Tumor-Associated Fibroblasts and their Matrix
The Tumor Microenvironment Series Editor: Isaac P. Witz
For further volumes: http://www.springer.com/series/7529
Margareta M. Mueller • Norbert E. Fusenig Editors
Tumor-Associated Fibroblasts and their Matrix
1 3
Editors Prof. Dr. Margareta M. Mueller Research Group Tumor and Microenvironment German Cancer Research Center, Heidelberg and HFU (Hoschschule Furtwangen University) campus Villingen-Schwenningen Villingen-Schwenningen Germany
[email protected] Prof. Dr. Norbert E. Fusenig Former Division Head at the German Research Cancer Center, Heidelberg and Emeritus of the University of Heidelberg Heidelberg Germany
[email protected];
[email protected] ISBN 978-94-007-0658-3 e-ISBN 978-94-007-0659-0 DOI 10.1007/978-94-007-0659-0 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2011925928 © Springer Science+Business Media B.V. 2011 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Cover design: deblik, Berlin Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)
Preface
During the last century cancer research was mainly focussed on the tumor cells alone which could be easily propagated in cell culture. During this time many important findings were obtained clearly demonstrating that cancer is a genetic disease, controlled by the activation and/or inactivation of critical control genes. However during the last two decades it has become increasingly clear that genetic alterations alone are not the sole driving force behind tumor development but that tumor growth and progression are rather intimately controlled by the microenvironment. One could almost speak of are “rediscovery” of the tumor as a highly complex tissue composed of carcinoma cells and surrounding stroma. Studies in different areas of biology including tumour biology have demonstrated that tissue structure, function and dysfunction are highly intertwined with the microenvironment and that during the development of cancer tissue biology and host physiology are subverted to drive malignant progression. It is now clear that the context is crucial and that the status of the cellular microenvironment plays a significant role in determining whether cells within a tissue retain their normal architecture or undergo tumor progression. The tumor stroma or microenvironment is made up of multiple non-malignant cell populations, including fibroblasts, adipocytes, endothelial and inflammatory cells that are embedded in a tumour specific extracellular matrix (ECM). Nowadays, there is a huge interest in tumor stroma research, and in understanding the contributions of the different stromal cell types to tumor growth and progression. One of the key components of the tumor microenvironment in carcinomas are activated fibroblasts termed cancer associated fibroblasts (CAFs). In the meantime our knowledge of CAFs has changed from being viewed as a passive bystander to becoming an important co-mediator of cancer progression. In response to cancer growth, host stromal fibroblasts undergo a dramatic morphologic and biochemical transition to form “reactive stroma” in a desmoplastic reaction much like the granulation tissues found at the site of wound healing. While the malignant cells activate fibroblasts in the tumor stroma by various stimuli, including growth factors and cytokines, cancer associated fibroblasts secrete growth factors and build a permissive soil in which the cancer cells thrive. CAFs are responsible for the elaboration of most of the connective tissue and ECM components v
vi
Preface
as well as, proteolytic enzymes and their inhibitors. The composition and structure of the ECM in the tumor microenvironment is essential for promoting tumor development and metastasis. The constituents of the ECM include collagens, laminins, fibronectin and several proteoglycans. They provide mechanical support for cells, facilitate cell communication and serve as substrates for cell migration. Changes in the composition or architecture of the extracellular matrix within tumors can alter integrin expression and function and promote metastatic progression, angiogenesis and lymphangiogenesis. In this unique textbook world leading experts of the area of tumor microenvironment review the most recent knowledge of the still growing complexity of the tumor microenvironment focussing on tumor associated stromal cells and the most important extracellular matrix components and summarize the role of these players in tumor progression. Moreover, novel therapeutic targets are discussed that have been discovered in the tumor microenvironment and are increasingly used in experimental and clinical tumor therapy. The message from their contributions is clear: the tumor microenvironment and its components are important and essential players in tumor progression and interesting targets for novel therapeutic strategies. However there are still many white areas on the map and we are just beginning to understand the complex interplay between tumor and stromal cells. We express our deepest gratitude to all our colleagues who have made this book the first comprehensive antology covering all major aspects of the role of the tumor microenvironment and its extracellular matrix components. Heidelberg
Margareta M. Mueller and Norbert E. Fusenig
Contents
Part I The Tumor Microenvironment 1 Critical Roles of Stromal Fibroblasts in the Cancer Microenvironments ���������������������������������������������������������������������� 3 Leland W. K. Chung Part II Stromal Cell Diversity 2 Functional Diversity of Fibroblasts ��������������������������������������������������������� 23 H. Peter Rodemann and Hans-Oliver Rennekampff 3 The Role of the Myofibroblast in Fibrosis and Cancer Progression ���� 37 Boris Hinz, Ian A. Darby, Giulio Gabbiani and Alexis Desmoulière 4 The Role of Myofibroblasts in Communicating Tumor Ecosystems ���� 75 Olivier De Wever, Astrid De Boeck, Pieter Demetter, Marc Mareel and Marc Bracke 5 Tumor Vessel Associated-Pericytes ���������������������������������������������������������� 91 Arne Bartol, Anna M. Laib and Hellmut G. Augustin 6 The Role of Cancer-Associated Adipocytes (CAA) in the Dynamic Interaction Between the Tumor and the Host ������������������������ 111 Marie-Christine Rio Part III The Tumor ECM 7 Hyaluronan: A Key Microenvironmental Mediator of Tumor-Stromal Cell Interactions �������������������������������������������������������� 127 Naoki Itano 8 Function of Tenascins in the Tumor Stroma ������������������������������������������� 145 Florence Brellier and Ruth Chiquet-Ehrismann vii
viii
Contents
9 Fibulins and Their Role in the ECM ��������������������������������������������������� 159 Helen C. M. Cooney and William M. Gallagher 10 Tumor Fibroblast-Associated Metalloproteases ��������������������������������� 175 Julie Lecomte, Anne Masset, Dylan R. Edwards and Agnès Noël Part IV Tumor Modulating-Fibroblast Interactions 11 Multiple Fibroblast Phenotypes in Cancer Patients: Heterogeneity in Expression of Migration Stimulating Factor ��������� 197 Ana M. Schor and Seth L. Schor 12 TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression in Adjacent Epithelia ������������������������������������������������ 223 Brian R. Bierie and Harold L. Moses 13 The SDF-1-Rich Tumour Microenvironment Provides a Niche for Carcinoma Cells ���������������������������������������������������������������� 245 Masayuki Shimoda, Kieran Mellody and Akira Orimo 14 Role of PDGF in Tumor-Stroma Interactions ������������������������������������ 257 Carina Hellberg and Carl-Henrik Heldin 15 Radiation-Induced Microenvironments and Their Role in Carcinogenesis ���������������������������������������������������������������������������������� 267 Mary Helen Barcellos-Hoff and David H. Nguyen Part V Tumor-Modulating ECM Interactions 16 The Extracellular Matrix as a Multivalent Signaling Scaffold that Orchestrates Tissue Organization and Function ��������� 285 Jamie L. Inman, Joni D. Mott and Mina J. Bissell 17 SPARC and the Tumor Microenvironment ���������������������������������������� 301 Stacey L. Thomas and Sandra A. Rempel 18 Integrin-Extracellular Matrix Interactions ���������������������������������������� 347 Christie J. Avraamides and Judith A. Varner 19 The Multifaceted Role of Cancer Associated Fibroblasts in Tumor Progression ���������������������������������������������������������������������������� 361 Hans Petter Eikesdal and Raghu Kalluri
Contents
ix
Part VI Therapeutic Application/Targeting 20 Cancer Associated Fibroblasts as Therapeutic Targets ��������������������� 383 Christian Rupp, Helmut Dolznig, Christian Haslinger, Norbert Schweifer and Pilar Garin-Chesa 21 Targeting Tumor Associated Fibroblasts and Chemotherapy ���������� 403 Debbie Liao and Ralph A. Reisfeld 22 Antibody-Based Targeting of Tumor Vasculature and Stroma ��������� 419 Katharina Frey and Dario Neri Index ��������������������������������������������������������������������������������������������������������������� 451
.
Contributors
Hellmut G. Augustin Vascular Oncology and Metastasis, German Cancer Research Center (DKFZ-ZMBH Alliance), 69120 Heidelberg, Germany Medical Faculty Mannheim,Vascular Biology and Tumor Angiogenesis, Heidelberg University, 68167 Mannheim, Germany Christie J. Avraamides Moores UCSD Cancer Center, University of California, San Diego, 3855 Health Sciences Drive #0819, La Jolla, CA 92093-0819, USA e-mail:
[email protected] Mary Helen Barcellos-Hoff Departments of Radiation Oncology and Cell Biology, NYU Langone Medical Center, 566 First Avenue, 10016 New York, USA e-mail:
[email protected] Arne Bartol Vascular Oncology and Metastasis, German Cancer Research Center (DKFZ-ZMBH Alliance), 69120 Heidelberg, Germany Medical Faculty Mannheim, Vascular Biology and Tumor Angiogenesis, Heidelberg University, 68167 Mannheim, Germany e-mail:
[email protected] Brian R. Bierie Department of Cancer Biology, Vanderbilt-Ingram Cancer Center, 691 Preston Research Building 2220 Pierce Ave., Nashville, TN 37232-6838, USA e-mail:
[email protected] Mina J. Bissell Life Science Division, Lawrence Berkeley National Laboratory, One Cyclotron Rd, Berkeley, CA 94720, USA e-mail:
[email protected] Astrid De Boeck Laboratory of Experimental Cancer Research, Department of Radiotherapy and Experimental Cancer Research, Ghent University Hospital, De Pintelaan 185, 9000 Ghent, Belgium Marc Bracke Laboratory of Experimental Cancer Research, Department of Radiotherapy and Experimental Cancer Research, Ghent University Hospital, De Pintelaan 185, 9000 Ghent, Belgium
xi
xii
Contributors
Florence Brellier Friedrich Miescher Institute for Biomedical Research, Novartis Research Foundation, Maulbeerstrasse 66, 4058 Basel, Switzerland Ruth Chiquet-Ehrismann Friedrich Miescher Institute for Biomedical Research, Novartis Research Foundation, Maulbeerstrasse 66, 4058 Basel, Switzerland e-mail:
[email protected] Leland W. K. Chung Uro-Oncology Research Program, Departments of Medicine and Surgery, Samuel Oschin Comprehensive Cancer Institute, Cedars-Sinai Medical Center and the University of California, Los Angeles, CA 90048, USA e-mail:
[email protected] Helen C. M. Cooney 73 Nutley Lane, Donnybrook, Dublin 4, Ireland e-mail:
[email protected] Ian A. Darby Cancer and Tissue Repair Laboratory, School of Medical Sciences, RMIT University, Bundoora, VIC 3083, Australia Pieter Demetter Department of Pathology, Erasme University Hospital, Université Libre de Bruxelles (ULB), Brussels, Belgium Alexis Desmoulière Faculty of Pharmacy, Cellular Homeostasy and Pathologies (EA 3842) and Department of Physiology, IFR 145, University of Limoges, 87025 Limoges, France Helmut Dolznig Institute of Pathology, Medical University of Vienna, Waehringer Guertel 18–20, 1090 Vienna, Austria Institute of Medical Genetics, Centre of Pathobiology and Genetics, Medical University of Vienna, Waehringer Strasse 10, 1090 Vienna, Austria e-mail:
[email protected] Dylan R. Edwards School of Biological Sciences, University of East Anglia, Norwich, NR4 7TJ, UK Hans Petter Eikesdal Department of Oncology, Haukeland University Hospital, 5021 Bergen, Norway Katharina Frey Department of Chemistry and Applied Biosciences, ETH Zurich, Wolfgang-Pauli-Str. 10, 8093 Zurich, Switzerland e-mail:
[email protected] Giulio Gabbiani Department of Pathology and Immunology, CMU, University of Geneva, Rue Michel-Servet 1, 1211 Geneva 4, Switzerland e-mail:
[email protected] William M. Gallagher UCD School of Biomolecular and Biomedical Science, UCD Conway Institute, University College Dublin, Dublin 4, Ireland e-mail:
[email protected] Pilar Garin-Chesa Institute of Pathology, Medical University of Vienna, Waehringer Guertel 18–20, 1090 Vienna, Austria
Contributors
xiii
Boehringer Ingelheim RCV GmbH & Co KG, Dr. Boehringer-Gasse 5–11, 1130 Vienna, Austria Christian Haslinger Boehringer Ingelheim RCV GmbH & Co KG, Dr. BoehringerGasse 5–11, 1130 Vienna, Austria Carl-Henrik Heldin Ludwig Institute for Cancer Research, Uppsala University, Box 595, SE 751 24 Uppsala, Sweden e-mail:
[email protected] Carina Hellberg Ludwig Institute for Cancer Research, Uppsala University, Box 595, SE 751 24 Uppsala, Sweden e-mail:
[email protected] Boris Hinz Faculty of Dentistry, Laboratory of Tissue Repair and Regeneration, Matrix Dynamics Group, University of Toronto, Toronto, ON M5S 3E2, Canada Jamie Inman Life Science Division, Lawrence Berkeley National Laboratory, One Cyclotron Rd, Berkeley, CA 94720, USA Naoki Itano Department of Molecular Oncology, Division of Molecular and Cellular Biology, Institute on Aging and Adaptation, Shinshu University Graduate School of Medicine, Matsumoto, Nagano 390-8621, Japan e-mail:
[email protected] Raghu Kalluri Division of Matrix Biology, Department of Medicine, Beth Israel Deaconess Medical Center & Harvard Medical School, 330 Brookline Ave, E/CLS Room #11-090, Center for Life Sciences, 02115 Boston, MA, USA Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA, USA Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA, USA e-mail:
[email protected] Anna M. Laib Vascular Oncology and Metastasis, German Cancer Research Center (DKFZ-ZMBH Alliance), 69120 Heidelberg, Germany Julie Lecomte Laboratory of Tumor and Development Biology, Groupe Interdisciplinaire de Génoprotéomique Appliqué-Cancer (GIGA-Cancer), University of Liège, 4000 Liège, Belgium Debbie Liao Department of Immunology and Microbial Sciences, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, USA Marc Mareel Laboratory of Experimental Cancer Research, Department of Radiotherapy and Experimental Cancer Research, Ghent University Hospital, De Pintelaan 185, 9000 Ghent, Belgium Anne Masset Laboratory of Tumor and Development Biology, Groupe Interdisciplinaire de Génoprotéomique Appliqué-Cancer (GIGA-Cancer), University of Liège, 4000 Liège, Belgium
xiv
Contributors
Kieran Mellody CR-UK Stromal-Tumour Interaction Group, Paterson Institute for Cancer Research, The University of Manchester, Wilmslow Road, Manchester, M20 4BX, UK Harold L. Moses Department of Cancer Biology, Vanderbilt-Ingram Cancer Center, 691 Preston Research Building 2220 Pierce Ave., Nashville, TN 372326838, USA e-mail:
[email protected] Joni D. Mott Life Science Division, Lawrence Berkeley National Laboratory, One Cyclotron Rd, Berkeley, CA 94720, USA Dario Neri Department of Chemistry and Applied Biosciences, ETH Zurich, Wolfgang-Pauli-Str. 10, 8093 Zurich, Switzerland e-mail:
[email protected] Agnès Noël Laboratory of Tumor and Development Biology, Groupe Interdisciplinaire de Génoprotéomique Appliqué-Cancer (GIGA-Cancer), University of Liège, 4000 Liège, Belgium e-mail:
[email protected] Akira Orimo CR-UK Stromal-Tumour Interaction Group, Paterson Institute for Cancer Research, The University of Manchester, Wilmslow Road, Manchester, M20 4BX, UK e-mail:
[email protected] Ralph A. Reisfeld Department of Immunology and Microbial Sciences, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, USA e-mail:
[email protected] Sandra A. Rempel Department of Neurosurgery, Hermelin Brain Tumor Center, Henry Ford Hospital, Detroit, MI 48202, USA e-mail:
[email protected] Marie-Christine Rio Institut de Génétique et de Biologie Moléculaire et Cellulaire, CNRS UMR 7104, INSERM U964, UDS, BP, 67404 Illkirch Cedex, 10142 C.U. de Strasbourg, France e-mail:
[email protected] H. Peter Rodemann Division of Radiobiology & Molecular Environmental Research, Department of Radiation Oncology, Eberhard Karls University Tübingen, Röntgenweg 11, 72076 Tübingen, Germany e-mail:
[email protected] Hans-Oliver Rennekampff Hospital for Plastic, Hand and Reconstructive Surgery, Medical School Hannover, Hannover, Germany Christian Rupp Institute of Pathology, Medical University of Vienna, Waehringer Guertel 18–20, 1090 Vienna, Austria
Contributors
xv
Ana M. Schor Unit of Cell and Molecular Biology, The Dental School, University of Dundee, Dundee DD1 4HR, UK e-mail:
[email protected] Seth L. Schor Unit of Cell and Molecular Biology, The Dental School, University of Dundee, Dundee DD1 4HR, UK e-mail:
[email protected] Norbert Schweifer Boehringer Ingelheim RCV GmbH & Co KG, Dr. BoehringerGasse 5–11, 1130 Vienna, Austria Masayuki Shimoda Department of Pathology, School of Medicine, Keio University, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan CR-UK Stromal-Tumour Interaction Group, Paterson Institute for Cancer Research, The University of Manchester, Wilmslow Road, Manchester, M20 4BX, UK Stacey L. Thomas Department of Neurosurgery, Hermelin Brain Tumor Center, Henry Ford Hospital, Detroit, MI 48202, USA e-mail:
[email protected] Judith A. Varner Moores UCSD Cancer Center, University of California, San Diego, 3855 Health Sciences Drive #0819, La Jolla, CA 92093-0819, USA Department of Medicine, University of California, San Diego, La Jolla, CA, 920930819, USA e-mail:
[email protected] Olivier De Wever Laboratory of Experimental Cancer Research, Department of Radiotherapy and Experimental Cancer Research, Ghent University Hospital, De Pintelaan 185, 9000 Ghent, Belgium e-mail:
[email protected] Part I
The Tumor Microenvironment
.
Chapter 1
Critical Roles of Stromal Fibroblasts in the Cancer Microenvironments Leland W. K. Chung
1.1 Introduction A mounting body of evidence suggests that the ability of cancer cells to interact reciprocally with the host microenvironment contributes to cancer growth, progression and resistance to therapeutic interventions. Interactions with stromal fibroblasts at the primary site and marrow stromal cells at metastatic sites could create a favorable microenvironment supporting cancer growth, survival, evasion of immune surveillance and resistance to therapy (Mueller and Fusenig 2004; Chantrain et al. 2008; Karnoub et al. 2007; Ronnov-Jessen and Bissell 2009; Chung et al. 2006). Changes in cancer microenvironments could also add selection pressures favorable to cancer cell evolution, increasing cancer cell heterogeneity and reciprocally causing the co-evolution of adjacent stromal fibroblasts, resulting in the development of organ- and stage-specific stromal fibroblasts capable of programming and reprogramming the fate of cancer cells (Hill et al. 2005; Sung et al. 2008; Franco et al. 2010). Over the past several years, active research has broadened our awareness of the plasticity of cancer-associated stroma, which undergo both morphologic and functional transitions supporting the pathogenesis of cancer cells (Sung et al. 2008). The dynamic presence of bone marrow-derived mesenchymal stem cells in localized and metastatic cancers contributes further to the diversity and heterogeneity of cancer-associated stroma and ultimately determines the site of cancer metastasis, while stromal fibroblasts have multiple functional roles modulating cancer growth either positively or negatively (Martin et al. 2010; Molloy et al. 2009; Rhodes et al. 2009; Zhao et al. 2009a). Figure 1.1 depicts how soluble factors, insoluble extracellular matrix proteins, and reactive oxygen species secreted by cancer cells and cells in cancer microenvironments can guide and maintain the growth and differentiation of local and distant cancers and their interactions with host microenvironments L. W. K. Chung () Uro-Oncology Research Program, Departments of Medicine and Surgery, Samuel Oschin Comprehensive Cancer Institute, Cedars-Sinai Medical Center and the University of California, Los Angeles, CA 90048, USA e-mail:
[email protected] M. M. Mueller, N. E. Fusenig (eds.), Tumor-Associated Fibroblasts and their Matrix, The Tumor Microenvironment 4, DOI 10.1007/978-94-007-0659-0_1, © Springer Science+Business Media B.V. 2011
3
4
L. W. K. Chung
Bone MET
Bone stroma
MSC Secondary metastasis
Hematogenous metastasis Blood vessel Cancer cells
BF Brain
MSC Activation
Macrophages Platelets Activated stroma T and B cells Dendritic cells
ROS GFs
Normal
Lymphatic metastasis
Transformation EMT
ECM NAF
Secondary metastasis
NAF
Benign
ROS GFs
CAF
Transformed
LF
ROS GFs CAF
Malignant
LNF Lymph node
Lung
Prostate
Fig. 1.1 Tumor-microenvironments interactions contribute to prostate cancer skeletal and soft tissue metastases. At the primary site, reciprocal cellular interactions between prostate cancer cells and local/migrating stromal fibroblasts, mediated by soluble and insoluble factors and ROS, promote the transition of normal/benign stromal fibroblasts to cancer-associated stromal fibroblasts which activate and transform prostate epithelial cells to gain increased growth and survival potential. Subsequently, through epithelial-to-mesenchymal transition (EMT), prostate cancer cells gain futher growth, survival, migratory, invasive and metastatic potential allowing them to metastasize to the skeleton via hematogeneous routes. Prostate cancer cells also frequently metastasize to lymph node and then reach the skeleton through lymphatic metastatic routes. Metastatic prostate cancer cells, after reaching the bone, adhere to and exit from marrow endothelium via transendothelial migration, undergo mesenchymal-to-epithelial transition (MET) and colonize in the bone by increased expression of cell–cell adhesive/junction proteins such as E-cadherin. The metastatic cascade is aided by the active participation of both resident fibroblasts, such as lymph node fibroblast (LNF), lung fibroblast (LF), and brain fibroblast (BF), and migrating stromal fibroblasts such as cells derived from mesenchymal stem cells (MHC) at primary and metastatic sites. The multipotent MHC can differentiate into inductive cell populations in the tumor microenvironment, including reactive stroma, macrophage, platelet, dendritic cell and T- and B-cells, in the primary and at metastatic sites, to ‘mark’ the site prior to the occurrence of secondary metastases
(Karnoub et al. 2007; Chung et al. 2006; Desgrosellier and Cheresh 2010; Ishikawa et al. 2008; Jung et al. 2009; Kaplan et al. 2005; Svineng et al. 2008). The therapeutic targeting of cancer cells, though necessary, is insufficient by itself to control the growth of localized and disseminated cancers. This has led to the acceptance of co-targeting both the cancer and its microenvironments, including stromal fibroblasts, the vascular endothelial network, immune cells and host humoral substances (Chung et al. 2005; Cress and Mohla 2004; Tu and Lin 2008; Vessella and Corey
1 Critical Roles of Stromal Fibroblasts in the Cancer Microenvironments
5
2006; Pollard 2009). The unique cell signaling networks linking the behavior of cancer cells (i.e. cell proliferation, resistance to apoptosis, cell migration, invasion and metastasis) with the activation of key downstream signaling pathways can be exploited as new and promising druggable targets (Chung et al. 2006; Anastasiadis et al. 2003; Kang and Altieri 2009; Tasseff et al. 2010). In this review, we use prostate cancer as a model to dissect the critical determinants that regulate growth, differentiation and progression of prostate cancer to the skeleton, the lethal human prostate cancer phenotype. Looking forward, the study of cancer-associated stromal fibroblasts is rapidly evolving and could take center stage in revealing the secrets of cancer cell evolution. Recent exciting discoveries include the reprogramming adult normal stromal fibroblasts to form induced pluripotent stem (iPS) cells capable of orchestrating the development and differentiation of an entire embryo (Takahashi and Yamanaka 2006; Yu et al. 2009; Okita et al. 2007; Park et al. 2008; Zhao et al. 2009b). These findings raise new questions about the pathways leading to stromal fibroblast heterogeneity and the potential of in situ reprogramming of adult stromal fibroblasts to undergo iPS transition in an organ-specific environment. iPS cells could serve as progenitor cells for a subsequent generation of derivative cells comprising the entire stromal microenvironments, raising the possibility of developing stroma-based cancer therapy in the future.
1.2 The Roles of Stromal Fibroblasts in the Context of Tumor Microenvironment 1.2.1 Local Microenvironment Cancer growth and evolution is intimately controlled by its microenvironment, and cancer cells also contribute reciprocally to the active process of remodeling their microenvironment (Ingber 2008; Rhee et al. 2001; Pathak et al. 1997). Cancer cells secrete soluble factors such as EGF, IGFs, PDGF, VEGF, HGF, FGFs, and TGFβs which collectively stimulate pleiotropic signaling in converging multi-signaling pathways that induce activated stromal fibroblasts or myofibroblasts, with potent growth-promoting effect on cancer cells. Local microenvironments are heterogeneous and composed of resident vascular endothelial cells, smooth muscle cells, basal/stem cells and, to a lesser extent, cells from neural and neuroendocrine lineages that could interact with cancer cells in a reciprocal manner through secreted soluble factors and insoluble extracellular matrices. In addition to resident stromal fibroblasts, migrating mesenchymal stem cells from bone marrow, with either hematopoietic or mesenchymal lineage, also contribute to stromal heterogeneity. These cells, macrophages, platelets, dendritic cells, T- and B-cells, and activated stroma, are likely to support local cancer growth, progression and distant metastasis through complex cancer–stroma, cancer–immune and cancer–stem cell interactions
6
L. W. K. Chung
(Josson et al. 2010; Singh et al. 2009; Jaganathan et al. 2007; McKeithen et al. 2010). The local heterogeneity of stromal fibroblasts could be determined by their anatomical location. Our laboratory recently showed that human prostate stromal fibroblasts derived from the peripheral zone of the prostate gland are more inductive than stromal fibroblasts derived from the transitional or central zones, and these results are consistent with the observation that progressive prostate cancer is derived predominantly from the peripheral rather than transitional or central zones (Thalmann et al. 2009).
1.2.2 Distant Microenvironments Because of the propensity of prostate cancer to metastasize to bone, which is considered lethal, much effort has focused on defining the bone microenvironment and the mechanisms of bone turnover, including enhanced bone resorption, that contribute to the ability of cancer cells to colonize bone. Several key cell types in the bone microenvironment are of particular importance. Among these are the osteoblasts (OBs), bone-forming cells derived from bone marrow mesenchymal stem cells (MSC). Upon interaction with soluble factors such as bone morphogenic proteins, TGFβs, EGF, FGFs, or PDGF, MSC can potentially differentiate into OBs, chondrocytes, or adipocytes. In addition, there are osteoclasts (OCs), bone resorbing cells derived from monocytes of hematopoietic mesenchymal cell lineage. OCs express receptor activator of NF-kappaB (RANK), a receptor responding to RANK ligand (RANKL), secreted by osteoblasts or prostate cancer cells, promoting maturation of OCs, inducing the fusion of monocytes to form activated multinucleated OCs. These matured multinucleated OCs contribute to bone resorption or bone turnover resulting in the release of soluble growth factors, nutrients, calcium ions, and extracellular matrices (Araujo and Logothetis 2009; Buckle et al. 2010; Mizutani et al. 2009). The actions of these soluble and matrix factors alter cancer cell adhesion, proliferation and survival and also the responsiveness of the host microenvironments toward factors secreted by both cancer cells and cells in cancer microenvironments. Collectively, the factors present in the cancer milieu could determine how cancer cellinduced osteoblastic or osteolytic lesions ultimately support cancer cell colonization in bone. In addition to the local action of cancer microenvironments, the factors secreted by cancer microenvironments could conceivably govern the propensity of secondary cancer metastases to organs such as the lung, liver, brain, and kidney.
1.3 The Plasticity of the Stromal Microenvironment 1.3.1 Reactive Stroma In response to cancer growth, host stromal fibroblasts undergo a dramatic morphologic and/or biochemical transition to form “reactive stroma” in a desmoplas-
1 Critical Roles of Stromal Fibroblasts in the Cancer Microenvironments
7
tic reaction much like the granulation tissues found at the site of wound healing (Tuxhorn et al. 2001, 2002; Malins et al. 2006). Although the desmoplastic reaction associated with human prostate cancer is less apparent than in human breast cancer and melanoma, the transition of stromal fibroblasts to myofibroblasts at the gene expression level is quite apparent. Upon transition to reactive stromal cells, they express more abundant and diverse classes of extracellular matrix proteins with altered expression of genes associated with myofibroblasts, such as α- and γ-smooth muscle actin, fibronectin, actin bundle, paladin, Thy1, and TGF-β1 (Sung et al. 2008; Untergasser et al. 2005; Dakhova et al. 2009). A number of soluble growth factors, when added directly to cultured stromal fibroblasts, have been shown to induce myofibroblast transition (Olaso et al. 2003; Cushing et al. 2008; Kennard et al. 2008; Kikuta et al. 2006). Direct interaction between reactive stroma and cancer cells has been observed to promote cancer growth. This is consistent with clinical observations where the detection of reactive stroma in the prostate cancer microenvironments, for example, predicts PSA recurrence and the clinical outcome in prostate cancer patients (Dakhova et al. 2009; Tothill et al. 2008; Yanagisawa et al. 2007). The engagement between cancer cells and reactive stroma could promote tissue reorganization involving the participation of the vascular network and migrating MSC and host stromal fibroblasts to lead to increased cancer growth (Zhao et al. 2009a; Santamaria-Martinez et al. 2009). The prevalence of myofibroblasts in the cancer environment has been shown in many different forms of cancer including colon, liver, lung, prostate, ovary, pancreas, and breast (Tuxhorn et al. 2001; Friedman et al. 1984; Garin-Chesa et al. 1990; Radisky and Przybylo 2008; Yao et al. 2009). The myofibroblastic appearance often precedes the onset of cancer invasion.
1.3.2 Plasticity of EMT and MET Dynamic bi-directional epithelial-to-mesenchyme (EMT) and mesenchymal-toepithelial (MET) transitions have been observed in embryonic development and in cancer progression (Chung et al. 2006; Prindull 2005; Birchmeier et al. 1996; Wells et al. 2008; Hugo et al. 2007). These transitions are commonly associated with a predictive switch of cancer behaviors by the affected cancer epithelium, which assumes increased migratory, invasive and metastatic potential, as assessed by changes in cell morphology and gene expression profiles. This is the rationale for designing novel EMT/MET-based targeting strategies (Sabbah et al. 2008; Moen et al. 2009; Ponzo et al. 2009). In the context of cancer microenvironments, these transitions offer a possible new explanation of the origins of the inductive stromal fibroblasts and the responding cancer epithelial cells, since both cancer epithelial cells and stromal fibroblasts can be derived from either resident or migrating pluripotent stem cells or from a selective population of transforming cancer epithelial or reactive stromal cells through interactions with specific cell types or factors in the cancer microenvironment (Santamaria-Martinez et al. 2009; Leber and Efferth 2009). A small side population of pancreatic cancer stem cells was recently reported to be particularly sensitive to EMT induction by TGF-β (Kabashima et al. 2009).
8
L. W. K. Chung
Likewise, circulating cancer cells were shown to be sensitive to growth factor induction to undergo EMT and MET (Vessella et al. personal communication).
1.3.3 Mesenchymal Stem Cells (MSC) MSC are a class of multipotent stem cells capable of differentiating into osteoblasts, chondrocytes, and adipocytes. They can be derived from bone marrow stromal cells and also from adult linage stem cells with self-renewal capability. These cells were found to be present in cancer microenvironments, with the potential of promoting cancer growth and progression (Roorda et al. 2009), exerting immunosuppression by interfering with dendritic and T-cell functions (Spaeth et al. 2008) and ‘marking’ the sites where cancer cells subsequently metastasize (Jung et al. 2009; Kaplan et al. 2005). These functions are generally accomplished by the ability of MSC to secrete specific factors which, via circulatory network and paracrine interaction, confer migratory, invasive and metastatic potential to cancer cells at the primary site of cancer growth. They also interact at the site of metastasis, for instance by increasing bone turnover to create a favorable microenvironment supporting the dissemination of cancer cells (Kaplan et al. 2007).
1.4 The Mediators and Cell Signaling Network Governing the Plasticity of Stromal Fibroblasts Soluble and insoluble mediators secreted by cancer cells and cells in cancer microenvironments are responsible for supporting the growth and progression of cancer by interacting with selective receptors that transmit signals orchestrating a switch in cancer cell morphology and function compatible with the survival of cancer cells. In this section, specific examples defining prostate cancer interactions with soluble and matrix proteins will be used to illustrate the importance of understanding the cell signaling network to identify relevant therapeutic targets for clinical translation and develop new drugs.
1.4.1 Soluble Growth Factors The general model depicted in Fig. 1.1 shows cancer cell-secreted soluble factors promoting the activation of both cancer-associated resident stromal fibroblasts and migrating MSC and/or their derivative stromal fibroblasts. This triggers additional remodeling of tumor microenvironments, reciprocally affecting the genotype and phenotype of both cancer cells and stromal fibroblasts in the cancer microenvironment via soluble factors including growth factors, cytokines, chemokines, and
1 Critical Roles of Stromal Fibroblasts in the Cancer Microenvironments
9
reactive oxygen species (ROS) released in the tumor microenvironment by resident and migrating host mesenchymal or stromal cells. These factors often work in concert to induce a primarily stromal reaction manifested by activated stromal fibroblasts or myofibroblasts producing either higher levels and/or more effective combinations of soluble growth factors capable of inducing cancer growth, invasion and metastasis directly, and also of promoting the reorganization of the vascular network favoring the dissemination of cancer cells to distant organs. In other words, soluble factors produced by cancer and host stroma can be disseminated to distant sites where they become responsible for remodeling the premetastatic niche and facilitating subsequent cancer cell dissemination (Kaplan et al. 2007). In some cases, soluble factors can act at long distance at metastatic sites to modulate tumor microenvironments by providing higher concentrations or more complementary growth factors via increased bone turnover, creating a less hostile environment for the growth of cancer cells via immune suppression, or promoting osteomimicry within cancer cells and creating a metastatic bone ‘niche’ favoring overall cancer cell growth and survival at metastatic sites (Chung et al. 2006; Cooper et al. 2003). Soluble factors secreted by cancer and activated stromal cell components within the tumor microenvironment can also modulate other cell signaling networks mediated by integrin-extracellular matrix interactions (Sangaletti et al. 2008; Chiarugi and Giannoni 2005) and androgen receptor signaling pathways (Huang et al. 2006; Olapade-Olaopa et al. 1999), activating the cell signaling network to upregulate integrins and/or androgen receptor expression (Liegibel et al. 2002; Bonaccorsi et al. 2006).
1.4.2 Extracellular Matrices (ECMs) Cancer cell proliferation and survival within the tumor microenvironment depends on the ability to adhere and attach to ECMs (Desgrosellier and Cheresh 2010). Through ECM-integrin interactions, cancer cells can also gain increased invasive, migratory and metastatic potential, mediated by the activation of converging cell signaling networks downstream that confer growth and survival advantages to cancer cells (Pontier and Muller 2009). ECM-integrin interactions, known to affect embryonic development (Armant 2005), also play a directive role in determining the gene expression profiles of cancer cells, the ability of cancer cells to degrade their surrounding ECMs by matrix metalloproteinases (MMPs), to extravasate into metastatic microenvironments by adhesion to organ-specific endothelial cells (Kargozaran et al. 2007), and the status of differentiation of cancer cells.
1.4.3 ROS and Oxidative Stress High levels of ROS and oxidative stress can be created in cancer microenvironments by the continued growth and expansion of solid tumors which deplete oxygen
10
L. W. K. Chung
supplies and build up metabolic waste at the site of tumor growth. Cancer cells with downregulated manganese superoxide dismutase (MnSOD) show increased levels of superoxide, which induce hypoxia inducing factor 1 (HIF-1α) and VEGF, causing increased angiogenesis and tumor growth (Kaewpila et al. 2008; Wang et al. 2005). An accumulation of hydrogen peroxide can also be induced by placing cancer cells under hypoxic conditions and in cells with defective catalase (Azad et al. 2009). Hydrogen peroxide is a potential mediator contributing to the co-evolution of the genotype and phenotype of both prostate cancer and bone stroma when they are in contact under 3-D co-culture conditions (Sung et al. 2006, 2008). Hypoxic conditions can also affect the transdifferentiation of cancer cells such as EMT and MET, where hypoxia induces EMT (Klymkowsky and Savagner 2009; Jiang et al. 2007) and hyperbaric oxygen treatment causes MET (Moen et al. 2009) with respective corresponding changes of either increased or decreased cancer growth and invasion in animals.
1.5 Overall Significance and Clinical Translation of Integrated Approaches to Cancer–Stromal Fibroblast Interaction Our laboratory has established a 3-dimensional (3-D) co-culture system to investigate how the information derived from cancer–stromal fibroblast interaction can be applied in the clinic and to understand the molecular pathways that determine the behaviors of cancer cells. This approach is based on our prior work showing that cancer cell phenotype and genotype can be irreversibly “programmed” when cancer cells are grown together with prostate or bone stromal cells under 3-D conditions as prostate organoids (Sung et al. 2008; Rhee et al. 2001) or in mice as tumor xenografts (Sung et al. 2008). We found several important features of these types of cellular interactions. (1) The irreversible “programming” of the phenotype and genotype of cancer cells by stromal fibroblasts is bi-directional. We observed that human stromal fibroblasts co-cultured with human prostate cancer cells under 3-D conditions can program the genotype (assessed by cytogenetics and genome-wide scan (Sung et al. 2008) and phenotype (measured by gene expression and ability to grow tumors with metastatic potential in mice (Sung et al. 2008). Remarkably, normal stromal fibroblasts from mouse, benign/normal human prostate stromal fibroblasts, and the MG-63 osteosarcoma cell line have also been observed to undergo irreversible and non-random genotypic and phenotypic changes when co-cultured with prostate cancer cells under 3-D conditions (Sung et al. 2008; Rhee et al. 2001). (2) Gene expression analyses revealed that stromal fibroblasts, after physical contact with prostate cancer cells, had increased levels of brain-derived neurotropic factor (BDNF), chemokines, CCL5 and CXCL5, versican, tenascin, connective tissue growth factor, stromal cell derived factor-1 (SDF-1/CXCL12), and HIF-1α (Sung et al. 2008). We have validated the overexpression of these biomarkers identified by our cell culture model in clinical tissue and serum samples collected from
1 Critical Roles of Stromal Fibroblasts in the Cancer Microenvironments
11
prostate cancer patients with confirmed bone metastasis (Sung et al. 2008). These studies highlight the bidirectional interactions and the co-evolution of tumor-stroma in prostate cancer progression.
1.6 Co-targeting Tumor and Stroma as an Effective Therapeutic Strategy for the Treatment of Cancer and Cancer Metastasis Cancer-host microenvironment communication in the primary and at distant metastatic sites, mediated by soluble factors and insoluble matrices, supports the growth and survival of cancer cells. This provides a sound rationale for co-targeting both the tumor and the host microenvironment to achieve better tumor growth control and improve overall patient survival. Cancer development is complex, involving multiple interactions between different cell types and pleiotropic signaling mechanisms leading to progression. Reciprocal interaction between prostate cancer cells and resident and migrating cells in the tumor microenvironments mediated by cell signaling networks should be considered viable targets. Prostate cancer frequently metastasizes to bone and Fibulin 4, Fibulin 1 > Fibulin 3
Fibulin 4 though not fibulins 3 and 5 interact with collagen IV collagen XV derived endostatin and nidogen-2. The second subgroup of fibulins does not interact with FN and most basement membrane proteins. Dieter P Reinhardt and coworkers (Reinhardt et al. 1996) demonstrated in vitro colocalization and high affinity binding between fibulin 2 and fibrillin 1. Fibrillins (1, 2 and 3) are similar glycoproteins to the fibulins. The Fibrillins are the major proteins of microfibrils which are found close to the basement membrane in the ECM. Hubmacher et al. (2008) demonstrated in vitro how fibrillin 1 was able to self assemble. They showed that the C-terminal half of fibrillin 1 assembles into multimeric globular proteins stabilized by disulphide bonds. Then these globules join head to tail and appear like beads-on-a-string. The fibrillin-1 functionally has 43 cbEGF domains interspersed with TGFβ binding protein
9 Fibulins and Their Role in the ECM
167
domains and hybrid domains. The N-C self interaction site of fibrillin is on cbEGF 41–43 and the binding site for interaction with fibulin-2 was previously narrowed down to the N terminus of fibrillin-1 at the amino acid positions range 45–450 (Reinhardt et al. 1996). Fibulin 2 was present in perichondrium, elastic intima of blood vessels and kidney glomerulus though not in tendon or lung alveoli. Localization of the fibulin-2 was present at the interface between elastin cores and microfibrils. Kobayashi et al. (2007) has demonstrated that fibulin-5 in present in a very similar distribution in relation to the microfibrils. Fibulin-4 is found on the other hand in the microfibril surrounding the elastin core. They concluded from their work using immunoelectron microscopy and recombinant techniques that in contrast to fibulins 1 and 2 which primarily bind to elastic fibers and basement membrane, members of the second subgroup (3, 4 and 5) primarily bind to elastic fiber components. Ablation of genes showed that fibulin 4 and 5 play an essential role in the assembly of elastic fibres during development. Figure 9.3 has been compiled from (Albig and Schiemann 2004) to depict the effect of fibulin-5 on angiogenic sprouting by endothelial cells and the interactions TGFβ tes ula ion im ess t S pr Ex
Down-regulation +
C
KS
–
BL
O
–
Thrombospondin
+
FIBULIN 5
VEGF
TSP I Expression
– Antagonize
–
– +
E
ND
OTHEIA
gen
ic S
pro
L
gio
L
An
LC
E l elia tion oth fera Endll proli Ce
l lia n the io do nvas n E ll i Ce
utin
g
Fig. 9.3 Factors controlling fibulin 5 activity and its own effect on cell activities
168
H. C. M. Cooney and W. M. Gallagher
of other mediators on the effects of fibulins 5. Fibulin 5 shown centrally antagonises the endothelial cell invasion and proliferation. These endothelial cells in turn cause down regulation of fibulin 5 expression. TGFβ stimulates expression of Fibulin 5 there by decreasing angiogenesis. VEGF, a pro-angiogenic factor directly promotes vascularization and indirectly by blocking the effect of TGFβ on fibulin-5. Fibulin 5 modulates the effect of the VEGF signal and also increases TSP I expression which antagonises endothelial cell from invasion and proliferation. VEGF and TSP I have a reciprocal relationship (i.e. where VEGF is low TSP I is high and visa versa). Fibulin 6 functions in the origination of specific epithelial cell junctions (Sisto et al. 2009) and it is located in the pericellular ECM of epithelial cells (Xu et al. 2007). Detachment of cells from their ECM leads to apoptosis known in this context as ANOIKIS (Sisto et al. 2009). Loss of appropriate attachment of epithelial cells to ECM in the salivary gland may give rise to Sjogren’s syndrome. Fibulin 7 is a glycoprotein secreted by 3 pre-odontoblasts and odontoblasts in developing teeth. It is also found in hair, cartilage and placenta. It is a cell adhesion molecule (de Vega et al. 2009), likely by an intregin receptor and a heparin sulphate cell surface receptor. It binds dental mesenchyme cells and odontoblasts but not dental epithelial cells (forerunner of enamel matrix secreting ameloblast cells). Its role appears to be in the development and maintenance of odontoblasts and dentin formation.
9.3 Role of Fibulins in Cancer The fibulins have a number of possible roles in promotion and inhibition of the neoplastic transformation of cells. They may facilitate or inhibit subsequent tissue changes such as angiogenesis and the breakdown of tissue restraints when tumour cells metastasize. Evidence for the involvement of fibulin-1 in tumourigenesis comes from a number of sources. 1. In vitro work showed that fibulin-1 supresses fibronectin regulated cell adhesion and mobility (Twal et al. 2001) and delays invasion and tumour progression in fibrosarcoma cells (Qing et al. 1997). 2. Fibulin-1 expression is elevated in breast and ovarian cancer (Roger et al. 1998; Greene et al. 2003). 3. The first source suggested that fibulin-1 suppressed carcinogenesis whereas the second source pointed to fibulin-1 being oncogenic. Gallagher et al. (2005) suggested the dicotomy could be explained by alternative splice variant production with fibulin 1-C being pro-oncogenic and fibulin 1-D being tumour suppressive. Estradiol has been shown to regulate fibulin-1 expression in ovarian cancers (Moll et al. 2002). Moreover further work has shown that fibulin-1 expression is regulated at transcriptional level and post-transcriptional level. The half life (t1/2) of mRNA translating to fibulin 1-D is decreased while the
9 Fibulins and Their Role in the ECM
169
mRNA for fibulin 1-C is not reduced. This results in larger amounts of fibulin-1C being available under these conditions as opposed to the less stable fibulin 1-D (i.e. the ratio of fibulin 1 C/fibulin 1 D is increased). Gallagher et al. (2005) also suggested that due to the close association of fibulin-1 with basement membranes and connective tissue matrix fibres, fibulin-1 might be degraded by proteinases and modulate its ability to regulate tumour invasion through the barrier of the basement membrane. Wlazlinski et al. (2007) found that fibulins as well as laminins were differentially expressed in prostate cancers partly brought about by differences in DNA methylation. Fibulin 1, 4 and 5 were down regulated in prostate cancers relative to controls. Fibulin-1 down regulation did not fit with that found in breast and ovarian cancers and the ratio of splice-variants fibulin 1C to fibulin 1D remained constant. However, fibulin-4 and especially fibulin 5 down regulation in prostate cancer agreed well with that found for other cancers. Xie et al. (2008) found using HT1080 tumour cells that Fibulin 1 and fibulin 5 both inhibited tumour angiogenesis. Fibulin-1 also inhibited tumour growth by induction of apoptosis of tumour cells. Furthermore they used the lysozomal enzyme Cathepsin D to partially degrade fibulin-1. A protein fragment was produced that contained the antiangiogenic activity of fibulin-1. This fragment had a similar molecular size to fibulin-5 and a possible cleavage site for Cathepsin D on fibulin-1 aligned with the N-terminal position of fibulin-5 sequence. Both fibulin-1 and fibulin-5 were derived from the basement membrane. Fibulin-4 has been shown by (Gallagher et al. 2001) to be upregulated in a large number of human colon carcinoma biopsies. They suggest this is due to enhanced translation or increased stability of the mRNA for fibulin-4. Gallagher et al. (1999) showed that fibulin-4 interacts with the tumour suppressor protein p53 and this interaction inhibits tumourigenesis. Independently p53 also inhibits tumourigenesis. The interesting features of p53 are that the wild-type protein causes decreased vascularisation, likely through repression at transcription level of factors such as thrombospondin I and VEGF (Dameron et al. 1994; Mukhopadhyay et al. 1995). A mutant of p53 on the other hand does the exact opposite and increases VEGF with resulting increased angiogenesis and thereby is an oncogene, (Kieser et al. 1994). The p53 mutants interestingly can behave as wild type at 32°C and revert to mutant type as the temperature is increased probably due to a conformation change in the protein. To implicate the interaction of mutant p53 with fibulin-4 as the oncogene stimulus one has to explain how they can interact if fibulin-4 is extracellular and p53 is intracellular. This may be explained by 1. accepting that fibulin-4 on occasion does not have an N-terminal signal peptide, thus making it likely that on these occasions fibulin-4 is not a secreted protein, or 2. accepting that the evidence points to mutant p53 and fibulin-4 meeting in the lumen of the endoplasmic reticulum. Either of these explanations would make it reasonable to suggest that fibulin-4 has mutant p53 oncogenic abilities. As discussed previously there is a large body of evidence that points to fibulin-5 having an anti tumourigenic effect. This is due to its effect on endothelial cells and
170
H. C. M. Cooney and W. M. Gallagher
particularly to its negative effect on angiogenesis. The decreased ability to form new blood vessels is thought to decrease a tumours ability to grow and metastasize. This is true of human tumours such as kidney, breast, ovary and colon (Schiemann et al. 2002). Along with in vitro inhibition of angiogenesis it has also been shown that fibulin-5 is down regulated when there is active endothelial cell tubulogenesis. In vivo studies (Spencer et al. 2005) have shown that fibulin-5 deficient mice showed increased cell proliferation, motility and invasion in response to a mutagenic stimulus. This could be reversed by overexpressing fibulin-5. However in the case of HT1080 fibrosarcoma cells fibulin-5 increases cell proliferation and invasion, which is at odds with its effects on the other cell types and tumours. There is evidence for an overall tendency for fibulin-5 to suppress malignancy as many primary and metastatic tumours are accompanied by a down regulation of fibulin-5 (Gallagher et al. 2005).
9.4 Fibulins in Other Disease States As previously mentioned fibulin-5 plays a key role in the development and in maintaining the integrity of elastic tissue. Mutations which cause dysfunctional elastic tissues cause loose skin, joint laxity, defective large arteries and heart valves and emphysematous lungs due to the loss of elastic recoil. Three forms of cutis laxa are known (Loeys et al. 2002) two of which are autosomal recessive and one is autosomal dominant, with loose skin and systemic manifestations of varying severity. One such disease is thought to be due to a homozygous missense mutation in the fibulin-5 gene causing autosomal recessive cutis laxa and pulmonary emyphysema. Loeys and coworkers described a very rare case of cutis laxa type 1 (Loeys et al. 2002). At one month old this child had loose skin of the face and neck with a “droopy” facial appearance which gave the illusion of the eyes being displaced downward (sunset phenomenon). Furthermore this child had supra-aortic valve stenosis with thickened aortic valve and was shown to have pulmonary emphysema at 6 months with recurrent lower respiratory tract infections. Histology of skin sections showed poor elastin fiber development compared to controls. These workers investigated a fibulin-5 gene mutation as the causative agent in this hereditary disease. A mutation (T-C) was found at position 998 of the fibulin-5 gene giving rise to a serine-to-proline substitution at 227 in the fourth Ca++ binding EGF-like domain of fibulin-5. This segregated with the disease phenotype in the pedigree of the child that was studied. The further evidence for cause and effect between the mutation and the disease was in four parts. This evidence supports the belief that serine 227 is necessary for the correct function of the fibulin-5 protein. 1. The T998C mutation was absent in 100 chosen controls. 2. The serine is located at position in the secondary and tertiary structure of the protein that makes it very important for maintenance of protein conformation.
9 Fibulins and Their Role in the ECM
171
It is the position of the fourth cbEGF-like domain between cysteines 3 and 4 which remains constant (highly conserved) between different species and in other human fibulins such as fibulin-3. 3. In a similar ECM protein, i.e. fibrillin-1, 31 of its 43 cbEGF-like domains have a serine residue in a corresponding place. 4. The mutation is considered to have important structural and thus functional consequences in other proteins with cbEGF-like modules in that the folding of the recombinantly expressed domain is altered in vitro (Wu et al. 1995). Marfan’s syndrome is a well known disease of elastic tissue and is regarded as a fibrillinopathy. It is caused by an analogous mutation in fibrillin-1, serine to proline in a position between the third and the fourth cysteine in a cbEGF-like domain. Such a change in an analogous position in another cbEGF-like module of fibrillin-1 has been shown to increase susceptibility to proteases and alter the packing of nearby protein modules. Proper folding ensures less proteolysis of the protein (Downing et al. 1996). Hu et al. (2006) provided a further insight into the mechanism by which mutation of fibulin-5 caused impaired elastic fiber development and recessive cutis laxa. Fibulin-5 mutation S227P and another mutation caused reduced synthesis and secretion of fibulin-5. Mutant fibulin-5 had impaired association with elastic fibers due to reduced binding to tropoelastin and fibrilliln-1 microfibrils. Cells that express fibulin-5 are subject to increased apoptosis. Interactions between elastic fibers and fibrillin-1 microfibrils are disrupted by missense mutant fibulin-5 and globular elastin deposits are seen instead of the mature continuous elastic fiber core. Synpolydactyly, a dominantly inherited congenital disorder of the hands and feet, where the metatarsals and metacarpals are fused, is due to a translocation in the last intron of the fibulin-1 gene. The alternative splicing that brings about fibulin 1-D results in its loss following the translocation. This insufficiency probably gives rise to digit malformation because fibulin 1-D is required for cell immigration and apoptosis. Another defect in fibulin 1-D expression is associated with the giant platelet syndromes that result in macrothrombocytopenia and also deafness, renal disease and other anomalies. Interestingly, a protein associated with fibulin-1 namely fibrillin-2 can be subjected to mutation, and this gives rise to the same congenital anomaly of synpolydactyly as a mutation of fibulin-1 itself (Chaudhry et al. 2001). A study by (Stone et al. 1999) showed that a mutation in fibulin-3, in the last cbEGF module, namely Arg 345 Trp, was associated with two autosomal dominant macular degenerative diseases. In support of this mutation being causative, (Chu and Tsuda 2004) argued that recombinantly produced mutant (Arg 345 Trp) fibulin-3 was shown to be misfolded and secreted inefficiently (Marmorstein et al. 2002). The two diseases are known as Malattia Leventineses and Doyne honeycomb retinal dystrophy and they map close to the fibulin-3 locus, Chap. 2 p16–p21. However several families had these diseases without the fibulin-3 mutation (Tarttelin et al. 2001) suggesting molecular genetic heterogeneity. Fibulin-2 being predominantly expressed in aorta and blood vessels has a large number of polymorphisms which may genetically modify this protein to
172
H. C. M. Cooney and W. M. Gallagher
be involved in common vascular diseases but this is improved (Chu and Tsuda 2004). Not unlike fibulin-1 and some of the other fibulins a missense mutation in the fibulin-4 protein gives rise to the skin condition cutis laxa. It also gives abnormalities in other tissues which have a high content of elastic tissue necessary for proper function. These are mainly the lungs and the large arteries. Hucthagowder et al. (2006) studied an autosomal recessive syndrome due to a mutation in the fibulin-4 gene. They found the missense mutation G169A that caused a collection of connective tissue disorders not unlike those due to mutations in the other fibulins. Additionally, other glycoproteins such as fibrillins and laminins also give rise to similar disorders when they undergo a mutation. Fibulin 6 dysfunction has thus far been associated with a least two diseases. 1. Blindness due to age-related macular degeneration. This is partly due to a gln5375arg transition in fibulin 6 (Schultz et al. 2003). 2. Sjogren’s Syndrome which is an autoimmune disease of the salivary and lacrimal glands (decreased salve and tear production) has been shown to be associated with a reduction or dysfunction in fibulin 6 (Sisto et al. 2009). Fibulin 7 dysfunction has not yet been shown to be related to a specific disease state.
References Albig AR, Schiemann WP (2004) Fibulin-5 antagonizes vascular endothelial growth factor (VEGF) signaling and angiogenic sprouting by endothelial cells. DNA Cell Biol 23(6):367–379 Argraves WS, Dickerson K et al (1989) Fibulin, a novel protein that interacts with the fibronectin receptor beta subunit cytoplasmic domain. Cell 58(4):623–629 Chaudhry SS, Gazzard J et al (2001) Mutation of the gene encoding fibrillin-2 results in syndactyly in mice. Hum Mol Genet 10(8):835–843 Chu ML, Tsuda T (2004) Fibulins in development and heritable disease. Birth Defects Res C Embryo Today 72(1):25–36 Cohen S, Elliott GA (1989) The stimulation of epidermal keratinization by a protein isolated from the submaxillary gland of the mouse. 1962. J Invest Dermatol 92(Suppl 4):157S; discussion 158S–159S Dameron KM, Volpert OV et al (1994) The p53 tumor suppressor gene inhibits angiogenesis by stimulating the production of thrombospondin. Cold Spring Harb Symp Quant Biol 59:483– 489 de Vega S, Iwamoto T et al (2007) TM14 is a new member of the fibulin family (fibulin-7) that interacts with extracellular matrix molecules and is active for cell binding. J Biol Chem 282(42):30878–88 de Vega S, Iwamoto T et al (2009) Fibulins: multiple roles in matrix structures and tissue functions. Cell Mol Life Sci 66(11–12):1890–1902 Downing AK, Knott V et al (1996) Solution structure of a pair of calcium-binding epidermal growth factor-like domains: implications for the Marfan syndrome and other genetic disorders. Cell 85(4):597–605 Gallagher WM, Argentini M et al (1999) MBP1: a novel mutant p53-specific protein partner with oncogenic properties. Oncogene 18(24):3608–3616
9 Fibulins and Their Role in the ECM
173
Gallagher WM, Greene LM et al (2001) Human fibulin-4: analysis of its biosynthetic processing and mRNA expression in normal and tumour tissues. FEBS Lett 489(1):59–66 Gallagher WM, Currid CA et al (2005) Fibulins and cancer: friend or foe? Trends Mol Med 11(7):336–340 Giltay R, Timpl R et al (1999) Sequence, recombinant expression and tissue localization of two novel extracellular matrix proteins, fibulin-3 and fibulin-4. Matrix Biol 18(5):469–480 Greene LM, Twal WO et al (2003) Elevated expression and altered processing of fibulin-1 protein in human breast cancer. Br J Cancer 88(6):871–878 Hu Q, Loeys BL et al (2006) Fibulin-5 mutations: mechanisms of impaired elastic fiber formation in recessive cutis laxa. Hum Mol Genet 15(23):3379–3386 Huber R, Scholze H et al (1980) Crystal structure analysis and molecular model of human C3a anaphylatoxin. Hoppe Seylers Z Physiol Chem 361(9):1389–1399 Hubmacher D, El-Hallous EI et al (2008) Biogenesis of extracellular microfibrils: multimerization of the fibrillin-1 C terminus into bead-like structures enables self-assembly. Proc Natl Acad Sci U S A 105(18):6548–6553 Hucthagowder V, Sausgruber N et al (2006) Fibulin-4: a novel gene for an autosomal recessive cutis laxa syndrome. Am J Hum Genet 78(6):1075–1080 Hugli TE et al (1986) Biochemistry and biology of anaphylatoxins. Complement 3(3):111–127 Kieser A, Weich HA et al (1994) Mutant p53 potentiates protein kinase C induction of vascular endothelial growth factor expression. Oncogene 9(3):963–969 Kobayashi N, Kostka G et al (2007) A comparative analysis of the fibulin protein family. Biochemical characterization, binding interactions, and tissue localization. J Biol Chem 282(16):11805– 11816 Loeys B, Van Maldergem L et al (2002) Homozygosity for a missense mutation in fibulin-5 (FBLN5) results in a severe form of cutis laxa. Hum Mol Genet 11(18):2113–2118 Marmorstein LY, Munier FL et al (2002) Aberrant accumulation of EFEMP1 underlies drusen formation in Malattia Leventinese and age-related macular degeneration. Proc Natl Acad Sci U S A 99(20):13067–13072 Moll F, Katsaros D et al (2002) Estrogen induction and overexpression of fibulin-1C mRNA in ovarian cancer cells. Oncogene 21(7):1097–1107 Mukhopadhyay D, Tsiokas L et al (1995) Wild-type p53 and v-Src exert opposing influences on human vascular endothelial growth factor gene expression. Cancer Res 55(24):6161–6165 Pan TC, Kluge M et al (1993) Sequence of extracellular mouse protein BM-90/fibulin and its calcium-dependent binding to other basement-membrane ligands. Eur J Biochem 215(3):733–740 Qing J, Maher VM et al (1997) Suppression of anchorage-independent growth and matrigel invasion and delayed tumor formation by elevated expression of fibulin-1D in human fibrosarcoma-derived cell lines. Oncogene 15(18):2159–2168 Reinhardt DP, Sasaki T et al (1996) Fibrillin-1 and fibulin-2 interact and are colocalized in some tissues. J Biol Chem 271(32):19489–19496 Roger P, Pujol P et al (1998) Increased immunostaining of fibulin-1, an estrogen-regulated protein in the stroma of human ovarian epithelial tumors. Am J Pathol 153(5):1579–1588 Sasaki T, Mann K et al (1997) Dimer model for the microfibrillar protein fibulin-2 and identification of the connecting disulfide bridge. EMBO J 16(11):3035–3043 Savage CR Jr, Inagami T et al (1972) The primary structure of epidermal growth factor. J Biol Chem 247(23):7612–7621 Schiemann WP, Blobe GC et al (2002) Context-specific effects of fibulin-5 (DANCE/EVEC) on cell proliferation, motility, and invasion. Fibulin-5 is induced by transforming growth factorbeta and affects protein kinase cascades. J Biol Chem 277(30):27367–27377 Schultz DW, Klein ML et al (2003) Analysis of the ARMD1 locus: evidence that a mutation in HEMICENTIN-1 is associated with age-related macular degeneration in a large family. Hum Mol Genet 12(24):3315–3323 Sisto M, D’Amore M et al (2009) Fibulin-6 expression and anoikis in human salivary gland epithelial cells: implications in Sjogren’s syndrome. Int Immunol 21(3):303–311
174
H. C. M. Cooney and W. M. Gallagher
Spencer JA, Hacker SL et al (2005) Altered vascular remodeling in fibulin-5-deficient mice reveals a role of fibulin-5 in smooth muscle cell proliferation and migration. Proc Natl Acad Sci U S A 102(8):2946–2951 Stone EM, Lotery AJ et al (1999) A single EFEMP1 mutation associated with both Malattia Leventinese and Doyne honeycomb retinal dystrophy. Nat Genet 22(2):199–202 Tarttelin EE, Gregory-Evans CY et al (2001) Molecular genetic heterogeneity in autosomal dominant drusen. J Med Genet 38(6):381–384 Timpl R, Sasaki T et al (2003) Fibulins: a versatile family of extracellular matrix proteins. Nat Rev Mol Cell Biol 4(6):479–489 Twal WO, Czirok A et al (2001) Fibulin-1 suppression of fibronectin-regulated cell adhesion and motility. J Cell Sci 114(Pt 24):4587–4598 Wlazlinski A, Engers R et al (2007) Downregulation of several fibulin genes in prostate cancer. Prostate 67(16):1770–1780 Wu YS, Bevilacqua VL et al (1995) Fibrillin domain folding and calcium binding: significance to Marfan syndrome. Chem Biol 2(2):91–97 Xie L, Palmsten K et al (2008) Basement membrane derived fibulin-1 and fibulin-5 function as angiogenesis inhibitors and suppress tumor growth. Exp Biol Med (Maywood) 233(2):155–162 Xu X, Dong C et al (2007) Hemicentins assemble on diverse epithelia in the mouse. J Histochem Cytochem 55(2):119–126
Chapter 10
Tumor Fibroblast-Associated Metalloproteases Julie Lecomte, Anne Masset, Dylan R. Edwards and Agnès Noël
10.1 Introduction Tumorigenesis and cancer progression rely on the acquisition by tumor cells of novel capacities through the accrual of mutations in genes critical for, at the very least, cell proliferation and survival (Vogelstein and Kinzler 2004). However, tumors are not isolated entities but rather depend on, interact with and react to the adjacent microenvironment. A tumor mass is not only composed of malignant cells but also includes several other cell types (fibroblastic cells, blood and lymphatic endothelial cells, inflammatory cells) which infiltrate the tumor and lead to the elaboration of a permissive stroma. In contrast to the initial view, genetic alterations accumulate not only in tumoral cells but also in stromal cells during cancer progression. The importance of the tumor microenvironment in cancer progression is now recognized (Joyce and Pollard 2009), but the critical molecular changes occurring in the tumor stroma accompanying and affecting cancer evolution remain largely unknown. New technological advances have been helpful recently in exploration of the stromal compartment. Indeed, microarray analysis of breast tumor stroma samples has given rise to some new stroma-derived prognostic predictor genes that can be used to identify sample clusters distinct from previously identified breast tumor subtypes (Beck et al. 2008; Finak et al. 2008). The stromal signature, alone or in combination with other molecular prognostic predictors, promises to improve molecular classification and outcome prediction in cancer, specifically by aiding the identification of patients who may benefit from aggressive therapies, or stratifying cancer subjects for clinical trials (van’t Veer et al. 2002; Santos et al. 2009). Desmoplasia, the stromal reaction associated with most carcinomas, is characterized by the local deposition of abundant extracellular matrix (ECM). Changes in the ECM may be of either a degradative or a productive nature varying as a function of time or localization. It is now recognized that the nature of the connective/stromal A. Noël () Laboratory of Tumor and Development Biology, Groupe Interdisciplinaire de Génoprotéomique Appliqué-Cancer (GIGA-Cancer), University of Liège, 4000 Liège, Belgium e-mail:
[email protected] M. M. Mueller, N. E. Fusenig (eds.), Tumor-Associated Fibroblasts and their Matrix, The Tumor Microenvironment 4, DOI 10.1007/978-94-007-0659-0_10, © Springer Science+Business Media B.V. 2011
175
176
J. Lecomte et al.
tissue is crucial, and that proteases that are involved in this tissue remodeling contribute to malignancy in early and late stages of primary and secondary tumor development (Mueller and Fusenig 2004). The cancer proteome relies not only on important changes in protein expression, but also on proteolytic post-translational modifications of the elaborated proteins, thus perturbing signaling pathways in both tumor and reactive stromal cells (Doucet et al. 2008; Overall et al. 2004). In this regard, the cancer degradome—the full repertoire of proteases implicated in cancerassociated tissue remodeling—is particularly important. The degradome includes five different families based on the nature of the chemical group responsible for the catalytic activity: serine proteases (plasmin/plasminogen activator), cysteine proteases (B-cathepsin, L-cathepsin), aspartic proteases (D-cathepsin), threonine proteases (proteasome components) and metalloproteinases (MMPs, ADAMs, ADAMTS) (Egeblad and Werb 2002; Lopez-Otin and Overall 2002; Overall et al. 2004). Some of these proteases exist as membrane-anchored forms on the cell surface or as soluble forms excreted into the extracellular surroundings. The relevance of the degradome as a molecular determinant of cancer progression has been recently underlined in studies based on gene expression signatures. Microarray analyses led to the classification of tumor sub-types according to their expression profiles associated with tumour metastasis or adverse outcome in several cancer types (van’t Veer et al. 2002; Santos et al. 2009; van de Vijver et al. 2002). Unsurprisingly, degradome genes are represented in most of these predictors, and the utility of several more proteases and related genes as biomarkers has been established by several profiling studies (Casey et al. 2009; Ma et al. 2009). In this review, we focus on metalloproteinases (MMPs and ADAMTSs) specifically expressed by tumor-associated fibroblasts which are important components of the tumor-host interplay (Kalluri and Zeisberg 2006). Stromal changes at the invasion front include the appearance of myofibroblasts, activated cells sharing characteristics with fibroblasts and smooth muscle cells. The transdifferentiation of fibroblasts into myofibroblasts is modulated by cancer cell-derived cytokines. These “activated” fibroblasts also called peritumoral or reactive fibroblasts, cancer-associated fibroblasts (CAF) and myofibroblasts (De Wever et al. 2008) are described in different chapters of the current issue. Despite being the predominant cell type within the tumor stroma, fibroblasts have received relatively little attention with regards to inflammationassociated tumorigenesis. The main complexity for studying fibroblast populations is the heterogeneity of fibroblastic cells within the tumor stroma (Sugimoto et al. 2006). Indeed, many fibroblast markers have been characterized but none is specific to all fibroblastic cell subsets (Kalluri and Zeisberg 2006). In addition, their cellular origins appear to be multiple, further increasing the complexity of the tumor stroma (Fig. 10.1). Resident fibroblasts surrounding cancer cells obviously contribute to the tumor stroma. However, another possible source of cancer-associated fibroblasts (CAFs) is their derivation from pericytes or vascular smooth muscle cells around vessels (Desmouliere et al. 2004). In the last few years, bone marrow-derived cells have been identified as cells that can be recruited into the tumor stroma and differentiate into myofibroblasts. It has also been suggested that CAFs are derived from malignant epithelial cells, or normal epithelial cells, undergoing epithelial–mesenchymal
10 Tumor Fibroblast-Associated Metalloproteases Cellular origin of CAF
177
CAF-associated MMPs and ADAM/TS
Protease function in tumor progression
Bone marrow
Tumor cell MSC
CAF
1
Cleavage of cell surface protein (E-cadherin, integrin, …)
SMC Pericytes
Resident fibroblast
Local connective tissue
MMP1 MMP2 MMP3 MMP9 MMP11 MMP13 MMP14 (MT1-MMP) ADAMTS 12
Release and/or generation of bioactive molecules
2 Degradation of ECM
3
inactive
active
active
inactive
EMT Others?
Cleavage of chemokines, cytokines or growth factors
Fig. 10.1 Multiple functions of stromal proteases produced by fibroblast activated cells ( CAF ). CAF can differentiate from a wide variety of cells such as resident fibroblasts ( local connective tissue), mesenchymal stem cells issued from bone marrow, vascular smooth muscle cells, pericytes, epithelial cells performing epithelial-to-mesenchymal transition. These activated fibroblasts constitute an important source of MMPs and likely ADAMTSs. These enzymes are endowed with multiple functions contributing thereby to cancer progression. These proteases can cleave cell surface proteins, thus controling tumor cell adhesion. By degrading extracellular matrix, they can not only favor cell migration, but also release growth factors from the matrix and generate bioactive molecules from fragments of matrix components. In addition, proteases can participate in a complex network of molecular messages by activating or inactivating chemokines, cytokines and growth factors
transition (Kalluri and Zeisberg 2006; Ostman and Augsten 2009). Endothelial cells were also recently identified as a candidate source of CAFs (Zeisberg et al. 2007). The present review aims at describing several MMPs and ADAMTSs produced by fibroblastic cells that have acquired a modified phenotype within the tumour stroma.
10.2 Metalloproteinases: MMPs and ADAMTSs MMPs are a family of 23 human zinc-binding endopeptidases that can degrade virtually all ECM components, and release and activate/inactivate a growing number of modulators of cell functions (Cauwe et al. 2007; Egeblad and Werb 2002; Overall and Dean 2006). MMPs are multidomain proteins characterized by at least
178
J. Lecomte et al.
three conserved regions: (1) a zinc-binding motif (HEXXHXXGXXH) required for proteolytic activity, (2) a propeptide cysteine site (PRCGXPD) whose cysteine residue interacts with the zinc ion in the zymogen form and (3) a ‘methionine turn’ (XXMXP) which likely maintains the zinc-binding site integrity (Egeblad and Werb 2002). Most of the MMPs are secreted as soluble enzymes but six of them are membrane-type MMPs (MT-MMPs) which associate with the cell membrane by either a COOH-terminal transmembrane domain (MT1-, MT2-, MT3-, MT5-MMPs) or a glycosylphosphatidylinositol (GPI) anchor (MT4-, MT6-MMPs) (Zucker et al. 2003). Proteolytic activity is inhibited by tissue inhibitors known as tissue inhibitors of metalloproteinase (TIMPs). ADAMTSs consist of 19 human MMP-related enzymes characterized by the presence of a disintegrin-like domain and at least one thrombospondin type I repeat (TSP-1) (Porter et al. 2005; Colige et al. 2005; Rocks et al. 2008; Roy et al. 2006). The thrombospondin motifs at the carboxy terminus of these secreted enzymes, together in some cases with an internal spacer region, facilitate their localization in the ECM in close proximity to their cognate substrates. Their multi-domain structure endows these proteins with various functions including the control of cell properties such as cell proliferation, apoptosis, adhesion and migration (Rocks et al. 2008). In contrast to their universal MMP-inhibitory properties, TIMPs appear to be more selective in inhibiting ADAMTSs. It is worth noting that TIMP3 which is a potent inhibitor of some ADAMs and ADAMTSs is expressed by stromal cells as assessed by laser-capture microdissection and in situ hybridization (Shukla et al. 2008).
