endothelial cells to cytokine-initiated cathepsin-dependent apoptosis. The Journal of biological chemistry 278, 21295-21306 (2003). 99. Madge, L.A. & Pober, J.S. A phosphatidylinositol 3-kinase/Akt pathway, activated by tumor necrosis factor or interleukin-1, inhibits apoptosis but does not activate NFkappaB in human endothelial cells. The Journal of biological chemistry 275, 15458-15465 (2000). 100. Matschurat, S. et al. Negative regulatory role of PI3-kinase in TNF-induced tumor necrosis. International journal of cancer 107, 30-37 (2003). 101. Song, G., Ouyang, G. & Bao, S. The activation of Akt/PKB signaling pathway and cell survival. Journal of cellular and molecular medicine 9, 59-71 (2005). 102. James, S.R. et al. Specific binding of the Akt-1 protein kinase to phosphatidylinositol 3,4,5- trisphosphate without subsequent activation. The Biochemical journal 315 ( Pt 3), 709-713 (1996). 103. Andjelkovic, M. et al. Role of translocation in the activation and function of protein kinase B. The Journal of biological chemistry 272, 31515-31524 (1997). 104. Alessi, D.R. et al. Characterization of a 3-phosphoinositide-dependent protein kinase which phosphorylates and activates protein kinase Balpha. Curr Biol 7, 261-269 (1997). 105. Milburn, C.C. et al. Binding of phosphatidylinositol 3,4,5-trisphosphate to the pleckstrin homology domain of protein kinase B induces a conformational change. The Biochemical journal 375, 531-538 (2003). 106. Anderson, K.E., Coadwell, J., Stephens, L.R. & Hawkins, P.T. Translocation of PDK-1 to the plasma membrane is important in allowing PDK-1 to activate protein kinase B. Curr Biol 8, 684-691 (1998). 107. Currie, R.A. et al. Role of phosphatidylinositol 3,4,5-trisphosphate in regulating the activity and localization of 3-phosphoinositide-dependent protein kinase-1. The Biochemical journal 337 ( Pt 3), 575-583 (1999). 108. Dutil, E.M., Toker, A. & Newton, A.C. Regulation of conventional protein kinase C isozymes by phosphoinositide-dependent kinase 1 (PDK-1). Curr Biol 8, 1366-1375 (1998). 109. Alessi, D.R., Kozlowski, M.T., Weng, Q.P., Morrice, N. & Avruch, J. 3-Phosphoinositide- dependent protein kinase 1 (PDK1) phosphorylates and activates the p70 S6 kinase in vivo and in vitro. Curr Biol 8, 69-81 (1998). 110. Alessi, D.R. et al. Mechanism of activation of protein kinase B by insulin and IGF-1. The EMBO journal 15, 6541-6551 (1996). 111. Persad, S. et al. Regulation of protein kinase B/Akt-serine 473 phosphorylation by integrin- linked kinase: critical roles for kinase activity and amino acids arginine 211 and serine 343. The Journal of biological chemistry 276, 27462-27469 (2001). 112. Balendran, A. et al. PDK1 acquires PDK2 activity in the presence of a synthetic peptide derived from the carboxyl terminus of PRK2. Curr Biol 9, 393-404 (1999). 113. Toker, A. & Newton, A.C. Akt/protein kinase B is regulated by autophosphorylation at the hypothetical PDK-2 site. The Journal of biological chemistry 275, 8271-8274 (2000). 114. Feng, J., Park, J., Cron, P., Hess, D. & Hemmings, B.A. Identification of a PKB/Akt hydrophobic motif Ser-473 kinase as DNA-dependent protein kinase. The Journal of biological chemistry 279, 41189-41196 (2004). 115. Partovian, C. & Simons, M. Regulation of protein kinase B/Akt activity and Ser473 phosphorylation by protein kinase Calpha in endothelial cells. Cellular signalling 16, 951-957 (2004). 116. Kawakami, Y. et al. Protein kinase C betaII regulates Akt phosphorylation on Ser-473 in a cell type- and stimulus-specific fashion. The Journal of biological chemistry 279, 47720-47725 (2004). 117. Viniegra, J.G. et al. Full activation of PKB/Akt in response to insulin or ionizing radiation is mediated through ATM. The Journal of biological chemistry 280, 4029-4036 (2005). 118. Sarbassov, D.D., Guertin, D.A., Ali, S.M. & Sabatini, D.M. Phosphorylation and regulation of Akt/PKB by the rictor-mTOR complex. Science (New York, N.Y 307, 1098-1101 (2005). Introduction: The PI3K/Akt pathway 115

119. Hresko, R.C. & Mueckler, M. mTOR.RICTOR is the Ser473 kinase for Akt/protein kinase B in 3T3-L1 adipocytes. The Journal of biological chemistry 280, 40406-40416 (2005). 120. Jacinto, E. et al. SIN1/MIP1 maintains rictor-mTOR complex integrity and regulates Akt phosphorylation and substrate specificity. Cell 127, 125-137 (2006). 121. Meier, R., Alessi, D.R., Cron, P., Andjelkovic, M. & Hemmings, B.A. Mitogenic activation, phosphorylation, and nuclear translocation of protein kinase Bbeta. The Journal of biological chemistry 272, 30491-30497 (1997). 122. Meier, R., Thelen, M. & Hemmings, B.A. Inactivation and dephosphorylation of protein kinase Balpha (PKBalpha) promoted by hyperosmotic stress. The EMBO journal 17, 7294- 7303 (1998). 123. Datta, S.R. et al. Akt phosphorylation of BAD couples survival signals to the cell-intrinsic death machinery. Cell 91, 231-241 (1997). 124. Brunet, A. et al. Akt promotes cell survival by phosphorylating and inhibiting a Forkhead transcription factor. Cell 96, 857-868 (1999). 125. Kops, G.J. et al. Direct control of the Forkhead transcription factor AFX by protein kinase B. Nature 398, 630-634 (1999). 126. Rena, G., Guo, S., Cichy, S.C., Unterman, T.G. & Cohen, P. Phosphorylation of the transcription factor forkhead family member FKHR by protein kinase B. The Journal of biological chemistry 274, 17179-17183 (1999). 127. Plas, D.R. & Thompson, C.B. Akt activation promotes degradation of tuberin and FOXO3a via the proteasome. The Journal of biological chemistry 278, 12361-12366 (2003). 128. Cardone, M.H. et al. Regulation of cell death protease caspase-9 by phosphorylation. Science (New York, N.Y 282, 1318-1321 (1998). 129. Fujita, E. et al. Akt phosphorylation site found in human caspase-9 is absent in mouse caspase-9. Biochemical and biophysical research communications 264, 550-555 (1999). 130. Datta, S.R., Brunet, A. & Greenberg, M.E. Cellular survival: a play in three Akts. Genes & development 13, 2905-2927 (1999). 131. Biggs, W.H., 3rd, Meisenhelder, J., Hunter, T., Cavenee, W.K. & Arden, K.C. Protein kinase B/Akt-mediated phosphorylation promotes nuclear exclusion of the winged helix transcription factor FKHR1. Proceedings of the National Academy of Sciences of the United States of America 96, 7421-7426 (1999). 132. Aoki, M., Jiang, H. & Vogt, P.K. Proteasomal degradation of the FoxO1 transcriptional regulator in cells transformed by the P3k and Akt oncoproteins. Proceedings of the National Academy of Sciences of the United States of America 101, 13613-13617 (2004). 133. Modur, V., Nagarajan, R., Evers, B.M. & Milbrandt, J. FOXO proteins regulate tumor necrosis factor-related apoptosis inducing ligand expression. Implications for PTEN mutation in prostate cancer. The Journal of biological chemistry 277, 47928-47937 (2002). 134. Dijkers, P.F., Medema, R.H., Lammers, J.W., Koenderman, L. & Coffer, P.J. Expression of the pro-apoptotic Bcl-2 family member Bim is regulated by the forkhead transcription factor FKHR-L1. Curr Biol 10, 1201-1204 (2000). 135. Rokudai, S., Fujita, N., Kitahara, O., Nakamura, Y. & Tsuruo, T. Involvement of FKHR- dependent TRADD expression in chemotherapeutic drug-induced apoptosis. Molecular and cellular biology 22, 8695-8708 (2002). 136. Tang, T.T. et al. The forkhead transcription factor AFX activates apoptosis by induction of the BCL-6 transcriptional repressor. The Journal of biological chemistry 277, 14255-14265 (2002). 137. Schmidt, M. et al. Cell cycle inhibition by FoxO forkhead transcription factors involves downregulation of cyclin D. Molecular and cellular biology 22, 7842-7852 (2002). 138. Medema, R.H., Kops, G.J., Bos, J.L. & Burgering, B.M. AFX-like Forkhead transcription factors mediate cell-cycle regulation by Ras and PKB through p27kip1. Nature 404, 782-787 (2000). 139. Dijkers, P.F. et al. Forkhead transcription factor FKHR-L1 modulates cytokine-dependent transcriptional regulation of p27(KIP1). Molecular and cellular biology 20, 9138-9148 (2000). 140. Shin, I. et al. PKB/Akt mediates cell-cycle progression by phosphorylation of p27(Kip1) at threonine 157 and modulation of its cellular localization. Nature medicine 8, 1145-1152 (2002). 141. Li, Y., Dowbenko, D. & Lasky, L.A. AKT/PKB phosphorylation of p21Cip/WAF1 enhances protein stability of p21Cip/WAF1 and promotes cell survival. The Journal of biological chemistry 277, 11352-11361 (2002). 116 Chapter 3

142. Mayo, L.D. & Donner, D.B. A phosphatidylinositol 3-kinase/Akt pathway promotes translocation of Mdm2 from the cytoplasm to the nucleus. Proceedings of the National Academy of Sciences of the United States of America 98, 11598-11603 (2001). 143. Ogawara, Y. et al. Akt enhances Mdm2-mediated ubiquitination and degradation of p53. The Journal of biological chemistry 277, 21843-21850 (2002). 144. Kim, A.H., Khursigara, G., Sun, X., Franke, T.F. & Chao, M.V. Akt phosphorylates and negatively regulates apoptosis signal-regulating kinase 1. Molecular and cellular biology 21, 893-901 (2001). 145. Barthwal, M.K. et al. Negative regulation of mixed lineage kinase 3 by protein kinase B/AKT leads to cell survival. The Journal of biological chemistry 278, 3897-3902 (2003). 146. Park, H.S. et al. Akt (protein kinase B) negatively regulates SEK1 by means of protein phosphorylation. The Journal of biological chemistry 277, 2573-2578 (2002). 147. Du, K. & Montminy, M. CREB is a regulatory target for the protein kinase Akt/PKB. The Journal of biological chemistry 273, 32377-32379 (1998). 148. Pugazhenthi, S. et al. Akt/protein kinase B up-regulates Bcl-2 expression through cAMP- response element-binding protein. The Journal of biological chemistry 275, 10761-10766 (2000). 149. Sizemore, N., Leung, S. & Stark, G.R. Activation of phosphatidylinositol 3-kinase in response to interleukin-1 leads to phosphorylation and activation of the NF-kappaB p65/RelA subunit. Molecular and cellular biology 19, 4798-4805 (1999). 150. Sizemore, N., Lerner, N., Dombrowski, N., Sakurai, H. & Stark, G.R. Distinct roles of the Ikappa B kinase alpha and beta subunits in liberating nuclear factor kappa B (NF-kappa B) from Ikappa B and in phosphorylating the p65 subunit of NF-kappa B. The Journal of biological chemistry 277, 3863-3869 (2002). 151. Koul, D., Yao, Y., Abbruzzese, J.L., Yung, W.K. & Reddy, S.A. Tumor suppressor MMAC/PTEN inhibits cytokine-induced NFkappaB activation without interfering with the IkappaB degradation pathway. The Journal of biological chemistry 276, 11402-11408 (2001). 152. Kane, L.P., Mollenauer, M.N., Xu, Z., Turck, C.W. & Weiss, A. Akt-dependent phosphorylation specifically regulates Cot induction of NF-kappa B-dependent transcription. Molecular and cellular biology 22, 5962-5974 (2002). 153. Ozes, O.N. et al. NF-kappaB activation by tumour necrosis factor requires the Akt serine- threonine kinase. Nature 401, 82-85 (1999). 154. Romashkova, J.A. & Makarov, S.S. NF-kappaB is a target of AKT in anti-apoptotic PDGF signalling. Nature 401, 86-90 (1999). 155. Gustin, J.A. et al. Cell type-specific expression of the IkappaB kinases determines the significance of phosphatidylinositol 3-kinase/Akt signaling to NF-kappa B activation. The Journal of biological chemistry 279, 1615-1620 (2004). 156. Cross, D.A., Alessi, D.R., Cohen, P., Andjelkovich, M. & Hemmings, B.A. Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B. Nature 378, 785-789 (1995). 157. Pap, M. & Cooper, G.M. Role of glycogen synthase kinase-3 in the phosphatidylinositol 3- Kinase/Akt cell survival pathway. The Journal of biological chemistry 273, 19929-19932 (1998). 158. Pap, M. & Cooper, G.M. Role of translation initiation factor 2B in control of cell survival by the phosphatidylinositol 3-kinase/Akt/glycogen synthase kinase 3beta signaling pathway. Molecular and cellular biology 22, 578-586 (2002). 159. Juhaszova, M. et al. Glycogen synthase kinase-3beta mediates convergence of protection signaling to inhibit the mitochondrial permeability transition pore. The Journal of clinical investigation 113, 1535-1549 (2004). 160. Tan, J., Geng, L., Yazlovitskaya, E.M. & Hallahan, D.E. Protein kinase B/Akt-dependent phosphorylation of glycogen synthase kinase-3beta in irradiated vascular endothelium. Cancer research 66, 2320-2327 (2006). 161. Bellacosa, A. et al. Molecular alterations of the AKT2 oncogene in ovarian and breast carcinomas. International journal of cancer 64, 280-285 (1995). 162. Cheng, J.Q. et al. Amplification of AKT2 in human pancreatic cells and inhibition of AKT2 expression and tumorigenicity by antisense RNA. Proceedings of the National Academy of Sciences of the United States of America 93, 3636-3641 (1996). 163. Nakatani, K. et al. Up-regulation of Akt3 in estrogen receptor-deficient breast cancers and androgen-independent prostate cancer lines. The Journal of biological chemistry 274, 21528- 21532 (1999). Introduction: The PI3K/Akt pathway 117

164. Kondapaka, S.B., Singh, S.S., Dasmahapatra, G.P., Sausville, E.A. & Roy, K.K. Perifosine, a novel alkylphospholipid, inhibits protein kinase B activation. Molecular cancer therapeutics 2, 1093-1103 (2003). 165. Du, K. & Tsichlis, P.N. Regulation of the Akt kinase by interacting proteins. Oncogene 24, 7401-7409 (2005). 166. Laine, J., Kunstle, G., Obata, T. & Noguchi, M. Differential regulation of Akt kinase isoforms by the members of the TCL1 oncogene family. The Journal of biological chemistry 277, 3743- 3751 (2002). 167. Laine, J., Kunstle, G., Obata, T., Sha, M. & Noguchi, M. The protooncogene TCL1 is an Akt kinase coactivator. Molecular cell 6, 395-407 (2000). 168. Pekarsky, Y. et al. Tcl1 enhances Akt kinase activity and mediates its nuclear translocation. Proceedings of the National Academy of Sciences of the United States of America 97, 3028- 3033 (2000). 169. Kim, A.H., Sasaki, T. & Chao, M.V. JNK-interacting protein 1 promotes Akt1 activation. The Journal of biological chemistry 278, 29830-29836 (2003). 170. Jahn, T., Seipel, P., Urschel, S., Peschel, C. & Duyster, J. Role for the adaptor protein Grb10 in the activation of Akt. Molecular and cellular biology 22, 979-991 (2002). 171. Yue, Y., Lypowy, J., Hedhli, N. & Abdellatif, M. Ras GTPase-activating protein binds to Akt and is required for its activation. The Journal of biological chemistry 279, 12883-12889 (2004). 172. Basso, A.D. et al. Akt forms an intracellular complex with heat shock protein 90 (Hsp90) and Cdc37 and is destabilized by inhibitors of Hsp90 function. The Journal of biological chemistry 277, 39858-39866 (2002). 173. Fontana, J. et al. Domain mapping studies reveal that the M domain of hsp90 serves as a molecular scaffold to regulate Akt-dependent phosphorylation of endothelial nitric oxide synthase and NO release. Circulation research 90, 866-873 (2002). 174. Du, K., Herzig, S., Kulkarni, R.N. & Montminy, M. TRB3: a tribbles homolog that inhibits Akt/PKB activation by insulin in liver. Science (New York, N.Y 300, 1574-1577 (2003). 175. Wu, M., Xu, L.G., Zhai, Z. & Shu, H.B. SINK is a p65-interacting negative regulator of NF- kappaB-dependent transcription. The Journal of biological chemistry 278, 27072-27079 (2003). 176. Maira, S.M. et al. Carboxyl-terminal modulator protein (CTMP), a negative regulator of PKB/Akt and v-Akt at the plasma membrane. Science (New York, N.Y 294, 374-380 (2001). 177. Paramio, J.M., Segrelles, C., Ruiz, S. & Jorcano, J.L. Inhibition of protein kinase B (PKB) and PKCzeta mediates keratin K10-induced cell cycle arrest. Molecular and cellular biology 21, 7449-7459 (2001). 178. Chen, W.S. et al. Growth retardation and increased apoptosis in mice with homozygous disruption of the Akt1 gene. Genes & development 15, 2203-2208 (2001). 179. Chen, J. et al. Impaired platelet responses to thrombin and collagen in AKT-1-deficient mice. Blood 104, 1703-1710 (2004). 180. Chen, J. et al. Akt1 regulates pathological angiogenesis, vascular maturation and permeability in vivo. Nature medicine 11, 1188-1196 (2005). 181. Cho, H. et al. Insulin resistance and a diabetes mellitus-like syndrome in mice lacking the protein kinase Akt2 (PKB beta). Science (New York, N.Y 292, 1728-1731 (2001). 182. Garofalo, R.S. et al. Severe diabetes, age-dependent loss of adipose tissue, and mild growth deficiency in mice lacking Akt2/PKB beta. The Journal of clinical investigation 112, 197-208 (2003). 183. George, S. et al. A family with severe insulin resistance and diabetes due to a mutation in AKT2. Science (New York, N.Y 304, 1325-1328 (2004). 184. Woulfe, D. et al. Defects in secretion, aggregation, and thrombus formation in platelets from mice lacking Akt2. The Journal of clinical investigation 113, 441-450 (2004). 185. Easton, R.M. et al. Role for Akt3/protein kinase Bgamma in attainment of normal brain size. Molecular and cellular biology 25, 1869-1878 (2005). 186. Tschopp, O. et al. Essential role of protein kinase B gamma (PKB gamma/Akt3) in postnatal brain development but not in glucose homeostasis. Development (Cambridge, England) 132, 2943-2954 (2005). 187. Peng, X.D. et al. Dwarfism, impaired skin development, skeletal muscle atrophy, delayed bone development, and impeded adipogenesis in mice lacking Akt1 and Akt2. Genes & development 17, 1352-1365 (2003). 118 Chapter 3

188. Yang, Z.Z. et al. Dosage-dependent effects of Akt1/protein kinase Balpha (PKBalpha) and Akt3/PKBgamma on thymus, skin, and cardiovascular and nervous system development in mice. Molecular and cellular biology 25, 10407-10418 (2005). 189. Dummler, B. et al. Life with a single isoform of Akt: mice lacking Akt2 and Akt3 are viable but display impaired glucose homeostasis and growth deficiencies. Molecular and cellular biology 26, 8042-8051 (2006). 190. Inoki, K., Li, Y., Zhu, T., Wu, J. & Guan, K.L. TSC2 is phosphorylated and inhibited by Akt and suppresses mTOR signalling. Nature cell biology 4, 648-657 (2002). 191. Potter, C.J., Pedraza, L.G. & Xu, T. Akt regulates growth by directly phosphorylating Tsc2. Nature cell biology 4, 658-665 (2002). 192. Li, Y., Inoki, K., Vacratsis, P. & Guan, K.L. The p38 and MK2 kinase cascade phosphorylates tuberin, the tuberous sclerosis 2 gene product, and enhances its interaction with 14-3-3. The Journal of biological chemistry 278, 13663-13671 (2003). 193. Garami, A. et al. Insulin activation of Rheb, a mediator of mTOR/S6K/4E-BP signaling, is inhibited by TSC1 and 2. Molecular cell 11, 1457-1466 (2003). 194. Zhang, Y. et al. Rheb is a direct target of the tuberous sclerosis tumour suppressor proteins. Nature cell biology 5, 578-581 (2003). 195. Tee, A.R., Manning, B.D., Roux, P.P., Cantley, L.C. & Blenis, J. Tuberous sclerosis complex gene products, Tuberin and Hamartin, control mTOR signaling by acting as a GTPase- activating protein complex toward Rheb. Curr Biol 13, 1259-1268 (2003). 196. Inoki, K., Li, Y., Xu, T. & Guan, K.L. Rheb GTPase is a direct target of TSC2 GAP activity and regulates mTOR signaling. Genes & development 17, 1829-1834 (2003). 197. Long, X., Lin, Y., Ortiz-Vega, S., Yonezawa, K. & Avruch, J. Rheb binds and regulates the mTOR kinase. Curr Biol 15, 702-713 (2005). 198. Corradetti, M.N. & Guan, K.L. Upstream of the mammalian target of rapamycin: do all roads pass through mTOR? Oncogene 25, 6347-6360 (2006). 199. Sarbassov, D.D. et al. Prolonged rapamycin treatment inhibits mTORC2 assembly and Akt/PKB. Molecular cell 22, 159-168 (2006). 200. Dormond, O., Madsen, J.C. & Briscoe, D.M. The effects of mTOR-Akt interactions on anti- apoptotic signaling in vascular endothelial cells. The Journal of biological chemistry 282, 23679-23686 (2007). 201. Holz, M.K., Ballif, B.A., Gygi, S.P. & Blenis, J. mTOR and S6K1 mediate assembly of the translation preinitiation complex through dynamic protein interchange and ordered phosphorylation events. Cell 123, 569-580 (2005). 202. Raught, B. et al. Serum-stimulated, rapamycin-sensitive phosphorylation sites in the eukaryotic translation initiation factor 4GI. The EMBO journal 19, 434-444 (2000). 203. Rogers, G.W., Jr., Richter, N.J. & Merrick, W.C. Biochemical and kinetic characterization of the RNA helicase activity of eukaryotic initiation factor 4A. The Journal of biological chemistry 274, 12236-12244 (1999). 204. Jefferies, H.B. et al. Rapamycin suppresses 5'TOP mRNA translation through inhibition of p70s6k. The EMBO journal 16, 3693-3704 (1997). 205. Fingar, D.C. et al. mTOR controls cell cycle progression through its cell growth effectors S6K1 and 4E-BP1/eukaryotic translation initiation factor 4E. Molecular and cellular biology 24, 200-216 (2004). 206. Gingras, A.C., Raught, B. & Sonenberg, N. Regulation of translation initiation by FRAP/mTOR. Genes & development 15, 807-826 (2001). 207. Hay, N. & Sonenberg, N. Upstream and downstream of mTOR. Genes & development 18, 1926-1945 (2004). 208. Mamane, Y. et al. eIF4E--from translation to transformation. Oncogene 23, 3172-3179 (2004). 209. Mayer, C., Zhao, J., Yuan, X. & Grummt, I. mTOR-dependent activation of the transcription factor TIF-IA links rRNA synthesis to nutrient availability. Genes & development 18, 423-434 (2004). 210. Redpath, N.T., Foulstone, E.J. & Proud, C.G. Regulation of translation elongation factor-2 by insulin via a rapamycin-sensitive signalling pathway. The EMBO journal 15, 2291-2297 (1996). 211. Wang, X. et al. Regulation of elongation factor 2 kinase by p90(RSK1) and p70 S6 kinase. The EMBO journal 20, 4370-4379 (2001). Introduction: The PI3K/Akt pathway 119

212. Browne, G.J. & Proud, C.G. A novel mTOR-regulated phosphorylation site in elongation factor 2 kinase modulates the activity of the kinase and its binding to calmodulin. Molecular and cellular biology 24, 2986-2997 (2004). 213. Hait, W.N., Wu, H., Jin, S. & Yang, J.M. Elongation factor-2 kinase: its role in protein synthesis and autophagy. Autophagy 2, 294-296 (2006). 214. Wu, H., Yang, J.M., Jin, S., Zhang, H. & Hait, W.N. Elongation factor-2 kinase regulates autophagy in human glioblastoma cells. Cancer research 66, 3015-3023 (2006). 215. Schmidt, E.V. The role of c-myc in cellular growth control. Oncogene 18, 2988-2996 (1999). 216. Inoki, K., Zhu, T. & Guan, K.L. TSC2 mediates cellular energy response to control cell growth and survival. Cell 115, 577-590 (2003). 217. Shaw, R.J. et al. The LKB1 tumor suppressor negatively regulates mTOR signaling. Cancer cell 6, 91-99 (2004). 218. Shaw, R.J. et al. The tumor suppressor LKB1 kinase directly activates AMP-activated kinase and regulates apoptosis in response to energy stress. Proceedings of the National Academy of Sciences of the United States of America 101, 3329-3335 (2004). 219. Kimball, S.R. & Jefferson, L.S. New functions for amino acids: effects on gene transcription and translation. The American journal of clinical nutrition 83, 500S-507S (2006). 220. Gulati, P. & Thomas, G. Nutrient sensing in the mTOR/S6K1 signalling pathway. Biochemical Society transactions 35, 236-238 (2007). 221. Klionsky, D.J. & Emr, S.D. Autophagy as a regulated pathway of cellular degradation. Science (New York, N.Y 290, 1717-1721 (2000). 222. Averous, J. & Proud, C.G. When translation meets transformation: the mTOR story. Oncogene 25, 6423-6435 (2006).

Chapter 4 Angiogenesis

Introduction: Angiogenesis 123

4. ANGIOGENESIS

4.1. Background and definitions

Mammalian cells require oxygen and nutrients for their survival and are therefore located within 100 to 200µm (the diffusion limit of oxygen) of blood vessels. Multicellular organisms develop and recruit new blood vessels by two distinct mechanisms: vasculogenesis and angiogenesis, generally called neovascularization. During embryonic development the first blood vessels arise from endothelial precursor cells (EPCs), which share an origin with hematopoietic progenitor cells (hemangioblasts). These EPCs assemble into a primitive vascular labyrinth of small capillaries, a process called vasculogenesis. Angiogenesis is the process whereby new vessels grow from pre-existing vessels through sprouting and remodeling. In the growing embryo the primitive vascular plexus further expands through angiogenesis into a highly organized vascular network of larger vessels branching into smaller ones, and with functional arteries, arterioles, capillaries, venules and veins. Capillaries are the smallest blood vessels, measuring only 5 to 10 µm in diameter, and enable the exchange of oxygen, nutrients and waste products between blood and surrounding tissues. Capillaries, together with post-capillary venules, are the vessels on which cytokines act to regulate vascular permeability and recruitment of leukocytes to inflamed tissue. The wall of capillaries consists of a single layer of endothelial cells (ECs). Larger vessels, however, become covered with mural cells in a process called vessel maturation or arteriogenesis. These mural cells, in essence pericytes and smooth muscle cells (SMCs), provide strength to the vessels and allow regulation of vessel perfusion 1.

After birth, angiogenesis still contributes to further growth, but in the adult most blood vessels remain quiescent and angiogenesis normally only occurs in wound healing, in the cycling ovary and in the placenta during pregnancy. However, in certain diseases, especially cancer, angiogenesis may become prominent (see below). Angiogenesis is tightly regulated by a balance of many pro- and anti-angiogenic factors (see Table 4.1.). In normal, healthy tissue, pro-angiogenic factors are delicately balanced by anti-angiogenic factors, keeping the angiogenic program off. During tissue repair or tumor growth on the other hand, pro-angiogenic factors are produced and/or anti-angiogenic factors are downregulated, thereby tipping the balance towards angiogenesis. This system is known as the angiogenic switch (Fig.4.1.) .

124 Chapter 4

Fig. 4.1. The balance hypothesis of the ‘angiogenic switch’. Angiogenesis is tightly controlled by the balance of two sets of counteracting factors: angiogenesis activators and inhibitors (or pro- and anti- angiogenic factors, respectively). In many pathological conditions this balance is disturbed, leading to either excessive or impaired angiogenesis.

Two main forms of angiogenesis are generally recognized: sprouting angiogenesis and intussusception. In sprouting angiogenesis, pro-angiogenic factors stimulate endothelial cells in pre-existing blood vessels to degrade the basement membrane and to start to proliferate into the surrounding matrix towards the angiogenic stimulus, in the form of a solid sprout that remains connected with the original vessel. By contrast, in intussusceptive angiogenesis, the endothelial cells extend into the lumen of the pre- existing blood vessel to divide a single vessel into two parallel vessels. It has now become clear that EPCs (also called angioblasts) from the bone marrow or present in circulation may also contribute to neovascularization in the adult (Fig.4.2.) 2.

Introduction: Angiogenesis 125

Fig. 4.2. Angiogenesis and vasculogenesis. (A) Genesis of the vascular system during embryonic development. Mesodermal cells differentiate into hemangioblasts leading to the formation of primitive blood islands. The peripheral hemangioblasts differentiate into angioblasts, the precursors of endothelial cells. Then, the angioblasts migrate allowing the fusion of blood islands and their remodeling into tubular structures, giving rise to a first primitive vascular plexus. This vascular plexus remodels into larger vessels, which then further branch through the process of angiogenesis. (B) Mechanisms of angiogenesis. In sprouting angiogenesis, a sprout of endothelial cells expands and migrates towards the angiogenic stimulus, guided by an endothelial tip cell. Intussusception involves the insertion of interstitial tissue columns into the lumen of pre-existing blood vessels. EPCs recruited from the bone marrow or from circulation may also contribute to the endothelial lining of new blood vessels. Adapted from Lamalice et al., 2007 and Auguste et al., 2005 3, 4.

126 Chapter 4

4.2. Regulation of angiogenesis

Table 4.1. Some naturally occurring angiogenesis activators and inhibitors. Activators Inhibitors VEGF family members VEGF-R1, soluble VEGF-R1 VEGF-Rs, NRP-1 soluble NRP-1 Ang-1, Tie-2 Ang-2* PDGF-BB, PDGF-Rs TSP-1, TSP-2 TGF-1*, endoglin, TGF-Rs Angiostatin, related plasminogen kringles FGF, HGF Endostatin (collagen XVIII fragment) MCP-1, IL-8, other chemokines* Vasostatin; Calreticulin

Integrins 11, v15, v3, v5, 51 Platelet factor-4 VE-cadherin, PECAM IFN-, IFN-, IFN- Ephrins Meth-1, Meth-2 Plasminogen activators, MMPs TIMPs, MMP inhibitors, PEX PAI-1* IP-10 NOS, COX-2 IL-4, IL-12, IL-18 AC133 Prothrombin kringle-2 Id1/Id3 Prolactin 16kDa fragment TNF* VEGI Proliferin Canstatin; Tumstatin Angiogenin Osteopontin fragment; Fibronectin fragment Pleiothrophin Fragment of SPARC; Antithrombin III fragment Leptin Maspin; Restin; Proliferin-related protein * opposite effects in some contexts. Adapted from Carmeliet et al., 2000 2.

4.2.1. Positive regulators of angiogenesis

In 1984, the first pro-angiogenic factor was isolated by Shing and co-workers from a rat chondrosarcoma, based on its affinity for heparin 5. After purification and sequencing it turned out to be basic fibroblast growth factor (bFGF) 6. The fibroblast growth factor (FGF) family comprises more than 20 members, of which two have been associated with angiogenesis: acidic FGF (aFGF) (or FGF-1), and bFGF (or FGF-2). All FGF family members have high affinity for heparin and can associate with cell-surface heparan sulfate proteoglycans. Four tyrosine kinase receptors for FGFs have been identified: FGF-R1, -R2, -R3, and -R4. Both aFGF and bFGF are potent inducers of angiogenesis in many angiogenesis models, such as the chick chorioallantoic membrane (CAM) assay and the rabbit corneal neovascularization assay, even more potent than VEGF 6-8. aFGF and bFGF induce proliferation and migration of endothelial cells, production of urokinase-type plaminogen activator (uPA) and matrix metalloproteinases (MMPs), as well as the formation of tubular structures in vitro 9-11. By contrast, genetic manipulations have revealed that aFGF and bFGF have no significant role in embryonic angiogenesis. Both aFGF and bFGF knockout mice have normal vascular development, and even double knockout mice Introduction: Angiogenesis 127

are viable, suggesting that aFGF and bFGF play only limited roles under physiological conditions 12-14. However, bFGF-deficient mice display mild defects in wound healing, and topical application of aFGF or bFGF has been shown to accelerate wound healing in a number of animal models, pointing to a role of these FGFs in wound healing and vascular repair 15, 16. aFGF and bFGF are potent mitogens for endothelial cells, fibroblasts, SMCs and keratinocytes, and hence these growth factors can induce the rapid formation of granulation tissue that fills up a wound space early in the wound healing process 17-20. Interestingly, aFGF and bFGF lack a signal peptide for secretion by the endoplasmic reticulum-Golgi pathway. It has been suggested that these growth factors are stored within the cell, and are only released in case of cell damage or death 21, 22. aFGF and bFGF may also be secreted through an alternative pathway 23, 24. The high affinity of aFGF and bFGF for heparin and heparan sulfate allows storage of these FGFs in the extracellular matrix (ECM) and on the surface of cells until they are released by the actions of proteinases and heparinases that are produced during the angiogenic process 25. Elevated levels of bFGF are found in many tumors and in certain tumor types, particularly melanoma, the bFGF levels appear to correlate with intratumoral microvessel density and cancer progression/prognosis 11.

VEGF is probably the most important activator of both physiological and pathological angiogenesis. The VEGF family consists of seven members, called VEGF (or VEGF- A), VEGF-B, VEGF-C, VEGF-D, VEGF-E, VEGF-F and placental growth factor (PlGF). Alternative exon splicing of the VEGF RNA results in the generation of at least six VEGF isoforms, of which VEGF165, the isoform consisting of 165 amino acids after cleavage of a signal sequence, is the predominant isoform 26. These VEGF family members exert their effects via three tyrosine kinase receptors: VEGF-R1 (or Flt-1), VEGF-R2 (or Flk-1/KDR), and VEGF-R3 (Fig.4.3.). VEGF binds to both

VEGF-R2 and VEGF-R1. In addition, VEGF165 also binds to neuropilin 1 (NRP1) and NRP2, which act as coreceptors for VEGF-R2, and cell-surface heparan sulfate proteoglycans 27. Genetic manipulations have revealed the fundamental importance of the VEGF signaling pathway for physiological blood vessel formation in mice. All VEGF, VEGF-R1, VEGF-R2 or NRP1 knockout mice die in utero, having incomplete or absent vascular structures 28-30. Even mice having only a single allele of VEGF are embryonic lethal 31, 32.

128 Chapter 4

Fig. 4.3. VEGF family members, their receptors, and most important signaling pathways. The VEGF-Rs have seven extracellular immunoglobulin-like domains. The extracellular domain of VEGF- R3 is once cleaved and joined with a disulfide (S-S) bond. VEGF-R2 is the main signaling receptor on ECs and is a major regulator of vasculogenesis, angiogenesis, vasodilation and vascular permeability. VEGF-R3 is the main signaling receptor on lymphatic ECs and is involved in lymphangiogenesis. VEGF-R3 has also been associated with tumor angiogenesis. The extracellular domain of VEGF-R1 is also expressed as a soluble form (sR1). This soluble form of VEGF-R1 is an important modulator for the placental vasculature. Besides on ECs, VEGF-R1 is also expressed on monocytes and its activation leads to chemotaxis and tissue factor production in monocytes. NRPs are coreceptors for VEGF-R2.

NRP1 binds VEGF165, PlGF-2, VEGF-B and VEGF-E; NRP2 binds VEGF145, VEGF165, PlGF-2 and VEGF-C. Adapted from Roy et al., 2006 27.

VEGF was originally described as a vascular permeability factor secreted by carcinoma cell lines that enhanced permeability of skin blood vessels 33. Indeed, VEGF can enhance vascular permeability by increasing the vesicular transport through the cytoplasm of the endothelial cells, by inducing endothelial fenestrations, and by weakening endothelial cell-cell junctions, at least in part through tyrosine phosphorylation of VE-cadherin 34-37. This increased permeability, together with the VEGF-induced vasodilation, may be considered as the first step in the angiogenic process 38. Vasodilation and increased permeability lead to leakage of fluid and plasma proteins such as fibrin into the surrounding tissue. Cross-linked fibrin deposits may function as a provisional ECM that supports endothelial cell migration and new blood vessel formation 39. VEGF induces the production of proteolytic enzymes by endothelial cells, including MMPs and the plasminogen activators uPA and tissue- type plasminogen activator (tPA) 40-42. MMPs and plasminogen activators mediate the Introduction: Angiogenesis 129

breakdown of the basement membrane that surrounds the blood vessel and ECM components, which allows the migration of endothelial cells. Plasminogen activators are serine proteases that convert the pro-enzyme plasminogen to plasmin, which then degrades fibrin and other ECM components. Plasmin may also activate MMPs 43. uPA is secreted as an inactive pro-enzyme, binds to the uPA receptor (uPAR) on endothelial cells and is proteolytically activated by MMP-2. The expression of uPAR on endothelial cells is also increased by VEGF 44. Plasminogen activators are inhibited by plasminogen activator inhibitor-1 (PAI-1), which is also induced by VEGF 40. It has been reported that uPAR promotes endothelial cell adhesion on the ECM component vitronectin, and that PAI-1 blocks this adhesion and thus enhances endothelial cell migration on vitronectin 45. MMPs are a family of zinc-dependent endopeptidases that are capable of degrading various components of the ECM and play an important role in tissue remodeling events such as angiogenesis and tumor invasion. MMPs are secreted as inactive pro-enzymes, which become activated by proteinases such as plasmin or other MMPs. VEGF induces the expression of the collagenase MMP-1 and the activation of the gelatinase MMP-2, which is constitutively secreted by endothelial cells 41, 42, 46. MMP-2 was shown to be responsible for the VEGF-induced activation of uPA bound to its cell-surface receptor uPAR 47. Besides by inducing ECM breakdown, VEGF also stimulates the migration of endothelial cells by directly increasing the motility of these cells 3, 48. VEGF furthermore induces proliferation and survival of endothelial cells through both the PI3K/Akt and MAPK pathways 49-52. Finally, VEGF induces sprouting and the formation of tubular structures in various in vitro and in vivo models 53-55. It has been shown in a retinal angiogenesis model that angiogenic sprouts are guided by tip cells at the invading front of angiogenic sprouts. These tip cells appear to cluster VEGF- R2s and follow a VEGF gradient 56. VEGF-R1 and VEGF-R2 are also expressed on EPCs and VEGF may induce mobilization and recruitment of these cells to the site of neovascularization 57, 58. In conclusion, VEGF mediates the whole angiogenic process up to the formation of immature, but functional neovessels. Interestingly, sustained expression of VEGF, as is the case in many tumors, actually inhibits the maturation of new blood vessels, resulting in vascular beds that are tortuous, unorganized, dilated and leaky 59, 60. The expression of VEGF is regulated by the oxygen concentration in the cell via the transcription factors hypoxia-inducible factor (HIF)-1 and HIF-2. HIF-1 is a heterodimer of HIF-1 and aryl hydrocarbon nuclear translocator (ARNT), also called HIF-1. Under normoxic conditions, HIF-1 is a target for HIF prolyl hydroxylase, which makes HIF-1 a target for E3 ubiquitin ligase activity, leading to quick degradation by the proteasome. Under hypoxic conditions, HIF prolyl hydroxylase is inhibited and HIF-1 is stabilized. HIF-1 can then bind to HIF-responsive elements (HREs) in the promoter regions of VEGF and other genes, thereby inducing expression 61. The von Hippel-Lindau tumor suppressor protein (pVHL) plays an 130 Chapter 4

important role in regulating HIF-1. pVHL binds to the modified HIF-1 and forms the recognition component for the E3 ubiquitin ligase complex that ubiquitinates HIF-1 62, 63. Mutations of the VHL gene are associated with increased VEGF expression and angiogenesis 64. The expression of VEGF is also induced by growth factors, such as EGF, FGF, PDGF, TGF- and TGF-, cytokines such as IL-1 and IL-6, and certain hormones 65. PlGF is a member of the VEGF family that was first identified in placenta but is now known to be also present in other tissues such as heart and lungs 66. Four human isoforms of PlGF have been described: PlGF-1, -2, -3, and -4 67. In mice, only PlGF-2 is expressed 68. All isoforms bind to VEGF-R1, and PlGF-2 also binds to NRP1 and NRP2. Like VEGF, PlGF stimulates endothelial cell growth, migration and survival, recruitment of EPCs, and angiogenesis under certain conditions 69-71. In addition, PlGF has also been shown to be a survival factor for monocytes and macrophages, and is capable of recruiting these cells to sites of inflammation and angiogenesis 72-74. In contrast to mice deficient for VEGF signaling, PlGF knockout mice develop normally and are healthy and fertile 75. These mice have normal physiological angiogenesis. However, PlGF knockout mice have impaired angiogenesis under pathological conditions such as ischemia, wound healing and cancer. Thus PlGF appears to be specific for pathological angiogenesis, making it an interesting target for anti-angiogenic therapy 71, 75, 76. Most of the angiogenic effects of PlGF appear to be dependent on VEGF. In the absence of PlGF, VEGF binds to VEGF-R1, with minimal signaling activity. However, when PlGF is expressed, it displaces VEGF from VEGF-R1, making it available for signaling via VEGF-R2. VEGF-R1/VEGF- R2 heterodimers have also been found that are activated by PlGF-2/VEGF heterodimers. Furthermore, binding of PlGF to VEGF-R1 has been shown to enhance VEGF-R2 phosphotyrosine levels, suggesting that there is cross-talk between these two receptors 77. To add to complexity, PlGF-1/VEGF heterodimers have been identified that are functionally inactive. Thus, PlGF-1 may function as a natural antagonist of VEGF 78.

Angiopoietins (Angs) are a family vascular-specific growth factors that play an important role in angiogenesis. The human Ang family consists of three members, called Ang-1, Ang-2 and Ang-4. Mice do not express Ang-4, but express a fourth Ang family member, called Ang-3, which is believed to be the mouse orthologue of human Ang-4 79. The Angs signal via the tyrosine kinase receptor Tie-2, which is mainly expressed on endothelial cells. There is also a closely related Tie-1, which is a so- called orphan receptor 80. Ang-1 is generally considered as an activator of Tie-2, while Ang-2 is considered to be a negative regulator of Tie-2 signaling. Ang-1 is mainly produced by mural cells, in essence pericytes and SMCs, and stabilizes the interaction of endothelial cells with mural cells. Mice overexpressing Ang-1 produce enlarged and more numerous vessels with highly regulated junctional complexes, that are resistant to VEGF-induced permeabilization 81, 82. Ang-1 has also been shown to Introduction: Angiogenesis 131

induce sprouting angiogenesis in vitro and this activity was synergistic with VEGF, suggesting that Ang-1 does not counteract VEGF-induced angiogenesis 83. Ang-1 confers survival signals to endothelial cells via the PI3K/Akt pathway and upregulation of anti-apoptotic proteins such as the IAP family member survivin 84. Ang-2 is mainly produced by endothelial cells at sites of vascular remodeling. Ang-2 thus regulates Tie-2 activity in an autocrine manner. Binding of Ang-2 to Tie-2 may render the receptor less responsive to Ang-1, leading to destabilization of the vessel 85. Mice overexpressing Ang-2 or deficient for Ang-1 or Tie-2 are embryonic lethal and display vascular defects attributed to defective vascular remodeling and disrupted interaction between endothelial cells and mural cells 86-88. In the absence of VEGF, Ang-2 induces regression of angiogenic sprouts, whereas in the presence of VEGF it promotes vascular sprouts 80, 89. The expression of Ang-1 and Ang-2 is inversely regulated. Ang-2 is upregulated by VEGF, hypoxia and inflammatory cytokines such as TNF 90, 91. By contrast, the expression of Ang-1 is downregulated by these stimuli 80, 92. Recently, Shim and colleagues have suggested that spatiotemporal changes of the Ang-1/Ang-2 balance are crucial to regulate the angiogenic cycle (Fig.4.4.) 80.

Certain integrins appear to have specific functions in angiogenesis. Integrins mediate the interactions of endothelial cells with the ECM. Several integrins were shown to be upregulated in endothelial cells during angiogenesis, including the integrins 11, 93-96 21, v3, v5 and 51 . 11 and 21 bind to collagen and laminin, v5 binds to vitronectin, 51 binds to fibronectin, while v3 binds to many ECM components containing an RGD (arginine, glycine, aspartic acid) sequence, such as vitronectin, fibronectin and fibrinogen. All these integrins have been shown to play crucial roles in angiogenesis and inhibition of these integrins using specific antagonists can completely block angiogenesis in experimental models 96-100. The activation of integrins transmits important survival signals and allows the functional connection between focal adhesions and the actin cytoskeleton that is required to drive cell migration 101. These integrins may also function as receptors for pro- angiogenic growth factors. The integrins 51 and v5 have been shown to bind Ang-1 and Ang-2, and this interaction promoted the migration of endothelial cells 102. Similarly it was shown that endothelial cells can adhere to the VEGF isoforms

VEGF165 and VEGF189 via the integrin v3 and this interaction also stimulated 103 endothelial cell migration and survival . v3 can also directly promote the migration and invasion of endothelial cells by binding to MMP-2 104. Furthermore, it has been shown that v3 associates with VEGF-R2 and participates in the full 105 activation of this receptor . Antagonists of v3 potently blocked angiogenesis induced by VEGF, but not by bFGF. By contrast, inhibition of v5 blocked bFGF- but not VEGF-induced angiogenesis 98. Together, these data suggest that there is a close relationship between v3 and VEGF-signaling.

132 Chapter 4

Fig. 4.4. The angiogenic cycle. (1) The angiogenic switch is triggered by hypoxia, inflammation, or genetic mutations that disrupt the balance between pro- and anti-angiogenic factors. (2) The upregulation of Ang-2 in response to these stimuli destabilizes the endothelium by decreasing the phosphorylation status of Tie-2. This results in increased vessel plasticity and more responsive endothelial cells. (3) The activated matrix proteases promote mural cell detachment and facilitate endothelial chemotaxis and sprouting in response to VEGF, Ang-1, FGF, PDGF, and hepatocyte growth factor (HGF). (4) As the angiogenic stimuli and the expression of Ang-2 gradually decrease, an equilibrium is reached that enables reinvestment of basement membrane and ECM, and mural cell recruitment to the nascent vasculature. Adapted from Shim et al., 2007 80.

Ephrins and Eph receptors are a large family of membrane-bound/transmembranous ligands and tyrosine kinase receptors that play essential roles in axon guidance in the developing nervous system, but also in angiogenesis. Mouse embryos deficient for ephrin-B2 or its EphB4 receptor suffer fatal defects in early angiogenic remodeling 106, 107. Ephrin-B2 and EphB4 display remarkably reciprocal expression patterns. Ephrin-B2 is specifically expressed on arterial endothelial cells and SMCs, while EphB4 is mainly expressed on venous endothelial cells. Ephrin-B2/EphB4 signaling is critical for the establishment of arterial and venous identities and participates in the formation of arteriovenous contact by arresting endothelial cell migration at the arterial-venous interface 108. Ephrin-B2 and EphB4 have also been linked to breast cancer. Breast cancer cells express EphB4, while the tumor vasculature expresses ephrin-B2, suggesting the ephrin-B2/EphB4 signaling can also support tumor angiogenesis 109, 110. Ephrin-A1 has also been implicated in angiogenesis. Both TNF and VEGF induce the expression of ephrin-A1 in HUVECs as an immediate early gene product 111, 112. Introduction: Angiogenesis 133

PDGF-BB and TGF-1 play essential roles in vessel maturation. Nascent vessels are stabilized by the recruitment of mural cells and the production of ECM. PDGF-BB is secreted by endothelial cells, presumably in response to VEGF, and induces the recruitment of PDGF-R-positive mesenchymal progenitor cells, which then differentiate to vascular mural cells 113. Mural cells in turn produce Ang-1, which then further stabilizes the vessel. PDGF-BB also induces endothelial cell proliferation and tube formation in vitro and has potent angiogenic activity in the CAM assay 114, 115. TGF-1 and its receptors (TGF-RI, TGF-RII and endoglin) promote vessel maturation by stimulating ECM production and by inducing differentiation of mesenchymal cells to mural cells 116, 117. TGF-1 is expressed by a number of cells, including endothelial cells and mural cells, and, depending on the context and concentration, can have pro- or anti-angiogenic effects. At low concentration, it contributes to angiogenesis by upregulating proteinases and Id1, a protein required for proliferation and migration. At high doses, TGF-1 inhibits endothelial cell growth and induces vessel maturation 59, 118. TGF-1 also induces the expression of PAI-1, which promotes vessel maturation by preventing degradation of the provisional matrix around the nascent vessel 59.

As endothelial cells are among the most important target cells of TNF, it is not surprising that TNF can also modulate the angiogenic process. Like it is the case for TGF-1, the angiogenic effect of TNF appears to be context and concentration dependent. TNF is a potent inducer of vessel permeabilization, and strongly synergizes with VEGF for this effects 119. It has even been shown that TNF knockout mice fail to respond to VEGF with an increase in vascular permeability, suggesting that TNF acts as a permissive factor for VEGF-induced vessel permeabilization, which is a crucial step early in the angiogenic process 120. Like VEGF, TNF induces expression of angiogenic factors such as ephrin-A1 and Ang-2 in cultured endothelial cells 90, 111. Ang-2 can in turn sensitize endothelial cells toward TNF and modulate TNF-induced expression of endothelial cell adhesion molecules 121. Low doses of TNF can stimulate chemotaxis and the formation of capillary tube-like structures of cultured endothelial cells in vitro, and induce capillary blood vessel formation in the rabbit cornea and CAM models of angiogenesis 122, 123. By contrast, TNF inhibits endothelial cell proliferation in response to bFGF in vitro, and at higher dose it can even induce endothelial cell apoptosis 123, 124. In particular the combination of TNF and IFN- appears to be capable of inducing endothelial cell death 125.

