THE EFFECT OF OXYGEN TENSION ON THE CYTOTOXIC ACTION OF TUMOUR NECROSIS FACTOR-ALPHA

by

Eileen M Lynch

A thesis submitted to the University of London for the degree of Doctor of Philosophy in the faculty of Science

1996

Department of Oncology, University College London, and The Gray Laboratory Cancer Research Trust, Mount Vernon Hospital, Northwood, Middlesex ProQuest Number: 10044562

All rights reserved

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest.

ProQuest 10044562

Published by ProQuest LLC(2016). Copyright of the Dissertation is held by the Author.

All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code. Microform Edition © ProQuest LLC.

ProQuest LLC 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 ABSTRACT

The influence of tumour relevant oxygen tensions on the cytotoxic action of TNFa on tumour cells, both in vitro and in vivo, was investigated. Tumour oxygen tensions were assessed in two murine tumours (SaF and CaNT), directly, using the Eppendorf pOi histograph, and indirectly, by determining the radiobiological hypoxic fraction. These tumours are largely hypoxic with median pOi values less than 2% oxygen (15 mmHg). Such oxygen tensions re-established in vitro induce resistance to TNFa cytotoxicity in murine and human tumour cell lines.

The mechanism(s) of hypoxia-induced resistance to TNFa cytotoxicity was studied using the SaF cell line. The role of cell cycle was assessed but was found to be negligible. Western blotting and ELISA techniques were used to assess putative protective proteins such as manganese superoxide dismutase (MnSOD) and TNFa itself. MnSOD is not implicated in hypoxia-induced TNF-resistance. Endogenous TNFa levels were assessed and found to be significantly induced during hypoxic preincubation. Indeed, both pre-treatment of oxic cells with exogenous TNFa and overexpression of endogenous TNFa, by gene transfection, induced resistance to TNFa challenge.

A growth delay assay and an in vivoHn vitro clonogenic assay determined that the direct cytotoxic action of TNFa in vivo is not the only mechanism implicated in its antitumour activity. Indeed, TNFa significantly reduces tumour blood flow, as demonstrated by laser Doppler flowmetry, and this may be largely responsible for the antitumour action of TNFa in vivo.

The cytotoxic action of TNFa is dramatically modulated by oxygen tensions known to exist in tumours. Thus, overcoming tumour hypoxia could potentially increase TNFa cytotoxicity and together with the effective anti-vascular effects could perhaps improve the antitumour efficacy of TNFa. ACKNOWLEDGEMENTS

This project was undertaken at The Gray Laboratory in the Tumour Microcirculation Group, the members of which I would like to thank for sharing their scientific expertise with me over the course of my research. In particular, I would like to thank my supervisor. Dr Dai Chaplin, for his contagious enthusiasm, profound advice, and constructive criticism of my project and thesis. I am indebted to Dr Sally Hill for her help with the in vivo aspects of this project and for her efficiency in thoroughly reviewing this thesis. I would also like to thank Dr Louise Sampson for introducing me to TNF.

I would like to express my appreciation to everyone at the Gray Lab for providing a stimulating working environment and answers to so many questions. Many thanks to Dr George Wilson and Dr Horst Lohrer, for their helpful advice and for critically reviewing chapters 4 and 5, respectively. I am grateful to Dr Boris Vojnovic and Ros Locke for the design and provision of analyses systems for laser Doppler data and autoradiograph scanning. Thanks are extended to the Biological Services staff for caring for the animals used in this project and to John Draper for the photographs presented.

Many thanks to Prof. Robert L. Souhami, my supervisor at UCL, particularly for his visits to the Gray Lab. I would like to acknowledge the financial support provided by a Scott of Yews scholarship, and the Cancer Research Campaign for the mnning costs of my project.

Finally, I would like to thank my family and friends for their encouragement and moral support over the past few years.

in CONTENTS

Abstract 11 Acknowledgements iii Abbreviations xi Illustrations xiv

Chapter 1 General introduction 1 1.1 Discovery of TNF 2 1.2 Cellular sources of TNF 6 1.3 Biosynthesis of TNF in vitro 8 1.4 Biosynthesis of TNF in vivo 9 1.5 Inducers of TNF 9 1.6 Inhibitors of TNF 10 1.7 Structure of TNF 12 1.7.1 Chromosomal location 12 1.7.2 TNF oligomerization 12 1.7.3 TNF processing 13 1.8 Species specificity 13 1.9 Regulation of TNF gene expression 14 1.9.1 Transcriptional regulation 14 1.9.2 Post-transcriptional regulation 15 1.9.3 Autoinduction of TNF gene expression 15 1.10 TNF receptors 17 1.10.1 Structure of TNF-Rs 18 1.10.2 Distribution of TNF-Rs 18 1.10.3 Regulation of TNF-R expression 18 1.10.4 Interaction between TNF and receptors 18 1.10.5 Mechanism of TNF-R action 19 1.10.6 Down-modulation of TNF-Rs 20 1.10.7 Function of TNF-Rs 21 1.10.8 Soluble TNF-Rs 21 1.11 TNF signal transduction 22 1.12 TNF-induced gene expression 23

IV 1.11 TNF signal transduction 22 1.12 TNF-induced gene expression 23 1.13 TNF in cancer 23 1.13.1 Cytotoxic effects of TNF 25 1.13.2 Immune effects of TNF 26 1.13.3 Vascular effects of TNF 26 1.14 Clinical role of TNF 27 1.15 T umour oxygenation 28 1.16 Therapeutic implications of tumour oxygenation 29 1.16.1 Radiotherapy 30 1.16.2 Chemotherapy 3 0 1.16.3 Immunotherapy 30 1.17 Oxygen tension in the cytotoxic action of TNF 31 1.18 Hypothesis 31

Chapter 2 Measurement of tumour oxygenation status

in two murine tumour models 32 2.0 Introduction 33 2.1 Methods 37 2.1.1 Mice 37 2.1.2 Tumours 37 2.1.2.1 Sarcoma F 37 2.1.2.2 Carcinoma NT 37 2.1.3 Tumour transplantation 37 2.1.4 Eppendorf ™ pOi histograph 38 2.1.4.1 Principle of measurement 38 2.1.4.2 Experimental procedure 39 2.1.4.3 Analysis of results: pOi 40 2.1.5 Radiobiological hypoxic fraction 41 2.1.5.1 Principle of measurement 41 2.1.5.2 Tumour irradiation procedure 41 2.1.5.3 /« v/vo/m virra colony forming assay 42 2.1.5.4 Preparation of heavily irradiated feeder cells 42 2.1.5.5 Analysis of results: hypoxic fraction 43 2.2 Results 43 2.3 Discussion 49 2.4 Summary 50

Chapter 3 The effect of tumour relevant oxygen tensions on the cytotoxic effects of TNF in vitro 51 3.0 Introduction 52 3.1 Materials and methods 55 3.1.1 Cell lines 55 3.1.2 TNF 57 3.1.3 Actinomycin D 57 3.1.4 Oxygen conditions 57 3.1.5 Assessment of TNF cytotoxicity 58 3.1.5.1 Colourimetric assay 58 3.1.5.2 Clonogenic assay 60 3.1.6 To determine the effect of 24 h priming at tumour 60 relevant oxygen tensions on the cytotoxic action of TNF in 21% oxygen 3.1.7 To determine the effect of tumour relevant oxygen 61 tensions during TNF treatment on the cytotoxic action of TNF 3.1.8 To determine the time-course of hypoxia-induced 61 resistance to the cytotoxic action of TNF 3.1.9 To determine the time-course of re-sensitization to the 61 cytotoxic action of TNF after 24 h priming in 2% oxygen 3.1.10 To determine the effect of chronic exposure to 2% oxygen 61 on the cytotoxic action of TNF 3.2 Results 62 3.3 Discussion 74 3.4 Summary 77 Chapter 4 Mechanism of hypoxia-induced resistance

to the cytotoxic action of TNF: cell cycle 78 4.0 Introduction 80 4.1 Materials and methods 81 4.1.1 Principle of BrdUrd / PI cell cycle analysis 81 4.1.2 Experimental plan 82 4.1.2.1 Effect of priming cells in 21%, 2% and 0.2% 82 oxygen for 24 h on SaF cell growth 4.1.2.2 Effect of priming cells in 21 % and 2% oxygen 82 for 0-4 days on SaF cell growth 4.1.2.3 Effect of priming cells in 0.2%, 1%, 2%, 5% 83 and 21 % oxygen for 24 h on the distribution of SaF cells in the cell cycle 4.1.2.4 SaF cell cycle analysis DURING priming in 83 21%, 2% and 0.2% oxygen 4.1.2.5 SaF cell cycle analysis AFTER priming in 21% 84 and 2% oxygen, with and without TNF/AMD or AMD only, treatment 4.1.3 Proeedure for BrdUrd/ PI staining of cells 84 4.1.4 Flow cytometry: The FACScan 85 4.1.5 Data acquisition and analysis 85 4.1.5.1 Lysysn 85 4.1.5.2 Cellfit 85 4.1.6 Cell cycle parameter analyses 86 4.1.6.1 Labelling index 88 4.1.6.2 Duration of cell cycle 88 4.1.6.3 Duration of S phase 88 4.1.6.4 Duration of G2/M phase 88 4.1.6.5 Duration of G 1 phase 89 4.1.6.6 G1 delay 89 4.1.6.6.1 Early S unlabelled 89 4.1.6.6.2 G1 unlabelled 89 4.1.6.7 G2 delay 90 4.1.6.7.1 G2 delay of S phase cells 90 4.1.6.7.2 G2 delay of G2 cells 90 4.2 Results 4.2.1/2 Effect of priming cells in 21%, 2% and 0.2% oxygen 90 on SaF cell growth 4.2.3 Effect of priming cells in 0.2%, 1%, 2%, 5% and 21% 91 oxygen for 24 h on the distribution of SaF cells in the cell cycle 4.2.4 SaF cell cycle analysis DURING priming in 21%, 2% 93 and 0.2% oxygen 4.2.5 SaF cell cycle analysis AFTER priming in 21%and 2% 100 oxygen, with and without TNF/AMD or AMD only, treatment 4.3 Discussion 105 4.4 Summary 110

Chapter 5 Mechanism of hypoxia-induced resistance to TNF cytotoxicity: protective proteins 111 5.0 Introduction 112 5.1 Materials and methods 115 5.1.1 Cell preparation 115 5.1.2 To determine whether freely diffusible protective/ 115 sensitising factors are present in medium of cells primed in 0.2% or 21% oxygen 5.1.3 To determine the effect of protein synthesis inhibition 116 during TNF treatment on SaF response to TNF 5.1.4 To determine the effect of pre-treatment with protein 116 synthesis inhibitors or TNF during priming on subsequent response to TNF 5.1.5 To determine the effect of rmuTNF on SaF cells 116 transfected with a murine TNFa gene 5.1.6 Western blotting protocol 117 5.1.6.1 SDS-polyacrylamide gel electrophoresis 117

V lll 5.1.6.2 Gel transfer to nitrocellulose membrane 117 5.1.6.3 Antibodies and standards 118 5.1.6.4 Antibody staining of membrane 118 5.1.6.5 Autoradiography and image analysis 118 5.1.7 TNF ELISA protocol 119 5.1.8 MnSOD ELISA protocol 119 5.2 Results 120 5.2.1 Hypoxia-induced resistance to TNF cytotoxicity 120 is not mediated by a freely diffusible factor 5.2.2 Partial TNF-resistance is induced in SaF cells by 121 a non-specific protein synthesis inhibitor during TNF treatment 5.2.3 Hypoxia-induced resistance to TNF cytotoxicity 123 appears to be independent of protein synthesis 5.2.4 Hypoxia-induced resistance to TNF cytotoxicity 123 requires DNA-dependent RNA synthesis 5.2.5 Pre-treatment with TNF during priming induces 126 resistance to TNF cytotoxicity 5.2.6 Murine TNFa gene transfection induces resistance 126 to TNF cytotoxicity 5.2.7 Hypoxia induces endogenous TNF production in 127 SaF cells 5.2.8 Manganese superoxide dismutase is not induced in 136 SaF cells by hypoxia 5.3 Discussion 140 5.4 Summary 148

Chapter 6 The cytotoxic action of TNF in vivo 149 6.0 Introduction 150 6.1 Materials and methods 152 6.1.1 Tumours 152 6.1.2 TNF 152 6.1.3 To determine the cytotoxic activity of TNF in vivo 152

IX 6.1.4 To determine the effect of TNF treatment on SaF 153 tumour growth 6.1.5 To determine the effect of TNF treatment on SaF 153 tumour blood flow 6.1.5.1 Laser Doppler flowmetry : The Oxford Array^'^ 153 6.1.5.2 Experimental procedure 154 6.1.5.3 Treatment groups 155 6.1.5.4 Data acquisition 156 6.1.5.5 Data analysis 156 6.2 Results 157 6.3 Discussion 166 6.4 Summary 170

Chapter 7 Concluding discussion 171

References 175

Appendix 221 I. General materials and methods 222 n. Western blotting solutions 226 ni. Details of transfected cell lines 227 IV. List of publications 230 Abbreviations

 Angstroms ACTH adrenocorticotropic hormone ADP adenosine diphosphate AIDS acquired immunodeficiency syndrome AMD actinomycin D AP-1 activator protein-1 ARDS adult respiratory distress syndrome ATP adenosine triphosphate BCG Bacille Calmette-Guérin BrdUrd 5-bromo-2 ’ -deoxyuridine BSA bovine serum albumin cAMP cyclic adenosine monophosphate CaNT carcinoma NT cDNA complementary deoxyribonucleic acid CFE colony forming efficiency e x cytotoxicity assay CXM cycloheximide Da Daltons DMSO dimethylsulphoxide DNA deoxyribonucleic acid DNAse deoxyribonuclease DTE dithiothreitol EBV Epstein Barr virus ECG Electrocardiogram ECU enhanced chemiluminescence EDTA ethylenediaminetetraacetic acid EGE epidermal growth factor EGR-1 early growth response gene-1 EIA enzyme immunoassay ELAM-1 endothelial leukocyte adhesion molecule-1 ELISA enzyme linked immunosorbant assay EMEM Eagle’s minimal essential medium enTNF endogenous TNF ES early S phase exTNF exogenous TNF FACScan fluorescence activated cell sorter FCS foetal calf serum fg fentogram Erre fluorescein isothiocyanate Fsa-N fibrosarcoma-N G gauge GO quiescent or resting phase of the cell cycle G1 first gap phase of the cell cycle G2 second gap phase of the cell cycle GM-CSF granulocyte-macrophage colony stimulating factor GMD geometric mean diameter h hour HF hypoxic fraction HIV human immunodeficiency virus

XI HLA-B human leukocyte antigen-B HN haemorrhagic necrosis hsp heat shock protein Ik B inhibitory factor kappa B i.v. intravenous ICAM-1 intercellular cell adhesion molecule-1 ICE interleukin-1 p converting enzyme IFN interferon Ig immunoglobulin IL interleukin ILP isolated limb perfusion IRE-1 interferon regulatory factor-1 IS international standard lU international units kb kilobase kDa kiloDalton kg kilogram kV kilovolt L litre LAK lymphokine activated killer LI labelling index log logarithmic LPL lipoprotein lipase EPS lipopolysaccharide M (phase) Mitosis M Molar yi{n) marker mA milliamps mg milligram MHC major histocompatibility complex min minute ml millilitre mM millimolar MMP matrix metalloproteinase MnSOD manganese superoxide dismutase mRNA messenger ribonucleic acid MT metallothionein MW molecular weight NF-IL6 nuclear factor-interleukin-6 NF-k B nuclear factor kappa-B ng nanogram NGS normal goat serum NIBSC National Institute of Biological Standards NK natural killer nm nanometer NO nitric oxide OD optical density OPD O-paraphenyldiamine hydrochloride ORPs oxygen regulated proteins PAI-1 platelet activating inhibitor-1 PBS phosphate buffered saline PC-PLC phosphatidylcholine-specific phospholipase C Pg picogram PGE2 prostaglandin F 2 PI propidium iodide PKA protein kinase A PKC protein kinase C PLA^ phospholipase A^ PMNCs polymorphonuclear cells PMSF phenylmethylsulphonyl fluoride pÛ2 partial pressure of oxygen poly A polyadenylation A R region rhuTNF recombinant human TNFa RIA radioimmunoassay rmu-TNF recombinant murine TNFa RNA ribonucleic acid ROS reactive oxygen species rpm revolutions per minute S S phase of the cell cycle S phase DNA synthesis phase s.c subcutaneous s.d standard deviation s.e.m standard error of the mean SaF sarcoma F SDS sodium dodecyl sulphate Ser/Thr serine/threonine SF surviving fraction SLF systemic lupus erythematosus SOBR sum of broadened rectangles SOD superoxide dismutase sTNF-RI soluble TNF receptor I (55 kDa) sTNF-RII soluble TNF receptor II (75 kDa) SV40 Simian virus 40 Svecs SV-40 transformed endothelial cells TBF tumour blood flow TBS tris buffered saline Tc length of cell cycle TCF TNF-converting enzyme TFMFD (N,N,N’, N’, -tetra-methyl-ethylene diamine) Tgi duration of G1 Tg 2/M duration of G2/ M phase TNF tumour necrosis factor TNF-R TNF receptor TRAF TNF receptor associated protein Tris Tris(hydroxymethyl) aminomethane Ts duration of S phase uPA urokinase type plasminogen activator VCAM-1 vascular cell adhesion molecule-1

Xlll Illustrations

Figures

Chapter 2 2.1 The dependence of radiosensitivity on oxygen concentration 35 2.2 The Eppendorr"^ pOi histograph 38 2.3 Experimental set-up for pOi measurement 40 2.4 pOi distribution in SaF dorsal and foot tumours 45 2.5 pO] distribution in CaNT dorsal and foot tumours 46 2.6 Survival curves of SaF dorsal and foot tumours 47 2.7 Survival curves of CaNT dorsal and foot tumours 48

Chapter 3 3.1 The cytotoxic effect of TNF on SaF cells after priming in 2% and 21% 63 oxygen for 24 h, determined by colourimetric and clonogenic assays 3.2 Amount of rhu- and rmu-TNF required to cause 50% SaF cell kill 65 after 24 h priming at tumour relevant oxygen tensions prior to TNF treatment in 21% oxygen 3.3 The effect of 24 h priming at tumour oxygen tensions on the cytotoxic 68 effects of TNF on MCF-7 and Fsa-N cells 3.4 The effect of treating SaF cells with TNF at tumour relevant oxygen 70 tensions after 24 h priming in 21% oxygen 3.5 Time-course of hypoxia-induced resistance to the cytotoxic effects of 71 TNF on SaF cells 3.6 Time-course of resensitization of SaF cells to TNF after 24 h priming 72 in 2% oxygen 3.7 The effect of chronic exposure to 2% oxygen on the response of SaF 73 cells to the cytotoxic action of TNF

Chapter 4 4.1 Bivariate dot-plot and corresponding histograms of SaF cells, showing 87 BrdUrd uptake and DNA content 4.2 Distribution of SaF cells in the cell cycle during priming in 21%, 94 2% and 0.2% oxygen 4.3 The movement of SaF cells, during priming in 0.2% and 21 % 97 oxygen, through a narrow region in mid S phase 4.4 The movement of SaF cells through mid S phase during priming in 98 21% and 2% oxygen; and in 21% oxygen, after 24 h priming, with and without AMD or TNF/AMD treatment 4.5 SaF cells during priming in 0.2% and 21% oxygen: showing the 101 movement of early S phase unlabelled cells into S phase and the % G1 labelled cells 4.6 SaF cells during priming in 0.2% and 21% oxygen showing: the 102 exit of G2 unlabelled cells into mitotic phase and the exit of unlabelled cells from G1 into S phase 4.7 Proliferation of SaF cells, after priming in 21% and 2% oxygen, 103 treated with TNF/AMD, AMD only, or medium 4.8 Labelling index of SaF cells during priming in 21% and 2% 104 oxygen for 24 h; and in 21% oxygen, after priming in 2% and 21% oxygen, with and without AMD or TNF/AMD treatment

Chapter 5 5.1 Effect of CXM, in the presence or absence of AMD, on TNF 122 cytotoxicity of SaF cells 5.2 Effect of CXM pre-treatment of SaF cells, during priming in 124 21% or 2% oxygen for 24 h, on the cytotoxic action of TNF 5.3 Effect of AMD pre-treatment of SaF cells, during priming in 125 0.2% and 21% oxygen for 24 h, on the cytotoxic action of TNF 5.4 Effect of rhu-TNFa pre-treatment of SaF cells, during priming in 128 0.2% and 21% oxygen for 24 h, on the cytotoxic action of TNF 5.5 Effect of murine TNFa gene transfection on SaF cell response to 129 rmuTNFa in 21% oxygen 5.6(a) Autoradiograph of Western blot showing TNF expression of SaF 130 cells during 24 h priming in 21% and 0.2% oxygen 5.6 (b) Effect of priming SaF cells in 0.2% or 21% oxygen for 24 h on 130

XV the relative induction of TNF, determined by Visilog analysis 5.7(a) Autoradiograph of Western blot showing TNF expression in SaF 132 cells during priming in 21% and 0.2% oxygen for 24 h and after priming, in 21% oxygen, for 24 h 5.7(b) The effect of priming SaF cells in 21% or 0.2% oxygen for 24 h 133 on the relative induction of TNF, assessed by Visilog analysis 5.8 Endogenous TNF production of SaF cells during priming in 21% 135 and 0.2% oxygen for 24 h, and after priming, in 21% oxygen, for 24 h, assessed by ELISA 5.9(a) Autoradiograph of Western blot showing MnSOD expression in 137 SaF cells during priming in 21% and 0.2% oxygen for 24 h, and after priming, in 21% oxygen, for 24 h 5.9(b) Visilog analysis of autoradiograph shown in Figure 5. 9 (a) 137 5.10 Endogenous MnSOD levels of SaF cells during priming in 21% 139 and 0.2% oxygen for 24 h, and after priming, in 21% oxygen, for 24 h, assessed by ELISA

Chapter 6 6.1 The Oxford Array^'^ laser Doppler system 154 6.2 Experimental set-up for measurement of tumour blood flow 155 6.3 Effect of rhu-TNF on SaF tumour cell survival 18 h after treatment 158 6.4 Growth delay of SaF rear dorsal tumours after a single i.v. dose of 159 rhu-TNF 6.5 Effect of 0.5% BS A/saline (TNF diluent) on blood flow of skin and 161 tumour of SaF-bearing mice 6.6 Effect of rhu- and rmu-TNF on blood flow of SaF dorsal tumours 163 6.7 Effect of rhuTNF on SaF tumour blood flow compared to skin 164 blood flow 6.8 Effect of TNF on blood flow of SaF dorsal tumours compared to skin 165

XVI Tables Chapter 1 1.1 Characteristics of TNFa and TNpp 4 1.2 Biological activities of TNF 5 1.3 Sources of TNF 7 1.4 Detection of TNF in serum of humans and mice 8 1.5 Inducers of TNF gene expression 10 1.6 Inhibitors of TNF gene expression 11 1.7 Regulation of gene expression by TNF 16 1.8 Biological activities of TNF receptors 17 1.9 Differential regulation of TNF-R55 and TNF-R75 19 1.10 Overview of TNF-activated signalling molecules and mediators 24

Chapter 2 2.1 Summary of pOi and hypoxic fraction data 44

Chapter 3 3.1 Factors that induce sensitivity/resistance to the cytotoxic action of TNF 53 3.2 Cell number seeded for TNF cytotoxicity assays 58 3.3 TNF concentrations used to treat cell lines 59 3.4 Amount of TNF, in the presence of AMD, required to kill 50% 66 of cells after 24 h priming at tumour relevant oxygen tensions 3.5 Amount of TNF required to kill 50% of Fsa-N cells after 24 h 67 priming in 21 % or 2% oxygen

Chapter 4 4.1 PI determination of the proportion of SaF cells in each phase of 92 the cell cycle after 24 h priming in tumour oxygen tensions 4.2 BrdUrd/PI determination of the proportion of SaF cells in each 92 phase of the cell cycle after 24 h priming in tumour oxygen tensions 4.3 Labelling index of SaF cells during priming in 0.2%, 2% and 21% 95 oxygen, and after priming in 2% and 21% oxygen, with and without AMD or TNF/AMD treatment 4.4 Summary of SaF cell cycle results 96 Chapter 1 General introduction

1.1 Discovery of TNF 1.2 Cellular sources of TNF 1.3 Biosynthesis of TNF in vitro 1.4 Biosynthesis of TNF in vivo 1.5 Inducers of TNF 1.6 Inhibitors of TNF 1.7 Structure of TNF 1.7.1 Chromosomal location 1.7.2 TNF oligomerization 1.7.3 TNF processing 1.8 Species specificity 1.9 Regulation of TNF gene expression 1.9.1 Transcriptional regulation 1.9.2 Post-transcriptional regulation 1.9.3 Autoinduction of TNF gene expression 1.10 TNF receptors 1.10.1 Structure of TNF-Rs 1.10.2 Distribution of TNF-Rs 1.10.3 Regulation of TNF-R expression 1.10.4 Interaction between TNF and receptors 1.10.5 Mechanism of TNF-R action 1.10.6 Down-modulation of TNF-Rs 1.10.7 Function of TNF-Rs 1.10.8 Soluble TNF-Rs 1.11 TNF signal transduction 1.12 TNF-induced gene expression 1.13 TNF in cancer 1.13.1 Cytotoxic effects of TNF 1.13.2 Immune effects of TNF 1.13.3 Vascular effects of TNF 1.14 Clinical role of TNF 1.15 Tumour oxygenation 1.16 Therapeutic implications of tumour oxygenation 1.16.1 Radiotherapy 1.16.2 Chemotherapy 1.16.3 Immunotherapy 1.17 Oxygen tension in the cytotoxic action of TNF 1.18 Hypothesis 1.1 Discovery of TNF In the late nineteenth century Dr. William Coley, a surgeon in New York, read a case report of a patient that had contracted erysipelas infection and subsequently showed regression of an inoperable sarcoma of the neck. Coley realised that the tumour regression was related to the infection and thus attempted to mimic the situation by infecting advanced cancer patients with Streptococcus pyrogenes, to induce erysipelas. However, infecting patients with Streptococci, and later, Serratia marcescens, which increased the level of infection, often led to severe and uncontrolled toxic infections which were sometimes fatal. He tried to control the level of infection by sterilising the bacterial mixtures and this became known as Coley's toxins. Coley developed fifteen different mixtures of toxins; some of these resulted in complete tumour cures and others showed dramatic albeit short-lived tumour regression (Coley, 1893). However, the clinical results were inconsistent and eventually this method of treating advanced cancer patients had to be abandoned. From these studies it was postulated that infectious agents or their products might somehow combat cancer (Coley-Nauts, 1953). Coley's toxins were the primary systemic form of cancer therapy preceding radiotherapy and chemotherapy. For decades there was a search for an agent that was induced by infection that could cause tumour regression. This included the discovery of lipopolysaccharide (LPS), the major component of endotoxin, in 1943 (Shear et al 1943). Endotoxin had anti­ tumour activity in vivo and could also induce infection but it did not affect tumour cells in vitro, implying that the antitumour activity of endotoxin was mediated by the host. In 1962 O'Malley and co-workers reported that serum from mice suffering from endotoxic shock could initiate necrosis of a tumour transplanted from another animal (O'Malley, 1962). Thus, apparently endotoxic shock induced the production of some factor that was released into the systemic circulation that had the ability to cause tumour necrosis. In 1975 Currie and Bosham demonstrated that the toxicity of macrophages to different tumour cell types was due to a soluble factor released by activated macrophages (Currie and Bosham, 1975). Seventy-five years after the discovery of Coley's toxins a factor distinct from bacterial endotoxin was discovered that induced the haemorrhagic necrosis and partial or complete regression of certain transplantable murine tumours (Carswell et al 1975). The serum activity of this factor mimicked that of endotoxin, but was a host factor rather than residual endotoxin, they called this factor “tumour necrosis factor” or “TNF”. TNF was originally described as “a soluble factor found in sera from mice that were sequentially treated with a reticulo-endothelial stimulator {Mycobacterium bovis, strain BCG) and bacterial endotoxin or LPS” (Gifford and Flick, 1987). TNF was isolated from rabbit and human sera (Matthews, 1978; 1981; Green et al. 1976) and in 1984 the TNF gene was cloned (Pennica et al. 1984). Soon after, the proteins amino acid sequence was identified and recombinant TNF produced (Aggarwal et al. 1985; Pennica et al. 1985; Shirai et al. 1985). TNF was rediscovered in 1980 by Rouzer and Cerami as a result of the search for a cachexia-inducing factor (Rouzer and Cerami, 1980). Whilst studying the biochemical mechanism of cachexia in rabbits infected with Trypanosoma brucei they noticed that the animals suffered severe weight loss that was greater than could be due to the parasite burden alone. They realised that the parasite induced some host factor which was directly responsible for the cachexia. Five years later Beutler and co-workers purified the product from a macrophage cell line and called it cachectin (Beutler et al. 1985a). The amino-terminal sequence of the 17 kDa murine protein was determined and was found to be almost identical to human TNF cDNA (Beutler et al. 1985b). Cachectin and TNF shared biological activities and subsequent amino acid sequencing of cachectin and cloning of the cachectin gene confirmed that it was identical to TNF (Fransen et al. 1985, Pennica et al. 1985).

The TNF family consists of three members: TNFa, (cachectin) which is produced by activation of a variety of cells, and the lymphotoxins (LT), LT-a (or TNFp) and LT-p, which are produced by activation of lymphoid cells only. TNFa and TNFP are genetically, structurally, and functionally related cytokines (Aggarwal, 1992) but differ in their signal transduction and in their biological responses (Oster et al. 1987). Two TNF-Rs, TNF-RI (or p55 (55 kDa)) and TNF-RQ (or p75 (75 kDa)), (Hohmann et al. 1989; Brockhaus et al. 1990) are shared with similar affinities, by TNFa and TNFp (Loetscher et al. 1990b) but not LT-p. LT-p has a specific receptor (LTpR, also known as TNFR-related protein or TNFR-III) (Crowe et al. 1994). The primary ligand for the LT-pR is a heteromeric complex formed by LT-a and LT-P on the cell surface (Browning et al. 1993). The characteristics of TNFa and TNFp are summarised in Table 1.1. TNF has many other effects apart from its antitumour activity and it is associated with both physiological and pathological conditions, as can be seen from Table 1.2. TNFa, hereafter called TNF, was studied for the purpose of this thesis. Table 1.1 Characteristics of TNFa and TNFp

Property TNFa TNFp

Gene Chromosomal location 6 6

Number of introns 3 3

Gene size (kb) 3 3

Protein Mature protein amino acids 157 residues 171 residues (mouse: 156 residues) (mouse: 169 residues) Signal sequence amino acids 76 residues 34 residues

Cysteine residues 2 0 (mouse: 2) (mouse: 1) Methionine residues 0 3

Carbohydrates N on-gly coprotein N-and 0- Glycosylation

Quaternary structure Trimer Trimer

Molecular weight 17 kDa monomer 18 kDa monomer 52 kDa trimer 55 kDa trimer

Isoelectric point 5.2-6.8 5.8 - 8.3

pH stability 4.8 -10.5 -

Temperature stability 37°C for 40 h -

Crystallisation Trigonal (3.5 A) Hexagonal (1.9 Â)

Except where indicated data are from experiments with human TNF

(Adapted from Van Ostade et al 1994; Aggarwal et al 1987) Table 1.2 Biological activities of TNF

Cell type Effect Reference Adipocytes Suppression of lipogenic enzymes Torti et al. 1985 Ded ifferen tiati on Torti et al. 1985 Adrenal cells Modulation of steroidogenesis Natarajan et al. 1989 Inhibition of ACTEl-induced gene expression Jààttelà er < 3/. 1991 B lymphocytes Proliferation and differentiation Jelinek and Lipsky, 1987 Augmentation of antibody production Becker era/. 1990 Endometrial cells Production of IL-6 Tabibzadeh er a/. 1989 Endothelial cells Surface expression of ICAM-1 Haskard et <3/. 1986 Surface expression of ELAM-1 Bevilacqua et al. 1987 Surface expression of VCAM-1 Osbom et al. 1989a Surface expression of MHC antigens Collins era/. 1986 Induction of procoagulant activity Bevilacqua et al. 1986 Production of IL-1 Nawroth et al. 1986a Production of GM-CSE Munk&r et al. 1986 Growth inhibition Frater-Schroder et al. 1987 Eosinophils Increased toxicity to pathogens Silberstein and David, 1986 Eibroblasts Growth stimulation Sugarman et al. 1985 Production of collagenase and PGE 2 Dayer et al. 1985 Inhibition of collagen synthesis Solis-Herruzo er a/. 1988 Expression of IFN-P, R eiser a/. 1989 Surface expression of EGF receptors Palombella er <3/. 1987 Surface expression of MHC antigens Collins era/. 1986 Induction of membrane bound IL-I Le et al. 1987 Production of IFN-P 2 Kohaseera/. 1986 Production of GM-CSF Munker er a/. 1986 Expression of oncogenes Lin and Vilcek, 1987 Granulosa cells Modulation of steroidogenesis Emoto and Baird, 1988 Hepatocytes Expression of acute phase proteins Perlmutter et al. 1986 Monocytes Increased cytotoxicity Philip and Epstein, 1986 Chemotaxis Ming et al. 1987 Muscle cells Glycogenolysis Lee et al. 1987a Depolarisation of plasma membrane Tracey et al. 1986a Neutrophils Increased toxicity to pathogens Shalaby er a/. 1985 Chemotaxis Mingera/. 1987 Osteoblasts Growth inhibition Yoshihara et al. 1990 Inhibition of collagen synthesis Bertolini er a/. 1986 Osteoclasts Activation Bertolini er < 3/. 1986 Pancreatic p cells Enhancement of IL-1 cytotoxicity Mandrup-Poulsen et al. 1987 Cytotoxicity Rabinovitch et al. 1990 Pituitary cells Production of pituitary hormones Milenkovic et al. 1989 Synovial cells Production of collagenase and PGE 2 D ayerera/. 1985 T lymphocytes Surface expression of IL-2 receptor Lee era/. 1987b Surface expression of MHC antigens Scheurich et al. 1987 Testicular cells Modulation of steroidogenesis Warren era/. 1989 Tumour cells Cytotoxicity Carswell er a/. 1975 Growth inhibition Sugarman er a/. 1985 Induction of type I PAI-I and u-PA Georg et al. 1989 Virus-infected cells Inhibition of virus replication Wong and Goeddel, 1986 Cytotoxicity Wong and Goeddel, 1986 Reactivation of infection Osbom et al. 1989b

(Adapted from Jââttelà, 1991) Seven years after pure recombinant TNF became available it was entered into clinical trials (reviewed by Jones and Selby, 1989; Fraker et al. 1995). Since then, it was concluded that TNF has a potent antitumour activity that can be exploited in locoregional treatment of human cancer patients such as isolated limb perfusion. Indeed, such treatment regimes result in 90% complete responses in metastatic malignant melanoma (Liénard et al. 1992a), the treatment of ascites (Rath et al. 1991) and other forms of local treatment (Taguchi and Sohmura, 1991). However, the induction of a systemic inflammatory response syndrome prevents the extension of these studies to treatments involving the administration of TNF in the general circulation where TNF proves to be very toxic causing side effects such as severe hepatotoxicity and hypotension. TNF toxicity is dose limiting and thus therapeutic effects are seldom observed (reviewed by Jones and Selby, 1989; Fraker et al. 1995). However, the shock inducing and antitumour activities of TNF can be partly segregated (Barbara et al. 1994; Vandenabeele et al. 1995b). Thus, it is anticipated that ways will be found to inhibit the shock-inducing effects in a selective manner.

1.2 Cellular sources of TNF The principal source of TNF in vivo is from cells derived from monocyte/macrophage lineage (Matthews et al. 1981a; Aggarwal et al. 1985). TNF is also produced by many other cell types (c/Table 1.3) and is found in blood of healthy adults at a concentration of 0-6.3 pg/ml (Stallmach et al. 1995). Most cells capable of producing TNF need to be activated by some type of stimulus to trigger its production. TNF production from peripheral blood monocytes and macrophages from different tissues is differentially regulated (Martinet fl/. 1988; Strieter < 2/. 1989). Table 1.3 Sources of TNF

Source Reference Monocytes Matthews, 1981 Lymphocytes Cuturi et al. 1987 T- and B-cells Sung et al. 1988a, b Tumour cells Spriggs et al. 1987 Natural killer cells Peters et al. 1986 Mast cells Young et al. 1987 Microglial cells Chiang and McBride, 1991 Kupffer cells Kutteh et al. 1991a Alveolar macrophages Streiter et al. 1989

(Adapted from Aggarwal, 1985)

The TNF gene is expressed constitutively in normal cells only by lymphocytes of the thymus (Giroir et al. 1992). Passive immunisation of neonatal mice against TNF leads to thymic involution (Grau et al. 1990), hence it was postulated that TNF plays a role in the development or regulation of immune response. Indeed, TNF has been detected in various disease states, as noted in Table 1.4. Many human lymphoblastoid cell lines and leukaemic B cells produce TNF constitutively (Sung et al. 1988b; Foa et al. 1992). Certain non-haematopoietic tumour cell lines resistant to TNF were shown to produce TNF constitutively (Rubin et al. 1986; Spriggs et al. 1987; Krônke et al. 1988). TNF is also produced constitutively in human thyroid epithelial cells of patients with autoimmune thyroiditis (Zheng et al. 1992) and increased levels of TNF have also been detected in patients with AIDS (Hess et al. 1991). Significantly elevated levels of TNF are found in amniotic fluid when infection (chorioamnionitis) or intrauterine growth retardation occurs although the fetal and maternal TNF levels remain unchanged

(Stallmach er <2/. 1995). Table 1.4 Detection of TNF in serum of humans and mice

Species Disease/ Detection TNFa Reference Status method (ng/ml) Human Normal adult ELISA 0-0.0063 Stallmach er a/. 1995 Pregnancy ELISA 0-0.055 Stallmach gr a/. 1995 AIDS RIA 0.06-0.28 Lahdevirta gr a/. 1988 Meningitis CX^ ND^-35 WsLSigQ et al. 1987 Tuberculosis RIA ND-0.24 Barnes et al. 1990 Leprosy ELISA ND-70 Fisa et al. 1990 Endotoxemia ELISA ND-0.7 Michie et al. 1988 Malaria RIA ND-2.8 Grau et al. 1989 Leishmaniasis ELISA ND-85 Fisa et al. 1990 Parasitemia ELISA ND-0.8 Scuderi et al. 1986 Cancer ELISA ND-40 Balkwill gf aZ. 1987a Colitis ' ELISA ND-0.15 MacDonald et al. 1990 Murine model Malaria CX 0.1-7.0 Clark and Chaudhri, 1988 Endotoxemia cx 8-10 Zuckerman et al. 1989 Cachexia CX 1.2-22.8 Oiiffetal. 1987 CX= cytotoxicity assay ; ND = nondetectable; ^ = plasma (Adapted from Jaattela, 1991 and Stallmach et al. 1995)

1.3 Biosynthesis of TNF in vitro LPS is the most potent inducer of TNF both in vitro and in vivo (Carswell et al. 1975; Matthews, 1981b). When cells and tissues are stimulated with LPS, TNF gene transcription is increased approximately 3-fold, TNF mRNA is immobilised for translation and levels increase by a factor of 50-100. Secretion of the bioactive peptide occurs minutes after stimulation and rises by a factor of about 10,000 (Beutler et al. 1985c and 1986). A brief exposure to LPS is not sufficient to induce significant TNF production, it must be present during the entire induction and secretion process (Gifford and Lohmann, 1986). Resident macrophages contain a pool of TNF mRNA that is not expressed as protein (Beutler et al. 1986) but once activated TNF can constitute 1-5% of the protein secreted (Beutler et al. 1985a). Macrophages release their TNF immediately after synthesis and do not keep TNF protein stores (Gifford and Flick, 1987). After TNF biosynthesis, the TNF gene product is down-regulated and macrophages become refractory to LPS and remain so for 12-24 hours, after which they become responsive (Cockfield gr a/. 1993).

1.4 Biosynthesis of TNF in vivo TNF does not exist in stored form but is synthesised following cell activation. The biosynthesis of TNF is tightly regulated by transcriptional and largely post- transcriptional mechanisms (Beutler et al 1986; Han et al 1990). In vivo, TNF appears almost immediately on LPS injection of mice and reaches a maximum from about 1.5-2 hours and then disappears rapidly, by 4-6 hours it is undetectable (Satomi et a l 1981). Mice become unresponsive to repeated LPS treatment and do not produce TNF again for at least one week (Gifford and Flick, 1987). Exogenous LPS in vivo causes TNF mRNA to rapidly accumulate in most tissues whilst TNF receptor mRNA levels are found in lung, spleen, lymph nodes and lymphocytes. In most cases TNF receptor mRNA levels do not parallel TNF mRNA levels (de Kossodo et al 1994) Regulation of gene transcription is multifactorial and a number of factors that either increase or decrease TNF production have been identified (reviewed by Pauli, 1994).

1.5 Inducers of TNF TNF production can be induced by a variety of diverse stimuli in vivo and in vitro as shown in Table 1.5. Different stimuli induce TNF to varying extents in different cell lines. The mechanism(s) by which these stimuli induce TNF expression has not been fully elucidated but appears to differ in cells of myeloid origin compared to cells of lymphoid origin. As previously mentioned, LPS is the best known inducer of TNF both in vivo and in vitro. It may be advantageous to induce TNF and its anti tumour activity in certain circumstances however it would be necessary to find non-toxic inducers of TNF that did not interfere with its activity. Inducing large quantities of TNF selectively in the tumour microenvironment may cause tumour regression without host toxicity, if the TNF acts locally. However, it has been shown that induced TNF protein is toxic (Tracey et a l 1986b). Therefore, regulating macrophages in situ for the endogenous production of TNF may be more beneficial as macrophages kill tumours by several mechanisms in addition to TNF activity (reviewed by Bonta and Ben-Efraim, 1993). Table 1.5 Inducers of TNF gene expression

Inducer Reference

LPS Carswell er a/. 1975 IFN-y Dannis^ra/. 1991 TNFa Spriggs et al. 1990 IL-1 Philip and Epstein, 1986 IL-ip Bethea 1992 IL-2 Kasid^ra/. 1990 IgE Borish a/. 1991 X-rays Hallahan et al. 1989 Mycobacterial proteins Valone a/. 1988 Cyanobacterium Nakano^tfl/. 1991 Mycoplasma membranes SherefaA 1990 Zymosan Sanguedolce et al. 1992 Malarial parasite antigens Bate gt aZ. 1988 Tumour cell proteins De Marco et al. 1992 Sendai virus Aderka er a/. 1986 Epstein-Barr virus Gosselin et al. 1992 Herpes simplex virus Berent gt aZ. 1986

(Adapted from Sidhu and Bollon, 1993)

1.6 Inhibitors of TNF There seem to be at least four distinct types of mechanism leading to TNF suppression: (a) inhibition of translation or secretion of TNF protein (b) inhibition of TNF mRNA transcription by increasing intracellular cAMP concentration (c) inhibition of transactivation of the TNF gene by alteration of the predominant composition of NF- kB complex (d) inhibition of activation of transcriptional factor NF-kB (Lieberman et al. 1992; Pauli, 1994).

10 There are a variety of agents that down-regulate TNF production (c/Table 1.6) including cytokines, antioxidants, glucocorticoids, steroids, and immunosuppressants. TNF production can be completely blocked by compounds that inhibit matrix metalloproteinases (MMP s) (Gearing et al. 1995). Prostaglandins inhibit TNF production in a dose-dependent manner (Dayer et al 1985; Kunkel et al 1986; Bachwich et al 1986), mediating regulation of TNF expression largely at a posttranscriptional level, probably via cAMP. cAMP analogs decrease synthesis of TNF almost certainly by inhibiting TNF transcription via a cAMP-dependent protein kinase (Taffet et al 1989). cAMP inhibits TNF production at both the mRNA and protein level (Spengler et al 1989).

Table 1.6 Inhibitors of TNF gene expression

Inhibitor Reference IL-4 Essnereta/. 1989 IL-6 Aderka eta/. 1989 IL-10 Ralph a/. 1992 PGEi Kunkel era/. 1988 Histamine Vannier eta/. 1991 Dexamethasone Beutler era/. 1986 Cortisol Kutteh era/. 1991b Danazol Mori era/. 1990 Ambroxol Bianchi era/. 1990 Pentoxifylline Waage era/. 1990 Thalidomide Sampaio era/. 1991 Lipoxygenase inhibitors Schade era/. 1989 Cyclosporin-A Remick era/. 1989 sTNF-Rl Seckinger er a/. 1988 STNF-R2 Olsson era/. 1989 Nicotinamide Heine era/. 1995

(Adapted from Sidhu and Bollon, 1993)

11 Synthesis of TNF is inhibited by actinomycin D (AMD) or cycloheximide (CXM), indicating that it is an inducible protein (Gifford and Flick, 1987). Changes in the stability of TNF mRNA have been reported when cells are treated with protein synthesis inhibitors or with phosphatase inhibitors (Zuckerman et al 1991; Chung et al 1992; Sung et al 1992; Tebo and Hamilton, 1994). It may be advantageous to inhibit TNF activity in circumstances where its presence can be detrimental to the host including cachexia, rheumatoid arthritis, systemic lupus erythematosus, and insulin- dependent diabetes mellitus.

1.7 Structure of TNF 1.7.1 Chromosomal location The genes encoding human TNFa and TNFP are approximately 3 kilobases in size, arranged tandemly, separated by 1.1 kilobases on the short arm of chromosome 6 (Nedwin et al 1985). The human TNFa gene is located within the class m locus of the major histocompatibility complex (MHC), proximal to the centromere (Spies et al 1989). It is linked to genes encoding for MHC class I antigens (210 kb from the HLA-B locus) on one side and to genes encoding for the major heat shock protein hsp 70, complement components and cytochrome p450 c21 on the other side (Jaattela, 1991). The murine TNF genes are found on chromosome 17 in mice (Aggarwal et al 1985) and are also located within the MHC complex, approximately 70 kb proximal to the H2- D locus (Muller et al 1987). The location of TNF genes in the MHC region suggests that TNF may play a role in some functions of the immune system.

1.7.2 TNF oligomerization The existence of TNF in a monomeric state is not favoured. It has been shown by cross-linking that biologically active TNF protein consists of a homotrimeric complex made up of three 17 kDa molecules (Smith and Baglioni, 1987; Wingfield et a l 1987). The three subunits associate noncovalently to form a compact, conical trimer. Oligomerization of TNF appears to depend on protein concentration as it can be dissociated into monomers by lowering the pH or by the presence of chaotropic agents. Thus, both hydrophobic and electrostatic interactions are responsible for trimer formation (Smith and Baglioni, 1987). The TNF trimer binds eight times more avidly to the TNF receptor than the 17 kDa monomer (Jaattela, 1991). The base of the trimer interacts with the receptor (Eck and Sprang, 1989).

12 1.7.3 TNF processing The exact molecular weight of TNF is 52,068 Da (Piers et al. 1987). Human TNF is synthesised as a 233 amino acid (26 kDa) type II cell-surface associated molecule. It possesses bioactivity and may function as a cytotoxic transmembrane protein by either cell-cell interaction or local release of the secretory non-glycosylated 17 kDa component (Kriegler et al. 1988; Luettig et al. 1989) consisting of 157 amino acid residues (Pennica et al. 1984). The 17 kDa monomer forms an elongated, anti­ parallel p-pleated sheet sandwich and has a 'jelly-roll' topology (Eck and Sprang, 1989). The processing of TNF precursor is unusual. It has been suggested that soluble, active TNF is released from the cell membrane-anchored precursor by proteolytic cleavage (Van Ostade et al. 1994). Several highly accessible protease sites are exposed near the bottom of the TNF molecule, where the supposed active site is located (Jones et al. 1991). The protease(s) that cleaves TNF has not been identified but is thought to be a matrix metalloproteinase (MMP). Purified matrix metalloproteinases (MMP s) can cleave a recombinant pro-TNF substrate to yield mature TNF (Gearing et al. 1995). MMP s are produced in an inactive form that can be activated by other enzymes including serine proteases. Thus, it is possible that in some situations serine proteases could contribute to pro-TNF processing by activating a key MMP enzyme. An enzyme that partially cleaves a TNF peptide substrate has been partially purified (Mohler et al. 1994) but the identity of TNF- converting enzyme (TCE) is still unknown. There may be one or more TNF-converting enzymes, possibly membrane bound, that process pro-TNF. Interleukin-ip converting enzyme (ICE) or other ICE-like proteases have recently been implicated in TNF mediated cell death (Tewari and Dixit, 1995; Enari et al. 1995). However, other data suggest that an MMP-like enzyme is directly involved in TNF processing (Gearing et al. 1995).

1.8 Species specificity TNF is almost non-species specific (Fransen et al. 1986), almost certainly because the TNF amino acid sequence is conserved between species. Human and mouse TNF show 76% homology and there is greater than 70% homology between TNF from humans and rabbits (Ito et al. 1986; Marmenout et al. 1985). It has been observed that antibodies against TNF from one species do not cross-react with TNF from another species despite the sequence similarities and the ability of TNF to bind to the TNF

13 receptor of other species (Fiers et al. 1987). However, the most sequence variable region of the TNF molecule corresponds to the antibody accessible surface loops (Jones et al. 1992). Murine TNF has a three-fold higher specific activity than human TNF and is also more toxic in mice (Brouckaert et al. 1992).

1.9 Regulation of TNF gene expression Regulation of TNF gene expression is multifactorial and is tightly regulated in a tissue specific manner. Transcription, translation, and secretion, all appear to be controlled in multiple-step processes (reviewd by Pauli, 1994). Such controlled regulation is critical for the expression of a gene product whose biological activity can result in both beneficial and harmful effects in vivo. The 5' flanking region of the TNF gene contains multiple potential regulatory sites. However, the functional importance and interactions of these regulatory elements remain undefined. The 3' untranslated region contains a sequence element affecting posttranslational control of TNF through mRNA stability and translation efficiency

1.9.1 Transcriptional regulation TNF transcription is thought to be controlled by short-lived repressor molecules (Collart et al. 1986). The basal level of TNF gene transcription is increased approximately 3-fold by LPS (Beutler et al. 1986) which induces protein kinase C (PKC) (Kovacs et al. 1988) and the binding of various DNA-binding proteins to enhancer elements in the regulatory region of the TNF gene (Akira et al. 1990; Shakov et al. 1990). LPS-induced TNF transcription in macrophages requires a Y-box (Dorn et al. 1987) and at least two kappa-B-like enhancer elements (Sen & Baltimore, 1986; Shakov et al. 1990), perhaps even four, (Collart et al. 1990) and constitutive as well as inducible forms of NFk-B (Sariban et al. 1988; Stone et al. 1988; Takasuka et al. 1995). To date, the only protein known to participate in the signal transduction mechanism of LPS-induced TNF gene expression is CD 14 (Takasuka et al. 1995). Other factors thought to be involved in the transcriptional regulation of TNF include calcium- dependent processes (Drysdale et al. 1983), arachidonic acid metabolites (Kunkel et al. 1989), cAMP (Spengler et al. 1989), and MAP kinase/ pp90^^^/ EGR-1 (Saleem et al. 1995).

14 1.9.2 Post-transcriptional regulation TNF gene expression is complex and is regulated only in part by transcriptional events. Many of the observed regulatory mechanisms function post-transcriptionally. The translation rate per unit TNF mRNA increases 200-fold after LPS-activation and this enhancement is linked to the 3'-untranslated region (Han et al 1990). This uridine and adenosine rich 3'-untranslated region contains a 33-nucleotide sequence composed of repeated and overlapping copies of the consensus sequence UUAUUUAU (Caput et al 1986). It is presumed that this sequence controls both the half life and the translation rate of mRNAs (Caput et al 1986; Shaw and Kamen, 1986), thus preventing inappropriate persistent production of a large amount of biologically active polypeptides. However, translational de-repression is necessary for TNF protein production after LPS-stimulation. A ribonuclease may play a crucial role in controlling the biosynthesis of TNF in macrophages and other mammalian cells by selectively degrading mRNAs at a rate depending on the number of AU sequences (Beutler et al 1988). It has been shown that degradation of murine TNF mRNA proceeds from 3’ to 5’ end in a two-step process: (1) removal of the poly A tail and (2) rapid degradation of the remaining mRNA (Lieberman et al 1992). It seems that inhibition of PKC enhances the rate of mRNA degradation by facilitating the removal of the poly A tail of the mRNA.

1.9.3 Autoinduction of TNF gene expression Exogenous TNF induces production of endogenous TNF, this has been established for monocytes, myeloid leukaemia cells, and epithelial tumour cells in vitro (Philip and Epstein, 1986; Spriggs et al 1987; Hensel et al 1987). Arachidonic acid metabolites and phospholipase A^ activation is thought to be involved for transcriptional induction of the TNF gene by TNF (Spriggs et al 1990). The TNF gene contains a TNF- responsive element which when stimulated with TNF produces an 8-10 fold activation of the human TNF promoter (Leitman et al 1991). The autoinduction of TNF transcripts is rapid and transient and is one of the earliest detectable changes in gene expression associated with exposure to TNF and has been associated with a development of resistance to its cytolytic effects (Niitsu et al 1988). In addition to TNF autoinduction, TNF regulates the expression of a wide range of genes (reviewed by Neta et a l 1992), a nonexhaustive list is given in Table 1.7.

15 Table 1.7 Regulation of gene expression by TNF

Gene Effect Transcription Reference factor Acute phase proteins Stimulation ND^ Perlmutter et al. 1986 GM-CSF Stimulation ND Mxivk&ï et al. 1986 c-fos Stimulation ND Lin and Vilcek,1987 c-myc Stimulation ND Lin and Vilcek, 1987 c-myc Suppression ND Krdnke gr a/. 1987 IL-la and p Stimulation ND Le et al. 1987 LPL Suppression ND Cornelius et al. 1988 al collagen Suppression ND Solis-Herruzo et al. 1988 Histone 3 Suppression ND Schütze et al. 1988 HIV Stimulation NF-KB-like Osborn a/. 1989b IFN-pl Stimulation IRF-1 Vuÿtidi et al. 1989 Collagenase Stimulation AP-1 Brenner a/. 1989 c-jun Stimulation AP-1 Brenner a/. 1989 Type 1 PAI Stimulation ND Georg et al. 1989 IL-2 receptor a Stimulation NF-KB-like Lowenthal et al. 1989 u-PA Stimulation ND Gœrg et al. 1989 IL-6 Stimulation NF-KB-like Zhang et al. 1990 MHC class I Stimulation NF-KB-like Pessara and Koch, 1990 Stimulation AP-1 Johnson and Pober, 1990 MHC class n Stimulation NF-KB-like Pessara and Koch, 1990 MAD6/A20 Stimulation ND Opipari et al. 1990 P450 oxidase Suppression ND Jââttelâ et al. 1991b acyl-CoA synthase Suppression ND Weiner gr aZ. 1991 Endothelin 1 Stimulation ND Marsden and Brenner, 1992 EGF receptor Stimulation ND AdsLchi et al. 1992 NO synthetase Suppression ND Yoshizumi et al. 1993 ^ND, not detected (Adapted from Jaattela, 1991a; Neta et al. 1992)

16 In brief, the mechanisms by which TNF synthesis is regulated at the transcriptional, translational and post-translational levels are slowly unravelling, as is the necessity of keeping tight control over TNF synthesis.

1.10 TNF receptors Almost all cell types interact with TNF through a single class of specific high- afflnity receptors (Aggarwal et ai 1985; Smith and Baglioni, 1992). Two different TNF receptors with molecular masses of 55-60 kDa and 75-80 kDa, referred to as TNF-R55 (or, p55, TNF-Rl, TNF-Rp or CD 120a) and TNF-R75 (or, p75, TNF-R2, TNF-Ra, or CD 120b) have been identified (Hohmann et al 1989; Brockhaus et a l 1990). It is not clear whether multiple biologic functions of TNF observed in a cell line are mediated through one receptor or both receptors. The functions assigned to both receptors are summarised in Table 1.8.

Table 1.8 Biological activities of TNF receptors

Activity TNF- TNF- Reference R55 R75

Cellular cytotoxicity Yes Heller gr aZ. 1992 Cellular cytotoxicity Yes Tartaglia gr a/. 1991 Proliferation of lymphocytes Yes Yes Tartaglia and Goeddel, 1992a Proliferation of primary Yes Yes Tartaglia and Goeddel, 1992b thymocytes Gene induction in endothelial cells Yes Barbara gf a/. 1994 Induced cytokine production Yes Yes Vandenabeele aZ. 1995b

NF-k B activation Yes Yes Vandenabeele aZ. 1995b Apoptosis Yes Yes Vandenabeele er a/. 1995a

17 1.10.1 Structure of TNF-Rs TNF-R55 and TNF-R75 consist of an extracellular, transmembrane and intracellular domain and share approximately 25% overall amino acid sequence identity. There is homology between the extracellular domain of both receptors but no homology between the intracellular domains. The extracellular part of the TNF-Rs is not required for signalling (Bazzoni et al 1995) which might explain why the receptors transduce signals via different pathways (Vandenabeele et al 1995b).

1.10.2 Distribution of TNF-Rs Receptors for TNF are widely distributed throughout most cells and tissues. The TNF-R55 is more prevalent on epithelial cells, whereas the TNF-R75 is more abundant on myeloid cells (Hohmann et al 1989). The highest amounts are found in liver, kidneys, and spleen, while modest amounts are found in stomach, duodenum, small bowel, pancreas, colon, and oesophagus. Red blood cells do not contain TNF-Rs and thus do not contribute to TNF binding in whole tissues (Munker and Koeffler, 1987). The number of receptors per cell ranges from 100 to 10,000 copies depending on cell type and activation status.

1.10.3 Regulation of TNF-R expression Murine TNF-R55 receptor expression is constitutive and is directed by a non­ inducible house-keeping-type promoter (Fuchs et al 1992; Rothe et al 1993). The wide variation in receptor number between different cell lines may be largely accounted for by the up- and down-regulation of TNF-R75 by various agents, (c/Table 1.9). Receptor expression does not always correlate with cellular response (Hass et al 1985; Chan and Aggarwal, 1994; Pandita et al 1992), this relationship varies depending on the nature of the TNF-dependent response and also on the cell line. However, in general, cells expressing TNF-R55 do not display TNF activities unless the receptor is triggered. In contrast to TNF-R55, TNF-R75 signalling seems to be highly dependent on receptor density (Bigda, 1994; Vandenabeele et al 1995b)

1.10.4 Interaction between TNF and its receptors TNF receptors are activated by clustering but it is still not clear if this accounts for all TNF-induced activities (Van Ostade et al 1994). Three receptor molecules bind symmetrically to one lymphotoxin trimer (Banner et al 1993). Four cysteine-rich

18 domains of the TNF-Rs are involved in optimal TNF binding (Hsu et al. 1993). Also, the N terminal amino acids are critical for both receptor binding and biological activity of TNF (Tchorzewski et al 1993). The half time of dissociation differs from 180 minutes for TNF-R55, which has a Kd value of 500 pM to 10 minutes for TNF-R75, which has a Kd of 100 pM (Fiers, 1991 and 1995). Oligomerization of receptors is often followed by receptor phosphorylation. TNF receptor-associated proteins (TRAP-1 and TRAP-2) have been cloned (Song et al. 1994). However, the involvement of associated proteins in signal transduction remains to be proven. Once TNF has bound to TNF-R55 the complex is rapidly internalised by clathrin coated pits (Porteu and Hieblot, 1994) and is degraded in lysosomes (Mosselmans et al 1988). TNF-R75 lacks a consensus sequence for rapid cellular internalisation through coated pits (Collawn, 1990) and probably mediates signalling via shed soluble TNF-R75. Murine TNF-R75 is species specific in that it binds murine TNF but not human TNF (Lewis et al 1991). In humans both murine and human TNF can trigger both the TNF-R75 and TNF-R55 receptor types. The receptors induce distinct intracellular signals but can have co-operative effects.

Table 1.9 Differential regulation of TNF-R55 and TNF-R75

Cell type Stimulus TNF-R55 TNF-R75 Reference

HL-60 cells cAMP Unaffected Upregulated Hohmann et al. 1990a B cells Mitogens Unaffected Upregulated Erikstein et al. 1991 T cells Mitogens Unaffected Upregulated Dembic et al. 1990 Macrophages LPS Unaffected Upregulated Tannenbaum et al. 1993 CD56^ NK cells IL-2 Unaffected Upregulated Naumeefa/. 1991

1.10.5 Mechanism of TNF-R action The large differences in the cytoplasmic domains of the two receptor types are in favour of a non-redundant signalling model for TNF-R (Lewis et al. 1991). The TNF-Rs do not have intrinsic kinase activity, suggesting that the associated molecules might

19 mediate signal transduction. Two novel putative signal transducers, called TRAFl and TRAF3 were cloned and characterised. As both TNF-Rs share identical TRAF members this might allow intracellular crosstalk between their signalling pathways (reviewed by Vandenabeele gr aZ. 1995b; Yim etal. 1995). Tartaglia and Goeddel originally proposed that TNF receptors operated according to a ligand passing model. They suggested that TNF first binds to TNF-R75, which has a higher affinity and dissociation rate than TNF-R55, resulting in a high local concentration of TNF at the cell surface. TNF then dissociates from the TNF-R75 and binds to the TNF-R55, which mediates intracellular signalling (Tartaglia et al 1993). Recently Leeuwenberg and co-workers have shown that in cells responding to TNF via the TNF-R55, the extracellular part of TNF-R75 captures TNF and delivers it to TNF-R55, resulting in an enhanced response to TNF (Leeuwenberg et al 1995). TNF-R75 preferentially binds TNF at low concentrations, thus, its primary function is to facilitate association between TNF and TNF-R55. Whether this passing is simply a result of a TNF-R75-dependent increase in local ligand concentration or whether the mechanism requires the formation of a transient heterocomplex is unknown. However, the ratios between the number of TNF-R75 and TNF-R55 molecules seem to be important in the process of ligand passing (Tartaglia et al 1993; Leeuwenberg et al 1995). It is likely that TNF-R75 plays a more important role in its soluble form where it is involved in the inactivation of TNF in the circulation (Bemelmans et a l 1993).

1.10.6 Down-modulation of TNF-Rs Down regulation of TNF receptors may occur through several mechanisms (Winzen et al 1992; Totpal et al 1995; Richter et al 1995). Indeed, TNF down- modulates its own receptors rapidly even in the presence of protein synthesis inhibitors (Tsujimoto and Vilcek, 1987; Watanabe et al 1988a). Thus, the process of receptor do wn-modulation is induced and does not involve changes in the synthesis of proteins but rather in their activity. The mechanisms appear to be transient as prolonged exposure to TNF causes cells to regain their ability to bind TNF (Winzen et al 1993). LPS, PKC and TNF cause a rapid down regulation of surface TNF-R expression on monocytes (Scheurich et al 1989) which in the case of TNF stimulation is associated with a subsequent increase in sTNF-R levels (Lantz et al 1990; Olsson et a l 1993).