10.3 Metalloproteinases as Key Molecular Determinants of Tumor-Associated Fibroblasts MMPs were initially claimed to be important in late stages of tumor progression by degrading connective tissue stroma and basement membrane (Egeblad and Werb 2002). However, due to the rapid development of innovative biochemical techniques (Greenlee et al. 2006; Lopez-Otin and Overall 2002; Overall et al. 2004) and the expanding use of transgenic mice (Cauwe et al. 2007; Page-McCaw et al. 2007), it became obvious that the action of MMPs is not restricted to the massive destruction of physiological matrix barriers (Overall and Dean 2006). MMPs are viewed as key regulators of the multiple cellular functions which dictate malignant growth. Although some MMPs are produced by tumor cells (e.g. MMP7, MT4-MMP) (Chabottaux et al. 2006; Lynch et al. 2007), most MMPs are rather produced by host cells and therefore might be considered as molecular determinants of the ‘seed and soil’ concept proposed by Paget in 1889 (Fidler 2003). Fibroblasts and myofibroblasts are principal producers of several types of MMPs, which highlight their crucial role in ECM remodeling (Kalluri and Zeisberg 2006). One prevailing view is that tumor-associated stroma is activated by the malignant epithelial cells to foster tumor growth—for example, by secreting growth factors, increasing angiogenesis, and
10 Tumor Fibroblast-Associated Metalloproteases
179
facilitating cell migration, ultimately resulting in metastasis to remote organ sites (Liotta and Kohn 2001; Mueller and Fusenig 2004). The importance of fibroblastic cells was initially demonstrated through in vivo experiments in which fibroblasts co-injected with tumor cells were shown to promote primary cancer growth (Noel et al. 1993, 1996; Picard et al. 1986). In these systems, the contribution of MMPs is demonstrated by the inhibition of this tumor promoting effect of fibroblasts by using synthetic MMP inhibitors (Maquoi et al. 2004). More recently, mesenchymal stem cells co-injected with tumor cells have been shown to promote breast metastasis (Karnoub et al. 2007). Cancer cells might stimulate fibroblasts to synthesize MMPs in a paracrine manner through the secretion of interleukins, interferons, growth factors and EMMPRIN (Heppner et al. 1996; Noel et al. 1994; Sternlicht and Werb 2001; Zigrino et al. 2005). Fibroblasts constitute therefore an important source of MMPs including mainly MMP1 (Ala-aho and Kahari 2005), MMP2 (Bisson et al. 2003), MMP3 (Sternlicht et al. 1999), MMP9, MMP11 (Basset et al. 1990; Rio 2005), MMP13 (Ala-aho and Kahari 2005) and MT1-MMP (MMP14) (Sounni and Noel 2005; Zucker et al. 2003). Of interest are the recent findings of several genes differentially expressed in invasive stroma compared to in situ stroma by combining laser capture microdissection and gene expression microarray (Ma et al. 2009). Among these genes, MMP11, MMP2 and MMP14 showed significant increases in invasive stroma. Other studies demonstrated that MMP13 is expressed by fibroblasts surrounding epithelial tumors and is associated with invasive and metastatic tumors (Pendas et al. 2000). Due to space constraints, we have decided to focus our interest on stromal MMPs identified by the recent microarray analyses performed on human samples: MMP2, MMP11, MMP13 and MMP14. In addition, we will consider members of the ADAMTS family that also appear to be involved in the stromal response to tumors, as is the case for ADAMTS1 and ADAMTS12.
10.3.1 Secreted Stromal MMPs MMP11 or Stromelysin-3 (ST3) is mainly expressed by cells of mesenchymal origin in association with remodeling processes occurring during embryogenesis (Lefebvre et al. 1995; Maquoi et al. 1997) and tissue development (Lefebvre et al. 1992). MMP11 is involved in epithelial homeostasis (Rio et al. 1996) but also in various non-cancerous pathological conditions such as repair processes (Okada et al. 1997; Wolf et al. 1992), human atherosclerosis (Schonbeck et al. 1999) and rheumatoid arthritis (Konttinen et al. 1999; Lubberts et al. 1999). In cancer, MMP11 is associated with tumor invasion and poor prognosis. Its expression is restricted to the stromal fibroblasts adjacent to cancer cells (Basset et al. 1990; Rio 2005), thus suggesting that MMP11 production by stromal cells in tumor tissues could correspond to an individual response by host stromal cells to stimulatory messages derived from cancer cells (Heppner et al. 1996). Indeed, breast cancer cell lines such as MCF7, MDA-MB-231 and ZR75 were found to induce MMP11 expression by human invasion front fibroblasts (Ahmad et al. 1997).
180
J. Lecomte et al.
Clinical data indicate that MMP11 promotes aggressive behavior in tumors of various origins. Increased MMP11 gene expression is correlated with the local invasiveness of cancer cells in head and neck squamous cell carcinoma (Muller et al. 1993). A high level of MMP11 is a strong independent prognostic indicator of shorter disease-free survival in breast cancers (Ahmad et al. 1998; Chenard et al. 1996) and a marker of aggressive clinical outcome in oesophageal cancer (Porte et al. 1998; Yamashita et al. 2004). Therefore, MMP11 expression is a useful diagnostic that is of clinical value and this enzyme has therefore been proposed as a potential target for new treatment strategies (Anderson et al. 1995). Recently, MMP11 has been identified as a novel broadly expressed tumor-associated antigen and a candidate target for cancer immunotherapy (Peruzzi et al. 2009). In experimental models, MMP11 appears as a stromal factor that promotes the primary implantation of cancer cells in a tissue environment initially not permissive for tumor growth (Noel et al. 1996; Rio 2005). In Mmp11-deficient mice, the number of apoptotic cancer cells is increased in primary tumors indicating that host MMP11 helps cancer cells in escaping apoptosis (Boulay et al. 2001). The contribution of fibroblast-derived MMP11 is supported by the lack of tumor promoting effect of fibroblasts generated from Mmp11-deficient mice as compared to their wild type counterpart (Masson et al. 1998). Surprisingly, no tumor phenotype was observed in transgenic mice ectopically expressing MMP11 in the mammary gland epithelial cells (WAP-MMP11), in fibroblasts (vimentin-MMP11), or ubiquitously in all cells (CMV-MMP11) (Noel et al. 1996; Rio et al. 1996; Rio 2005). The role of MMP11 in oncology is complex and unclear. In MMTV-ras transgenic mice, Mmp11-deficiency is associated with more metastases for a similar number and size of primary invasive tumors, indicating that cancer cells evolving in Mmp11-deficient stroma have increased potential for hematogenous dissemination (Andarawewa et al. 2003). Thus MMP11 appears to repress metastatic dissemination while it enhances primary tumor development. These paradoxical functions require further investigation and support the emerging roles of MMPs in tumour repression (Lopez-Otin and Matrisian 2007). An interesting insight into the potential role of MMP11 in the pathogenesis of primary mammary tumors is provided by the recent observation that MMP11 promotes differentiation of adipocytes into fibroblasts at the invasion front, contributing to desmoplasia and creating a more permissive environment for tumor growth (Andarawewa et al. 2003). MMP13 (collagenase-3) originally identified in human breast cancer tissue (Freije et al. 1994) is secreted from cells as an inactive zymogen that can be activated by the MMP2/MMP14/TIMP2 complex (Knauper et al. 1996) or by plasmin (Cowell et al. 1998). The overexpression of MMP13 in several types of malignancy (Balbin et al. 1999; Bartsch et al. 2003; Bostrom et al. 2000; Corte et al. 2005; Curran et al. 2004; Dunne et al. 2003; Leeman et al. 2002; Luukkaa et al. 2006; Pendas et al. 2000) is associated with shorter overall survival of the patients (Curran et al. 2004; Leeman et al. 2002; Luukkaa et al. 2006). It has been described as a potential new tumor marker for breast cancer diagnosis as it is overexpressed in breast cancer tissues compared to normal adjacent tissues (Chang et al. 2009). MMP13 expres-
10 Tumor Fibroblast-Associated Metalloproteases
181
sion was also correlated with metastasis formation (Ellsworth et al. 2009; Lee et al. 2009). Consistently, functional studies have demonstrated that MMP13 increases the invasive capacities of malignant cells (Airola et al. 1999; Balduyck et al. 2000; Culhaci et al. 2004; Daja et al. 2003; Johansson et al. 1999). MMP13 acts in the extracellular environment as a potent collagenase capable of degrading a variety of collagen types. While several studies concluded that MMP13 is synthesized predominantly by tumor cells (Balduyck et al. 2000), others claimed that MMP13 is expressed in a subpopulation of stromal myofibroblasts (Kleiner and Stetler-Stevenson 1999; Nielsen et al. 2001, 2007) in invasive breast carcinoma. Interestingly, the presence of microinvasion in ductal carcinoma in situ (DCIS) is associated with focal expression of MMP13 mRNA in stromal fibroblasts (Nielsen et al. 2001, 2007). Direct comparison of the MMP13 mRNA expression pattern with that of the MMP2, MMP11 and MMP14 mRNAs indicates that MMP13 is unique in this respect since the other MMPs are also present in DCIS in the absence of invasion (Nielsen et al. 2001). These observations raise the question as to whether MMP13 is a rate-limiting proteinase that mediates the initial steps in breast cancer invasion. The recent generation of Mmp13-deficient mice will be helpful in unravelling the function of host MMP13 during cancer progression (Inada et al. 2004; Nielsen et al. 2008). However, in the aggressive mouse mammary tumor virus-polyoma middle T-antigen (MMTV-PyMT) model of breast cancer, the absence of MMP13 did not influence tumor growth, vascularization, or metastasis to the lungs, suggesting that MMP13’s role in breast cancer may depend on the nature of the genetic lesions driving malignancy (Nielsen et al. 2008).
10.3.2 Membrane-Associated MMPs According to the structure of their C-terminal extension, MT-MMPs can be classified into two sub-groups: (1) type I transmembrane proteins including MT1-, MT2-, MT3- and MT5-MMP characterized by a long hydrophobic sequence followed by a short cytoplasmic tail and (2) glycosylphosphatidylinositol (GPI)-type MTMMPs (MT4- and MT6-MMP) containing a short hydrophobic signal anchoring to GPI (Zhao et al. 2008). The GPI-anchored MT4-MMP is exclusively expressed by breast tumor cells and contributes to tumor angiogenesis and metastatic dissemination (Chabottaux et al. 2006, 2009). In contrast, the cellular distribution of MT1-MMP is much more complex and depends on the cancer type considered. It was initially reported to be expressed in lung carcinoma cells and in the adjacent fibroblasts (Sato et al. 1994). Such patterns of MT1-MMP expression and MMP2 activation were also observed in many types of tumors, such as lung (Nawrocki et al. 1997; Polette et al. 1996; Sato et al. 1994; Tokuraku et al. 1995), gastric (Bando et al. 1998; Mori et al. 1997; Nomura et al. 1995), colon (Ohtani et al. 1996), breast (Ishigaki et al. 1999; Polette et al. 1996; Ueno et al. 1997), bladder (Kanayama et al. 1998), head and neck (Yoshizaki et al. 1997), thyroid (Nakamura et al. 1999), ovar-
182
J. Lecomte et al.
ian (Afzal et al. 1998; Fishman et al. 1996) and cervical carcinomas (Gilles et al. 1996), and brain tumors (Forsyth et al. 1999; Nakada et al. 1999; Yamamoto et al. 1996). Transcripts were also detected in both tumor cells and surrounding stroma cells (Afzal et al. 1998; Heppner et al. 1996; Ohtani et al. 1996; Polette et al. 1996). It is particularly noteworthy that co-cultivation of fibroblasts with tumor cells induced the up-regulation of MMP2 and MT1-MMP expression by fibroblasts (Noel et al. 1994; Polette et al. 1997). MT1-MMP has been first described as a specific activator of pro-MMP2, with TIMP2 acting as an adaptor molecule which mediates pro-MMP2 binding to MT1MMP (Strongin et al. 1995). MMP2 is frequently co-expressed with MT1-MMP in mesenchymal cells (Apte et al. 1997; Kinoh et al. 1996). In breast cancer, MT1MMP has been demonstrated to be confined to αSMA positive myofibroblasts in close contact to tumor cells whereas MMP2 is produced by different fibroblastic populations (Bisson et al. 2003) (Fig. 10.2). Carcinoma cells (Stetler-Stevenson et al. 1993; Tryggvason et al. 1993) in tissue and cancer cell lines (Sato et al. 1992) rarely express MMP2, even though they can use MMP2 that is derived from the surrounding fibroblasts by binding and activating it using MT1-MMP on the cell surface. Pro-MMP13 also can be activated by MT1-MMP in a cell-mediated manner (Knauper et al. 1996). This MT1-MMP/MMP2/MMP13 system could be important for tumor cells to invade the basement membrane by degrading type IV collagen and subsequently the stroma by degrading interstitial collagens. Activated MT1-MMP is a potent membrane proteinase with the ability to cleave type I collagen more efficiently than type II or III collagens (Sounni and Noel 2005). Consistently, MT1-Mmp knockout mice are characterized by a defect in cartilage remodeling and bone development (Holmbeck et al. 1999, 2003). MT1-MMP is currently recognized as a key regulator of connective tissue remodeling (Sabeh et al. 2009). Besides its capacity to activate pro-MMP2 and pro-MMP13, there is clear evidence that MT1-MMP and MMP2 are involved at different stages of tumor progression from initial tumor development, growth and angiogenesis to invasion, metastasis and growth at secondary sites (Deryugina et al. 2002; Sounni et al. 2002, 2004). MT1-MMP overexpression strongly promotes cellular invasion in vitro and experimental metastasis. The key role of fibroblast-derived MT1-MMP has been evidenced in co-transplantation of tumor cells and fibroblasts. In contrast to MT1MMP expressing fibroblasts, MT1-Mmp-null fibroblasts were unable to support in vivo growth of tumor cells (Zhang et al. 2006). However, co-operation between tumor-derived MT1-MMP and stromal MMP-2 in vivo has also been clearly shown using Mmp2-knockout mice and tumorigenic cell lines engineered to over-express MT1-MMP (Taniwaki et al. 2007). MT1-MMP is a multifunctional protein that can also cleave several soluble, cell surface and pericellular proteins, regulating cell behavior in tumor metastasis and angiogenesis (Sounni and Noel 2005). These regulation mechanisms include (1) the alteration of cell–cell interactions, cell–matrix interactions, (2) the release, activation or inactivation of autocrine, or paracrine signaling molecules, (3) the shedding or activation of surface receptors and (4) the activation of intracellular signaling pathways.
10 Tumor Fibroblast-Associated Metalloproteases
αSMA
183
MT1-MMP
T
MMP2
T
T
a
b
c
d
e
f
Fig. 10.2 Stromal localisation of MT1-MMP and MMP2 in human breast carcinomas. Immunostaining for αSMA (a) and in situ hybridization of MT1-MMP (b) and MMP2 (c) were performed on serial sections of human breast cancer. MT1-Mmp mRNA positive cells (b) are at the proximity of tumor cells, whereas Mmp2 mRNAs (c) and αSMA (a) are more widely distributed. Binary image of αSMA distribution in blue (d) was superimposed to that of MT1-MMP ( in green) (e) or MMP2 ( in green) (f). Colocalization appears in red in (e, f). T, tumor cell islet. Different sub-population of fibroblastic cells express MT1-MMP and MMP2
10.3.3 ADAMTSs The cancer-related functions of the AdamTS gene family are much less well understood than those of the MMPs. Indeed our knowledge of the basic enzymatic activities of these intriguing molecules is still in its infancy. One subgroup (ADAMTS1, 4, 5, 8, 9, 15 and 20) have been termed aggrecanases or hyalectanases based on their ability to cleave matrix proteoglycans; functions are also recognized for the pro-collagen N-proteinases (ADAMTS2, 3 and 14) constitute another group and
184
J. Lecomte et al.
the von Willebrand Factor-cleaving proteinase (ADAMTS13) (Porter et al. 2005; Rocks et al. 2006). However the activities of the other 8 ADAMTSs are unknown. What is clear is that expression of this gene family is altered dramatically during tumourigenesis and progression (Porter et al. 2004; Rocks et al. 2006). Also, several ADAMTS genes are primarily expressed by stromal fibroblasts in non-neoplastic tissues suggesting important roles in homeostasis (Porter et al. 2004). A further complexity is highlighted by ADAMTS1, which shows both pro- and anti-tumorigenic/metastatic activities depending on whether the enzyme is in full-length or cleaved forms (Liu et al. 2006).
10.4 Stromal MMPs and ADAMTS as Anti-Tumor Regulators Recently, the generation of animal models involving gain or loss of function of specific matrix metalloproteinases (MMPs) has led to the surprising discovery of tumour-suppressive function for some proteases (Lopez-Otin and Matrisian 2007; Lopez-Otin and Bond 2008). These host protective proteases are not produced by tumor cells themselves, but mainly by tumour infiltrating cells including inflammatory cells (MMP8) (Balbin et al. 2003; Gutierrez-Fernandez et al. 2007) and fibroblastic cells (MMP19) (Jost et al. 2006). An increasing body of evidence indicates that members of the ADAMTS family exhibit tumor inhibitory activities. Among them, ADAMTS1 and ADAMTS8 display anti-angiogenic properties (de Fraipont et al. 2001; Iruela-Arispe et al. 2003; Lee et al. 2006; Vazquez et al. 1999). In the case of ADAMTS1, anti-angiogenesis may involve multiple activities, including sequestration of vascular endothelial growth factor (Luque et al. 2003) and cleavage/activation of thrombospondins-1 and -2 (Lee et al. 2006). AdamTS1, AdamTS9, AdamTS15 and AdamTS18 genes have been found epigenetically silenced in several carcinomas (Jin et al. 2007; Lind et al. 2006; Lo et al. 2007; Viloria et al. 2009). An intriguing opposite regulation of ADAMTS12 expression in colon carcinomas has been recently described (Moncada-Pazos et al. 2009). While the gene is epigenetically inactivated in tumor cells, its expression is transcriptionally induced in surrounding fibroblastic cells. Further studies are required to clarify the putative protective role of this and other ADAMTS family members produced in the stromal compartment. Altogether, these recent findings have broken the dogma of proteases as simple positive regulators of cancer progression and emphasize the urgent need for identifying individual proteases as host protective partners or tumor-promoting agents.
10.5 Conclusion and Perspectives The key conceptual advance in the metalloproteinase field in the past decade has been the recognition that these enzymes are much more than matrix-degrading “scissors” that clear paths through tissue stroma for invading cancer cells. Some
10 Tumor Fibroblast-Associated Metalloproteases
185
metalloproteinases do indeed carry out this function, but many have a more significant role as regulators of the pericellular signaling environment, and this latter capacity is arguably their chief relevance in cancer biology. MMPs and ADAMTSs produced by CAFs and other stromal cells will thus shape the tumor microenvironment. These genes are proving to have utility as diagnostic, prognostic or predictive cancer biomarkers. However our knowledge of the precise roles of particular proteases in different types of cancers, and at different disease stages, is still very incomplete. There is thus a need for continuation of functional studies on the cancer degradome. Moreover, the use of mass spectrometry technologies for identification of critical protease substrates is helping to unlock biological mechanisms, which in turn will lead to new strategies for cancer therapy. Acknowledgments This work was supported by grants from Ministerio European projects (FP7 HEALTH-F2-2008-201279 “MICROENVIMET”), the Fondation contre le Cancer, the D.G.T.R.E. from the « Région Wallonne », the Interuniversity Attraction Poles Programme—Belgian Science Policy (Belgium). DE is grateful to support from the Breast Cancer Campaign, Cancer ResearchUK and the Big C Appeal.
Reference Afzal S, Lalani EN, Poulsom R, Stusbbs A, Rowlinson G, Sato H, Seiki M, Stamp GW (1998) MT1-MMP and MMP-2 mRNA expression in human ovarian tumors: possible implications for the role of desmoplastic fibroblasts. Hum Pathol 29:155–165 Ahmad A, Marshall JF, Basset P, Anglard P, Hart IR (1997) Modulation of human stromelysin 3 promoter activity and gene expression by human breast cancer cells. Int J Cancer 73:290–296 Ahmad A, Hanby A, Dublin E, Poulsom R, Smith P, Barnes D, Rubens R, Anglard P, Hart I (1998) Stromelysin 3: an independent prognostic factor for relapse-free survival in node-positive breast cancer and demonstration of novel breast carcinoma cell expression. Am J Pathol 152:721–728 Airola K, Karonen T, Vaalamo M, Lehti K, Lohi J, Kariniemi AL, Keski-Oja J, Saarialho-Kere UK (1999) Expression of collagenases-1 and -3 and their inhibitors TIMP-1 and -3 correlates with the level of invasion in malignant melanomas. Br J Cancer 80:733–743 Ala-aho R, Kahari VM (2005) Collagenases in cancer. Biochimie 87:273–286 Andarawewa KL, Boulay A, Masson R, Mathelin C, Stoll I, Tomasetto C, Chenard MP, Gintz M, Bellocq JP, Rio MC (2003) Dual stromelysin-3 function during natural mouse mammary tumor virus-ras tumor progression. Cancer Res 63(18):5844–5849 Anderson IC, Sugarbaker DJ, Ganju RK, Tsarwhas DG, Richards WG, Sunday M, Kobzik L, Shipp MA (1995) Stromelysin-3 is overexpressed by stromal elements in primary non-small cell lung cancers and regulated by retinoic acid in pulmonary fibroblasts. Cancer Res 55:4120–4126 Apte SS, Fukai N, Beier DR, Olsen BR (1997) The matrix metalloproteinase-14 (MMP-14) gene is structurally distinct from other MMP genes and is co-expressed with the TIMP-2 gene during mouse embryogenesis. J Biol Chem 272:25511–25517 Balbin M, Pendas AM, Uria JA, Jimenez MG, Freije JP, Lopez-Otin C (1999) Expression and regulation of collagenase-3 (MMP-13) in human malignant tumors. APMIS 107:45–53 Balbin M, Fueyo A, Tester AM, Pendas AM, Pitiot AS, Astudillo A, Overall CM, Shapiro SD, Lopez-Otin C (2003) Loss of collagenase-2 confers increased skin tumor susceptibility to male mice. Nat Genet 35:252–257 Balduyck M, Zerimech F, Gouyer V, Lemaire R, Hemon B, Grard G, Thiebaut C, Lemaire V, Dacquembronne E, Duhem T, Lebrun A, Dejonghe MJ, Huet G (2000) Specific expression of matrix metalloproteinases 1, 3, 9 and 13 associated with invasiveness of breast cancer cells in vitro. Clin Exp Metastasis 18:171–178
186
J. Lecomte et al.
Bando E, Yonemura Y, Endou Y, Sasaki T, Taniguchi K, Fujita H, Fushida S, Fujimura T, Nishimura G, Miwa K, Seiki M (1998) Immunohistochemical study of MT-MMP tissue status in gastric carcinoma and correlation with survival analyzed by univariate and multivariate analysis. Oncol Rep 5:1483–1488 Bartsch JE, Staren ED, Appert HE (2003) Matrix metalloproteinase expression in breast cancer. J Surg Res 110:383–392 Basset P, Bellocq JP, Wolf C, Stoll I, Hutin P, Limacher JM, Podhajcer OL, Chenard MP, Rio MC, Chambon P (1990) A novel metalloproteinase gene specifically expressed in stromal cells of breast carcinomas. Nature 348:699–704 Beck AH, Espinosa I, Gilks CB, van de Rijn M, West RB (2008) The fibromatosis signature defines a robust stromal response in breast carcinoma. Lab Invest 88:591–601 Bisson C, Blacher S, Polette M, Blanc JF, Kebers F, Desreux J, Tetu B, Rosenbaum J, Foidart JM, Birembaut P, Noel A (2003) Restricted expression of membrane type 1-matrix metalloproteinase by myofibroblasts adjacent to human breast cancer cells. Int J Cancer 105:7–13 Bostrom PJ, Ravanti L, Reunanen N, Aaltonen V, Soderstrom KO, Kahari VM, Laato M (2000) Expression of collagenase-3 (matrix metalloproteinase-13) in transitional-cell carcinoma of the urinary bladder. Int J Cancer 88:417–423 Boulay A, Masson R, Chenard MP, Fahime MEl, Cassard L, Bellocq JP, Sautes-Fridman C, Basset P, Rio MC (2001) High cancer cell death in syngeneic tumors developed in host mice deficient for the stromelysin-3 matrix metalloproteinase. Cancer Res 61:2189–2193 Casey T, Bond J, Tighe S, Hunter T, Lintault L, Patel O, Eneman J, Crocker A, White J, Tessitore J, Stanley M, Harlow S, Weaver D, Muss H, Plaut K (2009) Molecular signatures suggest a major role for stromal cells in development of invasive breast cancer. Breast Cancer Res Treat 114:47–62 Cauwe B, Van den Steen PE, Opdenakker G (2007) The biochemical, biological, and pathological kaleidoscope of cell surface substrates processed by matrix metalloproteinases. Crit Rev Biochem Mol Biol 42:113–185 Chabottaux V, Sounni NE, Pennington CJ, English WR, van den Brûle F, Blacher S, Gilles C, Munaut C, Maquoi E, Lopez-Otin C, Murphy G, Edwards DR, Foidart JM, Noel A (2006) Membrane-type 4 matrix metalloproteinase promotes breast cancer growth and metastases. Cancer Res 66:5165–5172 Chabottaux V, Ricaud S, Host L, Blacher S, Paye A, Thiry M, Garofalakis A, Pestourie C, Gombert K, Bruyere F, Lewandowsky D, Tavitian B, Foidart JM, Duconge F, Noel A (2009) MembraneType 4 Matrix Metalloproteinase (MT4-MMP) induces lung metastasis by alteration of primary breast tumor vascular architecture. J Cell Mol Med 13:4002–4013 Chang HJ, Yang MJ, Yang YH, Hou MF, Hsueh EJ, Lin SR (2009) MMP13 is potentially a new tumor marker for breast cancer diagnosis. Oncol Rep 22:1119–1127 Chenard MP, O’Siorain L, Shering S, Rouyer N, Lutz Y, Wolf C, Basset P, Bellocq JP, Duffy MJ (1996) High levels of stromelysin-3 correlate with poor prognosis in patients with breast carcinoma. Int J Cancer 69:448–451 Colige A, Ruggiero F, Vandenberghe I, Dubail J, Kesteloot F, Van Beeumen J, Beschin A, Brys L, Lapiere CM, Nusgens B (2005) Domains and maturation processes that regulate the activity of ADAMTS-2, a metalloproteinase cleaving the aminopropeptide of fibrillar procollagens types I-III and V. J Biol Chem 280:34397–34408 Corte MD, Gonzalez LO, Corte MG, Quintela I, Pidal I, Bongera M, Vizoso F (2005) Collagenase-3 (MMP-13) expression in cutaneous malignant melanoma. Int J Biol Markers 20:242– 248 Cowell S, Knauper V, Stewart ML, D’Ortho MP, Stanton H, Hembry RM, Lopez-Otin C, Reynolds JJ, Murphy G (1998) Induction of matrix metalloproteinase activation cascades based on membrane-type 1 matrix metalloproteinase: associated activation of gelatinase A, gelatinase B and collagenase 3. Biochem J 331(Pt 2):453–458 Culhaci N, Metin K, Copcu E, Dikicioglu E (2004) Elevated expression of MMP-13 and TIMP-1 in head and neck squamous cell carcinomas may reflect increased tumor invasiveness. BMC Cancer 4:42
10 Tumor Fibroblast-Associated Metalloproteases
187
Curran S, Dundas SR, Buxton J, Leeman MF, Ramsay R, Murray GI (2004) Matrix metalloproteinase/tissue inhibitors of matrix metalloproteinase phenotype identifies poor prognosis colorectal cancers. Clin Cancer Res 10:8229–8234 Daja MM, Niu X, Zhao Z, Brown JM, Russell PJ (2003) Characterization of expression of matrix metalloproteinases and tissue inhibitors of metalloproteinases in prostate cancer cell lines. Prostate Cancer Prostatic Dis 6:15–26 de Fraipont F, Nicholson AC, Feige JJ, Van Meir EG (2001) Thrombospondins and tumor angiogenesis. Trends Mol Med 7:401–407 De Wever O, Demetter P, Mareel M, Bracke M (2008) Stromal myofibroblasts are drivers of invasive cancer growth. Int J Cancer 123:2229–2238 Deryugina EI, Ratnikov BI, Postnova TI, Rozanov DV, Strongin AY (2002) Processing of integrin alpha(v) subunit by membrane type 1 matrix metalloproteinase stimulates migration of breast carcinoma cells on vitronectin and enhances tyrosine phosphorylation of focal adhesion kinase. J Biol Chem 277:9749–9756 Desmouliere A, Guyot C, Gabbiani G (2004) The stroma reaction myofibroblast: a key player in the control of tumor cell behavior. Int J Dev Biol 48:509–517 Doucet A, Butler GS, Rodriguez D, Prudova A, Overall CM (2008) Metadegradomics: toward in vivo quantitative degradomics of proteolytic post-translational modifications of the cancer proteome. Mol Cell Proteomics 7:1925–1951 Dunne AA, Sesterhenn A, Gerisch A, Teymoortash A, Kuropkat C, Werner JA (2003) Expression of MMP-2, -9 and -13 in cell lines and fresh biopsies of squamous cell carcinomas of the upper aerodigestive tract. Anticancer Res 23:2233–2239 Egeblad M, Werb Z (2002) New functions for the matrix metalloproteinases in cancer progression. Nat Rev Cancer 2:161–174 Ellsworth R, Seebach J, Field L, Heckman C, Kane J, Hooke J, Love B, Shriver C (2009) A gene expression signature that defines breast cancer metastases. Clin Exp Metastasis 26:205–213 Fidler IJ (2003) The pathogenesis of cancer metastasis: the ‘seed and soil’ hypothesis revisited. Nat Rev Cancer 3:453–458 Finak G, Bertos N, Pepin F, Sadekova S, Souleimanova M, Zhao H, Chen H, Omeroglu G, Meterissian S, Omeroglu A, Hallett M, Park M (2008) Stromal gene expression predicts clinical outcome in breast cancer. Nat Med 14:518–527 Fishman DA, Bafetti LM, Stack MS (1996) Membrane-type matrix metalloproteinase expression and matrix metalloproteinase-2 activation in primary human ovarian epithelial carcinoma cells. Invasion Metastasis 16:150–159 Forsyth PA, Wong H, Laing TD, Rewcastle NB, Morris DG, Muzik H, Leco KJ, Johnston RN, Brasher PM, Sutherland G, Edwards DR (1999) Gelatinase-A (MMP-2), gelatinase-B (MMP9) and membrane type matrix metalloproteinase-1 (MT1-MMP) are involved in different aspects of the pathophysiology of malignant gliomas. Br J Cancer 79:1828–1835 Freije JM, Diez-Itza I, Balbin M, Sanchez LM, Blasco R, Tolivia J, Lopez-Otin C (1994) Molecular cloning and expression of collagenase-3, a novel human matrix metalloproteinase produced by breast carcinomas. J Biol Chem 269:16766–16773 Gilles C, Polette M, Piette J, Munaut C, Thompson EW, Birembaut P, Foidart JM (1996) High level of MT-MMP expression is associated with invasiveness of cervical cancer cells. Int J Cancer 65:209–213 Greenlee KJ, Corry DB, Engler DA, Matsunami RK, Tessier P, Cook RG, Werb Z, Kheradmand F (2006) Proteomic identification of in vivo substrates for matrix metalloproteinases 2 and 9 reveals a mechanism for resolution of inflammation. J Immunol 177:7312–7321 Gutierrez-Fernandez A, Inada M, Balbin M, Fueyo A, Pitiot AS, Astudillo A, Hirose K, Hirata M, Shapiro SD, Noel A, Werb Z, Krane SM, Lopez-Otin C, Puente XS (2007) Increased inflammation delays wound healing in mice deficient in collagenase-2 (MMP-8). FASEB J 21:2580–2591 Heppner KJ, Matrisian LM, Jensen RA, Rodgers WH (1996) Expression of most matrix metalloproteinase family members in breast cancer represents a tumor-induced host response. Am J Pathol 149:273–282
188
J. Lecomte et al.
Holmbeck K, Bianco P, Caterina J, Yamada S, Kromer M, Kuznetsov SA, Mankani M, Robey PG, Poole AR, Pidoux I, Ward JM, Birkedal-Hansen H (1999) MT1-MMP-deficient mice develop dwarfism, osteopenia, arthritis, and connective tissue disease due to inadequate collagen turnover. Cell 99:81–92 Holmbeck K, Bianco P, Chrysovergis K, Yamada S, Birkedal-Hansen H (2003) MT1-MMPdependent, apoptotic remodeling of unmineralized cartilage: a critical process in skeletal growth. J Cell Biol 163:661–671 Inada M, Wang Y, Byrne MH, Rahman MU, Miyaura C, Lopez-Otin C, Krane SM (2004) Critical roles for collagenase-3 (Mmp13) in development of growth plate cartilage and in endochondral ossification. Proc Natl Acad Sci U S A 101:17192–17197 Iruela-Arispe ML, Carpizo D, Luque A (2003) ADAMTS1: a matrix metalloprotease with angioinhibitory properties. Ann N Y Acad Sci 995:183–190 Ishigaki S, Toi M, Ueno T, Matsumoto H, Muta M, Koike M, Seiki M (1999) Significance of membrane type 1 matrix metalloproteinase expression in breast cancer. Jpn J Cancer Res 90:516–522 Jin H, Wang X, Ying J, Wong AH, Li H, Lee KY, Srivastava G, Chan AT, Yeo W, Ma BB, Putti TC, Lung ML, Shen ZY, Xu LY, Langford C, Tao Q (2007) Epigenetic identification of ADAMTS18 as a novel 16q23.1 tumor suppressor frequently silenced in esophageal, nasopharyngeal and multiple other carcinomas. Oncogene 26:7490–7498 Johansson N, Vaalamo M, Grenman S, Hietanen S, Klemi P, Saarialho-Kere U, Kahari VM (1999) Collagenase-3 (MMP-13) is expressed by tumor cells in invasive vulvar squamous cell carcinomas. Am J Pathol 154:469–480 Jost M, Folgueras AR, Frerart F, Pendas AM, Blacher S, Houard X, Berndt S, Munaut C, Cataldo D, Alvarez J, Melen-Lamalle L, Foidart JM, Lopez-Otin C, Noel A (2006) Earlier onset of tumoral angiogenesis in matrix metalloproteinase-19-deficient mice. Cancer Res 66:5234–5241 Joyce JA, Pollard JW (2009) Microenvironmental regulation of metastasis. Nat Rev Cancer 9:239–252 Kalluri R, Zeisberg M (2006) Fibroblasts in cancer. Nat Rev Cancer 6:392–401 Kanayama H, Yokota K, Kurokawa Y, Murakami Y, Nishitani M, Kagawa S (1998) Prognostic values of matrix metalloproteinase-2 and tissue inhibitor of metalloproteinase-2 expression in bladder cancer. Cancer 82:1359–1366 Karnoub AE, Dash AB, Vo AP, Sullivan A, Brooks MW, Bell GW, Richardson AL, Polyak K, Tubo R, Weinberg RA (2007) Mesenchymal stem cells within tumour stroma promote breast cancer metastasis. Nature 449:557–563 Kinoh H, Sato H, Tsunezuka Y, Takino T, Kawashima A, Okada Y, Seiki M (1996) MT-MMP, the cell surface activator of proMMP-2 (pro-gelatinase A), is expressed with its substrate in mouse tissue during embryogenesis. J Cell Sci 109(Pt 5):953–959 Kleiner DE, Stetler-Stevenson WG (1999) Matrix metalloproteinases and metastasis. Cancer Chemother Pharmacol 43(Suppl):S42–S51 Knauper V, Will H, Lopez-Otin C, Smith B, Atkinson SJ, Stanton H, Hembry RM, Murphy G (1996) Cellular mechanisms for human procollagenase-3 (MMP-13) activation. Evidence that MT1-MMP (MMP-14) and gelatinase a (MMP-2) are able to generate active enzyme. J Biol Chem 271:17124–17131 Konttinen YT, Ainola M, Valleala H, Ma J, Ida H, Mandelin J, Kinne RW, Santavirta S, Sorsa T, Lopez-Otin C, Takagi M (1999) Analysis of 16 different matrix metalloproteinases (MMP-1 to MMP-20) in the synovial membrane: different profiles in trauma and rheumatoid arthritis. Ann Rheum Dis 58:691–697 Lee NV, Sato M, Annis DS, Loo JA, Wu L, Mosher DF, Iruela-Arispe ML (2006) ADAMTS1 mediates the release of antiangiogenic polypeptides from TSP1 and 2. EMBO J 25:5270–5283 Lee CF, Ling ZQ, Zhao T, Fang SH, Chang WC, Lee SC, Lee KR (2009) Genomic-wide analysis of lymphatic metastasis-associated genes in human hepatocellular carcinoma. World J Gastroenterol 15:356–365 Leeman MF, McKay JA, Murray GI (2002) Matrix metalloproteinase 13 activity is associated with poor prognosis in colorectal cancer. J Clin Pathol 55:758–762
10 Tumor Fibroblast-Associated Metalloproteases
189
Lefebvre O, Wolf C, Limacher JM, Hutin P, Wendling C, LeMeur M, Basset P, Rio MC (1992) The breast cancer-associated stromelysin-3 gene is expressed during mouse mammary gland apoptosis. J Cell Biol 119:997–1002 Lefebvre O, Regnier C, Chenard MP, Wendling C, Chambon P, Basset P, Rio MC (1995) Developmental expression of mouse stromelysin-3 mRNA. Development 121:947–955 Lind GE, Kleivi K, Meling GI, Teixeira MR, Thiis-Evensen E, Rognum TO, Lothe RA (2006) ADAMTS1, CRABP1, and NR3C1 identified as epigenetically deregulated genes in colorectal tumorigenesis. Cell Oncol 28:259–272 Liotta LA, Kohn EC (2001) The microenvironment of the tumour-host interface. Nature 411:375– 379 Liu YJ, Xu Y, Yu Q (2006) Full-length ADAMTS-1 the ADAMTS-1 fragments display pro- and antimetastatic activity, respectively. Oncogene 25:2452–2467 Lo PH, Leung AC, Kwok CY, Cheung WS, Ko JM, Yang LC, Law S, Wang LD, Li J, Stanbridge EJ, Srivastava G, Tang JC, Tsao SW, Lung ML (2007) Identification of a tumor suppressive critical region mapping to 3p14.2 in esophageal squamous cell carcinoma and studies of a candidate tumor suppressor gene, ADAMTS9. Oncogene 26:148–157 Lopez-Otin C, Bond JS (2008) Proteases: multifunctional enzymes in life and disease. J Biol Chem 283:30433–30437 Lopez-Otin C, Matrisian LM (2007) Emerging roles of proteases in tumour suppression. Nat Rev Cancer 7:800–808 Lopez-Otin C, Overall CM (2002) Protease degradomics: a new challenge for proteomics. Nat Rev Mol Cell Biol 3:509–519 Lubberts E, Joosten LA, van Den BL, Helsen MM, Bakker AC, van Meurs JB, Graham FL, Richards CD, van Den Berg WB (1999) Adenoviral vector-mediated overexpression of IL-4 in the knee joint of mice with collagen-induced arthritis prevents cartilage destruction. J Immunol 163:4546–4556 Luque A, Carpizo DR, Iruela-Arispe ML (2003) ADAMTS1/METH1 inhibits endothelial cell proliferation by direct binding and sequestration of VEGF165. J Biol Chem 278:23656–23665 Luukkaa M, Vihinen P, Kronqvist P, Vahlberg T, Pyrhonen S, Kahari VM, Grenman R (2006) Association between high collagenase-3 expression levels and poor prognosis in patients with head and neck cancer. Head Neck 28:225–234 Lynch CC, Vargo-Gogola T, Martin MD, Fingleton B, Crawford HC, Matrisian LM (2007) Matrix metalloproteinase 7 mediates mammary epithelial cell tumorigenesis through the ErbB4 receptor. Cancer Res 67:6760–6767 Ma XJ, Dahiya S, Richardson E, Erlander M, Sgroi DC (2009) Gene expression profiling of the tumor microenvironment during breast cancer progression. Breast Cancer Res 11:R7 Maquoi E, Polette M, Nawrocki B, Bischof P, Noel A, Pintiaux A, Santavicca M, Schaaps JP, Pijnenborg R, Birembaut P, Foidart JM (1997) Expression of stromelysin-3 in the human placenta and placental bed. Placenta 18:277–285 Maquoi E, Sounni NE, Devy L, Olivier F, Frankenne F, Krell HW, Grams F, Foidart JM, Noel A (2004) Anti-invasive, antitumoral, and antiangiogenic efficacy of a pyrimidine-2,4,6-trione derivative, an orally active and selective matrix metalloproteinases inhibitor. Clin Cancer Res 10:4038–4047 Masson R, Lefebvre O, Noel A, Fahime ME, Chenard MP, Wendling C, Kebers F, LeMeur M, Dierich A, Foidart JM, Basset P, Rio MC (1998) In vivo evidence that the stromelysin-3 metalloproteinase contributes in a paracrine manner to epithelial cell malignancy. J Cell Biol 140:1535–1541 Moncada-Pazos A, Obaya AJ, Fraga MF, Viloria CG, Capella G, Gausachs M, Esteller M, LopezOtin C, Cal S (2009) The ADAMTS12 metalloprotease gene is epigenetically silenced in tumor cells and transcriptionally activated in the stroma during progression of colon cancer. J Cell Sci 122:2906–2913 Mori M, Mimori K, Shiraishi T, Fujie T, Baba K, Kusumoto H, Haraguchi M, Ueo H, Akiyoshi T (1997) Analysis of MT1-MMP and MMP2 expression in human gastric cancers. Int J Cancer 74:316–321
190
J. Lecomte et al.
Mueller MM, Fusenig NE (2004) Friends or foes—bipolar effects of the tumour stroma in cancer. Nat Rev Cancer 4:839–849 Muller D, Wolf C, Abecassis J, Millon R, Engelmann A, Bronner G, Rouyer N, Rio MC, Eber M, Methlin G (1993) Increased stromelysin 3 gene expression is associated with increased local invasiveness in head and neck squamous cell carcinomas. Cancer Res 53:165–169 Nakada M, Nakamura H, Ikeda E, Fujimoto N, Yamashita J, Sato H, Seiki M, Okada Y (1999) Expression and tissue localization of membrane-type 1, 2, and 3 matrix metalloproteinases in human astrocytic tumors. Am J Pathol 154:417–428 Nakamura H, Ueno H, Yamashita K, Shimada T, Yamamoto E, Noguchi M, Fujimoto N, Sato H, Seiki M, Okada Y (1999) Enhanced production and activation of progelatinase A mediated by membrane-type 1 matrix metalloproteinase in human papillary thyroid carcinomas. Cancer Res 59:467–473 Nawrocki B, Polette M, Marchand V, Monteau M, Gillery P, Tournier JM, Birembaut P (1997) Expression of matrix metalloproteinases and their inhibitors in human bronchopulmonary carcinomas: quantificative and morphological analyses. Int J Cancer 72:556–564 Nielsen BS, Rank F, Lopez JM, Balbin M, Vizoso F, Lund LR, Dano K, Lopez-Otin C (2001) Collagenase-3 expression in breast myofibroblasts as a molecular marker of transition of ductal carcinoma in situ lesions to invasive ductal carcinomas. Cancer Res 61:7091–7100 Nielsen BS, Rank F, Illemann M, Lund LR, Dano K (2007) Stromal cells associated with early invasive foci in human mammary ductal carcinoma in situ coexpress urokinase and urokinase receptor. Int J Cancer 120:2086–2095 Nielsen BS, Egeblad M, Rank F, Askautrud HA, Pennington CJ, Pedersen TX, Christensen IJ, Edwards DR, Werb Z, Lund LR (2008) Matrix metalloproteinase 13 is induced in fibroblasts in polyomavirus middle T antigen-driven mammary carcinoma without influencing tumor progression. PLoS ONE 3:e2959 Noel A, Pauw-Gillet MC, Purnell G, Nusgens B, Lapiere CM, Foidart JM (1993) Enhancement of tumorigenicity of human breast adenocarcinoma cells in nude mice by matrigel and fibroblasts. Br J Cancer 68:909–915 Noel AC, Polette M, Lewalle JM, Munaut C, Emonard HP, Birembaut P, Foidart JM (1994) Coordinate enhancement of gelatinase A mRNA and activity levels in human fibroblasts in response to breast-adenocarcinoma cells. Int J Cancer 56:331–336 Noel AC, Lefebvre O, Maquoi E, VanHoorde L, Chenard MP, Mareel M, Foidart JM, Basset P, Rio MC (1996) Stromelysin-3 expression promotes tumor take in nude mice. J Clin Invest 97:1924–1930 Nomura H, Sato H, Seiki M, Mai M, Okada Y (1995) Expression of membrane-type matrix metalloproteinase in human gastric carcinomas. Cancer Res 55:3263–3266 Ohtani H, Motohashi H, Sato H, Seiki M, Nagura H (1996) Dual over-expression pattern of membrane-type metalloproteinase-1 in cancer and stromal cells in human gastrointestinal carcinoma revealed by in situ hybridization and immunoelectron microscopy. Int J Cancer 68:565–570 Okada A, Tomasetto C, Lutz Y, Bellocq JP, Rio MC, Basset P (1997) Expression of matrix metalloproteinases during rat skin wound healing: evidence that membrane type-1 matrix metalloproteinase is a stromal activator of pro-gelatinase A. J Cell Biol 137:67–77 Ostman A, Augsten M (2009) Cancer-associated fibroblasts and tumor growth–bystanders turning into key players. Curr Opin Genet Dev 19:67–73 Overall CM, Dean RA (2006) Degradomics: systems biology of the protease web. Pleiotropic roles of MMPs in cancer. Cancer Metastasis Rev 25:69–75 Overall CM, Tam EM, Kappelhoff R, Connor A, Ewart T, Morrison CJ, Puente X, Lopez-Otin C, Seth A (2004) Protease degradomics: mass spectrometry discovery of protease substrates and the CLIP-CHIP, a dedicated DNA microarray of all human proteases and inhibitors. Biol Chem 385:493–504 Page-McCaw A, Ewald AJ, Werb Z (2007) Matrix metalloproteinases and the regulation of tissue remodelling. Nat Rev Mol Cell Biol 8:221–233 Pendas AM, Uria JA, Jimenez MG, Balbin M, Freije JP, Lopez-Otin C (2000) An overview of collagenase-3 expression in malignant tumors and analysis of its potential value as a target in antitumor therapies. Clin Chim Acta 291:137–155
10 Tumor Fibroblast-Associated Metalloproteases
191
Peruzzi D, Mori F, Conforti A, Lazzaro D, De Rinaldis E, Ciliberto G, Monica NLa, Aurisicchio L (2009) MMP11: a novel target antigen for cancer immunotherapy. Clin Cancer Res 15:4104–4113 Picard O, Rolland Y, Poupon MF (1986) Fibroblast-dependent tumorigenicity of cells in nude mice: implication for implantation of metastases. Cancer Res 46:3290–3294 Polette M, Nawrocki B, Gilles C, Sato H, Seiki M, Tournier JM, Birembaut P (1996) MT-MMP expression and localisation in human lung and breast cancers. Virchows Arch 428:29–35 Polette M, Gilles C, Marchand V, Seiki M, Tournier JM, Birembaut P (1997) Induction of membrane-type matrix metalloproteinase 1 (MT1-MMP) expression in human fibroblasts by breast adenocarcinoma cells. Clin Exp Metastasis 15:157–163 Porte H, Triboulet JP, Kotelevets L, Carrat F, Prevot S, Nordlinger B, DiGioia Y, Wurtz A, Comoglio P, Gespach C, Chastre E (1998) Overexpression of stromelysin-3, BM-40/SPARC, and MET genes in human esophageal carcinoma: implications for prognosis. Clin Cancer Res 4:1375–1382 Porter S, Scott SD, Sassoon EM, Williams MR, Jones JL, Girling AC, Ball RY, Edwards DR (2004) Dysregulated expression of Adamalysin-Thrombospondin genes in human breast carcinoma. Clinical Cancer Res 10:2429–2440 Porter S, Clark IM, Kevorkian L, Edwards DR (2005) The ADAMTS metalloproteinases. Biochem J 386:15–27 Rio MC (2005) From a unique cell to metastasis is a long way to go: clues to stromelysin-3 participation. Biochimie 87:299–306 Rio MC, Lefebvre O, Santavicca M, Noel A, Chenard MP, Anglard P, Byrne JA, Okada A, Regnier CH, Masson R, Bellocq JP, Basset P (1996) Stromelysin-3 in the biology of the normal and neoplastic mammary gland. J Mammary Gland Biol Neoplasia 1:231–240 Rocks N, Paulissen G, Quesada Calvo F, Polette M, Gueders M, Munaut C, Foidart JM, Noel A, Birembaut P, Cataldo D (2006) Expression of a disintegrin and metalloprotease (ADAM and ADAMTS) enzymes in human non-small-cell lung carcinomas (NSCLC). Br J Cancer 94(5):724–730 Rocks N, Paulissen G, El Hour M, Quesada F, Crahay C, Gueders M, Foidart JM, Noel A, Cataldo D (2008) Emerging roles of ADAM and ADAMTS metalloproteinases in cancer. Biochimie 90(2):369–379 Roy R, Zhang B, Moses MA (2006) Making the cut: protease-mediated regulation of angiogenesis. Exp Cell Res 312:608–622 Sabeh F, Li XY, Saunders TL, Rowe RG, Weiss SJ (2009) Secreted versus membrane-anchored collagenases: relative roles in fibroblast-dependent collagenolysis and invasion. J Biol Chem 284:23001–23011 Santos ES, Perez CA, Raez LE (2009) How is gene-expression profiling going to challenge the future management of lung cancer? Future Oncol 5:827–835 Sato H, Kida Y, Mai M, Endo Y, Sasaki T, Tanaka J, Seiki M (1992) Expression of genes encoding type IV collagen-degrading metalloproteinases and tissue inhibitors of metalloproteinases in various human tumor cells. Oncogene 7:77–83 Sato H, Takino T, Okada Y, Cao J, Shinagawa A, Yamamoto E, Seiki M (1994) A matrix metalloproteinase expressed on the surface of invasive tumour cells. Nature 370:61–65 Schonbeck U, Mach F, Sukhova GK, Atkinson E, Levesque E, Herman M, Graber P, Basset P, Libby P (1999) Expression of stromelysin-3 in atherosclerotic lesions: regulation via CD40CD40 ligand signaling in vitro and in vivo. J Exp Med 189:843–853 Shukla CJ, Pennington CJ, Riddick ACP, Sethia KK, Ball RY, Edwards DR (2008) Laser capture microdissection in prostate cancer research: establishment and validation of a powerful tool for the assessment of tumour–stroma interactions. Brit J Urol 101:765–774 Sounni NE, Noel A (2005) Membrane type-matrix metalloproteinases and tumor progression. Biochimie 87:329–342 Sounni NE, Baramova EN, Munaut C, Maquoi E, Frankenne F, Foidart JM, Noel A (2002) Expression of membrane type 1 matrix metalloproteinase (MT1-MMP) in A2058 melanoma cells is associated with MMP-2 activation and increased tumor growth and vascularization. Int J Cancer 98:23–28
192
J. Lecomte et al.
Sounni NE, Roghi C, Chabottaux V, Janssen M, Munaut C, Maquoi E, Galvez BG, Gilles C, Frankenne F, Murphy G, Foidart JM, Noel A (2004) Up-regulation of vascular endothelial growth factor-A by active membrane-type 1 matrix metalloproteinase through activation of Src-tyrosine kinases. J Biol Chem 279:13564–13574 Sternlicht MD, Werb Z (2001) How matrix metalloproteinases regulate cell behavior. Annu Rev Cell Dev Biol 17:463–516 Sternlicht MD, Lochter A, Sympson CJ, Huey B, Rougier JP, Gray JW, Pinkel D, Bissell MJ, Werb Z (1999) The stromal proteinase MMP3/stromelysin-1 promotes mammary carcinogenesis. Cell 98:137–146 Stetler-Stevenson WG, Aznavoorian S, Liotta LA (1993) Tumor cell interactions with the extracellular matrix during invasion and metastasis. Annu Rev Cell Biol 9:541–573 Strongin AY, Collier I, Bannikov G, Marmer BL, Grant GA, Goldberg GI (1995) Mechanism of cell surface activation of 72-kDa type IV collagenase. Isolation of the activated form of the membrane metalloprotease. J Biol Chem 270:5331–5338 Sugimoto H, Mundel TM, Kieran MW, Kalluri R (2006) Identification of fibroblast heterogeneity in the tumor microenvironment. Cancer Biol Ther 5(12):1640–1646 Tokuraku M, Sato H, Murakami S, Okada Y, Watanabe Y, Seiki M (1995) Activation of the precursor of gelatinase A/72 kDa type IV collagenase/MMP-2 in lung carcinomas correlates with the expression of membrane-type matrix metalloproteinase (MT-MMP) and with lymph node metastasis. Int J Cancer 64:355–359 Taniwaki K, Fukamachi H, Komori K, Ohtake Y, Nonaka T, Sakamoto T, Shiomi T, Okada Y, Itoh T, Itohara S et al (2007) Stroma-derived matrix metalloproteinase (MMP)-2 promotes membrane type 1-MMP-dependent tumor growth in mice. Cancer Res 67(9):4311–4319 Tryggvason K, Hoyhtya M, Pyke C (1993) Type IV collagenases in invasive tumors. Breast Cancer Res Treat 24:209–218 Ueno H, Nakamura H, Inoue M, Imai K, Noguchi M, Sato H, Seiki M, Okada Y (1997) Expression and tissue localization of membrane-types 1, 2, and 3 matrix metalloproteinases in human invasive breast carcinomas. Cancer Res 57:2055–2060 van de Vijver MJ, He YD, van’t Veer LJ, Dai H, Hart AA, Voskuil DW, Schreiber GJ, Peterse JL, Roberts C, Marton MJ, Parrish M, Atsma D, Witteveen A, Glas A, Delahaye L, van der Velde T, Bartelink H, Rodenhuis S, Rutgers ET, Friend SH, Bernards R (2002) A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347:1999–2009 van’t Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA, Mao M, Peterse HL, van der Kooy K, Marton MJ, Witteveen AT, Schreiber GJ, Kerkhoven RM, Roberts C, Linsley PS, Bernards R, Friend SH (2002) Gene expression profiling predicts clinical outcome of breast cancer. Nature 415:530–536 Vazquez F, Hastings G, Ortega MA, Lane TF, Oikemus S, Lombardo M, Iruela-Arispe ML (1999) METH-1, a human ortholog of ADAMTS-1, and METH-2 are members of a new family of proteins with angio-inhibitory activity. J Biol Chem 274:23349–23357 Viloria CG, Obaya AJ, Moncada-Pazos A, Llamazares M, Astudillo A, Capella G, Cal S, LopezOtin C (2009) Genetic inactivation of ADAMTS15 metalloprotease in human colorectal cancer. Cancer Res 69:4926–4934 Vogelstein B, Kinzler KW (2004) Cancer genes and the pathways they control. Nat Med 10:789– 799 Wolf C, Chenard MP, Durand DG, Bellocq JP, Chambon P, Basset P (1992) Breast-cancer-associated stromelysin-3 gene is expressed in basal cell carcinoma and during cutaneous wound healing. J Invest Dermatol 99:870–872 Yamamoto M, Mohanam S, Sawaya R, Fuller GN, Seiki M, Sato H, Gokaslan ZL, Liotta LA, Nicolson GL, Rao JS (1996) Differential expression of membrane-type matrix metalloproteinase and its correlation with gelatinase A activation in human malignant brain tumors in vivo and in vitro. Cancer Res 56:384–392 Yamashita K, Tanaka Y, Mimori K, Inoue H, Mori M (2004) Differential expression of MMP and uPA systems and prognostic relevance of their expression in esophageal squamous cell carcinoma. Int J Cancer 110:201–207
10 Tumor Fibroblast-Associated Metalloproteases
193
Yoshizaki T, Sato H, Maruyama Y, Murono S, Furukawa M, Park CS, Seiki M (1997) Increased expression of membrane type 1-matrix metalloproteinase in head and neck carcinoma. Cancer 79:139–144 Zeisberg EM, Potenta S, Xie L, Zeisberg M, Kalluri R (2007) Discovery of endothelial to mesenchymal transition as a source for carcinoma-associated fibroblasts. Cancer Res 67:10123– 10128 Zhao H, Sohail A, Sun Q, Shi Q, Kim S, Mobashery S, Fridman R (2008) Identification and role of the homodimerization interface of the glycosylphosphatidylinositol-anchored membrane type 6 matrix metalloproteinase (MMP25). J Biol Chem 283:35023–35032 Zhang W, Matrisian LM, Holmbeck K, Vick CC, Rosenthal EL (2006) Fibroblast-derived MT1MMP promotes tumor progression in vitro and in vivo. BMC Cancer 6:52 Zigrino P, Loffek S, Mauch C (2005) Tumor-stroma interactions: their role in the control of tumor cell invasion. Biochimie 87:321–328 Zucker S, Pei D, Cao J, Lopez-Otin C (2003) Membrane type-matrix metalloproteinases (MTMMP). Curr Top Dev Biol 54:1–74
.
Part IV
Tumor Modulating-Fibroblast Interactions
.
Chapter 11
Multiple Fibroblast Phenotypes in Cancer Patients: Heterogeneity in Expression of Migration Stimulating Factor Ana M. Schor and Seth L. Schor
11.1 Introduction Advances in molecular biology during the past three decades lent support to a genetic model of carcinogenesis according to which the progressive accumulation of mutations in relevant oncogenes and tumour suppressor genes was deemed necessary and sufficient to result in cancer development. The contemporaneous recognition that carcinogenesis is a multi-step process resulted in the now widely held view that cancer development commences with the inception of an initiating genetic lesion affecting the expression of a cancer-critical gene (Knudson 2001). This genetic change is postulated to confer a relative growth advantage to the progeny of the initiated cell. Disease progression results from the subsequent random accumulation of complementary genetic lesions within the initiated clonal population and the selection of derivative sub-clones displaying ever increasing growth advantage (Nowell 1976). Successive iterations of this clonal selection process result in the emergence of progressively dysfunctional populations of pre-neoplastic and neoplastic cells, culminating in the appearance of an overt malignancy. A corollary of this genetic model is that cancer inception and progression are essentially deterministic processes inexorably driven by the progressive accumulation of genomic lesions. Although undoubtedly of major importance, this strictly genetic view is not sufficient in itself to account for all existing data. For example, numerous studies have documented the local and systemic presence in cancer patients of stromal cells displaying morphological, biochemical and behavioural features customarily used to define neoplastic transformation in vitro. These cells were, however, commonly reported to be non-tumorigenic when assayed in vivo and therefore classified as “partially-transformed”. Several genetic mechanisms were invoked to account for the systemic presence of aberrant stromal cells at sites distant from the tumour (usually skin), including the inheritance of a predisposing germ-line mutation in patients A. M. Schor () Unit of Cell and Molecular Biology, The Dental School, University of Dundee, Dundee DD1 4HR, UK e-mail:
[email protected] M. M. Mueller, N. E. Fusenig (eds.), Tumor-Associated Fibroblasts and their Matrix, The Tumor Microenvironment 4, DOI 10.1007/978-94-007-0659-0_11, © Springer Science+Business Media B.V. 2011
197
198
A. M. Schor and S. L. Schor
with familial cancer syndromes. Seen within the confines of the prevailing genetic and epithelio-centric model of cancer pathogenesis, these aberrant stromal cells were generally not considered to make a direct contribution to cancer development. Instead, their presence was interpreted as a convenient biomarker of an inherited (or incurred) genetic lesion which only contributed to carcinogenesis when concomitantly expressed by the cancer progenitor cell population. Tumours are, however, a complex community of different cell types, including (in carcinomas) malignant epithelial cells, fibroblasts, myofibroblasts, endothelial cells, pericytes and various types of immune/inflammatory cells. Reciprocal interactions between these diverse cell populations are mediated by the inter-dependent signalling of cell-produced soluble factors and matrix macromolecules, as well as by direct cell-cell contact. Recognition of this inter-dependency prompted earlier workers to postulate that the behaviour of tumour cells may result from their interactions with functionally aberrant tumour-associated stromal cells (Tarin 1972; Cunha et al. 1985; Bissell and Barcellos-Hoff 1987; Schor and Schor 1987; Schor et al. 1987; Skobe and Fusenig 1998). Although not initially meeting with universal approbation, this “contextual” model has recently gained more widespread support (Broxterman and Georgopapadakou 2007; Kiaris et al. 2008; Tsellou and Kiaris 2008; Pietras et al. 2008; Guturu et al. 2009; Anton and Glod 2009; Xu et al. 2009). This chapter will focus on the evidence demonstrating the presence of functionally aberrant fibroblasts in cancer patients and their possible contribution to disease pathogenesis. Our own work has indicated that fibroblasts obtained from a majority of cancer patients resemble fetal cells in terms of their persistent production of a migration stimulating factor (MSF) which is not made by their normal adult counterparts. The multiple bioactivities and target cell populations of MSF suggest several means whereby its inappropriate expression may promote cancer progression. We now specifically propose to review evidence concerning (1) the existence of local and systemic fibroblast heterogeneity and its implications for tumour progression, (2) the molecular characterisation and functionality of MSF, and (3) the potential utility of MSF as a target for developing novel clinical intervention strategies. Our intention is to incorporate these data into a broadened conceptual framework of cancer pathogenesis explicitly recognising the joint contribution of genetic and epigenetic causality.
11.2 Local and Systemic Fibroblast Heterogeneity in Cancer Patients: Implications for Tumour Progression 11.2.1 T he Presence of Functionally Aberrant Fibroblasts in Cancer Patients There is a substantial literature documenting the presence of functionally aberrant fibroblasts in cancer patients. Pathologists have commonly described aberrantly
11 Multiple Fibroblast Phenotypes in Cancer Patients
199
appearing fibroblasts in tumour stroma as “fetal-like”, “plump”, “reactive” or “activated” in an effort to distinguish them from their “resting” or “quiescent” counterparts (“fibrocytes”) in healthy adult tissues (McNeal 1984). In addition to this “local” manifestation of stromal abnormality, aberrantly appearing and behaving fibroblasts have also been detected “systemically” at distant, uninvolved, skin in patients with either apparently sporadic forms of the disease or clearly defined hereditary cancer syndromes. With respect to the former patient group, tumour-associated stromal fibroblasts in patients with sporadic forms of colorectal, breast, ovarian, bladder and lung cancers have been reported to stain with antibodies specific for a cell surface glycoprotein (F19) previously shown to be expressed by sarcomas and proliferating cultured fibroblasts, but not by normal resting cells (Garin-Chesa et al. 1990). These workers further commented on the co-distribution of F19 reactive fibroblasts with fetal isoforms of the matrix molecule tenascin. Fibroblasts displaying aberrant phenotypic characteristics have also frequently been detected at distant uninvolved sites in patients with hereditary cancer syndromes. In this regard, Kopelovich and colleagues noted that skin fibroblasts obtained from patients with hereditary adenomatosis of the colon and rectum (ACR) or neurofibromatosis exhibit a number of aberrant phenotypic characteristics in vitro, including disorganisation of the actin microfilament array, reduced cell substratum adhesion, anchorage-independent growth, elevated sensitivity to overt transformation by oncogenic agents and the expression of a transformation-associated antigen (Frankel et al. 1986; Higgins and Kopelovich 1991; Kopelovich 1987, 1988; Kopelovich and Bias 1980; Kopelovich et al. 1980). Independent studies have corroborated and extended these findings by noting that skin fibroblasts from patients with ACR and other hereditary cancer syndromes (such as aniridia, dysplastic nevus syndrome, von Hippel-Lindau syndrome and Li-Fraumini syndrome) display enhanced sensitivity to transformation by carcinogens and a reduced efficiency of DNA repair (Rasheed and Gardner 1981; Rhim et al. 1981; Abrahams et al. 1998). As all of these familial cancer syndromes result from a germ line mutation, it is reasonable to assume that the annotated fibroblast abnormalities result from the systemic expression of the affected gene. Seen in this context, the aberrant fibroblasts have commonly been classified as “partially-transformed” or “initiated” cells and considered to be a convenient biomarker of an inherited predisposition to develop an overt malignancy by the relevant target cell population. Similarly aberrant skin fibroblasts have been obtained from patients with apparently sporadic cancers. Thielmann et al. (1987) reported that dermal fibroblasts obtained from patients with squamous cell carcinoma (SCC) and (to a lesser extent) basal cell carcinoma exhibited impaired repair DNA synthesis. Similarly, skin fibroblasts obtained from patients with oral SCC (Danes et al. 1990) or sporadic colon cancer (Svendsen et al. 1989) exhibited an increased tendency to develop hyperdiploidy when cultured in vitro. Danes et al. (1990) further noted a correlation between the precise anatomical location of SCC within the oral cavity and the expression of this skin fibroblast abnormality. Explicit use of the term “fetal-like” was made by McNeal (1984) in a comprehensive early study of the pathogenesis of benign prostatic hyperplasia and its
200
A. M. Schor and S. L. Schor
subsequent malignant transformation in which he noted the early appearance of a “fetal-like” stroma and postulated that its interaction with the emerging population of aberrant epithelial cells was a key factor in disease progression. Tumour-associated fibroblasts have also been reported to resemble fetal cells in terms of their expression of a fetal-specific cell surface antigen (Bartal et al. 1986) and production of stromelysin-3 (Basset et al. 1990). In this regard, it should be noted that fetal cells also form colonies in semi-solid medium (Nakano and Ts’O 1981). Nicolo et al. (1990) have reported that the oncofetal ED-B isoform of fibronectin is preferentially associated with the stroma of several types of malignant neoplasms. We also employed the term fetal-like, noting that (1) tumour-associated fibroblasts obtained from approximately 50% of resected breast carcinomas displayed an elevated migratory phenotype in vitro, similar to that of fetal fibroblasts (Durning et al. 1984; Schor et al. 1994), (2) systemic fibroblasts from approximately 50% of uninvolved skin biopsies obtained from patients with a range of sporadic cancers (including breast and other common carcinomas, soft tissue sarcomas and melanomas) also displayed fetal-like phenotypic characteristics (Schor et al. 1985a, b, 1988b) and (3) systemic fibroblasts obtained from uninvolved skin from 100% of patients with hereditary breast cancer (Schor et al. 1986), and approximately 60–70% of their unaffected first degree relatives displayed these same functional aberrations (Schor et al. 1986; Haggie et al. 1987). Other workers also explicitly noted that skin fibroblasts obtained from sporadic breast cancer patients displayed a number of fetal-like characteristics in vitro, including anchorage-independent growth, colony formation on epithelial monolayers, extended lifespan and continued DNA synthesis at saturation cell density (Azzarone and Macieira-Coelho 1987; Azzarone et al. 1984, 1988; Wynford-Thomas et al. 1986; Schor et al. 1988b). This shift in nomenclature from “partially transformed” to “fetal-like” is more than just a question of semantics (Schor et al. 1987; Schor and Schor 1997). The designation “partially transformed” carries with it the implication that the observed behaviour results from acquisition of a genetic lesion. In contrast, the term “fetallike” implies that the particular phenotypic attributes which define this state (e.g. continued production of MSF) are inherently physiological and reversible, although their expression may be inappropriate in the adult. Taken together, the detection of local and systemic functionally aberrant, fetallike, fibroblasts in cancer patients raises two important questions, namely: • What mechanisms lead to their presence in cancer patients?, and • What, if any, contribution do these fibroblasts make to cancer pathogenesis?
11.2.2 P ossible Mechanisms Responsible for the Presence of Aberrant Fibroblasts in Cancer Patients Which factors may lead to the postulated inappropriate expression of fetal phenotypic characteristics by fibroblasts in cancer patients? Our hypothesis is based
11 Multiple Fibroblast Phenotypes in Cancer Patients
201
on the fact that fibroblasts (in spite of their similar morphology) are actually a highly diverse cell population, exhibiting a significant degree of inter-site, intra-site and developmental heterogeneity (Schor and Schor 1987; Bayreuther et al. 1988; Sappino et al. 1990; Mukaida et al. 1991; Schor et al. 1996). On the basis of this heterogeneity, we initially proposed a “clonal modulation” model to account for the presence of MSF-secreting fibroblasts in cancer patients(Schor and Schor 1987; Schor et al. 1987). In this model, we suggested that (1) distinct subpopulations of fetal-like fibroblasts pre-exist in the healthy adult, (2) the fetal-like characteristics of these cells do not result from an acquired mutation, but reflect epigenetically regulated changes in gene expression analogous to those which occur during embryonic development and (3) their relative numbers may be increased, both locally and systemically, as a consequence of clonal expansion and/or epigenetic induction in response to both internal and environmental cues. In this regard, studies by Dabbous et al. (1987) revealed the existence of significant clonal heterogeneity in the production of proteases by tumour-associated fibroblasts in response to inductive signals from co-cultured carcinoma cells. Published data confirm the existence of inter- and intra-site heterogeneity in MSF expression by fibroblast subpopulations in fetal skin (Schor et al. 1985a), as well as in the healthy adult (Irwin et al. 1994). In the latter study, fibroblasts obtained from 100% (12/12) gingival biopsies examined produced detectable amounts of MSF, whereas none (0/9) of paired forearm skin fibroblasts obtained from the same individuals did so. Interestingly, wound healing in the oral mucosa is clinically distinguished from dermal healing in terms of both its rapidity and lack of scar formation (McCallion and Ferguson 1996). It is possible that the persistence of MSF-producing fibroblasts in the oral mucosa contributes to this regenerative and characteristically “fetal-like” mode of wound healing. This study further revealed the existence of intra-site heterogeneity in the oral mucosa. This involved separation of gingival lamina propria (connective tissue) from its overlying epithelium by exposure to trypsin and the subsequent microdissection of the lamina propria to allow the selective culture of fibroblasts derived from the tips of the papillae and deeper reticular tissue. Only fibroblasts derived from the papillae produced MSF. Prolonged subculture of papillary fibroblasts resulted in their cessation in MSF production and their adoption of a reticular fibroblast phenotype. Staining of gingival tissue with anti-MSF antibody confirmed the preferential localisation of MSF in the papillae; interestingly, this particular pattern of MSF distribution is identical to that previously described for tenascin (Sloan et al. 1990), another molecule preferentially produced by fetal fibroblasts (Chiquet-Ehrismann et al. 1986). Intra-site heterogeneity with respect to the distribution of MSF-secreting fibroblasts has also been observed in the normal breast (Schor et al. 1994). Intra- and interlobular fibroblasts were isolated by controlled enzymatic digestion and differential sedimentation from normal breast tissue obtained from patients undergoing reduction mammoplasty. Results indicated that 91% (10/11) of the interlobular fibroblasts displayed a fetal-like migratory phenotype, compared to 0% (0/10) of the intralobular cells. Significant differences were also observed with respect to the production of MSF by these cells, with 100% (11/11) of the interlobular lines and none (0/10) of the intralobular lines secreting this factor. These observations
202
A. M. Schor and S. L. Schor
indicate the presence of a clearly defined intra-site heterogeneity with respect to MSF production by fibroblasts in the normal breast of non-cancer patients.