134 Chapter 4

4.2.2. Negative regulators of angiogenesis

Thrombospondin-1 (TSP-1) was the first naturally occurring inhibitor of angiogenesis to be described 126, 127. A few years later, a closely related protein, TSP-2, was also found to inhibit angiogenesis 128. TSPs are a family of ECM proteins that can bind to cell surface receptors and other ECM components. TSP-1 and TSP-2 are produced by several cell types, including platelets, endothelial cells, and fibroblasts. TSP-1 binds

to integrin v3 as well as to CD36, a member of the scavenger receptor family, on endothelial cells 129. TSP-1 inhibits endothelial cell migration and proliferation, induces apoptosis, and blocks tube formation 130, 131. Moreover, TSP-1 renders endothelial cells refractory to stimulation with VEGF or bFGF 132.

Many angiogenesis inhibitors are actually fragments of ECM components, basement membrane components or other proteins, and are generated by the actions of MMPs, plasmin or other proteinases. These include, among many others, the plasmin fragment angiostatin, the collagen fragments restin, tumstatin, canstatin, endostatin and arresten, the calreticulin fragment vasostatin, and the MMP-2 fragment PEX 133- 140. Angiostatin is a circulating angiogenesis inhibitor that was first isolated in 1994 by the group of Folkman from the serum and urine of Lewis lung carcinoma (LLC)- bearing C57BL/6 mice, while they were studying the interesting phenomenon that a primary tumor can inhibit the growth of its distant metastasis, and that removal of a primary tumor can lead to rapid growth of metastasis 133. Angiostatin can be generated from plasmin by the actions of several MMPs 141. Angiostatin can induce apoptosis and inhibit migration and tube formation of endothelial cells in vitro, and is a potent inhibitor of tumor growth in vivo 142, 143. Several mechanisms of action have been proposed for the anti-angiogenic activities of angiostatin. It has been shown to bind to the ATP synthase on the surface of HUVECs and thus might interfere with the cells’ 144 energy supply . In addition, angiostatin can interact with integrin v3 and might interfere with hepatocyte growth factor (HGF) signaling 145, 146. Endostatin was also first described by the group of Folkman, as a factor produced by a tumor cell line that can inhibit endothelial proliferation in vitro, angiogenesis in the CAM assay, and tumor growth in vivo 134. It is a carboxyterminal fragment of the basement membrane protein collagen XVIII, generated after cleavage by MMPs or cathepsin L 147, 148. Endostatin inhibits proliferation and migration of endothelial cells, 149 most likely through binding with integrin 51 . Other fragments of collagen, such as tumstatin, canstatin and arresten, have also been shown to inhibit endothelial cell proliferation and migration by binding to integrins 150-152. PEX, a fragment of MMP- 140 2, binds to v3 and blocks the binding of MMP-2 to this integrin .

Tissue inhibitors of MMPs (TIMPs) are endogenous inhibitors of MMPs. There are 4 TIMPs, TIMP-1 to -4, which bind tightly to and inhibit the activity of almost all MMPs. Expression of TIMPs has been shown to block endothelial migration and Introduction: Angiogenesis 135

angiogenesis in several models 153-155. However, for some TIMPs additional, non- inhibitory functions have been identified. For example, TIMP-2 can bind to pro- MMP-2 and facilitate its cell surface activation by membrane type 1-MMP (MT1- MMP) 156, 157.

The activity of important pro-angiogenic growth factors is tightly regulated by several mechanisms. Soluble forms of VEGF-R1, NRP1, Tie-1 and Tie-2 have been identified, which may negatively regulate the activity of VEGF and Angs 158-161. Platelet factor 4 (or CXCL4), a chemokine released by platelets, can bind to both VEGF and bFGF, thereby inhibiting the binding of these growth factors to their receptors 162, 163.

Finally, some cytokines have anti-angiogenic activity, for example IL-12, IFNs, and the TNF family member vascular endothelial growth inhibitor (VEGI) 164. The anti- angiogenic activity of IL-12 was shown to be largely due to induction of IFN- 165. IFN- induces the expression of the anti-angiogenic chemokine CXCL10 in endothelial cells 166. CXCL10 inhibits endothelial tube formation in response to VEGF or bFGF via activation of its receptor CXCR3 167, 168. IFN- in combination with TNF has also been shown to downregulate the expression of integrin v3 on HUVECs 169. Moreover, Rüegg and co-workers demonstrated that treatment of cultured endothelial cells leads to decreased v3-dependent cell adhesion, resulting in detachment of endothelial cells and apoptosis 170.

136 Chapter 4

4.3. Tumor vascularization

4.3.1. Tumor angiogenesis

Tumor growth depends on angiogenesis. The concept of tumor angiogenesis was first introduced by Folkman and colleagues in the early 1970s. They demonstrated that, in the absence of new blood vessel formation, tumors cannot grow beyond a size of 1 to 2 mm in diameter 171. A tumor starts with a single cell that loses control over its cell cycle and starts proliferating, thereby forming a small nodule. However, when the tumor reaches a certain critical size, the oxygen and nutrients supplied by pre-existing vessels can no longer sufficiently reach the centre of the nodule and the tumor will stop growing. Tumor cells are still proliferating, but this is balanced by high rates of tumor cell apoptosis. The tumor is in a so-called dormant state. This dormant state can persist for a long period of time, even years, until, eventually, tumor cells have accumulated the required mutations to turn on the angiogenic switch. From that point on, the tumor is capable of inducing the growth of its own blood vessels, which deliver the necessary oxygen and nutrients to support the rapid expansion of the tumor 172. The angiogenic switch in tumors is mostly established by oncogene-driven expression of pro-angiogenic growth factors, such as VEGF, bFGF, PlGF, TGF-1, and others 173-175. Tumor cells can also induce the expression of pro-angiogenic factors by stromal cells in or near the tumor 176. Tumor angiogenesis is furthermore promoted by the hypoxic conditions in the tumor which upregulate the expression of VEGF, Ang-2 and other angiogenic factors 61, 91. Most tumors acquire their blood supply by inducing sprouting angiogenesis, in essence the ingrowth of new blood vessels into the tumor from pre-existing ones. However, in recent years, some other mechanisms have been recognized. These include EPC recruitment, vessel cooption, mosaic blood vessels, and vasculogenic mimicry.

Introduction: Angiogenesis 137

Fig. 4.5. Tumor vascularization. (A) Most tumors acquire their blood supply by inducing sprouting angiogenesis. Tumor cells secrete pro-angiogenic growth factors such as VEGF, bFGF or PDGF, which then activate their receptors on the endothelial cells of nearby vessels, thereby inducing endothelial cell proliferation, migration and survival, resulting in the ingrowth of new vessels into the tumor. Angiogenic factors are not all secreted by tumor cells. Some angiogenic proteins (both pro- and anti- angiogenic) are scavenged by platelets, stored in alpha granules, and appear to be released within the tumor vasculature. New endothelial cells do not all originate from pre-existing vessels. A few are derived from circulating EPCs, which are recruited to the tumor by VEGF or other angiogenic growth factors, differentiate to endothelial cells, and become incorporated into the growing vessels. (B) Some alternative mechanisms of tumor vascularization have been identified. These include vessel cooption, mosaic vessels and vasculogenic mimicry. See text for further details. Adapted from Folkman, 2007 and Auguste et al., 2005 4, 177.

4.3.2. EPC recruitment

EPCs are derived from stem cells present in the bone marrow. These stem cells, termed multipotent adult progenitor cells (MAPCs), express a specific cell surface marker, CD133, that has been associated with primitive undifferentiated cells 178. In vitro, MAPCs can differentiate into many adult tissue types when placed in the proper cytokine milieu 179. For example, MAPCs cultured on fibronectin and stimulated with VEGF have been shown to differentiate into endothelial cells 180. In the bone marrow, MAPCs can differentiate into early EPCs, which express CD133, CD34 and VEGF- R2 181. Early EPCs within the osteoblastic zone of the bone marrow also express cKit 138 Chapter 4

182. cKit is the receptor for the glycoprotein Kit ligand (KitL), also known as stem cell factor. KitL exists in two forms: a membrane bound form, mKitL, and a soluble form, sKitL. It has been suggested that VEGF-R1 may play an important role in initiating the mobilization of EPCs. Stimulation of VEGF-R1 with PlGF or VEGF induces the secretion of MMP-9 by CD34+ bone marrow cells. MMP-9 subsequently cleaves mKitL from the surface of stromal cells in the bone marrow, thereby generating sKitL, which then binds to its receptor cKit on early EPCs. This results in EPC mobilization and release into the peripheral circulation 183-186. Early circulating EPCs further differentiate to late circulating EPCs which no longer express CD133, but express typical endothelial cell markers such as PECAM, VE-cadherin and von Willebrand factor (vWF) 181. There is a basal level of circulating EPCs in the blood, but when neovascularization is necessary and angiogenic stimuli such as VEGF or Ang-1 are released, the circulating pool of EPCs is increased 187, 188. The importance of EPC mobilization in tumor vascularization has been demonstrated in a mouse mutant for Id proteins. Id1 Id3double mutant mice display vascular defects and several tumors fail to grow in these mice due to impaired VEGF-induced mobilization and proliferation of EPCs, resulting in poor neovascularization 189, 190. In wild-type mice, injection of a neutralizing antibody against VEGF-R2, but not against VEGF- R1, could block the recruitment of EPCs to tumor vasculature. However, it has been shown that another type of bone-marrow-derived progenitor cell, termed hematopoietic progenitor cell (HPC), expressing CD34, VEGF-R1 and CD11b, is also recruited to sites of neovascularization and may contribute to tumor vascularization, growth and metastasis 58, 190, 191. The extent of EPC recruitment and incorporation into the tumor vasculature is dependent on the tumor type, varying from 5% or less in most tumors up to 90% in the B6RV2 lymphoma implanted subcutaneously in mice 190, 192. As EPCs appear to be consistently recruited to tumors, it has been suggested that EPCs might be powerful tools to deliver angiogenesis inhibitors to the tumor vasculature 4, 192.

4.3.3. Vessel cooption

Tumors arising in highly vascularized tissues, such as brain and lung, can use another mechanism to acquire their blood supply, named vessel cooption by Holash and colleagues in 1999 89, 193, 194. Instead of inducing the ingrowth of new blood vessels into the tumor, the tumor cells can grow in cuffs around pre-existing blood vessels. By coopting vessels, tumors can also invade in the surrounding tissue. Cooption appears to depend on the site of tumor development. Implantation of a rat C6 glioma into rat brain, injection of rat RBA mammary adenocarcinoma cells into rat brain, metastasis of a Mel57 melanoma cells to brain, or metastasis of LLC cells to lungs were all shown to result in tumors growing by vessel cooption, whereas these tumors can induce sprouting angiogenesis when injected at other sites 89, 195. Interestingly, Introduction: Angiogenesis 139

Holash and colleagues found that coopted vessels displayed striking upregulation of Ang-2, which, in the absence of VEGF, resulted in vessel destabilization, endothelial cell apoptosis, and a dramatic decrease in vessel density by two weeks after implantation in the rat C6 glioma model. By four weeks after implantation, the tumor core was highly necrotic and the remaining cells were organized in cuffs around the few surviving vessels. The tumor periphery on other hand displayed expression of VEGF and robust angiogenesis 89. Vessel cooption is also found in human glioblastoma 89.

4.3.4. Mosaic vessels

Chang and co-workers showed that tumors can have blood vessels in which both endothelial cells and tumor cells form the luminal surface, and termed these vessels mosaic blood vessels 196. Thus, in mosaic blood vessels, tumor cells are in direct contact with blood. Mosaic blood vessel can originate by intravasation of tumor cells into the lumen of tumor vessels, during which the tumor cells stay temporarily in the capillary vessel wall. Moreover, angiogenic growth factors, such as bFGF and VEGF induce MMP-2 activity, thereby increasing the degradation of vessel basement membranes, which increases intravasation of tumor cells and thus the number of mosaic blood vessels. Alternatively, mosaic blood vessels may also be a consequence of endothelial cell apoptosis 4, 196, 197.

4.3.5. Vasculogenic mimicry

The concept of vasculogenic mimicry was introduced by Maniotis et al. in 1999 198. They observed that in aggressive forms of melanoma angiogenic vessels were only present at the border of the tumors, while inside the tumors vessels delineated with endothelial cells could not be found, and, remarkably, these tumors were not necrotic. Histological analysis revealed that there was a network of channels present in these tumor that was connected to the blood vessels in the periphery. Like real blood vessels, these vessels expressed endothelial cell markers such as vWF, PECAM, CD34, and VEGF-R2, and were lined by a thin basal membrane. However, no endothelial cells could be found. Instead, these vessels were composed of a layer of ECM lined by aggressive, undifferentiated tumor cells, and a layer of red blood cells in the vessel lumen 198, 199. Maniotis and colleagues furthermore demonstrated that these highly aggressive melanoma cells are capable of forming blood vessel-like channels and networks in three-dimensional cultures of matrigel or collagen in vitro, in the absence of endothelial cells or fibroblasts, and without the addition of soluble growth factors such as VEGF, bFGF, TGF- or PDGF 198. Gene expression profiling showed that these aggressive melanoma cells express endothelial cell markers such as VEGF-R2, VE-cadherin, Tie-1 and EphA2, as well as ECM components such as 140 Chapter 4

collagens and laminin-5 2 chain 198, 200, 201. It has been suggested that laminin-5 2 chain, in cooperation with MMP-2 and MT1-MMP, is required for vasculogenic mimicry by aggressive melanoma cells 202. Recently, it has been reported that EphA2, the receptor for ephrin-A1, and VE-cadherin act together to regulate melanoma vasculogenic mimicry 203. Besides in melanoma, vasculogenic mimicry has also been found in other types of cancer, including breast, lung, prostate and ovarian cancer 4, 204-207.

4.3.6. Abnormalities in tumor blood vessels

The disorganized production of angiogenic factors makes tumor vessels structurally and functionally abnormal (Fig.4.6.). In contrast to normal vessels, tumor vessels are tortuous an dilated, with uneven diameter and excessive branching and shunts. Consequently, tumor blood flow is chaotic and variable, leading to hypoxic and acidic regions in tumors 2, 208. In addition, tumor vessel basement membrane, endothelial cell-cell junctions, endothelial cell shape, and pericyte coverage are abnormal, which make tumor vessels highly permeable compared to normal vessels 209, 210. The high permeability of tumor vessels, together with the absence of functional lymphatics, results in high interstitial fluid pressure in tumors 2.

Fig. 4.6. Tumor vasculature is structurally and functionally abnormal. (A) Normal vasculature, composed of mature vessels, is well-organized and maintained by a perfect balance of pro- and anti- angiogenic factors. (B) The disorganized production of angiogenic factors makes tumor vasculature abnormal. Tumor vasculature, which is largely composed of immature vessels, is characterized by tortuous vessels, with excessive branching and shunts, and increased permeability. From Jain et al., 2001 211.

Introduction: Angiogenesis 141

4.4. References

1. Carmeliet, P. Angiogenesis in life, disease and medicine. Nature 438, 932-936 (2005). 2. Carmeliet, P. & Jain, R.K. Angiogenesis in cancer and other diseases. Nature 407, 249-257 (2000). 3. Lamalice, L., Le Boeuf, F. & Huot, J. Endothelial cell migration during angiogenesis. Circulation research 100, 782-794 (2007). 4. Auguste, P., Lemiere, S., Larrieu-Lahargue, F. & Bikfalvi, A. Molecular mechanisms of tumor vascularization. Critical reviews in oncology/hematology 54, 53-61 (2005). 5. Shing, Y. et al. Heparin affinity: purification of a tumor-derived capillary endothelial cell growth factor. Science (New York, N.Y 223, 1296-1299 (1984). 6. Esch, F. et al. Primary structure of bovine pituitary basic fibroblast growth factor (FGF) and comparison with the amino-terminal sequence of bovine brain acidic FGF. Proceedings of the National Academy of Sciences of the United States of America 82, 6507-6511 (1985). 7. Folkman, J. & Klagsbrun, M. Angiogenic factors. Science (New York, N.Y 235, 442-447 (1987). 8. Gaudric, A., N'Guyen, T., Moenner, M., Glacet-Bernard, A. & Barritault, D. Quantification of angiogenesis due to basic fibroblast growth factor in a modified rabbit corneal model. Ophthalmic research 24, 181-188 (1992). 9. Montesano, R., Vassalli, J.D., Baird, A., Guillemin, R. & Orci, L. Basic fibroblast growth factor induces angiogenesis in vitro. Proceedings of the National Academy of Sciences of the United States of America 83, 7297-7301 (1986). 10. Slavin, J. Fibroblast growth factors: at the heart of angiogenesis. Cell biology international 19, 431-444 (1995). 11. Presta, M. et al. Fibroblast growth factor/fibroblast growth factor receptor system in angiogenesis. Cytokine & growth factor reviews 16, 159-178 (2005). 12. Miller, D.L., Ortega, S., Bashayan, O., Basch, R. & Basilico, C. Compensation by fibroblast growth factor 1 (FGF1) does not account for the mild phenotypic defects observed in FGF2 null mice. Molecular and cellular biology 20, 2260-2268 (2000). 13. Ortega, S., Ittmann, M., Tsang, S.H., Ehrlich, M. & Basilico, C. Neuronal defects and delayed wound healing in mice lacking fibroblast growth factor 2. Proceedings of the National Academy of Sciences of the United States of America 95, 5672-5677 (1998). 14. Dono, R., Texido, G., Dussel, R., Ehmke, H. & Zeller, R. Impaired cerebral cortex development and blood pressure regulation in FGF-2-deficient mice. The EMBO journal 17, 4213-4225 (1998). 15. Mellin, T.N. et al. Acidic fibroblast growth factor accelerates dermal wound healing in diabetic mice. The Journal of investigative dermatology 104, 850-855 (1995). 16. Okumura, M., Okuda, T., Nakamura, T. & Yajima, M. Acceleration of wound healing in diabetic mice by basic fibroblast growth factor. Biological & pharmaceutical bulletin 19, 530- 535 (1996). 17. Bennett, N.T. & Schultz, G.S. Growth factors and wound healing: biochemical properties of growth factors and their receptors. American journal of surgery 165, 728-737 (1993). 18. Bennett, N.T. & Schultz, G.S. Growth factors and wound healing: Part II. Role in normal and chronic wound healing. American journal of surgery 166, 74-81 (1993). 19. Winkles, J.A., Friesel, R., Alberts, G.F., Janat, M.F. & Liau, G. Elevated expression of basic fibroblast growth factor in an immortalized rabbit smooth muscle cell line. The American journal of pathology 143, 518-527 (1993). 20. Shipley, G.D., Keeble, W.W., Hendrickson, J.E., Coffey, R.J., Jr. & Pittelkow, M.R. Growth of normal human keratinocytes and fibroblasts in serum-free medium is stimulated by acidic and basic fibroblast growth factor. Journal of cellular physiology 138, 511-518 (1989). 21. McNeil, P.L., Muthukrishnan, L., Warder, E. & D'Amore, P.A. Growth factors are released by mechanically wounded endothelial cells. The Journal of cell biology 109, 811-822 (1989). 22. Jackson, A. et al. Heat shock induces the release of fibroblast growth factor 1 from NIH 3T3 cells. Proceedings of the National Academy of Sciences of the United States of America 89, 10691-10695 (1992). 23. Mignatti, P., Morimoto, T. & Rifkin, D.B. Basic fibroblast growth factor, a protein devoid of secretory signal sequence, is released by cells via a pathway independent of the endoplasmic reticulum-Golgi complex. Journal of cellular physiology 151, 81-93 (1992). 142 Chapter 4

24. Cao, Y. & Pettersson, R.F. Release and subcellular localization of acidic fibroblast growth factor expressed to high levels in HeLa cells. Growth factors (Chur, Switzerland) 8, 277-290 (1993). 25. Rusnati, M. & Presta, M. Interaction of angiogenic basic fibroblast growth factor with endothelial cell heparan sulfate proteoglycans. Biological implications in neovascularization. International journal of clinical & laboratory research 26, 15-23 (1996). 26. Neufeld, G., Cohen, T., Gengrinovitch, S. & Poltorak, Z. Vascular endothelial growth factor (VEGF) and its receptors. Faseb J 13, 9-22 (1999). 27. Roy, H., Bhardwaj, S. & Yla-Herttuala, S. Biology of vascular endothelial growth factors. FEBS letters 580, 2879-2887 (2006). 28. Shalaby, F. et al. Failure of blood-island formation and vasculogenesis in Flk-1-deficient mice. Nature 376, 62-66 (1995). 29. Fong, G.H., Rossant, J., Gertsenstein, M. & Breitman, M.L. Role of the Flt-1 receptor tyrosine kinase in regulating the assembly of vascular endothelium. Nature 376, 66-70 (1995). 30. Kawasaki, T. et al. A requirement for neuropilin-1 in embryonic vessel formation. Development (Cambridge, England) 126, 4895-4902 (1999). 31. Ferrara, N. et al. Heterozygous embryonic lethality induced by targeted inactivation of the VEGF gene. Nature 380, 439-442 (1996). 32. Carmeliet, P. et al. Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature 380, 435-439 (1996). 33. Senger, D.R. et al. Tumor cells secrete a vascular permeability factor that promotes accumulation of ascites fluid. Science (New York, N.Y 219, 983-985 (1983). 34. Kohn, S., Nagy, J.A., Dvorak, H.F. & Dvorak, A.M. Pathways of macromolecular tracer transport across venules and small veins. Structural basis for the hyperpermeability of tumor blood vessels. Laboratory investigation; a journal of technical methods and pathology 67, 596-607 (1992). 35. Kevil, C.G., Payne, D.K., Mire, E. & Alexander, J.S. Vascular permeability factor/vascular endothelial cell growth factor-mediated permeability occurs through disorganization of endothelial junctional proteins. The Journal of biological chemistry 273, 15099-15103 (1998). 36. Esser, S., Lampugnani, M.G., Corada, M., Dejana, E. & Risau, W. Vascular endothelial growth factor induces VE-cadherin tyrosine phosphorylation in endothelial cells. Journal of cell science 111 ( Pt 13), 1853-1865 (1998). 37. Esser, S. et al. Vascular endothelial growth factor induces endothelial fenestrations in vitro. The Journal of cell biology 140, 947-959 (1998). 38. Kroll, J. & Waltenberger, J. A novel function of VEGF receptor-2 (KDR): rapid release of nitric oxide in response to VEGF-A stimulation in endothelial cells. Biochemical and biophysical research communications 265, 636-639 (1999). 39. Bootle-Wilbraham, C.A., Tazzyman, S., Thompson, W.D., Stirk, C.M. & Lewis, C.E. Fibrin fragment E stimulates the proliferation, migration and differentiation of human microvascular endothelial cells in vitro. Angiogenesis 4, 269-275 (2001). 40. Pepper, M.S., Ferrara, N., Orci, L. & Montesano, R. Vascular endothelial growth factor (VEGF) induces plasminogen activators and plasminogen activator inhibitor-1 in microvascular endothelial cells. Biochemical and biophysical research communications 181, 902-906 (1991). 41. Unemori, E.N., Ferrara, N., Bauer, E.A. & Amento, E.P. Vascular endothelial growth factor induces interstitial collagenase expression in human endothelial cells. Journal of cellular physiology 153, 557-562 (1992). 42. Burbridge, M.F. et al. The role of the matrix metalloproteinases during in vitro vessel formation. Angiogenesis 5, 215-226 (2002). 43. van Hinsbergh, V.W., Koolwijk, P. & Hanemaaijer, R. Role of fibrin and plasminogen activators in repair-associated angiogenesis: in vitro studies with human endothelial cells. Exs 79, 391-411 (1997). 44. Mandriota, S.J. et al. Vascular endothelial growth factor increases urokinase receptor expression in vascular endothelial cells. The Journal of biological chemistry 270, 9709-9716 (1995). 45. Waltz, D.A., Natkin, L.R., Fujita, R.M., Wei, Y. & Chapman, H.A. Plasmin and plasminogen activator inhibitor type 1 promote cellular motility by regulating the interaction between the urokinase receptor and vitronectin. The Journal of clinical investigation 100, 58-67 (1997). 46. Nguyen, M., Arkell, J. & Jackson, C.J. Human endothelial gelatinases and angiogenesis. The international journal of biochemistry & cell biology 33, 960-970 (2001). Introduction: Angiogenesis 143

47. Prager, G.W., Breuss, J.M., Steurer, S., Mihaly, J. & Binder, B.R. Vascular endothelial growth factor (VEGF) induces rapid prourokinase (pro-uPA) activation on the surface of endothelial cells. Blood 103, 955-962 (2004). 48. Rousseau, S., Houle, F. & Huot, J. Integrating the VEGF signals leading to actin-based motility in vascular endothelial cells. Trends in cardiovascular medicine 10, 321-327 (2000). 49. Fujio, Y. & Walsh, K. Akt mediates cytoprotection of endothelial cells by vascular endothelial growth factor in an anchorage-dependent manner. The Journal of biological chemistry 274, 16349-16354 (1999). 50. Gupta, K. et al. VEGF prevents apoptosis of human microvascular endothelial cells via opposing effects on MAPK/ERK and SAPK/JNK signaling. Experimental cell research 247, 495-504 (1999). 51. Gerber, H.P. et al. Vascular endothelial growth factor regulates endothelial cell survival through the phosphatidylinositol 3'-kinase/Akt signal transduction pathway. Requirement for Flk-1/KDR activation. The Journal of biological chemistry 273, 30336-30343 (1998). 52. Ferrara, N. & Davis-Smyth, T. The biology of vascular endothelial growth factor. Endocrine reviews 18, 4-25 (1997). 53. Pepper, M.S., Ferrara, N., Orci, L. & Montesano, R. Potent synergism between vascular endothelial growth factor and basic fibroblast growth factor in the induction of angiogenesis in vitro. Biochemical and biophysical research communications 189, 824-831 (1992). 54. Leung, D.W., Cachianes, G., Kuang, W.J., Goeddel, D.V. & Ferrara, N. Vascular endothelial growth factor is a secreted angiogenic mitogen. Science (New York, N.Y 246, 1306-1309 (1989). 55. Phillips, G.D. et al. Vascular endothelial growth factor (rhVEGF165) stimulates direct angiogenesis in the rabbit cornea. In vivo (Athens, Greece) 8, 961-965 (1994). 56. Gerhardt, H. et al. VEGF guides angiogenic sprouting utilizing endothelial tip cell filopodia. The Journal of cell biology 161, 1163-1177 (2003). 57. Rafii, S., Heissig, B. & Hattori, K. Efficient mobilization and recruitment of marrow-derived endothelial and hematopoietic stem cells by adenoviral vectors expressing angiogenic factors. Gene therapy 9, 631-641 (2002). 58. Rabbany, S.Y., Heissig, B., Hattori, K. & Rafii, S. Molecular pathways regulating mobilization of marrow-derived stem cells for tissue revascularization. Trends in molecular medicine 9, 109-117 (2003). 59. Jain, R.K. Molecular regulation of vessel maturation. Nature medicine 9, 685-693 (2003). 60. Carmeliet, P. VEGF as a key mediator of angiogenesis in cancer. Oncology 69 Suppl 3, 4-10 (2005). 61. Forsythe, J.A. et al. Activation of vascular endothelial growth factor gene transcription by hypoxia-inducible factor 1. Molecular and cellular biology 16, 4604-4613 (1996). 62. Mole, D.R., Maxwell, P.H., Pugh, C.W. & Ratcliffe, P.J. Regulation of HIF by the von Hippel-Lindau tumour suppressor: implications for cellular oxygen sensing. IUBMB life 52, 43-47 (2001). 63. Jaakkola, P. et al. Targeting of HIF-alpha to the von Hippel-Lindau ubiquitylation complex by O2-regulated prolyl hydroxylation. Science (New York, N.Y 292, 468-472 (2001). 64. Na, X. et al. Overproduction of vascular endothelial growth factor related to von Hippel- Lindau tumor suppressor gene mutations and hypoxia-inducible factor-1 alpha expression in renal cell carcinomas. The Journal of urology 170, 588-592 (2003). 65. Ferrara, N. Vascular endothelial growth factor: basic science and clinical progress. Endocrine reviews 25, 581-611 (2004). 66. Persico, M.G., Vincenti, V. & DiPalma, T. Structure, expression and receptor-binding properties of placenta growth factor (PlGF). Current topics in microbiology and immunology 237, 31-40 (1999). 67. Yang, W., Ahn, H., Hinrichs, M., Torry, R.J. & Torry, D.S. Evidence of a novel isoform of placenta growth factor (PlGF-4) expressed in human trophoblast and endothelial cells. Journal of reproductive immunology 60, 53-60 (2003). 68. DiPalma, T. et al. The placenta growth factor gene of the mouse. Mamm Genome 7, 6-12 (1996). 69. Luttun, A. et al. Revascularization of ischemic tissues by PlGF treatment, and inhibition of tumor angiogenesis, arthritis and atherosclerosis by anti-Flt1. Nature medicine 8, 831-840 (2002). 144 Chapter 4

70. Marcellini, M. et al. Increased melanoma growth and metastasis spreading in mice overexpressing placenta growth factor. The American journal of pathology 169, 643-654 (2006). 71. Fischer, C. et al. Anti-PlGF inhibits growth of VEGF(R)-inhibitor-resistant tumors without affecting healthy vessels. Cell 131, 463-475 (2007). 72. Adini, A., Kornaga, T., Firoozbakht, F. & Benjamin, L.E. Placental growth factor is a survival factor for tumor endothelial cells and macrophages. Cancer research 62, 2749-2752 (2002). 73. Pipp, F. et al. VEGFR-1-selective VEGF homologue PlGF is arteriogenic: evidence for a monocyte-mediated mechanism. Circulation research 92, 378-385 (2003). 74. Roy, H. et al. Adenovirus-mediated gene transfer of placental growth factor to perivascular tissue induces angiogenesis via upregulation of the expression of endogenous vascular endothelial growth factor-A. Human gene therapy 16, 1422-1428 (2005). 75. Carmeliet, P. et al. Synergism between vascular endothelial growth factor and placental growth factor contributes to angiogenesis and plasma extravasation in pathological conditions. Nature medicine 7, 575-583 (2001). 76. Luttun, A. et al. Loss of placental growth factor protects mice against vascular permeability in pathological conditions. Biochemical and biophysical research communications 295, 428-434 (2002). 77. Autiero, M. et al. Role of PlGF in the intra- and intermolecular cross talk between the VEGF receptors Flt1 and Flk1. Nature medicine 9, 936-943 (2003). 78. Eriksson, A. et al. Placenta growth factor-1 antagonizes VEGF-induced angiogenesis and tumor growth by the formation of functionally inactive PlGF-1/VEGF heterodimers. Cancer cell 1, 99-108 (2002). 79. Valenzuela, D.M. et al. Angiopoietins 3 and 4: diverging gene counterparts in mice and humans. Proceedings of the National Academy of Sciences of the United States of America 96, 1904-1909 (1999). 80. Shim, W.S., Ho, I.A. & Wong, P.E. Angiopoietin: a TIE(d) balance in tumor angiogenesis. Mol Cancer Res 5, 655-665 (2007). 81. Suri, C. et al. Increased vascularization in mice overexpressing angiopoietin-1. Science (New York, N.Y 282, 468-471 (1998). 82. Thurston, G. et al. Leakage-resistant blood vessels in mice transgenically overexpressing angiopoietin-1. Science (New York, N.Y 286, 2511-2514 (1999). 83. Koblizek, T.I., Weiss, C., Yancopoulos, G.D., Deutsch, U. & Risau, W. Angiopoietin-1 induces sprouting angiogenesis in vitro. Curr Biol 8, 529-532 (1998). 84. Papapetropoulos, A. et al. Angiopoietin-1 inhibits endothelial cell apoptosis via the Akt/survivin pathway. The Journal of biological chemistry 275, 9102-9105 (2000). 85. Scharpfenecker, M., Fiedler, U., Reiss, Y. & Augustin, H.G. The Tie-2 ligand angiopoietin-2 destabilizes quiescent endothelium through an internal autocrine loop mechanism. Journal of cell science 118, 771-780 (2005). 86. Sato, T.N. et al. Distinct roles of the receptor tyrosine kinases Tie-1 and Tie-2 in blood vessel formation. Nature 376, 70-74 (1995). 87. Suri, C. et al. Requisite role of angiopoietin-1, a ligand for the TIE2 receptor, during embryonic angiogenesis. Cell 87, 1171-1180 (1996). 88. Maisonpierre, P.C. et al. Angiopoietin-2, a natural antagonist for Tie2 that disrupts in vivo angiogenesis. Science (New York, N.Y 277, 55-60 (1997). 89. Holash, J. et al. Vessel cooption, regression, and growth in tumors mediated by angiopoietins and VEGF. Science (New York, N.Y 284, 1994-1998 (1999). 90. Kim, I., Kim, J.H., Ryu, Y.S., Liu, M. & Koh, G.Y. Tumor necrosis factor-alpha upregulates angiopoietin-2 in human umbilical vein endothelial cells. Biochemical and biophysical research communications 269, 361-365 (2000). 91. Oh, H. et al. Hypoxia and vascular endothelial growth factor selectively up-regulate angiopoietin-2 in bovine microvascular endothelial cells. The Journal of biological chemistry 274, 15732-15739 (1999). 92. Enholm, B. et al. Comparison of VEGF, VEGF-B, VEGF-C and Ang-1 mRNA regulation by serum, growth factors, oncoproteins and hypoxia. Oncogene 14, 2475-2483 (1997). 93. Senger, D.R. et al. Angiogenesis promoted by vascular endothelial growth factor: regulation through alpha1beta1 and alpha2beta1 integrins. Proceedings of the National Academy of Sciences of the United States of America 94, 13612-13617 (1997). 94. Brooks, P.C., Clark, R.A. & Cheresh, D.A. Requirement of vascular integrin alpha v beta 3 for angiogenesis. Science (New York, N.Y 264, 569-571 (1994). Introduction: Angiogenesis 145

95. Stromblad, S. & Cheresh, D.A. Cell adhesion and angiogenesis. Trends in cell biology 6, 462- 468 (1996). 96. Kim, S., Bell, K., Mousa, S.A. & Varner, J.A. Regulation of angiogenesis in vivo by ligation of integrin alpha5beta1 with the central cell-binding domain of fibronectin. The American journal of pathology 156, 1345-1362 (2000). 97. Senger, D.R. et al. The alpha(1)beta(1) and alpha(2)beta(1) integrins provide critical support for vascular endothelial growth factor signaling, endothelial cell migration, and tumor angiogenesis. The American journal of pathology 160, 195-204 (2002). 98. Friedlander, M. et al. Definition of two angiogenic pathways by distinct alpha v integrins. Science (New York, N.Y 270, 1500-1502 (1995). 99. Brooks, P.C. et al. Integrin alpha v beta 3 antagonists promote tumor regression by inducing apoptosis of angiogenic blood vessels. Cell 79, 1157-1164 (1994). 100. Eliceiri, B.P. & Cheresh, D.A. The role of alphav integrins during angiogenesis. Molecular medicine (Cambridge, Mass 4, 741-750 (1998). 101. Giancotti, F.G. Complexity and specificity of integrin signalling. Nature cell biology 2, E13- 14 (2000). 102. Carlson, T.R., Feng, Y., Maisonpierre, P.C., Mrksich, M. & Morla, A.O. Direct cell adhesion to the angiopoietins mediated by integrins. The Journal of biological chemistry 276, 26516- 26525 (2001). 103. Hutchings, H., Ortega, N. & Plouet, J. Extracellular matrix-bound vascular endothelial growth factor promotes endothelial cell adhesion, migration, and survival through integrin ligation. Faseb J 17, 1520-1522 (2003). 104. Brooks, P.C. et al. Localization of matrix metalloproteinase MMP-2 to the surface of invasive cells by interaction with integrin alpha v beta 3. Cell 85, 683-693 (1996). 105. Soldi, R. et al. Role of alphavbeta3 integrin in the activation of vascular endothelial growth factor receptor-2. The EMBO journal 18, 882-892 (1999). 106. Wang, H.U., Chen, Z.F. & Anderson, D.J. Molecular distinction and angiogenic interaction between embryonic arteries and veins revealed by ephrin-B2 and its receptor Eph-B4. Cell 93, 741-753 (1998). 107. Gerety, S.S., Wang, H.U., Chen, Z.F. & Anderson, D.J. Symmetrical mutant phenotypes of the receptor EphB4 and its specific transmembrane ligand ephrin-B2 in cardiovascular development. Molecular cell 4, 403-414 (1999). 108. Zhang, X.Q. et al. Stromal cells expressing ephrin-B2 promote the growth and sprouting of ephrin-B2(+) endothelial cells. Blood 98, 1028-1037 (2001). 109. Noren, N.K., Lu, M., Freeman, A.L., Koolpe, M. & Pasquale, E.B. Interplay between EphB4 on tumor cells and vascular ephrin-B2 regulates tumor growth. Proceedings of the National Academy of Sciences of the United States of America 101, 5583-5588 (2004). 110. Kumar, S.R. et al. Receptor tyrosine kinase EphB4 is a survival factor in breast cancer. The American journal of pathology 169, 279-293 (2006). 111. Holzman, L.B., Marks, R.M. & Dixit, V.M. A novel immediate-early response gene of endothelium is induced by cytokines and encodes a secreted protein. Molecular and cellular biology 10, 5830-5838 (1990). 112. Cheng, N. et al. Blockade of EphA receptor tyrosine kinase activation inhibits vascular endothelial cell growth factor-induced angiogenesis. Mol Cancer Res 1, 2-11 (2002). 113. Hellstrom, M., Kalen, M., Lindahl, P., Abramsson, A. & Betsholtz, C. Role of PDGF-B and PDGFR-beta in recruitment of vascular smooth muscle cells and pericytes during embryonic blood vessel formation in the mouse. Development (Cambridge, England) 126, 3047-3055 (1999). 114. Oikawa, T., Onozawa, C., Sakaguchi, M., Morita, I. & Murota, S. Three isoforms of platelet- derived growth factors all have the capability to induce angiogenesis in vivo. Biological & pharmaceutical bulletin 17, 1686-1688 (1994). 115. Battegay, E.J., Rupp, J., Iruela-Arispe, L., Sage, E.H. & Pech, M. PDGF-BB modulates endothelial proliferation and angiogenesis in vitro via PDGF beta-receptors. The Journal of cell biology 125, 917-928 (1994). 116. Pepper, M.S. Transforming growth factor-beta: vasculogenesis, angiogenesis, and vessel wall integrity. Cytokine & growth factor reviews 8, 21-43 (1997). 117. Chambers, R.C., Leoni, P., Kaminski, N., Laurent, G.J. & Heller, R.A. Global expression profiling of fibroblast responses to transforming growth factor-beta1 reveals the induction of inhibitor of differentiation-1 and provides evidence of smooth muscle cell phenotypic switching. The American journal of pathology 162, 533-546 (2003). 146 Chapter 4

118. Carmeliet, P. Angiogenesis in health and disease. Nature medicine 9, 653-660 (2003). 119. Risau, W. Angiogenesis is coming of age. Circulation research 82, 926-928 (1998). 120. Clauss, M. et al. A permissive role for tumor necrosis factor in vascular endothelial growth factor-induced vascular permeability. Blood 97, 1321-1329 (2001). 121. Fiedler, U. et al. Angiopoietin-2 sensitizes endothelial cells to TNF-alpha and has a crucial role in the induction of inflammation. Nature medicine 12, 235-239 (2006). 122. Leibovich, S.J. et al. Macrophage-induced angiogenesis is mediated by tumour necrosis factor-alpha. Nature 329, 630-632 (1987). 123. Frater-Schroder, M., Risau, W., Hallmann, R., Gautschi, P. & Bohlen, P. Tumor necrosis factor type alpha, a potent inhibitor of endothelial cell growth in vitro, is angiogenic in vivo. Proceedings of the National Academy of Sciences of the United States of America 84, 5277- 5281 (1987). 124. Robaye, B., Mosselmans, R., Fiers, W., Dumont, J.E. & Galand, P. Tumor necrosis factor induces apoptosis (programmed cell death) in normal endothelial cells in vitro. The American journal of pathology 138, 447-453 (1991). 125. Li, J.H. & Pober, J.S. The cathepsin B death pathway contributes to TNF plus IFN-gamma- mediated human endothelial injury. J Immunol 175, 1858-1866 (2005). 126. Good, D.J. et al. A tumor suppressor-dependent inhibitor of angiogenesis is immunologically and functionally indistinguishable from a fragment of thrombospondin. Proceedings of the National Academy of Sciences of the United States of America 87, 6624-6628 (1990). 127. Iruela-Arispe, M.L., Bornstein, P. & Sage, H. Thrombospondin exerts an antiangiogenic effect on cord formation by endothelial cells in vitro. Proceedings of the National Academy of Sciences of the United States of America 88, 5026-5030 (1991). 128. Volpert, O.V. et al. Inhibition of angiogenesis by thrombospondin-2. Biochemical and biophysical research communications 217, 326-332 (1995). 129. Varner, J.A. The sticky truth about angiogenesis and thrombospondins. The Journal of clinical investigation 116, 3111-3113 (2006). 130. Bagavandoss, P. & Wilks, J.W. Specific inhibition of endothelial cell proliferation by thrombospondin. Biochemical and biophysical research communications 170, 867-872 (1990). 131. Jimenez, B. et al. Signals leading to apoptosis-dependent inhibition of neovascularization by thrombospondin-1. Nature medicine 6, 41-48 (2000). 132. Iruela-Arispe, M.L., Lombardo, M., Krutzsch, H.C., Lawler, J. & Roberts, D.D. Inhibition of angiogenesis by thrombospondin-1 is mediated by 2 independent regions within the type 1 repeats. Circulation 100, 1423-1431 (1999). 133. O'Reilly, M.S. et al. Angiostatin: a novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma. Cell 79, 315-328 (1994). 134. O'Reilly, M.S. et al. Endostatin: an endogenous inhibitor of angiogenesis and tumor growth. Cell 88, 277-285 (1997). 135. Maeshima, Y. et al. Distinct antitumor properties of a type IV collagen domain derived from basement membrane. The Journal of biological chemistry 275, 21340-21348 (2000). 136. Kamphaus, G.D. et al. Canstatin, a novel matrix-derived inhibitor of angiogenesis and tumor growth. The Journal of biological chemistry 275, 1209-1215 (2000). 137. Ramchandran, R. et al. Antiangiogenic activity of restin, NC10 domain of human collagen XV: comparison to endostatin. Biochemical and biophysical research communications 255, 735-739 (1999). 138. Colorado, P.C. et al. Anti-angiogenic cues from vascular basement membrane collagen. Cancer research 60, 2520-2526 (2000). 139. Pike, S.E. et al. Vasostatin, a calreticulin fragment, inhibits angiogenesis and suppresses tumor growth. The Journal of experimental medicine 188, 2349-2356 (1998). 140. Brooks, P.C., Silletti, S., von Schalscha, T.L., Friedlander, M. & Cheresh, D.A. Disruption of angiogenesis by PEX, a noncatalytic metalloproteinase fragment with integrin binding activity. Cell 92, 391-400 (1998). 141. Cornelius, L.A. et al. Matrix metalloproteinases generate angiostatin: effects on neovascularization. J Immunol 161, 6845-6852 (1998). 142. Claesson-Welsh, L. et al. Angiostatin induces endothelial cell apoptosis and activation of focal adhesion kinase independently of the integrin-binding motif RGD. Proceedings of the National Academy of Sciences of the United States of America 95, 5579-5583 (1998). 143. O'Reilly, M.S., Holmgren, L., Chen, C. & Folkman, J. Angiostatin induces and sustains dormancy of human primary tumors in mice. Nature medicine 2, 689-692 (1996). Introduction: Angiogenesis 147

144. Moser, T.L. et al. Angiostatin binds ATP synthase on the surface of human endothelial cells. Proceedings of the National Academy of Sciences of the United States of America 96, 2811- 2816 (1999). 145. Tarui, T., Miles, L.A. & Takada, Y. Specific interaction of angiostatin with integrin alpha(v)beta(3) in endothelial cells. The Journal of biological chemistry 276, 39562-39568 (2001). 146. Wajih, N. & Sane, D.C. Angiostatin selectively inhibits signaling by hepatocyte growth factor in endothelial and smooth muscle cells. Blood 101, 1857-1863 (2003). 147. Heljasvaara, R. et al. Generation of biologically active endostatin fragments from human collagen XVIII by distinct matrix metalloproteases. Experimental cell research 307, 292-304 (2005). 148. Felbor, U. et al. Secreted cathepsin L generates endostatin from collagen XVIII. The EMBO journal 19, 1187-1194 (2000). 149. Digtyar, A.V., Pozdnyakova, N.V., Feldman, N.B., Lutsenko, S.V. & Severin, S.E. Endostatin: current concepts about its biological role and mechanisms of action. Biochemistry 72, 235-246 (2007). 150. Sudhakar, A. et al. Human tumstatin and human endostatin exhibit distinct antiangiogenic activities mediated by alpha v beta 3 and alpha 5 beta 1 integrins. Proceedings of the National Academy of Sciences of the United States of America 100, 4766-4771 (2003). 151. Magnon, C. et al. Canstatin acts on endothelial and tumor cells via mitochondrial damage initiated through interaction with alphavbeta3 and alphavbeta5 integrins. Cancer research 65, 4353-4361 (2005). 152. Sudhakar, A. et al. Human alpha1 type IV collagen NC1 domain exhibits distinct antiangiogenic activity mediated by alpha1beta1 integrin. The Journal of clinical investigation 115, 2801-2810 (2005). 153. Johnson, M.D. et al. Inhibition of angiogenesis by tissue inhibitor of metalloproteinase. Journal of cellular physiology 160, 194-202 (1994). 154. Fernandez, H.A. et al. Inhibition of endothelial cell migration by gene transfer of tissue inhibitor of metalloproteinases-1. The Journal of surgical research 82, 156-162 (1999). 155. Murphy, A.N., Unsworth, E.J. & Stetler-Stevenson, W.G. Tissue inhibitor of metalloproteinases-2 inhibits bFGF-induced human microvascular endothelial cell proliferation. Journal of cellular physiology 157, 351-358 (1993). 156. Wang, Z., Juttermann, R. & Soloway, P.D. TIMP-2 is required for efficient activation of proMMP-2 in vivo. The Journal of biological chemistry 275, 26411-26415 (2000). 157. Rundhaug, J.E. Matrix metalloproteinases and angiogenesis. Journal of cellular and molecular medicine 9, 267-285 (2005). 158. Gagnon, M.L. et al. Identification of a natural soluble neuropilin-1 that binds vascular endothelial growth factor: In vivo expression and antitumor activity. Proceedings of the National Academy of Sciences of the United States of America 97, 2573-2578 (2000). 159. McCarthy, M.J. et al. Potential roles of metalloprotease mediated ectodomain cleavage in signaling by the endothelial receptor tyrosine kinase Tie-1. Laboratory investigation; a journal of technical methods and pathology 79, 889-895 (1999). 160. Reusch, P. et al. Identification of a soluble form of the angiopoietin receptor TIE-2 released from endothelial cells and present in human blood. Angiogenesis 4, 123-131 (2001). 161. Kendall, R.L. & Thomas, K.A. Inhibition of vascular endothelial cell growth factor activity by an endogenously encoded soluble receptor. Proceedings of the National Academy of Sciences of the United States of America 90, 10705-10709 (1993). 162. Gengrinovitch, S. et al. Platelet factor-4 inhibits the mitogenic activity of VEGF121 and VEGF165 using several concurrent mechanisms. The Journal of biological chemistry 270, 15059-15065 (1995). 163. Sato, Y., Abe, M. & Takaki, R. Platelet factor 4 blocks the binding of basic fibroblast growth factor to the receptor and inhibits the spontaneous migration of vascular endothelial cells. Biochemical and biophysical research communications 172, 595-600 (1990). 164. Zhai, Y. et al. VEGI, a novel cytokine of the tumor necrosis factor family, is an angiogenesis inhibitor that suppresses the growth of colon carcinomas in vivo. Faseb J 13, 181-189 (1999). 165. Voest, E.E. et al. Inhibition of angiogenesis in vivo by interleukin 12. J Natl Cancer Inst 87, 581-586 (1995). 166. Strieter, R.M., Kunkel, S.L., Arenberg, D.A., Burdick, M.D. & Polverini, P.J. Interferon gamma-inducible protein 10 (IP-10), a member of the C-X-C chemokine family, is an 148 Chapter 4

inhibitor of angiogenesis. Biochemical and biophysical research communications 210, 51-57 (1995). 167. Angiolillo, A.L. et al. Human interferon-inducible protein 10 is a potent inhibitor of angiogenesis in vivo. The Journal of experimental medicine 182, 155-162 (1995). 168. Bodnar, R.J., Yates, C.C. & Wells, A. IP-10 blocks vascular endothelial growth factor- induced endothelial cell motility and tube formation via inhibition of calpain. Circulation research 98, 617-625 (2006). 169. Defilippi, P. et al. Tumor necrosis factor alpha and interferon gamma modulate the expression of the vitronectin receptor (integrin beta 3) in human endothelial cells. The Journal of biological chemistry 266, 7638-7645 (1991). 170. Ruegg, C. et al. Evidence for the involvement of endothelial cell integrin alphaVbeta3 in the disruption of the tumor vasculature induced by TNF and IFN-gamma. Nature medicine 4, 408- 414 (1998). 171. Folkman, J. Tumor angiogenesis: therapeutic implications. The New England journal of medicine 285, 1182-1186 (1971). 172. Folkman, J. Angiogenesis and apoptosis. Seminars in cancer biology 13, 159-167 (2003). 173. Relf, M. et al. Expression of the angiogenic factors vascular endothelial cell growth factor, acidic and basic fibroblast growth factor, tumor growth factor beta-1, platelet-derived endothelial cell growth factor, placenta growth factor, and pleiotrophin in human primary breast cancer and its relation to angiogenesis. Cancer research 57, 963-969 (1997). 174. Rak, J., Yu, J.L., Klement, G. & Kerbel, R.S. Oncogenes and angiogenesis: signaling three- dimensional tumor growth. The journal of investigative dermatology. Symposium proceedings / the Society for Investigative Dermatology, Inc 5, 24-33 (2000). 175. Kerbel, R. & Folkman, J. Clinical translation of angiogenesis inhibitors. Nat Rev Cancer 2, 727-739 (2002). 176. Fukumura, D. et al. Tumor induction of VEGF promoter activity in stromal cells. Cell 94, 715-725 (1998). 177. Folkman, J. Angiogenesis: an organizing principle for drug discovery? Nat Rev Drug Discov 6, 273-286 (2007). 178. Yin, A.H. et al. AC133, a novel marker for human hematopoietic stem and progenitor cells. Blood 90, 5002-5012 (1997). 179. Jiang, Y. et al. Pluripotency of mesenchymal stem cells derived from adult marrow. Nature 418, 41-49 (2002). 180. Reyes, M. et al. Origin of endothelial progenitors in human postnatal bone marrow. The Journal of clinical investigation 109, 337-346 (2002). 181. Hristov, M., Erl, W. & Weber, P.C. Endothelial progenitor cells: isolation and characterization. Trends in cardiovascular medicine 13, 201-206 (2003). 182. Bauer, S.M., Bauer, R.J. & Velazquez, O.C. Angiogenesis, vasculogenesis, and induction of healing in chronic wounds. Vascular and endovascular surgery 39, 293-306 (2005). 183. Janowska-Wieczorek, A. et al. Growth factors and cytokines upregulate gelatinase expression in bone marrow CD34(+) cells and their transmigration through reconstituted basement membrane. Blood 93, 3379-3390 (1999). 184. Rafii, S. et al. Angiogenic factors reconstitute hematopoiesis by recruiting stem cells from bone marrow microenvironment. Annals of the New York Academy of Sciences 996, 49-60 (2003). 185. Hattori, K. et al. Placental growth factor reconstitutes hematopoiesis by recruiting VEGFR1(+) stem cells from bone-marrow microenvironment. Nature medicine 8, 841-849 (2002). 186. Heissig, B., Werb, Z., Rafii, S. & Hattori, K. Role of c-kit/Kit ligand signaling in regulating vasculogenesis. Thrombosis and haemostasis 90, 570-576 (2003). 187. Asahara, T. et al. VEGF contributes to postnatal neovascularization by mobilizing bone marrow-derived endothelial progenitor cells. The EMBO journal 18, 3964-3972 (1999). 188. Hattori, K. et al. Vascular endothelial growth factor and angiopoietin-1 stimulate postnatal hematopoiesis by recruitment of vasculogenic and hematopoietic stem cells. The Journal of experimental medicine 193, 1005-1014 (2001). 189. Lyden, D. et al. Id1 and Id3 are required for neurogenesis, angiogenesis and vascularization of tumour xenografts. Nature 401, 670-677 (1999). 190. Lyden, D. et al. Impaired recruitment of bone-marrow-derived endothelial and hematopoietic precursor cells blocks tumor angiogenesis and growth. Nature medicine 7, 1194-1201 (2001). Introduction: Angiogenesis 149

191. Kaplan, R.N. et al. VEGFR1-positive haematopoietic bone marrow progenitors initiate the pre-metastatic niche. Nature 438, 820-827 (2005). 192. Davidoff, A.M. et al. Bone marrow-derived cells contribute to tumor neovasculature and, when modified to express an angiogenesis inhibitor, can restrict tumor growth in mice. Clin Cancer Res 7, 2870-2879 (2001). 193. Nagano, N., Sasaki, H., Aoyagi, M. & Hirakawa, K. Invasion of experimental rat brain tumor: early morphological changes following microinjection of C6 glioma cells. Acta neuropathologica 86, 117-125 (1993). 194. Pezzella, F. et al. Non-small-cell lung carcinoma tumor growth without morphological evidence of neo-angiogenesis. The American journal of pathology 151, 1417-1423 (1997). 195. Kusters, B. et al. Vascular endothelial growth factor-A(165) induces progression of melanoma brain metastases without induction of sprouting angiogenesis. Cancer research 62, 341-345 (2002). 196. Chang, Y.S. et al. Mosaic blood vessels in tumors: frequency of cancer cells in contact with flowing blood. Proceedings of the National Academy of Sciences of the United States of America 97, 14608-14613 (2000). 197. Folkman, J. Can mosaic tumor vessels facilitate molecular diagnosis of cancer? Proceedings of the National Academy of Sciences of the United States of America 98, 398-400 (2001). 198. Maniotis, A.J. et al. Vascular channel formation by human melanoma cells in vivo and in vitro: vasculogenic mimicry. The American journal of pathology 155, 739-752 (1999). 199. Folberg, R., Hendrix, M.J. & Maniotis, A.J. Vasculogenic mimicry and tumor angiogenesis. The American journal of pathology 156, 361-381 (2000). 200. Bittner, M. et al. Molecular classification of cutaneous malignant melanoma by gene expression profiling. Nature 406, 536-540 (2000). 201. Hendrix, M.J. et al. Expression and functional significance of VE-cadherin in aggressive human melanoma cells: role in vasculogenic mimicry. Proceedings of the National Academy of Sciences of the United States of America 98, 8018-8023 (2001). 202. Seftor, R.E. et al. Cooperative interactions of laminin 5 gamma2 chain, matrix metalloproteinase-2, and membrane type-1-matrix/metalloproteinase are required for mimicry of embryonic vasculogenesis by aggressive melanoma. Cancer research 61, 6322-6327 (2001). 203. Hess, A.R. et al. VE-cadherin regulates EphA2 in aggressive melanoma cells through a novel signaling pathway: implications for vasculogenic mimicry. Cancer biology & therapy 5, 228- 233 (2006). 204. Passalidou, E. et al. Vascular phenotype in angiogenic and non-angiogenic lung non-small cell carcinomas. British journal of cancer 86, 244-249 (2002). 205. Shirakawa, K. et al. Vasculogenic mimicry and pseudo-comedo formation in breast cancer. International journal of cancer 99, 821-828 (2002). 206. Su, M. et al. Plasticity of ovarian cancer cell SKOV3ip and vasculogenic mimicry in vivo. Int J Gynecol Cancer (2007). 207. Zhang, S., Zhang, D. & Sun, B. Vasculogenic mimicry: current status and future prospects. Cancer letters 254, 157-164 (2007). 208. Helmlinger, G., Yuan, F., Dellian, M. & Jain, R.K. Interstitial pH and pO2 gradients in solid tumors in vivo: high-resolution measurements reveal a lack of correlation. Nature medicine 3, 177-182 (1997). 209. Hashizume, H. et al. Openings between defective endothelial cells explain tumor vessel leakiness. The American journal of pathology 156, 1363-1380 (2000). 210. Baluk, P., Morikawa, S., Haskell, A., Mancuso, M. & McDonald, D.M. Abnormalities of basement membrane on blood vessels and endothelial sprouts in tumors. The American journal of pathology 163, 1801-1815 (2003). 211. Jain, R.K. Normalizing tumor vasculature with anti-angiogenic therapy: a new paradigm for combination therapy. Nature medicine 7, 987-989 (2001).