20 1.10.7 Functions of TNF-Rs TNF-R55 and TNF-R75 are independently regulated and mediate distinct signalling pathways but cell type determines the final cellular response (Hohmann et al 1989; Erikstein et al 1991). The different roles of both receptors on TNF signal transduction, together with the differential regulation of receptor expression may provide multiple control of the biologic activities of TNF. Most of the activities of TNF seem to be carried out by the TNF-R55 (Thoma et al 1990). TNF-R75 seems to be involved in the late long-term effects of TNF, whereas the more rapid effects of TNF appear to be induced by TNF-R55. Signalling through both receptors seems to be necessary for the biologic actions of TNF and expression of TNF-R55 may be necessary but not sufficient to induce cytotoxicity (Higuchi and Aggarwal, 1993). The mechanisms by which TNF induces its cytotoxic action in cells expressing both receptors may be different from that in cells expressing TNF-R55 only. TNF-R55 can perform all of the functions attributed to TNF-R75, usually at much lower receptor density (Vandenabeele et a l 1994). TNF-R55 might monopolise TNF-mediated signalling whenever it is expressed in sufficient quantities, and TNF-R75 signalling may not lead to any unique biological activities that are different from those exerted by TNF-R55. Thus, TNF-R75 seems to fulfil mainly an accessory role in enhancing or synergizing TNF-R55 effects (reviewed by Vandenabeele et al 1995a,b).

1.10.8 Soluble TNF-Rs The genes encoding two naturally occurring inhibitors of TNF, the soluble TNF- RI (sTNF-Rl (55 kDa)) and sTNF-RII (75 kDa) have been cloned (Loetscher et al 1990b; Schall et al 1990; Kohno et al 1990). Soluble, secreted forms of TNF receptors consisting of the extracellular domain of TNF receptors were isolated as TNF binding proteins and can bind TNF with high affinity (Seckinger et al 1988; Olsson et al 1989; Engelman et al 1989). Soluble TNF-Rs have been found in urine (Olsson et al 1989) and serum (Schall et al 1990) of humans and identified in a number of tumour cell lines. An increase in TNF receptors on the cell surface may lead to an increase in the sTNF-Rs. sTNF-Rs inactivate TNF by the formation of high affinity complexes thereby reducing the binding of TNF to target cell membrane receptors and down-regulating response to TNF. In addition, TNF can be stabilised in complex with sTNF-Rs, and under certain conditions the complex may act as a slow releaser and consequently increase the long-term effects of TNF (reviewed by Olsson et a l 1993).

21 1.11 TNF signal transduction. When TNF binds to its receptors there is a series of events triggered which causes a signal to be transduced to the nucleus. However, the nature of these post­ receptor events is not well characterised. TNF is not directly involved in intracellular signalling as once it induces receptor oligimerization it itself is then internalised and degraded intracellularly (Pennica et al 1992). Aggregation of the intracellular domains of the TNF receptors appears to be important for signal transduction. TNF receptors are probably coupled to G-proteins (Branellac et al 1992) that regulate the activation of

PLA2 (Silk et al 1989) and synthesis of cycloxygenase resulting in the production of arachidonic acid and lysophosphatidylcholine (Godfrey et a l 1987; Suffys et al 1988b) TNF also activates phosphatidylcholine-specific phospholipase C (PC-PLC) which results in diacylglycerol production (Wiegmann et al 1994). Diacylglycerol, arachidonic acid and lysophosphatidylcholine are all activators of PKC. PKC might mediate some of the actions of TNF as shown by its translocation from the cytosol to the cell membrane (Schutze et al 1990). However, protein kinase inhibitors affect TNF-mediated functions (Utsumi et al 1992; Hohmann et al 1992). TNF induces superoxide production in human neutrophils (Richter et a l 1995) and this is increased by inhibitors of PKC and inhibitors of tyrosine kinase (Utsumi et a l 1992). TNF does not increase the level of cAMP (Scholz and Altman, 1989) but cAMP- dependent protein kinase A (PKA) may be involved in TNF signal transduction (Zhang et a l 1988; Baud era/. 1988). Sphingomyelin may be involved in TNF signal transduction (Kim et a l 1991; Candela et al 1991; Wiegmann et al 1994). TNF hydrolyses sphingomyelin which forms ceramide. Also, TNF induced kinase activation can be mimicked by the addition of sphingomyelinase but not by the addition of phospholipase A 2, C or D (Dressier et al 1992). Activation of transcription factors by TNF induced protein kinases may be important in TNF signal transduction from cell membrane to nucleus. TNF stimulates NF-kB, (Lowenthal et al 1989; Duh et al 1989) by causing a rapid activation of pre­ existing NF-kB and by inducing transcription of the NF-kB gene (Hohmann et al 1990a, b; Meyer et al 1991). The activation of NF-kB by TNF is very interesting as NF- kB is a pleiotropic transcription factor that controls the expression of a variety of cellular genes (Lenardo and Baltimore, 1989; Gilmore, 1990). NF-kB is the only transcription factor that is directly activated by TNF signal transduction, thus it may be considered as a third messenger of the TNF signal transduction cascade (Vandenabeele

22 et al. 1995b). TNF activation of NF-kB does not involve either PKC or PKA (Meichle et al 1990; Hohmann et al 1991). Other, transcription factors that are activated at the transcriptional level include AP-1 (Brenner et al 1989), IRF-1 (Fujita et al 1989), NF- GMa (Shannon et al 1990) and EGR-1 (Gao et al 1992). These in turn induce a secondary set of genes (Stephens et al 1992) not primarily induced by TNF. Table 1.10 gives an overview of TNF-activated signalling intermediates and transcription factors.

1.12 TNF-induced gene expression Although the cellular actions of TNF that mediate the diverse physiological responses and contribute to the pathogenesis of diseases are largely unknown, a major site where TNF probably acts is at the level of gene transcription. Torti et a l (1985) first reported that TNF inhibits the transcription of adipocyte genes. Since then, TNF has been shown to alter the transcription of numerous genes (c/Table 1.7) that certainly contribute to its physiological and pathological effects. TNF increases the transcription of proto-oncogenes, ferritin heavy chain, MHC antigens and cytokines (reviewed by Kronke et al 1992). The mechanism(s) of the gene regulatory effects of TNF have not been extensively examined but it is thought that TNF may regulate gene expression by activating a variety of DNA binding proteins.

1.13 TNF in cancer Several mechanisms may be involved in the anti-tumour effect of TNF: the direct cytotoxic effect, activation of the host immune system and effects on the tumour vasculature. TNF is cytotoxic to some murine and human melanoma cell lines (Helson et al 1975; Haranaka and Satomi, 1981; Sugarman et al 1985) and causes regression of murine tumours, human tumour xenografts and isolated localised tumours in vivo (Carswell et al 1975; Haranaka et al 1984; Creasey et al 1986; Balkwill et al 1986; Lejeune, 1995). Although the direct cytotoxic effect and indirect effects on stromal elements play a role in tumour response, the relative importance of each action may vary between tumour types.

23 Table 1.10 Overview of TNF-activated signalling molecules and mediators

Activation of phospholipases Neutral sphingomyelinase Wiegmann gr a/. 1994 Acidic sphingomyelinase Wiegmann gr a/. 1994 Phosphatidylcholine-specific PLC Wiegmann et al 1994 Phosphatidylinositol-specific PLC Beyaeit^ra/. 1993 PL A], release of arachidonic acid Suffys et al 1991 Phospholipase D De Valck et al 1993

Activation of transcription factors AP-1 Brenner et al 1989 NF-k B Osborn et al 1989b NF-jun Brach et al 1993 Proteolytic degradation of Ik B Henkel er aZ. 1993

Kinases and phosphatases Phosphorylation of hsp27 Guesdon et al 1993 Phosphorylation of cytosolic PLAi Hoeck et al 1993 Phosphorylation of a 28-kDa cap-binding Marino et al 1989 protein Activation of MAP kinase vanhint et al 1992 Activation of a ceramide activated kinase Wiegmann gr a/. 1994 Activation of PKC Wiegmann et al 1994 Inactivation of phosphatases Guy et al 1993

Oxygen radicals Perturbation of the electron-transport Schulze-Osthoff er a/. 1993 chain in mitochondria Free radical production Goossens er a/. 1995

Other signalling molecules Increase in intracellular Ca^^ concentration Bellomo er fl/. 1992 Increase in ADP ribosylation Heine era/. 1995 Proteolytic activation Ruggiero et al 1987a Activation of ICE or ICE-like proteases Tewari era/. 1995 Production of nitric oxide Hauschildt et al 1992 Activation of pertussis-toxin-sensitive G-protein Branellac et al 1992

(Adapted from Vandenabeele et al 1995b)

24 1.13.1 Cytotoxic effects of TNF Initially it was thought that TNF was selectively toxic for tumour cells (Ruff and Gifford, 1981; Sugarman et al 1985; Fransen et al 1986). However, TNF is toxic to normal cells such as endothelial cells, adipocytes, kératinocytes, fibroblasts, smooth muscle cells and haemopoietic progenitor cells (Schuger et al 1989; Kawakami et al 1989; Pillai et al 1989; Palombella and Vilcek, 1989; Sato et al 1986; Broxmeyer et al 1986). Although TNF is not selective in its toxicity there is some degree of selectivity when tumour cells and normal cells from the same tissue are tested, tumour cells showing greater TNF-sensitivity (Dealtry et al 1987; Dollbaum et al 1988). However, the toxicity of TNF to normal cells, particularly endothelial cells, may be an essential component of the antitumour effects seen in vivo (Rothstein and Schreiber, 1992; Malik and Balkwill, 1992). Tumour cells show a wide variation in their response to TNF; it can have growth stimulatory, growth inhibitory, cytotoxic or no effect (Lewis et al 1987; reviewed by Sidhu and Bollon, 1993). Some cell lines are sensitive to TNF cytotoxicity in vitro but not in vivo (Helson et al 1975) and vice versa (Manda et al 1987). This suggests that either the effect of TNF is related to the biological characteristics of different cell populations or that TNF has different modes of action. Response to TNF’s cytotoxic action is not determined by histological origin, tumour type, receptor number, receptor affinity or rate of internalisation (Tsujimoto et al 1985; Sugarman et a l 1985; Dollbaum et al 1988). However, TNF response can vary according to stage of cell growth i.e., degree of confluence, differentiation and rate of proliferation (Kirstein et al 1986; Gerlach et al 1989; Kawakami et al 1989; van de Wiel et al 1992). The effects of TNF are cell cycle specific in some tumours and cell lines (Darynkiewicz et al 1984; Watanabe et al 1987; Belizario and Dinarello, 1991; Pusztai et al 1993). Cell communication can be an important determinant in the response to TNF. Confluent cultures, compact colonies and gap junctions are associated with resistance to TNF cytotoxicity. By contrast subconfluent cultures show increased susceptibility to TNF cytotoxicity (Fletcher et al 1987; Matthews and Neale, 1989; Patek et a l 1989). Competition between induced protective factors and the destructive capacity of TNF might determine whether a response is cytotoxic. Considering in vitro studies, it is likely that several mechanisms are involved in TNF cytotoxicity that may manifest themselves under various conditions in different cells. However, the mechanisms of TNF cytotoxicity have not been fully elucidated. In

25 vitro it can manifest as either apoptosis or necrosis depending on the type of target cell, culture conditions and the presence of metabolic inhibitors (Laster et al. 1988; Grooten et al. 1993).

1.13.2 Immune effects of TNF In addition to direct antiproliferative effects on tumour cells, TNF is a mediator and activator of other cellular anti tumour mechanisms such as tumouricidal macrophages (Talmadge et al. 1987; Urban et al. 1986; Hori et al. 1987), cytotoxic T cells (Nakano et al. 1989), natural killer cells (0stensen et al. 1987), natural cytotoxic cell activity (Degliantoni et al. 1985; Ortaldo et al. 1986), and neutrophils (Shau 1988; Shalaby et al. 1985). TNF can affect the interaction of the host and tumour by effects on host stromal cell populations and induce the release of other cytokines (e.g., EL-1 and IL-6) that may have anti tumour effects. TNF may also stimulate production of tumour specific antibodies (Balkwill, 1989). The precise role for these TNF-activated cellular antitumour mechanisms in vivo, apart from the role of T-cell mediated immunity, is largely unknown. The importance of T-cell mediated immunity in the anti tumour effects of TNF in murine tumour models has been noted by several groups (Havell et al. 1988; Palladino et al. 1987b; Haranaka et al. 1984). Tumour immunogenicity is an important factor in determining the anti tumour effect of TNF (Asher et al. 1987). Administration of TNF to mice bearing immunogenic fibrosarcomas induces haemorrhagic necrosis and tumour regression, but significant tumour regression is not seen in mice bearing poorly immunogenic tumours. Although some degree of necrosis is seen in both immunogenic and non-immunogenic tumours, a peritumoural cellular infiltrate is only seen in the immunogenic tumours. Thus the lack of this infiltrate in non-immunogenic tumours almost certainly leads to subsequent regrowth at the periphery (Asher et al. 1987).

1.13.3 Vascular effects of TNF The development of vascularised stroma is essential for the growth of all solid tumours (Folkman, 1990) and for the formation of metastasis (Blood and Zetter, 1990). Folkman (1971) hypothesised that tumours cannot grow beyond 1-2 mm in diameter without developing a blood supply which is vital to supply oxygen and nutrients and to remove waste products of metabolism. Vascular effects of TNF have been noted in

26 many tumour models (Havell et al 1988; Proietti et al 1988; Kallinowski et al 1989a,b; Madhevan et al 1990; Naredi et al 1993; de Kossodo et al 1995). TNF damages the tumour vasculature by a direct cytotoxic effect and indirectly by activation of host cells, with simultaneous induction of intravascular thrombosis and fibrin deposition (Naworth et al 1988; Shimomura et al 1988). TNF induces an early influx of red cells into the tumour (Havell et al 1988) and high TNF doses lead to growth inhibition and reduction of tumour blood flow (Kallinowski et a l 1989a). Inducing thrombus formation in the tumour vasculature via TNF has been suggested as one means of exploiting the procoagulant activity of tumours. TNF has been shown to alter numerous endothelial responses in vitro. It downregulates thrombomodulin and causes endothelial cells to release tissue factor, platelet-activating factor and von Willebrand factor, which favour thrombosis. In addition it causes cell adhesion molecules to be either upregulated or expressed de novo on the endothelial cell surface (Gamble et al 1992). This increases the adhesion of leukocytes to endothelial cells, through increased expression of ICAM-1 (Dustin and Springer, 1989), VCAM-1 and ELAM-1 (Pober et al 1986; Pober et al 1987; Pober and Cotran 1990). It has been proposed that the production of tumour-derived soluble proteins may enhance the procoagulant action of TNF on tumour endothelium and contribute to the selective damage of tumour vasculature (Nawroth et al 1988; Clauss et al 1992). Overall, these effects in excess may foster the development of disseminated intravascular coagulation and vascular occlusion.

1.14 Clinical role of TNF TNF alone and in combination has been used as an anti-tumour agent. A broad range of inhibitory activity has been shown for recombinant TNF in vitro against human colon carcinoma, breast carcinoma, cervical carcinoma, melanoma, and myeloid leukaemia cell lines (Helson et al 1975; Ruggiero et al 1987b; Watanabe et a l 1985; Munker and Koeffler, 1987a; Schiller et al 1987). A variety of regimes of TNF have been tested in cancer patients in phase VU clinical trials (reviewed by Jones and Selby, 1989 and by Fraker et a l 1995). However, no consistent anti-tumour activity was observed at tolerated doses. Some of the major side effects of TNF encountered during its systemic administration in cancer patients include nephrotoxicity, hepatotoxicity, pulmonary toxicity and neurotoxicity. Other TNF side effects include hypotension, fever, rigors.

27 acute dyspnea, cyanosis with hypoxia, and myalgia (Saks and Rosenblum, 1992). Peripheral blood alterations induced by TNF include lymphocytopenia and neutrophilia. The neutrophilia is due to recruitment of new cells from the bone marrow and demargination of cells from blood vessel walls. The lymphocytopenia is due to upregulation of adhesion molecules on endothelial cells that bind the lymphocytes and could also be due to direct toxicity to lymphocytes (Remick and Kunkel, 1986; 1993). There are potential approaches to improving the therapeutic application of TNF in cancer treatment. One approach that has already elicited tumour response is to limit normal tissue exposure so that higher TNF doses can be administered. The isolated limb perfusion technique has produced dramatic tumour response (Liénard et al 1992a, b and Lejeune et a l 1995). However, such use is restricted to a select group of tumours. The utility of systemic TNF delivery would be much improved if ways could be found to enhance its activity selectively in tumours. It has already been shown that TNF cytotoxicity can be potentiated by several diverse agents. IFN-y enhances TNF cytotoxicity in vitro (Lewis et al 1987) and sensitises some resistant cell lines to the cytotoxic action of TNF (Brouckaert et al 1986). Similar results have been found in vivo (Balkwill et al 1987b) although systemic toxicity may also be increased (Talmadge et al 1987). The cytotoxicity of TNF to some malignant cell lines is increased when used in combination with cytokines such as IL-1 (Ruggiero et al 1987b) and IL-2 (Me Intosh, 1988) and with certain cytotoxic drugs including topoisomerase II inhibitors (Alexander et al 1987; Das et al 1989) and AMD (Fiers et al 1986). Other approaches to improving therapies may be identified if mechanisms of TNF resistance could be elucidated. One potential modulator of TNF action is the oxygenation status of malignant tissue.

1.15 Tumour oxygenation Tumour oxygenation is governed by the supply of oxygen and its rate of consumption. Oxygen availability in turn is determined by the arterial oxygen supply and nutritive blood flow. The inadequate and heterogeneous microvasculature characteristic of malignant growth can affect tumour microenvironment parameters, such as tumour oxygenation, nutrient supply, bioenergetic status and pH distribution. This can lead to a deterioration of the diffusive transport of oxygen to the tumour cells and microareas of hypoxia and anoxia develop. The development of hypoxic, nutrient depleted environments within the tumour slows tumour growth and allows tumour

28 survival against conventional therapies, for example, non-proliferating hypoxic cells become resistant to cell-cycle specific chemotherapeutic agents. Thomlinson and Gray originally proposed a model of how “diffusion limited” hypoxia occurs (Thomlinson and Gray, 1955). They suggested that tumour cells which were more than 180 |im away from a fully functional blood vessel would be hypoxic or even anoxic as oxygen would be utilised by cells close to blood vessels resulting in a gradual decline in the oxygenation and nutritional status of cells distant from the vessel. However, this is not the only cause of hypoxic clonogenic cells in tumours. There is indirect and direct evidence for “perfusion limited” hypoxia (Reinhold et al 1977; Yamaura and Matsuzawa, 1979; Brown, 1979; Chaplin and Trotter, 1991). Recent studies have shown that blood vessels within tumours can be subjected to temporary cessation in erythrocyte flux and/or blood flow which could render endothelium hypoxic or even anoxic (Chaplin and Acker, 1987; Trotter et a l 1991; Chaplin & Hill, 1995). Hypoxic regions may also occur if erythrocytes are deoxygenated (Mueller-Kliesser et 1980; Vaupel, 1994). Dynamic microcirculation subjects regions of the tumour to altered microenvironment such as hypoxia, acidosis and nutrient deprivation (Vaupel et al 1981; Kallinowski et al 1989a; Vaupel, 1994). Hypoxia is a feature of both experimental and human tumours. Regional pOz values of 15 mmHg (~2%02) appear to be common in human tumours but not in normal tissues. Values <2.5 mmHg have been detected in primary human tumours, a level of oxygenation that is consistent with radiobiological hypoxia (Vaupel et al 1994; Hockel et al 1993).

1.16 Therapeutic implications of tumour oxygenation There are several therapeutic implications of heterogeneous tumour oxygenation. Hypoxia can increase tumour metastatic potential (Young et al 1988) and has recently been shown to be associated with a more malignant phenotype (Stackpole et al 1994). The compromised oxygenation status existing in many regions of solid tumours can affect tumour response to radiation, chemotherapy and biological approaches to cancer therapy.

29 1.16.1 Radiotherapy The oxygen effect is well established in radiobiology (Crabtree and Cramer, 1933; Mottram, 1936; Thomlinson and Gray, 1955). Oxygen levels below 10 mmHg (1.3% oxygen) cause cells to become resistant to radiation (Hall, 1988) and radioresistant cells are able to repopulate the tumour after radiotherapy (Denekamp et al. 1983). Tumour oxygenation has been shown to be a prognostic factor for radiation treatment of advanced cancer of the uterine cervix (Hockel et al 1993).

1.16.2 Chemotherapy Heterogeneous tumour blood flow can affect chemotherapeutic drug delivery. Hypoxia can cause resistance to certain chemotherapeutic drugs (Roizen-Towle and Hall, 1978; Smith et al 1980; reviewed by Teicher, 1994). It has been shown that hypoxia causes resistance to antiproliferative chemotherapy via effects on metabolism and transit through the cell cycle (Teicher, 1994).

1.16.3 Immunotherapy Coley’s toxins was one of the first biological approaches to cancer therapy (Coley, 1893; Coley, 1906). However, it is only recently that the influence of oxygen tension in biological approaches to cancer treatment has been investigated. Loeffler and co-workers showed that oxygen tensions known to exist in solid tumours, i.e. <7.6mmHg (1% oxygen), significantly reduced lymphocyte proliferation in response to IL-2. This suggested that inhibition of IL-2 stimulated lymphocyte proliferation by reduced oxygen tension, may be a factor in the relatively poor success rate of IL-2/LAK cell immunotherapy (Loeffler et al 1992). This theory was supported by the fact that IL-2 induction of LAK cell activity is strongly suppressed under oxygen- limited conditions (5% and 2% oxygen) which pertain physiologically compared to those at 20% oxygen (Ishizaka et al 1992). Thus, oxygen tension has been shown to modulate tumour response to radiotherapy, chemotherapy, immunotherapy. It is known to influence cellular production and action of several cytokines and growth factors (Ishizaka et al 1992; Wing et al 1988; Shaul et al 1992). However, despite the pivotal role TNF plays in tumour biology, the influence of tumour microenvironment conditions, such as oxygen tension, on its pleiotropic activities has not been elucidated. Such information is needed

30 to understand the in vivo effects of TNF within solid tumours as well as its potential role in cancer therapy and could identify ways to improve its activity in malignant tissue.

1.17 Oxygen tension in the cytotoxic action of TNF Matthews et al. originally proposed that a free radical-induced process was the mechanism responsible for TNF cytotoxicity. They showed that TNF kills sensitive cells by an oxidative process and that anaerobic cells were resistant to TNF cytotoxicity. The reduction of TNF cytotoxicity under anaerobic conditions was interpreted as a requirement for oxygen in the cytotoxic action of TNF (Matthews et al. 1987). Indeed, tumour cell response to the cytotoxic action of rhuTNF in vivo correlates with the anti-oxidant buffering capacity of the target tumour, as measured by its in vitro intracellular glutathione content (Zimmerman et al. 1989). Thus, if oxygen is necessary for TNF cytotoxicity it would not be surprising if the hypoxic cells which occur in solid tumours influence the therapeutic response to systemically administered TNF. Most in vitro studies with TNF are performed in air (21% oxygen) which is not representative of oxygen tensions found in vivo. The majority of cells in experimental murine and xenografted human tumours exist in regions with oxygen tensions of 2% or less (Lartigau et al. 1993; Vaupel et al. 1991).

1.18 Hypothesis This research project was initiated to test the hypothesis that the oxygen tensions which exist in solid tumours modulate the cytotoxic activity of TNFa. Central to this was the need to measure tumour oxygenation in vivo and then to establish how such oxygen levels when recreated in vitro modify the cytotoxic action of TNF.

31 Chapter 2

Measurement of tumour oxygenation status in two murine tumour models

2.0 Introduction 2.1 Methods 2.1.1 Mice 2.1.2 Tumours 2.1.2.1 Sarcoma F 2.1.2.2 Carcinoma NT 2.1.3 Tumour transplantation

2.1.4 Eppendorf ™ pÛ 2 histograph 2.1.4.1 Principle of measurement 2.1.4.2 Experimental procedure 2.1.4.3 Analysis of results: pOi 2.1.5 Radiobiological hypoxic fraction 2.1.5.1 Principle of measurement 2.1.5.2 Tumour irradiation procedure 2.1.5.3 In vivo/in vitro colony forming assay 2.1.5.4 Preparation of heavily irradiated feeder cells 2.1.5.5 Analysis of results: hypoxic fraction 2.2 Results 2.3 Discussion 2.4 Summary

32 2.0 Introduction

Tissue oxygenation is a chief parameter defining the cellular microenvironment in addition to tissue pH distribution, bioenergetic status and nutrient supply. It has been hypothesised for some time that some cells in solid tumours exist in regions of low oxygen concentration (Mottram, 1936; Thomlinson and Gray, 1955). There is now convincing evidence that human tumours contain hypoxic cells (Guichard, 1989; Rockwell and Moulder, 1990; Moulder and Rockwell, 1987). Tumours contain two types of hypoxic cells: chronically hypoxic cells (Thomlinson and Gray, 1955) and transiently hypoxic cells (Reinhold et a l 1977; Brown et al 1979; Chaplin and Trotter, 1991). The former are most likely to result from an increase in intercapillary distances in tumours, such that some cells exist beyond the diffusion distance of oxygen and the latter occur as a result of a restriction of the microcirculation and thus of the oxygen availability to tumour cells in vivo. In tumours of differing structure the mean distance from a neighbouring blood vessel to a necrotic region has been found to be in the range of 100-200 |im (Thomlinson and Gray, 1955; Tannock, 1968) beyond which diffusion limited (chronically hypoxic) cells are assumed to reside. Transiently hypoxic cells occur due to temporary cessations or reductions in tumour blood flow and thus can exist even immediately adjacent to endothelial cells (Reinhold et al 1977; Chaplin and Trotter, 1991; Chaplin and Hill, 1995). It is likely that tumours exhibit a continuous distribution of oxygen tension gradients which constantly change throughout tumour growth and development. Vaupel et a l (1990) have shown that both the median pOi values and the oxygen consumption rates strongly depend on the oxygen availability and thus on the efficiency of tumour blood flow. Whereas in normal tissues the median pOz values range from 24-66 mmHg, the respective values in malignancies are usually < 20 mmHg (2.6% oxygen) (Vaupel et al 1990; Kallinowski et al 1990). Although hypoxia is a common feature of solid tumours, tumour oxygenation status varies markedly from tumour-to-tumour and more importantly intratumour heterogeneities also exist. The importance of tumour hypoxia as a determinant of tumour response to treatment has led to the use of a variety of methods to quantify hypoxia. The ideal method would be quick, convenient, reliable, non-invasive and would enable

33 measurements before, during, and after treatment, however, current methods do not match these criteria. The classical method of assessing tumour oxygenation status is by measurement of the radiobiological hypoxic fraction (HF). The hypoxic fraction can be calculated from a paired survival curve assay which determines the difference in response of tumours irradiated unperturbed in air breathing animals and those in which the tumour has been rendered completely hypoxic. The greater is this difference, the smaller the proportion of radiobiological hypoxic cells. This method assumes that naturally occurring and artificially induced hypoxic cells have similar radiation sensitivities (reviewed by Moulder and Rockwell, 1984). Cells which are well oxygenated are three times more sensitive to the effects of radiation than poorly oxygenated cells (Gray, 1957), as shown in Figure 2.1. Although the hypoxic fraction does not give any indication of the range of tumour oxygen levels, it assesses the proportion of radiobiologically hypoxic cells, which are known to range from 1-90% in experimental tumours (Moulder and Rockwell, 1984). In theory, if the hypoxic fraction of a tumour is 30% then 70% of the tumour must consist of oxygen tensions greater than 5.0 mmHg (c/Figure 2.1). The advent of the Eppendorf p02 histograph has enabled oxygen tensions in normal and tumour tissues to be determined in situ. The electrode which acts as the cathode is also used to measure the current produced, which is proportional to the oxygen tension, or p02 (p02 = partial pressure of oxygen). Polarographic oxygen electrodes offer a quick, reliable method of measuring the oxygenation status of tissues and easily accessible tumours, and, they do not have any adverse effects (Lartigau et al 1992). One disadvantage of this technique is that the oxygen tension of all cells sampled is measured, regardless of their viability. However, p02 can be monitored before, during, and after treatment and thus related to therapeutic outcome (Gatenby et al 1988; Vaupel et al 1991; Hôckel et al 1993). By measuring the oxygenation status of tumours using p02 electrodes and the radiobiological HF assay, the proportion of oxic and hypoxic cells can be determined, giving a better assessment than either method alone (Hirst et al 1985; Horsman et al 1994; Vaupel et al 1989; Thomas et a l 1994).

34 3.0

21 % 0 , 100% O,

■f' 2.5 (/) 0)c CO o ^ 2.0 CT3 CC

4—' 03 ŒCD

50 40 50 60 70 190 380 570 760 Oxygen Tension (mmHg at 37°C)

Figure 2.1 Illustration of the dependence of radiosensitivity on oxygen concentration. Assuming the radiosensitivity under anoxia is assigned a value of 1.0 then the radiosensitivity is about 3.0 under well-oxygenated conditions. The radiosensitivity changes as the oxygen tension increases from zero to 30 mmHg but further increase in oxygen tension, to that found in air (160 mmHg), or even pure oxygen at high pressure, has little further effect. A sensitivity halfway between anoxia and full oxygenation occurs for a partial pressure of oxygen about 3 mmHg, which corresponds to 0.4% oxygen (adapted from Hall, 1988).

35 Oxygen tension is critical in radiation therapy, chemotherapy and biological approaches to cancer therapy. Oxygen levels have been shown to be important in the action of many cytotoxic drugs. Hypoxic cells may be significantly more resistant than oxygenated cells to chemotherapeutic agents which have oxygen involved in their mechanism of toxicity (Teicher, 1994). Oxygen and reactive oxygen species are implicated in the cytotoxic action of TNF (Matthews et ah 1987; Yamauchi er al. 1989; Chang et al. 1992; Goossens et al 1995) hence, it is conceivable that oxygen tension could modulate its cytotoxic action in vivo. TNF has not been as successful in the clinic as was originally expected (reviewed by Jones and Selby, 1989; Fraker et al. 1995). One factor involved in determining response to TNF in vivo could be tumour oxygenation status. The cytotoxic effects of TNF in vitro have, with few exceptions, been studied using oxygen concentrations of 21% oxygen (160 mmHg). Evidence indicates that the majority of cells in experimental tumours and human tumour xenografts exist in a microenvironment of 2% oxygen or less (Lartigau et al. 1993; Vaupel et al. 1991). In order to establish a relevant range over which to investigate the importance of oxygen tension on the cytotoxic effects of TNF, studies were undertaken to determine the oxygenation status of two experimental tumour models in vivo.

The aim of this section of the project was to determine the oxygenation status of two murine tumour models, namely, the sarcoma F (SaF) and carcinoma NT (CaNT) using two independent methods:

(1) measurement of the oxygen partial pressure (p 0 2 ) distribution polarographically, using the Eppendorf pOi histograph (2) measurement of the radiobiological hypoxic fraction

An assessment of the tumour oxygenation status would enable in vitro experiments, investigating the cytotoxic action of TNF, to be carried out at relevant oxygen tensions (i.e. at the levels of oxygen found in vivo) and yield a more realistic view of what could be expected in vivo.

36 2.1 Methods

2.1.1 Mice All experiments were carried out on inbred female CBA/Gy f TO mice aged 12 to 16 weeks. The mice were bred in the Gray Laboratory mouse colony. They were housed in steel cages on sawdust bedding and given water and mouse chow ad lib.

2.1.2 Tumours 2.1.2.1 Sarcoma F Sarcoma F (SaF) tumour is a rapidly growing, anaplastic sarcoma that arose spontaneously on a CBA/Ht mouse in 1957 in the Gray Laboratory mouse colony. The tumour has been maintained in the murine strain of origin by serial transplantation of a crude cell suspension from a frozen stock originally derived in 1978. Since, tumours have been maintained from this frozen stock after 10 passages. It is rapidly growing and has a latent period of about 10 days.

2.1.2.2 Carcinoma NT Carcinoma NT (CaNT) arose spontaneously on the neck of a female CBA/Ht mouse in 1968 and was classified as a mammary carcinoma. It is a moderately differentiated carcinoma and grows as an encapsulated mass. It has a latent period of about 3 weeks.

2.1.3 Tumour transplantation The tumour bearing donor mouse was sacrificed by cervical dislocation. The tumour was aseptically excised, chopped with scissors and crossed scalpels and the resulting tumour mince diluted with 10 ml of saline. Mice were anaesthetised by inhalation of metofane, the backs shaved and wiped with 70% ethanol, or the foot wiped with 70% ethanol. Tumour cell suspension, 0.05 ml, (approximately 5x10^ cells) was injected subcutaneously into either the rear dorsum (back) or the foot of the right hind limb (foot) of 12 to 16 week old female mice. Experiments were carried out when the tumours had reached a tumour volume of 200 mm^, 5.5-6 mm geometric mean diameter (GMD), which generally occurred 10-14 days (SaF) or 21-30 days (CaNT) after implantation.

37 2.1.4 Eppendorf pOi histograph

Estimates of tumour pÛ 2 were obtained using a fine needle oxygen electrode probe (Eppendorf, Hamburg, Germany) shown in Figure 2.2. This work was carried out in collaboration with Dr M R Horsman’s group, Danish Cancer Society, Department of Experimental Clinical Oncology, Aarhus, Denmark.

Figure 2.2 The Eppendorf pOi histograph

P

2.1.4.1 Principle of measurement Oxygen partial pressure (tension) is measured by a polarographic micro­ electrode contained in the recessed tip of a bevelled steel needle probe which is 250 pm in diameter. The oxygen micro-electrode has an active gold cathode with a diameter of 12 pm which is covered by a teflon membrane. It is polarized to -700 mV towards a Ag/AgCl anode which is attached to the skin or underlying musculature near the site where the probe is inserted. Oxygen electrodes work on the principle of electrochemical reduction of oxygen at the cathode (probe surface). The process of reduction results in the production of OH ions and requires a source of electrons supplied by the anode:

O2 4- 4e' -f- 2 H2O 40H (Fatt, 1976)

38 The probe is inserted into the tissue and automatically moves in a "pilgrim-step" fashion (1 mm forward and 0.3 mm backwards) which is designed to relieve tissue from compression caused by forward motion of the needle, and thus compensates for possible artifact due to tissue compression and/or bleeding and also oxygen consumption by the probe. The probe consumes oxygen in the process of measuring oxygen tension thus an absolute measure of oxygen tension is not attainable. The resulting current (0.01 to 3.0 nA) is proportional to the oxygen partial pressure at the electrode tip. The probe can drift up to 5% in measurement thus the p02 value measurements are corrected for temperature (2.4% / °C) and atmospheric pressure; there is no pH dependence from pH 6.8 to pH 7.5. In order to be able to compare measured values from sample to sample, the needle probe must be calibrated (in 0.9% saline) before and after every sample. Recalibration in air and nitrogen after measurement automatically corrects the measured data by linear interpolation and thus increases the precision of the measurement.

2.1.4.2 Experimental procedure Sarcoma F and carcinoma NT, dorsal and foot tumours, were randomized into groups of 5-6 mice for pO] measurements. The tumour area was shaved (dorsal tumour) and any residual hair scraped off with a scalpel blade. The non-anaesthetized mouse was restrained, in a Lucite jig immobilized with tape, with the tumour-bearing area exposed (c/Figure 2.3) An ECG electrode was attached to smooth skin near the tumour and this was connected to the histograph. The probe was calibrated in saline solution immediately before each tumour measurement. It was then inserted into the tumour after making a small hole in the overlying skin with a 25 G needle if necessary. The tip of the probe was manually advanced about 1 mm into the tumour tissue and thereafter it automatically advanced in 1 mm forward steps, each with a 0.3 mm retraction, giving a net forward movement of 0.7 mm before taking a reading. Response time is 1.4 seconds and a single p02 value measurement is made in <40 milliseconds. This process was repeated until a sufficient number of p02 values (50-100) were collected; it takes approximately 6 min to record 199 values. All tracks started from almost the same position but took different directions. Path lengths were typically 5 mm but were modified according to the geometry of individual tumours if necessary. At the end of each complete set of measurements the probe was removed from the tissue and recalibrated. Tumour temperatures were individually measured after p02 measurement

39 using a needle thermoeouple. The pO? data were then recalibrated taking into account tumour temperature and air pressure.

Figure 2.3 Experimental set-up for pO% measurement

2.1.4.3 Analysis of results: p02 Data collection was completed within 3-5 minutes per tumour. The relative frequency of the pOi measurements was calculated and displayed as a histogram. Raw data values were transferred from the Eppendorf histograph and analysed using an ExeeF'^ software package. Median and mean values and values at or below 2.5 mmHg and 5.0 mmHg for individual tumours were calculated from the raw data. Data for all tumours of each group were also pooled and cumulative pO] histograms for each group were plotted.

40 2.1.5 Radiobiological hypoxic fraction

2.1.5.1 Principle of measurement The radiobiological hypoxic fraction was calculated from a paired survival curve assay. This compares the survival curves for tumours irradiated under normal aeration and 100% hypoxia (Hewitt and Wilson, 1959; van Putten and Kallman, 1968). Tumour cells were assayed for viability by an in vitro colony forming assay. When the oxic and hypoxic curves are parallel the best linear fit is found by linear regression.

2.1.5.2 Tumour irradiation procedure All irradiations were performed using a 250 kV Pantak X-ray machine, set at 240 kV and 15 mA. The mean dose rate at the tumour was 2.3 Gy/min for foot tumours and 3.6 Gy/min for dorsal tumours.

SaF and CaNT tumours were irradiated in a similar fashion to the method originally developed by Sheldon and Hill (1977). The mice were either unanaesthetized air-breathing or hypoxic during irradiation. Hypoxia was induced by sacrificing the host by cervical dislocation 10 min before irradiation. They were restrained in lead jigs with the aid of autoclave tape. Pieces of cardboard were placed under the posterior in order to fully expose the subcutaneous dorsal tumour which projected through a hole in the jig whilst the rest of the body was protected from irradiation. Four tumours were irradiated simultaneously. The mice were turned through 180° and exchanged between diagonal portals halfway through each irradiation to correct for decrease in dose across the tumour and for non-uniformity between the portals (Fowler et al. 1975). Doses given were 10, 15, 20 and 25 Gy, unirradiated control mice were restrained in jigs for the duration of the radiation procedure.

For irradiation of tumours implanted in the foot, mice were placed in boxes from which the left hind leg protruded. Four boxes were placed on a perspex plate with the feet held in position by a series of perspex posts (Douglas and Fowler, 1976). Only the feet of the mice were exposed to the radiation through a slit in the collimator. All tumour irradiations were performed in air at room temperature.

41 2.1.5.3 In vivo Un vitro colony forming assay After irradiation the tumours were either excised immediately or left for 18 hours before excision (air-breathing mice). The tumours were excised aseptically, weighed, chopped with crossed scalpels and placed in an enzyme-cocktail of protease (1 mg/ml), DNAse (0.5 mg/ml) and collagenase (0.5 mg/ml) made up in 10 ml of serum- free Eagle’s minimal essential medium (EMEM) {cf Appendix). The tumour/enzyme mixture was continuously mixed on a rotar wheel for 30 min or 1 h, for foot tumours and dorsal tumours respectively, at 37°C. Any remaining debris was allowed to settle to the bottom of the universal and the supernatant was carefully removed and neutralized with 10 ml of EMEM containing 10% ECS. This mixture was then centrifuged for 10 min at 2000 rpm. The supernatant was discarded and the pellet was resuspended in an appropriate volume of EMEM + 10% ECS. The mixture was vortexed and then syringed three times through a 21 G needle. A cell count was performed, using a haemocytometer and a light microscope, by taking 0.3 ml of the single tumour cell suspension and mixing this with 0.3 ml of trypan blue. The cell count was adjusted to give a known number of cells per ml from which a serial dilution was made. Tumour cell suspension was added to an equal volume of heavily irradiated feeder cell suspension {cf 2.1.5.4 below) and the petri dishes incubated for 7-10 days at 37°C in a humidified 2% oxygen / 5% CO 2 / balance nitrogen incubator or until colonies were visible and well formed. The colonies were fixed with 70% industrial methylated spirits for 5 minutes, stained with methylene blue for 5 minutes and then rinsed in tap water three times. The dishes were allowed to air dry and the colonies were counted using a colony counter.

2.1.5.4 Preparation of heavily irradiated feeder cells Although the SaF and CaNT tumour cells grow readily in culture, the plating efficiency can be increased by the provision of a heavily irradiated feeder layer of V79- 379A cells (c/Appendix) When required for an in vitro colony forming assay, 50-100 ml of V79 cell suspension was centrifuged at 2000 rpm for 5 minutes. The supernatant was discarded, the pellet resuspended in 20 ml of EMEM 4- 10% ECS and a single cell suspension was prepared. The cell number was counted and adjusted accordingly. The cells were vortexed and then irradiated by Cobalt^^ source for 5 min (>100 Gy). The heavily irradiated cells were added to the appropriate volume of EMEM + 10% ECS medium to give the required number of cells per ml, 2x10^ cells were plated per 9 cm

42 petri dish and allowed at least 2 h to attach before the tumour cell suspension was plated.

2.1.5.5 Analysis of results: hypoxic fraction Colony forming efficiency (CFE) was calculated by counting the number of colonies on the petri dish and dividing this by the number of cells originally seeded. Three dilutions were made of each sample and each dilution was duplicated, thus the average of 6 samples gives the average CFE for each group and a standard error on the mean was calculated. The surviving fraction (SF) was calculated by calculating the number of clonogens per 100 mg of treated tumour and dividing this by the average number of clonogens per 100 mg of control tumour. The SF of the aerobic and hypoxic treatments were plotted on a semi-log scale (radiation dose versus SF) and the best linear fit for the hypoxic curve was calculated from the equation:

Linear fit = {EXP(A + B(X))}

This fit was superimposed on the aerobic data and the hypoxic fraction was calculated from the vertical displacement of the aerobic curve from the hypoxic curve. The mathematical formula for the hypoxic fraction is as follows:

Hypoxic fraction = natural log e a(oxic))

In order to obtain an error on the hypoxic fraction the ratios of the four paired dose points were averaged and the standard deviation of the mean calculated. Significance levels were determined using the Students t test.

2.2 Results

pOi histograms for SaF and CaNT tumours grown subcutaneously on the rear dorsum (dorsal) and hind foot dorsum (foot) of mice are shown in Figures 2.4 and 2.5 respectively. The oxygenation status of tumours is generally expressed by the median pÛ 2 of all pooled measurements of a single tumour type as well as by the fraction of pO]

43 values less than 2.5 mmHg and 5.0 mmHg. The histograms represent the individual classified pOi values calculated from the raw data. Clearly, there is a distinct shift of the distribution curves to the left which

indicates that most cells are situated in an hypoxic environment. The median pÛ 2 values of the SaF and CaNT tumour types range from 2-12 mmHg and 2-5 mmHg, respectively, which is less than 2% oxygen (15 mmHg). The raw data summarised in Table 2.1 show the mean, median and percentage of values at or below both 2.5 mmHg and 5.0 mmHg, with 95% confidence limits.

Figures 2.6 and 2.7 show survival curves of SaF and CaNT, dorsal and foot tumours, respectively. Each survival curve was compiled from the combined data of tumours excised immediately and 18 hours post irradiation as there was no evidence for potential lethal damage repair. The survival curves clearly show that the SaF and CaNT dorsal tumours have similar hypoxic fractions, 67 ± 3 % (SaF) and 57 ± 13% (CaNT). By contrast, the SaF foot tumours have a significantly greater HF (at the 0.05 level) than CaNT foot tumours (72 ± 25% versus 24 ± 18%, respectively). However, further data analysis (c/Table 2.1) reveals that overall the SaF and CaNT models, regardless of tumour site, have hypoxic fractions within the same wide range: 7-70% (CaNT) and 48- 979% (SaF).

Table 2.1 Summary of pOz and hypoxic fraction data

Tumour Hypoxic Mean p O z Median p O z ( % ) p O z ( % ) p O z (site) Fraction (%) mmHg mmHg < 2.5 mmHg < 5.0 mmHg

CaNT(back) 57 (44-70)* 6.1 (3.8-8.4) 3.3 (2.6-4.1) 32.5 (8-57) 80.0 (67-94) CaNT (foot) 24 (7-41) 7.7 (5.8-9.6) 3.2(1.8-4.6) 41.4(17-66) 60.9 (47-74) SaF (back) 67 (64-69) 5.3 (3.4-7.2) 3.2 (2.4-3.9) 30.9 (9-53) 76.7 (62-92) SaF (foot) 72 (48-97) 11.3 (6.5-16.1) 7.4 (3.1-11.7) 29.9 (0-62) 45.5 (18-73) * = 95% confidence interval

44 MedianpO^ = 3.24 mmHg 50 n (a)

4 0 -

3 0 -

2 0-

^ 10 ë N=6;n=291

I °0 25 50 75 100 Vh P-H O) . > 5 0 1 -w M edianpO 2 = 7.36 mmHg (b)

CD 40H

3 0 -

2 0 -

10 - N=5;n=318

25 50 75 100 Oxygen Partial Pressure (mmHg)

Figure 2.4 pOi distribution in SaF (a) dorsal and (b) foot tumours. N= number of tumours; n= number of pOi values.

45 MedianpÛ 2 = 3.28 mmP% (a) 50-1

4 0 -

3 0 -

o 2 0 -

0 25 50 75 100

CD > * -W MedianpO 2 = 3.20 mmP% (b) j g 50 n ~0j

40

30

20

10

N=5; n=259

lluWmH . _ , —I----r- -I------1------1------r - 25 50 75 100 Oxygen Partial Pressure (mmHg)

Figure 2.5 pO] distribution in CaNT (a) dorsal and (b) foot tumours. N= number of tumours; n= number of pO? values.

46 10° HF = 66.6 ±2.8%

Hypoxic (n=8) Air Breathing (n=12)

LL 10 15 20 25 O) c > > 10° HF = 71.9 ±24.7% cn

Hypoxic (n=7) Air Breathing (n=11)

10 15 20 25 Dose (Gy)

Figure 2.6 Survival curves of SaF (a) dorsal and (b) foot tumours;

n = number of tumours.

47 10° HF = 56.8 ±13.4%

Hypoxic (n=7) Air Breathing (n=11)

D) 10 15 20 25 C > > 3 0 ° Cf) HF = 24.2 ±17.5%

Hypoxic (n= Air Breathing (n=

10 15 20 25 Dose (Gy)

Figure 2.7 Survival curves of CaNT (a) dorsal and (b) foot tumours;

n = number of tumours.

48 2.3 Discussion

It has been clearly established using two independent techniques that both the SaF and CaNT tumours, implanted in either the rear dorsum (dorsum) or hind foot dorsum (foot), are poorly oxygenated. However, the interpretation and significance of the results obtained warrants further discusssion. The classical method of assessing radiobiologically hypoxic cells is the paired survival curve assay, which determines the difference in response between tumours irradiated under aerated (air-breathing) conditions and those rendered completely hypoxic. The greater is this difference, the smaller the proportion of hypoxic cells. Recently, the Eppendorf pOi histograph has enabled more detailed assessments of oxygen tensions present in tissues. In this study tumour oxygenation status was assessed by both radiobiological hypoxic fraction measurement and Eppendorf pOi histography. Tumour oxygenation was assessed in tumours growing in different sites (rear dorsum and hind foot dorsum) to determine if differences occurred between tumour sites, which would probably be due to the differences in blood supply to the tumours. It can be seen from the pOz histograms that there is an accumulation of values in the lower pOi classes which clearly establishes that the oxygenation status of both the SaF and CaNT is inadequate. Nearly all pOi values lie below 35 mmHg (-5% oxygen). All tumours have at least 50% of pOi values <5.0 mmHg, which is the oxygen tension below which radiobiological hypoxia occurs (c/Figure 2.1). The median pO% of SaF and CaNT tumours ranges from 2-12 mmHg and 2-5 mmHg which is less than 2% oxygen (15 mmHg). There is less scattering of the pO: values in dorsal tumours than in foot tumours due to an absence of high pÛ 2 values. However, the Eppendorf is known to drift by up to 5% (usually 2 mmHg) in measurements, hence, median pOi values are not significantly different between tumour types or site of tumour implantation. There is no evidence presented in this study that HF varies between the SaF and CaNT, two tumours which differ in both growth rate and degree of differentiation. However, although the HF of SaF is constant between different tumour sites, this is not true of the CaNT. CaNT dorsal tumours have a significantly greater HF (p <0.05) than CaNT foot tumours although their pOa distributions are similar. Clearly, differences in the proportion of clonogenic to non-clonogenic hypoxic cells in two tumours with the same pO] distribution may lead to different fractions of radiobiologically hypoxic cells. In the case of foot tumours, although a large fraction of the tumour may be hypoxic, the

49 radiobiological HF would be low if a substantial proportion of non-clonogenic cells are present in the regions of low oxygen tension. Hypoxic fraction may differ between tumours (SaF versus CaNT foot tumours) because of differences in tumour vasculature, or intrinsic cellular characteristics such as the rate of oxygen consumption and ability to retain clonogenicity under hypoxia. However, in this study it has been clearly established that, overall, the SaF and CaNT exhibit similar, wide ranges, of hypoxic fractions suggesting that the aforementioned characteristics are similar in these tumour models. Table 2.1 summarises the pO] and hypoxic fraction data. The HF, if estimated from the p02 data could be taken as the percentage of values below 5.0 mmHg, the oxygen tension below which radiobiological hypoxia occurs. Overall, the data show that the proportion of p02 values <5.0 mmHg range between 18-92% for the SaF and between 47-94% for the CaNT. Thus, both tumour types have similar ranges of p02 values <5.0 mmHg which is in agreement with their hypoxic fractions which range from 48-97% (SaF) and 7-70% (CaNT) establishing the hypoxic nature of these tumour models. However, for CaNT foot tumours the p02 estimation of HF is marginally lower than that calculated by radiobiological methods. One explanation for this discrepancy is the inherent error of 2 mmHg on Eppendorf p02 measurements and its inclusion of necrotic, non-clonogenic hypoxic cells in the measurements.

2.4 Summary The tumour oxygenation status of the SaF and CaNT tumours has been assessed both directly, by p 0 2 measurements, and indirectly, by measuring the radiobiological hypoxic fraction. Clearly, the tumours are poorly oxygenated with most cells situated in an environment at or less than 15 mmHg (2% oxygen). This suggests that experiments investigating the cytotoxic action of TNF should be carried out at such oxygen tensions in vitro. However, it is likely that cells at the same oxygen tension in the two different tumour types studied may be exposed to entirely different microenvironments. Factors such as extracellular pH, pC02, level of cellular toxic products and availability of growth factors may contribute to a cell’s response to TNF, however, these have not been investigated in this study. The purpose of the present study was to determine if and how, tumour relevant oxygen tensions influence the cytotoxic action of TNF, and this first part of the study has provided the range of oxygen tensions which should be evaluated.

50 Chapter 3

The effect of tumour relevant oxygen tensions on the cytotoxic effects of TNF in vitro

3.0 Introduction 3.1 Materials and methods 3.1.1 Cell lines 3.1.2 TNF 3.1.3 Actinomycin D 3.1.4 Oxygen conditions 3.1.5 Assessment of TNF cytotoxicity 3.1.5.1 Colourimetric assay 3.1.5.2 Clonogenic assay 3.1.6 To determine the effect of 24 h priming at tumour relevant oxygen tensions on the cytotoxic action of TNF in 21% oxygen 3.1.7 To determine the effect of tumour relevant oxygen tensions during TNF treatment on the cytotoxic action of TNF 3.1.8 To determine the time-course of hypoxia-induced resistance to the cytotoxic action of TNF 3.1.9 To determine the time-course of re-sensitization to the cytotoxic action of TNF after priming in 2% oxygen 3.1.10 To determine the effect of chronic exposure to 2% oxygen on the cytotoxic action of TNF 3.2 Results 3.3 Discussion 3.4 Summary

51 3.0 Introduction

The direct cytotoxic effects of TNF on many types of malignant cells have been studied extensively. Evidence indicates that the therapeutic value of TNF as an anti­ tumour agent is limited by toxic side-effects which occur at high TNF doses, and by a wide variation in the sensitivity of tumour cells to TNF (Kirstein et al 1986; Gerlach et aA1989^

The cytotoxic effects of TNF can manifest as either apoptosis or necrosis depending on the type of target cell, culture conditions and the presence of metabolic inhibitors (Faster et al 1988; Grooten et al 1993). Although TNF is not selective in its toxicity there is some degree of selectivity when tumour cells and normal cells from the same tissue are studied (Dealtry et al 1987; Dollbaum et al 1988). It has been estimated that 40% of tumour cell lines are growth inhibited or lysed by TNF in vitro (Goeddel et al 1986). The most sensitive target cells can be lysed with just picograms of TNF. The majority of untransformed cells are resistant to the cytotoxic action of TNF as resistance is inherited as a dominant trait (Nophar et al 1988). It has been suggested that the toxicity of TNF to normal cells, particularly endothelial cells, may be an essential component of its anti-tumour activity in vivo (Malik and Balkwill, 1989; Rothstein and Schreiber, 1993). Indeed, it has been shown that TNF has a direct cytotoxic effect on normal, nascent and tumour endothelium (Remick and Kunkel, 1993; Sato et al 1986; Gerlach et al 1989; Watanabe et al 1988b)

The mechanism(s) involved in the cytotoxic action of TNF are slowly being unravelled. Factors known to induce resistance or sensitivity to the cytotoxic action of TNF are listed in Table 3.1.

52 Table 3.1(a) Factors that induce sensitivity to the cytotoxic action of TNF

Factor Reference Adenovirus El A protein Chen et al. 1987 Overexpression of c-myc oncoprotein Klefstrom et al. 1994 Proteases involved in TNF degradation Ohsawa and Natori, 1988 Lysosomal enzymes Watanabe gr a/. 1988a Reactive oxygen species Shoji et al. 1995 Goossens a/. 1995

Table 3.1(b) Factors that induce resistance to the cytotoxic action of TNF Factor Reference Serine protease inhibitors Suffys et al. 1988b

Inhibitors of phospholipase A 2 Mutch et al. 1992 Inhibitors of arachidonic acid metabolism Neale et al. 1988 Intracellular delivery of TNF Smith 1990 Anoxia Matthews et al. 1987 Oxidants Yamauchi etal. 1989 Antioxidants Satomi et al. 1988 Manganese superoxide dismutase (MnSOD) Wong and Goeddel, 1988 Hirose gr a/. 1993 TNF autoinduction Kxxhmetal. 1986 Endogenous TNF production Rubin gr aZ. 1988 Induction of heat shock proteins Jaattela aZ. 1989 Zinc finger protein A20 Opipari et al. 1992 Expression of BCL-2 protein ITennetgfd/. 1993a

53 Competition between induced protective factors and the damaging capacity of TNF might determine whether a response is cytotoxic. Initially it was thought that TNF receptors (TNF-Rs) might serve as determinants of cellular susceptibility to the cytotoxic action of TNF (Old, 1985). However, high affinity receptors for TNF have been demonstrated in various cell lines regardless of their sensitivity. In addition, it was shown that TNF binds to specific receptors on cells which are completely resistant to the cytotoxic action of TNF (Tsujimoto et al 1985; Yoshie et al 1986). It is possible that conformational change in TNF-Rs, the levels of soluble TNF-Rs and free TNF may affect the cells response to TNF (Peck et al 1989). It has been suggested that post­ receptor events, particularly TNF degradation processes, may determine the cytotoxic action of TNF (Oku et al 1987; Ohsawa and Natori, 1988). Although much effort has been expended investigating the cytotoxic mechanisms of TNF both in vitro and in vivo the influence of the cellular microenvironment on the actions of TNF has not been considered in the majority of studies. One microenvironmental factor which is known to differ between normal and malignant tissue is the oxygenation status. Oxygen tension has been shown to be an important modulator of the cellular action of several cytokines and growth factors (Wing et al 1988; Aune and Pogue, 1989; Loeffler et al 1992; Ishizaka et al 1992). Matthews et a l (1987) showed that TNF kills sensitive cells by an oxidative process as cells under anaerobic conditions were resistant to the cytotoxic action of TNF. It was hypothesised that tumour cell sensitivity to TNF in vivo is dependent on its capacity to buffer oxidative attack (Zimmerman et al 1989). A few years later it was found that the cytotoxic effects of TNF were eliminated under anoxic conditions suggesting a critical role for oxygen and reactive oxygen species (ROS) (Park e ta l 1992). There is now convincing evidence that TNF mediates its cytotoxic activity via reactive oxygen species (ROS) particularly superoxide (Goossens et a l 1995; Shoji et a l 1995; Hennet et al 1993b; Schulze-Osthoff et al 1992). Mitochondrial production of superoxide is a causal mechanism of TNF cytotoxicity as elevated superoxide levels induced by TNF lead to activation of dehydrogenase resulting in a fatal decrease in cellular ATP levels (Goossens et al 1995). Hydroxyl radicals generated from superoxide and hydrogen peroxide contribute to DNA damage, which occurs before cell lysis, in L929 cells treated with TNF and AMD (Shoji et al 1995). The mitochondrial production of superoxide increases linearly with oxygen tension as it is a non-enzymatic process (Turrens et al 1982). Indeed, at reduced oxygen tensions there are reduced

54 levels of ROS (Schultze-Osthoff et al. 1992). However, TNF accelerates leakage of ROS, particularly superoxide, from the mitochondrial electron transport chain (Hennet et al. 1993b; Shoji et al. 1995) and ROS scavengers block the formation of ROS and arrest TNF cytotoxicity (Goossens et al. 1995). These studies suggest a critical role for oxygen levels in determining susceptibility to TNF cytotoxicity. However, although studies to date indicate the potential importance of oxygen tension on the cytotoxic action of TNF, the dependence in detail has not been investigated. Such studies are needed to understand the importance of tumour relevant oxygen tensions on the activity of this cytokine. Most in vitro studies with TNF are performed in air (21% oxygen) (Meager, 1988) which is not representative of tumour oxygen tensions found in vivo. As mentioned previously, it is well established that the majority of cells in experimental murine and xenografted human tumours exist in regions with oxygen tensions of 2% or less (Vaupel et al. 1991; Lartigau et al. 1993). Understanding the cytotoxic effects of TNF at tumour relevant oxygen tensions (i.e., oxygen levels known to exist in tumours) in vitro can greatly contribute to the search for strategies to improve the therapeutic benefits of TNF.

The aim of this part of my investigations was to determine if tumour relevant oxygen tensions, as defined in Chapter 2, modulate the cytotoxic effects of TNF on a range of human and murine cell lines in vitro.

3.1 Materials and methods

3.1.1 Cell Lines All cell lines listed below were grown as monolayers. Details of the maintenance of cells lines are provided in Appendix. The Fsa-N, HT-29, MCF-7, and Svec cell lines were provided by Dr G.J. Dougherty, University of British Columbia, Vancouver, British Columbia,Canada.

Carcinoma NT (CaNT) Murine CaNT cells were derived from a CaNT tumour in 1990, by isolation of a single colony from a CaNT tumour, as described by Parkins et al. 1994.

55 Fibrosarcoma-N (Fsa-N) The Fsa-N is a non-immunogenic fibrosarcoma that arose in an irradiated C3Hf/Sed/Kam mouse, but in a site outside the irradiation field. It is well characterized with respect to the immuological responses it generates and its host cell populations (McBride, 1986).

HT-29: ATCC HTB 38 HT-29 cells were derived from a human colon adenocarcinoma which was moderately well-differentiated, grade II. This cell line was isolated from a primary tumour of a 44-year old white female in 1964 by J. Fogh. HT-29 cells have an epithelial- like morphology.

L929: ECACC No: 85011425 L929 is a subclone of the parental strain L, a cell line established by W.R. Earle in 1940. The L strain was derived from normal subcutaneous areolar and adipose tissue of a 100 day old male C34/An mouse. The L929 mouse fibroblast-like cell line used in these experiments was a gift from the National Institute of Biological Standards (NIBSC)

MCF-7: ATCC HTB 22 Human breast adenocarcinoma cells were originally derived by Soule et al. 1973 and have retained several characteristics of differentiated mammary epithelium.

Sarcoma F (SaF) Murine SaF cells were derived from a single colony of SaF cells which were obtained by enzymatic digestion and culture of an SaF tumour in 1992.

Svec Svec is a male mouse lymphoid endothelial cell line immortalised by Simian virus 40 (SV40) which retains functional characteristics of normal endothelial cells and binds lymphocytes. This cell line was originally derived by transient SV40 infection of mouse endothelial cells derived from lymph node stroma and denoted SVEC4-10 (O'Connell and Edidin, 1990).

56 3.1.2 Tumour Necrosis Factor-alpha (TNF) Recombinant human TNF (rhuTNF), international standard (IS) 87/650, and recombinant murine TNF (rmuTNF), IS 88/532, was provided by NIBSC as a sterile lyophilised powder. Each rhuTNF vial had a bioactivity of 40,000 lU (1 lU = 25 pg) and a protein content of 1 pg; a stock solution of 4,000 U/ml was stored at -20°C. Each rmu­ TNF vial contained 200,000 lU and was reconstituted as a stock solution of 40,000 U/ml and stored at -20°C. TNF was reconstituted in a solution of saline / 0.1% bovine serum albumin and further diluted in culture medium. The stock solution of TNF was thawed rapidly in a water bath (37°C) before use. TNF activity at room temperature was assessed over a one hour period and was found to be stable. In addition, the effect of repeated freeze/thaw cycles on the cytotoxic activity of TNF was assessed and was found to be stable for at least ten cycles.

3.1.3 Actinomycin D Actinomycin D (AMD) was reconstituted in EMEM, a stock solution of 1 mg/ml was stored at -20°C. AMD is light sensitive, therefore the concentration of the stock solution was confirmed by spectroscopy. Actinomycin D is a metabolic inhibitor that enhances the cytotoxic action of TNF in L929 murine fibroblasts (Meager et al. 1989). As the majority of cell lines employed were not innately sensitive to TNF, actinomycin D was used to sensitise the cells to TNF. AMD was used in all assays at a final concentration of 1 pg/ml (100 pl/well), unless otherwise stated, which caused 50% growth inhibition.

3.1.4 Oxygen conditions Oxygen concentration was controlled at 21%, 5% or 2% in the incubator gas phase using the combined C02\N2\02 QUEUE QWJ-700 incubator which is designed to electronically control oxygen tension as well as 5% CO 2 and temperature (37°C). Extreme levels of 1% or 95% oxygen and 95% nitrogen were achieved using premixed cylinders of 1% oxygen/5 % CÜ 2/balance nitrogen, 95% oxygen/5 % CO 2 or 95% nitrogen/5 % CO 2 respectively, connected to sealed containers at 37 °C. Containers were gassed at the required oxygen concentration for a total duration of two hours to render the atmospheres stable.

57 3.1.5 Assessment of TNF cytotoxicity SaF cells were seeded in 96-well tissue culture plates at IxloVml (100 pi/ well) in EMEM + 10% ECS unless otherwise stated. Table 3.2 shows the various concentrations of other cell lines employed.