11.2.3 C ontribution of Aberrant Fibroblasts to Cancer Pathogenesis The vast majority of early studies documenting the presence of aberrant fibroblasts in cancer patients have avoided ascribing any carcinogenic relevance to them. Indeed, as mentioned above, in those few communications in which this question was explicitly addressed, it was assumed that such fibroblasts resulted from the inheritance of a germ line mutation which only contributed to cancer pathogenesis when also expressed by the relevant epithelial cell population. According to this “innocent bystander” model, aberrant fibroblasts were considered to be convenient phenotypic markers of a putative genetic lesion and not an intrinsic driver of disease progression. This passive interpretation has gradually been replaced by a more proactive perspective explicitly recognising the important role played by interactions between different cell populations in determining their respective phenotypes (Broxterman and Georgopapadakou 2007; Kiaris et al. 2008; Tsellou and Kiaris 2008; Pietras et al. 2008; Guturu et al. 2009; Anton and Glod 2009). Such interactions have long been recognised to regulate epithelial growth, migration and differentiation during fetal development (Cunha et al. 1985; Gurdon 1988). Shekhar et al. (2001) presented data supporting the role played by aberrant stromal cells in the manifestation of malignant phenotypic characteristics by associated breast carcinoma cells. This conclusion is consistent with a large number of studies indicating that interactions between tumour cells and fibroblasts enhance tumour growth (Camps et al. 1990) and metastasis in vivo (Picard et al. 1986; Tanaka et al. 1988; Gärtner et al. 1992) as well as invasion and various other neoplastic characteristics in vitro (Picard et al. 1986; Tanaka et al. 1988; Matsumoto et al. 1989; Gärtner et al. 1992; Atula et al. 1997). The expression of growth factors and proteases has also been reported to be modulated by interactions between tumour and stromal cells (Wong and Wang 2000; Dong et al. 2001; Sung and Chung 2002; Koshida et al. 2006; Maeda et al. 2006; Sugimoto et al. 2005; He et al. 2007). Perhaps the most dramatic manifestation of such tissue level interactions relate to observations that an appropriate host environment is capable of suppressing carcinogenesis resulting from the expression of oncogenes ordinarily capable of inducing rapid tumour development (Mintz and Illmensee 1975; Stoker et al. 1990). These effects of fibroblast subpopulations on epithelial cells are complex and vary both as a function of developmental stage (fetal vs adult), disease progression (early vs late stages) and fibroblast site of origin (Cornil et al. 1991; Fabra et al. 1992; Lu et al. 1992). In addition to noting a stage-dependency of tumour cell interaction with their associated stromal fibroblasts, Tsellou and Kiaris (2008) further demonstrated that there appears to be a parallel temporal evolution towards manifestation of more neoplastic behavioural features in both cell populations.
11 Multiple Fibroblast Phenotypes in Cancer Patients
203
Interactions between epithelial cells and fibroblasts are mediated by the interdependent signalling of soluble regulatory molecules (such as cytokines) and insoluble constituents of the extracellular matrix (ECM). The effects of cytokines and matrix macromolecules are mutually interdependent in the sense that (1) the specific response of cells to cytokines is modulated by the nature of the surrounding matrix, and (2) the deposition of this matrix is modulated by the action of cytokines (Schor 1994; Schor and Schor 1987; Nathan and Sporn 1991). Interactions between tumour cells and fibroblasts have been reported to affect the production of both cytokines and matrix macromolecules in a bidirectional fashion. These interactions involve positive feedback loops which may result in an expansion in cell number and/or amplification of signal molecule synthesis. For example, Cullen et al. (1991) reported that mammary carcinoma cells synthesize PDGF which stimulates fibroblast proliferation and synthesis of IGF I and II; interestingly, the fibroblast-produced IGFs in turn stimulate mammary carcinoma cell proliferation and synthesis of PDGF. Recent studies have documented a reciprocal effect of such local cell-cell interactions on gene expression between breast epithelial cells (non-malignant and malignant) and different populations of fibroblasts (Rozenchan et al. 2009), including alterations in the expression of TGF-β regulated genes. The effects of such tumour-associated stromal cells may be mediated by several mechanisms, including cell-cell contact and the concerted signalling of cell-produced matrix macromolecules and soluble factors. The possible role of aberrant systemic (distant) fibroblasts remains more difficult to conceptualise. At one end of the spectrum is the notion that such fibroblasts make no direct contribution to cancer pathogenesis, but provide a biomarker of exposure to factors (endogenous and/or environmental) which induced similar changes to cells (stromal and epithelial) at the site of tumour development. At the other extreme, it is also possible that such systemic cells do indeed precede and make a direct contribution to cancer pathogenesis, possibly by generating bioactive soluble factors (such as MSF) or by providing a receptive microenvironment (e.g. hyluronan-rich) to support the seeding of metastatic tumour cells at distant sites. Further studies are required to discriminate between these possibilities.
11.2.4 Section Summary Fibroblasts obtained from cancer patients commonly display a number of phenotypic characteristics which distinguish them from their healthy adult counterparts. Significantly, such fibroblasts have been obtained from both the tumour-associated stroma (i.e. locally), as well as from distant, apparently uninvolved, sites (i.e. systemically). We suggest that (1) such fibroblasts may actively contribute to cancer pathogenesis by perturbing normative interactions with neighbouring epithelial cells, and (2) have arisen by changes in gene expression mediated by epigenetic mechanisms resulting from exposure to endogenous or environmental “epigenotoxins” and/or the clonal expansion of a pre-existing subpopulation. Considerably more remains to be learned about the extent of fibroblast functional heterogeneity
204
A. M. Schor and S. L. Schor
in normal and diseased tissues, the mechanisms responsible for their divergence and their consequences for disease progression.
11.3 MSF: A Novel Bioactive Oncofetal Protein 11.3.1 Molecular Characterisation We initially reported that fibroblasts derived from cancer patients (both tumourassociated and those obtained from distant uninvolved sites) differ from their healthy adult counterparts in terms of their persistent expression when cultured in vitro of a putative migration stimulating factor (MSF) (Schor et al. 1988a). MSF was later shown to be a truncated isoform of fibronectin (Schor et al. 2003). Fibronectin is a modular glycoprotein consisting of the following functional domains, so named on the basis of their binding affinities to other matrix molecules and cell surface integrins (Fig. 11.1): Hep-1/Fib-1 (N-terminal low affinity binding to heparin and type III
1
type II type I
Fibronectin
MSF+aa
3
4
5
6
7
8
9 10 11
12 13 14
15
1 2 1 2 3 4 5
6
I I I I I
I II II I I I
Hep1/Fib1
Gel- BD
I I I
I
I
Hep1/Fib1
MSF-aa
2
I I I
I
I
Hep1/Fib1
7 8 9
I II II
I
ED B
III III III III
Cell-BD
I
10 11 12
RGD III III III III III III III
ED A
III III III
IIICS
Hep2
I
III1a
intron 12 derived unique decamer
I
III1a
intron 12 derived unique decamer
III I I I
Fib 2
Gel- BD
I II II
I
I
Gel-BD -15 aa
Fig. 11.1 MSF is a genetically truncated isoform of fibronectin. Fibronectin is a modular glycoprotein consisting of the following “functional domains” defined on the basis of their characteristic binding affinities for other matrix macromolecules and cell surface receptors: Hep-1/ Fib-1 (N-terminal binding to heparin and fibrin), Gel-BD (binding to gelatin/collagen), Cell-BD (RGD-mediated binding to cell surface integrins), Hep-2 (high affinity binding to heparin) and Fib-2 (C-terminal binding to fibrin). Each of these functional domains is composed of several “homology modules”, respectively designated as types I, II and III. Alternative splicing at EDA, EDB and IIICS generates approximately 20 “full-length” fibronectin isoforms (molecular masses in the region of 250–280 kDa). MSF+aa is a truncated (70 kDa) isoform of fibronectin identical to its N-terminus, up to and including the amino acid sequence coded by exon III-1a. MSF message is transcribed from the fibronectin gene by a variation of standard alternative splicing involving read-through of intron 12 (separating exons III-1a and III-b), followed by intra-intronic cleavage. MSF+aa protein terminates in an MSF-unique 10 amino acid sequence (coded by the first 30 bp of intron 12) which is not present in any full-length fibronectin. Arrows indicate the location of the four IGD motifs in modules I-3, I-5, I-7 and I-9. MSF-aa is identical to MSF+aa, with the exception of a 15 amino acid deletion in module II-1
11 Multiple Fibroblast Phenotypes in Cancer Patients
205
fibrin), Gel-BD (binding to gelatin/collagen), Cell-BD (RGD-mediated binding to integrins), Hep-2 (high affinity heparin binding) and Fib-2 (C-terminal fibrin binding site). Each functional domain is composed of a different tandem array of three possible homology modules, respectively designated type I, II and III. Approximately 20 “full-length” fibronectin isoforms are produced by alternative splicing involving the inclusion or exclusion of two particular type III modules (EDA and EDB), and a more complex splicing repertoire at the IIICS region (Hynes 1990). All such isoforms consist of two peptide chains (250–280 kDa each) covalently linked by disulfide bonds at their respective C-termini. In the fibronectin gene, all type I and II homology modules are coded by correspondingly named single exons; in contrast, all type III modules are coded by two consecutive exons (designated “a” and “b”), with the exception of the individual exons coding for the “extra domain” type III modules (EDA and EDB) found in those isoforms displaying an oncofetal expression profile. MSF is a genetically truncated isoform (70 kDa) of fibronectin. It is identical to the N-terminus of the fibronectin monomer, up to and including the amino acid sequence coded by exon III-1a, followed by an MSF-specific intron-coded C-terminal decamer (Fig. 11.1) (Schor and Schor 2001; Schor et al. 2003). Murine MSF, truncated at the same exon as human MSF, has also been cloned (unpublished). Other truncated isoforms of fibronectin have been described in Zebrafish embryos (Zhao et al. 2001) goldfish and rainbow trout, as well as mouse and human liver, prostate, ovary, brain and spleen (Liu et al. 2003). Two isoforms of human MSF have been cloned by our group. Both contain the same MSF-specific intron-coded C-terminal decamer and, as will be discussed below, both isoforms also contain the same bioactive IGD amino acid motifs and consequently display the same spectrum of equipotent effects on target cells. The two isoforms differ solely in terms of a 45 bp deletion in exon II-1 and are consequently referred to as MSF+aa and MSF-aa to indicate the retention or deletion of a 15 amino acid sequence in module II-1 (Fig. 11.1). The global term MSF will be employed to denote both isoforms (i.e. total MSF). Based on this information, we have generated the following MSF-specific reagents: (1) a pan-MSF identification antibody raised against the shared MSF-specific C-terminal decamer, (2) a function-neutralising antibody effectively abrogating the bioactivities of both MSF isoforms, (3) an identification antibody specific for MSF-aa raised against the neoantigen generated by its 15 amino acid deletion, (4) riboprobes for MSF message, and (5) a murine-specific pan-MSF identification polyclonal antibody raised against its distinct intron-coded C-terminus. The identification antibodies and riboprobes have been optimised and validated for use in immunohistochemistry (IHC) and in situ hybridisation (ISH) with archival tissue blocks. The identification antibodies have also been validated for biochemical analyses (ELISA, western and dot blots). MSF+aa message is generated by a two-stage processing mechanism. This initially entails generating an MSF-specific primary transcript from the fibronectin gene by read-through of intron 12 (separating exons III-1a and III-1b), followed by intra-intronic cleavage to produce a 5.9 kb MSF pre-message (Kay et al. 2005). The pre-message remains sequestered within the nucleus, where it is rapidly degraded as a consequence of the presence of an AU rich instability element (Bakheet et al. 2001) in its 3′-UTR (Schor et al. 2003; Kay et al. 2005). Under standard tissue
206
A. M. Schor and S. L. Schor
culture conditions, this pre-message is produced by cells that express MSF protein (such as fetal fibroblasts), as well as by cells which do not (such as fibroblast from healthy adult tissues). This unexpected occurrence may reflect the induction of a “partially activated” (“wounded”) phenotype in vitro. In cells which do express MSF protein, the intron-derived 3′ UTR of the pre-message is cleaved a second time to produce a 2.1 kb mature MSF message. This has a significantly shorter (195 bp) intron-derived 3′ sequence containing a 30 bp in-frame coding sequence (immediately contiguous with exon III-1a), followed by a 165 bp 3′-UTR containing several in-frame stop codons and a cleavage/polyadenylation signal. The mature message is rapidly exported to the cytoplasm for translation. The generation of MSF-aa message presumably occurs by a similar mechanism, with the addition of the splicing out of a 45 bp sequence in exon II-1.
11.3.2 M SF: Oncofetal Expression Profile and the Contextual Control of its Expression by Epigenetic Mechanisms MSF is an oncofetal protein initially shown to be constitutively expressed in vitro by fibroblasts explanted from human fetal skin, but not from the majority of healthy adult skin (Fig. 11.2). MSF-expressing (“fetal-like”) fibroblasts were also derived from excised tumour tissue (Schor et al. 1994), as well as distant uninvolved skin in cancer patients (Schor et al. 1988a, b; Haggie et al. 1987). Subsequent ex-vivo studies confirmed and extended these initial findings by demonstrating that MSF message and protein are both expressed by at least three cell populations (epithelial,
MSF producer MSF non-producer
percentage
100 80 60 40 20 0 fetal
adult
patient (local)
patient (systemic)
Fig. 11.2 MSF expression by fibroblasts in vitro. Skin fibroblasts between passage 8–15 were assessed for their expression of MSF. Cells were derived from fetal skin (n = 37), healthy adult skin (n = 52), breast tumours (local, n = 55) and matching skin from these patients (systemic, n = 55). Serum-free conditioned media obtained from confluent cultures (72 h incubation) were assayed for the presence of MSF by assessing (1) their ability to stimulate the migration into 3D collagen gels of a target line of adult, non-MSF producing, adult skin fibroblasts (as described in Schor et al. 1988a), (2) abrogation of motogenic activity by an MSF-function neutralising antibody and, in some cases (3) removal of motogenic activity by affinity to anti-MSF identification antibody
11 Multiple Fibroblast Phenotypes in Cancer Patients
207
fibroblasts and endothelial) in fetal skin, but not by the majority of these cells in healthy adult tissue (Schor et al. 2003). These studies additionally indicated that MSF was also expressed by tumour and tumour-associated stromal cells in patients with cancer, as well as by skin cells in biopsies obtained from distal uninvolved sites (Schor et al. 2003; unpublished data). Recent findings further suggest that the MSF+aa isoform of MSF may be the only one systemically expressed by skin cells in cancer patients, whereas the MSF-aa appears to be the main isoform produced by tumour and tumour-asociated stromal cells (Fig. 11.3). More recent observations by ourselves and others confirm that MSF is expressed by tumour cells and tumour-associated stromal cells in a majority of cancers examined to date, including carcinomas (breast, colorectal, oral, oesophageal, skin), melanoma and glioma (Hu et al. 2009; unpublished observations).
Fig. 11.3 Expression of total MSF and MSF-aa by skin and tumour tissue. Serial sections of normal skin from a breast cancer patient (a, b) oral carcinoma (c, d) and breast carcinoma (e, f) were stained with antibodies that recognise total MSF (both isoforms) (a, c, e) or only MSF-aa (b, d, e). The two tumours stained positively with both antibodies, whereas staining in the skin was positive for total MSF, but negative for MSF-aa. These data suggest that skin from the breast cancer patient only expresses MSF+aa, whereas both tumours contain predominantly MSF-aa
208
A. M. Schor and S. L. Schor 40
MSF grade
number of specimens
0/1 2/3
30
20
10
0 NB
NB-T Tissues
T
Fig. 11.4 Expression of MSF in breast tissues. Paraffin-embedded archival specimens were stained with MSF-specific identification antibody raised against the MSF-unique C-terminal decamer (Schor et al. 2003). MSF staining was graded as negative (0), weak (1), moderate (2), or strong (3). Results were obtained by consensus of 2–3 independent observers. The tissues examined were: NB (n = 19): normal breast from reduction mammoplasties; NB-T (n = 19): histologically normal breast from breast tumour patients; and T (n = 23): breast carcinomas. MSF expression was significantly increased in a step-wise manner from NB to NB-T (Fisher’s exact test, p = 0.0031) and from NB-T to T (p = 0.0258)
In vitro studies initially indicated that intralobular fibroblasts isolated from histologically normal breast adjacent to a carcinoma expressed MSF, whereas such fibroblasts isolated from normal breast tissue from patients undergoing reduction mammoplasty did not (Schor et al. 1994). This finding has recently been confirmed in a semi-quatitative immuno-histochemical study using archival tissues (Perrier et al. unpublished). In this study, histologically normal breast tissue adjacent to a carcinoma displayed a significantly elevated level of MSF expression compared to normal breast obtained from reduction mammoplasty controls. MSF expression was significantly higher in carcinomas than in either of the normal breast tissues (Fig. 11.4). These findings have implications both for our understanding of the concept of field cancerisation, as well as for patient management (e.g. identification of “disease-free” margins in resected tumour tissues). In addition to being produced by a variety of cell populations in cancer patients (i.e. tumour cells, tumour-associated stromal cells and distal uninvolved host cells), bioactive MSF has also been detected in the serum of breast cancer patients (Picardo et al. 1991; unpublished data). These data indicate that MSF bioactivity (as detected by MSF-specific stimulation of fibroblast migration) is present in 90% of cancer patients compared to only 10% of healthy age-matched controls (Fig. 11.5), a value consistent with the incidence of MSF-secreting skin fibroblasts originally detected in these individuals (Fig. 11.2). Neutrophil gelatinase-associated lipocalin (NGAL) is a potent inhibitor of MSF+aa, but does not affect manifestation of MSFaa activity (Jones et al. 2007; unpublished data). NGAL is present in the serum of both healthy individuals and cancer patients, suggesting that the MSF-aa isoform is
11 Multiple Fibroblast Phenotypes in Cancer Patients negative for MSF activity positive for MSF activity
100 percentage of samples
Fig. 11.5 MSF bioactivity in serum. The presence of detectable levels of MSF bioactivity in the serum of breast cancer patients (n = 30) and age- and sex-matched healthy controls (n = 30) was ascertained as described in Picardo et al. (1991). Data are presented as the percentage of positive (detectable) and negative (not detectable) samples in the two subject groups
209
80 60 40 20 0 control
patients GROUPS
present in the latter. This conclusion has been confirmed in studies that additionally indicate that MSF+aa is in fact present in control sera, but is not able to manifest its motogenic bioactivity as a consequence of its association with an MSF inhibitor. The initial study by Picardo et al. (1991) further indicated that the bioactive MSF detected in patient sera is not a measure of tumour burden, as it persisted for many years after apparently successful resection of the presenting tumour. The cellular origin of MSF in cancer patient sera is not known. The programmed control of gene expression during development has long been recognised to be regulated by epigenetic mechanisms. In contrast to their genetic (e.g. mutational) counterparts, epigenetic mechanisms are mediated by reversible, although heritable (persistent), changes in DNA and/or histone methylation. We have recently demonstrated that MSF expression may be switched on and off in such a reversible and persistent fashion. A transient (2 h) exposure of normal adult skin fibroblasts (non-producers of MSF) growing on a “wounded” substratum in vitro (such as denatured type I collagen or fibrin) induces the expression of MSF protein. This “activated” phenotype is persistent for the entire subsequent lifespan of the treated cells cultured in the absence of TGF-β and irrespective of the nature of the substratum employed. MSF expression may, however, be switched off again at any time by a second transient exposure to TGF-β1, this time when the cells are growing on a “healthy” matrix of native type I collagen (Fig. 11.6). It is important to note that this “on-off” switch is repeatable and strictly requires the concerted action of TGF-β1 and the appropriate matrix: i.e. exposure to TGF-β1 of previously activated cells growing on a wounded matrix will not switch off MSF expression. These observations may shed light on the apparent stage-dependence of TGF-β influence on tumour progression: namely, TGF-β acts as a suppressor of early stage disease and as a promoter of late stage disease (Bierie and Moses 2006). We suggest that these and related perplexing observations might reflect the switching-off of MSF expression by activated fibroblasts associated with early stage disease (in which matrix degradation is minimal) and the converse switching-on of MSF in late
210
A. M. Schor and S. L. Schor ii. TGF-β
persistent MSF production healthy matrix MSF-np
MSF-p
persistent non-producer wounded matrix
i. TGF-β iii. 5-azaC
Fig. 11.6 Control of MSF expression. Quiescent adult skin fibroblasts do not produced MSF (MSF-np). They are, however, induced to do so by (i) a transient (2 h) exposure to TGF-β1 when cultured on a “wound” matrix, such as fibrin or denatured type I collagen. This switch-on is persistent and the resultant “activated” MSF-producing cells (MSF-p) continue to express MSF for the entire duration of their in vitro lifespan when cultured under standard tissue culture conditions in the absence of TGF-β and irrespective of the nature of their substratum. MSF expression may be persistently switched-off again by (ii) a subsequent transient exposure of activated cells to TGFβ1, this time when they are grown on a “healthy tissue” matrix (such as native type I collagen). This switch-on and -off is completely reversible and may be repeated for the entire duration of their in vitro life span. MSF-np cells may be similarly activated by a transient exposure to (iii) 5-azacytidine (5-azaC). The persistent MSF expression so induced is again switched off by a subsequent transient exposure of resultant MSF-p cells to TGF-β1 when adherent to a native collagen substratum
stage disease (in which matrix degradation is more extensive). This matrix-modulation of TGF-β functionality is consistent with the extensive literature documenting the functional interactions of cytokines and matrix (Nathan and Sporn 1991; Schor 1994; Pardali and ten Dijk 2009; Padua and Massague 2009). TGF-β1 also up-regulates the expression of full-length oncofetal (EDA and EDB) fibronectins (Hynes 1990). The possible involvement of epigenetic mechanisms in the switching-on and -off of MSF expression has initially been suggested by observations that a transient exposure of normal adult fibroblasts to 5-azacytidine (5-azaC), a pharmacological agent inducing changes in gene expression by CpG island demethylation, results in a persistent switch-on of MSF expression. Again, this induced expression is reversible and may be switched off by a subsequent exposure of the activated cells to TGF-β1 when grown on a native collagen substratum (Fig. 11.6). Cytotoxic/carcinogenic agents have been found to induce MSF expression by fibroblasts and carcinoma cells. For example, Yoshino et al. (2007) reported that exposure of a bronchioloalveolar carcinoma cell line to the tobacco carcinogen benzo[a]pyrene results in the induction of MSF expression. These findings raise
11 Multiple Fibroblast Phenotypes in Cancer Patients
211
the possibility that environmental exposure to genotoxins and/or epigenotoxins may induce the local and/or systemic inappropriate persistent switch-on of MSF expression which may then contribute to the development of an overt malignancy. We have also observed that transfection of adult skin fibroblasts with dominantnegative p53 results in the persistent switch-on of MSF-aa expression (unpublished data). Hill et al. (2005) reported that inactivation of tumour suppressor gene pRb in a human prostate carcinoma cell line resulted in the enrichment of mouse fibroblasts harbouring silenced p53 in a xenograft model of cancer progression, presumably as a consequence of selective pressure exerted by malignant prostate cells. These findings suggest that interactions between tumour and stromal cells may lead to the selection of a pre-existing inactivated p53 clonal population of stromal fibroblasts. Finally, the potential clinical relevance of the above findings is highlighted by observations that constitutive MSF expression by fetal and cancer patient fibroblasts may also be persistently switched-off by exposure of cells growing on a native collagen substratum to TGF-β1. These findings suggest that it may be possible to develop novel therapeutic strategies based on the modulation of MSF expression in a clinically desirable fashion by pharmacologic intervention.
11.3.3 M SF: Spectrum of Bioactivities and Their Contextual Control MSF exhibits a number of potent bioactivities. As its name implies, the first to be demonstrated was its stimulation of fibroblast migration into 3D gels of native type I collagen fibres (Schor et al. 1988a), a process apparently mediated by its stimulation of hyaluronan biosynthesis (Schor et al. 1989). These effects on target adult skin fibroblasts displayed a bell-shaped dose response and were exceptionally potent, with half-maximal activity elicited at femtomolar concentrations, i.e. 1 pg/ml (Fig. 11.7) (Schor et al. 2003). Subsequent studies revealed its potent motogenic effect on various other target cell types, including normal and tumour epithelial cells, melanoma cells, endothelial cells, and pericytes (Houard et al. 2005; Hu et al. 2009; migration HA synthesis
relative migration
5
3
4 2
3 2
1
1 0
–2
–1 0 1 2 3 4 concentration MSF (log pg/ml)
5
0
relative HA synthesis
Fig. 11.7 MSF stimulation of cell migration and hyluronan (HA) synthesis by adult skin fibroblasts. The effects of MSF on migration and HA synthesis by confluent adult skin fibroblasts growing on a 3D native type I collagen matrix were determined as previously described (Ellis et al. 1992). Data are expressed relative to control cultures incubated in the absence of MSF
A. M. Schor and S. L. Schor 50
MCF-7 MDA-MB231
40
MDA-MB435
30 20 10
00
0
MSF (ng/ml)
10
10
1
1
10
0.
0.
01
0
0
number of cells migrated
212
Fig. 11.8 The effect of MSF on the migration of breast carcinoma cell lines. The effect of various concentrations of MSF on the migration of three breast cancer lines was tested in the transmembrane assay using membranes coated with native type I collagen. The MSF-7 line displayed a bellshaped dose-response, whereas the two MDA lines displayed a plateau level of stimulation within the concentration range tested
Schor et al. 2005). Certain tumour cell lines exhibit a bell-shaped motogenic dose response to MSF, whereas others display a plateau level in the tested concentration range (Fig. 11.8). MSF induces endothelial cell activation in vitro, as manifest by the formation of a sprouting (angiogenic) phenotype (Schor et al. 2001; unpublished data). It also induces angiogenesis in vivo when implanted subcutaneously in rats, mice and pigs; as well as when evaluated in the chick embryo yolk sac assay (Fig. 11.9). In the latter, different quantities of MSF were incorporated within dried
angiogenic response (%)
100 positive controls PMA
80 60 40 20
negative controls MCG only 00 50 00
0
10
50
0 10
50
5
1
0. 5
0. 1
0. 01
0
concentration (ng MSF/MCG)
Fig. 11.9 The Induction of angiogenesis in the chick yolk sac assay by MSF. Different concentrations of MSF were incorporated into methyl cellulose gels (MCGs) and these were then dried and applied to the chick embryo yolk sac membrane. The induction of a radial disposition of vessels 24 h later was scored as a positive angiogenic response. Results are expressed as the percentage of positive angiogenic responses elicited by various concentrations. A significant response was induced over a broad concentration range of 0.5–500 ng/MCG. PMA (300 ng/MCG) was used as positive control. Upper and lower horizontal bars indicate values achieved by the positive and negative controls (excipient only)
11 Multiple Fibroblast Phenotypes in Cancer Patients
213
methyl cellulose gels (MCG) which were then applied to the yolk sac membrane. The induction of a subjacent radial array of blood vessels 24 h later was taken as a positive angiogenic response. MSF elicited a significant (bell-shaped) angiogenic response over a broad concentration range of 0.5–500 ng/MCG. Homogeneous monolayer cultures of “resting” endothelial cells (mimicking the endothelium lining mature blood vessels in vivo) maintain their phenotype under basal tissue culture conditions, i.e. in the absence of an exogenous angiogenic factor (Schor et al. 2001). Addition of such a factor (including FGF-2, MSF or VEGF) results in the rapid induction of a network of elongated cells displaying a “sprouting” phenotype, thereby forming a mixed culture of both resting and sprouting cells (Schor et al. 1983, 2001). We have demonstrated that sprouting endothelial cells produce bioactive MSF. Significantly, exposure of such mixed cultures to an MSF function-neutralising antibody rapidly results in the apoptotic death of the sprouting cell subpopulation without affecting the viability of the co-cultured resting cells (Fig. 11.10). As discussed below, this apparent dependence of sprouting endothelial cell survival on the maintenance of MSF functionality may provide a novel therapeutic strategy based on the selectively induction of tumour-induced angiogenic vessel involution. The IGD tripeptide motif ( isoleucine-glycine-aspartate) is a highly conserved feature of fibronectin type I modules to which no biological functionality was initially ascribed (Hynes 1990). MSF+aa and MSF-aa both contain four such IGD motifs (Fig. 11.1). In vitro mutagenesis studies indicated that the stimulation of fibroblast migration by MSF is mediated by the two IGD motifs present in type I modules I-7 and I-9 (Schor et al. 2003), whereas all four IGD motifs affect endothelial cell migration (unpublished data). Critically, small IGD-containing synthetic peptides mimic all MSF bioactivities, including the stimulation of cell migration, endothelial cell activation (Schor et al. 1999, 2003) and angiogenesis. It is important to note that none of these bioactivities are manifest by any full-length fibronectin isoform, apparently as a consequence of steric hindrance of their constituent IGD motifs (Millard et al. 2007; Vakonakis et al. 2009). Houard et al. (2005) have identified a second motif (HEEGH) in MSF module I-8 which mediates the
Fig. 11.10 MSF is required for the survival of sprouting, but not resting, endothelial cells. Endothelial cells were cultured on type I collagen-coated dishes and maintained under low serum basal conditions when they reached confluence. a Monolayer of resting confluent endothelial cells in control cultures. b The induction of a sprouting cell network by angiogenic stimulus (such as bovine fetal serum or VEGF). c Selective death of sprouting cells following 1–2 day incubation with MSF function-neutralising antibody
214
A. M. Schor and S. L. Schor native collagen
relative cell migration
3
denatured collagen 2
1
0 HSF
BREC
BRP
Fig. 11.11 Matrix modulation of the motogenic response of target cells to MSF. The effects of the substratum (native or denatured type I collagen) on the motogenic response of human skin fibroblasts (HSF), bovine retinal endothelial cells (BREC) and bovine retinal pericytes (BRP) to 10 pg/ml MSF were evaluated in the transmembrane (Boyden chamber) migration assay. The number of cells migrated in the presence of MSF is expressed relative to baseline (control) migration. BRP and HSF only responded to MSF when attached to a native collagen substratum. In contrast, the migration of BREC was stimulated by MSF irrespective of substratum
motogenic response of a breast cancer cell line (MCF7) and is additionally required for manifestation of proteinase activity. Our recent data indicate that the motogenic response of certain target cells may be mediated by both IGD and HEEGH, whereas other cells only respond to the IGD motif. The functional responses of target cells to MSF (and its IGD synthetic peptide mimetic) are also modulated by contextual parameters. For example, the stimulation of fibroblast migration is manifest by cells adherent to a native, but not denatured, type I collagen substratum (Schor et al. 1999, 2003). Native collagen may accordingly be considered to provide a “permissive” substratum, whereas denatured collagen is “non-permissive”. The motogenic response of pericytes, but not endothelial cells, exhibits a similar matrix-dependence (Fig. 11.11). The stimulation of fibroblast migration by MSF is also abrogated by a variety of soluble factors, including TGF-β and NGAL, an endogenous MSF inhibitor produced by normal “resting” keratinocytes (Ellis et al. 1992; Jones et al. 2007). Interestingly, a distinct MSF inhibitor is produced by resting endothelial cells (unpublished data).
11.3.4 Section Summary Distinctive features of MSF biology are summarised in Fig. 11.12 in which it is noted that (1) MSF may be expressed by several cell types, including epithelial (normal and tumour), fibroblasts and endothelial cells, (2) MSF is a pleiotropic effector, capable of eliciting multiple responses from a variety of target cell types, and (3) the expression of MSF and its precise effect on target cells are modulated by a hierarchy of contextual control networks involving the inter-dependent signalling
11 Multiple Fibroblast Phenotypes in Cancer Patients TARGET CELL
PRODUCER epithelium
fibroblast
215
migration
MSF
migration matrix remodelling migration angiogenesis
endothelium expression of and response to MSF are both regulated by complex interplay of stimulatory and inhibitory factors (soluble and matrix)
Fig. 11.12 Contextual control of MSF expression and effects on potential target cells. MSF may be produced by epithelial cells (normal and tumour), fibroblasts and endothelial cells. These same cell populations are also potential targets of MSF, capable of responding in terms of the stimulation of cell migration, matrix remodelling and angiogenesis. Control of MSF expression and manifestation of its potential bioactivities are both regulated by a complex hierarchy of inter-dependent “contextual” actions of ECM and soluble factors
of soluble factors and the ECM. In this complex framework, MSF may function in both an autocrine and paracrine fashion. The inter-dependent modulatory effects of the ECM and soluble factors on both MSF expression and target cell response are consistent with the now well recognised dynamic and reciprocal contextual control of cell behaviour by multiple tissue-level cues (Bissell and Barcellos-Hoff 1987; Nathan and Sporn 1991; Schor 1994; Schor and Schor 1997; Xu et al. 2009). An important corollary of this understanding is that the temporal and spatial expression of MSF by its various potential producer cells, as well as the precise response (or lack of it) by its potential target cells, is not invariant, but may change in response to alterations to the tissue microenvironment during disease progression.
11.4 Epilogue: Clinical Consequences and Novel Intervention Strategies 11.4.1 Postulated Role of MSF in Cancer Progression Tumour progression is an indolent process in which many decades may elapse between inception of the initiating genetic lesion and the emergence of a clinically recognizable malignancy. Precise information regarding what proportion of initiated cells proceed to develop into an overt malignancy is not available, although published data suggest that this figure may be quite low. For example, Nielsen et al. (1987) noted the presence of microfoci of carcinoma-in-situ in the breasts
216
A. M. Schor and S. L. Schor
of apparently healthy women who died in road traffic accidents; surprisingly, this study revealed that greater than 40% of women over the age of 40 had such histologically discernable lesions, although clearly only a relatively small proportion of these would presumable have developed into an overt malignancy during the lifespan of the individual. These observations suggest that factors which alter the kinetics and severity of disease progression may play an important, and perhaps decisive, role in determining the probability of developing a clinically detectable malignancy. It is in this postulated role of an “accelerator” of cancer progression that we view the potential contribution of MSF-secreting fibroblasts to disease pathogenesis. According to this proposal, we suggest that MSF produced by “fetal-like” fibroblasts (as well as other activated cell types) may contribute to the creation of a milieu (“soil”) which is conducive to the clonal expansion of the evolving population of (pre-)neoplastic cells (“seed”). These seed-soil interactions are likely to involve both permissive and inductive mechanisms, and continue to be influenced by the stochastic accumulation of genetic/epigenetic lesions and environmental agents. The diverse bioactivities of MSF are consistent with its postulated influence on cancer progression. In this regard, we draw specific attention to (1) its capacity to stimulate the migration/invasion of tumour and non-tumour cells, (2) its effects on matrix remodelling, including stimulation of HA synthesis, (3) its induction of angiogenesis, and (4) its possible role as a specific survival factor for angiogenic endothelial cells. The postulated deleterious effect of MSF on disease outcome is supported by our recent data indicating that high MSF expression by tumour and tumour-associated stromal cells is associated with poor survival in patients with breast cancer (Fig. 11.13). A similar inverse association has been observed in patients with oral squamous cell carcinoma (unpublished data). Independent evidence below median above median
Percent survival
100
90
80
70 0
50 100 survival in months
150
Fig. 11.13 Prognostic value of MSF expression in the stroma and of breast cancer patients. Archival sections of breast tumours collected at presentation (n = 75) were stained with an antibody that recognises MSF-aa. MSF staining in the tumour stroma was assessed by image analysis. Overall patient survival was analysed according to the percentage of stromal area stained (divided by the median). Results show Kaplan-Mayer survival curves. High MSF expression was significantly associated with poor survival (log rank test, p = 0.02)
11 Multiple Fibroblast Phenotypes in Cancer Patients
217
supporting a role for MSF in cancer pathogenesis has recently been published by Hu et al. (2009). Using an unbiased proteomic screen, these authors identified MSF as a critical angiogenic factor driving oesophageal cancer progression.
11.4.2 Implications for the Development of Novel Therapies Translational research targeting tumour-stromal interactions are now beginning to be evaluated and promise to yield clinically relevant results. Nakajima et al. (2004) reported that co-culture with fibroblasts reduced the sensitivity of a scirrhous gastric cancer cell line to 5-fluorouracil (a commonly used chemotherapeutic agent) and, more significantly, this protective effect of the stromal cells could be reversed by a pharmacologic agent (tranilast). With particular reference to MSF, our findings suggest that it may be possible to improve disease outcome by developing novel strategies variously designed to abrogate manifestation of its diverse functionality (such as by MSF-specific neutralising antibodies or small molecule inhibitors) and/or switch-off its inappropriate expression (such as by siRNA or epigenetic manipulation) (Aharinejad et al. 2009). The possible ability of such interventions to result in the specific regression of tumour-induced angiogenic blood vessels is a potentially exciting option. Other clinical applications of MSF may include population screening for individuals at elevated risk of developing cancer (by measuring its bioactivity in serum). Finally, quantitative assessment of MSF expression, either in serum and/or by ex-vivo immunohistological examination of the presenting tumour, may contribute to compiling a molecular profile of the individual cancer, thereby assisting in patient stratification and prognostic assessment. Acknowledgements Work from our laboratory presented in this review has been funded by Cancer Research UK, Breast Cancer Campaign, Scottish Enterprise Proof of Concept Programme, Engineering and Physical Sciences Research Council and Biotechnology and Biological Sciences Research Council. We thank the many collaborators that have contributed to this work by providing specimens and experimental data.
References Abrahams PJ, Houweling A, Cornelissen-Steijger PD, Jaspers NG, Darroudi F, Meijers CM, Mullenders LH, Filon R, Arwert F, Pinedo HM, Natarajan AP, Terleth C, Van Zeeland AA, van der Eb AJ (1998) Impaired DNA repair capacity in skin fibroblasts from various hereditary cancerprone syndromes. Mutat Res 407:189–201 Aharinejad S, Sioud M, Lucas T, Abraham D (2009) Targeting stromal-cancer cell interactions with siRNAs. Methods Mol Biol 487:243–266 Anton K, Glod J (2009) Targeting the tumor stroma in cancer therapy. Curr Pharm Biotechnol 10:185–191 Atula S, Grenman R, Syrjänen S (1997) Fibroblasts can modulate the phenotype of malignant epithelial cells in vitro. Exp Cell Res 235:180–187
218
A. M. Schor and S. L. Schor
Azzarone B, Macieira-Coelho A (1987) Further characterization of the defects of skin fibroblasts from cancer patients. J Cell Sci 87:155–162 Azzarone B, Mareel M, Billard C, Scemana P, Chaponnier C, Marciera-Coelho M (1984) Abnormal properties of skin fibroblasts from patients with breast cancer. Int J Cancer 33:759–764 Azzarone B, Chaponnier C, Krief P, Mareel M, Suarez H, Macieira-Coelho A (1988) Human fibroblasts from cancer patients: lifespan and transformed phenotype in vitro and role of mesenchyme in vivo. Mutat Res 199:313–325 Bakheet T, Frevel M, Williams BRG, Greer W, Khabar KSA (2001) ARED: human AU-rich element-containing mRNA database reveals an unexpectedly diverse functional repertoire of encoded proteins. Nuc Acids Res 29:246–254 Bartal AH, Lichtig C, Cardo CC, Feit C, Robinson E, Hirshaut Y (1986) Monoclonal antibody defining fibroblasts appearing in fetal and neoplastic tissue. J Natl Cancer Inst 76:415–419 Basset P, Bellocq JP, Wolf C, Stoll I, Hutin P, Limacher JM, Podhajcer OL, Chenard MP, Rio MC, Chambon P (1990) A novel metalloproteinase gene specifically expressed in stromal cells of breast carcinomas. Nature 348:699–704 Bayreuther K, Rodemann HP, Hommel R, Dittman K, Albiez M, Francz PI (1988) Human skin fibroblasts in vitro differentiate along a terminal cell lineage. Proc Natl Acad Sci U S A 85:5112–5116 Bierie B, Moses HL (2006) TGF-beta and cancer. Cytokine Growth Factor Rev 17:29–40 Bissell MJ, Barcellos-Hoff MH (1987) The influence of extracellular matrix on gene expression: is structure the message? J Cell Sci 8(Suppl 10):327–343 Broxterman HJ, Georgopapadakou NH (2007) Anticancer therapeutics: a surge of new developments increasingly target tumor and stroma. Drug Resist Update 10:182–193 Camps JL, Chang SM, Hsu TC, Freeman MR, Hong SJ, Zhau HE, von Eschenbach AC, Chung LW (1990) Fibroblast mediated acceleration of human epithelial tumor growth in vitro. Proc Nat Acad Sci U S A 87:75–79 Chiquet-Ehrismann R, Mackie EJ, Pearson CA, Sakaura T (1986) Tenascin: an extracellular matrix protein involved in tissue interactions during fetal development and oncogenesis. Cell 47:131–139 Cornil I, Theodorescu D, Man S, Herlyn M, Jambrosic J, Kerbel RS (1991) Fibroblast cell interactions with human melanoma cells affect tumor cell growth as a function of tumor progression. Proc Natl Acad Sci U S A 88:6028–6032 Cullen KJ, Smith HS, Hill S, Rosen N, Lippman ME (1991) Growth factor messenger RNA expression by human fibroblasts from benign and malignant lesions. Cancer Res 51:4978–4985 Cunha GR, Bigsby RM, Cooke PS, Sugimara Y (1985) Stromal-epithelial interactions in adult organs. Cell Differ 17:137–148 Dabbous MK, Haney L, Carter LM, Paul AK, Reger J (1987) Heterogeneity of fibroblast response in host-tumor cell-cell interactions in metastatic tumors. J Cell Biochem 35:333–344 Danes BS, De Angeles P, Traganos F, Ringborg U, Nielsen LH, Melamed MR (1990) Comparsion of anatomical location of squamous cell carcinoma within the oral cavity and oropharynx with the incidence of in vitro hyperdiploidly. Clin Genet 37:188–193 Dong Z, Nemeth JA, Cher ML, Palmer KC, Bright RC, Fridman R (2001) Differential regulation of matrix metalloproteinase-9, tissue inhibitor of metalloproteinase-1 (TIMP-1) and TIMP-2 expression in co-cultures of prostate cancer and stromal cells. Int J Cancer 93:507–515 Durning P, Schor SL, Sellwood RAS (1984) Fibroblasts from patients with breast cancer show abnormal migratory behaviour in vitro. Lancet 2:890–892 Ellis I, Grey AM, Schor AM, Schor SL (1992) Antagonistic effects of TGF-β1 and MSF on fibroblast migration and hyaluronic acid synthesis—possible implications for dermal wound healing. J Cell Sci 102:447–456 Fabra A, Nakajima M, Bucana CD, Fidler IJ (1992) Modulation of the invasive phenotype of human colon carcinoma cells by organ specific fibroblasts of nude mice. Differentiation 52:101– 110 Frankel JW, Bidot P, Kopelovich L (1986) Enhanced sensitivity of skin fibroblasts from neurofibromatosis patients to transformation by the Kirsten murine sarcoma virus. A potential laboratory assay for individuals at risk of cancer. Ann N Y Acad Sci 486:403–412
11 Multiple Fibroblast Phenotypes in Cancer Patients
219
Garin-Chesa P, Old LJ, Rettig WJ (1990) Cell surface glycoprotein of reactive stromal fibroblasts as a potential target in human epithelial cancers. Proc Nat Acad Sci U S A 87:7235–7239 Gärtner MFRM, Wilson EL, Dowdle EB (1992) Fibroblast-dependent tumorigenicity of melanoma xenografts in athymic mice. Int J Cancer 51:788–791 Gurdon JB (1988) A community effect in animal development. Nature 336:772–774 Guturu P, Shah V, Urrutia R (2009) Interplay of tumor microenvironment cell types with parenchymal cells in pancreatic cancer development and therapeutic implications. J Gastrointest Cancer 40:1–9 Haggie J, Schor SL, Howell A, Birch JM, Sellwood RAS (1987) Fibroblasts from relatives of hereditary breast cancer display foetal-like behaviour in vitro. Lancet 1:1455–1457 He Y, Liu XD, Chen ZY, Zhu J, Xiong Y, Li K, Dong JH, Li X (2007) Interaction between cancer cells and stromal fibroblasts is required for activation of the uPAR-uPA-MMP-2 cascade in pancreatic cancer metastasis. Clin Cancer Res 13:3115–3124 Higgins PJ, Kopelovich L (1991) Analysis of actin microfilaments and cell-to-substrate adhesive structures in human fibroblasts from individuals genetically predisposed to colonic carcinoma. Exp Cell Res 195:395–400 Hill R, Song Y, Cardiff RD, Van Dyke T (2005) Selective evolution of stromal mesenchyme with p53 loss in response to epithelial tumorigenesis. Cell 123:1001–1011 Houard X, Germain S, Gervais M, Michaud A, van den Brule F, Foidart J-M, Noel A, Monnot C, Corvol P (2005) Migration-stimulating factor displays HEXXH-dependent catalytic activity important for promoting tumor cell migration. Int J Cancer 116:378–384 Hu H, Ran Y, Zhang Y, Zhou Z, Harris SJ, Yu L, Sun L, Pan J, Liu J, Lou J, Yang Z (2009) Antibody library-based tumor endothelial cells surface proteomic functional screen reveals migration-stimulating factor as an anti-angiogenic target. Mol Cell Proteomics 8:816–826 Hynes R (1990) Fibronectins. Springer, New York Irwin CR, Picardo M, Ellis I, Sloan P, Grey AM, McGurk M, Schor SL (1994) Inter- and intrasite heterogeneity in the expression of fetal-like phenotypic characteristics by gingival fibroblasts: potential significance for wound healing. J Cell Sci 107:1333–1346 Jones SJ, Florence MM, Ellis IR, Kankova K, Schor SL, Schor AM (2007) Co-expression by keratinocytes of migration stimulating factor (MSF) and functional inhibitor of its bioactivity (MSFI). Exp Cell Res 313:4145–4157 Kay RA, Ellis IR, Jones SJ, Perrier S, Florence MM, Schor AM, Schor SL (2005) The expression of MSF, a potent oncofetal cytokine, is uniquely controlled by 3′-untranslated region dependent nuclear sequestration of its precursor messenger RNA. Cancer Res 65:10742–10749 Kiaris H, Trimis G, Papavassiliou AG (2008) Regulation of tumor-stromal fibroblast interactions: implications in anticancer therapy. Curr Med Chem 15:3062–3067 Knudson AG (2001) Two genetic hits (more or less) to cancer. Nat Rev Cancer 1:157–162 Kopelovich L (1987) Tissue culture assays in familial polyposis coli: observations and considerations. Semin Surg Oncol 3:159–164 Kopelovich L (1988) The transformed (initiated) human cell phenotype: studies on cultured skin fibroblasts from individuals predisposed to cancer. Mutat Res 199:369–385 Kopelovich L, Bias NE (1980) Tumor promoter induces loss of anchorage dependence in human skin fibroblasts from individuals genetically predisposed to cancer. Exp Cell Biol 48:207–217 Kopelovich L, Lipkin M, Blattner WA, Fraumeni JF Jr (1980) Organization of actin-containing cables in cultured skin fibroblasts from individuals at high risk of colon cancer. Int J Cancer 26:301–307 Koshida Y, Kuranami M, Watanabe M (2006) Interaction between stromal fibroblasts and colorectal cancer cells in the expression of vascular endothelial growth factor. J Surg Res 134:270–277 Liu X, Zhao Q, Collodi P (2003) A truncated form of fibronectin is expressed in fish and mammals. Matrix Biol 22:393–396 Lu C, Vickers MF, Kerbel RS (1992) Interleukin 6: a fibroblast-derived growth inhibitor of human melanoma cells from early but not advanced stages of tumor progression. Proc Nat Acad Sci U S A 89:9215–9219 Maeda T, Desouky J, Friedl A (2006) Syndecan-1 expression by stromal fibroblasts promotes breast carcinoma growth in vivo and stimulates tumor angiogenesis. Oncogene 25:1408–1412
220
A. M. Schor and S. L. Schor
Matsumoto K, Horikoshi M, Rikimaru K, Enomoto S (1989) A study of an in vitro model of squamous cell carcinoma. J Oral Pathol Med 18:498–501 McCallion RL, Ferguson MWJ (1996) Fetal wound healing and the development of antiscarring therapies for adult wound healing. In: Clark RAF (ed) The molecular and cellular biology of wound repair, 2nd edn. Plenum Press, New York, pp 561–600 McNeal JE (1984) Anatomy of the prostate and morphogenesis of BPH. In: Kimbal FA, Buhl AE, Carter DB (eds) New approaches to the study of benign prostatic hyperplasia. A.R. Liss, New York, pp 27–53 Millard CJ, Ellis IR, Pickford AR, Schor AM, Schor SL, Campbell ID (2007) The role of fibronectin IGD motif in stimulating fibroblast migration. J Biol Chem 282:35530–35535 Mintz B, Illmensee K (1975) Normal genetically mosaic mice produced from malignant teratocarcinoma cells. Proc Natl Acad Sci U S A 72:3585–3589 Mukaida H, Hirabayashi N, Hirai T, Iwata T, Saeki S, Toge T (1991) Significance of freshly cultured fibroblasts from different tissues in promoting cancer cell growth. Int J Cancer 48:423–427 Nakajima K, Okita Y, Matsuda S (2004) Sensitivity of scirrhous gastric cancer to 5-fluorouracil and the role of cancer cell-stromal fibroblast interaction. Oncol Rep 12:85–90 Nakano S, Ts’O PO (1981) Cellular differentiation and neoplasia: characterization of subpopulations of cells that have neoplasia related growth properties in Syrian hamster embryo cell cultures. Proc Natl Acad Sci U S A 78:4995–4999 Nathan C, Sporn M (1991) Cytokines in context. J Cell Biol 113:981–986 Nicolo G, Salvi S, Oliveri G, Borsi L, Castellani P, Zardi L (1990) Expression of tenascin and the ED-B containing oncofetal fibronectin isoform in human cancer. Cell Differ Dev 32:401–408 Nielsen M, Thomsen JL, Primdahl S, Dyreborg U, Andersen JA (1987) Breast cancer and atypia among young and middle-aged women. Br J Cancer 56:814–819 Nowell PC (1976) The clonal evolution of tumor cell populations. Science 194:23–28 Padua D, Massague J (2009) Role of TGF-β in metastasis. Cell Res 19:89–102 Pardali E, ten Dijke P (2009) Transforming growth factor-beta signaling and tumor angiogenesis. Front Biosci 14:4848–4861 Picard O, Rolland Y, Poupon MF (1986) Fibroblast-dependent tumorigenicity of cells in nude mice: implication for implantation and metastasis. Cancer Res 46:3290–3294 Picardo M, Schor SL, Grey AM, Howell A, Laidlaw I, Redford J, Schor AM (1991) Migration stimulating activity in serum of breast cancer patients. Lancet 337:130–133 Pietras K, Pahler J, Bergers G, Hanahan D (2008) Functions of paracrine PDGF signaling in the proangiogenic tumor stroma revealed by pharmacological targeting. PLoS Med 5:e19 Rasheed S, Gardner MB (1981) Growth properties and susceptibility to viral transformation of skin fibroblasts from individuals at high genetic risk for colorectal cancer. J Natl Cancer Inst 66:43–49 Rhim JS, Arnstein P, Huebner RJ (1981) Chemical transformation of cultured skin fibroblasts from humans genetically predisposed to cancer. Cancer Detect Prev 4:239–247 Rozenchan PB, Carraro DM, Brentani H, de Carvalho Mota LD, Bastos EP, Ferreira EN, Torres CH, Katayama ML, Roela RA, Lyra EC, Soares FA, Folgueira MA, Góes JC, Brentani MM (2009) Reciprocal changes in gene expression profiles of cocultured breast epithelial cells and primary fibroblasts. Int J Cancer 125:2767–2777 Sappino AP, Schurch W, Gabbiani G (1990) Differentiation repertoire of fibroblastic cells: expression of cytoskeletal proteins as marker of phenotypic modulations. Lab Invest 63:144–161 Schor AM, Schor SL, Allen TD (1983) The effects of culture conditions on the proliferation and morphology of bovine aortic endothelial cells in vitro: reversible expression of the sprouting cell phenotype. J Cell Sci 62:267–285 Schor AM, Rushton G, Ferguson JE, Howell A, Redford J, Schor SL (1994) Phenotypic heterogeneity in breast fibroblasts—functional anomaly in fibroblasts from histologically normal tissue adjacent to carcinoma. Int J Cancer 59:25–32 Schor AM, Ellis I, Schor SL (2001) Collagen gel assay for angiogenesis. Induction of endothelial cell sprouting. In: Murray JC (ed) Methods in molecular medicine, vol 46: angiogenesis protocols. Humana Press, Totowa, pp 145–162
11 Multiple Fibroblast Phenotypes in Cancer Patients
221
Schor SL (1994) Cytokine control of cell motility: modulation and mediation by the extracellular matrix. Prog Growth Factor Res 5:223–248 Schor SL, Schor AM (1987) Clonal heterogeneity in fibroblasts: implications for the control of epithelial-mesenchymal interactions. BioEssays 7:200–204 Schor SL, Schor AM (1997) Stromal acceleration of tumour progression: role of fetal-like fibroblast subpopulations. Pathol Update 4:75–95 Schor SL, Schor AM (2001) Tumour-stroma interactions—phenotypic and genetic alterations in mammary stroma: implications for tumour progression. Breast Cancer Res 3:373–379 Schor SL, Schor AM, Rushton G, Smith L (1985a) Adult, fetal and transformed fibroblasts display different migratory phenotypes on collagen gels: evidence for an isoformic transition during fetal development. J Cell Sci 73:221–234 Schor SL, Schor AM, Durning P, Rushton G (1985b) Skin fibroblasts obtained from cancer patients display a fetal-like migratory behaviour on collagen gels. J Cell Sci 73:235–244 Schor SL, Haggie J, Durning P, Howell A, Sellwood RAS, Crowther D (1986) The occurrence of a foetal fibroblast phenotype in familial breast cancer. Int J Cancer 37:831–836 Schor SL, Schor AM, Howell A, Crowther D (1987) Hypothesis: persistent expression of fetal phenotypic characteristics by fibroblasts is associated with an increased susceptibility to neoplastic disease. Exp Cell Biol 55:11–17 Schor SL, Schor AM, Grey AM, Rushton G (1988a) Fetal and cancer patient fibroblasts produce an autocrine migration-stimulating factor not made by normal adult cells. J Cell Sci 90:391–399 Schor SL, Schor AM, Rushton G (1988b) Fibroblasts from cancer patients display a mixture of both fetal and adult-like phenotypic characteristics. J Cell Sci 90:401–407 Schor SL, Schor AM, Grey AM, Chen J, Rushton G, Grant ME, Ellis I (1989) Mechanism of action of the migration stimulating factor produced by fetal and cancer-patient fibroblasts: effect on hyaluronic acid synthesis. In Vitro Cell Dev Biol 25:737–746 Schor SL, Ellis I, Banyard J, Dolman C, Seneviratne K, Gilbert AD, Chisholm DM (1996) Fetallike phenotypic characteristics of gingival fibroblasts: potential relevance to wound healing. Oral Dis 2:155–166 Schor SL, Ellis I, Banyard J, Schor AM (1999) Motogenic activity of the IGD amino acid motif. J Cell Sci 112:3879–3888 Schor SL, Ellis IR, Jones SJ, Baillie R, Seneviratne K, Clausen J, Motegi K, Vojtesek B, Kankova K, Furrie E, Sales MJ, Schor AM, Kay R (2003) Migration-stimulating factor: a genetically truncated onco-fetal fibronectin isoform expressed by carcinoma and tumor-associated stromal cells. Cancer Res 63:8827–8836 Schor SL, Schor AM, Keatch RP, Belch JFF (2005) Role of matrix macromolecules in the aetiology and treatment of chronic ulcers. In: Lee BY (ed) The wound management manual. McGraw Hill, New York, pp 109–121 Shekhar MP, Werdell J, Santner SJ, Pauley RJ, Tait L (2001) Breast stroma plays a dominant regulatory role in breast epithelial growth and differentiation: implications for tumor development and progression. Cancer Res 61:1320–1326 Skobe M, Fusenig NE (1998) Tumorigenic conversion of immortal human keratinocytes through stromal cell activation. Proc Natl Acad Sci U S A 95:1050–1055 Sloan P, Schor SL, Lopes V (1990) Immunohistochemical study of the heterogeneity of tenascin distribution within the oral mucosa of the mouse. Arch Oral Biol 35:67–70 Stoker AW, Hatier C, Bissell MJ (1990) The embryonic environment strongly attenuates v-src oncogenesis in mesenchymal and epithelial tissues, but not in endothelia. J Cell Biol 111:217– 228 Sugimoto T, Takiguchi Y, Kurosu K, Kasahara Y, Tanabe N, Tatsumi K, Hiroshima K, Minamihisamatsu M, Miyamoto T, Kuriyama T (2005) Growth factor-mediated interaction between tumor cells and stromal fibroblasts in an experimental model of human small-cell lung cancer. Oncol Rep 14:823–830 Sung SY, Chung LW (2002) Prostate tumor-stroma interaction: molecular mechanisms and opportunities for therapeutic targeting. Differentiation 70:506–521
222
A. M. Schor and S. L. Schor
Svendsen LB, Thorup J, Larsen JK, Norgard T, Willumsen H, Hansen OH (1989) Association between tumor DNA aneuploidy and in vitro tetraploidy of skin fibroblasts in patients with colorectal neoplasms. Scand J Gastroenterol 24:755–760 Tanaka H, Mori Y, Ishii H, Akedo H (1988) Enhancement of metastatic capacity of fibroblasttumor cell interaction in mice. Cancer Res 48:1456–1459 Tarin D (1972) Tissue interactions in carcinogenesis. Academic, London Thielmann HW, Edler L, Burkhardt MR, Jung EG (1987) DNA repair synthesis in fibroblast strains from patients with actinic keratosis, squamous cell carcinoma, basal cell carcinoma, or malignant melanoma after treatment with ultraviolet light, N-acetoxy-2-acetyl-aminofluorene, methyl methanesulfonate and N-methyl-N-nitrosurea. J Cancer Res Clin Oncol 113:171–186 Tsellou E, Kiaris H (2008) Fibroblast independency in tumors: implications in cancer therapy. Future Oncol 4:427–432 Vakonakis I, Staunton D, Ellis IR, Starkies P, Flanagan A, Schor AM, Schor SL, Campbell ID (2009) Motogenic sites in human fibronectin are masked by long range interactions. J Biol Chem 284:15668–15675 Wong YC, Wang YZ (2000) Growth factors and epithelial-stromal interactions in prostate cancer development. Int Rev Cytol 199:65–116 Wynford-Thomas D, Smith P, Williams ED (1986) Prolongation of fibroblast lifespan associated with epithelial rat tumor development. Cancer Res 46:3125–3127 Xu R, Boudreau A, Bissell MJ (2009) Tissue architecture and function: dynamic reciprocity via extra- and intra-cellular matrices. Cancer Metastasis Rev 28:167–176 Yoshino I, Kometani T, Shoji F, Osoegawa A, Ohba T, Kouso H, Takenaka T, Yohena T, Maehara Y (2007) Induction of epithelial-mesenchymal transition-related genes by benzo[a]pyrene in lung cancer cells. Cancer 110:369–374 Zhao Q, Liu X, Collodi P (2001) Identification and characterization of a novel fibronectin in Zebrafish. Exp Cell Res 268:211–221
Chapter 12
TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression in Adjacent Epithelia Brian R. Bierie and Harold L. Moses
12.1 Introduction TGF-β is an important regulator of carcinoma initiation, progression and metastasis. Over the past three decades, much of the research related to TGF-β signaling has been directed toward cell autonomous effects of stimulation. However, it is now known that TGF-β signaling regulates intrinsic cell autonomous signal transduction in addition to cross-talk between adjacent cell populations. The latter effect of TGF-β signaling in the tumor microenvironment has been elevated in priority with regard to investigating paracrine cross-talk that may targeted to manage human carcinoma recurrence and improve overall survival. At present several regulatory mechanisms have been identified in association with stromal fibroblast responses to TGF-β that can regulate adjacent epithelial tumor initiation, progression and metastasis. TGF-β can suppress the production of tumor promoting paracrine signals including HGF, Mst-1, TGF-α, WNT-2, WNT3A and WNT5A. When TGF-β signaling was lost in fibroblasts, which has been shown to occur during carcinoma progression, these paracrine ligands may be increased and thereby contribute to adjacent carcinoma progression. Conversely, TGF-β can cause a fibroblast to myofibroblast transition that has also been associated with adjacent carcinoma progression. Further, TGF-β production by fibroblasts has been shown to increase the sensitivity of carcinoma cells to signals such as SDF-1 that is abundantly expressed by carcinoma associated fibroblasts. At present, the literature suggests that the TGF-β response by fibroblasts and fibroblast production of TGF-β ligands can suppress or promote adjacent carcinoma progression depending upon the context of stimulation.
B. R. Bierie () Whitehead Institute for Biomedical Research, Nine Cambridge Center, Rm 309, Cambridge, MA 02142, USA e-mail:
[email protected] M. M. Mueller, N. E. Fusenig (eds.), Tumor-Associated Fibroblasts and their Matrix, The Tumor Microenvironment 4, DOI 10.1007/978-94-007-0659-0_12, © Springer Science+Business Media B.V. 2011
223
224
B. R. Bierie and H. L. Moses
12.2 TGF-β Signaling and Regulation in Cancer The transforming growth factor beta (TGF-β) ligands TGF-β1, TGF-β2 and TGF-β3 are potent regulators of cell behavior during carcinoma initiation, progression and metastasis (Feng and Derynck 2005; Derynck and Zhang 2003; Massague 2008; Moustakas and Heldin 2007; Lee et al. 2007; Yamashita et al. 2008). Carcinomas, which arise from epithelial cell populations, have been the predominant focus for much of the literature related to TGF-β signaling in this context but it is now evident that stromal-epithelial interactions regulated by TGF-β also significantly contribute to carcinoma progression. TGF-β ligands when expressed are secreted into the extra-cellular matrix where they remain as inactive complexes until subsequent activation (Stover et al. 2007). The ligands may be released from inactive complexes through interactions involving αvβ6 integrin, calpain, cathepsin D, chymase, elastase, endoglycosidase F, kallikrein, matrix metalloproteinase 9 (MMP-9), neuraminidase, plasmin and thrombospondin-1 (TSP1) (Abe et al. 1998; Munger et al. 1999; Lyons et al. 1988; Akita et al. 2002; Taipale et al. 1995; Miyazono and Heldin 1989; Yu and Stamenkovic 2000; Schultz-Cherry and Hinshaw 1996; SchultzCherry and Murphy-Ullrich 1993). In addition, ionizing radiation and reactive oxygen free radicals are known to activate latent TGF-β complexes (Barcellos-Hoff and Dix 1996; Jobling et al. 2006). Once activated, TGF-β1 and TGF-β3 are able to efficiently bind the Type II TGF-β receptor (TβRII) which results in recruitment and transactivation of the Type I TGF-β receptor (TβRI) (Derynck and Zhang 2003; Shi and Massague 2003). Alternatively, TGF-β2 requires the Type III TGF-β receptor (betaglycan) to efficiently bind TβRII and transactivate TβRI. Most of the signaling associated with TGF-β stimulation occurs via a glycine and serine rich region of the TβRI termed the GS domain (Derynck and Zhang 2003). The ligand bound receptor complex can initiate downstream SMAD dependent and SMAD independent pathways (Derynck and Zhang 2003; Shi and Massague 2003) (Fig. 12.1). The SMADs are transcription factors, and TGF-β has been shown to potently stimulate the phosphorylation of SMAD2 and SMAD3 that are often referred to as R-SMADs (Shi and Massague 2003). Once activated the R-SMADs change confirmation enabling them to homo- or hetero-dimerize with SMAD4 to mediate downstream signaling via co-activation or co-repression of transcription. The activated SMAD hetero- and homodimers form complexes in the cytoplasm and subsequently shuttle into the nucleus. The shuttling is mediated by intrinsic nuclear localization signals. SMAD4 nuclear localization has been associated with importin-alpha binding and SMAD3 is able to associate with importin-beta to mediate nuclear import (Brown et al. 2007; Massague et al. 2005). In addition, both SMAD2/3 can associate with CAN/ NUP214 and NUP153 nuclear pore proteins to mediate nuclear import (Brown et al. 2007; Massague et al. 2005). Smad3 and Smad4 are capable of bind DNA once they have entered the nucleus while SMAD2 is not able to directly bind DNA due to a 30 amino acid insertion in the MH1 domain (Brown et al. 2007). The SMAD complexes are thought to remain associated while actively regulating transcription. Once the Smads are inactivated and no longer part of a regulatory transcription complex
12 TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression
225
-β TGF-β
TβRII
TβRI
SMAD Signaling SMAD2, SMAD3, SMAD1, SMAD5
SMAD7, STRAP, YAP65, SMURF1/2, GADD34, PP1
Degradation Competitive Inhibition De-phosphorylation
Association with SMAD4 Nuclear Import
Co-regulation of Transcription
SKI SKIL
SMAD-Independent Signaling ShcA, RHO, RAC/CDC42, RAS, TRAF6, TAK1, PI3K, PAR6, MAP3K1, DAXX, PP2A
Fig. 12.1 TGF-β ligands when expressed are secreted into the extra-cellular matrix where they remain as inactive complexes until subsequent activation through interactions involving αvβ6 integrin, calpain, cathepsin D, chymase, elastase, endoglycosidase F, kallikrein, matrix metalloproteinase 9 (MMP-9), neuraminidase, plasmin and thrombospondin-1. Ionizing radiation and reactive oxygen free radicals are known to activate latent TGF-β complexes. Once activated, TGF-β can bind the Type II TGF-β receptor (TβRII) which results in recruitment and transactivation of the Type I TGF-β receptor (TβRI). The ligand bound receptor complex can initiate downstream SMAD dependent and SMAD independent pathways. The SMADs are transcription factors that mediate downstream signaling via co-activation or co-repression of transcription. Non-canonical SMAD dependent signaling is known to include activation of ShcA, RHO, RAC/CDC42, RAS, TRAF6, TAK1, PI3K, PAR6, MAP3K1, DAXX and PP2A signaling. Negative regulators of TGF-β signaling are known to include SMAD7, STRAP, YAP65, SMURF1/2, GADD34, PP1, SKI and SKIL. Together, the balance between SMAD dependent and independent signaling determines the impact of TGF-β stimulation in vitro and in vivo
they exit the nucleus as monomers (Brown et al. 2007). SMAD4 has a nuclear export signal (NES) that permits association with CRM1 to facilitate export (Brown et al. 2007). Alternatively, SMAD2 and SMAD3 are exported through a CRM1 independent mechanism. The inactive SMAD2 and SMAD3 monomers bind CAN/ NUP214 and NUP153 that are thought to mediate their export (Brown et al. 2007). Although SMAD2 and SMAD3 signaling has been studied in great detail, emerging evidence suggests that in some contexts SMAD1 and SMAD5 (also R-SMADs) can be activated by TGF-β in epithelial cells and fibroblasts (Daly et al. 2008; Liu et al. 2009). SMAD1 and SMAD5 have been traditionally considered bone morphogenic
226
B. R. Bierie and H. L. Moses
protein (BMP) dependent SMADs, however recent evidence supports direct activation of these SMAD family members in response to TGF-β stimulation (Daly et al. 2008; Liu et al. 2009). Further, the activation of SMAD1 and SMAD5 which also complex with SMAD4 to regulate transcription, may be clinically relevant as they have been functionally linked to the regulation of anchorage independent growth in response to TGF-β stimulation (Daly et al. 2008; Liu et al. 2009). Canonical SMAD activation is a significant component of TGF-β signaling, however the SMADs also have important non-canonical roles that significantly regulate cell behavior and global gene expression as illustrated by the non-canonical SMAD dependent regulation of micro-RNA processing (Davis et al. 2008). In addition to canonical and non-canonical SMAD dependent signaling, the SMAD independent pathways are also important for the regulation of tumor initiation, progression and metastasis. The SMAD independent pathways are known to include ShcA, RHO, RAC/ CDC42, RAS, TRAF6, TAK1, PI3K, PAR6, MAP3K1, DAXX and PP2A (Feng and Derynck 2005; Derynck and Zhang 2003; Massague 2008; Moustakas and Heldin 2007; Lee et al. 2007; Yamashita et al. 2008). Together, the balance between SMAD dependent and independent pathways determines the impact of TGF-β stimulation in vitro and in vivo (Fig. 12.1). In human cancer the role for TGF-β has been well characterized with regard to carcinoma initiation, progression and metastasis. TGFB1 overexpression has been demonstrated in human carcinomas including those that occur in the breast, colon, esophagus, stomach, lung, pancreas and prostate tissue (Levy and Hill 2006). However, within the carcinoma cells TGF-β signaling is often attenuated or completely abrogated during tumor progression as a result of TGFBR1, TGFBR2, SMAD2 and SMAD4 loss, mutation or attenuation of expression (Massague 2008; Levy and Hill 2006; Bierie and Moses 2006a). TGFBR2 mutations are frequently observed in micro-satellite instable (MSI+) carcinomas (Grady and Markowitz 2002; Grady et al. 1999; Markowitz et al. 1995). In MSI+ carcinoma cells, mis-match repair defects are common and as a result a 10 bp poly-adenine repeat region (Poly(A)10 tract) from the coding sequence of TGFBR2 are frequently observed. MSI associated Poly(A)10 tract mutations in TGFBR2 have been identified in biliary, colon, gastric, glioma, lung and pancreatic cancers (Massague 2008; Levy and Hill 2006; Bierie and Moses 2006a). The Poly(A)10 tract mutations often lead to frameshift mis-sense mutations or an early termination that prevents translation of a functional TβRII protein. In addition to Poly(A)10 tract mutations in MSI+ carcinomas, intragenic mutation, downregulation and loss of TGFBR2 expression has been observed in bladder, breast, colon, esophageal, lung, ovarian, pancreatic and prostate cancers (Massague 2008; Levy and Hill 2006; Bierie and Moses 2006a). Loss of expression, downregulation and mutation has been observed in TGFBR1 from biliary, bladder, breast, gastric, liver, ovarian, pancreatic and prostate cancers (Massague 2008; Levy and Hill 2006; Bierie and Moses 2006a). The SMADs are central mediators of TGF-β signaling and they are also subject to mis-regulation in human cancer. SMAD4 mutation, deletion and loss of expression is known to be associated with biliary, bladder, breast, cervical, colon esophageal, intestine, liver, lung, ovarian and pancreatic cancers (Massague 2008; Levy and Hill 2006; Bierie and Moses 2006a).