Chapter 5 Vascular-targeting anticancer therapy

Introduction: Vascular-targeting anticancer therapy 153

5. VASCULAR-TARGETING ANTICANCER THERAPY

5.1. Introduction

The development of a tumor vasculature is essential for the survival, growth and metastasis of solid tumors. Folkman and colleagues postulated already in the early 1970s that it should be possible to block tumor growth by inhibiting tumor angiogenesis, and that, by targeting the endothelial cells of newly formed blood vessels, it should be possible to cause sustained regression of established tumors to a size of 1-2 mm in diameter where survival is possible without a blood vessel supply 1. Targeting the tumor vasculature has clear-cut advantages over classical chemotherapeutics that aim at directly killing the tumor cells: (1) One small tumor vessel, composed of only a limited number of endothelial cells, supports the growth of thousands of tumor cells, thus destroying one endothelial cell might lead to starvation and death of many tumor cells. (2) Destroying only a small part of a microvessel affects the functionality of the whole vessel and downstream vessels. Thus, even if the vascular-targeting therapy reaches only a small percentage of endothelial cells, it can still be 100% effective. (3) As endothelial cells are readily accessible via blood, the delivery of a vascular-targeting drug to its target cells is not expected to be a problem. (4) One of the major problems associated with conventional chemotherapeutics is the development of drug resistance. In contrast to tumor cells, endothelial cells are genetically stable, and therefore should not become resistant to therapy 2. However, the latter is only partially true. Several mechanisms for the development of resistance to anti-angiogenic therapy have been described 3. For example, an anti-angiogenic therapy based on the inhibition of one pro-angiogenic factor might favor the selection of tumor cells that produce another pro-angiogenic factor, making the therapy no longer effective. One major drawback of a vascular-targeting therapy is that it is expected not to lead to a complete eradication of all tumor cells. At the rim of a tumor, tumor cells may be present that do not depend on the tumor vasculature, but on pre-existing nearby vessels. Continuous treatment with anti-angiogenic drugs might prevent the regrowth of the tumor from these few remaining tumor cells. Such a therapeutic approach is called a dormancy-inducing therapy and is meant to control cancer in a chronic fashion. Another, particularly promising solution is to combine a vascular-targeting therapy with other treatments which appear to be most effective in the well-oxygenated rim of the tumor, such as conventional chemotherapy, radiotherapy, immunotherapy, or radioimmunotherapy (RIT) 4, 5.

The vascular-targeting anticancer drugs can be divided into three classes: angiogenesis inhibitors (AIs), vascular-disrupting agents (VDAs), and ligand-directed vascular-targeting agents (VTAs). AIs aim at inhibiting the formation of new blood vessels, whereas VDAs aim at damaging or disrupting established tumor vessels. AIs and VDAs are considered to be respectively cytostatic and cytotoxic for endothelial 154 Chapter 5

cells. However, the distinction between these two groups is not always so clear. Some AIs may also induce regression of established tumor vessels. Ligand-directed VTAs use antibodies or peptides to target agents such as toxins, cytokines, or radio-isotopes specifically to the tumor vasculature. As ligand-based VTAs are usually intended to damage tumor blood vessels, these agents may be considered as a subclass of VDAs 6, 7.

5.2. Angiogenesis inhibitors

Angiogenesis is essential for the growth of almost all solid tumors, while it is largely dispensable for the healthy adult body, making the angiogenic process an ideal target for cancer therapy. As the angiogenic cascade is regulated by a great variety of cytokines, receptors, enzymes and adhesion molecules, there is a vast number of possible targets for therapeutic intervention. However, the majority of AIs that have been developed somehow interfere with VEGF signaling. The VEGF pathway has been targeted by a variety of strategies, including monoclonal antibodies, such as bevacizumab, CDP791 and IMC-1121B, soluble receptors such as aflibercept (VEGF Trap), VEGF receptor tyrosine kinase inhibitors such as cediranib (AZD2171), semaxanib (SU5416), vatalanib (PTK787/ZK222584) and vandetanib (ZD6474), VEGF antisense oligonucleotides, and anti-VEGF ribozymes. In 2004, the first drug developed solely as an AI, bevacizumab (trade name Avastin), which is a monoclonal anti-VEGF antibody, was approved by the United States Food and Drug Administration (FDA) for use in the treatment of colorectal cancer in combination with standard chemotherapy 8. In 2005 bevacizumab was also approved by the European Medicine Agency (EMEA) for use in colorectal cancer. In animal models, anti-VEGF therapies not only efficiently block angiogenesis and growth of experimental tumors, but have also been shown to induce regression of existing tumor vessels by increasing endothelial cell death 9. In addition, anti-VEGF therapies may increase the efficacy of classical chemotherapeutic drugs or radiotherapy by ‘normalizing’ the tumor vasculature, mainly by reducing vascular permeability and IFP, and increasing the intratumoral oxygen levels 10, 11. Besides the VEGF pathway, other promising targets of AIs are other RTKs, MMPs, and integrins. Many of these AIs were shown to be effective in animal models of cancer and are currently evaluated in clinical trials. The most important AIs are summarized in Table 5.1.

Introduction: Vascular-targeting anticancer therapy 155

Table 5.1. Angiogenesis inhibitors in clinical development or practice.

Class / Inhibitor Mechanism of action Developmental phase, cancer type*

Endogenous inhibitors

Angiostatin integrin v3 antagonist, endothelial Phase II: NSCLC ATP synthase inhibitor§ § Endostatin integrin 51 antagonist Phase III: NSCLC IL-12 upregulates IFN- and CXCL10 Phase II: melanoma, lymphoma, multiple myeloma, ovarian, cervical, breast, prostate cancer, other cancers Antibodies

Abergrin (MEDI-522 or anti-v3 antibody Phase II: melanoma, prostate cancer Vitaxin) Bevacizumab anti-VEGF antibody FDA-approved for colorectal cancer, breast cancer, NSCLC

CDP791 pegylated F(ab’)2 antibody fragment Phase II: NSCLC directed against VEGF-R2 IMC-1121B anti-VEGF-R2 antibody Phase II: hepatocellular carcinoma, RCC, melanoma Other proteins and peptides A6 uPAR antagonist Phase II: ovarian cancer, peritoneal carcinoma ABT-510 TSP-1 mimetic, binds to CD36 Phase II: melanoma, lymphoma, head and neck cancer, soft-tissue sarcoma, RCC, NSCLC Aflibercept (VEGF Trap) VEGF decoy receptor comprised of Phase III: pancreatic, colorectal, ovarian, extracellular segments of VEGF-R1 prostatic cancer, NSCLC and VEGF-R2

ATN-161 integrin 51 antagonist Phase II: RCC

Cilengitide (EMD121974) integrin v3 and v5 antagonist Phase II: leukemia, glioma, glioblastoma, prostate cancer, melanoma Small molecule inhibitors 2-methoxyestradiol inhibits HIF-1 and tubulin Phase II: glioblastoma, multiple polymerization myeloma, RCC, prostatic, ovarian cancer, other carcinomas AMG706 multi-targeted RTK inhibitor Phase III: NSCLC Axitinib (AG-013736) multi-targeted RTK inhibitor Phase III: pancreatic ductal carcinoma BIBF1120 multi-targeted RTK inhibitor Phase II: ovarian, fallopian tube, peritoneal cavity cancer BMS-275291 MMP inhibitor Phase III: NSCLC Cediranib (AZD2171) VEGF-R RTK inhibitor Phase III: colorectal, lung, fallopian tube, ovarian, peritoneal cavity cancer

E7820 reduces integrin 2 subunit Phase II: colorectal cancer expression Erlotinib (OSI-774) EGF-R RTK inhibitor FDA-approved for NSCLC and pancreatic cancer Neovastat (AE-941) MMP inhibitor Phase III: NSCLC, kidney, breast, colorectal cancer Prinomastat (AG3340) MMP inhibitor Phase III: prostate cancer, NSCLC Semaxanib (SU5416) VEGF-R2 RTK inhibitor Phase III: colorectal cancer

156 Chapter 5

Table 5.1. (continued) Sorafenib (BAY 43-9006) multi-targeted RTK inhibitor, Raf FDA-approved for RCC inhibitor Suramin blocks binding of various growth Phase III: prostate cancer factors to their receptors Thalidomide inhibits TNF production, inhibits FDA-approved for multiple myeloma VEGF and bFGF secretion of and cellular responses 12 TKI-258 inhibitor of FGF-R3 and other RTKs Phase I: melanoma, multiple myeloma, AML Vandetanib (ZD6474) VEGF-R2 and EGF-R RTK inhibitor Phase III: NSCLC Vatalanib VEGF-R1, -R2 and PDGF-R RTK Phase III: colorectal cancer (PTK787/ ZK222584) inhibitor XL184 multi-targeted RTK inhibitor Phase II: NSCLC XL880 inhibitor of c-Met and other RTKs Phase II: RCC, gastric cancer, squamous cell carcinoma XL999 multi-targeted RTK inhibitor, Src Phase II: multiple myeloma, AML, NSCLC, inhibitor RCC, ovarian cancer, colorectal cancer * only the highest phase of clinical development is mentioned. For lower phase clinical trials see www.clinicaltrials.gov. AML: acute myeloid leukemia, RCC: renal cell carcinoma, NSCLC: non- small-cell lung cancer. Based on Folkman et al., 2007 13.

5.3. Vascular-disrupting agents

5.3.1. Tubulin-binding agents

Already in the 1930s it was found that the classic tubulin-binding agent colchicine had damaging effects on tumor vasculature and caused haemorrhage and necrosis of both animal and human tumors 14. However, its toxicity prevented any further clinical evaluation. Also the Vinca alkaloids vincristine and vinblastine, two other potent tubulin-binding agents, were found to have damaging effects on tumor vasculature 15- 17. Vincristine and vinblastine are currently used and evaluated in the clinic for the treatment of many different tumor types. However, these drugs only produce significant anti-vascular effects at doses close to their maximum tolerated dose (MTD). Therefore other, safer tubulin-binding agents were sought. The combretastatins were first isolated from the South African bush willow tree Combretum caffrum 18. A large number of natural and synthetic combretastatin analogs were tested for tubulin binding activity and cytotoxicity against a panel of tumor cells and the most active compound was combretastatin A4 (CA4) 19. CA4 has a similar structure as colchicine and binds to tubulin near the colchicine binding site, causing the destabilization of the tubulin polymers of the cytoskeleton 20-22. To increase solubility CA4 is produced as a CA4 phosphate (CA4P) prodrug 23. After administration, CA4P is cleaved by endogenous phosphatases to the active CA4. One of the first in vivo studies with CA4P showed rapid and irreversible vascular damage and haemorrhagic necrosis of a murine NT adenocarcinoma after a single injection of Introduction: Vascular-targeting anticancer therapy 157

the drug. In the same study it was shown that ex vivo perfusions with CA4P caused a rapid increase in vascular resistance in an isolated rat P22 carcinosarcoma, but not in normal rat hind limb 24. In the years thereafter, the tumor-specific anti-vascular effects of CA4P have been demonstrated in many other animal models as well as in man 25-27. Importantly, the vascular-damaging effects of CA4P can be achieved with doses of one-tenth of the MTD, offering a wide therapeutic window 24. CA4P is currently evaluated in phase I and II clinical trials for the treatment of solid tumors, particularly anaplastic thyroid cancer. Also other potent tubulin-binding VDAs are now being developed, including combretastatin A1 phosphate (OXi4503), the CA4 derivate AVE8062, and the colchicine derivate ZD6126 28-33. The anti-vascular effects of CA4P have been extensively studied. In vitro studies on HUVECs have shown that CA4P is a potent inhibitor of endothelial cell proliferation by arresting cells at the G2/M phase of mitosis, leading to mitotic cell death 34. By contrast, non-proliferating endothelial cells appear to be relatively resistant to CA4P- induced cytotoxicity. Also HUVECs cocultured with SMCs are protected against CA4P-induced cell death, possibly through stabilization of the endothelial cell-cell interactions 35. These results may at least in part explain the specificity of CA4P for the immature tumor vasculature. However, other tubulin-binding agents such as the Vinca alkaloids are also known for their anti-mitotic properties but only produce anti- vascular effects at doses close to their MTD, suggesting that other mechanisms may be involved in the vascular-damaging effect of CA4P. Indeed, Vincent and co- workers found that CA4P inhibits endothelial migration and capillary tube formation, and increases vascular permeability predominantly by interfering with the VE- cadherin/-catenin/Akt signaling pathway 36. CA4P induces the disruption of VE- cadherin-mediated endothelial cell-cell contact, resulting in increased vascular permeability, increased IFP, and eventually collapse of the weakened tumor vessels. In addition, the loss of endothelial cell-cell contact results in exposure of the already abnormal basement membrane, which in turn can result in activation of the coagulation cascade and subsequent thrombus formation 36, 37. Vincent and colleagues furthermore demonstrated that CA4P synergizes with a monoclonal anti-VE-cadherin antibody for inducing regression of nascent tumor vessels and necrosis of a murine B16 melanoma, with minimal toxicity and without affecting normal, stabilized vessels 35.

5.3.2. TNF and TNF-inducing flavonoids

It is well-established that the TNF-induced tumor necrosis is characterized by a selective destruction of the tumor vasculature (see chapter 1.7). The mechanism responsible for this remarkable effect is, however, not yet completely understood. Several possible mechanisms have been proposed, which may all add up to cause the in vivo vascular-disrupting effect of TNF: (1) One obvious candidate mechanism is the induction of endothelial cell death. Several reports have shown that TNF alone or 158 Chapter 5

in combination with IFN- is capable of triggering apoptotic or cathepsin-mediated cell death of cultured endothelial cells 38-41. (2) TNF may activate the coagulation cascade. TNF has been shown to inhibit the endothelial expression of the anti- coagulant protein thrombomodulin and induce the expression of tissue factor, the principal initiator of coagulation 42-45. Intravascular thrombus formation may lead to vessel occlusion, endothelial damage and haemorrhage 46. (3) It has also been suggested that the infiltration of leukocytes through the activated endothelium might cause endothelial damage 46. (4) Rüegg and collegues have reported that treatment of

HUVECs with TNF and IFN- suppresses the activation of integrin v3, leading to decreased endothelial cell adhesion and survival. They furthermore observed

detachment and apoptosis of v3-positive endothelial cells of the tumor vasculature in melanoma patients treated with TNF and IFN-(via ILP) 47. (5) Recent evidence has suggested that TNF may induce vessel disruption by a very similar mechanism as CA4P. It has been reported that TNF increases the permeability of endothelial cell monolayers by inducing microtubule destabilization and actin rearrangement in a p38- dependent manner, leading to disruption of VE-cadherin-based cell-cell junctions and the formation of intercellular gaps 48. TNF has also been shown to induce Fyn- dependent tyrosine phosphorylation of VE-cadherin, which is known to facilitate the formation of intercellular gaps 49. Menon and colleagues reported that TNF treatment of endothelial cells, cultured under hypoxia and with melanoma conditioned medium to mimic ILP conditions, caused disappearance of VE-cadherin from cell-cell junctions, resulting in the appearance of gaps between endothelial cells from 3 hours after treatment. They furthermore demonstrated in human melanoma xenografts raised in mice that TNF selectively increased tumor vascular permeability by 3 hours after treatment and decreased tumor blood flow by 6 hours after treatment. The formation of gaps between endothelial cells and the shutdown of tumor blood flow by 3 to 6 hours post-treatment was also observed in human melanoma patients after ILP with TNF plus melphalan 50.

In the late 1980s it was found that the flavonoid flavone acetic acid (FAA) had unexpectedly high antitumor activity in murine models. FAA induced haemorrhagic necrosis of various established murine tumors 51-53. Subsequent experiments showed that FAA caused a selective shutdown of the tumor blood flow 54-56. As FAA has some pharmacological drawbacks, FAA-derivates were developed with higher dose- potency. Of these compounds, 5,6-dimethylxanthenone-4-acetic acid (DMXAA) was found to be more than 10-fold more potent than FAA 57, 58. DMXAA is currently in phase I and II clinical trials for the treatment of various solid tumors. The exact mechanism of action of FAA and DMXAA is still poorly understood. Several reports have suggested that these flavonoids exert their rapid anti-vascular effects via the localized release of TNF from activated macrophages within the tumor tissue 59. DMXAA has been shown to strongly induce TNF mRNA upregulation in both human and murine cells of the macrophage/monocyte lineage 60. Furthermore, it has been Introduction: Vascular-targeting anticancer therapy 159

shown that the antitumor activities as well as the toxic side-effects of FAA and DMXAA can be inhibited with anti-TNF antibodies and are strongly attenuated in TNF-R1 knockout mice, respectively 61, 62. By contrast, Ching and co-workers reported that DMXAA is capable of inducing apoptosis of cultured endothelial cells in the absence of TNF mRNA upregulation, and they did not observe an induction of TNF in one patient treated with DMXAA, suggesting that DMXAA may also have TNF-independent anti-vascular effects 63.

5.4. Ligand-directed vascular-targeting agents

The strategy behind ligand-directed VTAs is to join ligands that bind selectively to components of tumor blood vessels and agents that are capable of damaging blood vessels, in order to increase effectiveness and minimize side-effects. Ligand-directed VTAs, therefore, are composed of targeting and effector moieties that are linked together via chemical cross-linkers or peptide bonds. The targeting moiety is usually an antibody or a peptide directed against a marker that is selectively upregulated on tumor versus normal vasculature. The first experimental evidence for the efficacy of ligand-directed VTAs was published by Burrows and Thorpe in 1993 64. To obtain a tumor-specific vascular marker they used a neuroblastoma cell line that expressed IFN- to grow subcutaneous tumors in BALB/c nude mice. Endothelial cells do normally not express MHC class II unless stimulated with IFN-. Thus, the IFN- secreted by the tumor cells activated endothelial cells within the tumor to express MHC class II, which could then be targeted with an immunotoxin composed of a high-affinity anti-MHC class II antibody linked to the protein toxin ricin A-chain. Intravenous injection of this immunotoxin induced rapid occlusion of the tumor blood vessels followed by extensive haemorrhagic necrosis of the tumor, with no detectable damage to other tissues. Tumor-specific vascular markers currently evaluated in pre-clinical studies include molecules associated with angiogenesis and vascular remodeling such as VEGF and its receptors, integrins v3 and v5, fibronectin extradomain B, tenascin-C extradomain C, aminopeptidase-N (CD13), and the TGF- co-receptor endoglin (CD105); cell adhesion molecules induced by inflammatory mediators that are secreted by tumor cells or host cells that infiltrate the tumor such as VCAM-1 and E- selectin; and other markers such as phosphatidylserine (PS), prostate-specific membrane antigen (PMSA), annexin A1, and the CD44 isoform CD44v6 7. The effectors used include toxins such as ricin, gelonin and Diphtheria toxin; cytotoxic drugs such as doxorubicin and neocarzinostatin; cytokines such as TNF, IL-2 and IL- 12; coagulation-inducing proteins such as tissue factor; short-lived radio-isotopes such as Actinium-225, Bismuth-213 and Iodine-131; and liposome-encapsulated drugs 5. Some examples of pre-clinical studies with ligand-directed VTAs are summarized in Table 5.2 (see also Table 1.2.). 160 Chapter 5

Table 5.2. Examples of pre-clinical studies with ligand-directed VTAs.

VTA Description / Results Refs

K4-2C10-dgRA - conjugate of the toxin ricin A-chain (dgRA) and an anti-endoglin antibody [65] - cytotoxic for murine endothelial cells in vitro; induced long-lasting complete inhibition of growth of human breast cancer cells in SCID mice

anti-VCAM-1-tTF - fusion protein of an anti-VCAM-1 antibody and the extracellular domain of tissue [66] factor - induced thrombosis of tumor vessels and tumor growth retardation in mice bearing solid Hodgkin’s tumors

dox-RGD-4C - conjugate of doxorubicin and an Arg-Gly-Asp (RGD) sequence or an Asn-Gly-Arg [67]

dox-CNGRC (NGR) sequence -containing peptide, which selectively targets integrins v3 and

v5 or aminopeptidase-N, respectively - induced vascular damage in human breast carcinoma xenografts in nude mice without apparent toxicity in other organs, compared to often lethal toxicity with free doxorubicin; no tumor vascular damage induced by equivalent dose of free doxorubicin

ScFv(L19)-tTF - antibody fragment directed to the extradomain B of fibronectin fused to the [68] extracellular domain of tissue factor - induced complete and selective infarction of 3 different types of solid tumors in mice; complete tumor eradication was observed in 30% of treated mice without apparent side-effects

VEGF121/rGel - fusion protein of VEGF121 and the toxin gelonin [69] - cytotoxic for VEGF-R2 overexpressing endothelial cells in vitro; inhibited growth of human melanoma and prostate cancer xenografts in nude mice; induced selective damage to vascular endothelium in human prostate cancer xenografts in nude mice

(131)I-L19-SIP - Iodine-131-labeled small immunoprotein (SIP) directed to the extradomain B of [70] fibronectin - selective tumor uptake; inhibited tumor growth and improved survival in a colorectal tumor model

Introduction: Vascular-targeting anticancer therapy 161

5.5. References

1. Folkman, J. Tumor angiogenesis: therapeutic implications. The New England journal of medicine 285, 1182-1186 (1971). 2. Iga, A.M., Sarkar, S., Sales, K.M., Winslet, M.C. & Seifalian, A.M. Quantitating therapeutic disruption of tumor blood flow with intravital video microscopy. Cancer research 66, 11517- 11519 (2006). 3. Sweeney, C.J., Miller, K.D. & Sledge, G.W., Jr. Resistance in the anti-angiogenic era: nay- saying or a word of caution? Trends in molecular medicine 9, 24-29 (2003). 4. Kerbel, R.S. Tumor angiogenesis: past, present and the near future. Carcinogenesis 21, 505- 515 (2000). 5. Thorpe, P.E. Vascular targeting agents as cancer therapeutics. Clin Cancer Res 10, 415-427 (2004). 6. Patterson, D.M. & Rustin, G.J. Vascular damaging agents. Clinical oncology (Royal College of Radiologists (Great Britain)) 19, 443-456 (2007). 7. Neri, D. & Bicknell, R. Tumour vascular targeting. Nat Rev Cancer 5, 436-446 (2005). 8. Hurwitz, H. et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. The New England journal of medicine 350, 2335-2342 (2004). 9. Shaheen, R.M. et al. Effects of an antibody to vascular endothelial growth factor receptor-2 on survival, tumor vascularity, and apoptosis in a murine model of colon carcinomatosis. International journal of oncology 18, 221-226 (2001). 10. Jain, R.K. Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy. Science (New York, N.Y 307, 58-62 (2005). 11. Carmeliet, P. Angiogenesis in life, disease and medicine. Nature 438, 932-936 (2005). 12. Melchert, M. & List, A. The thalidomide saga. The international journal of biochemistry & cell biology 39, 1489-1499 (2007). 13. Folkman, J. Angiogenesis: an organizing principle for drug discovery? Nat Rev Drug Discov 6, 273-286 (2007). 14. Boyland, E. & Boyland, M.E. Studies in tissue metabolism: The action of colchicine and B. typhosus extract. The Biochemical journal 31, 454-460 (1937). 15. Baguley, B.C., Holdaway, K.M., Thomsen, L.L., Zhuang, L. & Zwi, L.J. Inhibition of growth of colon 38 adenocarcinoma by vinblastine and colchicine: evidence for a vascular mechanism. Eur J Cancer 27, 482-487 (1991). 16. Hill, S.A., Lonergan, S.J., Denekamp, J. & Chaplin, D.J. Vinca alkaloids: anti-vascular effects in a murine tumour. Eur J Cancer 29A, 1320-1324 (1993). 17. Chaplin, D.J., Pettit, G.R., Parkins, C.S. & Hill, S.A. Antivascular approaches to solid tumour therapy: evaluation of tubulin binding agents. The British journal of cancer 27, S86-88 (1996). 18. Pettit, G.R., Cragg, G.M. & Singh, S.B. Antineoplastic agents, 122. Constituents of Combretum caffrum. Journal of natural products 50, 386-391 (1987). 19. Lin, C.M. et al. Interactions of tubulin with potent natural and synthetic analogs of the antimitotic agent combretastatin: a structure-activity study. Molecular pharmacology 34, 200- 208 (1988). 20. Pettit, G.R. et al. Isolation and structure of the strong cell growth and tubulin inhibitor combretastatin A-4. Experientia 45, 209-211 (1989). 21. McGown, A.T. & Fox, B.W. Structural and biochemical comparison of the anti-mitotic agents colchicine, combretastatin A4 and amphethinile. Anti-cancer drug design 3, 249-254 (1989). 22. Woods, J.A., Hadfield, J.A., Pettit, G.R., Fox, B.W. & McGown, A.T. The interaction with tubulin of a series of stilbenes based on combretastatin A-4. British journal of cancer 71, 705- 711 (1995). 23. Pettit, G.R. et al. Antineoplastic agents 322. synthesis of combretastatin A-4 prodrugs. Anti- cancer drug design 10, 299-309 (1995). 24. Dark, G.G. et al. Combretastatin A-4, an agent that displays potent and selective toxicity toward tumor vasculature. Cancer research 57, 1829-1834 (1997). 25. Chaplin, D.J., Pettit, G.R. & Hill, S.A. Anti-vascular approaches to solid tumour therapy: evaluation of combretastatin A4 phosphate. Anticancer research 19, 189-195 (1999). 26. Tozer, G.M. et al. Combretastatin A-4 phosphate as a tumor vascular-targeting agent: early effects in tumors and normal tissues. Cancer research 59, 1626-1634 (1999). 27. Galbraith, S.M. et al. Combretastatin A4 phosphate has tumor antivascular activity in rat and man as demonstrated by dynamic magnetic resonance imaging. J Clin Oncol 21, 2831-2842 (2003). 162 Chapter 5

28. Hill, S.A., Toze, G.M., Pettit, G.R. & Chaplin, D.J. Preclinical evaluation of the antitumour activity of the novel vascular targeting agent Oxi 4503. Anticancer research 22, 1453-1458 (2002). 29. Holwell, S.E. et al. Combretastatin A-1 phosphate a novel tubulin-binding agent with in vivo anti vascular effects in experimental tumours. Anticancer research 22, 707-711 (2002). 30. Hori, K., Saito, S., Nihei, Y., Suzuki, M. & Sato, Y. Antitumor effects due to irreversible stoppage of tumor tissue blood flow: evaluation of a novel combretastatin A-4 derivative, AC7700. Jpn J Cancer Res 90, 1026-1038 (1999). 31. Hori, K., Saito, S. & Kubota, K. A novel combretastatin A-4 derivative, AC7700, strongly stanches tumour blood flow and inhibits growth of tumours developing in various tissues and organs. British journal of cancer 86, 1604-1614 (2002). 32. Davis, P.D. et al. ZD6126: a novel vascular-targeting agent that causes selective destruction of tumor vasculature. Cancer research 62, 7247-7253 (2002). 33. Blakey, D.C. et al. Antitumor activity of the novel vascular targeting agent ZD6126 in a panel of tumor models. Clin Cancer Res 8, 1974-1983 (2002). 34. Kanthou, C. et al. The tubulin-binding agent combretastatin A-4-phosphate arrests endothelial cells in mitosis and induces mitotic cell death. The American journal of pathology 165, 1401- 1411 (2004). 35. Vincent, L. et al. Combretastatin A4 phosphate induces rapid regression of tumor neovessels and growth through interference with vascular endothelial-cadherin signaling. The Journal of clinical investigation 115, 2992-3006 (2005). 36. Hinnen, P. & Eskens, F.A. Vascular disrupting agents in clinical development. British journal of cancer 96, 1159-1165 (2007). 37. Tozer, G.M. et al. Mechanisms associated with tumor vascular shut-down induced by combretastatin A-4 phosphate: intravital microscopy and measurement of vascular permeability. Cancer research 61, 6413-6422 (2001). 38. Petrache, I. et al. Caspase-dependent cleavage of myosin light chain kinase (MLCK) is involved in TNF-alpha-mediated bovine pulmonary endothelial cell apoptosis. Faseb J 17, 407-416 (2003). 39. Petrache, I., Crow, M.T., Neuss, M. & Garcia, J.G. Central involvement of Rho family GTPases in TNF-alpha-mediated bovine pulmonary endothelial cell apoptosis. Biochemical and biophysical research communications 306, 244-249 (2003). 40. Deshpande, S.S., Angkeow, P., Huang, J., Ozaki, M. & Irani, K. Rac1 inhibits TNF-alpha- induced endothelial cell apoptosis: dual regulation by reactive oxygen species. Faseb J 14, 1705-1714 (2000). 41. Li, J.H. & Pober, J.S. The cathepsin B death pathway contributes to TNF plus IFN-gamma- mediated human endothelial injury. J Immunol 175, 1858-1866 (2005). 42. Sohn, R.H. et al. Regulation of endothelial thrombomodulin expression by inflammatory cytokines is mediated by activation of nuclear factor-kappa B. Blood 105, 3910-3917 (2005). 43. Conway, E.M., Bach, R., Rosenberg, R.D. & Konigsberg, W.H. Tumor necrosis factor enhances expression of tissue factor mRNA in endothelial cells. Thrombosis research 53, 231- 241 (1989). 44. Bevilacqua, M.P. et al. Recombinant tumor necrosis factor induces procoagulant activity in cultured human vascular endothelium: characterization and comparison with the actions of interleukin 1. Proceedings of the National Academy of Sciences of the United States of America 83, 4533-4537 (1986). 45. Parry, G.C. & Mackman, N. Transcriptional regulation of tissue factor expression in human endothelial cells. Arteriosclerosis, thrombosis, and vascular biology 15, 612-621 (1995). 46. Pober, J.S. & Sessa, W.C. Evolving functions of endothelial cells in inflammation. Nat Rev Immunol 7, 803-815 (2007). 47. Ruegg, C. et al. Evidence for the involvement of endothelial cell integrin alphaVbeta3 in the disruption of the tumor vasculature induced by TNF and IFN-gamma. Nature medicine 4, 408- 414 (1998). 48. Petrache, I., Birukova, A., Ramirez, S.I., Garcia, J.G. & Verin, A.D. The role of the microtubules in tumor necrosis factor-alpha-induced endothelial cell permeability. American journal of respiratory cell and molecular biology 28, 574-581 (2003). 49. Angelini, D.J. et al. TNF-alpha increases tyrosine phosphorylation of vascular endothelial cadherin and opens the paracellular pathway through fyn activation in human lung endothelia. Am J Physiol Lung Cell Mol Physiol 291, L1232-1245 (2006). Introduction: Vascular-targeting anticancer therapy 163

50. Menon, C., Ghartey, A., Canter, R., Feldman, M. & Fraker, D.L. Tumor necrosis factor-alpha damages tumor blood vessel integrity by targeting VE-cadherin. Annals of surgery 244, 781- 791 (2006). 51. Corbett, T.H. et al. Activity of flavone acetic acid (NSC-347512) against solid tumors of mice. Investigational new drugs 4, 207-220 (1986). 52. Plowman, J. et al. Flavone acetic acid: a novel agent with preclinical antitumor activity against colon adenocarcinoma 38 in mice. Cancer treatment reports 70, 631-635 (1986). 53. Smith, G.P., Calveley, S.B., Smith, M.J. & Baguley, B.C. Flavone acetic acid (NSC 347512) induces haemorrhagic necrosis of mouse colon 26 and 38 tumours. European journal of cancer & clinical oncology 23, 1209-1211 (1987). 54. Evelhoch, J.L. et al. Flavone acetic acid (NSC 347512)-induced modulation of murine tumor physiology monitored by in vivo nuclear magnetic resonance spectroscopy. Cancer research 48, 4749-4755 (1988). 55. Hill, S., Williams, K.B. & Denekamp, J. Vascular collapse after flavone acetic acid: a possible mechanism of its anti-tumour action. European journal of cancer & clinical oncology 25, 1419-1424 (1989). 56. Zwi, L.J., Baguley, B.C., Gavin, J.B. & Wilson, W.R. Blood flow failure as a major determinant in the antitumor action of flavone acetic acid. J Natl Cancer Inst 81, 1005-1013 (1989). 57. McKeage, M.J., Kestell, P., Denny, W.A. & Baguley, B.C. Plasma pharmacokinetics of the antitumour agents 5,6-dimethylxanthenone-4-acetic acid, xanthenone-4-acetic acid and flavone-8-acetic acid in mice. Cancer chemotherapy and pharmacology 28, 409-413 (1991). 58. Baguley, B.C. Antivascular therapy of cancer: DMXAA. The lancet oncology 4, 141-148 (2003). 59. Baguley, B.C. & Ching, L.M. DMXAA: an antivascular agent with multiple host responses. International journal of radiation oncology, biology, physics 54, 1503-1511 (2002). 60. Ching, L.M., Joseph, W.R., Crosier, K.E. & Baguley, B.C. Induction of tumor necrosis factor- alpha messenger RNA in human and murine cells by the flavone acetic acid analogue 5,6- dimethylxanthenone-4-acetic acid (NSC 640488). Cancer research 54, 870-872 (1994). 61. Pratesi, G., Rodolfo, M., Rovetta, G. & Parmiani, G. Role of T cells and tumour necrosis factor in antitumour activity and toxicity of flavone acetic acid. Eur J Cancer 26, 1079-1083 (1990). 62. Zhao, L., Ching, L.M., Kestell, P. & Baguley, B.C. The antitumour activity of 5,6- dimethylxanthenone-4-acetic acid (DMXAA) in TNF receptor-1 knockout mice. British journal of cancer 87, 465-470 (2002). 63. Ching, L.M. et al. Induction of endothelial cell apoptosis by the antivascular agent 5,6- Dimethylxanthenone-4-acetic acid. British journal of cancer 86, 1937-1942 (2002). 64. Burrows, F.J. & Thorpe, P.E. Eradication of large solid tumors in mice with an immunotoxin directed against tumor vasculature. Proceedings of the National Academy of Sciences of the United States of America 90, 8996-9000 (1993). 65. Seon, B.K., Matsuno, F., Haruta, Y., Kondo, M. & Barcos, M. Long-lasting complete inhibition of human solid tumors in SCID mice by targeting endothelial cells of tumor vasculature with antihuman endoglin immunotoxin. Clin Cancer Res 3, 1031-1044 (1997). 66. Ran, S. et al. Infarction of solid Hodgkin's tumors in mice by antibody-directed targeting of tissue factor to tumor vasculature. Cancer research 58, 4646-4653 (1998). 67. Arap, W., Pasqualini, R. & Ruoslahti, E. Cancer treatment by targeted drug delivery to tumor vasculature in a mouse model. Science (New York, N.Y 279, 377-380 (1998). 68. Nilsson, F., Kosmehl, H., Zardi, L. & Neri, D. Targeted delivery of tissue factor to the ED-B domain of fibronectin, a marker of angiogenesis, mediates the infarction of solid tumors in mice. Cancer research 61, 711-716 (2001). 69. Veenendaal, L.M. et al. In vitro and in vivo studies of a VEGF121/rGelonin chimeric fusion toxin targeting the neovasculature of solid tumors. Proceedings of the National Academy of Sciences of the United States of America 99, 7866-7871 (2002). 70. El-Emir, E. et al. Characterisation and radioimmunotherapy of L19-SIP, an anti-angiogenic antibody against the extra domain B of fibronectin, in colorectal tumour models. British journal of cancer 96, 1862-1870 (2007).

Chapter 6 NKT cells

Introduction: NKT cells 167

6. NKT CELLS

6.1. Definition

Natural killer T (NKT) cells are a heterogenous group of lymphocytes that share properties of both T cells and NK cells (Fig.6.1.). They are often defined as cells that express a CD1d-restricted -TCR as well as the NK cell receptor NK1.1 (also known as NKR-P1C, Ly55, or CD161c). However, this definition is somewhat problematic as NK1.1 is only expressed in a limited number of mouse strains, including C57BL/6, but not in most other commonly used laboratory mouse strains such as 129, BALB/c, C3H/He, and many others; and even in C57BL/6 mice NKT cells are present that do not express NK1.1 1. Moreover, NKT-like cells have been identified that are not CD1d-restricted. NKT cells are also diversified in some other aspects such as the expression of CD4 and CD8 co-receptors, and the reactivity towards the glycolipid -galactosylceramide (GalCer) (see further). Thus, it has become clear that NKT cells are a heterogenous group of cells that vary in the expression of cell surface markers and most likely also in function. Therefore, NKT cells are nowadays classified, according to CD1d-dependence, NK1.1 expression, and reactivity towards GalCer, in at least three different classes (Table 6.1.).

Fig. 6.1. NKT cells share properties of both NK and T cells. This Venn diagram incorporates some of the key features of NKT, T and NK cells. Characteristics that are unique to each lineage are shown in the non-overlapping sections. Characteristics that are common to two of these lineages are indicated in the overlapping areas (purple or orange). The central area (white) indicates characteristics that can be exhibited by all three lineages. From Godfrey et al., 2000 2. 168 Chapter 6

Table 6.1. Categories of mouse NKT cells. Based on Godfrey et al., 2004 3. Type I Type II Type III Other names Classical NKT cells, Non-classical NKT cells NKT-like cells, CD1d- invariant NKT cells independent NK1.1+ T cells TCR -chain V14-J18 Diverse, but some Diverse V3.2-J9, V8 TCR -chain V8.2/7/2 Diverse, but some V8.2 Diverse Co-receptor CD4+ or DN CD4+ or DN CD4+, CD8+ or DN CD1d-dependence Yes Yes No GalCer reactivity Yes No No NK1.1 expression Yes* Heterogenous Yes IL-4 production Yes Yes No IFN- production Yes Yes Yes Location Thymus, liver Spleen, bone marrow Spleen, bone marrow (intestine, lymph node) * NK1.1 is expressed by resting, mature cells, but is low or absent in immature cells or post-activation. DN: CD4- CD8- double negative.

The most studied and best characterized type of NKT cell is the type I NKT cell, also called classical NKT cell or invariant NKT (iNKT) cell. iNKT cells comprise approximately 80% of all CD1d-dependent NKT cells 4, 5. CD1d is a member of the CD1 family of MHC class I-related antigen-presenting molecules. CD1d is expressed by professional APCs (dendritic cells, macrophages and B cells), thymocytes, some intestinal epithelial cells, hepatocytes, and some other cells 6-9. In contrast to other MHC molecules which are involved in the presentation of peptide antigens, CD1d is specialized in the presentation of glyco- and phospholipids, although highly hydrophobic peptides have also been shown to bind to CD1d 10. In 1997, Kawano and co-workers reported the identification of the first iNKT cell-specific ligand presented by CD1d: GalCer, a glycolipid extracted from the marine sponge Agelas mauritianus (Fig.6.2.). Co-culture of iNKT cells and dendritic cells pulsed with GalCer induced proliferation of the iNKT cells and rapid production of large amounts of IL-4 and IFN-. Moreover, GalCer-stimulated iNKT cells were found to be capable of killing YAC-1 and tumor cells by an NK-like mechanism 11. The identification of this first CD1d-dependent, iNKT cell-specific ligand led to the development of a new tool to identify iNKT cells: tetramers of CD1d loaded with GalCer, which allowed to further characterize these cells 12-15. iNKT cells represent in mice approximately 0.5% of the T cell population in blood and peripheral lymph nodes, 0.5% of thymocytes, 2.5% of T cells in the spleen, and up to 30% of T cells in the liver, where they reside in the sinusoids. Although the tissue distribution is less well studied in humans, iNKT cells appear to be approximately 10 times less frequent in all these locations (Table 6.2.) 16. In mice iNKT cells express an invariant TCR composed of a V14-J18 rearranged TCR - Introduction: NKT cells 169

chain combined with a V8.2, V7 or V2 TCR -chain (human iNKT cells express a homologues V24-J18/V11 TCR). The expression of the NK receptor NK1.1 (CD161) appears to depend on the maturity and activation state of the iNKT cells. Immature iNKT cells do not express NK1.1 and expression of this marker is downregulated for at least several days after the activation of mature iNKT cells 3, 17. Mature NKT cells also express cytotoxic effector molecules such as perforin, 18 granzyme B and FasL . Other NK markers such as the integrin subunit 2 (CD49b or DX5) and members of the Ly49 family of killer inhibitory receptors (KIRs) are not consistently expressed by iNKT cells. A high proportion of iNKT cells in the thymus expresses Ly49, whereas most liver iNKT cells lack this surface marker 19. On the other hand, most type III CD8+ NKT-like cells in the liver do express Ly49 20. Similarly, CD49b is expressed by only a minority of CD1d-dependent NKT cells, while most CD1d-independent NKT-like cells express this marker 21.

Table 6.2. A comparison of mouse and human iNKT cells. From Godfrey et al., 2007 22. Mouse Human TCR -chain V14-J18 V24-J18 TCR -chain V8.2, V7, V2 V11 CD1d-dependent Yes Yes Thymus dependent Yes Probably GalCer reactive Yes Yes iGb3 reactive Yes Yes Immature phenotype NK1.1low; CD4+ (or DN ?) CD161low; CD4+ Mature phenotype NK1.1+; CD4+ or DN CD161+; CD4+, DN, or CD8+ Frequency in thymus ~0,2-0,5% ~0,001-0,01% Frequency in blood ~0,2-0,5% ~0,001-1% Frequency in spleen ~1% ND Frequency in liver ~20-40% ~1% DN: CD4- CD8- double negative. ND: not determined.

6.2. Development

The origin of iNKT cells has long been controversial, particularly because of some early reports showing the presence of (mainly CD8+) NK1.1+ T cells in athymic or neonatally thymectomized mice 23-25. Nevertheless, it is now well-established that iNKT cells are thymus-dependent and share a common origin with other -TCR+ T cells. iNKT cells segregate from conventional T cells at the CD4+ CD8+ double positive (DP) stage. DP thymocytes expressing the invariant V14 TCR are positively selected by other, CD1d-expressing DP thymocytes in the cortex 22. It is still unknown which endogenous ligands are presented by CD1d for positive selection. Several reports have suggested that the lysosomal glycosphingolipid isoglobotrihexosylceramide (iGb3) is the endogenous ligand used for NKT cell 170 Chapter 6

selection (Fig.6.2.). Mice deficient in -hexosaminidase B, a lysosomal enzyme involved in the production of iGb3, display a 95% decrease in thymic NKT cell production, and iGb3 showed stimulatory activity towards both human and murine iNKT cells in vitro, albeit much weaker than GalCer 26. However, it was recently shown that mice deficient in iGb3 synthase, the enzyme directly responsible for the synthesis of iGb3, have normal development and function of iNKT cells, suggesting that iGb3 is not or at least not the only endogenous ligand involved in the thymic selection of iNKT cells 27. After positive selection, immature NKT cells downregulate CD8 and undergo several maturation steps to a NK1.1- CD44+ DX5+ stage, where most cells leave the thymus to further maturate in the periphery to NK1.1+ CD4+ or double negative (DN) mature iNKT cells 22.

