ANTISENSE OLIGONUCLEOTIDES AS THERAPEUTICS FOR THE TREATMENT OF AGGRESSIVE BREAST CANCER B y Dawn N . Waterhouse B.Sc. Simon Fraser University, 1995 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y In T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Department of Pathology and Laboratory Medicine) University of British Columbia November, 2001 © Dawn N . Waterhouse, 2001 We accept this thesis as conforming to the required standard U B C Special Collections - Thesis Authorisation Form Page 1 of 1 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the Un i v e r s i t y of B r i t i s h Columbia, I agree that the Lib r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r reference and study. I further agree that permission for extensive copying of th i s thesis f o r s c h o l a r l y purposes may be granted by the head of my department or by his or her representatives. It i s understood that copying or p u b l i c a t i o n of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of P n - H v n l ^ i j & LcJooccctcXy foeA^'Cme The U n i v e r s i t y of B r i t i s h Columbia Vancouver, Canada http://www.library.ubc.ca/spcoll/thesauth.html 3/11/02 ABSTRACT There are many drugs currently available for the treatment of aggressive breast cancer. These include anthracyclines, taxanes, alkylating agents, anti-metabolites, plant alkaloids, nucleoside analogues and anti-hormonal agents. Unfortunately, even armed with this impressive arsenal, there has been little ground gained in terms of disease free years or reduced mortality for what is an essentially incurable disease in the metastatic state. Clearly, we need to improve upon the therapies available for these patients. A key step toward this goal is the development of reproducible and relevant models in which newly developed drugs may be tested. This thesis outlines the characterisation of what is anticipated to be a powerful human xenograft model of aggressive breast cancer, that of the M D A 4 3 5 / L C C 6 cell line. This cell line may be grown easily in vitro, as well as an ascitic or as a solid tumour in mice. In order to have a major impact in the field of breast cancer treatment, it w i l l not suffice to develop yet another cytotoxic agent. Instead, we must turn to the newer technologies, including gene targeted therapies, which target the molecular root of the disease. This work includes the use of both free antisense oligonucleotides (ODN), as well as those formulated within a lipid carrier. These encapsulated O D N are retained in the circulation for a longer period of time, are less susceptible to the actions of nucleases, and due to pharmacokinetic and pharmacodistribution properties of the liposomal carrier, result in enhanced tumour cell uptake of the O D N . u Finally, O D N , both free and liposome encapsulated, were administered to tumour bearing female SCID/Rag2m mice, either singly or in combination with a commonly used anti-cancer agent (doxorubicin). It is shown that O D N are capable of mediating the specific down-regulation of the target protein as well as impacting the rate of tumour growth. It is the intention of the author to demonstrate the necessity for not only good models and newly developed and specifically targeted therapeutic agents, but that we must also consider the use of combination strategies in the treatment of aggressive breast cancer such that current mortality statistics may be improved. 111 TABLE OF CONTENTS Abstract i i Table of Contents iv List of Tables v i i List of Figures vi i i Abbreviations x i Acknowledgements x i i i Dedication xiv Chapter 1: Introduction 1 1.1 Breast Cancer 1 1.2 Prognostic and Predictive Factors in Breast Cancer 4 1.2.1 Gross and Histopathological Factors 5 1.2.2 Molecular Factors 6 1.2.2.1 CathepsinD 6 1.2.2.2 Vascular Endothelial Growth Factor 7 1.2.2.3 HER-2/neu 10 1.2.2.3.1 HER-2/neu Signalling Pathways 16 1.2.2.3.2 HER-2/neu in Normal Physiology 18 1.2.2.3.3 HER-2/neu Expression in Cancer 20 1.2.2.3.3.1 HER-2/neu in Breast Cancer 21 1.3 Therapeutic Strategies in the Treatment of Breast Cancer 22 1.3.1 Chemotherapeutics 22 1.3.2 Hormonal Strategies 24 1.3.3 Molecular Targeting 26 1.3.3.1 Antibodies 26 1.3.4 Gene Therapy 28 1.3.5 Antisense Oligonucleotides 28 1.3.5.1 Types of antisense oligonucleotides 29 1.3.5.2 Mechanisms of action 32 1.3.5.3 Pharmacokinetics and Toxicology 35 1.3.5.4 ODNs Targeting HER-2/neu 37 1.4 Liposomes 39 1.4.1 Liposomes as Drug Delivery Vehicles 40 1.4.2 Liposomes as Drug Delivery Vehicles in the Treatment of Solid Tumours, in Particular, Breast Cancer 41 1.4.2.1 Caelyx (Doxil) 46 1.4.2.2 Myocet 48 1.4.3 Liposomal Targeting 50 1.4.3.1 Liposomal Oligonucleotides 51 1.5 Thesis Objectives 52 1.5.1 Hypotheses 53 1.5.2 Summary of Novel Research and Impact in the Field of Breast Cancer Research. 53 IV Chapter 2: Materials and Methods 55 2.1 Chemicals and Oligonucleotides 55 2.2 Cel l Culture 56 2.3 In Vitro Growth Characterisation 57 2.4 In Vitro Cytotoxicity 58 2.5 In Vitro Oligonucleotide Delivery 58 2.6 In Vitro Oligonucleotide Mediated Downregulation of HER-2/neu 59 2.6.1 Western Blotting to Assess HER-2/neu Protein Levels 60 2.6.2 Flow Cytometric Analysis 62 2.7 In Vitro Oligonucleotide Mediated Downregulation of V E G F 63 2.7.1 Determination of V E G F Levels 64 2.8 Liposomal Oligonucleotide Formulation 64 2.9 Liposomal Oligonucleotide Characterisation 67 2.9.1 Size 67 2.9.2 Antisense to Lip id Ratio 68 2.9.3 Oligonucleotide Integrity 68 2.10 Liposomal Oligonucleotide Stability Over Time at 4°C, 21°C and 37°C 69 2.11 Immunohistochemical Visualisation of V E G F , p i85 , bcl-2 and Cathepsin D in Solid Tumour Sections 70 2.12 In Vivo Growth Characterisation of M D A - M B - 4 3 5 and M D A - M B - 4 3 5 H E R 2 Human Breast Carcinoma Cells 72 2.13 In Vivo Growth Characterisation of M D A 4 3 5 / L C C 6 and M D A 4 3 5 / L C C 6 H E R 2 Human Breast Carcinoma Cells 72 2.13.1 Solid Tumours 72 2.13.2 Ascites Tumours 73 2.14 Oligonucleotide Pharmacokinetics 73 2.14.1 Plasma Elimination of ant i -VEGF and anti-HER-2/neu Oligonucleotides 73 2.14.2 Tissue Distribution of ant i -VEGF and anti-HER-2/neu Oligonucleotides in Tumour-Free Mice 74 2.14.3 Peritoneal Fluid Oligonucleotide Accumulation and Tumour Cel l Association in M D A 4 3 5 / L C C 6 Ascites Tumour-Bearing Mice 75 2.15 In Vivo Efficacy of A n t i - V E G F Oligonucleotides in M D A 4 3 5 / L C C 6 Ascites Tumour Bearing Mice 76 2.16 In Vivo Efficacy of Anti-HER-2/neu Oligonucleotides in M D A - M B - 4 3 5 and M D A - M B - 4 3 5 H E R 2 Solid Tumour Xenografts 76 2.17 In Vivo Efficacy of Anti-HER-2/neu Oligonucleotides in M D A 4 3 5 / L C C 6 Ascites Tumour-Bearing Mice 77 2.18 Statistical Analysis 78 Chapter 3: Anti-HER-2/neu Antisense Oligonucleotides, Alone or in Combination with Conventional Cytotoxic drugs, in the Treatment of the MDA-MB-435 and MDA-MB-435HER2 Xenograft Models of Human Breast Cancer 79 3.1 Introduction 79 3.2 Results 82 3.2.1 In Vitro Growth of M D A - M B - 4 3 5 and M D A - M B - 4 3 5 H E R 2 Cells 82 3.2.2 In Vivo Growth of M D A - M B - 4 3 5 and M D A - M B - 4 3 5 H E R 2 Cells 82 v 3.2.3 pi85 Expression in Solid MDA-MB-435 and MDA-MB-435 Tumours 84 3.2.4 ODN Delivery to Cells In Vitro 84 3.2.5 In Vitro Downregulation of pi85 with H2A-ODN in both the M D A - M B -4 3 5HER2 C e U L m 6 ; a n d S K _ B R _ 3 ; a B r e a s t C e l l L i n e w h i c h Exhibits High HER-2/neu Levels Due to Over-Amplification of the Gene 86 3.2.6 Treatment with Free Doxorubicin in Mice Bearing MDA-MB-435 or M D A -M B - 4 3 5 H E R 2 Tumours 90 3.2.7 Treatment with Free H2A-ODN or SCR-H2A-ODN in Mice Bearing M D A -MB-435 or MDA-MB-43 5 H E R 2 Tumours 90 3.2.8 Liposomal ODN Characterisation 96 3.2.9 Treatment with Liposomal H2A-ODN or SCR-H2A-ODN in Mice Bearing MDA-MB-435 or M D A - M B - 4 3 5 H E R 2 Tumours 98 3.2.10 Combination Therapy: Liposomal ODN with Doxorubicin in the Treatment of MDA-MB-435 and M D A - M B - 4 3 5 H E R 2 Xenograft Tumours 107 3.3 Discussion 107 Chapter 4: Characterisation of the MDA435/LCC6 Human Breast Carcinoma Xenograft Tumour Model 113 4.1 Introduction 113 4.2 Results 116 4.2.1 In Vitro and In Vivo Growth Characteristics 116 4.2.2 Cytotoxicity of Taxol, Vincristine and Doxorubicin Against MDA435/LCC6 .... 119 4.2.3 Markers of Aggressive Breast Cancer 119 4.3 Discussion 128 Chapter 5: Pharmacodynamic Behaviour of Liposomal Antisense Oligonucleotides Targeting Her-2/neu and Vascular Endothelial Growth Factor as Measured in the Ascitic MDA435/LCC6 Human Breast Cancer Model 132 5.1 Introduction 132 5.2 Results 134 5.2.1 Plasma Elimination and Tissue Distribution of i.v. or i.p. Injected Free or Liposomal Antisense 134 5.3 Discussion 143 Chapter 6: Efficacy of Liposomal Antisense Oligonucleotides in the MDA435/LCC6 Human Breast Cancer Xenograft Model 145 6.1 Introduction 145 6.2 Results 148 6.2.1 Treatment of MDA435/LCC6 Ascites Tumour Bearing Mice with H2A-ODN.. . 148 6.2.2 Treatment of MDA435/LCC6 Ascites Tumour Bearing Mice with Trastuzumab 151 6.2.3 Treatment of MDA435/LCC6 Ascites Tumour Bearing Mice with V E G F - O D N 154 6.3 Discussion 156 Chapter 7: Conclusions 158 Bibliography 175 VI LIST OF TABLES Number Page 1.1 Mechanisms of action of chemotherapeutic agents used in the 25 treatment of aggressive breast cancer 2.1 Sequences and backbone modifications of antisense oligonucleotides 56 3.1 pl85 to actin ratios in DODAC/DOPE ± ODN treated M D A - M B - 89 435HER2 cells, and relative pi85 expression levels in SK-BR-3- cells following treatment with DODAC/DOPE ± ODN 3.2 Liposome size and ODN encapsulation of liposomal antisense 99 oligonucleotides at 4°C, 21°C or 37°C over 22 weeks 5.1 Mean A U C from 0 - 2 4 hours for free and liposomal O D N given both 136 intraperitoneally and intravascularly 5.2 Tissue antisense levels expressed as p.g ODN per g tissue following 140 i.v. injection of 10 mg/kg free or liposome encapsulated V E G F - O D N in the MDA435/LCC6 ascites tumour bearing mice 6.1 Tumour burden (total peritoneal fluid volume and tumour cell count), 155 V E G F levels per million peritoneal tumour cells, and serum V E G F (pg/ml) in control, untreated mice, and mice treated with 10 - three times weekly 10 mg/kg VEGF-ODN injections vn LIST OF FIGURES Number Title Page 1.1 Diagrammatic representation of the actions of V E G F isoforms on the 11 endothelium. 1.2 General structure of class I receptor tyrosine kinases. 13 1.3 Class I R T K ligands, homo- and heterodimerisation and signalling 15 pathways. 1.4 Native and common backbone modifications of antisense 30 oligonucleotides 1.5 Proposed mechanisms of action of antisense oligonucleotides in cells 33 1.6 Partitioning of drugs into the lipid bilayer or the central aqueous core 42 of liposomes 1.7 Ammonium sulphate gradient method of loading of doxorubicin into 47 liposomes. 1.8 pH gradient method of loading drugs into liposomes. 49 2.1 Lipids used in the formulation of liposomal antisense 65 oligonucleotides. 3.1 In vitro and in vivo growth of untreated MDA-MB-435 and M D A - 83 M B - 4 3 5 H E R 2 tumour cells 3.2 Immunohistochemical staining of pl85 in solid tumour sections from 85 MDA-MB-435 and MDA-MB-435 H E R 2 -derived tumours grown subcutaneously in female SCID/Rag2m mice. 3.3 In vitro delivery of FITC-labelled VEGF-ODN to MDA435/LCC6 87 human breast carcinoma cells. 3.4 pi85 expression in human breast carcinoma cells following treatment 88 with cationic liposome/ODN complexes, as assessed by flow cytometry 3.5 Doxorubicin and ODN treatment schedules in MDA-MB-435 and 91 MDA-MB-43 5 H E R 2 cell derived xenograft tumours 3.6 Impact of treatment with free doxorubicin in mice bearing M D A - M B - 92 435 or M D A - M B - 4 3 5 H E R 2 non-established or established tumours 3.7 Treatment of MDA-MB-435 or M D A-MB-43 5 H E R 2 cell derived 94 established or non-established xenograft tumours with SCR-H2A-ODN 3.8 Anti-HER-2/neu antisense treatment in MDA-MB-435 or M D A - M B - 95 435 H E R 2 cell derived tumour bearing mice. 3.9 Delivery of liposomal antisense oligonucleotides to tumour site via 97 systemic administration 3.10 Visualisation of antisense oligonucleotides in polyacrylamide gel by 100 silver stain vui 3.11 Treatment of MDA-MB-435 tumours with liposomal H2A-ODN, 101 liposomal SCR-H2A-ODN or a combination of liposomal H2A-ODN and doxorubicin. 3.12 Treatment of M D A-MB-43 5 H E R 2 tumours with liposomal H2A-ODN, 102 liposomal SCR-H2A-ODN or a combination of liposomal H2A-ODN and doxorubicin . 3.13 Immunohistochemical staining for pl85 on solid M D A - M B - 4 3 5 H E R 2 105 tumour sections following no treatment; treatment with free doxorubicin; liposomal SCR-H2A-ODN; liposomal H2A-ODN or liposomal H2A-ODN + free doxorubicin. 3.14 HER-2/neu protein expression in MDA-MB-435 and M D A - M B - 106 435 H E R 2 cells as assessed by Western blotting 4.1 Growth characteristics of MDA435/LCC6 cells in vitro and in vivo. 118 Grown as solid tumour on back; solid tumour in mammary fat pad or ascites tumour. 4.2 Cytotoxicity of vincristine, taxol and doxorubicin against M D A - M B - 120 435 and MDA435/LCC6 cells. 4.3 In vitro V E G F levels in supernatant collected from confluent M D A - 122 MB-435 and MDA435/LCC6 cells, measured by ELISA. 4.4 Immunohistochemical assessment of in vivo V E G F and HER-2/neu 124 levels in A431, SK-BR-3, MDA-MB-435 or MDA435/LCC6 solid tumours. 4.5 Bcl-2 protein levels in MDA-MB-435 and MDA435/LCC6 tumour 125 cells. 4.6 Bcl-2 staining on MDA-MB-435 and MDA435/LCC6 solid tumour 126 sections. 4.7 Cathepsin D staining on MDA-MB-435 and MDA435/LCC6 solid 127 tumour sections. 5.1 Plasma antisense, lipid and antisense to lipid ratios following i.v. or 135 i.p. injection of free or liposomal H2A-ODN at an antisense dose of 10 mg/kg into tumour-free female SCED/Rag2m mice 5.2 Serum antisense levels in MDA435/LCC6 ascites tumour-bearing 138 female SCID/Rag2m mice following injection of free or liposomal H2A-ODN or VEGF-ODN 5.3 Peritoneal fluid and peritoneal cell-associated antisense levels 142 following administration of free or liposomal V E G F - O D N at an antisense dose of 10 mg/kg into female SCID/Rag2m MDA435/LCC6 ascites tumour-bearing mice 6.1 HER-2/neu expression in peritoneal tumour cells taken from mice 149 bearing MDA435/LCC6 ascites tumours following treatment with saline, free H2A-ODN, free SCR-H2A-ODN or empty liposomes 6.2 Tumour burden and percent murine/human cells in MDA435/LCC6 150 tumour bearing female SCU)/Rag2m mice following treatment with free H2A-ODN or SCR-H2A-ODN, or liposomal H2A-ODN 6.3 Modified Wright-Giemsa stain of MDA435/LCC6 tumour cells 152 harvested from peritoneal cavity of an untreated mouse and cells v ix harvested from peritoneal cavity of a mouse inoculated with MDA435/LCC6 cells and treated with liposomal H2A-ODN Total volume and total number of peritoneal cells harvested from mice inoculated with MDA435/LCC6 cells and treated with saline, trastuzumab or empty liposomes ABBREVIATIONS aa Amino acids AR Amphiregulin p-CEL Beta-cellulin CHE Cholesteryl hexadecyl ether Choi Cholesterol DAB Diaminobenzidene DCIS Ductal carcinoma in situ dH 20 De-ionised water DMEM Dulbecco's modified eagle medium DMSO Dimethylsulphoxide DNA Deoxyribonucleic acid DODAC Dioleyl dimethylammonium chloride DODAP Dioleoyl dimethylammonium propane DSPC Distearoyl phosphatidylcholine DSPE Distearoyl phosphatidylethanolamine EDTA Ethylenediamine tetraacetic acid EGF Epidermal growth factor EGFR Epidermal growth factor receptor ELISA Enzyme linked immunosorbent assay EPC Egg phosphatidylcholine Epi Epiregulin ER Oestrogen receptor FBS Foetal bovine serum FITC Fluorescein isothiocyanate FIt-1 /ms-like tyrosine kinase H2A-ODN Anti-HER-2/neu antisense oligonucleotide HB-EGF Heparin binding EGF HBS HEPES buffered saline HBSS Hank's balanced salt solution HEPES Hydroxyethylpiperazine ethane sulphonic acid HSP Hydrogenated soya phosphatidylcholine IDC Infiltrating ductal carcinoma KDR kinase-insert-domain-containing receptor LCIS Lobular carcinoma in situ MAPK Mitogen activated protein kinase MPS Mononuclear phagocyte system mRNA Messenger ribonucleic acid MTT 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide NP-40 Nonidet P-40 ODN Antisense oligonucleotide PBS Phosphate buffered saline PCNA Proliferating-cell nuclear antigen P E G Poly [ethylene glycol] P E G - C , 4 C E R PEG2oooCi4Ceramide PI Propidium iodide PVDF Polyvinylidene fluoride Q E L S Quasi-elastic light scattering R N A ribonucleic acid RNase Ribonuclease SCR-H2A-ODN Scrambled control antisense oligonucleotide for HER-2/neu SDS Sodium dodecyl sulphate T E M E D N,N,N',N'-Tetramethylethylenediamine TGFa Transforming growth factor alpha tPA Tissue type plasminogen activator uPA Urokinase type plasminogen activator W O Vesicular vacuolar organelles V E G F Vascular endothelial growth factor V E G F - O D N Anti-VEGF antisense oligonucleotide V P F Vascular permeability factor ACKNOWLEDEGMENTS I would like to acknowledge the support of all the people in the department of Advanced Therapeutics. Foremost, my supervisor, Dr. Marcel Bal ly for his unflagging belief in my abilities, and advice and encouragement over the years. Dana Masin, Natashia Mcintosh and Rebecca N g for invaluable assistance in animal studies. Howard L i m (Howie) for holding my hand and showing me the ropes when I first arrived in the lab. Ellen Wasan for friendship and guidance. The wonderful atmosphere of camaraderie created by Nancy Dos Santos, Sheela Abraham, Daniel Menezes, Jennifer Shabbits, Ghania Chikh, Mac Parmar, Jason Sartor, Gig i Chiu, Maggie L i , and all the other members of the department. I would also like to acknowledge the work performed by Brenda Farnquist in the characterisation of the liposomal antisense formulation. xm DEDICATION This thesis is dedicated with much love to my daughters, Roxanne Nastassia and Rachel Emily. For surely the light and courage that emanate from their beings has guided my soul and lent me strength as I have ventured along this path. X I V Chapter 1 INTRODUCTION 1.1 Breast Cancer Breast cancer is the most common type of cancer of women both in North America and world-wide. It is second only to lung cancer in cancer related deaths in women. In British Columbia, in the year 2000, 2300 women were estimated to be diagnosed with some form of breast cancer, and 630 women estimated to have died of this disease (NCIC Canadian Cancer Statistics, 2000). There is no single causative factor in breast cancer, but rather a number of causes, ranging from low to high associated risk, and including both genetic and non-genetic factors. Those factors which increase relative risk of developing breast cancer by only a small amount include first menses before the age of twelve and rapid establishment of regular cycles, late menopause (women who experience menopause before the age of 45 have half the breast cancer risk of women who go through menopause after the age of 55 (Henderson & Feigelson, 1997)), family history of postmenopausal breast cancer, and prolonged hormone use. Moderate risk factors include postmenopausal obesity, as well as any hyperplasia or abnormality on a previous breast biopsy. Factors associated with a 4-fold or greater increased risk of developing breast cancer include being a female, ageing, previous cancer in one breast, a family history of premenopausal breast cancer, and severe hyperplasia or atypia on a previous breast biopsy (Olivotto et al, 1998; Henderson & Feigelson, 1997). In addition, there are environmental factors to be considered, such as exposure to x-ray radiation at an early age, which has been linked to 1 increased incidences of breast cancer (Doody et al., 2000). It is clear that there are more environmental factors involved in the development and progression of breast cancer. This may be simply demonstrated by a comparison of the incidence rates among North American women as compared to Japanese women (Japanese women have an incidence of breast cancer one half that of their North American counterparts). However, after only one generation, the incidence among Japanese women living in North America is 75% that of Caucasian American women, and after several generations, the incidence rates are equivalent among the two groups of women (Stanford et al., 1995). Conversely, there are factors associated with decreased risk of developing breast cancer, including increased parity, with the first birth at an early age, and lactation, with increased lactation resulting in decreased ovulatory cycles, and hence, decreased risk (Henderson & Feigelson, 1997), as well as partaking in regular exercise, and having a diet low in fat (Olivotto et al., 1998). A percentage of breast cancers are linked to specific genetic susceptibility, for example, mutations in the tumour suppressor genes B R C A 1 and B R C A 2 . For malignant transformation, both copies of the gene must be either lost, mutated or inactivated. Females with the B R C A 1 gene mutation have a 70% chance of developing breast cancer, compared to the approximately 11% risk of women in the general population, and this risk has been shown to be modified by non-genetic risk factors, such as parity (Narod et al, 1995). Genes are also involved in the progression of breast cancer, and its prognosis. Examples of this are the tumour suppressor gene p53 (mutations in 15 - 35% of breast cancers)(Bernstein et al, 1999), the overexpression of the epidermal growth factor receptor (EGFR) and the oncogene HER-2/neu (amplified and/or overexpressed in up to 30% of 2 breast cancers)(Slamon et al., 1987; Slamon et al., 1989). Alterations in these and other genes can be very important predictive and prognostic factors in the treatment of individual cases of breast cancer. They also provide unique targets for the development of therapeutics in the fight against this disease, as discussed later in this chapter and in this thesis. There are two main subcategories of malignant breast tumours; carcinomas and sarcomas. Of these two subtypes, carcinomas are by far the predominant tumour type. Currently, all breast carcinomas are classified on the basis of histological and/or cytological appearance, as well as gross findings such as site, size, shape, consistency, colour, and relationship to adjacent mammary and extramammary structures. Within the carcinomas, there are a variety of histological types, with most arising from the epithelial cells lining the terminal duct lobular units of the breast. Important among carcinomas is ductal carcinoma in situ (DCIS). In DCIS, the milk ducts become blocked with tumour cells, and calcium may be collected in the blocked ducts. There are two types of DCIS, namely solid or comedo, with comedo, identified by the large, irregular cells, being more likely to progress to invasion. Although not an infiltrating carcinoma, it is believed that the majority of infiltrating carcinomas arise from DCIS, especially from poorly differentiated, high grade DCIS. Of interest, the level of expression of the oncogene HER-2/neu within DCIS is also believed to be indicative of the likelihood of progression to infiltrating carcinoma of the breast (Olivotto et al., 1998). A second type of non-invasive carcinoma is lobular carcinoma in situ (LCIS). This is differentiated from DCIS on the basis of tumour formation in the glands rather than ducts within the breast (Olivotto et al., 1998). 3 Once tumour cells have broken through the epithelial lining of the ducts or glands and entered the fatty tissue of the breast, the tumour is classified as invasive, or infiltrating. The predominant type of infiltrating carcinoma is infiltrating ductal adenocarcinoma (IDC), which comprises 50 - 60% of all breast cancer cases (Rilke & Di Palma, 1998). These tumours are graded based on the degree of tubule formation (little or none being a higher grade tumour), nuclear pleomorphism, and mitotic index. With IDC, tumour emboli are often seen inside blood and lymph vessels within and outside the tumour area. The presence of such emboli within the lymphatics is a predictor of recurrence and metastatic spread (Henderson & Feigelson, 1997). IDC may be broken down into the subtypes tubular, colloid, mucinous, scirrhous and medullary, however treatment is generally the same for all types. (Olivotto et al., 1998). Other types of invasive breast cancer include i) inflammatory carcinoma, which is characterised by swollen, tender, reddened breasts, and is particularly aggressive, ii) lobular carcinomas, which account for approximately 15% of invasive breast cancer, and respond well to hormone based treatments, and iii) cancers that have their origin in parts of the breast other than the epithelium, including sarcomas, lymphomas and cystosarcoma phyllodes (Olivotto et al, 1998). 1.2 Prognostic and Predictive Factors in Breast Cancer Prognostic factors are those which relate to the risk of systemic recurrence and/or death following the primary therapy of a tumour. Predictive factors indicate how a tumour may 4 respond to a given therapy. A central aspect of this thesis concerns use of prognostic and predictive factors in guiding therapy. 1.2.1 Gross and Histopathological Factors The classic prognostic factors in breast cancer include the histological type of the tumour, e.g., DCIS, LCIS or invasive carcinoma, as well as the components of the T M N tumour grading system. The T M N grading scheme includes the tumour size (T), the presence of distant metastases (M), and the involvement of lymph nodes (N). To address the nodal status primarily, i f a patient has no cancer cells detectable in the axillary lymph nodes, this confers a good prognosis, and is referred to as node negative disease. Node positive disease encompasses any involvement of lymph nodes, whether it be only one node with detectable cancer cells, or more. The greater the number of positive lymph nodes, the poorer the prognosis for disease outcome. The presence of distal metastases is a poor prognostic factor, as is a larger tumour size at discovery. The mitotic index is a measure of the proliferation rate of the tumour. Other markers of proliferation are also employed, such as thymidine labelling index and measurement of cellular proteins expressed during specific parts of the cell cycle, for example, Ki67 (Pietilainen et al., 1996) and proliferating-cell nuclear antigen (PCNA) (Ioachim et al., 1996). As well, flow cytometry measurements of D N A are used to determine the proportion of tumour D N A in S-phase. With all of these measurements, the greater the proportion of the tumour that is proliferative, the poorer the overall prognosis. Other features of the T M N system include the presence of tumour fixation to the skin or underlying muscle, matting or fixation of the axillary nodes, and an inflammatory component, which also designate patients with a poor 5 prognosis. (Ravdin, 1997). As well as utilising the T M N system, pathologists rely on histologic grading, which refers to a classification of breast cancer based on the nuclear grade of a tumour, the mitotic rate, and architectural differentiation. Those tumours that have pleomorphic nuclei are associated with a poorer prognosis, as are tumours with high mitotic indices and lack of cellular differentiation. 