12 TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression
227
SMAD2 mutation and deletion has been observed in cervical, colon, liver and lung cancer (Massague 2008; Levy and Hill 2006; Bierie and Moses 2006a). In contrast with the other central TGF-β signaling components, SMAD3 is often maintained in human carcinomas suggesting that it may have a non-canonical role in the carcinoma cell that favors tumor progression (Levy and Hill 2006; Bierie and Moses 2006a; Sjoblom et al. 2006). Importantly, genetic mutation and loss of heterozygosity are not the only mechanisms known to regulate central components of the TGF-β signaling pathway during human carcinoma progression. Epigenetic alterations have been shown to silence the expression of TSP1, TGFB2, TGFBR1, TGFBR2 and SMAD4 (Kang et al. 1999; Kim et al. 2000; Shipitsin et al. 2007; Rojas et al. 2008; Aitchison et al. 2008; Hinshelwood et al. 2007). Alternatively, attenuation of TGF-β signaling can involve the expression of an inhibitory SMAD family member, SMAD7, that is known to be amplified in some human cancers (Levy and Hill 2006). SMAD7, when expressed can bind TβRI and inhibit downstream signaling (Nakao et al. 1997; Hayashi et al. 1997). SMAD7 bound to TβRI can inhibit activation of SMAD2 and SMAD3 due to competitive inhibition of the common active site (Inoue and Imamura 2008). SMAD7 also has the ability associate with GADD34 (growth arrest and DNA damage protein 34) to enable de-phosphorylation of TβRI by PP1 (protein phosphatase 1) (Shi et al. 2004). SMAD7 can associate with the SMAD ubiquitin regulatory factor (SMURF) E3 ubiquitin ligase proteins, SMURF1 and SMURF2 to promote ubiquitylation of the activated receptor complex and thereby target it for proteasomal degradation (Kavsak et al. 2000; Ebisawa et al. 2001). Importantly, SMAD7 may be enhanced by expression of other cofactors such as YAP65 and STRAP proteins that bind promote SMAD7 association with the receptor complex (Datta and Moses 2000; Ferrigno et al. 2002). In addition to SMAD7 expression, the SKI and SKIL (Ski-like; SnoN) proto-oncogenes that are known to be upregulated during human carcinoma progression are transcriptional co-factors that are known to attenuate SMAD activity (Levy and Hill 2006; Stroschein et al. 1999; Luo et al. 1999) (Fig. 12.1). TGF-β production within the tumor microenvironment is thought to have both positive and negative influences on carcinoma cell behavior. TGF-β is able to inhibit epithelial cell cycle progression or enhance proliferation depending on the context of stimulation. In general, non-transformed epithelial cells will respond to TGF-β with growth inhibition associated with activation of the canonical SMAD dependent signaling pathways. TGF-β is able to achieve the growth inhibition primarily through suppression of c-myc and upregulation of cell cycle dependent kinase inhibitors (CDKIs) such as p15INK4B, p16INK4a, p19ARF, p21CIP1 and p57 (Massague 2008; Vijayachandra et al. 2009). Importantly, p15INK4b expression allows the release of inactive p27KIP1 from CyclinD/CDK4 complexes to promote p27KIP1 dependent inhibition of CyclinE/CDK2 and CyclinA/CDK2 complexes (Massague 2008). Together, the changes in CDKI activity can contribute to p107 (Rb) hypophosphorylation thereby preventing G1/S phase cell cycle progression. However in some carcinoma cells, c-myc may be amplified or the expression of p15INK4b, p16INK4A, p19ARF, p21CIP1 and p107 (Rb) proteins can be attenuated or completely
228
B. R. Bierie and H. L. Moses
abrogated—as a result the growth inhibitory effect of TGF-β signaling may circumvented (Massague 2008; Vijayachandra et al. 2009; Chin et al. 1998). Importantly, in some carcinoma cells that circumvent SMAD dependent growth inhibition, TGF-β stimulation can promote proliferation through the non-canonical signaling pathways. In addition to the regulation of carcinoma cell proliferation, TGF-β can promote epithelial cell apoptosis and is known to promote an epithelial to mesenchymal transition that can enhance carcinoma cell motility and invasion. The importance of TGF-β signaling in the epithelium has been well documented; however, it is also known that TGF-β signaling can have an impact on many other cell types such as immune mediators and endothelium within the tumor microenvironment (Bierie and Moses 2006a, b). However, for many years the role for TGF-β signaling in fibroblast cell populations was largely unknown with regard to the impact on adjacent carcinoma initiation, progression and metastasis. It is now clear that the TGF-β response within fibroblasts can both initiate and promote adjacent carcinoma progression (Bhowmick et al. 2004; Cheng et al. 2005, 2007, 2008; Joseph et al. 1999). Emerging evidence suggests that TGF-β signaling within the stromal fibroblast population is altered in association with human carcinoma progression. In breast cancer, stromal LOH occurs frequently in human ductal carcinoma in situ (DCIS) and invasive ductal carcinoma (IDC). Importantly, LOH in the stromal population is distinct from the LOH observed in adjacent carcinoma epithelium (Moinfar et al. 2000). The distinct LOH profiles for carcinoma epithelium and adjacent stroma associated with DCIS and IDC tissues suggested that selection pressures in a tumor microenvironment can result in amplification of distinct mutant cell populations during tumor progression. Although the TβRII locus in this study was not directly analyzed, a microsatellite marker (D3S2432) that maps to the 3p22–24.2 region (in close proximity to TβRII at the chromosomal 3p22 locus) was shown to have no LOH in the carcinoma epithelium or stromal cells associated with human DCIS or IDC lesions in vivo (Moinfar et al. 2000). In addition to proximal LOH analyses, it has been shown, in a study using 72 breast cancer and 20 normal human tissues, that mutation and loss of coding sequences in the TGFBR2 gene associated with human breast cancer is rare (Barlow et al. 2003). However, in aggressive lymph node positive (LN+), estrogen and progesterone receptor negative (ER- and PR-) tumor stroma, it was shown that the stromal TβRII protein localization was predominantly cytoplasmic suggesting that downstream signaling was attenuated in vivo (Barlow et al. 2003). Notably, loss of TβRII protein was demonstrated in correlation with advanced stages of human prostate cancer progression (Li et al. 2008). In this study it was shown that out of 33 benign prostate lesions 84% scored positive for detectable stromal TGFBR2 by immunohistochemical staining, whereas in 107 cases of Gleason 6–10 prostate cancer only 31% scored positive for TβRII staining (Li et al. 2008). In colon cancer, a similar observation was made using 310 human tissues (Bacman et al. 2007). In this study, decreased TβRI and TβRII protein was detected in tumor stroma in association with increased lymph node metastasis and shorter survival. Importantly, stromal TβRII abundance was an independent prognostic factor for cancer related survival; loss of detectable TβRII in the tumor stroma correlated with poor survival (Bacman et al. 2007). Interestingly, in human head
12 TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression
229
and neck and colon cancer a somatic mutation, TGFBR1 * 6A, has been detected in the tumor stroma (Bian et al. 2007). This hypomorphic allelic variant is less effective at transmitting downstream pathway activation when compared with the wild type allele. It was shown that carcinoma stroma, even at a distance of 2 cm from the primary tumor edge, had evidence of the somatic mutation with the highest allelic ratios occurring near the tumor margin (Bian et al. 2007). Although the data related to TGF-β signaling in human tumor stroma is limited at present, these initial studies suggest that loss of the TGF-β response in human cancer stroma does correlate with and may contribute to tumor progression in several primary tumor organ sites.
12.3 Loss of the TGF-β Response in Fibroblasts Can Result in Adjacent Carcinoma Initiation, Progression and Metastasis Carcinoma initiation, progression and metastasis are processes often thought to be driven primarily by aberrant signaling within the carcinoma cells and supported by the surrounding stromal host cells. However, the host stromal cells are potent regulators of adjacent carcinoma initiation, progression and metastasis. Paracrine signals that are derived from host stromal cells significantly regulate adjacent epithelial cell populations. The first clear evidence that stromal fibroblasts had the ability to regulate adjacent epithelial carcinoma initiation, progression and metastasis was derived in mouse models engineered to attenuate and completely ablate TGF-β responses in fibroblasts. The first model that suggested stromal TGF-β responses may have an impact on adjacent epithelial cells involved expression of a dominant negative type II TGF-β receptor (dnTbRII) under control of the metallothionein promoter (Joseph et al. 1999). This promoter was expressed primarily in stromal cell populations and when active effectively attenuated TGF-β signaling. Initial observations in the mammary tissue suggested that loss of stromal TGF-β responses could result in adjacent epithelial cell proliferation. This result was found to correlate with increased HGF production by the fibroblasts that had an attenuated ability to respond to TGF-β (Joseph et al. 1999). The results were similar to those obtained from transgenic mice designed to express HGF from a mammary specific promoter (Gallego et al. 2003). Importantly, in subsequent studies that supported and extended results obtained via attenuation of TGF-β signaling in fibroblasts, it was shown that complete ablation of TGF-β signaling in stromal fibroblasts could initiate and promote adjacent carcinoma progression (Bhowmick et al. 2004; Cheng et al. 2005). Mice with a conditional deletion of exon 2 from Tgfbr2 in stromal fibroblasts (Tgfbr2FSPKO) developed intraepithelial neoplasia in the prostate (PIN) and invasive squamous cell carcinoma of the forestomach with 100% penetrance. The mice died around 8 weeks of age, most likely due to the extensive gastrointestinal disruption associated with the squamous cell carcinomas of the forestomach that extended into the fundus. The carcinomas in this mouse model were associated with elevated levels of HGF production and Met activation. It was also shown in this mouse model that modification
230
B. R. Bierie and H. L. Moses
of TGF-β signaling in stromal fibroblasts can lead to increased proliferation and increased apoptosis of adjacent mammary epithelium (Cheng et al. 2005). In addition HGF, Mst-1 and TGF-α expression was upregulated in Tgfbr2FSPKO fibroblasts when compared to the Tgfbr2fl/fl controls when analyzed using sub-renal tissue recombination involving the MMTV-PyVmT and 4T1 mammary carcinoma cell lines. The increased ligand expression in these experiments correlated with increased Met, Ron, ErbB1 and ErbB2 activation in the carcinoma cells that had been grafted with Tgfbr2FSPKO fibroblasts relative to the controls. In vitro, blocking HGF, Mst-1 and TGF-α signaling in the conditioned medium from Tgfbr2FSPKO fibroblasts resulted in a reduction of carcinoma cell proliferation (Cheng et al. 2005). This study complemented and extended previous results wherein expression of dnTβRII was targeted to mammary stromal fibroblasts using the metallothionein promoter (Joseph et al. 1999). Together, these results suggested that stromal fibroblasts significantly contribute to the regulation of adjacent epithelium and set precedence for investigation of molecular mechanisms related to TGF-β mediated stromal-epithelial interactions during tumor initiation, progression and metastasis.
12.4 Exploring the Paracrine Signals that Regulate TGF-β Associated Stromal-Epithelial Cross-Talk in the Tumor Microenvironment It is now clear that TGF-β regulates fibroblast mediated suppression of adjacent epithelial tumor initiation, progression and metastasis. Early work using mammary fibroblasts suggested that the aberrant regulation of several secreted proteins including HGF, Mst-1 and TGF-α could explain the tumor promoting ability of fibroblasts that lacked an ability to mount a TGF-β response. The paracrine signals derived from TGF-β signaling deficient fibroblasts in a primary tumor microenvironment were able to promote lung, liver and spleen metastasis of adjacent carcinoma cells (Cheng et al. 2007, 2008). In addition, it has now been shown using TGF-β signaling deficient prostate fibroblasts that Wnt signaling can contribute to enhanced carcinoma progression in adjacent epithelial cell populations and mediate their response to systemic hormones. Further, TGF-β has been shown to enhance the epithelial cell response to paracrine cross-talk within the tumor microenvironment. It has been known for many years that HGF expression could be suppressed by TGF-β in fibroblasts (Joseph et al. 1999; Gohda et al. 1992). However, the role of this signaling axis was not known with regard to the impact on carcinoma initiation or progression. It is now clear that TGF-β dependent fibroblast derived HGF plays a significant role in the regulation of adjacent carcinoma cells (Cheng et al. 2007). HGF was initially identified as a ligand that could enhance epithelial cell proliferation, dissociation and motility (Stoker et al. 1987; Naldini et al. 1991; Nakamura et al. 1986; Matsumoto et al. 1991). It is now known that HGF can promote many processes during tumor progression including carcinoma cell transformation, proliferation, migration, invasion, adhesion and resistance to apoptosis. HGF
12 TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression
231
expression is ubiquitous in vivo, however the level of expression and activation of this ligand is significantly enhanced in reactive stroma of tumor tissues (Aguirre Ghiso et al. 1999; Parr et al. 2004). The cognate receptor for the HGF ligand is MET, a proto-oncogene abundantly expressed in many carcinoma cell subtypes (Comoglio et al. 2008). When HGF binds MET, it may activate a number of downstream signaling pathways known to include PI3K/AKT, GRB2-SOS-RAS-RAF-MEKErk, STAT3, SRC and RAC1. (Comoglio et al. 2008; Ponzetto et al. 1994). Together, these signaling pathways contribute to tumor initiation and progression (Fig. 12.2). In cancer, paracrine HGF signaling has been clearly documented (Comoglio et al. 2008). The current literature also suggests that the receptor is upregulated in epithelium associated with hepatocarcinomas, gastrinomas and carcinomas in the colon, pancreas, stomach, prostate, ovary and breast (Boccaccio and Comoglio 2006). Importantly, elevated levels of MET expression may enhance the carcinoma cell sensitivity to HGF production in the adjacent stroma. This implicates the Epithelium
TGF-β Responsive Fibroblasts HGF, TGF-α, Mst-1, Wnt2, Wnt3A, Wnt5A
-β TGF-β
TGF-β Signaling Deficient Fibroblasts
Impact on adjacent epithelium
X
Increased Met, ErbB1, ErbB2, Ron phosphorylation Increased β-catenin stabilization Adjacent epithelial hyperplasia and carcinoma initiation Epithelium
Increased carcinoma cell proliferation, invasion and metastasis
Fig. 12.2 In fibroblasts, the abrogation of TGF-β signaling can promote adjacent epithelial hyperplasia followed by carcinoma initiation, progression and metastasis. Increased carcinoma cell proliferation, invasion and metastasis observed in the carcinoma cells is known to depend on paracrine signals derived from the TGF-β signaling deficient fibroblast cell population. At present, these paracrine signals are known to include increased secretion of HGF, Mst-1, TGF-α, Wnt2, Wnt3A and Wnt5A ligands. The expression of HGF, Mst-1 and TGF-α by TGF-β signaling deficient fibroblasts can result in adjacent epithelial MET, RON, ErbB1 and ErbB2 phosphorylation that contribute to adjacent carcinoma initiation, progression and metastasis. In addition, increased Wnt2, Wnt3A and Wnt5A expression by TGF-β signaling deficient fibroblasts has been shown to have an impact on adjacent carcinoma progression. Notably, the Wnt3A expression by TGF-β signaling deficient fibroblasts in the prostate been shown to result in carcinoma cell β-catenin stabilization and enhanced androgen independent carcinoma progression
232
B. R. Bierie and H. L. Moses
TGF-β regulated stromal-epithelial HGF-MET axis as an important factor that can regulate tumor progression. The Mst-1 receptor, Ron, is structurally related to Met and is known to enhance carcinoma cell scattering, motility and invasion. Due to the upregulation of Mst-1 in TβRII deficient fibroblasts it was thought that it may also play a role in the observed regulation of tumor progression. However, it was shown that when TGF-β signaling was attenuated in fibroblasts, HGF was responsible for a majority of the previously observed Ron activation in adjacent epithelium (Cheng et al. 2008). The mechanism of Ron transactivation in this model system remains unknown and was not altered in the presence of Met inhibition suggesting an indirect interaction between HGF and Ron (Cheng et al. 2008). Importantly, it was shown that Met had the ability to increase Stat3 and p44/42 activation in mammary carcinoma cells whereas Ron only contributed to Stat3 activation (Cheng et al. 2008). In recent studies that have extended the initial observations, it has now been shown that Stat3 and p44/42 activation were responsible for a significant proportion of the HGF associated carcinoma progression in the TGF-β regulated stromalepithelial recombination model system (Cheng et al. 2007, 2008). In prostate cancer, androgen ablation therapy is often used as a method to manage disease progression. Importantly, the androgen response has been shown to involve stromal-epithelial cross-talk. Specifically, it has been shown that the stromal response to androgen can regulate adjacent epithelium. Notably, abrogation of the fibroblast associated TGF-β response can significantly attenuate the ability to initiate cell death in adjacent epithelium during systemic androgen ablation therapy (Li et al. 2008). Further, it was shown that the enhanced cell survival in prostate epithelium juxtaposed with TGF-β signaling deficient stroma was dependent upon elevated Wnt signaling (Li et al. 2008). Wnt signaling has now emerged as an important paracrine mechanism for stromal-epithelial cross-talk in the tumor microenvironment. In the case of androgen receptor (AR) signaling it is known that canonical Wnt signaling through the frizzled receptor leads to inactivation of GSK3β (Mulholland et al. 2006). The inactivation of GSK3β permits accumulation of β-catenin that is known to work with the androgen receptor to promote androgen independent prostate cancer progression (Mulholland et al. 2006). Importantly, it has been shown that loss of TGF-β signaling in fibroblasts was able to enhance the basal fibroblast derived Wnt2, Wnt3A and Wnt5A expression (Li et al. 2008). The functional impact of enhanced Wnt expression was confirmed in vivo using fibroblast-carcinoma cell recombination in the presence or absence of SFRP expression. SFRP, a secreted factor known to inhibit Wnt3A and Wnt5A activity, restored androgen ablation sensitivity to carcinoma cells in the recombined grafts. Importantly, ablation of the TGF-β response in prostate epithelium had no effect on the epithelial response to androgen ablation (Placencio et al. 2008). It was subsequently shown that systemic attenuation of Wnt3A alone was sufficient to reduce the tumor progression of the androgen sensitive LNCaP prostate carcinoma cell line that had been co-grafted with TGF-β signaling deficient fibroblasts (Li et al. 2008). The enhanced expression of Wnt3A in the TGF-β signaling deficient fibroblasts correlated with increased binding of Stat3 in the Wnt3A promoter (Wojcik et al. 2006). Although it is not known how TGF-β inhibits Stat3 signaling, the data does recapitulate results in previous literature (Cheng et al. 2008; Bright and Sriram 1998; Walia et al. 2003).
12 TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression
233
Additional mechanisms will likely be revealed as this line of research continues and it is clear that the stromal-epithelial signaling regulated by TGF-β can significantly alter the outcome of disease progression. However, the regulation of stromal-epithelial interactions by TGF-β is not limited to the impact upon the fibroblasts within the tumor microenvironment. TGF-β also regulates the carcinoma cell response to stromal derived signaling. This has been clearly shown in the case of SDF-1 and its cognate receptor CXCR4 (Ao et al. 2007). In this model system carcinoma associated fibroblasts (CAFs) were recombined with BPH-1 carcinoma cells under the renal capsule. The BPH-1 cells are initiated but non-malignant human prostate carcinoma cells. In the presence of CAFs, however it was shown that the BPH-1 tumor progression is significantly enhanced. It was found that TGF-β derived from the carcinoma associated fibroblast was responsible for the tumor promotion. Interestingly, in this context TGF-β stimulated an increase in CXCR4 expression by the carcinoma cells (Fig. 12.3). The TGF-β dependent increase in carcinoma cell CXCR4 expression thereby enhanced the carcinoma cell response
CAF
TGF -β TGF-β SDF-1
SDF-1
SDF-1 SDF-1
SDF-1 CXCR4
AKT
SDF-1
SDF-1
SDF-1
CXCR4
CXCR4 CXCR4 CXCR4 AKT
AKT
AKT AKT
Carcinoma Cell
AKT AKT
AKT
Carcinoma Cell
Fig. 12.3 In prostate carcinoma cells it has been shown that TGF-β can upregulate CXCR4 expression. The increased CXCR4 expression has been shown to increase the carcinoma cell sensitivity to SDF-1. TGF-β and SDF-1 are both abundantly expressed by carcinoma associated fibroblasts (CAF). Increased CXCR4 expression was shown to increase AKT activation in prostate carcinoma cells. AKT activation has been linked to increased carcinoma cell proliferation and in some contexts an epithelial to mesenchymal transition that can enhance carcinoma cell migration and invasion
234
B. R. Bierie and H. L. Moses
to SDF-1 (Ao et al. 2007). CAFs are an abundant source of SDF-1 and stimulation of CXCR4 resulted in activation of AKT, and importantly, AKT signaling has been shown to enhance prostate carcinoma cell proliferation and promote context-dependent EMT (Ao et al. 2006, 2007). Further, CXCR4 signaling via SDF-1 production from CAFs has been associated with increased proliferation of human breast carcinoma cells in vitro and in vivo (Orimo et al. 2005).
12.5 TGF-β Can Promote a Fibroblast to Myofibroblast Transition It has been known for more than two decades that TGF-β can promote a fibroblast to myofibroblast conversion (Desmouliere et al. 1993; Ronnov-Jessen and Petersen 1993). This process has been characterized in a number of systems that include both tumor associated fibroblasts and those derived from disease-free tissue. Myofibroblasts are fibroblast-like cells that co-express vimentin, fibroblast activation protein (FAP) and alpha-smooth muscle actin (α-SMA) (Orimo and Weinberg 2007). In early studies it was shown that mammary fibroblasts exposed to TGF-β could respond with increased α-SMA expression (Ronnov-Jessen and Petersen 1993; Sieuwerts et al. 1998). The α-SMA marker is often used as a surrogate for fibroblast activation or a myofibroblast phenotype and TGF-β signaling has a direct positive regulatory role in its expression (Desmouliere et al. 1993; Ronnov-Jessen and Petersen 1993; Hautmann et al. 1997). The process of fibroblast to myofibroblast transdifferentiation associated with TGF-β stimulation has been studied in great detail and a number of genes have been identified at different stages of the process (Chambers et al. 2003). Importantly, myofibroblast cell populations are often found at the leading invasive edge of human carcinomas and it is thought that paracrine signals derived from this cell population can significantly promote disease progression (Sieuwerts et al. 1998; Lewis et al. 2004; De Wever et al. 2004a) (Fig. 12.4). Notably, in squamous cell carcinoma (SCC) it has been shown that HGF signaling is a central mediator of myofibroblast derived paracrine signaling that promotes tumor progression upon TGF-β dependent fibroblast to myofibroblast conversion (Lewis et al. 2004). In this study it was shown that SCC cells can produce enough TGF-β to induce myofibroblast conversion of adjacent fibroblast cell populations. As a result, the myofibroblasts produced more HGF than their fibroblast precursors and the HGF produced by myofibroblasts was shown to significantly enhance SCC invasion (Lewis et al. 2004). This double paracrine mechanism was one of the first to demonstrate a significant interaction directly associating carcinoma cell TGF-β production to myofibroblast production and induction of pro-invasive HGF expression. This is quite interesting in light of the role for TGF-β signaling in suppression of fibroblast HGF expression. At present it is thought that in fibroblasts, TGF-β can suppress HGF expression however once a fibroblast is converted to a myofibroblast the HGF expression is upregulated by virtue of being a new cell type. It is worth noting that some, but not all fibroblasts are able to become myofibroblasts
12 TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression
235
TGF -β TGF-β
Fibroblast
PGE2
FMT
TGF -β TGF-β
AS
Myofibroblast
HGF
TNC
Carcinoma
Fig. 12.4 In some fibroblast cell populations, TGF-β can promote a fibroblast to myofibroblast transition (FMT) that contributes to adjacent carcinoma progression. However, the FMT process can be balanced by expression of other paracrine factors, such as PGE2, known to be abundantly expressed within the tumor microenvironment. As a result, much like many other processes regulated by TGF-β, FMT is dependent on the cell type and context of stimulation. In addition, TGF-β can promote myofibroblast cell survival through upregulation of Bcl-2. The increased Bcl-2 expression can prevent cell death by apoptotic stimuli (AS) that would otherwise occur in the absence of TGF-β signaling. Further, TGF-β has been shown to suppress iNOS expression to promote myofibroblast cell survival. iNOS expression in response to IL-1b can promote apoptosis and the suppression of iNOS gene expression is another way in which TGF-β can promote myofibroblast accumulation within a tumor microenvironment. Importantly, relative to fibroblasts, myofibroblast cell populations abundantly express the extracellular matrix glycoprotein tenascinC (TNC) and HGF that together synergize to promote carcinoma cell invasion
236
B. R. Bierie and H. L. Moses
and therefore the mesenchymal lineage must be considered when interpreting the observed TGF-β response. Notable, the fibroblast to myofibroblast transition can be attenuated by other paracrine factors often present within a tumor microenvironment such as PGE2 (Thomas et al. 2007; Kolodsick et al. 2003). Interestingly, within the myofibroblast cell population TGF-β has been shown to upregulate extracellular matrix glycoprotein tenascin-C (TNC) that can synergize with HGF to promote carcinoma cell invasion (De Wever et al. 2004a) (Fig. 12.4). It was shown that TNC expression could suppress RhoA activity that was permissive for upregulation of HGF dependent pro-migratory Rac activation in the carcinoma cells (De Wever et al. 2004a). Further, carcinoma cell production of TGF-β was able to promote the invasive ability of myofibroblasts through JNK dependent upregulation of N-cadherin at the tips of myofibroblast filopodia (De Wever et al. 2004b). Interestingly, the induction of a myofibroblast phenotype by TGF-β does not seem to be the only role for this protein. It has been shown that TGF-β signaling can promote survival of myofibroblasts in circumstances that would otherwise promote myofibroblast cell death. Interleukin 1 beta (IL-1b) has been shown to induce cell death in fibroblast and myofibroblast cell populations (Zhang and Phan 1999). Cell death induced by IL-1b is known to involve induction of iNOS, and importantly TGF-β is able to abrogate IL-1b induced iNOS production. In addition, IL-1b has been shown to reduce anti-apoptotic Bcl-2 abundance while TGF-β stimulation was able to prevent loss of Bcl-2 in myofibroblasts. Together, abrogation of iNOS production and maintenance of Bcl-2 by TGF-β in this context can promote myofibroblast cell survival (Zhang and Phan 1999).
12.6 Differential Response to TGF-β in Unique Fibroblast Subpopulations It has recently been shown that TGF-β signaling in fibroblasts can contribute to suppression of tumor promoting stromal-epithelial interactions, however it is currently unknown whether this is a general mode of regulation or dependent on the unique molecular identity of specific fibroblast subpopulations in vivo. It is clear that there are differences between fibroblast cell populations with respect to regulation by TGF-β, however it is not clear which factors regulate the differential responses to TGF-β stimulation. It is likely that the distinct molecular profile, and microenvironment associated with an individual fibroblast population (e.g., prostate stromal fibroblast versus mammary stromal fibroblast or embryonic versus adult derived fibroblasts), determines the response to TGF-β stimulation in vivo. The concept of unique signaling in alternate fibroblast cell populations, has been previously addressed through global mRNA expression analyses, that indicated distinct molecular profiles could be used to identify the tissue from which individual fibroblast cell populations were derived (Chang et al. 2002). The alternate gene expression signatures observed in fibroblasts derived from different areas of the body has been termed positional memory. The effect of positional memory on TGF-β signaling,
12 TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression
237
may partially explain why carcinomas were observed specifically in the prostate and forestomach of mice expressing a fibroblast specific ablation of exon 2 from Tgfbr2 (Bhowmick et al. 2004). The results obtained by mRNA expression profiling that were used to define positional memory, suggested that much like epithelial or myeloid cells, there may be distinct fibroblast subpopulations present within each tissue type or state of tumor progression. In support of this concept, a recent study used vimentin, type I collagen, FSP (S100A4), α-SMA, PDGFRβ and NG2 as markers to examine fibroblast heterogeneity within mammary and pancreatic carcinomas (Sugimoto et al. 2006). The analyses indicated that several distinct fibroblast sub-populations could be identified and quantified within the tumor microenvironment. Interesting the FSP+ cell population was a minor fraction of the total fibroblast compartment (this is the cell population targeted in the previously mentioned Tgfbr2FSPKO model) (Bhowmick et al. 2004). Together these results also suggest that individual sub-populations of fibroblasts and myofibroblasts may play similar or alternate roles that together contribute to the regulation of tumor progression. A recent study, that demonstrated this type of differential response to TGF-β stimulation, contrasted fibroblasts derived from fetal and adult tissues. In fetal fibroblasts, stimulation with TGF-β resulted in growth inhibition, while in adult fibroblasts stimulation resulted in enhanced proliferation (Giannouli and Kletsas 2005). This study clearly demonstrated that individual subpopulations of fibroblasts initiate unique molecular programs in response to TGF-β stimulation. Specifically, in fetal fibroblasts TGF-β was able to activate protein kinase A (PKA) and upregulate cyclin-dependent kinase inhibitors p21CIP1 and p15INK4B. Alternatively, in adult fibroblasts TGF-β was shown to promote proliferation through activation of MEK/ ERK signaling (Giannouli and Kletsas 2005). Together these results indicate that the functional impact of TGF-β stimulation is ultimately dependent on the distinct lineage and molecular profile of the fibroblast cells present within a polyclonal stromal tumor microenvironment at the time of stimulation.
12.7 Summary At present, several roles for TGF-β signaling in the regulation of tumor associated fibroblasts have been identified. However this field of study is relatively new and much more work will be necessary to elucidate the true scope of this stromalepithelial signaling axis during carcinoma initiation and progression. The current data suggests that several significant paracrine signals derived from the fibroblast cell population are regulated by TGF-β signaling. These paracrine signals are known to include HGF, Mst-1, TGF-α, Wnt2, Wnt3A and Wnt5A. HGF is perhaps the best characterized ligand in this context and a large number of systemic inhibitors for HGF signaling are in pre-clinical stages of development or clinical trials for cancer therapy (Comoglio et al. 2008). Consequently, it has been shown that the carcinoma cell specific HGF response can promote early tumor cell engraftment and growth (Martin et al. 2003; Corso et al. 2008). In addition, when HGF signaling
238
B. R. Bierie and H. L. Moses
was attenuated after tumors were established, a significant reduction in tumor volume was observed. Notably, colonization and growth of pulmonary metastases were also shown to depend upon the response to HGF signaling specifically within metastatic carcinoma cells (Corso et al. 2008). According to these studies and many others in the current literature, it is clear that a significant impact can be ascribed to carcinoma cell specific HGF/MET signaling during primary and metastatic tumor progression in vivo. HGF has also been shown to promote carcinoma cell survival through activation of AKT and ERK signaling pathways (Xiao et al. 2001; Zeng et al. 2002), and may directly promote invasion through E1AF dependent stimulation of matrix metalloproteinase MMP1, MMP3 and MMP9 expression (Hanzawa et al. 2000). Importantly, once a carcinoma cell leaves the primary tumor tissue and arrives at a distant organ site, the ability to adhere may promote subsequent colonization. HGF signaling has been shown to increase the binding of MDA-MB-231 cells to laminin-1, laminin-5, fibronectin and vitronectin substrates through upregulation of β1, β3, β4 and β5 integrins (Trusolino et al. 2000). The mechanism of HGF dependent binding to these specific matrix substrates was shown to be PI3K dependent (Trusolino et al. 2000). In support of the mechanistic studies in vitro, functional regulation of metastasis by MET signaling has been linked to activation of both GRB2 and PI3K pathways in vivo (Bardelli et al. 1999). It is clear from these studies and many subsequent reports that fibroblast derived HGF is able to function as a mitogen, motogen, morphogen, inhibitor of apoptosis, stimulatory regulator of matrix degradation and pro-angiogenic factor that can support tumor initiation and progression in adjacent epithelial cell populations. As a result, the stimulation of integrin mediated adhesion by HGF may aid in homing of mammary carcinoma cells to tissues such as the lung that express laminin-5, fibronectin and vitronectin. Although the regulation of stromal HGF signaling mediated by TGF-β has been reported to involve Stat3 and p44/42 signaling in adjacent epithelium, it is likely that a number of the other downstream HGF dependent signaling components also contribute to the observed stromal-epithelial interaction carcinoma initiation, progression and metastasis. It is likely that we have yet to identify critical links between the stromal-epithelial axes regulated by TGF-β during carcinoma progression. In prostate cancer for example, it has been shown that Wnt3A was responsible for conferring resistance to androgen ablation therapy in the presence of TGF-β signaling deficient fibroblasts (Li et al. 2008). It was also shown that TGF-β could upregulate CXCR4 expression in prostate carcinoma cells that increased their sensitivity to SDF-1 (Ao et al. 2007). SDF-1 is abundantly expressed by carcinoma associated fibroblasts and the increased sensitivity to this ligand resulted in enhanced AKT activation in adjacent prostate carcinoma cells (Ao et al. 2007). Importantly, previous work that was not related to TGF-β signaling clearly demonstrated that Wnt3A could work together with AKT signaling in prostate carcinoma cells to promote androgen receptor independent carcinoma progression (Mulholland et al. 2006). These studies together suggest that TGF-β signaling within fibroblasts and TGF-β ligand production by fibroblasts can together regulate androgen independent prostate carcinoma progression. This type of molecular interaction is simple to address, however due this field
12 TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression
239
being in early stages of development the potential impact of such interactions have yet to be resolved. The complexity associated with stromal-epithelial interactions mediated by the TGF-β response within tumor derived fibroblasts is not limited to the scope of the molecular interactions within the tumor microenvironment. Our current view of the global complexity within the tumor microenvironment leaves many issues unresolved including the generalization of results attained in the FSP+ fibroblast cell population. It is likely that as we learn how to identify additional fibroblast subtypes and lineages additional analyses will be necessary to determine the impact of TGF-β signaling in each unique subpopulation. Although this issue also remains unresolved, it is clear that the TGF-β pathway can be attenuated within human carcinoma fibroblasts. In addition, it is clear that the loss of TGF-β signaling in fibroblasts can initiate and promote adjacent carcinoma progression through increased HGF, Wnt-2, Wnt-3A, Wnt-5A production. Further, CAF derived TGF-β production can enhance the sensitivity of carcinoma cells to paracrine signaling as demonstrated in the case of SDF-1. Alternatively, we also know that some TGF-β responsive fibroblasts exhibit a myofibroblast phenotype in response to TGF-β stimulation that can promote tumor progression through increased MMP, HGF and TNC production within a local tumor microenvironment. Together, the results presented within the current literature suggest that TGF-β signaling in stromal fibroblasts and TGF-β production by carcinoma associated fibroblasts can suppress or promote adjacent carcinoma initiation and progression depending upon the context of stimulation.
References Abe M, Oda N, Sato Y (1998) Cell-associated activation of latent transforming growth factor-beta by calpain. J Cell Physiol 174(2):186–193 Aguirre Ghiso JA et al (1999) Deregulation of the signaling pathways controlling urokinase production. Its relationship with the invasive phenotype. Eur J Biochem 263(2):295–304 Aitchison AA et al (2008) Promoter methylation correlates with reduced Smad4 expression in advanced prostate cancer. Prostate 68(6):661–674 Akita K et al (2002) Impaired liver regeneration in mice by lipopolysaccharide via TNF-alpha/ kallikrein-mediated activation of latent TGF-beta. Gastroenterology 123(1):352–364 Ao M et al (2006) Transforming growth factor-beta promotes invasion in tumorigenic but not in nontumorigenic human prostatic epithelial cells. Cancer Res 66(16):8007–8016 Ao M et al (2007) Cross-talk between paracrine-acting cytokine and chemokine pathways promotes malignancy in benign human prostatic epithelium. Cancer Res 67(9):4244–4253 Bacman D et al (2007) TGF-beta receptor 2 downregulation in tumour-associated stroma worsens prognosis and high-grade tumours show more tumour-associated macrophages and lower TGF-beta1 expression in colon carcinoma: a retrospective study. BMC Cancer 7:156 Barcellos-Hoff MH, Dix TA (1996) Redox-mediated activation of latent transforming growth factor-beta 1. Mol Endocrinol 10(9):1077–1083 Bardelli A et al (1999) Concomitant activation of pathways downstream of Grb2 and PI 3-kinase is required for MET-mediated metastasis. Oncogene 18(5):1139–1146 Barlow J et al (2003) Higher stromal expression of transforming growth factor-beta type II receptors is associated with poorer prognosis breast tumors. Breast Cancer Res Treat 79(2):149–159
240
B. R. Bierie and H. L. Moses
Bhowmick NA et al (2004) TGF-beta signaling in fibroblasts modulates the oncogenic potential of adjacent epithelia. Science 303(5659):848–851 Bian Y et al (2007) Somatic acquisition of TGFBR1 * 6A by epithelial and stromal cells during head and neck and colon cancer development. Hum Mol Genet 16(24):3128–3135 Bierie B, Moses HL (2006a) TGF-beta and cancer. Cytokine Growth Factor Rev 17(1–2):29–40 Bierie B, Moses HL (2006b) Tumour microenvironment: TGFbeta: the molecular Jekyll and Hyde of cancer. Nat Rev Cancer 6(7):506–520 Boccaccio C, Comoglio PM (2006) Invasive growth: a MET-driven genetic programme for cancer and stem cells. Nat Rev Cancer 6(8):637–645 Bright JJ, Sriram S (1998) TGF-beta inhibits IL-12-induced activation of Jak-STAT pathway in T lymphocytes. J Immunol 161(4):1772–1777 Brown KA, Pietenpol JA, Moses HL (2007) A tale of two proteins: differential roles and regulation of Smad2 and Smad3 in TGF-beta signaling. J Cell Biochem 101(1):9–33 Chambers RC et al (2003) Global expression profiling of fibroblast responses to transforming growth factor-beta1 reveals the induction of inhibitor of differentiation-1 and provides evidence of smooth muscle cell phenotypic switching. Am J Pathol 162(2):533–546 Chang HY et al (2002) Diversity, topographic differentiation, and positional memory in human fibroblasts. Proc Natl Acad Sci U S A 99(20):12877–12882 Cheng N et al (2005) Loss of TGF-beta type II receptor in fibroblasts promotes mammary carcinoma growth and invasion through upregulation of TGF-alpha-, MSP- and HGF-mediated signaling networks. Oncogene 24(32):5053–5068 Cheng N et al (2007) Enhanced hepatocyte growth factor signaling by Type II transforming growth factor-{beta} receptor knockout fibroblasts promotes mammary tumorigenesis. Cancer Res 67(10):4869–4877 Cheng N et al (2008) Transforming growth factor-beta signaling-deficient fibroblasts enhance hepatocyte growth factor signaling in mammary carcinoma cells to promote scattering and invasion. Mol Cancer Res 6(10):1521–1533 Chin L, Pomerantz J, DePinho RA (1998) The INK4a/ARF tumor suppressor: one gene–two products–two pathways. Trends Biochem Sci 23(8):291–296 Comoglio PM, Giordano S, Trusolino L (2008) Drug development of MET inhibitors: targeting oncogene addiction and expedience. Nat Rev Drug Discov 7(6):504–516 Corso S et al (2008) Silencing the MET oncogene leads to regression of experimental tumors and metastases. Oncogene 27(5):684–693 Daly AC, Randall RA (2008) Hill CS Transforming growth factor beta-induced Smad1/5 phosphorylation in epithelial cells is mediated by novel receptor complexes and is essential for anchorage-independent growth. Mol Cell Biol 28(22):6889–6902 Datta PK, Moses HL (2000) STRAP and Smad7 synergize in the inhibition of transforming growth factor beta signaling. Mol Cell Biol 20(9):3157–3167 Davis BN et al (2008) SMAD proteins control DROSHA-mediated microRNA maturation. Nature 454(7200):56–61 De Wever O et al (2004a) Critical role of N-cadherin in myofibroblast invasion and migration in vitro stimulated by colon-cancer-cell-derived TGF-beta or wounding. J Cell Sci 117(Pt 20):4691–4703 De Wever O et al (2004b) Tenascin-C SF/HGF produced by myofibroblasts in vitro provide convergent pro-invasive signals to human colon cancer cells through RhoA and Rac. Faseb J 18(9):1016–1018 Derynck R, Zhang YE (2003) Smad-dependent and Smad-independent pathways in TGF-beta family signalling. Nature 425(6958):577–584 Desmouliere A et al (1993) Transforming growth factor-beta 1 induces alpha-smooth muscle actin expression in granulation tissue myofibroblasts and in quiescent and growing cultured fibroblasts. J Cell Biol 122(1):103–111 Ebisawa T et al (2001) Smurf1 interacts with transforming growth factor-beta type I receptor through Smad7 and induces receptor degradation. J Biol Chem 276(16):12477–12480
12 TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression
241
Feng XH, Derynck R (2005) Specificity and versatility in TGF- signaling through Smads. Annu Rev Cell Dev Biol 21:659–693 Ferrigno O et al (2002) Yes-associated protein (YAP65) interacts with Smad7 and potentiates its inhibitory activity against TGF-beta/Smad signaling. Oncogene 21(32):4879–4884 Gallego MI, Bierie B, Hennighausen L (2003) Targeted expression of HGF/SF in mouse mammary epithelium leads to metastatic adenosquamous carcinomas through the activation of multiple signal transduction pathways. Oncogene 22(52):8498–8508 Giannouli CC, Kletsas D (2005) TGF-beta regulates differentially the proliferation of fetal and adult human skin fibroblasts via the activation of PKA and the autocrine action of FGF-2. Cell Signal 18(9):1417–1429 Gohda E et al (1992) TGF-beta is a potent inhibitor of hepatocyte growth factor secretion by human fibroblasts. Cell Biol Int Rep 16(9):917–926 Grady WM, Markowitz SD (2002) Genetic and epigenetic alterations in colon cancer. Annu Rev Genomics Hum Genet 3:101–128 Grady WM et al (1999) Mutational inactivation of transforming growth factor beta receptor type II in microsatellite stable colon cancers. Cancer Res 59(2):320–324 Hanzawa M et al (2000) Hepatocyte growth factor upregulates E1AF that induces oral squamous cell carcinoma cell invasion by activating matrix metalloproteinase genes. Carcinogenesis 21(6):1079–1085 Hautmann MB, Madsen CS, Owens GK (1997) A transforming growth factor beta (TGFbeta) control element drives TGFbeta-induced stimulation of smooth muscle alpha-actin gene expression in concert with two CArG elements. J Biol Chem 272(16):10948–10956 Hayashi H et al (1997) The MAD-related protein Smad7 associates with the TGFbeta receptor and functions as an antagonist of TGFbeta signaling. Cell 89(7):1165–1173 Hinshelwood RA et al (2007) Concordant epigenetic silencing of transforming growth factorbeta signaling pathway genes occurs early in breast carcinogenesis. Cancer Res 67(24):11517– 11527 Inoue Y, Imamura T (2008) Regulation of TGF-beta family signaling by E3 ubiquitin ligases. Cancer Sci 99(11):2107–2112 Jobling MF et al (2006) Isoform-specific activation of latent transforming growth factor beta (LTGF-beta) by reactive oxygen species. Radiat Res 166(6):839–848 Joseph H et al (1999) Overexpression of a kinase-deficient transforming growth factor-beta type II receptor in mouse mammary stroma results in increased epithelial branching. Mol Biol Cell 10(4):1221–1234 Kang SH et al (1999) Transcriptional repression of the transforming growth factor-beta type I receptor gene by DNA methylation results in the development of TGF-beta resistance in human gastric cancer. Oncogene 18(51):7280–7286 Kavsak P et al (2000) Smad7 binds to Smurf2 to form an E3 ubiquitin ligase that targets the TGF beta receptor for degradation. Mol Cell 6(6):1365–1375 Kim SJ et al (2000) Molecular mechanisms of inactivation of TGF-beta receptors during carcinogenesis. Cytokine Growth Factor Rev 11(1–2):159–168 Kolodsick JE et al (2003) Prostaglandin E2 inhibits fibroblast to myofibroblast transition via E. prostanoid receptor 2 signaling and cyclic adenosine monophosphate elevation. Am J Respir Cell Mol Biol 29(5):537–544 Lee MK et al (2007) TGF-beta activates Erk MAP kinase signalling through direct phosphorylation of ShcA. Embo J 26(17):3957–3967 Levy L, Hill CS (2006) Alterations in components of the TGF-beta superfamily signaling pathways in human cancer. Cytokine Growth Factor Rev 17(1–2):41–58 Lewis MP et al (2004) Tumour-derived TGF-beta1 modulates myofibroblast differentiation and promotes HGF/SF-dependent invasion of squamous carcinoma cells. Br J Cancer 90(4):822– 832 Li X et al (2008) Prostate tumor progression is mediated by a paracrine TGF-beta/Wnt3a signaling axis. Oncogene 27(56):7118–7130
242
B. R. Bierie and H. L. Moses
Liu IM et al (2009) TGFbeta-stimulated Smad1/5 phosphorylation requires the ALK5 L45 loop and mediates the pro-migratory TGFbeta switch. Embo J 28(2):88–98 Luo K et al (1999) The Ski oncoprotein interacts with the Smad proteins to repress TGFbeta signaling. Genes Dev 13(17):2196–2206 Lyons RM, Keski-Oja J, Moses HL (1988) Proteolytic activation of latent transforming growth factor-beta from fibroblast-conditioned medium. J Cell Biol 106(5):1659–1665 Markowitz S et al (1995) Inactivation of the type II TGF-beta receptor in colon cancer cells with microsatellite instability. Science 268(5215):1336–1338 Martin TA et al (2003) Growth and angiogenesis of human breast cancer in a nude mouse tumour model is reduced by NK4, a HGF/SF antagonist. Carcinogenesis 24(8):1317–1323 Massague J (2008) TGFbeta in cancer. Cell 134(2):215–230 Massague J, Seoane J, Wotton D (2005) Smad transcription factors. Genes Dev 19(23):2783–2810 Matsumoto K et al (1991) Marked stimulation of growth and motility of human keratinocytes by hepatocyte growth factor. Exp Cell Res 196(1):114–120 Miyazono K (1989) Heldin CH Role for carbohydrate structures in TGF-beta 1 latency. Nature 338(6211):158–160 Moinfar F et al (2000) Concurrent and independent genetic alterations in the stromal and epithelial cells of mammary carcinoma: implications for tumorigenesis. Cancer Res 60(9):2562–2566 Moustakas A, Heldin CH (2007) Signaling networks guiding epithelial-mesenchymal transitions during embryogenesis and cancer progression. Cancer Sci 98(10):1512–1520 Mulholland DJ et al (2006) PTEN and GSK3beta: key regulators of progression to androgenindependent prostate cancer. Oncogene 25(3):329–337 Munger JS et al (1999) The integrin alpha v beta 6 binds and activates latent TGF beta 1: a mechanism for regulating pulmonary inflammation and fibrosis. Cell 96(3):319–328 Nakamura T, Teramoto H, Ichihara A (1986) Purification and characterization of a growth factor from rat platelets for mature parenchymal hepatocytes in primary cultures. Proc Natl Acad Sci U S A 83(17):6489–6493 Nakao A et al (1997) Identification of Smad7, a TGFbeta-inducible antagonist of TGF-beta signalling. Nature 389(6651):631–635 Naldini L et al (1991) Scatter factor and hepatocyte growth factor are indistinguishable ligands for the MET receptor. EMBO J 10(10):2867–2878 Orimo A et al (2005) Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion. Cell 121(3):335– 348 Orimo A, Weinberg RA (2007) Heterogeneity of stromal fibroblasts in tumors. Cancer Biol Ther 6(4):618–619 Parr C et al (2004) The hepatocyte growth factor regulatory factors in human breast cancer. Clin Cancer Res 10(1, Pt 1):202–211 Placencio VR et al (2008) Stromal transforming growth factor-beta signaling mediates prostatic response to androgen ablation by paracrine Wnt activity. Cancer Res 68(12):4709–4718 Ponzetto C et al (1994) A multifunctional docking site mediates signaling and transformation by the hepatocyte growth factor/scatter factor receptor family. Cell 77(2):261–271 Rojas A et al (2008) The aberrant methylation of TSP1 suppresses TGF-beta1 activation in colorectal cancer. Int J Cancer 123(1):14–21 Ronnov-Jessen L, Petersen OW (1993) Induction of alpha-smooth muscle actin by transforming growth factor-beta 1 in quiescent human breast gland fibroblasts. Implications for myofibroblast generation in breast neoplasia. Lab Invest 68(6):696–707 Schultz-Cherry S, Hinshaw VS (1996) Influenza virus neuraminidase activates latent transforming growth factor beta. J Virol 70(12):8624–8629 Schultz-Cherry S, Murphy-Ullrich JE (1993) Thrombospondin causes activation of latent transforming growth factor-beta secreted by endothelial cells by a novel mechanism. J Cell Biol 122(4):923–932 Shi W et al (2004) GADD34-PP1c recruited by Smad7 dephosphorylates TGFbeta type I receptor. J Cell Biol 164(2):291–300
12 TGF-β Signaling in Fibroblasts Regulates Tumor Initiation and Progression
243
Shi Y, Massague J (2003) Mechanisms of TGF-beta signaling from cell membrane to the nucleus. Cell 113(6):685–700 Shipitsin M (2007) Molecular definition of breast tumor heterogeneity. Cancer Cell 11(3):259–273 Sieuwerts AM et al (1998) Urokinase-type-plasminogen-activator (uPA) production by human breast (myo) fibroblasts in vitro: influence of transforming growth factor-beta(1) (TGF beta(1)) compared with factor(s) released by human epithelial-carcinoma cells. Int J Cancer 76(6):829–835 Sjoblom T et al (2006) The consensus coding sequences of human breast and colorectal cancers. Science 314(5797):268–274 Stoker M et al (1987) Scatter factor is a fibroblast-derived modulator of epithelial cell mobility. Nature 327(6119):239–242 Stover DG, Bierie B, Moses HL (2007) A delicate balance: TGF-beta and the tumor microenvironment. J Cell Biochem 101(4):851–861 Stroschein SL et al (1999) Negative feedback regulation of TGF-beta signaling by the SnoN oncoprotein. Science 286(5440):771–774 Sugimoto H et al (2006) Identification of fibroblast heterogeneity in the tumor microenvironment. Cancer Biol Ther 5(12):1640–1646 Taipale J et al (1995) Human mast cell chymase and leukocyte elastase release latent transforming growth factor-beta 1 from the extracellular matrix of cultured human epithelial and endothelial cells. J Biol Chem 270(9):4689–4696 Thomas PE et al (2007) PGE(2) inhibition of TGF-beta1-induced myofibroblast differentiation is Smad-independent but involves cell shape and adhesion-dependent signaling. Am J Physiol Lung Cell Mol Physiol 293(2):L417–L428 Trusolino L et al (2000) HGF/scatter factor selectively promotes cell invasion by increasing integrin avidity. FASEB J 14(11):1629–1640 Vijayachandra K et al (2009) Induction of p16ink4a and p19ARF by TGFbeta1 contributes to growth arrest and senescence response in mouse keratinocytes. Mol Carcinog 48(3):181–186 Walia B (2003) TGF-beta down-regulates IL-6 signaling in intestinal epithelial cells: critical role of SMAD-2. Faseb J 17(14):2130–2132 Wojcik EJ et al (2006) A novel activating function of c-Src and Stat3 on HGF transcription in mammary carcinoma cells. Oncogene 25(19):2773–2784 Xiao GH et al (2001) Anti-apoptotic signaling by hepatocyte growth factor/Met via the phosphatidylinositol 3-kinase/Akt and mitogen-activated protein kinase pathways. Proc Natl Acad Sci U S A 98(1):247–252 Yamashita M et al (2008) TRAF6 mediates Smad-independent activation of JNK and p38 by TGFbeta. Mol Cell 31(6):918–924 Yu Q, Stamenkovic I (2000) Cell surface-localized matrix metalloproteinase-9 proteolytically activates TGF-beta and promotes tumor invasion and angiogenesis. Genes Dev 14(2):163–176 Zeng Q et al (2002) Hepatocyte growth factor inhibits anoikis in head and neck squamous cell carcinoma cells by activation of ERK and Akt signaling independent of NFkappa B. J Biol Chem 277(28):25203–25208 Zhang HY, Phan SH (1999) Inhibition of myofibroblast apoptosis by transforming growth factor beta(1). Am J Respir Cell Mol Biol 21(6):658–665
.
Chapter 13
The SDF-1-Rich Tumour Microenvironment Provides a Niche for Carcinoma Cells Masayuki Shimoda, Kieran Mellody and Akira Orimo
13.1 Introduction SDF-1, a member of the chemokine superfamily, was initially cloned from a bone marrow-derived stromal cell line (Tashiro et al. 1993), and identified as a pre-B-cell growth-stimulating factor involved in regulating B cell lymphopoiesis (Nagasawa et al. 1994). SDF-1 is secreted in abundance by mesenchymal cells present within sites of damaged tissues and the tumour stroma. This chemokine plays an essential role in promoting tissue repair and tumourigenesis. The main cognate receptor for SDF-1 is known as CXCR4 which belongs to the CXC chemokine receptor subfamily. This protein serves as an essential co-receptor of CD4, which facilitates the binding and entry of human immunodeficiency virus-1 (HIV-1) into T cells (Bleul et al. 1996; Oberlin et al. 1996). The SDF-1-CXCR4 signalling pathway is involved in embryonic development, tissue homeostasis, tissue inflammation, wound healing and tumourigenesis (Burger and Kipps 2006; Kryczek et al. 2007; Ratajczak et al. 2006). For a long time CXCR4 was considered the only available receptor for SDF1. However, recent studies have identified an additional receptor for this ligand known as CXCR7 (Balabanian et al. 2005; Burns et al. 2006). Studies show that CXCR4 and CXCR7 signalling both play a role in tumour progression (Balkwill 2004a; Burns et al. 2006), although little is known about the precise downstream signalling pathway that mediates their oncogenic functions. This chapter focuses on host stroma-derived SDF-1 signalling in cancer and its role in generating a tumourpromoting niche.
A. Orimo () CR-UK Stromal-Tumour Interaction Group, Paterson Institute for Cancer Research, The University of Manchester, Wilmslow Road, Manchester, M20 4BX, UK e-mail:
[email protected] M. M. Mueller, N. E. Fusenig (eds.), Tumor-Associated Fibroblasts and their Matrix, The Tumor Microenvironment 4, DOI 10.1007/978-94-007-0659-0_13, © Springer Science+Business Media B.V. 2011
245
246
M. Shimoda et al.
13.2 Physiological Role of SDF-1-CXCR4 Signalling Studies in transgenic and knockout mice have demonstrated the importance of the SDF-1-CXCR4 signalling pathway during embryonic development. Deletion of the gene encoding for either SDF-1 or CXCR4 results in an embryonic lethal phenotype due to multiple organ defects (Nagasawa et al. 1996; Tachibana et al. 1998; Zou et al. 1998). The shared phenotypes seen in the knockout mouse for either protein strongly suggest that CXCR4 is the main receptor for SDF-1. SDF1-CXCR4 signalling is especially important for the development of the stem cell niche, supporting and regulating the homing, retention, survival and quiescence of hematopoietic stem cells (HSCs) in the bone marrow (Heissig et al. 2002; Spiegel et al. 2008; Sugiyama et al. 2006). SDF-1 is produced by resident endothelial cells, CXCL12-abundant reticular (CAR) cells and osteoblasts. High levels of SDF-1 recruit CXCR4-expressing mesenchymal cell progenitors and various myeloid cells into the sites of wounded tissues in order to facilitate the repair and regeneration of damaged tissues (Spiegel et al. 2008).
13.3 Expression and Regulation of SDF-1 and CXCR4 in Cancer SDF-1-CXCR4 signalling has been implicated in over twenty different types of epithelial, mesenchymal and haematopoietic carcinomas (Balkwill 2004b). Within the tumour microenvironment, SDF-1 is produced by several different cell types that include endothelial cells, stromal fibroblasts, epithelial cells and carcinoma cells (Allinen et al. 2004; Orimo et al. 2005; Tait et al. 2007; Zou et al. 2001). CXCR4 is most commonly expressed on carcinoma cells but is rarely detected in normal tissue. Importantly, CXCR4 expression is progressively increased by carcinoma cells during tumour progression and it is actively regulated by crosstalk with several oncogenic signalling pathways (Fig. 13.1). As mentioned above, SDF-1 production is dramatically up-regulated at sites of tissue damage as well as in the tumour stroma. However, little is known about the molecular mechanisms by which the tissue damage response induces SDF-1 production.
13.4 Essential Roles for SDF-1-CXCR4 Signalling in Carcinoma Cells in Promoting Tumourigenesis In a tumour xenograft model, carcinoma cells expressing CXCR4 promote tumour growth, stimulate neoangiogenesis and increase spontaneous metastatic dissemination (Balkwill 2004a). Furthermore, these cells are able to resist anoikis and to colonise distant tissues and organs to form both micro- and macro-metastases
13 The SDF-1-Rich Tumour Microenvironment Provides a Niche for Carcinoma Cells
247
Fig. 13.1 Regulation and the roles of SDF-1-CXCR4 and SDF-1-CXCR7 signalling in carcinoma cells in promoting tumourigenesis. In tumour tissues, mesenchymal cells such as endothelial cells and fibroblasts are a main source of SDF-1. The latter binds to CXCR4 to activate signalling in the carcinoma cells. The expression of CXCR4 is up-regulated in hypoxia through activation of the hypoxia-inducible factor-1 (HIF-1) signalling pathway (Staller et al. 2003) and cytokines including TGF-β (Bertran et al. 2009) and VEGF (Bachelder et al. 2002). The expression of CXCR4 is stabilised by HER2 through prevention of the ligand-induced CXCR4 degradation (Li et al. 2004). Induction of CXCR4 signalling in turn stimulates transactivation of HER2 through Src kinase activation (Cabioglu et al. 2005). SDF-1-CXCR4 axis stimulates the distinct downstream signal pathways including AKT, MAPK, RhoA, Rac1 (Bartolome et al. 2004) and integrins (Hartmann et al. 2005), and elevates expression levels of proinvasive factors, such as MMPs, VEGF, IL-6 and IL-8 (Wang et al. 2005). Activation of CXCR4 signalling confers cells with increased proliferation, survival, migration, invasion and adhesion propensities. SDF-1-CXCR4 signalling may also be involved in the induction of an epithelial to mesenchymal transition (EMT) and the self-renewal properties of cancer stem cells (CSCs). Furthermore, CXCR7, an additional receptor for SDF-1, might form heterodimers with CXCR4 to activate the signalling pathway and facilitate tumourigenesis
(Zeelenberg et al. 2003). Concordantly, subpopulations of human breast carcinoma cells that demonstrate an increased metastatic propensity into bone or lung, also display increased levels of CXCR4 expression compared to their parental cells (Helbig et al. 2003; Kang et al. 2003). The increased capacity of CXCR4-expressing cancer cells to promote tumourigenesis is further highlighted by a study showing that oral squamous cell carcinoma cells, when exposed to recombinant SDF-1 protein, display a mesenchymal phenotype (Onoue et al. 2006). These cells undergo an epithelial to mesenchymal transition (EMT) that confers upon carcinoma cells the invasive and migratory phenotypes required for cancer progression. Conversely, different molecular or biochemical approaches to inhibit CXCR4 signalling within carcinoma cells substantially attenuates their tumourigenic and metastatic capabilities in vivo (Muller et al. 2001; Zeelenberg et al. 2003). Some studies have also shown that CXCR4 expression provides cancer cells with their self-renewing properties. Tentative CD133+CXCR4+ cancer stem cell (CSC) populations isolated from human pancreatic carcinoma cells, when orthotopically injected into mice, demonstrate a greater ability to form liver metastases
248
M. Shimoda et al.
compared to the control CD133+CXCR4− cell populations (Hermann et al. 2007). Furthermore, a putative breast carcinoma stem cell population, capable of forming mammospheres within an in vitro assay, displays a 10-fold increase in their expression of CXCR4 (Krohn et al. 2009). SDF-1 significantly influences the biological behaviour of carcinoma cells in a CXCR4-dependant fashion. Cancer cells cultured in the presence of SDF-1 increase their migratory and invasive propensities through the induction of actin polymerisation (Muller et al. 2001), and the production of pro-invasive factors such as matrix metalloproteinases (MMPs), VEGF, IL-6 and IL-8 (Bartolome et al. 2004; Wang et al. 2005) (Fig. 13.1). Collectively, these findings indicate an essential role for SDF-1-CXCR4 signalling in tumour progression.
13.5 Tumour-Promoting Role of SDF-1-CXCR7 Signalling CXCR7/RDC1 has been identified as an important gene which facilitates cell transformation induced by Kaposi sarcoma-associated herpes virus (Raggo et al. 2005). CXCR7 expression on carcinoma cells progressively increases during human prostate carcinoma development in patients (Wang et al. 2008). Cancer cells overexpressing CXCR7 also show accelerated tumour growth kinetics, stimulated neoangiogenesis and increased formation of metastases in vivo (Burns et al. 2006; Miao et al. 2007; Wang et al. 2008). Conversely, the inhibition or down-regulation of CXCR7 significantly reduces the growth rate of carcinoma cells. Moreover, in human prostate cancer cells, induction of CXCR7 signalling elevates IL-8 and VEGF expression and increases AKT signalling (Wang et al. 2008). These same molecules are also up-regulated by activation of the SDF-1-CXCR4 signalling pathway, suggesting a shared downstream pathway for both receptors. However, it is unclear whether CXCR7 can contribute to the following downstream signalling either by regulating CXCR4 activity through heterodimer formation or by signalling on its own (Fig.13.1). Further research is required in order to shed light on the precise downstream targets involved in SDF-1-CXCR7 signalling.
13.6 Host-Derived SDF-1 Signalling Promotes Tumour Growth, Invasion and Metastasis Recent accumulated evidence demonstrates that tumourigenesis is dependant upon contextual signals released from the apposing tumour-associated stroma that supports and sustains carcinoma cell growth throughout tumour progression (Bhowmick et al. 2004; Bissell and Radisky 2001; Coussens and Werb 2002; Kalluri and Zeisberg 2006; Mueller and Fusenig 2004; Orimo and Weinberg 2006; Polyak et al. 2009). Tumour-associated stroma includes numbers of fibroblasts and myofibroblasts. The latter is typically positive for α-smooth muscle actin (α-SMA) and these
13 The SDF-1-Rich Tumour Microenvironment Provides a Niche for Carcinoma Cells
249
fibroblasts are very frequently detected in the stroma of various types of human carcinomas including the breast (Fig. 13.2). The presence of large numbers of myofibroblasts is often associated with high-grade malignancies and poor prognoses in the cancer patients (Cardone et al. 1997; Kellermann et al. 2008; Maeshima et al. 2002). Importantly, unlike fibroblasts found within the non-cancerous stroma, myofibroblasts present within the tumour stroma actively produce SDF-1 (Orimo et al. 2005; Tait et al. 2007) (Fig. 13.2).
Fig. 13.2 Stromal myofibroblasts produce SDF-1 protein in human breast tumour. Paraffin sections of human invasive breast cancer tissue (a, b, d, e, and f) or non-cancerous tissue (c) were immunostained with anti-α-SMA (a) or anti-SDF-1 (b and c) antibodies. Note that SDF-1-positive fibroblast-like cells ( arrows in b) as well as α-SMA-positive myofibroblasts ( arrows in a) are present in the tumour-associated stroma. In contrast, in non-cancerous tissue SDF-1 protein is only detected in epithelial cells ( arrows in c) and not in any fibroblast-like cells ( arrowheads in c). Furthermore, the adjacent tumour section is double-stained with both anti-α-SMA (d) and antiSDF-1 (e) antibodies. Many α-SMA-positive myofibroblasts show positive staining for SDF-1 in carcinoma (f). (From Orimo et al. 2005)
250
M. Shimoda et al.
The tumour-promoting ability of CAFs, which are stromal myofibroblast-rich cell populations extracted from human carcinoma, has been previously investigated using tumour xenograft models (Hu et al. 2009; Hwang et al. 2008; Olumi et al. 1999; Orimo et al. 2005; Yang et al. 2006). Tumours that develop in the presence of CAFs show a more increased growth rate compared to those that develop in the presence of control fibroblasts isolated from non-cancerous tissues. Considerable numbers of CAFs or control fibroblasts stay alongside the carcinoma cells within advanced tumour xenografts (Orimo et al. 2005; Yang et al. 2005). CAF-secreted paracrine factors are therefore likely responsible for the observed increase in carcinoma growth during tumour development. Importantly, CAFs secrete SDF-1 that stimulates carcinoma cell proliferation, and promotes neoangiogenesis by recruiting circulating endothelial progenitor cells (EPCs) into the tumour site (Orimo et al. 2005) (Fig. 13.3). Other studies have also shown that CAFs secrete high levels of TGF-β1 (Ao et al. 2007; Rosenthal et al.
Fig. 13.3 Human breast tumours developing in the presence of CAFs are highly angiogenic. a Human breast MCF-7-ras carcinoma cells were co-injected with either CAFs or the control counterpart fibroblasts extracted from the non-cancerous tissue from the same individual, subcutaneously into immunodeficient mice. Sections prepared from the tumour xenografts were stained with Masson’s trichrome (A, B) and an anti-CD31 antibody (C, D). An extensive vascular formation is observed in tumours containing CAFs, whilst capillaries in tumours containing the control fibroblasts are less developed. b MCF-7-ras breast tumours derived from xenografts implanted with either CAFs or counterpart fibroblasts were dissociated into a single cell suspension at 60–63 days after injection. Cells were stained with antibodies specific to Sca1 and CD31 protein, both of which are makers of endothelial progenitor cells (EPCs). A far higher proportion (4.2-fold) of Sca1+CD31+ EPCs is detected in tumours admixed with CAFs compared to tumours containing the control counterpart fibroblasts. *p Breast cancer (Barth et al. 2005; Beck et al. 2008b; High in mammary Increased SPARC/poor Upregulated in: node negative; survival Stromal myoepithelial cells, Bergamaschi et al. 2008a; Jones b grades 1 and 3 fibroblasts, low level in interet al. 2004; Porter et al. 2003a; High SPARC/better > grade 2; ducECM and intra-lobular Sarrio et al. 2008b; Watkins et al. survival tal > lobular stromal fibroblasts, Increased with EMT-asso2005) present in arteriole aTumor cells are ciated genes and capillary endonegative thelial cells b Melanoma (Alonso et al. 2007b; Ledda et al. High SPARC correlates Capillary endothelium Highly upregulated Upregulated in: Fibroblasts 1997a; Wessel et al. 2008a) with metastases and in malignant Endothelial cells EMT genes melanoma a - Negative endo, stroma a Upregulated in: Prostate cancer (Lapointe et al. 2004a; Thomas et al. Luminal and basal Primary tumor High SPARC correlates Fibromuscular cells 2000) epithelial cells, with increased grade stromal cells low-moderate; fibromuscular and metastasis Mets highly stromal cells positive Gastric carcinoma (Junnila et al. 2009; Takeno et al. Tumor cells are Negative-weak Increased SPARC/ 2008; Wang et al. 2004) positive decreased survival a Meningioma (Bozkurt et al. 2009; Rempel et al. Absent in normal Tumor cells are Upregulated in: Increased SPARC/ 1999a; Schittenhelm et al. 2006b; cerebral cortex positive Reactive decreased survival a astrocytes Zeltner et al. 2007) Increased SPARC/ increased recurrence b No association with grade
306 S. L. Thomas and S. A. Rempel
(Che et al. 2006; Luo et al. 2004; Wong et al. 2009; Xue et al. 2006b; Yamashita et al. 2003a)
(Koukourakis et al. 2003; Sosa et al. 2007)
Esophageal cancer
Lung cancer
Upregulated in all grades; Heterogeneous expression in higher grades Positive
Tumor tissue expression
Negative in clear cell renal carcinoma Positive in sarcomatoid renal carcinomas a Negative Positive b Negative tumor cells (positive in a few cases) Majority negative Weak SPARC in to weak bronchial epithelial cells, negative for alveolar cells, positive in alveolar Mo,+ chrondrocytes
Negative
(Chin et al. 2005; Kato et al. 2005a)
Head and Neck cancer
SPARC-negative tumor cells Renal Cell carcinoma (Amatschek et al. 2004; Gieseg et al. 2002; Sakai et al. 2001)
Absent in normal cerebral cortex
(Pen et al. 2007; Rempel et al. 1998; Rich et al. 2005)
Normal tissue expression
Glioma
Table 17.1 (continued) Cancer type References
Upregulated in: Stromal cells Endothelial cells
Upregulated in: Stromal cells
Variably positive: Clear cell stroma Highly positive: Sarcomatoid stroma Increased SPARC/ decreased survival Increased SPARC/ increased tumor mets Increased SPARC/ decreased survival
Increased SPARC/ decreased survival (stage II tongue cancer patients)
Increased SPARC, doublecortex, and Semaphorin 3B/ decreased survival
Upregulated in: Endothelial cells Pericytes Reactive astrocytes Upregulated in: Stroma a
Immunohistochemical or cDNA and tissue microarray correlation with grade or survival
Tumor-associated stromal and endothelial cells
17 SPARC and the Tumor Microenvironment 307
(Brown et al. 1999b; Mok et al. Positive- Thecal cell layer, 1996a; Paley et al. 2000; Yiu et al. granulosa cells, 2001a) Negative- stroma Positivea/Negativebsurface epithelial cells
Ovarian cancer
Ductal epithelial cells
(Brune et al. 2008; Hong et al. 2008; Infante et al. 2007; Sato et al. 2003)
Pancreatic cancer
Epithelial cells
(Chan et al. 2008; Wiese et al. 2007; Yang et al. 2007)
Strong cleaved 24 kD cleaved protein in hepatocytes
Normal tissue expression
Colon carcinoma
Hepatocellular cancer (Lau et al. 2006; Le Bail et al. 1999)
Table 17.1 (continued) Cancer type References
24 kD Upregulated in: Stromal fibrous area and sinusoidal regions Upregulated in: Stroma cells
Negative
27.7% negative, 56.5% weak, 15.8% positive 66%- positive 69% -negative stoma, 34% 31% -positive -negative Positive in low and intermediate grade intraductal papillary mucinous neoplasms Negative in high grade Negative in infiltrating tumor cells Low-Mostly absent Upregulated in: Stromal cells, matrix Endothelial cells
Tumor-associated stromal and endothelial cells
Tumor tissue expression
Decreased with increasing grade
Survival: Tumor-/stroma- > tumor+/stroma- > tumor-/stroma+> tumor +/stroma+
Increased SPARC/ Increased survival
Immunohistochemical or cDNA and tissue microarray correlation with grade or survival Stromal/sinusoidal staining increases with increasing grade
308 S. L. Thomas and S. A. Rempel
(Yamanaka et al. 2001)
(Dalla-Torre et al. 2006)
Bladder cancer
Osteosarcoma
Negative
Normal tissue expression
Identifies the references related to the indicated observations
(Rodriguez-Jimenez et al. 2007) (Chlenski et al. 2002)
Uterine cancer Neuroblastoma
a,b
(DiMartino et al. 2006)
AML
Table 17.1 (continued) Cancer type References
Decreased in cell lines with MLL rearrangements Negative Negative to minimal in undifferentiated tumorsNegative in differentiating tumors
Tumor tissue expression
Increased SPARC/ decreased survival
Upregulated in: Stromal cells
Increases with grade; Increased SPARC/ decreased survival High SPARC in almost all tumors; Increased SPARC/ decreased event-free and relapse-free survival
Negative in undif- Decreased SPARC/ Increased grade ferentiated tumors Positive in differentiating tumors
Immunohistochemical or cDNA and tissue microarray correlation with grade or survival
Tumor-associated stromal and endothelial cells
17 SPARC and the Tumor Microenvironment 309
310
S. L. Thomas and S. A. Rempel
more, immunohistochemical analyses of SPARC in tumor tissue sections and in xenograft specimens from animal studies, especially in Sparc-null and Sparc-wt mice, indicate that SPARC is expressed in non-tumor cells, which can also influence tumor progression. Correlations of SPARC expression patterns in stroma and/ or tumor with patient survival suggest that the outcome may be different based on where SPARC is expressed. For example, pancreatic cancer patient survival is greatest when tumors are negative for SPARC in both tumor and stromal compartments and the poorest when tumors are positive for SPARC in both compartments (Infante et al. 2007) (Table 17.1). Indeed, even when tumor cells are negative, the positive expression in stroma correlates with poorer survival (Infante et al. 2007). Therefore, the role of SPARC in tumor progression must be expanded to include its functions in both stromal and tumor compartments, and a determination of how these functions influence tumor progression is essential. The treatment approach will need to be tailored by the particular expression pattern in both the stroma and tumor (not just by the oncogene or tumor suppressor role in the tumor cells), and this further emphasizes the need for individualized tumor assessment to ascertain the best treatment strategy. In summary, the array and immunohistochemical data along with the correlative survival analyses suggest that SPARC functionally contributes to tumor progression. In this chapter we will review the in vivo and in vitro data that assess the expression of tumor and stromal SPARC and the influences of that expression on tumor progression.