6.3. Function

iNKT are highly versatile cells at the interface between innate and adaptive immunity. They have been implicated in a broad array of immunological processes, from the control of infection and tumor immunity, to the development of autoimmune diseases and transplantation tolerance. The most striking effector function of iNKT cells is their ability to rapidly produce large amounts of both Th1 and Th2 cytokines upon activation, suggesting that these cells mainly have an immunoregulatory functions. Triggering of the TCR of iNKT cells, using anti-CD3 antibodies or GalCer-pulsed dendritic cells, results in rapid (within 1-2 hours) production of IFN-, TNF, IL-4 and IL-13 11, 21, 28. It has been reported that iNKT cells are able to release IL-4 and IFN- so rapidly upon activation because they posses some pre-formed mRNA for these cytokines in the resting state 29, 30. Stimulating iNKT cells by NK1.1 cross-linking has been shown to result in the production of IFN-, but not of IL-4 31. The cytokine production by iNKT cells is also influenced by the presence of other cytokines: IL-7 enhances IL-4 production, whereas IL-12 enhances IFN- production 32, 33. These results suggest that the pattern of cytokines produced by iNKT cells will differ depending on the microenvironment. The immunoregulatory cytokines produced by iNKT cells can, in turn, activate or recruit other cell types, including NK cells, T cells (thereby directing Th1/Th2 differentiation), B cells, macrophages, and dendritic cells 34-37.

iNKT cells have been shown to participate in host defense against some bacterial and parasitic infections. Some bacteria such as Sphingomonas and Ehrlichia muris, which are Gram-negative, LPS-negative members of the -Proteobacteria, contain cell wall components capable of specifically activating iNKT cells. In the case of Sphingomonas it was shown that the cell wall contains the glycosphingolipids - glucuronosylceramide and -galacturonosylceramide which are structurally related to Introduction: NKT cells 171

GalCer (Fig.6.2.). During infection, Sphingomonas is phagocytosed by macrophages and dendritic cells and elicits an activation cascade similar to exogenous GalCer 38- 40. The responses of iNKT cells to other bacteria, such as the Gram-negative, LPS- positive Salmonella, or even to parasitic infections by the helminth Schistosoma mansoni, were shown to depend on -hexosaminidase B activity, suggesting that in the absence of a suitable microbial NKT ligand, iNKT cells can still become activated by presentation of the endogenous ligand iGb3 40, 41. In the case of Salmonella it has been demonstrated that iNKT cell activation was absolutely dependent on TLR signaling and IL-12 production by the APC. The proposed mechanism is that TLR signaling leads to IL-12 secretion by the APC which then enhances the ability of dendritic cells to stimulate iNKT cells through presentation of endogenous ligand 16, 42.

Fig. 6.2. iNKT cell ligands. (A) Marine sponge GalCer (KRN7000). (B) Sphingomonas - glucuronosylceramide (GSL-1). (C) Mammalian iGb3. From Bendelac et al., 2007 16. iNKT cells have been implicated in a number of autoimmune and inflammatory diseases, including type I diabetes, lupus erythematosus, asthma and atherosclerosis. However, for each of these diseases contradictory studies have been reported, and at this point there is no decisive evidence for a specific role of iNKT cells in any of these diseases (recently reviewed by Bendelac et al. 16). For example, it was reported that CD1d- and J18-deficient mice, which both completely lack iNKT cells, exhibit decreased allergen-induced airway hyperreactivity in the alum-ovalbumin model of asthma 43, 44. However, similar studies by another group failed to observe differences 172 Chapter 6

between CD1d-deficient and wild-type mice 16. Using GalCer-CD1d tetramers one group reported that more than 60% of CD4+ T cells in bronchoalveolar-lavage fluid of a panel of human patients with chronic asthma were iNKT cells, which predominantly produced Th2 cytokines 45. Again, another group found that less than 2% of T cells in samples of chronic asthma patients were iNKT cells, which was not significantly different from healthy control subjects 46. Thus, it remains controversial whether iNKT cells play a pathologic role in asthma.

6.4. NKT cells in tumor immunity and cancer therapy

Several reports have suggested a natural role of iNKT cells in tumor surveillance. The first evidence came from tumor rejection experiments of the MHC-negative F9 embryonic carcinoma in mice. Depletion of either NK1.1+ cells,  T cells, or injection of anti-IL-4 antibodies rendered B6 mice significantly more susceptible to tumor formation. It was furthermore shown that IL-4-producing, NK1.1+ DN T cells accumulated in these F9 tumors, suggesting that these cells are involved in the tumor rejection in this model 47. More convincing data were published by Smyth and colleagues. They demonstrated that J18-deficient mice were two- to threefold more sensitive to tumor induction by the carcinogen methA. Moreover, they showed that tumors induced in J18-deficient mice grew after transplantation far more effectively in J18-deficient mice than in wild-type control mice. Smyth and co-workers furthermore demonstrated that host immune protection from methA-induced sarcoma is, besides iNKT cells, also dependent on NK cells, IL-12, IFN-, and perforin 48, 49. The most likely mechanism is then that IL-12, produced by CD1d-expressing APCs, stimulates iNKT cells to rapidly secrete IFN-, which then promotes secondary activation of NK cell proliferation and cytotoxicity (Fig.6.3.). A role of iNKT cells in tumor immunity is further supported by the observation that iNKT cells were reduced in number or functionally hyporeactive in human patients with advanced prostate cancer or progressive multiple myeloma 50, 51. A murine T cell lymphoma cell line has even been identified that sheds a glycolipid, gangliotriaosylceramide, that inhibited activation of iNKT cells in vitro 52. Together these data suggest that certain tumors may effectively evade the host’s antitumor immune response by inhibiting iNKT cell activation.

iNKT cells are promising targets for cancer immunotherapy. Already before it was known that GalCer is a specific NKT cell ligand, it was reported that this compound had potent tumor growth inhibitory and antimetastatic activities in B16 tumor-bearing mice 53, 54. Antitumor and antimetastatic effects of GalCer have now been observed in many murine tumor models, including melanoma, lymphoma, colon, lung, and renal cell carcinoma; and similar or even stronger antitumor effects have been Introduction: NKT cells 173

obtained by treatment with dendritic cells pulsed with GalCer in vitro 55-60. First clinical trials in cancer patients with GalCer as a soluble drug as well as with GalCer-pulsed dendritic cells have already been completed, with limited success 61, 62. As GalCer induces secretion of both IFN- and IL-4 by iNKT cells, it has been suggested that the net antitumor effect of GalCer may be limited due to the counteracting effects of these cytokines. Therefore, more potent GalCer analogs are currently developed that favor the production of Th1 cytokines by human iNKT cells 63. The antitumor effect of GalCer in murine tumor models was shown to be critically dependent on IFN- and NK cells 64. It was furthermore shown that the production of IFN- by iNKT cells in response to GalCer requires IL-12 production by dendritic cells and direct contact between NKT cells and dendritic cells through CD40-CD40 ligand interaction 65. It is now also clear that the potent antitumor effects of IL-12 therapy are at least partly mediated by iNKT cells (Fig.6.3.) 66-69. Smyth and co- workers demonstrated that the antitumor effect of GalCer requires subsequent IFN- production by iNKT cells and NK cells. The IFN- produced by iNKT cells is at least partly responsible for NK cell activation. The cytokines produced by iNKT and NK cells will most likely also induce bystander activation of other immune cells such as T cells, B cells, and macrophages (Fig.6.3.) 70. The activation of iNKT cells by GalCer appears to induce rapid downregulation of their TCR and rapid expansion, reaching maximum numbers around three days after GalCer administration, followed by a decline through apoptotic cell death, and long-lasting anergy of the remaining iNKT cells 71. It was recently shown that OX40 ligand (a member of the TNF superfamily) expressed on dendritic cells, through interaction with its receptor OX40 (CD134) on iNKT cells, strongly enhances iNKT cell activation, IFN- production, and antitumor immunity. Interestingly, the expression of OX40 ligand on dendritic cells can be upregulated by TNF 72.

174 Chapter 6

Fig. 6.3. The role of NKT cells in tumor rejection. (A) In IL-12-mediated tumor rejection, exogenously administrated IL-12 may induce IFN- production by both NK and NKT cells. In IL-12- induced, NKT cell-dependent tumor rejection, IFN- is known to be a key factor. Evidence exists for both IFN--dependent and perforin (Pfp)-dependent effector mechanisms. The relative roles of effector cells and associated molecules may vary in a tumor- and dose-dependent fashion. (B) In GalCer- mediated tumor rejection, exogenously administered GalCer is presented to iNKT cells by CD1d on APCs. The process involves subsequent IFN- production by both iNKT cells and NK cells, and is IL- 12-dependent but perforin-independent (indep.). The importance of CD8+ T cells may be model- dependent. (C) In natural tumor rejection, CD1d-restricted NKT cells may become activated by an endogenous, possibly tumor-derived, glycolipid antigen (endog. glycolipid Ag). Natural NKT cell- mediated tumor rejection is dependent on IL-12, IFN-, and perforin. Besides NKT cells and NK cells, also CD8+ T cells appear to play an important role in natural tumor rejection. M macrophage. Adapted from Smyth et al., 2002 73.

Introduction: NKT cells 175

6.5. References

1. Hammond, K.J. et al. CD1d-restricted NKT cells: an interstrain comparison. J Immunol 167, 1164-1173 (2001). 2. Godfrey, D.I., Hammond, K.J., Poulton, L.D., Smyth, M.J. & Baxter, A.G. NKT cells: facts, functions and fallacies. Immunology today 21, 573-583 (2000). 3. Godfrey, D.I., MacDonald, H.R., Kronenberg, M., Smyth, M.J. & Van Kaer, L. NKT cells: what's in a name? Nat Rev Immunol 4, 231-237 (2004). 4. MacDonald, H.R. Development and selection of NKT cells. Current opinion in immunology 14, 250-254 (2002). 5. Mercer, J.C., Ragin, M.J. & August, A. Natural killer T cells: rapid responders controlling immunity and disease. The international journal of biochemistry & cell biology 37, 1337-1343 (2005). 6. Bleicher, P.A. et al. Expression of murine CD1 on gastrointestinal epithelium. Science (New York, N.Y 250, 679-682 (1990). 7. Blumberg, R.S., Colgan, S.P. & Balk, S.P. CD1d: outside-in antigen presentation in the intestinal epithelium? Clinical and experimental immunology 109, 223-225 (1997). 8. Brossay, L. et al. Mouse CD1 is mainly expressed on hemopoietic-derived cells. J Immunol 159, 1216-1224 (1997). 9. Roark, J.H. et al. CD1.1 expression by mouse antigen-presenting cells and marginal zone B cells. J Immunol 160, 3121-3127 (1998). 10. Brutkiewicz, R.R. CD1d ligands: the good, the bad, and the ugly. J Immunol 177, 769-775 (2006). 11. Kawano, T. et al. CD1d-restricted and TCR-mediated activation of valpha14 NKT cells by glycosylceramides. Science (New York, N.Y 278, 1626-1629 (1997). 12. MacDonald, H.R. CD1d-glycolipid tetramers: A new tool to monitor natural killer T cells in health and disease. The Journal of experimental medicine 192, F15-20 (2000). 13. Benlagha, K., Weiss, A., Beavis, A., Teyton, L. & Bendelac, A. In vivo identification of glycolipid antigen-specific T cells using fluorescent CD1d tetramers. The Journal of experimental medicine 191, 1895-1903 (2000). 14. Matsuda, J.L. et al. Tracking the response of natural killer T cells to a glycolipid antigen using CD1d tetramers. The Journal of experimental medicine 192, 741-754 (2000). 15. Karadimitris, A. et al. Human CD1d-glycolipid tetramers generated by in vitro oxidative refolding chromatography. Proceedings of the National Academy of Sciences of the United States of America 98, 3294-3298 (2001). 16. Bendelac, A., Savage, P.B. & Teyton, L. The biology of NKT cells. Annual review of immunology 25, 297-336 (2007). 17. Chen, H., Huang, H. & Paul, W.E. NK1.1+ CD4+ T cells lose NK1.1 expression upon in vitro activation. J Immunol 158, 5112-5119 (1997). 18. Matsuda, J.L. et al. T-bet concomitantly controls migration, survival, and effector functions during the development of Valpha14i NKT cells. Blood 107, 2797-2805 (2006). 19. Robson MacDonald, H., Lees, R.K. & Held, W. Developmentally regulated extinction of Ly- 49 receptor expression permits maturation and selection of NK1.1+ T cells. The Journal of experimental medicine 187, 2109-2114 (1998). 20. Emoto, M., Zerrahn, J., Miyamoto, M., Perarnau, B. & Kaufmann, S.H. Phenotypic characterization of CD8(+)NKT cells. European journal of immunology 30, 2300-2311 (2000). 21. Pellicci, D.G. et al. DX5/CD49b-positive T cells are not synonymous with CD1d-dependent NKT cells. J Immunol 175, 4416-4425 (2005). 22. Godfrey, D.I. & Berzins, S.P. Control points in NKT-cell development. Nat Rev Immunol 7, 505-518 (2007). 23. Kikly, K. & Dennert, G. Evidence for extrathymic development of TNK cells. NK1+ CD3+ cells responsible for acute marrow graft rejection are present in thymus-deficient mice. J Immunol 149, 403-412 (1992). 24. Hammond, K., Cain, W., van Driel, I. & Godfrey, D. Three day neonatal thymectomy selectively depletes NK1.1+ T cells. International immunology 10, 1491-1499 (1998). 25. Coles, M.C. & Raulet, D.H. NK1.1+ T cells in the liver arise in the thymus and are selected by interactions with class I molecules on CD4+CD8+ cells. J Immunol 164, 2412-2418 (2000). 26. Zhou, D. et al. Lysosomal glycosphingolipid recognition by NKT cells. Science (New York, N.Y 306, 1786-1789 (2004). 176 Chapter 6

27. Porubsky, S. et al. Normal development and function of invariant natural killer T cells in mice with isoglobotrihexosylceramide (iGb3) deficiency. Proceedings of the National Academy of Sciences of the United States of America 104, 5977-5982 (2007). 28. Godfrey, D.I. & Kronenberg, M. Going both ways: immune regulation via CD1d-dependent NKT cells. The Journal of clinical investigation 114, 1379-1388 (2004). 29. Stetson, D.B. et al. Constitutive cytokine mRNAs mark natural killer (NK) and NK T cells poised for rapid effector function. The Journal of experimental medicine 198, 1069-1076 (2003). 30. Matsuda, J.L. et al. Mouse V alpha 14i natural killer T cells are resistant to cytokine polarization in vivo. Proceedings of the National Academy of Sciences of the United States of America 100, 8395-8400 (2003). 31. Arase, H., Arase, N. & Saito, T. Interferon gamma production by natural killer (NK) cells and NK1.1+ T cells upon NKR-P1 cross-linking. The Journal of experimental medicine 183, 2391-2396 (1996). 32. Hameg, A. et al. IL-7 up-regulates IL-4 production by splenic NK1.1+ and NK1.1- MHC class I-like/CD1-dependent CD4+ T cells. J Immunol 162, 7067-7074 (1999). 33. Leite-De-Moraes, M.C. et al. IL-4-producing NK T cells are biased towards IFN-gamma production by IL-12. Influence of the microenvironment on the functional capacities of NK T cells. European journal of immunology 28, 1507-1515 (1998). 34. Carnaud, C. et al. Cutting edge: Cross-talk between cells of the innate immune system: NKT cells rapidly activate NK cells. J Immunol 163, 4647-4650 (1999). 35. Singh, N. et al. Cutting edge: activation of NK T cells by CD1d and alpha-galactosylceramide directs conventional T cells to the acquisition of a Th2 phenotype. J Immunol 163, 2373-2377 (1999). 36. Nishimura, T. et al. The interface between innate and acquired immunity: glycolipid antigen presentation by CD1d-expressing dendritic cells to NKT cells induces the differentiation of antigen-specific cytotoxic T lymphocytes. International immunology 12, 987-994 (2000). 37. Kronenberg, M. & Gapin, L. The unconventional lifestyle of NKT cells. Nat Rev Immunol 2, 557-568 (2002). 38. Kinjo, Y. et al. Recognition of bacterial glycosphingolipids by natural killer T cells. Nature 434, 520-525 (2005). 39. Wu, D. et al. Bacterial glycolipids and analogs as antigens for CD1d-restricted NKT cells. Proceedings of the National Academy of Sciences of the United States of America 102, 1351- 1356 (2005). 40. Mattner, J. et al. Exogenous and endogenous glycolipid antigens activate NKT cells during microbial infections. Nature 434, 525-529 (2005). 41. Mallevaey, T. et al. Activation of invariant NKT cells by the helminth parasite schistosoma mansoni. J Immunol 176, 2476-2485 (2006). 42. Brigl, M., Bry, L., Kent, S.C., Gumperz, J.E. & Brenner, M.B. Mechanism of CD1d-restricted natural killer T cell activation during microbial infection. Nature immunology 4, 1230-1237 (2003). 43. Lisbonne, M. et al. Cutting edge: invariant V alpha 14 NKT cells are required for allergen- induced airway inflammation and hyperreactivity in an experimental asthma model. J Immunol 171, 1637-1641 (2003). 44. Akbari, O. et al. Essential role of NKT cells producing IL-4 and IL-13 in the development of allergen-induced airway hyperreactivity. Nature medicine 9, 582-588 (2003). 45. Akbari, O. et al. CD4+ invariant T-cell-receptor+ natural killer T cells in bronchial asthma. The New England journal of medicine 354, 1117-1129 (2006). 46. Vijayanand, P. et al. Invariant natural killer T cells in asthma and chronic obstructive pulmonary disease. The New England journal of medicine 356, 1410-1422 (2007). 47. Nakamura, E. et al. Involvement of NK1+ CD4- CD8- alphabeta T cells and endogenous IL-4 in non-MHC-restricted rejection of embryonal carcinoma in genetically resistant mice. J Immunol 158, 5338-5348 (1997). 48. Smyth, M.J. et al. Differential tumor surveillance by natural killer (NK) and NKT cells. The Journal of experimental medicine 191, 661-668 (2000). 49. Smyth, M.J., Crowe, N.Y. & Godfrey, D.I. NK cells and NKT cells collaborate in host protection from methylcholanthrene-induced fibrosarcoma. International immunology 13, 459-463 (2001). 50. Tahir, S.M. et al. Loss of IFN-gamma production by invariant NK T cells in advanced cancer. J Immunol 167, 4046-4050 (2001). Introduction: NKT cells 177

51. Dhodapkar, M.V. et al. A reversible defect in natural killer T cell function characterizes the progression of premalignant to malignant multiple myeloma. The Journal of experimental medicine 197, 1667-1676 (2003). 52. Sriram, V. et al. Inhibition of glycolipid shedding rescues recognition of a CD1+ T cell lymphoma by natural killer T (NKT) cells. Proceedings of the National Academy of Sciences of the United States of America 99, 8197-8202 (2002). 53. Kobayashi, E., Motoki, K., Uchida, T., Fukushima, H. & Koezuka, Y. KRN7000, a novel immunomodulator, and its antitumor activities. Oncology research 7, 529-534 (1995). 54. Morita, M. et al. Structure-activity relationship of alpha-galactosylceramides against B16- bearing mice. Journal of medicinal chemistry 38, 2176-2187 (1995). 55. Nakagawa, R. et al. Treatment of hepatic metastasis of the colon26 adenocarcinoma with an alpha-galactosylceramide, KRN7000. Cancer research 58, 1202-1207 (1998). 56. Nakagawa, R. et al. Antitumor activity of alpha-galactosylceramide, KRN7000, in mice with EL-4 hepatic metastasis and its cytokine production. Oncology research 10, 561-568 (1998). 57. Smyth, M.J. et al. Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) contributes to interferon gamma-dependent natural killer cell protection from tumor metastasis. The Journal of experimental medicine 193, 661-670 (2001). 58. Kawano, T. et al. Natural killer-like nonspecific tumor cell lysis mediated by specific ligand- activated Valpha14 NKT cells. Proceedings of the National Academy of Sciences of the United States of America 95, 5690-5693 (1998). 59. Toura, I. et al. Cutting edge: inhibition of experimental tumor metastasis by dendritic cells pulsed with alpha-galactosylceramide. J Immunol 163, 2387-2391 (1999). 60. Nagaraj, S. et al. Dendritic cells pulsed with alpha-galactosylceramide induce anti-tumor immunity against pancreatic cancer in vivo. International immunology 18, 1279-1283 (2006). 61. Giaccone, G. et al. A phase I study of the natural killer T-cell ligand alpha-galactosylceramide (KRN7000) in patients with solid tumors. Clin Cancer Res 8, 3702-3709 (2002). 62. Ishikawa, A. et al. A phase I study of alpha-galactosylceramide (KRN7000)-pulsed dendritic cells in patients with advanced and recurrent non-small cell lung cancer. Clin Cancer Res 11, 1910-1917 (2005). 63. Chang, Y.J. et al. Potent immune-modulating and anticancer effects of NKT cell stimulatory glycolipids. Proceedings of the National Academy of Sciences of the United States of America 104, 10299-10304 (2007). 64. Hayakawa, Y. et al. Critical contribution of IFN-gamma and NK cells, but not perforin- mediated cytotoxicity, to anti-metastatic effect of alpha-galactosylceramide. European journal of immunology 31, 1720-1727 (2001). 65. Kitamura, H. et al. The natural killer T (NKT) cell ligand alpha-galactosylceramide demonstrates its immunopotentiating effect by inducing interleukin (IL)-12 production by dendritic cells and IL-12 receptor expression on NKT cells. The Journal of experimental medicine 189, 1121-1128 (1999). 66. Cui, J. et al. Requirement for Valpha14 NKT cells in IL-12-mediated rejection of tumors. Science (New York, N.Y 278, 1623-1626 (1997). 67. Kawamura, T. et al. Critical role of NK1+ T cells in IL-12-induced immune responses in vivo. J Immunol 160, 16-19 (1998). 68. Takeda, K. et al. Relative contribution of NK and NKT cells to the anti-metastatic activities of IL-12. International immunology 12, 909-914 (2000). 69. Smyth, M.J., Taniguchi, M. & Street, S.E. The anti-tumor activity of IL-12: mechanisms of innate immunity that are model and dose dependent. J Immunol 165, 2665-2670 (2000). 70. Smyth, M.J. et al. Sequential production of interferon-gamma by NK1.1(+) T cells and natural killer cells is essential for the antimetastatic effect of alpha-galactosylceramide. Blood 99, 1259-1266 (2002). 71. Parekh, V.V. et al. Glycolipid antigen induces long-term natural killer T cell anergy in mice. The Journal of clinical investigation 115, 2572-2583 (2005). 72. Zaini, J. et al. OX40 ligand expressed by DCs costimulates NKT and CD4+ Th cell antitumor immunity in mice. The Journal of clinical investigation 117, 3330-3338 (2007). 73. Smyth, M.J. et al. NKT cells - conductors of tumor immunity? Current opinion in immunology 14, 165-171 (2002).

Chapter 7 Research objectives

Research objectives 181

7. RESEARCH OBJECTIVES

Cancer is one of the leading causes of death in our modern world. Last year alone, approximately 7.6 million people died of cancer worldwide, which is approximately 13% of all deaths. Cancer therapy, which classically comprises surgery, chemotherapy and radiotherapy, is nowadays curative in more than 60% of all cases (based on 5 year survival rates). However, for certain types of cancer, including lung, pancreatic, and liver cancer, current treatments are largely ineffective. Decades of research has made it possible that now new therapies are entering the field of cancer medicine, including anti-angiogenic therapy, vascular-disrupting therapy and immunotherapy. One particularly promising strategy may include the use of TNF. The anticancer potency of this inflammatory cytokine has been demonstrated in numerous animal models as well as in human patients, where it can induce rapid haemorrhagic necrosis of tumors. The therapeutic application of TNF is, however, severely hampered by potentially lethal side-effects, in essence severe hypotension, shock and organ failure, limiting the clinical application of TNF to administration via the ILP technique for the treatment of advanced melanomas and soft-tissue sarcomas of the limbs. Nevertheless, the high limb salvage rates obtained after ILP with TNF and melphalan clearly demonstrate the potential of TNF as an anticancer drug. The mechanism by which TNF exerts its tumor-destructive effect is not yet fully understood, but is known to involve an increased permeability of the tumor vasculature early after injection, which may be exploited to increase the tumor uptake of classically chemotherapeutics, followed by a selective destruction of this tumor vasculature. Moreover, TNF may induce the infiltration of immune cells in the tumor, although the importance of this effect is currently unknown. Previous work in our research group has demonstrated that the anticancer and shock-inducing effects of TNF are not inevitably linked. Mice can be protected against the TNF-induced shock with the sGC inhibitor methylene blue without impairing the antitumor effect. However, using potentially lethal doses of TNF together with an inhibitor of the toxic side-effects is not an attractive solution for human cancer therapy as failure of the inhibitor could lead to death of the patient. A safer strategy to lower the systemic toxicity of TNF would be to use lower doses of TNF or low-toxic TNF muteins in combination with sensitizers for the antitumor activity of TNF, such as IFN-. The work of An Goethals in our research group has confirmed that the direct target cells of TNF for its antitumor effect are the endothelial cells of the tumor vasculature. The abnormal structure (abnormal pericyte coverage, basement membrane, cell-cell junctions) and continuous remodeling of the tumor vasculature make this vasculature highly dependent on survival signaling, which is mainly conducted via the PI3K/Akt signaling pathway.

 The first objective of the present study was to determine whether interfering with endothelial survival could potentiate the antitumor activity of TNF. PI3K, having 182 Chapter 7

a bottleneck position in the survival pathway, was selected as a first target for a proof of principle. Other downstream targets, including Akt and mTOR, as well as the upstream target VEGF-R2 were also evaluated. We have furthermore examined the vascular-disrupting nature of the combination therapy of TNF and a PI3K inhibitor by using the dorsal skinfold window chamber model and several other histological approaches (see chapter 8).

 The second objective of this study was to determine whether immune cells are required for the antitumor effect of TNF. Previous experiments conducted in our research group using IRF-1 knockout mice have suggested that certain immune cells may be essential for the full antitumor response to TNF. By using various knockout mouse strains that lack specified types of immune cells as well as by selective depletion of immune cells, we have identified one type of immune cell that may contribute to the antitumor effect of TNF. Elucidation of the complete antitumor mechanism of TNF may reveal new targets for therapy. We have tested whether selective activation of this type of immune cell by exogenous drug administration could enhance the antitumor effect of TNF (see chapter 9).

 The third objective of this study was to further elucidate the mechanism underlying the synergistic antitumor effect of TNF and IFN-. The work of Jeroen Hostens in our research group has demonstrated that, like it is the case for TNF, the direct target cell of IFN- is the endothelial cell of the tumor vasculature. We have now determined using various knockout mouse strains and by selective depletion of immune cells whether immune cells are required for the synergism of TNF and IFN-, and which types of immune cells may be involved (see chapter 10).

Chapter 8 Sensitization of the tumor vasculature to the vascular-disrupting effect of TNF by

inhibition of the PI3K/Akt survival pathway

Experimental work: Sensitization to TNF by inhibition of survival signaling 185

8. SENSITIZATION OF THE TUMOR VASCULATURE TO THE VASCULAR-DISRUPTING EFFECT OF TNF BY INHIBITION OF THE PI3K/AKT SURVIVAL PATHWAY

Abstract

TNF is a cytokine with a remarkable antitumor activity. Both in animal models and in human patients, where it is used with high efficacy to treat regional tumors of the limbs via the ILP technique, it is capable of inducing haemorrhagic necrosis of tumors. Severe shock-inducing side-effects, however, preclude systemic treatments. Previous work in our research group has established that the endothelial cells of the tumor vasculature are the main target cells for the antitumor effect of TNF. The tumor vasculature is highly dependent on survival signals provided by a broad range of pro- angiogenic factors (e.g. VEGF family members) and by adhesion receptor signaling

(e.g. integrin v3, VE-cadherin), which are commonly mediated by the PI3K/Akt pathway. To determine whether inhibition of this pathway would sensitize to the antitumor effect of TNF, B16Bl6 melanoma bearing mice were treated with sub- therapeutic TNF alone or in combination with one of the PI3K inhibitors wortmannin or LY294002. The response to either PI3K inhibitor plus TNF was striking, with tumor regression in all treated animals. We determined that by PI3K inhibition a tenfold reduction of the effective dose of TNF could be achieved. Comparable results were obtained using the Akt inhibitor perifosine or the mTOR inhibitor rapamycin. Importantly, PI3K/Akt pathway inhibition sensitized to the antitumor effect of TNF without concomitantly increasing the toxic side-effects of TNF. Using intravital microscopy and by histological examination of tumors we could demonstrate that the antitumor effect of wortmannin and TNF is characterized by a pronounced destruction of the tumor vasculature. In vivo Hoechst staining revealed a complete shutdown of the tumor blood as early as 6 hours after injection of wortmannin and TNF, with no recovery by 48 hours after injection, thus effectively starving the tumor to death. Finally, we demonstrated that TNF therapy synergizes with anti-angiogenic therapy, using a VEGF-R2 blocking antibody, for the induction of tumor necrosis in the B16Bl6 tumor model. Combining TNF with anti-angiogenic therapy is a novel concept in anticancer therapy. Taken together, we have shown that interfering with endothelial survival allows a significant reduction of the effective dose of TNF.

This study was performed in collaboration with Dr. O. Feron and Dr. Pierre Sonveaux of the Unit of Pharmacology and Therapeutics (University of Louvain (UCL) Medical School, Brussels, Belgium) for the dorsal skinfold window chamber experiments, and with Dr. Bart Vanhaesebroeck (Queen Mary University of London, London, UK) who provided PI3K knock-in mice and intellectual input.

186 Chapter 8

8.1. Introduction

TNF is an inflammatory cytokine with a remarkable antitumor activity. Both in animal models and in human patients it can cause rapid haemorrhagic necrosis of tumors 1, 2. Its therapeutic application is, however, severely hampered by potentially lethal side-effects, i.e. severe hypotension which may lead to hypovolemic shock and organ failure 3, 4. Previous work in our research group has demonstrated that the antitumor and shock-inducing effects of TNF are not inevitably linked: mice can be protected against the TNF-induced shock with the sGC inhibitor methylene blue or by the induction of tolerance by repetitive administration of low doses of TNF, without impairing the antitumor effect 5, 6. The clinical application of TNF is currently limited to administration via the ILP technique for the treatment of irresectable melanomas and soft-tissue sarcomas of the limbs. The high limb salvage rate obtained after ILP with TNF and melphalan clearly demonstrates the potential of TNF as an anticancer drug 2, 7. Yet, despite these promising results, thus far no TNF-based therapy has proven to prolong patient survival. The reason for this discrepancy lies in the fact that patient survival is mainly determined by the presence of distant metastasis, which are not affected by the regional ILP treatment. Thus, to fully exploit the anticancer activities of TNF, systemic administration will be essential. Clinical studies with systemically administrated TNF have been performed, however, without success. Doses of 10 to 50 times lower than the estimated effective dose already caused severe toxic side-effects, and doses close to the maximum tolerated dose (~150-400µg/m2) produced only rare and minimal tumor responses 8-10. One strategy to allow systemic TNF application could be to use low, tolerable doses of TNF in combination with specific sensitizers for the antitumor effect of TNF. Another strategy may be to develop low-toxic TNF muteins. However, studies in mice with hTNF, which in the murine system can be considered as a low-toxic TNF mutein, have demonstrated that these TNF muteins will most likely also require sensitization to be effective 1, 11.

The precise mechanism underlying the TNF-induced tumor necrosis remains poorly understood. It has become clear that the antitumor effect of TNF is completely host- mediated and is independent of TNF-receptor expression on tumor cells. Stoelcker and colleagues demonstrated that the antitumor effect of TNF was maintained in TNF-R2-/- mice, but absent in TNF-R1-/- mice, suggesting that the antitumor effect of TNF is mediated by triggering the TNF-R1 on host cells 12. Previous work in our research group has extended these data by showing that TNF-R1 expression on neovasculature (controlled by a Flk-1 promoter) is sufficient for the antitumor effect of TNF, indicating that the endothelial cells of the tumor vasculature are the direct target cells of TNF 13. This is further supported by the observation, in experimental tumor models as well as in human ILP patients, that the TNF-induced tumor necrosis is characterized by a selective destruction of the tumor vasculature 2, 14-16. TNF may therefore be classified as a VDA. Experimental work: Sensitization to TNF by inhibition of survival signaling 187

The PI3K/Akt pathway is one of the main regulators of cell survival. Tumor vasculature, which is in contrast to normal vasculature not stabilized by normal pericyte coverage, basement membrane structure and endothelial cell-cell interactions, and is subject to constant remodelling due to imbalance of angiogenic factors produced in the tumor, may be greatly dependent on this survival pathway 17-20. The PI3K/Akt pathway is triggered by a broad range of pro-angiogenic factors, including growth factors, such as VEGF family members, and adhesion molecules, such as 21, 22 integrin v3 . Inhibition of PI3K has been shown to sensitize endothelial cells in vitro to the cytotoxic effects of TNF 23-25. Furthermore, Matschurat and colleagues reported a synergy of the PI3K inhibitor wortmannin and TNF for the induction of tumor necrosis in a murine MethA sarcoma model 26.

In the present study we have investigated whether inhibition of the PI3K/Akt pathway would sensitize to the antitumor effect of TNF in the murine B16Bl6 melanoma model and other tumor models. We found that inhibition of PI3K or other components of the survival pathway, including Akt and mTOR, allowed a 10-fold reduction of the effective dose of TNF. We have shown that the combination therapy of PI3K inhibition and TNF resulted in a pronounced and selective destruction of the tumor vasculature. Furthermore, we found a synergy of TNF and anti-VEGF-R2 therapy. The combination of TNF and anti-angiogenic therapy is a novel concept in anticancer therapy.

8.2. Materials and methods

8.2.1. Mice

Female C57BL/6J mice were purchased from Janvier, Le Genest-Saint-Isle, France. TNF-R1-/- mice were generated and backcrossed 6 generations to the C57BL6/J background by Dr. Bluethmann, Roche Research, Basel, Switzerland 27. Kinase-death p110 knock-in (p110ki/ki) mice on C57BL/6 background were provided by Dr. Bart Vanhaesebroeck, Centre for Cell Signaling, Institute of Cancer, Queen Mary University of London, London, UK 28. B6.p110-/- mice were generated by Dr. Emilio Hirsch, University of Torino, Turin, Italy 29. The animals were housed in 14-10 hours light/dark cycles in a temperature-controlled, air-conditioned room, and received food and water ad libitum. The mice were used for experimentation at the age of 8-12 weeks. All experiments were approved by and performed according to the guidelines of the local ethics committee of Ghent University, Belgium.

188 Chapter 8

8.2.2. Cells

B16Bl6 melanoma cells and EL4 lymphoma cells were cultured in RPMI-1640 medium supplemented with 10% FCS, 2mM L-glutamine, 0,4mM Na-pyruvate, 50 IU/ml penicillin G and 50µg/ml streptomycin sulphate 30. LLC-H61 carcinoma cells and PG19 melanoma cells were cultured in DMEM medium supplemented with 10% FCS, 2mM L-glutamine, 0,4mM Na-pyruvate, 50 IU/ml penicillin G and 50µg/ml streptomycin sulphate 31. B16Bl6 cells overexpressing a dominant-negative TNF-R1 mutant (B16.dnTNF-R1) were generated in our laboratory by Dr. An Goethals. For tumor inoculation, cells were detached from the culture flask by a short EDTA treatment, washed three times in endotoxin-free, sterile DPBS (Sigma), and resuspended in DPBS.

8.2.3. Reagents

Wortmannin and rapamycin were purchased from Sigma. LY294002 was purchased from LC Laboratories. Perifosine was provided by Keryx Biopharmaceuticals. Rat anti-mouse VEGF-R2 antibodies and mouse anti-human tPA antibodies (mouse IgG control) were provided by ThromboGenix. Recombinant mTNF and hTNF were produced in our laboratory and had a specific activity of 1.28 x 108 IU/mg and 6.8 x 107 IU/mg, and an endotoxin content (LAL assay) of 6.5 EU/mg and 36 EU/mg, respectively.

8.2.4. Tumor experiments

Mice were inoculated with 6 x 105 (B16Bl6 and EL4) or 5 x 106 (LLC and PG19) tumor cells s.c. in the back (day 0). Treatment was started when the tumor size index (TSI), the product of the largest perpendicular diameters in cm, was greater than 0.5 (≥day 10). Unless otherwise indicated, PBS, hTNF (30µg), mTNF (0.7-7µg), wortmannin (5µg), Ly294002 (500µg), rapamycin (20µg), and perifosine (50-200µg) were administered daily for 10 consecutive days via paralesional injection (s.c. injection near the tumor site but outside the nodule), or i.p. injection (rapamycin). Wortmannin, perifosine and rapamycin were injected 1h before TNF. LY294002 was injected 15min before TNF. Anti-VEGF-R2 (800µg) and mouse IgG control (500µg) antibodies were injected i.p. on days 12, 14, 16, 18 and 20 after tumor inoculation.

8.2.5. MTT cytotoxicity assay

5000 B16Bl6 melanoma cells per well were seeded into a 96-well plate and incubated

overnight in a culture oven (5% CO2). Serial dilutions of mTNF (0 to 10000 IU/ml in 1/10 steps in rows) and wortmannin (0 to 20µM in 1/3 steps in columns) were added to each well. The plates were incubated for 48 hours, after which 3-(4,5- Experimental work: Sensitization to TNF by inhibition of survival signaling 189

dimethyldiazol-2-yl)-2,5 diphenyl tetrazolium bromide (MTT) was added. 6 hours later SDS was added to dissolve the precipitate. OD was measured at 595 nm. The control well with the highest OD was set at 100% growth.

8.2.6. CD31 staining

B16Bl6 tumors and other tissues were frozen in cryo-embedding compound (Microm) on dry ice and stored at –70°C. Frozen sections were cut at 5µm. Sections were fixed in acetone for 10min at –20°C and washed in PBS. Nonspecific binding sites were blocked by incubation with 0.5% BSA in PBS for 30min. Rat anti-mouse CD31 antibodies (BD Pharmingen) were diluted 1/50 in PBS/0.5% BSA, added to the slides and incubated overnight at 4°C. After washing in PBS, the slides were incubated for 1h with AlexaFluor 568 anti-rat antibodies (BD Pharmingen), diluted 1/100 in PBS/0.5% BSA. Slides were washed twice in PBS and mounted with Vectashield with DAPI (Vector Labs).

8.2.7. TUNEL staining

Terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) staining was performed by using the DeadEnd Fluorometric TUNEL System, according to the manufacturer’s instructions (Promega). In short, 5µm frozen sections were fixed for 15min in 4% PFA, washed in PBS, and permeabilized by incubation for 10min with proteinases K (20µg/ml). After washing in PBS, the sections were fixed for 5min in 4% PFA, washed again in PBS, equilibrated, and incubated with the terminal deoxynucleotidyl transferase and Cy5-dUTP for 1h at 37°C. The reaction was stopped by incubation in 2X SSC for 15min. The slides were washed three times in PBS and mounted with Vectashield with DAPI.

8.2.8. In vivo Hoechst staining

B16Bl6 tumor-bearing mice were injected i.v. with the DNA-binding dye Hoechst 33342 (10mg/kg; Sigma). 1min later the mice were killed and tumors and other tissues were immediately frozen on dry ice in cryo-embedding compound and stored at –70°C. Sections were cut at 10µm and viewed on a Zeiss Axiophot photomicroscope under UV excitation.

8.2.9. Dorsal skinfold window chamber model

Dorsal skinfold window chambers were implanted in the back of adult male C57BL/6J mice by Dr. Pierre Sonveaux of the Unit of Pharmacology and Therapeutics, Department of Medicine, UCL Medical School, Brussels, Belgium. The window chambers were inoculated with 10 x 104 LLC carcinoma cells. 20 days later 190 Chapter 8

the mice were injected into the window chamber but outside the tumor nodule with PBS, mTNF (0.7µg) and/or wortmannin (5µg). For intravital microscopy, mice were anaesthetized with a mixture of xylazine and ketamine.

8.2.10. Statistics

For tumor experiments mice were stratified into treatment groups having non- significantly different mean TSI (one-way ANOVA, P-value > 0.95) using Graphpad Prism 4.0. Results were evaluated for statistical significance by the two-tailed Mann Whitney U test. Calculations were performed using SPSS v11.0. P-values below 0.05 were considered statistical significant.

8.3. Results

8.3.1. Inhibition of PI3K sensitizes to the antitumor effect of TNF

To determine whether inhibition of the PI3K/Akt pathway would sensitize to the antitumor effect of TNF, we used two different broad spectrum PI3K inhibitors: the potent, irreversible inhibitor wortmannin and the ATP-competitive, reversible inhibitor LY294002. One strategy to avoid lethal toxicity of TNF is to use low-toxic TNF muteins. In mice hTNF can be regarded as such a mutein: hTNF does not interact with the murine TNF-R2, is cleared much faster, and, consequently, causes only low systemic toxicity. In our B16Bl6 tumor model, treatment with hTNF results in a static effect, providing us a therapeutic window to test sensitizing agents. C57BL/6 mice bearing a subcutaneous B16Bl6 tumor were treated daily for ten consecutive days with wortmannin, LY294002, hTNF or a combination of hTNF and either PI3K inhibitor (Fig.8.1. and Table 8.1.). A range of LY294002 doses was tested (10µg – 500µg/mouse/day), and the dose that led to optimal tumor regression was determined to be 500µg per mouse per day. In control (PBS-treated) mice tumors grew exponentially. Treatment with LY294002 or wortmannin alone had respectively no or a marginal inhibiting effect on tumor growth. However, the response to either PI3K inhibitor plus hTNF was striking, with tumor regression in all treated animals. Moreover, complete regression (no palpable tumor at the 10th day of treatment) of the B16Bl6 tumor was obtained in four of seven treated animals after treatment with wortmannin plus hTNF and in six of seven treated animals after treatment with LY294002 plus hTNF. A regrowth of the tumor from the rim of the lesion was observed in most animals after the treatment was stopped, which is characteristic of a vascular-targeting therapy.

Experimental work: Sensitization to TNF by inhibition of survival signaling 191

Fig. 8.1. PI3K inhibition sensitizes to the antitumor effect of hTNF. (A) Tumor size and relative body weight after treatment with wortmannin and hTNF. (B) Idem after treatment with LY294002 and hTNF. The line under the graph represents the treatment period. Data are shown as mean ± SEM, n = 7. Statistical analyses (Mann Whitney U test, two-tailed) of hTNF vs. wortmannin + hTNF and hTNF vs. LY294002 + hTNF, * p < 0.05, # p < 0.01. Average body weight of the mice at day 10 was respectively 19.3g and 19.9g.

Treatment with wortmannin or LY294002 alone did not cause apparent systemic toxicity. However, severe toxicity was observed when mice were treated with wortmannin plus hTNF. All animals appeared cachectic. Average weight loss reached a maximum of approximately 19% of total body weight at the fifth day of treatment, and three out of seven mice died before the end of treatment. When LY294002 was used in combination with hTNF, no significant increase in systemic toxicity was observed, suggesting that the observed toxicity is wortmannin-specific and is not linked to PI3K inhibition per se (Fig.8.1. and Table 8.1.).

192 Chapter 8

Table 8.1. Responses to treatments with wortmannin, LY294002 and hTNF. Tumor response § Treatment CR (%) PR (%) S (%) PD (%) D (%) PBS (n = 14) – – – 93 7 Wortmannin (n = 7) – – – 100 – LY294002 (n = 7) – – – 100 – hTNF (n = 14) – 29 21 50 – Wortmannin + hTNF (n = 7) 57 – – – 43 LY294002 + hTNF (n = 7) 86 – 14 – – § TSI at day 19 after tumor inoculation was compared to that on day 10. Tumor response was classified as follows: PD = progressive disease, increase of more than 25%; S = static, tumor size between –25% and 25%; PR = partial regression, decrease between –25% and –99%; CR = complete regression, no palpable tumor. D = animal died during treatment. – = none.

We next evaluated whether inhibition of PI3K would allow a significant reduction of the effective dose of mTNF. When administered paralesional for ten consecutive days, a dose of 7µg of mTNF per day is required for inducing complete regression of a s.c. B16Bl6 tumor (Fig.8.2.). This treatment induces, however, marked systemic toxicity which becomes lethal for approx. 10% of treated mice. To determine the minimal effective dose of mTNF when combined with the PI3K inhibitor wortmannin, we performed a dose-response experiment in our B16Bl6 tumor model. mTNF doses ranging from 0.4µg to 7µg per day were tested. The response to low-dose mTNF plus wortmannin was synergistic# and the minimal effective dose of mTNF was determined to be 0.7µg per day. mTNF alone at this dose induced only on average a static effect, with complete response in one out of six treated animals, while in combination with wortmannin, this dose caused complete tumor regression in five out of six treated animals (one mouse died early during treatment) (Fig.8.2.). Interestingly, this treatment caused only a moderate increase in systemic toxicity compared to low-dose mTNF alone. When the reversible PI3K inhibitor LY294002 was given in combination with low-dose mTNF, a similar synergistic antitumor response was observed, inducing complete regression of the B16Bl6 tumor in four out of six treated animals, but here no corresponding increase in systemic toxicity was observed. These results suggest that a selective increase of the antitumor activity of TNF can be achieved by inhibition of PI3K signaling.

# As used in this thesis, the term ‘synergistic’ refers to the phenomenon in which the common effect generated by treatment with two molecules is greater than the effect achieved with either molecule alone. It does not necessarily mean that the generated effect is greater than the sum of the individual effects of these molecules. Experimental work: Sensitization to TNF by inhibition of survival signaling 193

Fig. 8.2. PI3K inhibition allows 10-fold reduction of the effective dose of mTNF. (A) Dose- response curve of B16Bl6 tumor growth in response to treatment with decreasing doses of mTNF. (B) Growth curve of B16Bl6 tumor after treatment with wortmannin, LY294002 and low-dose mTNF (0.7µg). All data are presented as mean + SEM, n = 6. The numbers between brackets indicate the number of mice with complete tumor regression on day 9 of treatment. § One mouse died during treatment with wortmannin + mTNF.

The synergistic antitumor activity of wortmannin plus low-dose mTNF was also confirmed in other tumor models: the melanin-negative PG19 melanoma, the EL4 lymphoma and the LLC carcinoma model (Fig.8.3.).

194 Chapter 8

Fig. 8.3. Treatment with wortmannin and low-dose mTNF is similarly effective in other tumor models. Tumor response of B16Bl6 melanoma, EL4 lymphoma, LLC carcinoma and PG19 melanoma to treatment with wortmannin and low-dose mTNF (0.875µg). Data are shown as mean ± SEM, n = 6.

8.3.2. The antitumor effect of TNF plus wortmannin is host-mediated

Although previous research had shown that the antitumor effect of TNF is completely host-mediated, this did not certify that our combined treatment of TNF plus PI3K inhibitors acted by the same mechanism. To determine whether inhibition of PI3K could sensitize the TNF-resistant B16Bl6 cells to the cytotoxic effects of TNF in vitro, an MTT cytotoxicity assay was performed (Fig.8.4.). The combination of wortmannin and TNF had a modest synergistic effect, causing a growth inhibition of the B16Bl6 cells of approx. 50% at the highest concentrations used. Interestingly, no cell death was observed. To investigate whether direct effects on the B16Bl6 tumor cells contribute to the in vivo synergistic antitumor effect of our combination therapy, wild-type C57BL/6 mice were inoculated with B16Bl6 cells rendered insensitive to TNF by overexpression of a dominant-negative TNF-R1 mutant (B16.dnTNF-R1), and TNF-R1-/- mice were inoculated with B16Bl6 cells with normal TNF-R1 expression (Fig.8.4.). In the wild-type mice bearing a TNF-insensitive B16Bl6 tumor treatment with wortmannin plus hTNF induced complete tumor regression in all treated mice, while, by contrast, treatment of TNF-R1-/- mice had no significant effect Experimental work: Sensitization to TNF by inhibition of survival signaling 195

on tumor growth. Taken together, these results indicate that the synergistic antitumor effect of TNF plus wortmannin is host-mediated, and that direct cytotoxic or cytostatic effects on the melanoma cells are unimportant.

Fig. 8.4. The antitumor effect of wortmannin plus TNF is host-mediated. (A) 48h MTT cytotoxicity assay. (B) Tumor growth curve of C57BL/6 mice inoculated with B16Bl6 cells rendered insensitive to TNF by overexpression of a dominant-negative TNF-R1 mutant (B16.dnTNF-R1). (C) Tumor growth curve of TNF-R1-/- mice inoculated with B16Bl6 cells. The line under the graph represent the treatment period. Tumor growth graphs are shown as mean ± SEM, n = 6.

8.3.3. Role of p110 isoforms in the sensitivity to the antitumor effect of TNF

PI3K is ubiquitously expressed and is as such not an interesting therapeutic target. However, only the class I PI3K family members are involved in survival signaling downstream of receptor tyrosine kinases, GPCRs and non-receptor tyrosine kinases. Structurally, class I PI3Ks consist of a p110 catalytic subunit that is constitutively associated with a regulatory subunit. There are four p110 isoforms: p110, -, - and -. The isoforms p110 and – are ubiquitously expressed and homozygous knockout mice are embryonic lethal. The isoforms p110 and – on the other hand have a limited tissue distribution and their gene-deficient mice are viable and healthy 32. The latter two isoforms are predominantly expressed in leukocytes, although their presence has also been reported in other cells types, including endothelial cells 33, 34. 196 Chapter 8

To determine whether one of these isoforms could confer survival signals to endothelial cells of the tumor vasculature which may protect these cells against the vascular-disrupting effect of TNF, p110-/- mice and p110ki/ki mice bearing a s.c. B16Bl6 tumor were treated with hTNF (Fig.8.5.). In both strains the response to hTNF treatment was static, on average, and not significantly different from the response to hTNF observed in wild-type mice, suggesting that neither p110 or p110 play a critical role in determining the sensitivity to the antitumor effect of TNF. Interestingly however, p110ki/ki mice were markedly more sensitive to the toxic side- effects of TNF: 5 out of 18 mice died due to treatment-induced toxicity, while lethality due to hTNF treatment was not observed in wild-type control mice. Moreover, treatment of p110ki/ki mice with either combination of hTNF and wortmannin or hTNF and IFN- resulted in 100% lethality.

Fig. 8.5. p110-/- and p110-/- mice are not sensitized to the antitumor effect of TNF. (A) Tumor size of B16Bl6 tumor growing in p110-/- mice after treatment with hTNF. (B) Idem in p110ki/ki mice. The line under the graph represents the treatment period. Data are shown as mean ± SEM. The numbers between brackets below the curve indicate the number of mice that died due to treatment-induced toxicity.