1.2.2 Molecular Factors The discovery of new prognostic and predictive markers for breast cancer is exciting, and has opened the way for a new generation of targeted therapeutics for treating this disease. Several proteins have now been associated with more aggressive breast, as well as some other types of cancer. By looking at the combination of gross, traditional histopathologic markers, and these new molecular markers of aggressive disease, pathologists are able to obtain a much clearer picture of a patient's cancer, and are able to provide more individualised care to that patient. The following sections will discuss a select few of these new markers. 1.2.2.1 CathepsinD The oestrogen regulated lysosomal protease Cathepsin D has been proposed to play an important role in breast cancer invasion and metastasis (Han et al., 1997). Because of this, patients with elevated cathepsin D levels, defined as those tumours with clearly detectable 6 cathepsin D levels (greater than or equal to 10% positive cells)(Gonzalez-Vela et ah, 1999) have an adverse prognosis (Solomayer et ai, 1998). As a protease, cathepsin D is likely to be involved in the degradation of extracellular matrix and therefore, tumour cell spread in circulatory and lymphatic systems. This same mechanism may also contribute to distant metastases. Within a tumour, positive staining of stromal cells for cathepsin D is often correlated with vascular density, another factor adversely related to prognosis (see section 1.2.2.2). Positive non-tumor cell staining is also associated with high grade tumour, high cell mitotic rate and oestrogen receptor (ER) negative status (all features of increased biological aggressiveness)(Gonzalez-Vela et al., 1999). There is a trend of increased cathepsin D in the tumour cells of ER positive breast cancer, but there is disagreement in the significance of this cathepsin D, with some researchers disputing the relation to other clinicopathologic parameters (Gonzalez-Vela et al, 1999), and others finding significant correlation between cathepsin D positive staining in tumour cells, and tumour size and nodal status in breast cancer metastases (Solomayer et al., 1998). An interesting observation has been made by Solomayer et al. (1998), in their study of bone metastases, in that cathepsin D is found only on tumour cells, not on stromal cells in the bone marrow of patients with primary breast cancer. They feel therefore, that in the final steps of metastasis, cathepsin D in tumour cells, but not stromal, is involved in invasion. 1.2.2.2 Vascular Endothelial Growth Factor Growing tumours are much more than passive clumps of cells. They are able to actively modulate their environment, inducing changes that allow advantageous conditions for growth. Since the continued growth of a tumour is dependent upon an adequate blood 7 supply, not surprisingly, one of the major microenvironment modifications that malignant cells have developed is a means of controlling their vasculature, including the secretion of factors that either inhibit or stimulate new blood vessel formation. The formation of new vessels from pre-existing ones is known as angiogenesis, and involves a complex cascade of events and factors. One of the most important of these angiogenic factors is vascular endothelial growth factor (VEGF), a multi-functional cytokine and endothelial cell mitogen also known as vascular permeability factor (VPF). This alternative nomenclature is a reflection of the ability of VEGF/VPF to induce vascular permeability with a potency some 50,000 times that of histamine (Keck et al., 1989; Senger etal, 1990). V E G F is a 34 - 42 kDa heparin-binding, dimeric, disulphide bonded glycoprotein that acts directly on endothelial cells by way of two specific class III receptor tyrosine kinases, fins-like tyrosine kinase (Flt-1) and kinase-insert-domain-containing receptor (KDR) (Senger et al., 1993). Many tumour cells express and secrete V E G F at high levels, with the effect of inducing the formation of new vasculature to promote tumour growth. The human V E G F gene is divided among eight exons, and as the result of alternative splicing of mRNA, yields four different VEGF isoforms of 121, 165, 189 and 206 amino acids (aa), respectively. Of these, the 121 and 165 aa isoforms are secreted, while the 189 and 206 aa isoforms remain bound to the cells. Finally, there are three additional forms of VEGF which have recently been described; VEGF-B, VEGF-C and VEGF-D. The 8 biological functions of these forms have not yet been fully elucidated, but may be involved in tumour spread through the lymphatics (Olufsson et al, 1996; Joukov et al., 1996; Lee et al, 1996 Kurebayashi etal, 1999). V E G F is able to induce angiogenesis in two ways. Primarily, it does so by means of specific mitogenic activity on endothelial cells, but it also indirectly contributes to angiogenesis by effects on vascular permeability. One of the ways in which V E G F affects permeability is to enhance the activity of vesicular-vacuolar organelles ( W O ) within the cytoplasm of endothelial cells. W O are complex structures that span the endothelial cell cytoplasm and allow the passage of macromolecules, providing a pathway from the circulation into the tissues. VEGF-induced hyperpermeability of tumour microvasculature allows for the accumulation of fluid and plasma proteins in the tumour interstitium. One of the extravasated proteins is plasma fibrinogen, the presence of which in the interstitium, leads to fibrin deposition and subsequent alteration of the tumour extracellular matrix. This altered matrix is ideal for, and promotes the ingrowth of fibroblasts, macrophages and endothelial cells, which serve to promote tumour growth (Bobik et al, 1997). The deposition of an extravascular fibrin matrix is in itself, a potent angiogenic stimulus (Brown et al, 1992), and also contributes to overall tumour growth. Angiogenesis has also been related to the formation of metastases, in that risk of metastatic tumour is dramatically increased with the formation of a dense new vasculature (Bobik et 9 al, 1997). It is even possible in some cases to statistically correlate the appearance of metastases with the number of microvessels detected in the primary tumour (Martiny-Baron & Maime, 1995). The primary signal for angiogenesis in tumours is believed to be the release by tumour cells of diffusible signals toward pre-existing capillaries which commit resting endothelial cells to actively dividing and migrating (Plouet & Bayard, 1994). A further component of this theory is that cells at the tip of the migrating bud of endothelial cells are not dividing, but those behind them are, helping to push the bud toward the angiogenic stimulus. Figure 1.1 summarises some of the actions of the different forms of V E G F on the vascular endothelium. Endothelial cells stimulated by hypoxia, transforming events or cytokines are able to elaborate collagenases and plasminogen activators to aid in the local degradation of the basement membrane. Sprouting endothelial cells, which form the bud of a developing new vessel are able to induce the expression of the serine proteases; tissue type and urokinase type plasminogen activator (tPA and uPA, respectively) and metalloproteinase interstitial collagenases. This co-induction of PA and collagenases forms a pro-degradative environment which facilitates the migration of the endothelial cells through the tissue (Mignatti & Rifkin, 1996). 1.2.2.3 HER-2/neu HER-2/neu (also known as c-erbB-2) is a proto-oncogene located on the long arm of chromosome 17, specifically in the area of 17q.12-21.32 (Popescu et al., 1989). The HER-10 Tumour Cells Hypoxia Transforming Events Y Y Y Y Y K K V E G F | 6 5 V E G F , 2 , ~ > Figure 1.1 Actions of tumour-derived vascular endothelial growth factor on endothelial cells. Tumour cells secrete V E G F in response to various stimuli, which exerts a mitogenic affect on endothelial cells, as well as increasing intracytoplasmic [Ca 2 +] levels. An additional effect is the formation of a prodegradative environment that facilitates migration and sprouting of endothelial cells by the co-induction of the serine proteases urokinase plasminogen activator (UPA) and tissue-type plasminogen activator (tPA) as well as collagenase. Plasminogen activation results in generation of plasmin, which may cleave extracellular matrix bound V E G F (VEGFigg or VEGF206) to release a diffusible proteolytic fragment (VEGFuo). The induction of plasminogen activator inhibitor (PAI-1) serves as a regulatory step to balance the proteolytic process. These actions are exerted through the interaction of V E G F with the class III receptor tyrosine kinases, K D R and Flt-1. (Adapted from Ferrara & Davis-Smyth, 1997). 11 2/neu oncogene was first isolated from D N A extracted from ethylnitrosurea-induced adrenal neuroglioblastomas of neonatal rats (Shih et al, 1981). This gene encodes a 185 kDa transmembrane receptor tyrosine kinase (pi85), which is a member of the Class I receptor tyrosine kinase (RTK) family (Fig. 1.2). Class I RTKs all possess a large, glycosylated extracellular ligand binding domain with two cysteine-rich sequences, a single hydrophobic transmembrane domain and a long cytoplasmic tail containing the kinase domain and C-terminally located autophosphorylation sites (Ben-Baruch et al, 1998). The prototype member of the class I R T K family is the epidermal growth factor receptor (EGFR). EGFR is expressed at high levels in approximately one third of all epithelial cancers, including breast cancer. EGFR overexpression has also been implicated in a poor prognosis in breast cancer patients (Pawlowski et al., 2000). The HER-2/neu protein shares extensive sequence homology with EGFR (also known as HER-1), and in fact, the amino acids in the tyrosine kinase domain of the two receptors are more than 80 % identical (Schechter et al, 1984; Yamamoto et al, 1986) in the kinase domain, and approximately 45% homologous in the ligand-binding domain (Yamamoto et al, 1986). At least in the case of EGFR, the ligand binding cleft of the receptor is confined to the portion linking the two cysteine-rich domains (Ben-Baruch et al, 1998). Within the class I R T K family, in addition to EGFR and HER-2/neu, there are two other family members, namely HER-3 and HER-4, which are known to form both homodimers and heterodimers with both HER-1 and HER-2/neu. 12 CD • l - H S-H 03 O cd X • i—i GO cd o cd isiii Cysteine-rich Domains Transmembrane Helix Tyrosine Kinase Domain COO" Figure 1.2 General structure of class I receptor tyrosine kinases. This class of RTKs includes the family members epidermal growth factor receptor (EGFR), HER-2/neu, HER-3 and HER-4. Indicated are the cysteine rich extracellular domain characteristic of this class, the transmembrane domain, and the tyrosine kinase domain of the intracellular portion of the receptor. 13 Although classified as a receptor, to date, there is no known ligand for pl85, although it is known to mediate lateral signal transduction in all HER family members. This is by means of heterodimerization with EGFR (see Fig 1.3), and activation by binding of ligand (discussed further in section 1.2.2.3.1) to its receptor, or heterodimerization with HER-3 or HER-4, and binding of ligands to these receptors. Ligand binding results in activation of the receptor complex, trans-phosphorylation, and subsequent signal transduction. HER-2/neu was first proposed to be a prognostic factor in breast cancer by Slamon et al. (1987), using a measure of the amplification of the gene itself. This initial study showed strong correlation between increased copy number of the gene and shortened disease free survival and overall survival in node positive patients. Many studies have been performed since 1987, using both gene amplification and immunohistochemical measurement of protein level as a measure of HER-2/neu level (Allred et al., 1992; Press et al., 1997; Ross et al, 1998; Paik et al, 1990; Pich et al, 2000). These above mentioned studies, for the most part, support the contention that amplification and/or overexpression of HER-2/neu is associated with more aggressive disease, and enhanced metastatic potential in node positive patients. Overexpression of HER-2/neu has been shown to correlate with resistance to a number of agents commonly used in the treatment of aggressive breast cancer, including cyclophosphamide, methotrexate and 5-FU (Piccart et al, 1988; Mourali et al, 1993). Intriguing data is available for pi85 overexpressing breast cancers and treatment with increasing doses of doxorubicin. These studies have demonstrated significant improvement 14 TGFcx A R EGF HB-EGF P-CEL Epi N D F a NDFP P-CEL Epi HB-EGF ^ 1 \ / \ \ — %1 HER-2 HER, ^ Dimerization Ras-GAP RAS Fig. 1.3 Class I R T K ligands, homo- and heterodimerisation and signalling pathways. The four class I R T K family members are able to form either homodimers or heterodimers with any of the other family members. This leads to phosphorylation of tyrosine residues on the intracellular portion of the receptor, and subsequent signal transduction, culminating in specific gene regulation and cell cycle progression as outlined in the text. 15 in survival with increased doxorubicin (Paik et al, 1998; Thor et al, 1998) dose only in HER-2/neu overexpressing tumours, and not those with low or absent pi85 expression. It is unclear as to whether this is due to increased sensitivity of p i 85 expressing tumours to doxorubicin, or whether the higher dose serves to overcome an intrinsic resistance in these tumours (Esteva-Lorenzo et al., 1998). 1.2.2.3.1 HER-2/neu Signalling Pathways There is a large family of EGF-related peptides which serve as ligands for the HER family of receptors. These ligands fall into three general groups: i) EGF, amphiregulin (AR) and transforming growth factor a (TGF-a), which bind EGFR; ii) beta-cellulin (P-CEL), epiregulin (Epi) and heparin binding EGF-like growth factor (HB-EGF), which bind both HER-1 and HER-4, and iii) neu differentiation factors (NDFs) or heregulins, which bind to both HER-3 and HER-4. Upon binding of ligand to receptor, there is promotion of receptor homo- and heterodimerization, and receptor activation (Fig 1.3). Dimerisation leads to autophosphorylation of receptors on specific tyrosine residues in their cytoplasmic domains, which then provide docking sites for Src Homology 2 (SH2) and phosphotyrosine binding domain-containing proteins, including She, Grb2 and the p85 sub-unit of phosphatidylinositol kinase. This leads to signalling pathways such as the mitogen activated protein kinase (MAPK) pathway and the S6 kinase cascade (Olayioye et al., 1998). There is, within the HER family of receptors, considerable overlap in the signalling pathways employed; however, there is some receptor preference for certain pathways, for 16 example, the Cbl protein appears to couple exclusively to EGFR, whereas the Csk-homologous kinase binds only to HER-2. EGF- and NDF-activated EGFR associate with She, but only EGF-activated EGFR couples to Grb2, an observation explained by differential phosphorylation of EGFR (Olayioye et al, 1998). This type of relationship applies to HER-2/neu as well, where phosphopeptide mapping of HER-2/neu has revealed dimerisation dependent phosphorylation (Olayioye et al, 1998). When these preferences are coupled with the different receptor heterodimerization possibilities, the range of signalling can become quite diverse. It has been shown that HER-2/HER-3 dimers are favoured over HER-2/HER-1 dimers (Olayioye et al., 1998), and that in all the possible dimerisations, HER-2/neu is the preferred partner of the other HER-family receptors (Graus-Porta, 1997; Olayioye et al, 1998). Because of this, HER-2/neu can be activated by EGFR, HER-3, HER-4 or by monoclonal antibody induced homodimerisation. Homodimer formation also leads to specific and distinct receptor phosphorylation patterns. As different adapter molecules have been mapped to specific tyrosine residues on the HER family members, and as a range of phosphorylation is possible due to the number of tyrosine residues which are phosphorylated at any given time, a broad range of signalling may be achieved via these different dimerisations (Olayioye et al, 1998). The diverse responses of HER-2/neu to activation include transformation, monoclonal antibody induced growth inhibition, ligand-induced growth stimulation or apoptosis. These varying results imply activation specific differences in the signalling capacity of HER-17 2/neu (Olayioye et al, 1998). These differences may result in unique responses of cells with overexpression of this protein to a given therapy, dependent upon the phosphorylation state, and the type of therapy, whether it be targeted to the protein, the mRNA or the DNA. It is critical that we develop an understanding of this potential variation of response before we can truly have a thorough knowledge of the mechanism by which the drugs that are being developed, are acting at a cellular level. In addition to this, although it is apparent that HER-2/neu is an ideal candidate for targeting of anti-cancer drugs, there is the potential for interaction of the protein with other members of the class I R T K family, thereby confounding the interpretation of results garnered with such therapies. 1.2.2.3.2 HER-2/neu in Normal Physiology HER-2/neu protein expression has been identified by immunohistochemistry on epithelial membranes in the gastrointestinal, respiratory, urinary and reproductive tracts, and in the skin of both adult and foetal specimens. Generally however, the expression levels are higher in foetal specimens (Press et al, 1990). Cells expressing pl85 have been demonstrated as being derived from all three germ layers, and of polarised cells attached to basement membrane, endothelial and mesothelial cells were the only cells which were found not to express pi85 (Press et al, 1990). In studies with knockout (HER-2/neu"7") mice, HER-2/neu was found to be critical in the development of trabeculae in the heart, as well as the formation of synapses between motor neurons and muscles (Ben-Baruch et al, 1998). In expression studies of human embryos and foetuses, from gestational week 5 to 8 there has been shown to be strongly positive staining for pi85 in the developing 18 myocardium, as well as in the plasma membrane regions of the developing epicardium, whereas after gestational week 9, staining was not observed (Miosge et al., 1997). A similar time dependent expression pattern was observed in muscle, with embryos of gestational weeks 4 through 10 demonstrating strong membrane staining for pi85, while from gestational week 11 onwards, the immunoreaction ceased (Miosge et al., 1997). Studies have also identified a key role for HER-2/neu in the development of trigeminal and facial motor nerves. Due to striking similarities between the HER-2/neu7" mice and N R G 7 " mice, and other knockout studies, these researchers concluded that the developmental role of HER-2/neu may be to augment signalling by NRGs, as well as to prolong the activity of EGFR ligands (Ben-Baruch et al, 1998). As a generalisation, pi85 expression during development is found mainly during the embryonic (to gestational week 9), and not the foetal (gestational week 10, on) period. The embryonic period is characterised by cell proliferation, cell differentiation and migration, whereas in the foetal period, growth and proliferation are not the predominant processes (Miosge et al., 1997), implicating a role for pi85 in these processes. HER-2/neu is also known to play a role during the differentiation of normal mammary glands, as mRNA levels increase at the onset of proliferation and during the lobulo-alveolar development, and then decrease sharply in milk-producing glands. During the differentiation phase, post-transcriptional mechanisms, such as altered turnover of the protein, maintain high levels of p i 85 (De Bortoli & Dati, 1997). With the above in mind, added to the cautions listed toward the end of section 1.2.2.3.1 is 19 that care needs to be taken in devising a therapeutic which targets proteins localised in normal, as well as cancerous tissues. Given the limited expression in the tissues of normal adults however, it can be suggested that HER-2/neu represents an almost ideal target for development of targeted therapy. 1.2.2.3.3 HER-2/neu Expression in Cancer Many researchers have investigated HER-2/neu expression levels in a variety of human neoplasms, with sometimes conflicting results. These differing results may in part be due to alternate methods of tissue processing, such as freezing, paraffin embedding, length of time to staining, type of antibody used, etc. In addition, researchers may score tissue positivity on different scales, therefore leading to higher or lower reported HER-2/neu expression values. There is a now general consensus among researchers, that only staining which is present on the membranes of cells, and not intracytoplasmic, should be scored as positive staining for HER-2/neu. In fact, some researchers feel that intracytoplasmic staining may not even be representative of an immature precursor of pi85, but rather, a distinct protein that is able to cross react with the antibodies used (Tannapfel et al, 1996). Having stated this, there have been reports of HER-2/neu amplification and/or overexpression in several tumour types, including gastric (Tannapfel et al, 1996), colorectal (Tannapfel et al., 1996; Yang et al., 1996), transitional cell bladder (Tannapfel et al., 1996), renal cell (Hofmockel et al, 1997), ovarian (Schmitt et al, 1996; Eltabbakh et al, 1997), endometrial (Schmitt et al, 1996), cervical (Nakano et al, 1997; Sharma et al, 20 1999), breast (Schmitt et al, 1996) and non-small cell lung (Giatromanolaki et al, 1996) carcinomas. Not all of these instances of overexpression are linked with an adverse prognosis; however, this does appear to be the case for at least breast and ovarian carcinomas. 1.2.2.3.3.1 HER-2 /neu in Breast Cancer The HER-2/neu gene is amplified and/or overexpressed in up to 30% of all human breast carcinomas and has been correlated with decreased overall survival and time to relapse in this patient population. This patient group has a significantly lower overall survival rate, and shorter time to relapse than those patients whose tumours do not overexpress HER-2/neu (Allred et al, 1992; Borg et al, 1990; Sjogren et al, 1998; Slamon et al, 1989). Moreover, this significance is retained even when adjustments are made for other prognostic factors (Slamon et al, 1987), and is true for both node positive (Slamon et al, 1987) and node negative (Andrulis et al, 1998) breast cancers. Research into a mechanism for HER-2/neu action has shown that overexpression of HER-2/neu in cell lines increases the migratory capabilities of the cells, an important step in metastasis formation (Verbeek etal, 1998). When the overexpression of HER-2/neu in a significant sub-set of breast cancer patients is paired with the adverse prognosis related to this overexpression, it is apparent that HER-2/neu presents an ideal therapeutic target. Many researchers have embarked on the development of anti-HER-2/neu therapeutic strategies. The foremost among these are 21 discussed in section 1.3. 1.3 Therapeutic Strategies in the Treatment of Breast Cancer Metastatic breast cancer is generally thought of as an incurable disease, with a median survival of 18 - 24 months (Miller & Sledge, 1999). With this in mind, it is sometimes difficult to decide whether to treat aggressively with curative intent, or to give palliative treatment only. Palliation may be achieved with several treatments, including hormonal manipulations, systemic chemotherapy, local radiotherapy, and immune-based biologic therapies (Miller & Sledge, 1999). These same treatments may, in a few cases, also lead to cures, however there is a definite need for development of novel agents which will expand the number of cures being obtained. Although there a number of prognostic markers such as HER-2/neu, that increase the potential for better treatment options, it is likely that these new therapeutics will only be used in combination with one or all of the existing treatment modalities. It is therefore useful to review these here. 1.3.1 Chemotherapeutics A number.of chemotherapeutic agents form the backbone of current therapy for aggressive breast cancer. These agents include anthracyclines, taxanes, cyclophosphamide, methotrexate and 5-fluorouracil (5-FU), and responses can be as high as 30% in previously untreated patients (Miller & Sledge, 1999). Newer agents include Navelbine (vinorelbine), gemcitabine and capecitabine. Table 1.1 lists the more commonly used agents, as well as briefly describing some of their mechanisms of action. 