Table 3.2 Cell number seeded for TNF cytotoxicity assays Cell line no. /ml

CaNT 3 x lC f SaF 1 X 10’ L929 1.5 X 10’ MCF-7 2 x 10’ HT-29 1 X 10’ Svecs 2.5 X lO’ Fsa-N 1 X 10’

3.1.5.1 Colourimetric assay Cells were seeded in 96-well tissue culture plates (0.1 ml per well) at different concentrations, depending on the cell line (Table 3.2). The cells were pre-incubated or “primed” at a specified oxygen tension for 24 h, during which time they adhered to the plastic. After 24 h the medium was discarded and the cells were treated with TNF (Table 3.3) either in the presence or absence of 1 pg/ml AMD. For each experiment control cells were included, which were treated with AMD only or medium only, and were exposed to identical experimental conditions as treated cells. Cells were then incubated for another 24 h, after which time the growth medium was carefully discarded and the remaining adherent, viable cells, were then fixed in methanol (70%) and stained with crystal violet (l%(w/v)). After staining the plates were thoroughly washed in tap water, allowed to air dry and the bound crystal violet solubilized with glacial acetic acid (33%). The absorbance of each well was read at 540 nm using a microtitre plate reader. The amount of dye taken up is proportional to the number of residual cells. It was previously determined that there is a linear relationship between optical density

58 (OD) and cell number up to 5 x 10 cells per well. The percentage of cell survival for a given TNF concentration was calculated as follows:

% cell survival =

Mean OD of TNF treated cells in the presence of AMD x 100 Mean OD of untreated cells in the presence of AMD

The TNF titre is calculated from a semi-log plot of TNF concentration (U/ml) versus

OD (% cell survival), cf Figure 3.1(a) and is expressed as the IC 50 value i.e., the concentration of TNF required to give 50% cell survival. Only the linear portion of the curve is used to estimate a titre for the TNF sample.

Table 3.3 TNF (U/ml) used to treat cell lines CaNT SaF L929 MCF-7 HT-29 SVEC Fsa-N rhu *rhu / rmu rhu rhu rhu rhu rhu / rmu

400 160 2 0 250 160 1 0 0 1 0 0 0 160 39

2 0 0 80 5 62.5 80 25 750 80 19.5

1 0 0 40 1.25 15.63 40 5 500 40 9.75

50 1 0 0.313 3.91 1 0 0.5 250 1 0 4.9

25 2.5 0.158 0.98 2.5 0.05 1 0 0 2.5 2.4 5 0.625 0.078 0.625 SxlO"* 31.25 0.156 0.61 0.5 0.158 0.009 0.156 5x10'® 7.8 0.039 0.3 0.078 0.039 3.9 0.019 0.15

0.009 0 . 0 1 0 .1 5x10^ 0.025 *rhu/rmu = recombinant human and murine TNF, respectively

59 3.1.5.2 Clonogenic assay SaF cells were seeded at 2x10^ cells per well, in 1 ml EMEM + 10% PCS, in 24- well plates which were preincubated in 21% and 2% oxygen (5% CO 2 / balance nitrogen) for 24 hours. The medium was discarded and the cells were treated with 5x10 ^ - 5.0 U rhu-TNF (0.5 ml of each concentration) and AMD, or AMD only or medium only. The plates were replaced in 21% oxygen for 24 h after which the cytotoxic action of TNF was tested by a colony forming assay: The medium was discarded and the cells were washed twice with 0.5 ml PBS. Trypsin-EDTA (0.5 ml) was used to detach cells after which it was neutralised by the addition of EMEM 4- 10% ECS. Trypsin treatment also inactivates TNF (Aggarwal et al 1985; Van Kessel et al. 1991). The samples were centrifuged at 1000 rpm for 5 minutes and then the cell pellet was resuspended in 10 mis of medium to give a single cell suspension. A cell count was performed using a haemocytometer and the cell number adjusted to give a known cell number. Three dilutions of the stock were made in duplicates. V79- 379A cells were used to prepare a heavily irradiated feeder layer which was seeded at 2x10^ cells per 9 cm petri dish (8 ml) (c/2.1.5.4). Tumour cells were added to the feeder layer and the dishes were incubated for 7-10 days, at 2% oxygen/5 % CO 2 /balance nitrogen, after which colonies were fixed and stained as described in 2.2.3. The surviving fraction was calculated from the colony forming efficiency (CFE) corrected for yield, as follows:

(CFE of TNF treated/ CFE of AMD control) x 100 = % cell survival

3.1.6 To determine the effect of 24 h priming in tumour relevant oxygen tensions on the cytotoxic action of TNF in 21% oxygen Cells were seeded at different concentrations depending on the cell line (see Table 3.2). Preincubation or “priming” was carried out under various oxygen tensions

(95%, 21%, 5%, 2%, 1% and 0.2% oxygen, all with 5% CO 2) for 24 h prior to incubation with TNF. Following this priming period the medium was discarded and cells were treated with TNF either in the presence or absence of AMD. Cells were then further incubated for 24 h under conditions of 21% oxygen. For each experiment control cells were included which were treated with either AMD or medium only and exposed to identical oxygenation conditions as TNF treated cells. At the end of the 24 h incubation period cytotoxicity was assessed according to 3.1.5.1

60 3.1.7 To determine the effect of tumour relevant oxygen tensions during TNF treatment on the cytotoxic action of TNF

SaF cells were preincubated in 21% oxygen/ 5% CO 2 / balance nitrogen for 24 hours. After preincubation, rhu-TNF was added in the presence of AMD to all wells except controls which were treated with either AMD or EMEM only. Instead of replacing all plates back in 21% oxygen after preincubation, plates were replaced in

95%, 21%, 5%, 2% and 0.2% oxygen, all with 5% CO 2. After 24 h all plates were removed and cells were fixed, stained, scanned and the cytotoxic effect assessed as in 3.1.5.1.

3.1.8 To determine the time-course of hypoxia-induced resistance to the cytotoxic action of TNF SaF cells were placed in the 21% oxygen incubator until all cells had attached

(~2 h). Plates were then preincubated in 2% and 21% oxygen (5% CO 2) for 0, 2, 4, 12, 14, 16, 18, 20 and 24 h. At each time point a plate was removed from both oxygen tensions, the medium discarded and cells were treated with TNF/AMD, AMD only or medium only. The plates were replaced in 21% oxygen for 24 h after which cytotoxicity was assessed ( c/3.1.5.1).

3.1.9 To determine the time-course of re-sensitization to the cytotoxic action of TNF after 24 h priming in 2% oxygen

SaF cells were preincubated in 2% and 21% oxygen (5% CO 2) for 24 hours. After 24 h, 2% oxygen samples were removed to 21% oxygen and after 0, 1,3, 5, 7, 14, 16, 18, 20 and 24 h, the medium was discarded and cells were treated with TNF and AMD, AMD only, or medium only, and the plates were replaced in 21% oxygen for a further 24 h before cytotoxicity was assessed (3.1.5.1).

3.1.10 To determine the effect of chronic exposure to 2% oxygen on the cytotoxic action of TNF SaF cells were continuously grown in T75 flasks in 2% and 21% oxygen, 5%

CO2 and were subcultured once weekly. On subculturing, 96-well plates were seeded as usual for a TNF cytotoxicity assay (c /3.1.5.1). Cells were preincubated for 24 h in the oxygen tension in which they were cultured. After 24 h cells were treated with TNF/

61 AMD, AMD only or medium only for 24 h, during which time all plates were incubated in 21% oxygen. Cytotoxicity was assessed as described in 3.1.5.1.

3.2 Results

Figure 3.1 shows the cytotoxic effect of rhu-TNF on SaF cells after 24 h pre­ incubation or “priming” in 21% or 2% oxygen, as assessed by both colourimetric and clonogenic assay. The colourimetric assay allows the calculation of an IC 50 value (x-axis value, where y=50), which is the concentration of TNF required to kill 50% of cells. It can be seen from Figure 3.1(a) that cells primed in 21% oxygen have an IC 50 value of 0.16 U/ml whereas for cells primed in 2% oxygen it is 0.79 U/ml. Thus, after 24 h priming in 2% oxygen there is a five-fold increase in the amount of TNF required to kill 50% of cells compared to cells primed in 21% oxygen.

Figure 3.1(b) shows that SaF cells surviving a TNF cytotoxicity assay are actually clonogenic and a surviving fraction can be calculated by obtaining an SF(0.5) value. It can be seen that the SF(0.5) value of cells primed in 2% oxygen for 24 h before TNF treatment, is over 10-fold greater than cells primed in 21% oxygen. Thus, over 10 times more TNF is required to kill 50% of cells after 24 h priming in 2% oxygen compared to 21% oxygen. This result supports that of the colourimetric assay and also ascertains the clonogenicity of cells treated with TNF in the presence of AMD. As the colourimetric assay is quick, convenient and less costly this was the method of choice for further assessments of the cytotoxic action of TNF.

6 2 100 - o C 80 - O U 60 - O * * * W 40 - Q

° 2 0 - N=5; n =15

10° rhu-TNF (U/ml)

* * * (b)

• ^ 4 o * * * c3 * * * tin * * * W) G > > V-l cr:

N = 3 ; n =9

0.1 I I I 11 ii|------1— I— I I I III]------1— I I I n il] T 1 I I I I III------1— I I I I I 10' 10-2 10' 10' 10 rhu-TNF (U/ml)

Figure 3.1 The cytotoxic effect of TNF on SaF cells after priming in 2% oxygen (-■-) and 21% oxygen (-0-) for 24 h; determined by (a) colourimetric assay; symbols show the mean ± s.d; and (b) clonogenic assay, symbols show the mean ± s.e.m. N = number of experiments; n = total number of samples. *p <0.05; ***p <0.01.

63 The effect of 24 h priming, under a range of tumour relevant oxygen tensions, on the cytotoxic action of both recombinant human and murine TNF on SaF cells (c/3.1.6) is shown in Figure 3.2 (a) and (b) respectively. Clearly, SaF cells which are sensitive to TNF in 21% oxygen become less sensitive (resistant) to TNF when primed at or below 5% oxygen. As previously determined for rhuTNF after 24 h priming in 2% oxygen, the

IC50 increases by a factor of at least five over the oxygen range of 2 1 % - 1 % and by a factor of two in the case of rmuTNF. If cells were gassed with 95% nitrogen/5% CO 2, which is known to reduce oxygen tension to 0.2% or less, prior to exposure to TNF in 21% oxygen, no detectable cytotoxicity was evident even with concentrations of 100 U/ml rhuTNF and 20 U/ml rmuTNF. This result indicates that the resistance to rhu- and rmu-TNF increases by a factor of greater than 500 and 100 respectively, compared to cells incubated at 21% oxygen. Priming cells in 95% oxygen does not alter cell sensitivity to TNF beyond that seen in 21% oxygen.

Table 3.4 shows the cytotoxic effects of TNF on a range of human and murine cell lines, after 24 h priming in tumour relevant oxygen tensions, including the SaF results previously described. Clearly, TNF exerts varied activity depending on individual cell lines. However, of the cell lines which are sensitive to the cytotoxic effects of TNF there is a trend of reduced sensitivity at tumour relevant oxygen tensions. The SaF, L929, MCF-7 and Svecs are sensitive to the cytotoxic action of TNF (in the presence of AMD) after 24 h priming in 21% oxygen whereas the HT-29, CaNT and Fsa-N cells are resistant. However, those cell lines which are sensitive become resistant by 24 h priming in tumour relevant oxygen tensions, the level of resistance varying between cell lines. One example of hypoxia-induced resistance to the cytotoxic effects of TNF is the MCF-7 human tumour cell line, illustrated in Figure 3.3(a). Clearly, tumour relevant oxygen tensions (0.2-5% oxygen) induce resistance to the cytotoxic action of TNF in cells which are normally sensitive, in the presence of AMD, in 21% oxygen.

64 100 - (a) Human TNFa

1.0 -

0.6 -

0.4-

^ 0. 2 -

^ 0.0 J 0.2

(b) Murine TNFa

0 . 8 -

0.6 -

0.4-

0 .2 - 0.0 0.2 2 521 95 Oxygen Tension (%)

Figure 3.2

Amount of TNFa required (+AMD 1 |Xg/ml) to cause 50% SaF cell kill (IC 50) after 24 h priming at tumour relevant oxygen tensions prior to TNF treatment in 21% oxygen. Columns show the mean ± s.e.m of five experiments performed in triplicate (n = 15).* p <0.05; ** p <0.01; ***p <0.001.

65 Table 3.4 Amount of TNF (U/ml), in the presence of AMD 1 pg/ml, required to kill 50% of cells (IC50) after 24 h priming at tumour relevant oxygen tensions

P0 2 SaF L929 MCF-7 SVEC HT-29 CaNT Fsa-N (%) *rhu /rmu rhu rhu rhu rhu rhu rhu 0.2 >100 >20 33.6 >150 >1000 -- - 1 0.79 - 12.7 - - -- - 2 0.79 0.35 11.6 22.7 942 >100 >400 >160 5 0.79 0.28 10.7 13.13 1004 - -- 21 0.16 0.14 1.9 1.22 566 >24.2 >400 >160

*rhu/rmu = recombinant human / murine TNF in respective columns. Where “>“ is shown cells were completely resistant to this concentration of TNF., an IC 50 value was unobtainable. Where blank cells are shown (-) experiments were not performed.

66 A cell line which does not require AMD to sensitise it to the cytotoxic action of TNF yet also demonstrates the hypoxia-induced resistance effect, seen in cell lines which require AMD to sensitise them to TNF, is the Fsa-N. Table 3.5 summarises the cytotoxic effects of both rhu- and rmu-TNF on Fsa-N cells, in the presence or absence of AMD, after 24 h priming in 21% or 2% oxygen. Results show that Fsa-N cells are resistant to both human and murine TNF when AMD is present but are sensitive to TNF alone when primed in 21% oxygen. Priming Fsa-N cells in 2% oxygen induces at least 175-fold resistance to the cytotoxic action of rhuTNF (-AMD), illustrated in Figure 3.3(b). This result clearly establishes that hypoxia-induced resistance to the cytotoxic action of TNF can occur in cells which are normally sensitive to TNF in the absence of a non-specific protein synthesis inhibitor such as AMD.

Table 3.5

Amount of TNF (U/ml) required to kill 50% of cells (IC 50) after 24 h priming in 21% or 2% oxygen

pOz Fsa-N Human TNF Murine TNF -f-AMD -AMD +AMD -AMD 2% Oxygen >160 96.1 >39 >39 21% Oxygen >160 0.54 >39 21.1

Where “>“ is shown cells were completely resistant to this concentration of TNF; AMD was used at a final concentration of 1 p-g/ml

67 (a)

100 -

o ü 80 -

Q 2 % Oxygen O 20 - 21% Oxygen 0.2% Oxygen

10° 10' rhuTNFa (U/ml) (b)

00 -

o Ü 80 c o U 60 -

40 - 21% oxygen (+AMD) Q 21% oxygen (-AMD) o 20 2% oxygen (+AMD) 2% oxygen (-AMD) I I I 111|------1— I— I I I 1111------1— I— I I 11111------1— I— I I I III]------1------1—I I I 1111------r , n-2 ,M-110-' inO 10^ m 10' l m 2 rhu-TNFa (U/ml)

Figure 3.3 The effect of 24 h priming at tumour oxygen tensions on the cytotoxic effects of rhu­ TNF on (a) MCF-7 cells (-i-AMD 1 pg/ml) and (b) Fsa-N cells. Symbols show the mean ± s.e.m of a single experiment (n =3).

68 Figure 3.4 illustrates the effect of tumour relevant oxygen tensions (0.2%-5% oxygen) during TNF treatment (c/3.1.7), after priming in 21% oxygen for 24 h. The data indicate that tumour relevant oxygen tensions during TNF treatment do not affect the response to TNF cytotoxicity. Thus, it is during the 24 h priming period before TNF treatment that oxygen tension is critical in determining TNF response.

The time-course of hypoxia-induced resistance to the cytotoxic action of TNF was assessed (c/3.1.8) and results are shown in Figure 3.5. The data show that it takes at least 14 h for cells to become significantly resistant to TNF and that resistance is further induced by 24 h priming in 2% oxygen. The time-course of resensitization of SaF cells to TNF after priming in 2% oxygen was assessed ( c /3.1.9) and results are shown in Figure 3.6. Cells were primed in 2% oxygen for 24 h, removed to 21% oxygen, and subsequently cells were treated with TNF at various intervals (0-24 h). Figure 3.6 shows that resensitization occurs after at least 5 h in 21% oxygen and by 24 h in 21% oxygen SaF cells are fully sensitive to TNF.

Having determined that a 24 h exposure of SaF cells to hypoxia (2% oxygen) causes a transient resistance to TNF which is reversible (within 5-24 h in 21% oxygen), it was then investigated whether exposure of SaF cells to hypoxia through several passages had the same effect (cf 3.1.10). Figure 3.7 shows the effect of continuous culture of SaF cells under conditions of 2% oxygen. TNF cytotoxicity assays were performed after each passage (weekly) for up to five passages. The results show that after five passages in 2% oxygen cells require at least sixty times more TNF to induce 50% cell kill than those exposed to 21% oxygen for 24 h. Thus, over the time-course investigated, hypoxia-induced resistance to TNF increases significantly (p <0.01).

69 110 - 21 % oxygen 5% oxygen 00 - 95% oxygen 90 - 0.2% oxygen 2% oxygen 80 - o h 7 0 - o 6 0 - o Ë: 5 0 - Q ^ 40 -

30 -

20 -

10 -

0.1 1 10 rhuTNF-a (U/ml)

Figure 3.4 The effect of treating SaF cells with rhu-TNF at tumour relevant oxygen tensions after 24 h priming in 21% oxygen. Symbols show the mean ± s.e.m of three experiments (n = 9).

70 * * *

* * * 0 .4 -

a 0 .3 - u PL4 % H * * * * * * U"c u

0.1 -

0.0 4 12 14 16 18 20 24 Control TNF Addition Time (h)

Figure 3.5 Time-course of hypoxia-induced resistance to the cytotoxic effects of rhuTNF on SaF cells. Cells were primed in 2% oxygen for a period of time (above) before treatment with TNF in 21% oxygen. Columns show the mean ± s.e.m of 3 experiments (n=9); *p <0.05; *** p <0.01.

71 4H

3 -

b

2 -

U

1 -

i X X X X

0 0 1 3 5 7 14 16 18 20 24 Control TNF Addition Time (h)

Figure 3.6 Time-course of resensitization of SaF cells to rhu-TNF after 24 h priming in 2% oxygen. Columns show the mean ± s.e.m of 6 experiments (n=18).***p<0.01.

72 * * * 14-

1 2 -

1 0 - * * *

o U I— ( 6 -

4 - * * * * * *

2 -

Control 1 2 3 4 5 Passage Number

Figure 3.7 The effect of chronic exposure to 2% oxygen on the response of SaF cells to the cytotoxic effects of rhu-TNF. SaF cells were continuously grown in 2% oxygen and tested for response to TNF after each passage (weekly). Columns show the mean ± s.e.m of 3 experiments (n=9); ***p<0.01.

73 3.3 Discussion

I The present studies demonstrate that pre-treatment oxygen tension is a key modulator of the cytotoxic action of TNF in a range of human and murine cell lines. Detailed studies on the SaF tumour cell line have shown that tumour cells cultured at oxygen tensions relevant to those found when they grow as solid tumours in vivo are more resistant to the cytotoxic effect of TNF compared to those grown under conventional tissue culture conditions (i.e. 21% oxygen). SaF cells cultured for 24 h in an atmosphere of 2% oxygen are 5-10 times more resistant to rhuTNF compared to cells grown in an atmosphere containing 21% oxygen, depending on the assay employed. Clearly, there are differences in quantifying the hypoxia-induced TNF-resistance to the cytotoxic action of TNF (Figure 3.1). The colourimetric assay assesses short-term cytotoxicity by determining the proportion of cells remaining attached to the plate according to their membrane permeability (ability to incorporate crystal violet) shortly after TNF treatment. By contrast, the clonogenic assay assesses the long-term cytotoxic effects by determining the retention of regenerative capacity, or clonogenicity. Thus, the clonogenic assay is not sensitive to short-term anti-proliferative effects of TNF whereas the colourimetric assay will have some sensitivity to such effects. As a consequence, cells primed in 2% oxygen appear to have a relatively increased sensitivity to TNF cytotoxicity when the colourimetric assay is employed {cf Figure 3.1). Thus, the colourimetric assay utilised throughout these studies gives an adequate assessment of the short-term anti-proliferative/cytotoxic effects of TNF at the expense of underestimating the protective effect of hypoxia on the long-term cytotoxicity of TNF. Clearly, this may indicate that hypoxia is more effective at protecting against cytotoxicity than against the anti-proliferative effect of TNF. However, the level of hypoxia-induced resistance appears to be dependent not only on the assay employed but more importantly, on both the level of oxygenation and the duration of exposure to a given level of oxygenation. The 5-fold increase in resistance to TNF by priming in 2% oxygen is increased 60-fold when cells are grown for five passages in 2% oxygen {cf Figure 3.7) and 500-fold if cells are primed in 0.2% oxygen for 24 hours. Hypoxia-induced resistance occurs to both human and murine TNF although the effect is more dramatic with human TNF. This difference in effect may be caused by receptor specificity: rhuTNF can activate only the p55 TNF-R in murine cells, which is

74 the receptor responsible for the cytotoxic effect of TNF, whereas rmuTNF activates both the p55 and p75 TNF-Rs. Thus, activation of the p75 receptor by rmuTNF may reduce the oxygen effect by increasing p75 TNF-R functions, such as cell proliferation. Clearly, hypoxia-induced resistance occurs to both murine and human TNF, albeit less pronounced with rmuTNF. It is the pre-treatment oxygenation status of cells which determines response to TNF since no change in response is observed if cells are gassed with between 0.2% and 21% oxygen during the 24 h incubation with TNF (c/Figure 3.4). These results are in contrast to earlier reports which indicate that anoxia during TNF exposure reduced the cytotoxic effect (Matthews et al. 1987; Park et al 1992). However, those studies were performed under true anoxia, where potential oxygen contamination was eradicated by all means, including the utilisation of glassware rather than plastic ware, in which oxygen is dissolved (Chapman et al 1970). By contrast, the present studies were undertaken using plastic 96-well plates in which oxygen is dissolved and thus it is not possible to achieve complete anoxia. It has been determined that gassing with 95% nitrogen/5% CO 2 for 1.5 hours results in a local oxygen tension of 0.2% oxygen (Chapman et al 1970). Thus, treating cells with TNF in an oxygen tension of 0.2% oxygen as opposed to complete anoxia could explain the apparent anomaly between this and previous reports. It has been established that resistance to TNF cytotoxic action occurs within 14 h of hypoxic priming {cf Figure 3.5). However, the increased resistance induced by a hypoxic environment is completely reversed by placing cells back in an environment of 21% oxygen for at least 5 h prior to TNF exposure {of Figure 3.6). A similar resensitization effect has been reported for L929 cells (Sampson and Chaplin, 1994). However, in the present study significant resensitization is achieved by re-exposing the cells to 21% oxygen for 5 h whereas with L929 cells no resensitization was observed at time points less than 7 hours. In this part of the study the mechanism responsible for the increased resistance to TNF mediated cytotoxicity in cells pre-incubated at tumour relevant oxygen tensions has not been investigated. It is possible that there may be both specific and non-specific protective mechanisms operating to induce resistance to TNF cytotoxicity. Although the mechanisms by which cells become resistant to TNF are still surrounded by considerable controversy several mechanisms can be postulated with respect to hypoxia- induced TNF-resistance:

75 There is convincing evidence that reactive oxygen species (ROS) mediate, at least in part, the cytotoxic activities of TNF (Yamauchi et al. 1989; Chang et al. 1992; Goossens et al. 1995). Thus, if hypoxia induces the production of proteins which protect the cell against ROS, increased TNF resistance would occur. The only enzyme that is practically always induced by TNF which is involved in ROS metabolism and has been associated with protection from TNF cytotoxicity, is MnSOD (Wong and Goeddel, 1988; Asoh et al. 1989; Kawaguchi et al. 1990; Shaffer et al. 1990; Warner et al. 1991; Tsan et al. 1992; Del Maestro et al. 1992; Chang et al. 1992; Hirose et al. 1993; Tsan, 1993). Interestingly, Ketis and Jones (1992) found an increase in superoxide dismutase levels during hypoxia and suggested that ROS are generated during hypoxia, however, this has not been generally observed. A possible protective protein is TNF itself, as endogenous TNF production has been reported to protect cells against exposure to exogenous TNF (Himeno et al. 1990; Vanhaesebroeck et al. 1992). Indeed, it has been shown that exposure to hypoxia increases TNF production and secretion by peripheral blood mononuclear cells stimulated by endotoxin (Ghezzi et al. 1991). In addition, it was shown recently that hypoxia induces an unstimulated human monocyte cell line to release TNF and its soluble receptors in vitro (Scanned et al. 1993). Another possible explanation is that receptor number, affinity, or rate of internalisation of the receptor-ligand complex is altered by prior exposure to low oxygen environments. Indeed, hypoxia can modulate receptor affinity, as shown for a l adrenergic receptors on cardiac myocytes (Corr et al. 1981). Moreover, Gerlach et al. (1993) found that hypoxia increases both affinity and number of TNF-Rs and thus causes an increase in the binding of TNF to endothelial cells. It has been shown that cell cycle plays a role in TNF cytotoxicity and that accumulation of cells in G1 induces resistance to TNF cytotoxicity (Watanabe et al. 1987; Belizario and Dinarello, 1991; Pusztai et al. 1993; Darzynkiewicz et al. 1984). It is also well established that hypoxia modulates cell cycle (Born et al. 1976; Shrieve et al. 1983) and that cells exposed to extreme hypoxia arrest in G1 (Koch et al. 1973a; Bedford and Mitchell, 1974; Merz and Schneider, 1983). Thus, if tumour relevant hypoxia arrests SaF cells in G1 this could be responsible for the induced resistance to TNF cytotoxicity. Regardless of the mechanism of hypoxia-induced resistance to TNFs cytotoxic action it is not surprising that high doses of TNF are necessary to have a therapeutic

76 effect in vivo as the majority of tumour cells exist in an environment which is largely hypoxic. This study suggests that improving tumour oxygenation status before TNF treatment may overcome resistance induced by hypoxia.

3.4 Summary

This study emphasises the importance of performing in vitro experiments at relevant oxygen tensions as it has been clearly established that oxygenation conditions relevant to those found in solid tumour systems can profoundly effect tumour cell sensitivity to the cytotoxic action of TNF. Exposure of cells to tumour oxygen tensions (0.2-5% oxygen), when compared to well oxygenated conditions, prior to TNF treatment induces resistance to the cytotoxic action of TNF in vitro. It is possible that tumour oxygen tensions may modulate many of the other activities of this pleiotropic cytokine which have not been investigated in this study but may have important consequences in therapy with TNF. It is also possible that tumour oxygen tensions may modulate the activity of many other cytokines and growth factors in vivo. Several of the mechanism(s) of resistance proposed above are investigated in the forthcoming sections to gain insight into the cytotoxic effects of TNF in vivo, in an attempt to explain its inadequacy as a therapeutic agent for the majority of tumours.

77 Chapter 4

Mechanism of hypoxia-induced resistance to the cytotoxic action of TNF: cell cycle

4.0 Introduction 4.1 Materials and methods 4.1.1 Principle of BrdUrd / PI cell cycle analysis 4.1.2 Experimental plan 4.1.2.1 Effect of priming cells in 21%, 2% and 0.2% oxygen for 24 h on SaF cell growth 4.1.2.2 Effect of priming cells in 21% and 2% oxygen for 0-4 days on SaF cell growth 4.1.2.3 Effect of priming cells in 0.2%, 1%, 2%, 5% and 21% oxygen for 24 h on the distribution of SaF cells in the cell cycle 4.1.2.4 SaF cell cycle analysis DURING priming in 21%, 2% and 0.2% oxygen 4.1.2.5 SaF cell cycle analysis AFTER priming in 21% and 2% oxygen, with and without TNF/AMD or AMD only, treatment 4.1.3 Procedure for BrdUrd/ PI staining of cells 4.1.4 Flow cytometry: The FACScan 4.1.5 Data acquisition and analysis 4.1.5.1 Lysysn 4.1.5.2 Cellfit 4.1.6 Cell cycle parameter analyses 4.1.6.1 Labelling index 4.1.6.2 Duration of cell cycle 4.1.6.3 Duration of S phase 4.1.6.4 Duration of G2/M phase 4.1.6.5 Duration of G 1 phase 4.1.6.6 G1 delay 4.1.6.6.1 Early S unlabelled 4.1.6.6.2 G1 unlabelled 4.1.6.7 G2 delay

78 4.1.6.7.1 G2 delay of S phase cells 4.1.6.7.2 G2 delay of G2 cells 4.2 Results 4.2.1/ 2 Effect of priming cells in 21%, 2% and 0.2% oxygen on cell growth 4.2.3 Effect of priming cells in 0.2%, 1%, 2%, 5% and 21% oxygen for 24 h on the distribution of cells in the cell cycle 4.2.4 SaF cell cycle analysis DURING priming in 21%, 2% and 0.2% oxygen 4.2.5 SaF cell cycle analysis AFTER priming in 21% and 2% oxygen, with and without TNF/AMD or AMD only, treatment 4.3 Discussion 4.4 Summary

79 4.0 Introduction

The cell cycle refers to the distinct successive stages required to replicate DNA that can be discerned in actively dividing cells. There are four discrete phases in the cell cycle which were originally described by Howard and Pelc (1951): G1 (the first Gap phase), S phase (DNA synthesis phase), postsynthetic G2 (the second Gap phase) and M (mitosis). In the presynthetic G1 phase the cell prepares and commits itself for subsequent DNA synthesis. The G1 phase is followed by the S phase in which the cell replicates its DNA. In the postsynthetic G2 phase the cell monitors whether DNA replication has been successfully completed and prepares for cell division by synthesising the appropriate proteins required in mitosis. The transition between successive phases of the cell cycle is strictly controlled by cell division control (GDC) genes and regulatory proteins. The cyclins and cyclin-dependent kinases (CDKs), which together form of CDK/cyclin complexes, are positive regulators of the cell cycle and ensure the alteration of S and M phases (reviewed by Karp and Broder, 1995). CDK/cyclin complexes are in turn regulated by CDK-inhibitors (CKIs) (reviewed by Nigg, 1995). It is well established that cell growth and cell cycle control are affected in a fundamental manner by hypoxia (Born et al. 1976; Shrieve et al 1983; Shrieve and Begg, 1985, Âmellem and Pettersen, 1991; Amellem et al 1994). In general, cells exposed to extreme hypoxic conditions in S-phase immediately arrest, while cells in other phases of the cell cycle proceed to late G1 before they become arrested (Koch et al 1973a; Bedford and Mitchell, 1974; LoefHer et al 1987; Pettersen and Lindmo, 1981). Hypoxia does cause some cells to arrest in G2 or mitosis (M) but most cells are able to complete mitosis and divide in the absence of oxygen (Shrieve et al 1983; Shrieve and Begg, 1985). However, when cells are arrested, the oxygen-dependent restriction point is late G l, where, it is thought, cells arrest to prevent hypoxia-induced damage in the critical S phase (Merz and Schneider, 1983; Spiro et al 1984; Amellem and Pettersen, 1991). The mechanism of Gl arrest has not been fully elucidated but it is not caused by cellular energy deficit (Loeffler, 1985) but may be due to hypoxia-induced gene expression. Hypoxia induces accumulation of p53 protein, but activation of a Gl- phase checkpoint by hypoxia is independent of p53 status (Graeber et a l 1994). Cell cycle position is also an important factor that determines sensitivity of cells to the cytotoxic effects of TNF (Darzynkiewicz et al 1984). Cells are highly sensitive to

80 TNF when they are in G2/M phase, which correlates with maximum TNF-R levels (Watanabe et al 1987). By contrast, cells are resistant to the cytotoxic effects of TNF when they are in Gl phase (Belizario and Dinarello, 1991; Pusztai et al 1993), the oxygen-dependent restriction point. Furthermore, it is known that if cells are in G l when they are re-oxygenated (after extreme hypoxia) then there is no delay in the initiation of DNA synthesis. However, if they are in G2, a delay is encountered on reoxygenation (Amellem and Pettersen, 1993). Thus, the aim of the following investigations was to ascertain whether priming SaF cells in oxygen tensions known to exist in vivo (hypoxia) arrests cells in G l, the phase in which resistance to TNF cytotoxicity has been reported. Central to this aim was the need to perform the TNF cytotoxicity experiments {cf Chapter 3) in the context of the cell cycle.

The cell cycle was followed by 5-bromo-2’-deoxyuridine (BrdUrd) pulse labelling, during a 24 h priming period in 0.2%, 2% and 21% oxygen, and also, in 21% oxygen, after priming in 2% and 21% oxygen, during which cells were treated with AMD or TNF/AMD. As resistance to the cytotoxic effects of TNF is significantly increased after 24 h priming in hypoxia, the proportion of cells in each phase of the cell cycle was assessed at this time point to ascertain if a higher proportion were residing in Gl. If this were established it may explain the mechanism of hypoxia-induced resistance to the cytotoxic effects of TNF in vitro.

4.1 Materials and methods

4.1.1 Principle of BrdUrd / propidium iodide cell cycle analysis BrdUrd is a thymidine analogue which is incorporated into DNA only during synthesis (S-phase). Propidium iodide (PI) is a fluorescent dye, which intercalates between the bases in double stranded nucleic acids and this produces red fluorescence when excited by blue light. The fluorescence is proportional to the nucleic acid content and thus is a measure of relative DNA content. A flow cytometric method which measures total DNA (PI) content and BrdUrd content simultaneously was employed. This method was originally developed by Dolbeare et û/.(1983).

In practice, cells are pulse labelled with BrdUrd for 20 minutes and then stained for their BrdUrd content at different time intervals using an anti-BrdUrd monoclonal

81 antibody. They are then reacted with a secondary antibody conjugated with fluorescein, (usually fluorescein isothiocyanate: FITC), and then counterstained for their DNA content with PL PI does not cross intact plasma membrane, therefore cells must be either dead, lysed, or fixed, for binding to occur. The double stained cells are then analysed in the flow cytometer (FACScan) for red (DNA) and green (BrdUrd) fluorescence. The results are displayed as a bivariate distribution of BrdUrd versus DNA content (c/Figure 4.1) and analysis ‘windows’ can be set anywhere in the cell cycle by choosing the appropriate part of the DNA histogram or dot plot. Such pulse-chase experiments allow the movement of BrdUrd-labelled cells through the cell cycle to be followed and the effect of any treatment accurately assessed.

4.1.2 Experimental plan To investigate: 1. Effect of priming cells in 21%, 2% and 0.2% oxygen for 24 h on SaF cell growth 2. Effect of priming cells in 21% and 2% oxygen for 1-4 days on SaF cell growth 3. Effect of priming cells in 0.2%, 1%, 2%, 5% and 21% oxygen for 24 h on the distribution of SaF cells in the cell cycle 4. SaF cell cycle analysis DURING priming in 21%, 2% and 0.2% oxygen 5. SaF cell cycle analysis AFTER priming in 21% and 2% oxygen, with and without TNF/AMD or AMD only, treatment

4.1.2.1 Effect of priming cells in 21%, 2% and 0.2% oxygen for 24 h on SaF cell growth Exponentially growing SaF cells were harvested and seeded at 1 x 10^ cells per 9 cm petri dish. Cells were primed in 0.2%, 2% and 21% oxygen for 24 h. During the priming period cell growth was measured by taking samples hourly, starting immediately after seeding. Samples were trypsinized, centrifuged at 2000 rpm for 5 minutes, resuspended in ice-cold PBS (0.2 ml) and fixed with 10 mis of ice cold 70% ethanol. Fixed cells were counted using a haemocytometer.

4.1.2.2 Effect of priming cells in 21% and 2% oxygen for 1-4 days on cell growth Exponentially growing SaF cells were harvested and seeded at 7.5 x lO"^, 1.25 x

10^, 2.5 X 10^, or 5 X 10^ cells per 9 cm petri dish in order to have approximately 1 x

82 10^ cells after 1-4 days priming, respectively. Cells were primed in 21% and 2% oxygen for 24 h (5 X 10^), 48 h (2.5 x 10^), 72 h (1.25 x 10^) and 96 h (7.5 x 10^). After each priming period samples were taken for the following 24 h and prepared for counting as described in 4.1.2.1 above.

4.1.2.3 Effect of priming cells in 0.2%, 1%, 2%, 5% and 21% oxygen for 24 h on the distribution of SaF cells in the cell cycle This was determined by two different staining procedures. Cells were either stained with PI only or with BrdUrd and PI. SaF cells were harvested, seeded at 1 x 10^ cells per 9 cm petri dish and triplicates were primed in 0.2%, 1%, 2%, 5% and 21% oxygen for 24 h. After priming, cells were trypsinized, pelleted, resuspended in PBS (0.2 ml), fixed in 10 ml of ice-cold ethanol (70%), and then stained with PI at a final concentration of 10 pg/ml. Alternatively, after priming cells were pulse labelled with BrdUrd (20 pM) for 20 min (except control samples). BrdUrd was prepared when required, at a stock solution of 200 pM in medium, and 1 ml of stock was added quickly to 9 ml of culture medium. Petri dishes were replaced in their respective oxygen tensions during the 20 min BrdUrd labelling period. Samples were then prepared for anti-BrdUrd staining: cells were washed twice with warm medium (5 ml), trypsinized, centrifuged at 2000 rpm for 5 min, resuspended in ice-cold PBS (0.2 ml), fixed with 10 mis of ice cold ethanol (70%) and stored at 4°C until required for staining, as described in 4.1.3 below.

4.1.2.4 SaF cell cycle analysis DURING priming in 21%, 2% and 0.2% oxygen SaF cells were harvested, seeded at 5 x 10^ cells per 9 cm petri dish and grown in EMBM 4- 10% FCS in 21% oxygen until exponential growth was achieved (1-2 days). At the beginning of the 24 h priming period, in 21%, 2% or 0.2% oxygen, cells were pulse labelled with BrdUrd (20 pM) for 20 minutes (except control samples). The petri dishes were replaced in their respective priming oxygen tensions (21%, 2% or 0.2% oxygen) during the 20 min BrdUrd labelling period. After 20 min the BrdUrd containing medium was discarded and the cells were washed carefully with PBS at 37°C (5 ml x 2). Cells were replenished with fresh warm medium and replaced in their respective priming oxygen tensions. Controls and 0 h treatments were sampled 40 minutes after BrdUrd treatment. Other samples were trypsinized every hour up to 16 h, then every 2 h up to 24 h. After trypsinization, cells were centrifuged at 2000 rpm for 5 min,

83 resuspended in ice-cold PBS (0.2 ml), fixed in 10 ml of ice-cold ethanol (70%) and stored at 4°C before staining with anti-BrdUrd and PI, as described in 4.1.3 below.

4.1.2.5 SaF cell cycle analysis AFTER priming in 21% and 2% oxygen, with and without TNF/AMD or AMD only, treatment Exponentially growing SaF cells were harvested and seeded at 5 x 10^ cells per 9 cm petri dish. Cells were primed in 2% and 21% oxygen for 24 hours. After priming all samples (except controls) were treated with BrdUrd (20 |iM) for 20 minutes. After BrdUrd labelling cells were washed carefully with PBS at 37°C (5ml x 2) and treated with rhuTNF (0.25 lU) and/or actinomycin D (1 |ig/ml) (AMD), or medium only, for 24 hours. All treatment groups were replaced in 21% oxygen at 37°C and samples were taken every hour up to 16 h, then every 2 h up to 24 h. Samples were prepared for both assessment of cell growth (c/4.1.4.1 above) and anti-BrdUrd/ PI staining :

4.1.3 Procedure for BrdUrd and PI staining of cells BrdUrd-pulse-labelled SaF cells were counted using a haemocytometer and an aliquot of one million cells removed. The cells were resuspended in 2.5 ml of 2 M HCL containing 0.1 mg/ml pepsin solution. The suspension was incubated, with occasional mixing, at room temperature for 20 min, to release the cell nuclei and partially denature DNA. Acid was neutralised with 3 ml of 0.1 M borax solution and then centrifuged at 2000 rpm for 5 minutes. The cell pellet was washed with 3 ml PBS and centrifuged at 2000 rpm for 5 min, this was repeated. The cell pellet was resuspended in 0.5 ml PBS containing 0.5% normal goat serum (NGS), 0.5% Tween-20 and mouse anti-BrdUrd monoclonal antibody, 25 pi, (dilution 1:20) and incubated for at least 1 h at room temperature, mixing occasionally. PBS (5 ml) was added, the cells were centrifuged at 2000 rpm for 5 min and then resuspended in 0.5 ml PBS/NGS/Tween-20 and anti­ mouse IgG FITC, 25 pi, (dilution 1:20) was added. The secondary antibody was incubated for 30-60 min at room temperature, mixing occasionally. The cells were then washed in 5 ml of PBS, centrifuged at 2000 rpm for 5 min then resuspended in 2 ml of PBS. Propidium iodide, was added at a final concentration of 10 pg/ml and the sample vortexed and analysed using a Becton-Dickinson FACscan.

84 4.1.4 Flow cytometry :The FACScan The cell cycle was assessed by flow cytometry using a Becton-Dickinson FACScan. The FACScan is an automated cell analyser attached to a computer. It is capable of measuring five optical parameters simultaneously:- forward scatter, side scatter, and three spectral regions of fluorescence. It has three high performance photomultipliers with band pass filters of 530 nm, 585 nm and >650 nm. The FACScan has an air-cooled 15 milliwatt Argon-ion laser with a single excitation wavelength of 488 nm.

The single cells enclosed within a pressurised saline solution pass through the flowcell generating five signal pulses simultaneously from the optical detectors. These pulses are processed by an analog-digital converter and then stored and processed by the computer system which enables single and multiple parameter data analysis.

4.1.5 Data acquisition and analysis 4.1.5.1 Lysys II Lysys II FACScan software, which allows the collection and analysis of the light emission data, was used to analyse samples stained with BrdUrd/PI as described in 4.1.3 above. The FITC signal was collected in FLl (515-545 nm) and PI in FL3 (>650 nm). This set-up negates spectral overlap between the fluorochromes. A double discrimination module was used to process the DNA signal into height, area and width to allow separation of cell doublets from the single cell population. Doublets were removed by gating around a singlet population on an area versus width plot of the DNA channel. Data for 10,000 single cells was collected for analysis of each sample. In analysis mode any combination of the collected data could be displayed as histograms or dot plots. Statistical analyses were carried out by analysis of computer generated regions of the data (c/Figure 4.1).

4.1.5.2 Cellfit Cellfit is a comparative data package to Lysys II but was used for analysis of cells stained with PI only (i.e. DNA profiles) as described in 4.1.2.3. DNA histograms were used to determine the distribution of cells within the cell cycle by calculating the percentage of cells in Gl, S and G2/M. The cellfit model which gave the best fit of the DNA profile was the SOBR (sum of broadened rectangles) model.

85 SOBR assumes that cells are distributed in all phases of the cell cycle. A complex formula fits the GO/Gl and G2/M populations with single Gaussian curves producing successive approximations to the actual histogram. A selected number of Gaussian convolved rectangles fit the S phase. The SOBR fit estimates the initial percentage coefficient of variance (% CV) by determining the peak width at the inflection point (60% of peak height) and then generates the final reported % CV during the iterated fit. The reported % CV is determined by the final Gaussian curve that is fitted to the GO/Gl peaks.

4.1.6 Cell cycle parameter analyses To determine if SaF cell cycle is modulated during or after priming in 21%, 2% or 0.2% oxygen or by TNF/AMD or AMD only treatment after priming (c /4.1.2.4 and 4.1.2.5), the following cell cycle parameters were analysed: labelling index, duration of cell cycle, G l, S and G2/M phases, delays in Gl and G2 phases. Data analysis was generally performed on bivariate dot plots (DNA versus BrdUrd content) with regions set around the various subsets of cells required to calculate the above parameters. Figure 4.1 shows a typical dot plot with appropriate regions and its corresponding histograms.

86 AB n M 3 1 = 1

M 2 _ M 4 M 5

1023

C D

M 3 oo I

Ml M 4

0 1023

Ml = Total cells M4 = 02 window M2 = Mid S window M5 = Early S window M3 = Gl window

Figure 4.1 Cell cycle parameters were calculated from: (A) Bivariate dot-plot of SaF cells showing BrdUrd uptake (vertical axis) and DNA content (horizontal axis) with a region set around total cells BrdUrd labelled (Rl). (B) Histogram corresponding to dot-plot (A) showing total DNA where the markers (M) denote the regions set (see above). (C) Histogram of BrdUrd labelled DNA (Rl of (A)); (D) Histogram of BrdUrd unlabelled DNA.

87 4.1.6.1 Labelling index (LI) The labelling index (LI) gives the fraction of the total cell population synthesising DNA at any instant. It can be calculated by setting a region or marker around the cells with significant BrdUrd uptake (Rlof (A)) and this is expressed as a percentage of the total number of cells. The LI was assessed during the pulse-chase experiments taking into account cells which divided and resided in G 1, as follows:

Corrected LI = (Rl - (0.5 x M3(C)) / (M1(B) - (0.5 x M3(C)) x 100

4.1.6.2 Duration of cell cycle (Tc) The cell cycle time, Tc, is the time taken for the cells to complete one cycle. It was calculated by measuring the passage of cells through a narrow window in mid-S phase (M2 in Figure 4.1). The ratio of BrdUrd labelled cells (M2(C)) to the total number of cells (M2(B)) in this window was plotted against time and Tc was estimated as the time interval between the first and second peaks. A more accurate method of determining Tc was performed using the Mackintosh^'^ IMP package which fits a curve to the graph of percentage mid S phase versus time and calculates the midpoint of the second peak using a fitting procedure designed specifically for the purpose of obtaining a Tc value. This also allows the calculation of a standard error on the Tc value obtained.

4.1.6.3 Duration of S phase (Ts) The length of S phase (Ts), the time taken for cells to complete DNA synthesis, was calculated using the IMP Mackintosh^'^ analysis from the same curve fitted to calculate Tc, as described above.

4.1.6.4 Duration of G2/M phase (Tg 2m)

The duration of G2/M (T g 2/m) was calculated by plotting the entry of BrdUrd labelled cells into Gl (M3(C)) versus time. From this graph the length of G2/M phase can be calculated by fitting a regression line to the data points on the linear portion of the curve, this is where cells are entering Gl from G2/M phase. The extrapolation of the regression line to the x-axis denotes Tg 2/m- 4.1.6.5 Duration of G l phase (T gi )

The duration of the remaining cell cycle phase, Gl, (T g i) was obtained by subtracting Tq 2/m and Ts from Tc.

T g i = T c - (Ts + T g 2 /m )

4.1.6.6 G l delay 4.1.6.6.1 Early S unlabelled Gl delay was assessed by analysing the movement of Gl cells into early S (ES) phase window (M5(D)) and expressing this as a proportion of the total unlabelled population (M1(D)). This shows any delay of cells moving into or out of ES phase. Immediately after labelling the proportion of ES phase unlabelled cells is low as most cells in this phase are labelled. As cells move out of Gl into ES phase the proportion in ES phase increases, thus if the proportion in ES phase does not increase a delay or block in Gl is evident. If there is a delay or block in ES phase itself the proportion will not decrease after all the unlabelled cells from Gl have moved into this window.

%ES unlabelled = (M5(D)/M1(D)) x 100

4.1.6.6.2 Gl unlabelled The proportion of Gl unlabelled cells can also be used to calculate a delay in cells either entering or leaving Gl (M3(D)). Immediately after labelling this population represents -75% of unlabelled cells. With time, they leave Gl and enter S phase, simultaneously, cells enter Gl from G2/M phase. Thus a delay in cells leaving Gl into S phase will result in a build up of cells in Gl and the proportion of Gl unlabelled will remain high for the duration of the delay. Cells blocked in G2/M phase will show a decrease in the proportion coming back into Gl either immediately after labelling or after the first cell division. In order to calculate a Gl delay an IC 50 value is calculated for each sample, this requires a maximum and a minimum value for each curve. From the IC 50 value the difference in time can be calculated. A delay in cells leaving Gl can be seen on a graph as the slope of the curve will be shifted to the right.

% G l unlabelled = (M3(D) / M1(D)) x 100

89 4.1.6.7 G2 delay 4.1.6.7.1 G2 delay of S phase cells The entry of BrdUrd labelled cells into Gl (M3(C)) was used to calculate a 02 delay of S phase cells. This graph will show delay of S phase cells entering G2/M phase or a delay of G2/M phase cells entering Gl. The formula below corrects for dividing cells. G2 / M phase delay = (0.5 x M3(C) / (M1(C) - (0.5 x M3(C))) x 100

4.1.6.7.2 G2 delay of G2 cells The proportion of G2 unlabelled cells can show any delay of cells entering or leaving G2 (M4(D)). Initially the proportion of unlabelled cells in G2 is quite low (-25%). It decreases to practically zero as cells leave G2 and enter mitosis. Unlabelled cells from Gl must traverse S phase before reaching G2 and increasing the proportion of cells in this window. The time taken for the unlabelled cells to decrease to 0% will be prolonged if there is a delay in cells leaving G2, and the graph will slope to the right. A delay in cells entering G2 will prolong the time take for the proportion to increase and the curve will be shifted to the right again. A delay in G2/M phase can also be seen from the entry of BrdUrd labelled cells into Gl after the first cell division (c/4.1.6.6)

%G2 unlabelled = (M4(D) / M1(D)) x 100

4.2 Results

4.2.1/ 2 Effect of priming cells in 21%, 2% and 0.2% oxygen on SaF cell growth The proliferation of SaF cells after 24 h priming in 0.2%, 2% and 21% oxygen was assessed according to 4.1.2.1 and was found to be similar irrespective of the priming oxygen tension (data not shown). In addition, continuous priming in 2% and 21% oxygen was assessed according to 4.1.2.2 It was determined that the proliferation rate of SaF cells is not affected for at least 4 days after which cells primed in 21% oxygen proliferate faster than cells primed in 2% oxygen (data not shown).

90 4.2.3 Effect of priming cells in 0.2%, 1%, 2%, 5% and 21% oxygen for 24 h on the distribution of SaF cells in the cell cycle Table 4.1 summarises the data on the proportion of SaF cells in each phase of the cell cycle, independent of their proliferative status (c/4.1.2.3), after 24 h priming in tumour relevant oxygen tensions. These data were analysed using Cellfit (c /4.1.5.2), as cells were stained with PI only, which determines the proportion of cells in each phase by total DNA content, i.e, both proliferating and non-proliferating cells. Clearly, after 24 h priming in 0.2%, 1% and 5% oxygen the proportion of cells in Gl and G2/M is significantly higher (p <0.01) than in 21% oxygen controls and the proportion in S phase significantly lower (p <0.01). For cells primed in 2% oxygen the proportion in S phase and G2 is significantly lower (p <0.01) and the %G1 significantly higher than cells primed in 21% oxygen. However, there is no correlation between the priming oxygen tension and the proportion of cells in each phase of the cell cycle. Moreover, this experiment includes all cells and does not differentiate between fully functional cycling cells and non-proliferating, metabolically compromised cells.

Therefore, this experiment was substantiated by BrdUrd labelling, which accounts for the growth status of cells, in addition to PI, these results are summarised in Table 4.2. There are no significant differences between the proportion of cells in each phase of the cell cycle when cells are primed in 0.2% or 5% oxygen compared to cells primed in 21% oxygen. However, the %G1 is significantly higher when cells are primed in 1% oxygen (p <0.05) or 2% oxygen (p <0.01) compared to 21% oxygen. In addition, the %S phase is significantly lower (p <0.05) for cells primed in 2% oxygen compared to 21% oxygen. However, overall there is no correlation between the priming oxygen tension and the distribution of cells throughout the cell cycle.

91 Table 4.1 PI determination of the proportion of SaF cells in each phase of the cell cycle after 24 h priming in tumour oxygen tensions, compared to 21% oxygen.

Oxygen %S % G l %G2/M

(%) phase phase phase 0.2 39.4 (0.03)"** 34.8 (0.05)** 25.7 (0.05)** 1.0 35.0 (0.05)** 37.8 (0.03)** 27.2 (0.04)** 2.0 52.9 (0.09)** 33.5 (0.03)** 13.6 (0.05)** 5.0 40.6 (0.08)** 37.1 (0.04)** 22.3 (0.04)** 21.0 57.5 (0.09) 26.1 (0.05) 16.3 (0.04) ^ (standard error of mean), where n = 3; (i.e. three separate experiments). **p <0.01.

Table 4.2 BrdUrd / PI determination of the proportion of SaF cells in each phase of the cell cycle after 24 h priming in tumour oxygen tensions, compared to 21% oxygen

% % S % G l % G2/M Oxygen phase phase phase 0.2 53.6 (0.8)" 37.9 (2.0) 5.1 (0.9) 1.0 53.9 (0.8) 38.5(1.2)* 6.4 (0.4) 2.0 51.8 (0.4)* 41.2 (0.06)** 5.8 (0.6) 5.0 55.9(1.3) 36.3(1.4) 6.3 (0.4) 21.0 56.9(1.5) 33.2 (0.9) 5.9 (0.6) " (standard error of mean), where n=3 (i.e. three separate experiments). * p <0.05; ** p <0.01

92 The main difference observed in cells stained with BrdUrd/PI compared to PI alone is the %G2/M phase cells. When cells are stained with PI they are distributed fairly evenly over the cell cycle phases. However, the proportion of proliferating cells is actually greater in both S and Gl phases than G2/M phase after 24 h priming, as determined by BrdUrd/PI analysis.

4.2.4 SaF cell cycle analysis DURING priming in 21%, 2% and 0.2% oxygen Figure 4.2 shows the results of BrdUrd pulse labelling experiments (c /4.1.2.4) of SaF cells during priming for 24 h in 21%, 2% and 0.2% oxygen. Each panel represents a two parameter dot plot of BrdUrd content (green fluorescence on vertical axis) versus DNA content (red fluorescence on horizontal axis). Immediately after labelling (0 h) the BrdUrd labelled S phase cells, clearly identified by their strong green fluorescence, are positioned between Gl (left) and G2 (right) populations which have weak green fluorescence. The progress of the labelled cells through the cell cycle is easily visualised with time.

As the labelled cells move out of S phase into G2/M phase the green fluorescence shifts from the centre to the right. In mitosis the cells divide and green fluorescing cells then appear in Gl on the left of the profiles with approximately half of their original FITC labelling, due to cell division. The cell cycle is completed as labelled cells move from Gl back into S phase and the green fluorescence is central again. Clearly, a similar pattern of movement through the cell cycle was seen for cells primed in 0.2%, 2% and 21% oxygen although movement was visibly slower during priming in 0.2% and 2% oxygen.

93 1023

0 1023

(N o ,fc F I ■ f t

o: o: 1023 1023 1023

1023 1023

Figure 4.2 Distribution of SaF cells in the cell cycle during priming in 21% (left column), 2% (middle) and 0.2% oxygen (right). Cells were pulse labelled with BrdUrd and samples taken hourly for 14 h then every 2 h until 26 h. Data shown is 0 h, 6 h, 12 h and 18 h (top to bottom).

94 Table 4.3 summarises the LI data of SaF cells for experiments 4.1.2.4 and 4.1.2.5. It can be seen that the fraction of the total population synthesising DNA (LI) varies between experiments. In experiment 4.1.2.4, the LI was slightly higher in cells primed in 2% oxygen (0.81) compared to cells primed in 21% oxygen (0.63) after 24 h priming. However, in experiment 4.1.2.5, the LI was lower in cells primed in 2% oxygen (0.42) compared to cells primed in 21% oxygen (0.53). By contrast, the LI of cells primed in 0.2% oxygen is similar to that of cells primed in 21% oxygen (-0.6) implying that priming in 0.2% oxygen does not affect the ability of cells to incorporate BrdUrd. Thus, there are no clear differences in the LI of cells during priming in 0.2%, 2% and 21% oxygen.

Table 4.3 Labelling index of SaF cells during priming in 0.2%, 2% and 21% oxygen and after priming in 2% and 21% oxygen, with and without AMD or TNF/AMD treatment

Expt. Treatment P02 (%) Oh 24 h 4.1.2.4 During priming: 21 0.58 0.63 2 0.67 0.81 0.2 0.55 0.58 4.1.2.5 After priming: Controls 21 0.53 0.55 2 0.42 0.53

AMD (1 pg/ml) 21 0.55 0.54 2 0.43 <0.14

TNF/AMD 21 0.58 0.52 2 0.43 0.37

95 Figure 4.3 shows the progression of 0.2% and 21% oxygen primed cells through mid S phase from which Tc and Ts were calculated using Mackintosh''''^ IMP analysis. Initially the majority of cells (90%) in mid S phase are labelled, from 4 h they start to leave the mid S window and by 10-12 h the proportion of cells in this region has reached a minimum. The rate of entry into G2/M phase is not affected by pÛ 2 but the 21% oxygen primed cells exit G2/M phase and traverse Gl faster than 0.2% oxygen primed cells. Thus, it appears that 0.2% oxygen primed cells are slightly delayed both in exiting G2/M phase and in traversing Gl phase. Figure 4.4(a) shows that priming cells in 2% oxygen for 24 h has little effect on the progression of cells through mid S phase. Thus, during priming, Tc varied from 15.5 ± 3.2 h for cells primed in 21% oxygen to 17.5 ± 5.2 h for cells primed in 2% oxygen to 20.4 ± 3.8 h for cells primed in 0.2% oxygen. Cell cycle results are summarised in Table 4.4.

Table 4.4 Summary of SaF cell cycle results calculated by non-linear fitting of % mid S phase curves by Mackintosh^'^ JMP

Priming PÛ 2 Tc *stc Ts **sts T g2/M T gi (%)

During 21 15.5 3.18 10.9 1.31 2.3 2.3 2 17.5 5.2 9.6 0.99 3.0 4.9

After 21 17.6 4.02 11.9 1.09 3.0 2.7 2 16.3 3.98 10.2 0.89 2.0 4.1

During 21 16.3 3.22 12.3 3.14 2.2 1.8 0.2 20.4 3.78 14.9 2.86 3.3 2.2 *stc = standard error on Tc **sts = standard error on Ts

96 8 0 -

4 0 -

2 0 -

0 5 10 15 20 25 Time (h)

Figure 4.3 The movement of SaF cells, during priming in 0.2% (-■-) and 21% oxygen (-0-), through a narrow region (M2(C) / M2(D)) in mid S phase from which Tc and Ts were calculated.

97 100 - (a)

100 -

8 0 - 8 0 -

60 — 6 0 -

40 - 4 0 -

2 0 -

T— '— I— '— I—'— r 0 510 15 20 25 0 5 10 15 20 25

100 - 100 -

80 - 8 0 -

60 — 60 -

4 0 - 4 0 -

2 0 - 2 0 -

0 5 10 15 20 25 0 5 10 15 20 25 Time (h)

Figure 4.4 The movement of SaF cells through mid S phase (M2) (a) during priming in 21% oxygen (-0-) and 2% oxygen (-■-); and in 21% oxygen, after 24 h priming, (b) controls (c) treated with AMD (d) treated with TNF/AMD.

98 Priming cells in 0.2% oxygen causes cell cycle delays in Gl and 02 which are shown in Figures 4.5 and 4.6. The percentage of unlabelled cells in early S (ES) phase is shown in Figure 4.5(a). Initially there are very few unlabelled cells in ES phase as most cells in this phase are BrdUrd labelled. The proportion of unlabelled cells increases, slightly faster for 21% oxygen primed cells than 0.2% oxygen primed cells, as cells from Gl move into ES phase. There is an initial delay of 0.2% oxygen primed cells moving from Gl into ES (~1 h) and once in ES, cells are further delayed by approximately 3 h before moving into S phase.

Figure 4.5(b) shows the % BrdUrd labelled Gl cells, from which the length of G2/M phase was calculated as 2.2h for cells primed in 21% oxygen and 3.3h for 0.2% oxygen primed cells (c /4.1.6.4). Immediately after labelling the %G1 labelled cells is low (-10%) but as S phase labelled cells traverse G2/M phase into Gl it increases to maximum values by 12 hours. There is a delay of 0.2% oxygen primed cells entering Gl from G2/M phase (-2 h) thus %G1 labelled cells accumulates with a 2-fold increase in maximum levels compared to 21% oxygen primed cells. S phase cells are probably affected by transition through G2/M phase as 0.2% oxygen primed cells accumulate in Gl because of blocks in Gl and G2. Although priming cells in 0.2% oxygen causes cell cycle delays in Gl and G2 compared to priming cells in 21% oxygen, priming cells in 2% oxygen does not cause similar delays. During priming in 2% oxygen cells enter and leave Gl without delays (data not shown).

Figure 4.6(a) shows the proportion of 0.2% and 21% oxygen primed cells in the G2 unlabelled window. In both 21% and 0.2% oxygen, the G2 population exit this phase at the same rate suggesting that there is no delay experienced by the G2 population. At later time intervals, there are differences between 21% and 0.2% oxygen primed cells, but this is due to delays experienced by the original Gl population moving through Gl and S phase. From the previous plots, it is most likely due to a Gl delay. Figure 4.6(b) shows the status of Gl cells. There was a clear delay in 0.2% oxygen primed cells exiting this phase which again reinforces the existence of a Gl delay. It would appear that this delay is about 3 to 4 hours. However, priming cells in 2% oxygen did not cause delays of either the Gl or G2 populations (data not shown).

99 4.2.5 SaF cell cycle analysis AFTER priming in 21% and 2% oxygen, with and without TNF/AMD or AMD only, treatment

Figure 4.7 shows the growth curves of SaF cells after priming in 2% and 21% oxygen. Clearly, untreated control cells have similar proliferation rates irrespective of the priming oxygen tension. TNF/AMD treated cells proliferate at the same rate as controls for the first 8 h after which the rate decreases steadily until 24 h, when it completely ceases. However, treatment of cells with AMD only causes the proliferation rate to differ between oxygen tensions from 7 to 14 hours when cells primed in 21% oxygen have a higher proliferation rate. However, from 15 h onwards there is no difference in the proliferation rate between 2% and 21% oxygen and by 24 h the difference is negligible.

The LI also assesses the proliferation status of cells as non-proliferating cells will be unable to incorporate BrdUrd and thus have a low LI. Table 4.3 shows that the LI decreases after priming in 2% oxygen compared to 21% oxygen. Indeed, Figure 4.8 shows that when cells which were primed in 2% oxygen are replaced in 21% oxygen the LI is rapidly reduced to almost half its previous value for at least 5 h, but by 24 h in 21% oxygen it reaches control levels (c/Figure 4.8(a) and (b)). By contrast, the LI of 21% oxygen primed cells remains constant during and after the priming period. Figure 4.8 (c) and (d) show the LI of cells which were treated with AMD or TNF/AMD in 21% oxygen, after 24 h priming in 2% and 21% oxygen. The LI of cells which were primed in 2% oxygen is 7-fold lower 24 h after AMD treatment than that of 21% oxygen controls, however, TNF appears to overcome this effect to some extent.