8.3.4. Inhibition of PI3K sensitizes to the vascular-disrupting effect of TNF

First signs of necrosis in subcutaneous B16Bl6 tumors treated with wortmannin plus hTNF were macroscopically visible at 48 hours after the first injection as a black discoloration of the skin overlaying the tumor. By 72 hours, central necrosis of the tumor became apparent, followed by a progressive collapse of tumor from the centre towards the rim, which was in some cases associated with bleeding. Tumor necrosis was in most cases complete by 5 days after the start of treatment. Similar observations were made in haematoxylin & eosin-stained tumor sections: extensive haemorrhagic necrosis of the tumor became apparent from 48 hours after the first injection of hTNF plus wortmannin and necrosis was complete at 72 hours. Destruction of the tumor vasculature appeared to coincide with the overall necrosis of the tumor. However, Experimental work: Sensitization to TNF by inhibition of survival signaling 197

palpation of tumors treated with wortmannin plus hTNF revealed that there was a marked loss of rigidity by 24 hours after injection, which is indicative of a decrease in interstitial fluid pressure in the tumors, suggesting that early vascular events are taking place before necrosis becomes apparent.

Fig. 8.6. Effects of wortmannin and hTNF on tumor vessel perfusion. Functional tumor (B16Bl6) vessels were visualized by in vivo Hoechst staining (blue) 6h and 48h after injection of hTNF. Wortmannin was injected 1h before hTNF. Adjacent pictures were taken and stitched to show the whole tumor section. Scale bar is approx. 2mm.

We performed an in vivo Hoechst staining to assess the functionality of the tumor blood vessels after injection of wortmannin and hTNF (Fig.8.6.). 6 hours after injection of hTNF alone the tumor blood flow was significantly reduced compared to control tumors. Functional blood vessels were present in the rim of the tumor, particularly near the skin. A similar reduction was observed at 24 hours after injection. However, at 48 hours after injection of hTNF, tumor blood flow was almost completely recovered, except for the necrotic centre of the tumor. When wortmannin was injected prior to hTNF, perfusion of the tumor vasculature was completely blocked at 6 hours after treatment, but in contrast to treatment with hTNF alone, no recovery could be observed at 48 hours after injection. This persistent block of tumor blood flow could effectively starve the tumor cells to death. Together, these results suggest that tumor necrosis may be secondary to a complete block in tumor blood flow, and that endothelial PI3K signaling may be essential to survive the TNF- induced vascular effects.

198 Chapter 8

To visualize the gross effects of PI3K inhibition and TNF on tumor blood vessels, we used the dorsal skinfold window chamber model (Fig.8.7.). 20 days after tumor inoculation, when the LLC tumors were visible as a small protrusion of the skin opposing the window, the mice were injected with wortmannin and low-dose mTNF inside the window chamber but outside the tumor nodule. The LLC tumors had a dense microvascular network at that time. Injection of low-dose mTNF induced transient dilation of larger vessels in the window chamber visible at 6 hours after injection, with signs of vascular leakage (swelling of tissue in the window chamber) in 2 of 4 treated mice. No haemorrhage was observed. The tumor microvasculature appeared to be largely unaffected by the treatment and tumors grew as in control mice. Injection of wortmannin caused local haemorrhage at the site of injection in 3 of 4 treated animals, which interfered with the visualization of the tumor vasculature. However, where the tumor microvasculature was visible, no effect could be observed and treatment did not affect tumor growth. Bleeding was never seen in PBS-treated animals, suggesting that the haemorrhage observed after injection of wortmannin is drug-related. The effect of the combined injection of wortmannin and low-dose mTNF was striking, with complete destruction of the vasculature in the tumor area. This was associated with swelling and haemorrhage which filled up the whole window chamber in all treated animals as early as 6 hours after injection. At 72 hours after injection, when haemorrhage was starting to get cleared, the tumor area was largely devoid of vasculature, while intact microvasculature was readily visible outside the tumor area, suggesting that the vascular-disrupting effect of our combination treatment is selective for the tumor neovasculature (Fig.8.8.). Tumors were undetectable at 72 hours after injections of wortmannin plus mTNF.

Experimental work: Sensitization to TNF by inhibition of survival signaling 199

Fig. 8.7. Effects of wortmannin and mTNF on tumor vasculature in the dorsal skinfold window chamber model. Pictures of the tumor area were taken before and 6h, 12h, 24h and 48h after injection of wortmannin and low-dose mTNF (0.7µg). Representative pictures are shown. Original magnification: ×5 objective.

200 Chapter 8

Fig. 8.8. Detail of a dorsal skinfold window chamber 72h after injection of wortmannin and low-dose mTNF showing intact microvasculature outside the tumor area (left of dashed line). Original magnification: ×5 objective.

The vascular-disrupting effect of treatment with wortmannin and low-dose mTNF was further investigated by staining sections of s.c. B16Bl6 tumors for the endothelial marker CD31 (Fig.8.9.A.). B16Bl6 tumors were well vascularized. Tumor vessels were narrow, having small or undetectable lumina, indicative of high interstitial fluid pressure in the tumor. Treatment with wortmannin had no apparent effect on the tumor vasculature. Blood vessels of mTNF-treated tumors had markedly enlarged lumina, which was already visible at 4 hours after injection. Blood clots were present inside the enlarged vessels. However, blood coagulation may have occurred during dissection as no perfusion was performed to remove the blood from the tumor vessels before excision of the tumor. The endothelial lining of most vessels in mTNF-treated tumors remained intact. At 7 hours after injection, some tumor vessels could be observed that appeared to be filled with leukocytes. At 24 hours, the majority of tumor vessels was intact. Necrotic areas that contained some endothelial cell debris were present near the centre of the tumor, suggesting that treatment with low-dose mTNF may induce disruption of some tumor vessels. After treatment with the combination of wortmannin and low-dose mTNF, tumor vessels were enlarged as after treatment with mTNF alone, and intravascular blood clots were also visible. However, treatment with wortmannin plus mTNF induced pronounced disruption of the vessel wall. First gaps in the endothelial lining of the vessels were apparent from 6 hours after treatment. Moreover, the space between the vascular endothelium and the surrounding tumor cells was markedly increased, suggesting that treatment may induce loss of both endothelial cell-cell and cell-matrix contact. The background staining surrounding the enlarged vessels is suggestive of vascular leakage. Disruption of the vessel wall was complete by 24 hours after treatment. Treatment affected the majority of tumor vessels. However, some intact vessels could be found in the periphery. Experimental work: Sensitization to TNF by inhibition of survival signaling 201

202 Chapter 8

Experimental work: Sensitization to TNF by inhibition of survival signaling 203

Fig. 8.9. (pp201-202) Treatment with wortmannin and mTNF induces selective disruption of the tumor vasculature. B16Bl6 tumor sections of 5h, 6h, 7h and 24h after injection of low-dose mTNF (Ld mTNF; 0.7µg) were stained for (A) endothelial cells by CD31 staining (bright red), and (B) cell death by TUNEL staining (bright green). Wortmannin (Wm) was injected 1h before mTNF. Sections were counterstained with DAPI (blue). Aspecific staining is visible in necrotic areas and clotted blood (dim red and green). Representative blood vessels are shown. Original magnification: ×20 objective.

We performed a TUNEL staining to visualize cell death in these tumor sections (Fig.8.9.B.). In all tumors sections necrotic areas were present near the centre of the tumor. Healthy blood vessels adjacent to these necrotic areas were surrounded by a cuff of healthy tumor cells, revealing a typical island structure. In mTNF-treated tumors, some TUNEL-positive tumor vessels could be seen at 24 hours after treatment, and some necrotic areas were present close to apparently intact tumor vessels, suggesting that cell death may be caused by insufficient vessel perfusion. After treatment with wortmannin and mTNF, the first TUNEL-positive vessels could be observed already 7 hours after treatment, and by 24 hours the majority of tumor vessels stained TUNEL-positive. Apoptotic endothelial cells were difficult to distinguish as dying vessels were surrounded by TUNEL-positive tumor cells. Cell death was mostly necrotic. At the time points investigated, endothelial cell death appeared to coincide with tumor necrosis, suggesting that endothelial cell death is secondary to vessel occlusion.

These spectacular vascular-disrupting effects prompted us to test whether a single injection of wortmannin and low-dose mTNF would be sufficient to induce complete regression of a B16Bl6 tumor. However, we found that this was not the case. One treatment with this combination therapy induced, on average, a static response. At least two consecutive treatments (24 hour interval) were required to induce complete tumor regression.

8.3.5. Inhibition of Akt or mTOR sensitizes to the antitumor effect of TNF

To evaluate whether this would be a generic solution in which PI3K inhibitors could be substituted by inhibitors of other components of the survival pathway, we first selected two targets downstream of PI3K: Akt and mTOR. Akt is directly downstream of class I PI3Ks and an obligatory component of the survival pathway. Perifosine is an alkylphospholipid that structurally resembles PI(3,4,5)P3, the principal product of class I PI3Ks. It was found to be a specific Akt inhibitor, most likely by interfering with the PH domain-mediated membrane recruitment and activation of Akt 35. Perifosine is currently being evaluated in clinical trials for the treatment of multiple tumor types. We treated B16Bl6 tumor-bearing mice with perifosine, sub-therapeutic TNF (hTNF or low-dose mTNF), or a combination of these drugs (Fig.8.10. and Table 8.2.). Treatment with perifosine had no significant effect on tumor growth. Treatment 204 Chapter 8

with a combination of perifosine and either hTNF or low-dose mTNF induced a pronounced synergistic antitumor effect, with tumor regression in five out of seven and six out of seven treated mice, respectively. Moreover, complete tumor regression was obtained in two out of seven mice for both combinations. Mildly increased systemic toxicity was observed after treatment with perifosine plus hTNF. Average weight loss reached a maximum of approximately 18% of total body weight at the fourth day of treatment. No lethality occurred. Increased toxicity was not observed after treatment with perifosine and low-dose mTNF.

Fig. 8.10. Inhibition of Akt sensitizes to the antitumor effect of TNF. (A) Growth curve of B16Bl6 tumor after treatment with perifosine (50µg) and hTNF. (B) Growth curve of B16Bl6 tumor after treatment with perifosine (200µg) and low-dose mTNF (0.7µg). The line under the graph represents the treatment period. All data are shown as mean ± SEM, n = 7.

We next tested the mTOR inhibitor rapamycin. mTOR was selected as a target because (1) although mTOR is not a component of the survival pathway sensu stricto, it is a key regulator of proteins synthesis, which may be important to survive TNF stimulation; (2) the anticancer potency of rapamycin has previously been demonstrated in numerous tumor models, and mTOR inhibitors are currently being tested in clinical trials; (3) the PI3K inhibitor LY294002 is also a potent inhibitor of mTOR; and (4) treatment with rapamycin has been shown to inhibit Akt activity in HUVECs stimulated with TNF 36-38. We evaluated the antitumor effect of rapamycin, alone or in combination with TNF, in the B16Bl6 tumor model (Fig.8.11. and Table 8.2.). In contrast to PI3K inhibitors, treatment with rapamycin alone caused a significant delay in tumor growth. This corresponded with the cytotoxic effect of rapamycin on B16Bl6 cells in vitro: a concentration of 20µM rapamycin caused a growth inhibition of approx. 60% in the 48 hour MTT cytotoxicity assay due to an almost complete inhibition of cell proliferation (no death cells or cell debris could be observed). However, in this in vitro assay no additional effects of TNF could be observed, while in vivo the response to rapamycin plus either hTNF or low-dose mTNF was striking, with regression of the B16Bl6 tumor in four out of six and all treated animals, respectively. Complete tumor regression was observed in respectively Experimental work: Sensitization to TNF by inhibition of survival signaling 205

two out of six treated animals and all treated animals. Notably, rapamycin did not increase the apparent toxic side-effects of TNF. The synergism of rapamycin and TNF was not existent in TNF-R1-/- mice, suggesting that this is also a host-mediated effect.

Fig. 8.11. Rapamycin sensitizes to the antitumor effect of TNF. (A) Growth curve of B16Bl6 tumor after treatment with rapamycin and hTNF, and (B) after treatment with rapamycin and low-dose mTNF (0.7µg). The line under the graph represents the treatment period. All data are shown as mean + SEM. For PBS and rapamycin alone n = 5, for other groups n = 6.

Table 8.2. Responses to treatments with perifosine, rapamycin, and TNF. Tumor response § Treatment CR (%) PR (%) S (%) PD (%) D (%) PBS (n = 24) – – – 92 8 Perifosine 50µg (n = 7) – – – 100 – Perifosine 200µg (n = 7) – – – 86 14 Rapamycin (n = 10) – – – 90 10 hTNF (n = 13) – 15 23 62 – Perifosine + hTNF (n = 7) 29 43 14 14 – Rapamycin + hTNF (n = 6) 33 33 33 – – Low-dose mTNF (n = 13) 15 8 8 61 8 Perifosine +mTNF (n = 7) 29 57 – 14 – Rapamycin +mTNF (n = 6) 100 – – – – § TSI before start of treatment was compared to that on the 10th day of treatment. Tumor response was classified as follows: PD = progressive disease, increase of more than 25%; S = static, tumor size between –25% and 25%; PR = partial regression, decrease between –25% and –99%; CR = complete regression, no palpable tumor. D = animal died during treatment. – = none.

206 Chapter 8

8.3.6. Synergy of TNF and anti-angiogenic therapy

We showed that general inhibition of the PI3K/Akt pathway, using broad-spectrum inhibitors of PI3K, Akt or mTOR, significantly increased the antitumor activity of TNF, without concomitantly increasing overt toxic side-effects of TNF. However, PI3K, Akt and mTOR are ubiquitously expressed and have important functions in healthy tissue, and therefore these enzymes may not be ideal therapeutic targets, unless isoform-selective inhibitors could be used to increase specificity. Endothelial cells of the tumor vasculature, the principal target cells of TNF for its antitumor effect, are nowadays specifically targeted by an increasing number of anti-angiogenic drugs, many of which target growth factors receptors or adhesion molecules capable of signaling via the PI3K/Akt pathway. The contribution of each target molecule to total PI3K activation is, however, likely to differ depending on the tumor type. The growth of s.c. B16Bl6 tumors has recently been shown to be significantly suppressed by treatment with neutralizing anti-VEGF-R2 antibodies 39. We tested whether anti- VEGF-R2 treatment would sensitize to the antitumor effects of TNF in the B16Bl6 tumor model (Fig.8.12. and Table 8.3.). Treatment with mouse IgG control antibodies had no effect on tumor growth or response to TNF. Treatment with anti-VEGF-R2 antibodies, however, induced a significant suppression of tumor growth, comparable to the effect reported by Fischer and colleagues 39. The antitumor effect of TNF was markedly increased by concomitant treatment with anti-VEGF-R2 antibodies. Combined treatment induced tumor regression in all treated animals, and complete tumor regression (no palpable tumor) was observed in three out of six animals after anti-VEGF-R2 plus hTNF treatment and two out of six animals after anti-VEGF-R2 plus low-dose mTNF treatment. Complete tumor regressions did not occur in hTNF- or low-dose mTNF-treated mice. A mildly increased toxicity was observed in mice treated with anti-VEGF-R2 antibodies and hTNF, which was not observed when anti- VEGF-R2 treatment was combined with low-dose mTNF. One animal died during treatment with anti-VEGF-R2 antibodies and low-dose mTNF, most likely due to extensive bleeding which is sometimes observed during haemorrhagic necrosis of the tumors. Our results suggest that VEGF is an important survival factor for endothelial cells of the tumor vasculature in B16Bl6 tumors, and that anti-angiogenic therapy may be used to increase the efficacy of sub-therapeutic TNF therapy.

Experimental work: Sensitization to TNF by inhibition of survival signaling 207

Fig. 8.12. Anti-VEGF-R2 treatment synergizes with TNF. Growth curve of B16Bl6 tumor after treatment with (A) mouse IgG (mIgG) control antibody and hTNF, (B) mIgG control antibody and low- dose mTNF (0.7µg), (C) anti-VEGF-R2 antibody and hTNF, and (D) anti-VEGF-R2 antibody and low- dose mTNF. The line under the graph represents the treatment period. All data are shown as mean + SEM, n = 6.

208 Chapter 8

Table 8.3. Responses to treatments with IgG control antibodies, anti-VEGF-R2 antibodies, and TNF. Tumor response § Treatment CR (%) PR (%) S (%) PD (%) D (%) PBS (n = 6) – – – 100 – Mouse IgG control (n = 6) – – – 100 – Anti-VEGF-R2 (n = 6) – – – 100 – hTNF (n = 6) – 50 – 50 – Mouse IgG control + hTNF (n = 6) – – – 83 17 Anti-VEGF-R2 + hTNF (n = 6) 50 50 – – – Low-dose mTNF (n = 6) – 33 17 33 17 Mouse IgG control + mTNF (n = 6) – 50 – 33 17 Anti-VEGF-R2 + mTNF (n = 6) 33 50 – – 17 § TSI at day 12 after tumor inoculation was compared to that on day 21. Tumor response was classified as follows: PD = progressive disease, increase of more than 25%; S = static, tumor size between –25% and 25%; PR = partial regression, decrease between –25% and –99%; CR = complete regression, no palpable tumor. D = animal died during treatment. – = none.

8.4. Discussion

The primary findings of this study are: (1) inhibition of the PI3K/Akt pathway, using broad spectrum inhibitors of PI3K, Akt or mTOR, significantly increases the antitumor activity of TNF in the B16Bl6 melanoma and other tumor models, allowing at least a tenfold reduction of the effective dose of TNF; (2) PI3K/Akt pathway inhibition may increase the antitumor activity of TNF without concomitantly increasing its toxic side-effects; (3) the sensitizing effect of PI3K or mTOR inhibition to TNF is host-mediated; (4) inhibition of PI3K sensitizes the tumor vasculature to the vascular-disrupting effect of TNF; and (5) anti-angiogenic therapy using VEGF-R2 blocking antibodies synergizes with TNF for the induction of tumor regression in the B16Bl6 tumor model.

The dose-limiting toxicities of TNF are hypotension, liver toxicity, thrombocytopenia, leukopenia, and multi-organ failure, generally called ‘septic shock-like syndrome’ or SIRS 2, 8, 9, 40-42. Mice can be completely protected against the TNF-induced shock and lethality with the sGC inhibitor methylene blue, suggesting that both the lethal shock- inducing and antitumor effects of TNF are vascular effects 6, 43. However, we have evidence that the antitumor and shock-inducing effects are mediated by different target cell populations. Transgenic Flk-1 TNF-R1 reactivation knockout mice, expressing the TNF-R1 solely on endothelial cells, were markedly less sensitive to the lethal shock-inducing effect of TNF, while the antitumor effect was preserved 13. We and others have also shown that the toxic side-effects and antitumor effects are not linked at the signaling level. Inhibition of the TNF-induced shock with methylene Experimental work: Sensitization to TNF by inhibition of survival signaling 209

blue did not affect the antitumor activity of TNF 6. Likewise, two acute phase proteins, 1-acid-glycoprotein and 1-antitrypsin, were shown to protect against the TNF-induced liver toxicity without affecting its antitumor activity 43. More recently, zinc-induced expression of HSP70 was shown to confer protection against TNF- induced toxicity without interfering with the antitumor activity of TNF/IFN- therapy 44. In the present study we have demonstrated that PI3K/Akt/mTOR pathway inhibition may also allow to discriminate between the antitumor and toxic side-effects of TNF. Increased systemic toxicity was only observed when the PI3K inhibitor wortmannin was used, which may be due to the irreversible nature of this inhibitor. Toxicity was not observed when using the reversible broad-spectrum PI3K inhibitor LY294002, indicating that toxicity is not linked to PI3K inhibition per se. Interestingly, p110ki/ki mice were more sensitive to the TNF-induced toxic side- effects, suggesting that a PI3K-dependent protective effect may exist. We have furthermore shown that the PI3K isoforms p110 and p110 are no determinants of the sensitivity to the antitumor effect of TNF. Together these data suggest that p110- or p110-selective inhibitors, which are currently being developed (e.g. PIK-75, PI- 103 and TGX-221 45, 46), have the best chance to mimic the sensitizing effect of broad- spectrum PI3K inhibition with minimal risk of toxicity.

It was previously shown that the antitumor effect of TNF is completely host-mediated, via triggering of the TNF-R1 on the tumor vasculature 12, 13. We have shown, by using TNF-R1-/- mice, that the combination therapy of TNF and PI3K inhibition is also a host-mediated effect and that this treatment results in a pronounced destruction of the tumor vasculature, suggesting a similar antitumor effect as TNF monotherapy. This is further supported by the finding that the fast-cleared PI3K inhibitor LY294002 was largely ineffective when injected one hour before TNF, while its effectiveness was comparable to that of the irreversible inhibitor wortmannin when injected 15 minutes before TNF, suggesting that the sensitizing effect to TNF is only existent when PI3K inhibition and TNF stimulation occur at the same time, and thus most likely in the same cell.

The mechanism underlying the vascular-disrupting effect of TNF is still largely unknown. We used intravital microscopy as well as several histological staining methods to visualize the gross effects of treatment with wortmannin and TNF on the tumor vasculature. Our data suggest that very rapidly after injection treatment leads to vascular leakage and haemorrhage, followed by complete shutdown of the blood flow in sensitive tumor vessels by 6 hours after injection. Destruction of the vessel wall was complete by 24 hours after injection. Matschurat and colleagues have shown that inhibition of PI3K led to enhanced TNF-induced expression of tissue factor, the prime initiator of blood coagulation, in HUVECs. They furthermore demonstrated that co- injection of wortmannin promoted the TNF-induced necrosis of methA tumors, 210 Chapter 8

thereby suggesting that the expression of tissue factor on endothelial cells of the tumor vasculature, leading to thrombus formation and occlusion of the tumor vessels, would be responsible for the antitumor effect of this combination therapy 26. Our findings, however, do not fully support this hypothesis. Intravascular thrombus formation is a protective mechanism to prevent haemorrhage in the case of vascular damage. The extensive vascular leakage and haemorrhage observed in the dorsal skinfold window chambers and sections of tumors treated with wortmannin and TNF suggest that vessel occlusion is secondary to disruption of the vessel wall. There is substantial in vitro data showing that PI3K inhibition can sensitize endothelial cells to TNF-induced cytotoxicity 24, 25, 47. Particularly interesting in this context is the finding that TNF itself is capable of activating PI3K and Akt 23. PI3K signaling was shown to be essential in certain cell types for the activation of NF-B, the main transcription factor responsible for TNF-induced anti-apoptotic signaling. However, this was not the case in HUVECs 48. Madge and co-workers demonstrated that TNF triggers a PI3K-mediated anti-apoptotic response in HUVECs that is independent of Akt and NF-B, and that inhibition of PI3K results in cathepsin B- dependent, but caspase-independent cell death of these endothelial cells 25. These findings, however, do not fit with our in vivo data where the sensitizing effect of PI3K inhibition could be mimicked by Akt inhibition. Moreover, after treatment with wortmannin and TNF DNA fragmentation in tumor endothelial cells, a hallmark of apoptotic cell death, was only observed after disruption of the vessel wall and vessel occlusion had occurred. At that time endothelial cell death may also be a consequence of cell detachment (anoikis) or mere starvation. We hypothesize that the initial event in the vascular-disrupting effect of treatment with wortmannin and TNF is a loss of endothelial cell-cell contact and/or cell-matrix contact. Gaps in the endothelial lining of the tumor vessels were readily visible at 6 hours after treatment. Similarly, Menon and colleagues reported the formation of gaps in the endothelial lining of the tumor vasculature between 3 and 6 hours after treatment with TNF in human melanoma xenografts in mice as well as in human melanoma patients after ILP, and they attributed this effect to a disappearance of VE- cadherin from endothelial cell-cell junctions 49. The adherens junction component VE- cadherin is known to play a critical role in endothelial cell-cell connections and tumor angiogenesis 50. It was recently shown that TNF can induce tyrosine phosphorylation of the intracellular domain of VE-cadherin, thereby weakening the VE-cadherin-based junctions and facilitating the formation of intercellular gaps 51. Disruption of VE- cadherin junctions may lead to a shutdown of tumor blood flow by a mechanism of vascular collapse, in analogy with the vessel-disrupting mechanism reported for the tubulin binding drug CA4P. Disassembly of VE-cadherin junction would cause haemorrhage, an increase of interstitial fluid pressure, and eventually the collapse of the weakened tumor vessels 52. It is, however, currently unclear whether PI3K inhibition could exacerbate the TNF-induced disassembly of VE-cadherin junctions.

The integrins v3 and v5 play an important role in endothelial cell-matrix adhesion Experimental work: Sensitization to TNF by inhibition of survival signaling 211

in neovasculature 53. Rüegg and colleagues showed that treatment of HUVECs with

TNF and IFN- suppressed the activation of integrin v3, leading to a decrease in endothelial cell adhesion and survival, and they observed detachment and apoptosis of

v3-positive endothelial cells in the tumor vasculature of melanoma patients after ILP with TNF plus IFN- 54. Furthermore, they demonstrated that Akt activity can strengthen integrin-dependent endothelial cell adhesion and can protect HUVECs 55, 56 against cell death induced by TNF and/or integrin v3/v5 inhibition . Thus, TNF therapy is known to interfere with both endothelial cell-cell and cell-matrix adhesions. PI3K/Akt pathway inhibition may exacerbate this effect and/or compromise endothelial cell survival during periods of cell detachment.

The observation that mTOR inhibition mimics the sensitizing effect of PI3K or Akt inhibition is not surprising as persistent mTOR inhibition was recently shown to inhibit Akt activation in HUVECs and other cell types 37. However, of all compounds tested, rapamycin was the most potent one, which may suggests that mTOR, as an important regulator of protein synthesis, could have an additional protective function not directly linked to Akt activation. It is long known that inhibition of protein synthesis can sensitize cultured endothelial cells to TNF-induced cytotoxicity 57, 58.

Finally, we demonstrated that TNF therapy may synergize with anti-angiogenic therapy, which is a novel concept in anticancer therapy. We tested a VEGF-R2 blocking antibody known to have anti-angiogenic activity in the B16Bl6 tumor model 39. This anti-VEGF-R2 antibody was able to mimic the sensitizing effect of broad- spectrum PI3K/Akt pathway inhibition, suggesting that VEGF is an important survival factor for endothelial cells in B16Bl6 tumors. Interestingly, VEGF-R2 has been shown to interact with both VE-cadherin and integrin v3, and in both cases these interactions were linked to downstream PI3K activation, suggesting an important survival function of these complexes in VEGF-driven angiogenesis 59, 60.

212 Chapter 8

8.5. References

1. Brouckaert, P.G., Leroux-Roels, G.G., Guisez, Y., Tavernier, J. & Fiers, W. In vivo anti- tumour activity of recombinant human and murine TNF, alone and in combination with murine IFN-gamma, on a syngeneic murine melanoma. International journal of cancer 38, 763-769 (1986). 2. Lejeune, F.J., Lienard, D., Matter, M. & Ruegg, C. Efficiency of recombinant human TNF in human cancer therapy. Cancer Immun 6, 6 (2006). 3. Tracey, K.J. et al. Shock and tissue injury induced by recombinant human cachectin. Science (New York, N.Y 234, 470-474 (1986). 4. Lin, W.J. & Yeh, W.C. Implication of Toll-like receptor and tumor necrosis factor alpha signaling in septic shock. Shock (Augusta, Ga 24, 206-209 (2005). 5. Takahashi, N., Brouckaert, P. & Fiers, W. Induction of tolerance allows separation of lethal and antitumor activities of tumor necrosis factor in mice. Cancer research 51, 2366-2372 (1991). 6. Cauwels, A. et al. Protection against TNF-induced lethal shock by soluble guanylate cyclase inhibition requires functional inducible nitric oxide synthase. Immunity 13, 223-231 (2000). 7. Eggermont, A.M., de Wilt, J.H. & ten Hagen, T.L. Current uses of isolated limb perfusion in the clinic and a model system for new strategies. The lancet oncology 4, 429-437 (2003). 8. Kimura, K. et al. Phase I study of recombinant human tumor necrosis factor. Cancer chemotherapy and pharmacology 20, 223-229 (1987). 9. Creaven, P.J. et al. Phase I clinical trial of recombinant human tumor necrosis factor. Cancer chemotherapy and pharmacology 20, 137-144 (1987). 10. Lenk, H., Tanneberger, S., Muller, U., Ebert, J. & Shiga, T. Phase II clinical trial of high-dose recombinant human tumor necrosis factor. Cancer chemotherapy and pharmacology 24, 391- 392 (1989). 11. Ameloot, P. et al. Bioavailability of recombinant tumor necrosis factor determines its lethality in mice. European journal of immunology 32, 2759-2765 (2002). 12. Stoelcker, B. et al. Tumor necrosis factor induces tumor necrosis via tumor necrosis factor receptor type 1-expressing endothelial cells of the tumor vasculature. The American journal of pathology 156, 1171-1176 (2000). 13. Goethals, A., Pasparakis, M., Hostens, J., Breier, G., Kollias, G., Brouckaert, P. Expression of TNF-R1 exclusively on endothelial cells is sufficient to obtain tumor destruction by TNF. (Article submitted) (2008). 14. Shimomura, K. et al. Recombinant human tumor necrosis factor-alpha: thrombus formation is a cause of anti-tumor activity. International journal of cancer 41, 243-247 (1988). 15. Van de Wiel, P.A., Bloksma, N., Kuper, C.F., Hofhuis, F.M. & Willers, J.M. Macroscopic and microscopic early effects of tumour necrosis factor on murine Meth A sarcoma, and relation to curative activity. J Pathol 157, 65-73 (1989). 16. Eggermont, A.M. et al. Isolated limb perfusion with high-dose tumor necrosis factor-alpha in combination with interferon-gamma and melphalan for nonresectable extremity soft tissue sarcomas: a multicenter trial. J Clin Oncol 14, 2653-2665 (1996). 17. Carmeliet, P. & Jain, R.K. Angiogenesis in cancer and other diseases. Nature 407, 249-257 (2000). 18. Hashizume, H. et al. Openings between defective endothelial cells explain tumor vessel leakiness. The American journal of pathology 156, 1363-1380 (2000). 19. Baluk, P., Morikawa, S., Haskell, A., Mancuso, M. & McDonald, D.M. Abnormalities of basement membrane on blood vessels and endothelial sprouts in tumors. The American journal of pathology 163, 1801-1815 (2003). 20. Jain, R.K. Normalizing tumor vasculature with anti-angiogenic therapy: a new paradigm for combination therapy. Nature medicine 7, 987-989 (2001). 21. Gerber, H.P. et al. Vascular endothelial growth factor regulates endothelial cell survival through the phosphatidylinositol 3'-kinase/Akt signal transduction pathway. Requirement for Flk-1/KDR activation. The Journal of biological chemistry 273, 30336-30343 (1998). 22. Vanhaesebroeck, B. & Alessi, D.R. The PI3K-PDK1 connection: more than just a road to PKB. The Biochemical journal 346 Pt 3, 561-576 (2000). 23. Zhang, L., Himi, T., Morita, I. & Murota, S. Inhibition of phosphatidylinositol-3 kinase/Akt or mitogen-activated protein kinase signaling sensitizes endothelial cells to TNF-alpha cytotoxicity. Cell death and differentiation 8, 528-536 (2001). Experimental work: Sensitization to TNF by inhibition of survival signaling 213

24. Clermont, F., Adam, E., Dumont, J.E. & Robaye, B. Survival pathways regulating the apoptosis induced by tumour necrosis factor-alpha in primary cultured bovine endothelial cells. Cellular signalling 15, 539-546 (2003). 25. Madge, L.A., Li, J.H., Choi, J. & Pober, J.S. Inhibition of phosphatidylinositol 3-kinase sensitizes vascular endothelial cells to cytokine-initiated cathepsin-dependent apoptosis. The Journal of biological chemistry 278, 21295-21306 (2003). 26. Matschurat, S. et al. Negative regulatory role of PI3-kinase in TNF-induced tumor necrosis. International journal of cancer 107, 30-37 (2003). 27. Rothe, J., Mackay, F., Bluethmann, H., Zinkernagel, R. & Lesslauer, W. Phenotypic analysis of TNFR1-deficient mice and characterization of TNFR1-deficient fibroblasts in vitro. Circulatory shock 44, 51-56 (1994). 28. Okkenhaug, K. et al. Impaired B and T cell antigen receptor signaling in p110delta PI 3- kinase mutant mice. Science (New York, N.Y 297, 1031-1034 (2002). 29. Hirsch, E. et al. Central role for G protein-coupled phosphoinositide 3-kinase gamma in inflammation. Science (New York, N.Y 287, 1049-1053 (2000). 30. Hart, I.R. & Fidler, I.J. Role of organ selectivity in the determination of metastatic patterns of B16 melanoma. Cancer research 40, 2281-2287 (1980). 31. van Lamsweerde, A.L., Henry, N. & Vaes, G. Metastatic heterogeneity of cells from Lewis lung carcinoma. Cancer research 43, 5314-5320 (1983). 32. Vanhaesebroeck, B., Ali, K., Bilancio, A., Geering, B. & Foukas, L.C. Signalling by PI3K isoforms: insights from gene-targeted mice. Trends in biochemical sciences 30, 194-204 (2005). 33. Puri, K.D. et al. Mechanisms and implications of phosphoinositide 3-kinase delta in promoting neutrophil trafficking into inflamed tissue. Blood 103, 3448-3456 (2004). 34. Puri, K.D. et al. The role of endothelial PI3Kgamma activity in neutrophil trafficking. Blood 106, 150-157 (2005). 35. Kondapaka, S.B., Singh, S.S., Dasmahapatra, G.P., Sausville, E.A. & Roy, K.K. Perifosine, a novel alkylphospholipid, inhibits protein kinase B activation. Molecular cancer therapeutics 2, 1093-1103 (2003). 36. Brunn, G.J. et al. Direct inhibition of the signaling functions of the mammalian target of rapamycin by the phosphoinositide 3-kinase inhibitors, wortmannin and LY294002. The EMBO journal 15, 5256-5267 (1996). 37. Sarbassov, D.D. et al. Prolonged rapamycin treatment inhibits mTORC2 assembly and Akt/PKB. Molecular cell 22, 159-168 (2006). 38. Dormond, O., Madsen, J.C. & Briscoe, D.M. The effects of mTOR-Akt interactions on anti- apoptotic signaling in vascular endothelial cells. The Journal of biological chemistry 282, 23679-23686 (2007). 39. Fischer, C. et al. Anti-PlGF inhibits growth of VEGF(R)-inhibitor-resistant tumors without affecting healthy vessels. Cell 131, 463-475 (2007). 40. Sherman, M.L. et al. Recombinant human tumor necrosis factor administered as a five-day continuous infusion in cancer patients: phase I toxicity and effects on lipid metabolism. J Clin Oncol 6, 344-350 (1988). 41. Creaven, P.J. et al. A phase I clinical trial of recombinant human tumor necrosis factor given daily for five days. Cancer chemotherapy and pharmacology 23, 186-191 (1989). 42. Gamm, H., Lindemann, A., Mertelsmann, R. & Herrmann, F. Phase I trial of recombinant human tumour necrosis factor alpha in patients with advanced malignancy. Eur J Cancer 27, 856-863 (1991). 43. Brouckaert, P. et al. Inhibition of and sensitization to the lethal effects of tumor necrosis factor. Journal of inflammation 47, 18-26 (1995). 44. Van Molle, W. et al. Protection of zinc against tumor necrosis factor induced lethal inflammation depends on heat shock protein 70 and allows safe antitumor therapy. Cancer research 67, 7301-7307 (2007). 45. Knight, Z.A. et al. A pharmacological map of the PI3-K family defines a role for p110alpha in insulin signaling. Cell 125, 733-747 (2006). 46. Jackson, S.P. et al. PI 3-kinase p110beta: a new target for antithrombotic therapy. Nature medicine 11, 507-514 (2005). 47. Madge, L.A. & Pober, J.S. A phosphatidylinositol 3-kinase/Akt pathway, activated by tumor necrosis factor or interleukin-1, inhibits apoptosis but does not activate NFkappaB in human endothelial cells. The Journal of biological chemistry 275, 15458-15465 (2000). 214 Chapter 8

48. Gustin, J.A. et al. Cell type-specific expression of the IkappaB kinases determines the significance of phosphatidylinositol 3-kinase/Akt signaling to NF-kappa B activation. The Journal of biological chemistry 279, 1615-1620 (2004). 49. Menon, C., Ghartey, A., Canter, R., Feldman, M. & Fraker, D.L. Tumor necrosis factor-alpha damages tumor blood vessel integrity by targeting VE-cadherin. Annals of surgery 244, 781- 791 (2006). 50. Liao, F. et al. Monoclonal antibody to vascular endothelial-cadherin is a potent inhibitor of angiogenesis, tumor growth, and metastasis. Cancer research 60, 6805-6810 (2000). 51. Angelini, D.J. et al. TNF-alpha increases tyrosine phosphorylation of vascular endothelial cadherin and opens the paracellular pathway through fyn activation in human lung endothelia. Am J Physiol Lung Cell Mol Physiol 291, L1232-1245 (2006). 52. Vincent, L. et al. Combretastatin A4 phosphate induces rapid regression of tumor neovessels and growth through interference with vascular endothelial-cadherin signaling. The Journal of clinical investigation 115, 2992-3006 (2005). 53. Friedlander, M. et al. Definition of two angiogenic pathways by distinct alpha v integrins. Science (New York, N.Y 270, 1500-1502 (1995). 54. Ruegg, C. et al. Evidence for the involvement of endothelial cell integrin alphaVbeta3 in the disruption of the tumor vasculature induced by TNF and IFN-gamma. Nature medicine 4, 408- 414 (1998). 55. Bezzi, M., Hasmim, M., Bieler, G., Dormond, O. & Ruegg, C. Zoledronate sensitizes endothelial cells to tumor necrosis factor-induced programmed cell death: evidence for the suppression of sustained activation of focal adhesion kinase and protein kinase B/Akt. The Journal of biological chemistry 278, 43603-43614 (2003). 56. Bieler, G. et al. Distinctive role of integrin-mediated adhesion in TNF-induced PKB/Akt and NF-kappaB activation and endothelial cell survival. Oncogene 26, 5722-5732 (2007). 57. Pohlman, T.H. & Harlan, J.M. Human endothelial cell response to lipopolysaccharide, interleukin-1, and tumor necrosis factor is regulated by protein synthesis. Cellular immunology 119, 41-52 (1989). 58. Polunovsky, V.A., Wendt, C.H., Ingbar, D.H., Peterson, M.S. & Bitterman, P.B. Induction of endothelial cell apoptosis by TNF alpha: modulation by inhibitors of protein synthesis. Experimental cell research 214, 584-594 (1994). 59. Soldi, R. et al. Role of alphavbeta3 integrin in the activation of vascular endothelial growth factor receptor-2. The EMBO journal 18, 882-892 (1999). 60. Carmeliet, P. et al. Targeted deficiency or cytosolic truncation of the VE-cadherin gene in mice impairs VEGF-mediated endothelial survival and angiogenesis. Cell 98, 147-157 (1999).

Invariant NKT cells contribute to tumor Chapter 9 destruction by TNF

Experimental work: NKT cells contribute to tumor destruction by TNF 217

9. INVARIANT NKT CELLS CONTRIBUTE TO TUMOR DESTRUCTION BY TNF

Abstract

TNF is a pro-inflammatory cytokine with a remarkable antitumor activity. Both in animal models and in human patients it is capable of inducing a rapid, haemorrhagic necrosis of tumors, characterized by a selective destruction of the tumor vasculature. Systemic treatment with TNF is, however, severely hampered by the pronounced shock-inducing properties of TNF. We and others have previously shown that the antitumor effect of TNF is completely host-mediated, and requires expression of the TNF-R1 only on the endothelial cells of the tumor vasculature. Experiments in our research group using IRF-1-/- mice, which are deficient in mature NK cells, NKT cells and CD8+ T cells, have suggested that certain immune cells may also be important for the full antitumor response to TNF. In the present study we investigated, using immune-deficient mice and selective leukocyte depletion methods, which lymphocytes are required for the antitumor response to TNF. We found that no lymphocyte population was strictly essential for the antitumor activity of TNF. Athymic C57BL/6nu/nu mice, lacking mature T lymphocytes and NKT cells, as well as NK cell-depleted mice showed similar responses to TNF as wild-type, immune- competent mice. However, we found that four independent NKT-deficient mouse strains (IRF-1-/-, IL-15-/-, J18-/- and CD1-/-) displayed markedly reduced antitumor response to TNF. In IRF-1-/- mice as well as in J18-/- mice this defect could be largely restored by transfer of leukocytes isolated from wild-type liver, which is particularly rich in NK1.1+ T cells. A role for NK1.1+ cells was also confirmed by selective depletion of these cells. Our results suggest that the recruitment and/or activation of invariant NKT cells is an important event in the cascade that leads to complete tumor destruction by TNF. Furthermore, we have extended this finding by showing that a single injection of the specific NKT cell ligand GalCer markedly increased the efficacy of a sub-therapeutic TNF treatment. Using IFN-R-deficient mice we showed that the latter effect is dependent on IFN- production. Taken together, we have shown that targeting invariant NKT cells may be a worthwhile approach to extend the therapeutic applicability of TNF.

This work was performed in collaboration with the research group of Prof. Dr. Dirk Elewaut (Department of Rheumatology, Ghent University Hospital, Ghent, Belgium) who provided transgenic mice, reagents and intellectual input. It builds on observations made by Dr. Jeroen Hostens in our research group and reported in his PhD thesis. 218 Chapter 9

9.1. Introduction

TNF is a pro-inflammatory cytokine with a remarkable antitumor activity. Both in animal models and in human patients, where it is used in combination with melphalan to treat regional tumors of the limbs via the ILP technique, it is capable of inducing a rapid haemorrhagic necrosis of tumors, which is characterized by a selective destruction of the tumor vasculature. We and others have previously shown that this antitumor effect is completely host-mediated, and requires expression of the TNF-R1 on the tumor endothelium 1, 2. These experiments have also indicated that TNF-R1 expression on other cells than endothelial cells, most importantly tumor cells and immune cells, is unimportant for the full antitumor response to TNF. Certain immune cells may, however, be important for the antitumor effect of TNF. Eggermont and colleagues demonstrated that total body irradiation of tumor-bearing rats, thereby inducing general leukopenia, markedly decreased the tumor response and skin necrosis at the tumor site following ILP with TNF and melphalan 3. After TNF-based ILP, an inflammatory response is observed which results in a rapid recruitment of polymorphonuclear cells to the tumor, followed by the infiltration of macrophages and lymphocytes after several days to weeks 4, 5. Experiments performed in our lab using B16Bl6 tumor-bearing IRF-1-/- mice, which fail to develop mature NK cells, NK1.1+ T cells and CD8+ T cells, revealed a markedly reduced antitumor response to TNF in these mice 6-8. Responsiveness could be partly restored by reconstitution with wild-type splenocytes, suggesting a role for lymphocytes in this model.

In the present study we have investigated, by using immune-deficient mouse strains and selective lymphocyte depletion methods, which lymphocytes are involved in the antitumor mechanism induced by TNF. We found that no lymphocyte population was indispensible for the antitumor activity of TNF. However, we found that NKT- deficient mice displayed a markedly reduced antitumor response to TNF. Furthermore, we have shown that pre-treatment with the specific NKT cell ligand GalCer markedly increased the efficacy of a sub-therapeutic TNF treatment. Taken together, our results suggest that targeting invariant NKT cells may be a worthwhile approach to extend the therapeutic applicability of TNF.

9.2. Materials and methods

9.2.1. Mice

Female C57BL/6J mice and athymic C57BL/6nu/nu mice were purchased from Janvier, Le Genest-Saint-Isle, France. IRF-1-/- (B6.129S2-Irf1tm1Mak/J, stock nr. 002762), IFN- R-/- (B6.129S7-Ifngr1tm1Agt/J, stock nr. 003288), and athymic C57BL/6nu/+ (B6.Cg- Foxn1nu/J, stock nr. 000819) breeding couples were purchased from The Jackson Experimental work: NKT cells contribute to tumor destruction by TNF 219

Laboratory, Bar Harbor, Maine, USA 7, 9. Athymic Swissnu/nu mice (ICO:Swiss- Foxn1nu) were purchased from Charles River Laboratories, Saint-Germain-sur- l’Arbresle, France. IL-15-/- mice on C57BL/6 background were generated by Dr. Peschon, Immunex Corporation, Seattle, Washington, USA 10. CD1-/- mice were generated and backcrossed to C57BL/6 background by Dr. Van Kaer, Howard Hughes Medical Institute, Vanderbilt University School of Medicine, Nashville, Tennessee, USA 11. J18-/- mice were generated and backcrossed to C57BL/6 background by Dr. Taniguchi, Chiba University, Chiba, Japan 12. Transgenic mice expressing a dominant-negative IFN-R mutant controlled by a Flk-1 promoter (Flk-1 dnIFN-R) were generated and backcrossed to C57BL/6 background in our laboratory by Dr. Jeroen Hostens 13. All animals were housed in 14-10 hours light/dark cycles in a temperature-controlled, air-conditioned room, and received food and water ad libitum. The mice were used for experimentation at the age of 8-12 weeks. All experiments were approved by and performed according to the guidelines of the local ethics committee of Ghent University, Belgium.

9.2.2. Cells

B16Bl6 melanoma and EL4 lymphoma cells were cultured in RPMI-1640 medium supplemented with 10% FCS, 2mM L-glutamine, 0,4mM Na-pyruvate, 50 IU/ml penicillin G and 50µg/ml streptomycin sulphate 14. LLC-H61 carcinoma cells were cultured in DMEM medium supplemented with 10% FCS, 2mM L-glutamine, 0,4mM Na-pyruvate, 50 IU/ml penicillin G and 50µg/ml streptomycin sulphate 15. For tumor inoculation, cells were detached from the culture flask by a short EDTA treatment, washed three times in endotoxin-free, sterile DPBS (Sigma), and resuspended in DPBS.

9.2.3. Reagents

The NKT cell ligand GalCer was produced by Dr. Serge Van Calenbergh, Laboratory for Medicinal Chemistry, Faculty of Pharmaceutical Sciences, Ghent University, Ghent, Belgium. Rat anti-mouse IL-2R -chain antibodies, clone Tm1, were produced in our laboratory. The Tm1 hybridoma was generated and kindly provided by Dr. Tanaka, Tokyo Metropolitan Institute of Medical Science, Tokyo, Japan 16. Mouse anti-mouse NK1.1 antibodies, clone PK136, were purchased from BioXCell. The fluorochrome-labeled antibodies anti-NK1.1-FITC, anti-CD49b-FITC (clone DX5), and anti-TCR-PE were purchased from BD Pharmingen. FITC-labeled mouse CD1d tetramers loaded with GalCer were provided by Dr. Dirk Elewaut, Department of Rheumatology, Ghent University Hospital, Ghent, Belgium. Recombinant mTNF and hTNF were produced in our laboratory and had a specific 220 Chapter 9

activity of 1.28 x 108 IU/mg and 6.8 x 107, and an endotoxin content (LAL assay) of 6.5 EU/mg and 36 EU/mg, respectively.

9.2.4. Tumor experiments

Mice were inoculated with 6 x 105 (B16Bl6 and EL4) or 5 x 106 (LLC) tumor cells s.c. in the back (day 0). Treatment was started when the TSI, the product of the largest perpendicular diameters in cm, was greater than 0,5 (≥day 10). Unless otherwise indicated, mTNF (7µg) was administered daily for 10 consecutive days via paralesional injection (s.c. injection near the site of the tumor but outside the nodule). For GalCer experiments, GalCer (1µg) was injected i.p. on the first day of treatment, followed by daily treatment with low-dose mTNF (0.7µg) or hTNF (30µg) via paralesional injection for 9 consecutive days.

9.2.5. Transfer of splenocytes and liver leukocytes

Livers or spleens of wild-type C57BL/6 mice were passed through a 70µm cell strainer in complete DMEM medium. Red blood cells were lysed using Ack buffer

(0.15M NH4Cl, 10mM KHCO3, 0.1mM Na2EDTA.2H2O, pH 7.2). Hepatocytes were removed by density gradient centrifugation using a 40%-70% Percoll gradient. Cells were washed twice in PBS and resuspended in PBS. For splenocyte transfer, IRF-1-/- mice were injected i.v. into the tail vein with 50 x 106 splenocytes on day 5 after tumor inoculation. For transfer of liver leukocytes, IRF-1-/- or J18-/- mice were injected i.v. into the tail vein with approx. 2 x 106 cells (1:2 donor/acceptor ratio) on day 7 after tumor inoculation. Approx. 1% of splenocytes and 10-16% of liver leukocytes were NK1.1+TCR+ cells.

9.2.6. In vivo depletion of NK and NK1.1+ cells

Selective, long-term depletion of NK cells was obtained by a single i.p. injection of 1mg of anti-IL-2R antibody (Tm1) on day 5 after tumor inoculation as described 17. Depletion was verified by a YAC-1 lysis assay and by flow cytometric analysis of leukocytes isolated from liver and spleen. For the YAC-1 lysis assay, splenocytes were isolated 6 days after injection of the anti-IL-2R antibody and cultured overnight in complete RPMI medium supplemented with 1000 IU/ml human IL-2. The cells were seeded into 96-well plates containing 20000 YAC-1 target cells per well with effector/target ratios of 100:1 to 0.2:1. The cells were co-incubated during 6 hours, after which lysis of YAC-1 cells was determined by measuring lactate dehydrogenase in the medium with the Cytotox96 non-radioactive cytotoxicity assay kit (Promega). For flow cytometric analysis, leukocytes from liver and spleen were isolated 6 days after injection of the anti-IL-2R antibody and stained with pairs of Experimental work: NKT cells contribute to tumor destruction by TNF 221

the following fluorochrome-labeled reagents to detect NK and NKT cells: anti- NK1.1-FITC, anti-CD49b-FITC, GalCer CD1d tetramer-FITC, and anti-TCR-PE. Flow cytometry was performed on a FACSCalibur and results were analyzed with CellQuest software (BD Biosciences). Depletion of NK1.1+ cells, i.e. NK and NKT cells, was performed by i.p. injection of 200µg anti-NK1.1 antibody on days 8, 12, 16, and 20 after tumor inoculation. Depletion was verified by flow cytometric analysis of liver leukocytes isolated 3 days after injection of the anti-NK1.1 antibody.