22 Because of the increased efficacy of anthracycline-containing regimens, they are the most commonly used regimens for both primary and metastatic breast cancer (Hortobagyi, 2000). The most common representative of the anthracyclines is doxorubicin. Therapy with this drug is limited acutely by myelosuppression, however both acute and chronic cardiotoxicity are of great concern. Since the cardiomyopathy is dose dependent, most protocols limit the lifetime dose of this drug to about 450 mg/m2. Doxorubicin is usually given in combination with other drugs, such as cyclophosphamide and 5-FU. The taxanes paclitaxel and docetaxel are relatively new anticancer agents which work by inhibiting microtubule disassembly by binding to dimeric tubulin. This results in cells being blocked in the G2 and M phases of the cell cycle (Miller & Sledge, 1999). Although these are plant alkaloids (derived from European and Western Pacific yew trees), taxanes used in cancer therapy are semi-synthetically manufactured. Importantly, these agents have anti-tumour activity in anthracycline resistant breast cancer. Response rates in metastatic breast cancer are equal to or greater than 30% with single agent paclitaxel, and over 40% with single agent docetaxel, and each of the taxanes are now being incorporated into combination chemotherapy regimens (Hortobagyi, 2000). Nucleoside analogues are relatively new agents in the treatment of breast cancer. This group includes the deoxycytidine analogue, gemcitabine (difluorodeoxycytidine), as well as cytarabine. These drugs block the progression of cells from the G l phase of the cell cycle through to the S phase (DNA synthesis). Additionally, cytarabine will act in the S phase to 23 kill cells undergoing D N A synthesis. The most commonly used alkylating agent is currently cyclophosphamide. One of the advantages of cyclophosphamide over its counterparts, nitrogen mustard, and chlorambucil, is that is it available in both parenteral and oral formulations, and therefore can be used in a wide variety of dosage schedules (Carter, 1972). Cyclophosphamide is commonly used in conjunction with an anthracycline and another agent such as 5-FU in the treatment of metastatic breast cancer. Vinorelbine, a norvinblastine analogue, also targets microtubules, and has reduced neurotoxicity as compared with vincristine due to its higher affinity for mitotic spindle microtubules, rather than axonal microtubules (Miller & Sledge, 1999). Vinorelbine is also better tolerated than vincristine, and has value as a salvage treatment (final treatment for patients non-responsive to other available therapies) in patients with prior anthracycline and taxane exposure (Miller & Sledge, 1999). 1.3.2 Hormonal Strategies There are several agents used in the treatment of breast cancer that fall into this category, including tamoxifen, toremifene, raloxifene, anastrozole, mifepristone and others. Tamoxifen is a synthetic anti-oestrogen agent coming into common use in both the treatment of early breast cancer (Early Breast Cancer Trialists' Collaborative Group, 1998), as well as for the prevention of breast cancer (Fisher et al, 1998). When all breast tumour 24 Table 1.1 Mechanisms of action of chemotherapeutic agents used in the treatment of aggressive breast cancer. Chemotherapeutic Agent Mechanism(s) of Action Anti-tumour Antibiotics Anthracyclines (e.g., doxorubicin, epirubicin) Intercalation within D N A ; inhibition of topoisomerase II Anti-tumour Antibiotics Mitomycin Inhibition of D N A synthesis; R N A and protein synthesis inhibition at high doses Anti-tumour Antibiotics Anthracenediones (e.g., mitoxantrone) Intercalation within D N A Plant alkaloids Taxanes (e.g., paclitaxel, docetaxel) Inhibition of microtubule disassembly by binding to dimeric tubulin Plant alkaloids Vinca alkaloids (e.g., vincristine, vinblastine, vinorelbine) Inhibition of microtubules of the mitotic spindle Nucleoside analogues Gemcitabine, Cytarabine Incorporation into D N A of intracellularly generated metabolite Anti-metabolites 5-fluorouracil Inhibition of thymidylate synthetase; Incorporation of FUTP into R N A and D N A Anti-metabolites Capecitabine Prodrug leading to high tumour concentrations of 5-FU Anti-metabolites Methotrexate Inhibition of purine synthesis Alkylating agents Cyclophosphamide Prevention of cell division by cross-linking D N A strands Alkylating agents Chlorambucil Cross-linking of D N A and R N A strands, inhibition of protein synthesis types are included in analyses, the overall response rate for tamoxifen as a single agent is approximately 30 - 35%. When evaluating ER and progestin receptor positive breast tumours only, the response rate is approximately 60 - 75% (Fornier et al, 1999). Unfortunately, most breast cancer patients eventually become resistant to tamoxifen, and must rely on other strategies to fight their cancer (Fornier et al., 1999). Other anti-oestrogens currently in clinical trial or approved for use in metastatic breast cancer are 25 toremifene, droloxifene and ICI 182780. Other hormonal strategies are the use of i) aromatase inhibitors, which block the peripheral conversion of androstenedione to estrone (e.g., anastrozole letrozole, formestane and exemestane), and achieve major reductions in oestrogen levels without suppressing adrenal function, and ii) anti-progestins, an example of which is the abortion drug, R U 486 (Fornier etal, 1999). 1.3.3 Molecular Targeting The strategies discussed in section 1.3.2 are those which essentially target a specific component of breast cancer, namely, the oestrogen or progestin receptors. These drugs, while having a very specific mechanism of action, do not truly target themselves to the tissue in question. Many approaches of molecular targeting have been considered, including the use of monoclonal antibodies and small molecule inhibitors. The following section describes the use of antibodies directed against two members of the class I R T K family, HER-1 (EGFR) and HER-2/neu. This type of therapy represents a true targeting, as the therapy is concentrated at the site of action by the nature of the drug binding to specific ligands expressed on the tumour cell surface. 1.3.3.1 Antibodies The most widely discussed antibody used in treatment of cancer is trastuzumab (Herceptin™). This is a humanised monoclonal antibody targeted to the extracellular 26 domain of the HER-2/neu protein, and has been shown in clinical trials to produce objective tumour responses in 15 - 21% of patients with HER-2/neu overexpression and who had relapsed following chemotherapy for metastatic breast cancer (Cobleigh et al., 1998). It has been postulated that this antibody works by antagonising the function of the growth signalling properties of the HER-2/neu system. However it has also been suggested that it may signal immune cells to attack and kill tumour cells, and in addition, may enhance the cytotoxicity of other chemotherapeutic agents (Fornier et ah, 1999). A potential concern with this therapy is that it may either cause cardiac dysfunction, or amplify the cardiac toxicity observed with the use of anthracyclines (toxicity observed in 6% of doxorubicin only patients versus 27% of patients treated with the combination of trastuzumab plus doxorubicin in a phase III clinical trial)(Ewer et al., 1999). Trastuzumab has recently been granted approval in the United States and several European countries, and it is likely that it will be paired with anthracycline therapy; therefore this potential for enhanced cardiac toxicity needs to be further explored. One of the objectives of this thesis is develop an antisense oligonucleotide therapeutic targeting HER-2/neu, as does trastuzumab, and to pair this new therapy with cytotoxic agents such as doxorubicin. Ideally, this combination will be without the additional toxicity currently noted with the trastuzumab/doxorubicin combination. Another antibody targeting one of the HER family members is currently in development. C225 is an anti-EGFR antibody designed for use in women whose breast tumours overexpress EGFR. 27 1.3.4 Gene Therapy There are several aspects of gene therapy, including the enhancement of tumour immunogenicity by insertion of cytokine or co-stimulatory molecule encoding genes, or direct killing of tumour cells by insertion of tumoricidal genes, tumour suppressor genes, and genes encoding prodrug activating enzymes (Dachs et ah, 1997). In the work presented in this thesis, I have focused on the gene therapy approach of antisense oligonucleotides (ODN), short, single stranded segments of DNA. O D N are not a typical type of gene therapy in that they are not designed to correct gene defects or achieve insertion of new genes into tumour cells. These ODN thus share the specificity of whole gene approaches without the potentially deleterious insertion into the genome, as well as having other benefits as discussed below. 1.3.5 Antisense Oligonucleotides Antisense oligonucleotides are relatively novel therapeutic drugs, and by their nature, are able to target the mRNA or D N A of many different proteins. O D N are short sequences of single stranded DNA, typically ranging from 12-30 bases in length. They are designed to prevent gene expression by interfering with some portion of the translation process of mRNA into protein. Theoretically, any transcribed gene is a candidate target for antisense inhibition, although any targeted sequence must be checked for homologies to other genes. The affinity of any given ODN for the targeted mRNA results from the nature of the 28 hybridisation between the two. This is largely a result of Watson-Crick base pairing, as well as base stacking within the resultant double helix that forms. As the hybridisation length is increased, the affinity of binding increases accordingly, however, as R N A is able to adopt both secondary and tertiary structures, the theoretical binding affinities are rarely reached (Crooke, 1999). Also contributing to this lack of affinity is the complex nature of the cytoplasmic environment in which the hybrids are formed, with many protein interactions possibly obstructing complete hybridisation. 1.3.5.1 Types of antisense oligonucleotides Native D N A has a phosphodiester backbone, and many O D N studies have utilised this same backbone. More recently, several backbone modifications have been made which improve the nuclease stability of the ODN, while still resembling natural nucleic acids as closely as possible. These include the replacement of one of the non-bridging oxygen molecules of the phosphodiester bond with a methyl group (methylphosphonate ODN), a methyl ester (phosphotriester ODN), a methyl ammonium group (phosphoroamidate ODN) or with a sulphur atom (phosphorothioate ODN) (Figure 1.4). There are also many modifications to antisense which involve the sugar moiety alone, or in combination with backbone modifications (Manoharan, 1999; Geary et al, 2001). These modifications are designed to decrease the degradation of ODN due to both endo and exonucleases, as well as optimising tissue, plasma and target organ accumulation. The variability of different ODN with respect to size, sugar or backbone modifications may lead to subtle or more notable changes in pharmacokinetic properties such as plasma circulation time, tissue or disease site uptake, toxicity, as well as ultimately, efficacy. 29 o I -p-o-II o .(X ^Base O _ I S ~ ~ n _ ° \ ^ ° - -Base O O I H X - P - 0 • 3 ~ || w \ /Base O O I l - L C O - P - 0 O I H X H N - P - O -B • 3 w w '|| ~ ^ ^ ^ O ^ / B a s e D O • 3 w ' " , ' '|| w \ ^ O ^ ^ B a s e E O O Fig 1.4 Native and common backbone modifications of antisense oligonucleotides (A) Native phosphodiester, (B) Phosphorothioate, (C) Methylphosphonate, (D) Phosphotriester, (E) Phosphoroamidate. 30 Phosphorothioate ODN, with which this thesis is concerned, are well absorbed from parenteral sites, display rapid distribution and prolonged (in comparison to phosphodiester ODN) elimination from tissue distribution sites, and are much more resistant to nuclease degradation (although some exonuclease chain-shortening of the antisense does occur). However, phosphorothioate ODN have other characteristics which may not be desirable, including a high capacity binding for proteins, in particular, albumin and alpha-2 macro globulin (Brown et al., 1994), are not orally available and are unable to cross the blood brain barrier (Nicklin et al., 1998). Similar to antisense oligonucleotides in both composition and specificity, are catalytic R N A molecules called ribozymes (Christoffersen & Marr, 1995; Probst, 2000). These molecules can be engineered to bind to specific R N A sequences and cleave them by one of two means; with a transesterification or a hydrolysis mechanism. They can also be designed to splice R N A and create new gene functions. Ribozymes tend to be much larger than antisense ODN due to the folding required to generate the catalytic sites. Thus, many nucleotides are in loops that do not interact with the target RNA. These molecules are being tested in several areas as therapeutics, including viral diseases such as HIV, and cancer (Christoffersen & Marr, 1995). 31 1.3.5.2 Mechanisms of action There are several proposed (but largely still theoretical) mechanisms of antisense oligonucleotide action in the inhibition of protein translation (Figure 1.5). In some cases, the mechanism of action relies on the specific site within the mRNA that is targeted. For example, i f an ODN is directed toward a sequence of pre-mRNA that is required for splicing, the mechanism of action would be described as being that of inhibition of the production of mature mRNA by sequence specific inhibition (Crooke, 1999). Other mechanisms of blocking the processing of pre-mRNA include the blocking of polyadenylation of the 3'- terminus or 5'-capping of pre-mRNA with methylguanosine (Crooke, 1999), which are necessary for stabilisation of the RNA, and transport into the cytoplasm (Schlingensiepen & Schlingensiepen, 1996). In a similar fashion, many ODN are targeted toward the coding region of mRNA, and are thought to sterically block the progression of the ribosome, and therefore halt translation (Schlingensiepen & Schlingensiepen, 1996). Initiation of protein translation may be blocked by antisense targeted toward the translation initiation codon, the promoter region, or a region close to either of these sites. In addition to these site specific inhibition mechanisms, both phosphodiester and phosphorothioate antisense oligonucleotides are thought to rely on the action of the ubiquitous enzyme, RNaseH. This enzyme specifically recognises a DNA-R N A hybrid, and cleaves the R N A portion of that hybrid. Sugar modified ODN such as 2'-fluoro or 2'-0-methyl, or methylphosphonate ODN are not thought to act as substrates for RNaseH activity (Crooke, 1999), additionally, RNaseH is believed to only cleave the R N A strand i f five to eight contiguous base pairs of an R N A - D N A duplex form (Woolf et al., 1992). 32 Figure 1.5 Proposed mechanisms of action of antisense oligonucleotides in cells: inhibition of processing of pre-mRNA and transport into the cytoplasm or complimentary base pairing with mature mRNA in the cytoplasm, leading to lack of protein production by steric hindrance, blockage of ribosomal machinery or RNaseH activation. 33 As antisense oligonucleotides are a relatively new form of therapy, there is still some disagreement in the literature as to what constitutes adequate control(s) for these agents. It has been argued that no suitable controls may be defined to conclusively demonstrate that a given biological effect results from a sequence specific ODN mechanism (Neckers & Lyer, 1998). These authors state that although appropriate controls include a sense control, a scrambled control and a mismatch control, these sequences may not contain biologically active motifs found in the antisense sequence. Further, should mismatch controls be incorporated into experiments, these sequences may not control for hybridisation specificity of the active ODN. They conclude by stating that each individual antisense sequence may be considered a unique drug, for which no adequate control exists. Other authors disagree with this viewpoint, listing controls other than altered O D N sequences, including agents such as small molecule inhibitors of protein function, monoclonal antibodies, expression of dominant negative proteins or the use of gene knockouts (Giles, 2000). The work presented in this thesis includes the use of a single antisense control sequence, specifically the scrambled control for the active anti-HER-2/neu ODN. It was felt that as an experimental end point, the definitive demonstration of lack of protein expression in the case of active, but not control, sequence antisense, should provide a measure of satisfaction that the antisense was having a sequence specific effect, an end point considered to be imperative by Stein and Krieg (1994). In addition, bearing in mind the above arguments, it was unclear as to whether additional sequence controls would strengthen the case for specificity of action. A later chapter (Chapter 6) does include the use of a monoclonal antibody against HER-2/neu as a control treatment, which, as stated above, may indeed be a more relevant control for this type of experiment. 34 1.3.5.3 Pharmacokinetics and Toxicology Antisense O D N are presumed to be taken up by cells through endocytosis, and accumulate in an endosomal-lysosomal compartment. Following this stage, some of the O D N is able, through an as of yet undefined mechanism, to gain access to the cytosol and then the nucleus of the cell, in which places it is able to interact with both cytosolic and nuclear R N A species. A s this thesis outlines experiments performed with phosphorothioate O D N , this section of O D N pharmacokinetics wi l l have the same focus. Following either intravenous or subcutaneous injections or infusions, phosphorothioate O D N have a biphasic plasma half life, with a T ' / i a (distributional phase) of a few minutes to one hour, depending on dose, and a T'AP (elimination phase) of up to several hours (Juliano et al., 1999). Phosphorothioate oligonucleotides circulate in plasma bound to plasma proteins that protect them from filtration (Levin, 1999). O D N are distributed widely throughout the tissues of the body (excluding the central nervous system and testes), primarily in the liver and kidney, accounting for the majority of O D N clearance (Levin, 1999), then to a lesser extent in the spleen, bone marrow and lymph nodes, then other tissues. Liver and kidney concentrations of O D N appear to saturate as the dose is increased; however this may be attenuated by prolonging the infusion time, suggesting that saturation is a function of the dose rate as well as the dose (Levin, 1999). Unlike traditional small molecule drugs, which are mainly metabolised in cytochrome P-450 dependent pathways, O D N are metabolised by nucleases. Although 3'-exonuclease removal of bases is the predominant pathway for metabolic degradation, both 5'- and 3'-exonuclease excision may occur in tissues, but endonuclease digestion is not commonly observed (for phosphorothioate O D N ) (Levin, 1999). Exonuclease digestion of at least a single nucleotide of intravenously administered, 35 naked phosphorothioate O D N has been observed in up to 40% of the administered dose after only 5 minutes, however this does not increase substantially over time, such that after 4 hours, the predominant form of O D N is the parent compound, and not the shortened chains (Glover et al., 1997). Elimination of antisense oligonucleotides is mainly via the kidneys (Juliano et al., 1999). Toxicity of antisense oligonucleotides may be sequence dependent or independent. Sequence dependent effects are for the most part, centred on sequences containing G quartets, and C p G motifs. O D N containing G quartets (4 contiguous guanosine bases) are known to bind proteins, and are able to form tetrameric structures, altering the activity of the antisense (Levin, 1999). Unmethylated C p G motifs, especially when flanked 5' by two purines and 3' by two pyrimidines, are known to be potently immunogenic, with the ability to induce B cell proliferation and release of cytokines IL-6, -12 and interferon-gamma (Krieg, 1995). Rodents are particularly sensitive to the effects of the inclusion o f a C p G motif in phosphorothioate O D N . In addition to mitogenic activation of B cells, this motif induces a dose-dependent splenomegaly, lymphoid hyperplasia and diffuse multi-organ mixed mononuclear cell infiltrates. At doses above 10 mg/kg, the kidney, heart, lung, thymus and pancreas have infiltrates o f monocytes, lymphocytes and fibroblasts. When plasma concentrations exceed 40 - 50 u.g/ml, there is a transient activation of complement, primarily through the alternative pathway, which is thought to be mediated through the binding of O D N to Factor H , a negative regulatory factor of this pathway (Levin, 1999). A t this dose, phosphorothioate O D N are able to transiently inhibit clotting times, as measured by the activated partial thrombin time. This effect is also highly transient, and 36 parallels the concentration of ODN in the plasma (Henry et al, 1997). It is important to note however, that these toxicities are observed at doses that are substantially higher than those required to achieve a pharmacologic response, and indeed, are higher than those used in clinical trials (Levin, 1999). The above discussion has focused primarily on the in vivo toxicities of ODN, in which carrier systems are not typically utilised. In vitro applications however, require the use of helper lipids to assist the ODN in entering the cultured cells. The lipid systems most commonly used for this purpose consist of a cationic lipid that is able to bind to negatively charged ODN, and a helper lipid which aids in binding of the lipid/ODN complex to the cell surface, and mediates internalisation of the complex. This thesis explores the use of carrier systems in in vivo applications to determine i f the encapsulation of ODN into liposomes will be beneficial in increasing the exposure of tumour tissue to the ODN, and also to increase internalisation of that ODN once the target site is achieved. 1.3.5.4 ODNs Targeting HER-2/neu As HER-2/neu presents such an attractive target for anti-cancer therapeutic strategies, it is not surprising that several groups have developed ODN targeting HER-2/neu mRNA. Published references to ODN mediated downregulation of HER-2/neu appear in the literature as early as 1994 (Colomer et al.), in which up to a 60% reduction in HER-2/neu protein levels was achieved in an in vitro setting with several breast cancer cell lines. Further in vitro experiments with the human breast carcinoma cell lines SK-BR-3 (Vaughn et al, 1995), BT474 and T47D (Liu & Pogo, 1996; Roh et al, 1998) confirmed these early 37 results, using antisense targeting different areas of the HER-2/neu mRNA, and achieving as high as 90% protein inhibition as well as growth inhibition and accumulation of cells in the G l phase of the cell cycle. More recently, HER-2/neu targeting antisense has been taken into in vivo models, with Uedo et al. (1999) demonstrating an inhibition of carcinogenesis in a rat model following treatment with anti-HER-2/neu ODN. These authors surmised that the inhibition was due to the combination of decreased cell proliferation and increased apoptosis resulting from ODN treatment. Roh et al (1999; 2000(a); 2000(b)) have performed extensive work with HER-2/neu targeted ODN, including in vivo studies with human breast cancer xenografts in nude mice. This work supports earlier experiments in that cells in in vitro studies are arrested in G0/G1 and appear to undergo apoptosis, and importantly, in vivo work showed synergism between the ODN treatment and cytotoxic (doxorubicin) therapy. They conclude that by downregulating the HER-2/neu protein, the tumour cells are rendered more sensitive to the actions of cytotoxic agents, therefore making anti-HER-2/neu targeted ODN a potentially powerful therapeutic. These studies were all performed with xenograft models which had a high level of HER-2/neu overexpression, and indeed, Liu and Pogo (1996) suggest that anti-HER-2/neu targeted ODN should only be effective in breast cancer cells with elevated HER-2/neu. The work presented in this thesis expands upon the work by the above researchers by the inclusion of xenograft models with low expression of HER-2/neu. I will demonstrate the ability of a HER-2/neu targeted ODN to be efficacious in such a model, thereby 38 demonstrating the potential for a beneficial therapeutic effect in the absence of elevated expression levels. 1.4 Liposomes Liposomes are phospholipid vesicles with bilayer membranes which, depending upon the lipid composition, are relatively impermeable to both amphipathic and highly water soluble molecules at physiological temperature. When liposomes were first explored as potential delivery vehicles, they were formed from synthetic and naturally occurring phospholipids with sufficient cholesterol to stabilise the structures in vivo. In general, these formulations were eliminated relatively quickly by the cells of the mononuclear phagocyte system (MPS). Later modifications to liposomes included the addition of a moiety that conferred steric stabilisation such as the ganglioside GM1 or the commonly used polyethylene glycol (PEG)-conjugated lipids. Liposomes with these additions to the formulation are eliminated less rapidly by the MPS due to the exclusion of macromolecules such as opsonins from the peri-liposomal space, and have different bio-distribution patterns. It has more recently been suggested that the primary effect of steric stabilisation is due to elimination of surface-surface interaction that can lead to liposome aggregation or liposome-cell interactions (Johnstone et al., 2001). This section of the introduction will review liposomes as being a clinically viable, well characterised delivery system that has been used extensively to deliver O D N in in vitro settings, as well as preliminary in vivo research. 