100 1)

Time (h)

(b)

V. 40

10 15 Time (h)

Figure 4.5 SaF cells during priming in 0.2% oxygen (-■-) and 21% oxygen (-0-). (a) The movement of early S phase unlabelled cells into S phase (b) The % Gl labelled cells, from which the length of G2/M phase (Tg 2/m ) was calculated, by fitting a line to the straight part of the curve and extrapolating to y =0, as 3.3h for cells primed in 0.2% oxygen and 2.2h for cells primed in 21% oxygen.

101 (a)

3 0 -

0 5 10 15 20 25 Time (h)

(b)

8 0 -

0 5 10 15 20 25 Time (h)

Figure 4.6 SaF cells during priming in 0.2% oxygen (-■-) and 21% oxygen (-0-) (a) shows exit of G2 unlabelled cells into mitotic phase. (b) shows exit of unlabelled cells from Gl into S phase

102 5

4

" o % Vh 1(D 3 13 U 2

0 0 5 10 15 20 25 Time (h)

Figure 4.7 Proliferation of SaF cells after priming in 21% oxygen (open symbols) and 2% oxygen (solid symbols). After priming, cells were treated with TNF/AMD (-A-), AMD (1 |Ltg/ml) (-0-) only, or medium (-□-).

103 (a)

0.6 - 0.4 -

0.2 - 0.0 0 5 10 15 20 (b)

c 0.6 - O ■a 0.4 - o

2 0.2 - ÜH c/5 0.0 - GO 0 5 10 15 20 (c) 0.6 00 0.4

0.2

0.0 0 510 15 20 (d)

0.6 -

0.4 -

0.2 -

0.0 0 5 10 15 20 Time (h)

Figure 4.8 Labelling index of SaF cells (a) during priming in 21% oxygen (-0-) and 2% oxygen (-■-) for 24 h, and, in 21% oxygen, after priming in 2% and 21% oxygen (b) controls (c) treated with AMD( 1 |ig/ml) or (d) TNF/AMD.

104 Figure 4.4 shows the results of BrdUrd labelling of SaF cells in 21% oxygen after priming in 2% and 21% oxygen. Figure 4.4(b) shows that re-aeration of 2% oxygen primed cells does not affect Tc significantly, although it is slightly reduced to 16.3 ± 4 hours whereas Tc for 21% oxygen primed cells is increased to 17.6 ± 4 hours. AMD treatment inhibits proliferation completely in 21% but not 2% oxygen primed cells, which are allowed to progress slowly through S phase after 10 hours, as shown in Figure 4.4(c). TNF/AMD treatment damages 21% oxygen primed cells inhibiting them from progressing through mid S phase. Figure 4.4(d). However, hypoxia-induced resistance to the cytotoxic action of TNF/AMD enables 2% oxygen primed cells to gradually traverse mid S phase after 7 h, and by 24 h 60-80% of cells have progressed through S phase. TNF/AMD treatment induces a slight delay in 21% oxygen primed Gl unlabelled cells moving through ES phase. Indeed, TNF/AMD treatment induces delays of cells moving out of ES phase and completely blocks cells in both Gl and G2/M phases independent of p 0 2 (data not shown). Table 4.4 summarises the calculated cell cycle parameters after priming in 2% and 21% oxygen. Clearly, re-aeration of cells that had been primed in 2% oxygen does not significantly alter the duration of the cell cycle phases compared to 21% oxygen primed controls.

4.3 Discussion It has been known for many decades that exposure to hypoxia may cause cells to be delayed to varying degrees in some or all phases of the cell cycle (Pace et al. 1962; Koch et al 1973; Born et al 1976). More recently it has been recognised that TNF is cell cycle specific (Darzynkiewicz et al 1984; Watanabe et al 1987). The cytotoxic action of TNF is maximal when cells are in G2/M phase and minimal when cells are in GI phase (Watanabe et al 1987; Belizario and Dinarello, 1991). This study set out to determine if the hypoxia-induced resistance to the cytotoxic effects of TNF, as characterised in Chapter 3, could be explained by modulation of the cell cycle by hypoxia. More specifically, to mimic the TNF cytotoxicity assay and to determine if hypoxia causes accumulation of cells in Gl at the time of onset of resistance (14 h) or when resistance is highly significant (24 h) as it was reported that Gl cells are resistant to TNF , this study was performed.

105 SaF cell cycle was characterised both during priming in 0.2% oxygen (>500-fold induction of TNF-resistance) and in 2% oxygen (5-10 fold induction). It was established that the growth of SaF cells is not affected by priming in 0.2% or 2% oxygen for 24 h. This was confirmed by the LI data (Table 4.3) which shows that although there is inter­ experiment variation (2% and 21% oxygen) in the DNA synthesis fraction, there is no significant difference in the LI of cells during priming in 0.2%, 2% or 21% oxygen. Cell proliferation was an important factor to consider as it could possibly affect BrdUrd incorporation and give misleading results. Figure 4.7 shows that after 24 h priming in 2% and 21% oxygen cells grow exponentially, at the same rate, for the following 24 h. Indeed, it was determined that continuous priming of SaF cells in 2% oxygen does not affect cell growth for at least 4 days after which it is reduced, probably due to nutrient deprivation. This suggests that the TNF-resistance seen when cells are continuously primed in 2% oxygen is not due to modulation of growth rate by hypoxia. Table 4.3 summarises the LI data of SaF cells for experiments 4.1.2.4 and 4.1.2.5. It can be seen that the fraction of the total population synthesising DNA (LI) varies between experiments. After 24 h priming the LI could be slightly higher (0.81 versus 0.63) or lower (0.42 versus 0.53) in cells primed in 2% oxygen compared to cells primed in 21% oxygen. By contrast, the LI of cells primed in 0.2% oxygen is similar to that of cells primed in 21% oxygen (-0.6) implying that priming in 0.2% oxygen does not affect the ability of cells to incorporate BrdUrd. Thus, there are no clear differences in the LI of cells during priming in 0.2%, 2% and 21% oxygen. Interestingly, Jones and Bonting (1956) found that cells grow almost normally for at least 8 days in the total absence of oxygen. Furthermore, it was determined that for oxygen tensions greater than 500 ppm (0.05%) cell growth is minimally affected (Koch et al. 1973). In fact, many studies have shown that cells can survive and proliferate during a prolonged period of time at low oxygen tensions (Littbrand and Revesz, 1968; Anderson et al 1968; Koch et al 1973). Having established that cell growth is not affected by priming in 0.2% or 2% oxygen, the proportion of cells in each phase of the cell cycle after 24 h priming in tumour relevant oxygen tensions (0.2%, 1%, 2%, 5%) was assessed by two methods. Firstly, by staining only for total DNA content with propidium iodide (PI) (Table 4.1), and secondly, by staining for DNA synthesis (BrdUrd incorporation) in addition to DNA content (Table 4.2).

106 Table 4.1 shows that the distribution of cells throughout the cell cycle is significantly different between cells primed in 21% oxygen for 24 h and those primed in 0.2-5% oxygen. Cells primed in 0.2%, 1% and 5% oxygen show significantly lower proportions in S phase (p <0.01) and correspondingly higher proportions in Gl and G2/M phases (p <0.01) than cells primed in 21% oxygen. Initially this would suggest a relationship between oxygen tension and distribution of cells throughout the cell cycle. However, cells primed in 2% oxygen have a lower %G2/M cells than cells primed in 0.2% or 21% oxygen and there is no differential in the proportion of cells in Gl phase between 0.2% and 2% oxygen. Hence, cell cycle cannot account for the 100-fold difference in TNF-resistance between 0.2% and 2% oxygen previously determined. Clearly, there is no correlation between the priming oxygen tension and the proportion of PI stained cells in each phase of the cell cycle. Assessing the proportion of cells in each phase by staining with PI includes all cells regardless of their metabolic status and thus does not give an accurate analysis. However, the proportion of fully functional cycling cells was discerned by BrdUrd and PI staining. Table 4.2 presents the results of the proportion of actively proliferating cells in S phase and the corresponding Gl and G2 unlabelled populations of the cell cycle after 24 h priming in tumour oxygen tensions. The data show that there is no significant difference in the percentage of cells in each phase for cells primed in 0.2% or 5% oxygen compared to 21% oxygen. After 24 h priming in 2% oxygen there is a significantly higher proportion of cells in Gl phase (p <0.01) and a correspondingly lower proportion in S phase (p <0.05) compared to 21% oxygen. Similarly, after priming in 1% oxygen for 24 h a significantly higher proportion of cells reside in Gl (p <0.05). However, as the proportion of Gl cells is not increased above controls when cells are primed in 0.2% oxygen, where a 500-fold increase in TNF-resistance is observed, it is unlikely that this accumulation of cells in G l, when cells are primed in 1% and 2% oxygen, is responsible for induced TNF-resistance. This suggests that accumulation of cells in G1, by priming in hypoxia, is not generally responsible for the mechanism of TNF-resistance. Although priming SaF cells in oxygen tensions ranging from 0.2-5% did not cause an accumulation of cells in Gl after 24 h, this does not exclude the possibility of a Gl delay or arrest. A Gl arrest could possibly account for TNF-resistance as it has been reported that if hypoxic cells are in G 1 when they are re-oxygenated there is no delay in the initiation of DNA synthesis (Amellem and Pettersen, 1993). Thus, the next approach

107 was to follow the cell cycle during the 24 h priming period in 0.2%, 2% and 21% oxygen and later analyse the data for delays or arrests in cell cycle phases. The cell cycle was followed by BrdUrd pulse labelling during the priming period. It was found that priming cells for 24 h in 0.2% oxygen causes delays in all phases of the cell cycle but the most significant delay was encountered when cells traverse G l, cf Figure 4.5. Cells are delayed or arrested in Gl for 3-4 hours, after which they are delayed in moving from Gl into early S (ES) phase and also delayed leaving this phase, by ~3 h. During priming in 0.2% oxygen the original Gl population encounters a delay of ~2 h traversing G2/M phase causing almost a 2-fold higher percentage of 0.2% oxygen (37%) than 21% oxygen primed cells (19%) to accumulate in G2/M phase. Indeed, the delay in 0.2% oxygen cells progressing through Gl may explain why there are more than twice as many of these BrdUrd labelled Gl cells (43%) than 21% oxygen primed cells (19%) and probably accounts for the increased duration of Tc. This 2-fold difference in the amount of Gl labelled and G2 unlabelled cells occurs at 14 h, the same time as induction of TNF-resistance. However, it is unlikely that such differences in %G1 labelled and %G2 unlabelled, which result from a 3-4 h Gl delay, is entirely responsible for the 500-fold difference in sensitivity to TNF cytotoxicity exhibited by SaF cells after priming in 0.2% oxygen. These differences do not occur when cells are primed in 2% oxygen which also induces TNF-resistance. Furthermore, it has not been determined whether SaF cells in Gl phase are resistant to the cytotoxic action of TNF. It was observed that during priming in 2% oxygen for 24 h there are no major delays in the progression of cells through the cell cycle yet these cells are resistant to TNF cytotoxicity. However, there was a marginal delay in 2% oxygen primed cells leaving S phase and unlabelled cells enter G2/M phase slightly more slowly than controls (data not shown). Overall, following the cell cycle during priming in 2% oxygen did not reveal any significant delays which could account for the hypoxia- induced TNF-resistance. These results re-inforce the concept that cell cycle delays induced by priming cells in 0.2% or 2% oxygen are not responsible for hypoxia-induced TNF-resistance. However, it was not possible to predict the effect of re-aeration of 2% oxygen primed cells on cell cycle from the previous studies. Indeed, it was established that re­ aeration of 2% oxygen primed cells causes reduced DNA synthesis, shown in Figure 4.6, which is supported by a decrease in the proliferation rate to control levels, as shown by the labelling index data in Table 4.3. In addition, after priming the %G2 unlabelled

108 cells is 10% lower than controls which implies that priming cells in 2% oxygen causes cells to be delayed in either S or Gl phase. Indeed, the duration of Gl is increased after priming in 2% oxygen compared to 21% oxygen (4.1 h versus 2.7 h, respectively) which probably accounts for the higher proportion of Gl unlabelled cells between 8-13 h and at 24 h. However, although the Tgi is increased Ts and Tg 2/m are reduced and thus overall Tc is slightly reduced. Although Tc was prolonged by 2 h during priming in 2% oxygen it was slightly reduced after priming as major delays were not encountered. Thus, overall priming SaF cells for 24 h in 2% oxygen does not adversely affect the progression of cells through the cell cycle either during or after the priming period. When cells are treated with TNF/AMD after priming it takes at least 8 h before the proliferation rate decreases, independent of p 0 2 , which suggests that the uptake of TNF/AMD is the same in 2% oxygen as in 21% oxygen. Interestingly, as shown in Figure 4.6(c), treatment of SaF cells with AMD inhibited the progression of 21% oxygen primed cells through mid S phase whereas 2% oxygen primed cells appeared to gradually leave S phase and by 24 h almost 60% of cells have traversed S phase. Cells primed in 21% oxygen enter ES but are delayed exiting and are completely blocked in S phase, more than 80% of cells were still in S phase after a further 24 h in 21% oxygen (data not shown). Figure 4.8 shows that AMD treatment after priming does not affect the incorporation of BrdUrd (during DNA synthesis) in cells primed in 21% oxygen although it is slightly reduced at 24 h in 2% oxygen primed cells. AMD completely blocks SaF cells in Gl and G2/M phases independent of pOi (data not shown). Thus, it was not possible to determine Tc or the length of individual cell cycle phases for cells treated with AMD alone or TNF/AMD. Treatment of SaF cells with TNF/AMD, after priming, causes damage to cells that were primed in 21% oxygen and thus completely inhibits their progression through mid S phase (Figure 4.6(d)). By contrast, cells which were primed in 2% oxygen are resistant to the cytotoxic effects of TNF/AMD and although initially arrested for 7.5 h they very gradually progress through S phase and by 24 h almost 65% have entered G2/M phase. Thus, it appears that as cells primed in 2% oxygen are resistant to the cytotoxic action of TNF/AMD they are allowed to continue cycling albeit very slowly. Independent of p02, TNF/AMD treatment completely blocks cells in G2/M phase after 5 hours. This result is consistent with previous evidence that TNF exerts an early cytostatic effect in G2/M phase in mouse L-M cells (Darzynkiewicz et al, 1984). The

109 percentage of Gl labelled and 02 unlabelled did not increase beyond starting levels which suggests that cells are not allowed to enter 02 from S phase and therefore a block probably occurs in S phase or in Ol. Indeed, cells were completely blocked in 01 as determined by % Ol unlabelled (data not shown). Recently, Wan et al (1993) have shown that TNF may interrupt the entry of Ol cells into S phase, in vivo, with a minimal change in the rate of DNA synthesis. Thus, TNF/AMD allows cells primed in 2% oxygen to progress through S phase but blocks cells in all other cell cycle phases independent of pOi. Interestingly, TNF alone causes growth arrest in Ol phase, induces TNF- resistance (Belizario and Dinarello, 1991) and increases the expression of both p53 and p21 proteins (Jeoung et a l 1995). Cells that possess mutant p53 do not exhibit a Ol phase block when challenged by DNA damaging agents (reviewed by Cox and Lane, 1995). Thus, although the p53 and CDK/cyclin status of SaF cells is unknown, hypoxia may induce endogenous TNF (mutant p53?) which in turn causes TNF-resistance, this hypothesis is tested in Chapter 5. 4.4 Summary It has been determined that SaF cells proliferate equally well in 0.2% and 2% oxygen as in 21% oxygen for the 24 h priming period. In addition, cells primed in 2% oxygen have a similar proliferation rate as cells primed in 21% oxygen for up to 4 days. The proportion of Ol cells is increased only slightly after 24 h priming in 1% and 2% oxygen but not at all after priming in 0.2% or 5% oxygen. Thus an accumulation of cells in Ol is not responsible for induced TNF-resistance in SaF cells. Priming SaF cells in 0.2% oxygen for 24 h causes an increase in the length of the individual phases and thus the total cell cycle time. However, the only delay which may be significant is that of Ol, but this 3-4 h delay does not result in a significantly higher proportion of cells in Ol and thus is unlikely to account for a 500-fold increase in TNF-resistance. Priming cells for 24 h in 2% oxygen does not significantly modulate SaF cell cycle duration during or after priming. However, cells accumulate in O l, but as this has not been found for cells primed in 0.2% oxygen it is almost certainly not responsible for induced TNF-resistance. In conclusion, modulation of cell cycle by priming SaF cells in 0.2-5% oxygen does not appear to be responsible for induced resistance to the cytotoxic action of TNF. Thus, it is likely that there are other mechanisms involved in hypoxia-induced TNF- resistance.

110 C hapter 5

Mechanism of hypoxia-induced resistance to TNF cytotoxicity: production of MnSOD and TNF

5.0 Introduction 5.1 Materials and methods 5.1.1 Cell preparation 5.1.2 To determine whether freely diffusible protective/ sensitising factors are present in medium of cells primed in 0.2% / 21% oxygen 5.1.3 To determine the effect of protein synthesis inhibition during TNF treatment on SaF response to TNF 5.1.4 To determine the effect of pre-treatment with protein synthesis inhibitors or TNF during priming on subsequent response to TNF/AMD treatment 5.1.5 To determine the effect of rmuTNF on SaF cells transfected with a murine TNFa gene 5.1.6 Western blotting protocol 5.1.6.1 SDS-polyacrylamide gel electrophoresis 5.1.6.2 Gel transfer to membrane 5.1.6.3 Antibodies and standards 5.1.6.4 Antibody staining of membrane 5.1.6.5 Autoradiography and image analysis 5.1.7 TNF ELISA protocol 5.1.8 MnSOD ELISA protocol 5.2 Results 5.3 Discussion 5.4 Summary

111 5.0 Introduction

Tumour cells vary in their sensitivity to TNF but the majority of tumour cell lines are resistant to TNF cytotoxicity (TNF-resistant) (Sugarman et al. 1985). Reactive oxygen species (ROS) are involved in the cytotoxic action of TNF which can manifest as apoptosis or necrosis. The molecular mechanisms by which cells become refractory to TNF cytotoxicity are unclear and remain a major field of investigation. It was suggested that the ability to resist TNF cytotoxicity is either through expression of a product that inhibits cytotoxicity or through a mechanism that repairs TNF-mediated damage, or both (Nophar et al. 1988; Niitsu et al. 1988).

Since protein synthesis is generally required for the maintenance of TNF- resistance, which takes several hours to develop, this suggests the involvement of some rescue protein(s) in the mechanism (Collins et al. 1981; Patek et al. 1989). Indeed, early studies on protein synthesis inhibition and cell fusion indicated the presence of putative intracellular counter-acting or “protective proteins” (Ruff and Gifford, 1981; Watanabe et al. 1987; Nophar, 1988).

Inhibiting the synthesis of putative protective proteins by the addition of a protein synthesis inhibitor, such as cycloheximide (CXM) or actinomycin D (AMD), causes most TNF-resistant cells to become sensitive to TNF cytotoxicity. Thus, the ability of a cell to resist TNF mediated lysis usually requires either new transcription (inhibited by AMD) or translation (inhibited by CXM), since inhibitors of either process generally result in sensitization of resistant cells. A combination of TNF with AMD or CXM enhances cytotoxicity of TNF on some human tumour cells (Kull and Cuatrecasas, 1981; Zuber et al. 1988; Reid et al. 1989; Nio et al. 1990).

Various diverse factors, including proteins, have been identified which cause TNF-resistance (c/Table 3.1(b)) suggesting the involvement of several mechanisms in this process. The proteins identified at present seem to act at different steps, inhibiting ROS, or counter-acting their deleterious effects (Mehlen et al. 1995) thus inhibiting TNF cytotoxicity (Jaattela 1993; Neuner et al. 1994).

112 Damage does not occur to cells which produce and secrete anti-tumour cytokines, such as TNF and IFNy, (Niitsu et al 1988) which suggests that some endogenous cytokines, such as TNF itself, play a role in the acquisition of TNF- resistance (Spriggs et al 1987). In fact, endogenous TNF (enTNF) is constitutively expressed in TNF-resistant cells (Vanhaesebroeck et al 1992; Himeno et al 1990; Okamoto et al 1992) whereas, TNF-sensitive cells do not express enTNF (Niitsu et al 1988). In addition, pre-treatment with TNF induces enTNF expression and protects cells from TNF lysis in the presence of inhibitors of transcription or translation (Hahn et al 1985; Kirstein et al 1986; Kirstein and Baglioni, 1986; Rubin et al 1986; Spriggs et al 1987). TNF autoinduction by exposure to exTNF has been established for monocytes, myeloid leukaemia cells and epithelial tumour cells in vitro (Philip and Epstein, 1986; Spriggs et al 1987; Hensel et al 1987) and in a breast cancer xenograft model in vivo (de Kossodo et al 1995). Gene transfer experiments determined that enTNF confers resistance to TNF cytotoxicity (Himeno et al 1990; Vanhaesebroeck et al 1992; Okamoto era/. 1992).

The mechanism of resistance induced by enTNF expression is unknown but was suggested to occur intracellularly (Okamoto et al 1992; Himeno et al 1990). TNF can induce many different genes (Rubin et al 1988), some of which have been associated with TNF-resistance, such as manganese superoxide dismutase (MnSOD) (Wong and Goeddel, 1988; Kawaguchi et al 1990; Del Maestro et al 1992). MnSOD specifically catalyses the dismutation of superoxide to the less reactive hydrogen peroxide and oxygen and thus serves to protect cells against ROS (Hassan, 1988).

TNF mediates its cytotoxic action via superoxide and possibly other ROS (Yamauchi et al 1989; Schulze Osthoff et al 1992; Chang et al 1992; Hennet et al 1993b; Goossens et al 1995) thus the detoxification of superoxide should reduce TNF cytotoxicity or induce TNF-resistance. Although there are different types of superoxide dismutase (SOD), copper/zinc SOD, which is located in the cytosol, and MnSOD in the mitochondrial matrix which is expressed by virtually all cell types (Marklund, 1990), only MnSOD has been associated with TNF-resistance. MnSOD detoxifies intracellular ROS generated by TNF (Wong et al 1989; Asoh et al 1989). Indeed, high intrinsic levels of MnSOD correlate with resistance to TNF cytotoxicity (Wong and Goeddel,

113 1988; Kawaguchi et al. 1990; Del Maestro et al. 1992) and overexpression of MnSOD promotes survival of tumour cells exposed to TNF (Hirose et al. 1993).

MnSOD is induced by a variety of oxidative insults even when other anti­ oxidant enzymes remain constant (Stevens and Autor, 1977; Oberley et al. 1987) suggesting that this induction may have evolved to prevent an increased mitochondrial production of ROS. The mitochondrial production of superoxide is a non-enzymatic process, therefore, it increases linearly with oxygen tension (Turrens et al. 1982) and at reduced oxygen tensions there are reduced levels of ROS (Schulze-Osthoff et al. 1992). In addition, hypoxic culture conditions significantly decrease the release of superoxide by macrophages (West et al. 1994). Warner (1991) suggested that in eukaryotic cells oxygen may regulate MnSOD expression as is the case in prokaryotic cells (Hassan, 1988). In fact, hypoxia reduces MnSOD levels in rat cardiac myocytes whereas re­ oxygenation of hypoxic cells induces MnSOD (Yamashita et al. 1994).

TNF induces MnSOD gene expression in tissues and cells by an unknown mechanism (Wong and Goeddel, 1988; Asoh et al. 1989; Kawaguchi et al. 1990; Shaffer et al. 1990; Warner et al. 1991; Tsan et al. 1992; Tsan, 1993). Interestingly, exposure to hypoxia increases TNF production and secretion by peripheral blood mononuclear cells (Ghezzi et al. 1991). However, there is no evidence for the induction of TNF production in tumour cells by hypoxia. If hypoxia does induce TNF production this alone may confer TNF-resistance, or alternatively, increased TNF production may induce MnSOD expression which has been reported to cause TNF-resistance.

The aim of this section of the project was to test the hypothesis that the hypoxia- induced resistance to the cytotoxic action of TNF is due to protein induction. It was decided to determine whether TNF or MnSOD were putative proteins involved in the mechanism of hypoxia-induced TNF-resistance since both of these proteins have been reported to be induced by hypoxia and are associated with resistance to the cytotoxic action of TNF.

114 5.1 Materials and methods

5.1.1 Cell preparation In the TNF cytotoxicity assays SaF cells were seeded at 1 x 10^ /ml, in EMEM/ 10% ECS, 100 |xl per well, in 96-well plates unless otherwise stated. All treatments were performed in triplicates. Cytotoxicity was assessed by the colourimetric assay outlined in 3.1.5.1. For ELIS As and Western blots SaF cells were harvested from exponential cultures and seeded at 5 x 10^ cells in T150 flasks with EMEM + 10% ECS (35 ml).

Cells were then primed in 0.2% or 21% oxygen with 5% CO 2, and samples were taken at times 0, 2, 6, 12, 18 and 24 hours. After 24 h priming in 0.2% or 21% oxygen (5%

CO2), flasks were incubated in 21% oxygen for a further 24 h and samples were taken during this period. Samples were taken by discarding growth medium, washing flask with 10 ml of ice-cold PBS (twice) and gently scraping cells into 10 ml of ice-cold PBS. Cells were then centrifuged at 4000 rpm for 10 minutes at 4°C. The supernatant was discarded and the cell pellet was resuspended by vortexing and the gradual addition of 100 pi of lysis buffer (c/Appendix). Samples were stored at -20°C until required.

5.1.2 To determine whether freely diffusible protective/sensitising factors are present in medium of cells primed in 0.2% or 21% oxygen SaF cells were preincubated in 0.2% and 21% oxygen for 24 hours. The supernatant from the 21% oxygen primed cells was removed and replaced by the supernatant (100 pi) of the 0.2 % oxygen primed cells. Conversely, 0.2% oxygen primed cells were treated with supernatant of 21% oxygen primed cells. Each plate was treated with rhuTNF and AMD in the usual manner, with or without the supernatant from the primed cells. All plates were incubated in 21% oxygen for 24 h before cytotoxicity was assessed.

115 5.1.3 To determine the effect of protein synthesis inhibition during TNF treatment on SaF response to TNF SaF cells, in 21% oxygen, were treated with CXM (0.5 |ig/ml) at the same time as rhuTNF ± AMD (1 pg/ml). Controls received rhuTNF ± AMD, or medium ± AMD. All plates were incubated in 21% oxygen for 24 hours after treatment at which time TNF cytotoxicity was assessed.

5.1.4 To determine the effect of pre-treatment with protein synthesis inhibitors or TNF during priming on subsequent response to TNF/AMD treatment CXM or AMD (0.5 pg/ml) or rhuTNFa (5 x lO'^^U (12.5 fg)) was added to each plate of SaF cells; 100 pi of medium was added to control wells. Plates were preincubated in 21%, 2% or 0.2% oxygen for 24 hours. Medium was carefully discarded and the plates washed twice with warm medium (200 pl/well). rhuTNF standards (5 x 10^-50 U/ml) and AMD (1 pg/ml) were added to all wells except controls which received medium ± AMD. All plates were incubated in 21% oxygen for 24 h after treatment at which time TNF cytotoxicity was assessed.

5.1.5 To determine the effect of rmuTNF on SaF cells transfected with a murine TNFa gene SaF cells transfected with either a mouse TNFa gene or a neomycin resistance gene only (c/Appendix), and non-transfected SaF control cells, were incubated in 21% oxygen for 24 h before treatment with rmuTNFa (3.6 x 10'^- 0.6 ng/ml)±AMD (1 pg/ml) or medium ± AMD. Cells were incubated in 21% oxygen for a further 24 h before cytotoxicity was assessed.

116 5.1.6 Western blotting protocol 5.1.6.1 SDS-polyacrylamide gel electrophoresis Glass plates were assembled and a 12% resolving gel for tris-glycine SDS- polyacrylamide gel electrophoresis was prepared {cf Appendix). The gel mix was carefully poured between vertical glass plates and a comb was inserted. It was left to set, for at least 1 h, at room temperature or stored overnight at 4°C covered with damp tissues. The upper chamber of the gel tank was filled with running buffer (c/Appendix). The comb was removed from the gel and the individual wells rinsed with running buffer. Samples were prepared for loading: Samples, standards and molecular weight markers were thawed on ice and then micro-centrifuged (13000 rpm) for 5 minutes. Supernatant 10-50 pi was placed in an eppendorf tube containing loading buffer 10-20 pi, (c/Appendix) and all samples were boiled for 5 min (100°C) then micro-centrifuged (13000 rpm for 5 min). The samples were loaded as quickly as possible to wells (20-50 pi). The gel was placed in the gel tank and the lower chamber was filled with running buffer. Any air bubbles were removed from the bottom of the gel plates. The power supply was connected to the gel tank and the gel was run at room temperature at a constant 30 mA for at least 8 hours.

5.1.6.2 Gel transfer to nitrocellulose membrane Transfer buffer was prepared in advance as described in Appendix. The gel was removed from the gel tank and rinsed with distilled water briefly. Sponge pads, 3 mm blotting paper and Hybond nitrocellulose membrane were soaked in transfer buffer. A gel sandwich was prepared in a cassette: the gel was placed on top of blotting paper (which was on top of sponge pad) and Hybond membrane was placed over the gel (air bubbles rolled out with a pipette) which was followed by blotting paper and another sponge pad. The cassette was closed and placed in the transfer tank. The tank was filled with transfer buffer, a small magnetic flea placed inside to stir the buffer and the lid placed on the tank. Transfer was performed either at 4°C overnight at 80 volts (250 mA) constant, or 3 hours at 800 mA using a cooling element (4°C). After transfer of proteins to membrane the gel was carefully disassembled and the nitrocellulose membrane removed with forceps with the protein side upwards. The size markers were visible at this stage. The membrane was orientated and marked with pencil accordingly. It was placed in

117 distilled water and washed briefly. The gel itself was stained with Coomasie blue to check that protein transfer was efficient. The membrane was stained with Ponceau's stain for 10 min and then washed with 5% acetic acid for 5, 5, and 10 minutes. It was left to air dry and then wells and size markers were marked with pencil and a photocopy made.

5.1.6.3 Antibodies and standards Recombinant murine TNFa, purchased from Genzyme (TNF-M), was used as standard at 10 jig/ml (in 0.1% BSA/PBS), 10 |il. The primary antibody, rabbit anti­ mouse TNFa polyclonal antibody, purchased from Genzyme (B4735), was used at 1:1000 dilution. The secondary antibody, goat anti-rabbit IgG peroxidase, purchased from Sigma (A9169) was used at 1:10,000 dilution.

MnSOD standard and primary antibody was a gift from Professor N. Taniguchi (Osaka University, Japan). MnSOD standard, was purified rat MnSOD (1000 ng/ml, containing BSA and 50% glycerol), of which 10 ng was loaded. The primary antibody, rabbit anti-rat MnSOD IgG polyclonal antibody, was used at 1:1000 dilution. The secondary antibody, goat anti-rabbit IgG peroxidase, purchased from Sigma (A9169) was used at 1:10,000 dilution.

5.1.6.4 Antibody staining of membrane The membrane was blocked with 10% non-fat milk for 2 h at room temperature (RT). Primary antibody was prepared in blocking solution at 1:1000 dilution and incubated with membrane for 1-1.5 h at RT. The membrane was washed with TBS/Tween-20 (0.1%) four times (5, 10, 10 and 10 min). The secondary antibody, goat anti-rabbit IgG peroxidase, was prepared in TBS/Tween-20 (0.1%) at 1:10,000 dilution and incubated for 45 min at RT. The membrane was washed with TBS/Tween-20 four times (5, 10, 10 and 10 min) and then air dried briefly on blotting paper.

5.1.6.5 Autoradiography and image analysis The reaction was detected by Amershams ECL kit and autoradiography. Autoradiographs were analysed using the in-house General Purpose (GEN 1 ) application which is an extension of Visilog, an image processing package. This system converts the autoradiogaph image to optical density (CD) before performing measurements. It

118 determines background levels, calculates a baseline and finally the OD x 100 of samples. The OD x 100 values are expressed as a ratio of the control sample.

5.1.7 TNF ELISA protocol An EIA/RIA flat bottomed / high binding 96-well plate was coated with 50 pi of 10 pg/ml Genzyme 1221 hamster anti-mouse TNF antibody in 0.1 M sodium bicarbonate (pH 8.2) and left overnight at 4°C. The plate was washed twice with PBS/Tween-20 (500 pl/1 PBS) 200 pi per well and then blocked with PBS/10% PCS, 200 pi per well for 30 min at 37°C. Blocking buffer was discarded and the plate washed twice with washing buffer, PBS/Tween-20 (200 pl/well). Samples to be tested were thawed and diluted in PBS/10% PCS where necessary. Murine TNP standards were prepared by serial dilution of 10,000 pg/ml stock solution. The range in pg/ml was: 10,000, 5000, 2000, 1000, 500, 125, 31.25, 7.8 and 1.9, made up in medium. Samples and standards were added 50 pl/well. The plate was covered and left overnight at 4°C and then washed four times. 50 pi of 1:500 dilution of Genzyme IP400 polyclonal rabbit anti-mouse TNP antibody (made up in PBS 4- 10% PCS) was added and incubated for 2 h at RT and then washed four times with PBS/Tween-20. The reaction was detected using O- paraphenyldiamine hydrochloride (OPD) dissolved in phosphate citrate buffer (pH 5.0) at a concentration of 0.4 mg/ml (20 mg / 50 ml). 20 pi of hydrogen peroxide (30%) was added to 50 ml of the OPD solution immediately before use and 100 pi of this solution was added to every well. The colour was allowed to develop and then the reaction stopped by the addition of 25 pi of 3 M sulphuric acid. The plate was scanned at 492 nm on a multiskan microtitre plate reader.

5.1.8 MnSOD ELISA protocol Rabbit anti-rat MnSOD IgG was diluted 1/4000 by the addition of 50 mM sodium bicarbonate (pH 9.6) and 50 pi was added to wells of EIA/RIA high-binding, flat bottomed polystyrene 96-well plates. The plate was covered and left at 4°C overnight. The primary antibody was discarded and the plate was washed three times with 200 pi of washing buffer, 20 mM PBS (pH 7.4) containing 0.05% Tween-20. A plate washer was used. Blocking buffer, 0.1% BSA in 20 mM PBS, was added (50 pi) and the plate kept at 37°C for 2 hours.

119 Standards were prepared from a purified rat MnSOD stock of 1000 ng/ml (containing BSA and 50% glycerol), a serial dilution of 50, 20, 10, 5, 2.5, 1.25 and 0.625 ng/ml was made using 0.1% BS A/PBS. Samples to be tested were thawed and diluted in 0.1% BSA/PBS where necessary. Blocking buffer was discarded and the plate was washed three times. Samples and standards, 50 pi, were added and incubated at RT for 2 h and the plate was washed three times. Secondary antibody, biotinylated rabbit anti-rat MnSOD IgG, 50 pi of 1/5000 dilution in 0.1% BSA/PBS was incubated at RT for 2 hours. Peroxidase labelled avidin, was diluted 1/10,000 with 0.1% BSA/PBS and 50 pi added and incubated at RT for 10- 15 minutes. The plate was washed 4-5 times and the substrate for peroxidase, OPD, 0.4 mg/ml, in phosphate citrate buffer was prepared and immediately before use, 20 pi of 30% hydrogen peroxide was added and 100 pi added per well. The colour change was fully developed after 10-15 min and the reaction was stopped by the addition of 3 M sulphuric acid (100 pi per well). The absorbance was read at 492 nm using a Multiskan microtitre plate reader.

5.2 Results

5.2.1 Hypoxia-induced resistance to TNF cytotoxicity is not mediated by a freely diffusible factor It is conceivable that SaF cells primed in hypoxia secrete a factor that induces resistance to TNF cytotoxicity. Such a factor might be, a protease that degrades TNF, or soluble TNF receptors that bind TNF extracellularly, or a factor that reduces the sensitivity to TNF cytotoxicity. To test whether cells primed in hypoxic conditions secrete any such factor, cells were primed in 95% N%/5% CO 2 (0.2% oxygen) for 24 h after which time the growth medium from these cells was added to cells that had been primed in 21% oxygen. These 21% oxygen primed cells containing hypoxic-conditioned medium were then treated with TNF/AMD and cytotoxicity was assessed after 24 hours. However, it was determined that cells were not resistant to TNF (data not shown) but equally as sensitive as controls. The reverse experiment was also performed i.e., to determine whether cells primed in 21% oxygen secreted a sensitising factor into the medium which induced

120 susceptibility to TNF cytotoxicity. It was found that cells primed in 0.2% oxygen were not sensitised to TNF/AMD in the presence of 21% oxygen-conditioned medium (data not shown). This suggested that the hypoxia-induced resistance to TNF was not caused by a secreted factor but was an intrinsic property of the target cells, as was the sensitivity of 21% oxygen primed cells to TNF/AMD cytotoxicity. However, it was possible that cells primed in hypoxia produced some diffusible protective factor but that it was short-lived, de-activated or required contact with cells for some time prior to TNF treatment and thus was not detected in this type of experiment. Alternatively, it was possible that a putative protective factor was produced but not secreted.

5.2.2 Partial TNF-resistance induced in SaF cells by a non-specific protein synthesis inhibitor during TNF treatment SaF cells are innately resistant to TNF cytotoxicity but are sensitised by AMD which is a DNA-dependent RNA synthesis inhibitor which also inhibits protein synthesis at the concentration used in this study (data not shown). Therefore, it was investigated whether protein synthesis inhibition would also sensitise SaF cells to the cytotoxic effects of TNF. SaF cells were treated with TNF in the presence of a non­ specific protein synthesis inhibitor, CXM, with and without AMD. Figure 5.1 shows that CXM alone failed to sensitise cells to TNF cytotoxicity (IC 50 > 100 U/ml). However, when CXM was added to TNF in the presence of AMD cells were sensitised to TNF cytotoxicity (IC 50 =5.6 U/ml) but were still significantly (p <0.01) more resistant than TNF/AMD alone (IC 50 = 0.3 U/ml). These results suggest that the inhibition of protein synthesis is not sufficient to fully sensitise SaF cells to TNF cytotoxicity and that the inhibition of transcription (DNA-dependent RNA synthesis) is required.

121 100 -

90-

80-

ê

30- #— TNF + AMD

2 0 - TOF + CXM A— TNF + AMD + CXM 10 -

■3 ■2 1 ,0 1 2

rhu-TNFa (U/ml)

Figure 5.1 Effect of CXM(0.5 jig/ml), in the presence or absence of AMD, on TNF cytotoxicity of SaF cells. Symbols show the mean ± s.d. where n = 3; *p <0.05; **p<0.01.

122 5.2.3 Hypoxia-induced resistance to TNF cytotoxicity appears to be independent of protein synthesis To determine whether a protective protein was involved in blocking or repairing TNF-induced damage, CXM, a non-specific protein synthesis inhibitor, was added to cells during priming in 2% and 21% oxygen. It was hypothesised that the addition of a protein synthesis inhibitor would prevent the synthesis of a protective or repair protein and therefore sensitise the post-hypoxic TNF-resistant cells. However, as shown in Figure 5.2, it was found that when cells were pre-treated with CXM during the priming period they became significantly more resistant (p <0.01) to TNF/AMD challenge irrespective of the oxygen tension in which they had been primed. Cells pre-treated with

CXM during priming in 21% oxygen (IC 50 = 3.4 U/ml) showed a 16-fold increase in

TNF-resistance compared to untreated controls (IC 50 = 0.2 U/ml). Cells pre-treated with

CXM during priming in 2% oxygen (IC 50 =18.1 U/ml) were 15-fold more resistant than controls. Thus, not only did CXM fail to sensitise the 2% oxygen primed resistant cells to TNF cytotoxicity but it further enhanced the hypoxia-induced resistance. Furthermore, CXM induced partial resistance in cells which are normally sensitive to TNF cytotoxicity after priming in 21% oxygen.

5.2.4 Hypoxia-induced resistance to TNF cytotoxicity requires DNA- dependent RNA synthesis The involvement of a putative protective factor induced at the RNA level was investigated. As shown in Figure 5.3, SaF cells when pre-treated with AMD during priming in 21% and 0.2% oxygen are sensitised to the cytotoxic action of TNF irrespective of the priming oxygen tension. The IC 50 (0.09 U/ml) of SaF cells pre-treated with AMD during priming in 21% oxygen is decreased 13-fold compared to controls (p

<0.01). The TNF-resistance of SaF cells primed in 0.2% oxygen (IC 50 > 10 U/ml) is halved when cells are pre-treated with AMD during priming (p <0.01). Thus the sensitivity of SaF cells is enhanced by pre-treatment with AMD, independent of the priming oxygen tension. These results suggest that DNA-dependent RNA synthesis may be required for hypoxia induced resistance to TNF cytotoxicity.

123 100 -

90-

80 -

^ 70- g o 60-

^ 50-

40-

30-

2 0 -

10 -

4 3 2 1 ,0 1 rhu-TNFa (U/ml)

Figure 5.2 SaF cells were pre-treated with CXM, 0.5 pg/ml, ( solid symbols) or medium, (open symbols) during priming in 21% oxygen (-0-) or 2% oxygen (-□-) for 24 h. Subsequently, cells were treated in 21% oxygen with TNF/AMD for 24 h before cytotoxicity was assessed. Symbols show mean ± s.d. where n=15; *p <0.05; **p <0.01.

124 100 -

90-

80-

70-

60-

50-

40-

30-

2 0 -

10 -

■2 1 ,0 rhu-TNFa (U/ml)

Figure 5.3 SaF cells were pre-treated with AMD, 0.5 jUg/ml, (solid symbols) or medium (open symbols) during priming in 0.2% oxygen (-□-) and 21% oxygen (-0-) for 24 h. Subsequently, cells were treated in 21% oxygen with TNF/AMD for 24 h before cytotoxicity was assessed. Symbols show the mean ± s.e.m where n=3. *p <0.05; **p <0.01 .

125 5.2.5 Pre-treatment with TNF during priming induces resistance to TNF cytotoxicity It was decided to determine if pre-treatment with exogenous TNF (exTNF) during the priming period induced resistance to TNF cytotoxicity which would implicate TNF itself in the mechanism of resistance. SaF cells were pre-treated with a very low dose (5 X 10'^ U/ml ~ 12.5 fg) of rhuTNF during priming. After priming cells were washed and then tested for sensitivity to TNF/AMD. Figure 5.4 shows that pre­ treatment with TNF during priming in 21% oxygen (IC 50 = 2.2 U/ml) induced resistance significantly (p <0.01) in the normally sensitive (IC 50 =0.1 U/ml) 21% oxygen primed cells. Indeed, cells pre-treated with TNF in 21% oxygen were 19-fold more resistant to TNF cytotoxicity compared to untreated controls. By contrast, TNF pre-treatment during priming in 2% oxygen enhanced the hypoxia-induced resistance normally seen in these cells (IC50 =1.0 U/ml), by a factor of two only, principally at the lower TNF doses tested (p <0.01). These results suggest that endogenous TNF (enTNF) may be responsible for the hypoxia-induced resistance to TNF cytotoxicity.

5.2.6 Murine TNFa gene transfection induces resistance to TNF cytotoxicity It is well established that endogenous TNF (enTNF) induces resistance to TNF cytotoxicity and that TNF-resistant cells express high levels of enTNF (Okamoto et al. 1992; Vanhaesebroeck et al 1992). Therefore, if enTNF was responsible for the hypoxia-induced resistance to TNF cytotoxicity, as suggested by the TNF pre-treatment experiment above, then SaF cells which were genetically engineered to produce enTNF should also be resistant to TNF challenge, irrespective of the oxygen tension. This was tested by transfecting SaF cells with a muTNFa gene and then testing these TNF producing cells for sensitivity to TNF cytotoxicity. Figure 5.5 shows that SaF transfected cells, which produce and secrete enTNF (7.7-19.2 ng/ml produced by 10^ cells after 24 h; J.A.Chadwick, personal communication) are completely resistant to

TNF cytotoxicity (IC 50 >0.1 U/ml). The transfection process is not responsible for the resistance seen in TNF transfected SaF cells as cells transfected with a neomycin resistance gene (IC 50 = 0.0036 U/ml) by the same procedure are not resistant to TNF cytotoxicity but are as sensitive as non-transfected controls (IC 50 = 0.0126 U/ml). These results clarify that enTNF does in fact induce resistance to exTNF in SaF cells and that the production of enTNF is a likely candidate involved in the hypoxia-induced TNF- resistance mechanism.

126 5.2.7 Hypoxia induces endogenous TNF production in SaF cells Western blotting experiments were used to determine whether hypoxia induced the endogenous production of TNF in SaF cells, according to 5.1.6. During priming cells were untreated or treated with either AMD, in an attempt to inhibit TNF production, or LPS, to stimulate TNF production. However, it should be noted that the concentrations of AMD and LPS required to inhibit and stimulate TNF production, respectively, were not previously determined for the concentration of SaF cells used in Western blotting experiments where a 50-fold higher cell number was used compared to cytotoxicity assays. The TNF standard used was rmuTNFa (200 pg) which appears as a 17 kDa monomer and a 52 kDa TNF trimer. A photograph of a representative Western blot is shown in Figure 5.6(a) which shows that protein reactive to anti-rhuTNF was clearly identified in nearly all samples tested. Initially, cells primed in 21% oxygen for 2 h appear to have a band at least twice the size of that for 0.2% oxygen primed cells. By 6 h TNF is induced by hypoxia above control levels and this induction appears to be significant by 12 h at which time suppression of TNF production appears to have occurred in 21% oxygen samples. This result supports that of the time-course for hypoxia-induced TNF-resistance which shows that by 14 h in hypoxia SaF cells show significant resistance to TNF cytotoxicity. Clearly, by 24 h there appears to be a signifieantly greater production of TNF in 0.2% oxygen primed cells compared to 21% oxygen primed cells which show reduced TNF levels.

127 100 -

90-

70-

p 60-

50-

40-

30-

2 0 -

10 -

4 3 2 1 ,0 1

rhu-TNFa (U/ml)

Figure 5.4 SaF cells were pre-treated with rhu-TNFa (5 x 10'^ U/ml) (solid symbols) or medium (open symbols) during priming in 0.2% oxygen (-□-) and 21% oxygen (-0-) for 24 h. After priming, cells were treated in 21% oxygen with TNF/AMD for 24 h. Symbols show the mean ± s.d. where n=9; ***p <0.01 .

128 110 -

100 -

90-

80-

, 70- § ^ 60- 'o & 50- Q o 40-

30-

2 0 -

10 -

5 4 3 •2 rmu-TNFa (U/ml)

Figure 5.5 Effect of murine TNFa gene transfeetion on SaF cell response (in 21% oxygen) to rmuTNFa in the presence of AMD (1 pg/ml): non-transfeeted, control SaF cells (-0-); SaF cells transfected with neomyein resistance gene (-A-); SaF cells transfected with murine TNFa gene (-□-). Symbols show the mean ± s.d where n=3.

129 2 Q 2 S Q ë Û c crt 1 C/0 c C/] oo 1 Î o i O cu o § a . O Cc § ÛH o U U < U U < U u < U < U U f r-; X X X x: X X X X CN CNCN tu x: -C CN CN CN X (N CNCN CN

« «

2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18

Figure 5.6(a) Autoradiograph of Western blot showing TNF expression in SaF cells during priming in 21% and 0.2% oxygen for 24 h.

The autoradiograph of this Western blot was analysed using Visilog (c /5.1.6.5) and the results are shown in Figure 5.6 (b). Priming cells in 0.2% oxygen for 24 h causes a 1.7 fold increase in TNF levels. By contrast, priming cells in 21% oxygen causes suppression of TNF production to approximately one-third of control levels by 24 h. Thus, after 24 h priming, cells in 0.2% oxygen produce 5.5 times more TNF than cells primed in 21% oxygen.

LL z I- O c o o 3

C

CD >

CO CD OC

0.0 0 6 1 2 1 8 24 Tim e (h) Figure 5.6(b) Autoradiograph of Figure 5.6(a) was analysed using Visilog to determine the effect of priming SaF cells in 0.2% oxygen (-■-) or 21% oxygen (-0-) for 24 h on the relative induction of TNF.

1 3 0 AMD failed to function as a negative control for TNF production. This is probably because the AMD concentration used was too low, as 25-fold more AMD was used in experiment 5.1.4 to inhibit TNF-resistance (c/Figure 5.3). Similarly, LPS does not stimulate significant TNF production, either because it is already maximally stimulated or because the concentration of LPS was too low. In the Western blot illustrated in Figure 5.7(a) positive and negative controls consisted of HT-29 cells and mouse brain homogenate, respectively, provided by Dr N.Hassan. HT-29 cells show mono-, di- and tri-meric forms of TNF at 17, 34 and 52 kDa respectively, whereas TNF is not found in brain homogenate. The rmu-TNF standard consists mainly of monomeric TNF but also the trimeric form which is found in nearly all samples tested. TNF is active in its trimeric form but can be broken down into its basic subunits (17 kDa) by changes in pH or chaotropic agents. Comparing 0.2% and 21% oxygen primed samples, it can be seen that initially (0 h) cells primed in 21% oxygen produce more TNF (lane 18) than cells primed in 0.2% oxygen (lane 10). However, it is possible that some of the positive control, HT-29 cells, which was overloaded (lane 19), contaminated lane 18. Furthermore, by 2 h, TNF levels are reduced from 0 h levels in 21% oxygen samples and remain so up to 12 h. By contrast, TNF appears to be induced in 0.2% oxygen cells by 2 h (lane 9) and by 18 h (lane 6) this induction appears to be significant. However, samples taken at 24 h show reduced levels of TNF in 0.2% oxygen cells, but on re-oxygenation for 12 h TNF is induced above 21% oxygen control levels again. It appears that TNF is induced in 21% oxygen samples taken at 36 h and 48 h, however, it was determined by Ponceau staining of total protein that these lanes (11 and 12) were overloaded.

131 OO

► • -

1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20

Figure 5.7(a) Autoradiograph of Western blot showing TNF expression (52 kDa trimer) in SaF cells (lanes 3-18), during priming in 21% and 0.2% oxygen for 24 h and after priming, in 21% oxygen, for 24 h. Lane 19 shows TNF (52 kDa, 34 kDa and 17 kDa) produced by HT-29 cells (positive control). TNF was not detected in mouse brain homogenate (lane

20).

132 2 0 -

LL ^ 15-

■f= 1 0 - ■D

0 6 1218 24 30 36 42 48 Time (h)

Figure 5.7(b) Autoradiograph of Figure 5.7(a) was analysed using Visilog to determine the effect of priming SaF cells in 21% oxygen (-0-) or 0.2% oxygen (-■-) for 24 h on the relative induction of TNF.

Figure 5.7(b) shows the analysis of autoradiograph. Figure 5.7(a), using Visilog. TNF is not induced by priming cells in 21% oxygen over the 24 h priming period. However, when cells are primed in 0.2% oxygen TNF is induced 17-fold after 2 h, approaches control levels by 6 h but by 12 h there is a 10-fold induction of TNF. Although TNF is reduced to control levels by 24 h priming in 0.2% oxygen, based on the previous data it is likely that this result is spurious. However, as there were discrepancies between repeat blots, it was necessary to quantitate these observations to confirm results, and this was achieved by ELISA.

Cell lysates and supernatants collected from cultures primed in 0.2% or 21% oxygen for the various time periods were tested for TNF production by ELISA as

133 described in 5.1.7. TNF was not found in the supernatant of primed cells, it was found in cell lysates and was thus produced intracellularly but not secreted. As shown in Figure 5.8, TNF is produced intracellularly in SaF cells irrespective of the oxygen tension in which they are primed. This may explain why SaF cells are innately resistant to TNF cytotoxicity. Constant levels of TNF (1-2 ng/2.5 x 10^ cells) were found at all time points in lysates of cells primed in 21% oxygen. TNF was not detected in the supernatant of cells irrespective of the priming oxygen tension which suggests that either TNF is not secreted from the cells or if secreted it is quickly mopped up by TNF- Rs. There was a transient, but statistically significant increase in TNF production by hypoxic cells at 2 h (p <0.001) and at 12 h (p <0.01). This result supports that shown by Western blotting experiments and closely follows the time of onset of resistance to TNF cytotoxicity. Clearly, by 24 h in hypoxia TNF production is significantly induced (p <0.01) above 21% oxygen controls and it is at this time that maximum TNF-resistance has been noted. Interestingly, re-aeration of hypoxic cultures reduces TNF production to 21% oxygen control levels within 2 h and it remains at this level for the following 24 hours.

134 10

c/3 8 d) o

r-H X vn (N 6

d Ph 4 CD ■g 3 AIR 2

0 0 12 18 24 30 36 42 48 Time (h)

Figure 5.8 SaF cells primed in 21% oxygen (-0-) and 0.2% oxygen (-■-) for 24 h were tested for endogenous TNF production by ELISA. After 24 h priming all samples were replaced in 21% oxygen (air) for a further 24 h to determine if re-aeration affected TNF production. Symbols show the mean ± s.e.m of 3 experiments, each sample tested in triplicate. **p <0.01; *** p <0.001.

135 5.2.8 Manganese superoxide dismutase is not induced in SaF cells by hypoxia It is well established that TNF can induce MnSOD (Visner et al. 1990; Fuji and Taniguchi, 1991). The induction of synthesis of MnSOD in cells carrying an inducible TNF gene was demonstrated by Himeno et al (1992). They suggested that enTNF exerts its protective effects against TNF cytotoxicity by inducing MnSOD production. Thus, as hypoxia induced TNF production in SaF cells this suggested that SaF cells may carry an inducible TNF gene which could possibly induce MnSOD synthesis. The intracellular levels of MnSOD were assessed in cells primed in 0.2% and 21% oxygen by Western blotting experiments. The standard was purified rat MnSOD (10 ng) which has a molecular weight of 21 kDa. It was determined by Ponceau staining of membrane that samples in lanes 4, 12, 13 and 14 were overloaded (c/"Figure 5.9(a)). As shown in Figure 5.9(a), all samples show positive bands of MnSOD protein irrespective of the priming oxygen tension. It can be seen that at the earlier time points, 0 h and 2 h, cultures primed in 21% and 0.2% oxygen have equal sized protein bands. However, it appears that by 24 h priming in 0.2% oxygen there is an induction of MnSOD above control levels. Samples were also tested post-hypoxia to determine if re­ aeration affected MnSOD levels. It appears that 12 h of re-aeration induces MnSOD in the post-hypoxic cells. These results suggest that MnSOD is intrinsically present in SaF cells and that hypoxia appears to enhance MnSOD production. The autoradiograph of Figure 5.9(a) was analysed using Visilog and the results are shown in Figure 5.9(b).

136 an an an oo fN Tf fN N? fN (N OJUi Q (N O an a= an an an an a: an an an c5 on oo

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Figure 5.9(a) Autoradiograph of Western blot showing MnSOD expression in SaF cells during priming in 21% and 0.2% oxygen for 24 h and after priming, in 21% oxygen, for 24 h.

0 . 8 -

•a 0 . 6 -

M 0.4-

0 .2 -

0.0 0 6 12 18 24 30 36 42 48 Time (h)

Figure 5.9(b) Visilog analysis of MnSOD Western blot. SaF cells were primed in 21% oxygen (-0-) and 0.2% oxygen (-■-) for 24 h and all samples were replaced in 21% oxygen for a further 24 h. 21% oxygen sample at 42 h, and 0.2% oxygen samples at 36 h, 42 h and 48 h. were overloaded.

1 3 7 It can be seen that MnSOD production is suppressed by priming SaF cells in 0.2% oxygen and by 24 h it reaches one-third of control levels. Re-aeration of 0.2% oxygen primed cells for 18 h induces an increase in MnSOD production 1.4 times control levels, however, as previously mentioned, this sample was overloaded. Initially, MnSOD is suppressed by priming cells in 21% oxygen but from 12 h there is a gradual induction of MnSOD production. Cells primed in 21% oxygen have twice as much MnSOD after 24 h priming compared to cells primed in 0.2% oxygen. However, MnSOD production by cells primed in 21% oxygen is still lower after 24 h than 0 h controls but increases slightly over the next 24 h. Although these results demonstrate the presence of MnSOD it was necessary to make quantitative measurements, this was achieved by ELISA.

Figure 5.10 shows that MnSOD is produced intracellularly by SaF cells during priming in 21% and 0.2% oxygen as quantitated by ELISA (c/5.1.8). Clearly, there are no significant differences in the levels of MnSOD produced over the 24 h priming period or for the following re-aeration period of 24 h. These results agree with those obtained by Western blotting experiments and suggest that MnSOD is not implicated in the mechanism of hypoxia-induced resistance to TNF cytotoxicity. However, as MnSOD is constitutively expressed by SaF cells it may be partly responsible for the inherent resistance to TNF cytotoxicity.

138 8

zn (D O VC 0 6 r-H tü) AIR

Q 1S4 (U •g

2

0 0 6 12 18 24 30 36 42 48 Time (h)

Figure 5.10 The levels of endogenous MnSOD in SaF cells during priming in 21% oxygen (-0-) and 0.2% oxygen (-■-) for 24 h were assessed by ELISA. After priming cells were re-aerated for 24 h to determine this effect on MnSOD levels. Three separate experiments were performed with samples tested in triplicate. Symbols show the mean ± s.e.m, where n=9.

139 5.3 Discussion

Hypoxia-induced TNF-resistance could occur as a result of blocking any one or more of the recognition or activation steps involved in the TNF cytotoxicity pathway, or by a mechanism which repairs damage mediated via TNF activation. It was previously shown that certain TNF-resistant cells express a protein synthesis-dependent TNF resistance mechanism (Patek et al. 1987; Collins et al 1981). This study was set up to determine if a protective protein was responsible for TNF-resistance of post-hypoxic SaF cells. The putative protective proteins under investigation were TNF itself and MnSOD, because, as previously mentioned, both of these proteins are associated with TNF-resistance and have been reported to be regulated by oxygen (Ketis and Jones, 1992; Pogrebniak et al 1991; Yamashita et al 1994; Lewis-Molock et a l 1994). The involvement of a putative protective protein was initially studied using non­ specific inhibitors of transcription and translation. SaF cells are intrinsically resistant to TNF cytotoxicity. However, actinomycin D (AMD), a DNA-dependent RNA synthesis inhibitor (Reich et al 1962) used at a concentration that inhibits de novo protein synthesis, sensitises SaF cells to TNF cytotoxicity, which confirms that de novo protein synthesis is not required for TNF cytotoxicity. Indeed, it has been suggested that AMD renders cells more sensitive to TNF cytotoxicity by blocking the synthesis of TNF- induced repair proteins and by arresting and synchronising cells in G2/M phase (Rosenblum and Donato, 1989; Flick and Gifford, 1984), the phase in which sensitivity to TNF has been reported (Watanabe et al 1987). It was shown that the intrinsic TNF-resistance of SaF cells is not abolished by treating cells with cycloheximide (CXM), a non-specific protein synthesis inhibitor. Initially this appears to contradict the results obtained with AMD and suggests that protein synthesis is required for the cytotoxic action of TNF. Indeed, it has been suggested that TNF acts through mobile protein mediators to kill cells (Wong et al 1992) thus, it is possible that CXM inhibits the production of a protein which induces susceptibility to TNF. In addition, it has been suggested that TNF cytotoxicity in the presence of CXM may be mediated by enzymes of the lipoxygenase pathway (Reid et al 1991) which, if true, would imply that this is not the major mechanism of TNF cytotoxicity in SaF cells. However, the role of CXM as a protein synthesis inhibitor involves blocking cells in the G1 phase of the cell cycle (Rossow et al 1979; Pardee et

140 a l 1978), the phase in which TNF-resistance has been reported to occur (Darzynkiewiczs et al 1984). Furthermore, Belizario and Dinarello (1991) have shown that pre-treatment with CXM blocks cells in G1 and S phase generating a cell population resistant to TNF cytotoxicity. This may explain why pre-treatment of hypoxic SaF cells with CXM further enhances TNF-resistance and induces partial resistance in normally sensitive oxic cells. In addition, this may explain the discrepancy between results obtained with AMD and CXM treatment. Interestingly, when AMD and CXM are used in combination with TNF they appear to compete, which suggests that opposing cell cycle blocks may occur. Overall, it was established that a putative protective protein is involved in the mechanism of hypoxia-induced TNF-resistance. However, a protective protein if produced, was either not secreted, or if secreted, was either short-lived, de-activated or required a certain period of time for cellular uptake prior to TNF treatment. Indeed, at least one protective protein was produced intracellularly but not secreted, this protein was TNF itself. It was clearly demonstrated that hypoxia-induced resistance to TNF cytotoxicity is due, at least in part, to the induction of intracellular endogenous TNF (enTNF) production during priming in 0.2% oxygen. It was determined by both Western blotting and ELISA techniques that intracellular enTNF is induced by priming SaF cells in an hypoxic environment of 0.2% oxygen for 24 hours. More importantly. Figure 5.8 shows that enTNF was significantly induced (p <0.01) by hypoxia (0.2% oxygen) close to the time of onset of TNF-resistance (12-14 h) and at the time when significant resistance to TNF cytotoxicity is seen (24 h). Exposure of SaF cells to 0.2% oxygen caused rapid production of TNF. The release of TNF was biphasic, there was an initial peak at 2 h, which was highly significant (p <0.001), and a less dramatic albeit significant (p <0.01) peak at 24 hours. Re-aeration of hypoxic cells caused a reduction in TNF to control levels, suggesting that hypoxia regulates enTNF expression directly. In addition, it was determined that enTNF is constitutively produced by SaF cells. This could partly explain their inherent resistance to TNF cytotoxicity. The present study has demonstrated that hypoxia can induce intracellular enTNF production in a murine tumour cell line which inherently produces TNF and is intrinsically resistant to TNF cytotoxicity. As shown in Figure 5.4, pre-treatment of SaF cells during priming in 21% oxygen with a very low dose (12.5 fg) of recombinant human TNF induces significant

141 resistance (p <0.01) to further TNF/AMD challenge. The level of resistance induced by TNF pre-treatment in normally sensitive 21% oxygen primed cells was equal to that seen in post-hypoxic cells. TNF-pre-treatment of hypoxic cells did not significantly enhance the level of resistance induced by 0.2% oxygen alone, suggesting that maximum resistance is already induced in these cells. These results suggest that, as TNF pre-treatment induces resistance in cells that were primed in 21% oxygen but not 0.2% oxygen, TNF itself may already be induced by hypoxia and may also be responsible for hypoxia-induced resistance to TNF cytotoxicity. The protective effect of pre-treatment with TNF has also been observed in vivo (Patton et al. 1987; Fraker et al. 1988; Fraker et al. 1990). However, recently it was reported that low dose TNF treatment induces other possible protective factors apart from TNF, such as MnSOD, IL-6, GM-CSF and IL-8 (Sakuma et al. 1993). Therefore, this experiment was substantiated by transfecting SaF cells with a murine TNF gene, which results in the endogenous production of TNF. Endogenous TNF production in SaF cells induces complete resistance to TNF cytotoxicity in the presence of AMD, as shown in Figure 5.5. Although there are many reports in the literature which have noted that enTNF induces resistance to exTNF cytotoxicity (Okamoto et al. 1992; Jia et al. 1995) the mechanism of resistance at the molecular level is as yet unknown. However, whether SaF cells are pre-treated with exTNF for 24 h or artificially engineered to produce enTNF they become resistant to TNF cytotoxicity. These results confirm that intracellular enTNF causes TNF-resistance in SaF cells in vitro and is a protective factor likely to be induced by hypoxia in vivo which could induce resistance to TNF cytotoxicity. Thus, enTNF induces resistance to TNF cytotoxicity and as hypoxia induces enTNF in vitro it is likely that enTNF is, at least partly, responsible for hypoxia-induced TNF-resistance of SaF cells. Clearly, enTNF is involved in the mechanism of hypoxia-induced TNF- resistance but the role it plays has not been elucidated by these studies. However, as previously mentioned, TNF may not be the only protective protein induced by hypoxia which is involved in inducing resistance to TNF cytotoxicity. Both hypoxia and TNF induce expression of a large number of proteins, some of which may contribute to TNF’s protective role in post-hypoxic cells. Exposure of cells to extreme hypoxia (0.01% oxygen) results in the enhanced synthesis of a number of specific proteins, the oxygen regulated proteins (ORP’s) (Wilson and Sutherland, 1989; Heacock and Sutherland, 1986; Sutherland et al. 1986).