9.2.7. Statistics

For tumor experiments mice were stratified into treatment groups having non- significantly different mean TSI (one-way ANOVA, P-value > 0.95). Calculations and graphs were made using Graphpad Prism 4.0.

9.3. Results

9.3.1. Lymphocytes influence the response to TNF

Previous work in our lab demonstrated that IRF-1-/- mice, which fail to develop mature NK cells, NKT cells and CD8+ T cells, displayed a markedly reduced antitumor response to TNF. Responsiveness could be partly restored when these mice were injected with wild-type splenocytes 48 hours prior to inoculation with the B16Bl6 tumor cells 8. However, this effect was temporary, most likely due to washing out of the splenocytes before the end of the ten-day treatment period, and tumors which initially responded to treatment with mTNF already relapsed during the treatment period, resulting in no significantly different tumor response rate at the end of treatment. Moreover, this experiment could not exclude that the reduced response of IRF-1-/- mice to treatment with mTNF would be due to absence of the IRF-1 transcription factor. IRF-1 is known to interfere with TNF/NF-B signaling 18, 19. To determine whether immunological defects are at the basis of the reduced antitumor activity of TNF in IRF-1-/- mice, we compared the antitumor response to TNF in B16Bl6 tumor-bearing IRF-1-/- mice and IL-15-/- mice, which have comparable immune deficiencies (Fig.9.1.A-C and Table 9.1.). IL-15-/- mice essentially lack mature NK cells and have markedly reduced numbers of NKT and memory phenotype CD8+ T cells 10. Treatment of wild-type C57BL/6 mice with mTNF resulted in complete regression of the B16Bl6 tumor (no palpable tumor at the 10th day of treatment) in six out of seven treated mice. In both IRF-1-/- and IL-15-/- mice, however, the tumor response was markedly reduced, resulting in complete tumor regression in no and one out of seven treated mice, respectively. This suggests that the reduced antitumor response to TNF in these strains is due to the absence or reduced numbers of either NK, NKT or CD8+ T cells. 222 Chapter 9

Fig. 9.1. Lymphocytes are important for the antitumor response to TNF. (A-D) Tumor growth curves of B16Bl6 melanoma in C57BL/6 wild-type (wt) mice, IRF-1-/- mice, IL-15-/- mice, and IRF-1-/- mice reconstituted with wild-type splenocytes. Data are shown as mean ± SEM. (E) Overlay of tumor growth curves after treatment with mTNF. Data are shown as mean + SEM. For C57BL/6 wt and IL- 15-/- n = 7, for IRF-1-/- n = 6. The line under the graph represents the treatment period.

To further substantiate this we reconstituted IRF-1-/- with wild-type splenocytes (Fig.9.1. and table 9.1.). Disappearance of the splenocytes before the end of the treatment period was avoided by transferring the splenocytes 5 days after tumor inoculation. We found that reconstitution of IRF-1-/- mice with wild-type splenocytes markedly increased the sensitivity to TNF, resulting in complete tumor regression in Experimental work: NKT cells contribute to tumor destruction by TNF 223

all surviving mice. However, three out of six treated mice died due to treatment- related toxicity, suggesting that both the TNF-induced antitumor effect and toxic side- effects are influenced by the presence of immune cells. Alternatively, the increased sensitivity to TNF may also be a consequence of the supranormal numbers of certain leukocytes after splenocyte transfer (see further), or the presence of contaminants (e.g. LPS) in the transferred cell preparation which could lead to inflammation.

Table 9.1. Responses to treatment with mTNF in wild-type, IRF-1-/- and IL-15-/- mice. Tumor response § Strain CR (%) PR (%) S (%) PD (%) D (%) C57BL/6 wt (n = 7) 86 14 – – – IRF-1-/- (n = 6) – 33 17 50 – IL-15-/- (n = 7) 14 43 – 43 – IRF-1-/- + wt splenocytes (n = 6) 50 – – – 50 § TSI before start of treatment was compared to that on the 10th day of treatment. Tumor response was classified as follows: PD = progressive disease, increase of more than 25%; S = static, tumor size between –25% and 25%; PR = partial regression, decrease between –25% and –99%; CR = complete regression, no palpable tumor. D = animal died during treatment. – = none.

The similar results obtained in both IRF-1-/- and IL-15-/- strains suggest a role for either NK cells, NKT cells or CD8+ T cells in the antitumor mechanism induced by TNF. To determine whether T cells are involved in this effect, we treated tumor- bearing athymic C57BL6nu/nu mice and Swissnu/nu mice, which lack mature T lymphocytes including NKT cells (Fig.9.2.). A remarkable difference was observed between both strains, which appeared to be independent of the tumor type. Similar results were obtained for tumors induced by s.c. injection of human HT29 colon carcinoma cells or murine B16Bl6 melanoma cells. The antitumor response to TNF in Swissnu/nu mice was comparable to that in IRF-1-/- or IL-15-/- mice, which may reflect the importance of (NK)T cells for the antitumor activity of TNF. Complete tumor regression or lethality was not observed in Swissnu/nu mice. By sharp contrast, treatment of B16Bl6 tumor-bearing C57BL/6nu/nu mice resulted in apparent tumor regression in all treated mice. However, five out of six treated animals succumbed due to toxic side-effects, suggesting that C57BL/6nu/nu mice are more sensitive to TNF per se. 224 Chapter 9

Fig. 9.2. The antitumor activity of TNF in athymic nude mice. (A-B) Growth curve of B16Bl6 melanoma in athymic C57BL/6nu/nu and Swissnu/nu mice. Data are shown as mean ± SEM, n = 6. The line under the graph represents the treatment period.

Analysis of spleen, liver, and blood leukocytes revealed that C57BL/6nu/nu have two- to fourfold increased numbers of NK cells compared to wild-type C57BL/6 mice (Fig.9.3.). To investigate whether the increased number of NK cells could determine the increased sensitivity to TNF, we depleted NK cells in these mice by injecting an anti-IL-2R antibody which was previously shown to induce selective, long-term elimination of NK cells 17. NK cell depletion of C57BL/6nu/nu mice resulted in a markedly reduced sensitivity to TNF, with complete regression of the B16Bl6 tumor in only one out of six treated animals, and only one of six animals died during treatment. These results indicate that the presence of NK cells may indeed influence the response to TNF in this model.

Fig. 9.3. NK cells determine the sensitivity to TNF in C57BL/6nu/nu mice. (A) Growth curve of B16Bl6 tumor in NK cell-depleted C57BL/6nu/nu mice. Data are shown as mean ± SEM, n = 6. The line under the graph represents the treatment period. (B) Flow cytometric analysis of splenocytes and liver leukocytes of C57BL/6 wild-type (wt), C57BL/6nu/nu (nu/nu), and NK cell-depleted C57BL/6nu/nu mice isolated 6 days after injection of the anti-IL-2R antibody.

Experimental work: NKT cells contribute to tumor destruction by TNF 225

To investigate the role of NK cells and NK1.1+ T cells in the antitumor activity of TNF in wild-type mice, we depleted these cells using either an anti-IL-2R antibody or an anti-NK1.1 antibody. Analysis of liver leukocytes showed that injection of the anti-IL-R2 antibody selectively depleted NK cells, not affecting hepatic invariant NKT cell numbers, while injection of the anti-NK1.1 antibody depleted both NK and NK1.1+ T cells, thus allowing us to make a distinction between both cell types (Fig.9.4.). Treatment of NK cell-depleted mice, after injection of the anti-IL-2R antibody, resulted in regression of the B16Bl6 tumor as in control mice, indicating that NK cells are not required for the antitumor activity of TNF. However, depletion of NK1.1+ cells resulted in a markedly reduced antitumor response to TNF, with complete tumor regression in only one out of six treated animals. Taken together, these results suggest a role for NK1.1+ T cells in the antitumor mechanism induced by TNF.

Fig. 9.4. NK1.1+ T cells, but not NK cells, contribute to the antitumor activity of TNF. (A, C) Growth curve of B16Bl6 tumor in NK cell-depleted (anti-IL-2R; n = 7) mice and NK1.1+ cell- depleted (anti-NK1.1; n = 6) mice. Data are shown as mean ± SEM. The line under the graph represents the treatment period. (B, D) Flow cytometric analysis of liver leukocytes isolated respectively 6 and 3 days after injection of anti-IL-2R or anti-NK1.1 antibodies. Tetramer: GalCer- loaded mouse CD1d tetramer.

226 Chapter 9

9.3.2. Invariant NKT cells contribute to the TNF-induced antitumor effect

To further establish the involvement of NKT cells in the antitumor mechanism elicited by TNF, we used the NKT-deficient mouse strains J18-/- and CD1-/-. Both these strains lack classical invariant NKT cells (type I 20) and have no other overt defects than in NKT cells 11, 12. Treatment of these NKT-deficient strains resulted in both cases in a markedly reduced response to TNF compared to wild-type C57BL/6 mice (Fig.9.5. and Table 9.2). For B16Bl6 melanoma, the complete response rate (complete regression) was 62% in wild-type C57BL/6 mice, while this was 19% and 10% in J18-/- and CD1-/- mice, respectively. The objective response rates (complete + partial regression) were respectively 73%, 29% and 34%. The difference was even further enlarged when we used lower doses of mTNF. Treatment of wild-type C57BL/6 mice with 3.5µg or 1µg of mTNF resulted in both cases in complete tumor regression in four of six treated animals, while no complete tumor regressions were observed in J18-/- mice treated with these doses. Furthermore, the reduced antitumor response in J18-/- appeared to be independent of the tumor type as similar results were obtained for EL4 lymphoma, although this tumor grew markedly slower in J18- /- mice than in control mice.

Experimental work: NKT cells contribute to tumor destruction by TNF 227

Fig. 9.5. NKT cells contribute to the antitumor activity of TNF. (A-B) Growth curve of B16Bl6 melanoma in C57B/6 wild-type (wt) and J18-/- mice, n = 6. (C) Growth curve of EL4 lymphoma in C57BL/6 wild-type and J18-/- mice, n = 7. (D) Growth curve of B16Bl6 melanoma in CD1-/- mice, n = 7. All data are shown as mean ± SEM. The line under the graph represents the treatment period.

To determine whether the responsiveness to TNF treatment in NKT-deficient mice could be restored by reconstitution with invariant NKT cells, we transferred leukocytes isolated from wild-type liver, which is particularly rich is invariant NK1.1+ T cells, to J18-/- and IRF-1-/- mice (Fig.9.6. and table 9.2.). We determined that approx. 10 to 16% of all liver leukocytes were NK1.1+ T cells. Strikingly, transfer of wild-type liver leukocytes to J18-/- mice increased the objective response rate from 29% to 100%. Transfer to IRF-1-/- mice increased the objective response rate from 17% to 50%. Interestingly, no increased toxicity was observed after transfer of wild- type liver leukocytes. Taken together, these results indicate that invariant NKT cells are involved in the antitumor mechanism elicited by TNF.

228 Chapter 9

Fig. 9.6. Reconstitution with wild-type liver leukocytes restores the antitumor activity of TNF in NKT-deficient mice. (A) Tumor response of B16Bl6 melanoma to treatment with mTNF in C57BL/6 wild-type (wt) mice, J18-/- mice, and J18-/- mice reconstituted with wild-type liver leukocytes. (B) Idem for IRF-1-/- mice. All data are shown as mean + SEM, n = 6. The line under the graph represents the treatment period.

Table 9.2. Responses to treatment with mTNF in wild-type and NKT-deficient mice. Tumor response § Strain CR (%) PR (%) S (%) PD (%) D (%) C57BL/6 wt (n =53) 62 11 4 15 8 J18-/- (n = 31) 19 10 10 52 10 CD1-/- (n = 21) 10 24 29 38 – IRF-1-/- (n = 12) – 17 25 50 8 J18-/- + wt liver leukocytes (n = 6) 83 17 – – – IRF-1-/- + wt liver leukocytes (n = 6) 33 17 17 17 17 § TSI before start of treatment was compared to that on the 10th day of treatment. Tumor response was classified as follows: PD = progressive disease, increase of more than 25%; S = static, tumor size between –25% and 25%; PR = partial regression, decrease between –25% and –99%; CR = complete regression, no palpable tumor. D = animal died during treatment. – = none.

9.3.3. GalCer sensitizes to the antitumor effect of TNF

The finding that invariant NKT cells may contribute to tumor destruction upon TNF treatment prompted us to test whether selective activation of these cells with the marine sponge glycolipid GalCer could increase the efficacy of a low-toxic, sub- therapeutic TNF treatment. B16Bl6 melanoma-bearing mice were injected with GalCer 24 hours before the start of a nine-day treatment with either hTNF or low- dose mTNF (Fig.9.7. and Table 9.3.). Injection of GalCer alone had no effect on tumor growth. Treatment with either hTNF or low-dose mTNF induced respectively a static effect or a marginal growth inhibiting effect on the B16Bl6 tumor. Complete tumor regression was not observed. However, when mice were pre-treated with GalCer, treatment with hTNF or low-dose mTNF resulted in pronounced tumor destruction, with complete tumor regression in three of six and two of seven treated Experimental work: NKT cells contribute to tumor destruction by TNF 229

animals, respectively. Similar results were obtained for LLC tumors, suggesting that this synergistic antitumor effect is not restricted to B16Bl6. Combination therapy was well tolerated, and mildly increased systemic toxicity (weight loss) was only observed after combination with hTNF.

Fig. 9.7. Pre-treatment with GalCer increases the efficacy of a sub-therapeutic TNF treatment. (A) Growth curves of B16Bl6 tumor after treatment with GalCer and/or hTNF, n = 6. (B) Idem for low-dose mTNF (0.7µg), n = 7. (C) Growth curves of LLC tumor after treatment with GalCer and/or low-dose mTNF, n = 6. Data are shown as mean ± SEM. The arrow indicates the injection point of GalCer. The line under the graph represent the treatment period with TNF.

230 Chapter 9

Table 9.3. B16Bl6 tumor response to treatments with GalCer and TNF. Tumor response § Treatment CR (%) PR (%) S (%) PD (%) D (%) PBS (n = 13) – – – 92 8 GalCer (n = 13) – – – 85 15 hTNF (n = 6) – 17 33 50 – GalCer + hTNF (n = 6) 50 33 – – 17 Low-dose mTNF (n = 7) – – – 86 14 GalCer + mTNF (n = 7) 29 29 14 14 14 § TSI before start of treatment was compared to that on the 10th day of treatment. Tumor response was classified as follows: PD = progressive disease, increase of more than 25%; S = static, tumor size between –25% and 25%; PR = partial regression, decrease between –25% and –99%; CR = complete regression, no palpable tumor. D = animal died during treatment. – = none.

GalCer-activated NKT cells are capable of rapidly producing large amounts of cytokines such as IFN- and IL-4 21. In particular IFN- may be important in our model because (1) this cytokine was previously shown to be critical for the antitumor activity of GalCer and (2) the synergy of TNF and IFN- in murine tumor models is well-established 22, 23. To investigate whether the synergistic antitumor activity of GalCer and TNF is dependent on endogenous IFN- production, we treated IFN-R-/- mice with GalCer and hTNF, and transgenic Flk-1 dnIFN-R mice, which express a dominant-negative IFN-R mutant controlled by a Flk-1 promoter, thereby rendering endothelial cells insensitive to IFN-, with GalCer and low-dose mTNF (Fig.9.8.). In both strains, pre-treatment with GalCer did not increase the response to TNF treatment, indicating that the synergy of GalCer and TNF is indeed IFN-- dependent. The absence of synergy in transgenic Flk-1 dnIFN-R mice suggests that the endogenous IFN- produced upon injection of GalCer acts by activating the endothelial cells of the tumor vasculature, as was shown previously for exogenously administered IFN- 13.

Experimental work: NKT cells contribute to tumor destruction by TNF 231

Fig. 9.8. The synergistic antitumor effect of GalCer and TNF is IFN--dependent. (A) Growth curve of B16Bl6 melanoma in IFN-R-/- mice after treatment with hTNF and GalCer, n = 5. (B) Growth curve of B16Bl6 melanoma in transgenic Flk-1 dnIFN-R mice after treatment with GalCer and low-dose mTNF (0.7µg), n = 6. Data are shown as mean ± SEM. The arrow indicates the injection point of GalCer. The line under the graph represent the treatment period with TNF.

9.4. Discussion

The primary findings of this study are: (1) tumor regression upon TNF treatment was observed in all immune-deficient and lymphocyte-depleted mice tested, suggesting that the antitumor activity of TNF is not strictly dependent on the presence of one particular lymphocyte population; however, (2) four independent NKT cell-deficient mouse strains (IRF-1-/-, IL-15-/-, J18-/- and CD1-/-) as well as NK1.1+ cell-depleted mice displayed similar, reduced antitumor response to TNF; (3) response to TNF in J18-/- and IRF-1-/- mice could be restored by reconstitution with wild-type liver leukocytes; and (4) pre-treatment with the specific NKT cell ligand GalCer markedly increased the efficacy of a sub-therapeutic TNF treatment.

As TNF is a potent pro-inflammatory cytokine, it is not surprising that TNF-based therapies were shown to result in an infiltration of immune cells in tumors 4, 5. However, it remained to be elucidated whether these immune cells actively contribute to tumor destruction after TNF therapy. Eggermont and colleagues demonstrated that the tumor response to TNF-based ILP was markedly reduced in total body irradiated rats, suggesting that immune cells play an important role in the antitumor activity of TNF 3. Our experiments in IRF-1-/- and IL-15-/- mice, which both lack mature NK cells and have strongly reduced numbers of NKT and CD8+ T cells, support this hypothesis, and point to a role for cytotoxic lymphocytes. However, tumor response to treatment with mTNF was similar in athymic C57BL/6nu/nu mice and NK cell-depleted mice as in immune-competent C57BL/6 wild-type mice, indicating that neither (NK)T nor NK cells are essential for the antitumor activity of TNF. One plausible 232 Chapter 9

explanation for this discrepancy is that there might be sufficient functional redundancy between these different cytotoxic lymphocyte populations. NK cells, NKT cells and CD8+ T cells share many important effector functions, including IFN- production, FasL expression, and perforin/granzyme B-mediated cytotoxicity. The absence of one lymphocyte population may affect the functionality or proportion of another. For example, in athymic C57BL/6nu/nu mice we observed a two- to four-fold increased number of NK cells in spleen and liver, and we found that this could be linked to an overall increased sensitivity to TNF. Interestingly, an increased sensitivity to TNF was also observed after transfer of wild-type splenocytes to IRF-1-/- mice. It would be interesting to investigate whether this effect was due to the supranormal number of NK cells in these mice after transfer. Taken together, our data indicate that the presence of certain cytotoxic lymphocyte populations may influence the sensitivity to both the antitumor effect and toxic side-effects of TNF.

We demonstrated that the antitumor activity of TNF was markedly reduced in NKT- deficient mouse strains as well as in NK1.1+ cell-depleted mice, suggesting that invariant NKT cells significantly contribute to the antitumor effect of TNF. However, even in NKT-deficient mice complete tumor regressions were observed after treatment with mTNF, albeit much less frequently than in wild-type mice, indicating that the antitumor activity is not strictly dependent on NKT cells but is facilitated by the presence of these cells. Invariant NKT cells have the unique ability to rapidly produce large amounts of immunoregulatory cytokines upon activation, most notably of IFN- 21, 24. The synergy of TNF and IFN- is well-known in our B16Bl6 tumor model 23. Previous experiments in our lab showed that the antitumor activity of TNF was unaffected in IFN-R-/- mice. However, treatment of IFN--/- mice with TNF resulted in the same initial tumor destruction as observed in wild-type mice, with complete regression of the B16Bl6 tumor, but this was followed by an early onset of relapse already during the ten-day treatment period. These results suggest that endogenous IFN- production, acting directly on the tumor cells, may contribute to the TNF-induced antitumor mechanism 8. We may now have identified invariant NKT cells as the source of this endogenous IFN-. However, endogenous IFN- production alone cannot explain the reduced response observed in NKT-deficient mice and therefore it is likely that other NKT cell-derived cytokines, potentially TNF, or cytotoxic effector functions are involved in this effect. As the antitumor activity of TNF requires only endothelial TNF-R1 expression, it is puzzling how NKT cells could be specifically recruited to and activated at the tumor site. 2. It was shown that murine NKT cells exhibit robust response to the inflammatory chemokines CXCL9 and CXCL10, suggesting that these chemokines could act to recruit NKT cells to sites of inflammation 25. However, we found that the antitumor activity of TNF was unaffected in CXCR3-deficient mice (data not shown). Thus, we can exclude a role for those chemokines in the antitumor effect of TNF. Experimental work: NKT cells contribute to tumor destruction by TNF 233

The elucidation of the complete antitumor mechanism elicited by TNF is important as this may reveal new targets for therapeutic intervention. The finding that NKT cells may actively contribute to tumor destruction after high-dose TNF treatment prompted us to test whether specific activation of these cells could increase the efficacy of a low-toxic, sub-therapeutic TNF treatment. We demonstrated that pre-treatment with GalCer markedly increased tumor destruction induced by treatment with hTNF or low-dose mTNF. Using IFN-R-/- and transgenic Flk-1 dnIFN-R mice we could show that this effect was completely IFN--dependent and required functional IFN-R expression on endothelial cells, suggesting that GalCer acts by a similar mechanism as exogenously administered IFN- 8. An important question is of course whether using GalCer would be superior to treatment with IFN-. We started treatment with TNF 24 hours after i.p. injection of GalCer, when elevated serum levels of IFN- were already largely declined 26. Increased response to TNF was observed for at least several days after injection of GalCer, suggesting that GalCer may trigger a longer- lasting sensitizing effect. It has been demonstrated that the antimetastatic effect of GalCer was dependent on the activation of NK cells and IFN- production by these cells, which is then most likely followed by bystander activation of other immune cells 22, 27. Although no overt toxicities were observed in our model, caution is required when combining TNF with GalCer as both these drugs are known to cause liver toxicity, which is one of the major dose-limiting toxicities of TNF 28, 29. More potent GalCer analogs are currently being developed with a Th1-biased cytokine profile 30. It will be most interesting to test whether these compounds would even further increase the antitumor activity of TNF. In conclusion, we have identified invariant NKT cells as a new, promising target to extend the therapeutic applicability of TNF. 234 Chapter 9

9.5. References

1. Stoelcker, B. et al. Tumor necrosis factor induces tumor necrosis via tumor necrosis factor receptor type 1-expressing endothelial cells of the tumor vasculature. The American journal of pathology 156, 1171-1176 (2000). 2. Goethals, A., Pasparakis, M., Hostens, J., Breier, G., Kollias, G., Brouckaert, P. Expression of TNF-R1 exclusively on endothelial cells is sufficient to obtain tumor destruction by TNF. (Article submitted) (2008). 3. Manusama, E.R. et al. Assessment of the role of neutrophils on the antitumor effect of TNFalpha in an in vivo isolated limb perfusion model in sarcoma-bearing brown Norway rats. The Journal of surgical research 78, 169-175 (1998). 4. Renard, N. et al. Early endothelium activation and polymorphonuclear cell invasion precede specific necrosis of human melanoma and sarcoma treated by intravascular high-dose tumour necrosis factor alpha (rTNF alpha). International journal of cancer 57, 656-663 (1994). 5. Lejeune, F.J. & Ruegg, C. Recombinant human tumor necrosis factor: an efficient agent for cancer treatment. Bulletin du cancer 93, E90-100 (2006). 6. Ohteki, T. et al. The transcription factor interferon regulatory factor 1 (IRF-1) is important during the maturation of natural killer 1.1+ T cell receptor-alpha/beta+ (NK1+ T) cells, natural killer cells, and intestinal intraepithelial T cells. The Journal of experimental medicine 187, 967-972 (1998). 7. Matsuyama, T. et al. Targeted disruption of IRF-1 or IRF-2 results in abnormal type I IFN gene induction and aberrant lymphocyte development. Cell 75, 83-97 (1993). 8. Hostens, J., Brouckaert, P. (unpublished data). (2003). 9. Huang, S. et al. Immune response in mice that lack the interferon-gamma receptor. Science (New York, N.Y 259, 1742-1745 (1993). 10. Kennedy, M.K. et al. Reversible defects in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice. The Journal of experimental medicine 191, 771-780 (2000). 11. Mendiratta, S.K. et al. CD1d1 mutant mice are deficient in natural T cells that promptly produce IL-4. Immunity 6, 469-477 (1997). 12. Cui, J. et al. Requirement for Valpha14 NKT cells in IL-12-mediated rejection of tumors. Science (New York, N.Y 278, 1623-1626 (1997). 13. Hostens, J., Goethals, A., Breier, G., Schreiber, R.D., Brouckaert, P. IFN- synergistically enhances the anti-tumour effect of Tumour Necrosis Factor by activating the endothelial cells of the neovasculature. (article in preparation). 14. Hart, I.R. & Fidler, I.J. Role of organ selectivity in the determination of metastatic patterns of B16 melanoma. Cancer research 40, 2281-2287 (1980). 15. van Lamsweerde, A.L., Henry, N. & Vaes, G. Metastatic heterogeneity of cells from Lewis lung carcinoma. Cancer research 43, 5314-5320 (1983). 16. Tanaka, T. et al. A novel monoclonal antibody against murine IL-2 receptor beta-chain. Characterization of receptor expression in normal lymphoid cells and EL-4 cells. J Immunol 147, 2222-2228 (1991). 17. Tanaka, T. et al. Selective long-term elimination of natural killer cells in vivo by an anti- interleukin 2 receptor beta chain monoclonal antibody in mice. The Journal of experimental medicine 178, 1103-1107 (1993). 18. Drew, P.D. et al. NF kappa B and interferon regulatory factor 1 physically interact and synergistically induce major histocompatibility class I gene expression. J Interferon Cytokine Res 15, 1037-1045 (1995). 19. Suk, K. et al. Interferon gamma (IFNgamma ) and tumor necrosis factor alpha synergism in ME-180 cervical cancer cell apoptosis and necrosis. IFNgamma inhibits cytoprotective NF- kappa B through STAT1/IRF-1 pathways. The Journal of biological chemistry 276, 13153- 13159 (2001). 20. Godfrey, D.I., MacDonald, H.R., Kronenberg, M., Smyth, M.J. & Van Kaer, L. NKT cells: what's in a name? Nat Rev Immunol 4, 231-237 (2004). 21. Kawano, T. et al. CD1d-restricted and TCR-mediated activation of valpha14 NKT cells by glycosylceramides. Science (New York, N.Y 278, 1626-1629 (1997). 22. Hayakawa, Y. et al. Critical contribution of IFN-gamma and NK cells, but not perforin- mediated cytotoxicity, to anti-metastatic effect of alpha-galactosylceramide. European journal of immunology 31, 1720-1727 (2001). Experimental work: NKT cells contribute to tumor destruction by TNF 235

23. Brouckaert, P.G., Leroux-Roels, G.G., Guisez, Y., Tavernier, J. & Fiers, W. In vivo anti- tumour activity of recombinant human and murine TNF, alone and in combination with murine IFN-gamma, on a syngeneic murine melanoma. International journal of cancer 38, 763-769 (1986). 24. Godfrey, D.I. & Kronenberg, M. Going both ways: immune regulation via CD1d-dependent NKT cells. The Journal of clinical investigation 114, 1379-1388 (2004). 25. Johnston, B., Kim, C.H., Soler, D., Emoto, M. & Butcher, E.C. Differential chemokine responses and homing patterns of murine TCR alpha beta NKT cell subsets. J Immunol 171, 2960-2969 (2003). 26. Elewaut, D. (personal communication). 27. Smyth, M.J. et al. Sequential production of interferon-gamma by NK1.1(+) T cells and natural killer cells is essential for the antimetastatic effect of alpha-galactosylceramide. Blood 99, 1259-1266 (2002). 28. Jones, A.L. & Selby, P. Tumour necrosis factor: clinical relevance. Cancer surveys 8, 817-836 (1989). 29. Biburger, M. & Tiegs, G. Alpha-galactosylceramide-induced liver injury in mice is mediated by TNF-alpha but independent of Kupffer cells. J Immunol 175, 1540-1550 (2005). 30. Chang, Y.J. et al. Potent immune-modulating and anticancer effects of NKT cell stimulatory glycolipids. Proceedings of the National Academy of Sciences of the United States of America 104, 10299-10304 (2007).

The synergistic antitumor effect of TNF and Chapter 10 IFN- is dependent on neutrophils

Experimental work: The synergistic antitumor effect of TNF and IFN- 239

10. THE SYNERGISTIC ANTITUMOR EFFECT OF TNF AND IFN- IS DEPENDENT ON NEUTROPHILS

Abstract

TNF, especially in combination with IFN-, is widely appreciated as being a promising anticancer therapy. The synergism of TNF and IFN- for inducing haemorrhagic necrosis of tumors in experimental models has led to the successful addition of IFN- to ILP with TNF and melphalan for the treatment of regionally advanced melanoma or sarcoma of the limbs. The mechanism underlying the in vivo synergistic antitumor effects of TNF and IFN-, however, remains poorly understood. In the current study we investigated whether immune cells, which appear to infiltrate tumors after TNF/IFN--based ILP, are involved in the synergistic antitumor effect of TNF and IFN-. We have previously shown that invariant NKT cells may contribute to tumor destruction upon treatment with TNF. Our present data indicate that the synergism of TNF and IFN- is independent of NKT cells or other lymphocytes. By contrast we found that the synergistic antitumor effect of TNF and IFN- was markedly reduced in PI3K-deficient mice, which display defective chemokine- induced responses of neutrophils and monocytes/macrophages. Specific depletion of macrophages did not affect the synergistic antitumor effect of TNF and IFN- in our murine B16Bl6 melanoma model, while depletion of neutrophils completely abolished this synergism. Our data imply that the synergistic antitumor effect of TNF and IFN- is critically dependent on neutrophils and may have important implications for the addition of IFN- to TNF-based anticancer treatments.

This work was performed in collaboration with Dr. Emilio Hirsch (University of Torino, Turin, Italy) 240 Chapter 10

10.1. Introduction

TNF, a potent pro-inflammatory cytokine, is currently used in the clinic in combination with melphalan and mild hyperthermia to treat regionally advanced melanomas and soft tissue sarcomas of the limbs via the ILP technique. The high limb salvage rates obtained after TNF-based ILPs clearly demonstrate its potency as an anticancer drug. Co-treatment with IFN- has been shown to synergistically enhance the antitumor properties of TNF in numerous experimental tumor models 1-3. Also, although less striking than in animal models, clinical studies using ILP with the triple combination of TNF, IFN- and melphalan appear to result in consistently higher response rates than those using ILP with TNF and melphalan alone, suggesting that the synergistic effect of TNF and IFN- is also present in this system 4. In vitro, IFN- has been shown to sensitize for cytokine-induced cytotoxicity in an IRF-1-dependent manner 5, 6. However, we found that the synergistic antitumor effect of TNF and IFN-  was apparent in IRF-1-deficient mice, suggesting that this effect is independent of IRF-1 7. Histological analysis of tumors revealed a selective destruction of the tumor- associated vasculature after treatment with TNF and IFN-, and previous work in our laboratory showed that the synergistic antitumor effect of TNF and IFN- is dependent on endothelial IFN-R expression, indicating that the endothelial cells of the tumor vasculature are the target cells of TNF/IFN- combination therapy 4, 8-10. Recently, Li and colleagues demonstrated that IFN- can trigger the release of cathepsin B from lysosomes in HUVECs, thereby sensitizing these cells to TNF- induced cathepsin B-dependent cell death 11. The mechanism underlying the in vivo synergistic antitumor effects of TNF and IFN- still remains to be elucidated. Our hypothesis is that IFN- may increase endothelial injury and tumor destruction by exacerbating the TNF-triggered inflammatory response. TNF and IFN- have been shown to synergistically upregulate the endothelial expression of cell adhesion molecules, such as ICAM-1 and VCAM-1, and monocyte/lymphocyte chemokines, such as CCL2, CCL5 and CXCL10 12-14. We have evidence that invariant NKT cells are implicated in the antitumor mechanism induced by TNF. However, previous work in our laboratory indicated that the synergism of TNF and IFN- was preserved in either IRF-1-deficient mice, which are largely devoid of mature NK cells, NKT cells and CD8+ T cells, athymic nude mice, and NK cell-depleted mice, suggesting that neither NK cells nor (NK)T cells are important for the synergistic antitumor effect of TNF and IFN-.

Herein we report that the synergistic antitumor effect of TNF and IFN- is markedly reduced in PI3K-deficient mice, which display defective chemotaxis of neutrophils and monocytes. Specific depletion of monocytes/macrophages did not affect the synergistic antitumor effect of TNF and IFN- in our murine B16Bl6 melanoma model, while depletion of neutrophils completely abolished this synergism. Our data Experimental work: The synergistic antitumor effect of TNF and IFN- 241

imply that the synergistic antitumor effect of TNF and IFN- is critically dependent on neutrophils. This finding may have important implications for the addition of IFN-  to TNF-based anticancer treatments.

10.2. Materials and methods

10.2.1. Mice

Female C57BL/6J mice were purchased from Janvier, Le Genest-Saint-Isle, France. CD1-/- mice were generated and backcrossed to C57BL/6 background by Dr. Luc Van Kaer, Vanderbilt University School of Medicine, Nashville, Tennessee, USA 15. B6.p110-/- (PI3K-/-) mice were generated by Dr. Emilio Hirsch, University of Torino, Turin, Italy 16. The animals were housed in 14-10 hours light/dark cycles in a temperature-controlled, air-conditioned room, and received food and water ad libitum. The mice were used for experimentation at the age of 8-12 weeks. All experiments were approved by and performed according to the guidelines of the local ethics committee of Ghent University, Belgium.

10.2.2. Cells

B16Bl6 melanoma cells were cultured in RPMI-1640 medium supplemented with 10% FCS, 2mM L-glutamine, 0,4mM Na-pyruvate, 50 IU/ml penicillin G and 50µg/ml streptomycin sulphate 17. For tumor inoculation, cells were detached from the culture flask by a short EDTA treatment, washed three times in endotoxin-free, sterile DPBS (Sigma), and resuspended in DPBS.

10.2.3. Reagents

Mouse anti-mouse NK1.1 antibodies (clone PK136) and rat anti-mouse Gr1 antibodies (clone RB6-8C5) were purchased from BioXCell. Cyclophosphamide was purchased from Sigma. Liposome-encapsulated clodronate was a kind gift of Dr. Nico Van Rooijen, Free University Medical Center, Amsterdam, The Netherlands 18. Recombinant mTNF, hTNF and mIFN- were produced in our laboratory and had a specific activity of 1.28 x 108 IU/mg, 6.8 x 107 IU/mg and 1.16 x 108 IU/mg, respectively. mTNF and hTNF had an endotoxin content (LAL assay) of 6.5 EU/mg and 36 EU/mg, respectively.

10.2.4. Tumor experiments

Mice were inoculated with 6 x 105 B16Bl6 cells s.c. in the back (day 0). Treatment was started when the tumor size index (TSI), the product of the largest perpendicular 242 Chapter 10

diameters in cm, was greater than 0.5 (≥day 10). PBS, hTNF (30µg), mTNF (7µg) and mIFN- (5000 IU) were administered daily for 10 consecutive days via s.c. injection near the tumor site but outside the nodule.

10.2.5. In vivo depletion of neutrophils, macrophages or NK1.1+ cells

Mice were depleted of neutrophils by i.p. injection of cyclophosphamide or anti-Gr1 antibody as previously described 19, 20. Mice were injected with 3mg cyclophosphamide on day 9, and 2mg cyclophosphamide on days 12, 15 and 18 after tumor inoculation. For depletion using the anti-Gr1 antibody, mice received 0.5mg antibody on days 7, 9, 11, 13, 15 and 19 after tumor inoculation. Depletion of neutrophils was verified by differential leukocyte counting of May-Grünwald Giemsa- stained blood smears taken of control (PBS-treated) mice on days 10, 13 and 19 after tumor inoculation. Monocytes/macrophages were depleted by injection of liposome- encapsulated clodronate as previously described 21. Mice were injected i.p. with 200µl of the clodronate liposome solution on day 8, and 100µl on days 13 and 18 after tumor inoculation. Macrophage depletion was confirmed by flow cytometric analysis of splenocytes stained for CD11b and F4/80, prepared 3 days after injection of clodronate liposomes. Depletion of NK1.1+ cells, i.e. NK and NKT cells, was achieved by i.p. injection of 200µg anti-NK1.1 antibody on days 8, 12, 16, and 20 after tumor inoculation. Depletion was verified by flow cytometric analysis of liver leukocytes prepared 3 days after injection of the anti-NK1.1 antibody and stained for NK1.1, CD49b (DX5) and TCR,.

10.2.6. Statistics

For tumor experiments mice were stratified into treatment groups having non- significantly different mean TSI (one-way ANOVA, P-value > 0.95). Calculations and graphs were made using Graphpad Prism 4.0.

Experimental work: The synergistic antitumor effect of TNF and IFN- 243

10.3. Results

10.3.1. Lymphocytes are not required for the synergistic antitumor effect of TNF and IFN-

We previously found that the synergistic antitumor effect of TNF and IFN- was still apparent in IRF-1-/-, C57BL/6nu/nu, Swissnu/nu, and NK cell-depleted mice, suggesting that NK cells and (NK)T cells are not required for the synergism of TNF and IFN- 7. The tumor response to treatment with hTNF and IFN- was, however, markedly reduced in IRF-1-/- mice compared to wild-type C57BL/6 mice. We previously attributed the reduced response to mTNF treatment in IRF-1-/- and other NKT- deficient mice to the absence of invariant NK1.1+ T cells (see chapters 9.3.1. and 9.3.2.). To further assess the involvement of NKT cells in the synergistic antitumor effect of TNF and IFN-, we treated NKT-deficient CD1-/- mice and NK1.1-depleted mice with a combination of IFN- and mTNF or hTNF, respectively (Fig.10.1.). The synergistic antitumor effect of TNF and IFN- was apparent in both cases, confirming that this synergistic effect is independent of NKT cells. However, the tumor response to treatment with hTNF and IFN- was apparently reduced in NK1.1-depleted mice compared to control mice, albeit statistically not significantly. Therefore, at this moment we cannot exclude that invariant NKT cells may contribute to tumor destruction upon co-treatment with TNF and IFN-. Interestingly, treatment of CD1-/- mice with mTNF and IFN- resulted in complete eradication of the tumor in all seven treated mice, without apparent signs of toxicity (judged by the general condition, body weight and viability of the mice), while six of seven wild-type control mice died due to treatment-related toxicity, suggesting that NKT cells may also contribute to the toxic side-effects of treatment with TNF and IFN-. Taken together, our present data suggest that neither NKT cells nor other lymphocytes are required for the synergistic antitumor effect of TNF and IFN-. 244 Chapter 10

Fig. 10.1. The synergistic antitumor effect of TNF and IFN- is independent of NKT cells. Growth curve of s.c. B16Bl6 melanoma in (A) C57BL/6 wild-type (wt; for PBS and hTNF n = 6, for other groups n = 7), (B) NK1.1-depleted (n = 6) and (C) CD1-/- mice (n = 7) treated with hTNF and mTNF alone or in combination with IFN-. All data are shown as mean ± SEM. The line under the graph represents the ten-day treatment period.

10.3.2. Neutrophils are required for the synergistic antitumor effect of TNF and IFN-

When we were investigating whether the absence of certain PI3K isoform would sensitize mice to the antitumor effect of TNF, we unexpectedly found that in PI3K-/- mice the synergism of hTNF and IFN- was largely impaired (Fig.10.2.). Progressive disease was observed in 60% of PI3K-/- mice treated with hTNF and IFN- compared to 82% of mice treated with hTNF alone. In wild-type control mice progressive disease rates were 10% and 70%, respectively. Complete regression of the B16Bl6 tumor upon treatment with hTNF and IFN- occurred in more than 50% of wild-type control mice, while this was rarely observed in PI3K-/- mice. These data suggest that a PI3K-dependent signaling event is required for the synergistic antitumor effect of TNF and IFN-. Experimental work: The synergistic antitumor effect of TNF and IFN- 245

Fig. 10.2. The synergism of hTNF and IFN- is abrogated in PI3K-/- mice. (A) Growth curve of B16Bl6 melanoma in PI3K-/- mice treated with hTNF alone or combination with IFN-, or mTNF. Data are shown as mean ± SEM, n = 6. The line under the graph represents the treatment period. (B, C) Tumor response to treatment with hTNF alone or in combination with IFN- in C57BL/6 wild-type (wt) and PI3K-/- mice. TSI before start of treatment was compared to that on the 10th day of treatment. Tumor response was classified as follows: PD = progressive disease, increase of more than 25%; S = static, tumor size between –25% and 25%; PR = partial regression, decrease between –25% and –99%; CR = complete regression, no palpable tumor. D = animal died during treatment.

PI3K-/- mice have a number immunological defects, most prominently impaired chemokine-induced recruitment of neutrophils and monocytes/macrophages, and oxidative burst of neutrophils 16, 22. To investigate whether these defects could be at the basis of the impaired synergistic effect of hTNF and IFN- observed in PI3K-/- mice, we depleted wild-type mice of neutrophils or macrophages before treatment with hTNF and IFN- (Fig.10.3.). Neutrophils were first depleted by i.p. injection of the cytotoxic drug cyclophosphamide. The mice were effectively neutropenic by the start of the ten-day treatment period (day 13 after tumor inoculation), however, general leukopenia was induced by repeated injections of cyclophosphamide. The synergistic antitumor effect of hTNF and IFN- was largely abrogated by treatment with cyclophosphamide, suggesting that leukocytes, most likely neutrophils, are 246 Chapter 10

required for the synergism of hTNF and IFN-. To further assess the involvement of neutrophils, we specifically depleted these cells by treatment with an anti-Gr1 antibody. The depletion was specific and significant (P < 0.01 in unpaired t-test) on day 10 after tumor inoculation (treatment with hTNF and IFN- was started on day 11). However, the depletion was not statistically significant on day 19 after tumor inoculation due to large interindividual variability in tumor-bearing mice. Nevertheless, in Gr1-depeleted mice the synergistic antitumor effect of hTNF and IFN- was markedly reduced compared to control mice. Complete tumor regressions were not observed in neutrophil-depleted mice.

Fig. 10.3. The synergistic antitumor effect of hTNF and IFN- is dependent on neutrophils. Growth curves of s.c. B16Bl6 melanoma in (A) C57BL/6 control mice, (B, C) neutrophil-depleted mice, and (D) monocyte/macrophage-depleted mice treated with hTNF alone or in combination with IFN-. All data are shown as mean ± SEM, n = 6. The line under the graph represents the treatment period.

Daley and co-workers recently reported that treatment with the anti-Gr1 antibody, clone RB6-8C5, may not only deplete neutrophils, but also Gr1+ monocytes/macrophages 23. To determine whether monocytes/macrophages are involved in the synergistic antitumor effect of hTNF and IFN-, we specifically depleted these cells by injection of liposome-encapsulated clodronate. Macrophage Experimental work: The synergistic antitumor effect of TNF and IFN- 247

depletion, however, did not affect the response to treatment with hTNF and IFN-, which resulted in complete tumor regression in four of six treated animals. Taken together, these data indicate that neutrophils, but not monocytes/macrophages, are critically involved in the synergistic antitumor effect of TNF and IFN-.

10.4. Discussion

TNF, especially in combination with IFN-, is widely appreciated as being a promising anticancer therapy. Clinically, ILP with TNF and melphalan, with or without IFN- has proven to be a very successful alternative for amputation in patients with advanced melanoma or sarcoma of the limbs 4. Pronounced toxic side- effects, particularly severe hypotension and organ toxicity, however, preclude systemic use of TNF in cancer patients 24-26. One strategy to allow systemic TNF application may be to develop low-toxic TNF muteins. Studies in mice using hTNF, which can be considered as a fast-cleared, low-toxic TNF mutein in the murine system, indicated that these low-toxic TNF muteins will most likely require combination with other drugs, such as IFN-, to be effective 1, 27. The mechanism underlying the remarkable synergistic antitumor effect of TNF and IFN-, however, remains poorly understood. We have now demonstrated that this effect is critically dependent on neutrophils.

Extensive in vitro data exist showing that co-incubation with IFN- sensitizes endothelial cells to TNF-induced cell death 11. Also, Rüegg and co-workers reported that treatment of HUVECs with TNF and IFN- suppressed the activation of integrin

v3, leading to decreased endothelial cell adhesion and survival. They furthermore observed detachment and apoptosis of v3-positive endothelial cells of the tumor vasculature in melanoma patients treated with TNF and IFN- 10. The finding that in our B16Bl6 melanoma model the synergistic antitumor effect of TNF and IFN- required the presence of neutrophils implies that IFN- did not significantly sensitize the tumor endothelium to the direct vascular-disrupting effects of TNF in this in vivo model. We hypothesize that IFN- rather increases endothelial injury by exacerbating the TNF-triggered inflammatory response. Neutrophils are recruited to sites of inflammation by expression on the activated endothelium of the adhesion molecules E-selectin and ICAM-1, both of which are synergistically upregulated by TNF and IFN- 13, 28, 29. The adhesion of neutrophils to activated endothelium may be sufficient to trigger cytotoxic effector functions, including respiratory burst and degranulation of granules containing proteolytic enzymes and myeloperoxidase 30. We can, however, exclude NO production as an important effector function as we previously found that the synergistic antitumor effect of TNF and IFN- was preserved in iNOS-deficient mice 31. 248 Chapter 10

The finding that the synergistic antitumor effect of TNF and IFN- was largely abrogated in PI3K-/- mice, suggests that a GPCR-mediated signaling event is important in this effect 16. Neutrophils can be activated by a large number of chemotactic factors that signal via GPCRs, including the Glu-Leu-Arg (ELR)-positive

chemokines CXCL1-8 (except CXCL4), and the lipid mediators PAF and LTB4, and it will be challenging to determine which factor(s) is/are involved 30. Another interesting observation made in PI3K-/- mice was that the antitumor effect of mTNF was preserved in these mice, suggesting that the antitumor mechanism of mTNF monotherapy is essentially different from that induced by hTNF and IFN- combination therapy. Similarly, the antitumor activity of mTNF was unaffected in cyclophosphamide-treated mice, indicating that the antitumor activity of mTNF is independent of neutrophils (data not shown). Taken together, our data suggest that IFN- does not sensitize for the intrinsic antivascular or antitumor effect of TNF, but allows to trigger an additional, neutrophil-dependent mechanism. This finding offers a new rationale for the addition of IFN- to TNF-based anticancer therapies.

Experimental work: The synergistic antitumor effect of TNF and IFN- 249

10.5. References

1. Brouckaert, P.G., Leroux-Roels, G.G., Guisez, Y., Tavernier, J. & Fiers, W. In vivo anti- tumour activity of recombinant human and murine TNF, alone and in combination with murine IFN-gamma, on a syngeneic murine melanoma. International journal of cancer 38, 763-769 (1986). 2. Balkwill, F.R., Ward, B.G. & Fiers, W. Effects of tumour necrosis factor on human tumour xenografts in nude mice. Ciba Foundation symposium 131, 154-169 (1987). 3. Rubin, S.C. et al. Synergistic activity of tumor necrosis factor and interferon in a nude mouse model of human ovarian cancer. Gynecologic oncology 34, 353-356 (1989). 4. Lejeune, F.J., Lienard, D., Matter, M. & Ruegg, C. Efficiency of recombinant human TNF in human cancer therapy. Cancer Immun 6, 6 (2006). 5. Fulda, S. & Debatin, K.M. IFNgamma sensitizes for apoptosis by upregulating caspase-8 expression through the Stat1 pathway. Oncogene 21, 2295-2308 (2002). 6. Suk, K. et al. Interferon gamma (IFNgamma ) and tumor necrosis factor alpha synergism in ME-180 cervical cancer cell apoptosis and necrosis. IFNgamma inhibits cytoprotective NF- kappa B through STAT1/IRF-1 pathways. The Journal of biological chemistry 276, 13153- 13159 (2001). 7. Hostens, J., Brouckaert, P. (unpublished data). (2003). 8. de Kossodo, S. et al. Changes in endogenous cytokines, adhesion molecules and platelets during cytokine-induced tumour necrosis. British journal of cancer 72, 1165-1172 (1995). 9. Hostens, J., Goethals, A., Breier, G., Schreiber, R.D., Brouckaert, P. IFN- synergistically enhances the anti-tumour effect of Tumour Necrosis Factor by activating the endothelial cells of the neovasculature. (article in preparation). 10. Ruegg, C. et al. Evidence for the involvement of endothelial cell integrin alphaVbeta3 in the disruption of the tumor vasculature induced by TNF and IFN-gamma. Nature medicine 4, 408- 414 (1998). 11. Li, J.H. & Pober, J.S. The cathepsin B death pathway contributes to TNF plus IFN-gamma- mediated human endothelial injury. J Immunol 175, 1858-1866 (2005). 12. Marfaing-Koka, A. et al. Regulation of the production of the RANTES chemokine by endothelial cells. Synergistic induction by IFN-gamma plus TNF-alpha and inhibition by IL-4 and IL-13. J Immunol 154, 1870-1878 (1995). 13. Paleolog, E.M., Aluri, G.R. & Feldmann, M. Contrasting effects of interferon gamma and interleukin 4 on responses of human vascular endothelial cells to tumour necrosis factor alpha. Cytokine 4, 470-478 (1992). 14. Hillyer, P., Mordelet, E., Flynn, G. & Male, D. Chemokines, chemokine receptors and adhesion molecules on different human endothelia: discriminating the tissue-specific functions that affect leucocyte migration. Clinical and experimental immunology 134, 431-441 (2003). 15. Mendiratta, S.K. et al. CD1d1 mutant mice are deficient in natural T cells that promptly produce IL-4. Immunity 6, 469-477 (1997). 16. Hirsch, E. et al. Central role for G protein-coupled phosphoinositide 3-kinase gamma in inflammation. Science (New York, N.Y 287, 1049-1053 (2000). 17. Hart, I.R. & Fidler, I.J. Role of organ selectivity in the determination of metastatic patterns of B16 melanoma. Cancer research 40, 2281-2287 (1980). 18. Van Rooijen, N. & Sanders, A. Liposome mediated depletion of macrophages: mechanism of action, preparation of liposomes and applications. Journal of immunological methods 174, 83- 93 (1994). 19. Ismail, H.F., Fick, P., Zhang, J., Lynch, R.G. & Berg, D.J. Depletion of neutrophils in IL-10(- /-) mice delays clearance of gastric Helicobacter infection and decreases the Th1 immune response to Helicobacter. J Immunol 170, 3782-3789 (2003). 20. Zuluaga, A.F. et al. Neutropenia induced in outbred mice by a simplified low-dose cyclophosphamide regimen: characterization and applicability to diverse experimental models of infectious diseases. BMC infectious diseases 6, 55 (2006). 21. van Rooijen, N. & van Kesteren-Hendrikx, E. "In vivo" depletion of macrophages by liposome-mediated "suicide". Methods in enzymology 373, 3-16 (2003). 22. Sasaki, T. et al. Function of PI3Kgamma in thymocyte development, T cell activation, and neutrophil migration. Science (New York, N.Y 287, 1040-1046 (2000). 250 Chapter 10

23. Daley, J.M., Thomay, A.A., Connolly, M.D., Reichner, J.S. & Albina, J.E. Use of Ly6G- specific monoclonal antibody to deplete neutrophils in mice. Journal of leukocyte biology 83, 64-70 (2008). 24. Tracey, K.J. et al. Shock and tissue injury induced by recombinant human cachectin. Science (New York, N.Y 234, 470-474 (1986). 25. Creaven, P.J. et al. Phase I clinical trial of recombinant human tumor necrosis factor. Cancer chemotherapy and pharmacology 20, 137-144 (1987). 26. Kimura, K. et al. Phase I study of recombinant human tumor necrosis factor. Cancer chemotherapy and pharmacology 20, 223-229 (1987). 27. Ameloot, P. et al. Bioavailability of recombinant tumor necrosis factor determines its lethality in mice. European journal of immunology 32, 2759-2765 (2002). 28. Doukas, J. & Pober, J.S. IFN-gamma enhances endothelial activation induced by tumor necrosis factor but not IL-1. J Immunol 145, 1727-1733 (1990). 29. Strindhall, J., Lundblad, A. & Pahlsson, P. Interferon-gamma enhancement of E-selectin expression on endothelial cells is inhibited by monensin. Scandinavian journal of immunology 46, 338-343 (1997). 30. Zarbock, A. & Ley, K. Mechanisms and consequences of neutrophil interaction with the endothelium. The American journal of pathology 172, 1-7 (2008). 31. Goethals, A., Brouckaert, P. (unpublished data). (2007).