17.2 Role of SPARC in the Tumor Microenvironment Determining the role of SPARC in the tumor microenvironment is complicated by the fact that SPARC is produced by multiple cell types and has multifaceted effects that regulate several processes involved in tumor formation and progression. SPARC can be produced and secreted by cancer cells, stromal cells, and immune cells resulting in autocrine and paracrine effects on the tumor microenvironment. The ability of SPARC to cause deadhesion of cells from the ECM results in tumor and stromal cell migration and dissemination. SPARC can also inhibit proliferation in cancer cells, fibroblasts, and endothelial cells by direct effects on the cells or indirectly by binding to various growth factors and inhibiting the ability of the growth factors to activate tyrosine kinase receptors. In addition, SPARC can regulate ECM deposition, assembly, and remodeling by regulating collagen processing and altering the secretion of matrix proteins and matrix degrading proteases from cancer cells and stromal cells. SPARC can also regulate the tumor microenvironment by suppressing immune cell infiltration into the tumor and the ability of the immune system to eliminate tumor cells. The evidence for each of these functions of SPARC in the regulation of the tumor microenvironment will be discussed in detail in the following sections.
17 SPARC and the Tumor Microenvironment
311
17.2.1 Effects on Cancer Cells There are many studies that have examined the role of SPARC in various types of cancers. The following sections will discuss the roles for SPARC in regulating cancer cell proliferation, cell cycle progression, survival and apoptosis, adhesion and migration, invasion and metastasis, ECM production, and signal transduction. 17.2.1.1 Tumor Growth in Animal Models In vivo xenograft models have been used to assess the effects of forced or suppressed SPARC expression on human cancer cell growth in immune compromised rodents (Table 17.2). For example, SPARC expression was forced in the U87 glioma cell line and clones secreting various amounts of SPARC were selected and implanted intracranially in nude rats. All SPARC-expressing clones had significantly reduced tumor volumes at day 7 compared to the parental glioma cell line (Schultz et al. 2002). When the high SPARC-secreting clone and the parental control cells were allowed to grow until the animals showed signs of neurological deficit, the animals with the SPARC-expressing tumors lived longer, 20 days compared to 9 days for the parental cell line (Schultz et al. 2002). In contrast, SPARC promoted tumor growth when a genetically engineered glioma cell line was injected subcutaneously (Rich et al. 2003). It is not known whether the difference in the proliferation effects reflects differences in the microenvironment (subcutaneous [sc] versus the brain) or the differences in cell lines used; one is derived from a glioma patient (U87) and the other from a human astrocyte genetically altered to express the T/tAg, hTERT, and H-rasV12G (THR) genes (Table 17.2). When the SKOV3 ovarian cancer cell line or the HEP3B hepatocellular carcinoma cell line with forced SPARC expression were injected subcutaneously in nude mice, SPARC significantly reduced tumor growth (Lau et al. 2006; Mok et al. 1996) (Table 17.2). In contrast, a group that suppressed SPARC in melanoma cell lines found the opposite effect. SPARC antisense strongly reduced tumor growth (Ledda et al. 1997b; Prada et al. 2007), with a complete inhibition of tumor formation resulting from tumor cell rejection in one study (Ledda et al. 1997b) (Table 17.2). These melanoma studies are interesting because the data suggest that the loss of tumor-expressing SPARC altered the microenvironment structurally in such a way as to permit neutrophil access, resulting in tumor regression. Therefore, in addition to having effects on tumor growth, SPARC might play a role in suppressing the immune system (see Sect. 17.2.4). As a consequence, several in vivo studies have examined the role of tumor and stromal SPARC on tumor growth by comparing cancer cells grown in immune competent Sparc-null versus Sparc-wt mice (Table 17.3). The results of these studies depend on the type of cancer cell used, with ovarian, prostate, pancreatic, lung, and T cell lymphoma showing increased cancer growth and progression in Sparc-null animals, whereas intestinal, mammary, and skin cancer cells have decreased growth and progression in Sparc-null animals.
U87 (Kunigal et al. 2006) SNB19 (Kunigal et al. 2006) U251MG (Seno et al. 2009) Ovarian SKOV3 (Mok et al. 1996) Hepatocellular HEP3B (Lau et al. 2006) Melanoma IIB-MEL-LES (Ledda et al. 1997b) IIB-LES-IAN (Ledda et al. 1997b) IIB-MEL-J (Prada et al. 2007)
Glioma U87 (Schultz et al. 2002) U87 (Yunker et al. 2008) THR (Rich et al. 2003) THR (Rich et al. 2003)
Forced
Suppressed sc
Suppressed sc
Forced
Positive
Positive
Positive
sc
sc
sc
?
Positive
Forced
Forced
Very low
Brain sc
Negative
Forced Forced
Very low Very low
Brain
Very low
Forced
Very low
Brain
Dorsal skin fold Forced Dorsal skin fold Suppressed Brain
Forced
Very low
Table 17.2 Animal models Nude/SCID rodents SPARC Status Xenograft Cancer type/ EndogManipulated site enous Cell lines Invasion
No effect
Suppressed
Suppressed
Decreased
Decreased
Increased
Decreased
Increased Increased mets
Decreased Increased (decreased)
Tumor Growth (proliferation)
Increased CNI
Increased CNI
ECM
Increased
Decreased Decreased
Increased
Increased
Increased
Decreased Decreased
Angiogen- MVD esis (VEGF)
survival
cells
Increased N Increased N Increased No effect F
Increased
Animal
Stromal
312 S. L. Thomas and S. A. Rempel
Suppressed sc
Forced
Positive
Negative
Intracardiac
Suppressed sc
Positive
sc
Forced
Positive
Decreased
Decreased
No effect
Tumor Growth (proliferation)
Decreased mets
Invasion Increased CNI
ECM Increased
Angiogen- MVD esis (VEGF)
survival
cells
Increased No effect F Increased Increased N Increased Increased N
Animal
Stromal
THR [T/t-Ag, hTERT, H-rasV12G], ?- unknown, sc subcutaneous injection, mets metastases, CN collagen ECM extracellular matrix, MVD microvascular density, VEGF vascular endothelial growth factor, N neutrophils, F fibroblasts
A375N (Prada et al. 2007) IIB-MEL-J (Prada et al. 2007) A375N (Prada et al. 2007) Breast cancer MDA-231 (Koblinski et al. 2005)
Table 17.2 (continued) Nude/SCID rodents SPARC Status Xenograft Cancer type/ EndogManipulated site enous Cell lines
17 SPARC and the Tumor Microenvironment 313
?
Positive
Null vs WT
Positive
Positive
Null vs WT
Null vs WT
Null vs WT
Null vs WT
Null vs WT
ID8-VEGF (Said and Socha 2007c) OSEID8 (Phelps et al. 2009) Prostate cancer TRAMP x Sparc-null/ WT (Said et al. 2009) Pancreatic PAN02 (Puolakkainen et al. 2004) PAN02 (Arnold et al. 2008)
?
ID-8 (Said and Null vs Socha 2007c) WT
Increased
Increased
Increased
Pancreas
sc Increased
Increased
Spontane- Increased ous
ip
ip, sc
ip, sc
Table 17.3 Xenograft and spontaneous animal models SPARC-null/ SPARC status Xenograft Tumor SPARC-wt site growth Cancer type/Cell Host Tumor Delivery lines cells Ovarian Null vs ? ip Increased ID8 (Said and WT Motamed 2005)
No change Decreased
Increaseda
Increased/ increased mets
Increased mets, MMP-2 & -9
Decreased Decreased
Decreased
Increased (increased)
Increased
Decreased P, increased Mo Decreased N
Decreased
Decreased
Decreased
Increased
Increased Mo
Decreased
(Increased) [increased]
MVD (VEGF) Host stromal Survival [R] cells
Decreased CNI, III
Decreased CNI
Decreased Decreased Increased CNl I, IIl nodular dissemination Increased MMP-2,-9; decreased TIMP1, 2
ECM
Increased
Invasion/ Metastasis
Apoptosis
Proliferation
aT1 (Sangaletti et al. 2008)
Null vs WT
Low
Table 17.3 (continued) SPARC-null/ SPARC status SPARC-wt Cancer type/Cell Host Tumor lines cells Positive PAN02 ±MMP9 Null vs WT (Arnold et al. 2008) Lewis lung Null vs Positive LLC (Brekken et al. 2003) WT T-cell lymphoma EL4 (Brekken Null vs Negative et al. 2003) WT Colon carcinoma Null vs Null vs APC Min/+ x WT WT Sparc-null/ WT (Sansom et al. 2007) Breast Positive N1G, N2C, N3D Null vs WT (Sangaletti et al. 2003) Positive leukocyte (San- Null vs WT galetti et al. 2003) Null vs Low 4T1 (Sangaletti WT et al. 2008) Mets-no change
Decreased CNI fibers and maturity
Mammary No change fat
iv
Restored ECM
iv Decreased
Decreased CN IV
Decreased CN
Decreased Increased L (decreased)
Decreased area Decreased [decreased] Mo
No change
Decreased
Decreased CN, LNI
MVD (VEGF) Host stromal Survival [R] cells
ECM
Mammary Decreased fat
Increased enterocyte migration
Increased
Increased
sc/iv
Spontane- Decreased #, ous not size
No change No change Increased
Increased
Invasion/ Metastasis
sc/iv
Apoptosis
Further No change No change Increased/ increased inhibited mets
Proliferation
Pancreas
Xenograft Tumor site growth Delivery
Null vs WT
UVDecreased induced
Xenograft Tumor site growth Delivery
Proliferation
Apoptosis
Invasion/ Metastasis
ECM
MVD (VEGF) Host stromal Survival [R] cells
WT wildtype, ?- unknown, sc subcutaneous injection, iv intravenous injection, ip intraperitoneal injection, # number, a increased cyclins A, D1 and decreased p21, 27, MMP matrix metalloproteinase, mets metastases, CN collagen, LN laminin, ECM extracellular matrix, MVD microvascular density, VEGF vascular endothelial growth factor, [R] VEGF-receptor(s), Mo macrophages, P pericytes, N neutrophils, L leukocytes
Squamous cell hairless x Sparc- Null vs WT null/WT (Aycock et al. 2004)
Table 17.3 (continued) SPARC-null/ SPARC status SPARC-wt Cancer type/Cell Host Tumor lines cells
17 SPARC and the Tumor Microenvironment
317
Three different studies found that the intraperitoneal implantation of ovarian cancer cell lines in Sparc-null mice resulted in increased tumor volume and reduced animal survival compared to cancer cells grown in Sparc-wt mice (Phelps et al. 2009; Said and Motamed 2005; Said and Socha 2007c) (Table 17.3). When a murine pancreatic adenocarcinoma cell line was injected either subcutaneously or into the pancreas, the resulting tumors grew larger in Sparc-null compared to Sparc-wt animals (Arnold et al. 2008; Puolakkainen et al. 2004). Another study used the transgenic adenocarcinoma of mouse prostate (TRAMP) model of primary prostate carcinogenesis interbred with Sparc-null and Sparc-wt mice. In this model, Sparc-null mice had accelerated cancer development and progression. A SPARCexpressing cell line derived from this model grew significantly larger tumors when subcutaneously implanted into Sparc-null mice and 100% of mice developed tumors compared to 60% in Sparc-wt mice (Said et al. 2009). In addition, two other types of cancer cells, Lewis lung carcinoma and T-cell lymphoma, grew larger and more rapidly when subcutaneously implanted into Sparc-null compared to Sparc-wt mice (Brekken et al. 2003). On the other hand, several studies found that absence of SPARC in the stroma hindered tumor development and progression (Table 17.3). In one study, a mouse mammary carcinoma cell line was injected into the mammary fat pads in Sparc-null and Sparc-wt mice. Sparc-null mice had smaller tumors with undefined lobules and reduced tumor outgrowth (Sangaletti et al. 2003). Another study used a model of spontaneous intestinal tumorigenesis in which Sparc-null mice had significantly fewer adenomas in both the small and large intestines (Sansom et al. 2007). A third report of reduced tumor formation in Sparc-null mice used UV irradiation to induce squamous cell carcinoma in hairless mice crossed with Sparc-null and Sparcwt mice. All 20 Sparc-wt mice developed papillomas and some developed larger squamous cell carcinomas; whereas only six Sparc-null mice developed a small number of papillomas resulting in a highly significant reduction in tumor formation in Sparc-null animals (Aycock et al. 2004). These data suggest that stromal SPARC suppresses tumor growth in the ovary, prostate, pancreas and lung, whereas it is advantageous for tumor growth in the colon, breast, and skin (Fig. 17.1). These combined in vivo studies indicate that SPARC positively or negatively contributes to many aspects of tumor progression, and in vitro and in vivo studies have been utilized to further characterize its role in regulating tumor cell proliferation, cell cycle progression, survival, adhesion, migration, invasion, and ECM deposition, as follows. 17.2.1.2 Cell Proliferation Overwhelmingly, in vitro studies provide evidence that SPARC reduces proliferation in cancer cells, including a Lewis lung carcinoma cell line (Brekken et al. 2003), pancreatic cancer cell lines (Arnold et al. 2008; Sato et al. 2003), a mouse ovarian cancer cell line (Said and Motamed 2005; Said and Socha 2007c), human ovarian cancer cell lines (Phelps et al. 2009; Mok et al. 1996; Yiu et al. 2001), and
318
S. L. Thomas and S. A. Rempel
AML cell lines with MLL gene rearrangements (DiMartino et al. 2006). This reduction occurs whether the SPARC expression was forced or SPARC was given exogenously (Table 17.4). Further investigation using ovarian cancer cell lines show that exogenous SPARC was also able to inhibit alpha v integrin and beta 1 integrin stimulated proliferation of human ovarian cancer cell lines and a mouse ovarian cancer cell line on different ECMs (Said and Motamed 2005), and decrease LPA and IL-6 stimulated proliferation of a mouse ovarian cancer cell line (Said and Socha 2007c). Interestingly, these cell lines correspond to the human tumor specimens that have low SPARC in the tumor cells but high SPARC in the stroma (Table 17.1), and for which in vivo studies suggest that stromal (exogenous) SPARC suppresses tumor growth (Table 17.2). Outcomes in proliferation are more varied (Table 17.5) when examining the effects of forced SPARC versus suppressed SPARC on cell lines from tumors that normally have high SPARC expression (Table 17.1). Generally, forced SPARC expression was associated with reduced proliferation in a breast cancer cell line (Koblinski et al. 2005), melanoma cell lines (Prada et al. 2007), prostate cancer cell lines (Said et al. 2009), and glioma cell lines (Golembieski et al. 2008; Rempel et al. 2001; Vadlamuri et al. 2003). However, the suppression was not always observed in cells grown as monolayers. In one report using a glioma cell line, SPARC expression significantly delayed growth during the log phase growth of the cells but did not inhibit overall cell proliferation (Golembieski et al. 2001). When these same glioma cells were grown in vivo, there was a reduced MIB-1 proliferation index in SPARCexpressing tumors with a greater reduction in proliferation at the invading tumor edge compared to the tumor core (Schultz et al. 2002) (Table 17.2). SPARC overexpression in melanoma cells (Prada et al. 2007) and breast cancer cells (Koblinski et al. 2005) inhibited proliferation in spheroids but not in monolayer. Although the results of adding SPARC to cells already expressing SPARC appears to produce more variable results regarding proliferation, the reduction of SPARC with antisense in some of these same cell lines produced more consistent results. Inhibition of SPARC increased melanoma proliferation as spheroids (Prada et al. 2007), increased prostate cancer cell proliferation in monolayer (Said et al. 2009), and increased brain tumor cell proliferation under stress conditions in vitro (Shi et al. 2007). Interestingly, a study found that, while exogenous SPARC did decrease normal endothelial cell proliferation, and had biphasic effects on fibroblast proliferation, the melanoma, glioma and colorectal cell lines having SPARC expression were relatively unaffected by exogenous SPARC (Haber et al. 2008). This suggests that human tumors positive for SPARC (Table 17.1) may be less affected by SPARC expression in the stroma. Indeed, it has been reported that SPARC expressed by melanoma tumor cells and not the stromal cells is important for tumor growth (Prada et al. 2007). However, other studies did not detect a difference in proliferation with SPARCexpression in cancer cells. A pancreatic cancer cell line that was treated with exogenous SPARC did not show a significant reduction in proliferation rate (Puolakkainen et al. 2004) (Table 17.4). In addition, it was reported that the forced expression
Table 17.4 In vitro effects of SPARC on proliferation, cell cycle, apoptosis or attachment, adhesion, migration and invasion of SPARC-negative tumor cells Tumor type Cell lines used EndogMe Manipulated Proliferation (cell cycle) Attachment/adhesion/migration/invasion enous SPARC [apoptosis] SPARC Status Forced Exogenous Lewis lung Brekken et al. LLC Positive X Decreased (2003) Pancreatic Positive Decreased Migration increased by MMP-9 Arnold et al. (2008) PAN02 (mouse) ± MMP9 (transwell) Puolakkainen et al. PAN02 (mouse) Positive X No change [increased with (2004) Staurosporine] Sato et al. (2003) AsPC1 Negative M X Decreased BxPC3 Negative M Panc1 Slight H X Decreased U Fibroblasts High CFPAC1 Negative Ovarian Mok et al. (1996) SKOV3 Negative M X Decreased Yiu et al. (2001) MESO High U (mesothelial) HOSE (epithelial) High U X Decreased [none] SKOV3 Negative M X Decreased more [increased] OVCA420 Negative OVCA429 Low OVCA433 Negative X Decreased more DOV13 Interme- U X Decreased more diate
17 SPARC and the Tumor Microenvironment 319
Said et al. (2007a)
Said et al. (2007a)
Said and Socha (2007c)
Said and Motamed (2005)
M
Negative
Negative
?
NIH:OVCAR3
IGROV1
M
U
? Positive
LP9 (mesothelial) Meso 301 (mesothelial) SKOV3
X
X
X
X X
X
M
Negative
NIH:OVCAR3
X
X
M
Negative
X
X
X
X
? Positive
M
X
ID8-VEGF Meso 301 (mesothelial) SKOV3
Negative
NIH:OVCAR3
M
?
Negative
SKOV
X
Inhibits integrin-induced (Pl, FN, CN I) Inhibits integrin-induced (Pl, FN, CN I) Inhibits integrin-induced Pl, FN, CN I)
Decreased; inhibits LPA and IL-6 induced Decreased; inhibits LPA and IL-6 induced
Inhibits VEGF-induced
inhibits
Decreased [increased]
Manipulated Proliferation (cell cycle) SPARC [apoptosis] Status Forced Exogenous
Me
ID8
?
Endogenous SPARC
ID8
Table 17.4 (continued) Tumor type Cell lines used
Decreased integrin-mediated attachment to FN, VN, CN I Decreased integrin-mediated attachment to FN, VN, CN I Decreased integrin-mediated attachment to FN, VN, CN I
decreased chemotaxis and invasion; decreased ERK 1/2 and pAKT decreased chemotaxis and invasion; decreased ERK 1/2 and pAKT
Decreased MMP-2 & MMP-9
Decreased adhesion/invasion FN, CN I, CN IV, LN, VN, HA (transwell) Decreased adhesion/invasion FN, CN I, CN IV, LN, VN, HA (transwell) Decreased adhesion/invasion FN, CN I, CN IV, LN, VN, HA (transwell)
Attachment/adhesion/migration/invasion
320 S. L. Thomas and S. A. Rempel
M
Negative
Negative
NIH:OVCAR3
X X
U M M M
Positive Low Negative
Negative
Negative
X
X
Positive
Kasumi-1 WT MLL ME-1 WT MLL KG1a WT MLL ML-2 rearranged MLL MV411 rearranged MLL THP-1 rearranged MLL
X
X
X
Positive Negative Positive Positive
X X X X
X
X
X
OSEID8 (mouse) SKOV3 ES2 HCC60 M
M
U
Positive
Decreased
No change (no change) Decreased (increase in G1)
No change
Decreased No change Decreased No change
Decreased MCP-I induced
Decreased MCP-I induced
Manipulated Proliferation (cell cycle) SPARC [apoptosis] Status Forced Exogenous
Me
Meso 301 (mesothelial) SKOV3
Endogenous SPARC
Suppressed MCP1-induced invasion (transwell); reduced MMP-9 and uPA activity Suppressed MCP1-induced invasion (transwell); reduced MMP-9 and uPA activity
Attachment/adhesion/migration/invasion
AML acute myelogenous leukemia, MLL mixed-lineage leukemia gene, ?- unknown, U unmethylated, M methylated, LPA lysophosphatidic acid, IL-6 interleukin-6, MCP-1 monocyte chemoattractant protein, VEGF vascular endothelial growth factor, CN collagen, LN laminin, VN vitronectin, HA hyaluronic acid, MMP matrix metalloproteinase, uPA urokinase plasminogen activator, ME methylation
AML DiMartino et al. (2006)
Phelps et al. (2009)
Said et al. (2008)
Table 17.4 (continued) Tumor type Cell lines used
17 SPARC and the Tumor Microenvironment 321
Melanoma/colon/glioma Haber et al. (2008)
Prada et al. (2007)
Melanoma Ledda et al. (1997b)
Breast cancer Koblinski et al. (2005)
X X X X X X X X
Positive Negative Positive Negative Positive Negative Negative Low
X X
HMEC-1 (endothelial) MEC(endothelial) HUVEC (endothelial) WI38 (fibroblasts) HLF1 (fibroblasts) LoVo (colorectal) Hct116 (colorectal) U87 (glioma)
X X
X
X
Positive Positive Positive Positive
Positive
IIB-MEL-IAN
X
A375N IIB-MEL-J A375N J IIB-MEL-J
Positive Positive
Negative
IIB-MEL-J IIB-MEL-LES
MDA-231-GFP
No change
Biphasic Biphasic No change No change
Decreased Decreased
Decreased
Decreased in spheroids Decreased in spheroids Increased in spheroids Increased in spheroids
No change
Decreased attachment, invasion (Matrigel) and migration (transwell) decreased MMP-9 Decreased attachment and invasion (Matrigel)
Decreased in spheroids Decreased adhesion; decreased inva[no change] sion (Matrigel); decreased MMP-2
Table 17.5 In vitro effects of SPARC on proliferation, cell cycle, apoptosis or attachment, adhesion, migration and invasion of SPARC-positive tumor cells Tumor type Cell lines used Endogenous Manipulated SPARC Proliferation (cell Attachment/adhesion/migration/invasion SPARC Forced SupExog- cycle) [apoptosis] pressed enous
322 S. L. Thomas and S. A. Rempel
Prostate Said et al. (2009)
Uveal Melanoma Maloney et al. (2009)
Table 17.5 (continued) Tumor type
Low Negative Positive
DU145
PC3
LNCaP
X
X
X
X
Positive Positive
X X
X
Positive Positive Positive
X
Positive
X
X
Positive
SP6.5, MKTBR, OCM-1
X
Positive
IIB-MEL-J (melanoma) IIB-MEL-LES (melanoma) IIB-MEL-LES (melanoma) BLAST IIB-MEL-LES (melanoma) shRNA A375 (melanoma) MEL-888 (melanoma) SB-2 (melanoma)
Attachment/adhesion/migration/invasion
Decreased (decreased Decreased invasion (Matrigel) cyclin A, DI increased p21, p27) Decreased (decreased Decreased invasion (Matrigel) cyclin A, DI increased p21, p27) Decreased (decreased Decreased invasion (Matrigel) cyclin A, DI increased p21, p27)
Decreased
No change
No change No change
No change (inhibits re-entry)
No change
No change
Manipulated SPARC Proliferation (cell Forced SupExog- cycle) [apoptosis] pressed enous
Endogenous SPARC
Cell lines used
17 SPARC and the Tumor Microenvironment 323
Low Negative Positive Positive Low Negative Positive Positive Low
Low Low Low
DU145 PC3 LNCaP TRAMP1, -2.-3 DU145 PC3 LNCaP TRAMP1, -2.-3
U87
U87
U87
U87
U87
SNB19 THR
Glioma Vadlamuri et al. (2003)
Golembieski et al. (2001)
Golembieski et al. (2008)
McClung et al. (2007)
Kunigal et al. (2006)
Rich et al. (2003)
Low
Low
Endogenous SPARC
Cell lines used
Table 17.5 (continued) Tumor type
X X
X
X
X
X
X X X X X X X
No change No change (no change)
No change
Decreased
Decreased (variable depending on ECM) Delayed
Increased Increased Increased Increased Decreased Decreased Decreased No change
Manipulated SPARC Proliferation (cell Forced SupExog- cycle) [apoptosis] pressed enous
Increased invasion (spheroids) and adhesion & migration LN,CN.HA; biphasic TN Increased invasion (Matrigel) and migration (wound assay, transwell); pP38/pHSP27 Increased MT1-MMP, MMP2, cleaved galectin-3 Increased invasion (Matrigel) increased MMP-9, uPA/uPAR & Rho activation Increased invasion (Matrigel) Increased (Matrigel); increased MMP-3
Variable depending on ECM
Increased invasion (Matrigel) Increased invasion (Matrigel) Increased invasion (Matrigel) Increased invasion (Matrigel)
Attachment/adhesion/migration/invasion
324 S. L. Thomas and S. A. Rempel
Positive Positive
U251MG
U373MG
X
X
X X X X
X
X
X
X
X
X
X
[Decreased under stress; increased pAKT] [Decreased under stress; increased pAKT] [Decreased under stress; increased pAKT] [Increased; decreased FAK & ILK activity] [Increased; decreased FAK & ILK activity]
Manipulated SPARC Proliferation (cell Forced SupExog- cycle) [apoptosis] pressed enous
Decreased invasion (Matrigel), decreased migration (transwell) Decreased invasion (Matrigel), decreased migration (transwell)
Decreased invasion (Matrigel)
Decreased invasion (Matrigel)
Increased (Matrigel) Increased (Matrigel) Increased (Matrigel)
Attachment/adhesion/migration/invasion
[THR T/t-Ag, hTERT, H-rasV12G], CN collagen, LN laminin, TN tenascin, HA hyaluronic acid, ECM extracellular matrix, MMP matrix metalloproteinase, uPA urokinase plasminogen activator, uPAR uPA receptor, HSP27 heat shock protein 27
Seno et al. (2009)
Positive
U373MG
Positive
D54MG Positive
Positive
THR
Shi et al. (2004)
D54MG
Low Positive Positive Low
U87 U251MG D54MG U87
Shi et al. (2007)
Endogenous SPARC
Cell lines used
Table 17.5 (continued) Tumor type
17 SPARC and the Tumor Microenvironment 325
326
S. L. Thomas and S. A. Rempel
of SPARC in glioma cell lines (Kunigal et al. 2006) (Table 17.5) did not result in changes in proliferation. A review of the methods used suggests that some of the variability of SPARC effects on proliferation may result from the different assays used, different time points analyzed, and the use or not of different ECM substrates. 17.2.1.3 Cell Cycle Progression Whereas studies of normal cells indicate that SPARC delays the onset of S phase (Funk and Sage 1991), reports of the effect of SPARC on cell cycle progression in cancer cell lines is less consistent. It was found that exogenous SPARC inhibited the progression from G1 phase to S phase in an AML cell line with an MLL gene rearrangement, but had no effect on an acute myeloid leukemia cell line with a wild-type MLL gene (DiMartino et al. 2006) (Table 17.4). A different study used SPARC purified from human melanoma cells to assess the effect on cell cycle in five melanoma cell lines, two colon cancer cell lines, and one glioma cell line. In all the cell lines tested, exogenous SPARC did not alter the onset of S phase (Haber et al. 2008). The same group reported that the use of shRNA or antisense to reduce the endogenous level of SPARC in a melanoma cell line resulted in an inhibition of cell cycle reentry (Haber et al. 2008) (Table 17.5). As discussed above, these cells appear to be refractory to exogenous SPARC. Another group reported a biphasic effect of SPARC on cell cycle that depended on the amount of forced SPARC expression in glioma cells (Table 17.5). SPARCexpressing clones with high and intermediate levels of SPARC had a higher percentage of cells in G0/G1 than parental control cells, whereas the SPARC-expressing clone with the lowest amount of SPARC had a lower percentage of cells in G0/G1 and a higher percentage of cells in G2/M than parental control cells (Golembieski et al. 2001), suggesting that the level of secreted SPARC may dictate how far a cell progresses through the cell cycle. The same group later reported that the regulation of cell cycle progression by SPARC was also dependent on the type of ECM that the glioma cells were grown on. SPARC-expressing clones with high and intermediate levels of SPARC had fewer cells in G0/G1 when grown on vitronectin, fibronectin, hyaluronic acid, or collagen (Vadlamuri et al. 2003). These results indicate that the amount of SPARC expressed endogenously by the cells, the substrate they are grown on, and the presence of genetic mutations that regulate cell cycle may determine whether or not SPARC has an effect on cell cycle progression. 17.2.1.4 Cell Survival/Apoptosis Since SPARC was found to reduce proliferation in various types of cancer cells, some investigators looked further to determine if decreased cell numbers resulted from a reduction in cell survival, as indicated by an increase in apoptosis. In a pancreatic cancer cell line, SPARC alone did not induce apoptosis in vitro, but when
17 SPARC and the Tumor Microenvironment
327
SPARC was combined with staurosporine the cells had increased active caspase-3, indicating that SPARC may prime the cell for apoptosis (Puolakkainen et al. 2004) (Table 17.4). The same pancreatic cancer cell line was grown subcutaneously in Sparc-null and Sparc-wt mice. A loss of stromal SPARC resulted in fewer tumor cells positive for active caspase-3, suggesting that stromal SPARC is supportive of apoptosis (Prada et al. 2007). Further support for this role was observed when exogenous SPARC treatment resulted in significantly increased apoptosis in an ovarian cancer cell line (Said and Motamed 2005; Yiu et al. 2001) (Table 17.4). In contrast, glioma cells with forced expression of SPARC were found to have a survival advantage under the stress of serum withdrawal. The increase in cell number was not due to increased proliferation. Rather, analysis revealed that SPARC significantly decreased apoptosis in three glioma cell lines, and sensitivity to apoptosis could be restored with a PI3K inhibitor or an Akt inhibitor, indicating that the Akt pathway is involved in the increased survival induced by SPARC (Shi et al. 2004) (Table 17.5). The same group later reported that reducing SPARC expression in glioma cell lines with siRNA resulted in increased apoptosis under conditions of serum withdrawal (Shi et al. 2007) (Table 17.5). Other reports indicate that SPARC does not have an effect on apoptosis in cancer cells. One study found that exogenous SPARC did not induce apoptosis in acute myeloid leukemia cell lines with MLL gene rearrangements even though SPARC reduced the growth of these same cell lines (DiMartino et al. 2006) (Table 17.4). Another study reported that tumors formed by Lewis lung carcinoma cells grown in Sparc-null mice did not show a difference in apoptosis compared to tumors in Sparc-wt mice (Brekken et al. 2003) (Table 17.4). Similarly, a breast cancer cell line with forced expression of SPARC had no increase in apoptosis even though colony size and formation were reduced when the cells were grown in Matrigel (Koblinski et al. 2005) (Table 17.5). Combined, the reports suggest that SPARC plays a role in survival under stressful conditions, such as with serum withdrawal or drug treatment (discussed further in Sect. 17.3), and that for some cancers, stromal SPARC may influence tumor cell survival. 17.2.1.5 Cell Adhesion/Migration Since SPARC was shown to have a counter-adhesive function in normal cells (Motamed and Sage 1998) many investigators have examined the ability of SPARC to alter adhesion and migration in cancer cells. For three ovarian cancer cell lines, exogenous SPARC produced a significant inhibition of adhesion to several ECM proteins including fibronectin, collagen I, collagen IV, vitronectin, hyaluronic acid, and laminin (Said and Motamed 2005) (Table 17.4). A later study by the same group found that SPARC significantly reduced the basal adhesion of three ovarian cancer cell lines to fibronectin and vitronectin. Agonists for integrins alpha v and beta 1 increased the adhesion of these cell lines to ECM proteins and mesothelial cells, and SPARC was able to inhibit the integrin-mediated adhesion (Said et al. 2007b)
328
S. L. Thomas and S. A. Rempel
(Table 17.4). In contrast, another group found that when melanoma cell lines express antisense for SPARC there is a strong reduction in the ability of the cells to adhere to Matrigel (Ledda et al. 1997b) (Table 17.5). These different results may be due to the differences in substrate used for these studies. For example, a study using glioma cells demonstrated that the effects of SPARC on attachment are ECM protein-specific. Glioma cells with forced expression of SPARC had increased attachment to laminin, collagen, and hyaluronic acid, a biphasic attachment to tenascin, and no change in attachment to vitronectin and fibronectin (Golembieski et al. 2001) (Table 17.5). The changes in cancer cell adhesion induced by SPARC may result in changes in the ability of the cells to migrate. A study using a glioma cell line with forced expression of SPARC found that SPARC expression increased migration in three different assays (Golembieski et al. 2008; Rempel 2001) (Table 17.5). However, when the migration of glioma clones that express different amounts of SPARC were compared, there was a biphasic effect on migration such that the clone that secreted the most SPARC migrated the farthest, the clone with intermediate SPARC migrated the least, and the clone with the lowest SPARC migrated the most like the parental control cells (Rempel et al. 2001). Therefore, other conditions being equal, the amount of expressed and secreted SPARC appears to influence the extent of migration as well. Corroborating data demonstrate that anti-sense suppression of SPARC in glioma cells resulted in decreased migration (Seno et al. 2009) (Table 17.5). Similarly, antisense suppression of SPARC reduced the migration of a melanoma cell line (Ledda et al. 1997b) (Table 17.5). These observations emphasize that the effects of SPARC on tumor cell attachment, adhesion, and migration are influenced by the integrin expression profile of the cells and the ECM in the immediate microenvironment. 17.2.1.6 Invasion/Metastasis The effects of SPARC on invasion appear to depend on the type of cancer being studied with the literature providing evidence that SPARC increases invasiveness of glioma and melanoma and decreases invasiveness of breast, prostate, and ovarian cancers (Fig. 17.1). A study with melanoma cells found that antisense for SPARC reduced the ability of the cells to invade Matrigel by 70–80% in two cell lines, and the third cell line, which had the greatest suppression of SPARC, was completely noninvasive (Ledda et al. 1997b) (Table 17.5). Similarly, four in vitro studies provide evidence that SPARC increases the invasive potential of glioma cells. Glioma cells with forced expression of SPARC were found to have an increased ability to invade rat fetal brain aggregates (Golembieski et al. 1999) (Table 17.5), and Matrigel (Golembieski et al. 2008; Kunigal et al. 2006; Rich et al. 2003) (Table 17.5). Another study found that when two glioma cell lines that naturally express SPARC were treated with SPARC siRNA, invasion through Matrigel was inhibited by approximately 50% (Shi et al. 2007) (Table 17.5). In addition, when SPARC-expressing glioma cell lines and the parental control cell line were xenografted intracranially
17 SPARC and the Tumor Microenvironment
329
in nude rats, control cells gave rise to a well-circumscribed tumor mass whereas SPARC-expressing cells produced invasive tumors with invasion into the adjacent brain and along the corpus callosum. SPARC expression was detected in invading single cells and distant tumor masses (Schultz et al. 2002) (Table 17.5). The opposite approach demonstrated that SPARC suppression decreased glioma invasion (Seno et al. 2009) in vitro and in vivo (Tables 17.2, 17.5, respectively) In contrast, studies of other cancer types provide evidence that SPARC can also suppress invasion. Forced endogenous expression of SPARC was found to significantly inhibit invasion through Matrigel in a breast cancer cell line (Koblinski et al. 2005) (Table 17.5) and several prostate cancer cell lines (Said et al. 2009) (Table 17.5). In agreement with the latter study, five prostate cancer cell lines were treated with siRNA resulting in a significant increase in invasion in vitro (Said et al. 2009) (Table 17.5). Furthermore, several studies with ovarian cancer cell lines reported that SPARC suppresses invasion. Forced SPARC expression in several cell lines significantly inhibited the ability of the cells to invade various ECM proteins (Said and Motamed 2005; Said and Socha 2007c) (Table 17.4). The balance of SPARC expression between the environment versus the tumor cells may play an important role in regulating the invasive phenotype. SPARC expression in ovarian cancer cell lines was able to inhibit invasion induced by macrophages or macrophage chemoattractant protein-1 (MCP-1) (Said et al. 2008) (Table 17.4), and Sparc-null ascitic fluid increased cell invasiveness compared to Sparc-wt ascitic fluid (Said and Socha 2007c). Further evidence that exogenous or stromal SPARC can suppress invasion comes from an in vivo study in which pancreatic cancer cells grown in SPARC-null mice were more invasive with greater involvement of local organs compared to tumors in Sparc-wt mice (Arnold et al. 2008) (Table 17.3). Although different tumor types displayed different invasive phenotypes depending on SPARC expression, there was good correlation between the effects of SPARC on migration and invasion. When measured, SPARC-induced migration correlated with SPARC-induced invasion and vice versa, SPARC-suppressed migration correlated with SPARC-suppressed invasion (Table 17.5). The effects of SPARC on metastasis also differ depending not only on the cancer type, but also on the SPARC status in the microenvironment, and on where the cancer cells are injected. It was reported that intracardiac injection of a breast cancer cell line that was forced to express a high level of SPARC resulted in a significant reduction in the overall number of metastases and a decrease in the number of metastases to all organs analyzed (Koblinski et al. 2005) (Table 17.2). Another group used two different models to assess the metastasis of a breast cancer cell line in Sparc-null and Sparc-wt mice. If the cells were injected into the mammary fat pad, then metastases to the lungs, liver, lymph nodes, and brain were significantly reduced in Sparc-null animals (Sangaletti et al. 2008). However, if the cells were injected into the tail vein, then numerous metastases were seen in the lungs of both Sparc-null and Sparc-wt mice suggesting that stromal SPARC is involved in the ability of the cells to leave the primary tumor but not to seed in a different organ (Sangaletti et al. 2008).
330
S. L. Thomas and S. A. Rempel
A group using the TRAMP model of primary prostate carcinogenesis crossed with Sparc-null versus Sparc-wt mice found that a higher percentage of Sparc-null animals had distant metastases in the lung, liver, and lymph nodes (Said et al. 2009) (Table 17.3), suggesting that stromal SPARC is suppressive to metastatic invasion. In contrast, when glioma cell lines with forced expression of SPARC, which are not metastatic when implanted in the brain, were exposed to a different microenvironment by subcutaneous injection into the flank of mice, 50% of mice with SPARCexpressing tumors developed intrathoracic and/or intraperitoneal metastases compared to only 1 out of 40 mice with control or parental tumors (Rich et al. 2003) (Table 17.2). These studies indicate that the ability of SPARC to induce a metastatic phenotype in cancer cells may depend on the initial microenvironment in which the tumor develops. 17.2.1.7 ECM Production and Deposition The proliferation and migration potential of cancer cells expressing or exposed to SPARC may be influenced by alterations in the ECM that are present in the tumor microenvironment. Sparc-null fibroblasts were found to have an increased amount of processed collagen I but less collagen I incorporated into the ECM (Rentz et al. 2007). In addition to fibroblasts, cancer cells can also produce ECM components and regulate ECM composition in the tumor environment. There are many studies that report alterations in ECM when cancer cells are grown in Sparc-null compared to Sparc-wt mice. By far the major finding in these studies is that the amount of collagen fibers is reduced and the collagen is less mature in the tumor microenvironment in Sparc-null animals (Arnold et al. 2008; Brekken et al. 2003; Puolakkainen et al. 2004; Said et al. 2009; Said and Motamed 2005; Sangaletti et al. 2008; Sangaletti et al. 2003) (Table 17.3). When specific types of collagen were examined, collagens I, III, and IV were found to be reduced in Sparc-null tumors (Said et al. 2009; Said and Motamed 2005; Sangaletti et al. 2003). Two other ECM components, laminin and fibronectin, were also found to have reduced expression in Sparc-null tumors. One study in which a breast cancer cell line was injected into the mammary fat pad in Sparc-null and Sparc-wt mice found that fibronectin was reduced in Sparc-null tumors in addition to a reduction in collagen (Sangaletti et al. 2008). Another study reported reduced staining for laminin 1 in the outer region of Lewis lung carcinoma tumors which coincided with the region of the tumor where differences in SPARC expression were seen in Sparc-null mice (Brekken et al. 2003). Further evidence that SPARC regulates collagen abundance in the ECM comes from studies in which SPARC is overexpressed in cancer cells. Glioma cells with forced expression of SPARC produced tumors with increased collagen I staining and more mature collagen fibers compared to tumors from parental cells (Yunker et al. 2008) (Table 17.2). SPARC hyperexpression in melanoma cells led to a twofold increase in collagen deposition at the tumor-host tissue interface (Prada et al. 2007) (Table 17.2).
17 SPARC and the Tumor Microenvironment
331
Additional studies analyzed the in vitro collagen expression by tumor cell lines. Exogenous SPARC reduced collagen I synthesis by an osteosarcoma cell line, whereas the production of laminin was downregulated by exogenous SPARC in several different cancer cell lines (Kamihagi et al. 1994). Antisense-mediated reduction of SPARC expression in a melanoma cell line resulted in increased collagen I expression and decreased collagen V expression as detected by proteomic analysis of CM (Sosa et al. 2007). These studies provide evidence that SPARC plays a key role in ECM production and composition during tumor development. Such changes can significantly affect the ability of immune cells to infiltrate the tumor microenvironment, which in turn can have profound effects on tumor growth and progression (see Sect. 17.2.4). 17.2.1.8 Signal Transduction Based on the above review, it is clear that SPARC can positively or negatively regulate many aspects of tumor growth and progression, and this has many consequences regarding its utility as a therapeutic agent or target (See Sect. 17.3). For example, in tumor cells where SPARC inhibits proliferation but promotes tumor migration and invasion, targeting SPARC may very well provide the desired reduction in invasion, but it may also promote an undesired increase in proliferation. An understanding of how SPARC regulates these different aspects of tumor progression may permit strategies to selectively inhibit one SPARC-regulated pathway, but retain its regulation of others. As reviewed below, many insights have been made in understanding SPARC-induced signal transduction in governing cell proliferation, cell cycle progression, apoptosis, cell attachment, migration and invasion. SPARC has been shown to regulate signaling pathways that are involved in cell proliferation. In a model of spontaneous prostate cancer progression, neoplastic tissue in Sparc-null mice showed a significant increase in cyclins A1 and D1 and a decrease in p21 and p27, which likely contributes to the increased proliferation index in these tissues (Said et al. 2009) (Table 17.3). In the inverse experiment, forced expression of SPARC in prostate cancer cells resulted in decreased cyclins A1 and D1 and increased p21 and p27 (Said et al. 2009) (Table 17.5). Thus, tumorand stromal-derived SPARC suppressed cell-cycle progression. The ability of SPARC to enhance cancer cell survival was found to involve the Akt signaling pathway. In glioma cell lines, exogenous SPARC induced Akt phosphorylation and kinase activity and PI3K inhibitors blocked the induction of Akt by SPARC (Shi et al. 2004) (Table 17.5). In addition, a PI3K inhibitor or an Akt inhibitor restored the sensitivity of SPARC-expressing glioma cells to apoptosis upon serum withdrawal (Shi et al. 2004). The forced endogenous expression of SPARC in glioma cells or exogenous SPARC-treated glioma cells had increased FAK phosphorylation and ILK activation (Shi et al. 2007). SPARC-mediated Akt activation, cell survival, and invasion were found to be dependent on FAK and ILK activation (Shi et al. 2007) (Table 17.5).
332
S. L. Thomas and S. A. Rempel
Since SPARC was found to regulate invasion in cancer cells and regulate the processing and deposition of collagen and other ECM proteins, several studies have analyzed the role of SPARC in modulating the expression of matrix degrading proteases. The literature demonstrates that SPARC is associated with an increase in matrix metalloproteinases (MMPs) and urokinase plasminogen activator (uPA) and its receptor (uPAR) expression in cancer types in which SPARC increases invasion and, conversely, is associated with a decrease in MMP and uPA expression in other cancers in which SPARC suppresses invasion and metastasis. Studies with glioma cell lines that are forced to express SPARC have found that the expression and activity of MMP-9 (Kunigal et al. 2006; Rich et al. 2003), MMP-2 (McClung et al. 2007), uPA/uPAR (Kunigal et al. 2006), as well as the expression of MMP-3 (Rich et al. 2003) and MT1-MMP (McClung et al. 2007) are increased with SPARC (Table 17.5). Glioma cells that express SPARC were also found to have increased cleavage of galectin-3 in conditioned media, which is an indictor of protease activity as galectin-3 can be cleaved by MMP-2, MMP-9, and MT1-MMP (McClung et al. 2007). An inhibitor of MMP-3 reduced the ability of SPARC-expressing glioma cells to invade Matrigel (Rich et al. 2003). Another group reported that in a melanoma cell line, as SPARC expression is reduced with antisense, MMP-2 mRNA levels and activity are reduced accordingly (Ledda et al. 1997b) (Table 17.5). However, SPARC was shown to decrease MMP expression and activity in ovarian cancer cells, which have reduced invasion with SPARC. In ovarian cancer cells, forced SPARC expression reduced MMP-9 and uPA activity in two different cell lines (Said et al. 2008) (Table 17.4) and treatment of cells with exogenous SPARC reduced MMP-2 and MMP-9 mRNA (Said and Socha 2007c) (Table 17.4). The inhibitory effect of SPARC on invasion was reinforced using SPARC-null mice, in which loss of stromal SPARC correlated with increased MMP expression. For example, ovarian tumors in Sparc-null mice had increased MMP-2 and MMP-9 mRNA and decreased MMP inhibitors TIMP-1 and TIMP-2 (Said and Socha 2007c) (Table 17.3), and Sparc-null ascitic fluid had enhanced MMP-2 and MMP-9 proteolytic activity (Said and Motamed 2005; Said and Socha 2007c). Furthermore, a model of spontaneous prostate adenocarcinoma also demonstrated increased MMP2 and MMP-9 activity in Sparc-null mice (Said et al. 2009) (Table 17.3). SPARC can also regulate the ability of cancer cells to adhere to and migrate on ECM components by modulating integrin activation. Preincubation of ovarian cancer cell lines with SPARC inhibited alpha v and beta 1 integrin-induced adhesion to ECM proteins and mesothelial cells (Said and Najwer 2007b) and proliferation (Said and Najwer 2007b; Said and Socha 2007c) (Table 17.4). In glioma cell lines that have forced expression of SPARC and an increased migration and invasion potential, SPARC was found to activate RhoA but not Rac1 or Cdc42 (Kunigal et al. 2006) (Table 17.5). Another study using glioma cell lines with forced SPARC found that SPARC increases activation of the p38MAPK/HSP27 pathway and SPARC-mediated increases in migration and Matrigel invasion could be inhibited with siRNA to HSP27 (Golembieski et al. 2008) (Table 17.5).
17 SPARC and the Tumor Microenvironment
333
It has been suggested that, prior to acquiring a migratory and invasive phenotype, tumor cells need to adopt a process similar to the developmentally regulated epithelial-mesenchymal transition (EMT), where epithelial cells lose their polarity and develop a mesenchymal phenotype. This hypothesis remains controversial (Tsuji and Ibaragi 2009) however it is interesting to note that SPARC expression downregulates epithelial markers of EMT such as E-cadherin, and increases the induction of the mesenchymal genes such as FAK and snail in melanoma (Craene et al. 2005; Said and Najwer 2007b; Smit and Gardiner 2007). In addition, SPARC upregulation was associated with other EMT-related markers by expression array in basal-like breast cancer (Sarrio et al. 2008). In summary, these studies demonstrate an increased understanding of SPARC signaling in cancer cells. As discussed in the next sections, SPARC can also regulate cellular function and signaling in endothelial cells, fibroblasts, and immune cells.
17.2.2 Effects on Endothelial Cells and Tumor Vascularity Several in vitro studies have shown that SPARC has an inhibitory effect on endothelial cell proliferation, migration and ECM production (Fig. 17.2). SPARC can suppress endothelial cell proliferation by binding to VEGF, and thereby blocking the VEGF-VEGFR1-induced phosphorylation of ERK (Kupprion et al. 1998). SPARC can also indirectly inhibit bFGF-induced proliferation and DNA synthesis of endothelial cells (Motamed and Blake 2003). Another in vitro study demonstrated that SPARC can inhibit bFGF-stimulated human umbilical vein endothelial cell migration (Chlenski et al. 2006). One study tested five different preparations of human SPARC (four recombinant and one derived from melanoma cells) on bovine aortic endothelial (BAE) cells and found that all five types of SPARC inhibited proliferation, adhesion, and migration (Haber et al. 2008). In addition, a study found that the addition of SPARC to endothelial cell cultures greatly reduced mRNA for fibronectin and thrombospondin-1, and increased mRNA for type 1-plasminogen activator inhibitor (Lane et al. 1992). Modulation of these ECM components by SPARC could change the ECM in a way that is suppressive of endothelial cell proliferation and migration and reduces angiogenesis. Fragments of SPARC protein have been assessed to determine which domains are important in these functions. MMP-3-induced cleavage fragments of SPARC were examined and found to regulate BAE cell proliferation and migration (Sage et al. 2003) (Fig. 17.2). It was found that two fragments (Z2 and Z3) had inhibitory effects on proliferation, whereas one fragment (Z1-containing the Cu2+ binding sequence KHGK) produced a biphasic effect, with stimulation of proliferation at low concentrations and inhibition at higher concentrations. SPARC fragments Z2 and Z3 stimulated endothelial cell migration, whereas fragment Z1 had no effect on migration. Therefore, while the effects of full-length SPARC on endothelial cell proliferation and migration are inhibitory, peptides of SPARC that are cleaved by matrix metalloproteinases in the tumor microenvironment may have stimulatory
334
S. L. Thomas and S. A. Rempel
effects on these processes. In support of this, a stimulatory effect was observed with the Z1 KHGK-containing fragment using the in vivo chick chorioallantoic membrane (CAM) assay. Whereas the Z2 and Z3 fragments had no effect, the Z1 fragment showed biphasic effects on angiogenesis, with low concentrations increasing vessel density and higher concentrations decreasing vessel density (Sage et al. 2003) (Fig. 17.2). Another study used a neuroblastoma xenograft model and administered SPARC continuously by osmotic pump after tumors were established, which resulted in a decreased vessel density in SPARC-treated tumors (Chlenski et al. 2002). The angiosuppressive effect of SPARC was contained in the FS-E peptide encompassing the EGF-like module of the follistatin domain (Chlenski et al. 2004) (Fig. 17.2). This peptide could inhibit neovascularization induced by bFGF or neuroblastoma cells in the Matrigel plug angiogenesis assay. Such fragment studies are thus beginning to define the regions of SPARC necessary for its different effects on endothelial cells. In vivo studies show that SPARC can be either pro- or anti-angiogenic depending on the cancer model system that is used. Several studies in which SPARCexpressing or -transfected cancer cells were implanted into nude rodents reported that SPARC expression leads to a significant reduction in tumor vascularity. Glioma cells with forced expression of SPARC had reduced VEGF165 transcript, and glioma xenograft tumors that expressed SPARC had reduced VEGF staining and decreased vascularity compared to control tumors (Yunker et al. 2008) (Table 17.2). Furthermore, subcutaneous injection of a hepatocellular carcinoma cell line (Lau et al. 2006) (Table 17.2) or a transformed human embryonic kidney cell line into mice (Chlenski et al. 2006) having forced SPARC expression resulted in tumors with decreased vascularity. The suppressive effects of SPARC on vascularity were also inferred using Sparcnull mice. Ovarian cancer cells grown in a Sparc-null environment had significantly upregulated mRNA and increased immunostaining for VEGF and the VEGF receptors VEGFR2 and neuropilin-1 in tumor nodules, as well as higher VEGF protein levels in ascitic fluid (Said and Motamed 2005) (Table 17.3). Further study correlated the loss of SPARC with increased mean vascular density (Said and Socha 2007c). Similarly, VEGF protein expression was increased in spontaneous prostate adenocarcinomas in Sparc-null mice, which correlated with increased vessel density (Said et al. 2009) (Table 17.3). In contrast, studies of other tumors grown in Sparc-null mice suggest that SPARC produced by the stroma can also stimulate angiogenesis. Two reports demonstrated a decrease in the number of vessels supplying the tumors in Sparc-null mice. One study used a pancreatic adenocarcinoma cell line injected into the pancreas of Sparcnull and Sparc-null mice (Arnold et al. 2008) (Table 17.3) and the other study used a SPARC-producing mammary carcinoma cell line derived from transgenic mice and injected into the mammary fat pad of Sparc-null and Sparc-wt mice (Sangaletti et al. 2003) (Table 17.3). In addition, when Lewis lung carcinoma cells were implanted in Sparc-null mice, tumors had reduced mRNA for all three VEGF receptors with no change in VEGF expression. Although the number of vessels was not reduced, the tumors had decreased vascular area (Brekken et al. 2003) (Table 17.3).
17 SPARC and the Tumor Microenvironment
335
The literature therefore indicates that SPARC secreted from cancer cells or present in the stroma can have differing effects dependent upon the cancer type examined, possibly due to cleavage of SPARC to produce active fragments that can either stimulate or suppress angiogenesis. However, the majority of in vivo studies indicate that SPARC can attenuate vascularity or angiogenesis by suppressing VEGF expression, by binding to VEGF growth factor to inhibit VEGF-VEGFR interactions, and/or by reorganizing the ECM in such a way as to diminish vascular formation.
17.2.3 Effects on Fibroblasts and Other Stromal Cells Evidence that SPARC plays a role in fibroblasts comes from reports that SPARC expression is altered during fibroblast pathology. There are several reports of reduced SPARC expression in fibroblasts that have been transformed with either a virus or an oncogene (Colombo et al. 1991; Kraemer et al. 1999; Vial and Castellazzi 2000). One study analyzed fibrosarcoma development after the injection of viral oncogene transformed chick embryo fibroblasts into the wing web of newborn chicks. Treatment with exogenous SPARC or reexpression of SPARC in transformed fibroblasts strongly reduced the number and size of tumors (Vial and Castellazzi 2000). In support of these studies, an in vivo model showed that fibrosarcomas induced by subcutaneous injection of a chemical carcinogen have between 32–89% less SPARC mRNA than normal fibroblasts (Colombo et al. 1991). These results indicate that SPARC expression represses tumorigenesis in fibroblasts. In contrast, when SPARC is expressed by peritumoral fibroblasts in pancreatic adenocarcinoma, stromal SPARC is associated with a significantly worse prognosis (Infante et al. 2007). Therefore, SPARC produced by fibroblasts may have different effects on cancers cells than on the fibroblasts themselves. Studies of Sparc-null fibroblasts compared to normal fibroblasts have shown changes in phenotype including altered activation, proliferation, migration, collagen processing, and metalloproteinase expression. Skin fibroblasts, aortic smooth muscle cells, and mesangial cells from Sparc-null mice exhibited accelerated proliferation in growth media. The addition of exogenous SPARC to fibroblasts and mesangial cells reduced the proliferation rates of both Sparc-null and Sparc-wt cells with a greater reduction in proliferation for Sparc-null cells (Bradshaw et al. 1999). SPARC may reduce the proliferation of fibroblasts in part by forming a complex with PDGFAB and PDGF-BB and inhibiting the binding of PDGF to fibroblasts (Raines et al. 1992). SPARC has also been shown to enhance the migration of bFGF-stimulated fibroblasts (Chlenski et al. 2007) which may explain the increased numbers of fibroblasts in SPARC-expressing tumor xenografts (Chlenski et al. 2007; Prada et al. 2007). In addition, SPARC was shown to inhibit the activation of fibroblasts with TGF-beta as assessed by smooth muscle actin expression (Chlenski et al. 2007). Inhibition of fibroblast activation may alter the processing and deposition of collagen by fibroblasts. Sparc-null fibroblasts were found to have an increased amount
336
S. L. Thomas and S. A. Rempel
of processed collagen I associated with cell layers and a reduced amount of collagen incorporated into an insoluble matrix suggesting that SPARC regulates the conversion of procollagen I to collagen I and reduces the incorporation of collagen I into the ECM (Rentz et al. 2007). Furthermore, the addition of exogenous SPARC to rabbit synovial fibroblasts induced the expression of the metalloproteinases collagenase, stromelysin, and the 92-kD gelatinase; and peptides from domains III and IV of SPARC could also induce collagenase expression. Further investigation revealed that SPARC-depleted conditioned media was sufficient to induce collagenase expression, indicating that the effect may be indirect through a secreted protein that is regulated by SPARC (Tremble et al. 1993). These studies provide evidence that SPARC is expressed by fibroblasts and is an important regulator of fibroblast activity. Future research should assess the contribution of fibroblast-derived SPARC to tumor formation and progression. One study found that SPARCexpressing fibroblasts had no effect on the in vitro or in vivo growth of melanoma cell lines; however, SPARC hyperexpression by melanoma cells also did not affect in vivo tumor growth in this model (Prada et al. 2007).
17.2.4 Effects on Immune Cells Studies of the immune system in Sparc-null mice reveal alterations in immune cell numbers and function. Sparc-null mice have larger spleens characterized by follicular lymphoid hyperplasia with prominent germinal centers, attenuated marginal zones, increased white pulp, and reduced marginal zone B cells (Rempel et al. 2007). A blood panel analysis showed that Sparc-null mice have no change in the percentage of monocytes, a lower percentage of neutrophils, and a higher percentage of lymphocytes; however, the total white blood cell counts were significantly lower in Sparc-null animals. Analysis of Sparc-null bone marrow demonstrated an increase in CD3+ T cells and a decrease in CD19+ B cells. Importantly, Sparc-null mice were unable to mount an immune response as evaluated by footpad swelling in response to LPS injection (Rempel et al. 2007). Another study tested the immune response in Sparc-null mice using a cutaneous contact hypersensitivity (CHS) model. This study found that after immune system challenge, sensitized skin in Sparc-null mice had increased edema and increased numbers of neutrophils and macrophages. The enhanced CHS response was due to faster migration and increased numbers of dendritic cells in draining lymph nodes, which resulted in the accelerated priming of T cells in Sparc-null mice (Sangaletti et al. 2005). This study suggests that, as with other cell types, SPARC regulates the proliferation and migration of immune cells. Additional evidence that SPARC regulates immune cell function comes from studies of tumors grown in Sparc-null mice or tumors expressing SPARC antisense. Two studies in which melanoma cell lines expressing high amounts of SPARC were transduced with SPARC antisense and then grown in vivo demonstrated that reduction of SPARC in the cancer cells resulted in a strong recruitment of polymorpho-
17 SPARC and the Tumor Microenvironment
337
nuclear leukocytes (PMNs) and inhibition of tumor growth resulting in complete cancer cell rejection in some animals (Alvarez et al. 2005; Prada et al. 2007). The use of neutralizing antibodies to deplete PMN abrogated the inhibitory effect of SPARC reduction on tumor growth (Prada et al. 2007). It was also found that human PMN have an increased capacity to adhere to and exert a cytotoxic effect on melanoma cells in which SPARC is knocked down with antisense (Alvarez et al. 2005). Similarly, a study in which a mouse mammary carcinoma cell line was grown in Sparc-null and Sparc-wt mice found that Sparc-null mice and mice receiving bone marrow from a Sparc-null donor had a significantly larger leukocyte population with a greater number of all types of leukocytes than Sparc-wt animals. In addition, Sparc-wt animals had leukocytes located only in the perilobular stroma, whereas Sparc-null mice had a strong infiltration of leukocytes throughout the tumor parenchyma and associated with necrotic areas (Sangaletti et al. 2003) (Table 17.3). A similar change in the distribution of macrophages was found when pancreatic cancer cells were grown subcutaneously. Tumors in Sparc-null mice had macrophages uniformly distributed throughout the tumor, whereas tumors in Sparc-wt mice had macrophages only at the tumor margin (Puolakkainen et al. 2004) (Table 17.3). Another study using a model of peritoneal ovarian carcinomatosis found that tumor tissue and ascites in Sparc-null mice had increased macrophage infiltration compared to Sparc-wt mice and that Sparc-null ascitic fluid had significantly higher levels of MCP-1 and increased expression of the proinflammatory mediators IL-6 and 8-isoprostane (Said and Socha 2007c). Further in vitro assessment of the chemotactic effect of ovarian cancer cells on macrophages demonstrated that overexpression of SPARC in cancer cell lines significantly attenuated MCP-1 production and the migration of macrophages toward cancer cells (Said et al. 2008). Both macrophages and cancer cells can produce proinflammatory mediators, and exogenous SPARC or forced endogenous SPARC was shown to reduce PGE2 and 8-iosprostane production by ovarian cancer cells or co-cultures of cancer cells with macrophages and/or mesothelial cells (Said et al. 2008). SPARC also significantly suppressed the expression of IL-6 in co-cultures of ovarian cancer cells and macrophages (Said et al. 2008). Co-culture of ovarian cancer cell lines with macrophages or PGE2 treatment resulted in NF-kappaB activation, which was significantly attenuated when the ovarian cancer cells were forced to express SPARC (Said et al. 2008). Just as SPARC can attenuate the migration of macrophages, macrophages can attenuate the effects of SPARC. Alternatively activated macrophages that express stabilin-1 can efficiently internalize SPARC and target it to the endocytic pathway for lysosomal degradation. SPARC was found to bind to the extracellular EGF-like domain of stabilin-1 (Kzhyshkowska et al. 2006). Furthermore, the regulation of leukocyte transendothelial migration by SPARC has also been investigated. It was found that the expression of SPARC by leukocytes is required for transendothelial migration because leukocytes derived from Sparc-null mice have a significantly reduced capacity to migrate through endothelial cell monolayers. Leukocytes may need SPARC to induce intercellular gaps and compromise barrier function in endothelial cells (Kelly et al. 2007).
338
S. L. Thomas and S. A. Rempel
These studies bring up the question of whether the increased immune cell infiltration of the tumor tissue when SPARC is absent is a result of alterations in ECM deposition that reduce the interference to immune cell migration or changes in the migration of the immune cells themselves because the cells are not expressing SPARC. Further studies are necessary to determine the molecular mechanisms involved in the effects of SPARC on immune cell function.
17.2.5 Summary Overall, the research into the role of SPARC in the tumor microenvironment indicates that SPARC plays a key role in regulating tumor growth and invasion, angiogenesis, ECM production and composition, and immune cell infiltration during tumor development and progression. The conflicting data regarding the roles of SPARC in various types of cancer shed light on the importance of the microenvironment in which the tumor develops. The interaction of cancer cells with stromal and immune cells is regulated by the expression of specific integrins, ECM components, matrix-degrading proteases, and growth factors within the tumor environment. As described above, SPARC can inhibit the activity of certain integrins and growth factors and increase the production of ECM constituents by cancer and stromal cells. SPARC regulates the production of matrix-degrading proteases, which can not only remodel the ECM but also cleave SPARC into fragments that can have opposing biological activity compared to full-length SPARC. SPARC can also alter the sensitivity of cancer cells to the cytotoxic effects of the immune system. It is important to consider whether the in vitro and the in vivo data reflect what is observed in the human tissues and the survival patterns of the patients. Interestingly, the studies using the Sparc-null and Sparc-wt mice suggest that stromal SPARC is suppressive to tumor growth, invasion and metastasis for the majority of tumor types (Fig. 17.1). An expectation might be then that patients having tumors with high stromal SPARC should have better survival outcome. However this is not the case (Table 17.1). In general, high SPARC expression correlates with poor patient survival. However, the emerging data suggests that stromal SPARC may suppress the infiltration of macrophages and inhibit the infiltration of leukocytes, which would reduce the ability of the immune system to fight cancer cells. These latter observations would support the immunohistochemical and clinical correlative data, such that high SPARC expression would correlate with poor patient survival. More research is required to understand the relationship between the SPARC-positive tumors and the SPARC-negative tumors and their interactions with SPARC-positive and SPARC-negative stroma. Regardless of the differences in the effects of SPARC in various cancers, the literature provides solid evidence that the regulation of SPARC has the potential to control cancer and therapeutic strategies should be devised.
17 SPARC and the Tumor Microenvironment
339
17.3 Potential for Utilizing or Targeting SPARC to Control Cancer In certain cancer types, such as breast cancer, head and neck cancer, glioma and melanoma, SPARC may provide a therapeutic target to inhibit cancer cell invasion. However in other cancers, like neuroblastoma and ovarian, lung, and prostate cancers, the reexpression of SPARC in the tumor may provide a treatment for cancer. Several groups are attempting to utilize SPARC in the development of new cancer therapies based on both strategies. For tumors having high expression of SPARC, one approach is to directly target SPARC using the immune system. For example, one group used a SPARC peptide to immunize mice against SPARC. When a breast cancer cell line was inoculated subcutaneously, 50% of immunized mice had complete tumor rejection and the other 50% had significantly smaller tumors compared to non-immunized mice (Ikuta et al. 2009). Because SPARC has a high binding affinity to albumin, it was proposed that SPARC secreted by tumors could bind to albumin-bound drug and accumulate the drug within the tumor (Desai et al. 2009). Albumin-bound paclitaxel is in clinical trials for the treatment of head and neck squamous cell carcinoma, and the data indicate that the response to treatment was significantly higher for patients with SPARC-positive cancer compared to SPARC-negative cancer patients (Desai et al. 2009). Targeting SPARC can also be done indirectly. The anti-cancer agent Ukrain was found to reduce SPARC protein expression in glioblastoma cells (Gagliano et al. 2006). Furthermore, the fact that cancer cells and stromal cells can express high amounts of SPARC provides a target for directing viral gene expression to the tumor microenvironment, by exploiting the use of the SPARC gene promoter sequence. One group used the SPARC gene promoter to specifically target suicidal gene expression in SPARC-expressing melanoma cancer and stromal cells. Melanoma cells expressing the suicidal thymidine kinase gene under the control of a SPARC promoter were implanted subcutaneously and then treated with ganciclovir 10 days later. Tumor regression was seen in all mice and in most cases no visible tumor was present (Lopez et al. 2006). Similar results were obtained when a mix of tumor cells with and without the suicidal gene was implanted or when endothelial cells expressing the suicidal gene were mixed with melanoma cells (Lopez et al. 2006). These results indicated the efficacy of the approach and that targeting stromal cells was as effective as targeting the tumor cells. This model was advanced by the development of a conditionally replicative oncolytic adenovirus (CRAd) based on the SPARC promoter. This virus was used to assess the cross-talk between the tumor and stromal cells. The SPARC-based CRAd had a potent anti-tumor effect on melanoma xenografts with complete tumor disappearance in some mice; however a delay in tumor growth with no cure was seen when melanoma cells were mixed with fibroblasts or endothelial cells, indicating that targeting the tumor cells alone was insufficient.
340
S. L. Thomas and S. A. Rempel
The CRAd was also effective in SPARC-negative prostate cancer cells; however, the presence of stromal cells enhanced viral activity, possibly due to the induction of cell cycle in the cancer cells (Lopez et al. 2009). These experiments emphasize the need to study both the tumor cells and stromal cells to define therapeutic strategies. For tumors having a methylated SPARC promoter, an approach has been to reactivate SPARC expression. Demethylating agents such as 5-azacytidine and decitabine have already been used in clinical trials for myelodysplastic syndrome, and results indicate that 5-azacytidine prolongs the overall survival in these patients (Gurion et al. 2009). Indeed, SPARC is one of the genes upregulated in multiple myeloma cells treated with demethylating agents (Heller et al. 2008). Since the suppression of SPARC correlates with poor overall survival of multiple myeloma patients, its reactivation is considered to be a therapeutic approach (Heller et al. 2008). Although these drugs have not been successful for glioma treatment, their use is also being investigated for the reactivation of SPARC in other solid tumors. For example, a demethylating agent could restore SPARC expression in prostate cancer cell lines (Sato et al. 2003). Other studies show that aberrant methylation of SPARC leads to chemotherapy resistance in colorectal cancer (Tai et al. 2005), and direct administration of SPARC was shown to enhance apoptosis and potentiate chemotherapy sensitivity (Tang and Tai 2007). The same effects were accomplished using a demethylating drug; 5-aza-2′-deocycytidine treatment increased SPARC expression and improved response to therapy (Cheetham et al. 2008). In addition, the nonsteroidal anti-inflammatory drug NS398 reactivated SPARC expression in lung cancer cells by promoter demethylation, and this treatment was found to decrease the invasiveness of lung cancer cells in vitro (Pan et al. 2008). Finally, the effects of SPARC were enhanced with vitamin D, which was found to act synergistically with SPARC to increase the susceptibility of therapy-resistant colorectal cancer cells to chemotherapy (Taghizadeh et al. 2007). For tumors having suppressed SPARC in the absence of a methylated promoter, such as neuroblastoma (Yang et al. 2007), an approach has been to administer SPARC directly. For example, SPARC delivered by osmotic pump resulted in significantly reduced neuroblastoma tumor volume and vascularity in vivo (Chlenski et al. 2002). This direct approach can also work for tumors having methylated SPARC. These preclinical studies provide promising data that research efforts into the function of SPARC will be translated into the use of SPARC for the targeting or treatment of various types of cancer.