142 However, it is unlikely that ORP’s are involved in hypoxia-induced TNF-resistance because of the moderate hypoxia (0.2% oxygen) investigated in this study. For the same reason, it is unlikely that heat shock proteins (hsp’s) are implicated as they generally are induced by exposure to re-oxygenation of extremely hypoxic cells (Sciandra et al. 1984; Heacock and Sutherland, 1986). Adenosine increases dramatically during hypoxia (Ramkumar et a l 1995) and this is believed to confer cytoprotection to ischaemic tissues (Berne and Rubio, 1974). However, although the adenosine cytoprotection mechanism has not been fully elucidated it could be due to increased SOD levels which protect cells from ROS generated during re-oxygenation (Yamashita et al 1994). However, in this study MnSOD was not induced (by hypoxia) which suggests that adenosine is not involved in the mechanism of hypoxia-induced TNF-resistance. There is now convincing evidence that TNF mediates its cytotoxic activity via ROS, particularly superoxide (Shoji et al 1995; Hennet et al 1993b; Schulze-Osthoff et al 1992). It has been shown that mitochondrial production of superoxide is a causal mechanism of TNF cytotoxicity as elevated superoxide levels, induced by TNF, lead to activation of dehydrogenase, resulting in a fatal decrease in cellular ATP levels (Goossens et al 1995). The only enzyme that is practically always induced by TNF, which is involved in ROS metabolism, is MnSOD (Wong and Goeddel, 1988; Asoh et al 1989; Kawaguchi et al 1990; Shaffer et al 1990; Warner et al 1991; Tsan et al 1992; Tsan, 1993). Other antioxidant enzymes, such as CuZnSOD, catalase, glutathione peroxidase and glutathione reductase are not induced by TNF even when MnSOD is induced (Wong and Goeddel, 1988; Kawaguchi et al 1990; Wong et al 1992; Himeno et al 1992). Although the majority of studies have shown that TNF-resistance is associated with increased levels of MnSOD (Wong et al 1989; Himeno et a l 1992; Kizaki et al 1993; Jia et al 1995) a few reports show no correlation between expression and or induction of MnSOD and susceptibility to TNF (Boss et al 1991; Park et al 1992). However, TNF-resistant cells treated with TNF markedly induce MnSOD protein, supporting a role for MnSOD as a rescue protein (Kawaguchi et a l 1990). Although MnSOD plays a major role in protecting cells from TNF cytotoxicity (Hirose et al 1993) it has also been shown that MnSOD is only partly responsible for TNF- resistance; hence, there must be other protective proteins involved (Wong et a l 1992). Many studies have shown that TNF induces MnSOD which in turn induces resistance to TNF cytotoxicity (Himeno et al 1992; Zyad et al 1994). Although

143 MnSOD was singled out in this study as a likely candidate for protecting post-hypoxic cells from TNF cytotoxicity, it has been determined by two independent techniques that in SaF cells MnSOD protein is not induced by hypoxia or even by re-aeration post­ hypoxia. Thus, MnSOD protein is not directly involved in the hypoxia-induced TNF- resistance mechanism. However, MnSOD may be involved in the intrinsic resistance of SaF cells to TNF cytotoxicity as it is constitutively expressed in this cell line. Interestingly, it has recently been demonstrated that after hypoxic pre-conditioning MnSOD induction is regulated at the transcriptional level (Yamashita et al. 1994). Induction of MnSOD mRNA is rapid, persists for up to 96 h, requires very low levels of TNF (0.1 ng/ml) and does not require new protein synthesis (Wong and Goeddel, 1988; Tsan et al 1990). It has recently been demonstrated that enTNF and MnSOD play a role in protecting cells from TNF cytotoxicity (Himeno et al. 1992; Jia et al. 1995). Ketis and Jones (1992) found an increase in SOD levels during hypoxia and suggested that ROS are generated during hypoxia, however this has not been generally observed. It has been shown that ROS are generated and MnSOD induced by re­ oxygenation of hypoxic cells and not by hypoxia per se (Yamashita et al. 1994). Recently, it was demonstrated that hypoxic culture conditions significantly decrease release of superoxide from macrophages (West et al 1994). It was also shown that TNF-resistant cells produce less superoxide in response to TNF treatment (Hennet et al. 1993b). Thus, it is possible that the balance of ROS and ROS scavengers determines cell susceptibility to TNF cytotoxicity. Although MnSOD was not induced in SaF cells by hypoxia, if the level of superoxide generated by hypoxic cells is lower than for aerated cells then this would make available more ROS scavengers (e.g., MnSOD) to counteract ROS produced by re-oxygenation and the subsequent TNF/AMD treatment of cells. TNF can induce metallothionein (MT) (De et al. 1990; Sato et al. 1992) which regulates metabolism of essential metals such as zinc and copper and has a strong ability to scavenge ROS both in vitro and in vivo (Sato and Bremuer, 1993). Data indicate that if sufficient MT is located near the site of ROS production then MT can act as a scavenger. Sato et al (1995) showed that pre-treatment with TNF, before generation of superoxide by paraquat, further enhances levels of MT. TNF induces MnSOD and MT but not other antioxidants which suggests that they may play co-operative roles in cells as ROS scavengers (Sato et al 1995). Thus although MnSOD was not induced by

144 enTNF it is possible that other protective factors such as MT are induced by TNF in hypoxic conditions. It is possible that there may be both specific and non-specific protective mechanisms operating to induce resistance to TNF cytotoxicity. It has been shown in this study that intracellular TNF (enTNF) induced by hypoxia is at least partly responsible for hypoxia-induced resistance to TNF cytotoxicity as the time-course of production corresponds to that of TNF-resistance. Endogenous TNF could potentially induce resistance directly or via other protective mediators. There are numerous possibilities whereby TNF could directly induce resistance to TNF cytotoxicity: (1) enTNF down-regulates TNF receptors (TNF-R) by binding to intracellular receptors and inhibiting their transfer to the cell surface (2) enTNF down-modulates receptors intracellularly by altering receptor function or affinity (3) enTNF is secreted extracellularly and binds to soluble TNF-Rs which has been shown to induce resistance to TNF cytotoxicity (Lesslauer et al 1991; Neuner et al 1994) (4) enTNF is secreted extracellularly becomes membrane-bound and induces resistance in an autocrine manner by binding to cell-surface receptors (5) enTNF influences cellular energy metabolism to enable cells to cope with oxidative challenges. It appears that cells respond to TNF by switching their energy metabolism away from efficient oxidative metabolism and towards relatively inefficient anaerobic metabolism, while maintaining or increasing energy output (Freeman and Crapo, 1982). These changes may have the effect of relieving the mitochondria from their normal burden of ROS thus enabling them to deal better with exogenously induced ROS, produced by exTNF/AMD treatment. (6) enTNF induces an intracellular signal other than cytotoxicity, thus regulating expression of protective factors or suppression of damaging factors. TNF inhibits c-myc (Kronke et al 1987), which induces cellular susceptibility to TNF cytotoxicity (Klefstrom et al 1994). It has been suggested that TNF may activate processes in the cell that remedy damage caused by ROS (Boveris, 1977; Freeman and Crapo, 1982). TNF induces many genes (c/Table 1.7) some of which are associated with resistance to TNF cytotoxicity (Rubin et al 1988; Beresini et al 1988; Lee et al 1990). If TNF is not directly involved in inducing resistance to TNF cytotoxicity, it may induce other protective factors which protect cells from TNF cytotoxicity

145 Although this is the first report showing that hypoxia can induce intracellular enTNF production in a murine tumour cell line a similar phenomenon has been noted in polymorphonuclear cells (PMNCs). Ghezzi et al. (1991) showed that hypoxia increases both intracellular and extracellular TNF production in PMNCs stimulated with endotoxin. Maximum TNF levels were found after 24 h of hypoxic exposure which is in agreement with results of the present study. They detected at least twice as much TNF in cell supernatants as was found intracellularly, whereas in SaF cells TNF was detected only intracellularly. This is not surprising as LPS was required to stimulate TNF production in PMNCs and a macrophage cell line would be expected to secrete TNF. In addition, they studied a range of oxygen tensions (0%, 3%, 10% and 21%) and found that TNF was maximally potentiated in 0% oxygen. This suggests that oxygen tension may determine the degree of resistance to TNF cytotoxicity in SaF cells by regulating the level of TNF induction. Interestingly, West et al (1994) showed that pre-exposure to hypoxia for 2 h did not result in detectable TNF release unless macrophages were stimulated with at least 10 ng/ml of LPS. More recently, it was shown that hypoxia induces an unstimulated human monocyte cell line to release TNF and its soluble TNF-Rs (sTNF-R) in vitro (Scannell et al 1993). In agreement with the present study they found that during hypoxia, TNF was released in a biphasic manner, with an initial peak at 1 h and a pronounced peak at 18 hours. Both 55- and 75-kDa sTNF-Rs were released progressively with maximum levels at 24 h, which corresponds to the time when maximum resistance is seen in SaF cells. In addition, Gerlach et al (1993) demonstrated that in endothelial cells hypoxia is able to induce high affinity TNF receptors in a time- and dose-dependent manner. However, hypoxia failed to induce TNF whereas re-oxygenation for 3-4 h after hypoxia significantly induced TNF which was mediated by ROS. They found that hypoxia increases both affinity and number of TNF-Rs and thus causes an increase in the binding of TNF to endothelial cells. Indeed, the affinity was maximal at 10-14 h of hypoxic exposure which corresponds to the time of onset of TNF-resistance in post-hypoxic SaF cells. Furthermore, re-oxygenation did not increase the number or affinity of TNF-Rs but induced a slow decrease of TNF binding which could explain the post-hypoxic TNF- resistance found in SaF cells. Interestingly, their data indicate that hypoxia does not modify TNF itself but changes TNF-Rs. Hypoxia is able to modulate receptor affinity, as shown for a l adrenergic receptors on cardiac myocytes (Corr et al 1981).

146 Thus, hypoxia induces TNF and its receptors in cells of the monocyte lineage and in endothelial cells. This study has shown that hypoxia induces TNF production in murine tumour cells and, although TNF-R levels were not investigated in this study, based on the literature it is likely that hypoxia also induces TNF-Rs in hypoxic SaF cells. The induction of TNF-Rs by hypoxia provides one explanation for the resistance induced by hypoxia to TNF cytotoxicity, as follows: Intracellular enTNF production is induced in SaF cells by exposure to hypoxia for 2 hours. However, by 6 h TNF production is reduced to control levels due to upregulation of intracellular TNF-Rs which are induced by hypoxia to counteract the increased TNF levels. Once the TNF-Rs bind the excess enTNF, hypoxia switches off TNF-R production which in turn causes TNF levels to increase significantly (12 h) due to persistent hypoxia and reduced receptor levels. This cycling effect of TNF and its receptors, regulated by hypoxia, continues only during exposure to hypoxia, as it can be seen from Figure 5.8 that when cells are re-aerated TNF production returns to control levels. In addition it takes at least 5 h for SaF cells to resensitise to TNF cytotoxicity post-hypoxia, by which time TNF-R levels are no longer induced as TNF production is constant. The biphasic nature of TNF production is probably due to an initial dramatic response of cells to the hypoxic insult (2 h) as cells prepare themselves for severe damage. However, as the degree of hypoxia does not change, TNF levels drop to control values but then continue to increase in a cyclical manner in preparation for some further insult. It is likely that if the duration of hypoxia was prolonged beyond 24 h the TNF levels would continue to increase, this would explain why resistance to TNF cytotoxicity increases 60-fold when cells are grown continuously in 2% oxygen for 5 passages (c/ Chapter 3). The present study supports the idea that intracellular enTNF induced by hypoxia is most likely to confer resistance to TNF cytotoxicity. Further studies are warranted to substantiate any of the afore-mentioned mechanisms of hypoxia-induced resistance to TNF cytotoxicity. The mechanisms by which cells become resistant to TNF are still surrounded by considerable controversy. Although the mechanism of hypoxia-induced resistance to TNF cytotoxicity has not been determined in the present study the finding that intracellular enTNF is involved in the mechanism is of major importance. There is considerable evidence for a role of acquisition of TNF resistance in the onset and growth of malignant tumours (Chouaib et al. 1992). The resistance of some tumour cells to the cytotoxic action of TNF remains one limitation for the clinical use of this cytokine

147 (Spriggs and Yates, 1992). In fact, in some experimental models as well as in human tumours, the development of multi-drug resistance has been reported to be accompanied by an acquisition of resistance to TNF (Fruehauf et al 1991; Hofsli and Nissen-Meyer, 1989; Dollbaum et al 1988). This study suggests that hypoxia is an important factor in tumour cell resistance to TNF cytotoxicity as it induces intracellular endogenous TNF which causes TNF-resistance. The improved understanding of the molecular basis of hypoxia-induced TNF-resistance and the elucidation of its relationship with resistance to cytotoxic drugs will play a key role in shaping future strategies for more adaptable therapeutic modalities using TNF.

5.4 Summary

It has been clearly demonstrated that intracellular enTNF is one of the protective factors induced by hypoxia that causes resistance to TNF cytotoxicity in SaF cells. It was shown that the TNF susceptibility of SaF cells can be varied by artificially inducing or inhibiting expression of endogenous TNF however, this does not rule out the involvement of other protective factors in the mechanism of hypoxia-induced TNF- resistance.

148 Chapter 6

The cytotoxic action of TNF in vivo

6.0 Introduction 6.1 Materials and methods 6.1.1 Tumours 6.1.2 TNF 6.1.3 To determine the cytotoxic activity of TNF in vivo 6.1.4 To determine the effect of TNF treatment on SaF tumour growth 6.1.5 To determine the effect of TNF treatment on SaF tumour blood flow 6.1.5.1 Laser Doppler flowmetry: The Oxford Array 6.1.5.2 Experimental procedure 6.1.5.3 Treatment groups 6.1.5.4 Data acquisition 6.1.5.5 Data analysis 6.2 Results 6.3 Discussion 6.4 Summary

149 6.0 Introduction

Tumour necrosis factor derived its name from the first function ascribed to it, which was the dramatic haemorrhagic necrosis (HN) of a panel of transplanted tumours in mice (Carswell et al. 1975). Since, it has been shown that TNF causes regression of a wide range of murine tumours, human tumour xenografts, and local tumour burdens in vivo (Carswell et al. 1975; Haranaka et al. 1984; Creasey et al. 1986; Balkwill et al. 1986; Liénard et al. 1992a,b). Although TNF is directly cytotoxic to tumour cells, can modulate activated host cells (Talmadge et al. 1987; Nakano et al. 1989; 0stensen et al. 1987; Shalaby et al. 1985; Shau, 1988), can act on tumour-specific immunity (Palladino et al. 1987b; Balkwill et al. 1989) and on energy metabolism of host and tumour cells, the anti­ tumour activity of TNF can also be mediated through changes in hemodynamic and vascular functions. Vascular effects of TNF have been noted in many tumour models and appear to be essential to the development of haemorrhagic necrosis (HN) (Havell et al. 1988; Proietti et al. 1988; Shine et al. 1989; Kallinowski et al. 1989a). TNF does not cause regression of subcutaneous tumours prior to the development of a vascular supply (Palladino et al. 1987a; Manda et al. 1987). A developed capillary network in the tumour is almost certainly required for pronounced anti-tumour activity in vivo (Shimomura et al. 1988; Mule et al. 1988). However, studies on the anti-tumour activity of TNF in human tumour xenografts have shown effects varying from growth inhibition to complete tumour regression without evidence of haemorrhagic necrosis (Balkwill et al. 1987b). The precise mechanisms of these effects are not known but are likely to involve indirect actions of TNF on the host-tumour interactions. However, the majority of studies suggest that the anti-tumour effect of TNF depends primarily on the ability of TNF to damage tumour vasculature, by direct cytotoxic effects on tumour vascular endothelium and indirectly by activation of host cells. TNF directly affects the tumour vascular endothelium, causing morphologic changes, growth inhibition of vascular endothelial cells, and a pronounced cytotoxic effect on capillary endothelial cells (Sato et al. 1986; Watanabe et al. 1988b). Fortunately, tumour endothelium is more sensitive to TNF cytotoxicity than normal endothelium, for unknown reasons. It has been suggested that tumour derived factors may sensitise tumour endothelium to TNF, and contribute to the selective damage of

150 tumour vasculature (Nawroth et al 1988; Clauss et al 1992). In addition, it has been reported that microvasculature is different in its pathophysiologic nature from that of large blood vessels (Zetter, 1981). TNF possibly suppresses neovascularization to restrict nutrient supply into tumour tissues and finally to achieve tumour regression. High TNF doses given systemically are followed by a reduction of perfusion pressure (Tracey et al 1986b) and increased vascular permeability of macromolecules 1-2 h after administration (Stolphen et al 1986; Brett et al 1989; Wheatley, 1993). Micro vascular leakage results in a reduction of tumour blood flow (TBF), tissue accumulation of neutrophils (Yi and Ulich, 1992), and an early influx of red cells into the tumour (MacPherson and North, 1986; Shimomura et al 1988; Kallinowski et al 1989a; Van de Wiel et al 1990; Aicher et a l 1990). Significant tumour necrosis is seen as early as 3 h after TNF infusion which is preceded by swelling of the tumour endothelium and over-expression of cell adhesion molecules on the tumour blood vessels. TNF modulates the haemostatic properties of endothelial cells promoting clot formation in tumour vessels (4-6 h) which are subsequently destroyed (Nawroth and Stern, 1986; Bevilacqua et al 1986; Shimomura et al 1988), causing haemorrhage, congestion, and blood circulation blockage at 24 hours (Watanabe et al 1988b). These studies suggest that direct toxicity of TNF on tumour vasculature is a factor in the overall anti-tumour mechanism of TNF. Indeed, it was reported that death of tumour cells is not observed until the vasculature is destroyed. However, reduction in TBF and resulting necrosis by TNF treatment can occur without subsequent tumour regression (Naredi a/. 1993). Several groups have shown that TNF can have anti-tumour activity in vivo against transplantable tumour cell lines that are completely resistant to TNF cytotoxic activity in vitro (Brouckaert et al 1986; Grosser et a l 1986; Palladino et al 1987 a, b; Shimomura et al 1988; Van de Wiel et al 1989). SaF cells in vitro are resistant to TNF cytotoxicity in the absence of AMD, probably due, at least in part, to the constitutive production of intracellular endogenous TNF and MnSOD. Moreover, hypoxia further induces intracellular enTNF production in vitro, which has been shown to induce resistance to exogenous TNF challenge (Chapter 5), and it is likely that this may also occur in vivo. It has been shown that pre-exposure to hypoxia induces resistance to the cytotoxic action of TNF in both human and murine cells lines in vitro (Chapter 3). It has also been determined that the oxygenation status of SaF tumours is inadequate, with most tumour cells situated in an hypoxic environment. Indeed, as previously mentioned.

151 hypoxia is a common feature of solid human tumours (Vaupel et al. 1994; Hockel et al. 1993). As pre-exposure to hypoxia induces resistance to the cytotoxic action of TNF in vitro it would be expected that hypoxia in vivo should have a similar affect. The aim of this section of the project was to determine the cytotoxic action of TNF on the SaF tumour in vivo and to assess its relative importance in the anti-tumour action of TNF. Central to this, was the need to determine the effect of TNF-induced host-mediated activity via tumour growth delay assay and the effect of TNF treatment on tumour vascular status.

6.1 Materials and methods

6.1.1 Tumours SaF subcutaneous rear dorsal and hind foot tumours were allowed to grow to 5.5-6 mm GMD before treatment was commenced.

6.1.2 TNF Human recombinant TNFa, (rhuTNF) BISS-14 was provided by Celltech Ltd as a lyophilised powder stabilised with 8 |ig mannitol/|Lig TNF, resuspended at 1 mg/ml. It had a specific activity of 3.2 x 10^ U/mg. Murine recombinant TNFa (rmuTNFa) was purchased from Genzyme (TNF-M) as a frozen liquid (in 0.1% BSA/PBS) and had a specific activity of -5x10^ U/mg.

6.1.3 To determine the cytotoxic activity of TNF in vivo An in vivo!in vitro clonogenic assay was used to assess the fraction of cells, in a suspension prepared from TNF treated dorsal tumours, which had unlimited proliferative potential i.e., were clonogenic. The mice were randomised into treatment groups, six mice were allocated per TNF dose group and four mice per control group. Human recombinant TNFa, was suspended in 5% BSA/saline solution to give doses of 50, 100, 250 and 500 pg/kg; these doses were chosen based on data from Pimm et al. (1991). TNF was administered by intravenous injection (0.1 ml). Control mice were treated with either 5% BSA/saline or saline alone. Tumours were excised 18 h after treatment and an in vitro colony forming assay was performed (c/2.1.5.3).

152 6.1.4 To determine the effect of TNF treatment on SaF tumour growth The delay in tumour growth induced by TNF treatment was used as a measure of the effectiveness of TNF on dorsal s.c. tumours in vivo. Mice were randomised into treatment groups of six mice and two control groups of three mice which served as TNF vehicle controls and non-treated controls. rhuTNFa, suspended in 5% BSA/saline solution, was administered i.v. at 50 and 100 pg/kg. Following TNF treatment, tumours were measured three times per week over three perpendicular diameters, using vernier callipers. All tumour measurements were continued until the mean diameter reached 12 mm, at which time mice were sacrificed. The time taken to reach GMD plus 3 mm (r + 3) was measured for each individual tumour in each treatment group. The growth curves were normalised by adjusting the mean diameter of each tumour at day zero to a value of 5.5 mm, and subsequent values were also adjusted by this difference in measured and calculated starting points. From the individual regrowth times of each tumour, a mean and standard error of the mean time was calculated for each dose group. Statistical significance (p <0.05) was determined between the r + 3 values of TNF treated compared to control tumours by the Students t-test.

6.1.5 To determine the effect of TNF on SaF tumour blood flow Laser Doppler flowmetry was employed to measure blood flow. The principle of laser Doppler is that the frequency of laser light, emitted by a probe into adjacent blood vessels, becomes “Doppler” shifted when photons hit red blood cells. The number of photons that are Doppler shifted is proportional to the number of moving red blood cells (or volume) and the shift in frequency is proportional to velocity. The product of volume and velocity gives a measure of blood flow.

6.1.5.1 Laser Doppler flowmetry: The Oxford Array^*^ The Oxford Array’’’'^ (Oxford Optronix) multi-channel laser Doppler system (Figure 6.1 below) was used to measure microvascular perfusion (blood flow). This allows simultaneous blood flow measurements in up to twelve sites and temperature measurements in eight sites via custom built probes. The laser probes consist of both emitting and collecting fibre optics each approximately 100 pm in diameter. The fibres have a sampling radius of 50 pm, sampling depth of 100-150 pm and a sampling volume of 5 X 10’^ mm^.

153 Figure 6.1 The Oxford Array^^ Laser Doppler

?

£ 1

_ S T U A ftt

6.1.5.2 Experimental procedure In the case of dorsal tumours, the back of the mouse was shaved prior to set-up. The non-anaesthetised mouse was restrained in a perspex jig with the tumour area exposed. In the case of foot tumours the tumour-bearing leg was delicately taped so that blood flow was not constricted. The jig was taped to a thermal blanket set at 37°C. This in turn sat on top of a stone slab and a thick piece of sponge to prevent movement from the surrounding bench. An open-ended box was placed over the jig which served to prevent external distractions to the mouse and to hold the probe cables in place via slits in the top of the box. The experimental set-up can be seen in Figure 6.2.

154 Figure 6.2 Experimental set-up for measurement of tumour blood flow

Tumour temperature was monitored by a very fine wire thermocouple which was inserted directly into the tumour and body temperature was monitored via a rectal temperature probe. The tail vein was dilated using an infra-red lamp and a catheter inserted which contained saline/heparin (1000 U/ml) (1:1). Four to five probes were inserted in the tumour after making small incisions with a 25 G needle, and one probe was placed, using tissue glue, between the toes, which served as normal tissue control.

6.1.5.3 Treatment groups SaF subcutaneous dorsal and left hind limb tumours were chosen for treatment. rhuTNF (100 |ig/kg) or rmuTNFa (10 jig/kg) was administered i.v. using 0.5% BSA/saline as diluent. Mice were randomly divided into treatment groups of untreated controls, TNF-diluent treated controls, and rhu- and rmu-TNF treated, with 6 rear dorsal or hind-foot tumour-bearing mice per group. There was no difference between blood flow of rear-dorsal or hind-foot tumour bearing untreated or TNF-diluent treated controls so this data was combined. Similarly, as there was no significant difference

155 between blood flow of rhu-TNF dorsal and foot tumours the effect of rmuTNF was assessed on dorsal tumours only.

6.1.5.4 Data acquisition Real-time, continuous visual display of blood flow was observed for 5-10 min, to ensure that all probes were stable and functioned efficiently, after which time data was recorded and stored to disk. A period of 10 min before treatment was recorded to determine normal “control” blood flow. Treatment involved switching the syringe to one which contained TNF/0.5% BSA in saline or diluent alone, injecting 0.1 ml and logging event time on the computer. Blood flow was monitored for 2 h after treatment. At this time a lethal dose of sodium pentobarbitone was injected via the tail vein catheter and data recorded for 10 min thereafter to ensure that blood flow had ceased.

6.1.5.5 Data analysis The Oxford Array^^"^ takes a sample reading every 50 milliseconds which gives 20 readings per second. The average of every 4 min was taken and a standard error of the mean calculated on this datum value. The 10 min period recorded after the death of the animal was considered to reflect the ‘background’ reading and was subtracted from all values recorded. Background readings are most likely due to Brownian motion of free red blood cells in front of the probe. The first 10 min recorded was taken as control blood flow and thereafter treatment data was expressed as a percentage relative to control. Laser Doppler signals are scaled 0-1000 Blood Perfusion Units (BPU). The maximum attainable value is therefore 1000 any value exceeding this (due to probe or mouse movement) is scaled at 1000 BPU, this can be detected as an abrupt change in the backscatter signal which is recorded simultaneously by every probe, and thus can be eliminated. Data are shown as a graph of relative blood flow (y-axis) versus time (minutes) (x-axis).

156 6.2 Results

The acute cytotoxic effect of rhuTNF on SaF rear dorsal tumours 18 h after treatment is illustrated in Figure 6.3. The data show that non-lethal i.v. doses of 50

|ig/kg TNF decreases the surviving fraction to -30% of controls and 100 |ig/kg TNF causes at least two logs of cell kill, which is not enhanced by increasing the dose to 250 |ig/kg. However, a dose of 500 M-g/kg causes 3.5 logs of cell kill but systemic toxicity was seen at this dose, four out of six mice died within the 18 h treatment period. In agreement with the present results Pimm et al (1991) reported that mice given 500

|Lig/kg rhuTNF all died within 24 h of a single i.v. injection. They found that doses of 100 p.g/kg and below were well tolerated daily for up to 3 days, with no deaths, tail damage or loss in body weight (Pimm et al 1991).

Based on the in vivo cytotoxicity data a tumour growth delay assay was performed using doses of 50 and 100 }Xg/kg rhu-TNF, the results are shown in Figure 6.4. It can be seen that 50 fig/kg rhuTNF gives a modest, but significant, (p <0.01) 3 days growth delay. Doubling the TNF dose to 100 |Ltg/kg enhances the growth delay to 4 days, (p <0.01). It was interesting to determine how much of the growth delay observed was due to direct cytotoxicity and how much was due to effects on the tumour vasculature. The effect of TNF on tumour vasculature was assessed by measuring microvascular perfusion or tumour blood flow (TBF).

157 0 100 200 300 400 500 rhu-TNF (|iig/kg)

Figure 6.3 Effect of rhu-TNF on SaF tumour cell survival 18 h after treatment. Symbols represent the mean ± s.e.m of 6 tumours.

158 12

11

10

V-H 4—»(D

8

s

'5 7 S o (Ü O 6 Control (5% BSA/saline) 5 rhuTNF (50 |ag/kg) rhuTNF( 1 OOgg/kg) 0 4 6 8 10 Days after treatment

Figure 6.4 Growth delay of SaF rear dorsal tumours after a single i.v. dose of rhu- TNF. Curves were normalised to 5.5 mm at day zero. Each symbol represents the mean ± s.e.m of 6 tumours.

159 Laser Doppler flowmetry (LDF) allows the continuous monitoring of the microcirculatory function in superficial tumours. It was initially determined that the procedure involved in measuring blood flow by laser Doppler did not adversely alter blood flow over the chosen 2 h monitoring period. Figure 6.5(a) shows that TBF of unanaesthetized, untreated, SaF tumour-bearing mice is relatively stable over a 2 h period. Treatment with 0.5% BSA in saline (TNF diluent) causes a transient, non­ significant, increase in TBF (-127% at t = 40-60 min) as illustrated in Figure 6.5 (b). Recently, it has been shown that SaF tumours exhibit transient fluctuations in blood flow (Chaplin and Hill, 1995). By contrast, skin blood flow although more erratic than TBF remains unchanged 2 h after TNF-diluent treatment. It should be noted that the errors on skin blood flow data are significantly greater than TBF, this probably reflects the reduced number of measurements made for this site. Overall, at the end of the 2 h period neither tumour or skin blood flow are significantly different from pre-treatment blood flow. Tumour and body temperature remained relatively stable at 28-30°C and 36- 37°C, respectively, over the 2 h sampling period (data not shown).

160 120 4 100 80- 60 - 4 0 - — Tumour (N=12; n=60) 2 0 -

0 1------1------1------1------1------1------1------1------1------1------1------1------r— -10 0 10 20 30 40 50 60 70 80 90 100 110 120

> 140- E 120- " O 1 0 0 - S 80-

Tumour (N=12; n=48)

2 0 -

140-

120 -

100 - 80- 6 0 - 4 0 - o— Skin (N=12; n=12) 2 0 -

10 0 10 20 30 40 50 60 70 80 90 100 110 120 Time after treatment (min)

Figure 6.5 Laser doppler blood flow of SaF-bearing mice (a) untreated tumour; treated with 0.5% BSA in saline (b) tumour (c) skin. Symbols show the mean ± s.e.m for 4 min period. N= number of mice ; n= number of probes.

161 Figure 6.6 shows the effect of rhu-TNF (a) and rmu-TNF (b) on rear dorsal SaF tumour blood flow. The data show that it takes 30 min before TNF, both human and murine, affects tumour vasculature and the resulting blood flow. Clearly, rhu-TNF significantly reduces TBF to 43 ± 7% (p <0.01) of pre-treatment values within one hour. After 2 h rhu-TNF has further reduced TBF to 16 ± 3.3% of controls. By contrast, there is an indication that skin blood flow increases within 16 min of TNF treatment, remains elevated for 1 h and then returns to control levels for the following hour. Murine TNF also reduces TBF, albeit less dramatically; within 30 min of treatment blood flow begins to drop progressively and is significantly reduced to 66 ± 5% of control levels after 1 h (p <0.01). Blood flow continues to decline reaching 48 ± 3% of control levels after 2 h. Skin blood flow initially increases within 15 min of rmuTNF treatment, to 125 ± 12% but at the end of the 2 h period it is not significantly different from pre-treatment levels. TNF-treatment did not adversely affect either tumour (foot or rear dorsal) or rectal temperature over the 2 h sampling period. A comparison of the effect of rhuTNF on blood flow in rear dorsal and foot tumours is shown in Figure 6.7. Treatment of SaF foot-tumours with rhuTNF causes vascular effects within 30 min which results in a significant reduction (p <0.01) in TBF within 1 h to 58 ± 13% of pre-treatment controls. Over the following hour TBF progressively decreases to 15 ± 3% which is similar to the effect of rhuTNF on rear dorsal TBF, as outlined above. Clearly, TNF reduces blood flow to the same extent in the SaF tumour irrespective of the site of tumour implantation and the resulting vascular structure. Figure 6.8 illustrates the difference in species specificity between rhu- and rmu­ TNF on SaF tumour compared to skin. Human TNF activates only the TNF-R55 in the mouse but has a more profound effect on TBF, at a 10-fold higher dose, than rmu-TNF. Thus, when rhu-TNF is used to treat SaF murine tumours, potentially important functions of the TNF-R75 are not directly manifested. However, both rmu- and rhu-TNF reduce SaF TBF to varying extents indicating involvement of the TNF-R55 in the mechanism.

162 Tumour (N=6; n =28) 350- Skin (N=6; n =6) 300-

250-

200 -

150-

^ 100-

50- O

O -10 10 30 50 70 90 110 O

160-

120 -

100 -

80-

60-

40-

2 0 - T umour (N=8 ; n =3 3) Skin (N=8;n=8)

10 10 30 50 70 90 110 Time after TNF treatment (min)

Figure 6.6

Effect of TNF on blood flow of SaF dorsal tumours (a) rhu-TNFa, 100 |ig/kg and

(b) rmu-TNFa, 10 |Lig/kg. Symbols represent the mean ± s.e.m of 4 min period. N= number of tumours; n= number of probes.

163 1 4 0 -

^ 100

Skin (N=12;n=48) Dorsal tumour (N=6; n =28) Foot tumour (N=6; n =23) T------'---- 1------'------1------'------r -10 0 10 20 30 40 50 60 70 80 90 100 110 120 Time after injection (min)

Figure 6.7 Effect of rhuTNF on SaF tumour blood flow compared to skin. Symbols represent the mean ± s.e.m of 4 minute period. N=number of tumours; n=number of probes.

164 N=12; n =48

& o E O O ÇQ d) N=6; n =28 >

N=8;n=33 Skin rhu-TNF (100 p-g/kg) rmu-TNF (lOiug/kg)

-10 0 10 20 30 40 50 60 70 80 æ 100 110 120 Time after treatment (min)

Figure 6.8 Effect of TNF on blood flow of SaF dorsal tumours compared to skin. Symbols represent the mean ± s.e.m of 4 min period. N = number of tumours; n=number of probes.

165 6.3 Discussion

It has been clearly demonstrated that SaF cells, although resistant to TNF cytotoxicity in vitro (in the absence of AMD or when pre-exposed to hypoxia), are susceptible to TNF cytotoxicity in vivo. Tumour response depends on TNF dose; a single i.v. injection of a relatively high but non-lethal dose of rhu-TNF (100 |Lig/kg) reduces SaF tumour cell survival significantly, causing two logs of cell kill and four days growth delay. It has been reported that growth reduction at such high TNF concentrations is caused by direct effects on tumour cells (Sugarman et al. 1985; Creasey et al. 1987). However, this is unlikely to be the case with SaF tumours because of their intrinsic TNF-resistance in vitro. In addition, it has previously been determined that SaF tumour cells exist in an environment in vivo that is largely hypoxic and pre­ exposure to hypoxia induces resistance to the direct cytotoxic action of TNF in vitro.

Early damage to the tumour vasculature was shown by the effect of TNF treatment on the blood flow of SaF tumours as assessed by multi-channel laser Doppler flowmetry. It was observed that systemic TNF treatment affects SaF tumour vasculature within 30 min, causing a dramatic, progressive, and significant reduction in tumour blood flow (TBF) compared to pre-treatment flow. Interestingly, a serum half-life after a single i.v. injection of TNF was determined to be 30 ± 2 min for rhu-TNF (Asher et al. 1987) and 12-26 min for rmu-TNF (Bemelmans et al. 1993). Despite the brief half-life reported for systemically administered TNF, single i.v. doses of rhuTNF were found to cause modest, but significant, growth delay (4 days) of SaF tumours.

Human recombinant TNF (100 |Xg/kg) causes a significant (p <0.01) reduction in TBF in both rear dorsal (-84%) and hind foot (-85%) tumours within two hours. Normal tissue (skin) blood flow was not affected by TNF treatment, which is in keeping with other studies (Kallinowski et al. 1989a; Kluge et al. 1992; Pimm et al. 1991). Clearly, blood flow data consistently show significant microvascular failure in both SaF foot and rear dorsal tumours after rhuTNF treatment but not total ischaemia over the 2 h sampling period.

166 Treatment with rmuTNF had similar, albeit less dramatic, affects as rhuTNF on blood flow of SaF dorsal tumours. TNF treatment (10 fig/kg, i.v.) resulted in a significant decrease in blood flow to 48% of pre-treatment values after 2 hours (p <0.01). A half-life of 12-26 min (depending on assay) was reported for muTNF (12.5 |uig/kg, i.v.) (Bemelmans et al 1993) which corresponds to the time-course of onset of vascular effects seen in the present studies. Overall, the actions of TNF on the vascular endothelium were not species specific, i.e., both rmu- and rhu-TNF cause a continuous decrease in the microcirculatory function with minimum SaF TBF values 2 h after treatment, this is in agreement with studies on other tumour systems (Sato et al 1986; Watanabe et al 1988b). It should be noted that muTNF has a 3-fold higher specific activity than huTNF and is also more toxic in mice (Brouckaert et al 1992).

Recently it was shown that the vascular effects of TNF in vivo, such as neutrophil adhesion to the endothelium and E-selectin expression, are mediated by TNF- R75 (Barbara et al 1994). The fact that rhuTNF, which can activate only the TNF-R55 on murine cells, reduces SaF TBF more than rmuTNF probably reflects the higher dose given but also suggests the involvement of TNF-R55 in the anti-vascular effects of TNF. However, treatment with exogenous TNF has been shown to induce endogenous muTNF mRNA within 30 min which in turn may account for specifically TNF-R75 mediated anti-vascular effects (de Kossodo et al 1995) and hence the greater reduction of TBF by rhuTNF in SaF tumours.

The Oxford Array^*^ multi-channel laser Doppler measures global TBF as probes are inserted in random sites in the tumour. In agreement with the present results, many studies have reported the acute vascular effects of TNF to be more marked in the center of the tumour with minimal damage at the tumour periphery (Watanabe et al 1988b; Carswell et al 1975; Haranaka et al 1984; Palladino et al 1987a; Havell et al 1988). Indeed, histological examination of TNF-treated SaF tumours showed extensive haemorrhagic necrosis (HN) throughout the tumour, with surviving cells found only at the periphery (data not shown), as reported for other tumour systems (Asher et al 1987; Sersa a/. 1988; Huang a/. 1995).

Overall, it was determined that TNF has a dramatic effect on SaF tumour blood flow within 2 h after administration, irrespective of tumour site, which results in severe

167 HN within the following 24 hours. The reduction in TBF was not caused by stress as blood flow of animals treated with TNF diluent only was unchanged over the 2 h sampling period. This suggests that the anti-tumour activity of TNF is mediated primarily through its anti-vascular effects on the SaF tumour.

In agreement with the present study, several investigators have reported that single administration of TNF (dose levels of 0.05-1 mg/kg) reduces blood flow significantly in rodent tumours and also in mice with human tumour xenografts (Watanabe et al 1988b; Kallinowski et al 1989a,b; Edwards et al 1991; Pimm et al 1991; Naredi et al 1993). Maximum TBF reduction was found 1 h after i.v. administration of TNF and it was reduced for the following 24 h period (Madhevan et al 1990) and only recovered to 50% of pre-treatment levels after 96 h (Naredi et al 1993). TNF-induced tumour microcirculatory failure results in continuously declining nutritive blood flow, ischaemic changes, and deterioration of nutrient and energy supply thus contributing to tumour cell death (Shine et al 1989; Kluge et al 1992). It is not clear whether the anti-vascular effect of TNF on SaF tumours in vivo is due to direct cytotoxic effects of TNF on tumour vasculature or the indirect result of vascular damage, ischaemia and activation of host stromal elements, or a combination of these effects.

The mechanism of TNF-induced reduction in TBF has not been investigated in this study. However, several mechanisms have been shown to be involved in this important anti-tumour activity of TNF including both systemic and tumour specific mechanisms (Tracey et al 1986b; MacPherson and North, 1986; Nawroth and Stern, 1986; Bevilacqua et al 1986; Shimomura et al 1988; Watanabe et al 1988b; Kallinowski et al 1989a; Van de Wiel et al 1990; Aicher et al 1990). These studies suggest that direct toxicity of TNF on tumour vasculature is a factor in the overall anti­ tumour mechanism of TNF. Indeed, the death of tumour cells is not observed until tumour vasculature is destroyed. However, whether the anti-vascular activity of TNF is direct or indirect it causes significant tumour cell cytotoxicity which results in modest but significant tumour growth delay. In any case, although TNF is selectively toxic to tumour vasculature, directly or indirectly, resulting in tumour necrosis, effects on tumour vasculature are not sufficient for complete and sustained tumour regression

168 (North and Havell, 1988). Thus, recovery of TBF may be one of the reasons for the modest growth delay seen in SaF tumours treated with TNF.

The immunogenicity of tumours is believed to be an important factor in TNF- induced tumour regression. It has been reported that the growth of non-immunogenic tumours was not inhibited by TNF despite vascular changes (Naredi et al 1993; Asher et al 1987). This suggests that the direct cytotoxic effect of TNF on tumours may be of primary importance in the treatment of human tumours which generally lack immunogenicity. TD 50 assays have shown no evidence of immunogenicity in the SaF tumour model (S.A.Hill, personal communication).

TNF alone and in combination has been used as an anti-tumour agent in phase I/n trials in a variety of regimes with approximately 1% response rate (Blick et al 1987; Creaven et al 1987; Kimura et al 1987; Selby et al 1987; Lenk et al 1989). In fact systemic therapy with TNF in human cancer has proved highly toxic and is inactive against the majority of tumours tested at tolerated doses. The tumour vasculature appears to be uniquely sensitive to TNF cytotoxicity. It has been shown that TNF has a differential effect on tumour and skin vasculature, tumour being more susceptible (Watanabe et al 1988b; Van de Wiel, 1989). As previously mentioned, it has been suggested that tumours produce substances that sensitise them to TNF (Asher et a l 1987).

One potential modulator of TNF anti-tumour activity is the oxygenation status of malignant tissue. It is well established that the median pOz is lower in malignant compared to normal tissue (Vaupel et al 1991; Lartigau et al 1993) and that hypoxia is a common feature of human tumours (Vaupel et al 1994; Hôckel et a l 1993). It has previously been established (c/ Chapter 3) that oxygen tensions known to exist in vivo induce resistance to the direct cytotoxic action of TNF on tumour cells in vitro. The limitations imposed by the toxicity of TNF at high doses and the rapid development of TNF-resistance by tumour cells in vivo are most likely to account for the poor results of this cytokine in clinical trials (Frei and Spriggs, 1989; Lattime and Stutman, 1989). Indeed, based on experimental animal data, the optimal doses of TNF needed for anti­ tumour activity have not been given in the clinic (Frei and Spriggs, 1989). The response rates of 40-50% seen in some clinical trials of intratumoural TNF confirm the dose

169 dependency of the anti-tumour effect of TNF in vivo (Taguchi and Sohmura, 1991). Improving tumour oxygenation status before TNF treatment could possibly increase the direct cytotoxic activity of TNF on tumour cells in vivo.

The effectiveness of TNF can be increased by reducing normal tissue exposure and thus increasing the maximum tolerated dose of TNF 20-fold results in cure rates as high as 90 percent in isolated limb perfusion (ILP) of advanced human tumours (Liénard et al. 1992a,b). Local therapy, such as ILP, has resulted in complete and long lasting tumour regressions in patients with melanoma and local tumour burdens, with necrotic activity confined solely to the tumour vascular bed, suggesting that optimal TNF doses if attainable can cause tumour cure (Liénard et al. 1992a,b; Bggermont et al. 1992). The utility of systemic TNF delivery would be much improved if ways could be found to enhance its activity selectively in tumours so that optimal TNF doses could be given. Improving tumour oxygenation status may reduce the TNF dose required to give beneficial effects. Thus, overcoming tumour hypoxia could potentially increase the cytotoxic action of TNF and together with the anti-vascular effects could perhaps improve the anti-tumour efficacy of TNF.

6.4 Summary

Systemic administration of relatively high non-toxic doses of rhuTNF causes considerable tumour toxicity and modest, but significant, growth delay of SaF tumours. TNF has a dramatic effect on SaF tumour vasculature: it reduces tumour blood flow significantly within 2 h which results in extensive haemorrhagic necrosis within 24 hours. The anti-tumour effect of TNF on SaF tumours in vivo appears to be based on the ability of TNF to selectively damage tumour vasculature.

170 Chapter 7 Concluding discussion

TNF has been given to cancer patients in phase I/n trials in a variety of regimes, however, no consistent anti-tumour activity was observed at tolerated doses (reviewed by Jones and Selby, 1989). Evidence indicates that the therapeutic value of TNF as an anti-tumour agent is limited by several factors, including toxic side-effects at high TNF doses, a wide variation in tumour cell sensitivity to TNF and, the acquisition of TNF- resistance (Kirstein et al 1986; Gerlach et al 1989; Frei and Spriggs, 1989; Lattime and Stutman, 1989; Chouaib et al 1992; Spriggs and Yates, 1992). In experimental models as well as in some human tumours the development of multi-drug resistance has been reported to be accompanied by an acquisition of resistance to TNF (Fruehauf et al 1991; Hofsli and Nissen-Meyer, 1989; Dollbaum et al 1988). Although the mechanisms by which cells become resistant to TNF are still surrounded by considerable controversy this study suggests that one factor involved in determining TNF response in vivo, which is known to differ between normal and malignant tissue, is tumour oxygenation.

Tumour oxygenation status was assessed in two experimental murine tumour models and it was shown that the majority of cells exist in an hypoxic microenvironment. These results are in agreement with several studies that have shown hypoxia to be a common feature of both experimental and human tumours but not of normal tissues (Guichard, 1989; Rockwell and Moulder, 1990; Lartigau et al 1993; Vaupel et al 1991; Vaupel et al 1994; Hôckel et al 1993). Moreover, this study clearly demonstrated that oxygen tension modulates the cytotoxic activity of TNF in vitro. Specifically, tumour cells pre-exposed to oxygen tensions known to exist in solid tumours in vivo, prior to TNF treatment, become resistant to the cytotoxic activity of TNF in vitro.

Hypoxia-induced TNF-resistance was demonstrated in vitro in a range of human and murine tumour cell lines and in one transformed endothelial cell line. It was shown to be dependent upon the cell line and both the level and duration of pre-exposure to hypoxia. In addition, it was shown that cells need to be pre-exposed to hypoxia for several hours before TNF-resistance is induced, which then increases proportionally with the duration of hypoxia. It was determined that hypoxia-induced TNF-resistance is

171 a transient process, as it is completely reversed by re-aeration of cells for several hours prior to TNF exposure. This suggests that tumour cells in vivo which are chronically hypoxic may exhibit TNF-resistance whereas transiently hypoxic cells may be more susceptible to TNF cytotoxicity. Thus, the compromised oxygenation status existing in many regions of human solid tumours could adversely affect tumour response to TNF cytotoxicity.

The mechanism of hypoxia-induced TNF-resistance was investigated by assessing the role of both cell cycle and protective protein expression. It was determined that hypoxia-induced modulation of cell cycle may play some role, albeit limited, in TNF-resistance of cells pre-exposed to < 0.2% oxygen. However, it has been clearly shown that for 2% oxygen cell cycle perturbations are not involved in the induction of TNF-resistance.

MnSOD is not involved in the hypoxia-induced TNF-resistance mechanism as it is not induced by hypoxia or even by re-aeration post-hypoxia. However, MnSOD may be involved in the intrinsic resistance of SaF cells to TNF cytotoxicity as it is constitutively expressed in this TNF-resistant murine cell line. One factor that was shown to be induced by hypoxia which is, at least partly, responsible for TNF-resistance, is intracellular endogenous TNF (enTNF). It was shown that the level of TNF-resistance corresponded with induction of enTNF production. In addition, enTNF produced by TNF-transfected cells was shown to induce resistance to exogenous TNF (exTNF) challenge. The precise role of hypoxia-induced enTNF in the mechanism of TNF resistance has not been determined. However, the finding that hypoxia-induced intracellular enTNF is involved in the TNF-resistance mechanism, either directly or indirectly, is of major importance but does not preclude the involvement of other protective factors.

Irrespective of the mechanism of hypoxia-induced resistance to TNFs’ cytotoxic action it is not surprising that high doses of TNF are necessary to have a therapeutic effect in vivo as the majority of tumour cells exist in an environment which is largely hypoxic. Anoxia protects tumour cells from most oxygen-dependent treatments (Freitas and Baronzio, 1991; Coleman, 1988). Furthermore, TNF mediates its cytotoxic activity via reactive oxygen species, particularly superoxide (Shoji et al 1995; Hennet et al

172 1993; Schulze-Osthoff et al. 1992), and anoxia at the time of treatment is known to protect cells from its cytotoxic action (Matthews and Neale, 1989). However, the present studies show that oxygen levels as low as 0.2% during TNF treatment have no effect on TNF cytotoxicity. By contrast, such tumour relevant oxygen tensions present subsequent to TNF exposure have potent effects in the development of resistance to its cytotoxic action. From a therapeutic viewpoint these results suggest that unlike radiation treatment where oxygen tension needs to be improved during treatment, such improvement during TNF exposure would result in little or no therapeutic benefit. However, improvement of tumour oxygenation status before TNF exposure may increase the direct cytotoxic activity of TNF on tumour cells in vivo. Increasing tumour oxygenation status may down-regulate enTNF production thus reducing the dose of exTNF required to give beneficial effects with a concomitant decrease in host toxicity.

SaF cells whilst resistant to TNF cytotoxicity in vitro in the absence of AMD, are susceptible to TNF cytotoxicity in vivo, which is in keeping with studies of other cell lines (Brouckaert et al. 1986; Palladino et al. 1987a,b; Van de Wiel et al. 1989). The susceptibility of SaF cells to TNF cytotoxicity was shown to be mediated mainly through the anti-vascular activity of TNF. Systemic TNF treatment has a profound effect on SaF tumour vasculature, causing significant reduction in tumour blood flow within a few hours which results in severe haemorrhagic necrosis within 24 h. However, it has not been determined whether the anti-tumour effect of TNF on SaF tumours in vivo is due to direct cytotoxic effect of TNF on tumour vasculature or the indirect result of changes in vascular permeability, procoagulant activity and adhesiveness of endothelium, or a combination of these effects. In any case, although TNF alters tumour vascular function, resulting in tumour necrosis, it has been reported that effects on tumour vasculature are not sufficient for complete and sustained tumour regression (North and Havell, 1988). This suggests that the cytotoxic effect of TNF on tumour cells and the immune stimulating effects are important in the anti-tumour activity of TNF. Thus, overcoming tumour hypoxia could, by increasing direct cytotoxic effects of TNF potentially increase the overall antitumour action of this cytokine. In summary, this study emphasizes the importance of performing in vitro experiments at relevant oxygen tensions as it has been shown that oxygenation conditions relevant to those found in solid tumours in vivo can adversely effect tumour cell sensitivity to the cytotoxic action of TNF in vitro. In addition it poses many

173 questions regarding other tumour microenvironmental conditions such as extracellular pH, pC02, levels of nitric oxide or cellular toxic products and availability of growth factors which may also contribute to a cells’ response to TNF action in vivo. It is possible that such tumour microenvironmental conditions, in addition to oxygen tension, may modulate both cytotoxicity and many of the other activities of this pleiotropic cytokine which have not been investigated in this study but may have important consequences in therapy with TNF. It is also possible that tumour oxygen tensions may modulate the activity of other cytokines and growth factors in vivo.

This study suggests that determining the level of tumour hypoxia may improve the understanding of the prognostic value of TNF in tumour therapy and in determining what approaches are required for its modulation. Furthermore, the improved understanding of the molecular basis of hypoxia-induced TNF-resistance may play a key role in shaping future strategies for more adaptable therapeutic modalities using TNF.

174 REFERENCES

Adachi, K., Belser, P., Bender, H., Li, D., Rodeck, U., Benveniste, E.N., Woo, D., Schmiegel, W.H., and Herlyn, D. (1992) Enhancement of epidermal growth factor receptor expression on glioma cells by recombinant tumor necrosis factor alpha. Cancer Immunol Immunother., 34, 370-376.

Aderka, D., Holtmann, H., Toker, L., Hahn, T., and Wallach, D. (1986) Tumor necrosis factor induction by Sendai virus. J Immunol, 136, 2938-2942.

Aderka, D., Le, J.M., and Vilcek, J. (1989) IL-6 inhibits lipopolysaccharide-induced tumor necrosis factor production in cultured human monocytes, U937 cells, and in mice. J Immunol, 143, 3517-3523.

Aggarwal, B.B. (1992) Comparative analysis of the structure and function of TNF-a and TNF-p. In Tumour Necrosis Factor: Structure, Function and Mechanism of Action (eds B.B. Aggarwal and J. Vilcek), pp 61-78. Marcel Dekker, New York.

Aggarwal, B.B., Kohr, W.J., Hass, P.E., Moffat, B., Spencer, S.A., Henzel, W.J., Bringman, T.S., Nedwin, G.E., Goeddel, D.V., and Harkins, R.N. (1985) Human tumour necrosis factor: Production, purification and characterisation. J Biol Chem., 260, 2345- 2354.

Aicher, K.P., Dupon, J.W., White, D.L., Aukerman, S.L., Moseley, M.E., luster, R., Rosenau, W., Winkelhake, J.L., and Brasch, R.C. (1990) Contrast-enhanced magnetic resonance imaging of tumor-bearing mice treated with human recombinant tumor necrosis factor alpha. Cancer Res., 50, 7376-7381.

Akira, S., Hirano, T., Taga, T., and Kishimoto, T. (1990) Biology of multifunctional cytokines: IL-6 and related molecules (IL-1 and TNF). FASEB J., 4, 2860-2867.

Alexander, R.B., Nelson, W.G., and Coffey, D.S. (1987) Synergistic enhancement by tumor necrosis factor of in vitro cytotoxicity from chemotherapeutic drugs targeted at DNA topoisomerase II. Cancer Res., 47, 2403-2406.

Amellem, O., and Pettersen, E. (1993) Cell cycle progression in human cells following re-oxygenation after extreme hypoxia: consequences concerning initiation of DNA synthesis. Cell Prolif, 26, 25-35.

Amellem, O., and Pettersen, E.G. (1991) Cell inactivation and cell cycle inhibition as induced by extreme hypoxia: the possible role of cell cycle arrest as a protection against hypoxia-induced lethal damage. Cell prolif, 24, 127-141.

Amellem, O., Loeffler, M., and Pettersen, E.G. (1994) Regulation of cell proliferation under extreme and moderate hypoxia: the role of pyrimidine (deoxy) nucleotides. Br J Cancer., 70, 857-866.

Anderson, V., Hellun-Larsen, P., and Sorensen, S.F. (1968) Gptimal oxygen tension for human lymphocytes in culture. J Cell Physiol, 72, 149-151.

175 Asher, A., Mule, J., Reichert, C., Shiloni, E., and Rosenberg, S. (1987) Studies on the anti-tumour efficacy of systemically administered recombinant tumor necrosis factor against several murine tumors in vivo. J Immunol, 138, 963-974.

Asoh, K., Watanabe, Y., Mizoguchi, H., Mawatari, M., Ono, M., Kohno, K., and Kuwano, M. (1989) Induction of manganese superoxide dismutase by tumor necrosis factor in human breast cancer MCF-7 cell line and its TNF-resistant variant. Biochem Biophys Res Commun., 162, 794-801.

Aune, T., and Pogue, S. (1989) Inhibition of tumor cell growth by interferon-y is mediated by two distinct mechanisms dependent upon oxygen tension: Induction of tryptophan degradation and depletion of intracellular nicotinamide adenine dinucleotide. J Clin Invest., 84, 863-875.

Bachwich, P R., Chensue, S.W., Larrick, J.W., and Kunkel, S.L. (1986) Tumor necrosis factor stimulates interleukin-1 and prostaglandin Ei production in resting macrophages. Biochem Biophys Res Commun., 136, 94-101.

Balkwill, F., Burke, F., Talbot, D., Tavernier, J., Osborne, R., Naylor, S., Durbin, H., and Fiers, W. (1987a) Evidence for tumour necrosis factor/cachectin production in cancer. Lancet., 2, 1229-1232.

Balkwill, F.R. (1989) Tumour necrosis factor and lymphotoxin. In Cytokines in Cancer Therapy, pp. 54-87. Oxford University Press, New York.

Balkwill, F.R., Lee, A., Aladam, G., Moodie, E., Thomas, J.A., Tavernier, J., and Fiers, W. (1986) Human tumor xenografts treated with recombinant human tumor necrosis factor alone or in combination with interferons. Cancer Res., 46, 3990-3993.

Balkwill, F.R., Ward, B.C. and Fiers, W. (1987b) Effects of tumour necrosis factor on human tumour xenografts in nude mice. In Tumour Necrosis Factor and Related Cytotoxins (eds G. Bock and J. Marsh), pp 154-169. Ciba Foundation Symposium, Wiley Chichester.

Banner, D.W., Darcy, A., Janes, W., Gentz, R., Schoenfeld, H.J., Broger, C., Loetscher, H., and Lesslauer, W. (1993) Crystal-structure of the soluble human 55 kd TNF receptor-human TNF-beta complex: Implications for TNF receptor activation. Cell, 73, 431-445.

Barbara, J.A., Smith, W.B., Gamble, J R., Van Ostade, X., Vandenabeele, P., Tavernier, J., Fiers, W., Vadas, M.A., and Lopez, A.F. (1994) Dissociation of TNF-alpha cytotoxic and proinflammatory activities by p55 receptor- and p75 receptor-selective TNF-alpha mutants. EMBO J., 13, 843-850.

Barnes, P.F., Fong, S.J., Brennan, P.J., Twomey, P.E., Mazumder, A., and Modlin, R.L. (1990) Local production of tumor necrosis factor and IFN-y in tuberculosis pleuritis. J Immunol, 145, 149-154.

Bate, C.A., Taverne, J., and Playfair, J.H. (1988) Malarial parasites induce TNF production by macrophages. Immunology., 64, 227-231.

176 Baud, L., Perez, J., Friedlander, G. and Ardaillou, R. (1988) Tumor necrosis factor stimulates prostaglandin production and cyclic AMP levels in rat cultured mesangial cells. FEBS Lett., 239, 50-54.

Bazzoni, P., Alejos, E., and Beutler, B. (1995) Chimeric tumor necrosis factor receptors with constitutive signaling activity. Proc Natl Acad Sci USA., 92, 5376-5380.

Becker, J.C., Kolanus, W., Lonnemann, C., and Schmidt, R.E. (1990) Human natural killer clones enhance in vitro antibody production by tumour necrosis factor alpha and gamma interferon. Scand J Immunol., 32, 153-162.

Bedford, J.S., and Mitchell, J.B. (1974) The effect of hypoxia on the growth and radiation response of mammalian cells in culture. Br J Radiol, 47, 687-696.

Belizario, J.E., and Dinarello, C.A. (1991) Interleukin 1, interleukin 6, tumour necrosis factor, and transforming growth factor beta increase cell resistance to tumour necrosis factor cytotoxicity by growth arrest in the G1 phase of the cell cycle. Cancer Res., 51, 2379-2385.

Bellomo, G., Perotti, M., Taddel, P., Mirabelli, P., Pinardi, G., and Nicotera, P. (1992) Tumor necrosis factor a induces apoptosis in mammary adenocarcinoma cells by an increase in intranuclear Ca^"^ concentration and DNA fragmentation. Cancer Res., 52, 1342-1346.

Bemelmans, M.H., Gouma, D.J., and Buurman, W.A. (1993) Influence of nephrectomy on tumor necrosis factor clearance in a murine model. J Immunol, 150, 2007-2017.

Berent, S.L., Torczynski, R.M., and Bollon, A.P. (1986) Sendai virus induces high levels of tumor necrosis factor mRNA in human peripheral blood leukocytes. Nucleic Acids Res., 14, 8997-9015.

Beresini, M.H., Lempert, M.J., and Epstein, L.B. (1988) Overlapping polypeptide induction in human fibroblasts in response to treatment with interferon-alpha, interferon-gamma, interleukin 1 alpha, interleukin 1 beta, and tumor necrosis factor. J Immunol, 140, 485-493.

Berne, R.M., and Rubio, R. (1974) Adenine nucleotide metabolism in the heart. Circ Res., 35, 109-120.

Bertolini, D R., Nedwin, G.E., Bringman, T.S., Smith, D.D., and Mundy, G.R. (1986) Stimulation of bone resorption and inhibition of bone formation in vitro by human tumour necrosis factors. Nature., 319, 516-518.

Bethea, J R., Chung, I.Y., Sparacio, S.M., Gillespie, G.Y., and Benveniste, E.N. (1992) Interleukin-1 beta induction of tumor necrosis factor-alpha gene expression in human astroglioma cells, y A^Mro/mmMrto/., 36, 179-191.

Beutler, B., Greenwald, D., Hulmes, J.D., Chang, M., Pan, Y. C., Mathison, J., Ulevitch, R., and Cerami, A. (1985b) Identity of tumour necrosis factor and the macrophage-secreted factor cachectin. Nature., 316, 552-554.

177 Beutler, B., Mahoney, J., LeTrang, N., Pekala, P., and Cerami, A. (1986) Control of cachectin (tumour necrosis factor) synthesis: mechanisms for endotoxin resistance. Science., 232, 977-980.

Beutler, B., Mahoney, J., LeTrang, N., Pekala, P., and Cerami, A. (1985a) Purification of cachectin, a lipoprotein lipase-suppressing hormone secreted by endotoxin-induced RAW 264.7 cells. J Exp Med., 161, 984-995.

Beutler, B., Milsark, I.W., and Cerami, A C. (1985c) Passive immunization against cachectin/tumor necrosis factor protects mice from lethal effect of endotoxin. Science., 229, 869-871.

Beutler, B., Thompson, P., Keyes, J., Hagerty, K. and Crawford, D. (1988) Assay of a ribonuclease that preferentially hydrolyses mRNAs containing cytokine-derived UA-rich instability sequences. Biochem Biophys Res Commun., 152, 973-980.

Bevilacqua, M., Pober, J., Majeau, G., Fiers, W., Cotran, R., and Gimbrone, M. (1986) Recombinant tumour necrosis factor induces procoagulant activity in cultured human vascular endothelium: Characterization and comparison with the actions of interleukin- 1. Proc Natl Acad Sci USA., 83, 4533-4537.