Chapter 11 General summary & discussion

General summary & discussion 253

11. GENERAL SUMMARY & DISCUSSION

11.1. Summary

TNF is a potent, pro-inflammatory cytokine that was originally discovered thanks to its remarkable antitumor activity. Via its two cell surface receptors, TNF-R1 and TNF-R2, it is involved in a wide range of (patho)physiological processes. As most information regarding TNF signaling, including the activation of the NF-B, MAPK, and cell death pathways, is derived from TNF-R1, this receptor is regarded as the primary signaling receptor. However, since TNF-R2 only fully responds to the membrane-bound form of TNF, but not the soluble form, its role may be underestimated. The major biological function of TNF appears to be that of a central mediator of inflammation. Tissue macrophages stimulated with LPS or other bacterial products rapidly produce TNF to activate the local endothelium. This results in increased blood flow due to the production of the vasodilators NO and PGI2, increased vascular permeability and vascular leakage due to weakening of endothelial cell-cell junctions and cytoskeletal reorganizations, and the recruitment of neutrophils, monocytes and other leukocytes to the site of inflammation through upregulation of chemokines and adhesion molecules, such as E-selectin, ICAM-1 and VCAM-1. Moreover, as an additional protective mechanism, TNF may shut off local blood supply by stimulating blood coagulation, or by causing direct or indirect endothelial cell death. The same mechanisms by which TNF so efficiently contains local infection may, however, become harmful, as is the case in sepsis or chronic inflammatory disorders such as rheumatoid arthritis, inflammatory bowel disease and psoriasis.

The in vivo antitumor activity of recombinant TNF and its synergy with IFN- were first demonstrated in the murine B16Bl6 melanoma model by Brouckaert and colleagues in 1986 1. TNF induces a rapid, haemorrhagic necrosis of experimental tumors, which is characterized by a selective destruction of the tumor vasculature. Despite promising results in animal models the first clinical trials with systemically administered TNF were disappointing. Doses of 10 to 50 times lower than the estimated effective dose already caused severe toxic side-effects, most importantly hypotension and organ failure. To circumvent these side-effects, Lejeune and co- workers set up a locoregional treatment, in which TNF in combination with melphalan and sometimes IFN- is administered via the ILP technique to treat patients with locally advanced melanomas and soft tissue sarcomas of the limbs 2. The high response rates and similarly successful limb salvage rates obtained with this method clearly demonstrate the potential of TNF as an anticancer drug. However, as locoregional therapy has no effect on distant metastases, TNF-based ILP had thus far little impact on the final survival of patients, stressing the need for an effective and safe systemic treatment.

254 Chapter 11

Previous work in our and other laboratories has established that the anticancer and toxic side-effects of TNF are not inevitably linked. Mice can be protected against the TNF-induced lethal shock with the sGC inhibitor methylene blue without impairing the antitumor effect 3. Also, two acute phase proteins, 1-acid-glycoprotein and 1- antitrypsin, were found to confer protection against TNF-induced liver toxicity 4. More recently, the addition of zinc to drinking water of mice was shown to protect against TNF-induced lethal shock and bowel toxicity without affecting the effectiveness of TNF/IFN- therapy 5. These results clearly demonstrate that it should be feasible to set up a safe systemic treatment with TNF. However, using potentially lethal doses of TNF together with an inhibitor of the toxic side-effects is not an attractive solution for human cancer therapy as failure of the inhibitor could lead to death of the patient. A safer strategy to lower the systemic toxicity of TNF would be to use lower doses of TNF or low-toxic TNF muteins in combination with sensitizers for the antitumor activity of TNF, such as IFN-. In light of the latter strategy, we addressed three major questions in this thesis:

(1) Can we sensitize for the antitumor effect of TNF by inhibiting endothelial survival signaling ? (2) Which immune cells are involved in the antitumor mechanism induced by TNF ? (3) How does IFN- sensitize for the antitumor effect of TNF ?

 Sensitization to the antitumor effect of TNF by inhibition of endothelial survival signaling

The precise mechanism underlying the TNF-induced tumor necrosis remains poorly understood. It has become clear that the antitumor effect of TNF is completely host- mediated and independent of TNF-receptor expression on tumor cells. Stoelcker and colleagues demonstrated that the TNF-induced antitumor effect is maintained in TNF- R2-/- mice, but absent in TNF-R1-/- mice, suggesting that the antitumor effect of TNF is mediated by triggering the TNF-R1 on host cells 6. Previous work in our research group has extended these data by showing that TNF-R1 expression on neovasculature (controlled by a Flk-1 promoter) is sufficient for the antitumor effect of TNF, indicating that the endothelial cells of the tumor vasculature are the direct target cells of TNF 7. This is further supported by the observation, in experimental tumor models as well as in human ILP patients, that the TNF-induced tumor necrosis is characterized by a selective destruction of the tumor vasculature 2, 8-10. TNF may therefore be classified as a VDA.

The PI3K/Akt pathway is one of the main regulators of cell survival. Tumor vasculature, which is in contrast to normal vasculature not stabilized by normal General summary & discussion 255

pericyte coverage, basement membrane structure and endothelial cell-cell interactions, and is subject to constant remodelling due to imbalance of angiogenic factors produced in the tumor, may be greatly dependent on this survival pathway 11-14. The PI3K/Akt pathway is triggered by a broad range of pro-angiogenic factors, including growth factors, such as VEGF family members, and adhesion molecules, such as 15, 16 integrin v3 . Inhibition of PI3K has been shown to sensitize endothelial cells in vitro to the cytotoxic effects of TNF 17-19. Furthermore, Matschurat and colleagues reported a synergy of the PI3K inhibitor wortmannin and TNF for the induction of tumor necrosis in a murine MethA sarcoma model 20.

To determine whether inhibition of the PI3K/Akt survival pathway would sensitize to the antitumor effect of TNF, we treated B16Bl6 or other tumor-bearing mice with sub-therapeutic TNF alone or in combination with one of the PI3K inhibitors wortmannin or LY294002. The response to either PI3K inhibitor plus TNF was striking, with tumor regression in all treated animals. We determined that by PI3K inhibition a tenfold reduction of the effective dose of TNF could be achieved. Comparable results were obtained using the Akt inhibitor perifosine or the mTOR inhibitor rapamycin. Importantly, PI3K/Akt pathway inhibition sensitized to the antitumor effect of TNF without concomitantly increasing the toxic side-effects of TNF. Using intravital microscopy and by histological examination of tumors we could demonstrate that the antitumor effect of wortmannin and TNF is characterized by a pronounced destruction of the tumor vasculature, with disruption of the endothelial lining of the tumor vessels and haemorrhage. In vivo Hoechst staining revealed a complete shutdown of the tumor blood flow as early as 6 hours after injection of wortmannin and TNF, with no recovery by 48 hours after injection, thus effectively starving the tumor to death. Finally, we demonstrated that TNF therapy synergizes with anti-angiogenic therapy using a VEGF-R2 blocking antibody for the induction of tumor regression in the B16Bl6 melanoma model. Taken together, we have shown that interfering with endothelial survival allows a significant reduction of the effective dose of TNF.

 Immune cells involved in the antitumor mechanism induced by TNF

We and others have previously shown that the antitumor effect of TNF is completely host-mediated, and requires expression of the TNF-R1 on the tumor endothelium 6, 7. These experiments have also indicated that TNF-R1 expression on other cells than endothelial cells, most importantly tumor cells and immune cells, is unimportant for the full antitumor response to TNF. Certain immune cells may, however, be important for the antitumor effect of TNF. Eggermont and colleagues demonstrated that total body irradiation of tumor-bearing rats, thereby inducing general leukopenia, markedly decreased the tumor response and skin necrosis at the tumor site following ILP with 256 Chapter 11

TNF and melphalan 21. After TNF-based ILP, an inflammatory response is observed which results in a rapid recruitment of polymorphonuclear cells to the tumor, followed by the infiltration of macrophages and lymphocytes after several days to weeks 22, 23. Experiments performed in our lab with B16Bl6 tumor-bearing IRF-1-/- mice, which fail to develop mature NK cells, NK1.1+ T cells and CD8+ T cells, revealed a markedly reduced antitumor response to TNF in these mice 24-26. Responsiveness could be partly restored by reconstitution with wild-type splenocytes, suggesting a role for lymphocytes in this model.

We investigated, using immune-deficient mice and selective lymphocyte depletion methods, which lymphocytes are involved in the antitumor mechanism induced by TNF. We found that no lymphocyte population was strictly required for the antitumor activity of TNF. Athymic C57BL/6nu/nu mice, lacking mature T lymphocytes and NKT cells, as well as NK cell-depleted mice showed similar responses to TNF as wild-type, immune-competent mice. However, we found that four independent NKT-deficient mouse strains (IRF-1-/-, IL-15-/-, J18-/- and CD1-/-) displayed markedly reduced antitumor response to TNF. In IRF-1-/- mice as well as in J18-/- mice we could largely restore this defect by transfer of leukocytes isolated from wild-type liver, which is particularly rich in NK1.1+ T cells. A role for NK1.1+ cells was also confirmed by selective depletion of NK1.1+ cells. Our results suggest that the recruitment and/or activation of invariant NKT cells is an important event in the cascade that leads to complete tumor destruction by TNF. Furthermore, we have extended this finding by showing that a single injection of the specific NKT cell ligand GalCer markedly increased the efficacy of a sub-therapeutic TNF treatment. Using IFN-R-deficient mice we showed that the latter effect is dependent on endogenous IFN- production. Taken together, we have shown that targeting invariant NKT cells may be a worthwhile approach to extend the therapeutic applicability of TNF.

 The mechanism underlying the synergistic antitumor effect of TNF and IFN-

Co-treatment with IFN- has been shown to synergistically enhance the antitumor activity of TNF in numerous experimental tumor models 1, 27, 28. Also, although less striking than in animal models, clinical studies using ILP with the triple combination of TNF, IFN- and melphalan appear to result in consistently higher response rates than those without IFN-, suggesting that the synergistic effect of TNF and IFN- is also present in this system 2. In vitro, IFN- has been shown to sensitize for cytokine- induced cytotoxicity in an IRF-1-dependent manner 29, 30. However, we found that the synergistic antitumor effect of TNF and IFN- was still apparent in IRF-1-deficient mice, suggesting that in vivo this effect is independent of IRF-1 26. Histological General summary & discussion 257

analysis of tumors revealed a selective destruction of the tumor-associated vasculature after treatment with TNF and IFN-, and previous work in our laboratory showed that the synergistic antitumor effect of TNF and IFN- is dependent on endothelial IFN-R expression, indicating that the endothelial cells of the tumor vasculature are the target cells of TNF/IFN- combination therapy 2, 31-33. Recently, Li and colleagues demonstrated that IFN- can trigger the release of cathepsin B from lysosomes in HUVECs, thereby sensitizing these cells to TNF-induced cathepsin B-dependent cell death 34. The mechanism underlying the in vivo synergistic antitumor effects of TNF and IFN-, however, still remains to be elucidated.

We investigated whether immune cells, which appear to infiltrate tumors after TNF/IFN--based ILP, are involved in the synergistic antitumor effect of TNF and IFN-. Our data indicate that the synergism of TNF and IFN- is independent of NKT cells or other lymphocytes, which were previously shown to contribute to tumor destruction upon treatment with high-dose TNF alone. By contrast we found that the synergistic antitumor effect of TNF and IFN- was markedly reduced in PI3K- deficient mice, which display defective chemokine-induced responses of neutrophils and monocytes/macrophages. Specific depletion of macrophages did not affect the synergistic antitumor effect of TNF and IFN- in our murine B16Bl6 melanoma model, while depletion of neutrophils completely abolished this synergism. Our data imply that the synergistic antitumor effect of TNF and IFN- is critically dependent on neutrophils. Therefore, we hypothesize that IFN- increases endothelial injury and tumor destruction by exacerbating the TNF-triggered inflammatory response.

11.2. Discussion

 The vascular-disrupting mechanism of TNF

TNF-based therapies are characterized by a selective destruction of the tumor vasculature. This is not only observed in experimental animal models, but also in human patients treated via ILP 2, 9. The mechanism underlying this remarkable vascular-disrupting effect, however, still remains to be elucidated. In the present study we have shown that broad-spectrum PI3K inhibition sensitizes for this vascular- disrupting effect of TNF. We used intravital microscopy as well as several histological staining methods to visualize the gross effects of treatment with wortmannin and TNF on the tumor vasculature. Our data suggest that very rapidly after injection treatment leads to vascular leakage and haemorrhage, followed by complete shutdown of the blood flow in most tumor vessels by 6 hours after injection. Destruction of the vessel wall was complete by 24 hours after injection. Matschurat and colleagues have shown that inhibition of PI3K leads to enhanced TNF-induced 258 Chapter 11

expression of tissue factor, the prime initiator of blood coagulation, in HUVECs. They furthermore demonstrated that co-injection of wortmannin promoted TNF-induced necrosis of methA tumors, thereby suggesting that the expression of tissue factor on endothelial cells of the tumor vasculature, leading to thrombus formation and occlusion of tumor vessels, would be responsible for the antitumor effect of this combination therapy 20. Our findings, however, do not fully support this hypothesis. Intravascular thrombus formation is a protective mechanism to prevent haemorrhage in the case of vascular damage. The extensive vascular leakage and haemorrhage observed in dorsal skinfold window chambers and sections of tumors treated with wortmannin and TNF suggest that vessel occlusion is secondary to disruption of the vessel wall.

There is substantial in vitro data showing that PI3K inhibition can sensitize endothelial cells to TNF-induced cytotoxicity 18, 19, 35. Particularly interesting in this context is the finding that TNF itself is capable of activating the PI3K/Akt pathway 17. In certain cell types PI3K signaling was shown to be essential for the activation of NF-B, the main transcription factor responsible for TNF-induced anti-apoptotic signaling. However, this was not the case in HUVECs 36. Madge and co-workers demonstrated that TNF triggers a PI3K-mediated anti-apoptotic response in HUVECs that is independent of Akt and NF-B, and that inhibition of PI3K results in cathepsin B-dependent, but caspase-independent cell death of these cultured endothelial cells 19. These findings, however, do not fit with our in vivo data where the sensitizing effect of PI3K inhibition could be mimicked by Akt inhibition. Moreover, after treatment with wortmannin and TNF DNA fragmentation in tumor endothelial cells, a hallmark of apoptotic cell death, was only observed after disruption of the vessel wall and vessel occlusion had occurred. At that time endothelial cell death may also be a consequence of cell detachment (anoikis) or mere starvation.

We hypothesize that the initial event in the vascular-disrupting effect of TNF is a loss of endothelial cell-cell contact and/or cell-matrix contact. Gaps in the endothelial lining of the tumor vessels were readily visible at 6 hours after treatment with wortmannin and TNF. Similarly, Menon and colleagues reported the formation of gaps in the endothelial lining of the tumor vasculature between 3 and 6 hours after treatment with TNF in human melanoma xenografts in mice and in human melanoma patients after ILP, and they attributed this effect to a disappearance of VE-cadherin from endothelial cell-cell junctions 37. The adherens junction component VE-cadherin is known to play a critical role in endothelial cell-cell connections and tumor angiogenesis 38. It was recently shown that TNF can induce tyrosine phosphorylation of the intracellular domain of VE-cadherin, thereby weakening the VE-cadherin-based junctions and facilitating the formation of intercellular gaps 39. Disruption of VE- cadherin junctions may lead to a shutdown of tumor blood flow by a mechanism of vascular collapse, in analogy with the vessel-disrupting mechanism reported for the General summary & discussion 259

tubulin binding drug CA4P. Disassembly of VE-cadherin junction would cause haemorrhage, an increase of interstitial fluid pressure, and eventually the collapse of the weakened tumor vessels 40. It is currently unclear whether PI3K inhibition could exacerbate the TNF-induced disassembly of VE-cadherin junctions. The integrins

v3 and v5 play an important role in endothelial cell-matrix adhesion in neovasculature 41. Rüegg and colleagues showed that treatment of HUVECs with TNF and IFN- suppressed the activation of integrin v3, leading to a decrease in endothelial cell adhesion and survival, and they observed detachment and apoptosis of

v3-positive endothelial cells in the tumor vasculature of melanoma patients after ILP with TNF plus IFN- 33. Furthermore, they demonstrated that Akt activity can strengthen integrin-dependent endothelial cell adhesion and can protect HUVECs 42, 43 against cell death induced by TNF and/or integrin v3/v5 inhibition . Thus, TNF therapy is known to interfere with both endothelial cell-cell and cell-matrix adhesions. PI3K/Akt pathway inhibition may exacerbate this effect and/or compromise endothelial cell survival during periods of cell detachment.

 Immune cells involved in the antitumor mechanisms of TNF monotherapy and TNF/IFN- combination therapy

As TNF is a potent pro-inflammatory and immunoregulatory cytokine, it is not surprising that TNF-based therapies were shown to result in an infiltration of immune cells in tumors 22, 23. However, it remained to be elucidated whether these immune cells actively contribute to tumor destruction after TNF/IFN- therapy. In the present study we have demonstrated that the antitumor activity of TNF was markedly reduced in NKT-deficient mice as well as in NK1.1+ cell-depleted mice, suggesting that invariant NKT cells significantly contribute to the antitumor effect of TNF. However, even in NKT-deficient mice complete tumor regressions were observed after treatment with mTNF, albeit much less frequently than in wild-type mice, indicating that the antitumor activity of TNF is not strictly dependent on NKT cells but is facilitated by the presence of these cells. Moreover, tumor response to treatment with TNF was similar in athymic C57BL/6nu/nu mice, which also lack invariant NKT cells, as in immune-competent C57BL/6 wild-type mice. One plausible explanation for this discrepancy is that there might be sufficient functional redundancy between different cytotoxic lymphocyte populations. NK cells, NKT cells and CD8+ T cells share many important effector functions, including IFN- production, FasL expression, and perforin/granzyme B-mediated cytotoxicity. The absence of one lymphocyte population may affect the functionality or proportion of another. For example, in athymic C57BL/6nu/nu mice we observed a two- to four-fold increased number of NK cells in spleen and liver, and we found that this could be linked to an increased sensitivity to TNF. Interestingly, a markedly increased sensitivity to TNF was also 260 Chapter 11

observed after transfer of wild-type splenocytes to IRF-1-/- mice. It would be interesting to investigate whether this effect was due to the supranormal number of NK cells in these mice after transfer. Taken together, we have shown that the presence of certain cytotoxic lymphocytes, particularly invariant NKT cells, increases the sensitivity to the antitumor effect of TNF.

Invariant NKT cells have the unique ability to rapidly produce large amounts of immunoregulatory cytokines upon activation, most notably of IFN- 44, 45. Previous experiments in our lab with IFN-R-/- mice and IFN--/- indicated that endogenous IFN- production, acting directly on the tumor cells, may contribute to the TNF- induced antitumor mechanism 26. We may now have identified invariant NKT cells as the source of this endogenous IFN-. However, endogenous IFN- production alone cannot explain the reduced responses observed in NKT-deficient mice and therefore it is likely that other NKT cell-derived cytokines, potentially TNF, or other effector functions are involved in this effect. As the antitumor activity of TNF requires only endothelial TNF-R1 expression, it is puzzling how NKT cells could be specifically recruited to and/or activated at the tumor site 7. One possible explanation is that the actions of TNF would result in the exposure of an endogenous, possibly tumor- derived glycolipid antigen capable of activating NKT cells.

The antitumor effects of mTNF monotherapy and hTNF/IFN- combination therapy are macroscopically as well as microscopically very similar. Both treatments induce a haemorrhagic necrosis of tumors that is characterized by a selective destruction of the tumor vasculature. The mechanisms underlying their antitumor effects are, however, essentially different: (1) Previous work in our lab, using IFN--/- and IFN-R-/- mice, demonstrated that the difference in antitumor activity of mTNF and hTNF cannot be explained by endogenous IFN- induction 32. (2) We have evidence that NKT cells contribute to the TNF-induced antitumor effect by a mechanism that is largely independent of endogenous IFN- production. However, in NKT-deficient IRF-1-/-, CD1-/- and athymic Swissnu/nu mice, which all display reduced response to mTNF, co- treatment of mTNF and IFN- induced complete tumor regression, suggesting that IFN- triggers an additional mechanism that circumvents the need for NKT cells. (3) We found that the synergistic antitumor effect of hTNF and IFN- was largely abrogated in PI3K-/- mice, while the antitumor effect of mTNF was unaffected in these mice. (4) Similarly, depletion of neutrophils using cyclophosphamide completely abolished the synergy of hTNF and IFN-, while depletion did not affect the antitumor activity of mTNF. Taken together, these data imply that the antitumor effect of hTNF/IFN- combination therapy is critically dependent on neutrophils, while this is not the case for mTNF monotherapy.

General summary & discussion 261

Extensive in vitro data exist showing that co-incubation with IFN- sensitizes endothelial cells to TNF-induced cell death 34. Also, Rüegg and co-workers reported that treatment of HUVECs with TNF and IFN- suppressed the activation of integrin

v3, leading to decreased endothelial cell adhesion and survival. They furthermore observed detachment and apoptosis of v3-positive endothelial cells of the tumor vasculature in melanoma patients treated with TNF and IFN- 33. The finding that in our B16Bl6 melanoma model the synergistic antitumor effect of TNF and IFN- required the presence of neutrophils implies that IFN- did not significantly sensitize the tumor endothelium to the direct vascular-disrupting effects of TNF in this in vivo model. We hypothesize that IFN- increases endothelial injury by exacerbating the TNF-triggered inflammatory response. Neutrophils are recruited to sites of inflammation by expression on the activated endothelium of the adhesion molecules E-selectin and ICAM-1, both of which are synergistically upregulated by TNF and IFN- 46-48. The adhesion of neutrophils to activated endothelium may be sufficient to trigger cytotoxic effector functions, including respiratory burst and degranulation of granules containing proteolytic enzymes and myeloperoxidase 49. We can, however, exclude NO production as an important effector function as we previously found that the synergistic antitumor effect of TNF and IFN- was preserved in iNOS-deficient mice 50. The finding that the synergistic antitumor effect of TNF and IFN- was largely abrogated in PI3K-/- mice, suggests that a GPCR-mediated signaling event is important 51. Neutrophils can be activated by a large number of chemotactic factors that signal via GPCRs, including the Glu-Leu-Arg (ELR)-positive chemokines

CXCL1-8 (except CXCL4), and the lipid mediators PAF and LTB4, and it will be challenging to determine which factor(s) is/are involved 49. In conclusion, our data suggest that IFN- does not sensitize for the intrinsic antivascular or antitumor effect of TNF, but allows to trigger an additional, neutrophil-dependent mechanism. This finding offers a new rationale for the addition of IFN- to TNF-based anticancer therapies.

 GalCer versus IFN-

The elucidation of the complete antitumor mechanism elicited by TNF is important as this may reveal new targets for therapeutic intervention. The finding that NKT cells may actively contribute to tumor destruction after TNF therapy prompted us to test whether specific activation of these cells could increase the efficacy of a low-toxic, sub-therapeutic TNF treatment. We demonstrated that pre-treatment with GalCer markedly increased tumor destruction induced by treatment with hTNF or low-dose mTNF. Using IFN-R-/- and transgenic Flk-1 dnIFN-R mice we could show that this effect was completely IFN--dependent and required functional IFN-R expression on endothelial cells, suggesting that GalCer acts by a similar mechanism as 262 Chapter 11

exogenously administered IFN- 26. An important question is of course whether using GalCer would be superior to treatment with IFN-. For example, GalCer may offer better pharmacokinetics or induce IFN- production at the tumor site. We started treatment with TNF 24 hours after i.p. injection of GalCer, when elevated serum levels of IFN- were already largely declined 52. Increased response to TNF was observed for at least several days after injection of GalCer, suggesting that GalCer may trigger a longer-lasting sensitizing effect. It has been demonstrated that the antimetastatic effect of GalCer is dependent on the activation of NK cells and IFN- production by these cells, which is then most likely followed by bystander activation of other immune cells 53, 54. Although no overt toxicities were observed in our model, caution is required when combining TNF with GalCer as both these drugs are known to cause liver toxicity, which is one of the major dose-limiting toxicities of TNF 55, 56. More potent GalCer analogs with a Th1-biased cytokine profile are currently being developed 57. It will be most interesting to test whether these compounds would even further increase the antitumor activity of TNF.

11.3. Conclusion & future perspectives

The ultimate goal of this research project was to identify new therapeutic targets to selectively increase the antitumor activity of TNF, so that a safe and effective systemic treatment would become possible with low, tolerable doses of TNF. In this thesis we presented two new approaches towards this goal: first we demonstrated that interfering with endothelial survival, through inhibition of the PI3K/Akt/mTOR pathway, sensitized for the selective vascular-disrupting effect of TNF, and second we showed that specific pharmacological activation of invariant NKT cells, which appear to contribute to tumor destruction upon treatment with TNF, markedly increased the antitumor efficacy of TNF. Both these approaches allowed approximately a tenfold reduction of the effective dose of TNF in our murine tumor model. It would be interesting to investigate whether combining these approaches would further reduce the required dose of TNF. We also provided a new rationale for the addition of IFN- to TNF-based anticancer therapies by showing that this may trigger an additional, neutrophil-dependent antitumor mechanism. One important remark is that all combination treatments described in this thesis were performed by paralesional injection of TNF and IFN-. Thus, to really assess to safety and efficacy of these combination treatments, a clinically more relevant, systemic treatment should be set up. In case of survival pathway inhibition, the risk of side-effects could be minimized by using more selective inhibitors, for example PI3K isoform-selective inhibitors which are currently being developed or inhibitors of upstream components of the survival pathway such as receptor tyrosine kinase inhibitors and other angiostatics General summary & discussion 263

which are already in clinical trials. Finally, our results may offer new insights into the antitumor mechanisms of TNF and IFN- (summarized in Fig.11.1.).

Fig. 11.1. Potential mechanism for the vascular-disrupting and antitumor effect of TNF, and its synergy with IFN-. TNF and IFN- both act by activating their receptors on endothelial cells of the tumor vasculature. We hypothesize that the initial event in the TNF-induced destruction of the tumor vasculature is the weakening of endothelial cell-cell and/or cell-matrix contact, which would cause leakage, haemorrhage and tissue exposure, and compromise endothelial survival. Blood flow would then rapidly be shut off by a mechanism of vascular collapse and/or thrombosis. How invariant NKT cells (red arrows) contribute to TNF-induced tumor destruction is still to be elucidated, but could involve the production of cytokines such as IFN- and TNF. Alternatively, NKT cells could induce endothelial or tumor cell death through Fas- or perforin/granzyme B-mediated cytotoxicity. NKT cells may become activated by presentation of endogenous, possibly tumor-derived, glycolipid antigens. Co- treatment with IFN- (blue arrows) synergistically upregulates the endothelial expression of adhesion molecules and chemokines. We hypothesize that this results in a massive infiltration of neutrophils, which would then exacerbate the TNF-induced endothelial damage by ROS production or degranulation.

264 Chapter 11

11.4. References

1. Brouckaert, P.G., Leroux-Roels, G.G., Guisez, Y., Tavernier, J. & Fiers, W. In vivo anti- tumour activity of recombinant human and murine TNF, alone and in combination with murine IFN-gamma, on a syngeneic murine melanoma. International journal of cancer 38, 763-769 (1986). 2. Lejeune, F.J., Lienard, D., Matter, M. & Ruegg, C. Efficiency of recombinant human TNF in human cancer therapy. Cancer Immun 6, 6 (2006). 3. Cauwels, A. et al. Protection against TNF-induced lethal shock by soluble guanylate cyclase inhibition requires functional inducible nitric oxide synthase. Immunity 13, 223-231 (2000). 4. Brouckaert, P. et al. Inhibition of and sensitization to the lethal effects of tumor necrosis factor. Journal of inflammation 47, 18-26 (1995). 5. Van Molle, W. et al. Protection of zinc against tumor necrosis factor induced lethal inflammation depends on heat shock protein 70 and allows safe antitumor therapy. Cancer research 67, 7301-7307 (2007). 6. Stoelcker, B. et al. Tumor necrosis factor induces tumor necrosis via tumor necrosis factor receptor type 1-expressing endothelial cells of the tumor vasculature. The American journal of pathology 156, 1171-1176 (2000). 7. Goethals, A., Pasparakis, M., Hostens, J., Breier, G., Kollias, G., Brouckaert, P. Expression of TNF-R1 exclusively on endothelial cells is sufficient to obtain tumor destruction by TNF. (Article submitted) (2008). 8. Shimomura, K. et al. Recombinant human tumor necrosis factor-alpha: thrombus formation is a cause of anti-tumor activity. International journal of cancer 41, 243-247 (1988). 9. Van de Wiel, P.A., Bloksma, N., Kuper, C.F., Hofhuis, F.M. & Willers, J.M. Macroscopic and microscopic early effects of tumour necrosis factor on murine Meth A sarcoma, and relation to curative activity. J Pathol 157, 65-73 (1989). 10. Eggermont, A.M. et al. Isolated limb perfusion with high-dose tumor necrosis factor-alpha in combination with interferon-gamma and melphalan for nonresectable extremity soft tissue sarcomas: a multicenter trial. J Clin Oncol 14, 2653-2665 (1996). 11. Carmeliet, P. & Jain, R.K. Angiogenesis in cancer and other diseases. Nature 407, 249-257 (2000). 12. Hashizume, H. et al. Openings between defective endothelial cells explain tumor vessel leakiness. The American journal of pathology 156, 1363-1380 (2000). 13. Baluk, P., Morikawa, S., Haskell, A., Mancuso, M. & McDonald, D.M. Abnormalities of basement membrane on blood vessels and endothelial sprouts in tumors. The American journal of pathology 163, 1801-1815 (2003). 14. Jain, R.K. Normalizing tumor vasculature with anti-angiogenic therapy: a new paradigm for combination therapy. Nature medicine 7, 987-989 (2001). 15. Gerber, H.P. et al. Vascular endothelial growth factor regulates endothelial cell survival through the phosphatidylinositol 3'-kinase/Akt signal transduction pathway. Requirement for Flk-1/KDR activation. The Journal of biological chemistry 273, 30336-30343 (1998). 16. Vanhaesebroeck, B. & Alessi, D.R. The PI3K-PDK1 connection: more than just a road to PKB. The Biochemical journal 346 Pt 3, 561-576 (2000). 17. Zhang, L., Himi, T., Morita, I. & Murota, S. Inhibition of phosphatidylinositol-3 kinase/Akt or mitogen-activated protein kinase signaling sensitizes endothelial cells to TNF-alpha cytotoxicity. Cell death and differentiation 8, 528-536 (2001). 18. Clermont, F., Adam, E., Dumont, J.E. & Robaye, B. Survival pathways regulating the apoptosis induced by tumour necrosis factor-alpha in primary cultured bovine endothelial cells. Cellular signalling 15, 539-546 (2003). 19. Madge, L.A., Li, J.H., Choi, J. & Pober, J.S. Inhibition of phosphatidylinositol 3-kinase sensitizes vascular endothelial cells to cytokine-initiated cathepsin-dependent apoptosis. The Journal of biological chemistry 278, 21295-21306 (2003). 20. Matschurat, S. et al. Negative regulatory role of PI3-kinase in TNF-induced tumor necrosis. International journal of cancer 107, 30-37 (2003). 21. Manusama, E.R. et al. Assessment of the role of neutrophils on the antitumor effect of TNFalpha in an in vivo isolated limb perfusion model in sarcoma-bearing brown Norway rats. The Journal of surgical research 78, 169-175 (1998). 22. Renard, N. et al. Early endothelium activation and polymorphonuclear cell invasion precede specific necrosis of human melanoma and sarcoma treated by intravascular high-dose tumour necrosis factor alpha (rTNF alpha). International journal of cancer 57, 656-663 (1994). General summary & discussion 265

23. Lejeune, F.J. & Ruegg, C. Recombinant human tumor necrosis factor: an efficient agent for cancer treatment. Bulletin du cancer 93, E90-100 (2006). 24. Ohteki, T. et al. The transcription factor interferon regulatory factor 1 (IRF-1) is important during the maturation of natural killer 1.1+ T cell receptor-alpha/beta+ (NK1+ T) cells, natural killer cells, and intestinal intraepithelial T cells. The Journal of experimental medicine 187, 967-972 (1998). 25. Matsuyama, T. et al. Targeted disruption of IRF-1 or IRF-2 results in abnormal type I IFN gene induction and aberrant lymphocyte development. Cell 75, 83-97 (1993). 26. Hostens, J., Brouckaert, P. (unpublished data). (2003). 27. Balkwill, F.R., Ward, B.G. & Fiers, W. Effects of tumour necrosis factor on human tumour xenografts in nude mice. Ciba Foundation symposium 131, 154-169 (1987). 28. Rubin, S.C. et al. Synergistic activity of tumor necrosis factor and interferon in a nude mouse model of human ovarian cancer. Gynecologic oncology 34, 353-356 (1989). 29. Fulda, S. & Debatin, K.M. IFNgamma sensitizes for apoptosis by upregulating caspase-8 expression through the Stat1 pathway. Oncogene 21, 2295-2308 (2002). 30. Suk, K. et al. Interferon gamma (IFNgamma ) and tumor necrosis factor alpha synergism in ME-180 cervical cancer cell apoptosis and necrosis. IFNgamma inhibits cytoprotective NF- kappa B through STAT1/IRF-1 pathways. The Journal of biological chemistry 276, 13153- 13159 (2001). 31. de Kossodo, S. et al. Changes in endogenous cytokines, adhesion molecules and platelets during cytokine-induced tumour necrosis. British journal of cancer 72, 1165-1172 (1995). 32. Hostens, J., Goethals, A., Breier, G., Schreiber, R.D., Brouckaert, P. IFN- synergistically enhances the anti-tumour effect of Tumour Necrosis Factor by activating the endothelial cells of the neovasculature. (article in preparation). 33. Ruegg, C. et al. Evidence for the involvement of endothelial cell integrin alphaVbeta3 in the disruption of the tumor vasculature induced by TNF and IFN-gamma. Nature medicine 4, 408- 414 (1998). 34. Li, J.H. & Pober, J.S. The cathepsin B death pathway contributes to TNF plus IFN-gamma- mediated human endothelial injury. J Immunol 175, 1858-1866 (2005). 35. Madge, L.A. & Pober, J.S. A phosphatidylinositol 3-kinase/Akt pathway, activated by tumor necrosis factor or interleukin-1, inhibits apoptosis but does not activate NFkappaB in human endothelial cells. The Journal of biological chemistry 275, 15458-15465 (2000). 36. Gustin, J.A. et al. Cell type-specific expression of the IkappaB kinases determines the significance of phosphatidylinositol 3-kinase/Akt signaling to NF-kappa B activation. The Journal of biological chemistry 279, 1615-1620 (2004). 37. Menon, C., Ghartey, A., Canter, R., Feldman, M. & Fraker, D.L. Tumor necrosis factor-alpha damages tumor blood vessel integrity by targeting VE-cadherin. Annals of surgery 244, 781- 791 (2006). 38. Liao, F. et al. Monoclonal antibody to vascular endothelial-cadherin is a potent inhibitor of angiogenesis, tumor growth, and metastasis. Cancer research 60, 6805-6810 (2000). 39. Angelini, D.J. et al. TNF-alpha increases tyrosine phosphorylation of vascular endothelial cadherin and opens the paracellular pathway through fyn activation in human lung endothelia. Am J Physiol Lung Cell Mol Physiol 291, L1232-1245 (2006). 40. Vincent, L. et al. Combretastatin A4 phosphate induces rapid regression of tumor neovessels and growth through interference with vascular endothelial-cadherin signaling. The Journal of clinical investigation 115, 2992-3006 (2005). 41. Friedlander, M. et al. Definition of two angiogenic pathways by distinct alpha v integrins. Science (New York, N.Y 270, 1500-1502 (1995). 42. Bezzi, M., Hasmim, M., Bieler, G., Dormond, O. & Ruegg, C. Zoledronate sensitizes endothelial cells to tumor necrosis factor-induced programmed cell death: evidence for the suppression of sustained activation of focal adhesion kinase and protein kinase B/Akt. The Journal of biological chemistry 278, 43603-43614 (2003). 43. Bieler, G. et al. Distinctive role of integrin-mediated adhesion in TNF-induced PKB/Akt and NF-kappaB activation and endothelial cell survival. Oncogene 26, 5722-5732 (2007). 44. Kawano, T. et al. CD1d-restricted and TCR-mediated activation of valpha14 NKT cells by glycosylceramides. Science (New York, N.Y 278, 1626-1629 (1997). 45. Godfrey, D.I. & Kronenberg, M. Going both ways: immune regulation via CD1d-dependent NKT cells. The Journal of clinical investigation 114, 1379-1388 (2004). 46. Doukas, J. & Pober, J.S. IFN-gamma enhances endothelial activation induced by tumor necrosis factor but not IL-1. J Immunol 145, 1727-1733 (1990). 266 Chapter 11

47. Paleolog, E.M., Aluri, G.R. & Feldmann, M. Contrasting effects of interferon gamma and interleukin 4 on responses of human vascular endothelial cells to tumour necrosis factor alpha. Cytokine 4, 470-478 (1992). 48. Strindhall, J., Lundblad, A. & Pahlsson, P. Interferon-gamma enhancement of E-selectin expression on endothelial cells is inhibited by monensin. Scandinavian journal of immunology 46, 338-343 (1997). 49. Zarbock, A. & Ley, K. Mechanisms and consequences of neutrophil interaction with the endothelium. The American journal of pathology 172, 1-7 (2008). 50. Goethals, A., Brouckaert, P. (unpublished data). (2007). 51. Hirsch, E. et al. Central role for G protein-coupled phosphoinositide 3-kinase gamma in inflammation. Science (New York, N.Y 287, 1049-1053 (2000). 52. Elewaut, D. (personal communication). 53. Hayakawa, Y. et al. Critical contribution of IFN-gamma and NK cells, but not perforin- mediated cytotoxicity, to anti-metastatic effect of alpha-galactosylceramide. European journal of immunology 31, 1720-1727 (2001). 54. Smyth, M.J. et al. Sequential production of interferon-gamma by NK1.1(+) T cells and natural killer cells is essential for the antimetastatic effect of alpha-galactosylceramide. Blood 99, 1259-1266 (2002). 55. Jones, A.L. & Selby, P. Tumour necrosis factor: clinical relevance. Cancer surveys 8, 817-836 (1989). 56. Biburger, M. & Tiegs, G. Alpha-galactosylceramide-induced liver injury in mice is mediated by TNF-alpha but independent of Kupffer cells. J Immunol 175, 1540-1550 (2005). 57. Chang, Y.J. et al. Potent immune-modulating and anticancer effects of NKT cell stimulatory glycolipids. Proceedings of the National Academy of Sciences of the United States of America 104, 10299-10304 (2007).

Chapter 12 Algemene samenvatting & discussie

Algemene samenvatting & discussie 269

12. ALGEMENE SAMENVATTING & DISCUSSIE

12.1. Samenvatting

TNF is een krachtig pro-inflammatoir cytokine dat oorspronkelijk ontdekt werd dankzij zijn opmerkelijke antitumorale activiteit. Via zijn twee celoppervlakte receptoren, TNF-R1 en TNF-R2, is het betrokken bij een groot aantal (patho)fysiologische processen. De meeste informatie over TNF signalisatie, zoals de activatie van de NF-B, MAPK en celdood signalisatiewegen, werd afgeleid van TNF-R1 en deze receptor wordt dan ook beschouwd als de belangrijkste. Het is echter mogelijk dat de rol van TNF-R2 onderschat wordt aangezien deze receptor alleen volledig geactiveerd wordt door de membraangebonden vorm van TNF, en niet door vrij TNF. Eén van de belangrijkste biologische functies van TNF is het mediëren van ontstekingsreacties. Weefsel macrofagen die gestimuleerd worden met LPS of andere bacteriële bestanddelen produceren snel TNF om de lokale bloedvaten te activeren. Dit resulteert in verhoogde doorbloeding door de productie van de vasodilatoren NO en PGI2, verhoogde vasculaire permeabiliteit door verzwakking van endotheliale cel- cel juncties en cytoskeletale reorganisatie, en de rekrutering van neutrofielen, monocyten en andere witte bloedcellen naar de plaats van ontsteking door de opregulatie van chemokines en adhesie moleculen zoals E-selectine, ICAM-1 en VCAM-1. Bovendien kan TNF, als een additioneel beschermingsmechanisme, de lokale bloedtoevoer afsluiten door bloedstolling te stimuleren of door (in)directe endotheelceldood te veroorzaken. Dezelfde mechanismen waarmee TNF zo efficiënt lokale infecties kan bedwingen kunnen echter ook schadelijk worden, zoals in sepsis en chronisch inflammatoire aandoeningen zoals reumatoïde artritis, inflammatoir darmlijden en psoriasis.

De in vivo antitumorale activiteit van recombinant TNF en zijn synergie met IFN- werden voor het eerst aangetoond in het muis B16Bl6 melanoma model door Brouckaert en medewerkers in 1986 1. TNF induceert een snelle, bloederige necrose van experimentele tumoren die gekenmerkt wordt door een selectieve vernietiging van de tumorvasculatuur. Ondanks veelbelovende resultaten in proefdiermodellen waren de eerste klinische studies met systemisch toegediend TNF echter teleurstellend. Dosissen van 10 tot 50 keer lager dan de geschatte effectieve dosis veroorzaakten al ernstige nevenwerkingen, met als belangrijkste lage bloeddruk en orgaanfalen. Om deze nevenwerkingen te vermijden, kwamen twee Belgische chirurgen, Lejeune en Liénard, op het idee om TNF in de ILP setting te gebruiken 2. ILP met TNF, melphalan en eventueel IFN- kan tegenwoordig gebruikt worden voor de behandeling van patiënten met vergevorderde tumoren (in-transit melanoma metastasen en zachte weefsel sarcomas) van de ledematen als alternatief voor amputatie. Het succes van deze methode demonstreert duidelijk het potentieel van TNF als een antikanker geneesmiddel. Aangezien locoregionale therapie echter geen 270 Chapter 12

effect heeft op verafgelegen metastasen, heeft TNF-gebaseerde ILP tot op heden vrijwel geen invloed gehad op de finale overleving van kankerpatiënten, wat de noodzaak voor een effectieve en veilige systemische TNF behandeling benadrukt.

Voorafgaand werk in onze en andere onderzoeksgroepen heeft duidelijk aangetoond dat de antitumorale en toxische effecten van TNF niet onlosmakelijk met elkaar verbonden zijn. Muizen kunnen beschermd worden tegen de TNF-geïnduceerde letale shock met behulp van de sGC inhibitor methyleenblauw zonder het antitumoraal effect nadelig te beïnvloeden 3. Ook zijn er twee acute fase eiwitten geïdentificeerd, 1-zuur glycoproteïne en 1-antitrypsine, die de TNF-geïnduceerde leverschade kunnen verhinderen 4. Bovendien werd recent aangetoond dat het toevoegen van zink aan het drinkwater van muizen kan beschermen tegen de TNF-geïnduceerde letale shock en darmtoxiciteit zonder de antitumorale werking van TNF en IFN- teniet te doen 5. Deze resultaten tonen duidelijk aan dat het mogelijk moet zijn om een veilige systemische behandeling met TNF te ontwikkelen. Het gebruik van een mogelijks letale dosis TNF in combinatie met een inhibitor van de toxische nevenwerkingen is echter geen aantrekkelijke oplossing voor humane kankertherapie aangezien falen van de inhibitor de dood van de patiënt tot gevolg zou kunnen hebben. Een veiligere strategie om de systemische toxiciteit van TNF te verlagen zou zijn om lage TNF dosissen of laag-toxische TNF muteïnes te gebruiken in combinatie met stoffen die selectief de antitumorale activiteit verhogen, zoals IFN-. In het kader van deze laatste strategie behandelden we in deze scriptie drie grote vragen:

(1) Kunnen we gevoeliger maken voor het antitumoraal effect van TNF door te interfereren met endotheliale overlevingssignalisatie ? (2) Welke immuuncellen zijn betrokken bij het antitumoraal mechanisme van TNF ? (3) Hoe versterkt IFN- het antitumoraal effect van TNF ?

 Sensitisatie voor het antitumoraal effect van TNF door inhibitie van endotheliale overlevingssignalisatie

Het preciese mechanisme achter de TNF-geïnduceerde tumor necrose is nog grotendeels onbekend. Het is duidelijk geworden dat het antitumoraal effect van TNF volledig gastheer-gemedieerd is en geen expressie van TNF-receptoren op de tumorcellen vereist. Stoelcker en medewerkers hebben bovendien aangetoond dat het antitumoraal effect van TNF behouden blijft in TNF-R2-/- muizen, maar volledig afwezig is in TNF-R1-/- muizen, wat suggereert dat de antitumorale werking van TNF gebeurt via de TNF-R1 op gastheercellen 6. Vroeger werk in onze onderzoeksgroep heeft dit verder onderbouwd door aan te tonen dat expressie van TNF-R1 alleen op de neovasculatuur (onder controle van een Flk-1 promotor) volstaat voor het volledig Algemene samenvatting & discussie 271

antitumoraal effect van TNF, wat aangeeft dat de endotheelcellen van de tumorvasculatuur de directe doelwitcellen zijn van TNF 7. Dit wordt bevestigd door observaties in proefdiermodellen en ILP patiënten van een selectieve vernietiging van de tumorvasculatuur na behandeling met TNF 2, 8-10. TNF kan dus als een VDA beschouwd worden.

De PI3K/Akt signalisatieweg is één van de belangrijkste regulatoren van cel overleving. Aangezien de tumorvasculatuur in tegenstelling tot de normale vasculatuur niet gestabiliseerd is door normale interacties met pericyten, structuur van de basale membraan en cel-cel juncties, en bovendien onderhevig is aan constante herschikking door onevenwichtigheid van angiogene factoren geproduceerd in de tumor, is het aannemelijk dat deze vasculatuur sterk afhankelijk is van deze overlevings-signalisatieweg 11-14. De PI3K/Akt signalisatieweg wordt geactiveerd door een groot aantal pro-angiogene factoren, o.a. groeifactoren, zoals leden van de 15, 16 VEGF familie, en adhesie molecules, zoals integrine v3 . Het werd aangetoond dat inhibitie van PI3K endotheelcellen in vitro gevoeliger maakt voor de cytotoxische effecten van TNF 17-19. Bovendien rapporteerden Matschurat et al. dat inhibitie van PI3K het antitumoraal effect van TNF versterkt in een muis MethA sarcoma model 20.

Om na te gaan of inhibitie van de PI3K/Akt signalisatieweg gevoeliger zou maken voor het antitumoraal effect van TNF behandelden we B16Bl6 melanoma- of andere tumor-dragende muizen met subtherapeutisch TNF, alleen of in combinatie met één van de PI3K inhibitoren wortmannin of LY294002. De combinatie van een PI3K inhibitor met TNF resulteerde in een uitgesproken antitumoraal effect, met tumorregressie in alle behandelde dieren. We bepaalden dat PI3K inhibitie een tienvoudige verlaging van de benodigde TNF dosis toeliet. Vergelijkbare resultaten werden bekomen met de Akt inhibitor perifosine of de mTOR inhibitor rapamycin. Belangrijk is dat inhibitie van de PI3K/Akt signalisatieweg het antitumoraal effect van TNF versterkte zonder een gelijktijdige verhoging van de toxische nevenwerkingen van TNF te veroorzaken. Met behulp van intravitaal-microscopie en histologische analyse van tumoren toonden we aan dat behandeling met wortmannin en TNF een selectieve vernietiging van de tumorvasculatuur induceerde, met verbreking van onderling contact tussen endotheelcellen en bloeding. In vivo Hoechst kleuring gaf aan dat reeds 6 uur na injectie van wortmannin en TNF de doorbloeding van de tumor bijna volledig afgesloten was. 48 uur na injectie was de doorbloeding nog steeds niet hersteld, zodat de tumor effectief zou kunnen afsterven door ontbering. Tenslotte demonstreerden we dat anti-angiogene therapie met een VEGF-R2 blokkerend antilichaam eveneens de antitumorale activiteit van TNF versterkt in het B16Bl6 melanoma model. Gezamenlijk tonen deze resultaten aan dat interfereren met endotheliale overlevingssignalisatie een significante verlaging van de effectieve TNF dosis toelaat.