39 1.4.1 Liposomes as Drug Delivery Vehicles As liposomes have an internal aqueous space, it is possible to entrap therapeutic agents such as small molecule drugs or diagnostic dyes, within them. The encapsulation of drugs in liposomes alters the pharmacokinetic and biodistribution properties of the entrapped agents (Gabizon et al., 1994; Gabizon et al., 1998). The extent and character of these alterations depends on the lipids used in the liposome formulation (acyl chain length and charge), as well as the overall size and lipid dose of the administered liposomes (reviewed in Tardi et al., 1996). For example, the inclusion of cholesterol in liposome formulations has been shown to prevent lipoprotein-induced vesicle destabilisation and concomitant release of the encapsulated drug (Scherphof et al., 1978; Kirby et al, 1980). This leads to increased circulation longevity of that drug and enhanced bioavailability. When negatively charged phospholipids, such as phosphatidylserine, phosphatidic acid or phosphatidylglycerol are included in the formulation, liposomes are rapidly cleared from the circulation by the cells of the mononuclear phagocyte system (Schroit et al., 1985); however, rapidly cleared formulations may still affect drug parameters such as, in the case of doxorubicin, reduced cardiotoxicity (Rahman et al, 1990). There is a need to balance the enhanced circulation longevity with the ability to make the drug bioavailable, or, to design liposomes such that once they have reached the target site, whether it be a site of inflammation, tumour or other, the drug is released, resulting in a local depot of drug where required. Liposomes represent ideal drug delivery systems for tumours as the microvasculature in tumours is typically discontinuous, having pore sizes varying between 100 to 780 nm 40 (Hobbs et al., 1998). This allows vesicles the size of liposomes (approximately 100 - 150 nm) to extravasate within the tumour tissue to provide locally concentrated drug delivery, a primary role of liposomal formulation. Accumulation within the tumour tissue has the additional benefit of decreased drug accumulation in non-target tissues, and hence lower drug associated toxicities. Drugs may be loaded into liposomes in several manners. A common, but relatively inefficient loading technique is that of passive loading. In this technique, lipids which have been dried down to a film are hydrated with a buffer solution containing the drug that is to be encapsulated. This will typically result in up to 20% of the drug in solution being encapsulated in the liposomal interior, but may also be substantially lower than this. Alternative mechanisms of drug loading have therefore been devised (refer to sections 1.4.2.1 and 1.4.2.2 for examples). Once loaded, drugs may partition into the aqueous interior of the liposome if they are hydrophilic, or may be sequestered within the lipid bilayer, in the case of hydrophobic drugs (Fig. 1.6). 1.4.2 Liposomes as Drug Delivery Vehicles in the Treatment of Solid Tumours, in Particular, Breast Cancer The anthracycline antibiotic doxorubicin is one of the most important and widely used anti-tumour agents. Produced by the fungus Streptomyces peucetius var. caesius, doxorubicin has broad activity against a range of human neoplasms, including leukemias and lymphomas, breast, ovarian, stomach, bladder and bronchogenic carcinomas, and soft tissue sarcomas. 41 Fig 1.6 Partitioning of drugs (D) into the lipid bilayer or the central aqueous core of liposomes. (A) Hydrophobic drugs partition into the lipid bilayer, (B) Hydrophilic drugs sequester in the central aqueous core. 42 Currently, anthracycline containing combination therapy is the standard of care for treatment of breast cancer, although there is much excitement in the areas of the taxanes and monoclonal antibodies (Lebwohl & Canetta, 1999). These are attractive agents for combination therapy due to minimal overlapping toxicities with anthracyclines. In clinical trials of advanced breast cancer, the objective response rates for conventional doxorubicin treatment are approximately 42% and 30% in untreated and previously treated patients, respectively (Henderson & Canellos, 1980). The clinical activity of doxorubicin, as with other anthracyclines, is limited by cardiomyopathy. This, in turn, is related to the total cumulative dose of the drug and is principally due to free radical damage to myocytes (Ferrans et al., 1997; Muller et al, 1997). Damage occurs due to build-up of the anthracycline within the heart muscle, D N A intercalation and free radical formation, which result in damage to myocardial cells. Pathologically, this is caused by lysis of both the thick and thin myofilaments leading to myofibrillar loss, and dilatation of the sarcoplasmic reticulum, resulting in vacuolisation of myocytes (Ferrans et al., 1997). The typical acute dose limiting toxicity for doxorubicin is that of a high incidence of myelosuppression, typically leukopenia and thrombocytopenia. This is usually a transient phenomenon, with the leukocyte nadir observed by 10 to 14 days post treatment, and recovery to normal cellularity by 21 days (Cancer Care Handbook, 1999). In severe cases, this may lead to neutropenic fever and sepsis, requiring hospitalisation. Mucositis, including both stomatitis and pharyngitis which may lead to ulceration has been associated 43 with continuous infusion of doxorubicin. An almost universal occurrence of alopecia accompanies treatment with conventional doxorubicin (Dorr & Fritz, 1982). This is typically evident 3 to 4 weeks post initiation of therapy, with re-growth occurring 2 to 5 months after therapy, or in some cases, during therapy. Acute nausea and vomiting are very common, with many patients requiring the use of antiemetics to control these side effects. As well, rare doxorubicin induced toxicities include ulceration and necrosis of the colon, and neuropathy. The primary aim of doxorubicin encapsulation in liposomes has been to decrease non-specific organ toxicity. Liposomes are able to direct the doxorubicin away from sites with tight capillary junctions such as the heart muscle, and the gastrointestinal tract. They are instead accumulated within organs rich in cells of the mononuclear phagocyte system (MPS). It should be noted however, that when liposomes contain doxorubicin the cells of the MPS are affected such that they can no longer accumulate liposomal systems. Although this "MPS toxicity" has been noted in animals, as measured by a change in liposome elimination profiles (Parr et al., 1993), similar effects have not been observed in humans. Regardless MPS toxicity may be of some concern in cases with increased risk of opportunistic infections such as in AIDS patients, especially due to depletion of liver macrophages, which have been shown in animal models to take up to 2 weeks to be restored (Daemen et al., 1995). Interestingly, the first formulation of doxorubicin approved for use in humans was based on its use in the treatment of Kaposi's Sarcoma, an AIDS related cancer. Liposomal formulations containing polyethylene-glycol (PEG) modified 44 lipids have even greater blood circulation time as a result of diminished liposome-cell interactions, however these formulations have also been shown to exhibit MPS toxicities. Regardless of whether enhanced circulation lifetime of a liposomal formulation of doxorubicin is due to subtle changes in lipid composition, inclusion of doxorubicin which affects MPS cell function, or the use of PEG-modified lipids, increased circulation longevity of the liposomes allows enhanced extravasation across the leaky endothelium of solid tumours, providing a local depot for drug release (Papahadjopoulos et al, 1991). In order for liposomal formulations of doxorubicin to present an advantage in therapy over the free agent they must be able to enhance anti-tumour activity at the maximum tolerated dose without a concurrent increase in toxicity. The increased accumulation of liposome-encapsulated doxorubicin in tumours, and decreased accumulation in sensitive non-target organs can enhance the therapeutic index of the drug, but it is argued here that the primary benefits occur as a consequence of reduced toxicity rather than enhanced therapeutic activity. It is important to recognise that such conclusions have been reached on the basis of data obtained in animal models of cancer and current Good Laboratory Practices (cGLP) toxicity studies. The real benefits of liposomal encapsulation of doxorubicin on the drug's therapeutic index will only be appreciated following extensive clinical studies. In this regard, it is notable that arguments in support of clinical use of the two approved liposomal formulations of doxorubicin are different (reviewed in Waterhouse et al., 2001). Advocates of one formulation highlight improved anti-tumour activity while advocates of the other emphasise reduced toxicity. 45 The above discussion of liposomal doxorubicin has highlighted the fact that liposomal drug formulations are pharmaceutical^ viable therapeutics. Importantly, recent efforts have combined liposomal doxorubicin with trastuzumab in the treatment of patients with HER-2/neu overexpressing breast cancer. It is hoped that the combination of trastuzumab with the liposomal doxorubicin will ameliorate the cardiotoxicity noted when trastuzumab is combined with free doxorubicin, while maintaining the increased efficacy of the combination over that of doxorubicin as a single agent (Slamon et al., 1998). Examination of the properties of liposomal doxorubicin assist in illustrating the benefits and limitations to lipid-based drug delivery as applied to formulations of antisense oligonucleotides, which will be explored further in the body of this thesis. 1.4.2.1 Caelyx (Doxil) The pegylated (Stealth) liposomal doxorubicin has been in use in the U.S. since 1995 under the brand name Doxil and is marketed in Canada and Europe under the name Caelyx. This formulation is comprised of hydrogenated soya phosphatidylcholine (HSP) and cholesterol (Choi). The Doxil liposomal formulation also contains the hydrophilic polymer polyethylene glycol. The PEG moiety, linked to a distearoyl phosphatidylethanolamine (DSPE) lipid anchor, extends out from the liposomal surface to create a hydrated barrier or shield which protects the liposomes (see section 1.4 for a review of the benefits of PEG inclusion in liposomal formulations). The Doxil formulation is loaded with doxorubicin in an active fashion using a (NH 4) 2S04 gradient across the lipid bilayer to draw the drug to the liposomal interior, where it then precipitates with sulphate (Fig. 1.7). These liposomes 46 Figure 1.7 Ammonium sulphate gradient method of loading of doxorubicin into liposomes. Liposomes are formed in the desired concentration of ammonium sulphate. An ammonium sulphate gradient across the bilayer is formed by dialysis. Doxorubicin hydrochloride is added to the liposomal dispersion, crosses into the interior of the liposome, where it is trapped in a gel-like precipitate form. 47 have an altered distribution pattern as compared to liposomes without PEG modified lipids. They have a tendency to exhibit decreased accumulation in the liver and spleen and a reduced plasma elimination rate. 1.4.2.2 Myocet The second liposomal doxorubicin formulation is comprised of egg phosphatidylcholine (EPC) and cholesterol (Myocet, given community marketing authorisation from the European Commission in August, 2000 for the treatment of metastatic breast cancer). In pre-clinical and phase I studies, Myocet was shown to have a circulation half life of approximately 3 times that of free doxorubicin (Cowens et al., 1993), although these data were derived from a large range of a plasma elimination half lives. In addition, the encapsulation of doxorubicin in liposomes was associated with decreased toxicity of the drug, notably decreased cardiomyopathy and gastrological toxicity (Cowens et al., 1989; Kanter et al, 1993). The Myocet formulation also employs an active entrapment process for doxorubicin (Fig. 1.8): liposomes are prepared in an acidic buffer (pH 4.0 300 m M citrate), then sodium carbonate is added to increase the pH outside the liposomes to approximately 7.3. When these liposomes are combined with doxorubicin and heated briefly, the doxorubicin crosses the lipid bilayer and becomes protonated in the liposome interior. Once doxorubicin becomes protonated it has difficulty crossing the lipid bilayer which results in trapping efficiencies of over 99% (Mayer et al, 1986). 48 < Q ft + X + Q t i 34 X o 1 & . —1 T3 O oo cd O ed g - H n oo o '5 8 -ts cd &. C u ^3 l-l ^ u u Q ^ £ oo ^ CD Q c o ^ ch a « da t3 _ T 3 o CN w X >/-> m 3 C J 4 3 ffl ~ ^ < H cj < Q C J c U PQ ^ 3 o ••s s cd T3 C oo cd cd -4—» o T 3 Q CD J 3 g ^) oo J3 2 OH ^ SO cj r—. C ^ 4 3 oo •a cd CJ 43 .3 £ O A 3 o 52 _cd J3 (3 cd » c w cj 00 =S O T3 Q cd 1 1 cj CJ s> >-CJ =<3 ^ U S3 4 3 ' ' r, % ^ CD Q (6) 9Z(s j n o m n i Q s § 1 'a I oj O fcj -a o cj Q =1 ^ 2 PQ c 2 80 days for control groups) over which the tumours grew. For this reason, it was important to establish a different breast cancer xenograft model with a more reasonable growth rate. Thus, the next chapter of this thesis will focus on the characterisation of the MDA435/LCC6 cell line recently described 111 by Leonessa et al. (1996). In conclusion, I have demonstrated that the H2A-ODN antisense sequence used in this report is effective in the downregulation of the HER-2/neu protein both in vitro and in vivo, and further, that this downregulation is maintained when combined with doxorubicin treatment. In a model of low pi85 expression, a liposomal formulation of this antisense was effective in reducing the rate of tumour growth in SCID/Rag2m mice. I have also shown, however, that in at least one breast tumour model exhibiting high pi85 expression, treatment with liposomal anti-HER-2/neu antisense in the absence of doxorubicin treatment resulted in an increased rate of tumour growth, despite significant reductions in HER-2/neu expression as measured by immunohistochemistry. This result is surprising and somewhat disappointing, however it may be a reflection of the challenges that will be faced when developing any agent selectively targeting a single protein associated with tumour progression. It is likely that multiple proteins must be targeted simultaneously in order to dysregulate the tumour cell's machinery sufficiently to alter its own inherent capacity to survive. Future studies may need to investigate levels of proteins downstream of HER-2/neu as well as proteins where increased expression has been associated with altered HER-2/neu, such as vascular endothelial growth factor (VEGF)(Petit et al., 1997). 112 Chapter 4 CHARACTERISATION OF THE MDA435/LCC6 HUMAN BREAST CARCINOMA XENOGRAFT TUMOUR MODEL 4.1 Introduction Many of the approximately 11% of all North American women and 1% of all North American men who will develop breast cancer will also develop metastatic disease. Approximately 35% of those who develop breast cancer will ultimately succumb to their disease. For this reason, there is an obvious need to develop more efficacious treatment strategies. Hand in hand with the need for novel strategies is a requirement for better models in which to test these new ideas, such that they may be more accurately assessed for efficacy in human disease. To be able to develop a good model, researchers must understand the diseases with which they are working, know which pathways or proteins are activated or deactivated, and seek to characterise their models with respect to these pathways and proteins. In this way, the models used will be more relevant to the disease, and results garnered with novel therapeutics will be more easily interpreted as to the mechanism of action. It was with this in mind, that I proceeded to characterise the MDA435/LCC6 model of human breast cancer for general in vitro and in vivo growth characteristics, sensitivity to anti-cancer drugs and expression of several proteins relevant to breast cancer. 113 Research has identified several specific prognostic and predictive factors to identify patients at high risk for more aggressive breast disease, metastasis and for recurrence of disease in order to combat the high mortality associated with these factors. Several studies have demonstrated, for example, that an increase in cytosolic levels of vascular endothelial growth factor (VEGF) in tumour tissue samples is indicative of poorer prognosis for patients with node-negative breast carcinoma (Anan et al, 1998; Gasparini, 1997). If a tumour is to exceed 1-2 mm in diameter, it must be able to initiate the formation of an adequate blood supply to receive sufficient oxygen and nutrients. The process by which this neovascularisation occurs is known as angiogenesis (Folkman, 1992), and the extent of tumour angiogenesis is now widely accepted as a prognostic factor in several cancer types (Araya et al, 1997; Axelsson et al, 1995; Fox, 1997; Hermann et al, 1996; Weidner et al, 1992), with higher levels being indicative of more aggressive disease. A possible link between V E G F and another marker of poor prognosis in both node positive and node negative breast cancers has also been postulated, namely, the overexpression of the protein product of the HER-2/neu oncogene (Petit et al, 1997). Research into a mechanism for HER-2/neu action has shown that overexpression of HER-2/neu in cell lines increases cell migration, an important step in metastasis formation (Verbeek et al, 1998). It has also been suggested that there may be a link between HER-2/neu expression and the up-regulation of anti-apoptotic Bcl-2 and Bc l -XL proteins, and suppression of tamoxifen-induced apoptosis (Kumar et al, 1996). More recent research however, has linked the absence of bcl-2 expression with a poorer overall survival (Bhatavdekar et al, 2000) or conversely, that increased bcl-2 conferred decreased risk (Jager et al, 2000). In light of 114 these observations, I felt it was important to characterise the human breast cancer xenograft model, MDA435/LCC6 to determine if it expressed the above markers. The final protein assessed in this study was Cathepsin D. This protein has been proposed to play an important role in breast cancer invasion and metastasis (Han et al, 1997). Because of this, patients with elevated cathepsin D levels, defined as those tumours with clearly detectable cathepsin D levels (greater than or equal to 10% positive cells)(Gonzalez-Vela et al., 1999) have an adverse prognosis (Solomayer et al., 1998). As a protease, cathepsin D is likely to be involved in the degradation of extracellular matrix and therefore, tumour cell spread in circulatory and lymphatic systems. This same mechanism may also contribute to distant metastases. There is a trend of increased cathepsin D in oestrogen receptor positive breast cancer, but there is disagreement in the significance of this tumour cell associated cathepsin D. Some researchers dispute the relation to other clinicopathologic parameters (Gonzalez-Vela et al., 1999), while others find significant correlation between Cathepsin D positive staining in tumour cells, and higher frequency of recurrence and shorter disease free survival (Lah et al., 2000). The discovery of new prognostic and predictive markers for breast cancer is exciting, and has opened the way for a new generation of targeted therapeutics for treating this disease. In order to exploit these markers, as stated above, researchers require predictive in vivo models that can be used to i) establish the role of these markers in breast cancer progression and ii) screen agents directed against these targets for therapeutic activity. With regard to genetic targets, those cell lines used to establish models must also express relevant targets 115 and or pathways. The development of a human xenograft ascites model fulfilling the first two of these requirements has been achieved in the MDA435/LCC6 model (Leonessa et al., 1996). This model has a reproducible growth curve in mice bearing these tumours, and responses to cytotoxic drugs are similar to those seen in many human breast cancer patients (Leonessa et al., 1996). An additional feature that supports the use of this cell line is its increased growth rate as compared to the parental line, which necessitated in vivo experiments extending past 100 days (section 3.2.2). The cells may also be easily maintained in vitro, thus allowing rapid assessment of drugs both in vitro and in vivo. The herein described characterisation of MDA435/LCC6 demonstrated that in vitro, these cells display a three-fold increase in VEGF expression over parental MDA-MB-435 cells. This pattern is also readily apparent in vivo as determined by immunohistochemical staining. The model also shows readily detectable levels of both bcl-2 and cathepsin D expression, therefore making the MDA435/LCC6 a useful model for therapies that target these proteins. 4.2 Results 4.2.1 In Vitro and In Vivo Growth Characteristics The purpose of these studies was to characterise the M D A 435/LCC6 tumour cell line to determine its value as a model of human breast cancer. Part of this assessment included the in vitro growth characteristics of M D A 435/LCC6 as compared to the parental M D A - M B -435 line. As seen in Figure 4.1 A, M D A 435/LCC6 was a more rapidly growing cell line in 116 vitro, with the growth rate from day three onward being more rapid than MDA-MB-435. The doubling time for LCC6 cells from day 3 onward was approximately 20 hours, while that of MDA-MB-435 was 72 hours. This suggests that the LCC6 variant may be intrinsically more aggressive (as defined by proliferation rate) than the parental line. Having confirmed a more aggressively growing phenotype in vitro, I sought to determine whether this was also the case in vivo (Figure 4.IB, C & D). When grown as a subcutaneous tumour on the back of female SCID/RAG2m mice, the in vivo doubling time of the LCC6 variant was roughly 3 fold faster than the parental MDA-MB-435 cell line, a result that is consistent with the data in Figure 4.1 A. There are two features of this in vivo growth curve worth noting. First, the LCC6 cell line tumour growth becomes grossly apparent on day 20 to 25 post inoculation, while the parental growth is not apparent until day 60 to 70. Second, once growth is initiated, the tumour doubling time from 0.2 to 0.4 grams is approximately 12 days for tumours arising from the LCC6 cell line and 15.5 days for tumours arising from the parental cell line. MDA435/LCC6 also grows as a subcutaneous tumour in the mammary fat pad. Tumour growth curves from tumours grown in this area in female SCID/Rag2m mice are shown in Figure 4.1C. These tumours were followed out to day 35 post-tumour cell inoculation, by which time they had reached approximately 0.2 g. This form of inoculation was not pursued further due to the frequency with which tumours in this location became ulcerated. 117 (uiu 0Z9) sjmn aoueqjosqv 118 4.2.2 Cytotoxicity of Taxol, Vincristine and Doxorubicin Against MDA435/LCC6 To further the characterisation of the LCC6 variant, I tested the susceptibility of both the parental and variant lines to several commonly used anticancer agents, namely taxol, doxorubicin and vincristine. Results from these in vitro tests are shown in Figure 4.2A and B, and are compared to the cytotoxicity profile exhibited by the parental MDA-MB-435. It was important to demonstrate that this newly derived breast cancer cell line maintained a response to common cytotoxic drugs as is demonstrated by MDA-MB-435. As is highlighted in Figure 4.2, the IC50 values for the two lines are not substantially different, indicating neither increased resistance nor sensitivity to these drugs. The growth of LCC6 as an ascites tumour, with total number of viable cells excluding any red blood cells in the peritoneum, plotted over time is shown in Figure 4.ID. The typical doubling time when grown as an ascites tumour is less than four days once the tumour becomes established (after day 14). Between days 21 and 28, the peritoneal cavity becomes so crowded with cells that the tumour cells start to die due to nutrient depravation, such that by day 28, the bulk of the tumour cells are non-viable. 4.2.3 Markers of Aggressive Breast Cancer Having demonstrated a more rapidly and aggressively growing tumour both in vitro and in vivo, an assessment of the levels of several proteins that have been associated with more aggressive breast was completed. I screened the two cell lines for V E G F , HER-119 2/neu and bcl-2. Cathepsin D expression in 5 um paraffin embedded sections of subcutaneous tumours was also determined. Initial assessment of V E G F was in supernatants collected from confluent cells. These data, shown in Figure 4.3A, indicate that there was almost a 3-fold increase in VEGF levels of MDA435/LCC6 cell supernatant when compared to levels produced by the parental line. The VEGF levels in the ascites fluid collected from mice bearing ascitic MDA435/LCC6 tumours (Figure 4.3B) as measured by ELISA, increased as the tumour progressed. Between day 21 and day 28 there was a decline in the number of viable tumour cells in the peritoneum and this corresponded with the decline in peritoneal V E G F levels. As indicated in the insert to Figure 4.3B, serum V E G F levels in these mice steadily increased over the course of tumour progression with a more dramatic increase after day 21. This may have been due, in part, to the progressive breakdown of vascular beds in the peritoneum as judged by a significant increase in red cell contamination in the peritoneal cavity. It should be noted that plasma VEGF was not detected in animals bearing solid tumours derived from the LCC6 cell line. To determine i f the increased VEGF level was also seen in solid tumours, I immunohistochemically stained tumour sections taken from mice bearing M D A -MB-435 and MDA435/LCC6 tumours for VEGF. The expression patterns for VEGF in the two tumours is shown in Figure 4.4, where sections derived from A431 tumours were used as a positive control for V E G F expression (panel A). V E G F protein staining of LCC6 tumour (panel E) was less than that observed for A431, but greater than the parental M D A -MB-435 (panel C). HER-2/neu expression was also determined using several methods. Western analysis of 121 MDA MB 435 MDA435/LCC6 600.0x103 r Days Post Inoculation Figure 4.3 (A) In vitro V E G F levels in supernatant collected from confluent M D A - M B -435 and MDA435/LCC6 cells, measured by ELISA. Results in pg/ml ± standard error of the mean, and (B) in vivo V E G F levels from peritoneal fluid of MDA435/LCC6 derived ascites tumours, collected 0, 7, 14, 21 or 28 days post tumour cell inoculation. Inset shows serum V E G F concentration from these same animals. Ascites and serum were assayed by ELISA; results show pg and pg/ml, respectively, ± standard error of the mean. 122 parental and MDA435/LCC6 was completed using the highly overexpressing SK-BR-3 as a positive control. These data suggested that there was a 1.3 fold increase in the level of HER-2/neu protein in the LCC6, as compared to parental. However, MDA-MB-435 HER-2/neu expression was 0.7% that of SKBR-3, and LCC6 HER-2/neu expression was 1 % that of SK-BR-3 as determined by flow cytometric analysis. These cell lines may be considered low expressers of p i 85. As may be seen in panels D (MDA-MB-435) and F (MDA435/LCC6) of Figure 4.4, when grown as a solid tumour neither cell type expressed sufficient levels of pi85 for detection by immunohistochemistry. Analysis of bcl-2 expression levels by flow cytometry (Figure 4.5, panels A M D A - M B -435, and B MDA435/LCC6) shows that these cells express equivalent levels of bcl-2. Western blot analysis (Figure 4.5, panel C), with the strongly bcl-2 expressing MCF-7 as a positive control, indicated an approximate two-fold increase in bcl-2 levels in MDA435/LCC6 as compared to parental as determined by densitometry. Immunohistochemical staining of solid tumour sections (Fig. 4.6) indicated that both MDA-MB-435 and MDA435/LCC6 have readily detectable levels of bcl-2. Tumours with increased expression of the oestrogen regulated lysosomal protease cathepsin D have been correlated with increased invasion and metastasis of breast cancer. It was apparent from immunohistochemical staining (Figure 4.7) that MDA-MB-435 and MDA435/LCC6 express significant and approximately equivalent quantities of cathepsin D. How this relates to aggressiveness, likelihood for recurrence or metastatic potential has 123 100 pm 100 (im Figure 4.4 Immunohistochemical assessment of in vivo V E G F and HER-2/neu levels. 