Bevilacqua, M.P., Pober, J.S., Mendrick, D.L., Cotran, R.S., and Gimbrone Jr, M.A. (1987) Identification of an inducible endothelial-leukocyte adhesion molecule. Proc Natl Acad Sci U SA., 84, 9238-9242.

Beyaert, R., Heyninck, K., De Valck, D., Boeykens, P., van Roy, P., and Piers, W. (1993) Enhancement of tumor necrosis factor cytotoxicity by lithium chloride is associated with increased inositol phosphate accumulation. J Immunol, 151, 291-300.

Bianchi, M., Mantovani, A., Erroi, A., Dinarello, C.A., and Ghezzi, P. (1990) Ambroxol inhibits interleukin-1 and tumor necrosis factor production in human mononuclear cells. Agents Actions., 31, 275-279.

Bigda, J., Beletsky, I., Brakebusch, C., Varfolomeev, Y., Engelmann, H., Bigda, J., Holtmann, H., and Wallach, D. (1994) Dual role of the p75 tumor necrosis factor (TNP) receptor in TNP cytotoxicity. J Exp Med., 180, 445-460.

Blick, M B., Sherwin, S.A., Rosenblum, M., and Gutterman, J. (1987) A phase I trial of recombinant human tumor necrosis factor in cancer patients. Cancer Res., 47, 2986- 2989.

Blood, C H., and Zetter, B.R. (1990) Tumor interactions with the vasculature: angiogenesis and tumor metastasis. Biochim Biophys Acta., 1032, 89-118.

Bonta, I.V., and Ben-Efraim, S. (1993) Involvement of inflammatory mediators in macrophage antitumor activity. J Leuk Biol, 54, 613-626.

Borish, L., Mascali, J.J., and Rosenwasser, L.J. (1991) IgE-dependent cytokine production by human peripheral blood mononuclear phagocytes. J Immunol, 146, 63- 67.

178 Born, R., Hug, O., and Trott, K.-R. (1976) The effect of prolonged hypoxia on growth and viability of Chinese hamster cells. Int J Radiat Oncol Biol Phys., 1, 687-697.

Boss, J.M., Laster, S.M., and Gooding, L.R. (1991) Sensitivity to tumour necrosis factor-mediated cytolysis is unrelated to manganous superoxide dismutase messenger RNA levels among transformed mouse fibroblasts. Immunology., 73, 309-315.

Boveris, A. (1977) Mitochondrial production of superoxide radical and hydrogen peroxide. Adv Exp Med Biol, 78, 67-82.

Brach, M.A., Gruss, H.J., Sott, C., and Herrmann, F. (1993) The mitogenic response to tumor necrosis factor alpha requires c-Jun/AP-1. Mol Cell Biol, 13 (7), 4284-4290.

Branellec, D., De Cremoux, P., Barreau, P., Calvo, P., and Chouaib, S. (1992) Tumor necrosis factor-mediated cell lysis in vitro: relationship to cAMP accumulation and guanine nucleotide-binding proteins. Eur J Immunol, 22, 963-967.

Brenner, D.A., O'Hara, M., Angel, P., Chojkier, M., and Karin, M. (1989) Prolonged activation of jun and collagenase genes by tumour necrosis factor-a. Nature., 337, 661- 663.

Brett, J., Gerlach, H., Nawroth, P., Steinberg, S., Godman, G., and Stern, D. (1989) Tumor necrosis factor/cachectin increases permeability of endothelial cell monolayers by a mechanism involving regulatory G proteins. J Exp Med., 169, 1977-1991.

Brockhaus, M., Schoenfeld, H-J., Schlaeger, E-J., Hunziker, W., Lesslauer, W., and Loetscher, H. (1990) Identification of two types of tumour necrosis factor receptors on human cell lines by monoclonal antibodies. Proc Natl Acad Sci USA., 87, 3127- 3131.

Brouckaert, P., Libert, C., Everaerdt, B., and Piers, W. (1992) Selective species specificity of tumor necrosis factor for toxicity in the mouse. Lymphokine Cytokine Res., 11, 193-196.

Brouckaert, P.G., Leroux Roels, G.G., Guisez, Y., Tavernier, J., and Piers, W. (1986) In vivo anti-tumour activity of recombinant human and murine TNP, alone and in combination with murine IPN-y, on a syngeneic murine melanoma. Int J Cancer., 38, 763-769.

Brown, J.M. (1979) Evidence for acutely hypoxic cells in mouse tumours, and a possible mechanism of reoxygenation. Br J Radiol, 52, 650-656.

Broxmeyer, H E., Williams, D.E., Lu, L., Cooper, S., Anderson, S.L., Beyer, G.S., Hoffman, R., and Rubin, B Y. (1986) The suppressive influences of human tumor necrosis factors on bone marrow hematopoietic progenitor cells from normal donors and patients with leukemia: synergism of tumor necrosis factor and interferon-gamma. J Immunol, 136, 4487-4495.

Candela, M., Barker, S.C., and Ballou, L.R. (1991) Sphingosine synergistically stimulates tumor necrosis factor alpha-induced prostaglandin E? production in human fibroblasts, y E'xp MeJ., 174, 1363-1369.

179 Cao, X.M., Guy, G.R., Sukhatme, V.P., and Tan, Y.H. (1992) Regulation of the Egr-1 gene by tumor necrosis factor and interferons in primary human fibroblasts. J Biol Chem., 267, 1345-1349.

Caput, D., Beutler, B., Hartog, K., Brown-Shimer, S., and Cerami, A. (1986) Identification of a common nucleotide sequence in the 3'-untranslated region of mRNA molecules specifying inflammatory mediators. Proc Natl Acad Sci USA., 83, 1670-1674.

Carswell, E.A., Old, L.J., Kassel, R.L., Green, S., Fiore, N., and Williamson, B. (1975) An endotoxin-induced serum factor that causes necrosis of tumors. Proc Natl Acad Sci USA., 72, 3666-3670.

Chan, H., and Aggarwal, B.B. (1994) Role of tumor necrosis factor receptors in the activation of nuclear factor kappa B in human histiocytic lymphoma U-937 cells. J Biol Chem., 269, 31424-31429.

Chang, D.J., Ringold, G.M., and Heller, R.A. (1992) Cell killing and induction of manganous superoxide dismutase by tumor necrosis factor-alpha is mediated by lipoxygenase metabolites of arachidonic acid. Biochem Biophys Res Commun., 188, 538-546.

Chaplin, D.J., and Acker, B. (1987) The effect of hydralazine on the tumour cytotoxicity of the hypoxic cell cytotoxin RSU-1069; evidence for therapeutic gain. Int J Rad Oncol Biol Phys., 13, 579-585.

Chaplin, D.J., and Hill, S.A. (1995) Temporal heterogeneity in microregional erythrocyte flux in experimental solid tumours. Br J Cancer., 71, 1210-1213.

Chaplin, D.J., and Trotter, M. (1991) Chemical modifiers of tumor blood supply. In Tumour Blood Supply and Metabolic Microenvironment (eds P. Vaupel and R.K. Jain) pp 65-85. Gustav Fischer Verlag, Stuttgart.

Chapman, J., Sturrock, J., Boag, J., and Crookall, J. (1970) Factors affecting the oxygen tension around cells growing in plastic Petri dishes. Int J Radiat Biol, 17, 305-328.

Chen, M.J., Holskin, B., Strickler, J., Gorniak, J., Clark, M.A., Johnson, P.J., Mitcho, M., and Shalloway, D. (1987) Induction by ElA oncogene“ expression of cellular susceptibility to lysis by TNF. Nature., 330, 581-583.

Chiang, C.S., and McBride, W.H. (1991) Radiation enhances tumor necrosis factor alpha production by murine brain cells. Brain Res., 566, 265-269.

Chirgwin, J.M., Przybyla., A.E., MacDonald, R.J., and Rutter, W. (1979) Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. J Biochem., 18, 5294-5297.

Chouaib, S., Robinet, E., Zyad, A., and Branellec, D. (1992) Tumor necrosis factor: pleiotropic cytokine. Bull Cancer Paris., 79, 935-949.

Chung, I.Y., Kwon, J., and Benveniste, E.N. (1992) Role of protein kinase C activity in tumor necrosis factor-alpha gene expression. Involvement at the transcriptional level. J Immunol, 149, 3894-3902.

180 Clark, I.A., and Chaudhri, G. (1988). Roles of tumor necrosis factor in malaria. In Tumor Necrosis Factor/Cachectin and Related Cytokines, (eds B. Bonavida, G.E. Gifford, H. Kirchner and L.J. Old), pp. 240. Krager, Basel.

Clauss, M., Ryan, J., and Stern, D. (1992). Modulation of endothelial cell hemostatic properties by TNF: Insights into the role of endothelium in the host response to inflammatory stimuli. In Tumour Necrosis Factors. The Molecules and Their Emerging Role in Medicine, (ed B. Beutler) pp 49-63. Raven Press, New York.

Cockfield, S.M., Ramassar, V., and Halloran, P.P. (1993) Regulation of IFN-y and tumor necrosis factor-a expression in vivo. Effects of cycloheximide and cyclosporine in normal and lipopolysaccharide-treated mice. J Immunol, 150, 342-352.

Coleman, C.N. (1988) Hypoxia in tumors: A paradigm for the approach to biochemical and physiologic heterogeneity. J Natl Cancer Inst., 80, 310-317.

Coley, W.B. (1891) Contribution to the knowledge of sarcoma. Ann Surg.,l<\, 199-220.

Coley, W.B. (1893) The treatment of malignant tumors by repeated inoculations of erysipelas: With a report of ten original cases. Am J Med Sci., 105, 487-511.

Coley, W.B. (1906) Late results of the treatment of inoperable sarcoma by the mixed toxins of erysipelas and bacillus prodigiousus. Am J Med Sci., 131, 375-430.

Coley-Nauts, H., Fowler, G.A.A., and Bogatko, F.H. (1953) A review of the influence of bacterial infections and bacterial products (Coley's toxins) on malignant tumours in man. Acta Medica Scand (Suppl.)., 274, 29-97.

Collart, M.A., Baeuerle, P., and Vassalli, P. (1990) 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. Mol Cell Biol, 10, 1498-1506.

Collart, M.A., Belin, D., Vassalli, J.D., de Kossodo, S., and Vassalli, P. (1986) Gamma interferon enhances macrophage transcription of the tumor necrosis factor/cachectin, interleukin 1, and urokinase genes, which are controlled by short-lived repressors. J Exp Med., 164,2113-2118.

Collawn, J.F., Stangel, M., Kuhn, L.A., Esekogwu, V., ling, S.Q., Trowbridge, I.S., and Tainer, J.A. (1990) Transferrin receptor internalization sequence YXRF implicates a tight turn as the structural recognition motif for endocytosis. Cell, 63, 1061-1072.

Collins, J.L., Patek, P.Q., and Cohn, M. (1981) Tumorigenicity and lysis by natural killers. J Exp Med., 152, 89-106.

Collins, T., Lapierre, L.A., Fiers, W., Strominger, J.L., and Pober, J.S. (1986) Recombinant human tumor necrosis factor increases mRNA levels and surface expression of HLA-A, B antigens in vascular endothelial cells and dermal fibroblasts in vitro. Proc Natl Acad Sci USA., 83, 446-450.

Cornelius, P., Enerback, S., Bjursell, G., Olivecrona, T., and Pekala, P H. (1988) Regulation of lipoprotein lipase mRNA content in 3T3-L1 cells by tumour necrosis factor. Biochem J., 249, 765-769.

181 Corr, P.B., Shayman, J.A., Kramer, J.B., and Kipnis, R.J. (1981) Increased alpha- adrenergic receptors in ischemic cat myocardium. A potential mediator of electrophysiological derangements. J Clin Invest., 67, 1232-1236.

Crabtree, H.G., and Cramer, W. (1933) The action of radium on cancer cells U. Some factors determining the susceptibility of cancer cells to radium. Proc Roy Soc London, Ser.B., 113, 238-250.

Creasey, A.A., Doyle, L.V., Reynolds, M.T., Jung, T., Lin, L.S., and Vitt, C.R. (1987) Biological effects of recombinant human tumour necrosis factor and its novel muteins on tumour and normal cells. Cancer Res., 47, 145-149.

Creasey, A.A., Reynolds, T. and Laird, W. (1986) Cures, and partial regression of murine and human tumors by recombinant human tumor necrosis factor. Cancer Res., 46, 5687-5690.

Creaven, P.J., Plager, J.E., Dupe re. S., Huben, R.P., Takita, H., Mittelma, A., and Proefrock, A. (1987) Phase I trial of recombinant human tumor necrosis factor. Cancer Chemother Pharmacol, 20, 317-344.

Currie, G.A., and Bosham, C. (1975) Activated macrophages release a factor which lyses malignant cells but not normal cells. J Exp Med., 142, 1600-1605.

Cuturi, M.C., Murphy, M., Costa-Giomi, M.P., Weinmann, R., Perussia, B., and Trinchieri, G. (1987) Independent regulation of tumor necrosis factor and lymphotoxin production by human peripheral blood lymphocytes. J Exp Med., 165, 1581-1594.

Danis, V.A., Franic, G.M., Rathjen, D.A., and Brooks, P.M. (1991) Effects of granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-2, interferon-gamma (IFN-y), tumor necrosis factor-alpha (TNF-a) and IL-6 on the production of immunoreactive IL-1 and TNF-a by human monocytes. Clin Exp Immun., 46, 143-150.

Darzynkiewicz, Z., Williamson, B., Carswell, E.A., and Old, L.J. (1984) Cell cycle- specific effects of tumour necrosis factor. Cancer Res., 44, 83-90.

Das, A.K., Walther, P.J., Buckley, N.J., and Poulton, S.H. (1989) Recombinant human tumor necrosis factor alone and with chemotherapeutic agents. Effect on nude mouse- supported human bladder cancer heterografts. Arch Surg., 124, 107-110.

Dayer, J.M., Beutler, B., and Cerami, A. (1985) Cachectin/tumour necrosis factor stimulates collagenase and prostaglandin E 2 production by human synovial cells and dermal fibroblasts. J Exp Med., 162,2163-2168. de Kossodo, S., Critico, B., and Grau, G.E. (1994) Modulation of the transcripts for tumor necrosis factor-a and its receptors in vivo. Eur J ImmiioL, 24, 769-772. de Kossodo, S., Moore, R., Gschmeissner, S., East, N., Upton, C., and Balkwill, F.R. (1995) Changes in endogenous cytokines, adhesion molecules and platelets during cytokine-induced tumour necrosis. Br J Cancer., 72, 1165-1172.

182 De Marco, R., Ensor, J E., and Hasday, J.D. (1992) Tumor-stimuiated release of tumor necrosis factor-alpha by human monocyte-derived macrophages. Cell Immunol, 140, 304-318.

De Valck, D., Beyaert, R., Van Roy, P., and Piers, W. (1993) Tumor necrosis factor cytotoxicity is associated with phospholipase D activation. Eur J Biochem., 212, 491- 497.

De, S.K., McMaster, M.T., and Andrews, G.K. (1990) Endotoxin induction of murine metallothionein gene expression. J Biol Chem., 265, 15267-15274.

Deal try, G.B., Naylor, M.S., Piers, W., and Balkwill, P.R. (1987) The effect of recombinant human tumour necrosis factor on growth and macromolecular synthesis of human epithelial cells. Exp Cell Res., 170, 428-438.

Degliantoni, G., Murphy, M., Kobayashi, M., Prancis, M.K., Perussia, B., and Trinchieri, G. (1985) Natural killer (NK) cell-derived hematopoietic colony-inhibiting activity and NK cytotoxic factor. Relationship with tumor necrosis factor and synergism with immune interferon. J Exp Med., 162, 1512-1530.

Del Maestro, R.P., Lopez Torres, M., McDonald, W.B., Stroude, E.G., and Vaithilingam, I.S. (1992) The effect of tumor necrosis factor-alpha on human malignant glial cells. J Neurosurg., 76, 652-659.

Dembic, Z., Loetscher, H., Gubler, U., Pan, Y-C.E., Lahm, H-W., Gentz, R., Brockhaus, M., and Lesslauer, W. (1990) Two human intracellular TNP receptors have extracellular but distinct intracellular domain sequences. Cytokine., 2, 231-237.

Denekamp, J., Hill, S.A., and Hobson, B. (1983) Vascular occlusion and tumour cell death. Eur J Cancer Clin Oncol, 19, 271-275.

Dolbeare, P., Gratznert, H.G., Pallavicini, M., and Gray, J.W. (1983) Plow cytometric measurement of total DNA content and incorporated bromodeoxyuridine. Proc Soc Natl AcadSci USA., 80, 5573-5577.

Dollbaum, C., Creasey, A.A., Dairkee, S.H., Hiller, A.J., Rudolph, A.R., Lin, L., Vitt, C., and Smith, H.S. (1988) Specificity of tumor necrosis factor toxicity for human mammary carcinomas relative to normal mammary epithelium and correlation with response to doxorubicin. Proc Natl Acad Sci USA., 85, 4740-4744.

Dorn, A., Durand, B., Marfing, C., Le Meur, M., Benoist, C., and Mathis, D. (1987) Conserved major histocompatibility complex class II boxes-X and -Y are transcriptional control elements and specifically bind nuclear proteins. Proc Natl Acad Sci USA., 84, 6249-6253.

Douglas, B.G., and Powler, J.P. (1976) The effect of multiple small doses of X rays on skin reactions in the mouse and a basic interpretation. Radiat Res., 66, 401-426.

Dressier, K.A., Mathias, S., and Kolesnick, R.N. (1992) Tumor necrosis factor-alpha activates the sphingomyelin signal transduction pathway in a cell-free system. Science., 255, 1715-1718.

183 Drysdale, B E., Zacharchuk, C.M., and Shin, H.S. (1983) Mechanism of macrophage- mediated cytotoxicity: production of a soluble cytotoxic factor. J Immunol, 131, 2362- 2367.

Duh, E.J., Maury, W.J., Folks, T.M., Fauci, A.S., and Rabson, A.B. (1989) Tumor necrosis factor alpha activates human immunodeficiency virus type 1 through induction of nuclear factor binding to the NF-kappa B sites in the long terminal repeat. Proc Natl Acad Sci USA., 86, 5974-5978.

Dustin, M.L., and Springer, T.A. (1989) T-cell receptor cross-linking transiently stimulates adhesiveness through LFA-1. Nature., 341, 619-624.

Eck, M.J., and Sprang, S.R. (1989) The structure of tumour necrosis factor-a at 2.6 Â resolution. J Biol Chem., 264, 17595-17605.

Edwards, H.S., Bremner, J.C., and Stratford, I.J. (1991) Induction of tumour hypoxia by FAA and TNF: interaction with bioreductive drugs. Int J Radiat Biol, 60, 373-377.

Eggermont, A.M., Goey, S.H., Bolhuis, R.L., and Stoter, 0. (1992) Regional application of cytokines: experience with various clinical protocols. Pharmacol Res., 26, 54-55.

Elkind, M.M., and Sutton, H. (1960) Radiation response of mammalian cells grown in culture. I. Repair of x-ray damage in surviving Chinese hamster cells. Radiat Res., 13, 556-593.

Emoto, N., and Baird, A. (1988) The effect of tumor necrosis factor/cachectin on follicle-stimulating hormone-induced aromatase activity in cultured rat granulosa cells. Biochem Biophys Res Commun., 153, 792-798.

Enari, M., Hug, H., and Nagata, S. (1995) Involvement of an ICE-like protease in Fas- mediated apoptosis. Nature., 375, 78-81.

Engelmann, H., Aderka, D., Rubinstein, M., Rotman, D., and Wallach, D. (1989) A tumor necrosis factor-binding protein purified to homogeneity from human urine protects cells from tumor necrosis factor toxicity. J Biol Chem., 264, 11974-11980.

Erikstein, B.K., Smeland, E.B., Blomhoff, H.K., Funderud, S., Prydz, K., Lesslauer, W., and Espevik, T. (1991) Independent regulation of 55-kDa and 75-kDa tumor necrosis factor receptors during activation of human peripheral blood B lymphocytes. Eur J Immunol, 21, 1033-1037.

Essner, R., Rhoades, K., McBride, W.H., Morton, D.L., and Economou, J.S. (1989) IL-4 down-regulates EL-l and TNF gene expression in human monocytes. J Immunol, 142, 3857-3861.

Fatt, P. (1976) Proceedings: The space-clamped retina. J Physiol Lond., 258, 52-54.

Fiers, W. (1991) Tumor necrosis factor. Characterization at the molecular, cellular and in vivo level. FEBS Lett., 285, 199-212.

184 Fiers, W. (1995) Biologie Therapy with TNF: Preclinical Studies. In Biologie Therapy of Cancer, (eds V.T. De Vita Jr., S. Heilman., and S.A. Rosenberg), pp 295-327. J.P. Lippincott Company, Philadelphia.

Fiers, W., Brouckaert, P., Devos, R., Fransen, L., Leroux Roels, G., Remaut, E. Suffys, P., Tavernier, J., Van der Heyden, J., and Van Roy, F. (1986) Lymphokines and monokines in anti-cancer therapy. Cold Spring Harb Symp Quant Biol, 51, 587-95.

Fiers, W., Brouckaert, P., Goldberg, A.L., Kettelhut, I., Suffys, P., Tavernier, J., Vanhaesebroeck, B., and Van Roy, F (1987). Structure - function relationship of tumour necrosis factor and its mechanism of action. In Tumour Necrosis Factor and Related Cytotoxins, pp 109-123. Ciba Foundation Symposium, Wiley Chichester.

Flethcher, W.H., Shiu, W.W., Ishida, T.A., Haviland, D.L., and Ware, G.F. (1987) Resistance to the cytolytic action of lymphotoxin and tumour necrosis factor coincides with the presence of gap junctions uniting target cells. J Immunol, 139, 956-962.

Flick, D.A., and Gifford, G.E. (1984) Comparison of in vitro cell cytotoxic assays for tumour necrosis factor. J Immunol Methods., 68, 167-175.

Foa, R., Guarini, A., Francia di Celle, P., Trentin, L., Gillio Tos, A., Bellone, G., Carbone, A., Attisano, C., Massaia, M., Raspadori, D., Pizzolo, G., Chisesi, T., Lauria, F., and Semenzato, G. (1992) Constitutive production of tumor necrosis factor-alpha in hairy cell leukemia: possible role in the pathogenesis of the cytopenia(s) and effect of treatment with interferon-alpha. J Clin Oncol, 10, 954-959.

Folkman, J. (1971) Tumor angiogenesis: therapeutic implications. N Engl J Med., 285, 1182-1186.

Folkman, J. (1990) Endothelial cells and angiogenic growth factors in cancer growth and metastasis. Introduction. Cancer Metastasis Rev., 9, 171-174.

Ford, D.K., and Yerganian, G. (1958) Observations on the chromosomes of Chinese hamster cells in tissue culture J Natl Cancer Inst., 21, 393-425.

Fowler, J.F., Sheldon, P.W., Begg, A C., Hill, S.A., and Smith, A.M. (1975) Biological properties and response to X-rays of first-generation transplants of spontaneous mammary carcinomas in C3H mice. Int J Radiat Biol, 27, 463-480.

Fraker, D.L., Alexander, H R., and Pass, H I. (1995) Biologic Therapy with TNF: Systemic Administrationm and Isolation-Perfusion. In Biologic Therapy of Cancer, (eds V.T. DeVita, Jr; S. Heilman and S.A. Rosenberg), pp 329-345. J.P. Lippincott Company, Philadelphia.

Fraker, D.L., Sheppard, B.C., and Norton, J.A. (1990) Impact of tolerance on antitumor efficacy of tumor necrosis factor in mice. Cancer Res., 50, 2261-2267.

Fraker, D.L., Stovroff, M.C., Merino, M.J., and Norton, J.A. (1988) Tolerance to tumor necrosis factor in rats and the relationship to endotoxin tolerance and toxicity. J E.xp Med., 168,95-105.

Fransen, L., Muller, R., Marmenout, A., Tavernier, J., Van der Heyden, J., Kawashima, E., Chollet, A., Tizard, R., Van Heuverswyn, H., Van Vliet, A., Ruysschaert, M R., and

185 Fiers,W.( 1985) Moleular cloning of mouse tumor necrosis factor cDNA and its eukaryotic expression. Nucleic Acids Res., 13, 4417-4429.

Fransen, L., Ruysschaert, M R., Van der Heyden, J., and Fiers, W. (1986) Recombinant tumor necrosis factor: species specificity for a variety of human and murine transformed cell lines. Cell Immunol, 100, 260-267.

Frater Schroder, M., Risau, W., Kallmann, R., Gautschi, P., and Bohlen, P. (1987) Tumor necrosis factor type alpha, a potent inhibitor of endothelial cell growth in vitro, is angiogenic in vivo. Proc Natl Acad Sci USA., 84, 5277-5281.

Freeman, B.A., and Crapo, J.D. (1982) Biology of disease: free radicals and tissue injury. Lab Invest., 47, 412-426.

Frei, E., and Spriggs, D. (1989) Tumor necrosis factor: Still a promising agent. J Clin Oncol, 7, 291-294.

Freitas, I., and Baronzio, G.F. (1991) Tumor hypoxia, reoxygenation and oxygenation strategies: possible role in photodynamic therapy. J Photochem Photobiol B., 11, 3-30.

Fruehauf, J.P., Mimnaugh, E.G., and Sinha, B.K. (1991) Doxorubicin-induced cross­ resistance to tumor necrosis factor (TNF) related to differential TNF processing. J Immunother., 10, 165-173.

Fuchs, P., Strehl, S., Dworzak, M., Himmler, A., and Ambros, P.F. (1992) Structure of the human TNF receptor 1 (p60) gene (TNFRl) and localization to chromosome 12pl3. Genomics., 13, 219-224.

Fujii, J., and Taniguchi, N. (1991) Phorbol ester induces manganese-superoxide dismutase in tumor necrosis factor-resistant cells. J Biol Chem., 266, 23142-23146.

Fujita, T., Revs, L.F.L., Wantanbe, N., Kimura, Y., Taniguchi, T., and Vilcek, J. (1989) Induction of the transcription factor IRF-1 and interferon-13 mRNAs by cytokines and activators of second messenger pathways. Proc Natl Acad Sci USA., 86, 9936-9940.

Gamble, J R., Smith, W.B., and Vadas, M.A. (1992) TNF modulation of endothelial and neutrophil adhesion. In Tumour Necrosis Factors. The Molecules and Their Emerging Role in Medicine, (ed B. Beutler) pp 65-86. Raven Press, New York.

Gatenby, R.A., Kessler, H.B., Rosenblum, J.S., Coia, L.R., Moldofsky, P.J., Hartz, W.H., and Broder, G.J. (1988) Oxygen distribution in squamous cell carcinoma métastasés and its relationship to outcome of radiation therapy. Int J Radiat Oncol Biol Phys., 14, 831-838.

Gearing, A.J.H., Beckett, P., Christodoulou, M., Churchill, M., Clements, J.M., Crimmin, M., Davidson, A.H., Drummond, A.H., Galloway, W.A., Gilbert, R., Gordon, J.L., Leber, T.M., Mangan, M., Miller, K., Nayee, P., Owen, K., Patel, S., Thomas, W., Wells, G., Wood, L.M., and Woolley, K. (1995) Matrix metalloproteinases and processing of pro-TNF-a. J Leuk Biol, 57, 774-777.

Georg, B., Helseth, E., Lund, L.R., Skandsen, T., Riccio, A., Dano, K., Unsgaard, G., and Andreasen, P.A. (1989) Tumor necrosis factor-alpha regulates mRNA for

186 urokinase-type plasminogen activator and type-1 plasminogen activator inhibitor in human neoplastic cell lines. Mol Cell Endocrinol., 61, 87-96.

Gerlach, H., Lieberman, H., Bach, R., Godman, G., Brett, J., and Stern, D. (1989) Enhanced responsiveness of endothelium in the growing/motile state to tumor necrosis factor/cachectin J Exp Med., 170,913-31.

Gerlach, H.G., Gerlach, M., and Clauss M (1993) Relevance of tumour necrosis factor- alpha and interleukin-1-alpha in the pathogenesis of hypoxia-related organ failure. Euro J Anaesthes., 10, 273-285.

Ghezzi, P., Dinarello, C.A., Bianchi, M., Rosandich, M.E., Repine, I.E., and White, C.W. (1991) Hypoxia increases production of interleukin-1 and tumor necrosis factor by human mononuclear cells. Cytokine., 3, 189-194.

Gifford, G.E., and Flick, D.A. (1987) Natural production of tumor necrosis factor. In Tumor Necrosis Factor and Related Cytotoxins, (eds G. Bock and J. Marsh) pp 3-20. Ciba Foundation Symposium 131, John Wiley and Sons, Chichester, U.K.

Gifford, G.E., and Lohmann-Matthes, M.L. (1986) The requirement for the continual presence of lipopolysaccharide for the production of tumor necrosis factor by thioglycollate-induced peritoneal murine macrophages. Int J Cancer., 78, 121-124.

Gilmore, T.D. (1990) NF-kappa B, KBFl, dorsal, and related matters. Cell, 62, 841- 843.

Giroir, B.P., Brown, T., and Beutler, B. (1992) Constitutive synthesis of tumor necrosis factor in the thymus. Proc Natl Acad Sci USA., 89, 4864-4868.

Godfrey, R.W., Johnson, W.J., and Hoffstein, S.T. (1987) Recombinant tumor necrosis factor and interleukin-1 both stimulate human synovial cell arachidonic acid release and phospholipid metabolism. Biochem Biophys Res Commun., 142, 235-241.

Goeddel, D.V., Aggarwal, B.B., Gray, P.W., Leung, D.W., Nedwin, G.E., Palladino, M. A., Patton, J.S., Pennica, D., Shepard, H.M., Sugarman, B.J., and Wong, G.H.W. (1986) Tumor necrosis factors: gene structure and biological activities. Cold Spring Harb Symp Quant Biol, 51,591-609.

Goossens, V., Grooten, J., De Vos, K., and Fiers, W. (1995) Direct evidence for tumor necrosis factor-induced mitochondrial reactive oxygen intermediates and their involvement in cytotoxicity. Proc Natl Acad Sci USA., 92, 8115-8119.

Gosselin, J., Flamand, L., D. Addairo, M., Hiscott, J., and Menezes, J. (1992) Infection of peripheral blood mononuclear cells by herpes simplex and Epstein-Barr viruses. Differential induction of interleukin 6 and tumor necrosis factor-alpha. J Clin Invest., 89, 1849-1856.

Grau, G., de kossodo. S., Pointaire, P., Ody, C., Bonnefoy, J.Y., Piguet, P.F., and Vassalli, P. (1990a) Involvement of tumor necrosis factor/cachectin in murine development. Proceedings of the Third International Conference on Tumor Necrosis Factor and Related Cytokines, Makuhari, Japan. November 21-25, 1990, p. 190.

187 Grau, G.E., Bieler, G., Pointaire, P., de Kossodo, S., Tacchini Cotier, F., Vassalli, P., Piguet, P.P., Lambert, P.H., Han, J., and Beutler, B. (1990b) Significance of cytokine production and adhesion molecules in malarial immunopathology The essential role of the UA-rich sequence in endotoxin-induced cachectin/TNF synthesis. Immunol Lett., 25, 189-194.

Grau, G.E., Piguet, P.P., Vassalli, P., and Lambert, P.H. (1989) Tumor-necrosis factor and other cytokines in cerebral malaria: experimental and clinical data. Immunol Rev., 112, 49-70.

Gray, L.H. (1957) Oxygenation in radiotherapy: I. radiobiological considerations. Br J Radiology., 30, 403-406.

Gray, P.W., Aggarwal, B., Benton, C.V., Bringman, T.S., Henzel, W.J., Jarret, J.A., Leug, D.W., Moffat, B., Ng, P., Svedersky, L.P., Palladino, M.A., and Nedwin, G.E. (1984) Cloning and expression of cDNA for human lymphotoxin, a lymphokine with tumour necrosis activity. Nature., 312, 712-724.

Gray, P.W., Barrett, K., Chantry, D., Turner, M., and Feldmann, M. (1990) Cloning of human tumour necrosis factor (TNF) receptor cDNA and expression of recombinant soluble TNF-binding protein. Proc Natl Acad Sci USA., 87, 7380-7384.

Green, S., Dobrjansky, A., Carswell, E.A., Kassel, R.L., Old, L.J., Fiore, N., and Schwartz, M.K. (1976) Partial purification of a serum factor that causes necrosis of tumors. Proc Natl Acad Sci USA., 73, 381-385.

Gresser, I., Belardelli, P., Tavernier, J., Piers, W., Podo, P., Federico, M., Carpinelli, G., Duvillard, P., Prade, M., Maury, C., et al. (1986) Anti-tumor effects of interferon in mice injected with interferon-sensitive and interferon-resistant Friend leukemia cells. V. Comparisons with the action of tumor necrosis factor. Int J Cancer., 38, 771-778.

Grooten, J., Goossens, V., Vanhaesebroeck, B., and Piers, W. (1993) Cell membrane permeabilization and cellular collapse, followed by loss of dehydrogenase activity: early events in tumour necrosis factor-induced cytotoxicity. Cytokine., 5, 546-555.

Guesdon, P., Preshney, N., Waller, R.J., Rawlinson, L., and Saklatvala, J. (1993) Interleukin 1 and tumor necrosis factor stimulate two novel protein kinases that phosphorylate the heat shock protein hsp27 and p-casein. J Biol Chem., 268, 4236-4243.

Guichard, M. (1989) Comparison of the radiobiological properties of human tumour xenografts and rodent tumours. Int J Rad Biol, 56, 583-586.

Guy, G.R., Cairns, J., Ng, S.B., and Tan, Y.H. (1993) Inactivation of a redox-sensitive protein phosphatase during the early events of tumor necrosis factor/interleukin-1 signal transduction. J Biol Chem., 268, 2141-2148.

Hahn, T., Toker, L., Budilovsky, S., Aderka, D., Eshnar, Z., and Wallach, D. (1985) Use of monoclonal antibodies to a human cytotoxin for its isolation and for examining the self-induction of resistance to this protein. Proc Natl Acad Sci USA., 82, 3814-3818.

Hall, E.J. ( 1988). Radiobiology for the radiologist. Harper and Row, Publishers, Hagerstown, Maryland.

188 Hallahan, D.E., Spriggs, D R., Beckett, M.A., Kufe, D.W., and Weichselbaum, R.R. (1989) Increased tumor necrosis factor alpha mRNA after cellular exposure to ionizing radiation. Proc Natl Acad Sci USA., 86, 10104-10107.

Han, J., Brown, T., and Beutler, B. (1990) Endotoxin-responsive sequences control cachectin/tumor necrosis factor biosynthesis at the transcriptional level. J Exp Med., I l l , 465-475.

Haranaka, K., and Satomi, N. (1981) Cytotoxic activity of tumor necrosis factor (TNF) on human cancer cells in vitro. Jpn J Exp Med., 51, 191-194.

Haranaka, K., Satomi, N., and Sakurai, A. (1984) Antitumor activity of murine tumor necrosis factor (TNF) against transplanted murine tumors and heterotransplanted human tumors in nude mice. Int J Cancer., 34, 263-267.

Hartwell, L.H., and Weinert, T.A. (1989) Checkpoints: Controls that ensure the order of cell cycle events. Science., 246, 629-635.

Haskard, D., Cavender, D., and Ziff, M. (1986) Phorbol ester-stimulated T lymphocytes show enhanced adhesion to human endothelial cell monolayers. J Immunol, 137, 1429- 1434.

Hass, P.E., Hotchkiss, A., Mohler, M., and Aggarwal, B.B. (1985) Characterization of specific high affinity receptors for human tumor necrosis factor on mouse fibroblasts. J Biol Chem., 260, 12214-12218.

Hassan, H.M. (1988) Biosynthesis and regulation of superoxide dismutases. Free Radie Biol Med., 5, 377-385.

Hattori, M., and Sakaki., Y. (1985) Dideoxy sequencing method using denatured plasmid templates. Anal Biochem., 152, 232-238

Hauschildt, S., Scheipers, P., Bessler, W.G., and Mulsch, A. (1992) Induction of nitric oxide synthase in L929 cells by tumour-necrosis factor alpha is prevented by inhibitors of poly(ADP-ribose) polymerase. Biochem J., 288, 255-600.

Havell, E.A., Fiers, W., and North, R.J. (1988) The anti-tumour function of tumour necrosis factor (TNF). 1. Therapeutic action of TNF against an established murine sarcoma is indirect, immunologically dependent, and limited by severe toxixity. J Exp Med., 167, 1067-1085.

Heacock, C.S., and Sutherland, R.M. (1986) Induction characteristics of oxygen regulated proteins. Int J Radiat Oncol Biol Phys., 12, 1287-1290.

Heine, H., Ulmer, A.J., Flad, H.D., and Hauschildt, S. (1995) Lipopolysaccharide- induced change of phosphorylation of two cytosolic proteins in human monocytes is prevented by inhibitors of ADP-ribosylation. J Immunol, 155, 4899-4908.

Heller, R.A., Song, K., Fan, N., and Chiang, D.J. (1992) The p70 tumor necrosis factor receptor mediates cytotoxicity. Cell, 70, 47-56.

189 Heller, R.A., Song, K., Onasch, M.A., Fischer, W.H., Chang, D., and Ringold, G.M. (1990) Complementary DNA cloning of a receptor for tumour necrosis factor and demonstration of a shed form of the receptor. Proc Natl Acad Sci USA., 87, 6151- 6155.

Helson, L., Green, S., Carswell, E., and Old, L.J. (1975) Effect of tumour necrosis factor on cultured human melanoma cells. Nature., 258, 731-732.

Henkel, T., Machleidt, T., Alkalay, I., Kronke, M., Ben Neriah, Y., and Baeuerle, P.A. (1993) Rapid proteolysis of I KB-a is necessary for activation of transcription factor NF- kB. Nature., 365, 182-185.

Hennet, T., Bertoni, G., Richter, C., and Peterhans, E. (1993a) Expression of BCL-2 protein enhances the survival of mouse fibrosarcoid cells in tumor necrosis factor- mediated cytotoxicity. Cancer Res., 53, 1456-1460.

Hennet, T., Richter., C., and Peterhans, E. (1993b) Tumour necrosis factor-alpha induces superoxide anion generation in mitochondria of L929 cells. Biochem 7.,289, 587-592.

Hensel, G., Mannel, D.N., Pfizenmaier, K., and Kronke, M. (1987) Autocrine stimulation of TNF-alpha mRNA expression in HL-60 cells. Lymphokine Res., 6, 119- 125.

Hess, G., Rossol, S., Rossol, R., and Meyer zum Buschenfelde, K.H. (1991) Tumor necrosis factor and interferon as prognostic markers in human immunodeficiency virus (HIV) infection. Infection., 19, 93-97.

Hewitt, H.B., and Wilson, C.W. (1959) The effect of tissue oxygen tension on the radiosensitivity of leukaemia cells irradiated in situ in the livers of leukaemic mice. Br J Cancer., 13, 675-684.

Higuchi, M., and Aggarwal, B.B. (1993) P80-form of the human tumor-necrosis-factor receptor is involved in DNA fragmentation. FEBS Lett., 331, 252-255.

Himeno, T., Watanabe, N., Yamauchi, N., Maeda, M., Okamoto, T., Tsuji, N., Tsuji, Y., Akiyama, S., Sasaki, H., and Niitsu, Y. (1992) Induction of synthesis of manganous superoxide dismutase in L-M9pTNF cells carrying an inducible TNF gene. Int J Cancer., 50, 458-462.

Himeno, T., Watanabe, N., Yamauchi, N., Maeda, M., Tsuji, Y., Okamoto, T., Neda, H., and Niitsu, Y. (1990) Expression of endogenous tumour necrosis factor as a protective protein against the cytotoxicity of exogenous tumour necrosis factor. Cancer Res., 50, 4941-4945.

Hirose, K., Longo, D.L., Oppenheim, J.J., and Matsushima, K. (1993) Overexpression of mitochondrial manganese superoxide-dismutase promotes the survival of tumor-cells exposed to interleukin-1, tumor-necrosis-factor, selected anticancer drugs, and ionizing- radiation. FASEB J., 7, 361-368.

Hdckel, M., Knoop, C., Schlenger, K., Vorndran, B., Baussmann, E., Mitze, M., Knapstein, P., and Vaupel, P. (1993) Intratumoral pO: predicts survival in advanced cancer of the uterine cervix. Radio Oncol, 26, 45-50.

190 Hoeck, W.G., Ramesha, C.S., Chang, D.J., Fan, N., and Heller, R.A. (1993) Cytoplasmic phospholipase A] activity and gene expression are stimulated by tumor necrosis factor: dexamethasone blocks the induced synthesis. Proc Natl Acad Sci USA., 90, 4475-4479.

Hofsli, E., and Nissen Meyer, J. (1989) Effect of erythromycin and tumour necrosis factor on the drug resistance of multidrug-resistant cells: reversal of drug resistance by erythromycin. Int J Cancer., 43, 520-525.

Hohmann, H.P., Brockhaus, M., Baeuerle, P.A., Remy, R., Kolbeck, R., and van Loon, A.P. (1990b) Expression of the types A and B tumor necrosis factor (TNF) receptors is independently regulated, and both receptors mediate activation of the transcription factor NF-kappa B. TNF alpha is not needed for induction of a biological effect via TNF receptors. J Biol Chem., 265, 22409-22417.

Hohmann, H P., Remy, R., Aigner, L., Brockhaus, M., and van Loon, A.P. (1992) Protein kinases negatively affect nuclear factor-kappa B activation by tumor necrosis factor-alpha at two different stages in promyelocytic HL60 cells. J Biol Chem., 267, 2065-2072.

Hohmann, H.P., Remy, R., Brockhaus, M., and van Loon, A.P. (1989) Two different cell types have different major receptors for human tumor necrosis factor (TNF alpha). J Biol Chem., 264, 14927-14934.

Hohmann, H.P., Remy, R., Poschl, B., and van Loon, A.P. (1990a) Tumor necrosis factors-alpha and -beta bind to the same two types of tumor necrosis factor receptors and maximally activate the transcription factor NF-kappa B at low receptor occupancy and within minutes after receptor binding. J Biol Chem., 265, 15183-15188.

Hohmann, H.P., Remy, R., Scheidereit, C., and van Loon, A.P. (1991) Maintenance of NF-kappa B activity is dependent on protein synthesis and the continuous presence of external stimuli. Mol Cell Biol, 11, 259-266.

Horsman, MR., Khalil, A.A., Siemann, D.W., Grau, C., Hill, S.A., Lynch, E.M., Chaplin, D.J., and Overgaard, J. (1994) Relationship between radiobiological hypoxia in tumors and electrode measurements of tumor oxygenation. Int J Radiat Oncol Biol Phys., 29, 439-442.

Howard, A., and Pelc, S.R. (1951) Nuclear incorporation of P as demonstrated by autoradiographs. J. Exp. Cell Res., 2, 178-188.

Hsu, H., Xiong, J. and Goeddel, D.V. (1995) The TNF receptor 1-associated protein TRADD signals cell death and NF-kB activation. Cell, 81, 495-504.

Hsu, K.C., and Chao, M.V. (1993) Differential expression and ligand binding properties of tumor necrosis factor receptor chimeric mutants. J Biol Chem., 268, 16430-16436.

Huang, P., Allam, A., Perez, L.A., Taghian, A., Freeman, J., and Suit, H.D. (1995) The effect of combining recombinant human tumor necrosis factor-alpha with local radiation on tumor control probability of a human glioblastoma multiforme xenograft in nude mice. Int J Radiat Oncol Biol Phys., 32, 93-98

191 Hughs, P.F.D., Thacker, J. D., Hogge, D., Sutherland, H.J., Thomas, T. E., Lansdorp, P. M., Eaves, C. J., and Humphries, R.K. (1992) Retroviral gene transfer into primitive normal and leukemic progenitor cells using clinically applicable procedures. J Clin Invest., 89, 1817-1824.

Ishizaka, S., Kimoto, M., and Tsujii, T. (1992) Defect in generation of LAK cell activity under oxygen-limited conditions. Immunol Lett., 32, 209-214.

Ito, H., Shirai, T., Yamamoto, S., Akira, M., Kawahara, S., Todd, C.W., and Wallace, R.B. (1986) Molecular cloning of the gene encoding rabbit tumor necrosis factor. DNA., 5, 157-165.

Jââttelâ, M. (1991) Biologic activities and mechanisms of action of tumor necrosis factor-o/ cachectin. Lab Invest., 64, 724-742.

Jââttelâ, M., and Wissing, D. (1993) Heat-shock proteins protect cells from monocyte cytotoxicity: possible mechanism of self-protection. J Exp Med., 177, 231-236.

Jââttelâ, M., Ilvesmâki, V., Voutilainen, T., Stenman, U.H., and Saksela, E. (1991) Tumor necrosis factor as a potent inhibitor of ACTH-induced cortisol production and steroidogenic P450 enzyme gene expression in cultured human fetal adrenal cells. Endocrinology., 128, 623.

Jââttelâ, M., Saksela, K. and Saksela, E. (1989) Heat shock protects WEHI 164 target cells from the cytolysis by tumor necrosis factors a and (3. Eur J Immunol, 19, 1413- 1417.

Jelinek, D.F. and Lipsky, P.E. (1987) Enhancement of human B cell proliferation and differentiation by tumor necrosis factor-a and interleukin-1. J Immunol, 139, 2970- 2976.

Jia, L., Jiang, X.-R., Razak, K., Wu, Y.-L., Newland, A.C., and Kelsey, S.M. (1995) TNF-mediated killing of human leukaemic cells: Effects of endogenous antioxidant levels and TNFa expression in leukaemic cell lines. Leuk Res., 19, 187-194.

Johnson, D.R., and Pober, J.S. (1990) Tumor necrosis factor and immune interferon synergistically increase transcription of HLA class I heavy- and light-chain genes in vascular endothelium. Proc Natl Acad Sci USA., 87, 5183-5187.

Jones, A., and Selby, P. (1989) Clinical applications of tumour necrosis factor. Prog Growth Factor Res., 1, 107-122.

Jones, E.Y., Stuart, D.I., and Walker, N.P. (1992) Crystal structure of TNF. Immunol Ser., 56, 93-127.

Jones, E.Y., Walker, N.P., and Stuart, D.I. (1991) Methodology employed for the structure determination of tumour necrosis factor, a case of high non-crystallographic symmetry. Acta Crystall, 47, 753-770.

Jones, M., and Bonting, S.L. (1956) Some relations between growth and carbohydrate metabolism in tissue culture. Exp Cell Res., 10, 631-639.

192 Kallinowski, F., Moehle, R., Schaefer, C., and Vaupel, P. (1989b) Effects of tumor necrosis factor-alpha on tumor blood flow and hyperthermic treatment. Onkologie.,12, 131-135.

Kallinowski, P., Schaefer, C., Tyler, G., and Vaupel, P. (1989a) In vivo targets of recombinant human tumour necrosis factor-alpha: blood flow, oxygen consumption and growth of isotransplanted rat tumours. Br J Cancer., 60, 555-560.

Kallinowski, P., Zander, R., Hdckel, M., and Vaupel, P. (1990) Tumor tissue oxygenation as evaluated by computerized-p02-histography. Int J Radiat Oncol Biol Phys., 19, 953-961.

Kasid, A., Director, E.P., Stovroff, M.C., Lotze, M.T., and Rosenberg, S.A. (1990) Cytokine regulation of tumor necrosis factor-alpha and -beta (lymphotoxin)-messenger RNA expression in human peripheral blood mononuclear cells. Cancer Res., 50, 5072- 5076.

Kawaguchi, T., Suzuki, K.-i., Matsuda, Y., Nishiura, T., Uda, T., Ono, M., Sekiya, C., Ishikawa, M., lino. S., Endo, Y., and Taniguchi, N. (1990) Serum-manganese- superoxide dismutase: normal values and increased levels in patients with acute myocardial infarction and several malignant diseases determined by an enzyme-linked immunosorbent assay using a monoclonal antibody. J Immunol Methods., 127, 249-254.

Kawakami, M., Watanabe, N., Ogawa, H., Kato, A., Sando, H., Yamada, N., Murase, T., Takaku, P., Shibata, S., and Oda, T. (1989) Cachectin/TNP kills or inhibits the differentiation of 3T3-L1 cells according to developmental stage. J Cell Physiol, 138, 1-7.

Ketis, N.V., and Jones, J. (1992) Expression of superoxide dismutase during endothelial cell stress. In Molecular Basis of Oxidative Damage by Leukocytes, (eds A.J. Jesaitis and E.A. Dratz) pp 311-315. CRC Press Inc.

Kim, M.Y., Linardic, C., Obeid, L., and Hannun, Y. (1991) Identification of sphingomyelin turnover as an effector mechanism for the action of tumor necrosis factor alpha and gamma-interferon. Specific role in cell differentiation. J Biol Chem., 266, 484-489.

Kimura, K., Taguchi, T., Urushizaki, I., Ohno, R., Abe, O., Purue, H., Hattori, T., Ichihashi, H., Inoguchi, K., Majima, H., Niitani, H., Ota, K., Saito, T., Suga, S., Suzuoki, Y., Wakui, A., and Yamada, K. (1987) Phase I study of recombinant human tumor necrosis factor. Cancer Chemother Pharmac., 20, 223-229.

Kirstein, M., and Baglioni, C. (1986) Tumour necrosis factor induces synthesis of two proteins in human fibroblasts. J Biol Chem., 261, 9565-9567.

Kirstein, M., Piers, W., and Baglioni, C. (1986) Growth inhibition and cytotoxicity of tumour necrosis factor in L929 cells is enhanced by high cell density and inhibition of mRNA synthesis. J Immunol, 137, 2277-2280.

Kizaki, M., Sakashita, A., Karmakar, A., Lin, C H., and Phillip Koeffler, H. (1993) Regulation of manganese superoxide dismutase and other antioxidant genes in normal

193 and leukemic hematopoietic cells and their relationship to cytotoxicity by tumour necrosis factor. Blood., 82, 1142-1150.

Klefstrom, J., Vastrik, I., Saksela, E., Valle, J., Eilers, M., and Alitalo, K. (1994) c-Myc induces cellular susceptibility to the cytotoxic action of TNF-alpha. EMBO J., 13, 5442- 5450.

Kluge, M., Eiger, B., Engel, T., Schaefer, C., Seega, J., and Vaupel, P. (1992) Acute effects of tumor necrosis factor alpha or lymphotoxin on global blood flow, laser Doppler flux, bioenergetic status of subcutaneous rodent tumors. Cancer Res., 52, 2167- 2173.

Koch, C.J., Kruuv, J., and Frey, H.E. (1973) The effect of hypoxia on the generation time of mammalian cells. Radiat Res., 53, 43-48.

Kohase, M., Henriksen DeStefano, D., May, L.T., Vilcek, J., and Sehgal, P.B. (1986) Induction of beta 2-interferon by tumor necrosis factor: a homeostatic mechanism in the control of cell proliferation. Cell, 45, 659-666.

Kohno, T., Brewer, M.T., Baker, S.L., Schwatz, P.E., King, M.W., Hale, K.K., Squires, C H., Thompson, R.C., and Vannice, J.L. (1990) A second tumour necrosis factor receptor gene product can shed a naturally occurring tumour necrosis factor inhibitor. Proc Natl Acad Sci USA., 87, 8331-8335.

Kovacs, E., Radzioch, D., Young, H.A., and Varesio, L. (1988) Differential inhibition of IL-1 and TNF-a mRNA expression by agents which block second messenger pathways in murine macrophages. J Immunol, 141, 3101-3105

Kriegler, M., Perez, C., Defay, K., Albert, I., and Lu, S.D. (1988) A novel form of TNF/ cachectin is a cell surface cytotoxic transmembrane protein: Ramification for the complex physiology of TNF. Cell, 53, 45-53.

Kronke, M., Hensel, G., Schluter, P., Schiitze, S., and Pfizenmaier, K. (1988) Tumor necrosis factor and lymphotoxin gene expression in human tumor cell lines. Cancer Res., 48, 5417-5421.

Kronke, M., Schluter, C. and Pfizenmaier, K. (1987) Tumor necrosis factor inhibits MYC expression in HL-60 cells at the level of mRNA transcription. Proc Natl Acad Sci USA., 84, 469-473.

Kronke, M., Schiitze, S., Scheurich, P., and Pfizenmaier, K. (1992) TNF signal transduction and TNF-responsive genes. Immunol Ser., 56, 189-216.

Kull Jr, F.C., and Cuatrecasas, P. (1981) Possible requirement of internalization in the mechanism of in vitro cytotoxicity in tumor necrosis serum. Cancer Res., 41, 4885- 4890.

Kunkel, S.L., Remick, D.G., Strieter, R.M., and Larrick, J.W. (1989) Mechanisms that regulate the production and effects of tumor necrosis factor-alpha. Crit Rev Immunol, 9, 93-117.

194 Kunkel, S.L., Spengler, M., May, M A., Spengier, R., Larrick, J., and Remick, D. (1988) Prostaglandin E 2 regulates macrophage-derived tumor necrosis factor gene expression. J Biol Chem., 263, 5380-5384.

Kunkel, S.L., Wiggins, R.C., Chensue, S.W., and Larrick, J. (1986) Regulation of macrophage tumor necrosis factor by prostaglandin E 2. Biochem Biophys Res Commun., 137, 404-410.

Kutteh, W.H., Rainey, W.E., and Carr, B.R. (1991b) Glucocorticoids inhibit lipopolysaccharide-induced production of tumor necrosis factor-a by human fetal Kupffer cells. J Clin Endocr Metab., 73, 296-301.

Kutteh, W.H., Rainey, W.E., Beutler, B., and Carr, B.R. (1991a) Tumor necrosis factor- a and interleukin-Ip production by human fetal Kupffer cells. Am J Obstet Gynecol, 165, 112-120.

Lahdevirta, J., Maury, C.P.J., Teppo, A.-M., and Repo, H. (1988) Elevated levels of circulating cachectin / tumor necrosis factor in patients with acquired immunodeficiency syndrome. Am J Med., 85, 289-291

Laker, C., Stocking, C., Bergholz, U., Hess, N,, de Lamarter, J. P., and Ostertag, W. (1987) Autocrine stimulation after transfer of the granulocyte / macrophage colony- stimulating factor gene and autonomous growth are distinct but independent steps in the oncogenic pathway. Proc Natl Acad Sci USA., 84, 8458-8462.

Lantz, M., Malik, S., Slevin, M L., and Olsson, I. (1990) Infusion of tumor necrosis factor (TNF) causes an increase in circulating TNF-binding protein in humans. Cytokine., 2, 402-406.

Lartigau, E., Le Ridant, A.M., Lambin, P., Weeger, P., Martin, L., Sigal, R., Lusinchi, A., Luboinski, B., Eschwege, P., and Guichard, M. (1993) Oxygenation of head and neck tumours. Cancer., 71, 2319-2325.

Lartigau, E., Lespinasse, P., Vitu, L., and Guichard, M. (1992) Does the direct measurement of oxygen tension in tumours have any adverse effects? Int J Radiat Oncol Biol Phys., 22, 949-951.

Laster, S.M., Wood, J.G., and Gooding, L.R. (1988) Tumor necrosis factor can induce both apoptotic and necrotic forms of cell lysis. J Immunol, 141, 2629-2634.

Lattime, E., and Stutman, O. (1989) Tumor growth in vivo selects for resistance to tumor necrosis factor. J Immunol, 143, 4317-4323.

Le, J.M., Weinstein, D., Gubler, U., and Vilcek, J. (1987) Induction of membrane- associated interleukin 1 by tumor necrosis factor in human fibroblasts. J Immunol, 138, 2137-2142.

Lee, I.e., Truneh, A., Smith Jr, M.P., and Tsang, K.Y. (1987b) Induction of interleukin 2 receptor (TAG) by tumor necrosis factor in YT cells. J Immunol, 139, 1935-1938.

195 Lee, M.D., Zentella, A., Vine, W., Pekala, P.H., and Cerami, A. (1987a) Effect of endotoxin-induced monokines on glucose metabolism in the muscle cell line L6. Proc Natl Acad Sci USA., 84, 2590-2594.

Lee, T.H., Lee, G.W., Ziff, E.B., and Vilcek, J. (1990) Isolation and characterization of eight tumor necrosis factor-induced gene sequences from human fibroblasts. Mol Cell Biol., 10, 1982-1982.

Leeuwenberg, J.F.M., van Tits, L.J.H., Jeunhomme, T.M.A.A., and Buurman, W.A. (1995) Evidence for exclusive role in signalling of tumour necrosis factor p55 receptor and a potentiating function of p75 receptor on human endothelial cells. Cytokine., 7, 457-462.

Leitman, D., Ribeiro, R., Mackow, E., Baxter, J., and West, B. (1991) Identification of a tumor necrosis factor-responsive element in the tumor necrosis factor a gene. J Biol Chem., 266, 9343-9346.

Lejeune, E.J. (1995) High dose recombinant tumour necrosis factor (rTNFa) administered by isolation perfusion for advanced tumours of the limbs: A model for biochemotherapy of cancer. Eur J Cancer., 31A, 1009-1016.

Lenardo, M.J., and Baltimore, D. (1989) NF-kB a pleiotropic mediator of inducible and tissue-specific gene control. Cell, 58, 227-229,

Lenk, H., Tannenberger, S., Müller, U., Ebert, J., and Shiga, T. (1989) Phase 11 clinical trial of high-dose recombinant human tumor necrosis factor. Cancer Chemother Pharmac., 24, 391-392.

Lesslauer, W., Tabuchi, H., Gentz, R., Brockhaus, M., Schlaeger, E.J., Grau, G., Piguet, P.F., Pointaire, P., Vassalli, P., and Loetscher, H. (1991) Recombinant soluble tumor necrosis factor receptor proteins protect mice from lipopolysaccharide-induced lethality. Eur J Immunol, 21, 2883-2886.

Lewis, G.D., Aggarwal, B.B., Eessalu, I.E., Sugarman, B.J., and Shepard, H.M. (1987) Modulation of the growth of transformed cells by human tumor necrosis factor-alpha and interferon-gamma. Cancer Res., 47, 5382-5382.

Lewis, M., Tartaglia, L.A., Lee, A., Bennett, G.L., Rice, G.C., Wong, G.H., Chen, E.Y., and Goeddel, D.V. (1991) Cloning and expression of cDNAs for two distinct murine tumor necrosis factor receptors demonstrate one receptor is species specific. Proc Natl Acad Sci USA., 88, 2830-2834.

Lewis-Molock, Y., Suzuki, K., Taniguchi, N., Nguyen, D-dH., Mason, R.J., and White, C.W., (1994) Lung manganese superoxide dismutase increases during cytokine- mediated protection against pulmonary oxygen toxicity in rats. Am J Respir Cell Mol Biol, 10, 133-141.

Lieberman, A.P., Pitha, P.M., and Shin, M.L. (1992) Poly(A) removal is the kinase- regulated step in tumor necrosis factor mRNA decay. J Biol Chem., 267, 2123-2126.

Liénard, D., Ewalenko, P., Delmotte, J.J., Renard, N., and Lejeune, F. (1992b) High- dose recombinant tumor necrosis factor alpha in combination with interferon gamma

196 and Melphalan in isolation perfusion of the limbs for melanoma and sarcoma. J Clin Oncol., 10, 52-60.

Liénard, D., Lejeune, F.J., and Ewalenko, P. (1992a) In transit métastasés of malignant melanoma treated by high dose rTNF alpha in combination with interferon-gamma and melphalan in isolation perfusion. World J Surg., 16, 234-240.

Lin, J.X., and Vilcek, J. (1987) Tumor necrosis factor and interleukin-1 cause a rapid and transient stimulation of c-fos and c-myc mRNA levels in human fibroblasts. J Biol Chem., 262, 11908-11911.

Littbrand, B., and Revesz, L. (1968) Survival of cells in anoxia. Br J Radiol, 41, 479- 480.

Loeffler, D., Juneau, P., and Masserant, S. (1992) Influence of tumour physico-chemical conditions on interleukin-2-stimulated lymphocyte proliferation. Br J Cancer., 66, 619- 622.

Loeffler, M. (1985) Towards a further understanding of the growth-inhibiting action of oxygen deficiency. Exp Cell Res., 157, 195-206.

Loeffler, M. (1987) Restimulation of cell cycle progression by hypoxic tumour cells with deoxynucleosides requires ppm oxygen tension. Exp Cell Res., 169, 255-261.

Loetscher, H., Pan, Y.C., Lahm, H.W., Gentz, R., Brockhaus, M., Tabuchi, H., and Lesslauer, W. (1990a) Molecular cloning and expression of the human 55 kd tumor necrosis factor receptor. Cell, 61, 351-359.

Loetscher, H., Schlaeger, E.J., Lahm, H.W., Pan, Y.C., Lesslauer, W., and Brockhaus, M. (1990b) Purification and partial amino acid sequence analysis of two distinct tumor necrosis factor receptors from HL60 cells. J Biol Chem., 265, 20131-20138.

Lowenthal, J.W., Ballard, D.W., Bogerd, H., Bohnlein, E., and Greene, W.C. (1989) Tumour necrosis factor-activation of the IL-2 receptor-gene involves the induction of K- B specific DNA binding proteins. J Immunol, 142, 3121-3128.

Luettig, B., Decker, T., and Lohmann, M.M. (1989) Evidence for the existence of two forms of membrane tumor necrosis factor: an integral protein and a molecule attached to its receptor. J Immunol, 143, 4034-4048.

MacDonald, T.T., Hutchings, P., Choy, M. Y., Murch, S., and Cooke, A. (1990) Tumour necrosis factor-alpha and interferon-gamma production measured at the single cell level in normal and inflamed human intestine. Clin Exp Immunol, 81, 301-305.

MacPherson, G.G., and North, R.J. (1986) Endotoxin-mediated necrosis and regression of established tumours in the mouse. A correlative study of quantitative changes in blood flow and ultrastructural morphology. Cancer Immunol Immunother., 21, 209-216.

Madhevan, V., Malik, S.T.A., Maeger, T., Fiers, W., and Hart, I. (1990) Role of tumor necrosis factor in flavone acetic acid-induced tumor vascular shutdown. Cancer Res., 50, 5537-5542.

197 Malik, S.T.A., and Balkwill, F.R. (1992) Antiproliferative and antitumor activity of TNF in vitro and in vivo. Immun Ser., 56, 239-268.

Manda, T., Shimomura, K., Mukumoto, S., Kobayashi, K., Mizota, T., Hirai, O., Matsumoto, S., Oku, T., Nishigaki, F., Mori, J. and Kikuchi, H. (1987) Recombinant human tumor necrosis factor-alpha: evidence of an indirect mode of antitumor activity. Cancer Res., 47, 3707-3711.

Mandrup Poulsen, T., Bendtzen, K., Dinarello, C.A., and Nerup, J. (1987) Human tumor necrosis factor potentiates human interleukin 1-mediated rat pancreatic beta-cell cytotoxicity. J Immunol, 139, 4077-4082.