References Alonso SR, Tracey L et al (2007) A high-throughput study in melanoma identifies epithelial-mesenchymal transition as a major determinant of metastasis. Cancer Res 67(7):3450–3460 Alvarez MJ, Prada F et al (2005) Secreted protein acidic and rich in cysteine produced by human melanoma cells modulates polymorphonuclear leukocyte recruitment and antitumor cytotoxic capacity. Cancer Res 65(12):5123–5132
17 SPARC and the Tumor Microenvironment
341
Amatschek S, Koenig U et al (2004) Tissue-wide expression profiling using cDNA subtraction and microarrays to identify tumor-specific genes. Cancer Res 64(3):844–856 Arnold S, Mira E et al (2008) Forced expression of MMP9 rescues the loss of angiogenesis and abrogates metastasis of pancreatic tumors triggered by the absence of host SPARC. Exp Biol Med 233(7):860–873 Aycock RL, Bradshaw AC et al (2004) Development of UV-induced squamous cell carcinomas is suppressed in the absence of SPARC. J Invest Dermatol 123(3):592–599 Barker TH, Baneyx G et al (2005) SPARC regulates extracellular matrix organization through its modulation of integrin-linked kinase activity. J Biol Chem 280(43):36483–36493 Barth PJ, Moll R et al (2005) Stromal remodeling and SPARC (secreted protein acid rich in cysteine) expression in invasive ductal carcinomas of the breast. Virchows Arch 446(5):532– 536 Beck AH, Espinosa I et al (2008) The fibromatosis signature defines a robust stromal response in breast carcinoma. Lab Invest 88(6):591–601 Bergamaschi A, Tagliabue E et al (2008). Extracellular matrix signature identifies breast cancer subgroups with different clinical outcome. J Pathol 214(3):357–367 Bornstein P, Sage EH (2002) Matricellular proteins: extracellular modulators of cell function. Curr Opin Cell Biol 14(5):608–616 Bozkurt SU, Ayan E et al (2009) Immunohistochemical expression of SPARC is correlated with recurrence, survival and malignant potential in meningiomas. APMIS 117(9):651–659 Bradshaw AD, Francki A et al (1999) Primary mesenchymal cells isolated from SPARC-null mice exhibit altered morphology and rates of proliferation. Mol Biol Cell 10(5):1569–1579 Brekken RA, Sage EH (2001) SPARC, a matricellular protein: at the crossroads of cell-matrix communication. Matrix Biol 19(8):816–827 Brekken RA, Puolakkainen P et al (2003) Enhanced growth of tumors in SPARC null mice is associated with changes in the ECM. J Clin Invest 111(4):487–495 Brown TJ, Shaw PA et al (1999) Activation of SPARC expression in reactive stroma associated with human epithelial ovarian cancer. Gynecol Oncol 75(1):25–33 Brune K, Hong SM et al (2008) Genetic and epigenetic alterations of familial pancreatic cancers. Cancer Epidemiol Biomark Prev 17(12):3536–3542 Bull Phelps SL, Carbon J et al (2009) Secreted protein acidic and rich in cysteine as a regulator of murine ovarian cancer growth and chemosensitivity. Am J Obstet Gynecol 200(2):180 e1–e7 Chan SK, Griffith OL et al (2008) Meta-analysis of colorectal cancer gene expression profiling studies identifies consistently reported candidate biomarkers. Cancer Epidemiol Biomark Prev 17(3):543–552 Che Y, Luo A et al (2006) The differential expression of SPARC in esophageal squamous cell carcinoma. Int J Mol Med 17(6):1027–1033 Cheetham S, Tang MJ et al (2008) SPARC promoter hypermethylation in colorectal cancers can be reversed by 5-Aza-2′deoxycytidine to increase SPARC expression and improve therapy response. Br J Cancer 98(11):1810–1819 Chin D, Boyle GM et al (2005). Novel markers for poor prognosis in head and neck cancer. Int J Cancer 113(5):789–797 Chlenski A, Liu S et al (2002) SPARC is a key Schwannian-derived inhibitor controlling neuroblastoma tumor angiogenesis. Cancer Res 62(24):7357–7363 Chlenski A, Liu S et al (2004) Neuroblastoma angiogenesis is inhibited with a folded synthetic molecule corresponding to the epidermal growth factor-like module of the follistatin domain of SPARC. Cancer Res 64(20):7420–7425 Chlenski A, Liu S et al (2006) SPARC expression is associated with impaired tumor growth, inhibited angiogenesis and changes in the extracellular matrix. Int J Cancer 118(2):310–316 Chlenski A, Guerrero LJ et al (2007) SPARC enhances tumor stroma formation and prevents fibroblast activation. Oncogene 26(31):4513–4522 Colombo MP, Ferrari G et al (1991) Down-regulation of SPARC/osteonectin/BM-40 expression in methylcholanthrene-induced fibrosarcomas and in Kirsten-MSV transformed fibroblasts. Eur J Cancer 27(1):58–62
342
S. L. Thomas and S. A. Rempel
Dalla-Torre CA, Yoshimoto M et al (2006) Effects of THBS3, SPARC and SPP1 expression on biological behavior and survival in patients with osteosarcoma. BMC Cancer 6:237 De Craene B, van Roy F et al (2005) Unraveling signalling cascades for the Snail family of transcription factors. Cell Signal 17(5):535–547 Desai N, Trieu V et al (2009) SPAR+C Expression correlates with tumor response to albuminbound paclitaxel in head and neck cancer patients. Transl Oncol 2(2):59–64 DiMartino JF, Lacayo NJ et al (2006) Low or absent SPARC expression in acute myeloid leukemia with MLL rearrangements is associated with sensitivity to growth inhibition by exogenous SPARC protein. Leukemia 20(3):426–432 Funk SE, Sage EH (1991) The Ca2(+)-binding glycoprotein SPARC modulates cell cycle progression in bovine aortic endothelial cells. Proc Natl Acad Sci U S A 88(7):2648–2652 Gagliano N, Moscheni C et al (2006) Effect of Ukrain on matrix metalloproteinase-2 and Secreted Protein Acidic and Rich in Cysteine (SPARC) expression in human glioblastoma cells. Anticancer Drugs 17(2):189–194 Gieseg MA, Cody T et al (2002) Expression profiling of human renal carcinomas with functional taxonomic analysis. BMC Bioinforma 3:26 Golembieski WA, Ge S et al (1999) Increased SPARC expression promotes U87 glioblastoma invasion in vitro. Int J Dev Neurosci 17(5–6):463–472 Golembieski WA, Thomas SL et al (2008) HSP27 mediates SPARC-induced changes in glioma morphology, migration, and invasion. Glia 56(10):1061–1075 Gurion R, Vidal L et al (2009) 5-azacitidine prolongs overall survival in patients with myelodysplastic syndrome – systematic review and meta-analysis. Haematologica 2009 Sep 22 [Epub ahead of print] Haber CL, Gottifredi V et al (2008) SPARC modulates the proliferation of stromal but not melanoma cells unless endogenous SPARC expression is downregulated. Int J Cancer 122(7):1465–1475 Hecht JT, Sage EH (2006) Retention of the matricellular protein SPARC in the endoplasmic reticulum of chondrocytes from patients with pseudoachondroplasia. J Histochem Cytochem 54(3):269–274 Heller G, Schmidt WM et al (2008) Genome-wide transcriptional response to 5-aza-2′deoxycytidine and trichostatin a in multiple myeloma cells. Cancer Res 68(1):44–54 Hong SM, Kelly D et al (2008) Multiple genes are hypermethylated in intraductal papillary mucinous neoplasms of the pancreas. Mod Pathol 21(12):1499–1507 Huynh MH, Hong H et al (2000) Association of SPARC (osteonectin, BM-40) with extracellular and intracellular components of the ciliated surface ectoderm of Xenopus embryos. Cell Motil Cytoskeleton 47(2):154–162 Ikuta Y, Hayashida Y et al (2009) Identification of the H2-Kd-restricted cytotoxic T lymphocyte epitopes of a tumor-associated antigen, SPARC, which can stimulate antitumor immunity without causing autoimmune disease in mice. Cancer Sci 100(1):132–137 Infante JR, Matsubayashi H et al (2007) Peritumoral fibroblast SPARC expression and patient outcome with resectable pancreatic adenocarcinoma. J Clin Oncol 25(3):319–325 Jones C, Mackay A et al (2004) Expression profiling of purified normal human luminal and myoepithelial breast cells: identification of novel prognostic markers for breast cancer. Cancer Res 64(9):3037–3045 Junnila S, Kokkola A et al (2009) Gene expression analysis identifies over-expression of CXCL1, SPARC, SPP1, and SULF1 in gastric cancer. Genes Chromosomes Cancer 2009 Sep 24. [Epub ahead of print] Kamihagi K, Katayama M et al (1994) Osteonectin/SPARC regulates cellular secretion rates of fibronectin and laminin extracellular matrix proteins. Biochem Biophys Res Commun 200(1):423–428 Kato Y, Nagashima Y et al (2005) Expression of SPARC in tongue carcinoma of stage II is associated with poor prognosis: an immunohistochemical study of 86 cases. Int J Mol Med 16(2):263–268 Kelly KA, Allport JR et al (2007) SPARC is a VCAM-1 counter-ligand that mediates leukocyte transmigration. J Leukoc Biol 81(3):748–756
17 SPARC and the Tumor Microenvironment
343
Koblinski JE, Kaplan-Singer BR et al (2005) Endogenous osteonectin/SPARC/BM-40 expression inhibits MDA-MB-231 breast cancer cell metastasis. Cancer Res 65(16):7370–7377 Koukourakis MI, Giatromanolaki A et al (2003) Enhanced expression of SPARC/osteonectin in the tumor-associated stroma of non-small cell lung cancer is correlated with markers of hypoxia/ acidity and with poor prognosis of patients. Cancer Res 63(17):5376–5380 Kraemer M, Tournaire R et al (1999) Rat embryo fibroblasts transformed by c-Jun display highly metastatic and angiogenic activities in vivo and deregulate gene expression of both angiogenic and antiangiogenic factors. Cell Growth Differ 10(3):193–200 Kunigal S, Gondi CS et al (2006) SPARC-induced migration of glioblastoma cell lines via uPAuPAR signaling and activation of small GTPase RhoA. Int J Oncol 29(6):1349–1357 Kupprion C, Motamed K et al (1998) SPARC (BM-40, osteonectin) inhibits the mitogenic effect of vascular endothelial growth factor on microvascular endothelial cells. J Biol Chem 273(45):29635–29640 Kzhyshkowska J, Workman G et al (2006) Novel function of alternatively activated macrophages: stabilin-1-mediated clearance of SPARC. J Immunol 176(10):5825–5832 Lane TF, Iruela-Arispe ML et al (1992) Regulation of gene expression by SPARC during angiogenesis in vitro. Changes in fibronectin, thrombospondin-1, and plasminogen activator inhibitor-1. J Biol Chem 267(23):16736–16745 Lapointe J, Li C et al (2004) Gene expression profiling identifies clinically relevant subtypes of prostate cancer. Proc Natl Acad Sci U S A 101(3):811–816 Lau CP, Poon RT et al (2006) SPARC and Hevin expression correlate with tumour angiogenesis in hepatocellular carcinoma. J Pathol 210(4):459–468 Le Bail B, Faouzi S et al (1999) Osteonectin/SPARC is overexpressed in human hepatocellular carcinoma. J Pathol 189(1):46–52 Ledda MF, Bravo AI et al (1997a) The expression of the secreted protein acidic and rich in cysteine (SPARC) is associated with the neoplastic progression of human melanoma. J Invest Dermatol 108(2):210–214 Ledda MF, Adris S et al (1997b) Suppression of SPARC expression by antisense RNA abrogates the tumorigenicity of human melanoma cells. Nat Med 3(2):171–176 Lopez MV, Blanco P et al (2006) Expression of a suicidal gene under control of the human secreted protein acidic and rich in cysteine (SPARC) promoter in tumor or stromal cells led to the inhibition of tumor cell growth. Mol Cancer Ther 5(10):2503–2511 Lopez MV, Viale DL et al (2009) Tumor associated stromal cells play a critical role on the outcome of the oncolytic efficacy of conditionally replicative adenoviruses. PLoS One 4(4):e5119 Luo A, Kong J et al (2004) Discovery of Ca2+-relevant and differentiation-associated genes downregulated in esophageal squamous cell carcinoma using cDNA microarray. Oncogene 23(6):1291–1299 Maloney SC, Marshall JC et al (2009) SPARC is expressed in human uveal melanoma and its abrogation reduces tumor cell proliferation. Anticancer Res 29(8):3059–3064 Mann K, Deutzmann R et al (1987) Solubilization of protein BM-40 from a basement membrane tumor with chelating agents and evidence for its identity with osteonectin and SPARC. FEBS Lett 218(1):167–172 McClung HM, Thomas SL et al (2007) SPARC upregulates MT1-MMP expression, MMP-2 activation, and the secretion and cleavage of galectin-3 in U87MG glioma cells. Neurosci Lett 419(2):172–177 Mok SC, Chan WY et al (1996) SPARC, an extracellular matrix protein with tumor-suppressing activity in human ovarian epithelial cells. Oncogene 12(9):1895–1901 Motamed K, Blake DJ (2003) Fibroblast growth factor receptor-1 mediates the inhibition of endothelial cell proliferation and the promotion of skeletal myoblast differentiation by SPARC: a role for protein kinase A. J Cell Biochem 90(2):408–423 Motamed K, Sage EH (1998) SPARC inhibits endothelial cell adhesion but not proliferation through a tyrosine phosphorylation-dependent pathway. J Cell Biochem 70(4):543–552 Paley PJ, Goff BA et al (2000) Alterations in SPARC and VEGF immunoreactivity in epithelial ovarian cancer. Gynecol Oncol 78(3 Pt 1):336–341
344
S. L. Thomas and S. A. Rempel
Pan MR, Chang HC et al (2008) The nonsteroidal anti-inflammatory drug NS398 reactivates SPARC expression via promoter demethylation to attenuate invasiveness of lung cancer cells. Exp Biol Med (Maywood) 233(4):456–462 Pen A, Moreno MJ et al (2007) Molecular markers of extracellular matrix remodeling in glioblastoma vessels: microarray study of laser-captured glioblastoma vessels. Glia 55(6):559–572 Porter D, Lahti-Domenici J et al (2003) Molecular markers in ductal carcinoma in situ of the breast. Mol Cancer Res 1(5):362–375 Prada F, Benedetti LG et al (2007) SPARC endogenous level, rather than fibroblast-produced SPARC or stroma reorganization induced by SPARC, is responsible for melanoma cell growth. J Invest Dermatol 127(11):2618–2628 Puolakkainen PA, Brekken RA et al (2004) Enhanced growth of pancreatic tumors in SPARC-null mice is associated with decreased deposition of extracellular matrix and reduced tumor cell apoptosis. Mol Cancer Res 2(4):215–224 Raines EW, Lane TF et al (1992) The extracellular glycoprotein SPARC interacts with plateletderived growth factor (PDGF)-AB and -BB and inhibits the binding of PDGF to its receptors. Proc Natl Acad Sci U S A 89(4):1281–1285 Rempel SA, Ge S et al (1999) SPARC: a potential diagnostic marker of invasive meningiomas. Clin Cancer Res 5(2):237–241 Rempel SA, Golembieski WA et al (1998) SPARC: a signal of astrocytic neoplastic transformation and reactive response in human primary and xenograft gliomas. J Neuropathol Exp Neurol 57(12):1112–1121 Rempel SA, Golembieski WA et al (2001) SPARC modulates cell growth, attachment and migration of U87 glioma cells on brain extracellular matrix proteins. J Neurooncol 53(2):149–160 Rempel SA, Hawley RC et al (2007) Splenic and immune alterations of the Sparc-null mouse accompany a lack of immune response. Genes Immun 8(3):262–274 Rentz TJ, Poobalarahi F et al (2007) SPARC regulates processing of procollagen I and collagen fibrillogenesis in dermal fibroblasts. J Biol Chem 282(30):22062–22071 Rich JN, Shi Q et al (2003) Bone-related genes expressed in advanced malignancies induce invasion and metastasis in a genetically defined human cancer model. J Biol Chem 278(18):15951– 15957 Rich JN, Hans C et al (2005) Gene expression profiling and genetic markers in glioblastoma survival. Cancer Res 65(10):4051–4058 Rodriguez-Jimenez FJ, Caldes T et al (2007) Overexpression of SPARC protein contrasts with its transcriptional silencing by aberrant hypermethylation of SPARC CpG-rich region in endometrial carcinoma. Oncol Rep 17(6):1301–1307 Sage H, Johnson C et al (1984) Characterization of a novel serum albumin-binding glycoprotein secreted by endothelial cells in culture. J Biol Chem 259(6):3993–4007 Sage EH, Reed M et al (2003) Cleavage of the matricellular protein SPARC by matrix metalloproteinase 3 produces polypeptides that influence angiogenesis. J Biol Chem 278(39):37849– 37857 Said N, Motamed K (2005) Absence of host-secreted protein acidic and rich in cysteine (SPARC) augments peritoneal ovarian carcinomatosis. Am J Pathol 167(6):1739–1752 Said N, Najwer I et al (2007a) Secreted protein acidic and rich in cysteine (SPARC) inhibits integrin-mediated adhesion and growth factor-dependent survival signaling in ovarian cancer. Am J Pathol 170(3):1054–1063 Said NA, Najwer I et al (2007b) SPARC inhibits LPA-mediated mesothelial-ovarian cancer cell crosstalk. Neoplasia 9(1):23–35 Said N, Socha MJ et al (2007c) Normalization of the ovarian cancer microenvironment by SPARC. Mol Cancer Res 5(10):1015–1030 Said NA, Elmarakby AA et al (2008). SPARC ameliorates ovarian cancer-associated inflammation. Neoplasia 10(10):1092–1104 Said N, Frierson HF Jr et al (2009) The role of SPARC in the TRAMP model of prostate carcinogenesis and progression. Oncogene 28(39):3487–3498 Sakai N, Baba M et al (2001) SPARC expression in primary human renal cell carcinoma: upregulation of SPARC in sarcomatoid renal carcinoma. Hum Pathol 32(10):1064–1070
17 SPARC and the Tumor Microenvironment
345
Sangaletti S, Gioiosa L et al (2005) Accelerated dendritic-cell migration and T-cell priming in SPARC-deficient mice. J Cell Sci 118(Pt 16):3685–3694 Sangaletti S, Stoppacciaro A et al (2003) Leukocyte, rather than tumor-produced SPARC, determines stroma and collagen type IV deposition in mammary carcinoma. J Exp Med 198(10):1475–1485 Sangaletti S, Di Carlo E et al (2008) Macrophage-derived SPARC bridges tumor cell-extracellular matrix interactions toward metastasis. Cancer Res 68(21):9050–9059 Sansom OJ, Mansergh FC et al (2007) Deficiency of SPARC suppresses intestinal tumorigenesis in APCMin/+mice. Gut 56(10):1410–1414 Sarrio D, Rodriguez-Pinilla SM et al (2008) Epithelial-mesenchymal transition in breast cancer relates to the basal-like phenotype. Cancer Res 68(4):989–997 Sato N, Fukushima N et al (2003) SPARC/osteonectin is a frequent target for aberrant methylation in pancreatic adenocarcinoma and a mediator of tumor-stromal interactions. Oncogene 22(32):5021–5030 Schittenhelm J, Mittelbronn M et al (2006) Patterns of SPARC expression and basement membrane intactness at the tumour-brain border of invasive meningiomas. Neuropathol Appl Neurobiol 32(5):525–531 Schultz C, Lemke N et al (2002) Secreted protein acidic and rich in cysteine promotes glioma invasion and delays tumor growth in vivo. Cancer Res 62(21):6270–6277 Seno T, Harada H et al (2009) Downregulation of SPARC expression inhibits cell migration and invasion in malignant gliomas. Int J Oncol 34(3):707–715 Shi Q, Bao S et al (2004) Secreted protein acidic, rich in cysteine (SPARC), mediates cellular survival of gliomas through AKT activation. J Biol Chem 279(50):52200–52209 Shi Q, Bao S et al (2007) Targeting SPARC expression decreases glioma cellular survival and invasion associated with reduced activities of FAK and ILK kinases. Oncogene 26(28):4084–4094 Smit DJ, Gardiner BB (2007) Osteonectin downregulates E-cadherin, induces osteopontin and focal adhesion kinase activity stimulating an invasive melanoma phenotype. Int J Cancer 121(12):2653–2660 Socha M, Said N et al (2009) Aberrant promoter methylation of SPARC in ovarian cancer. Neoplasia 11(2):126–135 Sosa MS, Girotti MR et al (2007) Proteomic analysis identified N-cadherin, clusterin, and HSP27 as mediators of SPARC (secreted protein, acidic and rich in cysteines) activity in melanoma cells. Proteomics 7(22):4123–4134 Suzuki M, Hao C et al (2005) Aberrant methylation of SPARC in human lung cancers. Br J Cancer 92(5):942–948 Taghizadeh F, Tang MJ et al (2007) Synergism between vitamin D and secreted protein acidic and rich in cysteine-induced apoptosis and growth inhibition results in increased susceptibility of therapy-resistant colorectal cancer cells to chemotherapy. Mol Cancer Ther 6(1):309–317 Tai IT, Dai M et al (2005) Genome-wide expression analysis of therapy-resistant tumors reveals SPARC as a novel target for cancer therapy. J Clin Invest 115(6):1492–1502 Takeno A, Takemasa I et al (2008) Integrative approach for differentially overexpressed genes in gastric cancer by combining large-scale gene expression profiling and network analysis. Br J Cancer 99(8):1307–1315 Tang MJ, Tai IT (2007) A novel interaction between procaspase 8 and SPARC enhances apoptosis and potentiates chemotherapy sensitivity in colorectal cancers. J Biol Chem 282(47):34457– 34467 Termine JD, Kleinman HK et al (1981) Osteonectin, a bone-specific protein linking mineral to collagen. Cell 26(1 Pt 1):99–105 Thomas R, True LD et al (2000) Differential expression of osteonectin/SPARC during human prostate cancer progression. Clin Cancer Res 6(3):1140–1149 Tremble PM, Lane TF et al (1993) SPARC, a secreted protein associated with morphogenesis and tissue remodeling, induces expression of metalloproteinases in fibroblasts through a novel extracellular matrix-dependent pathway. J Cell Biol 121(6):1433–1444 Tsuji T, Ibaragi S (2009) Epithelial-mesenchymal transition and cell cooperativity in metastasis. Cancer Res 69(18):7135–7139
346
S. L. Thomas and S. A. Rempel
Vadlamuri SV, Media J et al (2003) SPARC affects glioma cell growth differently when grown on brain ECM proteins in vitro under standard versus reduced-serum stress conditions. Neuro Oncol 5(4):244–254 Vial E, Castellazzi M (2000) Down-regulation of the extracellular matrix protein SPARC in vSrcand vJun-transformed chick embryo fibroblasts contributes to tumor formation in vivo. Oncogene 19(14):1772–1782 Wang CS, Lin KH et al (2004) Overexpression of SPARC gene in human gastric carcinoma and its clinic-pathologic significance. Br J Cancer 91(11):1924–1930 Watkins G, Douglas-Jones A et al (2005) Increased levels of SPARC (osteonectin) in human breast cancer tissues and its association with clinical outcomes. Prostaglandins Leukot Essent Fatty Acids 72(4):267–272 Wessel C, Westhoff CC et al (2008) CD34(+) fibrocytes in melanocytic nevi and malignant melanomas of the skin. Virchows Arch 453(5):485–489 Wiese AH, Auer J et al (2007) Identification of gene signatures for invasive colorectal tumor cells. Cancer Detect Prev 31(4):282–295 Wong FH, Huang CY et al (2009) Combination of microarray profiling and protein-protein interaction databases delineates the minimal discriminators as a metastasis network for esophageal squamous cell carcinoma. Int J Oncol 34(1):117–128 Xue LY, Hu N et al (2006) Tissue microarray analysis reveals a tight correlation between protein expression pattern and progression of esophageal squamous cell carcinoma. BMC Cancer 6:296 Yamanaka M, Kanda K et al (2001) Analysis of the gene expression of SPARC and its prognostic value for bladder cancer. J Urol 166(6):2495–2499 Yamashita K, Upadhay S et al (2003) Clinical significance of secreted protein acidic and rich in cystein in esophageal carcinoma and its relation to carcinoma progression. Cancer 97(10):2412–2419 Yan Q, Weaver M et al (2005) Matricellular protein SPARC is translocated to the nuclei of immortalized murine lens epithelial cells. J Cell Physiol 203(1):286–294 Yang E, Kang HJ et al (2007) Frequent inactivation of SPARC by promoter hypermethylation in colon cancers. Int J Cancer 121(3):567–575 Yiu GK, Chan WY et al (2001) SPARC (secreted protein acidic and rich in cysteine) induces apoptosis in ovarian cancer cells. Am J Pathol 159(2):609–622 Yunker CK, Golembieski W et al (2008) SPARC-induced increase in glioma matrix and decrease in vascularity are associated with reduced VEGF expression and secretion. Int J Cancer 122(12):2735–2743 Zeltner L, Schittenhelm J et al (2007) The astrocytic response towards invasive meningiomas. Neuropathol Appl Neurobiol 33(2):163–168
Chapter 18
Integrin-Extracellular Matrix Interactions Christie J. Avraamides and Judith A. Varner
18.1 Introduction The tumor microenvironment is comprised of extracellular matrix molecules, tumor cells, endothelial cells, immune cells and fibroblasts. Fibroblasts synthesize extracellular matrix, and secrete growth factors and matrix metalloproteinases, which contribute to tumor growth and metastasis. Extracellular matrix proteins include collagens, laminins, fibronectin and proteoglycans (reviewed in Marastoni et al. 2008). These proteins provide mechanical support for cells, facilitate cell communication and serve as substrates for cell migration (reviewed in Kalluri 2003). Extracellular matrix (ECM) remodeling promotes embryonic development, wound healing and tumor progression. The ECM plays a key role in tumor growth and metastasis by promoting the growth of neoplastic cells, new blood vessels (angiogenesis) and new lymphatic vessels (lymphangiogenesis). The ECM may also promote the arrest of the tumor cells in capillary beds of distant organs. Intracellular signals triggered by the interaction of the ECM with integrin receptors occur in macromolecular structures called focal adhesions (Marastoni et al. 2008). In this review, we examine the contribution of integrins to tumor cell-ECM interactions. Angiogenesis, the development of new blood vessels from pre-existing vessels, is important for tumor growth and metastasis (Carmeliet 2005). Fibroblasts, tumor cells and tumor-associated macrophages secrete growth factors, such as vascular endothelial growth factor (VEGF-A) that promote angiogenesis (Adams and Alitalo 2007; Lin and Pollard 2007; Schmid and Varner 2007). Integrins regulate tumor angiogenesis by enabling endothelial cell migration and survival (reviewed in Avraamides et al. 2008). In addition, lymphangiogenesis, the growth of new lymphatic vessels, promotes tumor metastasis (Adams and Alitalo 2007; Dadras et al. 2005; Hirakawa et al. 2005; Roma et al. 2006). Tumor cells secrete factors that stimJ. A. Varner () Moores UCSD Cancer Center, University of California, San Diego, 3855 Health Sciences Drive #0819, La Jolla, CA 92093-0819, USA e-mail:
[email protected] Department of Medicine, University of California, San Diego, La Jolla, CA 92093-0819, USA M. M. Mueller, N. E. Fusenig (eds.), Tumor-Associated Fibroblasts and their Matrix, The Tumor Microenvironment 4, DOI 10.1007/978-94-007-0659-0_18, © Springer Science+Business Media B.V. 2011
347
348
C. J. Avraamides and J. A. Varner
ulate tumor and lymph node lymphangiogenesis, leading to increased lymph node metastasis (Dadras et al. 2005; Hirakawa et al. 2005; Roma et al. 2006). Recent studies indicate that integrins modulate tumor lymphangiogenesis (reviewed in Avraamides et al. 2008). As integrins regulate tumor growth and metastasis, tumor angiogenesis and tumor lymphangiogenesis by interacting with ECM proteins, integrin and the ECM are considered excellent targets for potential clinical applications.
18.2 Fibroblasts in Tumor Progression Fibroblasts affect the tumor microenvironment by synthesizing ECM, secreting proteases, and secreting growth factors (Kalluri 2003; Nyberg et al. 2008). Normal fibroblasts are embedded within the ECM of normal connective tissue, which consists of primarily of type I collagen and fibronectin. Fibroblasts interact with their surroundings via integrins. Within tumors, fibroblasts acquire an activated phenotype. Activated fibroblasts exhibit increased proliferation and enhanced secretion of ECM proteins, such as type I collagen and cellular fibronectin, which contains alternatively spliced CS-1 and EDA domains. These activated fibroblasts are called carcinoma-associated fibroblasts (CAFs) or tumor associated fibroblasts (TAFs). Tumor cell secreted growth factors such as transforming growth factor-β (TGF-β) appear to activate fibroblasts in the tumor microenvironment (Kalluri and Zeisberg 2006). Once activated, TAFs may promote tumor growth. For example, TAFs in invasive human breast carcinomas promote the growth of mammary cells and enhance tumor angiogenesis by expressing VEGF-A (Kalluri and Zeisberg 2006). Fibroblasts also remodel the ECM within tumors, producing matrix-metalloproteinases that degrade collagen fibers and thereby promote changes in integrin expression and fibroblast motility (Perentes et al. 2009). Fibroblasts may promote collective tumor cell metastasis by digesting ECM in advance of migrating clusters of tumor cells (Gaggioli et al. 2007; Friedl and Gilmour 2009).
18.3 Fibronectin Fibronectin is encoded by a single gene on human chromosome 2 (Ayad et al. 1994), and alternative splicing generates multiple isoforms. Fibronectin is a large molecular weight extracellular glycoprotein present at low concentration in basement membranes, high concentrations in remodeling extracellular matrices and at 300 µg/ml in plasma (Mosher 1984). This ECM protein is a dimer composed of two subunits (220–250 kDa) joined by a disulfide bond near the carboxy terminus. Two forms of fibronectin exist, an insoluble form of disulfide bonded oligomers and fibrils and a soluble dimeric form. Each fibronectin monomer has twelve type I,
18 Integrin-Extracellular Matrix Interactions
349
Structure of Fibronectin EDB Domain EDA Domain 1 2
Type I domain
3
4
5
6
7
8
9 10 11
RGD
12 13 14
15
IIICS
Type II domain Type III domain
Fig. 18.1 Cellular fibronectin structure. Fibronectin is comprised of two chains, which are disulfide bonded at the carboxy terminus. Cellular fibronectin isoforms contain 12 type I, 2 type II and 15 type III domains. Fibronectin isoforms arise from alternative splicing, which can include or exclude EDA, EDB and IIICS domains. Plasma fibronectin lacks the EDA and EDB domain, while cellular fibronectin isoforms can contain variable proportions of these domains. One monomer of plasma fibronectin contains a IIICS variable region, while the other does not. However, IIICS domains are expressed cellular fibronectin secreted by endothelial cells. In humans, five variants of the IIICS region generate up to 20 fibronectin isoforms. The RGD-binding integrins, such as α5β1 and αvβ3, bind the Arginine-Glycine-Aspartic Acid ( RGD) moiety located in an extended loop in the tenth type III repeat, while the integrin α4β1 bind an Glutamic Acid-Isoleucine-Leucine-Aspartic Acid-Valine (EILDV) peptide located in the alternatively spliced IIICS domain
two type II and fifteen to seventeen type III protein domains (Fig. 18.1). Alternative type III domains also form the EDA (extra domain A) and EDB (extra domain B), which are alternatively spliced domains of cellular fibronectin which are absent in plasma fibronectin. The type III V (CS) domain is present in both subunits of cellular fibronectin but only in one subunit of plasma fibronectin (Magnusson and Mosher 1998). This alternatively spliced domain contains binding sites for integrins α4β1, α4β7 and α9β1.
18.4 Collagen The collagens form a family of structurally related, fibrous glycoproteins with high glycine and proline content. Collagens have high tensile strength and are mainly produced by fibroblasts. The main types of collagen found in connective tissue are type I, II, III, V and XI; type I is the most common isoforms in connective tissue (Shoulders et al. 2009). These isoforms are called fibrillar collagens, because they assemble into cable-like fibrils, which in turn assemble into thick fibers. Not all collagens form fibers. For example, type IV collagen, the most abundant constituent of the basement membrane is nonfibrillar and these collagen molecules are organized into a flattened network. Collagen molecules are trimers consisting of three polypeptide chains called α chains that wrap around each other to form a triple helix. Collagens are characterized by a unique primary sequence in which every third amino acid is a glycine (Gly-X-Y motif). The X and Y position amino acids of the
350
C. J. Avraamides and J. A. Varner
motif are usually hydroxylysine or hydroxyproline (Shoulders et al. 2009). Prolines stabilize the helical conformation of the α chain, while glycine is regularly spaced at every third residue throughout the central region of the α chain. This structure allows tight packing of three α-helical chains to form a collagen superhelix.
18.5 Extracellular Matrix Proteins and Cancer Overexpression of provisional ECM molecules that can promote cell survival enhances tumor growth and progression (Marastoni et al. 2008). Disruption of the normal tissue ECM composition or architecture usually precedes tumor formation and can trigger genomic changes in tumor cells (Ghajar and Bissell 2008). While quiescent endothelial cells express fibronectin poorly, endothelial cells express this ECM protein during embryonic development and tumor progression. Since fibronectin is highly upregulated in transformed tissues, it has been called an oncofetal antigen. Increased fibronectin expression has been demonstrated in many tumor types, including nonsmall cell lung carcinoma (NSCLC) (Marastoni et al. 2008). In small cell lung carcinomas (SCLC), fibronectin expression is also elevated; at least 25% of tumor cells express fibronectin (Ritzenthaler et al. 2008). The EDA isoform of fibronectin is normally expressed during embryogenesis, while in adult tissue it is present only in areas of wound healing or tumor growth. The alternatively spliced EDA fibronectin is present in 47% of breast carcinoma cells and in 69% of adjacent stromal cells (Ritzenthaler et al. 2008). Transforming growth factor β (TGFβ), a growth factor expressed by many tumor cells, promotes the inclusion of the EDA domain into oncofetal fibronectin (Muro et al. 2008). Endothelial cells in tumors also secrete fibronectin that contains EDA, EDB and IIICS domains. Integrins α4β1, α5β1, and α9β1, all of which are expressed by vascular endothelial cells in tumors (Avraamides et al. 2008), facilitate cell adhesion and spreading by binding to fibronectin (Liao et al. 2002; Manabe et al. 1997). Tumor and endothelial cells secrete fibronectin containing the EDB domain, an isoform that is absent in plasma and tissues of healthy adults (Kaspar et al. 2006). However, fibronectin containing the EDB domain are present around neovascular structures, on invasive ductal breast carcinoma and in brain tumors (Kaspar et al. 2006). Most studies indicate that increased fibronectin is important for tumor angiogenesis and tumor invasion. However, loss of fibronectin can also result in loss of contact inhibition and tumor metastasis, for example, in head and neck cancers (Beier et al. 2007). Fibronectin loss is associated with the expression of human papilloma virus early protein −2, a transcription factor that suppresses fibronectin expression (Beier et al. 2007). The ECM collagen also has a role in tumor pathogenesis. For example, type IV collagen found normally in the basement membrane, is associated with tumor fibrosis and accumulates in the tumor during tumor development (Kalluri 2003). Collagen VII is necessary for tumorigenesis of Ras-transformed keratinocytes in a model of squamous cell carcinoma (Ghajar and Bissell 2008). Breast tumors exhibit
18 Integrin-Extracellular Matrix Interactions
351
tissue stiffness due deposition of large amounts of collagen by activated fibroblasts (Ghajar and Bissell 2008). This matrix stiffness acts via β1 integrin clustering and sustained activation of Rho to disrupt mammary epithelial cell differentiation. The importance of integrin β1 in tumor progression is highlighted by the fact that treatment with integrin β1 antibody reduces tumor growth and genetic deletion of integrin β1 in murine breast cancer models reduces the number of tumors (Ghajar and Bissell 2008). The ECM is thus important in promoting tumor growth and metastasis and it communicates with the surrounding cells via integrins.
18.6 The Integrin Family of ECM Receptors Integrins are receptors for extracellular matrix proteins and immunoglobulin superfamily molecules (Hynes 2002). These adhesion proteins are divalent cationdependent heterodimeric membrane glycoproteins comprised of non-covalently associated α and β subunits that promote cell attachment on the extracellular matrix. Eighteen α and eight β subunits associate to form twenty-four integrin heterodimers. Each integrin subunit consists of an extracellular domain, a single transmembrane region, and a short (approximately thirty to forty amino acids) cytoplasmic region (Hynes 2002). Growth factor or chemokine receptor signaling can alter integrin conformation (“inside-out” signaling) and regulate integrin activity. In an inactive integrin, the cytoplasmic regions of the α and β subunits are closely associated with one another. The N-termini of the α and β chains form a globular head region that is bent towards the plasma membrane (Arnaout et al. 2005; Beglova et al. 2002; Lu et al. 2001; Vinogradova et al. 2002). Growth factors promote association of talin with the beta chain tail, thereby breaking a salt bridge between the alpha and beta tails and allowing conformational shifts to occur. Unbending and elongation of the dimer and separation of the cytoplasmic domains mark integrin activation (Arnaout et al. 2005; Beglova et al. 2002; Lu et al. 2001; Vinogradova et al. 2002), allowing interaction of α and β cytoplasmic domains with intracellular proteins. Once activated, integrins bind ligands, cluster, and initiate signaling cascades. On endothelial cells, the expression of integrins α1β1, α2β1, α4β1, α5β1 and αvβ3 can be induced by growth factors or chemokines (Brooks et al. 1994a; Friedlander et al. 1995). Receptor-mediated signaling can also directly activate integrins. In carcinoma cells, integrin αvβ5 is constitutively expressed in inactive form, but it is activated by insulin-like growth factor mediated signal transduction (Brooks et al. 1997). In monocytes and other leukocytes, α4 and β2 integrins are generally inactive until chemokines, hormones, or other growth factors stimulate the cells (Grabovsky et al. 2000). Thus, integrins roles in cancer may be controlled either by intracellular signaling (inside-out signaling) or by expression. In tumors, growth factors may continuously stimulate integrin expression and activity, thereby promoting tumor growth and invasion.
C. J. Avraamides and J. A. Varner
352
18.7 The Beta 1 Integrin Subfamily and Cancer The β1 integrin subfamily contains 11 members: αvβ1, α1β1, α2β1, α3β1, α4β1, α5β1, α6β1, α7β1, α8β1, α9β1, α10β1, and α11β1. β1 integrins activate many signaling pathways, which promote cell proliferation, migration, invasion and survival (Boudreau and Varner 2004; Vogel et al. 1990; Yao et al. 2007). Fibronectin receptors include α3β1 (Takada et al. 1988), α4β1 (Elices et al. 1990), α5β1 (Pytela et al. 1985), α8β1 (Schnapp et al. 1995), α9β1 (Weinacker et al. 1995) and αvβ1 (Vogel et al. 1990) (Table 18.1). Collagen receptors include α2β1, α1β1, α3β1, α10β1 and α11β (Mercurio 2002) (Table 18.1) Integrins α4β1, α5β1 and α9β1 play roles in tumor angiogenesis (Garmy-Susini et al. 2005; Kim et al. 2000; Staniszewska et al. 2007; Vlahakis et al. 2007), while integrins α3β1 (Morini et al. 2000; Tang et al. 2008), α4β1 (Gosslar et al. 1996; Matsuura et al. 1996; Okada et al. 1999) and α5β1 (Caswell et al. 2007) also play roles in tumor invasion and metastasis. Increased β1 integrin expression levels are found in a variety of human tumors (Yao et al. 2007). High integrin β1 levels correlate with poor prognosis in cancers of the lung, pancreas and cutaneous melanoma. In addition, high β1 integrin expression levels are associated with poorer overall survival rates in patients with early stage invasive breast cancer (Yao et al. 2007). In small cell lung carcinoma (SCLC), ECM surrounding primary and metastatic tumor cells may induce integrin β1 mediated cell survival, suppressing chemotherapy-induced apoptosis. Therefore, inactivation of integrin β1 mediated survival-signaling may provide therapeutic benefit in the treatment of SCLC (Marastoni et al. 2008). β1 integrins are also important in controlling the initial proliferation of micrometastatic cancer cells disseminated in Table 18.1 Collagen and fibronectin-binding integrins
Integrins
Ligands
α1β1 α2β1 α3β1 α4β1 α5β1 α8β1 α9β1 α10β1 α11β1 αvβ1 αvβ3
Collagen, laminin Collagen, laminin Fibronectin, collagen, laminin, epiligrin, entactin Fibronectin, VCAM-1 Fibronectin, fibrinogen, L1-CAM Fibronectin Tenascin, collagen, laminin Collagen Collagen Fibronectin, vitronectin Fibronectin, vitronectin, von Willebrand factor, thrombospondin, Del-1 Fibronectin Fibronectin, fibrinogen, von Willebrand factor Fibronectin, VCAM-1
αvβ6 aIIbβ3 α4β7
A number of integrin β1 family members, which are expressed by most cells, bind fibronectin and collagen. Several αv family members also bind fibronectin. The platelet integrin aIIbβ3 and the leukocyte integrin α4β7 also bind fibronectin
18 Integrin-Extracellular Matrix Interactions
353
the lungs. While intravenously injected, nonmetastatic cancer cells ceased to proliferate after extravasating into the parenchyma of the lungs, metastatic cells continued to proliferate (Shibue and Weinberg 2009). While non-metastatic cells were unable to trigger adhesion-related FAK activation, metastatic cells activated FAK, enabling their tumor cell proliferation within the lung parenchyma in vivo. Activation of FAK depended on expression of integrin β1, and proliferation of cancer cells was diminished by knocking down the expression of either FAK or integrin β1. These results demonstrated the critical role that integrin β1-FAK signaling can play in promoting proliferation of micrometastases (Shibue and Weinberg 2009). Expression of integrin α5β1 by lung cancer cells has also been associated with a worse prognosis in lung carcinoma patients; as this integrin is not found in normal lung tissue, it may serve as therapeutic target (Ritzenthaler et al. 2008). Recent studies show that integrin α5β1 activation can be triggered by fibronectin deposited in tumors. Integrin α5β1 mediated signaling reduces cell apoptosis in a COX-2 dependent manner and results in inhibition of cyclin-dependent kinase inhibitor p21 gene expression (Marastoni et al. 2008). Recent studies showed that an activating β1 integrin mutation increases the conversion of benign papillomas to malignant squamous cell carcinomas (Ferreira et al. 2009). T188I beta1 integrin is a heterozygous mutation identified in poorly differentiated squamous cell carcinoma (SCC) that activates extracellular matrix adhesion and inhibits keratinocyte differentiation in vitro. Overexpression of this mutant in the basal layer of mouse epidermis promoted the conversion of papillomas to SCCs after chemical carcinogenesis. T188T beta 1 papillomas showed increased Erk activity and reduced differentiation. These observations showed that a genetic integrin β1 variant affects tumor progression (Ferreira et al. 2009). Other integrins, including α2, α6 and β4 subunits may promote colorectal cancer cell extravasation. Using in vivo models and intravital video microscopy, function blocking antibodies directed against these integrin subunits significantly reduced colon carcinoma cell extravasation and migration (Robertson et al. 2009).
18.8 The αv Subfamily The αv integrin subunit can combine with several different beta subunits (β1, β3, β5, β6 and β8). Integrin αvβ3 is a receptor for RGD-containing proteins such as vitronectin, fibronectin, fibrinogen and osteopontin and was the first of the alpha v integrins to be characterized (reviewed in Stupack 2005). It was also the first integrin to be found to regulate angiogenesis (Brooks et al. 1994a). Aggressive tumors, including melanoma, carcinomas of the prostate, breast, cervix and pancreas express integrin αvβ3. In cervical carcinoma, expression of αvβ3 correlates with disease progression and shorter survival. In pancreatic ductal adenocarcima, 58% of human tumors express integrin αvβ3 and expression of this integrin is associated with increased lymph node metastasis (Desgrosellier et al. 2009). Another αv integrin, αvβ6, has been shown to promote invasion and metastasis in oral carcinoma (Ramos
354
C. J. Avraamides and J. A. Varner
et al. 2002). Increased integrin αvβ6 expression has been reported in carcinomas of the lung, breast, pancreas, stomach, colon, ovary and salivary gland. Integrin αvβ6 promotes tumor cells to invade and metastasize through an αvβ6-extracellular signal-related kinase (ERK) direct signaling pathway (Peng et al. 2009). Ligated integrins lead to anchorage-dependent survival and proliferation of tumor cells. However, recently it has been demonstrated that unligated integrins can influence the malignant properties of tumor cells by activating apoptotic pathways leading to integrin-mediated death (IMD). Mechanisms that tumor cells develop to escape IMD, contributes to their metastatic behavior. Integrin αvβ3 expressed in carcinoma cells enhanced anchorage-independent tumor growth in vitro and increased lymph node metastases in vivo. These effects required recruitment of c-Src to the β3 integrin cytoplasmic tail, leading to c-Src activation, Crk-associated substrate (CAS) phosphorylation and tumor cell survival that was independent of cell adhesion or focal adhesion kinase (FAK) activation. Therefore αvβ3 contributes to tumor progression by enhancing anchorage-independent growth and by activating c-src independent of tumor cell adhesion (Desgrosellier et al. 2009).
18.9 Integrins in Angiogenesis and Lymphangiogenesis Tumor growth and metastasis are dependent on the growth of blood vessels into or proximal to the tumor mass. Integrins promote angiogenesis. β1 integrins play key roles in angiogenesis as animals with an endothelial specific deletion of the β1 integrin die during embryogenesis by E10.5 with severe vascular defects (Tanjore et al. 2007). One beta 1 integrin that promotes angiogenesis is integrin α5β1. Integrin α5β1 is poorly expressed on quiescent endothelium but is upregulated during tumor angiogenesis in both mice and humans (Kim et al. 2000). Antagonists of integrin α5β1 inhibit tumor (Kim et al. 2000), corneal (Muether et al. 2007) and choroidal (Umeda et al. 2006) angiogenesis and suppress tumor growth. Integrin α4β1 expression is induced on neovessels in murine and human tumors in response to VEGF, bFGF, IL-8 and TNFalpha. Integrin α4β1 promotes the survival of both endothelial cells and pericytes during angiogenesis. (Garmy-Susini et al. 2005). Another β1 integrin with a role in angiogenesis is integrin α9β1 (Staniszewska et al. 2007; Vlahakis et al. 2007). The direct binding of VEGF-A by integrin α9β1 promotes VEGF-A stimulated angiogenesis and blocking antibodies to α9β1 suppress VEGF-A induced angiogenesis (Vlahakis et al. 2007). Integrin αvβ3 is expressed on tumor blood vessels, but is not expressed on vessels in normal human tissues. Growth factors such as bFGF and TNF-α stimulate its expression on endothelial cells and its expression is important in survival and migration during angiogenesis. Integrin αvβ3 plays a key role in endothelial cell survival and migration during angiogenesis (Brooks et al. 1994a, b). Antagonists of either αvβ3 or αvβ5 inhibit angiogenesis and tumor growth in a variety of animal models of cancer and block choroidal angiogenesis in animal models of ocular disease (Brooks et al. 1995; Friedlander et al. 1995, 1996; Fu et al. 2007).
18 Integrin-Extracellular Matrix Interactions
355
In contrast, embryonic deletion of alpha v integrin in mice suggested αv integrins are not absolutely required for embryonic vascular development (Bader et al. 1998). Most mice lacking αv integrins die in utero at E9.5, although 20% die shortly after birth. These 20% of mice have severely abnormal vessels in the brain and intestines although other organs appear normal. In contrast to these preceding findings, studies of animals with mutations in the integrin β3 cytoplasmic tail indicate that integrin β3 is required for angiogenesis (Mahabeleshwar et al. 2006). The point mutations Y747F and Y759F in the integrin β3 cytoplasmic tail impair β3 integrin signal transduction and cell migration. Knockin mice bearing these mutations survived embryogenesis but exhibited reduced growth factor and tumor induced angiogenesis in adult mice. These studies confirm that integrin αvβ3 plays an important role in angiogenesis. Lymphatic vessels drain fluid and cells from tissues and are lined by loosely associated endothelial cells without a covering of mural cells. These vessels promote trafficking of immune cells and tumor cells. Importantly, lymph nodes are the first organs in which metastases are detected. Recent identification of lymphatic markers, including Prox-1, LYVE-1 and podoplanin made it possible to study mechanisms regulating lymphangiogenesis. The growth factors VEGF-C and VEGF-D promote lymphangiogenesis by activating the lymphatic endothelial cell receptor VEGFR-3. Integrins also regulate lymphangiogenesis. Integrins that regulate lymphangiogenesis include α9β1 and α4β1. Quiescent lymphatic endothelial cells express integrin α9β1. Integrin α9β1 null mice die 6 to 12 days after birth due to chylothoraces, an accumulation of lymph in the pleural cavity, suggesting a role for α9β1 in developmental lymphangiogenesis (Huang et al. 2000; Vlahakis et al. 2005). Integrin α4β1 is highly expressed on tumor lymphatic endothelium and antagonists of this integrin can block lymphangiogenesis and tumor metastasis.
18.10 Clinical Applications Tumor cells and proliferating endothelial cells express alternatively spliced cellular fibronectin isoforms. The EDA domain of FN can be used as a marker of tumor angiogenesis as it accumulates around blood vessels of certain tumors (Villa et al. 2008). F8, B7 and D5 are human monoclonal antibodies that have been developed against the EDA domain. The F8, B7 and D5 antibodies do not stain normal tissue but selectively stain neovascular structures on freshly isolated frozen tumor sections (Villa et al. 2008). L19 is a human antibody that is specific for the EBD domain that has been tested in preclinical animal models (Kaspar et al. 2006). L19 targets lung, colorectal or brain cancer and is able to distinguish between quiescent and growing lesions. Fusion proteins and chemical derivatives based on L19 have also been produced. Fusion proteins with IL2 (L19-IL2) have anticancer activity in orthotopic murine models of hepatocellular carcinoma and pancreatic cancer (Kaspar et al. 2006).
356
C. J. Avraamides and J. A. Varner
Peptide and antibody antagonists of αvβ3 block tumor angiogenesis and growth (Brooks et al. 1994b). Vitaxin/MEDI-522 is a humanized version of the antiintegrin αvβ3 monoclonal antibody LM609, which blocks tumor angiogenesis by inducing apoptosis in newly formed endothelial cells (Gutheil et al. 2000; McNeel et al. 2005). When Vitaxin was tested on patients with metastatic cancer who had failed other treatments, the disease stabilized without any toxicity. A Phase II study on metastatic melanoma showed that 53% patients treated with Vitaxin survived greater than one year versus 27% of patients receiving standard therapy (Hersey et al. 2005). A humanized anti-α5β1 antibody, M200 (volociximab), developed by Protein Design Labs and now by Biogen-Idec Pharmaceuticals, has shown low toxicity in Phase I studies and was evaluated in Phase II trials for metastatic melanoma, renal cell carcinoma and non-small cell lung cancer (Figlin et al. 2006; Kuwada 2007). Antibodies against integrin α5β1 decrease angiogenesis in vitro and in vivo (Bhaskar et al. 2008). Additionally, Volociximab inhibits endothelial cell growth in vitro. A mouse antibody to against α5β1, antibody 339.1, inhibits murine EC migration and tube formation and promotes cell death (Bhaskar et al. 2007). In xenograft models it inhibited the growth of tumors 60% and this correlates with a decrease in vessel density and slows tumor growth in vivo (Bhaskar et al. 2007).
18.11 Conclusions The extracellular matrix and its receptors play crucial roles in tumor development and metastasis. Tumors alter the surrounding extracellular matrices and integrin expression levels. Targeting of integrins may provide therapeutic benefit, as integrins are expressed on tumor cells as well as endothelial cells, and suppressing their functions can decrease angiogenesis and tumor metastasis, thereby preventing tumor growth and metastasis.
References Adams RH, Alitalo K (2007) Molecular regulation of angiogenesis and lymphangiogenesis. Nature Rev Mol Cell Biol 8:464–478 Arnaout MA, Mahalingam B, Xiong JP (2005) Integrin structure, allostery, and bidirectional signaling. Annu Rev Cell Dev Biol 21:381–410 Avraamides CJ, Garmy-Susini B, Varner JA (2008) Integrins in angiogenesis and lymphangiogenesis. Nat Rev Cancer 8:604–617 Ayad S, Boot-Handford RP, Humphries MJ, Kadler KE, Shuttleworth CA (1994) The extracellular matrix. Academic, San Diego Bader BL, Rayburn H, Crowley D, Hynes RO (1998) Extensive vasculogenesis, angiogenesis, and organogenesis precede lethality in mice lacking all [alpha]v integrins. Cell 95:507–519 Beglova N, Blacklow SC, Takagi J, Springer TA (2002) Cysteine-rich module structure reveals a fulcrum for integrin rearrangement upon activation. Nature Struct Biol 9:282–287
18 Integrin-Extracellular Matrix Interactions
357
Beier UH, Holtmeier C, Weise JB, Gorogh T (2007) Fibronectin suppression in head and neck cancers, inflammatory tissues and the molecular mechanisms potentially involved. Int J Oncol 30:621–629 Bhaskar V, Zhang D, Fox M, Seto P, Wong MH, Wales PE, Powers D, Chao DT, Dubridge RB, Ramakrishnan V (2007) A function blocking anti-mouse integrin alpha5beta1 antibody inhibits angiogenesis and impedes tumor growth in vivo. J Transl Med 5:61 Bhaskar V, Fox M, Breinberg D, Wong MH, Wales PE, Rhodes S, DuBridge RB, Ramakrishnan V (2008) Volociximab, a chimeric integrin alpha5beta1 antibody, inhibits the growth of VX2 tumors in rabbits. Invest New Drugs 26:7–12 Boudreau NJ, Varner JA (2004) The homeobox transcription factor Hox D3 promotes integrin [alpha]5[beta]1 expression and function during angiogenesis. J Biol Chem 279:4862–4868 Brooks PC, Clark RA, Cheresh DA (1994a) Requirement of vascular integrin [alpha]v[beta]3 for angiogenesis. Science 264:569–571 Brooks PC, Montgomery AM, Rosenfeld M, Reisfeld RA, Hu T, Klier G, Cheresh DA (1994b) Integrin alpha v beta 3 antagonists promote tumor regression by inducing apoptosis of angiogenic blood vessels. Cell 79:1157–1164 Brooks PC, Stromblad S, Klemke R, Visscher D, Sarkar FH, Cheresh DA (1995) Antiintegrin alpha v beta 3 blocks human breast cancer growth and angiogenesis in human skin. J Clin Invest 96:1815–1822 Brooks PC, Klemke RL, Schon S, Lewis JM, Schwartz MA, Cheresh DA (1997) Insulin-like growth factor receptor cooperates with integrin alpha v beta 5 to promote tumor cell dissemination in vivo. J Clin Invest 99:1390–1398 Carmeliet P (2005) Angiogenesis in life, disease and medicine. Nature 42:932–936 Caswell PT, Spence HJ, Parsons M, White DP, Clark K, Cheng KW, Mills GB, Humphries MJ, Messent AJ, Anderson KI, McCaffrey MW, Ozanne BW, Norman JC (2007) Rab25 associates with alpha5beta1 integrin to promote invasive migration in 3D microenvironments. Dev Cell 13:496–510 Dadras SS, Lange-Asschenfeldt B, Velasco P, Nguyen L, Vora A, Muzikansky A, Jahnke K, Hauschild A, Hirakawa S, Mihm MC, Detmar M (2005) Tumor lymphangiogenesis predicts melanoma metastasis to sentinel lymph nodes. Mod Pathol 18:1232–1242 Desgrosellier JS, Barnes LA, Shields DJ, Huang M, Lau SK, Prevost N, Tarin D, Shattil SJ, Cheresh DA (2009) An integrin alpha(v)beta(3)-c-Src oncogenic unit promotes anchorageindependence and tumor progression. Nat Med (in press) Elices MJ, Osborn L, Takada Y, Crouse C, Luhowskyj S, Hemler ME, Lobb RR (1990) VCAM-1 on activated endothelium interacts with the leukocyte integrin VLA-4 at a site distinct from the VLA-4/fibronectin binding site. Cell 60:577–584 Ferreira M, Fujiwara H, Morita K, Watt FM (2009) An activating beta1 integrin mutation increases the conversion of benign to malignant skin tumors. Cancer Res 69:1334–1342 Figlin RA, Kondagunta GV, Yazji S, Motzer RJ, Bukowski RM (2006) Phase II study of volociximab (M200), an [alpha]5[beta]1 anti-integrin antibody in refractory metastatic clear cell renal cell cancer (RCC). J Clin Oncol ASCO Annu Meeting Proc 24:4535 Friedl P, Gilmour D (2009) Collective cell migration in morphogenesis, regeneration and cancer. Nat Rev Mol Cell Biol 10:445–457 Friedlander M, Brooks PC, Shaffer RW, Kincaid CM, Varner JA, Cheresh DA (1995) Definition of two angiogenic pathways by distinct alpha v integrins. Science 270:1500–1502 Friedlander M, Theesfeld CL, Sugita M, Fruttiger M, Thomas MA, Chang S, Cheresh DA (1996) Involvement of integrins alpha v beta 3 and alpha v beta 5 in ocular neovascular diseases. Proc Natl Acad Sci U S A 93:9764–9769 Fu Y, Ponce ML, Thill M, Yuan P, Wang NS, Csaky KG (2007) Angiogenesis inhibition and choroidal neovascularization suppression by sustained delivery of an integrin antagonist, EMD478761. Invest Ophthalmol Vis Sci 48:5184–5190 Gaggioli C, Hooper S, Hidalgo-Carcedo C, Grosse R, Marshall JF, Harrington K, Sahai E (2007) Fibroblast-led collective invasion of carcinoma cells with differing roles for RhoGTPases in leading and following cells. Nat Cell Biol 9:1392–1400
358
C. J. Avraamides and J. A. Varner
Garmy-Susini B, Jin H, Zhu Y, Sung RJ, Hwang R, Varner J (2005) Integrin [alpha]4[beta] 1-VCAM-1-mediated adhesion between endothelial and mural cells is required for blood vessel maturation. J Clin Invest 115:1542–1551 Ghajar CM, Bissell MJ (2008) Extracellular matrix control of mammary gland morphogenesis and tumorigenesis: insights from imaging. Histochem Cell Biol 130:1105–1118 Gosslar U, Jonas P, Luz A, Lifka A, Naor D, Hamann A, Holzmann B (1996) Predominant role of alpha 4-integrins for distinct steps of lymphoma metastasis. Proc Natl Acad Sci U S A 93:4821–4826 Grabovsky V, Feigelson S, Chen C, Bleijs DA, Peled A, Cinamon G, Baleux F, Arenzana-Seisdedos F, Lapidot T, van Kooyk Y, Lobb RR, Alon R (2000) Subsecond induction of [alpha]4 integrin clustering by immobilized chemokines stimulates leukocyte tethering and rolling on endothelial vascular cell adhesion molecule 1 under flow conditions. J Exp Med 192:495–506 Gutheil JC, Campbell TN, Pierce PR, Watkins JD, Huse WD, Bodkin DJ, Cheresh DA (2000) Targeted antiangiogenic therapy for cancer using Vitaxin: a humanized monoclonal antibody to the integrin [alpha]v[beta]3. Clin Cancer Res 6:3056–3061 Hersey P, Sosman J, O’Day S, Richards J, Bedikian A, Gonzalez R, Sharfman W, Weber R, Logan T, Kirkwood JM (2005) A phase II, randomized, open-label study evaluating the antitumor activity of MEDI-522, a humanized monoclonal antibody directed against the human [alpha] v[beta]3 ([alpha]v[beta]3) integrin, [plusmn] dacarbazine (DTIC) in patients with metastatic melanoma. J Clin Oncol 2005 ASCO Annu Meeting Proc 23:7507 Hirakawa S, Kodama S, Kunstfeld R, Kajiya K, Brown LF, Detmar M (2005) VEGF-A induces tumor and sentinel lymph node lymphangiogenesis and promotes lymphatic metastasis. J Exp Med 201:1089–1099 Huang XZ, Wu JF, Ferrando R, Lee JH, Wang YL, Farese RV Jr, Sheppard D (2000) Fatal bilateral chylothorax in mice lacking the integrin alpha9beta1. Mol Cell Biol 20:5208–5215 Hynes RO (2002) Integrins: bidirectional, allosteric signaling machines. Cell 110:673–687 Kalluri R (2003) Basement membranes: structure, assembly and role in tumour angiogenesis. Nat Rev Cancer 3:422–433 Kalluri R, Zeisberg M (2006) Fibroblasts in cancer. Nat Rev Cancer 6:392–401 Kaspar M, Zardi L, Neri D (2006) Fibronectin as target for tumor therapy. Int J Cancer 118:1331– 1339 Kim S, Bell K, Mousa SA, Varner JA (2000) Regulation of angiogenesis in vivo by ligation of integrin [alpha]5[beta]1 with the central cell-binding domain of fibronectin. Am J Pathol 156:1345–1362 Kuwada SK (2007) Drug evaluation: volociximab, an angiogenesis-inhibiting chimeric monoclonal antibody. Curr Opin Mol Ther 9:92–98 Liao YF, Gotwals PJ, Koteliansky VE, Sheppard D, Van De Water L (2002) The EIIIA segment of fibronectin is a ligand for integrins [alpha]9[beta]1 and [alpha]4[beta]1 providing a novel mechanism for regulating cell adhesion by alternative splicing. J Biol Chem 277:14467–14474 Lin EY, Pollard JW (2007) Tumor-associated macrophages press the angiogenic switch in breast cancer. Cancer Res 67:5064–5066 Lu C, Takagi J, Springer TA (2001) Association of the membrane proximal regions of the [alpha] and [beta] subunit cytoplasmic domains constrains an integrin in the inactive state. J Biol Chem 276:14642–14648 Magnusson MK, Mosher DF (1998) Fibronectin: structure, assembly, and cardiovascular implications. Arterioscler Thromb Vasc Biol 18:1363–1370 Mahabeleshwar GH, Feng W, Phillips DR, Byzova TV (2006) Integrin signaling is critical for pathological angiogenesis. J Exp Med 203:2495–2507 Manabe R, Ohe N, Maeda T, Fukuda T, Sekiguchi K (1997) Modulation of cell-adhesive activity of fibronectin by the alternatively spliced EDA segment. J Cell Biol 139:295–307 Marastoni S, Ligresti G, Lorenzon E, Colombatti A, Mongiat M (2008) Extracellular matrix: a matter of life and death. Connect Tissue Res 49:203–206 Matsuura N, Puzon-McLaughlin W, Irie A, Morikawa Y, Kakudo K, Takada Y (1996) Induction of experimental bone metastasis in mice by transfection of integrin alpha 4 beta 1 into tumor cells. Am J Pathol 148:55–61
18 Integrin-Extracellular Matrix Interactions
359
McNeel DG, Eickhoff J, Lee FT, King DM, Alberti D, Thomas JP, Friedl A, Kolesar J, Marnocha R, Volkman J, Zhang J, Hammershaimb L, Zwiebel JA, Wilding G (2005) Phase I trial of a monoclonal antibody specific for [alpha]v[beta]3 integrin (MEDI-522) in patients with advanced malignancies, including an assessment of effect on tumor perfusion. Clin Cancer Res 11:7851–7860 Mercurio AM (2002) Lessons from the alpha2 integrin knockout mouse. Am J Pathol 161:3–6 Morini M, Mottolese M, Ferrari N, Ghiorzo F, Buglioni S, Mortarini R, Noonan DM, Natali PG, Albini A (2000) The alpha 3 beta 1 integrin is associated with mammary carcinoma cell metastasis, invasion, and gelatinase B (MMP-9) activity. Int J Cancer 87:336–42 Mosher DF (1984) Physiology of fibronectin. Annu Rev Med 35:561–575 Muether PS, Dell S, Kociok N, Zahn G, Stragies R, Vossmeyer D, Joussen AM (2007) The role of integrin [alpha]5[beta]1 in the regulation of corneal neovascularization. Exp Eye Res 85:356– 365 Muro AF, Moretti FA, Moore BB, Yan M, Atrasz RG, Wilke CA, Flaherty KR, Martinez FJ, Tsui JL, Sheppard D, Baralle FE, Toews GB, White ES (2008) An essential role for fibronectin extra type III domain A in pulmonary fibrosis. Am J Respir Crit Care Med 177:638–645 Nyberg P, Salo T, Kalluri R (2008) Tumor microenvironment and angiogenesis. Front Biosci 13:6537–6553 Okada T, Hawley RG, Kodaka M, Okuno H (1999) Significance of VLA-4-VCAM-1 interaction and CD44 for transendothelial invasion in a bone marrow metastatic myeloma model. Clin Exp Metastasis 17:623–629 Peng C, Liu X, Liu E, Xu K, Niu W, Chen R, Wang J, Zhang Z, Lin P, Wang J, Agrez M, Niu J (2009) Norcantharidin induces HT-29 colon cancer cell apoptosis through the alphavbeta6extracellular signal-related kinase signaling pathway. Cancer Sci 100(12):2302–2308 Perentes JY, McKee TD, Ley CD, Mathiew H, Dawson M, Padera TP, Munn LL, Jain RK, Boucher Y (2009) In vivo imaging of extracellular matrix remodeling by tumor-associated fibroblasts. Nat Methods 6:143–145 Pytela R, Pierschbacher MD, Ruoslahti E (1985) Identification and isolation of a 140 kd cell surface glycoprotein with properties expected of a fibronectin receptor. Cell 40:191–198 Ramos DM, But M, Regezi J, Schmidt BL, Atakilit A, Dang D, Ellis D, Jordan R, Li X (2002) Expression of integrin beta 6 enhances invasive behavior in oral squamous cell carcinoma. Matrix Biol 21:297–307 Ritzenthaler JD, Han S, Roman J (2008) Stimulation of lung carcinoma cell growth by fibronectinintegrin signalling. Mol Biosyst 4:1160–1169 Robertson JH, Yang SY, Winslet MC, Seifalian AM (2009) Functional blocking of specific integrins inhibit colonic cancer migration. Clin Exp Metastasis (in press) Roma AA, Magi-Galluzzi C, Kral MA, Jin TT, Klein EA, Zhou M (2006) Peritumoral lymphatic invasion is associated with regional lymph node metastases in prostate adenocarcinoma. Mod Pathol 19:392–398 Schmid MC, Varner JA (2007) Myeloid cell trafficking and tumor angiogenesis. Cancer Lett 250:1–8 Schnapp LM, Hatch N, Ramos DM, Klimanskaya IV, Sheppard D, Pytela R (1995) The human integrin alpha 8 beta 1 functions as a receptor for tenascin, fibronectin, and vitronectin. J Biol Chem 270:23196–23202 Shibue T, Weinberg RA (2009) Integrin beta1-focal adhesion kinase signaling directs the proliferation of metastatic cancer cells disseminated in the lungs. Proc Natl Acad Sci U S A 106:10290– 10295 Shoulders MD, Raines RT et al (2009) Collagen structure and stability. Annu Rev Biochem 78:929–958 Staniszewska I, Zaveri S, Del Valle L, Oliva I, Rothman VL, Croul SE, Roberts DD, Mosher DF, Tuszynski GP, Marcinkiewicz C (2007) Interaction of [alpha]9[beta]1 integrin with thrombospondin-1 promotes angiogenesis. Circ Res 100:1308–1316 Stupack DG (2005) Integrins as a distinct subtype of dependence receptors. Cell Death Differ 12:1021–1030
360
C. J. Avraamides and J. A. Varner
Takada Y, Wayner EA, Carter WG, Hemler ME (1988) Extracellular matrix receptors, ECMRII and ECMRI, for collagen and fibronectin correspond to VLA-2 and VLA-3 in the VLA family of heterodimers. J Cell Biochem 37:385–393 Tang J, Wu YM, Zhao P, Yang XM, Jiang JL, Chen ZN (2008) Overexpression of HAb18G/CD147 promotes invasion and metastasis via alpha3beta1 integrin mediated FAK-paxillin and FAKPI3K-Ca2+ pathways. Cell Mol Life Sci 65:2933–2942 Tanjore H, Zeisberg EM, Gerami-Naini B, Kalluri R (2007) [beta]1 integrin expression on endothelial cells is required for angiogenesis but not for vasculogenesis. Dev Dyn 237:75–82 Umeda N, Kachi S, Akiyama H, Zahn G, Vossmeyer D, Stragies R, Campochiaro PA (2006) Suppression and regression of choroidal neovascularization by systemic administration of an [alpha]5[beta]1 integrin antagonist. Mol Pharmacol 69:1820–1828 Villa A, Trachsel E, Kaspar M, Schliemann C, Sommavilla R, Rybak JN, Rosli C, Borsi L, Neri D (2008) A high-affinity human monoclonal antibody specific to the alternatively spliced EDA domain of fibronectin efficiently targets tumor neo-vasculature in vivo. Int J Cancer 122:2405– 2413 Vinogradova O, Velyvis A, Velyviene A, Hu B, Haas T, Plow E, Qin J (2002) A structural mechanism of integrin [alpha]IIb[beta]3 [ldquo]inside-out[rdquo] activation as regulated by its cytoplasmic face. Cell 110:587–597 Vlahakis NE, Young BA, Atakilit A, Sheppard D (2005) The lymphangiogenic vascular endothelial growth factors VEGF-C and -D are ligands for the integrin [alpha]9[beta]1. J Biol Chem 280:4544–4552 Vlahakis NE, Young BA, Atakilit A, Sheppard D (2007) Integrin [alpha]9[beta]1 directly binds to vascular endothelial growth factor (VEGF)-A and contributes to VEGF-A-induced angiogenesis. J Biol Chem 282:15187–15196 Vogel BE, Tarone G, Giancotti FG, Gailit J, Ruoslahti E (1990) A novel fibronectin receptor with an unexpected subunit composition (alpha v beta 1). J Biol Chem 265:5934–5937 Weinacker A, Ferrando R, Elliott M, Hogg J, Balmes J, Sheppard D (1995) Distribution of integrins alpha v beta 6 and alpha 9 beta 1 and their known ligands, fibronectin and tenascin, in human airways. Am J Respir Cell Mol Biol 12:547–556 Yao ES, Zhang H, Chen YY, Lee B, Chew K, Moore D, Park C (2007) Increased beta1 integrin is associated with decreased survival in invasive breast cancer. Cancer Res 67:659–664
Chapter 19
The Multifaceted Role of Cancer Associated Fibroblasts in Tumor Progression Hans Petter Eikesdal and Raghu Kalluri
Abbreviations αSMA α-Smooth muscle actin bFGF Basic fibroblast growth factor BMH Bone marrow-derived hematopoietic precursor cells BMM Bone marrow-derived mesenchymal precursor cells CAF Cancer associated fibroblast CK5 Cytokeratin 5 DAB 3,3′-diaminobenzidine ECM Extracellular matrix EMT Epithelial-to-mesenchymal transition EndMT Endothelial-to-mesenchymal transition ER Estrogen receptor FAP Fibroblast activation protein FSP1 Fibroblast specific protein-1 HGF Hepatocyte growth factor IGF Insulin-like growth factor KGF Keratinocyte growth factor MMP Matrix metalloproteinase MMPI Matrix metalloproteinase inhibitor NG2 NG2 chondroitin sulfate proteoglycan R. Kalluri () Division of Matrix Biology, Department of Medicine, Beth Israel Deaconess Medical Center & Harvard Medical School, 330 Brookline Ave, E/CLS Room #11-090, Center for Life Sciences, 02115 Boston, MA, USA e-mail:
[email protected] Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA, USA Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA, USA H. P. Eikesdal Department of Oncology, Haukeland University Hospital, 5021 Bergen, Norway M. M. Mueller, N. E. Fusenig (eds.), Tumor-Associated Fibroblasts and their Matrix, The Tumor Microenvironment 4, DOI 10.1007/978-94-007-0659-0_19, © Springer Science+Business Media B.V. 2011
361
362
H. P. Eikesdal and R. Kalluri
PDGF Platelet derived growth factor SCC Squamous cell carcinoma TGFβ Transforming growth factor-β.