5 X 106 A431 (A), SK-BR-3 (B), MDA-MB-435 (C & D) or MDA435/LCC6 (E & F) were inoculated s.c. into female SCID/Rag2m mice. Tumours were collected in formalin before reaching 1 gram in size, paraffin embedded and stained immunohistochemically for V E G F (panels A , C and E) or HER-2/neu (panels B, D and F). Panels (A) and (B) are positive controls for V E G F and HER-2/neu staining, respectively. Visualisation was with diaminobenzidine, where brown areas indicate positive staining. Tissues were counterstained with Gill 's hematoxylin. 124 10° 10' CO < Q £ CO < Q CD O O • O actin bcl-2 Figure 4.5 Bcl-2 protein levels in MDA-MB-435 and MDA435/LCC6 tumour cells. (A) flow cytometric representation of bcl-2 protein levels in MDA-MB-435 cells; (B) flow cytometric representation of bcl-2 protein levels in MDA435/LCC6 cells; (C) western blot of bcl-2 protein levels in MDA-MB-435, MDA435/LCC6 and MCF7 (positive control) cells. 125 o CQ — CO u g | .-5 'I C- 00 CO < 3 — CL. CO CO s ' OH 00 s o CJ cn a I o «3 •a o co d GO i co o - * 3 cd £ *- - O c3 U ft V O " H rj U T 3 ^ o i l IT) ro < CO co CO > m .g CQ CO O ft CO X CO 1+3 J3 CO ? § 1 ? 00 H SS . £ o 01 O 'B J3 cd ft .2 a CN CO -op E C N 00 > •a x co *^ q '. . ***** ••'Si it, £7 4 *3 * "7,;. *- ^ 4 3 * 0 -CD C a f o C oo c3 ao o a .a oo r\ S3 Q . 2 c O oo CD , oo CD 3 cd o o 3 <2 -a & S3 CD o oo O oo oo CD 8 <2 i g, S 'a & oo P. CD u u 00 oo g 1 i—1 CD — CO oo 00 < -2 1 Q -o'o •a £ s a s r3 <: 5 CD ro 00 ea CJ 2 4 2 O g I < Q d 5 S3 S3 c 'a 03 g Q f^43 oo c ."S Q -a t CD I I I § g * .£ £ u T3 6 2 ^ S *J 2? 5 * E <2 s 127 not been addressed in this study. The MDA-MB-435 model is a much slower growing tumour compared to the MDA435/LCC6 model, thus the cathepsin D is not an indicator of aggression as measured by growth rate; however it may reflect the potential of these cells to metastasise. This may be demonstrated by the fact that the original MDA-MB-435 cell line was derived from a pleural effusion (Callieau et al, 1974) and the MDA435/LCC6 line is a metastatic variant of this line. 4.3 Discussion Assessment of the potential clinical suitability of novel anticancer therapeutics typically includes the determination of anti-tumour activity in preclinical animal models of cancer, most of which are mouse based. Ideally, the anti-tumour response to the agent being tested should assist in predicting activity when used in a clinical setting. An associated correlate to this ideal is that the model used should also mimic the human disease state or condition. Because of this, researchers continue to strive for improved models of human disease. Historically, commonly used models for rapid screening and efficacy tests of novel anti-cancer agents have been the murine ascites models, L1210 and P388. These models provided the benefits of reproducibility, speed and certain similarities of response to chemotherapy observed for the human disease. One of the predominant drawbacks to these models is that they are murine in origin. In addition, the molecular defects associated with these murine malignancies may not be of relevance when considering events in human disease. Consequently, when considering the development of drugs for the treatment of aggressive breast cancer, it would be ideal i f the model used expressed some or all of the attributes of an aggressive breast cancer in humans. 128 The MDA435/LCC6 ascites tumour model has been previously reported to be a fast growing and reproducible xenograft model of human cancer. This cell line is a direct derivative of the oestrogen receptor negative, invasive and metastatic MDA-MB-435 cell line (Cailleau et al, 1974; Price et al, 1990), and has been characterised as such by isotype and karyotype analysis (Leonessa et al, 1996). The cells were isolated from a spontaneous ascites that developed in a NCr nu/nu athymic nude mouse bearing a MDA-MB-435 tumour in the mammary fat pad. In this study, further characterisation of the MDA435/LCC6 cell line has been presented, focusing on markers associated with aggressive breast cancer in humans. Protein expression was characterised for VEGF, HER-2/neu, bcl-2 and cathepsin D, and correlating the parameters with tumour growth properties and response to chemotherapy. Many tumour cells express and secrete VEGF at high levels, with the effect of inducing the formation of new vasculature to promote tumour growth. V E G F is able to induce angiogenesis in two ways. Primarily, it does so by means of specific mitogenic activity on endothelial cells, but it also indirectly contributes to angiogenesis by its ability to induce vascular permeability with a potency some 50 000 times that of histamine (Collins et al., 1993), thereby influencing the egress of plasma proteins that are able to stimulate angiogenesis. The LCC6 variant has been shown here to have V E G F expression some three times higher than the parental MDA-MB-435. Therefore tumours derived from this cell line may have an increased ability to promote tumour vascularisation, with associated accelerated growth. This may be a contributing factor in MDA435/LCC6's intrinsically 129 more aggressive behaviour, as reflected in both in vitro and in vivo growth rates (Figure 4.1). Importantly, when combined with the ease of growth of this cell line as an ascitic model, a factor which may be attributable to overexpression of V E G F (Luo et al., 1998), the LCC6 variant may be seen as a powerful model for studies of tumour angiogenesis and anti-angiogenic agents. It has been suggested that the co-expression of molecular markers in neoplastic transformation functions as a significant prognostic indicator for recurrence, and that the number of co-expressed markers is directly related to this significance (Han et ah, 1997). I examined the expression of four markers of aggressive breast disease in this study; pi85, bcl-2, V E G F and cathepsin D. Neither the MDA-MB-435 or the MDA435/LCC6 tumours have detectable pi 85 expression in the solid tumours, as determined by immunohistochemistry, although western blot analysis data did indicate that these cell lines each express low levels of the protein. Although they are not strong overexpressers of bcl-2, both MDA-MB-435 and MDA435/LCC6 do express this protein, and therefore may be used as representative models for therapies targeting either the protein itself, or a component of the apoptotic pathway mediated by bcl-2. Although perhaps not significant when expressed singly, when the expression of cathepsin D is combined with p53, nm23 (not assessed) and HER-2/neu expression, there is an 80% agreement between positive expression of all of these markers, and recurrence of disease (increased from 47.1% when cathepsin D is not included) (Han et al, 1997). Both the parental MDA-MB-435 and the LCC6 variant have strong expression of this protein as 130 determined by immunohistochemistry. The LCC6 variant is an extremely versatile tumour model in that it grows more rapidly both in vitro and in vivo than the widely used parental MDA-MB-435 cell line and can also be employed as an ascites tumour model. In comparison, the parental line can only be grown as a solid tumour and expresses decreased levels of V E G F compared to the LCC6 line. The ease of growth as an ascitic tumour for MDA435/LCC6 cells lends the added benefit that cells can be easily removed from the peritoneum and assayed by various means, such as western blot or flow cytometric analysis. For the above reasons, the MDA435/LCC6 model of human breast cancer has the potential to be a powerful tumour model in the analysis of human response to breast cancer therapies. 131 Chapter 5 PHARMACODYNAMIC BEHAVIOUR OF LIPOSOMAL ANTISENSE OLIGONUCLEOTIDES TARGETING HER-2/NEU AND VASCULAR ENDOTHELIAL GROWTH FACTOR AS MEASURED IN THE ASCITIC MDA435/LCC6 HUMAN BREAST CANCER MODEL 5.1 Introduction The previous chapter of this thesis dealt with the characterisation of the MDA435/LCC6 cell line as a xenograft model of aggressive human breast cancer. It was shown that these cells grow well both in vitro and in vivo (Fig. 4.1), and importantly, will grow as both solid and ascitic tumours. Based upon the finding that the MDA435/LCC6 tumours express increased amounts of vascular endothelial growth factor (Figs. 4.3 & 4.4), it was reasonable to proceed to test antisense oligonucleotides targeted against this protein in the model. In addition, ODN targeted against the HER-2/neu protein were shown to be efficacious in the low HER-2/neu expressing tumour model, MDA-MB-435. It followed then, that this treatment modality may be effective in another model with low HER-2/neu expression, namely the MDA435/LCC6 model. Although significant efficacy, as judged by delays in tumour growth or increases in the percentage of long term survivors, is the ultimate goal to consider when developing new therapeutics, it is imperative to gain an understanding of the pharmacokinetic and pharmacodistribution behaviour of the agent being tested. With this in mind, the MDA435/LCC6 model was used to assess the pharmacodynamic behaviour of the two 132 antisense oligonucleotide sequences designed to inhibit expression of V E G F and HER-2/neu. ODNs are designed to be far more specific than traditional chemotherapeutics, yet they can be designed to have chemical and pharmaceutical attributes that make applications of this form gene therapy similar to conventional therapeutic agents. ODNs have been extensively researched as pharmaceutical agents, however even with the chemical modification of the phosphate backbone to a more stable phosphorothioate bond, they are rapidly cleared from the circulation, primarily by the kidneys (Levin, 1999). Although ODNs are efficacious when given in vivo, activity is often dependent on use of aggressive dose scheduling, involving daily injections for time periods in excess of two weeks, or use of micro-infusion pumps (Akhtar and Agrawal, 1997). There is therefore, a push to develop formulations which demonstrate enhanced circulation longevity, as well as increased delivery of the ODN to sites where the target cells are localised. A well established method to achieve this involves the use of lipid based carriers (see section 1.4.1), leading to the experiments detailed in this chapter, wherein lipid carriers were utilised for delivery of antisense oligonucleotides to tumour cells in vivo. The two antisense oligonucleotides utilised in these studies were a 21-mer with a phosphorothioate backbone, targeted against human VEGF with demonstrated activity against production of this protein in an in vitro model (Masood et al., 1997), and a 15-mer phosphorothioate ODN specific for the area of HER-2/neu mRNA immediately upstream of the initiation codon (Vaughn et al., 1996; Saxon et al., 2000). These antisense sequences were encapsulated in a novel liposomal formulation, recently described by Semple et al. (2001) (Section 2.8), and their pharmacodynamic behaviour, delivered in free or liposomal form was determined using the MDA435/LCC6 ascites model. 133 5.2 Results 5.2.1 Plasma Elimination and Tissue Distribution of i.v. or i.p. Injected Free or Liposomal Antisense The plasma elimination of antisense, free or liposomal, was determined after i.v. and i.p. administration. These data have been summarised in Figure 5.1, which incorporates plasma antisense levels as measured using tracer levels of [ 3H]-H2A-ODN and plasma liposomal lipid levels as determined by a [ 1 4C] labelled CHE. It is important to note that the antisense levels measured may not represent the intact sequence and may be due to a radiolabeled metabolite. The plasma antisense levels obtained following i.v. (Fig. 5.1 A) and i.p. (Fig. 5.IB) injection of free (open circles) or liposome encapsulated (open triangles) H2A-ODN at an antisense dose of 10 mg/kg, clearly illustrate the primary benefit, enhanced plasma A U C attributable to use of liposomal carrier systems. Free antisense oligonucleotides were eliminated rapidly following i.v. injection, with greater than 95% of the injected dose eliminated within the first time point (1 hr). When free ODN was given by i.p. injection, the blood levels were slightly higher at the first time point and this is reflected by a slight increase in mean AUCo-24h obtained when the drug was given i.p. (see Table 5.1). In contrast, when the liposomal formulation was given i.v., there were significant increases in circulating blood levels of ODNs at every time point evaluated and this was reflected by the substantial increase in plasma AUCo-24h when comparing the liposomal formulation to the free ODN (Table 5.1). Similar increases in plasma mean 134 Fig. 5.1. Plasma antisense, lipid and antisense to lipid ratios following i.v. or i.p. injection of free or liposomal H2 A - O D N at an antisense dose of 10 mg/kg into tumour-free female SCID/Rag2m mice. (A & B) free and liposomal antisense plasma levels following i.v. or i.p. injection, respectively. (C & D) plasma lipid levels following i.v. or i.p. injections, respectively. (E & F) plasma antisense to lipid ratios following i.v. or i.p. injections, respectively, of liposomal antisense. (V) liposomal ODN; (O) free ODN; (•) liposomal lipid. Results are the mean of 4 ± standard error of the mean. 135 AUCo-24h were observed when the liposomal formulation was given by i.p. injection (Fig. 5.IB). As expected however, the levels in the plasma at the lh time point following i.p injection were significantly (p < 0.005) lower than those obtained following i.v. administration. At 4h, the level of ODN in the plasma was equivalent for both the i.p. and i.v. route of administration and this is consistent with the observations that liposomes given i.p. rapidly enter the blood via the lymphatics (M. Bally, unpublished observation). Table 5.1 Mean AUC from 0 - 2 4 hours for free and liposomal O D N given both intraperitoneally and intravascularly. Values are derived by the standard trapezoidal formula and are from the mean of a minimum of 4 animals. Free O D N i.v. (Ugh/ml.) Free O D N i.p. (Ugh/ml.) Liposomal O D N i.v. (Ugh/ml.) Liposomal O D N i.p. (Ugh/ml.) 148.9 153.8 1565.7 1219.7 The plasma liposomal lipid levels (Figure 5.1 C and D) indicate that this liposomal formulation, when given i.v., exhibits long circulation lifetimes. As much as 20% of the injected lipid dose was still present in the plasma compartment 24h after injection. The plasma levels of lipid obtained after i.p. injection were lower and less than 10% of the injected lipid dose could be found in the plasma compartment at 24h. Figure 5.1 E and F includes calculated data based on the ratio of measured plasma O D N concentration and measured plasma liposomal lipid concentration. This ratio is used as an indicator of the stability of the liposomal formulation in the plasma compartment. A decrease in the ODN to lipid ratio is indicative of antisense release (dissociation) from the liposomes. The results suggest that over the 24h time course there was a steady decrease in this ratio indicating that the antisense sequence can be released from the liposomes within the plasma 136 compartment. It is assumed that the majority of ODN within the plasma compartment is associated with the liposomal carrier. Similar plasma elimination studies were conducted in MDA435/LCC6 tumour bearing SCID mice, where mice were injected i.p. with 2 X 106 MDA435/LCC6 cells and 17 days later they were injected i.v. with either free or liposomal antisense at an antisense dose of 10 mg/kg. The plasma ODN levels have been summarised in Figure 5.2 for both the H2A-ODN (Fig. 5.2A) and the VEGF-ODN (Fig. 5.2B). The data in Figure 5.2A is very comparable to the plasma elimination data obtained in tumour free mice, suggesting that the presence of an established tumour has little impact on the elimination of either the free or liposomal ODN. In addition, the data in Figure 5.2 suggests that the plasma elimination behaviour of both ODN sequences are also comparable, even though the H2A-ODN is a 15 nucleotide sequence and the VEGF-ODN is a 21 nucleotide sequence. As was the case for non-tumour-bearing mice, the encapsulated antisense exhibited a much greater plasma residence time for each of these oligonucleotides, with 4 hour levels being approximately 50-fold higher for liposomal formulation as compared to free. Mean AUCo-24 analysis of these data demonstrated that the liposomal formulation provides for a 10-fold increase in mean plasma AUCo-24 when compared to the free ODN (see Table 5.1). The tissue distribution of ODN and liposomal lipid following i.v. administration of the VEGF-ODN was also determined and the results have been summarised in Table 5.2 for both free and liposomal ODN. There are two important points that can be made on the basis of the data in Table 5.2. First, the highest tissue levels of O D N obtained after 137 Fig. 5.2 Serum antisense levels in MDA435/LCC6 ascites tumour-bearing female SCID/Rag2m mice following injection of free or liposomal (A) H2A-ODN or (B) V E G F -ODN. Antisense was administered on day 17 post tumour cell inoculation. (V) liposomal ODN; (O) free ODN. Results are the mean of 4 ± standard error of the mean. 138 administration of the unencapsulated form was in the kidney. This is consistent with previous studies (Levin, 1999) and is indicative of rapid renal clearance of this agent. Second, the levels of ODN in the spleen and liver were equivalent on a weight basis and the levels within these tissues were reasonably constant over the 24h time course. Liver accumulation of ODN accounted for less than 10% of the injected O D N dose, while the spleen accumulation accounted for less than 1% of the injected dose. When comparing these data to the liposomal formulation (Table 5.2) there are several notable differences. First, liver and spleen ODN levels were substantially higher when the ODN was given in liposomal form. Liver levels were almost 3-fold greater in animals given liposomal ODN and spleen levels were 7 to 10-fold greater, depending on the time point. This increase in spleen and liver delivery is a consequence of the natural tendency of liposomes to accumulate in these organs. Second, the level of ODN in the kidney was at least 5-fold lower when the ODN was given in the liposomal form, an observation that is consistent with the fact that liposomes do not readily accumulate in organs such as the kidney (liposomal lipid levels were below detectable limits at the 1 and 4 hour time points). Similarly, the muscle ODN and lipid levels were below detectable limits, a consequence of little or no liposome accumulation in muscle tissue. It can be concluded from the data in Table 5.2 that the distribution of ODN, when given in the liposomal formulation is dictated by the biodistribution behaviour of the liposomes. It is worth noting that ODN to lipid ratios, estimated on the basis of the biodistribution data in Table 5.2, were lower than those measured in the plasma compartment (see Figure 5.IE) and suggest that the ODN was lost from the liposomes both during and after tissue localisation. Interestingly, the ODN to 139 00 l-l 3 O X T f CN 3 O 4 3 T f 3 O -S3 T f CN 00 l-l 3 O 4 3 00 S C/5 t-i 3 O 4 3 3 oo 00 ro OS © -H oo ro O S OS r--o -H CN r-00 CN 41 Os CN Tf' t--O o -H SO CO CN T f wo O -H oo CN 4, r--o CN wo co I 4? ro I T f os 4H r-ro ro CJ > 3 3 oo O N wo 41 wo O T t SO T f sq I 4? oo WO SO r--© ' 41 T f 0 0 o 41 T f os 3 CJ _CJ 00 V wo ro © wo CJ 3 •3 ro ro O 4) CN O N C I 0 CN © T f CN 41 wo o 00 wo 41 00 © so ro -H Os wo wo -H ro T f wo T f 41 oo 00 T f -H T f CJ > T f CN "H , O^ T f +1 ro © ro Q so © -H ro I CN +1 , T f CN Q 00 3 3 i—) Os wo CN CN -H T f W0 T f oo I 41 Os wo W0 Os CN 00 41 T f oo 00 CN OS CN +1 r--00 ro I 4? wo 00 T f CN CN Os 3 CJ "a 00 O wo -H T f CN SO ro Q Q os 3 T f S O © -H ro Os CJ 3 -3 a Q Q Q Q Q Q o 00 3 N Q 0 F u i o s o d i q •3 § O 4 3 4 3 CJ -a CJ & 00 CJ CN CJ 3 15 > B cd » cd cj 00 CJ >-CJ 4 3 o 4 3 oo CJ I O CJ -4—» CJ -3 "o3 CQ 140 lipid ratio in the kidney, (estimated at the 24h time point to be 0.4), was almost 3-fold higher than that measured in the plasma compartment. This is consistent with the fact that the kidney is the primary organ for ODN removal, once the ODN has been released from the liposomes. One of the benefits of utilising a tumour model where the cells grow as an ascitic tumour is that it provides easy access to the cell populations residing within the peritoneal cavity and this, in turn, means that drug delivery and associated cellular effects can be readily measured. To assess antisense delivery to the peritoneal cavity, MDA435/LCC6 tumour bearing animals (prepared as described for the studies shown in Figure 5.2) were given an i.v. injection of free or liposomal VEGF-ODN on day 17 post-tumour cell inoculation. At 1, 4 and 24 hours post antisense injection, peritoneal fluid and associated cells were recovered by peritoneal lavage, as described in sections 2.14.4 and 2.14.5. The results from this study are summarised in Fig. 5.3. The levels of ODN in the cell-free peritoneal fluid were comparable when the ODN was given as a free or liposomal formulation. This result was surprising considering that there was a substantial increase in the blood levels and circulation lifetime for the liposomal formulation. For conventional small molecule anticancer drugs, delivery in a liposomal carrier typically results in substantial increases in peritoneal cavity delivery, both in ascites tumour bearing and tumour free animals (Bally et al, 1994). It should be noted that the level of delivery is quite low, with as little as 1% of the injected dose accessing the peritoneal cavity at 24h. Another important conclusion made regarding the data in Figure 5.3 concerns the increase in peritoneal cell delivery achieved when using the liposomal formulation. Although the error in this analysis is large, 141 0 5 10 15 20 25 Time Post-Injection (hours) Fig. 5.3 Peritoneal fluid (A) and peritoneal cell-associated (B) antisense levels following administration of free or liposomal V E G F - O D N at an antisense dose of 10 mg/kg into female SCID/Rag2m MDA435/LCC6 ascites tumour-bearing mice. Antisense was administered day 17 post tumour cell inoculation. (V) liposomal ODN; (O) free ODN. Results are the mean of 4 ± standard error of the mean. 142 and significant differences were only notable at the 4h time point (p<0.05), it is evident that there was a substantial increase in cell delivery. As much as 3 to 5 times greater peritoneal cell associated ODN was found following liposomal ODN administration than that achieved when the free ODN was given. 5.3 Discussion Liposomes are advantageous for the delivery of many therapeutics as they are able to extend the circulation lifetime as well as engendering significant increases in the plasma concentration of the associated agent (see section 1.4.1). The increased circulation lifetime and plasma concentration have been shown to effect increases in drug accumulation at sites of disease, particularly those diseases which cause damage to the surrounding blood vessels. This is particularly true for sites of tumour growth where there is induction of neovasculature and the release of factors, such as VEGF, which cause increases in blood vessel permeability to circulating macromolecules (Nagy et al, 1989; Bally et al., 1994). Studies demonstrating the therapeutic activity of ODNs in vivo have had to rely on prolonged administration of the drug in free form, therefore it was anticipated that liposomal formulations of ODN should increase delivery of the intact sequence to sites of disease and reduce the frequency of drug administration required with the free form. The pharmacokinetic benefits of administering ODNs in a liposomal formulation have been illustrated well by the data presented in this chapter, where the mean plasma AUCo-24h was more than 10-fold greater than that achieved following i.v. administration of free O D N (see 143 Table 5.1) and the biodistribution of the ODN was dictated by the biodistribution of the liposomal carrier (see Table 5.2). Importantly, the work presented in this chapter demonstrated the feasibility of systemic ODN delivery to a tumour site using a liposomal carrier formulation, and moreover, that a greater amount of ODN was delivered to tumour cells when encapsulated within the liposomes. Also of note is that the size of the ODN molecule did not alter the pharmacodynamics of the formulation (Fig. 5.2), with the plasma elimination of both the 15 and the 21 base ODN being equivalent. Having established enhanced delivery of encapsulated ODN to tumour cells in an ascites model of human breast cancer expressing increased levels of vascular endothelial growth factor, the logical progression of experiments was to proceed with efficacy studies using the liposomal VEGF-ODN as well as the liposomal H2A-ODN in this model for the reasons outlined in section 5.1. 