Marino, M.W., Pfeffer, L.M., Guidon, P.T., and Donner, D.B. (1989) Tumor necrosis factor induces phosphorylation of a 28 kDa mRNA cap-binding protein in human cervical carcinoma cells. Proc Natl Acad Sci, USA., 86, 8147-8421.

Marklund, S.L. (1990) Expression of extracellular superoxide dismutase by human cell lines. Biochem J., 266, 213-219.

Markowitz, D., Goff, S., and Bank, A. (1988) A safe packaging line for gene transfer: separating viral genes on two different plasmids. J Virol, 62, 1120-1124.

Marmenout, A., Fransen, L., Tavernier, J., van der Hayden, J., Tijard, R., Kawashima, E., Shaw, A., Johnson, M.J., Semon, D., Muller, R., Ruysschaert, M R., vanVliet, A., and Fiers, W. (1985) Molecular cloning and expression of human tumor necrosis factor and comparison with mouse tumor necrosis factor. Eur J Biochem., 152, 512-519.

Marsden, P. A., and Brenner, B.M. (1992) Transcriptional regulation of the endothelin-1 gene by TNF-alpha. Am J Physiol, 262, C854-C861.

Martinet, Y., Yamauchi, K., and Crystal, R.G. (1988) Differential expression of the tumor necrosis factor/cachectin gene by blood and lung mononuclear phagocytes. Am Rev Respir Dis., 138, 659-665.

Matthews, N. (1978) Tumour-necrosis factor from the rabbit. II. Production by monocytes. Br J Cancer., 38, 310-315.

Matthews, N. (1981) Production of an anti-tumour cytotoxin by human monocytes. Immunology., 44, 135-142.

Matthews, N., and Neale, M.L. (1989) Relationship between tumour cell morphology, gap junctions and susceptibility to cytolysis by tumour necrosis factor. Br J Cancer., 59, 189-193.

Matthews, N., Neale, M.L., Jackson, S.K., and Stark, J.M., (1987) Tumour cell killing by tumour necrosis factor: inhibition by anaerobic conditions, free-radical scavengers and inhibitors of arachidonate metabolism. Immunology., 62, 153-155.

McIntosh, J.K., Mule, J.J., Merino, M.J., and Rosenberg, S.A. (1988) Synergistic antitumor effects of immunotherapy with recombinant interleukin-2 and recombinant tumor necrosis factor-alpha. Cancer Res., 48, 4011-4017.

198 Meager, A. (1988) Biological standardization of cytokines. Dev Biol Stand., 69, 199- 206.

Meager, A., Leung, H., and Woolley (1989) Assays for tumour necrosis factor and related cytokines. 7/mmMAzo/116, 1-17.

Mehlen, P., Preville, X., Chareyron, P., Briolay, J., Klemenz, R., and Arrigo, A.-P. (1995) Constitutive expression of human hsp27, drosophila hsp27, or human ocB- crystallin confers resistance to TNF-a and oxidative stress-induced cytotoxicity in stably transfected murine L929 fibroblasts. J Immunol, 154, 363-374.

Meichle, A., Schiitze, S., Hensel, G., Brunsing, D., and Kronke, M. (1990) Protein kinase C-independent activation of nuclear factor kappa B by tumor necrosis factor. J Biol Chem., 265, 8339-8343.

Merz, R., and Schneider, F. (1983) Growth characteristics of anaerobically treated early and late S-period of Ehrlich ascites tumour cells after reaeration. Z. Naturforsch C., 38, 313-318.

Meyer, R., Hatada, E.N., Hohmann, H P., Haiker, M., Bartsch, C., Rothlisberger, U., Lahm, H.W., Schlaeger, E.J., van Loon, A.P., and Scheidereit, C. (1991) Cloning of the DNA-binding subunit of human nuclear factor kappa B: the level of its mRNA is strongly regulated by phorbol ester or tumor necrosis factor alpha. Proc Natl Acad Sci USA., 88, 966-70.

Michie, HR., Manogue, K.R., Spriggs, DR., Revhaug, A., O'Dwyer, S., Dinarello, C.A., Cerami, A., Wolff, S.M., and Wilmore, D.W. (1988) Detection of circulating tumor necrosis factor after endotoxin administration. N EnglJ Med., 318, 1481-1486.

Milenkovic, L., Rettori, V., Snyder, G.D., Beutler, B., and McCann, S.M. (1989) Cachectin alters anterior pituitary hormone release by a direct action in vitro. Proc Natl Acad Sci USA., 86, 2418-2422.

Ming, W.J., Bersani, L., and Mantovani, A. (1987) Tumor necrosis factor is chemotactic for monocytes and polymorphonuclear leukocytes. J Immunol, 138, 1469-1474.

Mohler, K.M., Sleath, P R., Fitzner, J.N., Cerretti, D.P., Aldersorl, M., Kerwar, S.S., Torrance, D.S., Otten Evans, C., Greenstreet, T., and Weerawama, K. (1994) Protection against a lethal dose of endotoxin by an inhibitor of tumour necrosis factor processing. Nature., 370, 218-220.

Mori, H., Nakagawa, M., Itoh, N., Wada, K., and Tamaya, T. (1990) Danazol suppresses the production of interleukin-1 beta and tumor necrosis factor by human monocytes. Am J Reprod Immunol, 24, 45-50.

Mosialos, G., Birkenbach, M., Yalamanchili, R., VanArsdale, T., Ware, C., and Kieff, E. (1995) The Epstein-Barr virus transforming protein LMPl engages signaling proteins for the tumor necrosis factor receptor family. Cell, 80, 389-399.

Mosselmans, R., Hepburn, A., Dumont, J.E., Fiers, W., and G aland, P. (1988) Endocytic pathway of recombinant murine tumor necrosis factor in L-929 cells. J Immunol, 141, 3096-3100.

199 Mottram, J.C. (1936) A factor of importance in the radiosensitivity of tumors. Br J Radiol., 9, 606-614.

Moulder, J., and Rockwell, S. (1984) Hypoxic fractions of solid tumours: Experimental techniques, methods of analysis, and a survey of existing data. J Radiat Oncol Biol Phys., 10, 695-712.

Moulder, I.E., and Rockwell, S. (1987) Tumor hypoxia: its impact on cancer therapy. Cancer Metastasis Rev., 5, 313-341.

Mueller-Klieser, W., Vaupel, P., Manz, R., and Grünewald, W.A. (1980) Intracapillary oxyhemoglobin saturation in malignant tumors with central or peripheral blood supply. Eur J Cancer., 16, 195-207.

Mule, J.J., Asher, A., McIntosh, J., Lafreniere, R., Shiloni, E., Lefor, A., Reichert, C M., and Rosenberg, S.A. (1988) Antitumor effect of recombinant tumor necrosis factor- alpha against murine sarcomas at visceral sites: tumor size influences the response to therapy. Cancer Immunol Immunother., 26, 202-208.

Muller, U., Jongeneel, C.V., Nedospasov, S.A., Lindahl, K.F., and Steinmetz, M. (1987) Tumour necrosis factor and lymphotoxin genes map close to H-2D in the mouse major histocompatibility complex. Nature., 325, 265-267.

Munker, R., and Koeffler, P. (1987a) In vitro action of tumor necrosis factor on myeloid leukemia cells. Blood., 69, 1102-1108.

Munker, R., Gasson, J., Ogawa, M., and Koeffler, H.P. (1986) Recombinant human TNF induces production of granulocyte-monocyte colony-stimulating factor. Nature (Lond.)., 323, 79-82.

Murray, A.W. (1992) Creative blocks: cell cycle checkpoints and feedback controls. Nature., 359, 599-604.

Mutch, D.G., Powell, C.B., Kao, M.-S., and Collins, J.L. (1992) Resistance to cytolysis by tumor necrosis factor a in malignant gynecological cell lines is associated with the expression of protein(s) that prevent the activation of phospholipase A? by tumor necrosis factor a. Cancer Res., 52, 866-872.

Nakano, K., Okugawa, K., Furuichi, H., Matsui, Y., and Sohmura, Y. (1989) Augmentation of the generation of cytotoxic T lymphocytes against syngeneic tumor cells by recombinant human tumor necrosis factor. Cell Immunol, 120, 154-164.

Nakano, Y., Shirai, M., Mori, N., and Nakano, M. (1991) Neutralization of microcystin shock in mice by tumor necrosis factor alpha antiserum. Appl Environ Microbiol, 57, 327-330.

Naredi, P.L., Lindner, P.G., Holmberg, S.B., Stenram, U., Peterson, A., and Hafstrom, L.R. (1993) The effects of tumour necrosis factor alpha on the vascular bed and blood flow in an experimental rat hepatoma. Int J Cancer., 54, 645-649.

200 Natarajan, R., Ploszaj, S., Horton, R., and Nadler, J. (1989) Tumor necrosis factor and interleukin-1 are potent inhibitors of angiotensin-II-induced aldosterone synthesis. Endocrinology., 125, 3084-3089.

Naume, B., Shalaby, R., Lesslauer, W. (1991) Involvement of the 55- and 75- kDa tumor necrosis factor receptors in the generation of lymphokine-activated killer cell activity and proliferation of natural killer cells. J Immunol, 146, 3045-3048.

Naworth, P.D., Handley, D., Matseueda, G., Dewaal, R., Gerlach, H., Blohm, D., and Stern D. (1988) Tumour necrosis factor cachectin-induced intravascular fibrin formation in meth A fibrosarcomas. J Exp Med., 168, 637-642.

Nawroth, P.P., and Stem, D.M. (1986) Modulation of endothelial cell hemostatic properties by tumor necrosis factor. J Exp Med., 163, 740-745.

Nawroth, P.P., Bank, I., Handley, D., Cassimeris, J., Chess, L., and Stern, D.M. (1986) Tumor necrosis factor/cachectin interacts with endothelial cell receptors to induce release of interleukin-1. J Exp Med., 161, 1363-1375.

Neale, M L., Fiera, R.A., and Matthews, N. (1988) Involvement of phospholipase A 2 activation in tumor cell killing by tumor necrosis factor. Immunology., 64, 81-85.

Nedwin, G.E., Naylor, S.L., Sakaguchi, A.Y., Smith, D., Jarrett-Nedwin, J., Pennica, D., Goeddel, D.V., and Gray, P.W. (1985) Human lymphotoxin and tumour necrosis factor genes: Structure, homology and chromosomal location. Nucleic Acids Res., 13, 6361- 6373.

Neta, R., Sayers, T.J., and Oppenheim, J.J. (1992) Relationship of TNF to interleukins. In Tumor Necrosis Factors. Structure, Function, and Mechanism of Action, (eds B.B. Aggarwal and J. Vilcek) pp. 499-566. Marcel Dekker, New York.

Neuner, P., Klosner, G., Pourmojib, M., Trautinger, F., and Knobler, R. (1994) Selective release of tumor necrosis factor binding protein II by malignant human epidermal cells reveals protection from tumor necrosis factor alpha-mediated cytotoxicity. Cancer Res., 54, 6001-6005.

Niitsu, Y., Watanabe, N., Neda, H., Yamauchi, N., Maeda, M., Sone, H., and Kuriyama, H. (1988) Induction of synthesis of tumour necrosis factor in human and murine cell lines by exogenous recombinant human tumor necrosis factor. Cancer Res., 48, 5407- 5410.

Nio, Y., Zighelboim, J., Berek, J., and Bonavida, B. (1990) Cycloheximide-induced modulation of TNF-mediated cytotoxicity in sensitive and resistant ovarian tumor cells. Cancer Chemother Pharmac., 26, 1-8.

Nophar, Y., Holtmann, H., Ber, R., and Wallach, D. (1988) Dominance of resistance to the cytocidai effect of tumor necrosis factor in heterokaryons formed by fusion of resistant and sensitive cells. J Immunol, 140, 3456-3460.

North, R.J., and Havell, E.A. (1988) The antitumor function of tumor necrosis factor (TNF) II. Analysis of the role of endogenous TNF in endotoxin-induced hemorrhagic necrosis and regression of an established sarcoma. J Exp Med., 167, 1086-1099.

201 O'Connell, K., and Edidin, M. (1990) A mouse lymphoid endothelial cell line immortalized by Simian virus 40 binds lymphocytes and retains functional characteristics of normal endothelial cells. J Immunol., 144, 521-525.

O'Malley, W.E., Achinstein, B., and Shear, M.J. (1962) Action of bacterial polysaccharide on tumors. II. Damage of sarcoma 37 by serum of mice treated with Serratia marcescens polysaccharide and induced tolerance. J Natl Cancer Inst., 29, 1169-1175.

Oberley, L.W., St. Clair, D.K., Autor, A.P., and Oberley, T.D. (1987) Increase in manganese superoxide dismutase activity in the mouse heart after X-irradiation. Arch Biochem Biophys., 254, 69-80.

Ohsawa, P., and Natori, S. (1988) Selective degradation of tumour necrosis factor in sensitive cells and production of membrane-active substance. J Biochem., 103, 730-734.

Okamoto, T., Watanabe, N., Yamauchi, N., Tsuji, Y., Tsuji, N., Itoh, Y., Neda, H., and Niitsu, Y. (1992) Endogenous tumor necrosis factor exerts its protective function intracellularly against the cytotoxicity of exogenous tumor necrosis factor. Cancer Res., 52, 5278-5281.

Oku, N., Araki, R., Araki, H., Shibamoto, S., Ito, P., Nishihara, T., and Tsujimoto, M. (1987) Tumor necrosis factor-induced permeability increase of negatively charged phospholipid vesicles. J Biochem Tokyo., 102, 1303-1310.

Old, L.J. (1985) Tumour Necrosis Pactor. Scientific American., 258, 59-75.

Oliff, A., Defeo-Jones, D., Boyer, M., Martinez, K., Kiefler, D., Vuocolo, 0., Wolfe, A., and Socher, S.H. (1987) Tumors secreting human TNP/cachectin induce cachexia in mice. Cell, 50, 555-563.

Olsson, I., Gatanga, T., Gullberg, U., Lantz, M., and Granger, G.A. (1993) Tumour necrosis factor (TNP) binding proteins (soluble TNP receptor forms) with possible roles in inflammation and malignancy. Eur Cytokine Netw., 4, 169-180.

Olsson, I., Lantz, M., Nilsson, E., Peetre, C., Thysell, H., Grubb, A., and Adolf, G. (1989) Isolation and characterisation of a tumour necrosis factor binding protein from urine. Eur J Haematol, ^2,210-215.

Opipari Jr, A.W., Boguski, M.S., and Dixit, V.M. (1990) The A20 cDNA induced by tumor necrosis factor-alpha encodes a novel type of zinc finger protein. J Biol Chem., 265, 14705-14708.

Opipari Jr, A.W., Hu, H.M., Yabkowitz, R., and Dixit, V.M. (1992) The A20 zinc finger protein protects cells from tumor necrosis factor cytotoxicity. J Biol Chem., 267, 12424- 12427.

Ortaldo, J R., Mason, L.H., Mathieson, B.J., Liang, S.M., Plick, D.A., and Herberman, R.B. (1986) Mediation of mouse natural cytotoxic activity by tumour necrosis factor. Nature., 321, 700-702.

202 Osborn, L., Hession, C., Tizard, R., Vassallo, C., Luhowskyj, S., Chi Rosso, G., and Lobb, R. (1989a) Direct expression cloning of vascular cell adhesion molecule 1, a cytokine-induced endothelial protein that binds to lymphocytes. Cell., 59, 1203-1211.

Osborn, L., Kunkel, S., and Nabel, G.J. (1989b) Tumor necrosis factor alpha and interleukin 1 stimulate the human immunodeficiency virus enhancer by activation of the nuclear factor kappa B. Proc Natl Acad Sci USA., 86, 2336-2340.

0stensen, M.E., Thiele, D.L., and Lipsky, P.E. (1987) Tumor necrosis factor-a enhances cytolytic activity of human natural killer cells. J Immunol, 138, 4185-4191.

Oster, W., Lindemann, A., Horn, S., Mertelsmann, R., and Herrmann, F. (1987) Tumor necrosis factor (TNF)-alpha but not TNF-beta induces secretion of colony stimulating factor for macrophages (CSF-1) by human monocytes. Blood., 70, 1700-1703.

Pace, D.M., Thompson, J R., and van Camp, W.A. (1962) Effects of oxygen on growth in several established cell lines. J Natl Cancer Inst., 28, 897-909.

Palladino, M.A., Patton, J.S., Figari, I.S., and Shalaby, M R. (1987a). Possible relationships between in vivo anti-tumour activity and toxicity of tumour necrosis factor- a. In Tumour Necrosis Factor and Related Cytotoxins, pp 21-38. Ciba foundation symposium 131, Wiley Chichester.

Palladino, M.A., Shalaby, MR., Kramer, S.M., Ferraiolo, B.L., Braughman, R.A., Deleo, A.B., Liggitt, D., and Patton, J.S. (1987b) Characterisation of the anti-tumour activities of human tumour necrosis factor-alpha and the comparison with other cytokines: Induction of tumour specific immunity. J Immunol, 138, 4023-4032.

Palombella, F.C., and Vilcek, J. (1989) Mitogenic and cytotoxic actions of tumour necrosis factor in Balb/C 3T3 cells. Role of phospholipase activation. J Biol Chem., 264,18128-18133.

Palombella, V.J., Yamashiro, D.J., Maxfield, F.R., Decker, S.J., and Vilcek, J. (1987) Tumor necrosis factor increases the number of epidermal growth factor receptors on human fibroblasts. J Biol Chem., 262, 1950-1954.

Pandita, R., Pocsik, E., and Aggarwal, B.B. (1992) Interferon-gamma induces cell surface expression for both types of tumor necrosis factor receptors. FEBS Lett., 312, 87-90.

Pardee, A.B., Dubrow, R., Hamlin, J.L., and Kletzien, R.F. (1978) Animal cell cycle. Ann Rev Biochem., 47, 715-750.

Park, Y.M., Anderson, R.L., Spitz, DR., and Hahn, G.M. (1992) Hypoxia and resistance to hydrogen peroxide confer resistance to tumor necrosis factor in murine L929 cells. Radiat Res., 131, 162-168.

Parkins, C.S., Hill, S.A., Lonergan, S.J., Horsman, M R., Chadwick, J.A., and Chaplin, D.J. (1994) Ischaemia induced cell death in tumors: importance of temperature and pH. Int J Radiat Oncol Biol Phys., 29, 499-503.

2 0 3 Patek, P., Lin, Y., and Case, P. (1989) Cell lines cultured at high density are resistant to lysis by tumour necrosis factor and natural cytotoxic cells (42854). Soc Exp Biol Med., 190,234-239.

Patton, J.S., Peters, P.M., McCabe, J., Crase, D., Hansen, S., Chen, A.B., and Liggitt, D. (1987) Development of partial tolerance to the gastrointestinal effects of high doses of recombinant tumor necrosis factor-alpha in rodents. J Clin Invest., 80, 1587-1596.

Pauli, U. (1994) Control of tumor necrosis factor gene expression. Crit Rev Euk Gene Exp., 4, 323-344.

Peck, R., Brockhaus, M., and Frey, J R. (1989) Cell surface tumor necrosis factor (TNF) accounts for monocyte- and lymphocyte-mediated killing of TNF-resistant target cells. Celllmmunol., 122, 1-10.

Pennica, D., Hayflick, J.S., Bringman, T.S., Palladino, M.A., and Goeddel, D.V. (1985) Cloning and expression in Escherichia coli of the cDNA for murine tumor necrosis factor. Proc Natl Acad Sci USA., 82, 6060-6064.

Pennica, D., Kohr, W.J., Fendly, B.M., Shire, S.J., Raab, H.E., Borchardt, P.E., Lewis, M., and Goeddel, D.V. (1992) Characterization of a recombinant extracellular domain of the type 1 tumor necrosis factor receptor: evidence for tumor necrosis factor-alpha induced receptor aggregation. Biochemistry., 31, 1134-1141.

Pennica, D., Nedwin, G., Hayflick, J., Seeburg, P., Derynck, R., Palladino, M.A., Kohr, W.J., Aggarwal, B.B., and Goeddel, D.V. (1984) Human tumour necrosis factor: precursor, structure, expression and homology to lymphotoxin. Nature., 312, 724-729.

Perlmutter, D.H., Dinarello, C.A., Punsal, P.I., and Colten, H R. (1986) Cachectin/tumor necrosis factor regulates hepatic acute-phase gene expression. J Clin Invest., 78, 1349- 1354.

Pessara, U., and Koch, N. (1990) Tumor necrosis factor alpha regulates expression of the major histocompatibility complex class Il-associated invariant chain by binding of an NF-kappa B-like factor to a promoter element. Mol Cell Biol, 10, 4146-4154.

Peters, P.M., Ortaldo, J R., Shalaby, M.R., Svedersky, L.P., Nedwin, G.E., Bringman, T.S., Hass, P.E., Aggarwal, B.B., Herberman, R.B., Goeddel, D.V., and Palladino, M.A.J. (1986) Natural killer-sensitive targets stimulate production of TNF-alpha but not TNF-beta (lymphotoxin) by highly purified human peripheral blood large granular lymphocytes. J Immun., 137, 2592-2598.

Pettersen, E.G., and Lindmo, T. (1981) Low concentrations of misonidazole counteract effects of extreme hypoxia on cells in S. Br J Cancer., 43, 355-366.

Pfizenmaier, K., Scheurich, P., Sluter, C. and Kronke, M. (1987) Tumor necrosis factor enhances HLA-A,B,C and HLA-DR gene expression in human tumor cells. J Immunol, 138, 975-980.

Philip, R., and Epstein, L.B. (1986) Tumour necrosis factor as immunomodulator and mediator of monocyte cytotoxicity induced by itself, gamma-interferon and interleukin- 1. Nature., 323, 86-89.

204 Pillai, S., Bikle, D.D., Eessalu, T.E., Aggarwal, B.B., and Elias, P.M. (1989) Binding and biological effects of tumor necrosis factor alpha on cultured human neonatal foreskin kératinocytes. J Clin Invest., 83, 816-821.

Pimm, M., Gribben, S., and Morris, T. (1991) Influence of recombinant tumour necrosis factor alpha on blood flow and antibody localisation in human tumour xenografts in nude mice. J Cancer Res Clin Oncol, 117, 543-548.

Pisa, P., Gennene, M., Soder, O., Ottenhoff, T., Hansson, M., Kiessling, R., Sher, T., Yamin, A., Matzliach, M., Rottem, S., and Gallily, R. (1990) Serum tumor necrosis factor levels and disease dissemination in leprosy and leishmaniasis. Partial biochemical characterization of spiroplasma membrane component inducing tumor necrosis factor alpha. J Infect Dis., 161, 988-991.

Pober, J., Lapierre, L., Stolpen, A., Brock, T., Springer, T., Fiers, W., Bevilacqua, M., Mendrick, D., and Gimbrone, M. (1987) Activation of cultured human endothelial cells by recombinant lymphotoxin: Comparison with tumor necrosis factor and interleukin-1 species. J Immunol, 138, 3319-3324.

Pober, J.S., and Cotran, R.S. (1990) Cytokines and endothelial cell biology. Physiological Rev., 70, 427-451.

Pober, J.S., Bevilacqua, M.P., Mendrick, D.L., Lapierre, L.A., Fiers, W., and Gimbrone Jr, M.A. (1986) Two distinct monokines, interleukin 1 and tumor necrosis factor, each independently induce biosynthesis and transient expression of the same antigen on the surface of cultured human vascular endothelial cells. J Immunol, 136, 1680-1687.

Pogrebniak, H.W., Prewitt, T.W., Matthews, W.A., and Pass, HI. (1991) Tumor necrosis factor-alpha alters response of lung cancer cells to oxidative stress. J Thorac Cardiovasc Surg., 102, 904-907.

Porteu, F., and Hieblot, C. (1994) Tumor necrosis factor induces a selective shedding of its p75 receptor from human neutrophils. J Biol Chem., 269, 2834-2840.

Proietti, E., Belardelli, F., Carpinelli, G. Di Vito, M., Woodrow, D., Mors, J., Sestilli, P., Fiers, W., Gresser, I., and Podo, F. (1988) Tumor necrosis factor a induces early morphologic and metabolic alterations in Friend leukemia cell tumors and fibrosarcomas in mice. Int J Cancer., 42, 582-591.

Pusztai, L., Lewis, C.F. and McGee, J.O. (1993) Growth arrest of the breast cancer cell line, T47D, by TNF alpha; cell cycle specificity and signal transduction. Br J Cancer., 67, 290-296.

Rabinovitch, A., Baquerizo, H., and Sumoski, W. (1990) Cytotoxic effects of cytokines on islet beta-cells: evidence for involvement of eicosanoids. Endocrinology., 126, 67- 71.

205 Ralph, P., Nakoinz, L, Sampson Johannes, A., Fong, S., Lowe, D., Min, H.Y., and Lin, L. (1992) IL-10, T lymphocyte inhibitor of human blood cell production of fL-1 and tumor necrosis factor. J Immunol, 148, 808-814.

Ramkumar, V., Wilson, M., Dhanraj, D.N., Gettys, T.W., and Ali, H. (1995) Dexamethasone up-regulates A3 adenosine receptors in rat basophilic leukemia (RBL- 2H3) cells. J Immunol, 154, 5436-5443.

Rath U., Kaufmann, M., Schmid, H., Hofmann, J., Wiedenmann, B., Kist, A., Kempeni, J., Schlick, E., Bastert, G., Kommerell, B., and Mannel, D. (1991) Effect of intraperitoneal recombinant human tumour necrosis factor alpha on malignant ascites. Eur J Cancer., 27, 121-125.

Reich, E., Franklin, R.M., Shatkin, A.J., and Tatum, E.L. (1962) Action of Actinomycin D on animal cells and viruses. Proc Natl Acad Sci (Wash.)., 48, 1238.

Reid, T., Ramesha, C.S., and Ringold, G.M. (1991) Resistance to killing by tumor necrosis factor in an adipocyte cell line caused by a defect in arachidonic acid biosynthesis. J Biol Chem., 266, 16580-16586.

Reid, T.R., Torti, F.M., and Ringold, G.M. (1989) Evidence for two mechanisms by which tumor necrosis factor kills cells. J Biol Chem., 264, 4583-4589.

Reinhold, H.S., Blachiwiecz, B., and Blok, A. (1977) Oxygenation and reoxygenation in 'sandwich' tumours. BiblAnat., 15, 270-272.

Reis, L.F.L., Lee, T.H., and Vilcek, J. (1989) Tumor necrosis factor acts synergistically with autocrine interferon-[3 and increases interferon-p mRNA levels in human fibroblasts. J Biol Chem., 264, 16351-16354.

Remick, D.G., and Kunkel, S.L. (1993) Pathophysiologic alterations induced by tumor- necrosis-factor. Int Rev Expl Pathol, 34, 7-25.

Remick, D.G., Nguyen, D.T., Eskandari, M.K., Strieter, R.M., and Kunkel, S.L. (1989) Cyclosporine A inhibits TNF production without decreasing TNF mRNA levels. Biochem Biophys Res Commun., 161, 551-555.

Richter, J., Gullberg, U., and Lantz M (1995) TNF-induced superoxide anion production in adherent human neutrophils involves both the p55 and p75 TNF receptor. J Immunol, 154,4142-4149.

Rockwell, S., and Moulder, J.E. (1990) Hypoxic fractions of human tumors xenografted into mice: A review. Int J Rad Oncol Biol Phys., 19, 197-202.

Roizin Towle, L., and Hall, E.J. (1978) Studies with bleomycin and misonidazole on aerated and hypoxic cells. Br J Cancer., 37, 254-260.

Rosenblum, M.G., and Donato, N.J. (1989) Tumor necrosis factor alpha: a multifaceted peptide hormone. Crit Rev Immun., 9, 21-44.

206 Rossow, P.W., Riddle, V.G., and Pardee, A.B. (1979) Synthesis of labile, serum- dependent protein in early G1 controls animal cell growth. Proc Natl Acad Sci USA., 76, 4446-4450.

Rothe, J., Bluethmann, H., Gentz, R., Lesslauer, W., and Steinmetz, M. (1993) Genomic organization and promoter function of the murine tumor necrosis factor receptor beta gene. Mollmmunoi, 30, 165-175.

Rothstein, J.L., and Schreiber, H. (1992). Tissue damage caused by TNF and complement. In Tumor Necrosis Factors: Structure, Function, and Mechanism of Action, (eds B.B. Aggarwal and J. Vilcek) pp 453-481. Marcel Dekker, New York.

Rouzer, C.A., and Cerami, A. (1980) Hypertriglyceridemia associated with Trypanosoma brucei brucei infection in rabbits: role of defective triglyceride removal. Mol Biochem ParasitoL, 2, 31-38.

Rubin, B Y., Anderson, S.L., Lunn, R.M., Richardson, N.K., Hellermann, G.R., Smith, L.J., and Old, L.J. (1988) Tumor necrosis factor and IFN induce a common set of proteins. J Immunol., 141, 1180-1184.

Rubin, B.Y., Anderson, S.L., Sullivan, S.A., Williamson, B.D., Carswell, E.A., and Old, L.J. (1986) Nonhematopoietic cells selected for resistance to tumour necrosis factor produce tumour necrosis factor. N J Exp Med., 164, 1350-1355.

Ruff, M R., and Gifford, G.E. (1981) Rabbit tumor necrosis factor: mechanism of action. Infect Immun., 31, 380-385.

Ruggiero, V., Johnson, S., and Baglioni, C. (1987a) Protection from tumor necrosis factor cytotoxicity by protease inhibitors. Cell Immun., 107, 317-325.

Ruggiero, V., Latham, K., and Baglioni, C. (1987b) Cytostatic and cytotoxic activity of tumour necrosis factor on human cancer cells. J Immunol, 138, 2711-2717.

Saks, S., and Rosenblum, M. (1992) Recombinant human TNF-alpha: preclinical studies and results from early clinical trials. Immun Ser., 56, 567-587.

Sakuma, S., Sawamura, Y., Tada, M., Aida, T., Abe, H., Suzuki, K., and Taniguchi, N (1993) Responses of human glioblastoma cells to human natural tumour necrosis factor- alpha: susceptibility, mechanism of resistance and cytokine production studies. J Neuro- OncoL, 15, 197-208.

Saleem, A., Yuan, Z.-M., Taneja, N., Rubin, E., Kufe, D.W., and Kharbanda, S.M. (1995) Activation of serine/threonine protein kinases and early growth response 1 gene expression by tumor necrosis factor in human myeloid leukemia cells. J Immunol, 154, 4150-4156.

Sampaio, E.P., Sarno, E.N., Galilly, R., Cohn, Z.A., and Kaplan, G. (1991) Thalidomide selectively inhibits tumor necrosis factor alpha production by stimulated human monocytes. J Exp Med., 173, 699-703.

Sampson, L.E., and Chaplin, D.J. (1994) The influence of microenvironment on the cytotoxicity of TNF-a in vitro. Int J Radiat Oncol Biol Phys., 29, 467-471.

207 Sanguedolce, M.V., Capo, C., Bongrand, P., and Mege, J.L. (1992) Zymosan-stimulated tumor necrosis factor-alpha production by human monocytes. Down-modulation by phorbol ester. J Immunol, 148, 2229-2236.

Sariban, E., Luebbers, R., and Kufe, D. (1988) Transcriptional and posttranscriptional control of c-fos gene expression in human monocytes. Mol Cell Biol, 8, 340-346.

Sato, M., and Bremner, I. (1993) Oxygen free radicals and metallothionein. Free Radie Biol Med., 14, 325-337.

Sato, M., Sasaki, M., and Hojo, H. (1995) Antioxidant roles of metallothionein and manganese superoxide dismutase induced by tumor necrosis factor-a and interleukin-6. Arch Biochem Biophys., 316, 738-744.

Sato, M., Sasaki, M., Hojo, H., Shaul, P.W., Farrar, M.A., and Magness, R.R. (1992) Tissue specific induction of metallothionein synthesis by tumor necrosis factor-alpha

Sato, N., Goto, T., Haranaka, K., Satomi, N., Nariuchi, H., Mano-Hirano, Y., and Sawasaki, Y. (1986) Actions of tumor necrosis factor on cultured vascular epithelial cells:morphologic modulation, growth inhibition and cytotoxicity. J Natl Cane Inst., 76, 1113-1117.

Satomi, N., Haranaka, K., and Kunii, O. (1981) Research on the production site of tumor necrosis factor (TNF). Jpn J Exp Med., 51, 317-322.

Scanned, G., Waxman, K., Kalm, G.J., loli, G., Gatanaga, T., Yamamoto, R., and Granger, G.A. (1993) Hypoxia induces a human macrophage cell line to release tumor necrosis factor-alpha and its soluble receptors in vitro. J Surg Res., 54, 281-285.

Schade, U.F., Burmeister, I., Engel, R., Reinke, M., and Wolter, D.T. (1989) Lipoxygenase inhibitors suppress formation of tumor necrosis factor in vitro and in vivo. Lymphokine Res., 8, 245-250.

Schall, T.J., Lewis, M., Roller, K.J., Lee, A., Rice, G.C., Wong, G.H.W., Gotanaga, T., Granger, G.A., Lentz, R., Raab, H., Kohr, W.J., and Goeddel, D.V. (1990) Molecular cloning and expression of a receptor for human tumour necrosis factor. Cell, 61, 361- 370.

Scheurich, P., Kobrich, G., and Pfizenmaier, K. (1989) Antagonistic control of tumour necrosis factor receptors by protein kinases A and C. Enhancement of TNF synthesis by protein kinase A and transmodulation of receptors by protein kinase C. J Exp Med., 170, 947-958.

Scheurich, P., Thoma, B., Ucer, U., and Pfizenmaier, K. (1987) Immunoregulatory activity of recombinant human tumor necrosis factor (TNF)-a: induction of TNF receptors on human T cells and TNFa mediated enhancement of T cell responses. J Immunol, 138, 1786-1790.

Schiller, J.H., Bittner, G., Storer, B., and Willson, J.K. (1987) Synergistic antitumor effects of tumor necrosis factor and gamma-interferon on human colon carcinoma cell lines. Cancer Res., 47, 2809-2813.

208 Schoenfeld, H.J., Poeschl, B., Frey, J R., Loetscher, H., Hunziker, W., Lustig, A., and Zulauf, M. (1991) Efficient purification of recombinant human tumor necrosis factor beta from Escherichia coli yields biologically active protein with a trimeric structure that binds to both tumor necrosis factor receptors. J Biol Chem., 266, 3863-3869.

Scholz, W., and Altman, A. (1989) Synergistic induction of interleukin 2 receptor (TAG) expression on YT cells by interleukin 1 or tumor necrosis factor alpha in combination with cAMP inducing agents. Cell Signal., 1, 367-375.

Schuger, L., Varani, J., Marks, R.M., Kunkel, S.L., Johnson, K.J., and Ward, P.A. (1989) Cytotoxicity of tumor necrosis factor-alpha for human umbilical vein endothelial cells. Lab Invest., 61, 62-68.

Schulze Osthoff, K., Beyaert, R., Vandevoorde, V., Haegeman, G., and Fiers, W. (1993) Depletion of the mitochondrial electron transport abrogates the cytotoxic and gene- inductive effects of TNF. EMBO J., 12, 3095-3104.

Schulze-Osthoff, K., Bakker, A C., Vanhaesebroeck, B., Beyaert, R., Jacob, W.A., and Fiers, W. (1992) Cytotoxic activity of tumor necrosis factor is mediated by early damage of mitochondrial functions. Evidence for the involvement of mitochondrial radical generation. J Biol Chem., 267, 5317-5323.

Schiitze, S., Nottrott, S., Pfizenmaier, K., and Kronke, M. (1990) Tumor necrosis factor signal transduction. Cell-type-specific activation and translocation of protein kinase C. J Immunol, 144, 2604-2608.

Schiitze, S., Scheurich, S., Schulter, C., User, U., Pfizenmaier, K., and Kronke, M. (1988) Tumor necrosis factor-induced changes of gene expression in U937 cells: Differentiation-dependent plasticity of the responsive state. J Immunol, 140, 3000- 3005.

Sciandra, J.J., Subjeck, J.R., and Hughes, C.S. (1984) Induction of glucose-regulated proteins during anaerobic exposure and of heat-shock proteins after reoxygenation. Proc Natl Acad Sci USA., 81, 4843-4847.

Scuderi, P., Sterling, K.E., Lam, K.S., Finley, P R., Ryan, K.J., Ray, C.G., Petersen, E., Slymen, D.J., and Salmon, S. (1986) Raised serum levels of tumor necrosis factor in parasitic infections. Lancet., i, 1364-1365.

Seckinger, P., Isaaz, S., and Dayer, J.M. (1988) A human inhibitor of tumor necrosis factor alpha. 7 167, 1511-1516.

Selby, P., Hobbs, S., Fearon, K., Viner, C., Jackson, E., Jones, A., Newell, D., Calvert, A.H., McElwain, T., Feason, K., Humphreys, J., and Shiga,T. (1987) Tumour necrosis factor in man: clinical and biological observations. Br J Cancer., 56, 803-808.

Sen, R., and Baltimore, D. (1986) Inducibility of kappa immunoglobulin enhancer- binding protein NF-kappa B by a posttranslational mechanism. Cell, 47, 921-928.

Sersa, G., Willingham, V., and Milas, L. (1988) Antitumor effects of tumor necrosis factor alone or combined with radiotherapy. Int J Cancer., 42, 129-134.

209 Shaffer, J.B., Treanor, C.P., and Del Vecchio, PJ. (1990) Expression of bovine and mouse endothelial cell antioxidant enzymes following TNF-alpha exposure. Free Radie Biol Med., 8, 497-502.

Shakhov, A.N., Collart, M.A., Vassalli, P., Nedospasov, S.A., and Jongeneel, V.C. (1990) KB-type enhancers are involved in lipopolysaccharide-mediated transcriptional activation of the tumor necrosis factor a gene in primary macrophages. J Exp Med., 171, 35-47.

Shalaby, M.R., Aggarwal, B.B., Rinderknecht, E., Svedersky, L.P., Pinkie, B.S., and Palladino Jr, M.A. (1985) Activation of human polymorphonuclear neutrophil functions by interferon-gamma and tumor necrosis factors. J Immunol, 135, 2069-2073.

Shannon, M.F., Pell, L.M., Lenardo, M.J., Kuczek, E.S., Occhiodoro, F.S., Dunn, S.M., and Vadas, M.A. (1990) A novel tumor necrosis factor-responsive transcription factor which recognizes a regulatory element in hemopoietic growth factor genes. Mol Cell Biol, 10, 2950-2959.

Shau, H. (1988) Characteristics and mechanism of neutrophil-mediated cytostasis induced by tumor necrosis factor. J Immunol, 141, 234-240.

Shaul, P.W., Campbell, W.B., Farrar, M.A., and Magness, R.R. (1992) Oxygen modulates prostacyclin synthesis in ovine fetal pulmonary arteries by an effect on cyclooxygenase. J Clin Invest., 90, 2147-2155.

Shaw, G., and Kamen, R. (1986) A conserved AU sequence from the 3' untranslated region of GM-CSF mRNA mediates selective mRNA degradation. Cell, 46, 659-667.

Shear, M.J., Turner, F.C., Preeault, A., and Shovelton, T. (1943) Chemical treatment of tumours.V.Isolation of the hemorrhage-producing fraction from Serratia marcescens {Bacillus prodigiosus) culture filtrates. J Natl Cancer Inst., 4, 81-97.

Sheldon, P.W., and Hill, S.A. (1977) Hypoxic cell radiosensitizers and local control by X-ray of a transplanted tumour in mice. Br J Cancer., 35, 795-808.

Sher, T., Rottem, S., and Gallily, R. (1990) Mycoplasma capricolum membranes induce tumor necrosis factor alpha by a mechanism different from that of lipopolysaccharide. Cancer Immunol Immunother., 31, 86-92.

Shimomura, K., Manda, T., Mukumoto, S., Kobayashi, K., Nakano, K., and Mori, J. (1988) Recombinant human tumour necrosis factor-a: thrombus formation is a cause of anti-tumour activity. Int J Cancer., 41, 243-247.

Shine, N., Palladino, M.A., Patton, J.S., Diesseroth, A., Karczmar, G.S., Matson, G.B., and Weiner, M.W. (1989) Early metabolic response to tumour necrosis factor in mouse sarcoma L: a phosphorus-31 nuclear magnetic resonance study. Cancer Res., 49, 2123- 2127.

Shirai, T., Yamaguchi, H., Ito, H., Todd, C.W., and Wallace, R.B. (1985) Cloning and expression in E.coli of the gene for human tumor necrosis factor. Nature., 313, 803-806.

210 Shoji, Y., Uedono, Y., Ishikura, H., Takeyama, N., and Tanaka, T. (1995) DNA damage induced by tumour necrosis factor-a in L929 cells is mediated by mitochondrial oxygen radical formation. Immunology., 84, 543-548.

Shrieve, D.C., and Begg, A.C. (1985) Cell cycle kinetics of aerated, hypoxic and re­ aerated cells in vitro using flow cytometric determination of cellular DNA and incorporated bromodeoxyuridine. Cell Tissue Kinet., 18, 641-651.

Shrieve, D C., Deen, D.F., and Harris, J.W. (1983) Effects of extreme hypoxia on the growth and viability of EMT6/SF mouse tumor cells in vitro. Cancer Res., 43, 3521- 3527.

Sidhu, R.S., and Bollon, A.P. (1993) Tumor necrosis factor activities and cancer therapy-a perspective. Pharmac Ther., 57, 79-128.

Silberstein, D.S., and David, J.R. (1986) Tumor necrosis factor enhances eosinophil toxicity to Schistosoma mansoni larvae. Proc Natl Acad Sci USA., 83, 1055.

Silk, S.T., Clejan, S., and Witkom, K. (1989) Evidence of GTP-binding protein regulation of phospholipase Az activity in isolated human platelet membranes. J Biol Chem., 264, 21466-21469.

Smith, E., Stratford, I.J., and Adams, G.E. (1980) Cytotoxicity of adriamycin on aerobic and hypoxic Chinese hamster V79 cells in vitro. Br J Cancer., 42, 568-573.

Smith, M R., Munger, W.E., Kung, H.-F., Takacs, L., and Durum, S.K. (1990) Direct evidence for an intracellular role for tumor necrosis factor-a. Microinjection of tumor necrosis factor kills target cells. J Immunol, 144, 162.

Smith, R.A., and Baglioni, C. (1987) The active form of tumor necrosis factor is a trimer. J Biol Chem., 262, 6951-6954.

Smith, R.A., and Baglioni, C. (1992) Characterization of TNF receptors. Immunol Ser., 56, 131-147.

Solis Herruzo, J.A., Brenner, D.A., and Chojkier, M. (1988) Tumor necrosis factor alpha inhibits collagen gene transcription and collagen synthesis in cultured human fibroblasts. J Biol Chem., 263, 5841-5845.

Song, H.Y., Dunbar, J.D., and Donner, D.B. (1994) Aggregation of the intracellular domain of the type 1 tumor necrosis factor receptor defined by the two-hybrid system. J Biol Chem., 269, 22492-22495.

Sorrentino, V. (1989) Growth factors, growth inhibitors and cell cycle control (Review). Anticancer Res., 9, 1925-1936.

Spengler, R.N., Spengler, M.L., Lincoln, P., Remick, D.G., Stricter, R.M., and Kunkel, S.L. (1989) Dynamics of dibutyryl cyclic AMP- and prostaglandin Eo. mediated suppression of lipopolysaccharide-induced tumor necrosis factor alpha gene expression. Infect Immun., 57, 2837-2841.

211 Spies, T., Blanck, G., Bresnahan, M., Sands, J., and Strominger, J.L. (1989) A new cluster of genes within the human major histocompatibility complex. Science., 243, 214- 217.

Spies, T., Morton, C.C., Nedospasov, S.A., Fiers, W., Pious, D., and Strominger, J.L. (1986) Genes for the tumor necrosis factors alpha and beta are linked to the human major histocompatibility complex. Proc Natl Acad Sci USA., 83, 8699-8702.

Spiro, I.J., Rice, G.C., Durand, R.E., Stickler, R., and Ling, C.C. (1984) Cell killing, radiosensitization and cell cycle redistribution induced by chronic hypoxia. Int J Rad Oncol Biol Phys., 10, 1275-1280.

Spriggs, D R., and Yates, S.W. (1992). Cancer chemotherapy. Experience with TNF administration in humans. In Tumor Necrosis Factors. The Molecules and Their Emerging Role in Medicine (ed B. Beutler), pp 383-406. Raven Press, New York.

Spriggs, D R., Imamura, K., Rodriguez, C., Horiguchi, J., and Kufe, D.W. (1987) Induction of tumour necrosis factor expression and resistance in a human breast tumour cell line. Proc Natl Acad Sci USA., 84, 6563-6566.

Spriggs, D R., Sherman, M.L., Imamura, K., Mohri, M., Rodriguez, C., Rubbins, G., and Kufe, DW (1990) Phospholipase A 2 activation and autoinduction of tumour necrosis factor gene expression by tumour necrosis factor. Cancer Res., 50, 7101-7101.

Stackpole, C.W., Groszek, L., and Kalbag, S.S. (1994) Benign-to-malignant B16 melanoma progression induced in two stages in vitro by exposure to hypoxia. J Natl Cancer Inst., 8 6 , 361-367.

Stallmach, T., Hebisch, G., Joller Jemelka, HI., Orban, P., Schwaller, J. and Engelmann, M. (1995) Cytokine production and visualized effects in the feto-matemal unit. Quantitative and topographic data on cytokines during intrauterine disease. Lab Invest., 73, 384-392.

Stephens, J.M., Butts, M.D., and Pekala, P.H. (1992) Regulation of transcription factor mRNA accumulation during 3T3-L1 preadipocyte differentiation by tumour necrosis factor-alpha. J Mol Endocrinol, 9, 61-72.

Stevens, J.B., and Autor, A.P. (1977) Induction of superoxide dismutase by oxygen in neonatal rat lung. J Biol Chem., 252, 3509-3514.

Stolpen, A., Guinan, E., Fiers, W., and Pober, J. (1986) Recombinant tumor necrosis factor and immune interferon act singly and in combination to reorganize human vascular endothelial cell monolayers. Am J Pathol, 123, 16-24.

Stone, R.M., Weber, B.L., Spriggs, D R., and Kufe, D.W. (1988) Phospholipase C activates protein kinase C and induces monocytic differentiation of HL-60 cells. Blood., 72, 739-744.

Strieter, R.M., Remick, D.G., Lynch, J.P.d., Genord, M., Raiford, C., Spengler, R., and Kunkel, S.L. (1989) Differential regulation of tumor necrosis factor-alpha in human alveolar macrophages and peripheral blood monocytes: a cellular and molecular analysis. Am J Respir Cell Mol Biol, 1, 57-63.

212 Suffys, P., Beyaert, R., De Valck, D., Vanhaesebroeck, B., Van Roy, F., and Piers, W. (1991) Tumour-necrosis-factor-mediated cytotoxicity is correlated with phospholipase- A% activity, but not with arachidonic acid release per se. Ear J Biochem., 195, 465-75.

Suffys, P., Beyaert, R., Van Roy, P., and Piers, W. (1988b) Involvement of a serine protease in tumour-necrosis-factor-mediated cytotoxicity. Eur J Biochem., 178, 257-65.

Suffys, P., Van Roy, P., and Piers, W. (1988a) Tumor necrosis factor and interleukin 1 activate phospholipase in rat chondrocytes. FEBS Lett., 232, 24-28.

Sugarman, B.J., Aggarwal, S.B., Hass, P.P., Pigari, I.S., Palladino, M.A.J., and Shephard, H.M. (1985) Recombinant human tumour necrosis factor-a: Effects on proliferation of normal and transformed cells in vitro. Science (Wash.)., 230, 943-948.

Sung, S.J., Bjomdahl, J.M., Wang, C.Y., Kao, H.T., and Pu, S.M. (1988a) Production of tumor necrosis factor/cachectin by human T cell lines and peripheral blood T lymphocytes stimulated by phorbol myristate acetate and anti-CD3 antibody. J Exp Med., 167, 937-953.

Sung, S.J., Walters, J.A., and Pu, S.M. (1992) Stimulation of tumor necrosis factor alpha production in human monocytes by inhibitors of protein phosphatase 1 and 2A. J Exp Med., 176, 897-901.

Sung, S.S., Jung, L.K., Walters, J.A., Chen, W., Wang, C.Y., and Pu, S.M. (1988b) Production of tumor necrosis factor/cachectin by human B cell lines and tonsillar B cells. J Exp Med., 168, 1539-1551.

Sutherland, R., Preyer, J., Mueller Klieser, W., Wilson, R., Heacock, C., Sciandra, J., and Sordat, B. (1986) Cellular growth and metabolic adaptations to nutrient stress environments in tumor microregions. Int J Radiat Oncol Biol Phys., 12, 611-615.

Tabibzadeh, S.S., Santhanam, U., Sehgal, P.B., and May, L.T. (1989) Cytokine-induced production of IPN-P 2 /IL- 6 by freshly explanted human endometrial stromal cells. Modulation by estradiol-17 p. J Immunol, 142, 3134-3139.

Taffet, S.M., Singhel, K.J., Overholtzer, J.P., and Shurtleff, S.A. (1989) Regulation of tumor necrosis factor expression in a macrophage-like cell line by lipopolysaccharide and cyclic AMP. Cell Immunol, 120, 291-300.

Taguchi, T., and Sohmura, Y. (1991) Clinical studies with TNP. Biotherapy., 3, 177- 186.

Takahashi, N., Brouckaert, P., and Piers, W. (1991) Induction of tolerance allows separation of lethal and antitumor activities of tumor necrosis factor in mice. Cancer Res., 51, 2366-2372.

Takasuka, N., Matsuura, K., Yamamoto, S., and Akagawa, K.S. (1995) Suppression of TNP-alpha mRNA expression in EPS-primed macrophages occurs at the level of nuclear factor-kappa B activation, but not at the level of protein kinase C or CD 14 expression. J Immunol, 154, 4803-4812.

2 1 3 Talmadge, J.E., Bowersox, O., Tribble, H., Lee, S.H., Shepard, M., and Liggit, D. (1987) Toxicity of tumour necrosis factor is synergistic with interferon and can be reduced with cyclooxygenase inhibitors. Am J Path., 128, 410-425.

Tannenbaum, C.S., Major, J.A., and Hamilton, T.A. (1993) IFN-gamma and lipopolysaccharide differentially modulate expression of tumor necrosis factor receptor mRNA in murine peritoneal macrophages. J Immunol, 151, 6833-6839.

Tannock, I.F. (1968) The relation between cell proliferation and the vascular system in a transplanted mouse mammary tum'our. Br J Cancer., 22, 258-273.

Tartaglia, L.A., and Goeddel, D.V. (1992a) Two TNF receptors. Immunol Today., 13, 151-153.

Tartaglia, L.A., and Goeddel, D.V. (1992b) Tumor necrosis factor receptor signaling. A dominant negative mutation suppresses the activation of the 55-kDa tumor necrosis factor receptor. J Biol Chem., 267, 4304-4307.

Tartaglia, L.A., Pennica, D., and Goeddel, D.V. (1993) Ligand passing: the 75-kDa tumor necrosis factor (TNF) receptor recruits TNF for signaling by the 55-kDa TNF receptor. J Biol Chem., 268, 18542-18548.

Tartaglia, L.A., Weber, R.F., Figari, I.S., Reynolds, C., Palladino, M.A., and Goeddel, D.V. (1991) The two different receptors for tumor necrosis factor mediate distinct cellular responses. Proc Natl Acad Sci USA., 88, 9292-9296.

Tchorzewski, H., Zeman, K., Paleolog, E., Brennan, F., Feldmann, M., Kahan, M., Guga, P., Kwinkowski, M., Szymanska, B., Jarosz, J., Parniewski, P., and Kocur, E. (1993) The effects of tumor-necrosis-factor (TNF) derivatives on TNF receptors. Cytokine., 5, 125-132.

Tebo, J.M., and Hamilton, T.A. (1994) Okadaic acid stimulates inflammatory cytokine gene transcription in murine peritoneal macrophages. Cell Immunol, 153, 479-491.

Teicher, B.A. (1994) Hypoxia and drug resistance. Cancer and Metastasis Rev., 13, 139-168.

Tewari, M., and Dixit, V.M. (1995) Fas- and tumor necrosis factor-induced apoptosis is inhibited by the poxvirus crmA gene product. J Biol Chem., 270, 3255-3260.

Thoma, B., Grell, M., Pfizenmaier, K., and Scheurich, P. (1990) Identification of a 60- kD tumour necrosis factor (TNF) receptor as a major signal transducing component in TNF responses. Id J Exp. Med., 172, 1019-1023.

Thomas, C D., Chavaudra, N., Martin, L., and Guichard, M. (1994) Correlation between radiosensitivity, percentage hypoxic cells and pOo measuerments in one rodent and two human tumour xenografts. Radiat Res., 139, 1-8.

Thomlinson, R.H. and Gray, L.H. (1955) The histological structure of some human lung cancers and the possible implications for radiotherapy. Br J Cancer., 9, 539-549.

2 1 4 Torti, F.M., Dieckmann, B., Beutler, B., Cerami, A., and Ringold, G.M. (1985) A macrophage factor inhibits adipocyte gene expression: an in vitro model of cachexia. Science., 229, 867-869.

Totpal, K., Chaturvedi, M.M., LaPushin, R., and Aggarwal, B.B. (1995) Retinoids downregulate both p60 and p80 forms of tumor necrosis factor receptors in human histiocytic lymphoma U-937 cells. Blood., 85, 3547-3555.

Tracey, K.J., Beutler, B., Lowry, S.F. Merryweather, J., Wolpe, S., Milsark, I.W., Harir, R.J., Fahey III, T.J., Zentella, A., Albert, J.D., Shires, G.T., and Cerami, A. (1986b) Shock and tissue injury induced by recombinant human cachectin. Science (Wash.)., 234, 470-474.

Tracey, K.J., Lowry, S.F., Beutler, B., Cerami, A., Albert, J.D., and Shires, G.T. (1986a) Cachectin/tumor necrosis factor mediates changes of skeletal muscle plasma membrane potential. J Exp Med., 164, 1368-1373.

Trotter, M.J., Chaplin, D.J., and Olive, P.L. (1991) Possible mechanisms for intermittent blood flow in the murine SCCVII carcinoma. Int J Radiat Biol, 60, 139-146.

Tsan, M.F. (1993) Superoxide dismutase and pulmonary oxygen toxicity. Proc Soc Exp Biol Med., 203, 286-290.

Tsan, M.F., White, I.E., Del Vecchio, P.J., and Shaffer, J.B. (1992) IL-6 enhances TNF- alpha- and IL-1-induced increase of Mn superoxide dismutase mRNA and O? tolerance. Am J Physiol, 263, 22-26.

Tsan, M.F., White, I.E., Treanor, C. and Shaffer, J.B. (1990) Molecular basis for tumor necrosis factor-induced increase in pulmonary superoxide dismutase activities. Am J Physiol, 259, 506-512.

Tsujimoto, M., and Vilcek, J. (1987) Tumor necrosis factor-induced downregulation of its receptors in HeLa cells, J Biochem Tokyo., 102, 1571-1577.

Tsujimoto, M., Yip, Y., and Vilcek, J. (1985) Tumour necrosis factor: Specific binding and internalization in sensitive and resistant cells. Proc Natl Acad Sci, USA., 82, 7626- 7630.

Turrens, J.F., Freeman, B.A., Levitt, J.G., and Crapo, J.D. (1982) The effect of hyperoxia on superoxide production by lung submitochondrial particles. Arch Biochem Biophys., 217, 401-410.

Urban, J., Shepard, H., Rothstein, J., Sugarman, B., and Schreiber, H. (1986) Tumor necrosis factor: A potent effector molecule for tumour cell killing by activated macrophages. Proc Natl Acad Sci USA., 83, 5233-5237.

Utsumi, T., Klostergaard, J., Akimaru, K., Edashige, K., Sato, E.F., and Utsumi, K. (1992) Modulation of TNF-alpha-priming and stimulation-dependent superoxide generation in human neutrophils by protein kinase inhibitors. Arch Biochem Biophys., 294, 271-278.

215 Valone, S.E., Rich, E.A., Wallis, R.S., and Ellner, J.J. (1988) Expression of tumor necrosis factor in vitro by human mononuclear phagocytes stimulated with whole Mycobacterium bovis BCG and mycobacterial antigens. Infect Immun., 56, 3313-3315.

Van de Wiel, P.A., Bloksma, N., Kuper, C.F., Hofhuis, P.M., and Willers, J.M. (1989) Macroscopic and microscopic early effects of tumour necrosis factor on murine Meth A sarcoma, and relation to curative activity. J Pathol, 157, 65-73.

Van de Wiel, P.A., Bouma, G.J., van der Pijl, A., Weitenberg, E.S., Lam, A.W., and Bloksma, N. (1990) Effect of tumour necrosis factor and Lipid A on functional and structural vascular column in solid murine tumours. Br J Cancer., 62, 718-723.

Van de Wiel, P.A., Weitenberg, E.S., van der Pijl, A., and Bloksma, N. (1992) Factors that define the susceptibility of endothelial cells to tumor necrosis factor and lipid A. Immunopharmacology., 23, 49-56.

Van Kessel, K.P., van Strijp, J.A., and Verhoef, J. (1991) Inactivation of recombinant human tumor necrosis factor-alpha by proteolytic enzymes released from stimulated human neutrophils. J Immunol, 147, 3862-3868.

Van Lint, J., Agostinis, P., Vandevoorde, V., Haegeman, G., Fiers, W., Merlevede, W., and Vandenheede, J.R. (1992) Tumor necrosis factor stimulates multiple serine/threonine protein kinases in Swiss 3T3 and L929 cells. Implication of casein kinase-2 and extracellular signal-regulated kinases in the tumor necrosis factor signal transduction pathway. J Biol Chem., 267, 25916-25921.

Van Ostade, X., Tavernier, J., and Fiers, W. (1994) Structure-activity studies of human TNF. Protein Engineering., 7, 5-22.

Van Putten, L., and Kallman, R. (1968) Oxygenation status of a transplantable tumor during fractionated radiation therapy. J Natl Cancer Inst., 40,441-451.

Vandenabeele, P., Declercq, W., Beyaert, R., and Fiers, W. (1995b) Two tumour necrosis factor receptors: structure and function. Trends in Cell Biol, 5, 392-399.

Vandenabeele, P., Declercq, W., Vanhaesebroeck, B., Grooten, J., and Fiers, W. (1995a) Both TNF receptors are required for TNF-mediated induction" of apoptosis in PC60 cells. J Immunol, 154, 2904-2913.

Vandenabeele, P., Goossens, V., Beyaert, R., Declercq, W., Grooten, J., Vanhaesebroeck, B., Van de Craen, M., Vercammen, D., Depuydt, B., and Denecker, G. (1994) Functional requirement of the two TNF receptors for induction of apoptosis in PC60 cells and the role of mitochondria in TNF-induced cytotoxicity. Circ Shock., 44, 196-200.

Vanhaesebroeck, B., Decoster, E., Van Ostade, X., Van Bladel, S., Lenaerts, A., Van Roy, F., and Fiers, W. (1992) Expression of an exogenous tumor necrosis factor (TNF) gene in TNF-sensitive cell lines confers resistance to TNF-mediated cell lysis. J Immunol, 148, 2785-2794.

216 Vannier, E., Miller, L.C., and Dinarello, C.A. (1991) Histamine suppresses gene expression and synthesis of tumor necrosis factor alpha via histamine H2 receptors. J Exp Med., 174, 281-284.

Vaupel, P., D. K. Kelleher, and Engel, T. (1994) Do changes in tumour blood flow necessarily lead to changes in tissue oxygenation and in bioenergetic status? Adv Exp Med BioL, 361, 607-611.

Vaupel, P., Kallinowski, F., and Okunieff, P. (1990) Blood flow, oxygen consumption and tissue oxygenation of human tumors. Adv Exp Med BioL, 277, 895-905.

Vaupel, P., Okenieff, P., Kallinowski, P., and Neuringer, L.K. (1989) Correlations between^ ^ P-NMR spectroscopy and tissue O 2 tension measurements in a murine fibrosarcoma. Radiat.Res., 120, 477-493.

Vaupel, P., Schlenger, K., Knoop, C., and Hockel, M. (1991) Oxygenation of human tumors: Evaluation of tissue oxygen distribution in breast cancers by computerized O 2 tension measurements. Cancer Res., 51, 3316-3322.

Vaupel, P.W., Frinak, S., and Bicher, H.I. (1981) Heterogeneous oxygen partial pressure and pH distribution in C3H mouse mammary adenocarcinoma. Cancer Res., 41, 2008- 2013.

Visner, G.A., Dougall, W.C., Wilson, J.M., Burr, I.A., and Nick, H.S. (1990) Regulation of manganese superoxide dismutase by lipopolysaccharide, interleukin-1, and tumor necrosis factor. Role in the acute inflammatory response. J Biol Chem., 265, 2856-2864.

Waage, A., Halstensen, A., and Espevik, T. (1987) Association of tumor necrosis factor- a and fatal outcome in patients with meningococcal meningitis. Lancet., i, 355-357.

Waage, A., Sorensen, M., and Stordal, B. (1990) Differential effect of oxypentifylline on tumour necrosis factor and interleukin-6 production. Lancet., 335, 543.

Wan, J.M., Fogt, P., Bistrian, B.R., and Istfan, N.W. (1993) Evaluation of antitumor effect of tumor-necrosis-factor in terms of protein-metabolism and cell-cycle kinetics. Am J Phys., 265, C365-C374.

Warner, B.B., Burhans, M.S., Clark, J.C., and Wispe, J.R. (1991) Tumour necrosis factor-a increases Mn-SOD expression: Protection against oxidant injury. Am J Phys., 260, L296- L301.

Warren, D., Pasupuleti, V., Plater, B., Lu, Y., and Horton, R. (1989). Tumor necrosis factor and interleukin 1 inhibit testosterone secretion in cultured rat testis cells but stimulate testosterone production in adult rat testis cells (abstr) Program of the American Endocrine Society Meeting, pp. 312.

Watanabe, N., Kuriyama, H., Sone, H., Neda, H., Yamauchi, N., Maeda, M., and Niitsu, Y. (1988a) Continuous internalization of TNF receptors in a human myosarcoma cell line. J Biol Chem., 263, 10262-10266.

217 Watanabe, N., Niitsu, Y., Neda, H., Sone, H., Yamauchi, N., Umetsu, T., and Urushizaki, I. (1985) Antitumor effect of the tumor necrosis factor against various types of human cancer cells. Gan To Kagaku Ryoho., 12, 2185-2189.

Watanabe, N., Niitsu, Y., Umeno, H., Kuriyama, H., Neda, H., Yamauchi, N., Maeda, M., and Urushizaki, I. (1988b) Toxic effect of tumor necrosis factor on tumor vasculature in mice. Cancer Res., 48, 2179-2183.

Watanabe, N., Niitsu, Y., Yamauchi, N., Neda, H., Sone, H., Maeda, M., and Urushizaki, I. (1987) Cell cycle specificity of tumour necrosis factor and its receptor. Cell Biol Int Rep., 11, 813-817.

Weiner, F.R., Smith, P.J., Wertheimer, S., and Rubin, C.S. (1991) Regulation of gene expression by insulin and tumor necrosis factor a in 3T3-L1 cells. J Biol Chem., 266, 23525-23528.