19.1 Introduction The non-random appearance of metastases from breast cancer was described by Stephen Paget already in 1889 (Paget 1889). His “seed-and-soil” hypothesis states that the tumor cells need a certain soil to grow. Today, our knowledge of the “seed” has increased immensely, but we are still almost as far away as Dr. Paget from understanding why metastases appear where they do (Mueller and Fusenig 2004). In this chapter we will discuss what is currently known about fibroblasts and their role in establishing a proper “soil” for tumor cells to grow in. A malignant epithelial tumor can consist of up to 90% stroma (Elenbaas and Weinberg 2001). In Hodgkin lymphoma the case is even more extreme, with the malignant Hodgkin and Reed-Sternberg cells accounting only 0.1–3%, and the activated tumor stroma making the rest (Aldinucci et al. 2004). The tumor stroma is made up of multiple non-malignant cell populations, including fibroblasts, adipocytes, endothelial and inflammatory cells (Mueller and Fusenig 2004). Additionally the tumor stroma contains the extracellular matrix, the scaffold upon which the cells attach. There is a huge interest in tumor stroma research today, and our knowledge of the cancer associated fibroblast (CAF) has changed from being viewed as a passive bystander to becoming an important comediator of cancer progression. As we will discuss below fibroblasts are actually a family of many different cell types, they can arise from different cell populations and they might even initiate epithelial cancers.
19.2 Fibroblast Subtypes The exact definition of a fibroblast is difficult to make. Historically, fibroblasts have been cathegorized based on their morphology—spindle shaped cells scattered in the tissue stroma (Strutz et al. 1995). Recently, an increasing number of molecular markers have been identified—enabling immunolabeling of fibroblasts (De Wever et al. 2008). CAFs were originally defined as α-smooth muscle actin (αSMA) positive cells (Elenbaas and Weinberg 2001). However, experiments carried out in our laboratory and elsewhere demonstrate that tumors contain many subpopulations of fibroblasts, with different immunoreactivity and phenotypic characteristics (Fries et al. 1994; Ostman and Augsten 2009; Sugimoto et al. 2006). Apart from αSMA, subpopulations of tumor fibroblasts show immunoreactivity for NG2 chondroitin sulfate proteoglycan (NG2), platelet derived growth factor receptor-β (PDGFRβ) or fibroblast specific protein-1 (FSP1) (Fig. 19.1a). These fibroblast subtypes have
19 The Multifaceted Role of Cancer Associated Fibroblasts in Tumor Progression
363
Fig. 19.1 a FSP1+ fibroblasts in a squamous cell carcinoma of the anus. Immunolabeling with an FSP1 antibody (3, 3′-diaminobenzidine (DAB), brown). Arrow points to a FSP1+ fibroblast. * indicates an FSP1+ fibroblast “misplaced” in an island of epithelial cancer cells. E: epithelial cancer cells. S: stroma. Scale bar: 20 µm. b Fluorescence image of a section of testis from a FSP1-GFP transgenic mouse, indicating that FSP1 ( green) and αSMA ( red) positive fibroblasts are to a large extent separate cell populations. Arrow points to a FSP1+ and αSMA− fibroblast in the stroma (S). No immunolabeling used for FSP-GFP, whereas TRITC-conjugated immunolabeling was used for aSMA. Blue: DAPI nuclear staining. Scale bars: 100 µm. c Squamous cell carcinoma (SCC) of the anus. Before: CK5 immunolabeling (DAB, brown) of the epithelial cancer cells in the tumor as they looked before in vitro culture. After: CK5 immunolabeling after reimplanting from in vitro culture a comixture of CAFs and epithelial cancer cells from the anal SCC. Scale bars: 100 µm. d Highly proliferative anal SCC fibroblasts in vitro. Immunolabeling with Ki67 antibody ( green). Scale bars: 100 µm. e Crosstalk between epithelial cancer cells and CAFs. The cancer cells secrete growth factors X to stimulate the fibroblasts, whereas the fibroblasts secrete growth factors Y to stimulate the cancer cells. f Estrogen receptors (ER) are highly expressed both on anal SCC cancer cells and fibroblasts. Arrow points to a fibroblast. E: epithelial cancer cells. S: stroma. Immunolabeling with phosphorylated ERα antibody (DAB, brown). Scale bar: 100 µm
different localization, as for instance αSMA, PDGFRβ and NG2 typically label pericytes, whereas fibroblasts scattered in the stroma are αSMA or FSP1 positive (Sugimoto et al. 2006). Another important marker specific for CAFs and pericytes in tumors is fibroblast activation protein-α (FAP), which is stroma-specific in 90% of epithelial cancers tested (Haslam and Woodward 2003; Santos et al. 2009). Immunolabeling experiments are prone to unspecific staining, and to assess further the distribution of αSMA and FSP1 fibroblast subtypes, we used a transgenic mouse, where the FSP1 promoter is coupled to green fluorescent protein (FSP1-GFP) whereas αSMA was visualized by immunolabeling with a TRITC
364
H. P. Eikesdal and R. Kalluri
fluorochrome1. Accordingly any cell that is FSP1 positive will be green and any cell that is αSMA positive will be red. In this experiment we found that αSMA and FSP1 positive fibroblasts are indeed two separate cell populations (Fig. 19.1b). We are currently in the process of examining the role of these fibroblast subtypes for primary tumor growth and metastasis. There has been a lot of discussion about the specificity of FSP1 as a fibroblast marker. Originally identified as a cytoplasmic protein in mesenchymal cells, present after embryonic day 8.5, FSP1 was described as a selective marker of fibroblasts or epithelial cells undergoing epithelial-to-mesenchymal transitition (EMT) during kidney fibrosis (Strutz et al. 1995). Later studies have shown that FSP1, also known as S100A4, is also expressed in metastatic tumor cells and in very early granulocytic lineages from the bone marrow—but not in normal epithelial cells (Bhowmick et al. 2004a; Inoue et al. 2005). The mts1 gene, encoding S100A4, was found upregulated in normal cells with high invasiveness like macrophages, neutrophils and Tlymphocytes (Grigorian et al. 1994). However, testing a set of different macrophage antibodies it was found that apparently FSP1 positive macrophages were rather a case of unspecific staining of fibroblasts using macrophage antibodies (Inoue et al. 2005). In a recent paper by Bhowmick et al. they find that FSP1 is expressed in dendritic cells, using FSP1 genetic tracing (Boomershine et al. 2009). They did not find FSP1 expressed in macrophages, B- and T-lymphocytes. The specificity of their analysis is hampered by the use of antibody staining for dendritic cells—a method inherently prone to unspecificity. What can probably be extracted from the above discussion about FSP1 is the fact that it is a good marker for stroma cells and epithelial cells that have or are in the process of transitioning to mesenchymal cells during EMT. Accordingly, in our hands we do not find FSP1 expressed in epithelial cells in any organ of the body.
19.3 How do Fibroblasts Differ in Malignant and Normal Tissues? Cancer cells alter the environment they sit in to generate a reactive stroma, and this includes the activation of fibroblasts mediated by growth factors such as transforming growth factor-β (TGFβ) and platelet derived growth factor (PDGF) (Arendt et al. 2009; Mueller and Fusenig 2004). CAFs differ from fibroblasts in normal tissues in at least two ways; they have a very different gene expression profile, and they have a profound stimulatory effect on the epithelial cancer cells (Haviv et al. 2009). There are several articles demonstrating a shift in gene expression from fibroblasts to CAFs (Allinen et al. 2004; Fiegl et al. 2006; Hanson et al. 2006; Hu et al. 2005; Ricci et al. 2005). Accordingly, stromal cells within an epithelial cancer have been found to exhibit a particular gene expression profile or signature (Farmer et al. 1
The FSP1-GFP mice were a kind gift from E.G. Neilson.
19 The Multifaceted Role of Cancer Associated Fibroblasts in Tumor Progression
365
2009; Finak et al. 2008). The CAF genome is generally hypomethylated, reflecting the fact that there are extensive epigenetic changes in tumor fibroblasts with a multitude of genes overexpressed (Alphonso and Alahari 2009; Arendt et al. 2009). Also, not only do the stromal cells in breast cancer differ in gene expression from the normal breast—but there is also a difference between the intratumoral and the juxtatumoral area (Iacobuzio-Donahue et al. 2002). The tumor promoting effect of CAFs on epithelial cells is well established (Elenbaas and Weinberg 2001; Orimo et al. 2005). In an elegant study from 1999, Olumi et al. demonstrated that CAFs could make a non-tumorigenic human prostatic epithelial cell line tumorigenic, whereas normal tissue fibroblasts were not able to elicit tumorigenesis (Olumi et al. 1999). A potential flaw in their set-up was the fibroblast isolation method where tissue adjacent to prostate cancer was taken out, and purified based on filtration and centrifugation. This method, along with karyotyping, strongly suggests that they had a pure CAF population, but there still could be a minor contamination of prostate carcinoma cells. We established separate primary cultures of fibroblasts and epithelial cancer cells from a spontaneous squamous cell carcinoma arising in the anus of a mouse. After expanding the fibroblasts in vitro we injected them subcutaneously in Nu/Nu mice and observed that no tumors appeared. We then injected the same Nu/Nu mice with the epithelial tumor cells without any tumors appearing. Thereafter fibroblasts from a normal anus was coinjected with epithelial tumor cells, and again no tumors developed. Finally, tumor fibroblasts and epithelial tumor cells were coinjected and seven out of seven mice developed squamous cell carcinomas with the same characteristics as the original tumor (Fig. 19.1c). The purity of the fibroblast and tumor cell primary cultures was only established by immunolabelling in this experiment, but again—the data strongly suggests there is a profound stimulatory effect of CAFs on carcinoma cells. Interestingly, normal tissue fibroblasts are also modified by physiologic processes, such as pregnancy. The stroma of multiparous rats were shown experimentally to inhibit tumor growth when breast cancer cells where injected, whereas tumors developed much faster when grown in virgin rats (Maffini et al. 2005). This suggests that inhibitory stromal factors in the breast of multiparous animals might explain why pregnancy protects from breast cancer (Kelsey et al. 1993). Also, the correctly activated fibroblasts or CAFs seems necessary to permit tumor progression. This points to the profound importance of the crosstalk between CAFs and epithelial cancer cells discussed below.
19.4 Where do Fibroblasts Come from? Fibroblasts in tumors originate from various cell compartments. Apart from resident fibroblasts, they can arise through recruitment of mesenchymal stem cells from the bone marrow and through transdifferentiation of epithelial and endothelial cells (Arendt et al. 2009; Shimoda et al. 2010).
366
H. P. Eikesdal and R. Kalluri
Resident fibroblasts are traditionally believed to be a major source of new fibroblasts during for instance scar formation and other fibrotic processes, but the contribution of resident fibroblasts in malignant tumors is less clear (Ostman and Augsten 2009; Shimoda et al. 2010). The extensive proliferative capacity of CAFs grown in vitro suggests that a large number of them derive from local expansion within the tumor stroma (Fig. 19.1d). In a study of liver metastases from patients with colorectal cancer, it was found that fibroblasts within the metastatic tumor tissue had a very high resemblence to the resident fibroblasts in the liver (Mueller et al. 2007). However, the lack of genetic tracing of the fibroblasts and their similarity assessment based on immunolabeling does not rule out a significant contribution from other sources. Epithelial-to-mesenchymal transition (EMT) is a process in which epithelial cells change phenotype to become mesenchymal (Polyak and Weinberg 2009). In this way mesenchymal cells such as fibroblasts, chondrocytes and muscle cells are generated (Kalluri and Neilson 2003). During the EMT process epithelial cells loose their polarity and cell–cell adhesions, and instead acquire increased motility, invasiveness and anchorage-independence—typical fibroblast features (Polyak and Weinberg 2009). Classically, epithelial cells loose E-cadherin expression and acquire mesenchymal features such as vimentin and α-smooth muscle actin when transitioning (Thiery and Sleeman 2006; Tse and Kalluri 2007). EMT is a highly regulated process, where certain signaling pathways, such as the TGFβ pathway, are switched on while others are silenced (Polyak and Weinberg 2009; Tse and Kalluri 2007). EMT is increasingly appreciated as an important process of fibroblast generation, in diseases such as cancer, kidney fibrosis and liver fibrosis (Kalluri and Neilson 2003; Thiery and Sleeman 2006; Zeisberg et al. 2007). Additionally, the EMT process makes epithelial cancer cells less anchorage-dependent and thus facilitates invasion and metastasis. Recently, endothelial-to-mesenchymal transition (EndMT), a process similar to EMT, was discovered as a source of fibroblasts in disease processes such as cardiac fibrosis and desmoplasia within malignant tumors (Zeisberg et al. 2007a, b). In this process, local endothelial cells transition to acquire mesenchymal features. EndMT, like EMT, was originally described as a developmental process during embryogenesis, but as it turns out, it is a major contributor of fibroblasts in adult tissues as well (Zeisberg et al. 2007b). Another source of CAFs are the recruitment of mesenchymal stem cells from the bone marrow (Ostman and Augsten 2009). Similar to bone marrow-derived hematopoietic precursor cells (BMH), bone marrow-derived mesenchymal precursor cells (BMM cells) are recruited to the primary tumor and to sites of metastasis (Hung et al. 2005; Klopp et al. 2007; Mishra et al. 2008; Studeny et al. 2004). These BMM cells increase the growth and metastatic potential of cancer cells, as demonstrated both for breast cancer and pancreatic cancer in mice (Hwang et al. 2008; Karnoub et al. 2007). A subpopulation of fibroblasts found in the cancer stroma are bone marrow-derived (Direkze et al. 2004, 2006; Ishii et al. 2003; LaRue et al. 2006; Mori et al. 2005; Studeny et al. 2004), and they
19 The Multifaceted Role of Cancer Associated Fibroblasts in Tumor Progression
367
can amount to as much as 25–40% of the myofibroblast population (Direkze et al. 2004; Ishii et al. 2003). Fibroblasts involved in wound healing and cancer stroma development seem to arise in part from BMM cells (Klopp et al. 2007; Mishra et al. 2008; Mori et al. 2005; Studeny et al. 2004), although BMH cells may also be fibroblast precursors (LaRue et al. 2006). When mesenchymal stem cells are isolated from the bone marrow and labelled ex vivo, they localize to the tumor stroma as CAFs after being injected i.v. (Hung et al. 2005; Mishra et al. 2008; Studeny et al. 2004). What remains to be assessed is the relevance of the different CAF sources, i.e. what happens if any one of these recruitment pathways are blocked?
19.5 Crosstalk Between Fibroblasts and Cancer Cells Crosstalk between the fibroblasts and epithelial cancer cells is important for tumor progression, and a large number of growth factors are known to be secreted by these two cell compartments to stimulate each other (Fig. 19.1e). For an extensive description of which growth factors that are involved in this crosstalk, several detailed reviews are available (Bhowmick et al. 2004b; De Wever et al. 2008; Elenbaas and Weinberg 2001; Kalluri and Zeisberg 2006; Matsumoto and Nakamura 2006; Mueller and Fusenig 2004; Zvaifler 2006). Cancer cells produce growth factors such as PDGF, TGFβ, basic fibroblast growth factor (bFGF) to activate the tumor-associated fibroblasts (Mueller and Fusenig 2004). In return, the fibroblasts secrete growth factors, such as hepatocyte growth factor (HGF), keratinocyte growth factor (KGF) and insulin-like growth factor 1 and 2 (IGF-1 and -2), to stimulate the epithelial cancer cells (Bhowmick et al. 2004b; Nakamura et al. 1997). Additionally, cancer cells stimulate the fibroblasts to produce matrix metalloproteinases (MMPs), which binds to the cancer cells and is used for degradation and invasion through the ECM (Pavlaki and Zucker 2003). Also, the crosstalk involves other cell populations, like the tumor endothelium and inflammatory cells (Mueller and Fusenig 2004). Besides secreting growth factors to stimulate the cancer cells, CAFs can quench the inhospitable microenvironment generated by hypoxia and low pH by removing acidic metabolites (Alphonso and Alahari 2009). In a recent interesting article, Pavlides et al. describe what they call a “reverse Warburg effect” (Pavlides et al. 2009). They hypothesize that tumor cells can induce aerobic glycolysis in CAFs causing them to produce lactate and pyruvate which the tumor cells utilize as an energy source. In mice that are caveolin-1 (cav-1) deficient they show that fibroblasts are constitutively activated, like CAFs, with upregulated TGFβ signaling and tumor promoting capabilities. Proteomic analysis demonstrated that these cav-1 null fibroblasts have several glycolytic enzymes upregulated which are key regulators of the Warburg effect (Pavlides et al. 2009).
368
H. P. Eikesdal and R. Kalluri
19.6 Fibroblasts and Interaction with the Extracellular Matrix The extracellular matrix (ECM) consists of long, filamentous proteins making up a mesh on which cells attach, migrate and communicate with each other (Kalluri 2003). Fibroblasts are situated, along with other stroma cells, within this scaffold, and they are the principal source of ECM proteins (Elenbaas and Weinberg 2001; Kalluri and Zeisberg 2006). In support of the crucial role of fibroblasts in producing the tumor ECM, one finds that malignancies that has little fibrotic tissue, like poorly differentiated small cell lung cancer, have very few CAFs as well (Elenbaas and Weinberg 2001). The constituents of the ECM are, besides fibrillar proteins like collagens and elastin, proteins such as fibronectin and laminin and various glycosaminoglycans (Strutz et al. 1995). Additionally, various growth factors are stored in the ECM, and these factors are liberated during matrix breakdown by cancer cells to stimulate angiogenesis and cancer progression (Kalluri 2003). The ECM differs between the organs of the body, and empirically the chance of tumor cells proliferating to develop a solid mass increases when tumor cells are injected orthotopically instead of ectopically, indicating that the soil (including the local mixture of ECM and stromal cells) matters for tumor growth (Elenbaas and Weinberg 2001). Basement membranes are a special type of ECM, with a dense, sheet-like structure consisting of type IV collagen and laminin, typically separating epithelial cells or endothelial cells from the stroma (Kalluri 2003). In the tumor stroma basement membranes are partly degraded, liberating both pro- and antiangiogenic proteins (Kalluri 2003; Nyberg et al. 2005). When tumor cells are coinjected with basement membrane proteins, like in Matrigel, it increases tumor take and metastasis frequency profoundly (Elenbaas and Weinberg 2001). Fibroblasts are an important source of ECM-degrading proteases such as MMPs, which highlights their crucial role in regulating ECM turnover (Marx 2008). It was originally thought that cancer cells were the major source of MMPs based on immunolabeling data, but in situ hybridization studies have later demonstrated that MMPs are synthetized by fibroblasts and immune cells in the tumor stroma (Pavlaki and Zucker 2003). However, the cancer cells stimulate the stroma cells to produce MMPs and make docking proteins that will facilitate adherence of MMPs to the cancer cell cytoplasmic membrane (Pavlaki and Zucker 2003). The MMP family consist of different enzymes, many of which are upregulated in malignant tumors (Pavlaki and Zucker 2003). Their role in breakdown of the ECM to facilitate cancer invasion and angiogenesis led to large scale clinical trials to test MMP inhibitors (MMPIs) as anti-cancer drugs. However, these trials failed, and many reasons have been proposed to explain this (Eikesdal et al. 2002; Eikesdal and Kalluri 2009; Marx 2008; Pavlaki and Zucker 2003). It seems that MMPs are most important in the early stages of tumor progression, whereas the clinical MMPI trials were undertaken in patients with very advanced disease. A crucial point in the MMPI trials were also the opposing roles that MMPIs can have. At one end
19 The Multifaceted Role of Cancer Associated Fibroblasts in Tumor Progression
369
they may inhibit invasion through the ECM, but at the same time they may inhibit the liberation of antiangiogenic proteins—thereby promoting angiogensis and thus tumor progression (Mueller and Fusenig 2004; Pavlaki and Zucker 2003). Furthermore, different MMPs have different roles in ECM turnover, and many of the tested MMPIs were broad-spectrum inhibitors blocking a whole series of the enzymes (Pavlaki and Zucker 2003). Fibroblasts attach to the ECM via integrins (Alphonso and Alahari 2009; Chiquet et al. 2003; Desgrosellier and Cheresh 2010; Galbraith and Sheetz 1998). The cytoskeleton is reorganized and integrins and focal adhesion contacts between the cell and the ECM dynamically change location as the fibroblast migrates (Galbraith and Sheetz 1998). The integrin receptors sense stretch and thereby activate, through focal adhesion complexes, Rho-ROCK and MAPK signaling and upregulate tenascin C (Chiquet et al. 2003). This tenascin C upregulation can be blocked by β1-integrin antibodies (Chiquet et al. 2003). Fibroblasts also express the Nischarin protein which binds to α5β1-integrin to facilitate cell migration on fibronectin (Alphonso and Alahari 2009). Integrin α11 was found to be upregulated in CAFs in non-small cell lung cancer, and overexpression of α11β1 on tumor fibroblasts increased IGF-1 secretion which stimulated tumor growth (Desgrosellier and Cheresh 2010). Interestingly, antibody-based therapy, directed at β1-integrin is currently being tested clinically (Desgrosellier and Cheresh 2010). The fibrotic response commonly seen in epithelial cancers is called desmoplasia and is found as a dense, amorphous tissue with few cells in it. The most intense desmoplastic response is typically seen at the advancing tumor periphery (Elenbaas and Weinberg 2001). The desmoplastic tumor tissue differs from fibrous tissue elsewhere by increased amounts of fibrillar collagens (in particular type I collagen), fibronectin, proteoglycans, and glycosaminoglycans (Elenbaas and Weinberg 2001). The hardness of tumors, probably due to the desmoplastic tissue response, is the focus of research as a potential marker of malignant as opposed to benign neoplasms—using elastography (Bilgen et al. 2003; Giovannini et al. 2009; Itoh et al. 2006; Pallwein et al. 2008). Furthermore, the increased desmoplastic reaction seen in certain epithelial cancers has been proposed to correlate with increased risk of recurrence and metastasis (see Sect. 19.8 below).
19.7 Fibroblast Mutations in Epithelial Malignancies The current understanding of epithelial carcinogenesis is dominated by the notion that cancer develops due to sequential accumulation of epithelial mutations (Hanahan and Weinberg 2000; Maffini et al. 2004; Vogelstein and Kinzler 1993). However, there is an ongoing debate whether fibroblasts and other stromal cells in epithelial malignancies can acquire mutations and whether this matters (Campbell et al. 2009; Eng et al. 2009). There are multiple studies, both in human and animal tumors, indicating that fibroblast mutations exist (Hardwick et al. 2008; Hill et al. 2005; Kurose et al. 2001, 2002; Moinfar et al. 2000; Patocs et al. 2007; Weber et al.
370
H. P. Eikesdal and R. Kalluri
2006, 2007). Contrary to this, others find that CAF mutations are very rare, or nonexistant (Campbell et al. 2009; Crawford et al. 2009; Hu et al. 2005; Qiu et al. 2008; Shimoda et al. 2010; Walter et al. 2008). The potential flaws in analyzing fibroblast mutations in human cancer material are many. In most studies tumor stroma is isolated from tissue sections using laser microdissection, followed by PCR and genomic analysis. With this method the DNA yield is frequently low and when this is compensated for by an extensive number of PCR-cycles, the risk of false positive “mutations” increases dramatically (Campbell et al. 2009). This is further exaggerated by fragmented DNA found in formalin-fixed, paraffin sections from archival material. Thus, there are strict recommendations regarding DNA quality and concentration when using this method (Campbell et al. 2009). However, even if the DNA yield is good, there is no exact anatomical border between carcinoma cells and stroma cells inside a tumor (Fig. 19.1a). Therefore, the commonly used method of laser microdissection of stromal areas is prone to include carcinoma cells as well. Alternatively, tumor tissue removed from the patient can be degraded enzymatically and sorted by flow cytometry based on fibroblast specific antigens. The problem with this method is unspecific staining, which can occur with all antibodies. Also, antigens like FSP1 are localized inside the cytoplasm (Skalli et al. 1989; Strutz et al. 1995), requiring permeabilization to immunolabel them—which makes unspecific staining an even larger problem. Apart from these issues, another one is even more challenging: EMT is potentially a frequent phenomenon in epithelial cancers. So fibroblasts harboring mutations could simply be carcinoma cells that have undergone EMT—in which case the same mutations would be found in stroma and epithelial cancer cells. However, the presence of different mutations in these two compartments indicates that stromal mutations occurs independently and not only through EMT of cancer cells (Hill et al. 2005; Kurose et al. 2002; Moinfar et al. 2000; Shiraishi et al. 2006). The issue of fibroblast marker specificity can be controlled for more thoroughly in animal models by genetic tracing. I.e. a traceable genetic change is inflicted once a cell expresses a fibroblast marker. We investigated this by crossing mice expressing Cre recombinase under the FSP1 promoter (FSP1-Cre mice) with R26R-EYFP reporter mice, which have the enhanced yellow fluorescent protein (EYFP) coupled to the constitutively active ROSA promoter2. In these mice any cell expressing FSP1 will express Cre recombinase, and delete the STOP codons that inhibit the transcription of the ROSA promoter—thus causing continous expression of EYFP in any cell that has been or is expressing FSP1 ever after. In our hands this genetic tracing showed that FSP1 was specifically expressed only in stromal cells. In another interesting experiment, Maffini et al. gave rats with cleared mammary fat pads the N-nitrosomethylurea (NMU) carcinogen or sham treatment, before injecting mammary epithelial cells exposed to NMU or vehicle into the breast (Maffini et al. 2004). Breast carcinomas developed only if the stroma had been exposed FSP1-Cre and R26R-EYFP reporter mice were kindly provided by E.G. Neilson and B. G. Neel respectively. 2
19 The Multifaceted Role of Cancer Associated Fibroblasts in Tumor Progression
371
to NMU, whereas NMU-treated mammary epithelial cells caused no tumorigenesis if the stroma had not been exposed to the carcinogen previously. A potential pitfall in this experimental set-up is whether the mammary fat pads were completely cleared of mammary duct epithelium before starting the NMU exposure. However, the investigators assured complete mammary gland removal by whole mount staining of the excised fat pads. The experiment suggests that fibroblast mutations are required for epithelial cancer growth. Clearly if mutations are present in fibroblasts, there are many ways in which this would change our understanding of epithelial cancers. One question is whether fibroblast mutations occur before mutations in the epithelial cell layer. A few papers indicate that certain mutations are found exclusively in the stroma, and not in the epithelial cancer cells (Hill et al. 2005; Jacoby et al. 1997; Kurose et al. 2002; Moinfar et al. 2000). Such independent fibroblast mutations suggests they can have an initiating role in carcinoma development. Furthermore, fibroblast mutations mean that therapy targeted at this stromal cell compartment could face the same inherent problems of acquired drug resistance as other malignant cells, due to genetic instability. Interestingly though, investigators found that the occurrence of p53 mutations in fibroblasts sensitize tumors to chemotherapy—not the opposite (Lafkas et al. 2008). Finally, it seems that the stimulatory role of CAFs to make non-tumorigenic epithelial cells tumorigenic persists even if the CAFs are removed—suggesting that CAFs have a mutagenic effect on the epithelial cells (Mueller et al. 2001; Mueller and Fusenig 2004; Tlsty 2001) From this assumption one can speculate that perhaps fibroblasts normally inhibit epithelial cell layers to avoid mutagenesis. Therefore, if fibroblasts acquire mutations that reduce their control function, the epithelium would hyperproliferate and undergo neoplastic changes as well.
19.8 The Prognostic Relevance of Fibrosis The increased desmoplastic reaction seen in certain epithelial cancers has been proposed to correlate with increased risk of recurrence and metastasis. This has been suggested by clinical studies of desmoplastic breast cancer, non-small cell lung cancer and scirrhous gastric cancer (linitis plastica) (An et al. 2008; Cardone et al. 1997; Chen et al. 2002; Maeshima et al. 2002). There is however a debate as to the prognostic relevance of desmoplasia. When cases of scirrhous gastric cancer are matched with non-scirrhous cases with the same tumor stage, the difference in survival disappears—indicating that the prognostic factor is not the fibrosis content per se, but rather the tumor stage (Park et al. 2009). Additionally, increased fibrosis in other cancers, such as desmoplastic malignant melanoma does not seem to infer a worse prognosis (Livestro et al. 2005). It is known that high breast density, detected by mammography, gives a 4–6 fold increased risk of breast cancer (Boyd et al. 2007). It is not known why this is so, but increased collagen I content and upregulated IGF-1 signaling have been proposed
372
H. P. Eikesdal and R. Kalluri
as tumor stimulatory factors (Arendt et al. 2009; Yang et al. 2010). There are only a few publications where breast density has been investigated in detail, on a molecular level—so hopefully future research will address this issue further (Arendt et al. 2009). In animal xenograft studies desmoplasia is a rare phenomenon, and in those cases where desmoplastic cancers have been observed, cancer growth was slower than in less desmoplastic tumors (Shao et al. 2000). Fibroblasts may inhibit tumor growth by secreting lysyl oxidase which crosslinks collagen and elastin to protect ECM from protease degradation (Elenbaas and Weinberg 2001). On the other hand, fibrotic tissue may also protect the tumor cells from immune cells attacking them, and facilitate tumor growth by physically shielding the malignant tissue (Elenbaas and Weinberg 2001). In an experimental tumor model, preventing desmoplasia therapeutically by blocking collagen synthesis led to increased tumor growth and metastasis (Elenbaas and Weinberg 2001). Integrin receptors are commonly upregulated in epithelial cancers, and the stimulatory effect of increased fibrillar collagens within desmoplastic tumors seems to be mediated via collagen-integrin interaction (Desgrosellier and Cheresh 2010; Sethi et al. 1999) Apart from promoting cell growth and survival, adherence of colorectal cancer cells to collagen I conferred resistance to 5-FU chemotherapy. Interestingly this resistance could be counteracted by therapy directed at the αvβ3 and αvβ5 integrins (Conti et al. 2008). Using laser microdissection, Finak et al. assessed stromal gene expression in 53 patients with breast cancer and was able to derive a stromal gene expression profile strongly associated with clinical outcome (Finak et al. 2008). The stroma of poor outcome patients showed upregulation of genes related to angiogenesis, hypoxia and macrophage recruitment and downregulation of for instance negative regulators of Wnt signaling. Genes upregulated in the good outcome group were typically related to eliciting an immune response of T-lymphocytes and natural killer (NK) cells in the stroma (Finak et al. 2008). The stroma gene expression signature also came out as an independent prognostic variable in multivariate analysis, alongside known prognostic factors such as estrogen receptor status, HER2 status and lymph node involvement. In another study it was also established that a particular stromal signature was linked to a worse disease outcome in breast, lung and gastric cancer patients (Chang et al. 2004). Furthermore, a recent study showed that a stromarelated gene signature was predictive of resistance to neoadjuvant chemotherapy in advanced breast cancer (Farmer et al. 2009). However, this last analysis was undertaken on whole tumor material, i.e. not a selective analysis of tumor stromal cells. After quantifying αSMA positive CAFs in 192 human colorectal carcinomas, it was established that an increased number of fibroblasts in the tumor was independently related to a shortened disease-free survival, and in multivariate analysis, the number of αSMA fibroblasts came up as an independent prognostic factor (Tsujino et al. 2007). Similarily, increased expression of fibroblast activated protein (FAP) by immunolabeling was associated with a higher risk of disease recurrence postoperatively in colon and pancreatic cancer patients (Cohen et al. 2008; Henry et al. 2007). Also, the expression of secreted protein acidic and rich in cysteine (SPARC)
19 The Multifaceted Role of Cancer Associated Fibroblasts in Tumor Progression
373
by peritumoral fibroblasts in pancreatic carcinoma is associated with a worse prognosis, when analyzed by multivariate analysis (Infante et al. 2007). The expression of SPARC by the tumor cells was of no importance.
19.9 Therapy Directed at Cancer Associated Fibroblasts Due to the stimulatory role of fibroblasts in malignant neoplasms, there is a natural interest in designing therapy directed at the CAF cell population. Such therapy is not yet implemented clinically, but based on the success of targeting tumor endothelial cells, another stroma component, the potential of such treatment is definitively there. Strategies tested so far focus on two different ways of targeting CAFs, one being the direct inhibition of fibroblast function and the other to inhibit paracrine signaling between fibroblasts and other cell populations in the tumor (Arendt et al. 2009). Therapy directed at the tumor fibroblasts is made feasible either by specific fibroblast markers or by overexpression of common antigens on CAFs which make them especially sensitive to treatment. FAP is a membrane-bound serine protease which is overexpressed on and relatively specific for CAFs (Lebeau et al. 2009). FAP-specific protoxins can be synthesized which are selectively activated in the tumor stroma to kill the fibroblasts, as well as nearby epithelial tumor cells due to a bystander effect (Lebeau et al. 2009). Antibodies to FAP, conjugated with an anti-mitotic toxin, also demonstrated therapeutic potential in vivo (Ostermann et al. 2008). Genetic inactivation or pharmacologic inhibition of FAP by different small molecular compounds was also shown to cause tumor growth inhibition (Santos et al. 2009). Estrogen receptors (ER) are classically known to be expressed by subgroups of breast cancer cells, but additionally these receptors are expressed in the tumor stroma (Fig. 19.1f). Moreover, stromal ER expression seems more important than epithelial ER expression for estrogen-induced epithelial cell proliferation (Haslam and Woodward 2003). Certain types of antihormonal therapy, such as tamoxifen and fulvestrant, function as antagonists of ER and are used clinically for breast cancer treatment. However, the high expression of ER on stromal cells suggests that the therapeutic response of such drugs in part could be due to CAF inhibition (Arendt et al. 2009). Aromatase inhibitors, another group of antihormonal therapy in breast cancer therapy, inhibit aromatase function in the tumor stroma and stromal cells elsewhere in the body. Interestingly ER positive cells in the tumor stroma are stimulated by estrogen even when the breast cancer cells are ER negative (Arendt et al. 2009; Gupta et al. 2007). However, the clinical importance of this is doubtful as antihormonal therapy has been shown repeatedly to have no effect in ER negative breast cancer. The other way of targeting tumor fibroblast function is by interference with paracrine growth factor signaling. As mentioned above there is a large number of signaling substances that are secreted between various cell population in the tumor
374
H. P. Eikesdal and R. Kalluri
stroma and between the cancer cells and the stroma. PDGF is secreted by epithelial cells and epithelial cancers to stimulate CAFs and pericytes through activation of PDGFR (Pietras et al. 2008). The crucial role of PDGFR signaling in fibroblasts is demonstrated by the increased fibrotic stroma response seen when cancer cells are transfected to overexpress PDGF (Arendt et al. 2009). Imatinib mesylate, a multiple tyrosine kinase inhibitor halts proliferation and modulates cytokine expression in human tumor fibroblasts from colorectal metastases (Shibue and Weinberg 2009). Imatinib was found experimentally to halt tumor growth in part through PDGFR signal inhibition on CAFs, and in part through angiogenesis inhibition (Pietras et al. 2008). There is currently a lot of interest in developing drugs directed at HGF and its receptor c-Met (Eder et al. 2009; Peruzzi and Bottaro 2006; Porter 2010). HGF is a growth factor primarily secreted by fibroblasts to activate c-Met on epithelial cells (Birchmeier et al. 2003; Christensen et al. 2003; Eder et al. 2009; Haslam and Woodward 2003). The c-Met receptor is commonly hyperactivated on epithelial cancer cells, either because of paracrine or autocrine stimulation, or because of acquiring activating mutations (Birchmeier et al. 2003). This activation makes the tumor cells proliferate and become more invasive and motile, as part of an induced EMT programme (Birchmeier et al. 2003). Most of the c-Met directed drugs currently undergoing clinical trials are receptor tyrosine kinase inhibitors (TKIs), but antibody based strategies are also being assessed (Eder et al. 2009). Additionally NK4 and geldanamycin antagonists of HGF and c-Met respectively are being tested therapeutically (Birchmeier et al. 2003). The advantage of antibodies are their selectivity, the long half life, and that they elicit an immune response against the targeted cell. However, the small molecule TKIs can usually be administered orally, have better tissue penetration and a lower overall production cost (Eder et al. 2009). Several c-Met TKIs are now in phase II/III randomized trials, showing promising response rates clinically (Eder et al. 2009; Porter 2010). Although >150 cancer cell lines expressed c-Met and showed sensitivity to c-Met TKIs in vitro, it seemed that overexpression or amplification of the receptor was required for an in vivo effect (Christensen et al. 2007). Only approximately 5% of the cell lines tested exhibited such increased c-Met expression (Christensen et.al. 2007). Also, many of current compounds are broad-spectrum TKIs, antagonizing not only c-Met but a series of other receptor tyrosine kinases as well, which makes it difficult to interpret the isolated role of c-Met inhibition (Eder et al. 2009). The extensive tumor necrosis induced by some of the c-Met inhibitors points for instance to additional antiangiogenic effects (Christensen et al. 2003; Zou et al. 2007). The tumor fibroblasts can also be targeted indirectly by drugs directed at the ECM and ECM turnover. Tenascin C is upregulated in the ECM of many malignant tumors, and is a potential target for radioimmunotherapy, where tenascin antibodies are conjugated with radioisotopes (De Santis et al. 2006). Such therapy has shown promising results in glioma patients, and is also being developed further for epithelial malignancies (De Santis et al. 2006; Goetz et al. 2003). Apart from killing
19 The Multifaceted Role of Cancer Associated Fibroblasts in Tumor Progression
375
the tumor cells, such targeted irradiation will have the potential to inflict collateral damage on tumor stromal cells. As we described above MMPIs have already been tested in clinical trials, and effort is still made to develop these drugs further despite negative results in the past. These compounds target the MMPs produced by CAFs. Yet another potential way of targeting the tumor stroma is by genetically engineering mesenchymal cells from the bone marrow to produce for instance interferon β—which is then delivered by transfected cells travelling via the blood stream to the tumor stroma (Marx 2008). Therapy affecting CAFs can also occur as side effects of treatment directed at other cell populations. Drugs such as sunitinib and sorafenib are marketed as angiogenesis inhibitors, but they also inhibit receptors expressed on fibroblast subpopulations, such as PDGFRα and PDGFRβ (Oudard et al. 2007). Drugs are being developed clinically which target integrin receptors, and the survival of fibroblasts as well as many other cell populations within a tumor could be reduced if integrinECM interactions are disrupted (Desgrosellier and Cheresh 2010; Sethi et al. 1999). Various integrin-directed therapies are currently being tested in cancer patients (Desgrosellier and Cheresh 2010). Finally, chemotherapy utilizes the increased proliferation rate in cancer cells to achieve selective cytotoxicity in the tumor. CAFs also proliferate extensively, and could therefore be targeted by chemotherapy and this may contribute to the tumor regression observed clinically (Fig. 19.1d).
19.10 Conclusion Fibroblasts play a crucial role in cancer progression. They provide the scaffold on which the epithelial tumor cells sit, and they produce growth factors which promote tumor cell proliferation. Our increasing knowledge of the crosstalk between fibroblasts and tumor cells points to a potential for fibroblast-directed therapy in the future, as a new way of treating epithelial malignancies. The current controversy of fibroblast mutations should be investigated further, as the occurrence of such genetic alterations within the non-malignant cell populations would dramatically alter our understanding of the carcinogenesis process. Additionally, the origin of fibroblasts occuring in primary tumors and metastases should be investigated further to better understand how tumor fibroblasts are recruited. Today, tumor fibroblasts have become key players in the generation of the tumor microenvironment, and we are starting to understand how the “soil” affects the “seed” during tumor progression. Acknowledgements This work was primarily supported by National Institutes of Health Grant DK62987 and partially by National Institutes of Health Grants DK55001, DK61688, AA13913, and CA12550 and funds from the Department of Medicine for the Division of Matrix Biology at Beth Israel Deaconness Medical Center. This work was also supported by the Champalimaud Foundation. HPE was supported by grants from the University of Bergen and the Eckbo Legacy, Norway.
376
H. P. Eikesdal and R. Kalluri
References Aldinucci D et al (2004) Interactions between tissue fibroblasts in lymph nodes and Hodgkin/ Reed-Sternberg cells. Leuk Lymphoma 45:1731–1739 Allinen M et al (2004) Molecular characterization of the tumor microenvironment in breast cancer. Cancer Cell 6:17–32 Alphonso A, Alahari SK (2009) Stromal cells and integrins: conforming to the needs of the tumor microenvironment. Neoplasia 11:1264–1271 An JY et al (2008) Borrmann type IV: an independent prognostic factor for survival in gastric cancer. J Gastrointest Surg 12:1364–1369 Arendt LM et al (2009) Stroma in breast development and disease. Semin Cell Dev Biol 21:11–18 Bhowmick NA et al (2004a) Stromal fibroblasts in cancer initiation and progression. Nature 432:332–337 Bhowmick NA et al (2004b) TGF-beta signaling in fibroblasts modulates the oncogenic potential of adjacent epithelia. Science 303:848–851 Bilgen M et al (2003) Elastography imaging of small animal oncology models: a feasibility study. Ultrasound Med Biol 29:1291–1296 Birchmeier C et al (2003) Met, metastasis, motility and more. Nat Rev Mol Cell Biol 4:915–925 Boomershine CS et al (2009) Autoimmune pancreatitis results from loss of TGFbeta signalling in S100A4-positive dendritic cells. Gut 58:1267–1274 Boyd NF et al (2007) Mammographic density and the risk and detection of breast cancer. N Engl J Med 356:227–236 Campbell I et al (2009) Clonal mutations in the cancer-associated fibroblasts: the case against genetic coevolution. Cancer Res 69:6765–6768, discussion 9 Cardone A et al (1997) Prognostic value of desmoplastic reaction and lymphocytic infiltration in the management of breast cancer. Panminerva Med 39:174–177 Chang HY et al (2004) Gene expression signature of fibroblast serum response predicts human cancer progression: similarities between tumors and wounds. PLoS Biol 2:E7 Chen CY et al (2002) Peritoneal carcinomatosis and lymph node metastasis are prognostic indicators in patients with Borrmann type IV gastric carcinoma. Hepatogastroenterology 49:874–877 Chiquet M et al (2003) How do fibroblasts translate mechanical signals into changes in extracellular matrix production? Matrix Biol 22:73–80 Christensen JG et al (2003) A selective small molecule inhibitor of c-Met kinase inhibits c-Metdependent phenotypes in vitro and exhibits cytoreductive antitumor activity in vivo. Cancer Res 63:7345–7355 Christensen JG et al (2007) Cytoreductive antitumor activity of PF-2341066, a novel inhibitor of anaplastic lymphoma kinase and c-Met, in experimental models of anaplastic large-cell lymphoma. Mol Cancer Ther 6:3314–3322 Cohen SJ et al (2008) Fibroblast activation protein and its relationship to clinical outcome in pancreatic adenocarcinoma. Pancreas 37:154–158 Conti JA et al (2008) The desmoplastic reaction surrounding hepatic colorectal adenocarcinoma metastases aids tumor growth and survival via alphav integrin ligation. Clin Cancer Res 14:6405–6413 Crawford Y et al (2009) PDGF-C mediates the angiogenic and tumorigenic properties of fibroblasts associated with tumors refractory to anti-VEGF treatment. Cancer Cell 15:21–34 De Santis R et al (2006) Low and high tenascin-expressing tumors are efficiently targeted by ST2146 monoclonal antibody. Clin Cancer Res 12:2191–2196 De Wever O et al (2008) Stromal myofibroblasts are drivers of invasive cancer growth. Int J Cancer 123:2229–2238 Desgrosellier JS, Cheresh DA (2010) Integrins in cancer: biological implications and therapeutic opportunities. Nat Rev Cancer 10:9–22 Direkze NC et al (2004) Bone marrow contribution to tumor-associated myofibroblasts and fibroblasts. Cancer Res 64:8492–8495
19 The Multifaceted Role of Cancer Associated Fibroblasts in Tumor Progression
377
Direkze NC et al (2006) Bone marrow-derived stromal cells express lineage-related messenger RNA species. Cancer Res 66:1265–1269 Eder JP et al (2009) Novel therapeutic inhibitors of the c-Met signaling pathway in cancer. Clin Cancer Res 15:2207–2214 Eikesdal HP, Kalluri R (2009) Drug resistance associated with antiangiogenesis therapy. Semin Cancer Biol 19:310–317 Eikesdal HP et al (2002) Hyperthermia exhibits anti-vascular activity in the s.c. BT4An rat glioma -lack of interaction with the angiogenesis inhibitor batimastat. Int J Hyperthermia 18:141–152 Elenbaas B, Weinberg RA (2001) Heterotypic signaling between epithelial tumor cells and fibroblasts in carcinoma formation. Exp Cell Res 264:169–184 Eng C et al (2009) Genomic alterations in tumor stroma. Cancer Res 69:6759–6764 Farmer P et al (2009) A stroma-related gene signature predicts resistance to neoadjuvant chemotherapy in breast cancer. Nat Med 15:68–74 Fiegl H et al (2006) Breast cancer DNA methylation profiles in cancer cells and tumor stroma: association with HER-2/neu status in primary breast cancer. Cancer Res 66:29–33 Finak G et al (2008) Stromal gene expression predicts clinical outcome in breast cancer. Nat Med 14:518–527 Fries KM et al (1994) Evidence of fibroblast heterogeneity and the role of fibroblast subpopulations in fibrosis. Clin Immunol Immunopathol 72:283–292 Galbraith CG, Sheetz MP (1998) Forces on adhesive contacts affect cell function. Curr Opin Cell Biol 10:566–571 Giovannini M et al (2009) Endoscopic ultrasound elastography for evaluation of lymph nodes and pancreatic masses: a multicenter study. World J Gastroenterol 15:1587–1593 Goetz C et al (2003) Locoregional radioimmunotherapy in selected patients with malignant glioma: experiences, side effects and survival times. J Neurooncol 62:321–328 Grigorian M et al (1994) Modulation of mts1 expression in mouse and human normal and tumor cells. Electrophoresis 15:463–468 Gupta PB et al (2007) Systemic stromal effects of estrogen promote the growth of estrogen receptor-negative cancers. Cancer Res 67:2062–2071 Hanahan D, Weinberg RA (2000) The hallmarks of cancer. Cell 100:57–70 Hanson JA et al (2006) Gene promoter methylation in prostate tumor-associated stromal cells. J Natl Cancer Inst 98:255–261 Hardwick JC et al (2008) Bone morphogenetic protein signalling in colorectal cancer. Nat Rev Cancer 8:806–812 Haslam SZ, Woodward TL (2003) Host microenvironment in breast cancer development: epithelial-cell-stromal-cell interactions and steroid hormone action in normal and cancerous mammary gland. Breast Cancer Res 5:208–215 Haviv I et al (2009) Origin of carcinoma associated fibroblasts. Cell Cycle 8:589–595 Henry LR et al (2007) Clinical implications of fibroblast activation protein in patients with colon cancer. Clin Cancer Res 13:1736–1741 Hill R et al (2005) Selective evolution of stromal mesenchyme with p53 loss in response to epithelial tumorigenesis. Cell 123:1001–1011 Hu M et al (2005) Distinct epigenetic changes in the stromal cells of breast cancers. Nat Genet 37:899–905 Hung SC et al (2005) Mesenchymal stem cell targeting of microscopic tumors and tumor stroma development monitored by noninvasive in vivo positron emission tomography imaging. Clin Cancer Res 11:7749–7756 Hwang RF et al (2008) Cancer-associated stromal fibroblasts promote pancreatic tumor progression. Cancer Res 68:918–926 Iacobuzio-Donahue CA et al (2002) The desmoplastic response to infiltrating breast carcinoma: gene expression at the site of primary invasion and implications for comparisons between tumor types. Cancer Res 62:5351–5357 Infante JR et al (2007) Peritumoral fibroblast SPARC expression and patient outcome with resectable pancreatic adenocarcinoma. J Clin Oncol 25:319–325
378
H. P. Eikesdal and R. Kalluri
Inoue T et al (2005) Antibodies against macrophages that overlap in specificity with fibroblasts. Kidney Int 67:2488–2493 Ishii G et al (2003) Bone-marrow-derived myofibroblasts contribute to the cancer-induced stromal reaction. Biochem Biophys Res Commun 309:232–240 Itoh A et al (2006) Breast disease: clinical application of US elastography for diagnosis. Radiology 239:341–350 Jacoby RF et al (1997) A juvenile polyposis tumor suppressor locus at 10q22 is deleted from nonepithelial cells in the lamina propria. Gastroenterology 112:1398–1403 Kalluri R (2003) Basement membranes: structure, assembly and role in tumour angiogenesis. Nat Rev Cancer 3:422–433 Kalluri R, Neilson EG (2003) Epithelial-mesenchymal transition and its implications for fibrosis. J Clin Invest 112:1776–1784 Kalluri R, Zeisberg M (2006) Fibroblasts in cancer. Nat Rev Cancer 6:392–401 Karnoub AE et al (2007) Mesenchymal stem cells within tumour stroma promote breast cancer metastasis. Nature 449:557–563 Kelsey JL et al (1993) Reproductive factors and breast cancer. Epidemiol Rev 15:36–47 Klopp AH et al (2007) Tumor irradiation increases the recruitment of circulating mesenchymal stem cells into the tumor microenvironment. Cancer Res 67:11687–11695 Kurose K et al (2001) Genetic model of multi-step breast carcinogenesis involving the epithelium and stroma: clues to tumour-microenvironment interactions. Hum Mol Genet 10:1907–1913 Kurose K et al (2002) Frequent somatic mutations in PTEN and TP53 are mutually exclusive in the stroma of breast carcinomas. Nat Genet 32:355–357 Lafkas D et al (2008) P53 mutations in stromal fibroblasts sensitize tumors against chemotherapy. Int J Cancer 123:967–971 LaRue AC et al (2006) Hematopoietic origins of fibroblasts: I. In vivo studies of fibroblasts associated with solid tumors. Exp Hematol 34:208–218 Lebeau AM et al (2009) Targeting the cancer stroma with a fibroblast activation protein-activated promelittin protoxin. Mol Cancer Ther 8(5):1378–1386 Livestro DP et al (2005) Biology of desmoplastic melanoma: a case-control comparison with other melanomas. J Clin Oncol 23:6739–6746 Maeshima AM et al (2002) Modified scar grade: a prognostic indicator in small peripheral lung adenocarcinoma. Cancer 95:2546–2554 Maffini MV et al (2004) The stroma as a crucial target in rat mammary gland carcinogenesis. J Cell Sci 117:1495–1502 Maffini MV et al (2005) Stromal regulation of neoplastic development: age-dependent normalization of neoplastic mammary cells by mammary stroma. Am J Pathol 167:1405–1410 Marx J (2008) Cancer biology. All in the stroma: cancer’s Cosa Nostra. Science 320:38–41 Matsumoto K, Nakamura T (2006) Hepatocyte growth factor and the Met system as a mediator of tumor-stromal interactions. Int J Cancer 119:477–483 Mishra PJ et al (2008) Carcinoma-associated fibroblast-like differentiation of human mesenchymal stem cells. Cancer Res 68:4331–4339 Moinfar F et al (2000) Concurrent and independent genetic alterations in the stromal and epithelial cells of mammary carcinoma: implications for tumorigenesis. Cancer Res 60:2562–2566 Mori L et al (2005) Fibrocytes contribute to the myofibroblast population in wounded skin and originate from the bone marrow. Exp Cell Res 304:81–90 Mueller MM, Fusenig NE (2004) Friends or foes—bipolar effects of the tumour stroma in cancer. Nat Rev Cancer 4:839–849 Mueller MM et al (2001) Tumor progression of skin carcinoma cells in vivo promoted by clonal selection, mutagenesis, and autocrine growth regulation by granulocyte colony-stimulating factor and granulocyte-macrophage colony-stimulating factor. Am J Pathol 159:1567–1579 Mueller L et al (2007) Stromal fibroblasts in colorectal liver metastases originate from resident fibroblasts and generate an inflammatory microenvironment. Am J Pathol 171:1608–1618 Nakamura T et al (1997) Induction of hepatocyte growth factor in fibroblasts by tumor-derived factors affects invasive growth of tumor cells: in vitro analysis of tumor-stromal interactions. Cancer Res 57:3305–3313
19 The Multifaceted Role of Cancer Associated Fibroblasts in Tumor Progression
379
Nyberg P et al (2005) Endogenous inhibitors of angiogenesis. Cancer Res 65:3967–3979 Olumi AF et al (1999) Carcinoma-associated fibroblasts direct tumor progression of initiated human prostatic epithelium. Cancer Res 59:5002–5011 Orimo A et al (2005) Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion. Cell 121:335–348 Ostermann E et al (2008) Effective immunoconjugate therapy in cancer models targeting a serine protease of tumor fibroblasts. Clin Cancer Res 14:4584–4592 Ostman A, Augsten M (2009) Cancer-associated fibroblasts and tumor growth—bystanders turning into key players. Curr Opin Genet Dev 19(1):67–73 Oudard S et al (2007) Treatment options in renal cell carcinoma: past, present and future. Ann Oncol 18 10:x25–x31 Paget S (1889) Distribution of secondary growths in cancer of the breast. Lancet 1:571–573 Pallwein L et al (2008) Prostate cancer diagnosis: value of real-time elastography. Abdom Imaging 33:729–735 Park JC et al (2009) Clinicopathological aspects and prognostic value with respect to age: an analysis of 3362 consecutive gastric cancer patients. J Surg Oncol 99:395–401 Patocs A et al (2007) Breast-cancer stromal cells with TP53 mutations and nodal metastases. N Engl J Med 357:2543–2551 Pavlaki M, Zucker S (2003) Matrix metalloproteinase inhibitors (MMPIs): the beginning of phase I or the termination of phase III clinical trials. Cancer Metastasis Rev 22:177–203 Pavlides S et al (2009) The reverse Warburg effect: aerobic glycolysis in cancer associated fibroblasts and the tumor stroma. Cell Cycle 8:3984–4001 Peruzzi B, Bottaro DP (2006) Targeting the c-Met signaling pathway in cancer. Clin Cancer Res 12:3657–3660 Pietras K et al (2008) Functions of paracrine PDGF signaling in the proangiogenic tumor stroma revealed by pharmacological targeting. PLoS Med 5:e19 Polyak K, Weinberg RA (2009) Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits. Nat Rev Cancer 9:265–273 Porter J (2010) Small molecule c-Met kinase inhibitors: a review of recent patents. Expert Opin Ther Pat 20:159–177 Qiu W et al (2008) No evidence of clonal somatic genetic alterations in cancer-associated fibroblasts from human breast and ovarian carcinomas. Nat Genet 40:650–655 Ricci F et al (2005) Stromal responses to carcinomas of the pancreas: juxtatumoral gene expression conforms to the infiltrating pattern and not the biologic subtype. Cancer Biol Ther 4:302–307 Santos AM et al (2009) Targeting fibroblast activation protein inhibits tumor stromagenesis and growth in mice. J Clin Invest 119:3613–3625 Sethi T et al (1999) Extracellular matrix proteins protect small cell lung cancer cells against apoptosis: a mechanism for small cell lung cancer growth and drug resistance in vivo. Nat Med 5:662–668 Shao ZM et al (2000) Human breast carcinoma desmoplasia is PDGF initiated. Oncogene 19:4337–4345 Shibue T, Weinberg RA (2009) Integrin beta1-focal adhesion kinase signaling directs the proliferation of metastatic cancer cells disseminated in the lungs. Proc Natl Acad Sci U S A 106:10290– 10295 Shimoda M et al (2010) Carcinoma-associated fibroblasts are a rate-limiting determinant for tumour progression. Semin Cell Dev Biol 21:19–25 Shiraishi H et al (2006) Early genetic instability of both epithelial and stromal cells in esophageal squamous cell carcinomas, contrasted with Barrett’s adenocarcinomas. J Gastroenterol 41:1186–1196 Skalli O et al (1989) Alpha-smooth muscle actin, a differentiation marker of smooth muscle cells, is present in microfilamentous bundles of pericytes. J Histochem Cytochem 37:315–321 Strutz F et al (1995) Identification and characterization of a fibroblast marker: FSP1. J Cell Biol 130:393–405 Studeny M et al (2004) Mesenchymal stem cells: potential precursors for tumor stroma and targeted-delivery vehicles for anticancer agents. J Natl Cancer Inst 96:1593–1603
380
H. P. Eikesdal and R. Kalluri
Sugimoto H et al (2006) Identification of fibroblast heterogeneity in the tumor microenvironment. Cancer Biol Ther 5:1640–1646 Thiery JP, Sleeman JP (2006) Complex networks orchestrate epithelial-mesenchymal transitions. Nat Rev Mol Cell Biol 7:131–142 Tlsty TD (2001) Stromal cells can contribute oncogenic signals. Semin Cancer Biol 11:97–104 Tse JC, Kalluri R (2007) Mechanisms of metastasis: epithelial-to-mesenchymal transition and contribution of tumor microenvironment. J Cell Biochem 101:816–829 Tsujino T et al (2007) Stromal myofibroblasts predict disease recurrence for colorectal cancer. Clin Cancer Res 13:2082–2090 Vogelstein B, Kinzler KW (1993) The multistep nature of cancer. Trends Genet 9:138–141 Walter K et al (2008) Pancreatic cancer associated fibroblasts display normal allelotypes. Cancer Biol Ther 7:882–888 Weber F et al (2006) Total-genome analysis of BRCA1/2-related invasive carcinomas of the breast identifies tumor stroma as potential landscaper for neoplastic initiation. Am J Hum Genet 78:961–972 Weber F et al (2007) Microenvironmental genomic alterations and clinicopathological behavior in head and neck squamous cell carcinoma. JAMA 297:187–195 Yang WT et al (2010) Decreased TGFbeta signaling and increased COX2 expression in high risk women with increased mammographic breast density. Breast Cancer Res Treat 119:305–314. Zeisberg EM et al (2007a) Discovery of endothelial to mesenchymal transition as a source for carcinoma-associated fibroblasts. Cancer Res 67:10123–10128 Zeisberg EM et al (2007b) Endothelial-to-mesenchymal transition contributes to cardiac fibrosis. Nat Med 13(8):952–961 Zeisberg M et al (2007) Fibroblasts derive from hepatocytes in liver fibrosis via epithelial to mesenchymal transition. J Biol Chem 282(32):23337–23347 Zou HY et al (2007) An orally available small-molecule inhibitor of c-Met, PF-2341066, exhibits cytoreductive antitumor efficacy through antiproliferative and antiangiogenic mechanisms. Cancer Res 67:4408–4417 Zvaifler NJ (2006) Relevance of the stroma and epithelial-mesenchymal transition (EMT) for the rheumatic diseases. Arthritis Res Ther 8:210
Part VI
Therapeutic Application/Targeting
.
Chapter 20
Cancer Associated Fibroblasts as Therapeutic Targets Christian Rupp, Helmut Dolznig, Christian Haslinger, Norbert Schweifer and Pilar Garin-Chesa
20.1 Introduction The tumor stroma is an essential, intrinsic part of epithelial cancers and plays a primary role during carcinogenesis. Extensive clinical evidence and a variety of experimental mouse models have demonstrated the active role of the tumor stroma in promoting tumor growth. The advances in our understanding of the molecular basis for cancer initiation and progression provide the basis for the design of novel targeted agents that selectively address deregulated pathways in malignant cells. Drugs that target the stromal component of tumors may represent a further important approach to the overall control of cancer. The discovery and development of molecularly targeted drugs requires translational research, which include the identification of new molecular targets, target validation and the development of appropriate models to test the new drugs with regard to their mechanism of action, safety and efficacy before translating these findings into the clinic. Here we will address the challenges for drug development of new therapeutic agents directed towards the tumor stroma, in particular those targeting the cancer associated fibroblasts (CAFs), the limitations in the available experimental models and the complexity of the model systems in which the new targets can be studied in detail.
P. Garin-Chesa () Boehringer Ingelheim RCV GmbH & Co KG Dr. Boehringer-Gasse 5-11, 1130 Vienna, Austria
[email protected] Institute of Pathology, Medical University of Vienna, Waehringer Guertel 18-20, 1090 Vienna, Austria
M. M. Mueller, N. E. Fusenig (eds.), Tumor-Associated Fibroblasts and their Matrix, The Tumor Microenvironment 4, DOI 10.1007/978-94-007-0659-0_20, © Springer Science+Business Media B.V. 2011
383
384
C. Rupp et al.
20.2 Histological and Molecular Heterogeneity of Human Cancers Carcinomas which comprise the majority of human cancers are composed of malignant epithelial cells as well as mesenchyme-derived stromal cells, such as fibroblasts, myofibroblasts, endothelial cells, pericytes, smooth muscle and hematopoietic cells embedded in a matrix of extracellular proteins (ECM). Different histological subtypes of carcinomas exist and the extent and composition of the stroma varies among tumors. Certain tumor types such as those occurring in the pancreas, breast, and colon (Fig. 20.1) are characterized by the presence of a prominent stromal reaction (desmoplasia). The fibroblasts in those tumors express markers of activated fibroblasts such as fibroblast activation protein alpha (FAPα) (Garin-Chesa et al. 1990) and alpha-smooth muscle actin (α-SMA) (Desmouliere et al. 2004) that differ from their expression in resting fibroblasts of the adjacent normal tissues (Fig. 20.1a), indicating the phenotypic differences between normal and tumor fibroblasts. Different subsets of CAFs have been observed to occur in different tumor types (Huber et al. 2006; Koperek et al. 2007) suggesting that the activation programs of CAFs in cancer may be linked to the tissue of origin and might indicate functional differences of CAFs in tumor invasion and metastasis (Sugimoto et al. 2006). Several clinicopathologic studies have shown that the characteristics of the tumor stroma correlate with prognostic factors and patient survival (Koperek et al. 2007; Hasebe et al. 2001; Kunz-Schughart and Knuechel 2002).
Fig. 20.1 a Cancer associated fibroblasts are the main cellular stromal component of carcinomas and display molecular heterogeneity. Carcinomas arising in the pancreas, breast and colon display prominent stroma reaction (desmoplasia) separating the clusters of tumor cells. The fibroblasts in those tumors express markers of activated fibroblasts such as fibroblast activation protein (FAPα) and alpha-smooth muscle actin (α-SMA). FAPα is selectively expressed in the tumor stroma and is absent in normal tissues (see normal colon vs. colon cancer). In contrast, expression of α-SMA can be seen in subsets of fibroblasts surrounding the crypts and in the muscularis mucosa of the normal colon as well as in activated tumor fibroblasts. FAPα and α-SMA expression ( brown) visualized by the ABC immunoperoxidase method with hematoxylin counterstaining. b Subcutaneous xenograft models in immunodeficient mice. Like in the majority of xenograft models, subcutaneous injection of Colo205 cells, a human colorectal cancer cell line, induces tumors with a highly atypical morphology, characterized by clusters of tumor cells, with very little tumor stroma, and absence of glandular differentiation (for comparison see the human colon cancer sample above). Certain human tumor cells such as FaDu cells, derived from a head and neck squamous cell carcinoma, are able to induce a more prominent stroma reaction, with FAPα positive activated stromal fibroblasts and histotypic features resembling the human counterpart. Therefore, a careful selection of the in vivo models is required to determine the efficacy of drugs targeting the activated tumor fibroblasts. Hemaoxylin-eosin and immunohistochemical staining for FAPα ( brown; bottom panel)
20 Cancer Associated Fibroblasts as Therapeutic Targets
Normal colon
Human cancers Pancreas ca.
α-SMA
FAP
a
Breast ca.
FAP
Colon ca.
FAP
HE
b
Colo205
Mouse models
FaDu
385
386
C. Rupp et al.
20.3 Cancer Associated Fibroblasts (CAFs): Molecular Characterization Fibroblasts represent the major cellular component of the stroma of epithelial cancers. Several in vitro and in vivo studies have demonstrated that the growth, differentiation and invasive behaviour of malignant epithelial cells are influenced by the surrounding stroma (Tlsty and Hein 2001; Bhowmick et al. 2004; Joyce 2005; Mueller and Fusenig 2004). Normal fibroblasts have been reported to prevent the progression of transformed epithelial cells (Hayashi and Cunha 1991), in contrast, the presence of activated tumor stromal fibroblasts was shown to enhance malignant epithelial transformation in several cancer models (Nakamura et al. 1997; Olumi et al. 1999; Orimo et al. 2005). Cancer associated fibroblasts (CAFs) differ considerably from normal resting fibroblasts, and display distinct molecular signatures which can be linked to clinical outcome (Finak et al. 2008). Recent studies have identified new functional roles for CAFs and the existence of different CAFs subsets in human cancers by gene expression analysis (Chang et al. 2002; Iacobuzio-Donahue et al. 2002).
20.4 In Vitro and In Vivo Models for Tumor Stroma Interaction Understanding the molecular mechanisms that control the heterotypic interactions between malignant cells and the surrounding stroma may help to develop new targeted therapies. However these studies have been hampered by the challenges in studying multi-cellular interactions in experimental models (Fig. 20.2). The in vitro study of freshly dissociated cancer cells or established tumor cell lines and fibroblasts in two dimensional (2D) cultures has provided important insights into basic tumor cell biology and has enabled the identification of common genetic alterations in cancer cells that can be targeted therapeutically (Cornil et al. 1991; Yashiro et al. 2005; Elenbaas et al. 2001; Jones et al. 2008). However, such in vitro approaches have proven somewhat limited in studying stromal targets. Only a limited number of stromal-derived cells are available in culture and phenotypic changes can be induced under culture conditions (Orimo et al. 2005). In addition, many physiological aspects of tumors such as cell–cell and cell–matrix interactions are lost under conventional 2D culture conditions. Cells grown in three-dimensional (3D) scaffolds or as 3D multicellular spheroids recapitulate the architecture of tissues and tumors in vivo to a higher extent. They offer new opportunities to analyze the activation of differentiation programs and the pathways involved in cell migration and invasion when cells are grown in a heterotypic and physiologically relevant context (Kunz-Schughart and Knuechel 2002; Schmeichel and Bissell 2003). Organotypic 3D co-culture models have been
20 Cancer Associated Fibroblasts as Therapeutic Targets
387
Fig. 20.2 Carcinomas are heterogeneous mixtures of malignant cells and stromal cells, such as fibroblasts, blood vessels and immune cells embedded in a matrix of extracellular proteins (ECM) and grow as three dimensional structures. An example for this heterogeneity is shown in a histologic picture from a human non-small cell lung cancer ( left panel). Tumor cell clusters ( blue) are separated by strands of activated fibroblasts which are PDGFR-β positive ( brown) and blood vessels with CD31 positive endothelial cells ( dark blue). In conventional cell culture models, cells are grown as homogeneous cultures on plastic surfaces. Under these culture conditions many features of the in vivo growth, such as tissue architecture, cell–cell contact, heterotypic cellular interactions and signalling networks are lost ( middle panel). Using 3D cultures in the presence of ECM components, the heterotypic interactions of tumor cells and stromal cells can be studied in detail. A phase contrast picture of this model shows a culture of tumor cells grown as a multicellular spheroid, co-cultured with fibroblast embedded in a collagen I gel, recapitulating the in vivo heterogeneity
used to study the functional interplay between genetically altered epithelial cells and fibroblasts (Okawa et al. 2007; Sadlonova et al. 2005) and to study fibroblastled invasion in models of skin, breast, pancreatic and brain cancers (Gaggioli et al. 2007; Froeling et al. 2009). A novel experimental set-up has been developed (Dolznig et al. manuscript in preparation) combining multi-cellular spheroids, 3D collagen gel cultures and cocultures of human epithelial cancer cells with normal human fibroblasts or CAFs in one assay (Fig. 20.2). Using this model system the feasibility to study the tumor– stroma interactions phenotypically and at the molecular level was demonstrated. Gene expression profiles from these 3D co-cultures have been obtained and ongoing studies are exploring the applicability of the model to study the role of these new stromal targets in tumor invasion using knock-in/knock-down experiments of selected genes in the tumor cells or in the fibroblast population.
388
C. Rupp et al.
Xenograft tumor models are commonly used to analyze new mechanisms of action and to validate the efficacy of novel drugs in preclinical studies. The majority of these in vivo assays are performed in immunodeficient mice following the inoculation of established tissue culture cell lines into ectopic sites. From a histopathological view these tumor models show a highly atypical morphology with very little stroma or histotypic features resembling the human cancer counterparts (Fig. 20.1b). A more authentic histological appearance is observed in carcinoma models derived by direct transplantation of surgical specimens, purified cell suspensions freshly obtained from surgical specimens or in orthotopically implanted tumors (Rubio-Viqueira et al. 2006; Shu et al. 2008; O’Brien et al. 2007; Ostermann et al. 2008). Nevertheless, most of the preclinical validation studies are carried out using the ectopic (mostly subcutaneous) in vivo models, that are relatively easy to set up and that can be generated in large numbers of similar-sized tumors for randomization as pre-requisite to assess the effects of drugs. However, in many cases, the results obtained from xenograft models do not translate well in subsequent clinical studies (Ostermann et al. 2008). Genetically engineered mouse models (GEM) are promising alternatives (Frese and Tuveson 2007; Gopinathan and Tuveson 2008). These models, generated through the introduction of genetic mutations associated with specific human malignancies closely recapitulate the human disease at the pathophysiological and molecular level. To date, models have been developed for many common tumor types (e.g. lung, breast, prostate, colon and pancreatic cancer). Evidence for the usefulness of GEM has been demonstrated in preclinical studies evaluating targeted therapies in models of lung and breast cancer (Bhowmick et al. 2004; Politi et al. 2006; Beppu et al. 2008; Rottenberg and Jonkers 2008; Perera et al. 2009; Santos et al. 2009). These studies suggest that GEM can more accurately predict the therapeutic responses to those observed in the clinic.
20.5 Therapeutic Opportunities Different molecular targets have been shown to distinguish the cancer associated fibroblasts (CAFs) and different strategies to target these molecules are under evaluation. Here we will focus on potential drug candidates with special attention to those in more advanced clinical development.