144 Chapter 6 E F F I C A C Y O F L I P O S O M A L A N T I S E N S E O L I G O N U C L E O T I D E S IN T H E M D A 4 3 5 / L C C 6 H U M A N B R E A S T C A N C E R X E N O G R A F T M O D E L 6.1 Introduction The humanised monoclonal anti-HER-2/neu antibody trastuzumab is available for treatment of those patients with a strongly positive (3+ category as defined by the F D A approved HercepTest) level of expression of HER-2/neu, which is also corroborated with HER-2/neu gene amplification. By the very nature of treatment acceptance criteria, there are many breast cancer patients who do have some degree of HER-2/neu overexpression, but who do not qualify for trastuzumab treatment. In addition, there are those patients whose tumours are not classified as being overexpressers of HER-2/neu, but who may benefit from HER-2/neu targeted therapy as evidenced by the data presented in chapter 3 of this thesis, in which it was demonstrated that antisense ODN directed against HER-2/neu mRNA was efficacious in the treatment of mice bearing human xenograft tumours with low expression of HER-2/neu. The use of antisense oligonucleotides in these settings is attractive due to some of the drawbacks of trastuzumab therapy. Some of these have been mentioned in section 1.3.3, however in addition, it is known that the external portion of the pl85 molecule can be shed into the circulation (Pupa et al., 1993). This shedding of the receptor has two main implications from a therapeutic perspective. Patients with high levels of circulating shed 145 HER-2/neu may be resistant to the antiproliferative effects of trastuzumab by virtue of antibody plasma sequestration by binding to the circulating shed receptor, thus limiting the number of receptors in the plasma membranes that are available to bind to the antibody. In addition, it may be possible that the truncated fragment of the receptor resulting from the cleavage process is more oncogenic than the full receptor itself. Each of these possibilities suggest that a treatment modality other than antibodies which bind to the extracellular portion of the pi85 molecule might be preferable. Antisense oligonucleotides fit this requirement, and may in addition be beneficial to a larger subset of patients than is trastuzumab by being beneficial to those patients with low pi85 expression. Experiments with liposomally encapsulated H2A-ODN in a model of low p i 85 expression have shown some degree of efficacy of this molecule (MDA-MB-435; chapter 3). As was discussed in chapter 4 however, the MDA-MB-435 model was inferior to MDA435/LCC6 in several regards, including time to experimental conclusion (see Fig. 4.1) and versatility of tumour growth (i.e., solid and ascites in the case of MDA435/LCC6). These features, and its low HER-2/neu expression combine to make the MDA435/LCC6 model an attractive and obvious choice for continuation of efficacy experiments with H2A-ODN. It was also important to verify the positive results obtained in the MDA-MB-435 model following treatment with liposomal H2A-ODN (Fig 3.12) by using another model with similar low levels of p i 85 expression. Her-2/neu is one oncogene implicated in aggressive and metastatic breast cancer. As discussed earlier, there are several proteins and/or pathways which may be involved in 146 tumorigenesis, tumour progression and metastatic spread. One of these, as detailed in section 4.1, is vascular endothelial growth factor. The MDA435/LCC6 model has a V E G F expression level approximately three times that of the parental MDA-MB-435, thus providing an ideal model for testing the efficacy of antisense targeted against VEGF. VEGF is also an ideal target as there have been several studies linking the expression and/or stimulation of HER-2/neu with that of VEGF (Petit et al, 1997; Xiong et al, 2001). Preliminary experiments with liposomally encapsulated antisense oligonucleotides (ODN) targeted toward the mRNA of the HER-2/neu proto-oncogene provided conflicting results, as detailed in chapter 3. It is probable that cancer cells do not rely upon a single protein or pathway in order to outgrow their surroundings and spread to other locations in the body. Rather, they are more likely to have in place redundant systems, such that if one pathway is shut down, they are still able to proliferate and spread by relying on another. This theory is supported in part by the well known fact that in many cases, a combination of drugs is more effective than a single chemotherapeutic agent in combating cancer. A mixture of different drugs will target different pathways within the cell, for example, mitotic spindle assembly or D N A replication, thus limiting the repertoire of the cancer cell, and improving the chances for total cancer cell kill. This reasoning led to the experiments described within this chapter, in which two different antisense oligonucleotides have been tested for efficacy in the MDA435/LCC6 model, namely H2A-ODN and VEGF-ODN. It is hoped that once all parameters of single ODN therapy have been established, it will be possible to combine two or more ODN in the 147 treatment of aggressive breast cancer. Prior to this happening however, a complete understanding of the effects and downstream events of therapy with a single O D N must be firmly established. 6.2 Results 6.2.1 Treatment of MDA43 5/LCC6 Ascites Tumour Bearing Mice with H2A-ODN Female mice were inoculated with MDA435/LCC6 cells as described in section 2.13.2 for formation of ascites tumours. On day 3 post inoculation, 5 times weekly treatment with either free or liposomal H2A-ODN was initiated. Following a total of 15 ODN injections at an antisense dose of 10 mg/kg, peritoneal fluid and cells were harvested by peritoneal lavage such that both tumour burden and pi85 expression could be evaluated. Flow cytometric analysis of HER-2/neu expression in peritoneal tumour cells (Fig. 6.1) indicated a decrease in expression of HER-2/neu such that tumour cells removed from free H2A-ODN-treated mice had a HER-2/neu expression equivalent to 60% that of control untreated mice. This benefit was further enhanced following treatment with liposomal H2A-ODN, where the expression level of HER-2/neu in the treated mice was equivalent to 21% that of HER-2/neu expression in tumour cells harvested from untreated mice. The total number of cells recovered from the peritoneal cavities of those mice treated with liposomal H2A-ODN was significantly less than the control, untreated mice (figure 6.2a). In addition, when cells were spun down onto slides and stained (modified Wright - Giemsa stain), those cells which were present in the peritoneal cavities of liposomal H2A-ODN 148 10 r Figure 6.1 HER-2/neu expression in peritoneal tumour cells taken from mice bearing MDA435/LCC6 ascites tumours following treatment with saline, free H2A-ODN or free SCR-H2A-ODN. Results represent the mean ± standard error of the mean of flow cytometric data derived from equivalent numbers of cells from a minimum of 4 female SCfD/Rag2m mice. 149 OJ CJ (o/0) X J I A B 3 [Bauojuad in s\pj oo o o + LU O O OO O O + UJ o m oo o o + LU O O 00 O O + UJ o m oo o o + LU O O o o + LU O O o o o + LU o o o o 6 £ =5 o T3 CJ 43 Q o 'cj' CN § CJ 3 CJ cj - c a * CJ •5* § ^ CJ w a o « . s 3 = l - l o V3 a 2 c3 T t -o £ 3 H CJ o s .s 2 CN CN 00 c3 0 PH 00 3 2 2 2 a o CJ 2 treated mice were predominantly mouse reactive cells, such as neutrophils, macrophages, and possibly natural killer cells (Figure 6.3), as determined by staining with the anti-CD90 antibody, Thy-1. In fact, when the total number of viable human cells from the peritoneal cavities of these mice was estimated (Figure 6.2b), those mice treated with liposomal H2A-ODN had only 5% the number of tumour cells as compared to untreated controls (p<0.05). This dramatic reduction in the number of peritoneal tumour cells was not observed in mice treated with free H2A-ODN, signifying the utility of the liposomal carrier employed in these studies. 6.2.2 Treatment of MDA435/LCC6 Ascites Tumour Bearing Mice with Trastuzumab It was not possible to correctly interpret the above results without the inclusion of an empty liposome control. As well, it was important to include a trastuzumab treatment group for a reference. Trastuzumab is known to be efficacious in the treatment of solid tumours with high levels of p i 85 expression (T. Denyssevych, personal communication), but it was not known how tumours with low expression of pi85 would respond. As such, it was important to be able to directly compare the effects of the two anti-HER-2/neu therapies in this new model of human breast cancer. In solid tumour models of high pi85 expression (MCF-7 H E R 2 ) , trastuzumab has been demonstrated to be efficacious at doses as low as 0.3 mg/kg on an every other day schedule for a total of 15 doses. This experiment used a dose of 10 mg/kg, a dose which, in the aforementioned model, completely halted tumour growth. The results from this experiment are shown in figure 6.4, in which it may be seen that trastuzumab was not effective in 151 Figure 6.3 Modified Wright-Giemsa stain of (A) MDA435/LCC6 tumour cells harvested from peritoneal cavity of an untreated mouse, and (B) cells harvested from peritoneal cavity of a mouse inoculated with MDA435/LCC6 cells and treated with liposomal H2A-ODN as outlined in the text. 152 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 4e+8 4e+8 3e+8 h J L v 8e+7 A 7e+7 6e+7 5e+7 r 4e+7 3e+7 2e+7 1e+7 0 Day 7 Day 14 Day 21 Figure 6.4 Total volume (A) and total number of peritoneal cells (B) harvested from mice inoculated with MDA435/LCC6 cells and treated with saline (black bars) or trastuzumab (clear bars). Results are the mean ± standard error of the mean of data from a minimum of four animals. 153 delaying tumour growth in these mice. While preliminary in vitro testing with VEGF-ODN did not indicate significant benefit to the use of the drug in terms of V E G F downregulation, the ODN was taken into in vivo experiments in a limited nature, simply due to the lack of complete understanding of the precise mechanism in which antisense oligonucleotides exert their actions, and the necessary cellular environments for optimisation of this action. 6.2.3 Treatment of MDA435/LCC6 Ascites Tumour Bearing Mice with V E G F - O D N Antisense efficacy experiments were initiated with VEGF-ODN, as the MDA435/LCC6 cells produced increased amounts of VEGF in comparison to the parental cell line. Ascites tumours were produced as before. Free VEGF-ODN was administered dissolved in sterile water at an antisense dose of 10 mg/kg three times weekly, for a total of 10 treatments, starting on day 3 post-tumour cell inoculation. The mice were terminated one day following the final antisense injection. When the total tumour burden in both volume and absolute peritoneal tumour cell numbers was compared between the free VEGF-ODN-treated, and untreated groups, there were no significant differences noted. This was also the case for V E G F protein levels measured in both the peritoneal fluid and in the plasma of each of these two groups. To assess the efficacy of a V E G F targeted antisense in the MDA435/LCC6 tumour model, groups of SCID/Rag2m mice were inoculated on day 0 with 2 X 106 MDA435/LCC6 cells, then beginning on day 3 post tumour cell inoculation, mice were injected daily for 10 days with liposomal VEGF-ODN at an antisense dose of 10 mg/kg. As noted in Table 6.1, approximately 2 X 10 cells were isolated from the 154 peritoneal cavity under the conditions employed in these studies and this represented a 20-fold increase in cellularity when compared to non-tumour bearing mice (the average peritoneal cell yield from 19 non tumour-bearing mice was 4.0 X 1 0 6 ± 2 . 8 X 105). Under conditions where the tumour bearing animals were treated with saline (control treatment groups), almost 40% of the cells isolated from the peritoneal cavity were murine (i.e. non-human) as judged by flow cytometric analysis. Importantly less than 10% of these murine cells could be identified as peritoneal macrophages. Also the influx of murine cells, primarily neutrophils, into the peritoneal cavity was probably a consequence of tumour growth and perhaps the associated production of VEGF (see Figure 4.3). Table 6.1 Tumour burden (total peritoneal fluid volume and tumour cell count), V E G F levels per million peritoneal tumour cells, and serum V E G F (pg/ml) in control, untreated mice, and mice treated with 10 — three times weekly 10 mg/kg V E G F - O D N injections. Tumours were initiated on day 0 with i.p. injections of 2 X 10r' MDA435/LCC6 cells. Treatment was started on day 3 post tumour cell inoculation. Numbers represent the mean of 8 ± standard error of the mean. Control V E G F - O D N Tumour Burden (cells) 1.66 ±.42 X 10s 2.02 ± .59 X l O 8 VEGF (pg/10b cells) 285.34 ±43.00 311.26 ±52.32 V E G F (pg/ml serum) 54.52 ± 18.46 65.89 ± 18.77 As may be noted from the data in Table 6.1, the treatment of LDA435/LCC6 ascites tumour-bearing mice with liposomal VEGF-ODN did not cause any reduction in tumour burden in these mice, nor did the levels of VEGF measured in the serum decrease. Due to the lack of activity seen with VEGF-ODN in free form in both in vitro and in vivo tests, I chose not to pursue the liposomal formulation or combinatorial studies with V E G F - O D N at that time. 155 6.3 Discussion Although the MDA435/LCC6 model does overproduce vascular endothelial growth factor, I was unable to achieve a reduction in tumour burden, or decrease in V E G F production with the chosen anti-VEGF antisense sequence. I feel it is important to note that there are many parameters in the use of therapeutic antisense oligonucleotides which have not been resolved satisfactorily. These include i) determination of the most efficacious ODN sequence, ii) determination of the most adequate delivery method and schedule, and iii) assessment of sequence independent non-specific host antisense reactions. Further work in these areas is required, with a focus on identification of whether V E G F is significant in tumour vascularisation, ascites growth and/or metastases of MDA435/LCC6 tumours. With these factors in mind, I believe this model will be beneficial for studies involving anti-angiogenic compounds, including those targeting VEGF. In the studies presented above, the efficacy of liposomally encapsulated anti-HER-2/neu ODN in the treatment of MDA435/LCC6 derived ascitic tumours was presented (Fig.6.2). This supported earlier observations that H2A-ODN is effective in delaying the growth of tumours with low expression of HER-2/neu protein (sections 3.2.7 and 3.2.10), (Saxon et al., 1999). This is an important observation given the number of breast cancer patients whose tumours express the HER-2/neu protein at levels lower than those currently set as the level for inclusion in a trastuzumab containing regimen. In addition, it was shown that when encapsulated in liposomes, the H2A-ODN was more effective in the downregulation of the HER-2/neu protein than when administered in free 156 form. This is not surprising considering the results shown in section 5.2 (Fig 5.3), where it was demonstrated that more antisense was delivered to the tumour cell population within the peritoneal cavity with the encapsulated form than with free. This work has set the groundwork for many future studies in the MDA435/LCC6 model, including further experiments targeting HER-2/neu, as well as those targeting other proteins associated with aggressive breast cancer in humans. 157 Chapter 7 CONCLUSIONS The diagnosis of breast cancer is one of the most dreaded events in the lives of most women. Nearly all North American women have known either a close friend or relative who have experienced this devastating disease, and have borne witness to both the emotional and physical challenges it brings. Of the approximately 11% of women who develop breast cancer, fully one quarter will die of their disease. It has long been known that some breast cancers are more aggressive than others. While certain gross and histopathological features such as tumour type (e.g., comedo, lobular, ductal) and mitotic index have been identifiable for some time to aid in prognosis, scientists are now turning with ever increasing frequency to the use of molecular markers. It has been found that increased (or in some cases, decreased) expression of certain proteins, whether singly or in combination with other molecular factors, can be indicative of the nature and course of a specific breast cancer, and can therefore aid in selection of treatment options on an individual patient basis. A marker which falls into the category of one which has prognostic significance when combined with expression of other markers, is the protease Cathepsin D (section 1.2.2.1). While overexpression of this protein alone is not indicative of more aggressive disease, when it is co-expressed with other proteins such as p53, the significance of its 158 overexpression is clearly linked to adverse disease outcome (Han et al, 1997; Solomayer et al, 1998). Vascular endothelial growth factor (section 1.2.2.2) is another protein identified in many types of cancer, including breast cancer (Gasparini et al., 1997; Anan et al, 1998). When overexpressed by tumour cells, the tumour is able to induce the formation of a rich blood supply, thereby ensuring adequate nourishment for further growth. V E G F has also been implicated in the process of lymphangiogenesis in tumour growth. This protein is an example of one in which overexpression of a single protein may be linked to more aggressive disease and increased metastatic events. Downregulation of V E G F has been shown to halt the progression of tumour cell growth both in vitro and in vivo (Masood et al, 2001). This type of therapy would be ideal in combination with a cytotoxic therapeutic, as a two-pronged approach to cancer care, in which further growth of the tumour is halted by the anti-VEGF agent, while cell kill of existing tumour cells is accomplished by the conventional drug. The protein with which the bulk of this thesis has been focused upon is the protein product of the HER-2/neu oncogene, namely p i 85 (section 1.2.2.3). This member of the epidermal growth factor receptor family has been linked to more aggressive cancer, increased metastatic potential (Allred et al, 1992; Press et al, 1997; Pich et al, 2000), short time to relapse (Slamon et al, 1987) and resistance to certain types of therapy (Piccart et al, 1988; Mourali et al, 1993). The overexpression of p i 85 has been noted in many types of cancer, however this discussion will centre around breast cancer both due to the content of this 159 thesis, and because the bulk of research directed against HER-2/neu has been conducted in either breast cancer models or breast cancer patients. In normal adult breast tissue, typical, non-malignant epithelial cells will have relatively few (less than 100) pi85 receptors on their surface. In some breast cancer cells this number can be increased to 2000 or more receptors per cell (V. Dragowska, personal communication). This huge increase in receptor number leads to increased interaction of the HER-2/neu receptor with other members of class I RTKs, as well as homodimerisation, and subsequently increased signalling events and increased transcription of nuclear factors, culminating in increased tumour cell growth and spread. With the greater understanding of tumour biology that is now being realised, including the molecular features linked to rapid growth, invasiveness, metastasis, chemo- and radiation-resistance, there will likely be improvements in treatment options. These new therapies will link gene discovery, proteonomics and drug discovery to define exquisitely specific drugs that target molecular defects inside diseased cells. It is critical that the drugs emerging from these efforts are used in a manner consistent with our enhanced understanding of the disease. Cancer may be considered as a process of dynamic derangement of both growth and death pathways, with malignancy arising from a progressive accumulation of genetic defects that allow for the cancer cell to escape normal regulatory controls and develop malignant potential. The implication of the above is that we can attempt to re-impose biologic control 160 over the malignant cells and restore a normal growth pattern, however treatment will likely involve more than one molecular target or pathway in order to ensure eradication of a malignant clone. Due to the complexity and heterogeneity of cancer, it is likely that many of the new molecularly targeted therapeutic agents may not have sufficient activity as a single agent, but rather that their value will be realised through identification, in pre-clinical animal models, of effective combination regimens. There are many currently available therapeutic options in the treatment of metastatic breast cancer. One of the foremost drugs in use is doxorubicin, typically in combination with one or more other cytotoxic drugs, such as cyclophosphamide and 5-fluorouracil. In addition to the conventional drugs, the arsenal of weaponry in the fight against this disease also contains agents such as anti-oestrogens, aromatase inhibitors and anti-progestins, as well as the newer strategies of monoclonal antibodies. The first approved therapy targeting pi85 is trastuzumab, a humanised monoclonal antibody targeting the extracellular portion of the p i 85 molecule. While this therapy has shown impressive results in a small number of breast cancer patients with overexpression of p i 85, it has been ineffective in many more, with an overall response rate of only 21% following use as a single agent (Cobleigh et al, 1998). While it is true that the location of pi85 on the cell surface makes it an ideal target for an antibody based therapy, what is not clear, is why this therapy appears to be selectively effective in breast cancer patients. It may be possible that in those patients where trastuzumab is non-effective, there remains another, as of yet unidentified protein which is able to function in the absence of HER-161 2/neu overexpression to maintain the aggressive features of the tumour. It is also possible that tumours from different individuals react in differing ways to the anti-pl85 antibody. It has been demonstrated in nude mice that treatment with some pi85 targeted antibodies actually caused an increased rate of tumour growth, perhaps by stimulating increased phosphorylation of the HER-2/neu receptor (Stancovski et al., 1991). These same researchers demonstrated that all antibodies tested, accelerated, albeit to differing degrees, the rate of receptor turnover (endocytosis upon antibody binding). They surmised that the different end results of antibody binding (either reduced, increased or unchanged tumour growth rate) was due to the different extracellular epitopes of the pi 85 molecule to which these antibodies bound. They concluded that the positive tumour growth resulted from stimulation of the tyrosine kinase function of pl85. The newcomer to the battle against cancer is the strategy of gene therapy (discussed in section 1.3.4), with its several variations, including that of antisense oligonucleotides. This modality brings to the field an extreme specificity for its target, rivalled only by the use of antibody therapy. These drugs, as discussed, may be targeted against a specific protein product of a gene, or at a point downstream in the signalling pathway. For the specific example of pi85 directed ODN, it is hoped that this will render the tumour cells less aggressive and more amenable to eradication by either the pi85 targeting therapy itself, or a conventional cytotoxic agent. Chapter three of this thesis addressed the treatment of MDA-MB-435 and HER-2/neu transfected MDA-B-435 human xenograft tumours grown in female SCfD/Rag2m mice, 162 with antisense oligonucleotides targeted against the mRNA for the HER-2/neu protein. It was thought to be important at this stage to use cell lines with differing HER-2/neu expression levels (low and high) that had a minimum of additional differing factors, which may have confounded any results obtained. By minimising differences other than HER-2/neu expression, the results would be rendered less ambiguous, and more easily linked to direct outcome of expression. As was demonstrated in section 3.2.1, the HER-2/neu transfected cells grew more rapidly in vitro than the parental cell line. This was consistent with other observations that overexpression of HER-2/neu conferred a growth advantage to cells (Liu et al., 1995). However, increased rate of tumour growth was not observed in vivo. It is possible that the similarity in tumour size may not be reflective of the content of live cells within the tumour mass. Rather, it is possible that there is a greater percentage of viable cells within the M D A - M B - 4 3 5 H E R 2 cell-derived tumours as compared to the parental cell-derived tumours. This has indeed been found to be true for other cell lines with direct comparisons between parental and HER-2/neu transfected derivatives (V. Dragowska, personal communication). It is also possible that a greater degree of disparity in the growth rates of the two tumour types may have been noted had the tumours been permitted to grow further, as while not significant, there is a slight separation between the growth curves of the two tumour types, with M D A-MB-43 5 H E R 2 being higher, by day 104 post tumour cell inoculation (Fig. 3.1). It is important to note that the commonly used cytotoxic agent, doxorubicin, was efficacious in both the MDA-MB-435 and M D A - M B - 4 3 5 H E R 2 tumour