272 Chapter 12

 Immuuncellen betrokken bij het antitumoraal mechanisme van TNF

Zoals reeds vermeld blijkt uit vroegere experimenten dat het antitumoraal effect van TNF volledig gastheer-gemedieerd is en enkel TNF-R1 expressie op het tumorendotheel vereist 6, 7. Deze experimenten toonden ook aan dat TNF-R1 expressie op andere cellen dan endotheelcellen, zoals tumorcellen en immuuncellen, onbelangrijk is voor het volledig antitumoraal effect van TNF. Bepaalde immuuncellen zouden echter wel betrokken kunnen zijn bij het antitumoraal mechanisme van TNF. Eggermont en collega’s toonden aan dat volledige lichaamsbestraling van tumordragende ratten, waarbij algemene leukopenie geïnduceerd werd, de tumorvernietiging en huidnecrose op de plaats van de tumor na ILP met TNF en melphalan merkbaar verminderde 21. Bovendien werd er na TNF- gebaseerde ILP een inflammatoire respons waargenomen die resulteerde in een snelle infiltratie van neutrofielen in de tumor, gevolgd door de infiltratie van macrofagen en lymfocyten na enkele dagen tot weken 22, 23. Experimenten uitgevoerd in onze onderzoeksgroep met B16Bl6 tumor-dragende IRF-1-/- muizen, die geen mature NK cellen, NK1.1+ T cellen en CD8+ T cellen ontwikkelen, lieten een sterk verminderde respons op TNF behandeling zien in deze immuun-deficiënte muizen 24-26. De antitumorale activiteit van TNF in IRF-1-/- muizen kon gedeeltelijk hersteld worden door reconstitutie met wild-type splenocyten, wat een rol voor lymfocyten suggereert in dit model.

We onderzochten, met behulp van immuun-deficiënte muizen en methodes om lymfocyten selectief te depleteren, welke lymfocyten betrokken zijn bij het antitumoraal mechanisme van TNF. We vonden dat geen enkele lymfocyt populatie strikt noodzakelijk was voor de antitumorale werking van TNF. Zowel athymische C57BL/6nu/nu muizen, die geen mature T lymfocyten en NKT cellen hebben, als NK cel-gedepleteerde muizen vertoonden een gelijkaardige respons op TNF behandeling als wild-type controle muizen. Daarentegen vonden we dat vier onafhankelijke NKT cel-deficiënte muis stammen (IRF-1-/-, IL-15-/-, J18-/- en CD1-/-) een merkbaar verminderde respons op TNF behandeling vertoonden. We konden de antitumorale activiteit van TNF in IRF-1-/- muizen en J18-/- muizen grotendeels herstellen door deze muizen voor de behandeling te injecteren met leukocyten geïsoleerd uit wild- type lever, een orgaan dat bijzonder rijk is aan NK1.1+ T cellen. Een rol voor NK1.1+ cellen werd ook bevestigd door selectieve depletie van deze cellen met behulp van een anti-NK1.1 antilichaam. Onze resultaten suggereren dat de rekrutering en/of activatie van invariante NKT cellen een belangrijke rol speelt in het antitumoraal mechanisme dat door TNF in gang wordt gezet. We hebben deze hypothese verder onderbouwd door aan te tonen dat een enkele injectie van het specifieke NKT cel ligand GalCer de doeltreffendheid van een subtherapeutische TNF behandeling merkbaar verhoogde. Met behulp van IFN-R-deficiënte muizen demonstreerden we dat het effect van Algemene samenvatting & discussie 273

GalCer volledig afhankelijk is van endogene IFN- productie. Gezamenlijk tonen deze resultaten aan dat invariante NKT cellen mogelijks een interessant doelwit zijn om de therapeutische toepasbaarheid van TNF uit te breiden.

 Het mechanisme achter het synergetisch antitumoraal effect van TNF en IFN-

In vele experimentele tumormodellen werd aangetoond dat medebehandeling met IFN- de antitumorale activiteit van TNF synergetisch kan versterking 1, 27, 28. Bovendien kon men in klinische studies eveneens een trend tot beter resultaat waarnemen wanneer IFN- deel uitmaakte van de ILP behandeling, wat aanwijst dat de synergie van TNF en IFN- ook bestaat in dit systeem 2. In vitro bestaan er aanwijzingen dat IFN- gevoeliger maakt voor cytokine-geïnduceerde cytotoxiciteit op een IRF-1-afhankelijke manier 29, 30. Wij vonden daarentegen dat het synergetisch antitumoraal effect van TNF en IFN- nog steeds aanwezig was in IRF-1-deficiënte muizen, wat aangeeft dat dit in vivo effect onafhankelijk is van IRF-1 26. Uit histologische analyses blijkt dat behandeling met TNF en IFN- een selectieve vernietiging van de tumor-geassocieerde vasculatuur veroorzaakt 2, 31, 32. Bovendien heeft onderzoek in ons laboratorium aangetoond dat het synergetisch antitumoraal effect van TNF en IFN- afhankelijk is van IFN-R expressie op endotheelcellen 33. Dit wijst erop dat de endotheelcellen van de tumorvasculatuur de doelwitcellen zijn van TNF/IFN- combinatietherapie. Li en medewerkers hebben recent aangetoond dat IFN- de vrijstelling van cathepsine B uit lysosomen kan induceren in HUVECs, en dat dit sensitiseert voor TNF-geïnduceerde, cathepsine B-afhankelijke celdood 34. Het mechanisme achter het in vivo synergetisch effect van TNF en IFN- is echter nog grotendeels onbekend.

We gingen na of immuuncellen, die in tumoren blijken te infiltreren na TNF/IFN-- gebaseerde ILP, betrokken zijn bij het synergetisch antitumoraal effect van TNF en IFN-. Onze resultaten geven aan dat de synergie van TNF en IFN- onafhankelijk is van NKT cellen of andere lymfocyten waarvan we voorheen hadden aangetoond dat ze kunnen bijdragen tot het antitumoraal effect van TNF. Daarentegen vonden we dat het synergetisch antitumoraal effect van TNF en IFN- sterk verminderd was in PI3K-deficiënte muizen, die defecten vertonen in chemokine-geïnduceerde responses van neutrofielen en monocyten/macrofagen. Specifieke depletie van macrofagen had geen effect op de synergie van TNF en IFN- in ons muis B16Bl6 melanoma model, terwijl depletie van neutrofielen deze synergie teniet deed. Onze data impliceren dat het synergetisch antitumoraal effect van TNF en IFN- strikt afhankelijk is van neutrofielen. Onze huidige hypothese is daarom dat IFN- endotheelschade en 274 Chapter 12

tumorvernietiging vermeerdert door de TNF-geïnduceerde ontstekingsreactie te versterken.

12.2. Discussie

 Het bloedvat-vernietigend mechanisme van TNF

TNF-gebaseerde behandelingen worden gekenmerkt door een selectieve vernietiging van de tumorvasculatuur. Dit wordt niet alleen in experimentele proefdiermodellen waargenomen, maar ook in patiënten die behandeld werden via ILP 2, 9. Het mechanisme achter dit opmerkelijk bloedvat-vernietigend effect moet echter nog ontrafeld worden. In deze studie hebben wij aangetoond dat breed-spectrum PI3K inhibitie sensitiseert voor dit bloedvat-vernietigend effect van TNF. We gebruikten intravitaal-microscopie en verscheidene histologische kleuringsmethodes om de grove effecten van behandeling met wortmannin en TNF op de tumorvasculatuur te visualiseren. Onze data geven aan dat deze behandeling zeer snel na injectie leidt tot lekkage van bloedvaten en bloeding, gevolgd door het volledig afsluiten van de doorbloeding van de meeste tumorbloedvaten tegen 6 uur na injectie. De vernietiging van de vaatwand was compleet tegen 24 uur na injectie. Matschurat en medewerkers hebben aangetoond dat inhibitie van PI3K de TNF-geïnduceerde expressie van tissue factor, de initiator van de extrinsieke bloedstollingscascade, verhoogt. Bovendien demonstreerden zij dat medebehandeling met wortmannin het antitumoraal effect van TNF versterkt in een muis MethA sarcoma model. Samen suggereren deze resultaten dus dat expressie van tissue factor op endotheelcellen van de tumorvasculatuur, wat bloedstolling en occlusie van bloedvaten zou veroorzaken, verantwoordelijk zou zijn voor het antitumorale effect van deze combinatietherapie 20. Onze bevindingen sluiten echter niet volledig aan bij deze hypothese. Intravasculaire bloedstolling is een beschermend mechanisme om in het geval van endotheelschade bloeding te verhinderen. De uitgesproken vasculaire lekkage en bloeding die we waarnamen na behandeling met wortmannin en TNF suggereren dat occlusie van bloedvaten pas gebeurt na vernietiging van de vaatwand.

Het werd meermaals aangetoond dat PI3K inhibitie endotheelcellen in vitro gevoeliger kan maken voor TNF-geïnduceerde cytotoxiciteit 18, 19, 35. Bijzonder interessant in deze context is dat TNF blijkbaar zelf in staat is om de PI3K/Akt signalisatieweg te activeren 17. Het is zelfs aangetoond dat PI3K signalisatie in bepaalde celtypes noodzakelijk is voor de activatie van NF-B, de belangrijkste transcriptiefactor verantwoordelijk voor TNF-geïnduceerde anti-apoptotische signalisatie. Dit bleek echter niet het geval in HUVECs 36. Madge en medewerkers rapporteerden dat TNF in HUVECs een PI3K-gemedieerde anti-apoptotische respons induceert die onafhankelijk is van Akt en NF-B, en dat PI3K inhibitie deze Algemene samenvatting & discussie 275

endotheelcellen sensitiseert voor een cathepsine B-afhankelijke, maar caspase- onafhankelijke celdood 19. Dit komt echter niet overeen met onze in vivo data waar het sensitiserend effect van PI3K inhibitie kon nagebootst worden door Akt inhibitie. Bovendien werd er na behandeling met wortmannin en TNF in endotheelcellen alleen DNA fragmentatie, een kenmerk van apoptose, waargenomen nadat disruptie van de vaatwand en occlusie van bloedvaten hadden plaatsgevonden. Endotheliale celdood kan dan ook het gevolg zijn van het verlies van cel-extracellulaire matrix contact (anoikis) of louter ontbering.

Onze huidige hypothese is dat een verlies van endotheliaal cel-cel en/of cel- extracellulaire matrix contact aan de basis ligt van het bloedvat-vernietigend effect van TNF. Openingen tussen de endotheelcellen van de tumorvasculatuur waren 6 uur na injectie van wortmannin en TNF al waarneembaar. Menon et al. rapporteerden eveneens het verschijnen van openingen in de vaatwand van tumorbloedvaten tussen 3 en 6 uren na behandeling met TNF, zowel in humane melanoma xenograften in muizen als in melanoma patiënten die behandeld werden via ILP. Ze schreven deze openingen toe aan het verdwijnen van VE-cadherine uit endotheliale cel-cel juncties 37. VE-cadherine is een essentieel onderdeel van adherens juncties tussen endotheelcellen en speelt een belangrijke rol in tumor angiogenese 38. Recent werd aangetoond dat TNF tyrosine fosforylatie van het intracellulair domein van VE- cadherine kan induceren, waarvan geweten is dat het de VE-cadherine juncties verzwakt en dus de vorming van intercellulaire openingen vergemakkelijkt 39. Disruptie van VE-cadherine juncties zou de doorbloeding van de tumor kunnen stilleggen via een mechanisme van vasculaire collaps, zoals beschreven werd voor het bloedvat-vernietigend mechanisme van het tubuline-bindend geneesmiddel CA4P. Disruptie van VE-cadherine juncties zou bloeding, een verhoging van interstitiële vloeistofdruk, en uiteindelijk het dichtslaan van de verzwakte bloedvaten veroorzaken 40. Het is op dit ogenblik echter onduidelijk of PI3K inhibitie het effect van TNF op

VE-cadherine juncties zou kunnen versterken. De integrines v3 en v5 spelen een belangrijke rol in endotheelcel-extracellulaire matrix adhesie in neovasculatuur 41. Rüegg en medewerkers hebben aangetoond dat behandeling van HUVECs met TNF en IFN- de activatie van integrine v3 onderdrukte, wat leidde tot een verminderde endotheelcel adhesie en overleving. Ook observeerden ze het loskomen en afsterven van v3-positieve endotheelcellen in de tumorvasculatuur van melanoma patiënten na ILP met TNF en IFN- 32. Recent demonstreerden ze bovendien dat Akt activatie de integrine-afhankelijke endotheelcel adhesie kan versterken en HUVECs kan beschermen tegen celdood geïnduceerd door TNF en/of door inhibitie van de 42, 43 integrines v3en v5 . Het is dus bekend dat TNF kan interfereren met zowel endotheliale cel-cel als cel-extracellulaire matrix interacties. Inhibitie van de PI3K/Akt signalisatieweg zou dit effect kunnen versterken of zou de overleving van de losgekomen endotheelcel kunnen ondermijnen. 276 Chapter 12

 Immuuncellen betrokken bij het antitumoraal mechanisme van TNF monotherapie en TNF/IFN- combinatietherapie

Aangezien TNF een krachtig pro-inflammatoir en immuun-regulerend cytokine is, is het weinig verrassend dat na TNF-gebaseerde behandelingen een infiltratie van immuuncellen in tumoren waargenomen werd 22, 23. Het was echter nog niet geweten of deze cellen ook actief bijdragen tot de tumorvernietiging na behandeling met TNF/IFN-. In deze studie toonden wij aan dat de antitumorale activiteit van TNF merkbaar verminderd was in NKT-deficiënte muizen en NK1.1+ cel-gedepleteerde muizen, wat suggereert dat invariante NKT cellen effectief bijdragen tot het antitumoraal effect van TNF. Desondanks werden volledige tumorregressies ook waargenomen in NKT-deficiënte muizen, zij het veel minder frequent dan in wild- type muizen. Dit wijst erop dat de antitumorale werking van TNF niet strikt afhankelijk is van NKT cellen maar wel vergemakkelijkt wordt door de aanwezigheid van deze cellen. C57BL/6nu/nu muizen, die eveneens geen invariante NKT cellen hebben, vertoonden echter een gelijkaardige respons op TNF behandeling als wild- type controle muizen. Een mogelijke verklaring voor deze paradox is dat er voldoende functionele redundantie kan zijn tussen verschillende types van cytotoxische lymfocyten. NK cellen, NKT cellen en cytotoxische CD8+ T cellen delen veel belangrijke effector functies, zoals IFN- productie, FasL expressie en perforine/granzyme B-gemedieerde cytotoxiciteit. Bovendien kan de afwezigheid van één lymfocyt populatie de functionaliteit of proportie van een andere beïnvloeden. Zo observeerden wij in C57BL/6nu/nu muizen een twee- tot viervoudig toegenomen percentage NK cellen in milt en lever, en we toonden aan dat dit gekoppeld kon worden aan een toegenomen gevoeligheid voor TNF. Een verhoogde gevoeligheid voor TNF werd ook geobserveerd na reconstitutie van IRF-1-/- muizen met wild-type splenocyten. Het zou interessant zijn om te na te gaan of dit effect te wijten was een supernormaal aantal NK cellen in deze muizen na transfer. Samen tonen deze resultaten aan dat de aanwezigheid van bepaalde cytotoxische lymfocyten, in het bijzonder NKT cellen, de gevoeligheid voor het antitumoraal effect van TNF verhoogt.

Invariante NKT cellen hebben de unieke eigenschap om snel grote hoeveelheden immuun-regulerende cytokines te produceren na activatie, met als meest opmerkelijke kandidaat IFN- 44, 45. Vroegere experimenten in ons lab met IFN-R-/- en IFN--/- muizen wezen erop dat endogene productie van IFN-, dat direct inwerkt op de tumorcellen, kan bijdragen tot het antitumoraal effect van TNF 26. Mogelijks hebben we nu invariante NKT cellen geïdentificeerd als de bron van dit endogeen IFN-. De productie van endogeen IFN- alleen kan echter de verlaagde respons waargenomen in NKT-deficiënte muizen niet volledig verklaren, en vermoedelijk spelen dus andere Algemene samenvatting & discussie 277

cytokines, mogelijks TNF, of andere effector functies van NKT cellen een rol in het antitumoraal mechanisme van TNF. Aangezien alleen TNF-R1 expressie op endotheelcellen vereist is voor de antitumorale activiteit van TNF, is het momenteel onduidelijk hoe NKT cellen specifiek gerekruteerd worden naar de tumor en/of geactiveerd worden na behandeling met TNF 7. Een mogelijke verklaring kan zijn dat de acties van TNF zouden resulteren in de blootstelling van een endogeen glycolipide antigen, mogelijks afkomstig van de tumorcellen, dat in staat is om NKT cellen te activeren.

De antitumorale effecten van mTNF monotherapie en hTNF/IFN- combinatietherapie zijn zowel macroscopisch als microscopisch sterk gelijkaardig. Beide behandelingen veroorzaken een bloederige necrose van tumoren die gekenmerkt wordt door een selectieve vernietiging van de tumorvasculatuur. De onderliggende mechanismen van deze antitumorale effecten zijn echter duidelijk verschillend: (1) Vroeger werk in onze onderzoeksgroep met IFN--/- en IFN-R-/- muizen heeft aangetoond dat het verschil in antitumorale activiteit tussen mTNF en hTNF niet verklaard kan worden door een verschil in endogene IFN- inductie 33. (2) We hebben aanwijzingen dat NKT cellen bijdragen tot het antitumoraal effect van TNF via een mechanisme dat grotendeels onafhankelijk is van endogene IFN- productie. In NKT-deficiënte IRF-1-/-, CD1-/- en athymische Swissnu/nu muizen, die allemaal een verlaagde respons vertonen ten opzichte van mTNF, veroorzaakt mTNF/IFN- combinatietherapie echter wel volledige tumorregressie, wat suggereert dat IFN- een additioneel mechanisme in gang zet dat onafhankelijk is van NKT cellen. (3) We vonden dat het synergetisch antitumoraal effect van hTNF en IFN- sterk verminderd was in PI3K-/- muizen, terwijl de antitumorale activiteit van mTNF onaangetast was in deze muizen. (4) Depletie van neutrofielen met behulp van cyclofosfamide deed het synergetisch effect van hTNF en IFN- teniet, terwijl dit geen effect had op de antitumorale activiteit van mTNF. Deze resultaten impliceren dat het synergetisch antitumoraal effect van hTNF/IFN- combinatietherapie strikt afhankelijk is van neutrofielen, terwijl dit niet het geval is voor mTNF monotherapie.

Er bestaan overtuigende in vitro gegevens die aantonen dat co-incubatie met IFN- endotheelcellen gevoeliger maakt voor TNF-geïnduceerde celdood 34. Rüegg en medewerkers hebben ook aangetoond dat behandeling van HUVECs met TNF en

IFN- de activatie van integrine v3 onderdrukte, wat leidde tot een verminderde endotheelcel adhesie en overleving. Ook observeerden ze het loskomen en afsterven van v3-positieve endotheelcellen in de tumorvasculatuur van melanoma patiënten na ILP met TNF en IFN- 32. De bevinding dat in ons B16Bl6 melanoma model het synergetisch antitumoraal effect van TNF en IFN- de aanwezigheid van neutrofielen vereist, impliceert dat IFN- in dit in vivo model het tumorendotheel niet significant gevoeliger maakte voor de intrinsieke bloedvat-vernietigende effecten van TNF. Onze 278 Chapter 12

hypothese is dat IFN- endotheelschade en tumorvernietiging vermeerdert door de TNF-geïnduceerde ontstekingsreactie te versterken. Neutrofielen worden gerekruteerd naar inflammatoire sites door opregulatie van de adhesie moleculen E-selectine en ICAM-1 op het geactiveerde endotheel. Zowel E-selectine als ICAM-1 worden synergetisch opgereguleerd door TNF en IFN- 46-48. De adhesie van neutrofielen op het geactiveerd endotheel zou kunnen volstaan om cytotoxische effector functies te activeren, zoals ‘oxidatieve burst’ en degranulatie van granules die proteolytische enzymes en myeloperoxidase bevatten 49. We kunnen NO productie echter uitsluiten als een belangrijke effector functie aangezien vroeger werk in onze groep aangetoond heeft dat het synergetisch antitumoraal effect van TNF en IFN- behouden was in iNOS-deficiënte muizen 50. De bevinding dat het synergetisch antitumoraal effect van TNF en IFN- sterk verminderd was in PI3K-/- muizen suggereert dat GPCR’s betrokken zijn in dit effect 51. Neutrofielen kunnen geactiveerd worden door een groot aantal chemotactische factoren die signaliseren via GPCR’s, zoals de Glu-Leu-Arg (ELR)-positieve chemokines CXCL1-8 (met uitzondering van CXCL4), en de lipide 49 mediatoren PAF en LTB4 . Het zal een uitdaging zijn om te bepalen welke factoren betrokken zijn. We kunnen besluiten dat onze resultaten aangeven dat IFN- niet sensitiseert voor het intrinsiek antivasculair of antitumoraal effect van TNF, maar toelaat om een additioneel, neutrofiel-afhankelijk mechanisme aan te slaan. Deze bevinding biedt een nieuwe rationale voor de toevoeging van IFN- aan TNF- gebaseerde antikanker behandelingen.

 GalCer versus IFN-

De ontrafeling van het volledige antitumorale mechanisme van TNF is belangrijk omdat het nieuwe doelwitten kan opleveren voor therapeutische interventie. De bevinding dat NKT cellen mogelijks actief bijdragen tot het antitumoraal effect van TNF zette ons ertoe aan om te testen of specifieke activatie van deze cellen de doeltreffendheid van een laag-toxische, subtherapeutische TNF behandeling zou kunnen verhogen. We demonstreerden dat voorbehandeling met GalCer het tumorvernietigend effect van een behandeling met hTNF of lage-dosis mTNF merkbaar versterkte. Aan de hand van IFN-R-/- muizen en transgene Flk-1 dnIFN-R muizen konden we aantonen dat dit effect volledig afhankelijk is van IFN- en de expressie van een functionele IFN-R vereist op endotheelcellen. Dit wijst erop dat GalCer volgens een gelijkaardig mechanisme werkt als exogeen toegediend IFN- 26. Een belangrijke vraag is uiteraard of het gebruik van GalCer superieur zou zijn aan behandeling met IFN-. GalCer zou, bijvoorbeeld, een betere farmacokinetiek kunnen bieden of zou IFN- productie kunnen induceren op de juiste plaats, met name ter hoogte van de tumor. We startten de behandeling met TNF 24 uur na i.p. injectie van GalCer, wanneer de verhoogde serum concentraties van IFN- al grotendeels Algemene samenvatting & discussie 279

afgenomen waren 52. We observeerden een verhoogde respons op TNF behandeling gedurende op zijn minst enkele dagen na injectie van GalCer, wat suggereert dat GalCer een langduriger sensitiserend effect veroorzaakt. Het is aangetoond dat het antimetastatisch effect van GalCer afhankelijk is van de activering van NK cellen en IFN- productie door deze cellen, wat dan hoogstwaarschijnlijk gevolgd wordt door de activatie van andere immuuncellen 53, 54. We hebben geen duidelijke toxische nevenwerkingen waargenomen na behandeling met GalCer en TNF. Desondanks is er toch enige voorzichtigheid geboden wanneer men deze twee geneesmiddelen wil combineren aangezien bekend is dat zowel TNF als GalCer leverschade kunnen veroorzaken, wat toch al één van de dosis-limiterende nevenwerkingen van TNF is 55, 56. Krachtigere GalCer analogen met een meer Th1-geörienteerd cytokine profiel worden momenteel ontwikkeld 57. Het zou bijzonder interessant zijn om te testen of deze analogen de antitumorale activiteit van TNF nog verder zouden kunnen versterken.

12.3. Conclusie & perspectieven

Het finale doel van dit onderzoeksproject was om nieuwe therapeutische doelwitten te identificeren die het mogelijk maken om selectief de antitumorale activiteit van TNF te versterken, zodat een veilige en doeltreffende systemische behandeling met lage, tolereerbare dosissen TNF mogelijk zou worden. In dit proefschrift beschreven we twee nieuwe benaderingen om dit doel te bereiken: ten eerste demonstreerden we dat interfereren met endotheelcel overleving, door inhibitie van de PI3K/Akt/mTOR signalisatieweg, gevoeliger maakte voor het selectief bloedvat-vernietigend effect van TNF, en ten tweede toonden we aan dat we de antitumorale activiteit van TNF merkbaar konden versterken door farmacologische activatie van invariante NKT cellen, waarvan we aanwijzingen hebben dat ze bijdragen tot de tumorvernietiging na behandeling met TNF. Beide benaderingen lieten een tienvoudige verlaging van de effectieve TNF dosis toe in ons muis tumor model. Het zou interessant zijn om te onderzoeken of een combinatie van deze benaderingen de benodigde TNF dosis nog verder zou verlagen. We leverden ook een nieuwe rationale voor de toevoeging van IFN- aan TNF-gebaseerde antikanker behandelingen door aan te tonen dat dit toelaat om een additioneel, neutrofiel-afhankelijk mechanisme te induceren. Een belangrijke opmerking bij al deze combinatietherapieën is dat alle behandeling beschreven in dit proefschrift gebeurden via paralesionale injectie van TNF en IFN-. Dus, om de doeltreffendheid en veiligheid van deze combinatietherapieën echt te kunnen beoordelen, zou een klinisch meer relevante, systemische behandeling opgezet moeten worden. In het geval van overlevingssignalisatie inhibitie zou het risico op nevenwerkingen sterk verlaagd kunnen worden door gebruik te maken van meer selectieve inhibitoren, zoals PI3K isovorm-selectieve inhibitoren die momenteel 280 Chapter 12

ontwikkeld worden, of inhibitoren van stroomopwaartste componenten van de overlevings-signalisatieweg, zoals receptor tyrosine kinase inhibitoren of andere angiostatica die al getest worden in klinische studies. Tot slot kunnen onze resultaten ook nieuwe inzichten bieden in de antitumorale mechanismen van TNF en IFN- (samengevat in Fig.12.1.)

Fig. 12.1. Potentieel mechanisme van het bloedvat-vernietigend en antitumoraal effect van TNF, en zijn synergie met IFN-. TNF en IFN- werken beide door activatie van hun receptoren op het tumorendotheel. Onze hypothese is dat het verlies van endotheliaal cel-cel en/of cel-extracellulaire matrix contact aan de basis ligt van het bloedvat-vernietigend effect van TNF. Dit zou lekkage, bloeding en blootstelling van het onderliggende weefsel veroorzaken, en zou de overleving van endotheelcellen ondermijnen. De bloedtoevoer zou dan snel afgesloten worden door een mechanisme van vasculaire collaps en/of trombose. Hoe invariante NKT cellen (rode pijlen) bijdragen tot de TNF- geïnduceerde tumorvernietiging is onbekend, maar zou de productie van cytokines zoals TNF en IFN- kunnen omvatten. Anderzijds zouden NKT cellen ook endotheel- of tumorceldood kunnen induceren via Fas- of perforine/granzyme B-gemedieerde cytotoxiciteit. NKT cellen zouden geactiveerd kunnen worden door de presentatie van endogene glycolipide antigenen, die mogelijks afkomstig zijn van tumorcellen. Medebehandeling met IFN- (blauwe pijlen) veroorzaakt synergetische opregulatie van chemokines en adhesie molecules op endotheelcellen. Onze hypothese is dat dit resulteert in een massale infiltratie van neutrofielen, die dan de TNF-geïnduceerde endotheelschade zouden versterken door ROS productie of degranulatie.

Algemene samenvatting & discussie 281

12.4. Referenties

1. Brouckaert, P.G., Leroux-Roels, G.G., Guisez, Y., Tavernier, J. & Fiers, W. In vivo anti- tumour activity of recombinant human and murine TNF, alone and in combination with murine IFN-gamma, on a syngeneic murine melanoma. International journal of cancer 38, 763-769 (1986). 2. Lejeune, F.J., Lienard, D., Matter, M. & Ruegg, C. Efficiency of recombinant human TNF in human cancer therapy. Cancer Immun 6, 6 (2006). 3. Cauwels, A. et al. Protection against TNF-induced lethal shock by soluble guanylate cyclase inhibition requires functional inducible nitric oxide synthase. Immunity 13, 223-231 (2000). 4. Brouckaert, P. et al. Inhibition of and sensitization to the lethal effects of tumor necrosis factor. Journal of inflammation 47, 18-26 (1995). 5. Van Molle, W. et al. Protection of zinc against tumor necrosis factor induced lethal inflammation depends on heat shock protein 70 and allows safe antitumor therapy. Cancer research 67, 7301-7307 (2007). 6. Stoelcker, B. et al. Tumor necrosis factor induces tumor necrosis via tumor necrosis factor receptor type 1-expressing endothelial cells of the tumor vasculature. The American journal of pathology 156, 1171-1176 (2000). 7. Goethals, A., Pasparakis, M., Hostens, J., Breier, G., Kollias, G., Brouckaert, P. Expression of TNF-R1 exclusively on endothelial cells is sufficient to obtain tumor destruction by TNF. (Article submitted) (2008). 8. Shimomura, K. et al. Recombinant human tumor necrosis factor-alpha: thrombus formation is a cause of anti-tumor activity. International journal of cancer 41, 243-247 (1988). 9. Van de Wiel, P.A., Bloksma, N., Kuper, C.F., Hofhuis, F.M. & Willers, J.M. Macroscopic and microscopic early effects of tumour necrosis factor on murine Meth A sarcoma, and relation to curative activity. J Pathol 157, 65-73 (1989). 10. Eggermont, A.M. et al. Isolated limb perfusion with high-dose tumor necrosis factor-alpha in combination with interferon-gamma and melphalan for nonresectable extremity soft tissue sarcomas: a multicenter trial. J Clin Oncol 14, 2653-2665 (1996). 11. Carmeliet, P. & Jain, R.K. Angiogenesis in cancer and other diseases. Nature 407, 249-257 (2000). 12. Hashizume, H. et al. Openings between defective endothelial cells explain tumor vessel leakiness. The American journal of pathology 156, 1363-1380 (2000). 13. Baluk, P., Morikawa, S., Haskell, A., Mancuso, M. & McDonald, D.M. Abnormalities of basement membrane on blood vessels and endothelial sprouts in tumors. The American journal of pathology 163, 1801-1815 (2003). 14. Jain, R.K. Normalizing tumor vasculature with anti-angiogenic therapy: a new paradigm for combination therapy. Nature medicine 7, 987-989 (2001). 15. Gerber, H.P. et al. Vascular endothelial growth factor regulates endothelial cell survival through the phosphatidylinositol 3'-kinase/Akt signal transduction pathway. Requirement for Flk-1/KDR activation. The Journal of biological chemistry 273, 30336-30343 (1998). 16. Vanhaesebroeck, B. & Alessi, D.R. The PI3K-PDK1 connection: more than just a road to PKB. The Biochemical journal 346 Pt 3, 561-576 (2000). 17. Zhang, L., Himi, T., Morita, I. & Murota, S. Inhibition of phosphatidylinositol-3 kinase/Akt or mitogen-activated protein kinase signaling sensitizes endothelial cells to TNF-alpha cytotoxicity. Cell death and differentiation 8, 528-536 (2001). 18. Clermont, F., Adam, E., Dumont, J.E. & Robaye, B. Survival pathways regulating the apoptosis induced by tumour necrosis factor-alpha in primary cultured bovine endothelial cells. Cellular signalling 15, 539-546 (2003). 19. Madge, L.A., Li, J.H., Choi, J. & Pober, J.S. Inhibition of phosphatidylinositol 3-kinase sensitizes vascular endothelial cells to cytokine-initiated cathepsin-dependent apoptosis. The Journal of biological chemistry 278, 21295-21306 (2003). 20. Matschurat, S. et al. Negative regulatory role of PI3-kinase in TNF-induced tumor necrosis. International journal of cancer 107, 30-37 (2003). 21. Manusama, E.R. et al. Assessment of the role of neutrophils on the antitumor effect of TNFalpha in an in vivo isolated limb perfusion model in sarcoma-bearing brown Norway rats. The Journal of surgical research 78, 169-175 (1998). 22. Renard, N. et al. Early endothelium activation and polymorphonuclear cell invasion precede specific necrosis of human melanoma and sarcoma treated by intravascular high-dose tumour necrosis factor alpha (rTNF alpha). International journal of cancer 57, 656-663 (1994). 282 Chapter 12

23. Lejeune, F.J. & Ruegg, C. Recombinant human tumor necrosis factor: an efficient agent for cancer treatment. Bulletin du cancer 93, E90-100 (2006). 24. Ohteki, T. et al. The transcription factor interferon regulatory factor 1 (IRF-1) is important during the maturation of natural killer 1.1+ T cell receptor-alpha/beta+ (NK1+ T) cells, natural killer cells, and intestinal intraepithelial T cells. The Journal of experimental medicine 187, 967-972 (1998). 25. Matsuyama, T. et al. Targeted disruption of IRF-1 or IRF-2 results in abnormal type I IFN gene induction and aberrant lymphocyte development. Cell 75, 83-97 (1993). 26. Hostens, J., Brouckaert, P. (unpublished data). (2003). 27. Balkwill, F.R., Ward, B.G. & Fiers, W. Effects of tumour necrosis factor on human tumour xenografts in nude mice. Ciba Foundation symposium 131, 154-169 (1987). 28. Rubin, S.C. et al. Synergistic activity of tumor necrosis factor and interferon in a nude mouse model of human ovarian cancer. Gynecologic oncology 34, 353-356 (1989). 29. Fulda, S. & Debatin, K.M. IFNgamma sensitizes for apoptosis by upregulating caspase-8 expression through the Stat1 pathway. Oncogene 21, 2295-2308 (2002). 30. Suk, K. et al. Interferon gamma (IFNgamma ) and tumor necrosis factor alpha synergism in ME-180 cervical cancer cell apoptosis and necrosis. IFNgamma inhibits cytoprotective NF- kappa B through STAT1/IRF-1 pathways. The Journal of biological chemistry 276, 13153- 13159 (2001). 31. de Kossodo, S. et al. Changes in endogenous cytokines, adhesion molecules and platelets during cytokine-induced tumour necrosis. British journal of cancer 72, 1165-1172 (1995). 32. Ruegg, C. et al. Evidence for the involvement of endothelial cell integrin alphaVbeta3 in the disruption of the tumor vasculature induced by TNF and IFN-gamma. Nature medicine 4, 408- 414 (1998). 33. Hostens, J., Goethals, A., Breier, G., Schreiber, R.D., Brouckaert, P. IFN- synergistically enhances the anti-tumour effect of Tumour Necrosis Factor by activating the endothelial cells of the neovasculature. (article in preparation). 34. Li, J.H. & Pober, J.S. The cathepsin B death pathway contributes to TNF plus IFN-gamma- mediated human endothelial injury. J Immunol 175, 1858-1866 (2005). 35. Madge, L.A. & Pober, J.S. A phosphatidylinositol 3-kinase/Akt pathway, activated by tumor necrosis factor or interleukin-1, inhibits apoptosis but does not activate NFkappaB in human endothelial cells. The Journal of biological chemistry 275, 15458-15465 (2000). 36. Gustin, J.A. et al. Cell type-specific expression of the IkappaB kinases determines the significance of phosphatidylinositol 3-kinase/Akt signaling to NF-kappa B activation. The Journal of biological chemistry 279, 1615-1620 (2004). 37. Menon, C., Ghartey, A., Canter, R., Feldman, M. & Fraker, D.L. Tumor necrosis factor-alpha damages tumor blood vessel integrity by targeting VE-cadherin. Annals of surgery 244, 781- 791 (2006). 38. Liao, F. et al. Monoclonal antibody to vascular endothelial-cadherin is a potent inhibitor of angiogenesis, tumor growth, and metastasis. Cancer research 60, 6805-6810 (2000). 39. Angelini, D.J. et al. TNF-alpha increases tyrosine phosphorylation of vascular endothelial cadherin and opens the paracellular pathway through fyn activation in human lung endothelia. Am J Physiol Lung Cell Mol Physiol 291, L1232-1245 (2006). 40. Vincent, L. et al. Combretastatin A4 phosphate induces rapid regression of tumor neovessels and growth through interference with vascular endothelial-cadherin signaling. The Journal of clinical investigation 115, 2992-3006 (2005). 41. Friedlander, M. et al. Definition of two angiogenic pathways by distinct alpha v integrins. Science (New York, N.Y 270, 1500-1502 (1995). 42. Bezzi, M., Hasmim, M., Bieler, G., Dormond, O. & Ruegg, C. Zoledronate sensitizes endothelial cells to tumor necrosis factor-induced programmed cell death: evidence for the suppression of sustained activation of focal adhesion kinase and protein kinase B/Akt. The Journal of biological chemistry 278, 43603-43614 (2003). 43. Bieler, G. et al. Distinctive role of integrin-mediated adhesion in TNF-induced PKB/Akt and NF-kappaB activation and endothelial cell survival. Oncogene 26, 5722-5732 (2007). 44. Kawano, T. et al. CD1d-restricted and TCR-mediated activation of valpha14 NKT cells by glycosylceramides. Science (New York, N.Y 278, 1626-1629 (1997). 45. Godfrey, D.I. & Kronenberg, M. Going both ways: immune regulation via CD1d-dependent NKT cells. The Journal of clinical investigation 114, 1379-1388 (2004). 46. Doukas, J. & Pober, J.S. IFN-gamma enhances endothelial activation induced by tumor necrosis factor but not IL-1. J Immunol 145, 1727-1733 (1990). Algemene samenvatting & discussie 283

47. Paleolog, E.M., Aluri, G.R. & Feldmann, M. Contrasting effects of interferon gamma and interleukin 4 on responses of human vascular endothelial cells to tumour necrosis factor alpha. Cytokine 4, 470-478 (1992). 48. Strindhall, J., Lundblad, A. & Pahlsson, P. Interferon-gamma enhancement of E-selectin expression on endothelial cells is inhibited by monensin. Scandinavian journal of immunology 46, 338-343 (1997). 49. Zarbock, A. & Ley, K. Mechanisms and consequences of neutrophil interaction with the endothelium. The American journal of pathology 172, 1-7 (2008). 50. Goethals, A., Brouckaert, P. (unpublished data). (2007). 51. Hirsch, E. et al. Central role for G protein-coupled phosphoinositide 3-kinase gamma in inflammation. Science (New York, N.Y 287, 1049-1053 (2000). 52. Elewaut, D. (personal communication). 53. Hayakawa, Y. et al. Critical contribution of IFN-gamma and NK cells, but not perforin- mediated cytotoxicity, to anti-metastatic effect of alpha-galactosylceramide. European journal of immunology 31, 1720-1727 (2001). 54. Smyth, M.J. et al. Sequential production of interferon-gamma by NK1.1(+) T cells and natural killer cells is essential for the antimetastatic effect of alpha-galactosylceramide. Blood 99, 1259-1266 (2002). 55. Jones, A.L. & Selby, P. Tumour necrosis factor: clinical relevance. Cancer surveys 8, 817-836 (1989). 56. Biburger, M. & Tiegs, G. Alpha-galactosylceramide-induced liver injury in mice is mediated by TNF-alpha but independent of Kupffer cells. J Immunol 175, 1540-1550 (2005). 57. Chang, Y.J. et al. Potent immune-modulating and anticancer effects of NKT cell stimulatory glycolipids. Proceedings of the National Academy of Sciences of the United States of America 104, 10299-10304 (2007).

Chapter 13 Addendum

Addendum: The response to mTNF treatment is dependent on the tumor size 287

13. ADDENDUM: The response to mTNF treatment is dependent on the tumor size

The response to anticancer treatment may differ depending on the tumor size. To determine whether this is also the case for TNF, we pooled data of B16Bl6 tumor- bearing wild-type C57BL/6J mice treated with mTNF of 18 individual experiments, each with 5 to 7 mice per group, making a total of 116 mice (Fig.13.1.). Treatment with mTNF (7µg/day) was started when the average TSI (the product of the largest perpendicular diameters in cm) of each group was equal to or larger than 0.5 (≥ day 10 after tumor inoculation). Tumor sizes (TSI) at the beginning of treatment varied from 0.16 to 1.1, with an average of 0.56. Treatment resulted in complete tumor regression (no palpable tumor; TSI = 0) in 69 of 116 mice (59%). Partial regression, static response or progressive disease (‘other’) was observed in 39 mice (33%). 8 mice died during the ten-day treatment period without complete regression and were not further taken into account. The average TSI at the start of treatment of mice that underwent complete tumor regression was 0.59, while the average TSI of mice that did not undergo complete regression was only 0.49. This difference was statistically highly significant (p = 0.0017 in two-tailed t-test), indicating that the response to mTNF treatment is indeed dependent on the tumor size. This was even more pronounced when mice were grouped according to their initial TSI and complete response rates (number of mice that underwent complete tumor regression/total number of mice) were compared. Only 36% of mice with a TSI lower than 0.4 displayed complete response, while the complete response rate was 80% or higher for mice with a TSI higher than 0.7.

Fig. 13.1. Response to mTNF treatment is tumor size-dependent. (A) Tumor size at the start of treatment of mice that underwent complete tumor regression (CR) and mice that did not undergo complete tumor regression (other) during the ten-day treatment period with mTNF. Statistical analysis (two-tailed t-test; Graphpad Prism 4.0), ** p = 0.0017. (B) Complete response rate versus tumor size at the start of treatment.

These results clearly show that mice with larger tumors have a higher chance of undergoing complete tumor regression when treated with mTNF than mice with smaller tumors. Interestingly, no such correlation could be found in the NKT-deficient 288 Chapter 13

mouse strains J18-/-, CD1-/- and IL-15-/-. This could reflect that a critical tumor size or infarct size is required to induce an adequate immune response. It is widely accepted that TNF selectively disrupt neovasculature, while leaving the normal vasculature unharmed. The insult of disrupting the tumor neovasculature will increase rapidly with tumor size. As tumors can grow up to a size of 1-2 mm in diameter without angiogenesis, and the proportion of tumor cells depending on neovasculature instead of on pre-existing vasculature is lower in smaller tumors than in larger tumors, it can be expected that small tumors will respond less well to mTNF treatment. In addition, the conditions within the tumor (e.g. oxygen pressure, acidity, interstitial fluid pressure) may differ depending on the tumor size and influence the sensitivity of the tumor vasculature to the disruptive effects of TNF. Whatever may be the cause, the finding that the response to mTNF treatment is tumor size-dependent has important implications for the interpretation and comparison of historic data, and for setting up new experiments. Curriculum Vitae 289

CURRICULUM VITAE

Personalia

Name: Leander Marcel Maria Huyghe Birthdate and place: February 27, 1980; Bruges Address: Deinsesteenweg 214 8700 Aarsele [email protected]

Education

1992-1998: Secondary education Greek-Latin, O.L.V. College, Assebroek 1998-2000: Bachelor („Kandidaat‟) Biology, Ghent University 2000-2002: Master („Licentiaat‟) Biotechnology, Ghent University, graduated with Great Distinction Master thesis: “Sensitization for the in vivo antitumor effect of TNF by inhibition of growth factor-receptor signaling”. Promotor: Prof. Dr. Peter Brouckaert

2002-present: Doctor in Sciences: Biotechnology, Ghent University

Postgraduate training

2005-2006: Basic course in laboratory animal science (FELASA Cat. C), Ghent University

October 27-28, 2005: Effective writing for life sciences research, University of Oxford continuing education – continuing professional development

October 6, 2006: Effective presenting for life sciences research, University of Oxford continuing education – continuing professional development

International conferences and symposia

3rd ESH Interdisciplinary Euroconference on Angiogenesis; Dublin, Ireland; October 24-27, 2003. Poster presentation: “Sensitization of the tumor neovasculature to TNF in vivo”; Huyghe, L., Hostens, J., Van den Hemel, M. and Brouckaert, P.

14th Endothelial Cell Research Symposium; Amsterdam, The Netherlands; December 4, 2003. 290

Annual Meeting of the Belgian Association for Cancer Research “Genomics and proteomics in cancer”; Liege, Belgium; January 24, 2004. Oral presentation: “Sensitization of the tumor neovasculature to TNF in vivo”; Huyghe, L., Goethals, A., Hostens, J., Van den Hemel, M. and Brouckaert, P.

15th Endothelial Cell Research Symposium; Maastricht, The Netherlands; November 23, 2004

Annual meeting of the Belgian Association for Cancer Research “Impact of hypoxia in cancer treatment”; Brussels, Belgium; January 22, 2005. Poster presentation: “Inhibition of PI3K allows a 10-fold dose reduction of Tumor Necrosis Factor”; Huyghe, L., Goethals, A., Hostens, J., Van den Hemel, M. and Brouckaert, P.

International meeting IAP “Key genes in hormone-sensitive human epithelial tumors”; Brussels, Belgium; February 9-10, 2006.

3rd Focused Meeting on PI3K Signaling and Disease; Bath, UK; November 6-8, 2006. Poster presentation: “Inhibition of PI3K sensitizes the tumor neovasculature to TNF”; Huyghe, L., Goethals, A., Hostens, J. and Brouckaert, P.

Angiogenesis: fundamental and clinical perspectives; Maastricht, The Netherlands; December 15, 2006

Annual meeting of the Belgian Association for Cancer Research “Use of transgenic mice in cancer research”; Ghent, Belgium; January 20, 2007; Oral presentation: “NKT cells contribute to tumor destruction by TNF”; Huyghe, L., Goethals, A., Hostens, J., Van den Hemel, M., Elewaut, D. and Brouckaert, P.

1st International Kloster Seeon Meeting of the DFG SPP1190 "The tumor-vessel interface: Cellular and molecular mechanisms of tumor progression and metastasis"; Seeon, Germany; September 22-25, 2007. Oral presentation: “Inhibition of the PI3K- Akt pathway sensitizes the tumor neovasculature to TNF”; Huyghe, L., Sonveaux, P., Goethals, A., Deckers, W., Feron, O. and Brouckaert, P.

Awards

1st prize for the scientific presentation at the Annual Meeting of the Belgian Association for Cancer Research; Liege, Belgium; January 24, 2004

3rd prize for the scientific presentation (poster) at the Annual Meeting of the Belgian Association for Cancer Research; Brussels, Belgium; January 22, 2005 Curriculum Vitae 291

2nd prize for the scientific presentation at the Annual Meeting of the Belgian Association for Cancer Research; Ghent, Belgium; January 20, 2007

292

DANKWOORD

Na zes intense jaren, waarin ik veel meer heb geleerd dan ik voor mogelijk hield, is hij eindelijk hier: de laatste pagina van mijn thesis. En waarmee kan ik die beter invullen dan met een dankwoord voor alle mensen die op een of andere manier bijgedragen hebben tot dit werk? En dat zijn er heel wat.

In de eerste plaats wil ik Peter Brouckaert bedanken. Niet alleen omdat hij, op geheel eigen wijze, een goeie promotor en groepsleider is, maar vooral omdat hij mij overtuigd heeft om te beginnen aan een doctoraat. Toen ik volop bezig was met mijn licentiaatsthesis kwam hij met het voorstel om te doctoreren in zijn labo. De literatuur over TNF leek me toen zo overweldigend dat ik het niet zag zitten om ooit op het niveau te komen van mensen die er al jaren mee bezig waren, en spreken voor publiek was zowat mijn grootste fobie. Nee, doctoreren dat was niets voor mij. Toch kon Peter mij motiveren om eraan te beginnen. Dus Peter, bedankt! Zonder jou was ik niet de wetenschapper die ik vandaag ben.

Ten tweede wil ik alle mensen bedanken die een materiële, praktische of intellectuele wetenschappelijke bijdrage geleverd hebben aan dit werk. Dit zijn: Bart Vanhaesebroeck, Olivier Feron, Dirk Elewaut, Georges Leclercq, Pierre Sonveaux (merci beaucoup pour quelques expériences des plus sensationelles), Dieter Deforce, Pieter Dewint, Serge Van Calenbergh, Nico van Rooijen, Rudy Cocquyt, Céline Steyt (bedankt voor de goeie zorgen voor mijn B6 nudes), Katrien Van Beneden, Ann- Sophie Franki, Anje Cauwels, An Goethals, Jeroen Hostens, Maureen Van den Hemel (bedankt voor vanalles), Erica Houthuys, Wies Deckers en Jo Van Ginderachter. Zonder twijfel vergeet ik nu nog een groot aantal mensen. Soms heeft een klein praatje in de wandelgangen of op een meeting, of een emailtje mij op het idee gebracht voor een doorslaggevend experiment. Aan al deze mensen, bedankt!

Ook zonder mijn schatje Virginie was deze thesis er niet geweest. Bedankt voor jouw eindeloos geduld, jouw steun, en nog zoveel meer. En, hoewel hij dit de eerste jaren nog niet zal kunnen lezen, wil ik ook mijn spruitje Simon bedanken omdat hij af en toe net iets langer heeft moeten wachten om getroost te worden omdat ik eerst nog een stukje van deze thesis wou afwerken, en ook omdat hij een enorme motivatie is om door te zetten en succesvol te willen zijn in het leven.

Tot slot wil ik Frank, Joost, Mattias, Kurt, Joey, Tom, alle huidige en voormalige MPET collega’s (Rob, Elke, Lode, Paul, Annoucka, Stefanie, Bram, Patrick, Jennyfer, Evelien, Nathalie), collega’s van Unit 9, en uiteraard mijn ouders bedanken voor hun steun en allerlei zaken die weinig of niets met deze thesis te maken hebben, maar die steeds voor de nodige afwisseling en ontspanning wisten te zorgen tijdens deze zes jaren.