20.5.1 Fibroblast Activation Protein Alpha Fibroblast activation protein alpha (FAPα) is an integral cell surface protein selectively expressed by activated stromal fibroblasts of several types of human epithelial cancers. In normal tissues, FAPα expression is highly restricted to developing organs, healing wounds, and tissue remodeling. Epithelial tumor cells and most normal adult human tissues lack FAPα expression (Garin-Chesa et al. 1990; Rettig
20 Cancer Associated Fibroblasts as Therapeutic Targets
389
et al. 1993; Niedermeyer et al. 1997; Niedermeyer et al. 2001; Huber et al. 2003; Dolznig et al. 2005). FAPα is a serine protease capable of degrading type I collagen which places FAPα into the group of enzymes involved in tumor tissue remodeling (Scanlan et al. 1994; Park et al. 1999; Niedermeyer et al. 1998). FAPα activity can be detected in tumor samples and shows a good correlation with FAPα expression detected by immunohistochemistry (Huber et al. 2003; Park et al. 1999). Based on the selective expression of FAPα in the reactive stroma of many epithelial cancers, the lack of expression in normal adult tissues, and its protease activity, FAPα is an ideally suited stroma target to be exploited in the clinic. Two different approaches have been used to target FAPα in tumors. The first was to employ FAPα-specific monoclonal antibodies. Initial studies with radiolabeled murine and humanized antibodies against human FAPα have shown highly specific tumor targeting properties (Welt et al. 1994; Scott et al 2003), however no clinical efficacy could be demonstrated using the unlabeled humanized antibody in a study in metastatic colorectal cancer (Hofheinz et al. 2003), probably due to the lack of effector-function properties of the naked antibody. More recently, a novel antibody-maytansinoid conjugate (FAP5-DM1), targeting a shared epitope of human, mouse and cynomolgus monkey fibroblast activation protein alpha, has been developed. Using this conjugate in stroma-rich histotypic cancer xenograft models we were able to induce long-lasting inhibition of tumor growth and complete regressions in models of lung, pancreas and head and neck cancers, with no evidence of toxicity (Ostermann et al. 2008). The second approach has been to target the enzymatic activity of FAPα with small molecule inhibitors. Using the peptidase inhibitor PT-100 (talabostat) a reduction in tumor growth rate was shown in a variety of tumor models in mice (Cheng et al. 2005; Adams et al. 2004). This particular compound, however, inhibits multiple intracellular and extracellular dipeptidyl peptidases (e.g. FAPα, DPPIV, DPP7), so that the anti-tumor effect could not be directly attributed to FAPα inhibition. More recently, using FAPα-null mice and a more selective inhibitor (PT-630), the endogenous role of FAPα in tumorigenesis and the control of tumor growth mediated by pharmacologic inhibition of FAPα enzymatic activity has been reported (Santos et al. 2009; Pure 2009). Deletion of FAPα resulted in a marked reduction of tumor growth in a LSL-K-rasG12D genetic mouse model of lung cancer and in a syngeneic colon cancer model, suggesting a tumor promoting activity of endogenous FAPα. Treatment with PT-630 of tumorbearing wild type animals resulted in a marked inhibition of tumor growth in both models, supporting further clinical studies.
20.5.2 Matrix Metalloproteinases (MMPs) Cancer associated fibroblasts are a major source of MMPs in tumors, including MMPs 1,2,3,9,11,13 and MT1-MMP (Bisson et al. 2003; Sternlicht et al. 1999). Fibroblast derived MMPs have been extensively investigated in xenograft models, demonstrating the important role for these proteases in promoting tumor growth, metastasis, and angiogenesis (Egeblad and Werb 2002). Based on the results of
390
C. Rupp et al.
the preclinical studies MMPs were seen as attractive anticancer targets and several inhibitors have been developed and tested in the clinic in a variety of cancer types (Coussens et al. 2002). These trials however had failed to demonstrate a survival benefit despite the promising activity shown in pre-clinical models (King et al. 2003). Possible explanations include differences in the biology of the MMPs between mice and humans, lack of anti-immune response in the xenograft models used pre-clinically, dose-limiting toxicities at least in part due to off-target effects, a narrow therapeutic window for some of the inhibitors and perhaps the challenging fact that MMPs can have tumor-promoting as well as tumor-suppressor activities. Thus a better understanding of the functional complexity of this family of proteases and the use of second generation inhibitors with improved selectivity profile may provide better therapeutic outcomes (Konstantinopoulos et al. 2008).
20.5.3 Endosialin/TEM1 Endosialin is a highly sialylated, C-type lectin-like surface receptor structurally related to thrombomodulin and complement receptor C1qRp (Rettig et al. 1992; Christian et al. 2001). First identified with a monoclonal antibody, mAb FB5, endosialin was discovered independently through a SAGE screen (serial analysis of gene expression) of human cancer endothelial cells, leading to the alternative designation of tumor endothelial marker 1 (TEM1) (St Croix et al. 2000). Endosialin/TEM1 is expressed to varying degrees by tumor endothelial cells, pericytes and stromal fibroblasts (MacFadyen et al. 2005; Brady et al. 2004; Rupp et al. 2006a). Endosialin expression has not been detected in capillary endothelium in most normal tissues. The physiological role of endosialin is unknown. Endosialin/Tem1 knock-out mice are fertile and develop normally, however, when human HCT116 colon carcinoma cells were implanted orthotopically onto the serosal surface of the large intestine of nude KO mice, both tumor take and growth rate were reduced (Nanda et al. 2006). Recent evidence suggests that endosialin/TEM1 might interact with extracellular matrix components, including collagen type I, IV and fibronectin in promoting cell adhesion and migration processes during tumor invasion and metastasis (Tomkowicz et al. 2007). A humanized Endosialin/TEM1 blocking antibody (MORAb-004) is currently in clinical studies and might provide a therapeutic benefit in a broad range of tumors, based on its ability to target the endothelial cells as well as the peri-vascular stromal component of the tumors.
20.5.4 PDGF/PDGFR Pathway Platelet-derived growth factors (PDGFs) play important roles during embryonic development and wound healing (Betsholtz 2004) and expression of their tyrosine kinase receptors (PDGFRs) in the tumor stroma is a common feature of human
20 Cancer Associated Fibroblasts as Therapeutic Targets
391
cancers (Ostman and Heldin 2007). During tumorigenesis, PDGFR can drive tumor growth directly by autocrine stimulation of receptor-expressing tumor cells or in a paracrine manner by acting on the tumor stroma fibroblasts and pericytes (Pietras et al. 2008). The importance of the paracrine signaling network for the recruitment of cancer associated fibroblasts and pericytes has been shown in a number of studies (Skobe and Fusenig 1998; Anderberg et al. 2009). Pericytes provide both survival signals and structural support to endothelial cells contributing to a mature, functional vasculature and thus to tumor growth (Carmeliet 2003). Multiple tyrosine kinase inhibitors with anti-PDGFR activity, such as imatinib, sorafenib, and sunitinib, have been approved and are presently under further clinical development (Levitzki 2004; Steeghs et al. 2007). The most commonly used, imatinib, is a bcr-abl inhibitor with additional PDGFR and c-kit kinase inhibitory activity (Carroll et al. 1997). In experimental cancer models, imatinib has been shown to inhibit PDGFR activity on fibroblasts and pericytes and to significantly decrease the stromal reaction which was accompanied by a reduction in tumor cell proliferation and pericyte coverage of tumor vessels (Pietras et al. 2008; Kitadai et al. 2006). Furthermore, inhibition of PDGFRs increases the uptake and therefore the antitumor effect of conventional chemotherapeutics like paclitaxel by lowering tumor interstitial pressure (Pietras et al. 2002). Other multi-kinase inhibitors, such as BIBF1120, a triple angiokinase inhibitor of the VEGFR, PDGFR and FGFR families, has been shown to decrease the pericyte coverage of tumor blood vessels in experimental cancer models (Hilberg et al. 2008) which together with the reduction in tumor microvessel density contributed to the pronounced anti-tumor effects of the inhibitor. BIBF1120 is in clinical development for several tumor indications. Collectively, these results indicate that inhibition of PDGF receptor signaling might provide a complementary approach to conventional treatments. To date, it is still unknown to what extent selective blockage of stromal PDGF signalling contributes to the observed anti-tumor effects of these multi-kinase inhibitors.
20.5.5 Transforming Growth Factor β (TGF-β) Pathway Transforming growth factor β (TFG-β) is recognized for its dual and opposing functions, a tumor-suppressor activity in the pre-malignant state and a tumor promoter activity during malignant progression (Bierie and Moses 2006; Massague 2008). This dual role has made the design and development of drugs targeting this signaling pathway in cancer particularly complex. In tumors, activation of TGF-β is linked to the activity of several oncogenic pathways linked to the induction of epithelial mesenchymal transition (EMT) that enhances tumor cell invasion (Oft et al. 1996). TGF-β can have an additional role in tumor growth that is mediated through its activity on the tumor stroma, facilitating tumor tissue remodeling and neoangiogenesis. Studies with fibroblast specific TGF-β type II receptor knock-out models provided evidence for the tumor suppressor role of TGF-β in fibroblasts,
392
C. Rupp et al.
by blocking the production of tumor cell growth-promoting paracrine factors such as hepatocyte growth factor (HGF) (Bhowmick et al. 2004; Cheng et al. 2007). On the other hand, it was demonstrated that TGF-β stimulates myofibroblast differentiation and that blocking of TGF-β signaling in stromal fibroblast leads to a significant reduction of tumor-growth in a co-transplantation xenograft model (Verona et al. 2007), suggesting that pro- or anti-tumoral effects of TGF-β signaling may very much depend on individual tumor models. Considering the direct effect of TGF-β on tumor cells and its indirect effect on the tumor stroma, TGF-β signaling appeared as an attractive therapeutic concept. Several approaches to inhibit the TGF-β pathway have been investigated in preclinical models and clinical studies. Neutralizing antibodies that inhibit the ligand-receptor interaction, antisense oligonucleotides and small molecule inhibitors of the TGF-β receptor kinase complex have been developed and are at different stages of clinical development (Lahn et al. 2005; Jones et al. 2009). It is expected that this class of agents will be active in a broad range of tumors but due to the complex roles of this growth factor receptor family in tumorigenesis a careful selection of patients will be required to address their therapeutic benefits in patients.
20.5.6 Hedgehog Pathway The Hedgehog (Hh) family of proteins have been shown to control cell growth, survival and fate during embryonic development and when mutated or misregulated to contribute to tumorigenesis. Aberrant activation of the Hh pathway by mutations are causally associated with basal cell carcinoma of the skin, medulloblastoma and rhabdomyosarcoma (Varjosalo and Taipale 2008). Furthermore, components of the Hh pathway have been described to play a role in the growth of a variety of epithelial cancer types, including small cell lung cancer, pancreatic and prostate cancer even in the absence of mutations (Watkins et al. 2003; Thayer et al. 2003; Karhadkar et al. 2004). Recent studies in experimental cancer models support a model in which Hh acts in a paracrine manner on stromal cells. Hh increases tumor growth by stimulating the expression of extracellular matrix proteins and factors like IGF or Wnt in the stroma and thereby promoting stromal desmoplasia (Yauch et al. 2008). The most commonly used Hh antagonists are the plant alkaloid cyclopamine and its derivatives (Taipale et al. 2000). The anti-tumor effect of the semisynthetic cyclopamine-derivative IPI-926 was investigated in a mouse model of pancreatic ductal adenocarcinoma refractory to gemcitabine (a drug commonly used in the clinic). Mice treated with IPI-926 alone or in combination with gemcitabine were depleted of desmoplastic stroma reaction in the tumors and displayed increased intratumoral vascular density. These changes correlated with a more effective delivery of the co-administrated gemcitabine, resulting in enhanced efficacy of the drug (Olive et al. 2009). This study has identified a potential novel mechanism for anti-stroma therapy.
20 Cancer Associated Fibroblasts as Therapeutic Targets
393
20.6 Identification of Novel Therapeutic Targets by Gene Expression Profiling The recent technological advances for high-throughput DNA and RNA detection have shown that specific germline and somatic mutations, loss of heterozygosity, and DNA amplifications occur during cancer progression. Oncogenome signatures of human tumors have been shown to correlate with metastatic behaviour and clinical outcome in different cancer types (Alizadeh et al. 2000; Perou et al. 2000; Ramaswamy et al. 2003). However, the specific contribution of malignant epithelial cells and stromal cells to these genetic signatures is in most cases unclear, since most of the studies have used bulk tumor samples. Our approach to establish the molecular differences between CAFs and normal resting fibroblasts has been to generate gene expression signatures from microdissected cancer and corresponding normal tissues. We focussed on colorectal cancer and developed a protocol for laser capture microdissection guided by antibodies against FAP to separate epithelial cells from activated stromal fibroblasts (Rupp et al. 2006b). We performed whole genome Affymetrix GeneChip® analysis and obtained transcriptional signatures from tumor cells and activated tumor stroma that were compared with the expression profiles from microdissected normal colonic epithelium and normal fibroblasts, obtained from the same patients (Fig. 20.3, Rupp et al manuscript in preparation). Bioinformatic analysis comparing the tumor stroma vs. the normal stroma signatures identified a number of selectively up-regulated genes. Well characterized tumor stroma markers such as FAPα, MMP-2, PDGFR-β and FGFR1 among others appeared specifically up-regulated in the stroma compartment (Fig. 20.3) and served as a validation parameter for our screen. To further analyze the functional significance of these gene signatures in the context of tumorigenesis we performed a similar genetic screen in our above described 3D co-culture model of tumor cell spheroids and fibroblasts (normal and cancer-derived) grown in collagen gels. We established transcriptional profiles from the different cellular components grown in collagen gels in mono-cultures and compared the gene expression responses induced upon co-cultivation (Dolznig et al. manuscript in preparation). We observed a remarkable concordance between the gene sets obtained in our ex vivo study (colorectal cancer study from human samples) and this in vitro co-culture system. Examples of commonly regulated genes included COL11A1 and MMP3 (Fig. 20.3), well characterized markers of activated fibroblasts. Gene-Set Enrichment Analysis (GSEA) (Mootha et al. 2003) using the gene-set collections from the Molecular Signature Database (Broad Institute) (Subramanian et al. 2005) and Pathway analysis (Ingenuity®) revealed datasets and gene-networks that were significantly enriched in both screens. Gene-sets involved in extracellular matrix deposition, angiogenesis, wound healing and EMT were significantly up-regulated in both studies. Interestingly, many of the genes identified in our study have been reported in studies performed in vitro including the “wound response signature” of fibroblasts in response to serum stimulation (Chang et al. 2004), a hypoxia-associated response (Chi et al. 2006) as well as the signatures obtained from co-cultures of cancer cells
394
C. Rupp et al.
and fibroblasts cell lines of different origins (Sato et al. 2004; Gallagher et al. 2005). Using independent datasets from human cancers, it was shown that the “woundresponse signature” was strongly predictive of metastasis and progression in breast, lung and gastric cancers and was an independent predictor of outcome in a followup study in breast cancer (Chang et al. 2004). Other in vivo signatures have been described (West et al. 2005) comparing the expression patterns of good versus poor outcome in fibroblastic tumors. A subsequent comparison of these signatures with a breast cancer data set suggested that distinct patterns of stroma reaction defined two groups of breast cancers with significant differences in overall survival, indicating that the stromal response varies significantly in different subtypes of carcinomas and may be clinically relevant. Expression signatures from different tumor compartments have also been established using serial analysis of gene expression on antibody-sorted stromal components in breast cancers (Allinen et al. 2004) or laser capture microdissection in breast cancer and basal cell carcinoma of the skin (Casey et al. 2009; Micke et al. 2007). Using a set of genes expressed by the microdissected tumor stroma, a stroma-derived prognostic predictor signature (SDPP) was developed and shown to separate primary breast cancers into three distinct groups associated with different clinical outcomes (Finak et al. 2008). In another study, a stromal signature was shown to predict the response of estrogen-receptor negative breast tumors to chemotherapy (Farmer et al. 2009). The authors used a novel bioinformatics method that decomposes gene expression signals from a mixture of tumor and stromal cells into multiple independent signatures. They obtained a 50-gene stromal signature including FAPα, MMP2, MMP14, PDGFR-β which predicted resistance to chemotherapy. Fig. 20.3 Identification of novel tumor stroma markers by expression profiling analysis. a Antibody-guided laser capture microdissection allows the separation of epithelial cells from the activated stromal compartment in colon cancer samples. Activated tumor stromal fibroblasts were visualized by immunohistochemical staining with an antibody to FAPα. In the figure, the borders between epithelial and stromal structures are indicated by red dotted lines. Normal fibroblasts were isolated from normal colonic tissue after hematoxylin staining and morphological examination. After RNA isolation, whole genome Affymetrix GeneChip analysis was performed. Bioinformatic evaluation identified novel tumor stroma targets by comparing the tumor stroma vs. the normal stroma signatures. b Well characterized tumor stroma markers such as FAPα, MMP2, PDGFRß and FGFR1 were significantly up-regulated in the tumor stroma compartment. The expression levels are indicated by whisker box plots, the bold centre-line indicates the median; the box represents the interquartile range (IQR). Whiskers extend to 1.5 times the IQR. TC, tumor cells; NS, normal stroma; TS, tumor stroma. c Comparison of the transcriptional profiles obtained in our ex vivo screen in colorectal cancer samples with those obtained in an in vitro screen with a colon cancer cell line (LS174T) cultured in the presence colon-derived human CAFs in a 3D coculture assay. Gene-Set Enrichment Analysis (GSEA) revealed gene sets involved in extracellular matrix deposition, angiogenesis and wound healing significantly upregulated in both studies.Two representative examples, Collagen 11A1 (COL11A1) and matrix metalloprotease 3 (MMP3) are shown. TC, tumor cells; NS, normal stroma; TS, tumor stroma; blue whisker box blots indicate the expression levels after 3.5 days of LS174T/CAF co-cultivation (TC/CoCult); yellow box blots show the levels of expression of individually cultured LS174T cells and CAFs mixed together after cultivation (TC/CAF Mix). Whisker box plot as in b
20 Cancer Associated Fibroblasts as Therapeutic Targets
395
tumour stroma
tumour cells
LCM
normal stroma RNA isolation LCM
GeneChip analysis Bioinformatics New targets
a MMP 2 °
800
PDGFRß
2500
1000
° 800
2000
800
1500
600
600
1000
400
400
600
400
200
°
500
200
200
0 TC
NS
0
TS
TC
NS
0
TS
TC
NS
COL11A1
TS
TC
MMP3 in vitro
ex vivo
in vitro
ex vivo 1500 6000
4000
200 1000
2000 500
100
°
3000
ix
oC
M
AF
C
AF
oC
/C
/C
t
TC NS TS
C
AF
AF
/C
/C TC
TC
ul
ix
TS
M
TC NS
ul t
0
0
TC
mRNA expression level (a.u.)
°
° 0
b
c
FGFR1 1000
TC
Expression level (arbitrary units)
FAP
NS
TS
396
C. Rupp et al.
Taken together, these studies demonstrated that tumors express a variety of functionally different genes in their tumor stroma, representing different activation stages or different subtypes of CAFs that may be relevant for the invasiveness or clinical behavior of the tumors. The gene expression signatures derived from this type of analysis appear to have clinical significance in different cancer types and have provided new genetic markers in the tumor stroma that may serve as targets for novel therapeutic approaches.
20.7 Conclusions The rapid progress of research in molecular cancer biology has contributed to a better understanding of the role of the tumor stroma during tumor growth and metastasis formation and has led to the identification of selected tumor stroma markers that serve as targets for novel therapies. A number of monoclonal antibodies, small-molecule inhibitors and anti-sense approaches have been developed and investigated in pre-clinical models, some of these molecules have entered clinical development. Future approaches to stroma-targeted therapy will have to be based on further refinement of our understanding of the molecular mechanisms that control the tumor–stroma interaction, improved preclinical models that adequately reproduce the complexity of the tumor tissue, and a biomarker-based selection of patients most likely to benefit from the novel therapies.
References Adams S, Miller GT, Jesson MI, Watanabe T, Jones B, Wallner BP (2004) PT-100, a small molecule dipeptidyl peptidase inhibitor, has potent antitumor effects and augments antibody-mediated cytotoxicity via a novel immune mechanism. Cancer Res 64:5471–5480 Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, Rosenwald A, Boldrick JC, Sabet H, Tran T, Yu X et al (2000) Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403:503–511 Allinen M, Beroukhim R, Cai L, Brennan C, Lahti-Domenici J, Huang H, Porter D, Hu M, Chin L, Richardson A et al (2004) Molecular characterization of the tumor microenvironment in breast cancer. Cancer Cell 6:17–32 Anderberg C, Li H, Fredriksson L, Andrae J, Betsholtz C, Li X, Eriksson U, Pietras K (2009) Paracrine signaling by platelet-derived growth factor-CC promotes tumor growth by recruitment of cancer-associated fibroblasts. Cancer Res 69:369–378 Beppu H, Mwizerwa ON, Beppu Y, Dattwyler MP, Lauwers GY, Bloch KD, Goldstein AM (2008) Stromal inactivation of BMPRII leads to colorectal epithelial overgrowth and polyp formation. Oncogene 27:1063–1070 Betsholtz C (2004) Insight into the physiological functions of PDGF through genetic studies in mice. Cytokine Growth Factor Rev 15:215–228 Bhowmick NA, Chytil A, Plieth D, Gorska AE, Dumont N, Shappell S, Washington MK, Neilson EG, Moses HL (2004) TGF-beta signaling in fibroblasts modulates the oncogenic potential of adjacent epithelia. Science 303:848–851
20 Cancer Associated Fibroblasts as Therapeutic Targets
397
Bierie B, Moses HL (2006) Tumour microenvironment: TGFbeta: the molecular Jekyll and Hyde of cancer. Nat Rev Cancer 6:506–520 Bisson C, Blacher S, Polette M, Blanc JF, Kebers F, Desreux J, Tetu B, Rosenbaum J, Foidart JM, Birembaut P et al (2003) Restricted expression of membrane type 1-matrix metalloproteinase by myofibroblasts adjacent to human breast cancer cells. Int J Cancer 105:7–13 Brady J, Neal J, Sadakar N, Gasque P (2004) Human endosialin (tumor endothelial marker 1) is abundantly expressed in highly malignant and invasive brain tumors. J Neuropathol Exp Neurol 63:1274–1283 Carmeliet P (2003) Angiogenesis in health and disease. Nat Med 9:653–660 Carroll M, Ohno-Jones S, Tamura S, Buchdunger E, Zimmermann J, Lydon NB, Gilliland DG, Druker BJ (1997) CGP 57148, a tyrosine kinase inhibitor, inhibits the growth of cells expressing BCR-ABL, TEL-ABL, and TEL-PDGFR fusion proteins. Blood 90:4947–4952 Casey T, Bond J, Tighe S, Hunter T, Lintault L, Patel O, Eneman J, Crocker A, White J, Tessitore J et al (2009) Molecular signatures suggest a major role for stromal cells in development of invasive breast cancer. Breast Cancer Res Treat 114:47–62 Chang HY, Chi JT, Dudoit S, Bondre C, van de Rijn M, Botstein D, Brown PO (2002) Diversity, topographic differentiation, and positional memory in human fibroblasts. Proc Natl Acad Sci U S A 99:12877–12882 Chang HY, Sneddon JB, Alizadeh AA, Sood R, West RB, Montgomery K, Chi JT, van de Rijn M, Botstein D, Brown PO (2004) Gene expression signature of fibroblast serum response predicts human cancer progression: similarities between tumors and wounds. PLoS Biol 2:E7 Cheng JD, Valianou M, Canutescu AA, Jaffe EK, Lee HO, Wang H, Lai JH, Bachovchin WW, Weiner LM (2005) Abrogation of fibroblast activation protein enzymatic activity attenuates tumor growth. Mol Cancer Ther 4:351–360 Cheng N, Chytil A, Shyr Y, Joly A, Moses HL (2007) Enhanced hepatocyte growth factor signaling by type II transforming growth factor-beta receptor knockout fibroblasts promotes mammary tumorigenesis. Cancer Res 67:4869–4877 Chi JT, Wang Z, Nuyten DS, Rodriguez EH, Schaner ME, Salim A, Wang Y, Kristensen GB, Helland A, Borresen-Dale AL et al (2006) Gene expression programs in response to hypoxia: cell type specificity and prognostic significance in human cancers. PLoS Med 3:e47 Christian S, Ahorn H, Koehler A, Eisenhaber F, Rodi HP, Garin-Chesa P, Park JE, Rettig WJ, Lenter MC (2001) Molecular cloning and characterization of endosialin, a C-type lectin-like cell surface receptor of tumor endothelium. J Biol Chem 276:7408–7414 Cornil I, Theodorescu D, Man S, Herlyn M, Jambrosic J, Kerbel RS (1991) Fibroblast cell interactions with human melanoma cells affect tumor cell growth as a function of tumor progression. Proc Natl Acad Sci U S A 88:6028–6032 Coussens LM, Fingleton B, Matrisian LM (2002) Matrix metalloproteinase inhibitors and cancer: trials and tribulations. Science 295:2387–2392 Desmouliere A, Guyot C, Gabbiani G (2004) The stroma reaction myofibroblast: a key player in the control of tumor cell behavior. Int J Dev Biol 48:509–517 Dolznig H, Schweifer N, Puri C, Kraut N, Rettig WJ, Kerjaschki D, Garin-Chesa P (2005) Characterization of cancer stroma markers: in silico analysis of an mRNA expression database for fibroblast activation protein and endosialin. Cancer Immunol 5:10 Egeblad M, Werb Z (2002) New functions for the matrix metalloproteinases in cancer progression. Nat Rev Cancer 2:161–174 Elenbaas B, Spirio L, Koerner F, Fleming MD, Zimonjic DB, Donaher JL, Popescu NC, Hahn WC, Weinberg RA (2001) Human breast cancer cells generated by oncogenic transformation of primary mammary epithelial cells. Genes Dev 15:50–65 Farmer P, Bonnefoi H, Anderle P, Cameron D, Wirapati P, Becette V, Andre S, Piccart M, Campone M, Brain E et al (2009) A stroma-related gene signature predicts resistance to neoadjuvant chemotherapy in breast cancer. Nat Med 15:68–74 Finak G, Bertos N, Pepin F, Sadekova S, Souleimanova M, Zhao H, Chen H, Omeroglu G, Meterissian S, Omeroglu A et al (2008) Stromal gene expression predicts clinical outcome in breast cancer. Nat Med 14:518–527
398
C. Rupp et al.
Frese KK, Tuveson DA (2007) Maximizing mouse cancer models. Nat Rev Cancer 7:645–658 Froeling FE, Mirza TA, Feakins RM, Seedhar A, Elia G, Hart IR, Kocher HM (2009) Organotypic culture model of pancreatic cancer demonstrates that stromal cells modulate E-cadherin, betacatenin, and Ezrin expression in tumor cells. Am J Pathol 175:636–648 Gaggioli C, Hooper S, Hidalgo-Carcedo C, Grosse R, Marshall JF, Harrington K, Sahai E (2007) Fibroblast-led collective invasion of carcinoma cells with differing roles for RhoGTPases in leading and following cells. Nat Cell Biol 9:1392–1400 Gallagher PG, Bao Y, Prorock A, Zigrino P, Nischt R, Politi V, Mauch C, Dragulev B, Fox JW (2005) Gene expression profiling reveals cross-talk between melanoma and fibroblasts: implications for host–tumor interactions in metastasis. Cancer Res 65:4134–4146 Garin-Chesa P, Old LJ, Rettig WJ (1990) Cell surface glycoprotein of reactive stromal fibroblasts as a potential antibody target in human epithelial cancers. Proc Natl Acad Sci U S A 87:7235–7239 Gopinathan A, Tuveson DA (2008) The use of GEM models for experimental cancer therapeutics. Dis Model Mech 1:83–86 Hasebe T, Sasaki S, Imoto S, Ochiai A (2001) Highly proliferative fibroblasts forming fibrotic focus govern metastasis of invasive ductal carcinoma of the breast. Mod Pathol 14:325–337 Hayashi N, Cunha GR (1991) Mesenchyme-induced changes in the neoplastic characteristics of the Dunning prostatic adenocarcinoma. Cancer Res 51:4924–4930 Hilberg F, Roth GJ, Krssak M, Kautschitsch S, Sommergruber W, Tontsch-Grunt U, Garin-Chesa P, Bader G, Zoephel A, Quant J et al (2008) BIBF 1120: triple angiokinase inhibitor with sustained receptor blockade and good antitumor efficacy. Cancer Res 68:4774–4782 Hofheinz RD, al-Batran SE, Hartmann F, Hartung G, Jager D, Renner C, Tanswell P, Kunz U, Amelsberg A, Kuthan H et al (2003) Stromal antigen targeting by a humanised monoclonal antibody: an early phase II trial of sibrotuzumab in patients with metastatic colorectal cancer. Onkologie 26:44–48 Huber MA, Kraut N, Park JE, Schubert RD, Rettig WJ, Peter RU, Garin-Chesa P (2003) Fibroblast activation protein: differential expression and serine protease activity in reactive stromal fibroblasts of melanocytic skin tumors. J Invest Dermatol 120:182–188 Huber MA, Kraut N, Schweifer N, Dolznig H, Peter RU, Schubert RD, Scharffetter-Kochanek K, Pehamberger H, Garin-Chesa P (2006) Expression of stromal cell markers in distinct compartments of human skin cancers. J Cutan Pathol 33:145–155 Iacobuzio-Donahue CA, Argani P, Hempen PM, Jones J, Kern SE (2002) The desmoplastic response to infiltrating breast carcinoma: gene expression at the site of primary invasion and implications for comparisons between tumor types. Cancer Res 62:5351–5357 Jones E, Pu H, Kyprianou N (2009) Targeting TGF-beta in prostate cancer: therapeutic possibilities during tumor progression. Expert Opin Ther Targets 13:227–234 Jones S, Zhang X, Parsons DW, Lin JC, Leary RJ, Angenendt P, Mankoo P, Carter H, Kamiyama H, Jimeno A et al (2008) Core signaling pathways in human pancreatic cancers revealed by global genomic analyses. Science 321:1801–1806 Joyce JA (2005) Therapeutic targeting of the tumor microenvironment. Cancer Cell 7:513–520 Karhadkar SS, Bova GS, Abdallah N, Dhara S, Gardner D, Maitra A, Isaacs JT, Berman DM, Beachy PA (2004) Hedgehog signalling in prostate regeneration, neoplasia and metastasis. Nature 431:707–712 King J, Zhao J, Clingan P, Morris D (2003) Randomised double blind placebo control study of adjuvant treatment with the metalloproteinase inhibitor, marimastat in patients with inoperable colorectal hepatic metastases: significant survival advantage in patients with musculoskeletal side-effects. Anticancer Res 23:639–645 Kitadai Y, Sasaki T, Kuwai T, Nakamura T, Bucana CD, Fidler IJ (2006) Targeting the expression of platelet-derived growth factor receptor by reactive stroma inhibits growth and metastasis of human colon carcinoma. Am J Pathol 169:2054–2065 Konstantinopoulos PA, Karamouzis MV, Papatsoris AG, Papavassiliou AG (2008) Matrix metalloproteinase inhibitors as anticancer agents. Int J Biochem Cell Biol 40:1156–1168 Koperek O, Scheuba C, Puri C, Birner P, Haslinger C, Rettig W, Niederle B, Kaserer K, Garin Chesa P (2007) Molecular characterization of the desmoplastic tumor stroma in medullary thyroid carcinoma. Int J Oncol 31:59–67
20 Cancer Associated Fibroblasts as Therapeutic Targets
399
Kunz-Schughart LA, Knuechel R (2002) Tumor-associated fibroblasts (part I): active stromal participants in tumor development and progression? Histol Histopathol 17:599–621 Lahn M, Kloeker S, Berry BS (2005) TGF-beta inhibitors for the treatment of cancer. Expert Opin Investig Drugs 14:629–643 Levitzki A (2004) PDGF receptor kinase inhibitors for the treatment of PDGF driven diseases. Cytokine Growth Factor Rev 15:229–235 MacFadyen JR, Haworth O, Roberston D, Hardie D, Webster MT, Morris HR, Panico M, SuttonSmith M, Dell A, Van Der Geer P et al (2005) Endosialin (TEM1, CD248) is a marker of stromal fibroblasts and is not selectively expressed on tumour endothelium. FEBS Lett 579:2569–2575 Massague J (2008) TGFbeta in Cancer. Cell 134:215–230 Micke P, Kappert K, Ohshima M, Sundquist C, Scheidl S, Lindahl P, Heldin CH, Botling J, Ponten F, Ostman A (2007) In situ identification of genes regulated specifically in fibroblasts of human basal cell carcinoma. J Invest Dermatol 127:1516–1523 Mootha VK, Lindgren CM, Eriksson KF, Subramanian A, Sihag S, Lehar J, Puigserver P, Carlsson E, Ridderstrale M, Laurila E et al (2003) PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet 34:267–273 Mueller MM, Fusenig NE (2004) Friends or foes—bipolar effects of the tumour stroma in cancer. Nat Rev Cancer 4:839–849 Nakamura T, Matsumoto K, Kiritoshi A, Tano Y (1997) Induction of hepatocyte growth factor in fibroblasts by tumor-derived factors affects invasive growth of tumor cells: in vitro analysis of tumor–stromal interactions. Cancer Res 57:3305–3313 Nanda A, Karim B, Peng Z, Liu G, Qiu W, Gan C, Vogelstein B, St Croix B, Kinzler KW, Huso DL (2006) Tumor endothelial marker 1 (Tem1) functions in the growth and progression of abdominal tumors. Proc Natl Acad Sci U S A 103:3351–3356 Niedermeyer J, Scanlan MJ, Garin-Chesa P, Daiber C, Fiebig HH, Old LJ, Rettig WJ, Schnapp A (1997) Mouse fibroblast activation protein: molecular cloning, alternative splicing and expression in the reactive stroma of epithelial cancers. Int J Cancer 71:383–389 Niedermeyer J, Enenkel B, Park JE, Lenter M, Rettig WJ, Damm K, Schnapp A (1998) Mouse fibroblast-activation protein-conserved Fap gene organization and biochemical function as a serine protease. Eur J Biochem 254:650–654 Niedermeyer J, Garin-Chesa P, Kriz M, Hilberg F, Mueller E, Bamberger U, Rettig WJ, Schnapp A (2001) Expression of the fibroblast activation protein during mouse embryo development. Int J Dev Biol 45:445–447 O’Brien CA, Pollett A, Gallinger S, Dick JE (2007) A human colon cancer cell capable of initiating tumour growth in immunodeficient mice. Nature 445:106–110 Oft M, Peli J, Rudaz C, Schwarz H, Beug H, Reichmann E (1996) TGF-beta1 and Ha-Ras collaborate in modulating the phenotypic plasticity and invasiveness of epithelial tumor cells. Genes Dev 10:2462–2477 Okawa T, Michaylira CZ, Kalabis J, Stairs DB, Nakagawa H, Andl CD, Johnstone CN, Klein-Szanto AJ, El-Deiry WS, Cukierman E et al (2007) The functional interplay between EGFR overexpression, hTERT activation, and p53 mutation in esophageal epithelial cells with activation of stromal fibroblasts induces tumor development, invasion, and differentiation. Genes Dev 21:2788–2803 Olive KP, Jacobetz MA, Davidson CJ, Gopinathan A, McIntyre D, Honess D, Madhu B, Goldgraben MA, Caldwell ME, Allard D et al (2009) Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science 324:1457–1461 Olumi AF, Grossfeld GD, Hayward SW, Carroll PR, Tlsty TD, Cunha GR (1999) Carcinomaassociated fibroblasts direct tumor progression of initiated human prostatic epithelium. Cancer Res 59:5002–5011 Orimo A, Gupta PB, Sgroi DC, Arenzana-Seisdedos F, Delaunay T, Naeem R, Carey VJ, Richardson AL, Weinberg RA (2005) Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion. Cell 121:335–348 Ostermann E, Garin-Chesa P, Heider KH, Kalat M, Lamche H, Puri C, Kerjaschki D, Rettig WJ, Adolf GR (2008) Effective immunoconjugate therapy in cancer models targeting a serine protease of tumor fibroblasts. Clin Cancer Res 14:4584–4592
400
C. Rupp et al.
Ostman A, Heldin CH (2007) PDGF receptors as targets in tumor treatment. Adv Cancer Res 97:247–274 Park JE, Lenter MC, Zimmermann RN, Garin-Chesa P, Old LJ, Rettig WJ (1999) Fibroblast activation protein, a dual specificity serine protease expressed in reactive human tumor stromal fibroblasts. J Biol Chem 274:36505–36512 Perera SA, Li D, Shimamura T, Raso MG, Ji H, Chen L, Borgman CL, Zaghlul S, Brandstetter KA, Kubo S et al (2009) HER2YVMA drives rapid development of adenosquamous lung tumors in mice that are sensitive to BIBW2992 and rapamycin combination therapy. Proc Natl Acad Sci U S A 106:474–479 Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA et al (2000) Molecular portraits of human breast tumours. Nature 406:747–752 Pietras K, Rubin K, Sjoblom T, Buchdunger E, Sjoquist M, Heldin CH, Ostman A (2002) Inhibition of PDGF receptor signaling in tumor stroma enhances antitumor effect of chemotherapy. Cancer Res 62:5476–5484 Pietras K, Pahler J, Bergers G, Hanahan D (2008) Functions of paracrine PDGF signaling in the proangiogenic tumor stroma revealed by pharmacological targeting. PLoS Med 5:e19 Politi K, Zakowski MF, Fan PD, Schonfeld EA, Pao W, Varmus HE (2006) Lung adenocarcinomas induced in mice by mutant EGF receptors found in human lung cancers respond to a tyrosine kinase inhibitor or to down-regulation of the receptors. Genes Dev 20:1496–1510 Pure E (2009) The road to integrative cancer therapies: emergence of a tumor-associated fibroblast protease as a potential therapeutic target in cancer. Expert Opin Ther Targets 13:967–973 Ramaswamy S, Ross KN, Lander ES, Golub TR (2003) A molecular signature of metastasis in primary solid tumors. Nat Genet 33:49–54 Rettig WJ, Garin-Chesa P, Healey JH, Su SL, Jaffe EA, Old LJ (1992) Identification of endosialin, a cell surface glycoprotein of vascular endothelial cells in human cancer. Proc Natl Acad Sci U S A 89:10832–10836 Rettig WJ, Garin-Chesa P, Healey JH, Su SL, Ozer HL, Schwab M, Albino AP, Old LJ (1993) Regulation and heteromeric structure of the fibroblast activation protein in normal and transformed cells of mesenchymal and neuroectodermal origin. Cancer Res 53:3327–3335 Rottenberg S, Jonkers J (2008) Modeling therapy resistance in genetically engineered mouse cancer models. Drug Resist Updat 11:51–60 Rubio-Viqueira B, Jimeno A, Cusatis G, Zhang X, Iacobuzio-Donahue C, Karikari C, Shi C, Danenberg K, Danenberg PV, Kuramochi H et al (2006) An in vivo platform for translational drug development in pancreatic cancer. Clin Cancer Res 12:4652–4661 Rupp C, Dolznig H, Puri C, Sommergruber W, Kerjaschki D, Rettig WJ, Garin-Chesa P (2006a) Mouse endosialin, a C-type lectin-like cell surface receptor: expression during embryonic development and induction in experimental cancer neoangiogenesis. Cancer Immun 6:10 Rupp C, Dolznig H, Puri C, Schweifer N, Sommergruber W, Kraut N, Rettig WJ, Kerjaschki D, Garin-Chesa P (2006b) Laser capture microdissection of epithelial cancers guided by antibodies against fibroblast activation protein and endosialin. Diagn Mol Pathol 15:35–42 Sadlonova A, Novak Z, Johnson MR, Bowe DB, Gault SR, Page GP, Thottassery JV, Welch DR, Frost AR (2005) Breast fibroblasts modulate epithelial cell proliferation in three-dimensional in vitro co-culture. Breast Cancer Res 7:R46–59 Santos AM, Jung J, Aziz N, Kissil JL, Pure E (2009) Targeting fibroblast activation protein inhibits tumor stromagenesis and growth in mice. J Clin Invest 119:3613–3625 Sato N, Maehara N, Goggins M (2004) Gene expression profiling of tumor-stromal interactions between pancreatic cancer cells and stromal fibroblasts. Cancer Res 64:6950–6956 Scanlan MJ, Raj BK, Calvo B, Garin-Chesa P, Sanz-Moncasi MP, Healey JH, Old LJ, Rettig WJ (1994) Molecular cloning of fibroblast activation protein alpha, a member of the serine protease family selectively expressed in stromal fibroblasts of epithelial cancers. Proc Natl Acad Sci U S A 91:5657–5661 Schmeichel KL, Bissell MJ (2003) Modeling tissue-specific signaling and organ function in three dimensions. J Cell Sci 116:2377–2388
20 Cancer Associated Fibroblasts as Therapeutic Targets
401
Scott AM, Wiseman G, Welt S, Adjei A, Lee FT, Hopkins W, Divgi CR, Hanson LH, Mitchell P, Gansen DN et al (2003) A Phase I dose-escalation study of sibrotuzumab in patients with advanced or metastatic fibroblast activation protein-positive cancer. Clin Cancer Res 9:1639– 1647 Shu Q, Wong KK, Su JM, Adesina AM, Yu LT, Tsang YT, Antalffy BC, Baxter P, Perlaky L, Yang J et al (2008) Direct orthotopic transplantation of fresh surgical specimen preserves CD133+ tumor cells in clinically relevant mouse models of medulloblastoma and glioma. Stem Cells 26:1414–1424 Skobe M, Fusenig NE (1998) Tumorigenic conversion of immortal human keratinocytes through stromal cell activation. Proc Natl Acad Sci U S A 95:1050–1055 St Croix B, Rago C, Velculescu V, Traverso G, Romans KE, Montgomery E, Lal A, Riggins GJ, Lengauer C, Vogelstein B et al (2000) Genes expressed in human tumor endothelium. Science 289:1197–1202 Steeghs N, Nortier JW, Gelderblom H (2007) Small molecule tyrosine kinase inhibitors in the treatment of solid tumors: an update of recent developments. Ann Surg Oncol 14:942–953 Sternlicht MD, Lochter A, Sympson CJ, Huey B, Rougier JP, Gray JW, Pinkel D, Bissell MJ, Werb Z (1999) The stromal proteinase MMP3/stromelysin-1 promotes mammary carcinogenesis. Cell 98:137–146 Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES et al (2005) Gene set enrichment analysis: a knowledgebased approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 102:15545–15550 Sugimoto H, Mundel TM, Kieran MW, Kalluri R (2006) Identification of fibroblast heterogeneity in the tumor microenvironment. Cancer Biol Ther 5:1640–1646 Taipale J, Chen JK, Cooper MK, Wang B, Mann RK, Milenkovic L, Scott MP, Beachy PA (2000) Effects of oncogenic mutations in Smoothened and Patched can be reversed by cyclopamine. Nature 406:1005–1009 Thayer SP, di Magliano MP, Heiser PW, Nielsen CM, Roberts DJ, Lauwers GY, Qi YP, Gysin S, Fernandez-del Castillo C, Yajnik V et al (2003) Hedgehog is an early and late mediator of pancreatic cancer tumorigenesis. Nature 425:851–856 Tlsty TD, Hein, PW (2001) Know thy neighbor: stromal cells can contribute oncogenic signals. Curr Opin Genet Dev 11:54–59 Tomkowicz B, Rybinski K, Foley B, Ebel W, Kline B, Routhier E, Sass P, Nicolaides NC, Grasso L, Zhou Y (2007) Interaction of endosialin/TEM1 with extracellular matrix proteins mediates cell adhesion and migration. Proc Natl Acad Sci U S A 104:17965–17970 Varjosalo M, Taipale J (2008) Hedgehog: functions and mechanisms. Genes Dev 22:2454–2472 Verona EV, Elkahloun AG, Yang J, Bandyopadhyay A, Yeh IT, Sun LZ (2007) Transforming growth factor-beta signaling in prostate stromal cells supports prostate carcinoma growth by up-regulating stromal genes related to tissue remodeling. Cancer Res 67:5737–5746 Watkins DN, Berman DM, Burkholder SG, Wang B, Beachy PA, Baylin SB (2003) Hedgehog signalling within airway epithelial progenitors and in small-cell lung cancer. Nature 422:313–317 Welt S, Divgi CR, Scott AM, Garin-Chesa P, Finn RD, Graham M, Carswell EA, Cohen A, Larson SM, Old LJ et al (1994) Antibody targeting in metastatic colon cancer: a phase I study of monoclonal antibody F19 against a cell-surface protein of reactive tumor stromal fibroblasts. J Clin Oncol 12:1193–1203 West RB, Nuyten DS, Subramanian S, Nielsen TO, Corless CL, Rubin BP, Montgomery K, Zhu S, Patel R, Hernandez-Boussard T et al (2005) Determination of stromal signatures in breast carcinoma. PLoS Biol 3:e187 Yashiro M, Ikeda K, Tendo M, Ishikawa T, Hirakawa K (2005) Effect of organ-specific fibroblasts on proliferation and differentiation of breast cancer cells. Breast Cancer Res Treat 90:307–313 Yauch RL, Gould SE, Scales SJ, Tang T, Tian H, Ahn CP, Marshall D, Fu L, Januario T, Kallop D et al (2008) A paracrine requirement for hedgehog signalling in cancer. Nature 455:406–410
.
Chapter 21
Targeting Tumor Associated Fibroblasts and Chemotherapy Debbie Liao and Ralph A. Reisfeld
21.1 Introduction Effective treatment of solid tumors by chemotherapy depends on several key factors including: (1) systemic delivery of the chemotherapeutic drug to the tumor, (2) effective distribution of active drug in sufficient quantities to kill tumor cells, and (3) sensitivity of tumor cells to the chemotherapeutic drug (Fig. 21.1). Many tumor-cell-intrinsic mechanisms that contribute to chemoresistance have been identified such as: expression of drug efflux transporters (multi-drug resistanceassociated proteins) and detoxifying enzymes (glutathione S-transferase) by tumor cells, as well as defects in apoptosis regulatory proteins in tumor cells (Tannock 2001). However, non-transformed cells that reside in the tumor microenvironment can also contribute to chemoresistance of tumors. In particular, cancer associated fibroblasts (CAFs) are key mediators of tumor growth and can contribute to chemoresistance (Ostman and Augsten 2009). CAFs express and secrete many cytokines, growth factors, and extracellular matrix proteins that enhance survival of tumor cells, promote tumor angiogenesis and alter the composition of the extracellular matrix (ECM) in the tumor microenvironment (TME); all factors that can influence the delivery, uptake and activity of chemotherapeutic drugs to be favorable for tumor growth (Ostman and Augsten 2009). In this chapter, we will review the ways in which CAFs can influence the efficacy of chemotherapy by discussing its effects on drug delivery, activity, and drug sensitivity of tumor cells. Lastly, we will summarize the current approaches being developed by our laboratory and others for the therapeutic targeting of CAFs to improve cancer chemotherapy.
R. A. Reisfeld () Department of Immunology and Microbial Sciences, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, USA e-mail:
[email protected] M. M. Mueller, N. E. Fusenig (eds.), Tumor-Associated Fibroblasts and their Matrix, The Tumor Microenvironment 4, DOI 10.1007/978-94-007-0659-0_21, © Springer Science+Business Media B.V. 2011
403
404
D. Liao and R. A. Reisfeld Tumor cells Drug hypoxia C A
IFP
B
D
E ECM
ECM
Blood vessel
Fig. 21.1 Drug distribution in solid tumors. ( A) Systemic delivery of chemotherapeutic drugs to the tumor involves transport through the blood vessels of the circulatory system. ( B) Once drugs have reached the tumor, they must transport across the microvascular wall into the tumor interstitium. Dispersion of drugs across the microvascular wall and through the tumor can be impeded by elevated interstitial fluid pressure ( IFP). ( C) In areas that are distal to blood vessels, local hypoxia can result in acidosis, which can decrease the cellular uptake and pharmacological activity of certain drugs. ( D) Proteins of the tumor ECM, including collagens and proteoglycans, are a source of physical resistance and can directly impede drug penetration through the interstitial space. ( E) ECM components, including collagen, can also directly bind chemotherapeutic drugs, thus acting as a sink to prevent further drug distribution
21.2 Cancer Associated Fibroblasts and Effects on Drug Delivery 21.2.1 Vascular Physiology Systemic delivery of a chemotherapeutic drug involves transport through blood vessels of the circulatory system (Jain 1989). The efficiency of this processes is governed by vascular morphology and the blood flow rate (Jain 2001a). Normal blood vessels are highly organized and strictly regulate the movement of molecules across the vascular wall (Jang et al. 2003). The formation of new blood vessels, or angiogenesis, occurs normally in adults during situations such as wound healing, where a new vascular bed is required to oxygenate an area of regenerating tissue, and during the ovarian cycle (Papetti and Herman 2002). In these cases, angiogenesis occurs
21 Targeting Tumor Associated Fibroblasts and Chemotherapy
405
by a well-orchestrated and tightly regulated process that results in the formation of new functional vessels (Bergers and Benjamin 2003). In contrast, tumor-associated angiogenesis occurs in a rapid and disorganized fashion, and leads to the formation of tortuous vessels with excessive branching and arteriolar-venous shunts (Jain 2001b). These abnormal vessels are characterized by an absence of pericytes, vessel leakiness, and often exhibit irregular blood flow (Brown and Giaccia 1998; Less et al. 1991). This, in turn, leads to formation of necrotic and avascular regions, in addition to stabilized regions with microcirculation within the same tumor (Jain 2001a). This chaotic formation of blood vessels can result in regional differences in perfusion rates and heterogeneous spatial distribution of therapeutic agents (Jain 2001a; Endrich et al. 1979). For example, blood flow rates in necrotic and semi-necrotic regions are relatively poor while such rates in stabilized areas can be extremely variable (Jain 2001a). Since the systemic delivery of chemotherapeutic drugs depends on a functional circulatory system, dysfunctional vessels produced by tumor-associated angiogenesis can significantly impede the delivery of anti-cancer drugs to tumor cells and thus reduce their anti-tumor effects. CAFs have been shown to promote and mediate tumor angiogenesis and thus can negatively affect drug delivery in this way. CAFs can directly stimulate tumor angiogenesis by production of growth factors, including vascular endothelial growth factor (VEGF) and platelet-derived growth factor (PDGF) (Mueller and Fusenig 2004). Interestingly, normalization of tumor vasculature by blocking VEGF signaling has been shown to improve drug penetration in mouse xenograft models (Jain 2001b; Tong et al. 2004; Winkler et al. 2004). CAFs also secrete cytokines such as stromal-derived factor-1 (SDF-1), which promotes the recruitment of CXCR4 expressing endothelial cells that participate in angiogensis (Chometon and Jendrossek 2009). Additionally, using CAFs extracted from human breast carcinomas, Orimo et al. showed that co-implantation of CAFs with breast carcinoma cells significantly enhanced tumor growth in a xenograft model, compared to carcinoma cells implanted with normal fibroblasts (Orimo et al. 2005). In this model, enhanced tumor growth resulted from increased recruitment of endothelial progenitor cells and increased angiogenesis in response to SDF-1 secreted by CAFs (Orimo et al. 2005). These studies demonstrate that CAFs mediate tumor angiogenesis and thus can negatively affect the efficiency of systemic drug delivery.
21.2.2 Tumor Interstitial Fluid Pressure Once the chemotherapeutic drug has traversed the circulatory system and reached the tumor, drug transport out of blood vessels into the interstitial space relies on transcapillary pressure gradients (Jain 1987a). These pressure gradients are, in turn, determined by the hydrostatic and colloid osmotic pressures between the capillaries and interstitial space (Heldin et al. 2004). After exiting the vasculature, penetration of drugs into the tumor mass through extracellular space depends on concentration gradients and convection (Jain 2001a). In general, small molecules, such as oxygen and
406
D. Liao and R. A. Reisfeld
glucose, are transported by diffusion whereas larger molecules, such as soluble proteins and chemotherapeutic drugs, are transported by convection (Heldin et al. 2004). Typically in normal tissues, the overall transcapillary pressure gradient is slightly negative in the interstitium and results in a net outward transcapillary flow (Heldin et al. 2004). However, in solid tumors the hydrostatic and osmotic pressures of the interstitium are often increased, resulting in elevated interstitial fluid pressure (IFP) (Jain 1987a, b; Milosevic et al. 1998; Stohrer et al. 2000). The net effect of increased IFP is decreased diffusion and convection of systemically delivered drug compounds out of the circulation into the tumor. Additionally, after exiting the circulation, further penetration of drugs into the tumor mass by convection is dependent on the interstitial fluid velocity (Jain 1987b), which can also be reduced by IFP thus limiting the amount of drug delivered to distal tumor cells. Clinically, increased tumor IFP is commonly associated with worsened disease prognosis. For example, in patients with metastatic melanoma or non-Hodgkin’s lymphoma, the IFP of metastatic tumor nodules was shown to increase with disease progression (Curti et al. 1993). Additionally, a study on patients with cervical cancer showed that high IFP was a predictor of disease recurrence following radiotherapy, and was associated with increased mortality due to disease progression (Milosevic et al. 2001). CAFs can contribute to increased tumor IFP in several ways. First, CAFs promote tumor angiogenesis, as described previously, that leads to formation of dysfunctional and leaky vessels. These dysfunctional vessels allow increased outflow of macromolecules and proteins from the circulation thus causing an overall rise in the colloid osmotic pressure of the tumor interstitium (Curti et al. 1993). This increase in IFP due to leaky blood vessels can also be magnified by dysfunctional tumor lymphatics, which impedes the drainage of fluid and proteins from the interstitial space (Minchinton and Tannock 2006). Second, it has been proposed that fibroblasts can actively modulate IFP by directly regulating the amount of tension applied to the the ECM (Reed et al. 2001). This model stipulates that fibroblasts exert tension on the collagen microfibrillar network through collagen-binding integrins that, in turn, restrains the intrinsic swelling pressure of hyaluronan and proteoglycans in the ECM thus increasing IFP (Heldin et al. 2004; Meyer 1983). This kind of fibroblast contraction has been shown to occur in vitro in response to platelet-derived growth factor (PDGF) and transforming growth factor-β (TGF-β) (Clark et al. 1989; Montesano and Orci 1988). In this way, it is possible that CAFs can actively regulate IFP by controlling the amount of tension applied to the tumor ECM.
21.2.3 Extracellular Matrix Drug penetration into solid tumors can be directly impeded by components of the tumor ECM. Since CAFs produce a majority of the constituents that make up the ECM, they can significantly influence the rate of drug transport through the
21 Targeting Tumor Associated Fibroblasts and Chemotherapy
407
interstitial space (Kalluri and Zeisberg 2006). Components of the ECM include fibrous proteins (collagen and elastin) and polysaccharides (proteoglycans and hyaluronans) that link together to form a permeable mesh (Jang et al. 2003). Thus, by acting as a source of physical resistance, the ECM can directly influence the rate of macromolecular trafficking (Jang et al. 2003). For example, presence of glycosaminoglycan (GAG) can increase the resistance of water and solute transport through tissue (Jang et al. 2003). Additionally, collagen can also inhibit the transport of macromolecules as shown by xenograft studies in mice, where tumors with a welldefined collagen network exhibited increased rigidity and were more resistant to penetration of high molecular-weight drugs when compared to tumors with poorly organized and loose collagen networks (Netti et al. 2000). Intriguingly, in tumors with a well-defined collagen network, drug delivery could be improved by treatment with collagenase (Netti et al. 2000). Drugs can also bind directly to components of the ECM, thus preventing further penetration into more distal regions of the tumor (Berk et al. 1997). Studies using three-dimensional spheroids have shown that binding to ECM can directly affect drug penetration and distribution. For example, cisplatin and 5-fluorouracil, which do not bind readily to cellular macromolecules, diffused readily into spheroids whereas drugs which readily bind cellular macromolecules, such as doxorubicin and paclitaxel, were concentrated at the periphery of spheroids (Erlanson et al. 1992; Nederman and Carlsson 1984; Nicholson et al. 1997). In this way, the ECM can act as a sink and prevent further drug penetration into the tumor.
21.3 Cancer Associated Fibroblasts and Effects on Drug Activity and Sensitivity Effective chemotherapy is critically dependent on the delivery of bioactive drugs to tumor cells, which in turn depends on delivery and diffusion times. Therefore, conditions that delay drug delivery and diffusion, including compromised bloodflow and increased IFP, can significantly hinder the delivery of bioactive drugs. Since CAFs have significant influence on both tumor bloodflow and IFP, as described previously, they can significantly influence the delivery of bioactive drugs. However, there are other mechanisms by which CAFs can affect drug activity and sensitivity of tumor cells; these mechanisms will be discussed in this section and are summarized in Fig. 21.2.
21.3.1 Hypoxia Hypoxia, or low oxygen levels, is a hallmark of solid tumors (Hanahan and Weinberg 2000). During tumor growth, hypoxia occurs when rapidly proliferating tumor cells outgrow their blood supply, resulting in increased diffusion distances between
408
D. Liao and R. A. Reisfeld Cancer associated fibroblast
TGFβ HGF
Tumor cell proliferation
Hypoxia
Tumor cell
HIF-1α MDR genes Anti-apoptosis genes Cell cycle arrest
VEGF PDGF SDF-1 Tumor Angiogenesis
Collagen type 1 Fibronectin
CAM-DR
IFP
Reduced drug penetration and activity.
Fig. 21.2 How cancer associated fibroblasts affect chemotherapy. CAFs express and secrete many growth factors and ECM proteins that result in chemoresistance by promoting tumor angiogenesis, altering the composition of the tumor ECM, and enhancing the survival of tumor cells. Increased deposition of ECM proteins (collagen type 1 and fibronectin) and secretion of pro-angiogenesis proteins ( VEGF, PDGF and SDF-1) can increase interstitial fluid pressure ( IFP), which inhibits drug transport and penetration into the tumor. ECM proteins can also directly impede drug penetration or cause cell adhesion-mediated drug resistance ( CAM-DR). Growth factors produced by CAFs, such as TGFb and HGF, cause tumor growth and expansion that can result in areas of hypoxia, due to increased distances from blood vessels. Hypoxia, in turn, can increase chemoresistance of tumor cells by increasing the expression of transcription factor HIF-1 which, in turn, increases the expression of multidrug resistance ( MDR) and anti-apoptotic genes, as well as inducing cell cycle arrest
blood vessels (Liao and Johnson 2007). CAFs can mediate the development of tumor hypoxia by producing growth factors that promote tumor cell proliferation (e.g. TGFβ and HGF) as well as tumor angiogenesis (e.g. VEGF and PDGF) (Kalluri and Zeisberg 2006). Hypoxia can influence the efficacy of chemotherapy because the activity of many chemotherapeutic drugs is affected by the oxygenation status of the microenvironment and is therefore influenced by hypoxia. For example, the cytotoxic effects of cyclophosphamide and doxorubicin were shown to be oxygen dependent, both in vitro and in vivo, and decreased with hypoxia (Harrison and Blackwell 2004).
21 Targeting Tumor Associated Fibroblasts and Chemotherapy
409
Additionally, hypoxia often results in acidosis, or low extracellular pH, due to increased rates of anaerobic respiration that generate lactic acid, while intracellular pH remains unchanged (Tredan et al. 2007). This results in significant extracellularintracellular pH gradients that can affect the uptake of certain chemotherapeutic drugs (Gerweck and Seetharaman 1996). For example, lowered extracellular pH can decrease the cellular uptake and activity of drugs that are weak bases such as mitoxantrone, doxorubicin, and bleomycin (Vukovic and Tannock 1997). An acidic extracellular pH can also inhibit cellular uptake of drugs, such as methotrexate, across the cell membrane due to inhibition of active transmembrane transporters (Cowan and Tannock 2001). Hypoxia can also decrease the availability of free radicals, which reduces the DNA damaging effects of drugs such as etoposide, bleomycin, and anthracyclins (Harrison and Blackwell 2004). In addition to affecting the chemical activity of drugs, hypoxia can also directly influence the sensitivity of tumor cells to cytotoxic drugs. Because most chemotherapeutic drugs preferentially kill cells that are actively dividing, conditions in the tumor microenvironment that decrease cell cycling can affect the sensitivity of tumor cells to chemotherapeutic agents. For example, hypoxia often causes cells to undergo G0/G1 cell cycle arrest, thus limiting the cytotoxicity of drugs that specifically target rapidly proliferating cells such as mitoxantrone, paclitaxel and topotecan (Vukovic and Tannock 1997; Au and WaJ 2005). Hypoxia can also have marked effects on the gene expression of tumor cells and induce the expression of genes that confer resistance to chemotherapeutic drugs. In response to hypoxia, cells alter the expression of many oxygen-regulated genes that encode protein products involved in increasing oxygen delivery and activating alternate metabolic pathways that do not require oxygen (Semenza 2004). A key mediator of the cellular response to hypoxia is the transcription factor hypoxia inducible factor (HIF)-1. Stabilization of the HIF-1α subunit under low oxygen tensions results in transcription of many oxygen dependent genes that mediate diverse biological functions including angiogenesis, glycolytic metabolism, and cell survival (Semenza 2004). For example, under hypoxic conditions the pro-apoptotic genes Bid and Bad are down regulated in a HIF-1 dependent and independent manner, respectively (Erler et al. 2004). In contrast, hypoxia results in upregulation of anti-apoptotic proteins such as Bcl-2, Bcl-XL and IAP family members (Park et al. 2002). This hypoxia-induced alteration in the balance of anti- and pro- apoptotic genes protects tumor cells from drug-induced apoptosis. Hypoxia can also induce increased expression of multidrug resistance (MDR) genes. Classical MDR genes encode for ATP-dependent pumps belonging to a family of ATP-binding cassette (ABC) transporters that either exclude or extrude anticancer drugs from cells (Gottesman et al. 2002; Gottesman et al. 1996). Drugs that have been shown to be effluxed by ABC transporters include: actinomycin-D (an RNA transcription inhibitor), paclitaxel (a microtubule-stabilizing drug), vinca alkaloids (vinblastine and vincristine), anthracyclines (doxorubicin and daunorubicin) and taxanes (Gottesman et al. 2002; Leonessa and Clarke 2003). In response to hypoxia, HIF-1 was shown to upregulate the expression of the MDR1 gene product P-glycoprotein (Pgp) in human prostate cancer cells (Wartenberg et al. 2003). This
410
D. Liao and R. A. Reisfeld
hypoxic upregulation of Pgp expression resulted in increased resistance to doxorubicin in cultures of multicellular spheroids (Comerford et al. 2002). MDR1 expression has also been demonstrated in human leukemias, oesophegeal carcinoma, non-small-cell lung cancers, and breast carcinomas (Leonessa and Clarke 2003; Nooter et al. 1995). Similarly, hypoxia induces expression of breast cancer resistance protein (BCRP), also known as ATP-binding cassette, subfamily G, member 2 (ABCG2), in a HIF-1 dependent manner (Krishnamurthy et al. 2004). Like MDR1, BCRP encodes an ATP-dependent transporter of xenobiotic drugs (Doyle et al. 1998). BCRP has been shown to transport indolocarbazole topoisomerase I inhibitors (NB-506 and J-107088) and anthracyclins (mitoxantrone, daunorubicin, and doxorubicin), thus inducing drug resistance in human cancer cell lines (Doyle et al. 1998; Komatani et al. 2001; Miyake et al. 1999). In summary, CAFs mediate the development of tumor hypoxia by production of growth factors that promote both tumor cell proliferation and tumor angiogenesis. As a consequence of hypoxia, the uptake and activity of chemotherapeutic drugs is decreased. Additionally, the sensitivity of tumor cells to cytotoxic effects of chemotherapeutic drugs is also decreased by hypoxia through induction of cell cycle arrest and changes in gene expression that favor survival of tumor cells.
21.3.2 CAM-DR: Cell Adhesion-Mediated Drug Resistance Drug resistance acquired by tumor cell adhesion to extracellular matrix (ECM) is classified as cell adhesion-mediated drug resistance (CAM-DR). CAM-DR involves signaling through integrins, which are cell-surface molecules, consisting of a family of 18-α and eight-β subunits, that heterodimerize to form transmembrane receptors which bind to ECM proteins like collagen, fibronectin (FN) and laminin (Hehlgans et al. 2007). Ligand binding to the extracellular domain of integrins potentiates “outside-in” signaling and results in diverse intracellular responses that can affect drug resistance (Shain and Dalton 2001). CAFs are the main producers of ECM proteins in the TME, including FN and collagen type 1, and therefore mediate CAM-DR (Kalluri and Zeisberg 2006). Cell adhesion-mediated therapy resistance was first demonstrated in 1972 using monolayer and spheroid cultures, which were shown to be sensitive and resistant to radiation therapy, respectively (Hazlehurst and Dalton 2001). Since that time, CAM-DR to varying chemotherapeutic drugs has been demonstrated in many different cancers. For example, the human myeloma cell line 8226/S, that is normally sensitive to doxorubicin, was shown to become drug resistant when plated in direct contact with immobilized fibronectin (FN) (Damiano et al. 1999). Conversely, sensitivity of these cells to doxorubicin was restored by removal of FN (Damiano et al. 1999). Similarly, adhesion of U937 human histiocytic lymphoma cells and LiM6 colon cancer cells to ECM induced resistance in these cells to mitoxantrone and etoposide therapy, respectively (Hazlehurst et al. 2006; Kouniavsky et al. 2002). Also, adhesion to collagen VI conferred resistance to cisplatin in human ovarian
21 Targeting Tumor Associated Fibroblasts and Chemotherapy
411
cancer cells (Sherman-Baust et al. 2003). Additionally, adhesion to collagen type 1 caused resistance to paclitaxel in human colon, lung, and breast carcinoma cells (Ohbayashi et al. 2008). Futhermore, in a study of 1286 primary breast cancer specimens, high mRNA expression of FN was shown to be associated with shorter distant metastasis-free survival after adjuvant tamoxifen treatment of ER-positive lymph node-positive patients (Helleman et al. 2008). However, the mechanism by which adhesion to ECM proteins confers CAMDR differs depending on cell type. For example, in the human prostate cancer cell line PC3, β1 integrin-dependent adhesion to extracellular FN protected these cells from tumor necrosis factor-α (TNF-α) induced apoptosis through upregulation of survivin, a member of the inhibitor of apoptosis (IAP) family (Fornaro et al. 2003). In contrast, adhesion of myeloma cells to FN induced G1 cell cycle arrest and conferred drug resistance to etoposide through increased p27Kip1 protein levels (Hazlehurst et al. 2000). Further, adhesion of human histiocytic lymphoma cells to FN induced resistance to mitoxantrone though it diminished topoisomerase II levels and activity (Hazlehurst et al. 2006). In contrast, resistance to etoposide in human cervical carcinoma cells, grown as spheroids, correlated with a redistribution of topoisomerase IIα from the nucleus to the cytoplasm (Oloumi et al. 2000). Finally, CAM-DR to doxorubicin, cyclophosphamide and etoposide in human small cell lung cancer was caused by β1 integrin-stimulated tyrosine kinase activation leading to inhibition of caspase activation and apoptosis (Sethi et al. 1999). These examples illustrate the diverse mechanisms of how CAM-DR occurs in tumor cells. However, despite this diversity, the ultimate result of adhesion to ECM components in these studies is resistance to chemotherapy.
21.4 Improving Chemotherapy by Therapeutic Targeting of Cancer Associated Fibroblasts As evidenced above, CAFs are key mediators of therapeutic drug resistance in tumors and thus are attractive targets for improving chemotherapy. In general, three main avenues of attack are conceivable: (1) blocking recruitment and expansion of CAFs, (2) disruption of CAF-associated pro-tumorigenic signals involved with chemoresistance, and (3) abolishing CAF-associated interactions entirely by directly eliminating CAFs themselves in the TME. In this section, we will review the current strategies being developed that are aimed at targeting CAFs to improve chemotherapy.
21.4.1 Modulators of IFP Many strategies to improve chemotherapy have focused on therapeutically reducing tumor IFP in order to improve drug delivery and penetration in solid tumors.
412
D. Liao and R. A. Reisfeld
As described previously, fibroblast contraction in response to PDGF was proposed as a mechanism by which CAFs can directly control IFP. PDGF was first demonstrated to control IFP in the skin using a rat model of anaphylaxis and was subsequently shown to control the IFP in tumors using experimental rat colonic carcinomas (Pietras et al. 2001; Rodt et al. 1996). In the latter study, treatment of rats with STI571, a selective PDGF receptor kinase inhibitor, decreased tumor IFP and increased transcapillary transport in tumors grown subcutaneously (Pietras et al. 2001). Additionally, STI571 treatment caused increased uptake of the drug Taxol and resulted in enhanced anti-tumor effects (Pietras et al. 2002). Similar to STI571, prostaglandin (PG) E1 has also been shown to reduce intradermal IFP in vivo (Berg et al. 1998). Additionally, in experimental carcinomas, administration of PGE1 was shown to transiently lower tumor IFP thus facilitating the uptake of 5-fluorouracil (Salnikov et al. 2003). Protease-mediated digestion of ECM proteins produced by CAFs has also been tested as a strategy for reducing tumor IFP. In this context, systemic or intratumoral administration of collagenase, which digests collagen, reduced IFP of human osteosarcoma xenografts and was shown to improve the uptake of monoclonal antibodies (Eikenes et al. 2004). Additionally, treatment of penetration-resistant tumors, containing an extended collagen network, with collagenase improved interstitial diffusion rates in human colon adenocarcinoma xenografts (Netti et al. 2000). Similarly, it was shown that spread of the oncolytic herpes simplex virus vector MGH2 within human melanoma xenografts was inhibited by fibrillar collagen, and could be improved by co-injection with collagenase (McKee et al. 2006).
21.4.2 A DNA-Vaccine Targeting Cancer Associated Fibroblasts In our own laboratory, efforts to improve drug delivery in solid tumors have focused on the targeted elimination of CAFs within the TME. To achieve this, we have designed a DNA vaccine against fibroblast activation protein (FAP), which specifically targets CAFs for elimination by the host immune system. FAP is a type II transmembrane protein that functions as a serine protease and is specifically overexpressed on 90% of CAFs in colon, breast, and lung carcinomas (Scanlan et al. 1994). We have shown that oral vaccination of mice with doubly attenuated RE88 S. typhimurium transduced with full-length cDNA encoding murine FAP (pFAP), elicits a specific host cellular immune response against FAP-expressing CAFs (Loeffler et al. 2006). As a result, CAFs are effectively eliminated in the TME by cytotoxic T lymphocytes through a MHC Class 1 antigen response (Loeffler et al. 2006). We have shown that elimination of CAFs by vaccination with pFAP reduced collagen type 1 expression in the stroma of murine breast and colon carcinomas grown in immune competent mice (Loeffler et al. 2006) (Fig. 21.3). This reduction of collagen type 1 in the tumor stroma resulted in a significant increase in uptake of systemically administered doxorubicin by primary murine breast tumors (Loeffler et al. 2006) (Fig. 21.3). Additionally, the increased uptake of doxorubicin also
Anti-FAP Anti-collagen (type I) Actin
b
Dox/tumor tissue (ng/g)
Anti-collagen (type I)
a
Vector
pFap
Anti-FAP
Vector
413
pFap
21 Targeting Tumor Associated Fibroblasts and Chemotherapy
c
450 400 350 300 250 200 150 100 50 0
Vector
pFap
Fig. 21.3 A DNA vaccine targeting cancer associated fibroblasts improves drug uptake in experimental tumors. Expression of FAP, collagen type 1 and intratumoral uptake of doxorubicin. a Vaccination with pFAP markedly reduced the expression of both FAP and collagen type 1 by the tumor interstitium. b Analysis of protein expression by western blotting showed a decrease in collagen type 1 expression by tumors from mice vaccinated with pFap. c Following pFap vaccination, the reduction in collagen type 1 correlated with a significnat increase in doxorubicin uptake by the primary tumor (P