models, thereby demonstrating that these cells respond in an expected fashion to drugs used in the treatment 163 of aggressive breast cancer. Also importantly, neither of the two tumour types responded in any fashion to the scrambled control sequence of anti-HER-2/neu antisense, whether free (Fig. 3.7) or liposomal (Figs. 3.11 & 3.12). Of note, the free, active antisense, H2A-ODN, was efficacious in the established MDA-MB-435 cell-derived tumours, but not in the HER-2/neu transfected counterpart (Fig 3.8). It was anticipated that, given the extended circulation longevity typical of liposomal drug preparations (Boman et al., 1994; Webb et al., 1998c), there would be more potential for response in non-established tumours if the ODN were administered in liposomal form, rather than in free form. This was indeed the case, as was demonstrated in Figure 3.11, clearly defining the benefit of having an encapsulated form of anti-HER-2/neu antisense. While the free H2A-ODN had no impact on tumour growth in either established or non-established M D A - M B - 4 3 5 H E R 2 tumours, it was a surprising result to see an increased rate of tumour growth in the non-established M D A-MB-43 5 H E R 2 cell derived tumours (Fig. 3.12) following liposomal H2A-ODN treatment. Clearly, when one considers the difference in the free versus liposomal antisense in the tumour growth rate of the parental line, one can understand that it is probable that an increased amount of antisense has been delivered to the cells of the tumour mass. This was substantiated by immunohistochemical studies wherein visible expression of p i 85 was reduced to levels of the parental cells (Fig. 3.13) . This presented a challenging set of results for interpretation. While it was apparent that the H2A-ODN had resulted in the desired effect of reduced p i 85 protein expression, it was unclear as to why this would consummate in an increased rate of tumour growth. To expand on the earlier theory that perhaps the M D A-MB-43 5 H E R 2 cell-derived tumours have 164 a higher percentage of viable cells, there exists the possibility that by bringing the pi85 levels down in these cells, they are now able to enter into an alternate pathway for enhanced growth, and as a result of high viability, the growth rate is able to increase substantially. An intriguing set of results has been reported by Koscielny et al. (1998, 1999), which have been independently confirmed by several researchers (Dittadi et al., 1997; Pawlaowski et al., 1999). These results demonstrated that tumour HER-2/neu levels lower than those found in normal breast tissue, resulted in a negative prognostic outcome, as compared to tumours with normal HER-2/neu expression. They suggest that this observation may rely on the interaction between several oncogenes, for which the lack of expression of one may be counterbalanced by a concomitant overexpression of another, e.g., EGFR. They further suggest that perhaps it is not the overexpression or low HER-2/neu values that should be focused upon, but rather the deviation from normality in expression, in assessing prognosis in breast cancer patients. It would be of great interest in any future experiments, to include measurements of all the HER family members when downregulating HER-2/neu, as it may be possible that by having decreased HER-2/neu protein levels to lower than normal levels, the cell counterbalances by increasing expression of other proteins, as discussed by these researchers. Finally, it is possible that results presented here using HER-2/neu transfected M D A - M B -435 cells may differ from cells which have endogenous expression of the gene. This is because the HER-2/neu proximal promoter may not contain all the regulatory sequences for 165 the gene and other important control elements may be found both upstream and downstream of the sequence inserted into the transfection vector. For this reason, it is important to include tumours with endogenous overexpression of p i 85 in the experiments. Historically, these tumours, for example, SK-BR-3 and BT-474, have proven to be very difficult to establish in mice, and therefore were not included in these studies. There are however, new cell lines being established on an ongoing basis, and it is hoped that such a line will be available soon, as research into HER-2/neu is dependent upon it. A possible contender in this category is the newly described cell line, KPL-4, a human breast carcinoma line with endogenous overexpression of HER-2/neu (kurebayashi et al., 1999). In the experiments presented above, I believe that the most plausible explanation for enhanced tumour growth in the face of reduced HER-2/neu protein expression, follows from an enhanced number of viable cells in those tumours with overexpression of HER-2/neu, as compared to tumours with baseline HER-2/neu expression. Subsequent to the downregulation of pi85 in these tumour cells, I believe that an alternate pathway of enhanced cellular growth was entered into, plausibly involving HER-1, VEGF, a combination of these, or some other growth signalling pathway. To expand upon the requirements in breast cancer research, new, more relevant models, are required. Since the beginnings of what we consider as biomedical research, millions of laboratory animals have been experimented on and sacrificed for the purposes of research and education; their role purportedly the betterment of human health and welfare. There is however, a strong movement in most industrialised countries toward finding alternatives to 166 these animal experiments or towards the reduction in the numbers of animals required to achieve a particular aim. The rationale for this effort is based on scientific, economic and ethical considerations. Alternatives are any procedures which can completely replace the need for animal experiments (replacement), reduce the number of animals required (reduction), or diminish the amount of pain or distress suffered by the animals in meeting the essential needs of man and other animals (refinement). In the replacement of animal procedures, researchers have turned to methods such as cell culture, mathematical modelling, computer simulations and other such techniques. If experiments performed in cell lines derived from non-human sources cannot be assumed to model human disease accurately, neither can human-derived cells, growing in non-human hosts. There are a host of interactions between tumour cells and the environment in which they are growing, many of which may be governed by species specific proteins and pathways. Hence, since we are by necessity, limited to using non-human hosts in the modelling of human cancer, any results garnered in these models must be interpreted with extreme caution before moving on to designing clinical trials. This makes it imperative that any models we do use, are exquisitely carefully selected on the basis of their relevance to the specific disease, and that responses to established chemotherapeutic agents in these models mimic those responses seen in the clinic. Only in this way will the models chosen serve to reduce the overall burden of animal suffering. 167 Important aspects in the development of murine models of human cancer, specifically those models surrounding the development of therapeutics targeting HER-2/neu, include a more complete understanding of the complexity of signal transduction which commences following activation of the HER-2/neu receptor. We do not yet have a full understanding of the different environments in which the different heterodimers of the HER-2 family are formed, i.e., why in certain physiological conditions, certain dimers may be preferred over others. As well, further research needs to be conducted into the multiplicity of different signalling molecules which occur downstream in the signal transduction pathways, and how these molecules interact with each other, and why some pathways may take preference over others, again, triggered by specific physiological conditions. Until these parameters are met, we cannot hope to completely understand the interaction of any novel pharmaceutical agent, whether it be a small molecule, peptide, antibody or D N A based therapeutic, with the varying components of the HER-2/neu signal transduction pathway. Although some of these relationships can be studied in vitro, it is impossible to develop this understanding without considering in vivo parameters. The work presented in this thesis outlines quite clearly, the requirement for this understanding, as it has been demonstrated that antisense oligonucleotides targeted against the HER-2/neu mRNA, are in fact, quite capable of downregulating the absolute amount of pi85 expression on the cell surface, but at the same time, this reduction in expression may lead to increased tumour growth rates. It is not clear whether the apparent anomaly is due to the interaction of human tumour cells with the murine environment, or if this response is 168 intrinsic to the therapeutic used. It will be impossible to test this in the current state of human tumour modelling. With the above in mind, chapter four of this thesis was concerned with the development and characterisation of a new model for human breast cancer, the MDA435/LCC6 xenograft model. As the main thrust of this research was the development of therapeutics for breast cancer with overexpression of certain molecular markers which rendered that cancer more aggressive and more likely to metastasise, it was these targets which were initially characterised, along with such parameters as tumour growth rate and responsiveness to chemotherapeutic drugs. As was shown in section 4.2.2, the MDA435/LCC6 cell line was responsive to the drugs taxol, vincristine and doxorubicin in degrees comparable to the parental cell line, M D A -MB-435, and as would be anticipated for aggressive human breast disease (Fig. 4.2). This line was shown to be higher in VEGF expression than the parental line (Figs. 4.3 and 4.4), while retaining similar levels of HER-2/neu, bcl-2, and Cathepsin D expression (Figs. 4.4, 4.6 and 4.7, respectively), features which made it a good model for V E G F targeting therapeutics. Although considered a low expressor of HER-2/neu, this cell line also provided an opportunity to confirm results obtained when treating the parental M D A - M B -435 cell line with the liposome formulation of anti-HER-2/neu antisense. The MDA435/LCC6 tumour model was found to grow reproducibly as an ascites model (Fig. 4.1), making it amenable for rapid screening of anti-cancer drugs such as antisense 169 oligonucleotides, either singly, or in combination with other anti-cancer agents. In summary, having characterised this model as discussed, it was found to be a rapidly growing, reproducible and relevant model, suitable for further testing of the agents with which this thesis is concerned. Having the MDA435/LCC6 model in place, and having characterised the liposomal formulations of ODN, I was then able to turn to a more thorough examination of the pharmacodynamic properties of the antisense oligonucleotides, both free and liposomal. It was anticipated that by virtue of encapsulating antisense oligonucleotides into liposomal carriers, that greater circulation longevity would be achieved, and hence enhanced tumour cell delivery. This was indeed found to be the case, as was seen in Figures 5.1 and 5.2 and Table 5.1. Importantly, the enhanced longevity was similar for either the 15 or the 21-mer ODNs, indicating that the length of the molecule did not affect the circulation time, but rather, that it was the nature of the liposomal formulation which engendered this enhancement. Plasma elimination studies of free O D N administered either intraperitoneally or intravenously, showed essentially no difference in the plasma AUCs over 24 hours (Fig. 5.1), therefore allowing alternate routes of administration of the free ODN. This was an important consideration given the extended nature of the treatment schedule with either free or liposomal ODN, with typically over ten injections being administered over a short course of time. It was essential to show at this point that the increased circulation longevity afforded the ODN by encapsulation was translated to enhanced tumour cell delivery. This was 170 demonstrated in the MDA435/LCC6 ascites tumour model by measuring both the peritoneal fluid ODN levels as well as the peritoneal tumour-cell associated ODN. As was anticipated, those mice which received liposomal ODN rather than free, had a greater amount of tumour cell associated ODN (Fig. 5.3), although the total level of ODN delivery to the peritoneal cavity was comparable in animals given free and liposomal ODN. This was unexpected given previous studies demonstrating enhanced peritoneal delivery of liposome associated drugs, such as doxorubicin or vincristine. It is well known that liposomal delivery of drugs often results in increased uptake of both lipid and encapsulated drug in the organs of the MPS. With this in mind, increased levels of ODN were expected to be seen in the spleens and livers of those mice receiving the liposomal drug. This was indeed the case, as was shown in Tables 5.2 and 5.3 for MDA435/LCC6 ascites tumour bearing mice. It is important to note at this point that there was no toxicity associated with these increased uptake levels, as would be evidenced by general appearance or behaviour of the animals, or by gross pathological observations. This was not expected, given the relatively low toxicity of phosphorothioate O D N and the level of dosing used in these studies (Levin, 1999). Enhanced levels of O D N in the kidneys was likely a consequence of enhanced levels of ODN in the animals given the liposomal formulations and the slow dissociation of the O D N from lipid carriers in the plasma compartment or tissues. Renal function was not assessed in these studies. With the MDA435/LCC6 model in place and characterised with respect to tumour growth parameters, protein expression, and pharmacodynamic behaviour of the injected drugs, it 171 was possible to re-address the question of efficacy of the H2A-ODN in a model with low expression of the HER-2/neu protein. The initial observation of efficacy in the M D A - M B -435 model was entirely unexpected, and yet was thought provoking and exciting. There are many breast cancer patients whose HER-2/neu expression levels are too low to allow them to qualify for treatment with trastuzumab, the monoclonal anti-HER-2/neu antibody, and yet who may indeed benefit from a therapy targeting this protein. The lack of reduction in tumour burden associated with treatment of MDA435/LCC6 ascites tumour bearing mice with free VEGF-ODN was perhaps not an entirely unanticipated result, as VEGF levels in both the ascites fluid and plasma also were not impacted. It may of worth to remember as well, that ascites tumours are by their nature, bathed in fluid in the peritoneal cavity, and those clumps which are present, are small and attached at points along the peritoneal wall. Due to the dispersed nature of these cells, they would not be expected to rely heavily upon angiogenesis for supply of oxygen and nutrients. This is work which would be important for future studies of V E G F targeted therapy. In all, the experiments presented in chapter 6 have laid groundwork for future studies in which the combination of two or more strategies could be assessed. I believe that the combination of strategies targeting HER-2/neu and V E G F could have major impact on the growth inhibition of tumours, but that before this can be accomplished, there needs to be a careful study into the optimal method for selection of antisense sequences such that maximal protein downregulation is achieved. The field of antisense oligonucleotide therapy is relatively young, and advances in this area are being made on a 172 constant basis. One has also to be aware of some sequence independent toxicity properties of antisense oligonucleotides that are related to the chemical and physical properties of the molecules. However, these toxicities, primarily transient activation of the complement cascade as well as an inhibition of the clotting cascades, are only observed when ODN are administered at very high levels (LD50 = 750 mg/kg in primates) (Levin, 1999). It is hoped that in the very near future there will be in place a strategy for ODN sequence selection that fulfils the above requirements. To return to the objectives of this thesis, I have successfully developed and characterised the MDA435/LCC6 xenograft model of aggressive human breast cancer, and used that model to assess the actions of an ODN sequence, targeting HER-2/neu, both in free and in liposomal form. In addition, the combination of antisense oligonucleotide drugs with conventional cytotoxic agents (doxorubicin) was explored. I believe these combinations will maximise the efficacy achievable in the treatment of aggressive disease, while minimising morbidity, and are the direction that future cancer therapy must take in order to make gains in both extension and quality of life. I have also shown that the encapsulation of ODN into a liposomal carrier is an effective means of extending the circulation lifetime of the drug, as well as maximising tumour cell exposure, and hence increasing efficacy. As a concluding comment, I believe it is important to remember that the evaluation of the primary target (e.g., p i 85) may be insufficient to account for the overall effect of molecular targeted therapies on the growth of a tumour. It becomes essential to examine other, downstream effects of these therapies, and to examine redundant pathways within the cells, 173 as many signalling pathways interact at various points and appear to be multifactorial (Nieves-Neira and Pommier, 1999). With the advent of more powerful techniques for assessing multiple points within a pathway, and indeed, multiple pathways, and as our understanding of the diversity and interactions of these pathways increases, it will be possible in the not too distant future, to screen individual patients and tailor therapy regimens to best suit their individual disease. This work may aid in the identification of more relevant targets, for example, perhaps the ideal strategy for targeting HER-2/neu overexpressing tumours is not to target the mRNA for the protein itself, but to target instead, the HER-2/neu promoter binding factor, an action which would serve to halt the transcription of the gene itself. This remains to be experimentally determined. I anticipate that these regimens will consist of both targeted agents, such as the HER-2/neu or V E G F ODN or agents such as tyrosine kinase or farnesyl transferase inhibitors in combination with cytotoxic agents with known effects. In this manner, we will be able to make further advances in the care and treatment of cancer patients, both those with aggressive breast cancer and those with other, equally devastating forms of this disease. 174 BIBLIOGRAPHY Allred DC, Clark G M , Tandon A K , Molina R, Tormey DC, Osborne CK, Gilchrist K A W , Mansour EG, Abeloff M , Eudey L, McGuire WL. HER2/neu in node-negative breast cancer: prognostic significance of overexpression influenced by the presence of in situ carcinoma. J Clin Oncol 10:599-605, 1992. Anan K, Morisaki T, Katano M , Ikubo A , Tsukahara Y , Kojima M , Uchiyama A, Kuroki S, Torisu M , Tanaka M . Assessment of c-erbB2 and vascular endothelial growth factor mRNA expression in fine-needle aspirates from early breast carcinomas: pre-operative determination of malignant potential. Eur J Surgical Oncol, 24, 28-33, 1998. Andrulis IL, Bull S B, Blackstein M E, Sutherland D, Mak C, Sidlofsky S, Pritzker K P, Hartwick RW, Hanna W, Lickley L, Wilkinson R, Qizilbash A , Ambus U , Lipa M , Weizel H , Katz A, Baida M , Mariz S, Stoik G, Dacamara P, Strongitharm, D, Geddie W, McCready D. neu/erbB-2 amplification identifies a poor-prognosis group of women with node-negative breast cancer. Toronto Breast Cancer Study Group. J Clin Oncol. 16: 1340-9, 1998. Araya M , Terashima M , Takagane A, Abe K , Nishizuka S, Yonezawa H , Irinoda T, Nakaya T, Saito K. Microvessel count predicts metastasis and prognosis in patients with gastric cancer. J Surg Oncol, 65, 232-6, 1997. Ben-Baruch N , Alroy I, Yarden Y . Developmental and physiologic roles of ErbB receptors and their ligands in mammals. In: Hormones and growth factors in development and neoplasia. Dickson RB and Salomon DS, Eds. Pp. 145- 168 Wiley-Liss, Inc., 1998. Bernstein JL, Lopez-Carrillo L, Wang L. The epidemiology of HER-2/neu and P53 in breast cancer, salud publica de mexico 41 :S114-S123, 1999. Bobik A , Agrotis A, Little PJ. Vascular-derived growth factors: potential role in the development of the tumour vasculature. In: Tumour Angiogenesis. Editors: Roy Bicknell, Claire E. Lewis and Napoleone Ferrara. Oxford University Press. Pages 169-184, 1997. Boman N L , Masin D, Mayer LD, Cullis PR, Bally M B . Liposomal vincristine which exhibits increased drug retention and increased circulation longevity cures mice bearing P388 tumors. Cancer Res 54:2830-2833, 1994. Brown DA, Kang S-H, Gryaznov SM, DeDionisio L, Heidenreich O, Sullivan S, Xu X , Nerenberg MI. Effect of phosphorothioate modification of oligodeoxynucleotides on specific protein binding. J Biol Chem 269:26801-26805, 1994. Brown LF, Yeo K-T, Berse B, Yeo T-K, Senger DR, Dvorak HF, Van De Water L. Expression of vascular permeability factor (vascular endothelial growth factor) by epidermal keratinocytes during wound healing. J Exper Med 176:1375-1379, 1992. Cailleau R, Young R, Olive M , Reeves WJ. Breast tumor cell lines from pleural effusion. J Natl Cancer Inst. 53:661-674, 1974. Cancer Care Handbook, 6th edition. Kingston Regional Cancer Centre pp. 50 - 51, 1999 175 Carter SK. Single and combination nonhormonal chemotherapy in breast cancer. Cancer 30:1543-1555, 1972. Cefai D, Morrison BW, Sckell A, Favre L, Balli M , Leunig M , Gimmi CD. Targeting HER-2/neu for active-specific immunotherapy in a mouse model of spontaneous breast cancer. Int J Cancer 83:393-400, 1999. Chi K N , Gleave M E , Klasa R, Murray N , Bryce C, Tolcher A . A phase I trial of an antisense oligonucleotide to BCL-2 (G3139, Genta) and mitoxantrone in patients with metastatic hormone refractory prostate cancer (HRPC). Proc Amer Soc Clin Oncol, Abstract #1299, 2000. Christoffersen RE, Marr J. Ribozymes as human therapeutic agents. J Medicinal Chem 38:2023-2037, 1995. Cobeigh M A , Vogel CL, Tripathy D, Robert NJ, Scholl S, Fehrenbacher L, Paton V, Shak S, Lieberman G, Slamon D. Efficacy and safety of Herceptin "(humanized anti-HER2 antibody) as a single agent in 222 women with HER2 overexpression who relapsed following chemotherapy for metastatic breast cancer. Proc Amer Soc Clin Oncol Abstract #376, 1998. Collins PD, Connolly DT, Williams TJ. Characterization of the increase in vascular permeability induced by vascular permeability factor in vivo. Br J Pharmacol, 109, 195-9, 1993. Colomer R, Jupu R, Bacus SS, Gelmann EP. ErbB-2 antisense oligonucleotides inhibit the proliferation of breast carcinoma cells with erbB-2 oncogene amplification. Br J Cancer 70:819-825, 1994. Cowens JW, Kanter P, Brenner DE Ginsberg R, Mayer L, Bally M , Cullis P, Pavelic Z, Douglass H, Petrelli N , Karakousis C, Creaven PJ. Phase I study of doxorubicin encapsulated in liposomes. Proc Am Soc Clin Oncol 1989; 8:69. Cowens JW, Creaven PJ, Greco WR, Brenner DE, Tung Y , OstroM, Pilkiewicz F, Ginsberg R, Petrelli N . Initial Clinical (Phase I) Trial of TLC D-99 (Doxorubicin Encapsulated in Liposomes). Cancer Res 53,2796-2802,1993. Crooke ST. Molecular mechanisms of action of antisense drugs. Biochim Biophys Acta 1489:31-44, 1999. Dachs GU, Dougherty GJ, Stratford IJ, Chaplin DJ. Targeting gene therapy to cancer: A review. Oncology Research 9:313-325, 1997. Daemen T, Hofstede G, Ten Kate MT, Bakker-Woudenberg IA, Scherphof GL. Liposomal doxorubicin-induced toxicity: depletion and impairment of phagocytic activity of liver macrophages. Int J Cancer 61:716-21,1995. Davila E, Celis E. Repeated administration of cytosine-phosphorothiolated guanine-containing oligodeoxynucleotides together with peptide/protein immunization results in enhanced CTL responses with anti-tumor activity. J Immunol. 165: 539-47, 2000. De Bortoli M , Dati C. Hormonal regulation of type I receptor tyrosine kinase expression in the mammary gland. J Mammary Gland Biology & Neoplasia 2:175-85, 1997. Devine D V , Marjan JMJ. The role of immunoproteins in the survival of liposomes in the circulation. Crit Rev in Therapeutic Drug Carrier Systems 14:105-131, 1997. Dittadi R, Brazzale A , Pappagallo G, Salbe C, Nascimben O, Rosabian A , Gion M . ErbB2 assay in breast cancer: possibly improved clinical information using a quantitative method. Anticancer Res, 17:1245-1247, 1997. Dorr R, Fritz W. Drug data sheets, doxorubicin. In: Cancer chemotherapy handbook. 