West, M.A., Li, M.H., Seatter, S.C., and Bubrick, M.P. (1994) Pre-exposure to hypoxia or septic stimuli differentially regulates endotoxin release of tumor necrosis factor, interleukin-6, interleukin-1, prostaglandin Ei, nitric oxide, and superoxide by macrophages. J Trauma., 37, 82-89.

Wheatley, E.M., McKeown Longo, P.J., Vincent, P.A., and Saba, T.M. (1993) Incorporation of fibronectin into matrix decreases TNF-induced increase in endothelial monolayer permeability. Am J Physiol, 265, L148-L157.

Wiegmann, K., Schiitze, S., Machleidt, T., Witte, D., and Kronke, M. (1994) Functional dichotomy of neutral and acidic sphingomyelinases in tumor necrosis factor signaling. Cell, 78, 1005-1015.

Wilson, R.E., and Sutherland, R.M. (1989) Enhanced synthesis of specific proteins, RNA, and DNA caused by hypoxia and reoxygenation. Int J Radiat Oncol Biol Phys., 16, 957-961.

Wing, D., Talley, G., and Storch, T. (1988) Oxygen concentration regulates EGF- induced proliferation and EGF-receptor down regulation. Biochem Biophys Res Comm., 153, 952-958.

Wingfield, P., Pain, R.H., and Craig, S. (1987) Tumour necrosis factor is a compact trimer. FEBS Lett., 211, 179-184.

Winzen, R., Wallach, D., Engelmann, H., Nophar, Y., Brakebusch, C., Kemper, O., Resch, K., and Holtmann, H. (1992) Selective decrease in cell surface expression and mRNA level of the 55-kDa tumor necrosis factor receptor during differentiation of HL- 60 cells into macrophage-like but not granulocyte-like cells. J Immunol, 148, 3454- 3460.

Winzen, R., Wallach, D., Kemper, O., Resch, K., and Holtmann, H. (1993) Selective up-regulation of the 75-kDa tumor necrosis factor (TNF) receptor and its mRNA by TNF and IL-1. J Immunol, 150, 4346-4353.

Wong, G., Kamb, A., Elwell, J., Oberley, L., and Goeddel, D. (1992) MnSOD induction by TNF and its protective role. In Tumour Necrosis Factors: The Molecules and Their Emerging Role in Medicine (ed B. Beutler), pp 473-484. Raven Press, Ltd, New York.

2 1 8 Wong, G.H., and Goeddel, D.V. (1986) Tumour necrosis factors alpha and beta inhibit virus replication and synergize with interferons. Nature., 323, 819-822.

Wong, G.H.W., and Goeddel, D.V. (1988) Induction of manganous superoxide dismutase by tumor necrosis factor: Possible protective mechanism. Science., 242, 941- 944.

Wong, G.H.W., Elwell, J.H., Oberley, L.W., and Goeddel, D.V. (1989) Manganous superoxide dismutase is essential for cellular resistance to cytotoxicity of tumor necrosis factor. Cell., 58, 923-931.

Yamashita, N., Nishida, M., Hoshida, S., Kuzuya, T., Hori, M., Taniguchi, N., Kamada, T., and Tada, M. (1994) Induction of manganese superoxide dismutase in rat cardiac myocytes increases tolerance to hypoxia 24 hours after preconditioning. J Clin Invest., 94,2193-2199.

Yamauchi, N., Kuriyama, H., Watanabe, N., Neda, H., Maeda, M., and Niitsu, Y. (1989) Intracellular hydroxyl radical production induced by recombinant human tumor necrosis factor and its implication in the killing of tumor cells in vitro. Cancer Res., 49, 1671- 1675.

Yamaura, H., and Matsuzawa, T. (1979) Tumor regrowth after irradiation; an experimental approach. Int J Radiat Biol Relat Stud Phys Chem Med., 35, 201-219.

Yi, E.S., and Ulich, T.R. (1992) Endotoxin, interleukin-1 and tumor necrosis factor cause neutrophil-dependent microvascular leakage in postcapillary venules. Am J Path., 140, 659-662.

Yoshie, O., Tada, K., and Ishida, N. (1986) Binding and crosslinking of^“^I-labelled recombinant human tumor necrosis factor to cell surface receptors. J Biochem Tokyo., 100, 531-541.

Yoshihara, R., Shiozawa, S., Imai, Y., and Fujita, T. (1990) Tumor necrosis factor alpha and interferon gamma inhibit proliferation and alkaline phosphatase activity of human osteoblastic SaOS-2 cell line. Lymphokine Res., 9, 59-66.

Yoshizumi, M., Perrella, M.A., Burnett Jr., J.C., and Lee, M E. (1993) Tumor necrosis factor downregulates an endothelial nitric oxide synthase mRNA by shortening its half- life. Circ Res., 73, 205-209.

Young, J.D., Liu, C.C., Culter, G., Cohn, Z.A., and Gali, S.J. (1987) Identification, purification and characterization of a mast cell-associated cytolytic factor related to tumor necrosis factor. Proc Natl Acad Sci USA., 84, 9175-9179.

Young, S.D., Marshall, R.S., and Hill, R.P. (1988) Hypoxia induces DNA overreplication and enhances metastatic potential of murine tumour cells. Proc Natl Acad Sci USA., 85, 9533-9537.

Zetter, B.R. (1981) The endothelial cells of large and small blood vessels. Diabetes., 30, 24-28.

219 Zhang, Y.H., Lin, J.X., and Vilcek, J. (1990) Interleukin-6 induction by tumor necrosis factor and interleukin-1 in human fibroblasts involves activation of a nuclear factor binding to a kappa B-like sequence. Mol Cell BioL, 10, 3818-3823.

Zhang, Y.H., Lin, J.X., Yip, Y.K., and Vilcek, J. (1988) Enhancement of cAMP levels and of protein kinase activity by tumor necrosis factor and interleukin 1 in human fibroblasts: role in the induction of interleukin 6. Proc Natl Acad Sci USA., 85, 6802- 6805.

Zheng, R.Q., Abney, E.R., Chu, C.Q., Field, M., Maini, R.N., Lamb, J R., and Feldmann, M. (1992) Detection of in vivo production of tumour necrosis factor-alpha by human thyroid epithelial cells. Immunology., 75, 456-462.

Zimmerman, R., Marafmo, B., Chan, A., Landre, P., and Winkelhake, J. (1989) The role of oxidant injury in tumor cell sensitivity to recombinant human tumor necrosis factor in vivo. J Immunol., 142, 1405-1409.

Zuber, P., Accolla, R.S., Canel, S., Diserens, A C., and de Tribolet, N. (1988) Effects of recombinant human tumor necrosis factor-alpha on the surface phenotype and the growth of human malignant glioma cell lines. Int J Cancer., 42, 780-786.

Zuckerman, S.H., Bendele, A.M., Natarajan, R., Ploszaj, S., Horton, R., Nadler, J., Tabibzadeh, S.S., and Shah, K.D. (1989) Regulation of serum tumor necrosis factor in glucocorticoid-sensitive and -resistant rodent endotoxin shock models. Infect Immun., 57, 3009-3013.

Zuckerman, S.H., Evans, G.E., and Guthrie, L. (1991) Transcriptional and post­ transcriptional mechanisms involved in the differential expression of LPS-induced IL-1 and TNF mRNA. Immunology., 73, 460-465.

Zyad, A., Benard, J., Tursz, T., Clarke, R., and Chouaib, S. (1994) Resistance to TNF- alpha and Adriamycin in the human breast cancer MCF-7 cell line: Relationship to MDRl, MnSOD, and TNF gene expression. Cancer Res., 54, 825-831.

220 APPENDIX

I. General materials and methods 222 II. Western blotting solutions 226 ni. Details of transfected cell lines 227 IV. List of publications 230

221 APPENDIX

I. General materials and methods:

Materials Supplier

Antibodies Anti-rat IgG fluorescein isothiocyanate (FITC) Sigma Biotinylated rabbit anti-rat MnSOD IgG pAb Prof. N. Taniguchi Goat anti-mouse IgG FITC eonjugate (F-2012) Sigma Goat anti-rabbit IgG peroxidase conjugate A-9169 Sigma Hamster anti-mouse TNF MoAb 1221 Genzyme Purified rat anti-mouse TNF MoAb (1812D) Pharmingen Rabbit anti-mouse TNF pAb IP400 Genzyme Rabbit anti-rat MnSOD IgG pAb Prof. N.Taniguehi

Medium constituents Benzylpenecillin sodium BP (125 units/ml) Britannia Eagle's minimal essential medium (BMBM) (lOx) Gibeo BRL, Scotland BMBM suspension medium (lOx) Gibco BRL, Scotland Foetal calf serum TCS Biologicals Geneticin Sigma Glutamine (2mM) Integra Biosciences RPMI-1640 medium (1 Ox) Gibco BRL, Scotland Sodium bicarbonate solution 7.5% Integra Biosciences Streptomycin sulphate BP (lOOpg/ml) Evans Medical Ltd

All reagents were purchased from Sigma Chemical Co, Dorset, Poole, except those listed: P -mere aptoethanol BDH Copper II sulphate-5-hydrate (cupric sulphate) BDH ECL markers Amersham Ethanol BDH

222 Ethylenediamine tetraaceticacid (EOTA) BDH Glacial acetic acid Fisons Hanks balanced salt solution Gibco BRL, Scotland Heparin Injection BP (5000 units/ml) Pharmaceuticals Ltd Hydrochloric acid (Analar grade) BDH Chemicals Methanol BDH Mctofanc C-vet Non-fat milk Waitrose PBS tablets (Oxoid) Unipath Ltd Pentobarbitone sodium BP (Vet) May & Baker Ltd Saline solution Antigen Pharmaceuticals Sodium dodecyl sulphate (SDS) BDH Sterile water for injection Antigen Pharmaceuticals Sulphuric acid BDH Trypsin/EDTA Integra Biosciences

Equipment Butterfly-25, 4506, Venisystems Abbott Laboratories, Ltd Colony Counter Stuart Scientific Durapore surgical tape, 1538-0 3M Health Care, USA ECG electrodes for Eppendorf Medicotest EIA/RIA 96-well plates Costar Eppendorf ™ p02 Histograph Eppendorf, Hamburg, Germany FACscan Becton Dickinson Gases BOC Intradermic polyethylene tubing PE-20, 7406 Becton Dickinson Laser Doppler: The Oxford Array Oxford Optronix Microtitre plate reader: Multiskan MCC/340 Labsystems Mr Frosty Cryo 1 °C Freezing Container Nalgene NalgeneT"^ Freezing vials Nalgene 96-well plate sealers, 3095 Costar 96-well plates Greiner, Laboratory Supplies

223 Maintenance of cell lines: Culture medium All monolayer cells, except L929 cells, were grown in Eagle's minimal essential medium (BMBM) supplemented with foetal calf serum (PCS) 10%, glutamine (2mM), penicillin (125 units/ml, Crystapen 600mg) and streptomycin (lOOpg/ml) and buffered to pH 7.4 with sodium bicarbonate. This was modified for suspension cultures by replacing the BMBM with minimal essential medium for suspension cultures and by reducing the PCS to 7.5%. In the case of L929 cells BMBM was replaced with RPMI medium, all other medium constituents were the same as for BMBM. Medium was made up in 500 ml batches using distilled sterilized water which was autoclaved before use, all other reagents were kept sterile and added directly to sterile water. The medium was stored at 4°C and generally warmed to 37°C in a waterbath prior to use.

Subculture of adherent cell lines Cells were routinely grown in 75 cm^ surface plastic culture flasks (T75) in a humidified incubator with a gassing mixture of air (i.e., 21% oxygen) and 5% CO 2 and passaged every 2-4 days unless otherwise stated. When confluent, the cells were removed from the flask by trypsinization. The growth medium was discarded and the flask was rinsed in serum-free medium (10 ml). Trypsin-BDTA was added (10 ml) and cells were incubated for up to 5 min at 37°C, with the exception of SaP cells which were gently aggitated for a few minutes at room temperature. When cells had become dislodged from the flask, the trypsin was inactivated by the addition of 10 ml of medium containing 10% PCS. The resulting cell suspension was then centrifuged at 1,000 rpm for 5 minutes. The supernatant was discarded and the cell pellet resuspended by passing cells gently through a 21 G syringe three times before transfer of an appropriate aliquot to a fresh tissue culture flask.

Cell stocks Reserve stocks of all cell lines were kept in liquid nitrogen in case cells in use were transfected, showed abnormal growth characteristics or became infected. Cells were frozen by trypsinizing a confluent T75 flask of cells, neutralizing trypsin action by the addition of fresh medium containing PCS, then centrifuging at 1000 rpm for 5 minutes. Preezing vials were pre-cooled on ice and medium containing 10% (v/v) dimethyl sulfoxide (DMSO) was prepared. After centrifugation the supernatant was

224 discarded and the cell pellet was resuspended in 1 ml of medium. This suspension was vortexed and syringed three times through a 25 gauge needle to obtain a single cell suspension which was kept on ice and slowly diluted with medium containing 10% (v/v) DMSO. The suspension was then aliquoted into the freezing vials, 1.5 ml per vial. The vials were either placed in a Nalgene^^ Cryo 1°C freezing container filled with isopropyl alcohol (250 ml) for up to at least 4 h before liquid nitrogen storage or placed directly into the mouth of the liquid nitrogen cylinder for 12 h before immersing the vials, stacked in canes, directly into the liquid nitrogen. Cells remained viable in this frozen state indefinitely.

Recovery of frozen cells Cells were removed from liquid nitrogen and thawed rapidly by immersing the vial in a waterbath at 37°C. The cells were then diluted gradually with warm medium (up to 10 ml) and centrifuged at 1000 rpm for 5 minutes. The supernatant was discarded and the cell pellet resuspended in 5 ml of medium by vortexing. A single cell suspension was prepared and placed in a 25 cm^ surface flask for 12 h before changing medium.

Mycoplasma detection Cells (10^) were placed in antibiotic-free medium (0.5 ml) on a coverslip in a 6 cm petri dish and allowed to grow for 24-72 hours. The growth medium was not decanted but the cells were fixed with a mixture of methanol (100%) and glacial acetic acid (1:1) (5 ml). The fixative was decanted after 2 min and replaced by 5 ml of fresh fixative; this was repeated. The coverslip was left to air dry and was then stained with 0.05 pg/ml Hoechst 33258 (reconstituted in Hanks balanced salt buffer without phenol red, pH 7.0) for 10 min and then rinsed twice in distilled water. The coverslip was mounted in a solution of glycerol: PBS (1:1), sealed with nail polish to preserve, if necessary, and stored at 4°C.

V79-379A cell line V79 Chinese hamster lung fibroblasts were originally isolated in 1958 from lung tissue of mature male Chinese hamster, Cricetulus griseus (Ford & Yerganian, 1958). They were later subcloned and redesignated V79-1 (Elkind & Sutton, 1960). A subclone was donated to Dr L Revesz at the Karolinska Institute, Stockholm and an aliquot given

225 to Dr O.C.A. Scott at the Gray Laboratory in 1968 which is the V79-379A strain. This subline was recloned in 1972. V79 cells were grown in EMEM suspension medium supplemented with 7.5% PCS in a spinner flask at 37°C in a waterbath. Cells were subcultured twice weekly by the addition of fresh medium to an aliquot of cell suspension, thus they were maintained in the log phase of growth at a density of 2-5 x 10 cells/ml.

II. Western blotting solutions

12% Resolving gel (50 ml) ml Distilled water 16.5 30% acrylamide mix 20.0 1.5 M Tris (pH 8.8) 12.5 10% SDS 0.5 10% ammonium persulfate 0.5 TEMED 0.02

Running buffer (11) lOx Glycine 188 g Tris Buffer pH 8.3 30 g 10% SDS 50 ml Water 950 ml

Loading buffer (10 ml) 10% SDS 0.4 ml p-mereaptoethanol 0.5 ml 100 mM Tris HCl (pH 8.8) 121 mg 60 mM DTE 92.5 mg 0.5 M EDTA 0.8 ml 25% Glycerol 2.5 ml 0.1% w/v Bromophenol blue 1.0 ml Distilled water 4.8 ml

226 Transfer buffer (2.5 1) Glycine 28 g Tris 6 g Water 2 1 Methanol 500 ml 10% SDS 5 ml

Lysis buffer (20 ml) Distilled water 18.5 ml 1 M Magnesium chloride 1 ml Triton x-100 0.2 ml 100 mM PMSF 0.2 ml Aprotinin (10 mg/ml) 40 |J,1 Leupeptin (10 mg/ml) 1 |ll Pepstatin (1 mg/ml) 14 p.1 DNAse (10 mg/ml) 0.2 ml

Phosphate citrate buffer (pH 5.0) 0.2 M Sodium phosphate 25.7 ml 0.1 M Citric acid 24.3 ml Distilled water 50.0 ml

III. Details of transfected cell lines Cell lines were transfected by Dr G.J. Dougherty, University of British Columbia, Vancouver, British Columbia, Canada.

SaFJmTNF Sarcoma F cells originally derived at The Gray Laboratory were transfected with a mouse TNFa gene and were then selected for high TNFa expressing cells as described below.

227 SaFJneoII Sarcoma F cells from The Gray Laboratory stocks were transfected with a neomycin resistance gene, as described below.

Maintenance of transfected cell lines Transfected cell lines were grown for 6 passages only before returning to frozen stocks. They were grown in EMEM + 10% ECS, supplemented with geneticin (0.5 mg/ml, active weight). The geneticin ensures selectivity of the transfected cell line only, and cells which throw out the inserted gene cannot survive. Geneticin was reconstituted as a stock solution of 50 mg/ml. The pH was adjusted to pH 7 by the addition of a few drops of 10 mM sodium hydroxide, so that the colour changed from yellow to media- colour. Aliquots of 0.5 ml were stored at -20°C.

Isolation of a murine TNF-a cDNA Full length murine TNF-a cDNA was generated by RT-PCR. Briefly, total cellular RNA was isolated from Fsa-N tumors grown in a C3H/HeN mice as previously described (Chirgwin et al. 1979). cDNA was synthesized using the Pharmacia First Strand Synthesis Kit (Pharmacia, Baie d'Urfe, Quebec) and PGR was carried out exactly as described by the manufacturer using the following primer pair: 5' mTNF-a (5 -GGTCTAGACACCATGAGCACAG-3 ) and 3' mTNF-a (5-GGTCTAGAACACCCATTCCCTTCAC-3). Both primers contain added 5' Xbal restriction sites. The samples were placed in a thermal cycler (Biosycler, BIOS, New Haven, CT) and cycled 30 times at 95°C for 30 s, 55°C for 30 s, and 12°C for 1 min. PGR products were separated on a 1% agarose gel, purified using a Geneclean II Kit (BIO 101 Inc., Vista, GA), digested with Xbal and ligated into the Xbal site of pUG19. Double stranded templates were sequenced by the dideoxy chain termination method (Hattori and Sakaki, 1985), using T7 DNA polymerase and the reaction conditions suggested by the manufacturer (Sequenase Version 2.0 DNA Sequencing Kit, USB, Gleveland, OH).

Generation of retroviral vector JmTNF-a To construct the retroviral vector JmTNF-a, the full length murine TNF-a cDNA was first of all inserted into the Xbal site of the plasmid pTZ19R/Tk-neo. A

228 Smal-Hindni fragment containing both the TNF-a cDNA and Tk-neo was then isolated and cloned into Hpal-Hindlll cut Jzen.l (Laker et al. 1987). The plasmid obtained was transfected into the ecotropic packaging line GP+E-86 (Markowitz et al. 1988) by calcium phosphate precipitation and transfected cells selected using G418 (0.5 mg/ml active weight) (Gibco Life Technologies Inc., Grand Island, New York, USA).

Infection of SaF with JmTNF-a SaF cells were infected with the JmTNF-a virus or with control Jneo virus by incubating subconfluent monolayers of tumor cells with cell-free viral supernatants containing 5 mg/ml polybrene, as previously described (Hughs et al. 1992). Infected cells were selected and maintained in medium containing G418 (0.3 mg/ml active weight). To maintain any heterogeneity present within the starting population groups of at least 100 clones were pooled for further study.

229 IV. List of publications:

Lynch, E.M., Sampson, L.E., Khalil, A.A., Horsman, M R., and Chaplin, D.J. (1995) Cytotoxic effect of tumour necrosis factor-alpha on sarcoma F cells at tumour relevant oxygen tensions. Acta Oncol, 34 (3), 423-427.

Horsman, M.R., Khalil, A.A., Nordsmark, M., Siemann, D.W., Hill, S.A., Lynch, E.M., Chaplin, D.J., Stern, S., Thomas, C., Guichard, M., Grau, C., and Overgaard, J. (1995). The use of oxygen electrodes to predict radiobiological hypoxia in tumours. In Tumour Oxygenation, (eds P. Vaupel, D.K. Kelleher and M. Gunderoth) pp. 49-57. Stuttgart, Gustav Fisher Verlag.

Horsman, MR., Khalil, A.A., Siemann, D.W., Grau, C., Hill, S.A., Lynch, E.M., Chaplin, D.J., and Overgaard, J. (1994) Relationship between radiobiological hypoxia in tumors and electrode measurements of tumor oxygenation. Int J Radiat Oncol Biol Phys., 29 (3), 439-442.

2 3 0 Àcrû Oncologica Vol. 34, No. 3, pp. 423-427, 1995

CYTOTOXIC EFFECT OF TUMOUR NECROSIS FACTOR -ALPHA ON SARCOMA F CELLS AT TUMOUR RELEVANT OXYGEN TENSIONS

E il e e n M . L y n c h , L o u is e E. S a m p s o n , A z z a A. K h a l il , M ic h a e l R. H o r s m a n an d D a v id J. C h a p l in

We have investigated the response of a murine tumour cell line, the sarcoma F (SaF), to tumour necrosis factor-a (TNF) at oxygen tensions known to occur in vivo. Using the Eppendorf pOz histograph, the oxygen status of SaF tumours grown in situ was assessed. The median pOz of the SaF is less than 1% oxygen with over 90% of values at or below 15 mmHg ( <2% Oz). SaF cells primed in vitro for 24 h at tumour relevant oxygen tensions required at least four times more TNF to reduce cell number to 50% of controls following a 24 h incubation period in 21% oxygen. Chronic exposure of SaF cells to hypoxia during several passages increased resistance to TNF more than 60-fold. The oxygen sensitizing effect is transient as the resistance of hypoxic cells to TNF was reversed after 24 h incubation in air. These data clearly show that oxygen tension is a key modulator of the cytotoxic action of this important cytokine.

Tumour necrosis factor (TNF) is a pleiotropic cytokine Recent studies in our laboratory have shown that the with cytotoxic activity both in vitro and in vivo (I). It plays oxygen content of the cellular microenvironment prior to a pivotal role in tumour function affecting both tumour and treatment has a profound influence on the cytotoxic action endothelial cells and macrophages. However, the influence of TNF on L929 normal murine fibroblasts (5). of the tumour microenvironment on the activities of TNF The current investigations were undertaken to determine has not been elucidated. Such studies are needed to under­ the oxygen tension in a murine tumour, the sarcoma F stand the potential in vivo effects of this cytokine. Most in (SaF), in order to define the oxygen levels over which the vitro studies with TNF are performed in air (21% oxygen) cytotoxicity of TNF should be assessed in vitro and to (2) which is not representative of oxygen tensions found in determine whether these oxygen levels modulate the cyto­ vivo. It is well established that the majority of cells in toxic action of TNF on these tumour cells in vitro. experimental murine and xenografted human tumours exist in regions with oxygen tensions of 5% or less (3, 4). Material and Methods Oxygen tension is an important modulator of the cellular action of several cytokines and growth factors (5-11). Mice and tumours. pOj assessments were made on sar­ coma F (SaF) tumours grown in female CBA/Gy f TO mice. The SaF is a rapidly growing, anaplastic sarcoma Received 7 O ctober 1992. that arose spontaneously on the leg of a CBA/Ht mouse in Accepted 13 December 1994. 1957 in the Gray Laboratory mouse colony. The tumour From the Tumour Microcirculation Group, CRC Gray Labora­ has been maintained in the murine strain of origin by serial tory, Mount Vernon Hospital, North wood. England (E. M. Lynch, L. E. Sampson, D. J. Chaplin), and Danish Cancer transplantation as a crude cell suspension from a frozen Society, Department of Experimental Clinical Oncology, Aarhus, stock originally derived in 1978. Since, tumours have been Denmark (A. A. Khalil, M. R. Horsman). obtained from this frozen stock after 10 passages. A Correspondence to: Eileen M. Lynch, Tumour Microcirculation suspension of SaF tumour was prepared in saline and Group, CRC Gray Laboratory, PO Box 100, Mount Vemon approximately 5 x 10* cells were injected subcutaneously Hospital, Northwood, Middlesex, HA 6 2JR, England. Paper presented at the Tumour Microenvironment Workshop in into either the foot of the right hind limb (foot) or the rear Granada, Spain. September 23-25. 1994. dorsum (back) of 12 to 16-week-old female mice. Experi-

© Scandinavian University Press 1995. ISSN 0284-186X 423 424 E. M. LY N C H ET AL menls were carried out when the tumours had reached a Assessment of cytotoxicity. Cells were seeded in 96-well tumour volume of 200 mm^ (5.5-6 mm geometric mean tissue culture plates (Greiner, Laboratory Supplies, UK) diameter), which generally occurred 10-14 days after im­ at 1 X 10'* cells (0.1 ml) per well. Preincubation was car­ plantation. ried out under various oxygen tensions (95%, 21%, 5%, Tumour oxygenation measurements. Estimates of tu­ 2%, 1% and 0% oxygen, all with 5% CO;) for 24 h prior mour pOj were obtained using a One-need le autosensitive to incubation with TNF unless otherwise stated. Follow­ probe (Eppendorf, Hamburg, Germany). Non-anaes- ing this incubation period, during which the cells adhered thetized mice were restrained in Lucite jigs with the tu­ to the plastic, cells were treated with TNF. TNF was mour-bearing area exposed. The needle was inserted 1 mm added in the presence of I /ig/ml actinomycin D (AMD) into the tumour and then moved automatically through (Sigma) and the cells were then further incubated for 24 h the tissue in 0.7 mm increments, followed each time by a under conditions of 21% oxygen. Actinomycin D is a 0.3 mm backward step prior to measurement. Response metabolic inhibitor that enhances the cytotoxic action of time was 1.4 s. Repeated parallel insertions were per­ TNF in L929 murine fibroblasts (2). As the strain of SaF formed in each tumour until a total of 50-100 measure­ cells employed was not innately sensitive to TNF, actino­ ments was obtained. The relative frequency of the pO; mycin D was used to sensitize the cells to TNF. For each measurements was automatically calculated and displayed experiment control wells were included which were ex­ as a histogram. The median pOj values were calculated posed to identical oxygenation conditions in the absence from the raw data. of TNF. At the end of the 24 h incubation the growth Cells and culture conditions. Murine sarcoma F (SaF) medium containing TNF was discarded. The remaining cells, employed in these experiments, were derived from a adherent, viable cells, were then fixed in methanol (70%) single colony of SaF cells which were obtained by enzymatic and stained with crystal violet ( l%(w/v)). After staining digestion and culture of an SaF tumour. Cells were grown the plates were thoroughly washed, allowed to air dry and as a monolayer in Eagle’s minimal essential medium the bound crystal violet solubilized with glacial acetic acid (EMEM) (Gibco BRL, Scotland), supplemented with ( 33%). The absorbance of each well was read at 540 nm foetal calf serum (10%) (TCS Biologicals), glutamine using a microtitre plate reader (Labsystems Multiskan (2 mM), penicillin ( 125 units/ml, Crystapen 600 mg) and MCC/340). The amount of dye taken up is proportional streptomycin (100 //g/ml), Evans Medical Ltd) and buffered to the number of residual cells. We have previously deter­ to pH 7.4 with sodium bicarbonate. Cells were routinely mined that there is a linear relationship between optical grown in 75 cm’ surface plastic culture flasks in a humi­ density (O D ) and cell num ber up to 5 x 10^ cells per well. dified incubator with a gassing mixture of air (i.e. 21% The percentage of cell survival for a given TNF concen­ oxygen) and 5% CO; and passaged every 2-4 days unless tration was calculated as follows; otherwise stated. Oxygen conditions. Oxygen concentration was controlled Mean OD of TNF treated cells in Percentage presence of AMD at 21%, 5% or 2% in the incubator gas phase using the o f cell = ------X 100 combined CO 2 /N 2 /O 2 QUEUE QWJ-700 incubator which survival Mean OD of untreated cells in the is designed to electronically control O , tension as well as 5% presence of AMD CO; and temperature (37°C). Extreme oxygen levels of 1% or 95% oxygen and 95% nitrogen were achieved using The TNF titre is calculated from a semi-log plot of TNF premixed cylinders (BOC, UK) of 1% oxygen/5% CO; concentration (U/ml) versus OD (% cell survival) and is balance nitrogen, 95% oxygen/5% CO; or 95% nitrogen/5% the reciprocal of the TNF dilution at the point of 50% CO; respectively, connected to sealed containers at 37"C. cell survival (IC%). Only the linear portion of the curve is Containers were gassed at the required oxygen concentra­ used to estimate a titre for the TNF sample. This method tion for a total duration of 2 h to render the atmosphere has been used routinely by many investigators to assess stable. the cytotoxic effects of drugs and cytokines, including Tumour necrosis factor-alpha (TNF). Recombinant TNF (2, 7). It should be emphasised that survival deter­ human TNFa (International Standard 87/650) was pro­ mined using such a method may not necessarily be a vided by the National Institute of Biological Standards measure of clonogenicity. (NIBSC) as a sterile lyophilized powder. Each vial had a bioactivity of 40 000 lU and a protein content of 1 pg. Results TNF was reconstituted in a solution of saline/0.5% bovine serum albumin (Sigma, Dorset, UK) and further pO; histograms for SaF tumours grown on the backs diluted in culture medium. Human TNF binds to only the and feet of mice are illustrated in Fig. 1. The histograms p55 receptor on murine cells. However, there is strong represent the relative frequencies of the individual evidence to suggest that the cytotoxic action of TNF is classified pO; values. It can be seen from the data that the primarily transduced via the p55 receptor (12). median pO; of SaF tumours grown on the foot or back is TNh CYTOTOXICITY AND OXYGEN TENSION 425

5 0 . 7.4 mmHg and 3.2 mmHg respectively. The data indicate (A) that most of the cells are situated in an environment at or 40- less than 15 mmHg ( < 2% oxygen). Once the oxygen tension of the SaF tumour was deter­ 30- mined it was then possible to perform in vitro experiments 20 at tumour relevant oxygen tensions. The effect of incuba­ tion of SaF cells with TNF under various oxygen tensions 10 is show n in Fig. 2. SaF cells were ‘p n m e d ’ (in cu b ated ) N = 6 ; n = 2 9 1 0 under vanous oxygen tensions for 24 h before addition of L TNF. On addition of TNF, cells were re-incubated in air 0 25 50 75 100 (21% oxygen) for a further 24 h before staining. The data show that SaF cells that are sensitive to TNF in 21% (B) oxygen become less sensitive (resistant) to TNF when ^ 40- primed at or below 5% oxygen. The amount of TNF 30- required to reduce cell number to 50% of controls (IC 5 0 ) increases by a factor of at least four over this oxygen range

2 0 - (i.e. 21% to 1%). If cells were incubated under anoxic conditions prior to exposure to TNF in aerobic conditions N = 5 ; n = 3 1 8 no detectable cytotoxicity was evident even with TNF concentrations of lOOU/ml. This indicates that the resis­ 0 25 50 75 100 tance increased by a factor of greater than 500 compared to cells incubated at 2 1 % oxygen. Oxygen Partial Pressure (mmHg) We decided to investigate if SaF cells regain their sensi­ Fig. I. Histograms showing the oxygen profiles for Sa F tumours tivity to TNF after hypoxic priming. It can be seen from implanted s c. on the back (median pO^ = 3.24 mmHg) (A) and on the feet (median pO; = 7.36 mmHg) (B) of CB.A mice N = number of mice: n = total number of measurements made. 4

3

2

^ 0 .6 - 1 0 .4 -

0 0.0 0 1 3 5 7 24 Control 0 1 2 5 21 95 TNF Addition Time (h) Oxygen Tension (%) Fig. 3. Amount of TN F (plus .AMD) required to cause 50% cell Fig. 2. Amount of TNF required to cause 50% cell kill (ICju) in k ill ( I C 5 0 ) in SaF cells. Cells were incubated in 2% oxygen for Sa F cells in the presence of actinomycin D after 24 h pre-incuba- 24 h then placed in an atmosphere of 21% oxygen for periods up tion under various oxygen tensions prior to the addition of TNF to 24 h prior to the addition of TNF. Following addition of TN F in 21% oxygen. Results show the mean ± standard error of the cells were incubated at 21% oxygen for a further 24 h. Results mean (se) for 5 experiments. show mean + se for five experiments. 426 E M LYNCH ET AL.

14-1 TNF in the sarcoma F tumour cell line. Tumour cells cultured at oxygen tensions relevant to those found when they grow as solid tumours in vivo are more resistant to 1 2 - the cytotoxic effect of TNF compared to those grown under conventional tissue culture conditions (i.e. 21% oxy­ gen). The level of resistance seems to be dependent on both 1 0 - the level of oxygenation and the duration of exposure to a given level of oxygenation. For example, cells cultured for 24 h in an atmosphere of 2% oxygen are four times more k resistant to TNF compared to cells grown in an atmo­ Z sphere containing 21% oxygen, whereas cells grown for 5 H passages at 2% oxygen are 60 times more resistant. The increased resistance induced by a hypoxic environment is U completely reversed by placing the cells back in an envi­ ronment of 21% oxygen for 24 h prior to TNF exposure. A similar re-sensitization effect has been reported for L929 cells (5). However, in the present study significant re-sensi­ tization was achieved by re-exposing the cells to 21% oxygen for 3 h whereas with L929 cells no re-sensitization was observed at time points less than 5 h. It is the pretreat­ ment oxygenation status of the cells that determines the 2 3 4 5 response to TNF since no changes in response are ob­ Passage Number served if cells are gassed with between 1% and 21% oxygen Fig. 4. Effect o f chronic exposure to 2% oxygen on the sensitivity during the 24 h incubation with TNF (Lynch, unpublished of SaF cells to TN F. Cells were continually grown in 2% oxygen studies). and tested for TNF sensitivity after each passage (weekly). Results In the current study we have not investigated the mech­ are expressed as the mean ± standard error for each passage. anism responsible for the increased resistance to TNF mediated cytotoxicity in cells incubated at tumour relevent Fig. 2 that similar results are obtained over the range of oxygen tensions. However, several mechanisms can be 1% and 5% oxygen and as the majority of pOj values postulated. There is convincing evidence that reactive oxy­ were at or below 2% oxygen this was the oxygen tension gen species mediate, at least in part, the cytotoxic activities chosen for the following experiments. Cells were primed of TNF (6. 13, 14). Thus, if hypoxia induces the produc­ in 2% oxygen for 24 h, removed to air, and subsequently tion of proteins that protect the cell against such species, TNF was added at various intervals (0-24 h). It can be increased resistance would be evident. One such protein seen from Fig. 3 that re-sensitization does indeed occur. could be manganese superoxide dismutase (MnSOD), high Clearly, there is a time course of re-sensitization and levels of which are known to protect cells from TNF-in- by 24 h in 21% oxygen SaF cells are fully sensitive to duced cytotoxicity (15. 16). Endogenous TNF production TNF. has been reported to protect cells against exposure to Having determined that a 24 h exposure of SaF cells to exogenous TNF (17, 18). Interestingly, recent reports indi­ hypoxia causes a transient resistance to TNF that is re­ cate that incubating macrophages in an atmosphere con­ versible, we investigated whether exposure of SaF cells taining 2% oxygen can dramatically up-regulate the to hypoxia through several passages had the same effect. endogenous production of both TNF and its soluble recep­ Fig. 4 shows the effect of continuous culture of SaF cells tors (p55 and p75) (19). Another possible explanation is in 2% oxygen. TNF cytotoxicity assays were performed that receptor number, affinity, or rate of internalisation of after each passage (weekly) for up to 5 passages. The results the receptor-ligand complex is altered by prior exposure to show that cells after 5 passages in 2% oxygen require at low oxygen environments. Cell cycle has been shown to least 60 times m ore T N F to induce 50% cell kill than play a role in TNF cytotoxicity. Accumulation of cells in those exposed to air for 24 h. Thus, over the time course G1 induces resistance to TNF (20-23). However, our investigated, hypoxia induced resistance to TNF increases evidence to date indicates that incubation of cells in an significantly. environment of 2% oxygen compared to 21% oxygen does not induce significant changes in cell cycle parameters ( Lynch, unpublished studies). However, it is possible Discussion that some of the large protection observed following incu­ The present studies demonstrate that pretreatment oxy­ bation of cells under anoxic conditions is related to cell gen tension is a key modulator of the cytotoxic action of cycle. TNF CYTOTOXICITY AND OXYGEN TENSION 427

In summary, the results indicate that oxygenation condi­ 10. Wing DA, Talley CD. Storch TO. Oxygen concentration tions relevant to those found in solid tumour systems can regulates EGF-induccd proliferation and EGF-receptor down regulation. Biochem Biophys Res Commun 1988; 153: profoundly affect tumour cell sensitivity to the cytotoxic 952-8. action of TNF. The present study, and previous studies 11. Naldini A, Cesari S, Bocci V Effects o f hypoxia on the from other investigators (5-8, 11) emphasise the critical cytotoxicity mediated by tumour necrosis factor a. role tumour oxygenation can have on actions of certain Lymphokine and Cytokine Res 1994; 13: 233-7. cytokines. Further efforts are needed to determine in more 12. Tartaglia LA, Rothe M, Hu Y-F, Goeddel DV. Tumour detail how the tumour microenvironment alters the cellular necrosis factor's cytotoxic activity is signaled by the p55 TNF receptor. Cell 1993; 73: 213-6. effect of TNF and other cytokines. 13. Zimmerman RJ, Chan .4, Leadon SA. Oxidative damage in murine tumour cells treated in vitro by recombinant human tumour necrosis factor. Cancer Res 1989; 49; 1644-8. ACKNOWLEDGEMENTS 14. Yamauchi N, Kuriyama H, Watanabc N, Neda H, Maeda M, This investigation was supported by a grant from the Cancer Niitsu Y. Intracellular hydroxyl radical production induced Research Campaign and the Danish Cancer Society. EM L is by recombinant tumour necrosis factor and its implication funded by a Scott of Yews scholarship. in the killing of tumour cells in vitro. Cancer Res 1989; 49: 1671-5. 15. Wong CHW . Elwell JH, Oberley LW, Geoddel DV. Manganous superoxide dismutase is essential for cellular resis­ REFERENCES tance, to cytotoxicity of tumour necrosis factor. Cell 1989; 58: 1. Carswell EA, Old U, Kassel RL, Green S, Fiore N, 923-31. Williamson B. An endotoxin-induced serum factor that causes 16. Zyad A, Benard J, Tursz T, Clarke R. Chouaib S. Resistance necrosis of tumours. Proc Natl Acad Sci USA 1975; 72: to TNFa and Adriamycin in the human breast cancer MCF-7 3666-70. cell line: relationship to M D RI, MnSOD and TN F gene 2. Meager A, Leung H, Wooley J. Assays for tumour necrosis expression. Cancer Res 1994; 54: 825 - 31. factor and related cytokines. J. Immunol Methods 1988; 116; 17. Vanhaesebroeck B, Dccoster E. Van Ostade X, et al. Expres­ 1-17. sion of an exogenous tumour necrosis factor (TNF) gene in 3. Lartigau E. Le Ridant AM, Lambin P, et al. Oxygenation of T N F sensitive cell lines confers resistance to T N F mediated head and neck tumors. Cancer 1993; 71: 2319-25. cell lysis. J Immunol 1992; 148: 2785-94. 4. Vaupel P, Schlenger K. Knoop C, Hoeckel M. Oxygenation 18. Himeno T, Watanabe N, Yamauchi N, et al. Expression of human tumours: Evaluation of tissue oxygen distribution of endogenous TNF as a protective protein against the in breast cancers by computerised O^ tension measurements. cytotoxicity o f exogenous TN F Cancer Res 1990; 50: 4941 - Cancer Res 1991; 51: 3316-22. 5. 5. Sampson LE, Chaplin DJ. The influence of microenvironment 19. Scannel G, Waxman K. KaemI GJ, et al. Hypoxia induces a on the cytotoxicity of TNFa in vitro. Int J Radiat Oncol Biol human macrophage cell line to release tum our necrosis factor Phys 1994; 29: 467-71. a and its soluble receptors in vitro. J. Surg Res 1993; 54: 6. Park YMK, Anderson RL, Spitz DR, Hahn GM. Hypoxia 281-5. and resistance to hydrogen pieroxide confer resistance to tu­ 20. Belizario JE, Dinarello CA. Interleukin I, interleukin 6, tu­ mour necrosis factor in murine L929 cells. Radiat Res 1992; mour necrosis factor and transforming growth factor-^ in­ 131: 162-8. crease cell resistance to tumour necrosis factor cytotoxicity by 7. Matthews N, Neale ML. Jackson SK, Stark JM. Tum our cell growth arrest in the G I phase of the cell cycle. Cancer Res killing by tumour necrosis factor: inhibition by anaerobic 1991; 51. 2379-85. conditions, free radical scavengers and inhibitors of arachi- 21. W atanabe N, Niitsu Y. Yamauchi N. et al. Cell cycle specifi­ donate metabolism. Immunology 1987; 62: 153-5. city o f tum our necrosis factor and its receptor. Cell Biol Int 8. LoefBer DA, Juneau PL, Masserant S. Influence of tum our Reports 1987; II: 813-7. physio-chemical conditions in interleukin-2-stimulated lym­ 22. Coffman FD, Haviland DL, Green LM, Ware CF. Cytotoxic­ phocyte proliferation. Br J Cancer 1992; 66: 619-22. ity by tumor necrosis factor is linked with the cell cycle but 9. Aune TM, Pogue S. Inhibition of tumour cell growth by does not require DNA synthesis. Growth factors 1989; I: Interferon-'/ is mediated by two distinct mechanisms, depen­ 357-64. dent on tumour oxygenation, induction of tryptophan degra­ 23. Darzynkiewicz Z, Williamson B. Carswell EA, Old LJ. Cell dation and depletion of intracellular nicotinamide adenine cycle specific effects of tumor necrosis factor. Cancer Res dinucleotide. J Clin Invest 1989; 84: 863-75. 1984; 44: 83-90. Int. J. Radiation Oncology Biol. Phys., Vol. 29, No. 3, pp. 439-442, 1994 Pcrgamon C opyright @ 1994 Elsevier Science Ltd Printed in the USA. All rights reserved 0360-3016/94 $6.00+ .00 0360-3016(93)E0085-K

# Manipulation of Tumor Oxygen and Physiology

RELATIONSHIP BETWEEN RADIOBIOLOGICAL HYPOXIA IN TUMORS AND ELECTRODE MEASUREMENTS OF TUMOR OXYGENATION

M ichael R. Horsman, Ph.D.,* Azza A. K halil, M.D.,* Dietm ar W. Siemann, Ph.D.,^ Cai Grau, M.D.,* S ally A. H ill, Ph.D.,* Eileen M. Lynch, B.Sc.,* D a v id J. Chaplin, Ph.D.* and Jens O vergaard, M.D.*

•Danish Cancer Society, Department of Experimental Clinical Oncology, Denmark; ^Tumor Biology Division, University of Rochester Cancer Center, NY; and *CRC Gray Laboratory, Mount Vernon Hospital, Northwood, Middlesex, England

Purpose; To determine whether electrode measurements of tumor oxygenation, made in a variety of murine tumor models, correlate with estimates of radiobiological hypoxia in the same tumor systems. Methods and Materials: The tumor models used were a C3H mammary carcinoma grown in the feet of CDFl mice; the SCCVII, KHT and RIF-1 tumors grown in the feet or flanks of C3H/Km mice; and the CaNT and SaF tumors grown on the backs of CBA mice. All treatments were performed when tumors were about 200 mm^ in size. Radiobiological hypoxic fractions were determined using either a paired survival curve assay, with survival measured 0-24 h after irradiation, or using a clamped tumor control assay, with percent local tumor control estimated 90 days after treatment. Measurements of tumor oxygen partial pressure (pO%) distributions were performed using Eppendorf oxygen electrodes. Results: The hypoxic fractions determined from the radiation response data were about 1% in RIF-1 and SCCVll, 12% in C3H and KHT, 28% in CaNT and up to 38% in SaF tumors. When this data was compared with the tumor oxygenation measurements it was found that as hypoxic fraction increased the mean, median, and the percentage of pO] values ^ 5 mmHg showed a trend towards poorer oxygenation status. However, none of these pO% changes were significantly correlated with hypoxia. Moreover, the pOy values ^ 2.5 mmHg indicated an improvement in oxygen status with increasing hypoxic fraction. Conclusion: Electrode measurements of tumor oxygenation alone may, therefore, not be a good indicator of tumor hypoxia across different tumor cell lines.

Murine tumors. Radiobiological hypoxia, Eppendorf oxygen electrodes. Tumor oxygenation status.

INTRODUCTION ical tum or hypoxia (8, 9, II, 23-25), the real predictive value of this procedure has not been fully established. The The presence of hypoxic cells in human tumors is believed aim of the current study was to use an Eppendorf oxygen to be one of the major reasons for failure to locally control electrode system to determine the oxygenation status of certain tumor types with conventional radiation therapy a variety of murine tumor models and to see if these results (17). Numerous attempts have now been made to try and correlated with estimates of radiobiological hypoxia made identify those human tumors that contain hypoxic cells in the same tumor systems. and therefore, most likely to benefit from therapies which can overcome hypoxia (7). METHODS AND MATERIALS Several clinical studies have indicated that electrode measurements of oxygen partial pressure (pO]) distribu­ Tumor models tions in tumors may be a reliable method for identifying The tumor models used in this study were a C3H hypoxia (1 ,4 , 10, 13, 18). Although a num ber o f exper­ mammary carcinoma grown in the feet of CDFl mice, imental studies have strongly suggested the existence of a an SCCVll tumor grown in the feet of C3H/Km mice, correlation between pOz measurements and radiobiolog­ KHT and RIF-1 tumors both grown in the flanks of C3H/

Presented at the Eighth International Congress on Chemical Baden for expert technical assistance, and to Ms L. Spliid for Modifiers of Cancer Treatment, Kyoto, Japan, 16-19 June 1993. preparation of the manuscript. This investigation was supported Reprint requests to: Michael R. Horsman, Ph.D., Danish by a grant from the Danish Cancer Society, the British Cancer Cancer Society, Department of Experimental Clinical Oncology, Research Campaign, the American NIH (grant number CA Norrebrogade 44, BIdg. 5, DK-8000 Aarhus C, Denmark. 55300) and “Fonden til Fremme af Dansk Radiologi.” Acknowledgements —The authors would like to express their Accepted for publication 24 November 1993. thanks to Ms. 1. M. Johansen, Ms 1. M. Thuesen and Ms A.

439 440 I. J. Kadiuiion Oncology • Biology • Physics Volume 29, Number 3. 1994

Km mice, and the CaNT and SaF grown on the backs of 10 CBA mice. Derivation and maintenance of all these tu­ o • mors have been described previously (3, 5. 6. 12, 16, 21). 10' All treatments were carried out when tumors had reached 1 50-250 mm^ in size. 10

Radiation treatments 10 Irradiations were given either locally to the tumors, us­ ing 250 kV X rays at a dose rate of 2.3 Gy/min (C3H, 10 O'Qy SCCVII) or 240 kV X rays at a dose rate of 3.9 Gy/min (CaNT, SaF), or whole body using '^’Ce with a dose rate 10 of 4.3 Gy/min (KHT, RIF-1). Foot tumors were made 0 10 20 30 40 totally hypoxic by clamping the tumor-bearing leg with a RADIATION DOSE (G y) rubber tube 5 min before and during the irradiation. For Fig. I. Survival of SCCVII tum or cells irradiated in vivo under the other tumors, hypoxia was achieved by killing the either normal air breathing (O) or clamped (•) conditions. Cell mice (cervical dislocation) 5-10 min before irradiating. surviving fraction was determined by plating cells in vitro 24 h after irradiating. Lines are the best parallel fit for radiation doses of 10 Gy and above. Points represent individual tumors. Hypoxic fraction determinations For all tumors, except the C3H line, the hypoxic frac­ tion was determined using a paired survival curve pro­ breathing conditions results in a typical survival curve cedure. This involved measuring tumor cell survival with showing both a shoulder and a hypoxic tail (Fig. I). an in vivo/in vitro colony forming assay (20), either im­ Clamping tumors before and during irradiation radiopro­ mediately or 24 h after giving a range of radiation doses tects the tumor and the vertical displacement between the under normal air breathing or hypoxic conditions. The parallel regions of the aerobic and hypoxic curves gives a hypoxic fraction was then calculated from the vertical hypoxic fraction of < 1%. The pOz profile measured in distance between the aerobic and hypoxic survival curves identical SCCVII tumors using an oxygen electrode is (22). Since the C3H tumor does not grow in vitro, the shown in Figure 2. Some 37% of the measurements ob­ hypoxic fraction was estimated with a clamped tumor tained were ^2.5 mmHg, while over 50% of the values were control assay. For this, tumors were irradiated under ^ 5 mmHg. The mean and median pOz values were 7.5 clamped or undamped conditions and the percentage of and 4.5 mmHg, respectively. animals showing local tumor control 90 days later re­ The results taken from Figures I and 2 for the SCCVII corded. The hypoxic fraction was then determined from tumor and those obtained with all the other tumor models, direct analysis of the radiation dose-response data (2). are summarized in Figure 3. It was obvious from the mean and median pOz results that the C3H tumor gave results Tumor oxygenation measurements that were not comparable to those of the other tumors. Estimates of tumor pO; were obtained using a fine nee­ dle autosensitive electrode probe. ' The needle was inserted up to a depth of 1 mm into the tumor and then moved automatically through the tissue in 0.7 mm increments, followed each time by a 0.3 mm backward step prior to measurement. Response time was 1.4 s and repeated par­ allel insertions were performed in each tumor until a total of 50-100 measurements were obtained. The relative fre­ quency of the pO] measurements was automatically cal­ culated and displayed as a histogram, although the pO: parameters we used in this study, namely the percentage of values ^ 2.5 or 5 mmHg, and the mean and median pOz were actually calculated from the raw data. mTTTM 0 5 10 15 20 25 30 35 RESULTS OXYGEN PRESSURE (m m H g)

Figures 1 and 2 show representative examples of both Fig. 2. Histogram showing the oxygenation profile for SCCVII tumors. Measurements were made using an Eppendorf oxygen the radiation dose response curves and pO; histograms electrode under normal air breathing conditions. Results show obtained in this study for one of the tumor models, the the relative frequency of each tumor oxygen pressure obtained SCCVII. Irradiating these tumors under normal air from a total of 381 measurements in six tumors.

Eppendorf, Hamburg, Germany. Hypoxia in tumors • M. R. H o r s m a n c l a l. 441

16 12

14 to

12 a

10 6

5 4

6 2

4 0

60

&

f

: I I

0 10 20 30 40 50 0 10 20 30 40 50 HYPOXIC FRACTION (%) HYPOXIC FRACTION (%)

Fig. 3. Relationship between tumor oxygenation status and radiobiological hypoxia measured in a variety of murine tumor models. Results were obtained from C3H (•), SCCVII (O), RIF- I (A), KHT (A), SaF (■), and CaNT (□) tumors. Points show means (± I SE) with the lines fitted by regression analysis using all the tumors except the C3H mammary carcinoma.

Since the hypoxic fraction for this tumor was calculated increase in tumor hypoxia (25). More recently, we re­ using a local control assay and not a paired survival curve ported that by injecting CDFl mice with different drugs assay as used with all the other tumors, we therefore or by allowing these animals to breathe different gas mix­ elected not to include the C3H data in the final analysis. tures, we could not only substantially change the hypoxic For the remaining five tumor models there are obvious fraction in a C3H mammary carcinoma, but these mod­ trends between the pOz data and the hypoxic fractions, ifications of hypoxia were significantly correlated to in that as hypoxia increased so the mean and median pOi changes in tumor oxygenation measured with the Eppen­ decreased, while the percentage of values ^ 5 mmHg in­ dorf electrode (8, 9). creased. With the pOz values ^ 2.5 mmHg there was an In our current study, we found that some of the results apparent decrease with increasing hypoxia. The correla­ with the C3H mammary carcinoma did not actually fit tion coefficients for the regression lines fitted to these data with those obtained with the other tumor models. The were calculated to be 0.88, 0.49, 0.85, and 0.61, for the hypoxic fraction in thisC3H tumor was determined using mean, median, pOz < 5 and 2.5 mmHg, respectively. a clamped local tumor control assay and not the paired survival curve assay, as was used with all the other tumors. Since it is known that different assay procedures in the DISCUSSION same tumor model can result in vastly different levels of Previous experimental studies have always strongly radiobiological hypoxia (14. 15, 19), the inconsistency be­ suggested that pOz measurements with electrodes are a tween the C3H tumor results and all the other tumors is good indicator of clonogenic hypoxic cells in tumors. not totally unexpected. However, even when we removed Vaupel and colleagues (11, 23-25) reported that as tumor the C3H data from our analysis, although we observed size increased, oxygenation status decreased. Furthermore, trends between the pOz measurements and hypoxic frac­ with the FSall tumor they found that the decrease in mean tion, there were no significant correlations in the data. pOz with increasing tumor size correlated well with the This could be because our study only contained a few 442 1. J. Radiation Oncology # Biology # Physics Volume 29, Number 3, 1994

tumor models; we are now trying to obtain data with these oxia. The tumor sites have included cervix (4, 10), head and other tumors grown in different sites and treated at and neck (I, 4). breast (I, 18), lung, bladder, and lym­ different volumes. But, the failure to see a correlation may phomas (I). In general, the results showed that those tu­ simply be that while pO; and hypoxia can be correlated mors that were the least responsive to radiation therapy within tumor cell lines (8,9, 25), no such correlation exists were in fact those that had the lowest level of oxygenation. across tumor cell lines. This is perhaps not surprising be­ But, the failure to show a significant correlation between cause with the pOz electrode low oxygen measurements pO] measurements and radiobiological hypoxia in our may be associated with areas which do not necessarily current study clearly suggests that additional experimental contain viable hypoxic cells. studies need to be performed before such electrode mea­ Several limited clinical studies have attempted to find surements can be categorically used as a predictive assay a correlation between pOz measurements and tumor hyp­ of tumor sensitivity to radiation therapy.

REFERENCES

1. Badib, A. O.; Webster, J. H. Changes in tumor oxygen ten­ 13. Kolstad, P. Intercapillary distance, oxygen tension and local sion during radiation therapy. Acta Radiol. Ther. Phys. Biol. recurrence in cervix cancer. Scand. J. Clin. Lab. Invest. 106: 8:247-257; 1969. 145-157; 1968. 2. Bentzen, S. M.; Grau, C. Direct estimation of the fraction 14. McNally, N. J.; Sheldon, P. W. The effect of radiation on of hypoxic cells from tumour-control data obtained under tum our growth delay, cell survival, and cure of the animal aerobic and clamped conditions. Int. J. Radiat. Biol. 59: using a single tumour system. Br. J. Radiol. 50:321-328; 1435-1440;1991. 1977. 3. Denekamp, J.; Hill, S. A.; Hobson, B. Vascular occlusion 15. Moulder, J. E.; Martin, D. P. Hypoxic fraction determi­ and tumour cell death. Eur. J. Cancer Clin. Oncol. 19:271- nations in the BA 1112 rat sarcoma: Variations within and 275; 1983. among assay techniques. Radiat. Res. 98:536-548; 1984. 4. Gatenby, R. A.; Kessler, H. B.; Rosenblaum, J. S.; Coia, 16. Overgaard, J. Simultaneous and sequential hyperthermia L. R.; Moldofsky, P. J.; Hartz, W. H.; Broder, G. J. Oxygen and radiation treatment of an experimental tumor and its distribution in squamous cell carcinoma métastasés and its surrounding normal tissue in vivo. Int. J. Radiat. Oncol. relationship to outcome of radiation therapy. Int. J. Radiat. Biol. Phys. 6:1507-1517; 1980. Oncol. Biol. Phys. 14:831-838; 1988. 17. Overgaard, J. Sensitization of hypoxic tum our cells—clinical 5. Hill, S. A.; Fowler, J. P.; M inchinton, A. 1.; Stratford, experience. Int. J. Radiat. Biol. 56:801-811; 1989. M. R. L.; Denekamp, J. Radiosensitization of a mouse tu­ 18. Pappova, N.; Siracka, E.; Vacek, A.; Durkovsky, J. Oxygen mour by Ro-03-8799 acute and protracted administration. tension and prediction of the radiation response. Polaro- Int. J. Radiat. Biol. 44:143-150; 1983. graphic study in human breast cancer. Neoplasma 29:669- 6. Hirst, D. G.; Brown, J. M.; Hazlehurst, J. L. Enhancement 674; 1982. of CCNU cytotoxicity by misonidazole: Possible therapeutic 19. Siemann, D. W. Tum our size: A factor influencing the iso­ gain. Br. J. Cancer 46:109-116; 1982. effect analysis of tumour response to combined modalities. 7. Horsman, M. R. Hypoxia in tumours: Its relevance, iden­ Br. J. Cancer 41(Suppl. lV):294-298: 1980. tification and modification. In: Beck-Bomholt, H. P., ed. 20. Siemann, D. W. Enhancement of chemotherapy and nitro- Current Topics in Clinical Radiobiology of Tumours. Hei­ imidazole-induced chemopotentiation by the vasoactive delberg: Springer-Verlag; 1993:92-112. agent hydralazine. Br. J. Cancer 62:348-353; 1990. 8. Horsman, M. R.; Khalil, A.; Nordsmark, M.; Grau, C.; 21. Twentyman, P. R.; Brown, J. M., Gray, J. W.; Fran ko, A. J.; Overgaard, J. Measurement of pO; in a murine tumour and Scoles, M. A.; Kallman. R. P. A new mouse tumor model its correlation with hypoxic fraction. Adv. Exp. Biol. Med. system (RlP-1) for comparison end-point studies. JNCI 64: (In press). 595-604; 1980. 9. Horsman, M. R.; Khalil, A.; Nordsmark, M.; Grau, G.; 22. Van Putten, L. M.: Kallman, R. P. Oxygenation status of Overgaard, J. Relationship between radiobiological hypoxia a transplantable tumour during fractionated radiation ther­ and direct estimates of tumour oxygenation in a mouse tu­ apy. JNCl 40:441-451; 1968. mour model. Radiother. Oncol. 28:69-71; 1993. 23. Vaupel, P. Oxygen supply to malignant tumors. In: Peterson, 10. Hockel, M.; Knoop, C.; Schlenger, K.; Vomdran, B.; Mitz, H. 1., ed. Tumor Blood Circulation: Angiogenesis, Vascular M.; Knapstein, P. G.; Vaupel, P. Intratumoral pO^ predicts Morphology and Blood Plow of Experimental and Human survival in advanced cancer of the uterine cervix. Radiother. Tumors. Boca, Raton, PL; CRC Press Inc.; 1979:143-168. Oncol. 26:45-50; 1993. 24. Vaupel, P.; Portmeyer, H. P.; Runkel, S.: Kallinowski, P. 11. Kallinowski, P.; 2Lander, R.; Hockel, M.; Vaupel, P. Tum or Blood flow, oxygen consumption and tissue oxygenation of tissue oxygenation as evaluated by computerised—pO:— human breast cancer xenografts in nude mice. Cancer Res. histography. Int. J. Radiat. Oncol. Biol. Phys. 19:953-961; 47:3496-3505; 1987. 1990. 25. Vaupel, P.; Okunieff, P.; Kallinowski, P.; Neuringer, L. K. 12. Kallman, R. P.; Silini, G.; van Putten, L. M. Factors influ­ Correlations between ^'P-NMR spectroscopy and tissue O; encing the quantitative estimation of the in vivo survival of tension measurements in a murine fibrosarcoma. Radiat. cells from solid tumors. JN Cl 39:539-549; 1967. Res. 120:477-493; 1989. Corrigenda: to be inserted on page 80 after line 15: CKIs provide a crucial mechanism by which diverse agents, DNA damage and oxidative stress exert cell cycle suppression: S phase entry, and DNA replication, is prevented by a CKI called p21 (also known as WAFl or CIPl) (Nigg, 1995). Other negative regulatory factors of the cell cycle include: p53 and its variant protein p53as, the retinoblastoma protein (pRb), proliferating cell nuclear antigen (PCNA) and replication protein A (RPA) (Karp and Broder, 1995; Lohrer, 1996). p53, leads to cell cycle arrest in Gl, at the G 1/S boundary and in S phase. Premature Gl/S transition is also prevented by pRb, whereas, DNA replication is blocked by PCNA and RPA. p53 and pRb are regulated by CKIs, typically (though not always) p53 induces p21, and pRb is dependent on pl6 (reviewed by Cox and Lane, 1995). It has recently been suggested that Rb, p53 and p21 proteins mediate TNF antimitogenic activity (Jeoung gr a/. 1995).

Additional references:

Browning, J.L., Ngam-ek, A., Lawton, P., De Marinis, J., Tizard, R., Pingchang Chow, E., Hession, C., O’Brien-Grenco, B., Foley, S.F., and Ware, C.F. (1993) Lymphotoxin p, a novel member of the TNF family that forms a heteromeric complex with lymphotoxin on the cell surface. Cell, 72, 847-856.

Cox, L.S., and Lane, D.P. (1995) Tumour suppressors, kinases and clamps: how p53 regulates the cell cycle in response to DNA damage. Bioessays., 17, 501-508.

Crowe, P.D., VanArsdale, T.L., Walter, B.N., Ware, C.F., Hession, C., Ehrenfels, B., Browning, J.L., Din, W.S., Goodwin, R.G., and Smith, C.A. (1994) A lymphotoxin-p-specific receptor. Science., 264,707-710.

Graeber, T.G., Peterson, I.E., Tsai, M., Monica, K., Fomace, A.J., and Giaccia, A.J. (1994) Hypoxia induces accumulation of p53 protein, but activation of a G1-phase checkpoint by low-oxygen conditions is independent of p53 status. Molecular and Cellular Biol, 14, 6264- 6277.

Jeoung, D-i., Tang, B., and Sonenberg, M. (1995) Effects of tumor necrosis factor-a on antimitogenicity and cell cycle-related proteins in MCF-7 cells. J Biol Chem., 270, 18367- 18373.

Karp, J.E, and Broder, S. (1995) Molecular foundations of cancer: New targets for intervention. Nature medicine.,1, 309-320.

Lohrer, H.D. (1996) Regulation of the cell cycle following DNA damage in normal and Ataxia telangiectasia cells. Experientia., 52, 316-328.

Nigg, E.A. (1995) Cyclin-dependent protein kinases: Key regulators of the eukaryotic cell cycle. Bioessays., 17,471-480.