176 Elsevier Science, New York, pp. 388-401, 1982. Drummond DC, Meyer O, Hong K , Kirpotin DB, Papahadjopoulos D, Optimizing liposomes for delivery of chemotherapeutic agents to solid tumors. Pharmacological Reviews 51:691-743, 1999. Eltabbakh GH, Belinson JL, Kennedy AW, Biscotti C V , Casey G, Tubbs RR. p53 and HER-2/neu overexpression in ovarian borderline tumors. Gynecologic Oncology 65:218-224, 1997. Ferrans VJ , Clark JR, Zhang J Yu Z X , Herman EH. Pathogenesis and prevention of doxorubicin cardiomyopathy. Tsitologiia 39(10):928-37, 1997. Ferrara N , Davis-Smyth T. The biology of vascular endothelial growth factor. Endocrine Reviews 18:4-25, 1997. Folkman J. The role of angiogenesis in tumor growth. Semin Cancer Biol, 3, 65-71, 1992. Fornier M , Munster P, Seidman AD. Update on the management of advanced breast cancer. Oncology 13:647-664, 1999. Gabizon A, Catane R., Uziely B kaufman B, Safra T, Cohen R, Martin F, Huang A, Barenholz Y . Prolonged circulation time and enhanced accumulation in malignant exudates of doxorubicin encapsulated in polyethylene-glycol coated liposomes. Cancer Res 54:987-992,1994. Gabizon A , Goren D, Cohen R Barenholz Y . Development of liposomal anthracyclines: from basics to clinical applications. J Controlled release 53(l-3):275-279, 1998. Gascoyne RD, Adomat SA, Krajewski S, Krajewski M , Horsman DE, Tolcher A W , O'Reilly SE, Paul H, Coldman AJ, Reed JC, Connors JM. Prognostic Significance of Bcl-2 Protein Expression and Bcl-2 Gene Rearrangement in Diffuse Aggressive Non-Hodgekin's Lymphoma. Blood 90:244-251, 1997. Gasparini G. Prognostic significance of vascular endothelial growth factor protein in node-negative breast carcinoma. J Natl Cancer Inst, 89,139-147, 1997. Giatromanolaki A , Koukourakis MI, O'Byrne K, Kaklamanis L, Dicoglou C, Trichia E, Whitehouse R, Harris A L , Gatter K C . Non-small cell lung cancer: c-erbB-2 overexpression correlates with low angiogenesis and poor prognosis. Anticancer Research 16:3819-3826, 1996. Gibbs JB. Mechanism-based target identification and drug discovery in cancer research. Science 287:1969-73, 1969. Giles RV. Antisense oligonucleotide technology: From EST to therapeutics. Curr Opin Mol Ther, 2:238-252, 2000. Glover JM, Leeds JM, Mant TGK, Kisner DL, Zuckerman J, Levin A A , Shanahan WR. Phase I safety and pharmacokinetic profile of an intercellular adhesion molecule-1 antisense oligodeoxynucleotide (ISIS 2302). J Pharmacol Exp Ther 282:1173-1180, 1997. Gonzalez-Vela M C , Garijo MF, Fernandez F, Buelta L, Val-Bernal JF. Cathepsin D in host stromal cells is associated with more highly vascular and aggressive invasive breast carcinoma. Histopathology, 34, 35-42, 1999. Graus-Porta D, Beerli RR, Daly J M and Hynes NE. ErbB-2, the preferred heterodimerization partner of all ErbB receptors, is a mediator of lateral signaling. E M B O J. 16:1647-1655. 1997. Hacker H. Signal transduction pathways activated by CpG-DNA. Curr Top Microbiol Immunol 247:77-92, 2000. 177 Han S, Yun IJ, Noh D Y , Choe KJ , Song SY, Chi JG. Abnormal expression of four novel molecular markers represents a highly aggressive phenotype in breast cancer. Immunohistochemical assay of p53, nm23, erbB-2, and cathepsin D protein. J Surg Oncol, 65,22-7, 1997. Hermann R, Ferguson D, Powers C, Recant W M , Weichselbaum RR, Hellman S. Angiogenesis as a predictor of long-term survival for patients with node-negative breast cancer. J Natl Cancer Inst, 88, 1764-9, 1996. Henderson IC, Canellos GP. Cancer of the breast: The past decade. N Eng J Med 302:78-90, 1980. Henderson BE, Feigelson HS. Epidemiology and Screening In: Textbook of Breast Cancer: A Clinical Guide to Therapy. Bonadonna F, Hortobagyi G N and Gianni A M , Eds. Martin Dunitz Ltd., pp 1-16, 1997. Hengstler JG, Lange J, Kett A , Dornhofer N , Meinert R, Arand M , Knapstein PG, Becker R, Oesch F, Tanner B. Contribution of c-erbB-2 and topoisomerase JJalpha to chemoresistance in ovarian cancer. Cancer Res 59:3206-14, 1999. Henry SP, Monteith D, Levin A A . Antisense oligonucleotide inhibitors for the treatment of cancer: 2. Toxicological properties of phosphorothioate oligodeoxynucleotides. Anti-Cancer Drug Design 12:395-408, 1997. Hobbs SK, Monsky WL, Yuan F, Roberts WG, Griffith L, Torchilin VP, Jain RK. Regulation of transport pathways in tumor vessels: Role of tumor type and microenvironment. Proc Natl Acad Sci USA 95:4607-4612, 1998. Hofmockel G, Riess S, Bassukas ED, Darnmrich J. Epidermal growth factor family and renal cell carcinoma: expression and prognostic impact. European Urology 31:478-484, 1997. Hortobagyi GN. Developments in the chemotherapy of breast cancer. Cancer supplement 88:3073-3079, 2000. Ioachim E, Kamina S, Athanassiadou S, Agnantis NJ. The prognostic significance of epidermal growth factor receptor (EGFR), c-erbB-2, Ki-67 and P C N A expression in breast cancer. Anticancer Res 16:3141-7,1996. Johnstone SA, Masin D, Mayer L, Bally M B . Surface-associated serum proteins inhibit the uptake of phosphatidylserine and poly(ethylene glycol) liposomes by mouse macrophages Biochim Biophys Acta 1513:25-37, 2001. Joukov V , Pajusola K, Kaipainen A, Chilov D, Lahtinen I, Kukk E, Saksela O. A novel vascular endothelial growth factor, VEGF-C, is a ligand for the FLT4 (VEGFR-3) and K D R (VEGFR-2) receptor tyrosine kinases. E M B O J 15:290-298, 1996. Juhl H , Downing SG, Wellstein A , Czubayko F. HER-2/neu is rate-limiting for ovarian cancer growth. Conditional depletion of HER-2/neu by ribozyme targeting. Journal of Biological Chemistry 272:29482-6, 1997. Juliano RL, Alahari S, Yoo H, Kole R, Cho M . Antisense pharmacodynamics: Critical issues in the transport and delivery of antisense oligonucleotides. Pharm Res 16:494-502, 1999. Kanter P M , Bullard GA, Ginsberg RA, Pilkiewicz FG, Mayer LD, Cullis PR, Pavelic ZP. Comparison of the cardiotoxic effects of liposomal doxorubicin (TLC D-99) versus free doxorubicin in beagle dogs. In Vivo 7:17-26, 1993. Keck PJ, Hauser SD, Krivi G, Sanzo K, Warren T, Feder J, Connolly DT. Vascular permeability factor, an endothelial cell mitogen related to PDGF. Science 246:1309, 178 1989. Kirby C, Clark J, Gregoriadis G. Cholesterol content of small unilamellar liposomes controls phospholipid loss to high density lipoprotein. FEBS Lett 111:324-328, 1980. Kirpotin K, Park JW, Hong K, Zalipsky S, L i W-L, Carter P, Benz CC, Papahadjopoulos D. Sterically stabilized anti-HER2 immunoliposomes: Design and targeting to human breast cancer cells in vitro. Biochemistry 36:66-75, 1997. Kluin-Nelemans HC, Limpens J, Meerabux J, Beverstock GC, Jansen JH, de Jong D, Kluin PM. A new non-Hodgkin's B-cell line (DoHH2) with a chromosomal translocation t(14;18)(q32;q21). Leukemia, 5, 221-4, 1991. Koscielny S, Terrier P, Spielmann M , Delarue J-C. Prognostic importance of low c-erbB2 expression in breast tumors. J Natl Cancer Inst, 90:712, 1998. Koscielny S, Terrier P, Spielmann M , Delarue J-C. RESPONSE: re: prognostic importance of low c-erbB2 expression in breast tumors. J Natl Cancer Inst, 91:1585, 1999. Krieg A M . Mechanisms and applications of immune stimulatory CpG oligodeoxynucleotides. Biochim Biophys Acta. 1489: 107-16, 1999. Kurebayashi J, Otsuki T, Kunisue H, Mikami Y , Tanaka K , Yamamoto S, Sonoo H. Expression of vascular endothelial growth factor (VEGF) family members in breast cancer. JpnJ Cancer Research 90:977-81, 1999. Kurebayashi J, Otsuki T, Tang CK, Kurosumi M , Yamamoto S, Tanaka K , Mochizuki M , Nakamura H , Sonoo H. Isolation and characterization of a new human breast cancer cell line, KPL-4, expressing the Erb B family receptors and interleukin-6. Br. J. Cancer 79: 707-17, 1999 Lebwohl DE, Canetta R. New developments in chemotherapy of advanced breast cancer. Ann Oncol 10(suppl.6):S139-S146, 1999. Lee J, Gray A , Yuan J, Luoh S-M, Avraham H, Wood WI. Vascular endothelial growth factor related protein: A ligand and specific activator of the tyrosine kinase receptor Flt4. PNAS 93:1988-1992, 1996. Levin A A . A review of issues in the pharmacokinetics and toxicology of phosphorothioate antisense oligonucleotides. Biochim Biophys Acta 1489:69-84, 1999. Lim HJ, Masin D, Madden TD, Bally M B . Influence of drug release characteristics on the therapeutic activity of liposomal mitoxantrone. J Pharmacol Exp Ther. 281: 566-73, 1997. Lui X , Pogo BG-T. Inhibition of erbB-2-positive breast cancer cell growth by erbB-2 antisense oligonucleotides. Antisense Nucleic Acid Drug Development 6:9-16, 1996. Mannon RB, Nataraj C, Pisetsky DS. Stimulation of thymocyte proliferation by phosphorothioate D N A oligodeoxynucleotides. Cell Immunol. 201: 14-21, 2000. Manoharan M . 2'-Carbohydrate modifications in antisense oligonucleotide therapy: importance of conformation, configuration and conjugation. Biochim Biophys Acta 1489:117-130, 1999. Martiny-Baron G, Marine D. VEGF-mediated tumour angiogenesis: a new target for cancer therapy. Current Opinion in Biotechnology 6:6750680, 1995. Masood R, Cai J, Zheng T, Smith DL, Hinton DR, Gil l PS. Vascular endothelial growth factor (VEGF) is an autocrine growth factor for VEGF receptor-positive human tumors. Blood 98:1904-1913,2001. Maurer N , Mori A , Palmer L, Monck M A , Mok K W C , Mui B, Akhong QF, Cullis PR. Lipid-based systems for the intracellular delivery of genetic drugs. Molecular 179 Membrane Biology 16:129-140, 1999. Mayer LD, Bally M B , Cullis PR. Uptake of adriamycin into large unilamellar vesicles in response to a pH gradient. Biochim Biophys Acta 857:123-126, 1986. Mayer LD, Masin D, Nayar R, Boman NL, Bally M B . Pharmacology of liposomal vincristine in mice bearing L1210 ascitic and B16/BL6 solid tumours. Br J Cancer. 71: 482-488, 1995. Meyer O, Kirpotin D, Hong K, Sternberg B, Park JW, Woodle M C , Papahadjopoulos D. Cationic liposomes coated with polyethylene glycol as carriers for oligonucleotides. J Biol Chem 273:15621-15627, 1998. Mignatti P, Rifkin DB. Plasminogen activators and matrix metalloproteinases in angiogenesis. Enzyme Protein 49:117-37,1996. Miosge N , Schneider W, Gotz W, Herken R. The oncoproteins c-erb-B2, c-fos and the tumour suppressor protein p53 in human embryos and fetuses. Anat Embryol 195:345-352, 1997. Monia BP, Johnston JF, Geiger T, Muller M , Fabbro D. Antitumor activity of a phosphorothioate antisense oligodeoxynucleotide targeted against C-raf kinase. Nat Med 2:668-75, 1996. Monteith DK, Levin A A . Synthetic oligodeoxynucleotides: the development of antisense therapeutics. Toxicol Pathol. 27: 8-13, 1999. Mourali N , Tabbane F, Muenz LR, Behi J, Ben Moussa F, Jaziri M , Levine PH. Ten-year results utilizing chemotherapy as primary treatment in nonmetastatic, rapidly progressing breast cancer. Cancer Invest 11L363-370, 1993. Morin Doody M , Lonstein JE, Stovall M , Hacker DG, Luckyanov N , Land CE. Breast cancer mortality after diagnostic radiography: findings from the U.S. scoliosis cohort study. Spine 25:2052-63, 2000. Muller I, Niethammer D, Bruchelt G. Anthracycline-derived chemotherapeutics in apoptosis and free radical cytotoxicity (Review). Int J Mol Med 1:491-494, 1997. Nagy P, Jenei A , Damjanovich S, Jovin T M , Szollosi J. Complexity of signal transduction mediated by ErbB2: Clues to the potential of receptor-targeted cancer therapy. Pathology Oncology Res. 5(4): 255-71, 1999. Nakano T, Oka K, Ishikawa A, Morita S. Correlation of cervical carcinoma c-erb B-2 oncogene with cell proliferation parameters in patients treated with radiation therapy for cervical carcinoma. Cancer 79:513-520, 1997. Narod SA, Goldgar D, Cannon-Albright L, Weber B, Moslehi R, Ives E, Lenoir G, Lynch H. Risk modifiers in carriers of BRCA1 mutations. Int J Cancer; 64:394-398, 1995. Neckers L M , Lyer K. Non-antisense effects of antisense oligonucleotides. In: Applied Antisense Oligonucleotide Technology. Stein CA, Krieg A M (Eds), Wiley-Liss Inc. 147-159, 1998. NCIC Canadian Cancer Statistics, 2000. Nicklin PL, Bayley D, Giddings J, Craig SJ, Cummins L L , Hastewell JG, Phillips JA. Pulmonary bioavailability of a phosphorothioate oligonucleotide (CGP 64128A: Comparison with other delivery routes. Pharm Res 15:583-591, 1998. Nieves-Niera W, Pommier Y . Apoptotic response to camptothecin and 7-hydroxystaurosporine (UCN-01) in the 8 human breast cancer cell lines of the NCI anticancer drug screen: multifactorial relationships with topoisomerase 1, protein kinase C, Bcl-2, p53, MDM-2 and caspase pathways. Int J Cancer. 82:396-404, 1999. 180 Olayioye M A , Graus-Porta D, Beerli RR, Rohrer J, Gay B, Hynes NE. ErbB-1 and ErbB-2 acquire distinct signaling properties dependent upon their dimerization partner. Mol Cell Biol. 18: 5042-51, 1998. Olofsson B, Pajusola K, Aipaineen A, Von Euler G, Joukov V , Saksela O. Vascular endothelial growth factor B, a novel growth factor for endothelial cells. PNAS 93:2576-81, 1996. Olivotto I, Gelmon K , Kuusk U . Breast Cancer. Glegg C, Ed. Intelligent Patient Guide Ltd. 1998. Paik S, Bryant J, Park C, Fisher B, Tan-Chiu E, Hyams D, Fisher ER, Lippman M E , Wickerham DL, Wolmark N . erbB-2 and response to doxorubicin in patients with axillary lymph node-postive, hormone receptor-negative breast cancer. J Nat Cancer Inst. 90:1361-1370, 1998. Paik S, Hazan R, Fisher ER, Sass RE, Fisher B, Redmond C, Schlessinger J, Lippman M E , King CR. Pathologic findings from the National Surgical Adjuvant Breast and Bowel Project: prognostic significance of erbB-2 protein overexpression in primary breast cancer. J Clin Oncol 8:103-112, 1990. Paik S, King CR, Simpson S, Lippman M E . Quantification of erbB-2/neu levels in tissue. Methods Enzymol, 198:290-300, 1991. Papahadjopoulos D, Allen T M , Gabizon A, Mayhew E, Matthay K , Huang SK, Lee K , Woodle M C , Lasic DD, Redemann C, Martin FJ. Sterically stabilized liposomes: improvements in pharmacokinetics and antitumor therapeutic efficacy. Proc Natl Acad Sci USA. 88:11460-11464,1991. Park JW, Hong K, Kirpotin DB, Meyer O, Papahadjopoulos D, Benz CC. Anti-HER2 immunoliposomes for targeted therapy of human tumors. Cancer Letters 118:153-160, 1997. Parr MJ , Bally M B , Cullis PR. The presence of GM1 in liposomes with entrapped doxorubicin does not prevent RES blockade. Biochim Biophys Acta 1168:249-252, 1993. Pawlowski V , Revillion F, Hebbar M , Hornez L, Peyrat JP. Prognostic value of the type I growth factor receptors in a large series of human primary breast cancers quantified with a real-time reverse transcription-polymerase chain reaction assay. Clin Cancer Res 6:4217-4225, 2000. Pawlowski V , Revillion F, Hornez L, Peyrat JP. Low concentrations of c-erbB-2 mRNA (real time RT-PCR) are associated with a poor prognosis in 404 unselected primary breast cancers. Proc Am Assoc Cancer Res, 40:1355, 1999. Petit A M , Rak J, Hung M C , Rockwell P, Goldstein N , Fendly B, Kerbel RS. Neutralizing antibodies against epidermal growth factor and ErbB-2/neu receptor tyrosine kinases down-regulate vascular endothelial growth factor production by tumor cells in vitro and in vivo: angiogenic implications for signal transduction therapy of solid tumors. Amer J Pathology 151:1523-30, 1997. Piccart MJ , de Valeriola D, Paridaens R, Balikdjian D, Mattheiem W H , Loriaux C, Arrigo C, Cantraine F, Heuson JC. Six-year results of a multimodality treatment strategy for locally advanced breast cancer. Cancer 15:2501-2506, 1988. Pich A , Margaria E, Chiusa L. Oncogenes and male breast carcinoma: c-erbB-2 and p53 coexpression predicts a poor survival. J Clin Oncol 18:2948-2956, 2000. Pietilainen T, Lipponen P, Aaltomaa S, Eskelinen M , Kosma V M , Syrjanen K. The 181 important prognostic value of Ki-67 expression as determined by image analysis in breast cancer. J Cancer Res Clin Oncol 122:687-92, 1996. Pool GL, French M E , Edwards RA, Huang L, Lumb RH. Use of radiolabeled hexadecyl cholesteryl ether as a liposome marker. Lipids. 17:448-52, 1982. Popescu NC, King CR, Kraus M H . Localization of the human erB-2 gene on normal and rearranged chromosomes 17 to bands ql2-21.32. Genomics 4:362-366, 1989 Plouet J, Bayard F. Regulations of vasculotropin/vascular endothelial growth factor bioavailability. Hormone Res 42:14-19,1994. Press MF, Cordon-Cardo C, Slamon DJ. Expression of the HER-2/neu proto-oncogene in normal human adult and fetal tissues. Oncogene 5:953-62, 1990. Press MF, Pike M C , Chazin VR, Hung G, Udove JA, Markowicz M , Danyluk J, Godolphin W, Sliwkowski M , Akita R, Paterson M C , Slamon DJ. Her-2/neu expression in node-negative breast cancer: Direct tissue quantitation by computerized image analysis and association of overexpression with increased risk of recurrent disease. Cancer Res 53:4960-4970, 1993. Probst JC. Antisense oligodeoxynucleotide and ribozyme design. Methods 22:271-281, 2000. Pupa SM, Menard S, Morelli D, Pozzi B, De Palo G, Colnaghi MI. The extracellular domain of the c-erbB-2 oncoprotein is released from tumor cells by proteolytic cleavage. Oncogene 8:2917-2923, 1993. Puyal C, Milhaud P, Bienvenue A, Philippot JR. A new cationic liposome encapsulating genetic material: A potential delivery system for polynucleotides. Eur J Biochem 228:697-703, 1995. Rahman A, Treat J, Roh JK, Potkul L A , Alvord WG, Forst D, Woolley PV. A phase I clinical trial and pharmacokinetic evaluation of liposome-encapsulated doxorubicin. J Clin Oncol 8:1093-1100,1990. Ravdin P M . Prognostic Factors in Breast Cancer In: Textbook of Breast Cancer: A Clinical Guide to Therapy. Bonadonna F, Hortobagyi G N and Gianni A M , Eds. Martin DunitzLtd., pp 35-64, 1997. Rilke F, Di Palma S. Pathology. Pathology. In: Textbook of Breast Cancer: A Clinical Guide to Therapy. Bonadonna F, Hortobagyi G N and Gianni A M , Eds. Martin Dunitz Ltd., pp 17-33, 1997. Roh H, Hirose CB, Boswell CB, Pippin JA, Drebin JA. Synergistic antitumor effects of HER2/neu antisense oligodeoxynucleotides and conventional chemotherapeutic agents. Surgery 126:413-421, 1999. Roh H, Pippin J, Boswell C, Drebin JA. Antisense oligonucleotides specific for the HER2/neu oncogene inhibit the growth of human breast carcinoma cells that overexpress HER2/neu. J Surg Res 77:85-90, 1998. Roh H, Pippin JA, Green DW, Boswell CB, Hirose CT, Mokadam N , Drebin JA. HER2/neu antisense targeting of human breast carcinoma. Oncogene 19:6138-6143, 2000(a). Roh H , Pippin J, Drebin JA. Down-regulation of ER2/neu expression induces apoptosis in human cancer cells that overexpress HER2/neu. Cancer Res 60:560-565, 2000(b). Saxon DN, Lutwyche P, Webb M , Bally M B . Downregulation of HER-2/neu by an anti-HER-2/neu phosphorothioate oligonucleotide and its use in combination with doxorubicin for treatment of breast cancer xenograft models. Proceedings of the 182 American Association for Cancer Research 40:480, 1999. Saxon D, Chia SK, Bally M B . Drug Combinations for Therapy of Aggressive Breast Cancer: Antisense Oligonucleotides Against HER-2/neu Physicians' Education Resource; HER2 in Oncology; 1(3), 2000. Schechter A L , Stern DF, Vaidyauathan L, Decker SJ, Drebin JA, Greene MI, Weinberg RA. The neu oncogene: an erbB-related gene encoding a 185 OOOMw tumour antigen. Nature 312:513, 1984. Scherphof G, Roerdink F, Waite M , Parks J. Disintegration of phosphatidylcholine liposomes in plasma as a result of interaction with high-density lipoproteins. Biochim Biophys Acta 542:296-307, 1978. Schmitt JF, Susil BJ, Hearn MTW. Aberrant FGF-2, FGF-3, FGF-4 and C-ERB-B2 gene copy number in human ovarian, breast and endometrial tumours. Growth Factors. 13:19-35, 1996. Schroit AJ , Madsen JW, Tanka Y . In vivo recognition and clearance of red blood cells containing phosphatidylserine in their plasma membranes. J Biol Chem 260:5131-5138,1985. Semple SC, Klimuk SK, Harasym TO, Dos Santos N , Ansell SM, Wong K F , Maurer N , Stark H, Cullis PR, Hope MJ, Scherrer P. Efficient encapsulation of antisense oligonucleotides in lipid vesicles using ionizable aminolipids: formation of novel small multilamellar vesicle structures. Biochim Biophys Acta 78005:1-15, 2000. Semple SC, Klimuk SK, Harasym TO, Hope MJ. Lipid-based formulations of antisense oligonucleotides for systemic delivery applications. Methods Enzymol 313:322-341, 1999. Senger DR, Connolly DT, Van De Water L, Feder J, Dvorak HF. Purification and NH2-terminal amino acid sequence of guinea pig tumour-secreted vascular permeability factor. Cancer Res 50:1774, 1990. Senger DR, Van De Water L, Brown LF, Bag JA, Yeo K-T, Yeo T-K, Berse B, Jackman RW, Dvorak A M , Dvorak HF. Vascular permeability factor (VPF/VEGF) in tumour biology. Cancer and Metastasis Reviews 12:303-324, 1993. Sharma A, Pratap M , Sawhney V M , Khan IU, Bhambhani S, Mitra A B . Frequent amplification of C-erbB2 (HER-2/Neu) oncogene in cervical carcinoma as detected by non-fluorescence in situ hybridization technique on paraffin sections. Oncology 56:83-87, 1999. Shih C, Padhy L C , Murray M , Weinberg RA. Transforming genes of carcinomas and neuroblastomas introduced into mouse fibroblasts. Nature 290:261-264, 1981. Siegel P M , Muller WJ. Tyrosine kinases and signal transduction in mouse mamamry tumorigenesis. In: R. B. Dickson and D. S. Salomon (eds.), Hormones and Growth Factors in Development and Neoplasia, pp. 397-419. New York: Wiley-Liss, Inc., 1998. Skorski T, Nieborowska-Skorska M , Nicolaides NC, Szczylik C, Iversen P, Iozzo RV, Zon G, Calabretta B. Suppression of Philadelphia 1 leukemia cell growth in mice by BCR-A B L antisense oligodeoxynucleotide. Proc Natl Acad Sci U S A 91:4504-8, 1994. Slamon DJ, Clark G M , Wong SG, Levin WJ, Ullrich A, McGuire WL. Human breast cancer: Correlation of relapse and survival with amplification of the Her-2/neu Oncogene. Science 235:177-82, 1987. 183 Slamon DJ, Godolphin W, Jones L A , Holt JA, Wong SG, Keith DE, Levin WJ, Stuart SG, Udove J, Ullrich A . Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science 244:707-712, 1989. Slamon D, Leyland-Jones B, Shak S Paton V, Bajamonde A, Fleming T, Eiermann W, Wolter J, Baselga J, Norton L. Addition of Herceptin™ (humanized anti-HER2 antibody) to first line chemotherapy for HER2 overexpressing metastatic breast cancer (HER2+/MBC) markedly increases anticancer activity: a randomized, multinational controlled phase III trial. Proc Am Soc Clin Oncol. 17:98a Abstract #377, 1998. Solomayer EF, Diel IJ, Meyberg C, Gollan C, Bode S, Wallwiener D, Bastert G. Prognostic relevance of cathepsin D detection in micrometastatic cells in the bone marrow of patients with primary breast cancer. Breast Cancer Res Treat, 49, 145-54, 1998. Sparwasser, T., Miethke, T., Lipford, G., Erdmann, A., Hacker, H. , Heeg, K. , and Wagner, H. Macrophages sense pathogens via D N A motifs: induction of tumor necrosis factor-alpha-mediated shock. Eur J Immunol. 27: 1671-9, 1997. Stanford JL, Herrington LJ Schwartz SM, Weiss NS. Breast cancer incidence in Asian immigrants to the United States and their descendants. Epidemiology 6:181-3, 1995. Stein CA, Krieg A M . Problems in interpretation of data derived from in vitro and in vivo use of antisense oligodeoxynucleotides. Antisense Res Dev 4:67-69, 1994. Stuart DD, Kao GY, Allen T M . A novel long-circulating, and functional liposomal formulation of antisense oligodeoxynucleotides targeted against MDR1. Cancer Gene Ther 7:466-475, 2000. Tan M , Yao J, Y u D. Overexpression of the c-erbB-2 gene enhanced intrinsic metastasis potential in human breast cancer cells without increasing their transformation abilities. Cancer Res 57:1199-1205, 1997. Tandon A K , Clark G M , Chamness GC, Ullrich A , McGuire W L HER-2/neu oncogene protein and prognosis in breast cancer. J Clin Oncol 7:1120-8, 1989. Tannapfel A , Kiihn R, KeBler H, Wittekind C. Expression of c-erbB2 oncogene product in different tumours and its standardised evaluation. Analytical Cellular Pathology 10:149-160, 1996. Tari A , Khodadadian M , Ellerson D, Deisseroth A, Lopez-Berenstein G. Liposomal delivery of oligodeoxynucleotides. Leukemia and Lymphoma 21:93-97, 1996. Tari A M , Tucker SD, Deisseroth A , Lopez-Berenstein G. Liposomal delivery of methylphosphonate antisense oligodeoxynucleotides in chronic myelogenous leukemia. Blood 84:601-607, 1994. Thor A D , Berry DA, Budman DR, Muss HB, Kute T, Henderson IC, Barcos M , Cirrincione C, Edgerton S, Alfred C, Norton L, Liu ET. erbB-2, p53, and efficacy of adjuvant therapy in lymph node-positive breast cancer. J Nat Cancer Inst. 90: 1346-1360, 1998. Toikkanen S, Helin H, Isola J, Joensuu H. Prognostic significance of HER-2 oncoprotein expression in breast cancer: a 30-year follow-up. J Clin Oncol 10:1044-1048, 1992. Tardi PG, Boman N L , Cullis PR. Liposomal Doxorubicin. J Drug Targeting 4(3): 129-140, 1996. Uedo N , Tatsuta M , Baba M , Hirasawa R, Iishi H , Yano H , Sakai N , Uehara H, Nakaizumi A. Inhibition by rat C-ernN-2/neu antisense oligonucleotide of gastric carcinogenesis induced by N-methyl-N'-nitro-N-nitrosoguanidine in Wistar rats. Int J Cancer 83:670-184 673, 1999. Vaughn JP, Iglehart JD, Demirdji S, Davis P, Babiss L E , Caruthers M H , Marks JR. Antisense D N A downregulation of the ERBB2 oncogene measured by a flow cytometric assay. Proc Natl Acad Sci USA 92:8338-8342, 1995. Vaughn JP, Stekler J, Demirdji S, Mills JK, Caruthers M H , Iglehart JD, Marks JR. Inhibition of the erbB-2 tyrosine kinase receptor in breast cancer cells by phosphoromonothioate and phosphorodithioate antisense oligodeoxynucleotides. Nucleic Acids Res 24: 4558-64, 1996. Waterhouse DN, Tardi PG, Mayer LD, Bally M B . Changing toxicity profiles: A comparison of liposomal formulations of doxorubicin to drug administered in free form. Drug Safety 24: 903-920, 2001. Webb MS, Harasym TO, Masin D, Bally M B , Mayer LD. Sphingomyelin-cholesterol liposomes significantly enhance the pharmacokinetic and therapeutic properties of vincristine in murine and human tumour models. Br J Cancer 72: 896-904, 1995. Webb MS, Logan P, Kanter P M , St.-Onge G, Gelmon K, Harasym T, Mayer L D , Bally M B . Preclinical pharmacology, toxicology and efficacy of sphingomyelin/cholesterol liposomal vincristine for therapeutic treatment of cancer. Cancer Chemother Pharmacol. 42: 461-70, 1998(a). Webb MS, Saxon D, Wong F M , Lim HJ, Wang Z, Bally M B , Choi LS, Cullis PR, Mayer LD. Comparison of different hydrophobic anchors conjugated to poly(ethylene glycol): effects on the pharmacokinetics of liposomal vincristine. Biochim Biophys Acta. 1372: 272-82, 1998(b). Webb MS, Boman NL, Wiseman DJ, Saxon D, Sutton K , Wong K F , Logan P, Hope MJ . Antibacterial efficacy against an in vivo Salmonella typhimurium infection model and pharmacokinetics of a liposomal ciprofloxacin formulation. Antimicrob Agents Chemother 42:45-52, 1998c. Weidner N , Folkman J, Pozza F, Bevilacqua P, Allred EN, Moore D H , Meli S, Gasparini G. Tumor angiogenesis: a new significant and independent prognostic indicator in early-stage breast carcinoma. J Natl Cancer Inst 84:1875-87, 1992. Woolf T M , Melton DA, Jennings CG. Specificity of antisense oligonucleotides in vivo. Proc Natl Acad Sci USA 7305, 1992. Xiong S, Grijalva R, Zhang L, Nguyen NT, Pisters PQ, Pollock RE, Y u D. Up-regulation of vascular endothelial growth factor in breast cancer cells by the heregulin-betal-activated p38 signaling pathway enhances endothelial cell migration. Cancer Res 61:1727-1732,2001. Yamamoto T, Dcawa S, Akiyama T, Semba K , Nomura N , Miyajima N , Kaito T, Toyoshima K. Similarity of protein encoded by the human c-erb-B-2 gene to epidermal growth factor receptor. Nature 319:230-234, 1986. Yang JL, Ow KT, Russell PJ, Ham JM, Crowe PJ. Higher expression of oncoproteins c-myc, c-erbB-2/neu, PCNA and p53 in metastasizing colorectal cancer than in nonmetastasizing tumors. Annals of Surgical Oncology, 3:574-579, 1996. 185